GUT MICROBIOME ANALYSIS IN DOGS WITH LYMPHOMA UNDERGOING CHOP PROTOCOL: A CORRELATIVE ANALYSIS By Keita Kitagawa A THESIS Submitted to Michigan State University in partial fulfillment of the requirements for the degree of Comparative Medicine and Integrative Biology—Master of Science 2022 ABSTRACT Canine multicentric large cell lymphoma shares many similarities to Non-Hodgkin’s Lymphoma in humans. Both human and canine large cell lymphoma require chemotherapy and chemotherapeutic toxicities can result in limiting factors for the compliance of the treatment and the better outcomes. The gut microbiome is the assembly of genomes of the microorganisms in the GI tract. The gut microbiome is compositionally changed by many factors, such as GI diseases, diet, and chemotherapy administration; the clinical significance of the changes remains unclear. Therefore, we sought to describe the change in gut microbiome in a clinically well-characterized population. Also, we tried to explore the correlations between the changes in the gut microbiome and chemotherapeutic toxicities. Twenty dogs were included. In this study, 32 GI toxicities and 42 neutropenia events were identified, but there was no correlation between the relative abundance of the gut microbiome and chemotherapy toxicities. We observed a dynamic compositional change in the gut microbiome over the first 10 weeks of the CHOP protocol. The relative abundance of Lachnospiraceae in the GI toxicity (P=0.0205) and Both (P=0.0089) groups significantly decreased and the relative abundance of Fusobacterium.uncultured significantly decreased (P=0.0197) in the Both group, compared to the No toxicity group. Further data analysis of the compositional change in the gut microbiome during chemotherapy is needed. Copyright by KEITA KITAGAWA 2022 ACKNOWLEDGMENTS First and foremost, I would like to express sincere gratitude to my primary advisor, Dr. Paulo Vilar Saavedra for his encouragement, constructive feedback, and continuous support and patience throughout my master’s degree despite the pandemic of COVID-19. His solid scientific expertise, curiosity for translational medicine, and acceptance of new ideas have trained me for the right goal; it is a privilege that I could receive his guidance. My gratitude extends to the rest of my committee, Dr. Linda Mansfield, Dr. Nikolaos Dervisis, and Dr. Vilma Yuzbasiyan-Gurkan who dedicated their time to giving me their insights and guidance with their expertise. I am grateful to the former and current members of the Medical Oncology Service of Michigan State University Veterinary Teaching Hospital (i.e., former senior clinician – Dr. Honkisz –, interns, students who were rotating into our service, and veterinary technicians), the members of the Mansfield lab, other graduate students and interns/residents, and faculty and staff at the Department of Comparative Medicine and Integrated Biology for their support, collaborations, and friendships. In addition, I would like to thank my former supervisors and colleagues at Animal Medical Center, Gifu University. Especially, Dr. Takashi Mori, a professor of Gifu University, and the head of Veterinary Oncology Service, who contributed to my personal and professional development as a veterinary clinician specializing in veterinary oncology and triggered me to pursue an advanced degree in the field. Last not but least, I would also like to thank my family – my parents and grandparents, who supported me and believed in me throughout the program. I have been very fortunate to pursue my MS degree at Michigan State University, and I could not have done this work without all the people mentioned above in this COVID era. iv TABLE OF CONTENTS LIST OF ABBREVIATIONS ........................................................................................................ vi INTRODUCTION ........................................................................................................................... 1 Non-Hodgkin Lymphoma in humans ...................................................................................... 1 Canine multicentric large cell lymphoma ................................................................................ 2 The association of the gut microbiome multiple morbidities in humans ................................. 4 Characteristics of the intestinal microbiome in dogs ............................................................... 6 Methodology for the analysis of the intestinal microbiome..................................................... 7 Knowledge gap and the purpose of the study .......................................................................... 8 MATERIALS AND METHODS .................................................................................................... 9 Case selections ........................................................................................................................ 9 CHOP based protocol............................................................................................................... 9 Clinical data assessment......................................................................................................... 10 Fecal sample collection .......................................................................................................... 10 DNA extraction from fecal samples....................................................................................... 11 16S rRNA gene V4 region PCR analysis............................................................................... 12 Illumina 16S rRNA gene sequencing and data analysis ........................................................ 12 Correlation analysis................................................................................................................ 13 Statistical analyses ................................................................................................................. 13 RESULTS ...................................................................................................................................... 15 Study population .................................................................................................................... 15 Chemotherapy dose modulation............................................................................................. 16 GI toxicity and bone marrow suppression ............................................................................. 16 Intestinal microbiome profile ................................................................................................. 17 Correlation analysis................................................................................................................ 17 DISCUSSION................................................................................................................................ 19 CONCLUSIONS ........................................................................................................................... 23 Future directions ................................................................................................................ 23 ACKNOWLEDGEMWNTS ......................................................................................................... 24 REFERENCES .............................................................................................................................. 25 APPENDIX A: TABLES AND FIGURES ................................................................................... 31 APPENDIX B: SUPPLEMENTAL TABLES AND FIGURES ................................................... 53 v LIST OF ABBREVIATIONS CBC Complete blood cell count CBD Cannabidiol CTX Cytoxan: cyclophosphamide dH2O Distilled water DLBCL Diffuse large B cell lymphoma DMAC Dexamethasone, melphalan, actinomycin-D, cytosine arabinoside protocol DNA Deoxyribonucleic acid DNTP Deoxyribose nucleotide triphosphate DSS Dextran sodium sulfate DXR Doxorubicin (i.e., hydroxydaunorubicin, Adriamycin) GI Gastrointestinal Hct Hematocrit IBD Inflammatory bowel disease IM Intact male IV Intravenous LOPP Lomustine, vincristine, procarbazine, prednisone protocol MDR1 Multi-Drug Resistance 1 MgCl2 Magnesium chloride MOPP Mechlorethamine, vincristine, procarbazine, prednisone protocol MSU Michigan State University NF-κB Nuclear factor κB NHL Non-Hodgkin’s Lymphoma vi NM Neutered male OUT(s) Operational Taxonomic Unit(s) PBS Phosphate buffered saline PCR Polymerase Chain Reaction PE Physical examination PO Per os PTCL Peripheral T-cell lymphoma rRNA Ribosomal ribonucleic acid SEER Surveillance, Epidemiology, and End Results Program SF Spayed female VCOG-CTCAE Veterinary cooperative oncology group – common terminology criteria for adverse events VCR Vincristine WHO World Health Organization vii INTRODUCTION Canine and human lymphoma share a significant number of similarities. A review of human and canine lymphoma is provided below to set the stage for the goals and significance of this research endeavor for both species. Non-Hodgkin Lymphoma in humans Non-Hodgkin’s Lymphoma (NHL) is a lymphoid neoplasm that is derived from B cell, T cell, or their precursors1. According