3D - PRINTED IN VITRO ANALYTICAL DEVICES FOR DIABETES THERAPEUTICS AND BLOOD BANKING STUDIES By Chengpeng Chen A DISSERTATION Submitted to Michigan State University in partial fulfillment of the requirements for the degree of Chemistry - D octor of Philosophy 2015 ABSTRACT 3D - PRINTED IN VITRO ANALYTICAL DEVICES FOR DIABETES THERAPEUTICS AND BLOOD BANKING STUDIES By Chengpeng Chen Erythrocytes (ERYs) play an important role in regulating blood flow via a pathway involving ERY - derived adenosi ne triphosphate (ATP) and endothelium - derived nitric oxide (NO). I mpaired ATP release from ERYs of diabetic patients is potential ly re sponsible for diabetic complications, thereby therapies involving stimulation of ERY - derived ATP release might limit such complications . This work explores 3D - printing to fabricate a novel microfluidic device to mimic the physiology of ERYs. 3D - printing enables fabricat ion of the device following a standard 96 - well plate geometry for efficient and high throughput readout with a plate reader. This 3D - printed rugged device was reusable after simple rinsing, which enables the detection of a batch of samples on the same device during a long - term experiment. T his 3D - printed fluidic device facilitated the investigation of the effica cy of C - peptide on stimulating ERY - derived ATP. W ithout albumin, C - peptide and Zn 2+ cannot increase ERY - derived ATP, suggesting the indispensable role of albumin in the process. The glutamic acid at the 27 position of C - peptide participated in the binding to albumin. Collectively, the ensemble of albumin, C - peptide, and Zn 2+ enhances ERY - derived ATP, which may reduce diabetes complications. To make the results more physiologically conclusiv e , an Organs - on - a - Chip platform that combined pancreatic - cells, ERY s and endothelial cells as a blood barrier mimic was developed. The secretion profiles of the - cells on the device simulate the physiological secreting process well. Subsequent cell - cell communication investigations show ed that - cell secretions do n ot af fect the endothelial cells but increase ATP release from ERYs, which in turn, exert s a downstream effect on endothelial cells by stimulating NO production. C urrently approved hyperglycemic ERY storage solutions impair ATP release from ERYs . This work cont inues to investigate the reversibility of ATP release from stored ERYs and shows that 15 day s of storage is a turning point, after which the ATP release is no longer reversible . This result is consistent with the clinical reports that blood stored longer t han 2 weeks is more likely to result in transfusion complications. The mechanism by which hyperglycemia impairs ATP release was also explore d by studying ERY deformability using a 3D - printed demand - based cell filter, finding that h yperglycemia permanently alters the deformability of ERYs after 5 days of storage. A 3D - printed intravenous device was developed to mimic a transfusion process in vitro . Addition of - cells on the platform showed that hyperglycemi a - stored ERYs failed to respond to the endocrine cells. Summarily, 3D - printing yields reusable, robust and reproducible microfluidic devices, and demand - based devices. T he ensemble of albumin, C - peptide and Z n 2+ can be a potential therapy for diabetes c omplications, and the current ERY storage protocol adversely alter the physiology of stored ERYs. A normoglycemic alternative may avoid this problem. iv ACKNOWLEDGEMENT S First, I would like to thank the organ izations responsible for funding this research work, including Michigan State University for travel funding supports, and National Institute of Health (NIH) for research funding. I greatly appreciate the investments made by the organizations in my developm ent as a scientist. My undergraduate education at Ocean University of China (OUC) laid a solid foundation for my scientific career. With the abundant research funding provided by OUC, I was able to start real scientific research when I was a sophomore, and I gradually recognized my interest and potential in research, which served as the initial motivation to continue my education in a graduate school. The knowledge, skills and scientific habits I acquired from these research experiences also facilitate my g raduate study. After I joined Michigan State University (MSU), the faculty of the department of chemistry has provided me with a great graduate education that has enabled me to think critically and independently. Importantly, I would like to thank Dr. Dan a M. Spence, my advisor, for his guidance in the past three and a half years. He does not push his students too much, but let them develop independently. He will check your feel comfortable, but also trained me to conduct research in a more independent manner. I believe the training I have received from him is invaluable, which will undoubtedly benefit me throughout life. I would also like to thank Dr. L. Karl Olson v from the department of physiology at MSU, who trained me how to culture INS - 1 cells in his laboratory, without which, this research work cannot be completed. I also appreciated the help provided by Dr. Merlin Bruening and Dr. David P. Weliky in the past a few years . I sincerely thank the past and current group members in the Spence group. When I first joined the group, Yueli Liu, Paul Vogel, Adam Giebink and Steve Halpin provided their help and encouragement generously, which helped me get through the transition pa in in a very short time. I would also like to thank the other members in this group: Yimeng Wang, Kari Brook - Anderson Engelkemier, Suzanne Summers, Bell, Hamideh Keshavarz , Cody Pinger and Andrew Heller . I would like to thank my family and friends for their generous support and help. My parents have been working hard for tens of years, which set an excellent example for me in my life. I am also grateful that I always have k ind and supportive friends around me. Some activities such as cards, mahjong, BBQ and ho t pots with these friends also left a joyful memory in the chap ter of my graduate school life. vi T ABLE OF C ONTENT S LIST OF TABLE S .................. ....... ..... ................................................................................x LIST OF FIGURE S .................. ........... ..............................................................................xi KEY TO ABBREVIATIONS ......... .................... ............................................. . ............. . .... xvi Chapter 1 Introduction ................................ ................................ ................................ ....... 1 1.1 Erythrocytes Regulated Vasodilation ................................ ................................ ............ 1 1.2 Microfluidics ................................ ................................ ................................ ......... 5 1.2.1 Concept and Development ................................ ................................ .................. 5 1.2.2 On - chip Fluidic Component Integration ................................ ............................ 8 1.2.3 Applications of Microfluidics ................................ ................................ ............. 11 1.2.3 Advantages and Disadvantages of PDMS Based Microfluid ics ..................... 20 1.3 Three Dimensional (3D) - printing and Microfluidics ................................ ........... 22 1.3.1 3D - printing ................................ ................................ ................................ ........... 22 1.3.2 3D - printing Methods ................................ ................................ .......................... 22 1.3.3 3D - printed Microfluidics ................................ ................................ .................... 27 1.3.2 Current Challenges and Future Directions ................................ ...................... 31 1.4 Dissertation Motivation and Overview ................................ .............................. 32 REFERENCES ................................ ................................ ................................ ........................... 36 Chapter 2 Study of C - peptide Stimulated ATP Release from Erythrocytes on a 3D - printed Circulation - mimic Fluidic Device ................................ ................................ ......... 47 2.1 Background ................................ ................................ ................................ ......... 47 2.1.1 Diabetes Complications ................................ ................................ ...................... 47 2.1.2 C - peptide and Vasodilation ................................ ................................ ............... 48 2.1.3 Chapter Motivation and Overview ................................ ................................ ... 52 2.2 Methods ................................ ................................ ................................ ............. 53 2.2.1 The Design and Fabrication of the Fluidic Device ................................ ........... 53 2.2.2 ATP Measurement Optimization ................................ ................................ ...... 56 2.2.3 Evaluation of Analytical Features of the Device ................................ ............. 57 2.2.4 ERY Sample Preparation ................................ ................................ ..................... 59 2.2.5 ERY Derived ATP Determination ................................ ................................ ....... 62 2.3 Results ................................ ................................ ................................ ................ 63 2.3.1 Device Design and Fa brication ................................ ................................ .......... 63 2.3.2 ATP Measurement Optimization ................................ ................................ ...... 70 2.3.3 Evaluation of Analytical Features of the Device ................................ ............. 73 2.3.4 Investigation of ERY Derived ATP Release stimulated by C - peptide ........... 75 2.4 Discussion ................................ ................................ ................................ ........... 77 2.5 Conclusion ................................ ................................ ................................ .......... 80 REFERENCES ................................ ................................ ................................ ........................... 81 Chapter 3 An Organs - on - a - Chip platform to Study the Interactions between - cells and Blood Vessels ................................ ................................ ................ 86 3.1 Background ................................ ................................ ................................ ......... 86 3.1.1 Organs - on - a - Chip ................................ ................................ ................................ 8 6 3. 1.2 Cell Selection for Organs - on - a - Chip ................................ ................................ . 88 3.1.3 Reconstruction of in vivo Microenvironments ................................ ................ 91 vii 3.1.4 Current development of Organ s - on - a - Chip ................................ .................... 93 3.1.5 Challenges and Future Directions ................................ ................................ ..... 95 3.1.6 Chapter Overview ................................ ................................ ............................... 96 3.2 Methods ................................ ................................ ................................ ............. 98 3.2.1 Culture of INS - 1 Cells in Membrane Inserts ................................ .................... 98 3.2.2 Culture of Endothelial Cells in Membrane Inserts ................................ .......... 99 3.2.3 Integration of Cells onto the Fluidic Device ................................ .................... 99 3.2.4 INS - 1 Cell Secretion Profile Measurement ................................ .................... 102 3.2.5 ATP Quantification ................................ ................................ ............................ 102 3.2.6 NO Quantification ................................ ................................ ............................. 103 3.3 Results ................................ ................................ ................................ .............. 104 3.3.1 Cell culture in Membrane Inserts ................................ ................................ ... 104 3.3.2 INS - 1 cell Secretion Profile ................................ ................................ .............. 106 3.3.3 Cell - cell Interaction Studies ................................ ................................ ............. 108 3.4 Discussion ................................ ................................ ................................ ......... 114 3.5 Conclusion ................................ ................................ ................................ ........ 119 REFERENCES ................................ ................................ ................................ ......................... 121 Chapter 4 - Evaluating Stored Erythrocytes Used in Transfusion Medicine using 3D - Printed Analytical Devices ................................ ................................ ................................ 129 4.1 Background ................................ ................................ ................................ ....... 129 4.1.1 Blood Banking ................................ ................................ ................................ .... 129 4.1.2 Transfusion Related Complications ................................ ................................ 130 4.1.3 Chapter Motivation and Overview ................................ ................................ . 132 4.2 Methods ................................ ................................ ................................ ........... 133 4.2.1 ERY Collection and Storage ................................ ................................ .............. 133 4.2.2 Sample Preparation ................................ ................................ .......................... 134 4.2.3 Determination of ATP Release Reversibility ................................ .................. 135 4.2.4 Study of Reversibility of ERY Deformability by a 3D - printed Cell Filter ..... 137 4.2.5 Design and Fabrication of an i.v. Injection Device ................................ ....... 141 4.2.6 Characterization of the i.v. Injection Device ................................ ................. 143 4.2.6 Study of Responsiveness of Stored ERYs to Pancreatic - cells ................... 144 4.3 Results ................................ ................................ ................................ .............. 146 4.3.1 ERY Collection and Storage, Sample Preparation ................................ ......... 146 4.3.2 Study of Reversibility of ATP Release on the 3D - Printed Fluidic Device ... 148 4.3.3 Study of Reversibility of ERY Deformability ................................ ................... 151 4.3.4 Characterization of the i.v. Injection Device ................................ ................. 154 4.3.5 Study of Responsiveness of Stored ERYs to Pancreatic - cells ................... 157 4.4 Discussion ................................ ................................ ................................ ......... 160 4.5 Conclusion ................................ ................................ ................................ ........ 164 REFERENCES ................................ ................................ ................................ ......................... 166 Chapter 5 Conclusions and Future Directi ons ................................ ............................ 170 5.1 Conclusions ................................ ................................ ................................ ....... 170 5.1.1. C - peptide Stimulated ATP Release from Erythrocytes on a 3D - printed Circulation - mimic Fluidic De vice ................................ ................................ ............... 170 5.1.2 An Organs - on - a - Chip platform to Study the Interactions between Pancreatic - cells and Blood Components ................................ .............................. 172 5.1.3 Ev aluating Stored Erythrocytes used in Transfusion Medicine by 3D - Printed Analytical Devices ................................ ................................ ................................ ....... 173 5.1.4 Summarized Conclusion ................................ ................................ ................... 175 viii 5.2 Future Directions ................................ ................................ .............................. 175 5.2.1 C - peptide Stimulated ATP Release from Erythrocytes on a 3D - printed Circulation - mimic Fluidic Device ................................ ................................ ............... 175 5.2.2 An Organs - on - a - Chip platform to Study the Interactions between Pancreatic - cells and Blood Components ................................ .............................. 177 5.2.3 Evaluating Stored Erythrocytes used in Transfusion Medicine by 3D - Print ed Analytical Devices ................................ ................................ ................................ ....... 178 REFERENCES ................................ ................................ ................................ ......................... 180 ix LIS T OF TABLES T able 2.1. Remaining Volumes of Water in Table 2.2 . Optimization of ATP Collectio Table 2.3. Optimization of Luciferin/luciferase Table 2.4. Observed quantitative detection of 150 nM and 250 nM ATP from channel 1, channel 3 and channel 5 . Table 3.1 . Summary of currently developed Organs - on - a - Chip models Table 4.2. The concentration of fluorescein in each loop after an in jection . Table 4.3. . Table 4.4. The hematocrit of ERYs in each loop after an injection . Table 4.5. . x LIST OF FIGURES Figure 1.1. The process by which ERYs help regulat e blood flow. ERYs - released ATP binds to the P2y receptor on endothelial cells and evokes NO synthesis and release. NO diffuses to smooth muscle layers, the main structural co mponent of a vessel wall, resulting in relaxation of Figure 1.2. A scheme for rapid prototyping and replica molding of microfluidic devices in PMDS. (A) , a photomask with desired transparent features ( white parts) is placed on a photoresist layer coated on a silicon wafer, which is exposed to UV light. (B), after the photoresist under the tra nsparent feature areas is cured and the rest photoresist removed (called a master), PDMS prepolymer (the yellow p art) is pour ed onto the master. After curing , the PDMS layer is peeled off the master, with negative replica tion of the features F igure 1.3. (A), a scheme for an on - chip pneumatic pump. A top channel in a thin PD MS layer is placed across a bottom channel in another PDMS layer. When proper pressure is applied in the top channel, it inflate s and deflect s down to occlude the bottom eumatic pumps are integrated to create peristaltic force that drive s sample forward into the separation channel. (C), in s accurate injection of a sample plug into the electrophoresis channel from a continuous sample stream. The design of a pushback ch annel prevents stagnant samples from entering th e channel in the next injection Figure 1.4. Electrokinetic injection of a sample on an electrophoresis microfluidic platform. The sample channel crosses the elec trophoresis channels where a potential will applied across. After the sample channel is filled, the plug at the crossing area will be driven into the electrophoresis channel when the potential is applied across the Figure 1.5. A microfluidic electrophoresis device that contains valves to manipulate sample injection. When Valve 1 is closed and Valve 2 open, a sample is injected to the electrophoresis channel. By applying the two valves, accurate sa mple introduction can be achieved . .. F igure 1.6. A sandwich PDMS microfluidic device for blood analysis. A), device design. Porous membrane was placed between two layers of PDMS, the bottom channel layer of channe ls and the top layer of wells. B), the schematic side view of a channel. While RBCs were flowing through the channel, produced molecules such as NO will diffuse through the membrane and be collected in the above well, where fluorescence detection can be pe xi Figure 1.7. Current 3D - printing methods. A), stereolithography; B), inkjet printing; C), selective laser sintering; D), fused deposition modeling; E), laminated object manufacturing; F) continuous printing m odel Figure 2.1. P roinsulin. Each sphere represent an amino acid. The labeled amino acids represent C - peptide. Upon cleavage, the two chains connected with disulfide bond forms insulin, while the conn ecting segment forms C - peptide Fig ure 2.2. The engineering sketch of the fluidic dev Figure 2.3. ERY sample preparation and measurement. A), four ERY samples were prepared with different treatments: without C - peptide or Zn 2+ ; only with C - peptide; only with Zn 2+ ; and with both C - peptide and Zn 2+ , the concentrations of C - peptide and Zn 2+ were both 10 nM. B), the four samples were introduced in to four channels of the fluidic device. After the sampl es circulated for 1.5 hours in the fluidic d evice that was placed in a 37 o C incubator, aliquots of 50 µ L o f PSS were loaded in the wells in row E to collect ATP from the flowing via diffusion, which was then quantified with the 1 Figure 2.4. (A): The 3D - printed fluidic device (bottom) used in this study is modeled after the dimensions of a 96 - well plate (top). Column and row markers make it convenient to localize wells. Six channels were printed on the odd number columns, with th ree wells corresponding to each channel. Static wells, printed on the even number columns allow for internal standards or calibration. (B): Membrane inserts (top), which have a semi - permeable polyester membrane, are inserted into wells (bottom) of the 3 - D printed device. (C): Threads, printed at both ends of the channel, connect external tubing through a finger tight adaptor. (D): A schematic cross section of a channel and the membrane inserts. ( E ) : The device locks into the sample holder of the commercial 4 Figure 2.5. Validation of the alignment of the device and printed wells in the plate reader. Six f luorescein solutions (concentrations ranging from 0 to 20 µ M ) were prepared and were loaded into the six inserts in wells B above each channel, respectively. Fluorescence intensity above these inserts was detected by plate reader and was plotted versus fluorescein concentration. The good linearity (r 2 = 0.99) shows that the device and wells are aligned well with the plat 6 6 Figure 2.6. A fluorescence image of a channel after flowing fluorescein. The dotted circles indicate the well areas. The image implies that leakage of liquid from channel across the membranes did not occur. Due to the limited scope view, images on separate parts of the channel were obtained and integrated into a single image 9 Figure 2.7. ATP release from circulating ERYs with different treatments. The data summarized in black bar s were collected from ERY s flowing in PSS, which suggest that xii C - peptide and Zn 2+ together significantly increased ATP release, while C - peptide or Zn 2+ alone did not show this effect. Without albumin in the ERY streams, however, C - peptide and Zn 2+ fail to stimulate ATP release (grey bar s ), which confirmed the role of albumin in the ATP stimulation process. In this experiment, 10 nM C - peptide and Zn 2+ were used to treat ERY samples. (N=5, error=S.E.M) 6 Figure 2.8. ATP release from ERYs treated with C - peptide/Zn 2+ and E27A /Zn 2+ . Unlike C - peptide/Zn 2+ , E27A/Zn 2+ 7 F igure 3.1. The concept of Human - on - a - Chip and its potential applications in 7 Figure 3.2 . The setup of the Organs - on - a - Chip device. This study used four channels , above which, inserts cultured with INS - 1 cells ( - cells) and bPAECs (endothelial cells) were integrated. The static wells in columns 8 and 10 contained ATP and NO standards for on - c hip simultaneous calibration Figure 3.3. (A), microscope image s cells cultured in a membrane insert. (A) INS - 1 cells; (B) bPAECs Figure 3.4. (A), INS - 1 cell secretion profile in a lbumin - containing stimulation buffer. (B), INS - 1 cell secretion profile in albumin - free stimulation solution. (C), the amount of C - peptide diffused to the following solution underneath flowing on the device, with albumin presented in the system. (D), witho ut albumin in the system, the amount o f C - Figure 3.5. (A), the top view of the four channels with different treatments. Channel 1 contains INS - 1 cells, bPAECs cultured in the membrane inserts , and ERYs flowing in PSS; channel 2 contains all three cells, but all buffers are albumin - free; channel 3 does not contain INS - 1 cells, with the rest identical as channel 1; channel 4 is almost identical to channel 1, except the endothelial cells were pretr eated with PPADS, a P2y receptor inhibitor. B), a sid Figure 3.6. The amount of ATP collected in the insert above each channel, which is proportional to the ATP amount in corresponding flowing. Compar ing cha nnel s 1 and 3, INS - 1 cell stimulated ATP release from ERYs, as is shown by increased ATP amount in the collecting well. Without albumin (channel 2), however, INS - 1 cells failed to stimulate ATP release from ERYs, which proved the indispensable role of albu min in this process. Channel 4 is identical to channel 1, except the endothelial cells were inhibited, which did not affect the ATP released Figure 3.7. The amount of ATP in ERYs streams in the four channels , showing t he same trend Figure 3.8. The released NO from endothelial cells culture in the last insert above each channel. The data from channels 1 to 3 are consistent with the ATP observation, which xiii is with more ATP releas ed, more NO will be produced in endothelial cells. In channel 4, however, though there was a higher ATP release from ERYs, the endothelial cells were pretreated with PPADs, a P2y receptor (ATP receptor in endothelial cells) inhibitor, and thus NO was not i ncreased Figure 4.1. The 3D - printed cell filter for ERY deformability measurements. (A): the two slabs with O rings on the binding sides. A piece of semipermeable membrane is the key part of the filter device, which was 5 m pore size polycarbonate membrane. (B): A view of the assembled device, with the membrane between the slabs. The assembling was achieved by simply binding the two slabs on the four wings with binder clips. (C): The view of a real filtration experiment. A sample was introduced by a pump into the top slab, which would be forced to go through the membrane and was then collected for subsequent cell counting. (D): The schematic cross view of a filtration process. A pressure was applied to push cells to deform to pass through the membrane between the O rings. However, only cells with sufficient deformability can go through, otherwise they will be blocked by the membrane 9 Figure 4.2. The technical drawings of the top slab, bottom slab and O rings of the filtration device , achieved by the autodesk inven tor software. Before sending the files into the printer, the O rings were assembled to both slabs in the software so that they can be printed simultaneously. The un it for all the dimensions is mm . 140 Figure 4.3. The 3D - printed i.v. injection device. (A): the device comprises three parts: a flow splitter, a flow controller, and a stop cock. Screw threads were printed on each part for easy but tight connection. When the hole that goes through the stopcock was adjusted along the channel in the flow controller, liquid can be injected through. Otherwise, the flow will be stopped. (B): Injection of fluorescein into a water circulation. Water was circulating in the loop that connects the low branch of the splitter and the other end of a channel, while fluorescein was injected in by a syringe pump 2 Figure 4.4. The strategy to study the responsiveness of stored ERYs to - cells. Fresh ERYs with a hematocrit of 5% were circulating in the six channels of the fluidic device, while fresh, AS - 1N stored, and AS - 1 stored ERYs were injected into the circulations via the i.v. injection device. INS - 1 cells were included on the left three channels (green ovals). A 2 - hour incubation at 37 o C was applied after the injection, after which, ATP amount in each channel was quantified by the luciferase/luciferin assay in the blue ovals ..14 5 Figure 4.5. Extracellular glucose levels in AS - 1 and AS - 1N storage solutions. Throughout the whole 36 - day storage, glucose in AS - 1 solutions (open circles) stayed at an extremely high level around 40 mM, while AS - 1N solutions (filled circles) showed physiologica l glucose level s around 5 mM. Data present the mean of 7 trials ± S. E.M .. 7 xiv Figure 4.6. ATP release reversibility data at different storage days. AS - 1N - PSSN trial referred to ERYs stored in AS - 1N (normoglycemic ) transfused into normoglycemic PSS; AS - 1 - PSSH trial represented ERYs stored in AS - 1 (hyperglycemic) transfused into hyperglycemic PSS; AS - 1 - PSSN trial was ERYs stored in AS - 1 but transfused into normoglycemic PSS. During the 36 storage days, AS - 1N - PSSN ER Ys released higher ATP at a constant level, while AS - 1 - PSSH ERYs released less ATP. The AS - 1 - PSSN ERYs, however, could reverse the ATP release before 15 days (complete reverse before Day 5 and partial reverse on Day 8 and 12). Seven biological replica were analyzed and the data represent mean ± S.E.M 50 Figure 4.7. Deformability reversibility data at different storage days. Black bars were constant around 100% during 36 days storage, which indicate normoglycemia stored ERYs did not lose deformability after being transfused to normoglycemic PSS. The AS - 1 - PSSH cells, however, lost their deformability even on the first storage day. The hyperglycemia stored ERYs could reverse their deformability after being transfused into normog lycemic PSS for the first five days of storage. From Day8, these cells also lost the ability to reverse their deformability. Data represent mean values of 4 biological replica ± S.E.M 3 Figure 4.8. ATP release from ERYs a fter being transfused via the 3D - printed i.v. injection device. (A): with INS - 1 cells on the channels; (B): without INS - 1 cells on the channels. f - f: fresh ERYs transfused to fresh ERYs; AS - 1N - f: AS - 1N stored ERYs transfused to fresh ERYs; AS - 1 - f: AS - 1 sto 9 xv KEY TO ABBREVIATIONS ATP: A denosine triphosphate AS - 1: A dditive solution 1 AS - 1N: A dditive solution 1 (normoglycemic) bPAECs: B ovine pulmonary artery endothelial cells CPD: C itrate - phosphate - dextrose solution CPD - N: C itrate - phosphate - dextrose ( normoglycemic ) CE: C apillary electrophoresis CFTR: Cystic fibrosis transmembrane conductance regulator ERY: E rythrocyte HBSS: Hank i.v.: Intravenous INS - 1: R at insulinoma beta cells LOM: Laminated object manufacturing NO: N itric oxide PBS: P hosphate buffered salt solution PDMS: P oly - (dimethyl siloxane) PPADS: pyridoxalphosphate - 6 - azophenyl - 2',4' - disulfonic acid PSS: Buffered p hysiologica l salt solution 1 Ch apt er 1 Introduction 1.1 Erythrocytes Regulated Vasodilation Erythrocytes (ERYs), also referred to as red blood cells or RBCs, are the most abu ndant cells in the blood stream and play an important role in oxygen delivery to other tissues and cells 1 . However, in addition to their oxygen delivery, ERYs also help regula te blood flow by controlling vasodilation via a pathway involving adenosine triphosphate (ATP) and nitric oxide (NO). The layer of endothelial cells (called endotheli um ) lining the inside of a vessel can synthesize NO through an endothelial nitric ox ide synthase (eNOS) pathway. This NO can diffuse to smooth muscle layers that serve as the main structure of a vessel wall, resulting in a rise in cyclic guanosine monophos phate (cGMP), which leads to a decrease in endoplasmatic Ca 2+ concentration that results in smooth muscle cell relaxation, and subsequent vasodilation 2 . I ncrease d shear stress is a major stimulus for NO release from the endothelium 3 . However, Sprague and colleagues found that in isolated perfused rabbit lungs, in the absence of ERYs, alterat ion of sheer stress alone did not evoke NO release. They concluded that in the rabbit pulmonary circulation there must be some interactions between the ERYs and the endothelium, resulting in NO - mediated vasodilation 4 . The authors reported that shear stress induces deformability of ERYs, which increases ATP release from ERYs through a pathway involving a G protein coupled receptor (GPCR) and the cystic fibrosis transmembrane conductivity regulator (CFTR ) 5 . They also 2 proposed that the ATP released from ERYs plays a major role in stimulating NO production in endothelial cells; numerous studies proved the role of AT P as a stimulus of NO production 6 . For example, Bogel and colleagues found that ERY derived ATP can bind the P 2y receptor on endothelial cells resulting in Ca 2+ inf lux and the formation of calmodulin, which activates endothelial nitric oxide synthase (eNOS); the eNOS then converts arginine to citrulline, with NO as a byproduct 7 . Figure 1.1 depicts the interaction between ERYs and endothelial cells in the process of NO production and release. Disturbance of this NO induced vasodilation can cause pathophysiological consequences. People with any form of diabetes will likely experience at least one chronic complication, such as nerve damage (neuropathy), blindness (retinopathy), kidney damage (nephropathy), or cardiovascular disease, which may result in endless pain or mortality 8 . A common theme in many of these conditions is the dysfunction that occurs at the interface betwee n the blood stream and peripheral tissues, or the vessels. For example, Low et al. suggested that poor blood flow results in endoneurial hypoxia, decreasing nerve signaling and leading to neuropathy 9 . In the retina of diabetic patients, restricted vasodilation causes poor oxygen circulation, which stimulates retinal cells to form new but fragile vessels via a vascular endothelial growth factor m echanism. The breaking of the new, weak vessels can cloud the vitreous, and high vessel density may lead to detachment of the retina and eventually result in blindness 10 , 11 . In diabetic nephropathy, the glomerulus undergoes partial sclerosis or failure to dilate, which causes hyper filtration and protein leakage . Some 3 other diseases such as hypertension, atherosclerosis, and restenosis also relate to impaired NO - induced vasodilation 12 . I nterestingly, the release of ATP from ERYs obtained from diabetic patients is significantly lower than th at from healthy controls 13 , 14 . Given the role of ERY - derived ATP in regulating vasodilation via stimulating endothelium derived NO production, therapies involving stimulation of ATP release from these cells might ameliorate some diabetic complications . However, under in vivo conditions, ERYs circulat e under multiple physic al (e.g. shear stress) 15 and biochemical (e.g. cell - cell interaction) 16 f actors. Therefore, a dynamic platform that can mimic the physiological conditions of ERYs will be needed. Microfluidic technology enables precise control of sample ( e.g. ERYs) and reagent flow and , can recapitulat e a vascular microenvironment for in vitro ERY studies 17 . In recent years, many studies used microfluidics for vasculature investigations, including hemodynamics 18 , 19 , ERYs under shear stress 20 , 21 , endothelial barrier function 22 , and angiogenesis 23 . 