‘ . ~.......
'3’“: .svr‘a. '3'?‘“3."{nm-
ft-LLII I
fit' .. r- 3:.
HAS“ .., V
U. .. ,,
n
-.
l
‘1,
'1'.
MN
1:.
'1"
u
"N inner,
‘3' ‘; 3.1'Ju.s{"?
—'. I";

1N?" ”fa

€111?

v‘.‘.f~] IL

 

am
'61:: '52? v

1

5|

.1.“ fl.

 

 

 

I
"fit;
- l . . . 2‘ 5 .1“
'AH: - q'I' J;'\ 5,,
" ' . . - ,1" “I
V ' ‘.:gsy a . .€,;:.,W1£a 53:
v u v % 11c
’ ~ 1 3 I‘I'I’ a.
“$2333.”, n». w w.
>run: r-z». mm . u
“3V3.”
, e
7
7
3
r
.
11'

w- :. “gamma,“ .5.” --

1“ |'1\¥ ‘

 

4
‘
11-
9

AL '11“: 1N!" ‘1...

1h

{‘1

 

3

‘ i1"

"1"}. x}.
,x_

:3;
71319,

.. .331,

:4
(:1!

11“!

:11-

.1'
(4573969.. '0

“'7”,
h4g4."
me? 33"
thug" ‘l‘

- .1

ZFQIF '

{H

u-t f

Q

n 9
Adv»:
.1l-m1
n.

r

("51'
7* 1.

nkrn

5

.u...

v.4“ «Will—
1 . .
Emu?“ ' ‘A'x :‘

I ,.. ,3) .1.
15““ u. ”x ‘g‘ .‘f “13::
"1". C wt: 37:15.)

in]; .4.

’fitk
hmfigfii‘»
1 .

‘

'tf!

35-1113

:1'fiM1

'1:

-t_

~‘f'l \l
”5213' 154;.
fir

U 4;” -
”“1.
.1,.

1|

"<2:
ifim‘

m
‘u

‘ .1?!”
"'J‘l'x ’ AV
’1‘?

-. “um.

Kill 0

21

.

. ... ,...s

w“ Fang-)9
.

_ In"
#4115?

I 3335? :
grglt: 3531;11:7444 #9 (Erf '
'1 ”‘7‘ ”'52:? -v"::'v;7
1 '1: '

X"

. .. 4 F
5"1 :3

’1’,
“'4 715‘» I

53 :‘r

.35' -

"1-f- ‘ 3’3”: -. 1‘5?
5" I

' ‘:S"Q'“-v;:

“I" I

 

" - 59.8

This is to certify that the

thesis entitled
A SELECTIVE EXTRACTION TECHNIQUE FOR HYDROXYLATED

FATTY ACIDS IN MAMMARY TISSUE BASED ON ANALYSIS
BY GAS CHROMATOGRAPHY -- MASS SPECTROMETRY

presented by

ELLEN MARIE YUREK

has been accepted towards fulfillment
of the requirements for

MASTER OF SCIENCE . CHEMISTRY
degree 1n

Date (mm/m O

 

0-7639 MS U is an Affirmative Action/Equal Opportunin Institution

 

LIBRARY '
Michigan State
University

 

 

 

'5'. .__ , ..‘a.’.‘.

PLACE IN RETURN BOX to remove this checkout from your record.
TO AVOID FINES return on or before date due.

___________——__—_1

DATE DUE DATE DUE DATE DUE

____i
:; ,
fir—fl

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

ASE
__W___

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

; [—7 j l.

MSU Is An Affirmative Action/Equal Opportunity Institution
chS-pd

 

A SELECTIVE EXTRACTION TECHNIQUE FOR HYDROXYLATED
FATTY ACIDS IN MAMMARY TISSUE BASED ON ANALYSIS
BY GAS CHROMATOGRAPHY - MASS SPECTROMETRY

By

Ellen Marie Yurek

A THESIS

Submitted to
Michigan State University
in partial fulfillment of the requirements
for the degree of

MASTER OF SCIENCE

Department of Chemistry

1990

ABSTRACT

A SELECTIVE EXTRACTION TECHNIQUE FOR HYDROXYLATED
FATTY ACIDS IN MAMMARY TISSUE BASED ON ANALYSIS
BY GAS CHROMATOGRAPHY - MASS SPECTROMETRY
By
Ellen Marie Yurek

Dietary polyunsaturated fatty acids have been shown to increase the rate and
incidence of mammary tumorigenesis in animal and epidemiological studies. One
possible mechanism involves formation of oxygenated metabolites of the polyunsaturated
fatty acids, yet no direct evidence of such oxygenated fatty acids has been reported.
Published methodologies for the extraction of oxygenated fatty acids with subsequent
analysis by gas chromatography with flame ionization or mass spectral detection were
determined to be unsuitable for analysis of the mammary gland matrix. In this study, a
new procedure for the isolation of hydroxylated fatty acids from mammary tissue was
developed. Analysis by gas chromatography - mass spectrometry detected some
hydroxylated C18 fatty acids at the parts per billion level in many of the tissue extracts,
indicating that the developed methodology is appropriate for the determination of
hydroxylated fatty acids in mammary glands from mice fed diets of different fatty acid

content.