to the Surveillance, Epidemiology, and End Results Program (SEER), the annual incidence rate is approximately 20 per 100,000 and the mortality rate is up to 5 per 100,000 in the United States2. There are various subtypes of NHL and each subtype has many distinct characteristics: clinical features, epidemiology, etiology, genetic immunophenotype, and response to therapy1. Based on the histopathological characteristics, Armitage and Weisenburger classified NHL into thirteen subtypes3. Among these, diffuse large B cell lymphoma (DLBCL) and peripheral T-cell lymphoma (PTCL) are reported to be the most frequently diagnosed with B cells comprising 31% and T cell 6% of the total3. The clinical signs of NHL include fever, weight loss, or night sweats (as known as B- symptoms), and more than two-thirds of the patients have painless peripheral lymphadenomegaly1. However, the clinical presentation of NHL can vary, depending on the anatomical site that is affected, the subtype, and/or the presence or absence of B-symptoms4. Complete blood cell count (CBC), chemistry profile, imaging diagnostics (e.g., CT scan), and tissue biopsy are commonly needed as part of the diagnostic workup of the patient1. The pathologic diagnosis and classification of lymphoma are made following the WHO classification system for lymphoid neoplasms5. Malignant lymphomatous cells grow diffusely with a high mitotic activity, destroy the normal architecture of lymph nodes, and sometimes have an interfollicular or intrasinusoidal growth 1 pattern6. The Ann Arbor staging classification system, which was originally used for Hodgkin's lymphoma staging, has been widely used to stage NHL (Table 1.1)7. The current standard of care for patients with DLBCL is a combination therapy of rituximab, cyclophosphamide (CTX), and doxorubicin (DXR, generic name: hydroxydaunorubicin), vincristine (VCR, also known as Oncovin), and prednisone (R-CHOP) regimen8. Using this protocol, about 60–70% of patients with DLBCL are cured but approximately 25-30% of them relapse within 3 years after the R-CHOP regimen8,9. On the other hand, PTCL is less frequent (6% of NHL cases) and has a poor prognosis than DLBCL. Greer et al. reported that among patients with PTCL who underwent various combination chemotherapy (e.g., CHOP, BCOP or COMLA); only 24% of the patients achieved a complete remission and the median survival was 11 months10. Canine multicentric large cell lymphoma Canine lymphoma is the most common hematopoietic malignancy diagnosed in dogs, which consists of 7 to 24% of all canine neoplasms reported in the veterinary pathology database11. Several comparative studies have used canine lymphoma as a spontaneous tumor-bearing animal model of non-Hodgkin’s lymphoma in humans because canine lymphoma shares many significant similarities, including histopathologic classification, diagnostic work-up, and treatment11,12. Typical histopathological images of DLBCL and PTCL both in humans and in dogs are summarized in Figure 1.1. The most common clinical signs include non-painful generalized lymphadenomegaly and non-specific clinical signs, such as vomiting, diarrhea, hypo- or anorexia, and lethargy11. The diagnosis of canine multicentric lymphoma is made comprehensively, including physical examination, CBC, chemistry profile, cytology or tissue biopsy, and either PCR or flow cytometry is used for the immunophenotyping to determine B or T cell origin11. 2 Multiagent chemotherapy is used as the current standard of care for canine lymphoma, and the treatment goal is to maximize anticancer effect and minimize adverse effects of chemotherapy. CHOP protocol ( CTX, DOX, VCR, and prednisolone) with or without L-asparaginase, is reported as the most effective treatment regimen against multicentric, large cell lymphoma in dogs11. The complete response rates in canine multicentric B cell and T cell lymphoma, are approximately 80- 90% and 39-88%, respectively11,13-18. The median disease-free interval is approximately 8-13 months for B cell lymphoma11,13-15. Simon et al. reported that the overall response rate of CHOP based protocol was 89%; median 1st remission duration, which corresponds to progressive free interval, in dogs with complete remission was 243 days (range: 19–1,191 days, 95% confidence interval: 199–287 days), and 85% of the dogs that had achieved complete remission relapsed15. About 60% of the dogs that relapsed after being treated with CHOP protocol, were treated with either the same CHOP based protocol or other rescue protocols15. The overall response rate in these relapsed dogs was 79% and the median duration of the second remission was 130 days (range: 17–606 days, 95% confidence interval:76–184 days) in dogs that achieve the complete response15. In the three dogs that achieve the partial response, the duration of the second remission lasted 39, 182, and 486 days, respectively15. When compared to B cell lymphoma, large T cell lymphoma has a poorer prognosis, with a disease-free interval of 52-200 days, and a median survival time of approximately 5-7 months15,16,19-22. When relapse occurs, various rescue protocols (e.g., MOPP, LOPP, DMAC, and a single agent rabacfosadine) have been reported (Table 1.2). However, there are no standardized guidelines and recommendations on protocols to follow, and which rescue protocol is used depends on the clinicians’ preference. CHOP based protocol consists of classic chemotherapeutic agents that are characterized by nonspecific cytotoxic activity23. Since the classic chemotherapeutic agents target rapidly dividing 3 cells, the toxicity (chemotherapy-induced toxicity) occurs frequently as gastrointestinal (GI) toxicity (e.g., vomiting, diarrhea, and hypoxia) and bone marrow suppression (e.g., neutropenia)23. Other than these, liver and kidney damage are also reported as chemotherapy-induced toxicity23. Chemotherapy-induced toxicities can be challenging for the management of multicentric large cell lymphoma. The occurrence of chemotherapy-induced toxicities at any grade is high in dogs with multicentric large cell lymphoma; approximately 50-70% of dogs have experienced at least one toxicity14,15,19. Specifically, GI toxicity (e.g., vomiting) and bone marrow suppression (e.g., neutropenia) are frequently found at approximately 40-50% and 60%, respectively, with differing severities14,15. Toxicity levels are graded according to the criteria outlined (Table 1.3). Grade 3 or greater toxicities (either GI toxicity or neutropenia) occurs at approximately up to 30%. Moreover, dose modulation is required in approximately 40-67% of dogs with multicentric, large cell lymphoma due to chemotherapy-induced toxicities 13,19,24. Approximately 10% of them require hospitalization because of either febrile neutropenia or severe GI toxicity13-15,24. The association of the gut microbiome with multiple morbidities in humans Gut microbiota, which is composed of commensal microorganisms (e.g., bacteria, viruses, and fungi), has a stable and highly diverse ecosystem among individuals25-27. The gut microbiome is the assembly of genomes of the microorganisms that live in the GI tract28. The number of publications referring to the gut microbiome in humans has been explosively increasing since it emerged (Figure 1.2, A). Changes in the gut microbiome have been associated with many diseases (e.g., Inflammatory bowel disease -IBD-, obesity, and cancer)29. Dysbiosis is defined as the compositional imbalance in the gut microbiome that relates to a pathologic state distinct from a healthy state30. The gut microbiome diversity has two major components: richness and evenness31. 4 Richness is the number of phylotypes/taxa in the community, and evenness explains the difference in the relative abundance of species in the community31,32. For example, in IBD, Frank and colleagues reported that there were two distinct subsets (i.e., IBD subset and Control subset)33. The IBD subset predominantly consists of IBD, ulcerative colitis, and Crohn's disease, while the Control subset consists of non-IBD and that sequences representative of the Bacteroidetes and Lachnospiraceae were significantly depleted in IBD subset, and those of the Actinobacteria and Proteobacteria is more abundant in the IBD subset samples than the Control subset33. In obesity, Ley et al. used C57BL/6 mice and investigated the change of gut microbiome diversity by 16S rRNA analysis34. showed that the cecal microbiome in obese mice had a statistically significant decrease of Bacteroidetes, and a significantly higher concentration of Firmicutes, compared to lean mice34. In studies of carcinogenesis of colorectal cancer in C57BL/6 mice, Zackular et al. reported that gut inflammation leads to the development of colorectal cancer35. Using dextran sodium sulfate (DSS) administration to induce inflammation, they showed that a significant decrease in the diversity occurs in the gut microbiome after the first round of DSS administration inducing dysbiosis and resulting in developing more colorectal cancer than the control group with a healthy gut microbiome35. Moreover, gut microbiome is of key relevance in chemotherapy; various interactions between chemotherapy drugs and the gut