4 Figure 1.1. The process by which ERYs help regulat e blood flow. ERYs - released ATP binds to the P2y receptor on endothelial cells and evokes NO synthesis and release. NO diffuses to smooth muscle layers, the main str uctural component of a vessel wall, resulting in relaxation of these cells, and subsequent vasodilation. 5 1.2 Microfluidics 1.2.1 Concept and Development Microfluidics is the science and technology of miniaturized devi ces for analysis of small a mounts of liquids in channels with dimensions at micrometer levels. Microfluidic devices have many benefits in chemical and biological analyses, including reducing the amount of sample and reagents, decreasing experiment time, and enabling experiments to o ccur under dynamic conditions 24 . The ultimate goal of microfluidic technologies is to develop a lab - on - a - chip or total analysis system, where miniaturized sample pretreatment, separation and detection occur in an integrated fashion on a single device 25 . In 1979, Terry and colleagues invented t he first microfluidic device , a gas chromatograph fabricated on a silicon wafer for air analysis 26 . Ruzicka developed this concept by creating flow inject ion analysis systems on plastic milled devices in the 27 , 28 major terrorist threat, the US Department of Defense initiated and supported a series of studies on microfluidic devices for in - field biological and chemical detection, which greatly facilitated the development of the field of microfluidics 24 . D uring this period, the use o f a polymer, poly(dimethyl siloxane) (PDMS) greatly simplified fabrication of microfluidic devices 29 . Although the first generation of microfluidic devices relied on hard substrates such as sili con wafers and glass, fabrication was costly and time consuming. The use 6 of soft polymers as device substrates potentially reduces operation time and cost, and introduces new material characteristics, such as elasticity, transparency and easy molding to a microfluidic device 30 . After its introduction by the Whitesides group at Harvard Un iversity, PDMS beca me the most common substrate for microfluidic devices, because of its superior advantages, which include: (i) the simplicity of introducing micro scale features on PMDS with soft lithography; (ii) a relatively inert surface ; (iii) curing at relatively low temperatures; (iv) transparency , which enables optical detection; and (v) no toxic ity to biological cells 31 , 32 . PDMS based microfluidic devices are commonly fabricated by soft lithography, enabling rapid prototyping based on replica molding 33 . Rapid prototyping create s a master (on a silicon wafer) with a convex feature a s a mold for subsequent PDMS casting. This process starts by drawing desired features using computer aided design (CAD) software and printing these features on a t ransparent plastic film, which serves as a photomask. The photomask is placed on a silicon wa fer pre - coated with a photoresistant polymer. SU - 8 is a commonly used photoresist that cures upon exposure to ultraviolet (UV) light 34 . The photo mask allows UV light to pass through the transparent region only , which leads to selective curing of photoresist on the silicon wafer, thus createing the desired features. The featured master serves as a mold for PDMS device fabrication. PDMS prepolymer (mixed with curing reagent) is cast onto the master. After curing , the PDMS replica is peeled from the master, and creating a neg ative replica of the master can be fabricated in PDMS 35 . The low instrument requirements for this process make it widely used in many laboratories. 7 Moreover, i t is rapid and low cost, which makes it possible to test or optimize different device designs in a short time. (Figure 1.2) Figure 1.2. A scheme for rapid prototyping and replica molding of microfluidic devices in PMDS. (A) , a photomask wit h desired transparent features (white parts) is placed on a photoresist layer coated on a silicon wafer, which is exposed to UV light. (B), after the photoresist under the tra nsparent feature areas is cured and the rest photoresist removed (called a master ), PDMS prepolymer (the yellow part) is pour ed onto the master. After curing , the PDMS layer is peeled off the master, with negative replica tion of the features. Silicon wafer Photoresist Photomask UV UV A B 8 1.2.2 On - chip Fluidic Component Integration A functional microfluidic device must contain no t only channels, but also components such as mixers and valves necessary to assist liquid delivery, facilitate reaction and enhance experimental efficiency. The following is a brief review of each of these critical components of a microfluidic system. Liq uid flow in a micro scale channel is usually laminar and without convective mixing. When two fluids are flowing in a microchannel simultaneously, they flow in parallel, forming a clear interface between them , without turbulent mixing. The only mixing that will occur result s f rom diffusion at the interface of the two reagents. The tendency for a fluid to develop turbulent flow pattern s can be characterized by the Reynolds number (Re), which is proportional to flow speed and channel dimensions and inversely p roportional to liquid viscosity, as shown in Equation 1.1 36 . Re = Equation 1.1 is the density of the fluid (kg/m 3 ), v is the mean velocity of the fluid (m/s), L is the travelled length of the fluid (m), and µ is the viscosity of the fluid (kg/(m s)) Because of the small dimensions of microfluidic channels, the Re for microfluidic devices is almost always less than 1, which makes it impossible to be turbulent in character (which requires Re > 4000) 37 , 38 . For applications such as on - chip chemical reactions or biological studies, laminar flow is often not very useful, and thus on - chip 9 mixers that can homogenize flows are often requir ed. There have been several on - chip mixers reported, which predominantly rely on a chaotic advection effect that is realized by increasing contact area and contact time between flows 39 , 40 . A classical microfluidic mixer developed by the Whitesides group applies asymmetric herringbone grooves on the floor of a microchannel. A flow will be stretched in a groove and t hen folded to one another, which significantly increases contact area and time between the flowing species, and thus achieves high mixing efficiency 41 . There are some other on - chip mixer designs such as zigzag channels 42 or twisted channels 43 , or embedding barriers 44 in channels, all of which make it possible to increase mixing on a microfluidic platform. Valves are important components in any fluidic system to control and regulate flow patterns, and this is also true for microfluidi cs 45 . On - chip pneumatic valves were advantage of the elasti c property of PDMS 46 . The main idea of the valve design is pressure driven occlusion between stacked channels in two PDM S layers. The top channel, which is made within a thin (40 m) layer of PDMS, is crossed with a bottom channel that is in a thicker PDMS layer (4 mm). The liquid flows in the bottom channel, while the top one will be connected to a pneumatic pressure sourc e. When proper pneumatic pressure is applied through the top channel, it will be inflated and deflect down to occlude the bottom channel (Figure 1.3A). An array of such pneumatic valves w as also used to produce peristaltic force across a microfluidic chann el to drive flow on a chip 47 . For example, the Landers group used on - chip pneumatic pump arrays to 10 drive samples through a side channel, directly into an electrophoretic channel for separation and analysis (Figure 1.3B) 48 . Martin further optimized this application by incorporating a reduced - volume pneumatic valve that actuates (on the order of hundreds of milliseconds) to allow analytes from a continuously flowing sampling channel to be injected into a separation channel for electrophoresi s (Figure 1.3 C) 49 . F igure 1.3. (A), a scheme for an on - chip pneumatic pump. A top channel in a thin PDMS layer is placed across a bottom channel in another PDMS layer. When proper pressure is applied in the top c hannel, it inflate s and deflect s down to occlude the bottom eumatic pumps are integrated to create peristaltic force that drive s sample forward into the separation channel. (C), in tic pump enable s accurate injection of a sample plug into the electrophoresis channel from a continuous sample stream. The design of a pushback channel prevents stagnant samples from entering the channel in the next injection. Porous membranes integrated in PDMS microfluidic devices can act as molecule exchange media and cell culture support s 50 . By simply plac ing a piece of organic membrane between two PDMS slabs, the membrane binds tightly with PDMS by inter molecular forces, to form a sandwiched microfluidic device 51 . Membranes ha ve A B C 11 greatly expanded the functions of microfluidic devices in applications such as bio - sample analysis and on - line sample treatment. 52 , 53 . The ultimate goal of b - on - a - on - chip detection methods. Due to the miniaturized features of microfluidic devices, optical and electrochemical detectors are most common . O ptical fibers enable detection based on absorbance, fluorescence and chemiluminescence 54 . Chab inyc developed on - chip fluorescence detection by embedding an optical fiber in the microfluidic device and using an offline blue LED and a polymeric light filter for protein detection 55 . Compared to optical method s, which requires external components (light source, filters, photomultiplier tubing) and a dark environment 5 6 . Embedding of electrodes in PDMS layers during the casting process provides a much simpler detection platform 57 . Biologically relevant analytes such as cocaine 58 , antioxidants 59 , and neurotransmitters have been detected on electrochemically - based integrated microfluidic platforms. 1.2.3 Applications of Microfluidics The development of soft lithography and on - chip integration greatly reduced the time and cost of fabricating microfluidic devices, and enhanced the function of microfluidics, making this technology a powerful tool in many research areas. It also allowed microfluidic devices to be fabricated in any laboratory, as s ophisticated etching and ablation equipment was no longer required. T his chapter discusses the applications of microfluidics in biochemical engineering. 12 A complete analytical cycle includes sample pretreatment, separation and detection, which is usually time consuming and labor intensive. The concept of micro total analytical systems ( µ TAS) that integrate the multiple analytical steps on a miniaturized flow system for faster and automated analysis has gained increasing interest. Sample separation can be a key part in an analytical cycle, and one of the earliest examples of µ TAS was the integration of electrophoresis on an etched glass chip invented by Manz and colleagues in 1992, which shows the great potential of electrophoresis - based separation on a micr ofluidic device 60 . E lectrophoresis is easy to implement in microfluidic technology and thus has remained the primary separation technique on microfluidic devices. With integrated electrodes, an electrical field can drive flow electr okinetically, without the assistance of external bulky pumps, which makes such microfluidic devices a step forward towards automated analytical systems. With a shorter channel on a microfluidic device, the electrophoretic separation can be more efficient t han traditional capillary electrophoresis . Ramsey once reported a microfluidic device used for sub - millisecond electrophoresis, which was ~100 times faster than conventional methods 61 . Because a microfluidic channel can be simplified as a thick wall capillary without polyimide coating , the heat dissipation on a microfluidic device is more efficien t that in a capillary. Microfluidics also enables parallel and high throughput electrophoresis on one device, making it attractive when proteomics and genomics analyses are needed. For example, Simpson reported a h igh throughput genetic analysis on a micro fluidic electrophoresis platform 62 . 13 A nother advantage of on - chip electrophoresis is the short and controllable sample injection plug. For on - chip sample introd uction, electrokinetic injection is the preferred method, which is achieved by two steps, loading and dispensing 63 , 64 . The basic principle of this technique is to use two crossing channels, one for sample loading and the other for separation. After a sample is injected in the sample channel, the sample plug at the crossing zone will be driven through the separ ation channel by an applied voltage (Figure 1.4). This sample injection method was first reported by Manz on his etched glass device 60 . To increase separation efficiency, Ramsey minimized the injection plug by narrowing the sample c hannel size at the crossing on his device to separate rhodamine B and dichlorofluorescein 61 . 14 Figure 1.4. Electrokinetic injection of a sample on an electrophoresis microfluidic platform. The sample channel crosses the electrophoresis channels where a potential will applied across. After the sample channel is filled, the plug at the crossing a rea will be driven into th e electrophoresis channel when the potential is applied across the electrophoresis channel. 15 H owever, this injection method possesses limitations. First, sample dispersion into the electrophoresis channel occurs during sample loading, leading to uncontrolled plug length. The other problem is continuous sample leakage into the electrophoresis channel during the separation process, which will result in peak tailing 65 . To avoid these problems, Ramsey developed the electric pinched injection that can inhibit leakage of samples by using the electrical field from the separation channel to the sample waste channel as a virtual valve 66 . Furthermore, pneumatic pumps were applied on electrophoresis devices for more accurate and frequent sample injections. For example, Mecker reported a coupled on - chip electrophoresis device that enables online continuous sampling and injection 67 . As shown in Figure 1.5, microdialy sed samples were introduced to the sample inlet of the electrophoresis chip. Two pneumatic valves control the injection of sample into the separation channel. For example, if valve 2 is open while valve 1 closed, sample is injected. Then both valves are cl osed and separation will occur. This work was the first microfluidic system that coupled microdialysis and continuous injection of samples, electrophoretic separation, and detection on a single device. Another important feature of on - chip elelctrophoresis is the detection technique. Multiple optical detection methods such as laser induced fluorescence (LIF) and UV absorbance have been developed 68 , 69 . Because electrode embedding in a microfluidic device has become a mature technique that can be easily performed in many laboratories, electrochemical detection (e.g. amperometry) is widely used for on - chip 16 electrophoresi s detection, without the requirements for bulky optical detectors or dark environments, which potentially makes it more portable 70 . Besides pure engineering and technique development, on - chip elelctrophoresis has become a practical tool for many applications. It has been applied to the separation of small molecules such as neurotransmitters 71 , amino acids 69 , phenolic compounds 72 , and flavonoids 73 ; and macromolecules such as DNA 74 , 75 , proteins 75 , and glycans 76 . Cell and cell content separation by on - chip electrophoresis has also been reported 77 , 78 . Figure 1.5. A microfluidic electrophoresis device that contains valves to manipulate sample injection. When Valve 1 is closed and Valve 2 open, a sample is injected to the electrophoresis cha nnel. By applying the two valves, accurate sample introduction can be achieved. B Valve 1 B SW Valve 2 PB S Hydrodynamic Flow Channel Electrophoresis Channel B: buffer BW: buffer waste S: sample SW: sample wast e 17 Particle separation is an important tool in biological and chemical studies, including tumor cell screening, cell sorting and tissue engineering. However, conventional part icle separation methods such as centrifugation, chromatography, and immunoisolation, are time consuming, are not efficient, non - continuous and low throughput. Microfluidic devices have been developed for particle separation, to overcome the shortcomings of existing methods. Acoustic wave based separation, also referred to as acoustophoresis, is a commonly used on - chip particle separation method. The first microfluidic acoustophoresis device was invented by Kapishnikov and colleagues to separate blood cells 79 . Acoustophoresis applies an axial acoustic primary radiation force so that larger particles will move closer to the center of the channels while smaller particles move further from the center, and thus particles of different sizes can be separated 80 . Other forces such as gravity 81 and inertia 82 have also been reported to apply particle separation on microfluidic devices. 1.2.3.2 Analytical Science In contrast to lab - scale chemical analyses, microfluidics are miniaturized systems with continuous but controllable flow and integration capabilities, which have made it an ideal platform for rapid, high throughput chemical analysis. Immunoassay techniques have been widely used on microfluidic devices for protein detection 83 . Microliter and nanoliter scale channels can serve as an immunoreaction chamber that can significantly decrease experimental time and the usage of expensive antibodies. The high surface - to - v olume ratio in microchannels, as well as the ease of modifying 18 the PDMS surface, makes it feasible to immobilize antibodies on the surface of a channel. There have been several antibody immobilization methods reported, among which, Sebra and colleagues mad e an impressive immobilization by grafting acrylated antibodies on PDMS in the presence of poly(ethylene glycol) 84 . Due to the transparent characteristic of PDMS, optical detection (fluorescence and absorbance) can be performed. By conjugating electrochemical sensitive probes on the secondary antibodies, electrochemical detection with embedded electrodes can also be applied. Enzyme assays, cell - based assays, DNA det ection and amplification have also been successfully developed on microfluidic devices 85 . Since Martinez and colleagues developed several methods to fabricate microfluidic devices on paper, the concept of paper - based microfluidic diagnostic tools has become an active research area 86 . Previous work mainly focused on integrating assays (mainly colo rimetric assays) on paper fabricated devices. Although glucose and albumin have be en successfully detected, which can potentially serve as a diagnostic tool for diabetes and kidney disease 87 , more research regarding assay development and on - chip detection needs to be performed before paper microfluidics can realize its ultimate goal of providing an inexpensive diagnosis tool for global health, especially for resource limited areas 88 . Another attractive feature of performing analyses using microfluidic device technology is its ability i n sample preparation and for complex biological sample analysis. Kim and colleagues have reviewed microfluidic devices for on - chip cell lysis 19 and intracellular molecule extraction and purification 89 . With membrane integrated sandwich microfluidic devices, some complex samples do not need to be pretreated prior to detection. Figure 1.6 shows a blood analysis device developed by Halpin 52 . The porous membrane, which only allows molecules such as ATP and nitric oxide (NO) to diffuse through, rather than blood cells, serves as a natural separation unit so that molecules of interest can be directly detected in the well above membrane, without sample separation or pretreatment. F igure 1.6. A sandwich PDMS microfluidic device for b lood analysis. A), device design. Porous membrane was placed between two layers of PDMS, the bottom channel layer of channels and the top layer of wells. B), the schematic side view of a channel. While RBCs were flowing through the channel, produced molecu les such as NO will diffuse through the membrane and be collected in the above well, where fluorescence detection c an be performed to quantify NO. A B 20 1.2.3 Advantages and Disadvantages of PDMS Based Microfluidics The development of microfluidics in the past two decades has turned this field into a very active and dynamic research area, which is reflected by the abundant publications. A Web of Science search revealed that there are more than 13,000 ever, this large body of publications has not helped realize the initial goal of building true micro total analysis systems. Real - world application of the technology in industry is one method to judge the maturity or true beneficial utility of a technology . In the case of microfluidics, the amount of practical and applicable devices does not indicate utility. The market for microfluidics is fairly small, with few simple commercialized devices 90 . Althou gh PDMS has many advantages, it also has some inevitable disadvantages 91 . For example, microfluidic devices made from PMDS usually lack of rugged ness , causing fluid management difficulties due to leaking and/or uneven pressure. PDMS devices are typically singl e - use, which reduces the reproducibility of measurements by introducing inter - device errors. These disadvantages also include lack of fabrication standardization ( e.g. there are no standard production protocols for a specific microfluidic device; most curr ent devices are highly lab - dependent) and practical integration, which enlarges the gap between academic research and real - life applications. The reproducibility of PDMS microfluidic device fabrication is usually poor between laboratories. Even researcher s within the same laboratory, but with different 21 skill proficiency, can sometimes produce different results. The main reasons behind this is the lack of reusability of PDMS devices and the complicated multiple fabrication steps, as well as the fluid introd uction problems via the soft and weak punched ports 92 . For the former problem, a researcher has to fabricate multiple devices for a single measurement (a typical calibration and triplicate measurement will use eight devices), which increases the possibility of introducing error. For the latter problem, even though a different laboratory can have identical CAD designs of the device, the multiple fabrication steps ca n still inevitably introduce errors, which leads to inter - lab irreproducibility. Though many academic devices have integrated functional parts on a chip, each of them focused on integrating one part ( e.g. either liquid manipulation or detection). There hav e been few devices that can integrate several units on one device to realize a real Lab - on - a - Chip platform. Summarily, microfluidics has immense potential due to its unique characteristics of miniaturized reagent/sample usage, flow based measurement and in tegration. However, to move out of the academic realm and into applications, novel fabrication methods need to be developed to overcome the inherent drawbacks of PDMS or other fabrication methods involving multi - step production. Other polymers such as poly (methyl methacrylate) (PMMA) 93 and polystyrene (PS) 94 have been studied to make microfluidic devices, which showed advantages over PDMS, but did not essentially solve the problems. 3D - printing, a novel prototyping technology, has emerged as an alternative to fabricate microfluidic devices, and can potentially solve ir reproducibility and integration issues seen with PDMS, PMMA and PS - based devices. 22 1.3 Three Dimensional (3D) - printing and Microfluidics 1.3.1 3D - printing 3D - printing, which is achieved using an additive process, where successive layers of materials are la id down to build a three dimensional object, has been recently used in academic research laboratories 95 . The idea of 3D - printing was initiated in the early Hideo Kodama of Nagoya Municipal I ndustrial Research Institute created two additive manufacturing methods for 3D plastic model fabrication with photo - curable plastics 96 . Charles Hull further improved this manufacturing process by inventing the classical 3D - printing method, stereolithography, in which layers were created by c uring photopolymers with ultraviolet (UV) lasers. Another contribution by Hull to the field of 3D - printing is the invention of the .STL file format that has become the gold standard for data transfer to 3D - printers 97 . A design that is created by computer aid design (CAD) programs (e.g. AutoCAD, AutoDesk, SolidWorks) can be exported to .STL format, where the surface of a 3D model will be broken down to close - packed arrays of triangles, where the coordinates of the vertices can be defined in a text file, and thus be interpreted by a 3D - printer 98 . The resolution can be controlled by adjusting the numbers of the triangles that compose a sur face 99 . 1.3.2 3D - printing Methods D uri ng the past three decades, several 3D - printing methods have been commercialized. Stereolithography (SLA), created by Hull, was the first commercialized 3D - printing method 100 . Although there are different apparatus 23 designs for SLA from different companies, all of them share the core parts including a polymer reservoir, curing laser optics, and a support stage (Figure 1.7 A). The stage is immersed in a polymer res ervoir, which is on the top (surface) of the reservoir at the beginning, and will lower as the printing process continues. A UV light beam traces a 2D cross section on the stage, which will cure a layer of polymer into a specific shape. The stage then lowe rs by a predefined distance, and another layer of polymer will be cured in the same way on top of the previous layer. This process will be repeated, layer by layer, until the 3D object is finished 101 , 102 . Inkjet printing is another prototyping method for 3D - refers to the material being used to build a 3D object, such as a polymer 103 . Figure 1.7B illustrates the principle of an inkjet 3D - printer that uses a liquid binding material as its ink. A layer of polymer particles will be distributed evenly from the powder supply state onto the fabrication stage by a roller, after wh ich, the printing head drops the liquid binding material at desired areas to solidify it. After the first layer is finished, the fabrication stage will drop by a predefined distance, and a second layer will be distributed and selectively solidified by the binding liquid. The steps will be repeated until the 3D model is generated 104 - 106 . Compared to SLA, this technique has a wider material choice. For example, in addition to polymers, some inorganic materials, such as ceramic powders can also be combined with proper binding materials to create a 3D object. However, this powder - based method also possesses some obvious drawbacks. The introduction of liquid binding ma terials, which are often biologically toxic, may disable such 3D - printed devices for biological applications, such as cell 24 culture and tissue engineering. The degree of binding of the particles is also questionable, which may result in porosity problems an d surface roughness 107 . To overcome these disadvantages, powder - free inkjet printing methods have been developed 108 . For example, Stratasys produces several 3D - printer s that can use liquid photopolymers as the ink, which however, will need UV lights for curing. There is another powder - based, yet glue free 3D - printing metho d, known as selective laser sintering (SLS) 109 . The main setup of this method is similar to that of inkjet printing, except that a laser system is used to bind the powders through thermal sintering, instead of injecting liquid glue materials 110 . (Figure1.7C) This method can be applied with a very wide range of materials, including metals. However, a main problem is the shrinkage and deformation of printed objects, caused by laser heating and immediate cooling afterwards 11 1 , 112 . The most widely used 3D - Printing method today is a technique called fused deposition modeling (FDM), which fabricates a 3D model by depositing semi - fused materials (e.g. thermoplastics) on a sta ge layer by layer 113 . (Figure 1.7D) The setups of this method are more simplified than SLA and inkjet printing, which enables the production of compact, benchtop 3D - printers. Another advantage of this method is it can be applie d to any thermoplastic material , such as glass and certain ceramics. However, the resolution of objects fabricated with this method is fair, which is mainly determined by the size of the fila ments of feeding materials 114 . 25 He lysis developed a method of directly stacking sheet materials such as paper, plastic and metal to create a 3D model, which is called laminated object manufacturing (LOM) 115 . As shown in Figure 1.7E, two rollers keep supplying a sheet material under a laser optics setup, which can selectively cut an outline on a layer that is stacked on the previous layer by a subsequent heating/glue treatment. LOM can only be used with materials that can be processed into sheets. Delamination is also a common problem with this method 116 , 117 . In 2015, DeSimmone reported a modification to 3D - printing that can achieve high resolution (below 100 m) with shorter fabrication time 118 . This method is derived from SLA (Figure 1.7F). In this technique, an oxygen - permeable window is incorporated between the polymer reservoir and UV light, to form a continuous liquid interfa ce, which enables a continuous sequence of UV light to cure the polymer. Some examples show that hundreds of millimeters can be built within an hour, which significantly shortens 3D fabrication time. 26 Figure 1.7. Current 3D - printing methods. A), stereoli thography; B), inkjet printing; C), selective laser sintering; D), fused deposition modeling; E), laminated object manufacturing; F) continuous printing model 27 1.3.3 3D - printed Microfluidics A s discussed previously, microfluidics are powerful tools for many scientific applications that can be potentially developed to miniaturized lab - on - a - chip interfaces. However, certain disadvantages of PDMS and fabrication methods have prohibited this technique from being practically useful by the end - user. In additi on to the poor ruggedness and mechanical strength of PDMS, most of the devices were prepared during a multi - step fabrication, which not only increases fabrication time (especially given the fact that most of the devices are not reusable), but also makes it difficult to reproduce results between different laboratories. Therefore, some researchers have advocated to use one - step manufacturing methods, such as 3D - printing, to fabricate the next generation of microfluidic devices. The first 3D - printed microfluid ic device was created by Cronin in 2012, which is a flow - based micro reactor for the synthesis of imine. Subsequently, this concept has been further developed by Spence and other researchers in the past a few years, with the generation of a number of novel microfluidic devices that demonstrate superior features over PDMS devices. 1.3.3.1 Advantageous Features of 3D - printed Microfluidic Devices The poor ruggedness of PDMS usually results in uneven pressure distribution and flow problems ( e.g. accumulation of liquid around inlet ports ) . Even though hard polymers such as polystyrene have been used to avoid these problems, the fabrication process is time consuming, and the failure rate during experiments is high. 3D - printing, however, enables fabrication of a ru gged microfluidic device in one step. With proper 28 materials, such as acrylates, 3D - printed microfluidic devices can be reusable because they can be treated with harsh chemicals (e.g. bleach) and cleaning methods (e.g. ultra sonication), which significantly increases the reliability of results acquired on these devices 102 . 3D - printed microfluidic devices can also facilitate robust analytical method development, a s evidenced by the device reported by Chen and colleagues 119 . 