ACKNOWLEDGEMENTS

I wish to thank my committee members, Dr. J. Throck Watson, Dr. John Allison,
Dr. Charles Sweeley, and Dr. Clifford Welsch, for their assistance during this project.
Dr. Watson, my research preceptor, was particularly helpful in refining the writing style
in this and other works throughout my graduate career.

As a member of the Watson Lab and the MSU/NIH Regional Mass Spectrometry
Facility, 1 benefited from the expertise and camaraderie of its many members during the
past five years. In particular, I must thank Kathleen Kayganich, Dr. Doug Gage, Mel
Micke, Mike Davenport, Melinda Berning, Bev Chamberlin, Helen Mayer, and Ben
Gardner for their advice, assistance and instruction. Dr. George Yefchak deserves
recognition for his contribution of Figure 3-1.

Gratitude must be extended to my Michigan support team - Chris Evans, Leighta
Johnson, Karen Light, and Jim DeFrancesco - for their friendship during our graduate
careers.

No list of acknowledgements would be complete without recognition of the love
and support my parents have always given me and the supreme patience and devotion of
my husband, Scott Yawger. Their belief in me has helped me weather any storm.

I am indebted to the Public Health Service/NIH and MSU's College of Natural
Science for financial support for this project and a fellowship, respectively.

TABLE or CONTENTS

List of Tables
List of Figures
List of Abbreviations
1. Chapter One: Biochemical Significance of Oxygenated Fatty Acids
A. Introduction .
B. Effects of Dietary Polyunsaturated Fat on Tumorigenesis
C. Proposed Mechanisms for the Effect of Dietary Fat on
Mammary Tumorigenesis
1. Endocrine System
2. Intercellular Considerations
3. Dietary Factors
D. Antioxidant Suppression of Fat-Enhanced Tumorigenesis
E. Lack of Direct Evidence for Oxygenated Intermediate
II. Chapter Two: Adaptation of Existing Analytical Techniques
A. Identification of Hydroxylated Fatty Acids
l. Esterification of Fatty Acids
2. Trimethylsilyl Ethers from Hydroxylated FAMEs
B. Isolation Techniques for Hydroxylated Fatty Acids from Tissue
1. General Lipid Methods
2. Pace-Asciak Isolation of Keto-FAs from Kidney
3. Use of Funk and Powell Extraction Methodology
4. Hydrogenation of Extracts
111. Chapter Three: Development of New Isolation Procedure

iv

\O\O\OOOUI#UJU)U)

mNNHHHr-d
wooom—v—o

A. Development and Description of New Isolation Procedure
1. Solvent Extractions
2. Methylation
3. Hydrogenation
4. Normal-Phase Chromatography
5. Trimethylsilylation of Alcohols
6. Minimization of Contaminants
B. Detection Techniques for Analysis of Extracts
1. Electron Ionization/Scanning Mass Spectrometry
2. EI/MS with Selected Ion Monitoring
3. Determination of Limit of Detection
C. Recovery of Hydroxylated Fatty Acid
1. 12-OH 18:0 as the Model Compound
2. 2-OH 12:0 as the Model Compound
3. 3H—5-HETE as the Model Compound
D. Conclusions
IV. Chapter Four: Analysis and Quantitation of Hydroxy Fatty Acids in
Mammary Tissues
A. Analysis of Mammary Tissues for C18 OH-Fatty Acids
B. Use of Internal Standard for Quantitation
V. Chapter Five: Directions for Future Work
A. Determination of Other Hydroxylated Fatty Acids
B. Analysis for Other Oxygenated Fatty Acids
C. Minimization of Backgron Contaminants
D. Separation of Extract by High Performance
Liquid Chromatography

List of References

33
33
34
34
35
39
39
41
41
49
50
55
55

63

70
77
77
78
78

80
81

Table 2-1:

Table 2-2:
Table 3-1:

Table 3-2:
Table 3-3:
Table 4-1:
Table 4-2:

Table 4-3:

LIST OF TABLES

Fatty Acid Composition of Free Fatty Acid Extract Based
on Injection of Standards

Efficiency of Hydrogenation Reaction

List of m/z Values at which Peaks for (Jr-cleavage Ions
Would Occur for Isomers of OTMS-18:0 ME

Limit of Detection for OTMS 18:0 ME Isomers
Recovery of Hydroxylated Fatty Acids
Fatty Acid Composition of Oils and Butter

Amounts of OTMS 18:0 ME in Extracts as Analyzed by
GC-EI/MS with SIM

Amounts of OTMS 18:0 ME in Extracts as Analyzed by
GC-EI/scanning MS with Internal Standard

page

27
32

47
53
56
65

69

73

Figure 1-1:

Figure 2-1:

Figure 2-2:

Figure 2-3:

Figure 2-4:

Figure 2-5:

Figure 2-6:

Figure 2-7:

Figure 2-8:

Figure 2-9:

LIST OF FIGURES

Structures of the different families of unsaturated fatty
acids, demonstrating the different positional arrangements
of the conjugated double bonds, counting from the methyl
terminus. Source: Reference 16.