microbiota relating to the efficacy, toxicity, and metabolism of the drugs have been reported. Those interactions resulted in variations of response to treatment, severity and or frequency of the chemotherapy-induced toxicities26,36,37. Chemotherapy causes an imbalance of the gut microbiome both in function and in composition, leading to or exacerbating gut inflammation38. For example, CTX, a cytotoxic chemotherapy drug frequently used to treat human and canine hematopoietic malignancies 5 including NHL, disrupts the intestinal barrier by shortening the villi, and increasing intestinal permeability by loosening tight junction between enterocytes, and inducing inflammation and accumulation of mononuclear cells in the lamina propria39,40. The damaged intestinal mucosa sometimes referred to as a “leaky gut”, facilitates the bacterial translocation into the mesenteric lymph nodes and the spleen40.This translocation is selective of Gram-positive bacterial species and secondary to the dysbiosis caused by CTX administration40. Many clinical studies also have shown the importance of the compositional change in the gut microbiome during chemotherapy. Alexander et al. reported that the gut microbiome interacts with many chemotherapeutic agents in various mechanisms: translocation, immunomodulation, metabolism, enzymatic degradation, and reduced diversity26. Galloway-Peña et al. attempted to predict the risk of the infection during chemotherapy by the gut microbial community profiling41. In addition, Montassier et al. reported that chemotherapy has been shown to decrease bacterial diversity, richness, and metabolic functions and that severe dysbiosis in the gut microbiome by chemotherapy is associated with chemotherapy-induced GI toxicities38. Characteristics of the intestinal microbiome in dogs In veterinary medicine, like in humans, the publications on the gut microbiome have been an increasing trend since 2009 (Figure 1.2, B). In healthy dogs, the composition of the gut microbiome is relatively stable within the individual but more unstable among different individuals (interindividual) 27. The intestinal microbiome is altered in composition by several factors, such as diet, antibiotics, probiotics, and co-morbidities. Raw food diets alter the abundance of Lactobacillales, Enterobacteriaceae, Enterococcusona, Fusobacterium, and Clostridium at different taxonomic levels42,43. Also, raw food diets significantly increase the abundance of E.coli and Streptococcus and decreased the abundance of Faecalibacterium in the fecal microbiome43. 6 In some GI diseases (e.g., chronic inflammatory enteropathy), dysbiosis is often observed. Suchodolski et al. investigated the canine gut microbiome in dogs with diarrhea and reported that diseased dogs tended to have a decreased richness and diversity of gut microbiome in the fecal sample44. They also found dysbiosis in dogs with acute and chronic diarrhea, suggesting E. coli, Isospora, Giardia/Cryptosporidium, enterotoxigenic C. perfringens, and toxigenic C. difficile as potential pathogenic bacteria44. Gavazza et al. reported that dysbiosis was found in dogs with multicentric lymphoma with decreased abundance of Faecalibacterium, Fusobacterium, and Turicibacter45. In addition, dogs that received chemotherapy have been reported to have significantly increased pathogenic bacteria (i.e., E.coli and Streptococcus) in the gut microbiome at 8 weeks after the start of chemotherapy in a small study of 12 lymphoma dogs, compared to healthy dogs46. The dogs that have chemotherapy-induced toxicity, such as GI toxicity, might have a specific compositional change in the gut microbiome and different toxicities may cause different signatures of the gut microbiome profile. Therefore, profiling the gut microbiome is crucial for the better understanding, prediction, and management of chemotherapy-induced toxicity. Methodology for the analysis of the intestinal microbiome Many methodologies can be used to analyze the gut microbiome, including amplicon sequencing, shotgun sequencing (i.e., metagenomics), and metabolomics47. Amplicon analysis utilizes 16S rRNA sequencing and has been most commonly used in the past 15 years48. The differences between amplicon sequencing and shotgun sequencing are summarized by Allaband et al. and are presented in Table1.447. The16S rRNA amplicon analysis can detect one or more of 9 hypervariable regions (V1–V9) that have sequence diversity in otherwise highly conserved 16S rRNA gene. Using primers that are targeted to the highly conserved regions flanking the variable regions, amplicons are obtained from a wide range of bacterial targets, and sequencing of the 7 amplicons can reveal nearly all bacterial taxa present. While the amplicon sequencing approach has been employed to detect one specific gene (16S rRNA gene for archaea), the shotgun sequencing can sequence all DNA fragments from a sample and then integrate and analyze these fragments, revealing in greater detail the bacterial community47. Thus, amplicon sequencing reveals which bacterial taxa are in a sample and the relative abundance of the bacteria in a sample, while shotgun sequencing reveals all genes that are coded by the bacteria in a sample. Knowledge gap and the purpose of the study The overarching hypothesis is chemotherapeutic agents alter the gut microbiota, linking to diarrhea and/or neutropenia, and the stability of the gut microbiome reflects the damage produced by chemotherapeutic agents. Although some studies have already shown an association between chemotherapy-induced toxicity, including GI toxicities and bone marrow suppression (e.g., neutropenia) and changes in gut microbiome resulting in dysbiosis in both humans and dogs, limited information is yet available on the canine microbiome, especially as it changes with chemotherapy. Therefore, the primary purpose of this study is to characterize the baseline gut microbiome in the dog with multicentric lymphoma and to describe the longitudinal changes in the gut microbiome induced by chemotherapy in dogs with large cell lymphoma undergoing CHOP protocol. The secondary purpose is to investigate if there is any correlation between the change in the gut microbiome and clinical and laboratory parameters (e.g., grade of chemotherapy-induced toxicity, neutrophil count, and percentile change of neutrophil count). The tertiary purpose is to identify which chemotherapeutic agent would cause the most severe dysbiosis in canine lymphoma. 8 MATERIALS AND METHODS Case selections Cases were enrolled between 2018 to 2021 at Michigan State University (MSU) Veterinary Teaching Hospital (MSU-VTH) with informed client consent under the IBR (VTH) approval for sample collection and completion of the study. Client-owned dogs with lymphoma were included in the study based on the following inclusion criteria:1) dogs confirmed with large cell lymphoma based on cytology or histopathology, 2) fecal samples were to be collected and available for at least 7 out of the first 10 weeks of CHOP protocol (Table S1.1)13, 3) medical records were available for review and data analysis, and 4) no previous cytotoxic chemotherapy were administered in the past three months. The use of other concomitant medication (e.g., antibiotics, corticosteroids, and probiotics) at the registration for this study were accepted and recorded for analysis. CHOP based protocol The first 10 weeks of CHOP are delivered according the previous report13. Briefly, VCR was administered intravenously at 0.5-0.7 mg/m2 on Week 1, 3, 6 and 8, CTX was administered orally at 250-300 mg/m2 on Week 2 and 7, DXR was administered intravenously either at 30 mg/m2 if the dog is > 15kg or at 1mg/m2 if the dog is < 15kg. Prednisone was prescribed at 2 mg/kg daily and tapering off over 4 weeks and the discontinuation was decided by the attending clinician. Regarding concomitant medications, owners followed the instruction of clinicians when moderate to severe chemotherapy-induced toxicity occurs. Antibiotics, metronidazole, and/or antinausea medications, such as maropitant, were used to manage GI toxicity. For neutropenia, antibiotics were also used based on clinicians’ judgment. Anti-anxiety medications were also used for the safe administration of chemotherapy, depending on dogs. Supplements were allowed to use 9 unless clinicians prohibited them. Clinical data assessment For clinical assessment, patient characteristics collected included: breed, gender, age, stage and substage of lymphoma at initial presentation, immunophenotype, body condition score, the presence or absence of cardiac murmur, the presence or absence of chemotherapy-induced toxicities (i.e., diarrhea, vomiting, anorexia, anemia, neutropenia, and/or thrombocytopenia), any other concomitant medication, serum chemistry profile and chemotherapeutic agent administered. Staging diagnostics consisted of history, palpation, and cytology of blood smear, affected lymph node, and/or the spleen and liver. At initial presentation, the clinical stage was determined by the attending clinician following the WHO staging system (Table 1.1) 49. The chemotherapeutic agent that was administered at last hospital visit was considered a “contributor” when developing chemotherapy-induced toxicity. The interval between chemotherapies, the percentile changes in white blood cell count, the percentile changes in neutrophil count, the percentile changes in platelet count, and the percentile change in the hematocrit were calculated. Chemotherapy-induced toxicities were evaluated and graded according to the VCOG – CTCAE v1.150. All clinical data were reviewed in both electronic and paper-based form. Fecal sample collection Fecal samples were collected from the enrolled cases between 2018 and 2021 at every hospital visit on a weekly basis by attending clinicians and/or KK. The rectal examination, as part of the routine physical examination, was done in a minimally invasive manner, wearing examination gloves, using lubricant, and gently restraining the dog. A fecal sample was collected before chemotherapy administration during each hospital visit. To avoid bacterial contamination of the fecal samples from the hospital environment, fecal samples collected were transferred 10 promptly into a 2ml sterile, cryogenic vial. Then, the samples were immediately frozen at -20 °C, and then, transported to the laboratory after being placed on ice packs in a Styrofoam box, stored at -80 °C, and frozen until further microbiome analysis was done. DNA extraction from fecal samples Fecal DNA was extracted from all fecal samples by using the FastDNATM SPIN Kit for Soil (Mp Biomedicals LLC, California, USA) (Figure S1.1) following the kit manufacturer’s instructions. In brief, 0.5 grams of each defrosted fecal sample, 978 μl of phosphate buffered saline (PBS), and 122 μl of MT buffer were added to a purple top tube provided by the kit, and the fecal samples were homogenized for 40 seconds at a speed setting of 6.0. Once the homogenization is done, the purple tube was centrifuged at 14,000g for 5 minutes. Then, the supernatant was transferred to a microcentrifuge tube, and 250 μl of Protein Precipitation Solution (PPS) was added and mixed well by inverting the microcentrifuge tube 10 times. After mixing the supernatant and PPS, the microcentrifuge tube was centrifuged again at 14,000g for 5 minutes. After the centrifuging, the supernatant was transferred to a clean 15ml tube. Then, 1ml of nucleic acid binding buffer was added to the 15 ml tube, and the supernatant and this binding buffer were mixed well by inverting the 15ml tube. After the reaction with the nucleic acid binding buffer, the supernatant of this mixture was discarded and the residue (i.e., DNA containing solution) was transferred to another tube with a spin column. The spin column was centrifuged at 14,000g for 1 minute. After the centrifuging, 500 μl of nucleic acid wash solution was added to wash away impurities, and then, the spin column was centrifuged at 14,000g minute. After that, the spin column was air dried for 5 minutes at room temperature. Then, DNA was eluted by adding 100 μl of elution buffer followed by centrifugation at 14,000 g for 1 min. Once DNA extraction was done, DNA concentration and purity were measured by NanoDrop (Thermo Fisher Scientific Inc., 11 Waltham, Massachusetts, USA) and Qubit fluorometer (Thermo Fisher Scientific Inc., Waltham, Massachusetts, USA). The extracted DNA was stored at -80 °C prior to the amplification steps (e.g., PCR sequencing). 16S rRNA gene V4 region PCR analysis Subsequently, 10 μl of extracted DNA and the primer set (Forward Primer = 5′ GTGCCAGCMGCCGCGGTAA; Reverse Primer = 5′TTAATCTWTGGGVHCATCAGG ) were used to amplify the V4 region of the bacterial 16S rRNA by PCR51. This amplification was done for each fecal sample using the Master Mix. A total of 50 μl of reactants containing: 2.5 μl of forward and reverse primer, 5 μl of 10X PCR buffer, 1.5 μl of 50mM MgCl2, 4 μl of 2.5 mM DNTPs, 0.5 μl of Taq polymerase, 26.5 μl of dH2O, and 10 μl of the extracted DNA. The amplification process was done at 94 ℃ for 3 minutes. This was followed by 35 cycles of 94℃ for 45 seconds, 50 ℃ for 60 seconds and 72 ℃ for 90 seconds and cooling at 72 ℃ for 10 minutes. Sterile-filtered, PCR grade water (Water – PCR reagent, Sigma-Aldrich, St. Louis, MO, USA) was used as negative template controls. DNA concentration was adjusted to ensure samples were approximately equal in concentration. Final concentrations were calculated using Qubit. Illumina MiSeq Amplicon sequencing was done by MSU Research and Technology Support Facility. Illumina 16S rRNA gene sequencing and data analysis 16S rRNA gene amplicon analysis was performed using QIIME2 (v. 2019.1) and protocols available at reference or URL SOP https://docs.qiime2.org/2019.1/ accessed May 2019. Alignment was accomplished using the Silva 16S ribosomal gene database52. Chimeric sequences and any sequences classified as chloroplast, mitochondria, Archaea, or Eukaryota, were removed from the dataset using UCHIME. Sequences were clustered in Operational Taxonomic Units (OTUs) of 97% sequence identity yielding 79 OTUs. Analyses were performed 12 in PAST 3.0753 and R v.4.0654. Following processing of sequences and chimera removal in QIIME2. In the sequence clean-up process, forward and reverse primers, reads with ambiguous bases or homopolymers greater than eight base pairs and chimeras were removed in QIIME2. As a result, high-quality reads remained. Sequence read data has been made available in the National Center for Biotechnology Information (NCBI) Sequence Read Archive (SRA) as documented in “Availability of data and materials”. Correlation analysis The heatmap was generated on the relative abundance of 79 OTUs yielded to see if there is any difference in the gut microbiome between the dogs with and without chemotherapy- induced toxicity. Among the 79 OTUs, Lachnospiraceae, Bacteroides.uncultured, Prevotella.uncultured, Fusobacterium.uncultured., Streptococcus.uncultured, and Enterococcus were selected because they had a dynamic change in the relative abundance analysis. Then, the 6 selected OTUs were used for the correlation analysis. In the correlation analysis, regarding the chemotoxicity parameters, we categorized into four groups according to the chemotherapy toxicity status: No toxicity, GI toxicity, Neutropenia, and Both. No toxicity means the dogs didn't have either GI toxicity or neutropenia, GI toxicity means the dogs had GI toxicity but not neutropenia, Neutropenia means the dogs had chemotherapy induced neutropenia but not GI toxicity, and Both means the dogs had both GI toxicity and neutropenia. To determine whether there is a statistically significant difference in the relative abundance of the selected taxa, the relative abundance of each category was compared. Statistical analyses Kruskal-Wallis test was used for the Shannon Diversity Index. Principal Component 13 Analysis (PCA) was done using the Bray-Curtis test. The Pearson correlation coefficient was calculated for the correlation between the relative abundance of selected bacteria species in the gut microbiome and the numerical clinical parameters. The previous statistical analyses were done with the R software package (ver. 4.2.0). Mann-Kendall test was used for trends by using PAST software package (ver.4.03) Brown-Forsythe and Welch's ANOVA test were done followed by Games-Howell's multiple comparisons test to compare the means of the relative abundance of the gut microbiome in the No toxicity group with that of each other group. This ANOVA test and following multiple comparisons were done by using GraphPad Prism 9.0 for Mac OS X (GraphPad Software, La Jolla, California USA). P-value < 0.05 was considered as significant. 14 RESULTS Study population A total of 20 dogs matched our inclusion criteria. The medical records of all study dogs were reviewed. Signalment and pretreatment conditions (i.e., breed, age, sex, the WHO stage and substage, immunophenotype) were summarized in Table 1.5. At presentation, the study dogs had a median age of 7 years (range: 3–14) with 13 castrated males, 1 intact male, and 6 spayed females. The male to female ratio was 2.3:1. Fifteen different breeds were identified with Mixed breed dog (3) as the most common breed, followed by Golden Retrievers (2), Goldendoodles (2), Rottweilers (2), and one each of Boxer, Bullmastiff, Chesapeake Bay Retriever, Coonhound, English Cocker Spaniel, Great Dane, German shepherd dog, Jack Russell Terrier, Labrador Retriever, Miniature Schnauzer, and Siberian Husky. Regarding the staging of lymphoma, most dogs were categorized as stage III or higher. There were none of dogs with stage I, 1 with stage II, 7 with stage III, 5 with stage IV, and 7 with stage V. All dogs received palpation, abdominal ultrasound, and fine needle aspirate of affected lymph node for the diagnosis; not all dogs received the full diagnostic workup for the staging (e.g., fine needle aspirate in the spleen and the liver, or bone marrow aspirate)because of the discretion of the pet owner and the clinician49. There were 9 cases categorized as substage (a), and 11 as substage (b). In regards of cell immunophenotype, there were 13 of B cell lymphoma, 2 of T cell lymphoma, and 5 were not tested. Regarding concurrent therapies, there were 6 dogs that had received corticosteroid prior to