3D - printing enabled the authors to shape the fluidic device in the for mat of a standard 96 - well plate, which is amenable to plate readers for high throughput analysis. With static wells printed on the device, simultaneous calibration can also be acquired on the same device. Inter - lab reproducibility can also be enhanced by 3D - printed microfluidic devices. By sharing the engineering sketches of a design, with detailed data information, another researcher can easily achieve an identical device. Moreover, as more and more laboratories are using 3D - printing to fabricate their de vices, a library database that collects all the device designs and parameters can be built, which can potentially leverage academic sharing, communication and mutual development. A lthough integrated components have been realized on PDMS microfluidic devic es, the poor ruggedness of PDMS increases the difficulty, time, and cost to integrate units on a device. Integrated devices using PDMS also suffer because the device itself is typically discarded after single use. Because 3D - printing can fabricate 29 any feat - chip functional features such as connectors and mixers can be easily achieved 120 . Microfluidic devices usually need to be connected to external tubing for liquid transport. A common technique for connecting tubing to a PDMS microfluidic device include directly inserting tubing into holes punched into PDMS and using adhesive to seal tubing into PDMS slabs 121 , 122 . Frazier also created a locking mechanism by soft lithography that can form removable and tight connections, although this method was not widely implemented, partially because of the comp lexity of the fabrication process 123 . However, the lack of reliability and flexibility of these protocols can cause fatal problems for the devices. For example, inserting tubing directly into punched holes is susceptible to leakage or even ejection of the tubing. For 3D - printed devices, these problems and concerns can be removed by designing proper connection threads that fit standard finger tight adapters, via which external tubing can be connected. H ardin and colleagues also performed a thorough study by physical modeling and calculations that proved the reliability of printed threads 124 . Moreover, with multi - material printing, O rings can also be printed on a microfluidic device for tight sealing or connection 125 . Lee and colleagues recently reported an integrated 3D - printed microfluidic device for biomarker detection. This single device contains mixers, reaction cham bers, gradient generators and connectors, which enables sample pretreatment, reaction 30 and detection on the same chip, without skilled handling. Though this device was designed with different modules, 3D - printed alignments and O rings make it easy to combin e, which meets the starting point of the research to develop an end - user detection device 120 . W ith soft lithography, only channels with rectangular cross sections can be fabricated in PDMS. In the case of on - chip vasculatu re studies, because leukocytes preferentially migrate at the corners of a rectangular channel, the behaviors of these cells on such a device do not represent the in vivo conditions 126 . However, 3D - printing can easily fabricate channels w ith circular cross sections. Electrophoresis and on - chip electrochemical detection is widely used as a step closer to the goal of lab - on - a - chip. Although electrodes have been successfully incorporated or deposited in PDMS microfluidic devices, the non - reu sability of these devices increases research cost (especially when using expensive metals as electrodes), but reduces reliability of results due to inter - device error propagation. A 3D - printed electrochemical device that was developed by the Spence group a nd the Martin group avoided these problems by applying a flexible electrode system 127 . To be specific, electrodes were embedded in a hollow plastic screw nut with Nafion, which were then inte grated into a female threaded port in a microfluidic device. After a measurement, the electrodes and the microfluidic device can be separated by screwing the nut off, and cleaned for reusing. Krejcova et al. recently also reported a 3D - printed 31 microfluidic device that combines electrodes and magnetic beads for the detection of in fl uenza hemagglutinin 128 . 1.3.4 Current Challenges and Future Directions A lthough 3D - printed microfluidic devices possess a number of advantages as discussed above, current 3D - printing technologies are not advanced enough to fabricate perfect microfluidic devices. The resolution of current 3D - printers is usually at the level of tens to hu ndreds of microns, which is not sufficient for certain applications that require smaller sizes. Another problem resulting from the relatively poor resolution is surface roughness. Many 3D - printed objects need further polishing prior to use, which is not a problem for bulk surfaces, but may represent drawbacks for surfaces of small features and in terior structures. The void spaces (channels, chambers etc.) are usually printed with pre - filling of supporting materials (e.g. wax - like compounds) that need to be completely removed before using, which is time an easy way to clean the supporting materials. Also, most materials used in current 3D - printers are not elastomers. For cer tain studies that need to include mechanical factors, such as stretching, these hard devices will not be the appropriate choices. There is also a lack of studies evaluating the bio - toxicity of these materials. The development of novel 3D - printing techniqu es can naturally solve some of the problems. However, the philosophy of combining PDMS and 3D - printing can be an immediate solution to the problems, which combines the strong points of other 32 fabrication methods to make up for its own shortcomings. For exam ple, to avoid the surface roughness problem and the bio - toxicity concern, Gross recently reported a 3D - printed microfluidic device with a PDMS coated channel for cell culture 129 recently reported a creative concept, which is to use 3D - printing to fabricate complicated masters that cannot be easily achieved by soft lithography, to cast PDMS devices. With this method, they demonstrated several designs that include complicated 3D channel systems for flow regulating. This innovation enlightens the applications that have to be achieved on PDMS devices 130 . Also, 3D - printing can be used to produce rigid connection parts and sample pret reatment units that can be combined with PDMS devices. 1.4 Dissertation Motivation and Overview In vivo , ERYs circulate under shear stress, and interact with other cell types. Therefore, to investigate ERY in a physiologically relevant way , an in vitro vas culature mimic device that can integrate other cell types and hemodynamics needs to be developed. Because microfluidic devices allows the samples to flow or even circulate, and micro - channels can be fabricated with dimensions <10 µ m , which potentially repl icates the geometry and dimensions of microcirculation, microfluidic devices can be a powerful tool to study in vivo physiology of ERYs on an in vitro device. PDMS based microfluidic devices have been widely used to create vasculature models that can mimic certain physiological factors such as hemodynamics, physiological relevance, and cell int egration. However, due to the in herent drawbacks of PDMS 33 devices, such as lack of ruggedness and reproducibility, most of the currently developed microfluidic vascula ture devices only include one physiological factor. Also, these complicated devices are not reusable, which not only increases fabrication cost and time, but also reduces r eproducibility. One - step fabrication of microfluidic devices with more rugged materi als has thus been proposed by researchers, and 3D - printing appears to be a promising technique to fabricate the next generation of microfluidics. Some pioneering research work by the Spence group and others has proven 3D - printed microfluidic devices can be more integrated, rugged, reusable, and reproducible than PDMS based devices . More importantly, features such as threads and connection ports can also be easily achieved with 3D - printing. The design of a device with CAD software can also be shared between laboratories for device reproduction. Therefore, there are two goals for the research work presented in this dissertation. The first goal is relevant to engineering and attempts to develop a standard in vitro circulation/vasculature device by 3D - printing, to investigate ERY physiology. This 3D - printed fluidic device demonstrates some important advantages over its PDMS counterpart. For example, screw threads are printed at the ends of each channel for simple, yet tight connection to external tubing. The des ign of static wells enables simultaneous on - chip calibration, which promotes experiment efficiency and increases reliability. Moreover, the dimensions of the device, as well as the locations of the wells follow a standard 96 - well plate geometry so that the whole device can be 34 placed directly in a plate reader for high content readout. The device is rugged and reusable, which enables the valuation of stored ERY samples on the same device during an experiment period of six weeks. The device also shows its ver satility to be developed to an Organs - on - a - Chip model for inter - tissue communication studies. By simply plugging membrane inserts that contain cultured cells into the wells on the device, different cell types can be combined to investigate cell - cell intera ctions. In this study, pancreatic - cells (INS - 1 cell line) that can secret C - peptide, and endothelial cells that can mimic a blood vessel barrier will be integrated, and be connected by circulating ERYs. Although PMDS - based Organs - on - a - Chip platforms have been developed, they were mainly single - cell models, and had to be discarded after single use. This 3D - printed model demonstrates an innovative way to fabricate more integrated ( e.g. multiple cells), standard ( e.g. using membrane inserts to culture cells), and versatile ( e.g. any cell typ e that can be cultured in membrane inserts can be integrated and studied) Organs - on - a - Chip platforms. In this research work, 3D - printing also shows its capability in fabrication of demand - based and customized analytical devices. From a physiological perspe ctive, these 3D - printed devices facilitate the investigation of ERYs, with the ultimate goals of developing a therapy for diabetic complications, and improving the quality of stored ERYs. By circulating ERYs under different conditions through different cha nnels on the 3D - printed fluidic device, the efficacy of C - peptide, Zn 2+ and albumin in stimulating ATP release from ERYs is confirmed, which can be potentially developed to a therapy for diabetic complications. 3 5 This single device also enables long - term eva luation of stored ERYs, suggesting that a normoglycemic storage condition is more beneficial to stored ERYs than currently approved hyperglycemic solutions. The results acquired from a 3D - printed homemade cell filtration device proved that hyperglycemia ca n permanently diminish ERY deformability. 36 REFERENCES 37 REFERENCES 1. Fairbanks, G., Steck, T.L. & Wallach, D.F. Electrophoretic analysis of the major polypeptides of the human erythrocyte membr ane. Biochemistry 10, 2606 - 2617 (1971). 2. Cooke, J.P. & Dzau, V.J. Nitric oxide synthase: Role in the genesis of vascular disease. Annu Rev Med 48, 489 - 509 (1997). 3. Buga, G.M., Gold, M.E., Fukuto, J.M. & Ignarro, L.J. Shear - Stress Induced Release of Nit ric - Oxide from Endothelial - Cells Grown on Beads. Hypertension 17, 187 - 193 (1991). 4. Sprague, R.S. et al. Effect of L - NAME on pressure - flow relationships in isolated rabbit lungs: Role of red blood cells (vol 38, pg H1941, 1995). Am J Physiol - Heart C 270, U33 - U33 (1996). 5. Sprague, R.S., Ellsworth, M.L., Stephenson, A.H., Kleinhenz, M.E. & Lonigro, A.J. Deformation - induced ATP release from red blood cells requires CFTR activity. Am J Physiol - Heart C 275, H1726 - H1732 (1998). 6. Sprague, R.S., Ellsworth, M.L ., Stephenson, A.H. & Lonigro, A.J. ATP: The red blood cell link to NO and local control of the pulmonary circulation. Am J Physiol - Heart C 271, H2717 - H2722 (1996). 7. Bergfeld, G.R. & Forrester, T. Release of Atp from Human Erythrocytes in Response to a B rief Period of Hypoxia and Hypercapnia. Cardiovasc Res 26, 40 - 47 (1992). 8. Rich, S.S. Genetics of diabetes and its complications. J Am Soc Nephrol 17, 353 - 360 (2006). 9. Low, P.A., Schmelzer, J.D., Ward, K.K., Curran, G.L. & Poduslo, J.F. Effect of Hyperb aric Oxygenation on Normal and Chronic Streptozotocin Diabetic Peripheral - Nerves. Exp Neurol 99, 201 - 212 (1988). 10. Pierce, E.A., Avery, R.L., Foley, E.D., Aiello, L.P. & Smith, L.E.H. Vascular Endothelial Growth - Factor Vascular - Permeability Factor Expres sion in a Mouse Model of Retinal Neovascularization. P Natl Acad Sci USA 92, 905 - 909 (1995). 11. Sheetz, M.J. & King, G.L. Molecular understanding of hyperglycemia's adverse effects for diabetic complications. Jama - J Am Med Assoc 288, 2579 - 2588 (2002). 12. Craven, P.A., Caines, M.A. & Derubertis, F.R. Sequential Alterations in Glomerular Prostaglandin and Thromboxane Synthesis in Diabetic Rats - 38 Relationship to the Hyperfiltration of Early Diabetes. Metabolism 36, 95 - 103 (1987). 13. Carroll, J. et al. An al tered oxidant defense system in red blood cells affects their ability to release nitric oxide - stimulating ATP. Mol Biosyst 2, 305 - 311 (2006). 14. Sprague, R.S., Stephenson, A.H., Bowles, E.A., Stumpf, M.S. & Lonigro, A.J. Reduced expression of G(i) in eryt hrocytes of humans with type 2 diabetes is associated with impairment of both cAMP generation and ATP release. Diabetes 55, 3588 - 3593 (2006). 15. Chiu, J.J. & Chien, S. Effects of Disturbed Flow on Vascular Endothelium: Pathophysiological Basis and Clinica l Perspectives. Physiol Rev 91, 327 - 387 (2011). 16. Abaci, H.E., Hanjaya - Putra, D. & Gerecht, S. Hypoxia and Matrix Manipulation for Vascular Engineering. Biol Med Phys Biomed , 127 - 165 (2011). 17. Wong, K.H.K., Chan, J.M., Kamm, R.D. & Tien, J. Microfluidi c Models of Vascular Functions. Annual Review of Biomedical Engineering, Vol 14 14, 205 - 230 (2012). 18. Shevkoplyas, S.S., Gifford, S.C., Yoshida, T. & Bitensky, M.W. Prototype of an in vitro model of the microcirculation. Microvasc Res 65, 132 - 136 (2003). 19. Shevkoplyas, S.S., Yoshida, T., Gifford, S.C. & Bitensky, M.W. Direct measurement of the impact of impaired erythrocyte deformability on microvascular network perfusion in a microfluidic device. Lab Chip 6, 914 - 920 (2006). 20. Wan, J.D., Ristenpart, W .D. & Stone, H.A. Dynamics of shear - induced ATP release from red blood cells. P Natl Acad Sci USA 105, 16432 - 16437 (2008). 21. Sprung, R., Sprague, R. & Spence, D. Determination of ATP release from erythrocytes using microbore tubing as a model of resistan ce vessels in vivo. Anal Chem 74, 2274 - 2278 (2002). 22. Shao, J.B. et al. Integrated microfluidic chip for endothelial cells culture and analysis exposed to a pulsatile and oscillatory shear stress. Lab Chip 9, 3118 - 3125 (2009). 23. Bouhadir, K.H. & Mooney , D.J. Promoting angiogenesis in engineered tissues. J Drug Target 9, 397 - 406 (2001). 24. Whitesides, G.M. The origins and the future of microfluidics. Nature 442, 368 - 373 (2006). 39 25. Abgrall, P. & Gue, A.M. Lab - on - chip technologies: making a microfluidic network and coupling it into a complete microsystem - a review. J Micromech Microeng 17, R15 - R49 (2007). 26. Terry, S.C., Jerman, J.H. & Angell, J.B. Gas - Chromatographic Air Analyzer Fabricated on a Silicon - Wafer. Ieee T Electron Dev 26, 1880 - 1886 (1979). 27. Ruzicka, J. & Hansen, E.H. Flow - Injection Analysis and Its Early History. Talanta 29, 157 - 157 (1982). 28. Ruzicka, J., Hansen, E.H. & Ramsing, A.U. Flow - Injection Analyzer for Students, Teaching and Research - Spectrophotometric Methods. Anal Chim Acta 134, 55 - 71 (1982). 29. Whitesides, G.M. & Stroock, A.D. Flexible methods for microfluidics. Phys Today 54, 42 - 48 (2001). 30. Mijatovic, D., Eijkel, J.C.T. & van den Berg, A. Technologies for nanofluidic systems: top - down vs. bottom - up - a review. Lab Chip 5, 492 - 500 (2005). 31. McDonald, J.C. et al. Fabrication of microfluidic systems in poly(dimethylsiloxane). Electrophoresis 21, 27 - 40 (2000). 32. McDonald, J.C. & Whitesides, G.M. Poly(dimethylsiloxane) as a material for fabricating microfluidic devices. Accounts Chem Res 35, 491 - 499 (2002). 33. Xia, Y.N. & Whitesides, G.M. Soft lithography. Annu Rev Mater Sci 28, 153 - 184 (1998). 34. Lorenz, H. et al. SU - 8: a low - cost negative resist for MEMS. J Micromech Microeng 7, 121 - 124 (1997). 35. Whitesides, G.M., O stuni, E., Takayama, S., Jiang, X.Y. & Ingber, D.E. Soft lithography in biology and biochemistry. Annu Rev Biomed Eng 3, 335 - 373 (2001). 36. Fredsoe, J. Turbulent Boundary - Layer in Wave - Current Motion. J Hydraul Eng - Asce 110, 1103 - 1120 (1984). 37. Suh, Y.K . & Kang, S. A Review on Mixing in Microfluidics. Micromachines - Basel 1, 82 - 111 (2010). 38. Squires, T.M. & Quake, S.R. Microfluidics: Fluid physics at the nanoliter scale. Rev Mod Phys 77, 977 - 1026 (2005). 39. Hessel, V., Lowe, H. & Schonfeld, F. Micromix ers - a review on passive and active mixing principles. Chem Eng Sci 60, 2479 - 2501 (2005). 40. Nguyen, N.T. & Wu, Z.G. Micromixers - a review. J Micromech Microeng 15, R1 - R16 (2005). 40 41. Stroock, A.D. et al. Chaotic mixer for microchannels. Science 295, 64 7 - 651 (2002). 42. Mengeaud, V., Josserand, J. & Girault, H.H. Mixing processes in a zigzag microchannel: Finite element simulations and optical study. Anal Chem 74, 4279 - 4286 (2002). 43. Jen, C.P., Wu, C.Y., Lin, Y.C. & Wu, C.Y. Design and simulation of th e micromixer with chaotic advection in twisted microchannels. Lab Chip 3, 77 - 81 (2003). 44. Kim, D.S., Lee, S.W., Kwon, T.H. & Lee, S.S. A barrier embedded chaotic micromixer. J Micromech Microeng 14, 798 - 805 (2004). 45. Hong, J.W. & Quake, S.R. Integrated nanoliter systems. Nat Biotechnol 21, 1179 - 1183 (2003). 46. Thorsen, T., Maerkl, S.J. & Quake, S.R. Microfluidic large - scale integration. Science 298, 580 - 584 (2002). 47. Quake, S. The chips are down - microfluidic large - scale integration. Trac - Trend Anal Chem 21, Xii - Xiii (2002). 48. Karlinsey, J.M., Monahan, J., Marchiarullo, D.J., Ferrance, J.P. & Landers, J.P. Pressure injection on a valved microdevice for electrophoretic analysis of submicroliter samples. Anal Chem 77, 3637 - 3643 (2005). 49. Li, M.W., Huynh, B.H., Hulvey, M.K., Lunte, S.M. & Martin, R.S. Design and characterization of poly(dimethylsiloxane) - based valves for interfacing continuous - flow sampling to microchip electrophoresis. Anal Chem 78, 1042 - 1051 (2006). 50. de Jong, J., Lammertink, R.G .H. & Wessling, M. Membranes and microfluidics: a review. Lab Chip 6, 1125 - 1139 (2006). 51. Srigunapalan, S., Lam, C., Wheeler, A.R. & Simmons, C.A. A microfluidic membrane device to mimic critical components of the vascular microenvironment. Biomicrofluid ics 5 (2011). 52. Halpin, S.T. & Spence, D.M. Direct Plate - Reader Measurement of Nitric Oxide Released from Hypoxic Erythrocytes Flowing through a Microfluidic Device. Anal Chem 82, 7492 - 7497 (2010). 53. Vogel, P.A., Halpin, S.T., Martin, R.S. & Spence, D. M. Microfluidic Transendothelial Electrical Resistance Measurement Device that Enables Blood Flow and Postgrowth Experiments. Anal Chem 83, 4296 - 4301 (2011). 54. Kuswandi, B., Nuriman, Huskens, J. & Verboom, W. Optical sensing systems for microfluidic devi ces: A review. Anal Chim Acta 601, 141 - 155 (2007). 41 55. Chabinyc, M.L. et al. An integrated fluorescence detection system in poly(dimethylsiloxane) for microfluidic applications. Anal Chem 73, 4491 - 4498 (2001). 56. Erickson, D. & Li, D.Q. Integrated microfl uidic devices. Anal Chim Acta 507, 11 - 26 (2004). 57. Li, X.H., Klemic, K.G., Reed, M.A. & Sigworth, F.J. Microfluidic system for planar patch clamp electrode arrays. Nano Lett 6, 815 - 819 (2006). 58. Swensen, J.S. et al. Continuous, Real - Time Monitoring of Cocaine in Undiluted Blood Serum via a Microfluidic, Electrochemical Aptamer - Based Sensor. J Am Chem Soc 131, 4262 - 4266 (2009). 59. Kovachev, N., Canals, A. & Escarpa, A. Fast and Selective Microfluidic Chips for Electrochemical Antioxidant Sensing in Comp lex Samples. Anal Chem 82, 2925 - 2931 (2010). 60. Harrison, D.J., Manz, A., Fan, Z.H., Ludi, H. & Widmer, H.M. Capillary Electrophoresis and Sample Injection Systems Integrated on a Planar Glass Chip. Anal Chem 64, 1926 - 1932 (1992). 61. Jacobson, S.C., Culb ertson, C.T., Daler, J.E. & Ramsey, J.M. Microchip structures for submillisecond electrophoresis. Anal Chem 70, 3476 - 3480 (1998). 62. Simpson, P.C. et al. High - throughput genetic analysis using microfabricated 96 - sample capillary array electrophoresis micr oplates. P Natl Acad Sci USA 95, 2256 - 2261 (1998). 63. Jacobson, S.C., Ermakov, S.V. & Ramsey, J.M. Minimizing the number of voltage sources and fluid reservoirs for electrokinetic valving in microfluidic devices. Anal Chem 71, 3273 - 3276 (1999). 64. Raymon d, D.E., Manz, A. & Widmer, H.M. Continuous Sample Pretreatment Using a Free - Flow Electrophoresis Device Integrated onto a Silicon Chip. Anal Chem 66, 2858 - 2865 (1994). 65. Wu, D.P., Qin, J.H. & Lin, B.C. Electrophoretic separations on microfluidic chips. J Chromatogr A 1184, 542 - 559 (2008). 66. Jacobson, S.C., Hergenroder, R., Koutny, L.B., Warmack, R.J. & Ramsey, J.M. Effects of Injection Schemes and Column Geometry on the Performance of Microchip Electrophoresis Devices. Anal Chem 66, 1107 - 1113 (1994). 6 7. Mecker, L.C. & Martin, R.S. Integration of Microdialysis Sampling and Microchip Electrophoresis with Electrochemical Detection. Anal Chem 80, 9257 - 9264 (2008). 42 68. Mayers, B.T., Vezenov, D.V., Vullev, V.I. & Whitesides, G.M. Arrays and cascades of fluor escent liquid - liquid waveguides: Broadband light sources for spectroscopy in microchannels. Anal Chem 77, 1310 - 1316 (2005). 69. Ma, B. et al. A hybrid microdevice with a thin PDMS membrane on the detection window for UV absorbance detection. Electrophoresi s 28, 2474 - 2477 (2007). 70. Wang, J., Ibanez, A., Chatrathi, M.P. & Escarpa, A. Electrochemical enzyme immunoassays on microchip platforms. Anal Chem 73, 5323 - 5327 (2001). 71. Sandlin, Z.D., Shou, M.S., Shackman, J.G. & Kennedy, R.T. Microfluidic electroph oresis chip coupled to microdialysis for in vivo monitoring of amino acid neurotransmitters. Anal Chem 77, 7702 - 7708 (2005). 72. Shiddiky, M.J.A., Park, H. & Shim, Y.B. Direct analysis of trace phenolics with a microchip: In - channel sample preconcentration , separation, and electrochemical detection. Anal Chem 78, 6809 - 6817 (2006). 73. Schulze, P., Ludwig, M., Kohler, F. & Belder, D. Deep UV laser - induced fluorescence detection of unlabeled drugs and proteins in microchip electrophoresis. Anal Chem 77, 1325 - 1329 (2005). 74. Effenhauser, C.S., Paulus, A., Manz, A. & Widmer, H.M. High - Speed Separation of Antisense Oligonucleotides on a Micromachined Capillary Electrophoresis Device. Anal Chem 66, 2949 - 2953 (1994). 75. Blazej, R.G., Kumaresan, P. & Mathies, R.A. Microfabricated bioprocessor for integrated nanoliter - scale Sanger DNA sequencing. P Natl Acad Sci USA 103, 7240 - 7245 (2006). 76. Dang, F.Q. et al. Hybrid dynamic coating with n - dodecyl beta - D - maltoside and methyl cellulose for high - performance carbohydra te analysis on poly(methyl methacrylate) chips. Anal Chem 78, 1452 - 1458 (2006). 77. Ye, N.N., Qin, J.H., Shi, W.W. & Lin, B.C. Characterizing doxorubicin - induced apoptosis in HepG2 cells using an integrated microfluidic device. Electrophoresis 28, 1146 - 115 3 (2007). 78. Qin, J.H. et al. Simultaneous and ultrarapid determination of reactive oxygen species and reduced glutathione in apoptotic leukemia cells by microchip electrophoresis. Electrophoresis 26, 1155 - 1162 (2005). 79. Kapishnikov, S., Kantsler, V. & Steinberg, V. Continuous particle size separation and size sorting using ultrasound in a microchannel. J Stat Mech - Theory E (2006). 43 80. Shi, J.J., Huang, H., Stratton, Z., Huang, Y.P. & Huang, T.J. Continuous particle separation in a microfluidic channel v ia standing surface acoustic waves (SSAW). Lab Chip 9, 3354 - 3359 (2009). 81. Huh, D. et al. Gravity - driven microfluidic particle sorting device with hydrodynamic separation amplification. Anal Chem 79, 1369 - 1376 (2007). 82. Kuntaegowdanahalli, S.S., Bhagat , A.A.S., Kumar, G. & Papautsky, I. Inertial microfluidics for continuous particle separation in spiral microchannels. Lab Chip 9, 2973 - 2980 (2009). 83. Phillips, K.S. & Cheng, Q. Microfluidic immunoassay for bacterial toxins with supported phospholipid bi layer membranes on poly(dimethylsiloxane). Anal Chem 77, 327 - 334 (2005). 84. Jiang, H., Weng, X.A. & Li, D.Q. Microfluidic whole - blood immunoassays. Microfluid Nanofluid 10, 941 - 964 (2011). 85. Livak - Dahl, E., Sinn, I. & Burns, M. Microfluidic Chemical Ana lysis Systems. Annu Rev Chem Biomol 2, 325 - 353 (2011). 86. Martinez, A.W. et al. Simple telemedicine for developing regions: Camera phones and paper - based microfluidic devices for real - time, off - site diagnosis. Anal Chem 80, 3699 - 3707 (2008). 87. Ellerbee, A.K. et al. Quantifying Colorimetric Assays in Paper - Based Microfluidic Devices by Measuring the Transmission of Light through Paper. Anal Chem 81, 8447 - 8452 (2009). 88. Martinez, A.W., Phillips, S.T., Whitesides, G.M. & Carrilho, E. Diagnostics for the D eveloping World: Microfluidic Paper - Based Analytical Devices. Anal Chem 82, 3 - 10 (2010). 89. Kim, J., Johnson, M., Hill, P. & Gale, B.K. Microfluidic sample preparation: cell lysis and nucleic acid purification. Integr Biol 1, 574 - 586 (2009). 90. Volpatti, L.R. & Yetisen, A.K. Commercialization of microfluidic devices. Trends Biotechnol 32, 347 - 350 (2014). 91. Mukhopadhyay, R. Hard - soft microfluidic device bypasses drawbacks of PDMS. Anal Chem 81, 5108 - 5108 (2009). 92. Coltro, W.K.T., Lunte, S.M. & Carrilho , E. Comparison of the analytical performance of electrophoresis microchannels fabricated in PDMS, glass, and polyester - toner. Electrophoresis 29, 4928 - 4937 (2008). 93. Galloway, M. et al. Contact conductivity detection in poly(methyl methacylate) - based mi crofluidic devices for analysis of mono - and polyanionic molecules. Anal Chem 74, 2407 - 2415 (2002). 44 94. Shadpour, H. et al. Multichannel microchip electrophoresis device fabricated in polycarbonate with an integrated contact conductivity sensor array. Anal Chem 79, 870 - 878 (2007). 95. Gross, B.C., Erkal, J.L., Lockwood, S.Y., Chen, C.P. & Spence, D.M. Evaluation of 3D Printing and Its Potential Impact on Biotechnology and the Chemical Sciences. Anal Chem 86, 3240 - 3253 (2014). 96. Kodama, H. Automatic Method for Fabricating a 3 - Dimensional Plastic Model with Photo - Hardening Polymer. Rev Sci Instrum 52, 1770 - 1773 (1981). 97. Hull, C.W. Apparatus for production of three - dimensional objects by stereolithography. us patent 4,929,402 (1986). 98. Landers, R. et al. Fabrication of soft tissue engineering scaffolds by means of rapid prototyping techniques. J Mater Sci 37, 3107 - 3116 (2002). 99. Pandey, P.M., Reddy, N.V. & Dhande, S.G. Slicing procedures in layered manufacturing: a review. Rapid Prototyping J 9, 274 - 288 (2003). 100. Melchels, F.P.W., Feijen, J. & Grijpma, D.W. A review on stereolithography and its applications in biomedical engineering. Biomaterials 31, 6121 - 6130 (2010). 101. Stereolithography: Materials, Processes and Applications. Stereolithography: Ma terials, Processes and Applications , 1 - 340 (2011). 102. Waldbaur, A., Rapp, H., Lange, K. & Rapp, B.E. Let there be chip - towards rapid prototyping of microfluidic devices: one - step manufacturing processes. Anal Methods - Uk 3, 2681 - 2716 (2011). 103. Calvert, P. Inkjet printing for materials and devices. Chem Mater 13, 3299 - 3305 (2001). 104. Pfister, A. et al. Biofunctional rapid prototyping for tissue - engineering applications: 3D bioplotting versus 3D printing. J Polym Sci Pol Chem 42, 624 - 638 (2004). 105. Ke yser, E.L., Sears, S. B. Method and apparatus for recording with writing fluids and drop projection means therefor. us patent 3,946,398 (1976). 106. Breen, J., Nottrot, R. & Stellingwerff, M. Tangible virtuality - perceptions of computer - aided and physical modelling. Automat Constr 12, 649 - 653 (2003). 107. Lam, C.X.F., Mo, X.M., Teoh, S.H. & Hutmacher, D.W. Scaffold development using 3D printing with a starch - based polymer. Mat Sci Eng C - Bio S 20, 49 - 56 (2002). 108. de Gans, B.J., Duineveld, P.C. & Schubert , U.S. Inkjet printing of polymers: State of the art and future developments. Adv Mater 16, 203 - 213 (2004). 45 109. Beaman, J.J., Deckard, C. R. Selective laser sintering with assisted powder handling. US PATENT 4,938,816 (1990). 110. Kumar, S. Selective lase r sintering: A qualitative and objective approach. Jom - J Min Met Mat S 55, 43 - 47 (2003). 111. Kruth, J.P., Mercelis, P., Van Vaerenbergh, J., Froyen, L. & Rombouts, M. Binding mechanisms in selective laser sintering and selective laser melting. Rapid Proto typing J 11, 26 - 36 (2005). 112. Song, Y.A. & Koenig, W. Experimental study of the basic process mechanism for direct selective laser sintering of low - melting metallic powder. Cirp Annals 1997 Manufacturing Technology, Volume 46/1/1997 46, 127 - 130 (1997). 1 13. Zhong, W.H., Li, F., Zhang, Z.G., Song, L.L. & Li, Z.M. Short fiber reinforced composites for fused deposition modeling. Mat Sci Eng a - Struct 301, 125 - 130 (2001). 114. Jafari, M.A. et al. A novel system for fused deposition of advanced multiple ceramic s. Rapid Prototyping J 6, 161 - 174 (2000). 115. Feygin, M.S., A.; Diamond M. N.; Dvorskiy, E Laminated object manufacturing system. US patent 5,730,817 (1998). 116. Mueller, B. & Kochan, D. Laminated object manufacturing for rapid tooling and patternmaking in foundry industry. Comput Ind 39, 47 - 53 (1999). 117. Rengier, F. et al. 3D printing based on imaging data: review of medical applications. Int J Comput Ass Rad 5, 335 - 341 (2010). 118. Tumbleston, J.R. et al. Continuous liquid interface production of 3D o bjects. Science 347, 1349 - 1352 (2015). 119. Chen, C.P., Wang, Y.M., Lockwood, S.Y. & Spence, D.M. 3D - printed fluidic devices enable quantitative evaluation of blood components in modified storage solutions for use in transfusion medicine. Analyst 139, 3219 - 3226 (2014). 120. Lee, K.G. et al. 3D printed modules for integrated microfluidic devices. Rsc . Adv . 4, 32876 - 32880 (2014). 121. Tsai, J.H. & Lin, L.W. Micro - to - macro fluidic interconnectors with an integrated polymer sealant. J Micromech Microeng 11, 577 - 581 (2001). 122. Lo, R. & Meng, E. Integrated and reusable in - plane microfluidic interconnects. Sensor Actuat B - Chem 132, 531 - 539 (2008). 123. Han, K.H. & Frazier, A.B. Reliability aspects of packaging and integration technology for microfluidic systems. Ieee T Device Mat Re 5, 452 - 457 (2005). 46 124. Hardin, J.O., Ober, T. J., Valentine, A. D., Lewis, J. A Microfluidic Printheads for Multimaterial 3D Printing of Viscoelastic Inks. Adv Mater 10.1002/adma.201500222 (2015). 125. Paydar, O.H. et al. Characteriza tion of 3D - printed microfluidic chip interconnects with integrated O - rings. Sensor Actuat a - Phys 205, 199 - 203 (2014). 126. Burns, J.M., Yang, X.X., Forouzan, O., Sosa, J.M. & Shevkoplyas, S.S. Artificial microvascular network: a new tool for measuring rheo logic properties of stored red blood cells. Transfusion 52, 1010 - 1023 (2012). 127. Erkal, J.L. et al. 3D printed microfluidic devices with integrated versatile and reusable electrodes. Lab Chip 14, 2023 - 2032 (2014). 128. Krejcova, L. et al. 3D printed chip for electrochemical detection of influenza virus labeled with CdS quantum dots. Biosens Bioelectron 54, 421 - 427 (2014). 129. Bethany C Gross , K.B.A., Jayda E Meisel , Megan I McNitt , Dana M Spence Polymer Coatings in 3D Printed Fluidic Device Channels f or Improved Cellular Adherence Prior to Electrical Lysis. Anal Chem DOI: 10.1021/acs.analchem.5b01202 (2015). 130. Chan, H.N., Chen, Y., Shu, Y., Chen, Y., Tian, Q., Wu, H. Direct, one - step molding of 3D - printed structures for convenient fabrication of tru ly 3D PDMS microfluidic chips. Microfluidics and Nanofluidics DOI 10.1007/s10404 - 014 - 1542 - 4 (2015). 47 Chapter 2 Study of C - peptide Stimulated ATP Release from Erythrocytes on a 3D - printed Circulation - mimic Fluidic Device 2.1 Background 2.1.1 Diabetes Complications Diabetes mellitus is a medical condition that is characterized by uncontrolled and elevated blood glucose levels. It currently affects more than 20 million people in the United States, but this number may increase to 48.3 million peo ple by the year 2050 1 . Based on different pathologies, there are three main types of diabetes: type 1, type 2 and gestational diabetes. Type 1 diabetes stems from a lack of insulin production due - cells. These patients are dependen t on exogenous insulin therapy and t ype 1 diabetics make up about 10 % of all reported cases of diabetes. Type 2 diabetes, or non - insulin dependent diabetes, resulting from a resistance to insulin. Overall, it accounts for about 85 % of all reported cases of diabetes. The third type of diabetes is gestational diabetes, or pregnancy - induced diab etes, which accounts for about 5 % of diabetes in the United States 2 - 4 . Diabetes typically leads to chronic complications, which may result in endless pain and increased mortality. These complications may include nerve damage (neurop athy), blindness (retinopathy), kidney damage (nephropathy), or cardiovascular disease 5 . N early all people with diabetes in the United States suffer some form of neuropathy 6 . Some symptoms of the onset of neuropathy include numbne ss, loss of dexterity, burning, hypersensitivity and difficulty in rising from a 48 seated position 7 . Diabetic retinopathy (DR) is reported to be the leading cause of new cases of blindness 8 , and p eople with diabet ics are about 30 times more likely than a h ealthy individual to become blind 9 . Kidney disease is anot her severe result of diabetes, which leads to leakage of protein into urine and eventually kidney failure 10 . The direct cause of most diabetic complications lies in the dysfunction of the interface between the blood stream and peripheral tissues, or blood vessel barriers 11 . Low and colleagues suggested that neuropathy res ults from poor blood flow that lead to endoneurial hypoxia, which decreases nerve signaling 12 . Retinopathy stems from diminished blood flow that caus es poor oxygen supply, which triggers retinal cells to form new but fragile vessels via a vascular endothelial growth factor mechanism. The will eventually de tach the retina and cause blindness 13 , 14 . Nephropathy occurs when the glomerulus undergoes partial sclerosis or failure to dilate, resulting in hyper filtration and resulting protein leakage 15 . All of these studies indicate that restoration of vasodilation and/or blood flow can be a potential therapy for diabetic complications. 2.1.2 C - peptide and Vasodilation C - peptide and insulin are produced simultaneously from proinsulin in pancreatic - cells 16 . As shown in Figure 2.1, in a proinsulin molecule, the insulin chains A and B are linked by a connecting segment, which will subsequently form C - peptide. Two endopeptidases are involved in cleaving the proinsulin molecule to produce insulin, one cleaving exclusively on the C - terminal side of Arg 31/Arg 32 (B - chain/C - peptide 49 junction), the other on the C - terminal side of Lys 64/ Arg 65 of proinsulin (C - peptide/A - chain junction) 17 . Therefore, the 31 amino acid peptide (C - peptide) is produced and secreted in equimolar amounts with insulin. Since this process was discovered in 1967 18 , C - peptide had long been regarded as a by - produc t of insulin production, with no physiological relevance . Due to the longer half - life of C - peptide (30min) compared with insulin (3min), the main application of C - peptide is its use in m onitoring the - cells of diabetic patients 19 . In the past two decades, studies have shown that C - peptide can reduce diabetic complications. The Diabetes Control and Complications Trial Research Group reported that treatment of hyperglycemia decrease s diabetic complications, but it is not sufficient to fully reverse diabetes, suggesting that there is a missing therapy for diabetic complications 20 . Sjoberg and colleagues found that some type 1 diabetic patients who maintained - cell activity after the onset of the disease experienced less severe - cells were completely destroyed 21 . The evidence - cells are important for control of diabetic complications, and thus C - peptide attracted the attention of a few pioneer researchers. In a recent study, diabetic patients who retain ed detectable amounts of C - peptide after the onset of the disease exhibited lower levels of complications such as retinopathy, microalbuminuria, and neuropathy 22 . 