chemotherapy. Other concomitant medications at initial presentation were summarized in Table S1.2 Fourteen out of 20 dogs received concurrent corticosteroids as part of the chemotherapy. Of them, 8 dogs discontinued corticosteroids within Week 5 and the remainder of 6 dogs discontinued 15 it within Week 10. Multi Drug Resistance 1 (MDR1) gene, which encodes p-glycoprotein that efflux VCR and prednisolone that are used in CHOP based protocol55, and therefore, it contributes one of the mechanisms of drug resistance11. The mutation of MDR1 gene is a prognostic indicator in canine large cell lymphoma. In this cohort, only one dog tested MDR1 gene mutation, and the result was negative (-/-) for the gene mutation. Chemotherapy dose modulation Thirteen of the 20 dogs (65%) experienced dose modulation (i.e., either dose delay or dose reduction) at any point within Week 10. Of them, 11 dogs had at least one dose delay, 7 dogs had at least one dose reduction, and 5 dogs had both dose delay and reduction. The cause of dose modulation is due to chemotherapy-induced toxicity (neutropenia or GI toxicity). None of the dogs experienced discontinuation of CHOP based protocol in the first 10 week. GI toxicity and bone marrow suppression There was 32 episodes of GI toxicities identified. Of them, three episodes resulted in dose delay. The grade and frequency of the vomiting, diarrhea, and anorexia were summarized in Table 1.6. The frequency of GI toxicity per Week is summarized in Figure 1.3. In addition to these 32 episodes of GI toxicity, one dog was diagnosed with a gray zone between GI toxicity and food intolerance/allergy. This dog had severe GI signs at 8 days after the previous vincristine administration and was given a different dog food the day before, and therefore, the true cause of the GI signs in this dog was undetermined. Regarding chemotherapy-induced neutropenia, 42 episodes were identified. Most of them were low grade and self-limiting (Table 1.7). Of them, 10 episodes caused dose delay and dose delay was done at all grades. 16 Intestinal microbiome profiling Relative abundances were analyzed longitudinally in 8 dogs (Figure 1.4), and there are currently fecal samples from 13 dogs undergoing abundance analysis. Each dog showed a dynamic compositional change in the gut microbiome over the first 10 weeks of CHOP based protocol. The average of relative abundance of the gut microbiome is shown in Figure 1.5. When compared to the dog with vs without GI toxicity, dogs without GI toxicity tended to have a quicker stabilization of the gut microbiome composition than those with GI toxicity (Figure 1.6 and 7). In the Shannon diversity index, there was not statistically significant in each arm grouped by chemotherapeutic agent (Figure 1.8). The dogs with GI toxicity tended to have decreased diversity in the gut microbiome compared to those without GI toxicity (Figure 1.9). Also, the dogs receiving antibiotics have a significant decrease of diversity in the gut microbiome than those without antibiotics (Figure 1.10). When compared to the sample on Week 1, 6 of 10 top bacterial taxa contributing to PCA components 1 and 2 showed significant trends over the entire course of CHOP protocol (Table 1.8). Correlation analysis The heatmap was created to compare the microbiome composition between the dogs with and without GI toxicity, and those with and without neutropenia. There did not appear to be a distinct difference in the gut microbiome that is associated with either GI toxicity or chemotherapeutic neutropenia (Figure 1.11). Next, the relative abundance of selected taxa was paired with the myelosuppression parameters (e.g., neutrophil count) to calculate a correlation coefficient. A total of 58 pairs of observations were available for the correlation analysis. The correlation coefficient was summarized in Table 1.9. There was no strong correlation found between the relative abundance 17 of selected taxa and the numerical clinical parameters. We summarized the comparison of the relative abundance of selected taxa among the four groups that were categorized according to the chemotherapy-induced toxicity status in Figure 1.12. The relative abundance of Lachnospiraceae in GI toxicity (P=0.0205) and Both (P=0.0089) significantly decreased and the relative abundance of Fusobacterium.uncultured significantly decreased (P=0.0197), compared to No toxicity group. There was not any significant difference of the relative abundance of Bacteroides.uncultured, Prevotella.uncultured Streptococcus.uncultured, and Enterococcus when comparing No toxicity group to other categories. Although Streptococcus.uncultured and Enterococcus didn’t show any statistical significance, there was the increasing trend in the chemo toxic groups (i.e., GI toxicity, Neutropenia, and Both). 18 DISCUSSION In this study, we documented a dynamic compositional change of the gut microbiome over the first 10 weeks of CHOP based protocol. Each dog had a different diverse composition in the gut microbiome. Our study showed that interindividual diversity is greater than intraindividual diversity even in the state of lymphoma. This is in agreement with other reports where interindividual diversity in healthy dogs was shown to be greater than intraindividual diversity in the gut microbiome27 When compared to the dog with vs without GI toxicity, dogs without GI toxicity had a quicker stabilization of the gut microbiome composition than those with GI toxicity. Chemotherapy can cause taxonomic shift and increases the pathogenic bacteria such as E. coli and Streptococcus38,46. The relative abundance of Bacteroides nordii, Ruminococcus sp, and Gardnerella vaginalis are associated with severe toxicity of CTX56. Therefore, we expected that the relative abundance of pathogenic bacteria would increase compared to the baseline, and the compositional change in Bacteroides, Ruminococcus, and/or Gardnerella should be found in those dogs with severe toxicity of CTX in this cohort, too. Also, given that dogs with GI toxicity needed more time to stabilize the gut microbiome composition. This persistent destabilization in the gut microbiome may correlate with treatment response and/or prognosis. To test this, further studies are needed. The dogs receiving antibiotics have a significant decrease in Shannon diversity index than those without antibiotics. Gut inflammation promotes Enterobacteriaceae and the combination of insult by enterobacteria and the gut inflammation reduces the richness of bacteria57. Also, in the study of human colorectal cancer, Fei et al. reported that decreased diversity in patients with GI toxicity was observed58. They considered that the decrease of microbial diversity may be related 19 to the imbalance of gut microbiome because the dominant pathogenic bacteria consume nutrients or produce bacterial toxic metabolites58. Clinically, the use of antibiotics at hospital visits suggests that the dogs receiving antibiotics had a recent moderate to severe GI toxicity and/or chemotherapy-induced neutropenia. Therefore, the decrease of Shannon diversity in antibiotic use may reflect the dogs' state that antibiotics were needed because of the severe GI toxicity or neutropenia, and like in humans, chemotherapy can reduce the diversity in the canine gut microbiome. Although there was no statistical difference in the Shannon diversity index when comparing dogs with vs without GI toxicity, we noticed a decreasing trend in the diversity. In our study, most of the GI toxicity was low grade and self-limited. The degree of decrease in diversity may be related to the severity of GI toxicity. Another possible reason is that this may be a type 1 error due to the small sample size. Additional 13 dogs will shed light on the relationship between GI toxicity and microbiome diversity. In our study, the relative abundance of Lachnospiraceae significantly decreased in the GI toxicity and Both groups. Based on the OTU assignment used in our study, Lachnospiraceae includes butyrate-producing bacteria. In human medicine, there was a decreased relative abundance of Lachnospiraceae in the IBD group, compared to the control33. Butyrate has an anti- inflammatory effect by the inhibition of nuclear factor κB (NF-κB) activation in human colonic epithelial cells and reinforces the colonic defense barrier59. Thus, these suggest that a decreased relative abundance of Lachnospiraceae is predominantly associated with GI toxicity in our study. The possible mechanism is that the decrease of butyrate producing Lachnospiraceae in the gut microbiome reduces the anti-inflammatory effect by failing to reinforce the gut barrier. This leads to GI inflammation, which results in GI toxicity (mainly diarrhea). 