50 Johansson pe rformed multiple studies evaluating the effects of the short - term replacement of C - peptide on renal function in type 1 diabetic patients, finding that C - peptide infusion significantly decreased glomerular hyperfiltration. They also observed 40% and 55% dec reases in albuminuria (urine protein leakage) at 2 and 4 weeks, respectively, indicating that C - peptide potentially ameliorated diabetic nephropathy 23 - 25 . Als o, Forst reported that diabetes - induced cardiovascular complications, such as decreased blood flow in the extremities, were ameliorated by C - peptide therapy 26 . Furthe rmore, C - peptide prevents diabetic neuropathy via improving endoneural blood flow and preventing axonal swelling 27 . Evidence indicates that C - peptide increases blood flow and decrease diabetic complications. However, the mechanism of how C - peptide enhances blood flow remains unknown. Previous studies from the Spence group reported that C - peptide may regulate blood flow via stimulating ATP release from ERYs. However, t he first attempts to correlate C - peptide with increased ATP release were not reproducible. It was later discovered that only when C - peptide was administered with a metal ion, such as Fe 2+ and Cr 3+ , which are common impurities in commercial C - peptide, could an increase of ATP release from ERYs be observed 28 . Subsequent experiments with HPLC - purified C - peptide further confirmed this discovery 29 . Considering the physiological process by which C - peptide is produced and secreted, another metal ion, Zn 2+ , gained interest because of its high abundance 30 (at mM level) in pancreatic - c ell granules. Our group initially hypothesized that upon secretion from - c ells , C - peptide and Zn 2+ were somehow bound and it was this ensemble of C - peptide and Zn 2+ that 51 increases ERY derived ATP. While our group was able to confirm that C - peptide and Zn 2+ were indeed required for biological activity on ERYs, we were never able to confirm that the two species bind to each other. Figure 2.1. P roinsulin. Each sphere repre sent an amino acid. The labeled amino acids represent C - peptide. Upon cleavage, the two chains connected with disulfide bond forms insulin, while the conn ecting segment forms C - peptide. 52 2.1.3 Chapter Motivation and Overview Previous studies about C - experiments, in which, ERYs were incubated in a static buffer under different conditions (e.g. with/without C - peptide). However, under in vivo conditions, ERYs are circulating under shear stress, which is an im portant physical factor that affects ATP release from these cells 31 . Therefore, a fluidic device that enables long - term (e.g. hours) in vitro ERY circulation is needed to mimic in vivo conditions. As discussed in the first chapter, given the drawbacks of PDMS - bas ed microfluidic devices, 3D - printing will be explored to fabricate such a device. Although previous research showed the indispensable role of Zn 2+ in this process, numerous binding studies of C - peptide and Zn 2+ were unsuccessful, which indicates that there is not direct binding between the two species. However, under in vivo conditions, albumin is a common carrier protein, which can transport numerous molecules to cells 32 . Therefore, the hypothesis that albumin carries C - peptide and Zn 2+ to ERYs will also be examined. 53 2.2 Methods 2.2.1 The Design and Fabrication of the Fluidic Device The device was printed on an Objet Connex 350 printer (Stratasys Ltd, Eden Pra irie, MN) in the department of Electrical and Computer Engineering at Michigan State University. The 3 - D design (Figure 2.2) was created using the CAD based engineering software package AutoDesk Inventor Professional (Autodesk, Inc., San Francisco, CA). Th e device was printed using Objet VeroClear material (Stratasys Ltd, Eden Prairie, MN) whose exact composition is proprietary, but approximately contains isobornyl acrylate (15 - 30%), acrylic monomer (15 - 30%), urethane acrylate (10 - 30%), acrylic monomer (5 - 1 0; 10 - 15%), epoxy acrylate (5 - 10; 10 - 15%), acrylate oligomer (5 - 10; 10 - 15%), and a photoinitiator (0.1 - 1;1 - 2%). Once printed, the device is translucent and rigid, but becomes optically transparent and ready to use after a simple polish and cleaning with sa nd paper and water. Commercially available membrane inserts (6.5 mm diameter; Corning, Inc., Horseheads, NY) with polyester membranes (0.4 µm pore diameter ; 10 µ m membrane thickness; 10 8 pores/cm 2 ) were inserted into the dynamic wells above the channels. The inserts function as a semi - permeable barrier between flowing cells or reagents in the channel and reagents that are loaded in the insert (on the opposite side of the membrane). This configuration enables molecular transport through the pores by diffus ion. The side of each membrane insert was wrapped by a layer of PTFE seal tape (PL Sourcing, Inc., Newport News, VA) to enhance the seal between the inserts and the wells. 54 Fig ure 2.2. The engineering sketch of the fluidic device with detai led dimensions 55 Tygon tubing (Saint - Gobain PPL Corp, Jackson, MI) to the ends of a piece of 15 cm Ismatec tubing (Cole - Parmer Instrument Company, Vernou Hills, IL). The two en ds of the tubing were connected to male finger tight adapters (IDEX Health & Science LLC, Oak Harbor, WA), which can be integrated with the two threaded ends of a channel in the printed device, thus forming a loop. The threads were printed directly from th e device design (as opposed to being tapped post - printing) as 10 - 32 type threads, thus enabling the use of the aforementioned male finger tight adaptors. Fluids or samples were driven through the loop by a 12 - roller peristaltic pump (IDEX Health & Science LLC, Oak Harbor, WA). To confirm the absence of bulk fluid movement through the membrane pores into a channel, a 50 µ L aliquot of DDW was added to dynamic inserts in wells B1, E1 and G1, while DDW was delivered through the closed loop system at a flow rat e of 50 µ L/min for 2 hours. The amount of water in each insert after pumping was determined by mass measurement. As a control group, the same amount of DDW was loaded into each insert in static wells A2, D2 and F2 (no channels beneath and thus no possible water delivery) for 2 hours. There is no channel underneath the static inserts, thus representing a control group used to account for any change in volumes in the inserts due to evaporation. All other channels were examined in the same way, using static in serts in the next right column as a control set. 56 To investigate the extent of any leakage of fluids from a channel to the area above membrane inserts, a 30 µ M fluorescein solution in DDW was circulated at a rate of 50 µ L/min in the loop for 2 hours. The de vice was then detached from the pumping system and placed under the objective lens of a fluorescence microscope (Olympus, Japan). Leakage was investigated for all six channels on the device; any detectable quantity of fluorescein on the outside of the tran s - well inserts was considered to be indicative of leaking. 2.2.2 ATP Measurement Optimization The well - established luciferin/luciferase chemiluminescence assay was used for ATP assays 33 . Reagents were prepared by dissolving 2.0 mg of D - luciferin (Sigma Aldrich, St. Louis, MO) in 5 mL of DDW, and adding the resultant solution into a single 100 mg vial of firefly extract (Sig ma Aldrich, St. Louis, MO). Five ATP (Sigma Aldrich, St. Louis, MO) standards ranging in concentration from 0 nM to 800 nM were prepared by dissolved in DDW and diluting in a physiological salt solution (PSS, containing 4.7 mM KCl, 2.0 mM CaCl 2 , 1.2 mM MgS O 4 , 140.5 mM NaCl, 21.0 mM tris - hydroxymethyl aminomethane, 5.5 mM glucose, and 5% bovine serum albumin at pH = 7.4; all reagents were from Sigma Aldrich). A 50 µ L aliquot of PSS was loaded in the insert located in well B. The PSS served as a solution to c ollect ATP diffusing from the channel through the membrane of the insert. After 10 min, 20 min or 30 min of pumping the various ATP standards, the device was detached and placed in the sample holder of the plate reader (Molecular Devices LLC, Sunnyvale, CA ). A 50 µ L aliquot of 20 nM ATP 57 standard solution was pipetted into insert in static well 2A as a calibrator, thereby minimizing possible indeterminate error. A 10 µ L of luciferin/luciferase was then added into inserts in wells 2A and 1B, simultaneously. A fter 15 s, the chemiluminescence intensity from both inserts was detected simultaneously, followed by the evaluation of the detection background, detection limit, sensitivity and linearity for the different pumping times under investigation. To determine t he optimal volume of luciferin/luciferase mixture for each assay, ATP standards were prepared and circulated in a channel, as described above, with 50 µ L of PSS loaded in the insert in well 1B. ATP standards circulated for 20 min at a rate 50 µ L/min. Next, 10 µL to 30 µL aliquots of the luciferin/luciferase mixture were added to the membrane inserts in dynamic well 1B, as well as static well 2A (containing 50 µ L of 20 nM ATP standard as a calibrator) simultaneously. The device was detached and placed in the plate reader sample holder and the optimal assay amount was determined by an evaluation of detection limit, linearity and sensitivity. 2.2.3 Evaluation of Analytical Features of the Device ATP solutions of 0, 0.2, 0.3, 0.4, 0.5, 0.8 µM were prepared in PS S and circulated randomly through 6 channels, simultaneously, at a rate of 50 µ L/min for 20 min after 50 µ L of PSS were loaded in the membrane inserts in wells B above each channel. The device was then detached from the tubing and placed in the plate reade r for chemiluminescence detection with the optimal ATP quantitation parameters (Figure 2.4) . A curve of 58 chemiluminescence intensity versus ATP concentration was obtained, and the resulting regression statistics (slope, y - intercept, and coefficient of determination) were calculated. ATP standards of 150 nM and 250 nM were prepared and circulated in channel 1 at a rate of 50 µ L/min for 20 min, with 50 µ L of PSS buffer loaded in the membrane insert in we ll B (dynamic wells) above the channel to collect ATP by diffusion. A 50 µL aliquot of 20 nM ATP was then added into the membrane insert in wells found in row the sample holder of the plate reader. Chemiluminescence was detected 15 s after adding 10 µ L of the luciferin/luciferase assay into dynamic and static wells to quantify the ATP, using calibration curves prepared in advance in a similar manner as described above. Two more identical measurements were performed on channel 3 and channel 5. ATP standards having concentrations ranging from 0 to 800 nM were prepared and simultaneously circulated through channels 1, 3 and 5 at a rate 50 µ L /min for 20 min, while 50 µ L of PS S buffer were loaded in the membrane inserts in wells B above each channel. After adding 50 µ L of 20 nM ATP to wells 2A, 4A and 6A (as static calibrators for measurements on channels 1, 3, and 5, respectively), the device was detached and placed in the sam ple holder of the plate reader. An aliquot of 10 µ L of the luciferin/luciferase mixture was simultaneously added into inserts above the dynamic and static wells, and chemiluminescence was detected by the plate reader 59 after 15 s. A comparison of sensitivity , detection limit and linearity of results from the 3 channels was performed to determine if the device channels and inserts are statistically equivalent. Precision was also investigated by performing identical measurements as those described above while c hanging well inserts in between studies. 2.2.4 ERY Sample Preparation Whole blood was obtained from healthy human donors by venipuncture and collected into heparinized tubes. Blood was then centrifuged at 500 g for 10 min and the plasma and buffy coat were removed by aspiration. ERYs were washed 3 times in PSS 34 . For all albumin free experiments, ERYs were washed in albumin - free PSS. The hematocrit was determined using a hematocrit measurement device (CritSpin, Iris Sample Processing, Westwood, MA). Four samples were prepared containing ERYs at a 7% hematocrit in differe nt conditions: buffer only, buffer with C - peptide, buffer with both C - peptide and Zn 2+ (ZnCl 2 ), and buffer with Zn 2+ . Buffer choices included PSS and albumin - free PSS in order to study the effect of albumin on C - peptide and Zn 2+ delivery to the ERYs, as we ll as subsequent ATP release from the ERYs. Human C - peptide and Zn 2+ stock solutions (800 samples (10 nM) were then pumped to fill four randomly chosen channels on the device, after which each channel and corresponding tubing were closed to form a loop. The samples with different C - peptide treatments were then introduced into the device 60 to form four circulating loops. The whole setup was then placed in a 37 °C incubator for 2 h ours, during which, C - peptide/Zn 2+ /albumin that were mixed in circulating ERY samples would exert their effect on the cells. The entire setup of pump, device, and closed loops with circulating samples were placed in an incubator at 37 °C. The samples were allowed to circulate through the system for 1.5 h, after which, aliquot s of 50 µL of PSS were loaded in the inserts in row E to collect ATP from samples via diffusion through the pores in the insert membrane. ERY - derived ATP was determined in the same mann er as the ATP standards, other than the fact that a stream of ERYs was delivered through the channels as opposed to ATP standards. (Figure 2.3) 61 Figure 2.3. ERY sample preparation and measurement. A), four ERY samples were prepared with different treatments: without C - peptide or Zn 2+ ; only with C - peptide; only with Zn 2+ ; and with both C - peptide and Zn 2+ , the concentrations of C - peptide and Zn 2+ were both 10 nM. B), the four samples were introduced in to four channels o f the fluidic de vice. After the sampl es circulated for 1.5 hours in the fluidic device that was placed in a 37 o C incubator, a liquots of 50 µ L o f PSS were loaded in the wells in row E to collect ATP from the flowing via diffusion, which was then quantified with the lucife rin/luciferase assay by a plate reader. + - + + - + C - peptide Zn 2+ Erythrocyte in PSS - - A B 62 2.2.5 ERY Derived ATP Determination ATP was quantified by the luciferase/luciferin chemiluminescence assay on the 3 - D printed fluidic device. The device design was modeled as a 96 - well plate for subsequent measur ement on a standard plate reader. Six channels were fabricated in the device, with wells above these channels that align with the internal robotics of the plate reader for measurement convenience. A 6.5 mm diameter transwell insert whose bottom is porous p olyester (0.4 micron) is placed in each of the wells. During use, ERY samples flow through the channels and released ATP is able to diffuse through the membrane pores into buffer that was pre - loaded in the insert above the membrane. Aliquots (10 µ L) of the luciferin/luciferase mixture were then injected into each well, and the entire device was placed in the plate reader for chemiluminescence measurement. Before measuring ERY samples, a calibration curve was obtained by circulating four ATP standards prep ared in PSS (concentrations of 0, 100, 200 and 400 nM) in four randomly chosen channels at a flow rate of 50 µL/min using a peristaltic pump (IDEX Health & Science LLC, Oak Harbor, WA) as part of the pumping mechanism. To facilitate ATP diffusion through t he membrane pores and into the well insert, 50 µL of PSS was loaded in the membrane inserts on row E above each channel. After flowing for 20 minutes, the 3D - printed device was detached from the peristaltic pump tubing and was placed on the holder of a pla te reader (Molecular Devices LLC, Sunnyvale, CA), followed by the simultaneous addition of 10 µL of L/L assay mixture into each insert to measure chemiluminescence intensity values resulting from the reaction of the L/L 63 mixture with ATP. A standard curve w as generated by plotting the chemiluminescence intensity values against the known concentrations of ATP. 2. 3 Results 2. 3.1 Device Design and Fabrication As shown in Figure 2.4A, the device is modeled after the dimensions of a standard 96 - well plate, making it suitable for direct analysis on a plate reader and use with automated fluidic handling systems. The printed column (1 to 12) and row (A to H) markers make it easy to identify and label wells. This particular device enables parallel analyses as it consi sts of six channels, each with 3 wells. Membrane inserts, which are removable trans - well inserts often used in cell culture applications, were plugged into the wells (Figure 2.4B). 3D - printed threads were designed for both ends of each channel to allow for an amenable connection to external tubing via male finger tight adapters (Figure 2.4C). The static wells between channels enable simultaneous calibration and/or internal standards when necessary. In this current application, all 3 wells above each channel were used for dynamic determination of ATP release from ERYs flowing through the channel, while the 4 wells not over a channel were used for calibration by performing static measurements of ATP standards. In this construct, quantitative determinations of ATP could be performed, facilitated by a generated wo rking curve on a single device. Figure 2.4D displays the cross section of a single channel, enabling a view of each membrane insert after placement into the device. An evaluation of Figure 2.4D also show s how diffusion of the analytes could occur, moving 64 from the channel, across the membrane, and to the area above the insert for eventual measurement. Figure 2.4. (A): The 3D - printed fluidic device (bottom) used in this study is modeled after th e dimensions of a 96 - well plate (top). Column and row markers make it convenient to localize wells. Six channels were printed on the odd number columns, with three wells corresponding to each channel. Static wells, printed on the even number columns allow for internal standards or calibration. (B): Membrane inserts (top), which have a semi - permeable polyester membrane, are inserted into wells (bottom) of the 3 - D printed device. (C): Threads, printed at both ends of the channel, connect external tubing throu gh a finger tight adaptor. (D): A schematic cross section of a channel and the membrane inserts. ( E ) : The device locks into the sample holder of the commercial plate reader. 65 In this study, ATP diffuses from channels to wells across the porous membrane, wh ere it is collected into buffer already loaded into the well insert prior to commencement of pumping. Based on diffusion, the amount of ATP accumulated in the well (at some fixed pumping time) is proportional to the concentration in the channel, thereby en abling ATP that is in the channel to be quantitatively determined. After flow of samples in the channels, the device was detached from tubing and directly placed into the plate reader for measurements (Figure 2.4E). Since this device was designed amenable to a plate reader for convenient and high content readouts, the alignment of the device and the wells with the plate reader was characterized and validated prior to studies involving ATP measurements. Six fluorescein standard solutions with concentrations ranging from 0 to 20 µ M were loaded in the six wells in rows B, E and G, respectively, followed by fluorescence detection over the wells by the plate reader (ex. 494 nm; em. 521 nm). Figure 2.5 shows the fluorescence intensity over the wells in row B, as a function of fluorescein concentration. The fluorescence intensity is linearly proportional to the concentration of fluorescein loaded in the wells, which indicates correct alignments of the wells with the optical detectors in the plate reader. 66 Figur e 2.5. Validation of the alignment of the device and printed wells in the plate reader. Six f luorescein solutions (concentrations ranging from 0 to 20 µ M ) were prepared and were loaded into the six inserts in wells B above each channel, respectively. Fluor escence intensity above these inserts was detected by plate reader and was plotted versus fluorescein concentration. The good linearity (r 2 = 0.99) shows that the device and wells are aligned well with the plate reader. n = 3, Error = stdev. 67 The tran s - well inserts are not fixed into place by any type of glue or epoxy adhesives that lead to permanent combination. There is, however, a wrap of PTFE tape around the side of inserts to help them seal onto the bulk device tightly, yet remain removable if nee ded. However, to confirm that bulk fluid movement is not occurring from the channel to the area above the membrane insert, or vice - versa, 50 µ L of doubly deionized water (DDW) were added into membrane inserts in wells B1, E1 and G1 above channel 1. After c irculating DDW through the channel for 2 hours at a rate of 50 µ L/min, the remaining volumes of water in the inserts were measured. A control set was performed by adding 50 µ L of DDW in inserts placed in static wells A2, D2 and F2, and then measuring the r emaining volumes after two hours of loading. As shown in Table 2.1, though a minimal loss of liquid volume in the dynamic inserts (B1, E1 and G1, with flowing underneath) was observed, the lost volume was not likely due to liquid transfer into the channel for a minimal liquid loss was also shown in static inserts (A2, D2 and F2), beneath which, there is no channel. There is no significant difference in remaining volumes between dynamic and static inserts. The results suggest that loaded liquid can be firmly held in dynamic inserts without being drawn into the underlying flow channel for the time duration investigated. The other five channels were examined in the same way, and no bulk liquid transfer was observed. The leakage of fluids from the flowing channe l across the membrane and into any pre - loaded reagents on the other side of the membrane has the potential to be detrimental to the quality of analysis using this fluidic device. For example, the device detects ATP that has diffused from the channel to the inserts, any direct leakage of 68 fluids across the membrane will cause false signals and thus make the detected results unreliable. Also, dilution effect caused by leakage can potentially hinder detection limits of ATP. To determine if flowing fluids leak a cross membrane (as opposed to movement of an analyte by diffusion alone), a 20 µ M fluorescein solution was circulated through channel 1, with nothing loaded in the membrane inserts above the channel. After 2 hours of pumping this solution, fluorescence ima ges above the channel were obtained. Figure 2.6 was integrated by images observed on separate parts of the channel because of the limited camera view. As shown, there is some fluorescein at the bottom of each well shaped in rectangular. This fluorescence w as on the apical side of the membrane (the bottom side, touching the channel), on the channel area, resulting from remaining fluorescein in the channel. If there is any liquid ld be expected. Due to the lack of fluorescein entering the well, it was concluded that leakage of fluids was not occurring. The other five channels were verified with the same method and none showed any leakage of liquid across membrane. This is important because it will help ensure that only molecules can diffuse or move through the membrane to the other side of inserts. 69 Table 2.1. Remaining Volumes of Water in Inserts a: n=10, error=stdev; b: n=5, error=stdev Figure 2.6. A fluorescence image of a channel after flowing fluorescein. The dotted circles indicate the well areas. The image implies that leakage of liquid from channel across the membranes did not occur. Due to the limited scope view, images on separate parts of the channel were obtained and integrated into a single image. Well Remaini ng volume after 2 hour pu mping/ µL a Well Remaini ng volume after 2 hour b B1 49.5 ± 4.5 A2 48.7 ± 3.8 E1 48.6 ± 3.5 D2 49.4 ± 1.3 G1 46.8 ± 6.2 F2 47.6 ± 2.5 70 2. 3.2 ATP Measurement Optim ization M ultiple factors can affect the overall quantitative determination of ATP release from the stored ERYs flowing through the device channels. One of the first factors we investigated was the amount of time the ERYs flow ed through the device channel p rior to determination of ATP. As the ERYs pass through the channel, they release ATP that diffuses through the porous membrane on the bottom of the trans - well insert. Thus, an increase in the amount of time the ERYs pass through the device channels would i ncrease the collection time and amount of ATP in the well above the channel. To determine the optimal ATP collection time, ATP standards (0 to 800 nM) were circulated for 10 min, 20 min or 30 min through a device channel having a well insert that was loade d with 50 µ L of PSS. This circulation time enabled ATP from the ERYs to diffuse through the membrane pores. After the various ci rculation times of the ERYs (or the collection time of ERY - derived ATP), an aliquot of a luciferin/luciferase mixture was added to the well insert and the resultant chemiluminescence was recorded using the multi - well plate reader. Calibration curves were prepared to investigate such figures of merit as analytical sensitivity and limits of detection for each collection time. The res ultant data, which are summarized in Table 2.2, show that lower limits of detection and higher sensitivity can be achieved as the collection time increased. Although quantifiable results were obtained even after 10 minutes of pumping, the 20 minute collect ion period was utilized for subsequent studies because of its precision and significant reduction in the limit of detection in comparison to the 10 minute 71 pumping period and its suitability for measurements of ERY - derived ATP, which is typically in a range of 90 to 400 nM, depending on the health status of the donor. Table 2.2. Optimization of ATP Collection Time Collection time/min Y intercept Slope r 2 Detection limit/nM 10 20.9 ± 5.5 517.5 ± 33.6 0.98 ± 0.00 5 136.6 ± 24.0 20 15.8 ± 1.8 565.4 ± 38.6 0.9 9 ± 0.00 1 52.4 ± 7.3 30 19.8 ± 1.0 883.8 ± 69.9 0.99 ± 0.00 7 33.3 ± 6.8 (n =3; all errors represent standard deviations) A mixture consisting of luciferin/luciferase is required in the chemiluminescent determination of ATP. However, this mixture also co ntributes to high background luminescence during the measurement portion of the analysis. Thus, a study was performed to determine the volume of luciferin/luciferase mixture added during the assay that provided the best analytical features (detection limit , sensitivity, etc.). To perform this study, ATP standards of identical concentration were circulated in a channel for 20 min and ATP was allowed to diffuse through the membranes of the trans - well inserts that had already been filled with 50 µL of buffer, after which, the device was detached and placed on the sample holder of the plate reader. Aliquots of 10, 20 or 30 µ L of the luciferin/luciferase mixture were added to the wells and the 72 chemiluminescence from each trial was acquired. The data from these st udies, shown in Table 2.3, suggest that 10 L of the luciferin/luciferase mixture yielded the lowest background (y - intercept), and the best linearity, as measured by the coefficient of determination, r 2 . In fact, larger volumes of the luciferin/luciferase mixture added to the wells lowered analysis quality, likely due to higher background emission. Additionally, when adding the 20 or 30 L volumes of the luciferin/luciferase mixture, the signals from 100 nM or 200 nM ATP standards were not statistically different and exhibited reduced linearity and higher detection limits. Table 2.3. Optimization of Luciferin/luciferase Assay Volume Assay volume/ L Y intercept Slope R 2 Detection limit/nM 10 15.8 ± 1.8 565.4 ± 38.6 0.99 ± 0.0 01 52.4 ± 7.3 20 36.8 ± 3.7 515.9 ± 21.0 0.97 ± 0.0 2 120.1 ± 13.6 30 43. 8 ± 1.9 549.0 ± 32.0 0.97 ± 0.0 1 167.7± 39.4 (Each assay volume was performed on 5 ATP standards; the resultant calibration curves were evaluated for y - intercept, slope, and linearity which, in turn, were used to determine the detection limit. n =3; all e rrors represent standard deviations) 73 2. 3.3 Evaluation of Analytical Features of the Device With six channels integrated into the 3 - D printed device, all of which were amenable to a commercial plate reader, high - throughput applications were explored. S ix ATP standards (0 to 0.8 M) were circulated in the six channels, but in a random order. The membrane inserts in wells B on each channel were loaded with 50 L of PSS to collect ATP by diffusion. The chemiluminescence intensity detected from each insert was plotted as a function of the concentration of ATP flowing in the channels. The linearity (r 2 =0.99) and precision of measurement demonstrated the device can perform 6 quantitative analyses, simultaneously. Under optimal conditions, ATP standards of kno wn concentrations (150 nM and 250 nM) were evaluated with the device. These two concentrations were determined because a 7% solution of ERYs release ~200 nM ATP. Such measurements were performed on channels 1, 3 and 5. Data shown in Table 2.4 indicate that the levels of ATP can be detected quantitatively, and the detection results from the three channels did not show any statistically significant difference, which further suggests that quantitation on different channels will yield statistically similar resu lts. Such results are expected due to all channels being printed on the same printer using the same prototype dimensions. To confirm precision between channels, ATP standards were circulated in three channels and resultant calibration curves were compared. As shown in Table 2.4 , these channels were statistically equal in terms of background, sensitivity, linearity, detection limit and accuracy. This intra - channel reproducibility is another outstanding advantage of the 3 - D printed device, as it helps reduce variability 74 of devices from different manufactures and fabrication protocols. Another key advantage is the reusability of the 3 - D printed device; the same device was used for all reported experiments and was cleaned with a simple rinse with DDW. Data in Fi gure 2.5, Table 2.2 and Table 2.3 were obtained using the same channel on the same device. The error bars, which represent standard deviation, support the reusability of the device, which not only reduces use of materials, but also enables repeated measure ments of the same sample on the same device in a reproducible manner. Table 2. 4. Observed quantitative detection of 150 nM and 250 nM ATP from channel 1, channel 3 and channel 5 ATP standards of known concentrations were introduced to channels 1, 3, and 5 (Ch1, Ch3, and Ch5) in the fluidic device, and were subsequently quantified with the luciferin/l uciferase assay by a plate reader. The observed concentrations are the average ± stdev of 5 experiments. Calibration curve was made each time for quantitat ion purpose. Known concentration/nM 150 250 Observed Concentration/nM Ch1 Ch3 Ch5 Ch1 Ch 3 Ch5 140.6 ± 5.8 154.4 ± 7.2 159.4± 8.6 245.5 ± 15.6 253.7 ± 18.7 251.6± 14.0 75 2.3.4 Investigation of ERY Derived ATP Release stimulated by C - peptide As shown b y the black bars in Figure 2. 7 ERYs incubated with C - peptide and Zn 2+ in PSS stimulated the highest ATP release (319.8 ± 15.2 nM), which was significantly higher than the ERYs in the absence of C - peptide/ Zn 2+ in PSS as a control group (194.9 ± 19.7 nM, p < 0.005). ERYs incubated with C - peptide or Zn 2+ alone, however, did not show significant ATP release increase. These results further confirmed that, under hemodynamic conditions, C - peptide can increase ATP release from ERYs, only with the presence of Zn 2+ . However, binding studies using isothermal titration calorimetry (ITC) did not reveal any binding between C - peptide and Zn 2+ , it was then proposed that albumin, which is a common transporter for molecules in a blood stream, delivers C - peptide and Zn 2+ to E RYs. Subsequent ITC experiments proved the binding between albumin and C - peptide and Zn 2+ . However, the effect of C - peptide and Zn 2+ on ERY derived ATP release, in the absence of albumin, remained unknown. Thus, the same sample as aforementioned were prepa red, but in albumin - free PSS. The ATP measurements from these samples are summarized in Figure 2. 7 (grey bars). The data suggest that without albumin, C - peptide and Zn 2+ failed to increase ATP release from ERYs. 76 Figure 2. 7. A TP release from circulating ERYs with different treatments. The data summarized in black bar s were collected from ERY s flowing in PSS, which suggest that C - peptide and Zn 2+ together significantly increased ATP release, while C - peptide or Zn 2+ alone did not show this effe ct. Without albumin in the ERY streams, however, C - peptide and Zn 2+ fail to stimulate ATP release (grey bar s ), which confirmed the role of albumin in the ATP stimulation process. In this experiment, 10 nM C - peptide and Zn 2+ were used to treat ERY samples. (N=5, error=S.E.M) It has been reported that the amino acid at position 27 in C - peptide, glutamic acid, plays a crucial role in its physiological efficacy. Therefore, the mutant E27A (replaced glutamic acid with alanine) was applied to observe if it can in crease ATP release from ERYs. Figure 2. 8 shows with or without albumin, E27 A failed to enhance ERY - released ATP. 77 Figure 2.8. ATP release from ERYs treated with C - peptide/Zn 2+ and E27A/Zn 2+ . Unlike C - peptide/Zn 2+ , E27A/Zn 2+ can not stimulat e ATP release from ERYs. 2.4 Discussion A 3D - printed fluidic device was designed and implemented for in vitro studies of ATP release from circulating ERYs. Compared to conventional PDMS - based fluidic devices, the 3D - printed device is more rugged and robust, an d is capable of multiple uses over a month - long study, which enhances reproducibility of experimental results. To be specific, the leakage examination showed that there was no leakage or bulk liquid transfer occurred across porous membranes on the device, which in contrast , is a common problem/concern for PDMS based devices. Also, the six channels on the device and its reusab ility significantly enhance throughput and experiment efficiency. 78 With 3D - printing, the device was easily modeled after a 96 - well plat e, which is amenable to a commercial plate reader for efficient and high throughput analysis. Moreover, the design of static wells on the device enables simultaneous on - chip calibration, which not only simplifies experimental procedures, but also guarantee s result reliability. The device also possesses some other features that cannot be easily achieved in PDMS. For example, threads were directly printed at the ends of each channel, allowing it to be connected to external tubing by standard finger tight adap ters . This 3D - printed device facilitates the progress of developing C - peptide to be a potential therapy for type 1 diabetic (T1D) complications. In the past few decades, numerous animal tests and small scale clinical surveys have proven C - peptide is effe ctive to alleviate T1D complications. However, the underlying mechanism remained unknown, which has hindered subsequent research and clinical studies for the development of C - peptide therapy. The Spence group has found that C - peptide can increase ATP relea se from ERYs, with the presence of Zn 2+ , which can then enhance vasodilation by stimulating NO production from endothelial cells. However, previous experiments were performed by mixing ERYs with C - peptide/Zn 2+ in a test tube, followed by incubation at 37 º C. In other words, ERYs in these experiments were in a static condition, which cannot represent the actual physiological condition of ERYs, which are circulating under shear stress. With the 3D - printed fluidic device, however, an in vitro ERY circulation m imic can be easily achieved. The results shown in Figure 2. 7 confirmed previous observations that C - peptide and Zn 2+ are able to stimulate ATP 79 release from ERYs, while C - peptide or Zn 2+ alone does not show the effect. It was then anticipated that there mus t be a binding between C - peptide and Zn 2+ . Unfortunately, numerous attempts to demonstrate C - peptide binding to Zn 2+ were unsuccessful, therefore suggesting that another component in the system may be participating in the delivery of C - peptide and Zn 2+ to ERYs. Because of its well - known capability of carrying proteins, drugs and other molecules to cells, albumin was hypothesized to be the missing carrier of C - peptide and Zn 2+ . As discovered by Liu and colleagues, ITC experiments confirmed that albumin can bind C - peptide and Zn 2+ in a two - phase manner. It was also discovered that without albumin in the system, C - peptide and Zn 2+ cannot be delivered to ERYs 35 . All the evidence indicates that albumin is indispensable for C - peptide and Zn 2+ to play the role of increasing ERY - released ATP, which can be confirmed by the data in Figure 2. 7 that C - peptide and Zn 2+ failed to stimulate ERY released ATP if albumin is absent in the system. There were also numerous reports showing that the 27 position amino acid of C - peptide, glutamic acid, is required for C - peptide to af fect cells 36 , 37 . Therefore, an E27 mutant to alanine (E27A) was prepared and ITC experiments suggested this m utant cannot bind albumin. Efficacy studies indicated that this mutant cannot stimulate ATP release from ERYs, even if both Zn 2+ and albumin were present. Collectively, it is the ensemble of albumin, C - peptide, and Zn 2+ that enhances ATP release from ERYs. Although the binding structure is not completely understood yet, the 27 position glutamic acid has proven to be involved. 80 2.5 Conclusion Microfluidics has become a powerful tool in the field of bioanalytical chemistry. However, traditional microfluidic de vices fabricated in PDMS possess some disadvantages, such as lack of ruggedness and reusability, making it almost impossible to move microfluidics out of the academic realm and into real life applications. In this work, however, 3D - printing has been explor ed to fabricate a novel and functional microfluidic device. This device possesses numerous advantages over PDMS devices. It is rugged, reusable, high throughput and amenable to a plate reader, which significantly enhances the quality of data acquired on th e device. Moreover, unlike PDMS microfluidics, device examination experiments showed that no leakage or liquid bulk transfer occurred on the 3D - printed device. The printed threads that fit commercial finger tight adapters also prevent the problems of dead volumes and leakage at connection ports. The device was then applied to realize an in vitro circulation to study the effect of C - peptide and Zn 2+ on circulating hemodynamic ERYs. The results suggest that the three - molecule ensemble that contains C - peptide , Zn 2+ , and albumin can increase ATP release from ERYs, which can then stimulate vasodilation via stimulating NO production from the endothelium. Though the binding structure of the ensemble is not fully understood, the 27 position glutamic acid of C - pepti de has proven to be involved in the process. These results reveal a potential therapy for diabetic complications. 81 REFERENCES 82 REFERENCES 1. Narayan, K.M.V., Saaddine, J.B., Boyle, J.P., Thompson, T.J. & Geiss, L .S. Impact of recent increase in incidence on future diabetes burden - US, 2005 - 2050. Diabetes Care 29, 2114 - 2116 (2006). 2. Inzucchi, S. et al. Diagnosis and Classification of Diabetes Mellitus. Diabetes Care 33, S62 - S69 (2010). 3. Medvescek, M. Classific ation and Criteria for the Diagnosis of Diabetes - Mellitus and Other Categories of Glucose - Intolerance. Zdrav Vestn 52, 183 - 187 (1983). 4. Harris, M. Classification and Diagnosis of Diabetes - Mellitus and Other Categories of Glucose - Intolerance. Diabetes 28, 1039 - 1057 (1979). 5. Rich, S.S. Genetics of diabetes and its complications. J Am Soc Nephrol 17, 353 - 360 (2006). 6. Backonja, M. et al. Gabapentin for the symptomatic treatment of painful neuropathy in patients with diabetes mellitus - A randomized contro lled trial. Jama - J Am Med Assoc 280, 1831 - 1836 (1998). 7. Airey, M., Bennett, C., Nicolucci, A. & Williams, R. Aldose reductase inhibitors for the prevention and treatment of diabetic peripheral neuropathy. Cochrane Db Syst Rev (2005). 8. Wilkinson, C.P. e t al. Proposed international clinical diabetic retinopathy and diabetic macular edema disease severity scales. Ophthalmology 110, 1677 - 1682 (2003). 9. Rubanyi, G.M. & Vanhoutte, P.M. Superoxide Anions and Hyperoxia Inactivate Endothelium - Derived Relaxing F actor. Am J Physiol 250, H822 - H827 (1986). 10. Bennett, P.H. et al. Screening and Management of Microalbuminuria in Patients with Diabetes - Mellitus - Recommendations to the Scientific - Advisory - Board of the National - Kidney - Foundation from an Ad - Hoc - Committe e of the Council - on Diabetes - Mellitus of the National - Kidney - Foundation. Am J Kidney Dis 25, 107 - 112 (1995). 11. Sprague, R.S., Stephenson, A.H., Bowles, E.A., Stumpf, M.S. & Lonigro, A.J. Reduced expression of G(i) in erythrocytes of humans with type 2 di abetes is 83 associated with impairment of both cAMP generation and ATP release. Diabetes 55, 3588 - 3593 (2006). 12. Low, P.A., Schmelzer, J.D., Ward, K.K., Curran, G.L. & Poduslo, J.F. Effect of Hyperbaric Oxygenation on Normal and Chronic Streptozotocin Diab etic Peripheral - Nerves. Exp Neurol 99, 201 - 212 (1988). 13. Pierce, E.A., Avery, R.L., Foley, E.D., Aiello, L.P. & Smith, L.E.H. Vascular Endothelial Growth - Factor Vascular - Permeability Factor Expression in a Mouse Model of Retinal Neovascularization. P Nat l Acad Sci USA 92, 905 - 909 (1995). 14. Sheetz, M.J. & King, G.L. Molecular understanding of hyperglycemia's adverse effects for diabetic complications. Jama - J Am Med Assoc 288, 2579 - 2588 (2002). 15. Craven, P.A., Caines, M.A. & Derubertis, F.R. Sequential Alterations in Glomerular Prostaglandin and Thromboxane Synthesis in Diabetic Rats - Relationship to the Hyperfiltration of Early Diabetes. Metabolism 36, 95 - 103 (1987). 16. Kitabchi, A.E. Proinsulin and C - Peptide - Review. Metabolism 26, 547 - 587 (1977). 1 7. Marques, R.G., Fontaine, M.J. & Rogers, J. C - peptide - Much more than a byproduct of insulin biosynthesis. Pancreas 29, 231 - 238 (2004). 18. Steiner, D.F., Cunningh.D, Spigelma.L & Aten, B. Insulin Biosynthesis - Evidence for a Precursor. Science 157, 69 7 - & (1967). 19. Wahren, J., Johansson, B.L. & Wallberghenriksson, H. Does C - Peptide Have a Physiological - Role. Diabetologia 37, S99 - S107 (1994). 20. Grp, D.R. Effect of Intensive Diabetes Treatment on the Development and Progression of Long - Term Complicati ons in Adolescents with Insulin - Dependent Diabetes - Mellitus - Diabetes Control and Complications Trial. J Pediatr - Us 125, 177 - 188 (1994). 21. Sjoberg, S. et al. Residual Insulin Production, Glycemic Control and Prevalence of Microvascular Lesions and Polyn europathy in Long - Term Type - 1 (Insulin - Dependent) Diabetes - Mellitus. Diabetologia 30, 208 - 213 (1987). 84 22. Keenan, H.A. et al. Residual Insulin Production and Pancreatic beta - Cell Turnover After 50 Years of Diabetes: Joslin Medalist Study. Diabetes 59, 2846 - 2853 (2010). 23. Johansson, B.L., Sjoberg, S. & Wahren, J. The Influence of Human C - Peptide on Renal - Function and Glucose - Utilization in Type - 1 (Insulin - Dependent) Diabetic - Patients. Diabetologia 35, 121 - 128 (1992). 24. Johansson, B.L., Kernell, A., Sjobe rg, S. & Wahren, J. Influence of Combined C - Peptide and Insulin Administration on Renal - Function and Metabolic Control in Diabetes Type - 1. J Clin Endocr Metab 77, 976 - 981 (1993). 25. Johansson, B.L. et al. Beneficial effects of C - peptide on incipient nephr opathy and neuropathy in patients with Type 1 diabetes mellitus. Diabetic Med 17, 181 - 189 (2000). 26. Forst, T., Kunt, T., Pfutzner, A., Beyer, J. & Wahren, J. New aspects on biological activity of C - peptide in IDDM patients. Exp Clin Endocr Diab 106, 270 - 276 (1998). 27. Sima, A.A.F., Zhang, W.X. & Grunberger, G. Type 1 diabetic neuropathy and C - peptide. Exp Diabesity Res 5, 65 - 77 (2004). 28. Meyer, J.A., Froelich, J.M., Reid, G.E., Karunarathne, W.K.A. & Spence, D.M. Metal - activated C - peptide facilitates g lucose clearance and the release of a nitric oxide stimulus via the GLUT1 transporter. Diabetologia 51, 175 - 182 (2008). 29. W. Medawala, P.M., A. Giebink, J. Meyer, C. J. Ku, D. M. Spence A Molecular Level Understanding of Zinc Activation of C - peptide and its Effects on Cellular Communication in the Bloodstream. Rev Diabet Stud 6, 148 - 158 (2009). 30. Scott, D.A. & Fisher, A.M. The insulin and the zinc content of normal and diabetic pancreas. J Clin Invest 17, 725 - 728 (1938). 31. Chien, S., Usami, S., Dellen ba.Rj & Gregerse.Mi Shear - Dependent Deformation of Erythrocytes in Rheology of Human Blood. Am J Physiol 219, 136 - & (1970). 32. Elsadek, B. & Kratz, F. Impact of albumin on drug delivery - New applications on the horizon. J Control Release 157, 4 - 28 (2012) . 33. Sevin, B.U. et al. Application of an Atp - Bioluminescence Assay in Human - Tumor Chemosensitivity Testing. Gynecol Oncol 31, 191 - 204 (1988). 85 34. Carroll, J.S., Ku, C.J., Karunarathne, W. & Spence, D.M. Red blood cell stimulation of platelet nitric oxide production indicated by quantitative monitoring of the communication between cells in the bloodstream. Anal Chem 79, 5133 - 5138 (2007). 35. Liu, Y.L., Chen, C.P., Summers, S., Medawala, W. & Spence, D.M. C - peptide and zinc delivery to erythrocytes requires the presence of albumin: implications in diabetes explored with a 3D - printed fluidic device. Integr Biol - Uk 7, 534 - 543 (2015). 36. Wahren, J. & Larsson, C. C - peptide: New findings and therapeutic possibilities. Diabetes Res Clin Pr 107, 309 - 319 (2015). 37 . Pramanik, A., Olsson, M., Langel, U., Bartfai, T. & Rigler, R. Fluorescence correlation spectroscopy detects galanin receptor diversity on insulinoma cells. Biochemistry - Us 40, 10839 - 10845 (2001). 86 Chapter 3 An Organs - on - a - Chip platform to Study the Interactions between - cells and Blood Vessels 3.1 Background T he results in chapter 2 show that the ensemble of albumin, C - peptide and Zn 2+ elicits biol ogical activity, increasing ERY - released ATP. However, in these experiments, t hough ERYs were subject to flow, the C - peptide and Zn 2+ were added exogenously, which may not represent the physiological process by which C - peptide/Zn 2+ interacts with ERYs. Under in vivo conditions, C - peptide and Zn 2+ are secreted from pancreatic - cell s continuously into the blood stream 1 . It is difficult to control the amount of exogenous ly added C - peptide and Zn 2+ to mimic the gradually secreted amount in vivo . Furthermore, the structure of C - peptide and Zn 2+ upon secretion cannot be represented well in static experiments. These shortcomings led us to an in vitro platform that combines C - peptide secreting cells and a flowing stream of blood cells. The recent concept of Organs - on - a - Chip facilitates the development of such a device. 3.1.1 Organs - on - a - Chip The development of microfluidic technology has made it possible to integrate and cultu re cells on microfluidic devices, with manageable liquid or gas flow; these advances gradually developed into a new research field, Organs - on - a - Chip 2 . The concept of Or gans - on - a - chip represents a class of micro devices for in vitro tissue 87 culture that can mimic key structures and functions of in vivo tissues and organs 3 , 4 . This concept can be further developed to Human - on - a - Chip, which combines different in vitro organ models on one device through a circulation mimic, for pharmacology and fundamental physiology studies 5 - 8 (Figure 3.1). F igure 3.1. The concept of Human - on - a - Chip an d its potential applications in fundamental physiological studies and in pharmaceutical research . 88 3.1.2 C ell Selection for Organs - on - a - Chip T o construct an on - chip organ that is functional and representative of in vivo conditions, the first technical issue that needs to be addressed is the selection of cell/tissue type to build the organ of interest 5 . There are three types of cells/tissues that are widely used. Each of them has its own advantages and disadvantages, and the most appropr iate selection will be based on the specific goals of a study. Because cell - cell communication, and cell - extracellular matrix (ECM) interaction play an important role in cell physiology, a fresh biopsy, which preserves the original tissue architecture, is the most reliable in vitro organ mimic s . Ex vivo pancreatic islets have been successfully adapted to Organs - on - a - Chip devices by different researchers. For example, Roper monitored hormone secretion from islets on microfluidic chips, with simultaneous on - c hip detection 9 . Mohammed re ported an islet - on - chip device to characterize islets before transplantation 10 . Some other biopsies such as brain slice and heart tissue have also been successfully integrated onto microfluidic devices 11 - 14 . Though biopsies are more representative of in vivo environments, they only remain viable for a short time ( a few hours) after being extracted 15 , 16 . Therefore, Organ - on - Chip integrated with ex vivo tissues are limited to assays that do not require much time. Also, the availability of human ex vivo tissues is limited, which restric ts its application in human - oriented physiological and pharmaceutical studies. 89 P rimary cells are directly isolated from animals, making them a good mimic of in vivo conditions. Therefore, primary cells have become the most common cell source for Organ - o n - a - Chip platforms. Using proper protocols that have been developed, primary cells can remain viable and functional for weeks after isolation, which enables long term physiological observation and studies 17 , 18 . However, primary cells also possess noticeable disadvantages. Although a single primary cell can represent its in vivo counterpart, the lack of cell - cell, an d cell - ECM interaction s , and three dimensional (3D) architectures can compromise the relevance of the results from corresponding studies. Several advances in the field of cell biology, such as 3D cell culture and cell co - culture assays help overcome some o f the drawbacks. 3D cell culture models apply porous ECM gels as scaffolds for living cell cultures, to mimic tissue - and organ - specific micro architectures 2 , 19 . Goral and colleagues have successfully cultured hepatocytes in a 3D manner and integrated them onto a microfluidic platform 20 . An in vivo tissue/organ is comprised of different types of cells, the interaction between which usually plays an important role in maintaining normal functionality and activity 21 . Cell co - culture assays enable interactive culture of primary cells that can be applied to microfluidic platforms , as shown by the recapitulation of an on - chip blood vessel. A blood vessel consists of two layers, the endothelial cells (EC) that form a continuous lining for blood flow, and the smooth muscle cell (SMC) layer 22 . The interaction between EC and SMC regulates vessel dilation and thus is associated with 90 complications of several diseases including hypertension and diabetes. Wallace and colleagues have developed differen t co - culture assays for EC and SMC 23 . Though primary cells can mimic in vivo conditions, there are still limitations. The process to isolate primary cells can be labor intensive, time consuming and costly. Moreover, the availability of primary c ells is also limited, which is especially true in the case of human studies. Therefore, cell lines, which are less representative than primary cells, but are suitable for large volume pre - tests in a short time period, can be also used for Organs - on - a - Chip fabrication. B oth tissue biopsy and primary cells have the problem of limited availability. With the development of stem cell technology, especially human induced pluripotent stem cells (hiPSCs), it is possible to imagine collecting a body cell from a huma n and inducing it to hiPSCs, which then differentiate to different tissue/organ models on specific 3D culture scaffolds 24 . The Organ - on - a - Chip studies starting f rom stem cells are still in their infancy, yet appear to be very promising 25 - 27 . The pioneering work in this field started from on - chip stem cell culture and diff erentiation 28 . It is not easy to culture and differentiate stem cells because many factors such as media flow rate, chemical environments ( e.g. , gradient of growth factor), and mechanical factors are hard to regulate in conventional cell culture setups 29 . Microfluidics, however, have shown success in stem cell culture by controlling media flow and manipulating the chemical environment well. Recently, Wan and colleagues 91 obtained functional cardiomyocyte tiss ue derived from embryonic stem cells on a microfluidic device 30 . 3.1.3 Reconstruction of in vivo Microenvironments M echanical factors, such as shear stress, have long been recognized as a key determinant of tissue/organ function by regulating cell morphological polarization, molecule transportation and pathway activation 31 , 32 . Therefore, mechanical factors should be taken into full consideration when establishing an Organ - on - a - Chip system. The elasticity property of PDMS enables the integration of mechanical effects on a microflu idic chip, a well - known example of which is the lung - on - chip model developed by Ingber 33 . Gas exchange in the lungs occurs at the interface of alveoli and surroundin g capillaries. The alveoli are lined with epithelial cells, while the capillaries are lined with endothelial cells. During a respiratory process, with the expansion and contraction of an alveolus, both epithelial cells and endothelial cells experience cycl ic stretching, while the endothelial cells are also under forward shear stress caused by blood flow 34 model mimicked all these mechanical factors. This device mainly consisted of a center channel and two side chambers. The centered channel was used to mimic the alveoli - capillary interface, with a piece of elastic membrane between the upper and lower halves of the channel. Epithelial and endothelial cells were cultured on the top and bottom sides of the membrane, which was pre - coated with fibronectin and collagen. Liquid was pumped through the lower half of the channel, in contact with the 92 endothelial cells, to represent blood flow, while the upper half was filled with air in order to mimic the alveoli - capillary interface. A respiratory process was achiev ed by stretching the cell loaded membrane, using vacuum through the side chambers. With this device, the effect of mechanical shear stress on nanoparticle translocation was investigated. Another example is an on - chip blood vessel model reported by Zheng th at can apply both forward shear stress and cyclic shear stress on endothelial cells 32 . B ecause of the difference between in vitro experimental conditions and in viv o physiological conditions, Organs - on - a - Chip models need to provide appropriate chemical microenvironments for the tissues to perform normally, which is especially true for ex vivo biopsy and primary cell culture. These chemical microenvironments mainly in clude cell nutrient supply, growth factor manipulation and chemotaxis. The viability of extracted biopsies is usually short (within hours), which makes it difficult to perform long term or time - consuming studies. Microfluidic perfusion can be used to simul ate hemodynamics in a tissue biopsy and thus reduce the risk of necrosis in a short time. For example, Sankar and colleagues reported a microfluidic perfusion culture system for isolated mouse pancreatic islet studies that used hydrodynamic traps to mainta in the morphology and functions of the tissue 35 . Additionally, oxygen levels are a crucial determinant of the physiology of many cell types. An on - chip oxygen control device has been reported by Gor al and colleagues where they applied an oxygen controlling unit in the microfluidic device for human hepatocyte culture 36 . 93 Most on - chip stem cell studies are still focused on the early st age of characterizing chemical conditions that preserve self - renewal and differentiation, for which on - chip chemical gradient manipulation can be a powerful tool. Chung and colleagues developed a microfluidic device with a concentration gradient control me chanism, to study the effect of growth factor concentrations on neuronal - stem cell behavior; using this system, they discovered a correlation between growth factor concentration and differentiation rates. In some research cases, such as cancer studies, che motaxis needs to be evaluated 37 . Torisawa reported a double layer PDMS microfluidic device that applied chemical gradients along the channel to study immigration of cancer cells. 38 As mentioned previously, cell - cell and cell - ECM interaction in the correct micro architecture determines the degree of resemblance of in vitro models to in vivo condition s. Many 3D cell culture protocols and cell co - culture assays have been invented that have advanced the development of Organs - on - a - Chip models. 3.1.4 Current development of Organs - on - a - Chip Combining the advances in microfluidics and tissue engineering, dif ferent Organs - on - a - Chip models have been developed for pharmaceutical and fundamental science studies, as summarized in Table 3.1. 94 Table 3.1 . Summary of currently developed Organs - on - a - Chip models System O rgan model Study Emphasis Cardiovascular H eart Build contractible cardiomyocytes 39 stem cells to cardiomyocytes 40 B lood vessel Recapitulation of artery structure 41 Angiogenesis 4 2 N anoparticle translocation 43 EC and SMC co - culture 23 C ell interaction and on - chip analysis 44 Nervous B rain E lectrophysiology; neural circuits 14 Nerv e A xon - glia interaction 45 Tumor angiogenesis 46 Neuromuscular junction Neuromuscular contact 47 Blood - Brain Barrier T he barrier function and drug passage 48 Skeletal and Muscular B one L acuna - canalicular network 49 M uscle Recapitulation of contractible muscle 50 E xercise inducible muscle response 51 Digestive Gastro - intestine Re capitulation of intestine tissue 52 D rug toxicity 53 S ubstance intestinal absorption 54 Recapitulation of intestinal villi 55 Liver Recapitulation of sinusoid 56 Serum protein synth esis 57 , 58 Bi le canaliculi 59 , 60 Assessment of metabolites and toxicity 61 L iver intestinal interactions 62 L ong term repeat dose experiment 63 E ndocrine P ancreatic islet I slet characterization before transplantation 10 M onitoring of secretion 16 R espiratory Lung M ech anophysiological factors 33 Lymphatic Spleen Recapitulation of spleen tissue 64 drainage Lymphatic drainage system on chip 65 Disease Models Breast Malignant tumor invasion 66 Cancer metastasis 38 , 67 3D cultured tumor c ells B lood supply in tumor 68 D rug assessment 69 Drug dosing 70 Photody namic therapy 71 95 3.1. 5 Challenges and Future Directions Previously developed microfluidic Organs - on - a - Chip models have proven capable of simulating parts of huma n physiology, which can potentially facilitate fundamental studies such as physiology, pathogenesis and organogenesis, as well as pharmaceutical studies such as drug screening, assessment, and toxicity tests. Although some organ - organ interaction proof - of - concept models have been developed, most of the reported Organ - on - a - Chip systems are single organ/tissue models 5 . To enhance the applicability of such devices, multiple organs that are physiologically relevant need to be integrated to realize the Human - on - a - Chip concept 72 . Paradoxically, almost all Organs - on - a - Chip device s are currently fabricated in PDMS (some are in other soft materials such as hydrogel s ), which limi ts the integration capacity of multiple organs on a device because of the poor rigidness of the material. Other disadvantages of PDMS microfluidic devices, such as poor reproducibility and single use, make it nearly impossible to develop a standard and uni versal method to fabricate Organs - on - a - Chip devices, leading to inter - lab irreproducibility and prevention of such devices from being used for practical and industrial applications 2 . Therefore, new techniques that can fabricate more rugged and standard microfluidic devices, such as 3D - printing, need to be explored to make Organs - on - a - Chip models. Organs - on - a - Chip is an exciting research area that is highly interdiscipli nary between biology, chemistry and engineering. Thus, advances in these fields need to 96 be effectively leveraged to build novel Organs - on - a - Chip systems. The integration of stem cells on microfluidic devices is a good example, with the potential to produce physiologically representative tissues/organs on a chip. The development of hiPSCs makes it possible to establish patient - specific Human - on - a - Chip systems, for more accurate and personalized drug assessment and dosing determination. Also, to achieve the f ull potential of Organs - on - a - Chip devices, bioanalytical sensors such as optical sensors, electrodes and label free indicators need to be incorporated, all controlled by the appropriate software. Summarily, Organs - on - a - Chip can be a revolutionary technique in the fields of physiology and human healthcare because of its potential applications in fundamental biological research and pharmaceutical sciences. The development of Organs - on - a - Chip in the past decade has progressed based on different proof - of - concep t organ models. However, it is time to advance this technique from cutting - edge regeared towards real applications. 3.1.6 Chapter Overview As discussed above, the platform that will be used to study the interactions between endogenous C - peptide and flowing blood will contain C - peptide secreting cells, blood cells (ERYs) and endothelial cells as a blood barrier mimic. Currently developed Organs - on - a - Chip models, based on PDMS microfluidic devices, are mainly single organ models that cannot achieve the desira ble features of the multiple organ platform. Also, the lack of ruggedness and robustness of PDMS will significantly hinder 97 the establishment of such a platform. Therefore, the development of a novel Multi - Organs - on - a - Chip model based on 3D - printed fluidic devices will be investigated. H uman ERYs will be used as part of the blood flow mimic, and commonly used bovine pulmonary artery endothelial cells (bPAECs) will be used to mimic the endothelium 73 . Rat insulinoma cells (INS - 1 cells) will be used for the C - peptide se creting cells. Although great efforts have been exert ed into isolating a pancreatic - cell line that can respond to glucose normally, a normal - cell line of human origin has not yet been established. The most widely used - cell lines, including RIN (rat insulinoma cell line), HIT (hamster pancreatic beta cells), MIN (transgenic C57BL/6 i nsulinoma cell line), INS - 1, and and rats 74 . Among these cell lines, the INS - 1 represents many imp ortant characteristics of the pancreatic - cells, including high insulin content and responsiveness to glucose at physiological range 75 . This cell line was isola ted from rat insulinoma induced by X - ray irradiation by Asfari and colleagues in 1992 76 . Rat - cells, however, contain two types of proinsulins, which will subsequently produce two C - peptides (the 8 position amino acid is either proline or alanine) in equimolar amounts 75 . Though these C - peptides are different than human C - peptide, the acidic amino acids with carboxyl groups, which have proven to be vital for C - biological effects 44 , are at the same position in both forms of the secreted peptide. 98 A multi - Organs - on - a - Chip model, based o n a 3D - printed fluidic device is reported. Specifically, INS - 1 cells will be integrated as a mimic of endocrine - cells that will secrete molecules such as C - peptide and Zn 2+ into an ERY circulation where cell - cell interactions will be explored. An endothelium mimic will also be included to study endothelial cell - derived molecules that may be determinants of blood f low and vessel dilation in vivo . 3.2 Methods 3.2.1 Culture of INS - 1 Cells in Membrane Inserts Trans - well membrane inserts (Corning Inc., Horseheads, NY) were prepared for cell culture by coating the membrane with 30 L of a fibronectin solution (50 g/mL ), and subsequent drying in a cell hood and sterilization by exposure to UV light. R at INS - 1 cells were cultured in RPMI - 1640 medium (Life Technologies, Carlsbad, CA) supplemented with 1 mM sodium pyruvate, 100 U/mL penicillin, 100 g/mL streptomycin, 55 M mercaptoethanol, 10% fetal bovine serum, 2 mM L - glutamine and 10 mM HEPES. When the cells were confluent, as verified by optical microscopy, they were detached from the flask with trypsin/EDTA and centrifuged at 1000 g for 4 min. The pellet was resuspend RPMI - 1640 medium . Cell density was measured by a hemacytometer (Reichert, Buffalo, NY) and further diluted to a cell density of 10 5 / membrane insert (6.5 mm diameter with 5% CO 2 at 37 °C . After two hours, the medium in the inserts was removed and 99 before use on the 3D - printed fluidic device , which takes about 24 hours . 3.2.2 Culture of Endothelial Cells in Membrane Inserts Endothelial cell culture was accomplished using bovine pulmonary arterial endothelial cells (bPAECs) that were cultured in T - 75 flasks with DMEM media containing 5.5 mM glucose, 10% (v/v) fetal bovine serum, and penicillin/streptomyocin. When the cells were confluent, as verified with optical microscopy, they were detached from the flask with trypsin/EDTA and centrifuged at 1500 g for 5 min. The pellet was then resuspende density was determined by a hemacytometer, and the cell suspension was then diluted to a density of 0.05 million/ added into a coated membrane insert and placed in an in cubator with 5% CO 2 at 37 °C . were added. The cells were grown to confluence before use on the 3D - printed fluidic device. 3.2.3 Integration of Cells onto the Fluidic Device Prior to use, the medium for INS - containing 12 mM glucose and 0.1 % albumin were added to stimulate secretion of these cells for 1 hour. 100 Before integrating endothelial cell inserts into the fluidic device, the c ell culture PPADS (pyridoxalphosphate - 6 - azophenyl - - disulfonic acid) was used fo r the experiments involving inhibition of P2y receptors on endothelial cells. A PPADS L of the PPADS working solution was added to the insert. After a 30 min incubation at 37 °C , the cells were then rinsed with HBSS to remove extra PPADS solution, after which, 200 Membrane inserts with INS - 1 cells and bPAECs were then placed in the first and third wells above a channel, respectively. A clean insert with no cells was plugged in the middle well for ATP collection via diffusion. ERY samples circulated through the channels at a flow rate of 200 L/min (Figure 3.2) . ATP and NO standards were added to the static wells in columns 8 and 10, for on - chip calibration. The entire setup was then placed in a 37 o C incubator with the inserts covered by moist napkins to prevent evaporation of the solutions inside. An experimen t time of 2 hours was employed to observe the physiological interactions between the cells. 101 Figure 3.2. The setup of the Organs - on - a - Chip device. This study used four channels , above which, inserts cultured with INS - 1 cells ( - cells) and bPAECs (endothelial cells) were integrated. The static wells in columns 8 and 10 contained ATP and NO standards for on - chip simultaneous calibration. 102 3.2.4 INS - 1 Cell Secretion Profile Measurement The secretion profiles of INS - 1 cells were measured by an ELISA kit for rat C - peptides (Mercodia, Uppsala, Sweden 0.1% albumin was used to stimulate the INS - 1 cells. Because albumin - free buffer will be used in one experimental trial, albumin - free st with 12 mM glucose only) were also prepared. After adding 200 L of stimulating solution into each INS - 1 cell cultured insert, an aliquot of 5 L of the solution was sampled at 0, 10, 30, 60, 90, 120, and 180 min, which was subsequently diluted to 50 L with DI water. Following the instructions of the ELISA kit, the concentration of C - peptide in the samples at different times were quantified. INS - 1 cell inserts were integrated onto the device, followed by the circulation of PSS through the channels to measure the amount of C - peptide that diffused into the ERY circula tion. At 0, 10, 30, 60, 90, 120, and 180 min, the tubing was opened and 5 L of the flowing solution were collected, diluted in DI water (to 50 L) and quantified by the ELISA kit. For the cells that were stimulated by albumin - free solution, albumin - free PSS was used to measure diffused C - peptides. 3.2.5 ATP Quantification For ATP measurements, 50 L of PSS were loaded in the middle inserts above each channel, 20 min prior to measurement to collect ATP. Next, 10 L of luciferase/luciferin assay mixture were added into the ATP collection inserts 103 simultaneously by a multi - channel pipet, followed by chemiluminescence detection in a plate reader. A calibration curve was obtained before integrating the cells on the device, by circulating ATP standards of 0, 100, 200, 300, and 400 nM in five channels, and taking the measurements the sa me way as described above. 