20 Also, there was a significantly decreased relative abundance of Fusobacterium.uncultured in the Both group. Jugan et al. reported that the abundance of Fusobacterium decreased in dogs with lymphoma that underwent CHOP based protocol60. Also, Gavazza et al. reported that the abundance of Fusobacterium decreased in dogs with multicentric lymphoma45. Our finding regarding the relative abundance of Fusobacterium is supportive of these previous reports. However, the relationship between the decreased relative abundance of Fusobacterium and chemotherapy-induced toxicity remains unclear. Moreover, the Both group had a very small sample size, so this may be a type 1 error. Therefore, further studies are needed to unveil the relationship between the relative abundance of Fusobacterium and chemotherapy-induced toxicity using a larger sample size. Interestingly, there was an increasing trend in the relative abundance of Streptococcus.uncultured and Enterococcus in the chemo toxic groups in this study. Alessandra et al. showed that the pathogenic bacteria (e.g., Streptococcus and E. coli) increase after chemotherapy in dogs with lymphoma46. Stringer et al. reported that irinotecan administration to mice increases Enterococcus, and also, van Vliet et al. reported that a multiagent chemotherapy regimen used for pediatric patients with acute myeloid leukemia showed a drastically increased in Enterococcus61,62. These research on the increase of pathogenic bacteria by chemotherapy is consistent with our findings. Therefore, chemotherapy may increase the relative abundance of pathogenic bacterial taxa in canine gut microbiome. Specifically, CHOP protocol used for dogs with multicentric large cell lymphoma may also increase in Streptococcus and Enterococcus. One limitation of the study is the clinical staging of lymphoma. In this study, not all dogs received the full work-up for the staging, including bone marrow aspirate. Although 9 dogs with stage II and III received an abdominal ultrasound, most of them did not receive fine needle aspirate 21 in the liver and spleen. Thus, these cases might be underscored in staging. We failed to collect some fecal samples during this study because dogs had defecated just before the hospital visit or had severe GI toxicity. In these cases, it was impossible to collect relevant fecal samples of interest. Given the study design, the sample in Week 1 serves as a control for every dog included in this study. However, with Dogs 6 and 21, we failed to obtain samples in Week 1. Thus, for these two cases, the baseline gut microbiome status remains unclear. A bigger longitudinal study (sample size) which includes may contribute to the completeness of the data in this type of design that analyze the compositional change of the fecal microbiome over time. 22 CONCLUSIONS In conclusion, there was a dynamic compositional change in the gut microbiome of dogs during the first 10 weeks of the CHOP based protocol. During CHOP protocol, dogs on antibiotics had a significant decrease in diversity in the gut microbiome. Also, there was a significant decrease in relative abundance in some bacterial taxa in relation to chemotherapy-induced toxicity. We haven’t found any statistical significance in the correlation between the relative abundance in the gut microbiome and clinical laboratory parameters. Future directions The current level of analysis on 8 dogs studied longitudinally was not able to provide clear correlations with chemotherapy related toxicities. The microbiome analysis is being carried out in 13 additional dogs. By adding the samples from additional 13 dogs, the specific findings in the gut microbiome that correlate to or be associated with the chemotherapy-induced toxicity in dogs that underwent CHOP protocol would be found. For further studies, the compositional change in the gut microbiome during chemotherapy should be examined for their relationship with prognosis and be explored for use as predictive markers for chemotherapy-induced toxicity. 23 ACKNOWLEDGEMENTS This work was supported by Michigan State University College of Veterinary Medicine Endowed Research Fund 2018 (award $25,000). The author thanks veterinary technicians, veterinary students, and interns working in Medical Oncology Service for the support in fecal sample collections. The author -Keita Kitagawa- also thank Dr. Jason Couto, members and summer research students – Drs. Azam Sher, Eric Spilker, Hinako Terauchi, Nicole Mulready, Julia Bell, and Linda Mansfield – in Dr. Mansfield’s lab for their fundamental work for this project and 16S rRNA sequencing. Their efforts will be clearly recognized as first co-authorship of the upcoming scientific manuscript. 24 REFERENCES 1. Sapkota S SH. Non-Hodgkin Lymphoma. [Updated 2021 Dec 5]. In: StatPearls [Internet]. In: StatPearls Publishing; 2022 Jan. 2. Surveillance E, and End Results (SEER) Program (www.seer.cancer.gov) SEER*Stat Database. Recent Trends in SEER Incidence and U.S. Mortality Rates of Non-Hodgkin’s Lymphoma, 2000-2019. In, Released in April, 2022 ed. 3. Armitage JO, Weisenburger DD. New approach to classifying non-Hodgkin's lymphomas: clinical features of the major histologic subtypes. Non-Hodgkin's Lymphoma Classification Project. J Clin Oncol 1998;16:2780-2795. 4. Shankland KR, Armitage JO, Hancock BW. Non-Hodgkin lymphoma. The Lancet 2012;380:848-857. 5. Swerdlow SH, Campo E, Pileri SA, et al. The 2016 revision of the World Health Organization classification of lymphoid neoplasms. Blood 2016;127:2375-2390. 6. Korkolopoulou P, Vassilakopoulos T, Milionis V, et al. Recent Advances in Aggressive Large B-cell Lymphomas: A Comprehensive Review. Adv Anat Pathol 2016;23:202-243. 7. Carbone PP, Kaplan HS, Musshoff K, et al. Report of the Committee on Hodgkin's Disease Staging Classification. Cancer Res 1971;31:1860-1861. 8. Li S, Young KH, Medeiros LJ. Diffuse large B-cell lymphoma. Pathology 2018;50:74-87. 9. Coiffier B, Thieblemont C, Van Den Neste E, et al. Long-term outcome of patients in the LNH-98.5 trial, the first randomized study comparing rituximab-CHOP to standard CHOP chemotherapy in DLBCL patients: a study by the Groupe d'Etudes des Lymphomes de l'Adulte. Blood 2010;116:2040-2045. 10. Greer JP, York JC, Cousar JB, et al. Peripheral T-cell lymphoma: a clinicopathologic study of 42 cases. J Clin Oncol 1984;2:788-798. 11. Vail DM, Thamm DH, Liptak JM. 33 - Hematopoietic Tumors. In: Vail DM, Thamm DH, Liptak JM, eds. Withrow and MacEwen's Small Animal Clinical Oncology (Sixth Edition). St. Louis (MO): W.B. Saunders; 2019:688-772. 12. Khanna C, Lindblad-Toh K, Vail D, et al. The dog as a cancer model. Nat Biotechnol 2006;24:1065-1066. 13. Garrett LD, Thamm DH, Chun R, et al. Evaluation of a 6-Month Chemotherapy Protocol with No Maintenance Therapy for Dogs with Lymphoma. Journal of Veterinary Internal Medicine 2002;16:704-709. 25 14. Wang S-L, Lee J-J, Liao AT. Comparison of efficacy and toxicity of doxorubicin and mitoxantrone in combination chemotherapy for canine lymphoma. Can Vet J 2016;57:271-276. 15. Simon D, Nolte I, Eberle N, et al. Treatment of dogs with lymphoma using a 12-week, maintenance-free combination chemotherapy protocol. J Vet Intern Med 2006;20:948-954. 16. Rebhun RB, Kent MS, Borrofka SA, et al. CHOP chemotherapy for the treatment of canine multicentric T-cell lymphoma. Vet Comp Oncol 2011;9:38-44. 17. Curran K, Thamm DH. Retrospective analysis for treatment of naïve canine multicentric lymphoma with a 15-week, maintenance-free CHOP protocol. Vet Comp Oncol 2016;14 Suppl 1:147-155. 18. Goodman IH, Moore AS, Frimberger AE. Treatment of canine non-indolent T cell lymphoma using the VELCAP-TSC protocol: A retrospective evaluation of 70 dogs (2003- 2013). Vet J 2016;211:39-44. 19. Sorenmo K, Overley B, Krick E, et al. Outcome and toxicity associated with a dose- intensified, maintenance-free CHOP-based chemotherapy protocol in canine lymphoma: 130 cases. Veterinary and Comparative Oncology 2010;8:196-208. 20. Ruslander DA, Gebhard DH, Tompkins MB, et al. Immunophenotypic characterization of canine lymphoproliferative disorders. In Vivo 1997;11:169-172. 21. Deravi N, Berke O, Woods JP, et al. Specific immunotypes of canine T cell lymphoma are associated with different outcomes. Vet Immunol Immunopathol 2017;191:5-13. 22. Starrak GS, Berry CR, Page RL, et al. Correlation between thoracic radiographic changes and remission/survival duration in 270 dogs with lymphosarcoma. Vet Radiol Ultrasound 1997;38:411-418. 23. Gustafson DL, Bailey DB. 12 - Cancer Chemotherapy. In: Vail DM, Thamm DH, Liptak JM, eds. Withrow and MacEwen's Small Animal Clinical Oncology (Sixth Edition). St. Louis (MO): W.B. Saunders; 2019:182-208. 24. Benjamin SE, Sorenmo KU, Krick EL, et al. Response-based modification of CHOP chemotherapy for canine B-cell lymphoma. Veterinary and Comparative Oncology 2021;19:541- 550. 25. Suchodolski JS. Diagnosis and interpretation of intestinal dysbiosis in dogs and cats. Vet J 2016;215:30-37. 26. Alexander JL, Wilson ID, Teare J, et al. Gut microbiota modulation of chemotherapy efficacy and toxicity. Nat Rev Gastroenterol Hepatol 2017;14:356-365. 26 27. Garcia-Mazcorro JF, Dowd SE, Poulsen J, et al. Abundance and short-term temporal variability of fecal microbiota in healthy dogs. Microbiologyopen 2012;1:340-347. 28. Stefanaki C. Chapter 3 - The Gut Microbiome Beyond the Bacteriome—The Neglected Role of Virome and Mycobiome in Health and Disease. In: Faintuch J, Faintuch S, eds. Microbiome and Metabolome in Diagnosis, Therapy, and other Strategic ApplicationsAcademic Press; 2019:27-32. 