3.2.6 NO Quantification NO was determined by using the fluorescent probe DAF - FM ( 4 - Amino - 5 - methylamino - - difluorofluorescein ), which shows enhanced fluorescence after reacting with NO, with an excitation wavelength of 485 nm and emission wavelength of 515 nm. In this study, 1 mg of DAF - FM was dissolved in DMSO to make a 5 mM soluti Organs - on - a - Chip setup was placed in the 37 o - FM were added into each endothelial cell insert, to react with released NO for the whole experiment period. A bla ck box was applied to cover the device, to prevent photobleaching of the fluorescent probe. The NO donor Spermine NONOate (Cayman Chemicals, Ann Arbor, MI) has a half - life of 37 min at 37 o C and was used to prepare NO standard solutions . Spermine NONOate was dissolved in 0.01 M NaOH as a 1 mM stock solution, which can be stored at - 20 °C for months. This stock solution was diluted in HBSS to four standard solutions were the n kept in the dark and placed in a 37 °C incubator for 37 min. 104 NOate (with DAF - FM) were added into four static wells in column 10 of the fluidic device, which was then placed in a plate reader. Fluoresc ence intensity (ex. 485, em. 515) above the endothelial cell inserts and the four static inserts were detected simultaneously. A calibration curve was obtained by plotting the fluorescence intensity of the four static wells versus the concentration of NO s tandards, by which, the amount of NO released from endothelial cells was quantitatively determined . 3.3 Results 3.3.1 Cell culture in Membrane Inserts The INS - 1 cells and endothelial cells cultured in membrane inserts were examined under a microscope. Bo th cells can grow to a high degree of confluence on the membrane inserts. Figure 3.3 shows that the morphology of the INS - 1 cells (A) and bPAECs (B) are normal in the membrane inserts. To count the number of cells that were cultured in membrane inserts, th e medium cell suspension (in trypsin) was removed and cell numbers were determined by hemacytometer. In a membrane insert used in this study (6.5 mm diameter, polyester mem - 1 cells and 0.036±0.0011 million bPAECs can be cultured (results represent average of 5 measurements ± S.E.M.). The reason that there are more that INS - 1 cells than bPAECs on the same area 105 in the membrane insert is INS - 1 cells can form pseudo islets, or multi - layer aggregates (Figure 3.3A) while bPAECs can only grow along the surface. Figure 3.3. (A), microscope image s cells cultured in a membrane insert. (A) INS - 1 cells; (B) bPAECs . A 20 µ m B 20 µ m 106 3 .3.2 INS - 1 cell Secretion Profile Figure 3.4 A (with albumin in the stimulating solution) and Figure 3.4 B (without albumin in the stimulating solution) show the secretion pattern of C - peptide after INS - 1 cells were stimulated. The results suggest that album in does not affect the stimulation of C - peptide secretion from INS - 1 cells. After 60 min of stimulation, the amount of C - peptide measured reached a maximum at 26.04 ± 1.93 nM (n=4 inserts, error=S.E.M.) . After these INS - 1 cells were stimulated for 60 min, th e inserts were then plugged into the fluidic device, above different channels. For the inserts that had albumin containing stimulation solution, albumin containing PSS was circulated underneath. For INS - 1 cells stimulated with albumin free solution, albumi n free PSS was pumped through the corresponding channels. C - peptide in channels at different time points was quantified by the ELISA kit. Figure 3.4 (C, D) show the diffusion profiles of C - peptide from stimulated INS - 1 cells to the flowing stream under the membrane. The diffusion profiles did not show a significant difference between albumin containing (C) and albumin free (D) trials. The concentration range of C - peptide in the flow was at sub - to single digit nanomolar levels, which is consistent with repor ted physiological concentration range. 107 Figure 3.4. (A), INS - 1 cell secretion profile in albumin - containing stimulation buffer. (B), INS - 1 cell secretion profile in albumin - free stimulation solution. (C), the amount of C - peptide diffused to the fo llowing solution underneath flowing on the device, with albumin presented in the system. (D), without albumin in the system, the amount of C - peptide diffused to flowing. A B C D 108 3.3.3 Cell - cell Interaction Studies After the establishment of the multi - Organs - on - a - Chip model, the interactions between INS - 1 cells, ERYs, and endothelial cells was studied. Shown in Figure 3.5(A) is the top - down view of the strategy employed to study the interaction between the three cell types, while Figure 3.5 (B) is a side - view of the system, showing ERYs circulating underneath the various inserts. Once a 7% solution of ERYs was circulating through the system, C - peptide and Zn 2+ secreted from the INS - 1 cells (located in the top insert, represented by the blue ovals) diffuse through the membrane and into the channel. ERYs were pumped through the channels in albumin - containing PSS, except for channel 2, which was albumin - free. After 2 hours of circulating the ERYs through the system, measurements of ATP in the middle inserts (orange c ircles) and NO in the bottom inserts (green circles) were performed by placing the fluidic device in a plate reader. The data in Figure 3.6 clearly show ATP levels are only increased when stimulated INS - 1 cells are present in the ERYs flowing that contains albumin (channel 1). We presume that this increase in ATP release from the ERYs is due to the INS - 1 secretion of C - peptide and Zn 2+ , which are subsequently carried to the ERYs by albumin. Note in Figure 3.6 that the absence of albumin in the stream of ERY s (channel 2) results in no significant increase in ATP release. There is also no increase in ATP release in the absence of INS - 1 cells in the first well (clear oval in channel 3). In channel 4, the system contained INS - 1 cells and albumin while the third well containing endothelial cells that were incubated with PPADs, a P2y purinergic receptor inhibitor. This antagonist was added to block ATP binding to the endothelial cells for inhibition 109 of NO production; therefore, it has no effect on ATP release. Figu re 3.7 shows similar trends for measurement of ERY - derived ATP, although these measurements represent the amount of ATP determined in the channels underlying the membrane - based inserts. Figure 3.8 contains results from measurements of NO from the bottom w ells on each channel. These wells contained endothelial cells which, upon stimulation with ATP, will produce NO. As expected, based on data in Figure 3.6, the most NO production is measured when glucose is used to stimulate INS - 1 cells in the presence of a n albumin - containing buffered stream of ERYs (channel 1). In the absence of albumin (channel 2), there is reduced ATP release that result s in decreased amounts of NO produced. Reduced levels of NO were also measured when there were no INS - 1 cells in the fi rst well (channel 3); again, reduced ATP in channel 3 (both insert and channel) suggests that reduced levels of NO would be measured in this channel, and this is confirmed in Figure 3.8. Finally, the last bar in Figure 3.8 shows that the addition of the AT P receptor blocker, PPADS, blocks ATP released by the ERY from binding to the endothelial cells, thus inhibiting NO production. To exclude the possibility that INS - 1 cells , rather than ERYs, can release large amounts of ATP diffusing to the ERY streams , after INS - 1 inserts were integrated, PSS without ERYs was flowed in the channels, and ATP was measured as described above. It was found that the ATP amount from INS - 1 cells was below the detection limit of the device, which is negligible compared to the ATP amount measured in Figure 3.7. 110 The possibility that INS - 1 cells may directly stimulate endothelial cell derived NO production was also excluded. Figure 3.5. (A), the top view of the four channels with different treatments. Channel 1 contai ns INS - 1 cells, bPAECs cultured in the membrane inserts , and ERYs flowing in PSS; channel 2 contains all three cells, but all buffers are albumin - free; channel 3 does not contain INS - 1 cells, with the rest identical as channel 1; channel 4 is almost identi cal to channel 1, except the endothelial cells were pretreated with PPADS, a P2y receptor inhibitor. B), a side schematic view of channel 1. B A B - bPAECs Eryt - cells bPAECs ERYs 111 Figure 3.6. The amount of ATP collected in the insert above each channel, which is proportional to th e ATP amount in corresponding flowing. Compar ing channel s 1 and 3, INS - 1 cell stimulated ATP release from ERYs, as is shown by increased ATP amount in the collecting well. Without albumin (channel 2), however, INS - 1 cells failed to stimulate ATP release fr om ERYs, which proved the indispensable role of albumin in this process. Channel 4 is identical to channel 1, except the endothelial cells were inhibited, which did not affect the ATP released. 112 Figure 3.7. The amount of ATP in ERYs st reams in the four channels , showing the same trend as in Figure 3.6. 113 Figure 3.8. The released NO from endothelial cells culture in the last insert above each channel. The data from channels 1 to 3 are consistent with the ATP observation, wh ich is with more ATP released, more NO will be produced in endothelial cells. In channel 4, however, though there was a higher ATP release from ERYs, the endothelial cells were pretreated with PPADs, a P2y receptor (ATP receptor in endothelial cells) inhib itor, and thus NO was not increased. 114 3.4 Discussion Organs - on - a - Chip has become a promising tool for in vitro physiological and pharmaceutical studies, as a powerful alternative to existing 2D and 3D cell culture models 2 . Although many Organs - on - a - Chip platforms have been developed by researchers, these models are mainly single organ models that cannot meet the requirements of research that needs to systematically study the interactions between different tissues/cells 5 . Also, most of the currently developed models are not easy to use (with co mplicated fabrication processes), not reproducible (because the microfluidic devices are basically single use), and lab - dependent (with poor inter - lab reproducibility). One important reason for this impasse lies in the fabrication of microfluidic devices, which serve as the basis for an Organs - on - a - Chip model. As discussed in chapter 1, despite the desirable properties of PDMS, the most common material in microfluidic systems, PDMS has poor mechanical properties and the soft lithography process cannot integ rate complicated features in a PDMS device. In this work, however, a 3D - printed fluidic device was applied to build an Organs - on - a - Chip model, which showed great integration and robustness. This model contains three types of cells INS - 1 cells, ERYs, and en dothelial cells, as a mimic of on - chip endocrine tissue, blood flow, and blood vessel barrier, respectively. When an ERY stream flows through the endocrine mimic, the secreted molecules enter the blood stream mimic, where downstream physiological responses will be elicited and observed. From an engineering perspective, this 3D - printed multi - Organs - on - a - Chip model avoids several problems that can occur when using PDMS based platforms. First of all, 115 the INS - 1 cells and endothelial cells were cultured on comme rcial membrane inserts that can be integrated on the fluidic device, instead of being cultured directly on the device. This process allows for cell culture examination and optimization before taking a measurement. For example, if the morphology of cells in an insert tends to be abnormal, the insert can be discarded and a new one will be examined and implemented. If the cells are cultured directly on the device, it will be hard to examine the cells if the device has complicated features. Even though the cell s can be examined, once a researcher finds abnormality of the cells, the whole device will likely have to be discarded. Also, this philosophy of cell culture and integration can be easily adopted for investigating other cells. Any cell that can be cultured in a membrane insert can be integrated on the fluidic device for cell - cell interaction studies. Another important feature of the Organs - on - a - Chip platform is the realization of simultaneous on - chip detection. As discussed in chapter 2, the fluidic device was modeled following the dimensions of a standard 96 - well plate, with all the wells aligned with the detectors in a plate reader. After running an experiment, the whole device can be directly placed in a plate reader to detect the optical signals above m ultiple wells almost simultaneously. More importantly, static wells (without a channel beneath) were also included in the device, for on - chip calibration purpose. In this study, ATP and NO standards were added in two columns of static wel ls to calibrate th e amount of A T P and NO in dynamic wells (Figure 3.2, column 8 and 10 ), which not only made the measurements more efficient, but also enhanced the reliability of the results. Some other advantageous features of 3D - printed fluidic 116 devices, such as reusabili ty and ruggedness also facilitate the buildup of the Organs - on - a - Chip platform. From a biological perspective, this Organs - on - a - Chip platform mimics the physiological process of - cell secretion entering a blood stream. Data in Figure 3.4 shows that durin g the experimental period (2 hours), single digit nanomolar C - peptide was secreted into the blood stream mimic, which represents the physiological concentration in a blood stream. Moreover, the secretion from - cells significantly increases ATP release fro m ERYs, which is consistent with the discovery in chapter 2 that C - peptide/Zn 2+ can stimulate ERYs to release ATP. Though the - cells also secrete other molecules such as insulin, it can still be initially concluded it is the C - peptide and Zn 2+ that increa s e ATP release from ERYs , because a previous study reported that insulin does not stimulate ATP release from ERYs 77 . However, to consolidate this theory, more re search regarding on - chip separation needs to be conducted, as will be discussed in Chapter 5. The role of albumin was also examined on the Organs - on - a - Chip platform. As shown in Figure 3.6 and 3.7, without albumin in the system, the - cells failed to stimulate ATP release from ERYs, which confirmed the discovery in a more physiologically relevant way, that albumin delivers C - peptide and Zn 2+ to ERYs, thus playing a role in ATP release. To verify that the increased ATP was from ERYs, i nstead of from the INS - 1 cells (which may contain ATP in cytoplasm), PSS without ERYs was 117 flowing under INS - 1 cell inserts and the ATP amount in the flowing was quantified. The amount of ATP released under such condition s is not detectable. Endothelial ce lls were integrated on the platform to mimic a blood vessel barrier. The endothelial production of NO, which is a well - recognized vessel dilator, was - cells and endothelial cells was - cells cannot stimulate NO production from the endothelial cells, which indicated there is not a direct - cells and endothelial cells, in terms of NO production. However, with ERYs in the flowing channels, a higher ATP release can always elicit a high er NO production. As a control, different concentrations of ATP standards were circulate d under endothelial cell inserts, showing that higher ATP levels in the flow lead to more NO production from endothelial cells. These results reveal that secreted molec - cells can increase ATP release form ERYs, which in turn, exerts a downstream effect on endothelial cells. To prove that NO production occurs after ATP binding to endothelial cells, PPADS, an inhibitor of P2y receptor on endotheli al cells was used to treat endothelial cells on channel 4 (Figure 3.5 A). Even though the ATP level in this channel is high (statistically the same as channel 1), NO production did not increase (Figure 3. 8). This further proves that ATP binds onto endothel ial cells via the P2y receptors, which then elicits NO synthesis in the cells. 118 The interaction between ERYs and endothelial cells also enlightens some possibilities for drug discovery and development. For example, for drug development to enhance endotheli al cell NO production, ERYs can serve as a n alternative target to endothelial cells. By increasing ERY - released ATP with a certain drug, endothelium derived NO can also be indirectly enhanced. Keenan has reported that diabetic patients who still retain det ectable C - peptide in their blood stream exhibited a lower level of diabetic complications 78 . The results presente d in this chapter can potentially explain this cli nical observation in a scenario that C - - cells enters the blood stream and interact s with ERYs in the presence of albumin and Zn 2+ , subsequently increasing ERY - derived ATP that in turn, exerts a downstream effect on endothe lial cells by increasing endothelium - derived NO, which promotes vasodilation and thus ameliorates diabetic complications resulting from restricted blood flow. This conclusion can also be supported by the observations of Wahren, Sjoberg, Johansson and other researchers that C - peptide reduces diabetes complications. All the evidence indicates that C - peptide can potentially be a missing link in diabetes treatment 79 - 84 . The results from Chapter 2 and this chapter proved the indispensable role of albumin and Zn 2+ in the process that C - peptide affects ERYs, which represents a step forward of the development of C - peptide therapy, but also evokes further questions including the binding structure of the three molecules, the binding site of C - peptide on ERYs, and the pathway by which ATP release is stimulated. Such basic research work needs to be performed before turning C - peptide in to a therapeutics for diabetes. 119 3.5 Conclusio n In this chapter, a multi - Organs - on - a - Chip model was established on a 3D - pr inted fluidic device, to study claimed - cells, ERYs, and endothelial cells, in a physiological manner. This is the first Organs - on - a - Chip model that contains three interactive cell types. The philosophy of culturing cells in membrane inserts that will be plugged in the 3D - printed fluidic device guarantees the quality (morphology, confluence etc.) of the cells to be studied. It also provides a universa l way to study other cells/tissue as long as the cells can be cultured in an insert. The 3D - printed fluidic device is amenable to a plate reader for simultaneous and efficient optical measurements. Also the static wells on the device enable simultaneous on - chip calibration for ATP and NO, which can enhance the reliability of - cells mimic the physiological secretion process well - cells do not affect the endothelial cells , but increase ATP release from ERYs, which in turn, plays a downstream role on endothelial cells, stimulating NO production. Also, albumin is indispensable in these processes, without which, none of the above response could be observed. Although more on - c hip separation work needs to be performed to eventually conclude which - cell secretion exert the observed effects, it can be initially concluded that C - peptide and Zn 2+ are required. These results further confirmed the efficacy of C - p eptide/Zn 2+ on vessel dilation, and the indispensable role of albumin, in a more physiologically relevant way. The philosophy applied in establishing the multi - 120 Organs - on - a - Chip platform can also be used for other studies involving cell interactions and dru g discovery. 121 R EFERENCES 122 REFERENCES 1. Brandenburg, D. History and Diagnostic Significance of C - Peptide. Exp Diabetes Res (2008). 2. Huh, D., Hamilton, G.A. & Ingber, D.E. From 3D cell culture to organs - on - chips. Trends Cell Biol 21 , 745 - 754 (2011). 3. van der Meer, A.D. & van den Berg, A. Organs - on - chips: breaking the in vitro impasse. Integr Biol - Uk 4 , 461 - 470 (2012). 4. Beebe, D.J., Ingber, D.E. & den Toonder, J. Organs on Chips 2013. Lab Chip 13 , 3447 - 3448 (2013). 5. Luni, C., Serena, E. & Elvassore, N. Human - on - chip for therapy development and fundamental science. Curr Opin Biotech 25 , 45 - 50 (2014). 6. Ghaemmaghami, A.M., Hancock, M.J., Harrington, H., Kaji, H. & Khademhosseini, A. Biomimetic tissues on a chip for drug discovery. Drug Discov Today 17 , 173 - 181 (2012). 7. Sung, J.H. et al. Microfabricated mammalian organ systems and their integration into models of whole animals and humans. Lab Chip 13 , 1201 - 1212 (2013). 8. Williamson, A., Singh, S., Fernekorn, U. & Schober, A. The future of the patient - specific Body - on - a - chip. Lab Chip 13 , 3471 - 3480 (2013). 9. Yi, L. et al. Integrated perfusion and separation systems for entrainment of insulin secretion from islets of Langerhans. Lab Chip 15 , 823 - 832 (2015). 10. Mohammed, J.S., Wang, Y., Harvat, T.A., Oberholzer, J. & Eddington, D.T. Microfluidic device for multimodal characterization of pancreatic islets. Lab Chip 9 , 97 - 106 (2009). 11. Cheah, L.T. et al. Microfluidic perfusion system for maintain ing viable heart tissue with real - time electrochemical monitoring of reactive oxygen species. Lab Chip 10 , 2720 - 2726 (2010). 12. Huang, Y., Williams, J.C. & Johnson, S.M. Brain slice on a chip: opportunities and challenges of applying microfluidic technolo gy to intact tissues. Lab Chip 12 , 2103 - 2117 (2012). 123 13. Mauleon, G. , Fall, C.P. & Eddington, D.T. Precise spatial and temporal control of o ygen within in v itro b rain s lices via m icrofluidic g as c hannels. Plos One 7 (2012). 14. Scott, A. et al. A microflui dic microelectrode array for simultaneous electrophysiology, chemical stimulation, and imaging of brain slices. Lab Chip 13 , 527 - 535 (2013). 15. Wang, Y. et al. Application of microfluidic technology to pancreatic islet research: first decade of endeavor. Bioanalysis 2 , 1729 - 1744 (2010). 16. Dishinger, J.F., Reid, K.R. & Kennedy, R.T. Quantitative Monitoring of Insulin Secretion from Single Islets of Langerhans in Parallel on a Microfluidic Chip. Anal Chem 81 , 3119 - 3127 (2009). 17. Jang, K.J. et al. Human k idney proximal tubule - on - a - chip for drug transport and nephrotoxicity assessment. Integr Biol - Uk 5 , 1119 - 1129 (2013). 18. Achyuta, A.K.H. et al. A modular approach to create a neurovascular unit - on - a - chip. Lab Chip 13 , 542 - 553 (2013). 19. van der Meer, A.D ., Orlova, V.V., ten Dijke, P., van den Berg, A. & Mummery, C.L. Three - dimensional co - cultures of human endothelial cells and embryonic stem cell - derived pericytes inside a microfluidic device. Lab Chip 13 , 3562 - 3568 (2013). 20. Giobbe, G.G. et al. Human p luripotent stem cell derived hepatocyte - like cells integrated in microfluidic platform for drug screening applications. Hepatology 60 , 1187a - 1188a (2014). 21. Lee, S.A. et al. Spheroid - based three - dimensional liver - on - a - chip to investigate hepatocyte - hepat ic stellate cell interactions and flow effects. Lab Chip 13 , 3529 - 3537 (2013). 22. Wong, K.H.K., Chan, J.M., Kamm, R.D. & Tien, J. Microfluidic Models of Vascular Functions. Annu Rev Biomed Eng 14 , 205 - 230 (2012). 23. Truskey, G.A. Endothelial vascular smo oth muscle cell coculture assay for high throughput screening assays to identify antiangiogenic and othe therapeutic molecules. International Journal of High Throughput Screening 99 , 171 - 181 (2010). 124 24. Tibbitt, M.W. & Anseth, K.S. Hydrogels as Extracellul ar Matrix Mimics for 3D Cell Culture. Biotechnol Bioeng 103 , 655 - 663 (2009). 25. Figallo, E. et al. Micro - bioreactor array for controlling cellular microenvironments. Lab Chip 7 , 710 - 719 (2007). 26. Lee, J.M., Kim, J.E., Kang, E., Lee, S.H. & Chung, B.G. A n integrated microfluidic culture device to regulate endothelial cell differentiation from embryonic stem cells. Electrophoresis 32 , 3133 - 3137 (2011). 27. Fung, W.T., Beyzavi, A., Abgrall, P., Nguyen, N.T. & Li, H.Y. Microfluidic platform for controlling t he differentiation of embryoid bodies. Lab Chip 9 , 2591 - 2595 (2009). 28. Kawada, J., Kimura, H., Akutsu, H., Sakai, Y. & Fujii, T. Spatiotemporally controlled delivery of soluble factors for stem cell differentiation. Lab Chip 12 , 4508 - 4515 (2012). 29. El - Ali, J., Sorger, P.K. & Jensen, K.F. Cells on chips. Nature 442 , 403 - 411 (2006). 30. Wan, C.R., Chung, S. & Kamm, R.D. Differentiation of Embryonic Stem Cells into Cardiomyocytes in a Compliant Microfluidic System. Ann Biomed Eng 39 , 1840 - 1847 (2011). 31. Jang, K.J. & Suh, K.Y. A multi - layer microfluidic device for efficient culture and analysis of renal tubular cells. Lab Chip 10 , 36 - 42 (2010). 32. Zheng, W.F. et al. A microfluidic flow - stretch chip for investigating blood vessel biomechanics. Lab Chip 12 , 3441 - 3450 (2012). 33. Huh, D. et al. Reconstituting Organ - Level Lung Functions on a Chip. Science 328 , 1662 - 1668 (2010). 34. Weibel, E.R. Morphological Basis of Alveolar - Capillary Gas - Exchange. Physiol Rev 53 , 419 - 495 (1973). 35. Sankar, K.S. et al. Cultu ring Pancreatic Islets in Microfluidic Flow Enhances Morphology of the Associated Endothelial Cells. Plos One 6 (2011). 36. Goral, V.N. & Yuen, P.K. Microfluidic Platforms for Hepatocyte Cell Culture: New Technologies and Applications. Ann Biomed Eng 40 , 1 244 - 1254 (2012). 125 37. Chung, B.G. et al. Human neural stem cell growth and differentiation in a gradient - generating microfluidic device. Lab Chip 5 , 401 - 406 (2005). 38. Torisawa, Y.S. et al. Microfluidic platform for chemotaxis in gradients formed by CXCL1 2 source - sink cells. Integr Biol - Uk 2 , 680 - 686 (2010). 39. Kim, J.H., Kang, T.J. & Yu, W.R. Mechanical modeling of self - expandable stent fabricated using braiding technology. J Biomech 41 , 3202 - 3212 (2008). 40. Alford, P.W., Feinberg, A.W., Sheehy, S.P. & Parker, K.K. Biohybrid thin films for measuring contractility in engineered cardiovascular muscle. Biomaterials 31 , 3613 - 3621 (2010). 41. Gunther, A. et al. A microfluidic platform for probing small artery structure and function. Lab Chip 10 , 2341 - 2349 (20 10). 42. Franco, C. & Gerhardt, H. TISSUE ENGINEERING Blood vessels on a chip. Nature 488 , 465 - 466 (2012). 43. Kim, Y. et al. Probing nanoparticle translocation across the permeable endothelium in experimental atherosclerosis. P Natl Acad Sci USA 111 , 1078 - 1083 (2014). 44. Liu, Y.L., Chen, C.P., Summers, S., Medawala, W. & Spence, D.M. C - peptide and zinc delivery to erythrocytes requires the presence of albumin: implications in diabetes explored with a 3D - printed fluidic device. Integr Biol - Uk 7 , 534 - 543 (2 015). 45. Park, J., Koito, H., Li, J.R. & Han, A. Microfluidic compartmentalized co - culture platform for CNS axon myelination research. Biomed Microdevices 11 , 1145 - 1153 (2009). 46. Chung, S. et al. Cell migration into scaffolds under co - culture conditions in a microfluidic platform. Lab Chip 9 , 269 - 275 (2009). 47. Das, M., Rumsey, J.W., Bhargava, N., Stancescu, M. & Hickman, J.J. A defined long - term in vitro tissue engineered model of neuromuscular junctions. Biomaterials 31 , 4880 - 4888 (2010). 48. Hawkins, B.T. & Davis, T.P. The blood - brain barrier/neurovascular unit in health and disease. Pharmacol Rev 57 , 173 - 185 (2005). 49. You, L.D. et al. Osteocytes as mechanosensors in the inhibition of bone resorption due to mechanical loading. Bone 42 , 172 - 179 (2008 ). 126 50. Nagamine, K. et al. Spatiotemporally controlled contraction of micropatterned skeletal muscle cells on a hydrogel sheet. Lab Chip 11 , 513 - 517 (2011). 51. Nedachi, T., Fujita, H. & Kanzaki, M. Contractile C2C12 myotube model for studying exercise - ind ucible responses in skeletal muscle. Am J Physiol - Endoc M 295 , E1191 - E1204 (2008). 52. Kimura, H., Yamamoto, T., Sakai, H., Sakai, Y. & Fujii, T. An integrated microfluidic system for long - term perfusion culture and on - line monitoring of intestinal tissue models. Lab Chip 8 , 741 - 746 (2008). 53. Mahler, G.J., Esch, M.B., Glahn, R.P. & Shuler, M.L. Characterization of a Gastrointestinal Tract Microscale Cell Culture Analog Used to Predict Drug Toxicity. Biotechnol Bioeng 104 , 193 - 205 (2009). 54. Imura, Y., Sa to, K. & Yoshimura, E. Micro Total Bioassay System for Ingested Substances: Assessment of Intestinal Absorption, Hepatic Metabolism, and Bioactivity. Anal Chem 82 , 9983 - 9988 (2010). 55. Domansky, K. et al. Perfused multiwell plate for 3D liver tissue engin eering. Lab Chip 10 , 51 - 58 (2010). 56. Lee, P.J., Hung, P.J. & Lee, L.P. An artificial liver sinusoid with a microfluidic endothelial - like barrier for primary hepatocyte culture. Biotechnol Bioeng 97 , 1340 - 1346 (2007). 57. Carraro, A. et al. In vitro analy sis of a hepatic device with intrinsic microvascular - based channels. Biomed Microdevices 10 , 795 - 805 (2008). 58. Kane, B.J., Zinner, M.J., Yarmush, M.L. & Toner, M. Liver - specific functional studies in a microfluidic array of primary mammalian hepatocytes. Anal Chem 78 , 4291 - 4298 (2006). 59. Nakao, Y., Kimura, H., Sakai, Y. & Fujii, T. Bile canaliculi formation by aligning rat primary hepatocytes in a microfluidic device. Biomicrofluidics 5 (2011). 60. Sudo, R. et al. Transport - mediated angiogenesis in 3D e pithelial coculture. Faseb J 23 , 2155 - 2164 (2009). 61. van Midwoud, P.M., Verpoorte, E. & Groothuis, G.M.M. Microfluidic devices for in vitro studies on liver drug metabolism and toxicity. Integr Biol - Uk 3 , 509 - 521 (2011). 127 62. van Midwoud, P.M., Merema, M. T., Verpoorte, E. & Groothuis, G.M.M. A microfluidic approach for in vitro assessment of interorgan interactions in drug metabolism using intestinal and liver slices. Lab Chip 10 , 2778 - 2786 (2010). 63. Khetani, S.R. & Bhatia, S.N. Microscale culture of hum an liver cells for drug development. Nat Biotechnol 26 , 120 - 126 (2008). 64. Baker, M. A Living System on a Chip. Nature 471 , 661 - 665 (2011). 65. Wong, K.H.K., Truslow, J.G., Khankhel, A.H., Chan, K.L.S. & Tien, J. Artificial lymphatic drainage systems for vascularized microfluidic scaffolds. J Biomed Mater Res A 101 , 2181 - 2190 (2013). 66. Sung, K.E. et al. Transition to invasion in breast cancer: a microfluidic in vitro model enables examination of spatial and temporal effects. Integr Biol - Uk 3 , 439 - 450 (20 11). 67. Song, J.W. et al. Microfluidic Endothelium for Studying the Intravascular Adhesion of Metastatic Breast Cancer Cells. Plos One 4 (2009). 68. Elliott, N.T. & Yuan, F. A Review of Three - Dimensional In Vitro Tissue Models for Drug Discovery and Trans port Studies. J Pharm Sci - Us 100 , 59 - 74 (2011). 69. Wlodkowic, D. & Cooper, J.M. Tumors on chips: oncology meets microfluidics. Curr Opin Chem Biol 14 , 556 - 567 (2010). 70. Wang, S.Y. et al. Application of microfluidic gradient chip in the analysis of lung cancer chemotherapy resistance. J Pharmaceut Biomed 49 , 806 - 810 (2009). 71. Jedrych, E., Pawlicka, Z., Chudy, M., Dybko, A. & Brzozka, Z. Evaluation of photodynamic therapy (PDT) procedures using microfluidic system. Anal Chim Acta 683 , 149 - 155 (2011). 72. Huh, D., Torisawa, Y.S., Hamilton, G.A., Kim, H.J. & Ingber, D.E. Microengineered physiological biomimicry: Organs - on - Chips. Lab Chip 12 , 2156 - 2164 (2012). 73. Vogel, P.A., Halpin, S.T., Martin, R.S. & Spence, D.M. Microfluidic Transendothelial Electrical Resistance Measurement Device that Enables Blood Flow and Postgrowth Experiments. Anal Chem 83 , 4296 - 4301 (2011). 128 74. Skelin, M., Rupnik, M. & Cencic, A. Pancreatic Beta Cell Lines and their Applications in Diabetes Mellitus Research. Altex - Altern Anim E x 27 , 105 - 113 (2010). 75. Merglen, A. et al. Glucose sensitivity and metabolism - secretion coupling studied during two - year continuous culture in INS - 1E insulinoma cells. Endocrinology 145 , 667 - 678 (2004). 76. Asfari, M. et al. Establishment of 2 - Mercaptoet hanol - Dependent Differentiated Insulin - Secreting Cell - Lines. Endocrinology 130 , 167 - 178 (1992). 77. Giebink, A.W., Vogel, P.A., Medawala, W. & Spence, D.M. C - peptide - stimulated nitric oxide production in a cultured pulmonary artery endothelium is erythrocy te mediated and requires Zn2+. Diabetes - Metab Res 29 , 44 - 52 (2013). 78. Keenan, H.A. et al. Residual Insulin Production and Pancreatic beta - Cell Turnover After 50 Years of Diabetes: Joslin Medalist Study. Diabetes 59 , 2846 - 2853 (2010). 79. Wahren, J. & Lar sson, C. C - peptide: New findings and therapeutic possibilities. Diabetes Res Clin Pr 107 , 309 - 319 (2015). 80. Wahren, J., Johansson, B.L. & Wallberghenriksson, H. Does C - Peptide Have a Physiological - Role. Diabetologia 37 , S99 - S107 (1994). 81. Sjoberg, S. e t al. Residual Insulin Production, Glycemic Control and Prevalence of Microvascular Lesions and Polyneuropathy in Long - Term Type - 1 (Insulin - Dependent) Diabetes - Mellitus. Diabetologia 30 , 208 - 213 (1987). 82. Johansson, B.L., Sjoberg, S. & Wahren, J. The Inf luence of Human C - Peptide on Renal - Function and Glucose - Utilization in Type - 1 (Insulin - Dependent) Diabetic - Patients. Diabetologia 35 , 121 - 128 (1992). 83. Johansson, B.L., Kernell, A., Sjoberg, S. & Wahren, J. Influence of Combined C - Peptide and Insulin Adm inistration on Renal - Function and Metabolic Control in Diabetes Type - 1. J Clin Endocr Metab 77 , 976 - 981 (1993). 84. Johansson, B.L. et al. Beneficial effects of C - peptide on incipient nephropathy and neuropathy in patients with Type 1 diabetes mellitus. Di abetic Med 17 , 181 - 189 (2000). 129 Chapter 4 - Evaluating Stored Erythrocytes Used in Tra nsfusion Medicine u sing 3D - Printed Analytical Devices 4.1 Background 4.1.1 Blood Banking Blood component transfusion has become a critical part of modern healthcare. In 2011, approximately 15.7 million units of whole blood were collected and 87% of them were transfused in the United States alone 1 , 2 . The history of blood banking dates back to 1915, when Rous and Turner invented a mixed solution containing citrate and glucose for blood storage 3 . This idea was further developed by Ro bertson who applied citrate as an anticoagulant and, using glucose as an erythrocyte (ERY) preservative, subsequently became the basis for modern blood banking protocols 4 . Currently, the FDA approved blood banking protocol consists of two steps: collection of bloo d in a solutions known as CPD (Citrate - Phosphate - Dextrose), and 5 - 7 . As shown in Table 4.1, there are two types of CPD and three variants of AS solutions that are widely used for ERY storage 7 - 10 . CPD is usually used with AS - 1, while CP2D, which contains double the amount of d extrose (glucose) as CPD is mainly combined with AS - 3 or AS - 5. A typical ERY collection procedure involves 450 mL of whole blood drawn into a primary bag made of polyvinyl chloride (PVC) that contains 63 mL of the anticoagulant solution, CPD. After centri fugation, ERYs (~ 220 mL, containing ~ 40 mL of CPD/plasma) are sedimented and separated from the buffy coat (which contains platelets and 130 leukocytes) and plasma. With a semiautomated extractor, plasma is then exported out of the top port, and the ERYs are exported out of the bottom port into a satellite bag containing 100 mL of AS - 1. This procedure results in a final, stored ERY solution at a 55% - 60% hematocrit 11 , 12 . Based on the ingredient information from Table 4.1, the final concentration of dextrose (glucose) in a storage bag will be ~ 40 mM, which is much higher than the physiological level of 5.5 mM 13 . 