29. Guinane CM, Cotter PD. Role of the gut microbiota in health and chronic gastrointestinal disease: understanding a hidden metabolic organ. Therap Adv Gastroenterol 2013;6:295-308. 30. Shreiner AB, Kao JY, Young VB. The gut microbiome in health and in disease. Curr Opin Gastroenterol 2015;31:69-75. 31. Young VB, Schmidt TM. Overview of the gastrointestinal microbiota. Adv Exp Med Biol 2008;635:29-40. 32. Lozupone CA, Stombaugh JI, Gordon JI, et al. Diversity, stability and resilience of the human gut microbiota. Nature 2012;489:220-230. 33. Frank DN, St Amand AL, Feldman RA, et al. Molecular-phylogenetic characterization of microbial community imbalances in human inflammatory bowel diseases. Proc Natl Acad Sci U S A 2007;104:13780-13785. 34. Ley RE, Bäckhed F, Turnbaugh P, et al. Obesity alters gut microbial ecology. Proc Natl Acad Sci U S A 2005;102:11070-11075. 35. Zackular JP, Baxter NT, Iverson KD, et al. The gut microbiome modulates colon tumorigenesis. mBio 2013;4:e00692-00613. 36. Ma W, Mao Q, Xia W, et al. Gut Microbiota Shapes the Efficiency of Cancer Therapy. Frontiers in Microbiology 2019;10. 37. Gori S, Inno A, Belluomini L, et al. Gut microbiota and cancer: How gut microbiota modulates activity, efficacy and toxicity of antitumoral therapy. Critical Reviews in Oncology/Hematology 2019;143:139-147. 38. Montassier E, Gastinne T, Vangay P, et al. Chemotherapy-driven dysbiosis in the intestinal microbiome. Aliment Pharmacol Ther 2015;42:515-528. 39. Keefe DM, Brealey J, Goland GJ, et al. Chemotherapy for cancer causes apoptosis that precedes hypoplasia in crypts of the small intestine in humans. Gut 2000;47:632-637. 40. Viaud S, Saccheri F, Mignot G, et al. The intestinal microbiota modulates the anticancer immune effects of cyclophosphamide. Science 2013;342:971-976. 27 41. Galloway-Peña JR, Shi Y, Peterson CB, et al. Gut Microbiome Signatures Are Predictive of Infectious Risk Following Induction Therapy for Acute Myeloid Leukemia. Clin Infect Dis 2020;71:63-71. 42. Bermingham EN, Maclean P, Thomas DG, et al. Key bacterial families (Clostridiaceae, Erysipelotrichaceae and Bacteroidaceae) are related to the digestion of protein and energy in dogs. PeerJ 2017;5:e3019. 43. Schmidt M, Unterer S, Suchodolski JS, et al. The fecal microbiome and metabolome differs between dogs fed Bones and Raw Food (BARF) diets and dogs fed commercial diets. PLOS ONE 2018;13:e0201279. 44. Suchodolski JS, Markel ME, Garcia-Mazcorro JF, et al. The fecal microbiome in dogs with acute diarrhea and idiopathic inflammatory bowel disease. PLoS One 2012;7:e51907. 45. Gavazza A, Rossi G, Lubas G, et al. Faecal microbiota in dogs with multicentric lymphoma. Vet Comp Oncol 2018;16:E169-e175. 46. Alessandra G, Giacomo R, Lubas G, et al. Fecal microbiota differences in canine lymphoma treated with chemotherapy and probiotics. In: 3° Convegno a cura delle Piattaforme Tematiche di Ateneo su “Alimenti e Nutrizione” e “Salute Umana e Animale” 2018;57-58. 47. Allaband C, McDonald D, Vázquez-Baeza Y, et al. Microbiome 101: Studying, Analyzing, and Interpreting Gut Microbiome Data for Clinicians. Clin Gastroenterol Hepatol 2019;17:218- 230. 48. Tringe SG, Hugenholtz P. A renaissance for the pioneering 16S rRNA gene. Curr Opin Microbiol 2008;11:442-446. 49. Owen LN, World Health Organization. Veterinary Public Health U, Oncology WHOCCfC. TNM Classification of Tumours in Domestic Animals/ edited by L.N. Owen. In. Geneva: World Health Organization; 1980. 50. Veterinary cooperative oncology group - common terminology criteria for adverse events (VCOG-CTCAE) following chemotherapy or biological antineoplastic therapy in dogs and cats v1.1. Vet Comp Oncol 2016;14:417-446. 51. Kozich JJ, Westcott SL, Baxter NT, et al. Development of a dual-index sequencing strategy and curation pipeline for analyzing amplicon sequence data on the MiSeq Illumina sequencing platform. Appl Environ Microbiol 2013;79:5112-5120. 52. Hall M, Beiko RG. 16S rRNA gene analysis with QIIME2. In: Microbiome analysisSpringer; 2018:113-129. 53. Hammer Ø, Harper DA, Ryan PD. PAST: Paleontological statistics software package for education and data analysis. Palaeontologia electronica 2001;4:9. 28 54. (2021) RCT. R: A Language and Environment for Statistical Computing. In: 2021. 55. Mealey KL, Fidel J. P-glycoprotein mediated drug interactions in animals and humans with cancer. Journal of veterinary internal medicine 2015;29:1-6. 56. Oh B, Boyle F, Pavlakis N, et al. Emerging Evidence of the Gut Microbiome in Chemotherapy: A Clinical Review. Frontiers in Oncology 2021;11. 57. Lupp C, Robertson ML, Wickham ME, et al. Host-Mediated Inflammation Disrupts the Intestinal Microbiota and Promotes the Overgrowth of Enterobacteriaceae. Cell Host & Microbe 2007;2:119-129. 58. Fei Z, Lijuan Y, Xi Y, et al. Gut microbiome associated with chemotherapy-induced diarrhea from the CapeOX regimen as adjuvant chemotherapy in resected stage III colorectal cancer. Gut Pathog 2019;11:18-18. 59. Canani RB, Costanzo MD, Leone L, et al. Potential beneficial effects of butyrate in intestinal and extraintestinal diseases. World J Gastroenterol 2011;17:1519-1528. 60. Jugan MC, Wouda RM, Higginbotham ML. Preliminary evaluation of probiotic effects on gastrointestinal signs in dogs with multicentric lymphoma undergoing multi‐agent chemotherapy: A randomised, placebo‐controlled study. Veterinary record open 2021;8:e2. 61. Stringer AM, Gibson RJ, Logan RM, et al. Chemotherapy-induced diarrhea is associated with changes in the luminal environment in the DA rat. Exp Biol Med (Maywood) 2007;232:96- 106. 62. van Vliet MJ, Tissing WJ, Dun CA, et al. Chemotherapy treatment in pediatric patients with acute myeloid leukemia receiving antimicrobial prophylaxis leads to a relative increase of colonization with potentially pathogenic bacteria in the gut. Clin Infect Dis 2009;49:262-270. 63. Rassnick KM, Mauldin GE, Al-Sarraf R, et al. MOPP chemotherapy for treatment of resistant lymphoma in dogs: a retrospective study of 117 cases (1989-2000). J Vet Intern Med 2002;16:576-580. 64. Fahey CE, Milner RJ, Barabas K, et al. Evaluation of the University of Florida lomustine, vincristine, procarbazine, and prednisone chemotherapy protocol for the treatment of relapsed lymphoma in dogs: 33 cases (2003-2009). J Am Vet Med Assoc 2011;239:209-215. 65. Alvarez FJ, Kisseberth WC, Gallant SL, et al. Dexamethasone, melphalan, actinomycin D, cytosine arabinoside (DMAC) protocol for dogs with relapsed lymphoma. J Vet Intern Med 2006;20:1178-1183. 29 66. Parsons-Doherty M, Poirier VJ, Monteith G. The efficacy and adverse event profile of dexamethasone, melphalan, actinomycin D, and cytosine arabinoside (DMAC) chemotherapy in relapsed canine lymphoma. Can Vet J 2014;55:175-180. 67. Saba CF, Vickery KR, Clifford CA, et al. Rabacfosadine for relapsed canine B-cell lymphoma: Efficacy and adverse event profiles of 2 different doses. Vet Comp Oncol 2018;16:E76-e82. 68. Seelig DM, Avery AC, Ehrhart EJ, et al. The Comparative Diagnostic Features of Canine and Human Lymphoma. Vet Sci 2016;3:11. 69. Xie Y, Pittaluga S, Jaffe ES. The histological classification of diffuse large B-cell lymphomas. Semin Hematol 2015;52:57-66. 70. Kato T, Miyoshi H, Kobayashi S, et al. Clinicopathological analysis in PTCL-NOS with CADM1 expression. Virchows Arch 2017;471:659-666. 30 APPENDIX A: TABLES AND FIGURES Table 1.1 The comparison of the staging system in human and canine NHL7,49 Ann Arbor Staging system for human NHL The WHO staging system for canine NHL Principal Definition Stages Definition stages I Involvement of one lymph node or one extra nodal organ or site I Involvement limited to a single node or (IE) lymphoid tissue in a single organ II Involvement of two or more lymph node regions on the same II Involvement of many lymph nodes in a regional side of the diaphragm, or localized involvement of an extra area (+/- tonsils) nodal site or organ (IIE) and one or more lymph node regions on the same side of the diaphragm III Involvement of lymph node regions on both sides of the III Generalized lymph node involvement diaphragm, which might be accompanied by localized involvement of an extra nodal organ or site (IIIE) or spleen (IIIS) or both (IIISE); the spleen is regarded as nodal IV Diffuse or disseminated involvement of one or more distant IV Liver and/or spleen involvement (+/- Stage III) extra nodal organs with or without associated lymph node involvement V Manifestation in the blood and involvement of bone marrow and/or other organ systems (+/- Stages I-IV) Modifiers Definition Substage Definition A Absence of B symptoms (listed below) a Without systemic signs B Temperature >38°C, night sweats, and weight loss of greater b With systemic signs than 10% of bodyweight in the 6 months preceding admission are defined as systemic symptoms 31 Table 1.2 The summary of the selected rescue protocol for canine lymphoma Chemotherapeutic agent(s) Overall Response Median Duration of Reference Rate Complete Response MOPP mechlorethamine, vincristine, procarbazine, 65 % 63 days Rassnick et al.63 prednisone LOPP lomustine, vincristine, procarbazine, prednisone 61 % not Reported Fahey et al.64 DMAC dexamethasone, melphalan, actinomycin-D, 72 % 61 days Alvarez et al.65 cytosine arabinoside 43 % not Reported Parsons-Doherty