4.1.2 Transfusion Related Complications Although blood transfusion has proven successful in the care of critically ill patients to replace lost hemoglobin and maintain nor mal oxygen delivery, transfusion related complications still exist 14 . Survival of patients should be the first priority in all medical treatments. Unfortunately, studies have also demonstrated that ERY transfusion is associated with increased mortality, in comparison to those patients who did not receive a transfusion 15 - 18 . An increased risk of complications such as pneumonia 19 , 20 , infections 21 , and multi - organ failure 18 , 22 ( e.g. , renal failure) has also been reported. Interesting ly, some of these reports have indicated that the aging of stored ERYs may be significantly associated with the clinical consequences. For example, Koch reported that patients who received ERYs stored more than two weeks had higher rates of in - hospital mor tality, complications, and even a higher mortality rate one year after a transfusion 23 . Zallen and colleagues found that blood units stored longer than 14 days incre ased the risk of multiple organ failure and serious infections 22 . The adverse clinical consequences that occur after a transfusion potentially result from physiological alterations of ERYs, known as storage lesions, during storage in an 131 environment that is dramatically different from in vivo conditions 24 , 25 . The hallmark feature of storage conditions is hyperglycemia, which may adversely affect the physiology of stored ERYs, especially when considering the fact that in diabetic patients, 7 - 9 mM blood glucose already results in noticeable ERY lesions, including impaired ATP release 26 . The extremely high glucose level in ERY stor age solutions may be a key factor related to the ERY storage lesions. Table 4.1. Constituents of CPD and AS Used for Blood Banking C onstituents (mM) CPD AS - 1 CP2D AS - 3 AS - 5 Sodium citrate 101.9 - 101.9 23 - M onobasic sodium phosphate 16.1 - 16.1 23 - C itric acid 15.6 - 15.6 2 - D extrose 129 111 258 55 45 S odium chloride - 154 - 70 150 Adenine - 2 - 2 2.2 M annitol - 41 - 45.5 - In addition to oxygen delivery, ERYs also participate in regulating blood flow by releasing ATP into the blood stream. AT P binds to endothelial cells on the inner side of vessel walls, leading to endothelium NO synthesis. As discussed previously, NO 132 participat es in the relaxation of smooth muscle cells surrounding vessel walls, enabling vessel dilation and increased blood fl ow 27 . Interestingly, some reports show that people who received a transfusion suffer from insufficient nitric oxide bioavailability (INOBA) 28 , 29 . In relation to decreased NO availability, Tsai and colleagues reported a 63% decrease in microvascular flow after transfusion wi th stored ERYs 30 . Inspired by these results, the Spence group has studied the effects of hyperglycemia on stored ERYs, discovering that high glucose level lead s to diminished ERY release of ATP, and thus decreases downstream endothelium derived NO. Furthermore, an alternative storage solution AS - 1N (normoglycemic version of AS - 1) was proposed to be beneficial to maintain a relatively normal ATP level from stored ERYs 31 . 4.1.3 Chapter Motivation and Overview Previous research has shown that hyperglycemic storage conditions can diminish ATP release from stored ERYs 31 , which may lead to insufficient NO bioavailability after a transfusion, and eventually cause severe complications. However, the performance of stored ERYs and their overall condition after being transfused into the normoglycemic blood stream of a patient remained unknown. For example, can the hyperglycemia stored ERYs reverse their function of releasing normal amount o f ATP? Therefore, the reversibility of ATP release from stored ERYs will be first studied in this chapter. Also, the mechanism by which hyperglycemia decreases ATP release will be explored in this work. 133 As discussed in chapter 3, healthy ERYs respond to p ancreatic - cells by releasing ATP that can stimulate NO production from endothelial cells. The results from Chapter 3 provided motivation to examine the responsiveness of stored ERYs to - cell secretion, especially C - peptide. Here, a 3D - printed intravenous (i.v.) i njection device will be described that mimics an in vitro transfusion process, with - cells integrated to examine how stored ERYs respond to a mimic of normal endocrine function. 4.2 Methods 4.2.1 ERY Collection and Storage The collection process consist s of preparing 6 non - siliconized and untreated (i.e., no heparin or other anticoagulant) 10 mL glass Vacutainer tubes (BD, Franklin Lakes NJ); 1 mL of CPD was injected into 3 of the tubes using a syringe, while the other 3 contained 1 mL of CPD - N. Next, ap proximately 7 mL of whole blood were collected into each tube, resulting in a total volume of 8 mL. The blood remained in the collection solutions for at least 30 minutes, but not more than 2 hours at room temperature (~ 20 ° C) prior to processing. Tubes we re centrifuged at 2000 g for 10 minutes followed by aspiration of the plasma and buffy coat layers. Importantly, an additional 2 - mm layer off the top of the packed ERYs was also removed to minimize leukocyte presence during subsequent storage in the AS - 1 o r AS - 1N solutions. The purified ERYs from the 3 tubes containing CPD were then combined into a single 15 mL tube, followed by the addition of AS - 1 such that the ratio of packed ERYs to AS - 1 was 2:1. The same protocol was followed for ERYs collected in CPD - N and stored in AS - 1N. Finally, 2 mL of the ERYs (stored in the AS - 1 or AS - 1N) were added to PVC bags 134 and stored at 4 ° C. Prior to use, the PVC bags were sterilized under UV light overnight. The PVC bags were prepared in - house using rolled PVC (ULINE, Pleas ant Prairie WI) and a heat sealer. All blood collection and storage processes were performed under sterile conditions. All blood collection procedures from informed and consented donors were approved by the Biomedical and Health Institutional Review Board at Michigan State University. The ingredients of CPD, AS - 1 are listed in Table 4.1. The glucose concentration is 5.5 mM in CPD - N and AS - 1N, with other constituents identical to CPD and AS - 1. The ERYs stored in the normoglycemic AS - 1N solution required peri odic glucose every week by opening the PVC storage bag and adding 20 µL of 400 mM glucose in saline to the ERYs, and then re - sealing the bag in a sterile environment. The gluc ose saline was prepared by dissolving 0.72 g of dextrose in 10 mL of saline (a solution of 0.90% (weight/volume) of sodium chloride (NaCl)), in order to prevent cell lysis. This helped to maintain the glucose concentrations in the stored cells at around 5 mM, while maintaining a constant volume into which the cells were stored. 4.2.2 Sample Preparation The hematocrit of stored ERYs was first determined for quantitative re - suspension by collecting ERYs in a microcapillary tube, spinning in a microhematocrit centrifuge (CritSpin M960 - 22, Statspin, Westwood, MA ) and visually quantifying the percentage of packed cells using a microcapillary reader (Statspin, Westwood, MA). The glucose 135 levels in corresponding storage solutions were also monitored by a portable A ccu - Chek Aviva glucose meter (Indianapolis, IN). Buffered physiological salt solution (PSS, contains 4.7 mM KCl, 2.0 mM CaCl 2 , 1.2 mM MgSO 4 , 140.5 mM NaCl, 21.0 mM tris - hydroxymethyl aminomethane, 5.5 mM glucose, and 5% bovine serum albumin at pH = 7.4) wa s prepared to re - suspend the ERYs. A high glucose version of PSS (referred to as PSSH) was also prepared as a control, whose glucose concentration equals the extracellular glucose level of ERYs to be suspended. Three samples were prepared for a single exp eriment. AS - 1N - PSSN cells refer to those stored in normoglycemic AS - 1N and transfused into normoglycemic PSS; AS - 1 - PSSH refers to cells stored in hyperglycemic AS - 1 and transfused to hyperglycemic PSS; and the AS - 1 - PSSN cells were those stored in hyperglyc emic AS - 1 but transfused to normoglycemic PSS. The final hematocrit of ERYs in all three samples was 5 % in corresponding PSS, which was preheated to 37 o C before mixing with ERYs. 4.2.3 Determination of ATP Release Reversibility The 3D - printed fluidic de vice that mimics the in vitro circulation, as described in the previous chapter, was applied to measure ATP release from stored ERYs after resuspension in PSS. Briefly, the device consists of two parts, channels that allow samples to flow through, and well s above the channels where membrane inserts can be plugged in to form a membrane barrier between the flow and the well. The three ERY samples were pumped to fill all six channels on the device in duplicate, after which each channel and corresponding tubing were closed to form a loop of 450 µ L . The 136 circulation in the loops was driven by a peristaltic pump (IDEX Health & Science LLC, Oak Harbor, WA) at a flow rate of 200 L/min. An aliquot of 50 L PSS was then loaded in wells in Row E above each channel to collect ATP from the flowing ERY samples via diffusion. The entire setup of the pump, the device and closed loops with circulating samples was then placed in an incubat or at 37 °C for 20 min to better mimic circulation conditions, after which, the bulk device was detached from tubing and placed in the sample holder of a plate reader (SpectraMax 4, Molecular Devices, Sunnyvale, CA). The well - established luciferase/lucifer in chemiluminescence assay was used to detect ATP in the wells. Specifically, 10 L of L/L assay (dissolving 2.0 mg of D - luciferin in 5 mL of DDW, and adding 100 mg of firefly extract) were added into the wells simultaneously by a multichannel pipet. After 15 s, the chemiluminescence intensity above all six wells was detected by the pl ate reader, simultaneously. To confirm that the increase of released ATP was not due to cell lysis, an absorbance measurement was performed to evaluate if free hemoglobin was detected in the supernatant after flow was concluded. If hemoglobin was detected, that particular sample would be discarded due to indication of lysis. In the studies reported here, there were no samples discarded due to lysis. A calibration curve was obtained for quantification purpose by circulating five ATP standards prepared in PSS (concentrations of 0, 100, 200, 300 and 400 nM) in five randomly chosen channels with identical subsequent detection processes as described above. 137 4.2.4 Study of Reversibility of ERY Deformability by a 3D - printed Cell Filter Deformability of stored ERYs w as detected by forcing the cells through a piece of porous membrane in a filter. The number of cells that passed through the filter, relative to the number that were pumped to the filter, were counted on a hemacytometer (Reichert, Buffalo, NY), as a measur ement of deformability. The filter was fabricated by 3D - printing on the Objet Connex 350 printer (Stratasys Ltd, Eden Prairie, MN) with XY resolution of 100 m and Z of 16 m, housed in the Engineering Department of Michigan State University. As shown in F igure 4.1, the main parts of the filter are two flange - shaped slabs that were printed by VeroClear material (Stratasys Ltd, Eden Prairie, MN) whose exact composition is proprietary, but approximately contains isobornyl acrylate (15 - 30%), acrylic monomer (1 5 - 30%), urethane acrylate (10 - 30%), acrylic monomer (5 - 10; 10 - 15%), epoxy acrylate (5 - 10; 10 - 15%), acrylate oligomer (5 - 10; 10 - 15%), and a photoinitiator (0.1 - 1;1 - 2%). A sample inlet was fabricated on the top slab, which was a female fitting with printed s crew threads on the inside. The two slabs can be clamped together with a piece of semipermeable membrane in between to construct the filter, and the simultaneously printed O rings by rubber - like TangoBlack material (Stratasys Ltd, Eden Prairie, MN, approxi mately contains 30 - 60% urethane acrylate oligomer, 1 - 20% exo - 1,7,7 - trimethylbicyclo[2.2.1]hept - 2 - yl acrylate, 1 - 20% methacrylate oligomer, 1 - 20% resin, polyurethane, and 0.1 - 1% photoinitiator) on the binding sides of both slabs ensured tight holding of the membrane without liquid leaking. Figure 4.2 shows the engineering sketch of the filtration device, with detailed dimensions. 138 A piece of polycarbonate membrane with a pore size of 5 m, which is smaller than the mean diameter (6 - 8 m) of ERYs, was used in the filter. After the slab - membrane - slab device was constructed by binding the two slabs with binder clips, a section of tubing for sample introduction was connected to the top slab by screwing the male finger tight adapter at the end of the tubing into th e female inlet. The tubing was forced by a peristaltic pump (IDEX Health & Science LLC, Oak Harbor, WA) to drive samples into the filter, with a driving pressure of 12 cm H 2 O column (0.17 psi). All three samples of AS - 1N - PSSN, AS - 1 - PSSH and AS - 1 - PSSN were measured through the device in a random order. The effluent was collected for 10 min, after which, the number of cells in the effluent was counted on the hemacytometer. Cell lysis in the effluent was also checked and in the studies reported here, no filtra tion induced lysis was observed. 139 Figure 4.1. The 3D - printed cell filter for ERY deformability measurements. (A): the two slabs with O rings on the binding sides. A piece of semipermeable membrane is the key part of the filter device, which w as 5 m pore size polycarbonate membrane. (B): A view of the assembled device, with the membrane between the slabs. The assembling was achieved by simply binding the two slabs on the four wings with binder clips. (C): The view of a real filtration experime nt. A sample was introduced by a pump into the top slab, which would be forced to go through the membrane and was then collected for subsequent cell counting. (D): The schematic cross view of a filtration process. A pressure was applied to push cells to de form to pass through the membrane between the O rings. However, only cells with sufficient deformability can go through, otherwise they will be blocked by the membrane. 140 Figure 4.2. The technical drawings of the top slab, bottom slab and O rings of the filtration device , achieved by the autodesk inven tor software. Before sending the files into the printer, the O rings were assembled to both slabs in the software so that they can be printed simultaneously. The unit for all the dimensions is m m. Bottom Slab O ring 141 4.2.5 Design and Fabrication of an i.v. Injection Device A Y shaped adapter was designed and 3D - printed by the same printer as mentioned above, to mimic an in vitro i.v. injection process. As shown in Figure 4.3A, this adapter consists of three parts: a flow splitter, a controller and a stopcock. Screw threads were printed on the parts for easy connection. Just prior to use, the three parts were assembled and then the Y shaped flow splitter was connected to a channel on the fluidic device. The low branc h of the flow splitter is connected with the other end of the channel by soft tubing, which goes around the rollers of a peristaltic pump to drive the flow. The top branch and the flow controller were connected to a syringe via a male finger tight adapter. Figure 4 .3 B displays the injection of fluorescein by the syringe into a water circulation. Upon injection, the stopcock is adjusted to make the hole (through the stopcock) align with the channel so that liquid can be injected through. When an injection wa s finished, the stopcock was rotated 90 o to stop further injection. Teflon tape may be needed for better sealing between the stopcock and the controller in some cases. The top branch and the controller need to be prefilled with the liquid to be injected at the beginning of an experiment. Otherwise, some trapped air in this part will be injected into the circulation and thus the injection amount cannot be easily quantified. 1 42 Figure 4.3 . The 3D - printed i.v. injection device. (A): the device comp rises three parts: a flow splitter, a flow controller, and a stop cock. Screw threads were printed on each part for easy but tight connection. When the hole that goes through the stopcock was adjusted along the channel in the flow controller, liquid can be injected through. Otherwise, the flow will be stopped. (B): Injection of fluorescein into a water circulation. Water was circulating in the loop that connects the low branch of the splitter and the other end of a channel, while fluorescein was injected in by a syringe pump. Flow Splitter Controller Stopcock A B 143 4.2.6 Characterization of the i.v. Injection Device This device will be eventually used to study i.v. injection (transfusion) of stored ERYs, and quantitative injection will be required to acquire reliable experimental results. There fore, the injection accuracy was first characterized. Fluorescein injection was conducted to investigate the injection accuracy of small molecules, and for ease of detection. An aliquot of 50 µ L of fluorescein (300 µ M) was injected to a loop of water with a volume of 450 µ L. The injection was performed at 100 µ L/min for 30 s by a syringe pump. The liquid was allowed to circulate for another 2 min after the injection, after which, the liquid in the loop (tubing and channel) was collected and the fluorescein concentration was quantified on a plate reader (ex. 494 nm; em. 521 nm). Based on the injection ratio (1 to 9 ), the concentration of fluorescein in a loop was expected to be 30 µ M. The detected fluorescein concentration was compared with this calculated nu mber, the difference between which, can indicate injection accuracy. The possibility that the the fluidic device material may absorb fluorescein molecules was examined by detecting the loss of fluorescence signal intensity of a fluorescein solution after b eing circulated in a channel for hours. The injection accuracy of ERYs was also measured. A 5% ERY sample suspended in PSS was circulating in a loop, while 50 µ L of 50 % ERYs were injected at 100 µ L/min for 30 s. The calculated hematocrit of ERYs in the l oop after an injection was expected to be 10%. By comparing the detected hematocrit and the calculated value of 10%, injection accuracy of ERYs can be determined. In these experiments, counting of ERYs on a hemacytomer was performed for accurate hematocrit measurements. 144 The distribution of injected fluorescein molecules and ERYs in the circulating loops was also characterized. After fluorescein was injected, 2 min of circulation were allowed, after which, the liquid in the loop was divided into three random sections that were collected in three vials, respectively. If the fluorescein molecules distribute evenly in the loop after an injection, the concentrations in the three parts were expected to be the same. The distribution of ERYs after an injection was c haracterized with the same method, except 5 min of circulating time were applied before measuring the hematocrit in the three sections of tubing. 4.2.6 Study of Responsiveness of Stored ERYs to Pancreatic - cells INS - 1 cells were integrated on the fluidi c device, with 5 % fresh ERYs circulating in the channels. As shown in Figure 4.4 , INS - 1 cells cultured in membrane inserts were integr ated on the left three channels of the fluidic device, while the other three did not contain the endocrine - like cells (em pty ovals). Fresh, AS - 1 and AS - 1N stored ERYs (hematocrit 50 %) from the same donor were injected via the i.v. injection device into the fresh ERY streams. After the ERYs were injected by the procedures discussed previously, the entire setup was placed in a 37 o C incubator for 2 hours, with moisture napkins covering the wells to minimize evaporation of buffers, after which, the ATP amount in each channel was quantified on a plate reader, with the method introduced in Chapter 2. 145 Figure 4.4 . The st rategy to study the responsiveness of stored ERYs to - cells. Fresh ERYs with a hematocrit of 5% were circulating in the six channels of the fluidic device, while fresh, AS - 1N stored, and AS - 1 stored ERYs were injected into the circulations via the i.v. injection device. INS - 1 cells were included on the left three channels (green ovals). A 2 - hour incubation at 37 o C was applied after the injection, after which, ATP amount in each channel was quantified by the luciferase/luciferin assay in the blue ovals . 5 % fresh ERYs are flowing in the channels With - cells Without - cells 146 4.3 Results 4.3.1 ERY Collection and Storage, Sample Preparation Two blood storage solutions, one hyperglycemic (AS - 1, currently FDA approved) and one normoglycemic (AS - 1N, Spence modified), were used to study the effect of glucose on stored ERYs. Although the normoglycemic storage solution AS - 1N was reported beneficial to stored ERYs, it is not conducive for a storage period over 1 week due to the exhaustion of glucose by cell metabolism. Therefore, the ERYs stored in AS - 1N were fed with a 20 µ L of 200 mM glucose saline solution every week to maintain the glucose level at around 5.5 mM. Before ATP and deformability of the ERY samples were characterized, glucose levels in corresponding storage solutions were first monitored. Figure 4.5 shows that the glucose level in AS - 1 stays at an extremely high leve l around 40 mM for the whole storage period of 36 days, while in AS - 1N, the glucose concentration maintains at around 5 mM with periodic feeding of glucose, which falls into the healthy physiological range. 147 Figure 4. 5 . Extracellular glucose levels in AS - 1 and AS - 1N storage solutions. Throughout the whole 36 - day storage, glucose in AS - 1 solutions (open circles) stayed at an extremely high level around 40 mM, while AS - 1N solutions (filled circles) showed physiological g lucose level s around 5 mM. Data present the mean of 7 trials ± S. E. M. 148 4.3.2 Study of Reversibility of ATP Release on the 3D - Printed Fluidic Device ATP is an important molecule in the blood stream because of its participation in regulating blood flo w. It was previously reported that ERYs stored in hyperglycemic AS - 1 solution released significantly less ATP than those stored in AS - 1N 31 . However, the performance of hyperglycemia stored ERYs transfused into a normoglycemic environment, remains unknown. Therefore, three experimental trials were designed and conducted to determine if the ATP release capability of hyperglycemia stored ER Ys is reversible. A physiological salt solution (PSS) was used as the transfusion media, into which the stored ERYs would be transfused. The AS - 1N - PSSN trial was to transfuse AS - 1N stored ERYs to normoglycemic PSS (glucose 5.5 mM), the AS - 1 - PSSH was to tra nsfuse AS - 1 stored ERYs to hyperglycemic PSS (glucose concentration was equal to that in AS - 1) and the AS - 1 - PSSN trial was to transfuse AS - 1 stored ERYs to normoglycemic PSS to investigate the reversibility of ATP release. To better mimic a real transfusio n process, the stored ERYs were transfused into corresponding PSS by a 1:10 v/v ratio (with a final hematocrit of 5%), and the blood samples were then introduced into the six channels, in duplicate, on the 3D - printed fluidic device that allows the blood sa mples to circulate for 20 min at 37 °C . ATP in each flowing stream was collected from loaded buffer in wells E above the channels for subsequent quantification on a plate reader. As shown in Figure 4.6 , ATP release from AS - 1N - PSSN cells maintains constant at around 200 nM for the whole storage period, while AS - 1 - PSSH released ATP stays at a much lower level and the AS - 1 - PSSN released ATP keeps decreasing with time. 149 Comparably speaking, on Day 1, or after several hours of storage, ATP released from the AS - 1 - PSSH cells is already 40% less than AS - 1N - PSSN cells, while the AS - 1 - PSSN cells release even a little more ATP than AS - 1N - PSSN, though the difference between the latter two strategies is not significant. However, on Day 5, AS - 1 - PSSN released ATP begins to decrease and on Day 8, it is significantly lower than AS - 1N - PSSN, yet still higher than AS - 1 - PSSH. The same trend is observed until Day 15, when the ATP release from AS - 1 - PSSN have decreased to an extent that is not significantly different from AS - 1 - PSSH. After Day 15, ATP release from both AS - 1 - PSSH and AS - 1 - PSSN are statistically the same (p>0.5), but are dramatically lower than AS - 1N - PSSN. These results confirmed our previous results that the high glucose levels in the storage solutions impair ATP releas e from ERYs. Furthermore, Figure 4.6 clearly indicates that ATP release from hyperglycemia stored ERYs is conditionally reversible after ERYs are transfused back to a normoglycemic environment. After 8 days of storage, the ATP release reversibility is part ially impaired and after Day 15, the reversibility is no longer observed, suggesting permanent adverse alteration has occurred on these cells after 15 days of storage in hyperglycemic environments. 150 Figure 4.6 . A TP release reversibility data at different storage days. AS - 1N - PSSN trial referred to ERYs stored in AS - 1N (normoglycemic) transfused into normoglycemic PSS; AS - 1 - PSSH trial represented ERYs stored in AS - 1 (hyperglycemic) transfused into hyperglycemic PSS; AS - 1 - PSSN trial was ERYs stored in AS - 1 but transfused into normoglycemic PSS. During the 36 storage days, AS - 1N - PSSN ERYs released higher ATP at a constant level, while AS - 1 - PSSH ERYs released less ATP. The AS - 1 - PSSN ERYs, however, could reverse the ATP release before 15 days (complete reverse before Day 5 and partial reverse on Day 8 and 12). Seven biological replica were analyzed and the data represent mean ± S.E.M. 151 4.3.3 Study of Reversibility of ERY Deformability The reversibility of stored ERY d eformability was also studied due to the causal relationship between deformability and ATP release 32 . A deformability investigation was completed by a filtration test that has been widely used to study ERY deformability due to its low instrument requirements and high reproducibility 33 , 34 . Because commercial devices were not found proper for this application, 3D - printing, used to fabricate a filtration device that minimizes dead volumes, saves th e usage of filtering membranes and simplifies experimental operation. As shown in Panel A in Figure 4.1, the device is comprised of two slabs, with simultaneously 3D - printed O rings to seal a piece of semipermeable membrane in between for cell filtration. Due to the small sizes of the O rings, the dead volume of the filter is minimized to about 55 L. Panel B shows the device assembly, which was constructed by clamping the slabs with four binder clips on the wings. Panel C demonstrates a real ERY filtration experiment on the device. ERY samples were pumped into the inlet on the top slab, and were fo rced to go through the membrane into the cone shaped chamber in the bottom slab, and were collected for cell counting. Cells that had sufficient deformability would deform to pass through the filter, whose pore size (5 m) was ter (6 - 8 m), while less deformable cells would not be able to go through (Panel D). Resuspended 5% AS - 1N - PSSN, AS - 1 - PSSH and AS - 1 - PSSN cells were evaluated with this method. For a clearer data presentation, the cell number in the filter effluent 152 of AS - 1N - PSSN trial on Day 1 was set to 100%, and all other collected data were normalized to this 100% value. As shown in Figure 4.7 , the deformability of AS - 1N - PSSN ERYs maintains constant at around 100% during the whole 36 days storage, while AS - 1 - PSSN trial kee ps decreasing from 100% on Day 1 to 60% on Day 36. The deformability of AS - 1 - PSSH ERYs also maintains constant, but at a significantly lower level than AS - 1N - PSSN. Comparably speaking, on Day 1, AS - 1 - PSSH cells lose much deformability (about 25%) compared with AS - 1N - PSSN, while AS - 1 - PSSN cells completely reverse their deformability after being transfused to the normoglycemic PSS. The trend continues until Day 8, when deformability of AS - 1 - PSSN trial drops to AS - 1 - PSSH level. After Day 8, however, the deform ability of AS - 1 - PSSN cells keeps decreasing, though at a slow rate, without any observed reversibility. 153 Figure 4.7 . Deformability reversibility data at different storage days. Black bars were constant arou nd 100% during 36 days storage, which indicate normoglycemia stored ERYs did not lose deformability after being transfused to normoglycemic PSS. The AS - 1 - PSSH cells, however, lost their deformability even on the first storage day. The hyperglycemia stored ERYs could reverse their deformability after being transfused into normoglycemic PSS for the first five days of storage. From Day8, these cells also lost the ability to reverse their deformability. Data represent mean values of 4 biological replica ± S.E.M . 154 4 .3.4 Characterization of the i.v. Injection Device After 50 µ L of 300 µ M fluorescein were injected into a water circulation of 450 µ L, the expected concentration of fluorescein in the loop is 30 µ M. As shown in Table 4.2, the measured fluorescein con centrations in the six loops of the fluidic device after an injection were around 30 µ M, suggesting the injection of small molecules such as fluorescein via the 3D - printed i.v. injection device can be very accurate. On the other hand, the detected fluoresc ein concentrations do no show significant difference between loops, which indicates the reproducibility between different injection devices is high. Although an extra fluorescein aliquot (50 µ L) was injected into a loop, due to the elasticity of the tubing , no leakage or burst was observed due to the additional fluorescein added. To examine if the injected fluorescein molecules mixed well in a flowing stream, the liquid in a loop was divided into three parts randomly, and each of them was collected in a se parate vial. The concentrations of fluorescein in the three parts were then quantified by a plate reader. If the molecules distribute evenly in a loop after injection, the concentrations of fluorescein in the three parts are expected to be the same. The da ta summarized in Table 4.3 clearly show that the concentration of fluorescein in each part of a loop is around 30 µ M, which indicates that the injected fluorescein molecules are able to distribute evenly along a whole loop after 2 min mixing after the inje ction. 155 Table 4.2. The concentration of fluorescein in each loop after an injection N=5; error=stdev Table 4.3. The concentrations of fluorescein in each part of a loop Fluorescein l oop 1 l oop 2 l oop 3 P 1 P 2 P 3 P 1 P 2 P 3 P 1 P 2 P 3 30.6± 1.2 30.5± 0.1 30.6± 1.1 29.5± 1.7 29.6± 1.7 30.1± 1.2 29.8± 0.9 30.5± 1.3 29.8± 1.1 Fluorescein l oop 4 l oop 5 l oop 6 P 1 P 2 P 3 P 1 P 2 P 3 P 1 P 2 P 3 2 9.5± 1.4 29.5± 1.7 29.9± 1.5 29.8± 0.3 28.7± 0.5 28.3± 0.1 30.4± 0.6 30.4± 0.7 31.0± 1.3 N = 5; error=stdev loop 1 loop 2 loop 3 loop 4 loop 5 loop 6 Fluorescein 30.3±2.4 29.3±1.1 30.7±0.9 29.2±0.8 30.5±1.2 30.0±1.0 156 T he injection accuracy of ERYs and the distribution of injected ERYs were also characterized using a similar method as fluorescein. ERYs with a hematocrit of 50 % were injected into 5 % ERY streams, with an expected final hematocrit of 10 % in the loops. Data in Table 4.4 and Table 4.5 prove that ERYs can be injected accurately and can distribute evenly along a loop after being injected. Table 4.4. The hematocrit of ERYs in each loop after an injection N=5 ; error=stdev Table 4.5. The hematocrit of ERYs in each part of a loop Hct, % l oop 1 l oop 2 l oop 3 P 1 P 2 P 3 P 1 P 2 P 3 P 1 P 2 P 3 9.9±0. 4 10.2± 0.5 10.7± 0.2 10.4± 0.3 10.4± 0.3 10.2± 0.3 10.3± 0.1 10.4± 0.7 10.4± 0.5 Hct, % l oop 4 l oop 5 l oop 6 P 1 P 2 P 3 P 1 P 2 P 3 P 1 P 2 P 3 9.9±0. 1 10.0± 0.3 10.3± 0.3 10.3± 0.3 10.4± 0.1 10.3± 0.5 9.7±0. 5 10.1± 0.5 10.3± 0.3 N=5; er ror=stdev loop 1 loop 2 loop 3 loop 4 loop 5 loop 6 Hct, % 10.3±0.6 9.7±0.5 10.1±1.0 10.7±1.3 10.7±0.8 10.4±0.5 157 4.3.5 Study of Responsiveness of Stored ERYs to Pancreatic - cells The results in Tables 4.4 and 4.5 show that ERYs can be transfused quantitatively via the i.v. injection device, and the injected ERYs can mix evenly in a circulation stream. In t his part, the response of the injected ERYs to pancreatic - cells was evaluated. INS - 1 cells were cultured in membrane inserts and integrated on the fluidic device. Next, a 5% hematocrit of fresh ERYs were delivered through the six channels, while three ty pes of ERYs were injected into the circulation: fresh, AS - 1 stored, and AS - 1N stored ERYs. After incubation at 37 o C for 2 hours, the ATP in each channel was quantitatively determined by a plate reader measurement of luciferase/luciferin assay. In Figure 4 . 8 , f - f refers to fresh ERYs transfused to a fresh ERY stream; AS - 1N - f refers to AS - 1N stored ERYs transfused to fresh ERYs; and AS - 1 - f refers to AS - 1 stored ERYs transfused to fresh ERY streams. The f - f trial is to transfuse fresh ERYs to fresh ERYs in ci rculation, AS - 1N - f is to transfuse AS - 1N stored ERYs to fresh ERYs in circulation, and AS - 1 - f trial is to transfuse AS - 1 stored ERYs to fresh ERY in circulation. By comparing Figure 4. 8 A and Figure 4. 8 B, it can first be concluded that with INS - 1 cells, the ATP release in the trials of f - f and AS - 1N - f is significantly higher (A) than those without the endocrine cells (B), suggesting that after being transfused, like fresh ERYs, AS - 1N stored ERYs can respond to - cells by releasing ATP. The AS - 1 stored ERYs, however, failed to respond to the endocrine cells. Interestingly, though a loop after an injection consists of 5 % fresh ERYs and the same amount of AS - 1 stored ERYs, the fresh cells did not respond to - cell s either. Though more evidence will be needed, it appears that the transfusion of AS - 1 stored ERYs had an adverse effect on fresh ERYs. 158 Considering the reported clinical observation that patients who received transfused blood are prone to suffer from insuf ficient NO bioavailability, which may be a determinant in post transfusion complications, these results indicate that ERYs stored in a normoglycemic condition (AS - 1N) retained their ability to respond to - cells, which can potentially prevent the problems of insufficient NO bioavailability and related complications. 159 Figure 4.8 . ATP release from ERYs after being transfused via the 3D - printed i.v. injection device. (A): with INS - 1 cells on the channels; (B): without INS - 1 cells on the channels. f - f: fresh ERYs transfused to fresh ERYs; AS - 1N - f: AS - 1N stored ERYs transfused to fresh ERYs; AS - 1 - f: AS - 1 stored ERYs transfused to fresh ERYs. In these experiments, 50 µ L of f/AS - 1N/AS - 1 E RYs were transfused to a circulation of 450 µ L . N=4, error=S.E.M. A B 160 4.4 Discussion The Spence group has reported that the currently approved hyperglycemic ERY storage solutions impair ATP release from the cells and proposed an alternative storage solution (AS - 1N) with a physiological glucose level, in which the ERYs have proven to release significantly more ATP. This study continues to investigate the adverse effects of hyperglycemia on stored ERYs by 3D - printed devices, with an emphasis on biochemical reve rsibility of the ERYs and the response of stored ERYs to endocrine - cells. The reversibility of ATP release from stored ERYs was studied by applying the 3D - printed fluidic device as discussed in Chapter 2. This device can better mimic the real conditions of ERYs by circulating samples at 37 o C . Furthermore, it allows for simultaneous measurements of all three samples under a plate reader in a high throughput fashion. All these features make it an ideal platform for flow induced ATP measurements from stored ERYs. Data shown in Figure 4.6 reaffirmed our previous findings tha t high glucose in storage solutions reduced ERY ATP release, and extended that observation to demonstrate how reversible these ERYs are, in terms of ATP release, during the storage period. Day 15 was observed as a turning point, before which, the hyperglyc emia stored ERYs are able to reverse (completely or at least partially) the ATP release ability, but after which the reversibility cannot be observed. ERYs participate in regulating blood flow through releasing ATP that has proven to be able to stimulate endothelium - derived NO production, a well - established vessel 161 dilator. Numerous studies and surveys have demonstrated the existence of adverse complications after transfusion and the 15 storage day as a turning point of the quality of stored ERYs. It is not to be assumed that the failed ATP release reversibility after 15 days storage in hyperglycemia is the root cause of the clinical consequences. However, the ATP reversibility trend against storage time was consistent with the clinical observations, both sh owed an adverse alteration after two weeks. And given the important role of ERY released ATP in a blood stream, though more direct evidence will be required, it is rational to hypothesize a link between irreversible ATP release and the post transfusion hea lth problems in such a scenario that insufficient ATP release from long - term stored (e.g. two weeks or longer) ERYs causes insufficient NO bioavailability, which restricts blood transportation to tissues and thus adversely How ever, the mechanism that describes how hyperglycemia leads to failed reversibility of ATP release still remains unknown. The mechanism would lie in one of the two possible scenarios: hyperglycemia induces less intracellular ATP production or affects the AT P release process. Wang reported that intracellular ATP in hyperglycemia stored ERYs was higher than those stored in normoglycemic solutions, and it did not significantly decrease throughout 36 days storage 31 , which inferred hyperglycemia may adversely alter the ATP release process, rather than reducing intracellular ATP production. 162 Sprague reported that sheer stress induced deformabilit y of ERYs leads to ATP release through a CFTR involved mechanism, suggesting that ATP release is related to cell deformability 32 . These findings provided a clue t o potentially explain the failed ATP reversibility by studying the deformability of stored ERYs. A user - friendly 3D - printed cell filter with a small dead volume was fabricated and applied to study the deformability of stored ERYs. Since the filter membrane pore size is smaller than human ERY mean diameter, only cells with sufficient deformability can go through. Therefore, the number of cells in the filter effluent can be used as a measurement of deformability. As shown in Figure 4.7 , compared with the AS - 1 N - PSSN cells, the AS - 1 - continued to decrease beyond that time point. By combining the ATP and deformability data ( Figure 4.6 and Figure 4.7 ), a correlation between deformability and ATP re lease can be preliminarily revealed. During the first five days of storage in AS - 1, ERY deformability could completely reverse after being transfused into a normoglycemic condition, as did the ATP release. On Day 8 and 12, however, the deformability of ERY s could not reverse, and the ATP release could only partially reverse, which was significantly lower than AS - 1N - PSSN cells. After 15 Days storage, a permanent alteration of cell deformability had occurred, as did the ATP release. This correlation suggests that hyperglycemia may permanently and adversely alter the deformability of ERYs via a certain mechanism, which eventually leads to decreased ATP release. 163 Although the results discussed above suggest ERYs stored in hyperglycemic conditions can reverse thei r deformability and ATP release in the first a few days of storage, ERYs stored in normoglycemic condition never lost their deformability or ATP release during the whole 36 - day storage, which reinforced our hypothesis that an alternative normoglycemic solu tion should be considered for ERY storage used in transfusion medicine. To better study a transfusion process in vitro , a 3D - printed i.v. injection device was invented, which consists of a flow splitter, a controller and a stopcock. 3D - printing enables the fabrication of threads on the device and precisely controlled dimensions, which contribute to the high inter - device reproducibility. The characterization experiments of the device showed that it enables quantitative injection of solutions and ERY suspensi ons without leakage, and the injected parts can be mixed evenly in a circulation in a short time after an injection. Pancreatic - cells (INS - 1 cells) were integrated on the platform for investigation of the responsiveness of stored ERYs to these endocrine cells. It was found that AS - 1N stored ERYs respond as fresh ERYs to the - cells by releasing significantly more ATP than those w ithout the - cells. However, after the AS - 1 stored ERYs were transfused to a stream of fresh ERYs, no response to the - cells was observed. It appears that even the preexisting fresh ERYs in the flowing stream did not respond to the endocrine cells, which indicated that transfusion of hyperglycemia (AS - 1) stored ERYs may adversely affect the original fresh ERYs in a patient. More evidence will be needed to confirm this hypothesis, which will be 164 discussed in Chapter 5. All these results suggest that AS - 1N st ored ERYs can potentially be healthier than those stored in currently approved AS - 1. 4.5 Conclusion In this study, the reversibility of ATP release and deformability of hyperglycemia stored ERYs was studied with 3D - printed devices. Measurements on a 3D - pri nted circulation mimic device showed that hyperglycemia - stored ERYs cannot reverse or restore ATP release to normal levels after 15 days of storage, whil e normoglycemia - stored ERYs are robust to release ATP throughout the whole storage period. Deformabilit y reversibility of stored ERYs was also examined by a 3D - printed cell filter, which indicated hyperglycemia induces a permanent deformability alteration on stored ERYs after 1 to 2 weeks storage, which may explain the failure of ATP release reversibility. An i.v. injection device was designed and 3D - printed to investigate the responsiveness of stored ERYs to pancreatic - c ells, in a more clinically relevant manner. The results suggest that AS - 1N stored ERYs can respond to - c ells like fresh ERYs, by releasi ng increased ATP. The AS - 1 stored ERYs, however, failed to respond to the - c ells even after only 1 day storage, indicating some lesions has occurred on the stored cells by hyperglycemia. More importantly, after being mixed with AS - 1 stored ERYs, original fresh ERYs cannot respond to the - c ells either, which initially suggests that AS - 1 stored ERYs may adversely affect the original ERYs of a patient by a certain mechanism. All these results, combined with the previous reports involving ERY storage and post transfusion complications, strongly suggest that normoglycemic 165 storage conditions for ERYs may help maintain normal properties of the cells and potentially reduce the risk of post - transfusion complications. 166 REFERENCES 167 REFERENCES 1. Witaker, B.I. 2011 National Blood Collection and Utilization Survey. American Association of Blood Banks (2011). 2. Rubin, O., Crettaz, D., Tissot, J.D. & Lion, N. Pre - analytical and methodological challenges in red bloo d cell microparticle proteomics. Talanta 82 , 1 - 8 (2010). 3. Rous, P. & Turner, J.R. The preservation of living red blood cells in vitro. I. Methods of preservation. J Exp Med 23 , 219 - 237 (1916). 4. Robertson, O.H. & Rous, P. Autohemagglutination experime ntally induced by the repeated withdrawal of blood. J Exp Med 27 , 563 - U566 (1918). 5. Mollison, P.L. The introduction of citrate as an anticoagulant for transfusion and of glucose as a red cell preservative. Brit J Haematol 108 , 13 - 18 (2000). 6. Hogman, C.F., Hedlund, K. & Sahlestrom, Y. Red - Cell Preservation in Protein - Poor Media .3. Protection against Invitro Hemolysis. Vox Sang 41 , 274 - 281 (1981). 7. Hogman, C.F., Hedlund, K., Akerblom, O. & Venge, P. Red Blood - Cell Preservation in Protein - Poor Media .1. Leukocyte Enzymes as a Cause of Hemolysis. Transfusion 18 , 233 - 241 (1978). 8. Moore, G.L., Ledford, M.E. & Brooks, D.E. Distribution and Utilization of Adenine in Red Blood - Cells during 42 Days of 4c Storage. Transfusion 18 , 538 - 545 (1978). 9. Moore, G.L., Ledford, M.E. & Peck, C.C. The Invitro Evaluation of Modifications in Cpd - Adenine Anti - Coagulated - Preserved Blood at Various Hematocrits. Transfusion 20 , 419 - 426 (1980). 10. Moore, G.L. Additive Solutions for Better Blood Preservation. Crit Rev Cl Lab Sci 25 , 211 - 229 (1987). 11. Hogman, C.F., Eriksson, L., Hedlund, K. & Wallvik, J. The Bottom and Top System - a New Technique for Blood Component Preparation and Storage. Vox Sang 55 , 211 - 217 (1988). 168 12. Hill, H.R., Oliver, C.K., Lippert, L.E., Green walt, T.J. & Hess, J.R. The effects of polyvinyl chloride and polyolefin blood bags on red blood cells stored in a new additive solution. Vox Sang 81 , 161 - 166 (2001). 13. Said, O. et al. Maintaining A Physiological Blood Glucose Level with Glucolevel, A C ombination of Four Anti - Diabetes Plants Used in the Traditional Arab Herbal Medicine. Evid - Based Compl Alt 5 , 421 - 428 (2008). 14. Tinmouth, A., Fergusson, D., Yee, I.C., Hebert, P.C. & Investigators, A. Clinical consequences of red cell storage in the cri tically ill. Transfusion 46 , 2014 - 2027 (2006). 15. Fratantoni, J.C. Points to Consider in the Safety Evaluation of Hemoglobin - Based Oxygen Carriers. Transfusion 31 , 369 - 371 (1991). 16. Purdy, F.R., Tweeddale, M.G. & Merrick, P.M. Association of mortality with age of blood transfused in septic ICU patients. Can J Anaesth 44 , 1256 - 1261 (1997). 17. Basran, S. et al. The association between duration of storage of transfused red blood cells and morbidity and mortality after reoperative cardiac surgery. (vol 10 3, pg 15, 2006). Anesth Analg 108 , 1953 - 1953 (2009). 18. Basran, S. et al. The association between duration of storage of transfused red blood cells and morbidity and mortality after reoperative cardiac surgery. Anesth Analg 103 , 15 - 20 (2006). 19. Hoyt, D. et al. A 12 - year prospective study of postinjury multiple organ failure - Has anything changed? Discussion. Arch Surg - Chicago 140 , 439 - 440 (2005). 20. Ciesla, D.J. et al. A 12 - year prospective study of postinjury multiple organ failure - Has anything c hanged? Arch Surg - Chicago 140 , 432 - 439 (2005). 21. Offner, P.J., Moore, E.E., Biffl, W.L., Johnson, J.L. & Silliman, C.C. Increased rate of infection associated with transfusion of old blood after severe injury. Arch Surg - Chicago 137 , 711 - 716 (2002). 22. Zallen, G. et al. Age of transfused blood is an independent risk factor for postinjury multiple organ failure. Am J Surg 178 , 570 - 572 (1999). 23. Koch, C.G. et al. Duration of red - cell storage and complications after cardiac surgery. New Engl J Med 358 , 1229 - 1239 (2008). 24. Szymanski, I.O. & Odgren, P.R. Studies on the Preservation of Red Blood - Cells - Attachment of the 3rd Component of Human - Complement to Erythrocytes during Storage at 4 - Degrees - C. Vox Sang 36 , 213 - 224 (1979). 169 25. Valtis, D.J. & Kenne dy, A.C. Defective Gas - Transport Function of Stored Red Blood - Cells. Lancet 1 , 119 - 125 (1954). 26. Carroll, J. et al. An altered oxidant defense system in red blood cells affects their ability to release nitric oxide - stimulating ATP. Mol Biosyst 2 , 305 - 31 1 (2006). 27. Sprague, R.S., Ellsworth, M.L., Stephenson, A.H. & Lonigro, A.J. ATP: The red blood cell link to NO and local control of the pulmonary circulation. Am J Physiol - Heart C 271 , H2717 - H2722 (1996). 28. Roback, J.D., Neuman, R.B., Quyyumi, A. & Sutliff, R. Insufficient nitric oxide bioavailability: a hypothesis to explain adverse effects of red blood cell transfusion. Transfusion 51 , 859 - 866 (2011). 29. Allen, B.W. & Piantadosi, C.A. How do red blood cells cause hypoxic vasodilation? The SNO - hem oglobin paradigm. Am J Physiol - Heart C 291 , H1507 - H1512 (2006). 30. Tsai, A.G., Hofmann, A., Cabrales, P. & Intaglietta, M. Perfusion vs. oxygen delivery in transfusion with "fresh" and "old" red blood cells: The experimental evidence. Transfus Apher Sci 43 , 69 - 78 (2010). 31. Wang, Y.M., Giebink, A. & Spence, D.M. Microfluidic evaluation of red cells collected and stored in modified processing solutions used in blood banking. Integr Biol - Uk 6 , 65 - 75 (2014). 32. Sprague, R.S., Ellsworth, M.L., Stephenson, A.H., Kleinhenz, M.E. & Lonigro, A.J. Deformation - induced ATP release from red blood cells requires CFTR activity. Am J Physiol - Heart C 275 , H1726 - H1732 (1998). 33. Juhan, I. et al. Abnormalities of Erythrocyte Deformability and Platelet - Aggregation in In sulin - Dependent Diabetics Corrected by Insulin Invivo and Invitro. Lancet 1 , 535 - 537 (1982). 34. Srour, M.A., Bilto, Y.Y., Juma, M. & Irhimeh, M.R. Exposure of human erythrocytes to oxygen radicals causes loss of deformability, increased osmotic fragility , lipid peroxidation and protein degradation. Clin Hemorheol Micro 23 , 13 - 21 (2000). 170 Chapter 5 Conclusions and Future Directions 5.1 Conclusions 5.1.1. C - peptide Stimulated ATP Release from Erythrocytes on a 3D - printed Circulation - mimic Fluidic Devi ce Microfluidics technologies have become a powerful tool in the field of bioanalytical chemistry because of their ability to control flow and reduce the consumption of reagents and samples 1 . Most microfluidic devices are fabricated in PDMS, a nontoxic and transparent elastomer 2 , 3 . However, despite the advantages of this polymer, it possesses some drawbacks that prevent microfluidics from becoming a common tool in the laboratory. For example, PDMS devices are not rigid, which may cause flow problems such as leakage and uneven pressure distribution 4 . These devices are single use, which not only increases research cost (time and money), but also compromises the reproducibility of results. Although previous research has successfully integrated multiple functional units, such as on - ch ip mixers and detectors on PDMS microfluidic devices, the fabrication is lab - dependent, with poor inter - lab reproducibility 5 . Finally, although there has been a large number of publications studies did not elici t many commercialized products 6 , 7 . The lack of standardization and ruggedness of PMDS - based devices are probably pa rtially responsible for this impasse 8 . In this work, 3D - printing was employed to fabricate a more standard and rugged microfluidic device. 3D - printing is a novel ra pid prototyping technique using an 171 additive process to build a 3D object 9 . Instead of the mul tiple soft lithography steps and binding multiple layers of PDMS slabs, 3D - printing can fabricate a microfluidic device in one step, with precisely controlled shapes, features and dimensions 10 - 12 . The device developed in this study is based on the dimensions of a standard 96 - well plate, thereby enabling convenient and high content readout by a plate reader. Moreover, the ruggedness of the device allows for long te rm experiments ( e.g., >2 hours circulation of ERYs), which cannot be easily achieved on soft PDMS devices. The reusability of the device also enhances the precision of the results. Some unique features are also implemented on the device. For example, threa ds are printed on the ends of each channel to connect a tubing via standard male finger tight adapters. The design of static wells enables on - chip simultaneous calibration. All these results suggest that with 3D - printing, a microfluidic device can be fabri cated in a more standard and simple way, with integration of complicated features. This 3D - printed device facilitates the study of the effect of C - peptide on circulating ERYs. First of all, our previous conclusion that C - peptide and Zn 2+ can stimulate ATP release from ERYs was confirmed on the device. More importantly, this result has further advanced by discovering the indispensable role of albumin in this process. Without albumin, C - peptide and Zn 2+ cannot exert their effect on ERYs, suggesting that album in may be the molecule that delivers C - peptide and Zn 2+ to the ERY. Without albumin in the solutions, C - peptide cannot be delivered onto ERYs 13 . These results support our previous theory involving C - peptide and Zn 2+ and provide further evidence that this effect requires an ensemble of albumin, C - peptide, and Zn 2+ . 172 5.1.2 An Organs - on - a - Chip platform to Study the Interactions between Pancreatic - cells and Blood Components The 3D - printed fluidic device described in this dissertation also shows its versatility in an Organs - on - a - Chip platform. To study the effe ct of endogenous C - peptide on ERYs in a more physiologically relevant manner, this platform contains a - cell line, INS - 1 cells, to secret C - peptide gradually into an ERY circulation. The - cells and endothelial cells were successfully integrated on the de vice by using membrane inserts. The application of membrane inserts presents a novel philosophy of Organs - on - a - Chip construction. Any cells that can be cultured in a membrane insert can be directly plugged into a fluidic device, before which the cell prope rties can be thoroughly examined. The secretion profiles of the - cells on the platform resemble a real C - peptide secretion process, by producing C - peptide at the same levels as those believed to occur in vivo . Interactions between the - cells, ERYs, and endothelial cells were studied. It was found that the - cells do not have any effect on the endothelial cells, in terms of NO production. However, ERYs are affected by the endocrine cell secretion, the result of which is an increase of ERY - derived ATP. The ERY released ATP will then exert a downstream effect on the endo thelial cells by stimulating NO production and release. By using PPADS, a P2y receptor on endothelial cells that ATP binds, it is confirmed that ATP acts as an indirect vasodil ator by stimulating endothelium - heory that ERYs need to be present for shear stress to enhance NO production from endothelial cells 14 . 173 5.1.3 Evaluating Stored Erythrocytes used in Transfusio n Medicine by 3D - Printed Analytical Devices Stored ERYs suffer from adverse physiological alterations, or storage lesions that can lead to serious complications for transfusion patients, including increased morbidity 15 - 17 . Previous studies show that the hyperglycemia (40 mM) in a storage bag can diminish ERYs release d ATP, which is an important molecule in regulating blood flow. However, in a common transfusion proc ess, about 1/10 ERYs of the whole volume of blood (400 mL v.s. 5 L) are transfused to a normoglycemic circulation. The performance of the stored ERYs after being transfused back to a normoglycemic environment remains unknown. Therefore, a primary focus of this work is to determine if hyperglycemia - stored ERYs can restore or reverse their capability to release ATP after suspension in a normoglycemic environment. PSS was used as the transfusion media, into which stored ERYs were transfused at a 1:10 ratio. Af ter this in vitro transfusion, the samples were allowed to circulate in the 3D - printed fluidic device followed by on - chip ATP determination. The results show that after about 5 days storage in hyperglycemic AS - 1, the ability of the ERYs to release ATP can recover back to normal levels. However, after about 1 week, the ATP release can only partially reverse, and after about 2 weeks, the reversibility cannot be observed anymore, suggesting a permanent adverse alteration has occurred on the ERYs after 2 weeks of storage. P atients who receive stored ERYs older than two weeks are more likely to suffer the complications 18 , 19 , which is consistent with our observation that the 174 hyperglycemic solutions impair the ATP release of ERYs after 15 days storage. A reduction in ATP release correlates with impaired vessel dilation and restricted blood flow 20 . Though more direct evidence will be needed, given the important role of ATP in the blood stream, and the 15 day clinical observations, it is rational to hypothesize a correlation between im paired ATP release reversibility and post - transfusion complications. To better mimic a real transfusion process in vitro , an i.v. injection device was developed and fabricated using 3D - printing. Initial characterization shows that this device enables quant itative injection of solutions and ERY suspensions. The injected parts can distribute evenly along a loop in a short time after an injection. The INS - cells were integrated on the i.v. injection platform to study the response of stored ERYs to the endocrine cells. AS - 1N stored ERYs can respond to the - cells by releas ing ATP, as fresh ERYs do. AS - 1 - stored ERYs, however, cannot respond to the - cell secretions as wells as ERYs in normoglycemic environ ments. Moreover, after the AS - 1 - stored ERYs were transfused to preexisting fresh ERYs, the original ERYs cannot respond to the endocrine cells either, indicating that the transfused ERYs may affect those fresh ERYs in a n unknown mechanism. More evidence will be needed to confirm this. A possible mechanism by which hyperglycemia impairs ATP release from ERYs was also investigated. It has been shown that there exists a causal relationship between ERY deformability and ATP release 21 . Therefore the effect of hyperglycemia on ERY deformability was investigated using an in - house prepared 3D - printed cell filter. 175 Because 3D - printing enables the fabrication of desired features for a user, this 3D - printed cell filter has a minimized dead volume of 55 µ L. Moreover, this device does not require complicated instruments and its use is very straightforward. The deformability data show that hyperglycemia can permanently stiffen stored ERYs after 5 days of storage, which is consistent with t he ATP reversibility observation. Based on these results, it can be initially concluded that hyperglycemia impairs ERY deformability, which leads to a diminished ATP release. 5.1.4 Summarized Conclusion I n this dissertation, several 3D - printed devices wer e invented to study the role of ERYs in a variety of biological fields. From an engin eering perspective, 3D - printing can fabricate functional, reusable, and reproducible microfluidic devices used for bioanalysis, and provide s superior advantages over PDMS - based microfluidics. Also, with 3D - specific demands. From a physiological perspective, the results acquired on these 3D - printed device revealed new information towards our ulti mate goal of developing an additional replacement therapy for people with type 1 diabetes. The adverse effects of currently approved ERY storage solutions on ERYs were also evaluated and verified. 5.2 Future Directions 5.2.1 C - peptide - Stimulated ATP Relea se from Erythrocytes on a 3D - printed Circulation - mimic Fluidic Device 176 The main problem of the 3D - printed fluidic device is the dimensions (mm) of the channels. It appears to be a common problem that the resolution of current 3D - printers needs to be improve d. Taking the Connex 350 that was used in this study for example, the best resolution on the XY plane is 100 µ m. The relatively poor (compared with soft lithography) resolution can also cause high surface roughness that will lead to resistive flow and cell adhesion. The further development of 3D - printing technology can naturally solve these problems in the future. However, for now, other strategies may be used to avoid these problems. For example: PDMS coating on 3D - printed fluidic devices can effectively r educe the dimensions of channels and smooth the surfaces 22 . The shape of channels on the device can also be optimized to better mimic blood circulation. For exampl e, bifurcations can be applied to study the physiology of ERYs under different flow conditions (bifurcated smaller channels will increase shear stress for ERYs). The results shown in this work reveal the important role of albumin in delivering C - peptide an d Zn 2+ to ERYs. However, the binding structure of the three molecule ensemble remains unknown. Crystallography combined with X - ray diffraction, or NMR spectroscopy can be applied to further study the binding structure. Also, the ERY receptors for C - peptide and Zn 2+ are not understood. 177 5.2.2 An Organs - on - a - Chip platform to Study the Interactions between Pancreatic - cells and Blood Components This Organs - on - a - Chip platform that integrates pancreatic - cells, ERYs, and endothelial cells enable s studies of interactions between cell types. However, future research may make it a better physiological mimic. First, the culture of the INS - 1 cells can be improved. These cells were cultured in a plan a r way along the membrane in this study. Though the INS - 1 cells ar e prone to aggregate to f or m pseudo islets, this culture cannot completely mimic in vivo - cells, which are aggregated in an islet. The technique of 3D cell culture can be applied to culture a spheroid of the - cells, which can then be placed in a membrane insert on the fluidic device. The tumor cell property of the INS - 1 cells increases the f easibility of culturing an islet in vitro . In such a manner, the secretion profiles of the cells need to be re - measured because the interactions between the INS - cells in a spheroid may change the secretion patterns. Another study that needs to be conducte d is on - chip separation. The results presented in this work can only directly prove that the secretions of INS - 1 cells stimulate ATP release from ERYs. However, in addition to C - peptide, these endocrine cells also secrete other molecules, such as insulin a nd amylin. Though previous results in our group show that insulin does not stimulate ERY - derived ATP, our results will be more conclusive if a separation process can be applied before the secretions entering an ERY circulation. 178 Chelates and affinity reage nts immobilized o n a piece of porous membrane have become a commonly used technique for membrane based separations 23 . He re, results will be improved if on the membr ane where INS - 1 cells are cultured, an anti - C - peptide antibody can be immobilized to prevent the secreted C - peptide from entering ERY streams. If no increased ATP release is observed after this, it can be concluded that it is C - peptide that plays the impor tant role (as opposed to, say, insulin). In a similar way, the role of Zn 2+ , which is also secreted from the - cells, insulin and amylin can be examined. 5.2.3 Evaluating Stored Erythrocytes used in Transfusion Medicine by 3D - Printed Analytical Devices The results in this work proved the irreversibility of hyperglycemia stored ERYs after 15 days of storage, whic h is potentially caused by the alteration of cell deformability during storage. However, the mechanism by which hyperglycemia impairs ERY deformability remains unknown. ERY deformability is mainly regulated by structural proteins such as actin and spectrin on the inside of the cell membrane 24 . We suspect that hyperglycemic storage changed the configuration of the binding pattern between the proteins. A characterization method that can detect these proteins needs to be developed to study how hyperglycemia affects the se structures. It is also possible that the hyperglycemia causes glycation of certain proteins in ERYs. There are reports showing that the oxygen delivery of stored ERYs is compromised during storage 25 , 26 , which indicates that the hemoglobin may have been 179 affected by the high glucose levels. Some research focusing on this part can help to understand the lesions of stored ERYs from a broader perspective. The response of stored ERYs to - cells needs to be further explored too, especially the effect of transfused ERYs on original ERYs. Results have shown that hyperglycemia stored ERYs may foul preexisting fresh ERYs . If this is true, the mechanism behind it should be elucidated , which can potentially change the game rules of the field of blood banking in a profound way. The i.v. injection device was used to inject stored ERYs into fresh ERYs in this study. However, t his device can also be used for other applications such as drug administration and drug dosing. For example, diabetic ERYs, which release less ATP than ERYs from health y people 27 , can be circulating in the fluidic de vice, while a drug candidate, such as the ensemble of albumin, C - peptide, and Zn 2+ can be injected via the side branch, to study the effect of the drug candidate. 180 REFERENCES 181 REFERENCES 1. Whitesides, G.M . The origins and the future of microfluidics. Nature 442 , 368 - 373 (2006). 2. McDonald, J.C. et al. Fabrication of microfluidic systems in poly(dimethylsiloxane). Electrophoresis 21 , 27 - 40 (2000). 3. McDonald, J.C. & Whitesides, G.M. Poly(dimethylsiloxan e) as a material for fabricating microfluidic devices. Accounts Chem Res 35 , 491 - 499 (2002). 4. Chen, C.P., Wang, Y.M., Lockwood, S.Y. & Spence, D.M. 3D - printed fluidic devices enable quantitative evaluation of blood components in modified storage solutio ns for use in transfusion medicine. Analyst 139 , 3219 - 3226 (2014). 5. Mukhopadhyay, R. Hard - soft microfluidic device bypasses drawbacks of PDMS. Anal Chem 81 , 5108 - 5108 (2009). 6. Volpatti, L.R. & Yetisen, A.K. Commercialization of microfluidic devices. Trends Biotechnol 32 , 347 - 350 (2014). 7. Chin, C.D., Linder, V. & Sia, S.K. Commercialization of microfluidic point - of - care diagnostic devices. Lab Chip 12 , 2118 - 2134 (2012). 8. Coltro, W.K.T., Lunte, S.M. & Carrilho, E. Comparison of the analytical perf ormance of electrophoresis microchannels fabricated in PDMS, glass, and polyester - toner. Electrophoresis 29 , 4928 - 4937 (2008). 9. Gross, B.C., Erkal, J.L., Lockwood, S.Y., Chen, C.P. & Spence, D.M. Evaluation of 3D Printing and Its Potential Impact on Bio technology and the Chemical Sciences. Anal Chem 86 , 3240 - 3253 (2014). 10. Waldbaur, A., Rapp, H., Lange, K. & Rapp, B.E. Let there be chip - towards rapid prototyping of microfluidic devices: one - step manufacturing processes. Anal Methods - Uk 3 , 2681 - 2716 (2 011). 182 11. Lee, K.G. et al. 3D printed modules for integrated microfluidic devices. Rsc Adv 4 , 32876 - 32880 (2014). 12. Kitson, P.J., Rosnes, M.H., Sans, V., Dragone, V. & Cronin, L. Configurable 3D - Printed millifluidic and microfluidic 'lab on a chip' rea ctionware devices. Lab Chip 12 , 3267 - 3271 (2012). 13. Liu, Y.L., Chen, C.P., Summers, S., Medawala, W. & Spence, D.M. C - peptide and zinc delivery to erythrocytes requires the presence of albumin: implications in diabetes explored with a 3D - printed fluidic device. Integr Biol - Uk 7 , 534 - 543 (2015). 14. Dietrich, H.H., Ellsworth, M.L., Sprague, R.S. & Dacey, R.G. Red blood cell regulation of microvascular tone through adenosine triphosphate. Am J Physiol - Heart C 278 , H1294 - H1298 (2000). 15. Zallen, G. et al . Age of transfused blood is an independent risk factor for postinjury multiple organ failure. Am J Surg 178 , 570 - 572 (1999). 16. Purdy, F.R., Tweeddale, M.G. & Merrick, P.M. Association of mortality with age of blood transfused in septic ICU patients. Ca n J Anaesth 44 , 1256 - 1261 (1997). 17. Basran, S. et al. The association between duration of storage of transfused red blood cells and morbidity and mortality after reoperative cardiac surgery. (vol 103, pg 15, 2006). Anesth Analg 108 , 1953 - 1953 (2009). 1 8. Koch, C.G. et al. Duration of red - cell storage and complications after cardiac surgery. New Engl J Med 358 , 1229 - 1239 (2008). 19. Offner, P.J., Moore, E.E., Biffl, W.L., Johnson, J.L. & Silliman, C.C. Increased rate of infection associated with transfu sion of old blood after severe injury. Arch Surg - Chicago 137 , 711 - 716 (2002). 20. Sprague, R.S., Stephenson, A.H., Bowles, E.A., Stumpf, M.S. & Lonigro, A.J. Reduced expression of G(i) in erythrocytes of huma ns with type 2 diabetes is associated with impairment of both cAMP generation and ATP release. Diabetes 55 , 3588 - 3593 (2006). 183 21. Sprague, R.S., Ellsworth, M.L., Stephenson, A.H., Kleinhenz, M.E. & Lonigro, A.J. Deformation - induced ATP release from red blo od cells requires CFTR activity. Am J Physiol - Heart C 275 , H1726 - H1732 (1998). 22. Gross, B.A., Storey, A., Orbach, D.B., Scott, R.M. & Smith, E.R. Microsurgical treatment of arteriovenous malformations in pediatric patients: the Boston Children's Hospita l experience. J Neurosurg - Pediatr 15 , 71 - 77 (2015). 23. Ning, W.J., Wijeratne, S., Dong, J.L. & Bruening, M.L. Immobilization of Carboxymethylated Polyethylenimine - Metal - Ion Complexes in Porous Membranes to Selectively Capture His - Tagged Protein. Acs Appl Mater Inter 7 , 2575 - 2584 (2015). 24. Gokhin, D.S. et al. Dynamic actin filaments control the mechanical behavior of the human red blood cell membrane. Mol Biol Cell 25 (2014). 25. Chaplin, H., Beutler, E., Collins, J.A., Giblett, E.R. & Polesky, H.F. Cu rrent Status of Red - Cell Preservation and Availability in Relation to Developing National Blood Policy. New Engl J Med 291 , 68 - 74 (1974). 26. Dawson, R.B. Hemoglobin Function in Stored - Blood .19. Inosine Maintenance of 2,3 - Dpg for 35 Days in a Cpd - Adenine Preservative. Transfusion 17 , 525 - 528 (1977). 27. Carroll, J. et al. An altered oxidant defense system in red blood cells affects their ability to release nitric oxide - stimulating ATP. Mol Biosyst 2 , 305 - 311 (2006).