et al.66 Rabacfosadine rabacfosadine 74%, 107 days Saba et al.67 32 Table 1.3 Selected toxicity grading items defined by VCOG-CTCAE50 Grade Anorexia Vomiting Diarrhea Neutropenia 1 Coaxing or dietary change required to <3 episode in 24 h Increase of up to 2 stools Neutrophil count: maintain appetite per day over baseline 1500 /ul to lower limit of normal 2 Oral intake altered (≤3 days) without 3 – 10 episodes in 24 h ;<5 Increase of 3–6 stools per significant weight loss; oral nutritional episodes/day for ≤48 h; parenteral day over baseline; supplements/appetite stimulants may be fluids (IV or SC) indicated ≤48 h; medications indicated; 1000-1499 /ul indicated medications indicated parenteral (IV or SC) fluids indicated ≤48 h; not interfering with ADL 3 Of >3 days duration; associated with Multiple episodes >48 h and IV Increase of >6 stools per significant weight loss (≥10%) or fluids indicated>48h day over baseline; malnutrition; IV fluids, tube feeding or incontinence >48 h; IV fluids 500-999 /ul force feeding indicated >48 h; hospitalization; interfering with activities of daily living 4 Life-threatening; TPN indicated; >5 days Life-threatening Life-threatening <499 /ul duration 5 Death Death Death Death 33 Table 1.4 The summary of the differences between amplicon and shotgun sequencings Amplicon sequencing Shotgun sequencing Technical Most selected organisms present, depending on Every organism present will have most of the genome sequenced: all features method used (no viruses) bacteria, fungi, viruses etc. Target 16S – bacteria and some archaea All organisms (including host tissues) 18S – eukaryotes ITS – fungi Method High throughput High throughput 34 Table 1.5 Demographic data of the enrolled dogs Dog number Breed Age Sex WHO Stage Substage Immunophenotype 1 Goldendoodle 3 SF 4 b N/A 2 Mixed Breed 11 NM 5 a B 3 Miniature Schnauzer 13 SF 4 b B 4 Mixed Breed 5 NM 4 b N/A 5 Mixed Breed 8 SF 5 a B 6 Bullmastiff 4 NM 4 a B 7 Goldendoodle 8 NM 5 b T 8 Golden Retriever 8 NM 3 b T 9 Rottweiler 6 NM 3 b B 10 Chesapeake Bay Retriever 10 SF 3 b B 11 Coonhound 7 NM 3 a B 12 Great Dane 7 NM 2 a B 13 German Shepherd Dog 7 NM 3 a B 14 Siberian Husky 6 NM 5 a N/A 15 Jack Russel Terrier 14 NM 5 b B 16 Golden Retriever 6 IM 5 a N/A 17 Rottweiler 3 NM 3 a B 35 Table 1.5 (cont’d) 18 English Cocker Spaniel 7 NM 3 b B 19 Boxer 6 SF 4 b N/A 20 Labrador Retriever 12 SF 5 b B 36 Table 1.6 Summary of GI toxicity (events) Grade anorexia vomiting diarrhea 1 6 6 6 2 0 6 11 3 0 0 0 4 0 0 0 5 0 0 0 37 Table 1.7 Summary of chemotherapy-induced neutropenia Grade Definition (neutrophil count/ul) Events Dose delay 1 1500- 2,999 34 4 2 1000-1499 4 3 3 500-999 3 3 4 <499 1 1 Total 42 10 38 Table 1.8 The summary of bacterial taxa that contributed to PCA components 1 and 2 Taxa Showing Trends Dog Antibiotic GI toxicity Increase Decrease number at any point 1 No No Lacnospiraceae (P=0.0036*) Enterococcus (P=0.0092*) 2 No Yes Bacteroides.uncultured None 5 No Yes Prevotella.uncultured (P=0.0166*) Streptococcus.uncultured (P=0.0187*) Fusobacterium.uncultured (P=0.023*) Lacnospiraceae (P=0.0166*) 9 Yes No None None 3 Yes Yes None Enterococcus (P=0.0046*) 6 Yes Yes None None 7 Yes Yes None None 8 Yes Yes Lacnospiraceae (P=0.023*) Prevotella.uncultured (P=0.046*) Note: The dogs are ordered by first the presence or absence of GI toxicity, and then, the antibiotic use. It shows then if there was any significant increasing/decreasing trend in the gut microbiome taxa compared to the sample on Week 1. 39 Table 1.9 Correlation coefficient between relative abundance of the selected OTUs and numerical clinical variables Lachnospir Bacteroides.unc Prevotella_unc Fusobacterium.un Streptococcus_un Enteroco aceae ultured ultured cultured cultured ccus. Neutrophil count -0.0611517 0.21075359 -0.00969 0.04540302 -0.10473 -0.06718 Platelet count 0.2575669 0.06335429 -0.25824 -0.0864545 -0.26436 0.168598 8 Hct value -0.1942566 0.10650308 0.333786 0.12566997 0.117599 -0.03016 Percentile change of -0.0203185 0.12595068 -0.0611128 0.38534031 -0.0684224 -0.06779 neutrophil count Percentile change of 0.0656166 0.10747033 -0.07893 -0.0771521 0.093838 0.048584 platelet count 3 Percentile change of Hct 0.1200166 0.16635226 0.061099 0.03902454 -0.03104 -0.30184 value 6 Note: 1st row means the selected OTUs, and 1st column means numerical clinical variables, respectively. The intersection of each row and column is the corresponding correlation coefficient [r]. 40 DLBCL, Human PTCL, Human DLBCL, Dog PTCL, Dog Figure 1.1 The comparative aspects of DLBCL and PTCL in histopathology68-70 The pictures show typical histopathologic finding of DLBCL and PTCL both in humans and in dogs 41 The comparison of publication in the gut microbiome between humans and dogs 10000 8000 No. of publications 6000 4000 2000 0 20 20 20 20 20 20 20 2000 02 04 05 06 07 08 09 20 20 20 20 20 20 20 2010 11 12 13 14 15 16 17 20 20 20 20 2018 19 20 21 22 Year Figure 1.2 The comparison of publication in the gut microbiome between humans and dogs Blue: the number of publications of the gut microbiome in humans Orange: the number of publications the gut microbiome in dogs 42 The number of GI toxicity 7 6 5 Episodes 4 3 2 1 0 1 2 3 4 5 6 7 8 9 10 N/A Week Figure 1.3 The number of GI toxicity identified in the study X-axis is Week (based on CHOP based protocol) and Y axis is the number of episodes. N/A means hospital visit out of the first week of CHOP protocol. 43 Figure 1.4 Longitudinal relative abundance of the gut microbiome in 8 dogs (Dog 1-8) Each column shows the relative abundance of the gut microbiome in each dog. X axis means Week (based on CHOP based protocol). N/A means the week which was not applicable for the protocol. 44 Figure 1.5 The average relative abundance of the gut microbiome in 8 dogs that underwent CHOP based protocol X-axis means the Week of CHOP protocol, and Y-axis means the percentage of the gut microbiome. 45 Figure 1.6 The relative abundance of the dog without GI toxicity 46 Figure 1.7 The relative abundance of the dog with GI toxicity 47 Figure 1.8 The Shannon Diversity grouped by treatment received 48 Figure 1.9 Shannon Diversity Index and PCA analysis in the dogs with vs without GI toxicity (A) Shannon Diversity Index grouped by the presence or absence of GI toxicity showed a trend towards a difference in microbiome diversity. (B) PCA plot of the presence or absence of GI toxicity. 49 Figure 1.10 Shannon Diversity Index and PCA analysis in the dogs with vs without antibiotic use (A) Shannon Diversity Index grouped by the administration of antibiotics during treatment and shows a significant difference, p- value<0.05 by Kruskal-Wallis test. (B) PCA plot of the administration of antibiotics. 50 Figure 1.11 Heat map of relative OTU abundance across samples (A) heatmap of GI toxicity (B) heatmap of neutropenia Abundances were measured as proportions of samples and all the 79 abundant OTUs are shown. Samples and OTUs were clustered hierarchically based on relative abundance profiles. On the left y-axis labels represent individual sample code. The color-coded on the right y axis groups the presence or absence of toxicity. OTUs are represented on the x-axis with corresponding relative abundances shown in the heatmap grid with increasing abundance from white to dark green. 51 Figure 1.12 The comparison of the means of the relative abundance between dogs with chemotherapy induced toxicity and without it X axis means chemotherapy induced toxicity categories (No toxicity, GI toxicity, Neutropenia, or Both). Y axis means the relative abundance of the gut microbiome. 52 APPENDIX B: SUPPLEMENTAL TABLES AND FIGURES Table S1.1 The summary of the first 10 week of CHOP based protocol Drug Route Week 1 Week 2 Week 3 Week 4 Week 5 Week 6 Week 7 Week 8 Week 9 Week 10 VCR IV x x x x CTX PO x x DXR IVa x x Prednisoneb PO x x x x Fecal sample collection x x x x (x)c x x x x (x)c CBC x x x x x x x x x x PE x x x x x x x x x x Note: a. Doxorubicin is intravenously administered by free dripping for 15 minutes or longer. Dose is adjusted at 30 mg/m2 when weight is >15kg, at 1mg/kg when weight is <15kg. b. Prednisolone is prescribed at 2mg/kg daily, gradually decrease to 0.5mg/kg daily, and discontinue at Week 4. c. Fecal sample collection at this point is optional. 53 Table S1.2 The summary of concomitant medications in 20 dogs included in the study Week Medications 0 1 2 3 4 5 6 7 8 9 10 Antibiotics 0 1 4 2 3 0 2 0 0 0 0 Steroid 0 3 12 14 11 1 5 0 0 0 0 Metronidazole 0 0 1 3 1 0 1 0 0 0 0 Probiotics 1 1 1 1 1 0 3 0 0 0 0 Maropitant 0 2 2 1 0 0 2 0 3 1 0 Omeprazole 0 1 1 1 3 0 1 1 1 1 0 Pantoprazole 0 1 0 0 1 0 0 0 0 0 0 Ondansetron 0 0 0 0 1 0 2 1 1 1 0 Famotidine 0 0 4 4 3 0 1 1 2 1 0 Diphenhydramine 0 1 0 0 0 0 0 0 0 0 0 Trazodone 0 0 2 2 1 0 3 4 4 7 0 Fluoxetine 0 1 0 0 0 0 0 0 0 0 0 Phenylpropanolamine 0 1 1 1 1 0 1 0 0 0 0 levothyroxine 0 1 1 1 1 0 1 1 1 1 0 Vitamin B12 0 0 0 1 1 0 1 1 1 1 0 CBD 0 1 0 2 1 0 0 1 1 0 0 Clopidogrel 0 0 0 0 0 0 0 0 1 1 0 Total 1 14 29 33 29 1 23 10 15 14 0 Note: the row means the type of the concomitant drugs. The column means the Week of CHOP based protocol. 54 Figure S1.1 The picture of FastDNA SPIN Kits for Soil 55