Acid catalyzed methylation of carboxylic acid by
diazomethane.

The electron ionization mass spectrum of the TMS ether
derivative of the methyl caster of 12-hydroxy stearic acid.
Source: Reference 37.

Negative ion chemical ionization mass spectrum of the
TMS ether, pentafluorobenzyl ester of 5-HETE using
methane as the auxiliary gas. Source: Reference 49.

Pace-Asciak procedure for the extraction of oxygenated
fatty acids from rat kidney.

ECD chromatogram of PFB ester, TMS ether of Pace-
Asciak extract of mammary tissue. 15 m DB-l megabore
(0.52 mm i.d., 1.0 pm film) column, 180° - 280°, 8°/min.
16:0 PFB ester and 26:0 PFB ester were added as internal
standards.

FID chromatogram of methyl ester, TMS ether of extract of
mammary tissue processed by Pace-Asciak method. 60 m
DB-l capillary (0.25 mm i.d., 0.25 um film) column, 180° -
280°, 4°lmin. Hydrocarbons HC 10, HC 12, HC 14, HC
16, HC 18 and HC 20 were added for retention indexing.

Funk and Powell procedure for the extraction of
oxygenated fatty acids from rat and rabbit aortae, as
modified in this laboratory.

FID chromatogram (top) and reconstructed total ion
chromatogram (bottom) of the methyl ester, TMS ether of a
free fatty acid extract processed by the modified Funk and
Powell method. Both analyses - 15-m DB-l fused silica
megabore (0.52 mm i.d., 1.0 um thick film) column.

FID chromatogram of the methyl ester, TMS ether of a free
fatty acid extract processed by the modified Funk and
Powell method. 60-m DB-l capillary (0.25 mm i.d., 0.1
ttm thick film) column, 180° -325°, 4°lmin,
attentuation=1.

I3389

10

12

14

16

18

19

21

24

25

Figure 2-10:

Figure 2-11:

Figure 2-12:

Figure 3-1:

Figure 3-2:

Figure 3—3:

Figure 3-4:

Figure 3-5:

FID chromatogram of the methyl ester, TMS ether of a
toral lipid extract processed by the modified Funk and
Powell method. 60-m DB-l capillary (0.25 mm i.d., 0.1
pm thick film) column, 180° -325°, 4°lmin, attentuation=3.

Reconstructed total ion and mass chromatograms at m/z

301, 187, 313, 175 and 73 from GC-EI/MS (scanning m/z
50-500), illustrating potential chromatographic overlap of
13-O'TMS 18:0 ME (10 ng inj.) and 20:0 ME (50 ng inj.).
Other peaks in the TIC are unsaturated C20 methyl esters.

FID chromatograms of non-hydrogenated (top) and
hydrogenated (bottom) aliquots of total lipid extract. 60 m
DB—l capillary (0.25 mm i.d., 0.1 pm film) column, 180° -
325°, 4°lmin.

Schematic representation of the gas manifold used to
hydrogenatc four samples simultaneously.

Megabore column (DB-l, 15-m long, 0.52 mm i.d., 1.5 um
thick film) GC-FID chromatograms resulting from analysis
of spiked mammary gland extract clucnts from selected
mobile phases through silica SPE columns. The bottom
chromatogram is 12-OH 18:1 ME and 12-OH 18:0 ME in
the elucnt from 100% benzene (1.2% of the cluent injected;
attenuation=3). The top chromatogram represents a
mixture of 14:0, 16:0, 18:1, and 18:0 MEs (1.7% of the
elucnt injected; attentuation=8). GC conditions: N 2 flow
rate 15.0 mL/min; temperature program-160° for 4 min,
then 8°/min to 280°.

Summary of new procedure to isolate hydroxylated fatty
acids from mammary glands. See text for detailed
description.

Megabore column (DB-1, 15-m long, 0.52 mm i.d., 1.5 pm
thick film) GC-FID chromatograms of chloroform solvent
blanks through silica SPE columns from Burdick &
Jackson (bottom) and J .T. Baker (top) with elution times of
some FAMEs designated. GC conditions: N2 flow rate
15.0 mIJmin; temperature program-140° for 4 min, then
8°lmin to 280°; 2 mL injection volumes. Attenuation=5.
The marked peak has an area equivalent to the area of the
signal from 0.1 ug 18:0 methyl ester.

Capillary column (DB-1, 60-m long, 0.25 mm i.d., 0.1 pm
thick film) GC-FID chromatograms of 4.0% of extract from
new isolation procedure (top) and 0.5% of total lipid
extract from modified Funk and Powell Method (bottom)
with elution times of some FAMEs designated. GC
conditions: He flow rate 1 mL/min; temperature program-
180°, 4°lmin to 320°; attentuation=3.

26

29

31

36

38

42

43

Figure 3-6:

Figure 3-7:

Figure 3-8:

Figure 3-9:

Figure 3-10:

Figure 3-11:

Figure 4-1:

Figure 4-2:

Figure 4-3:

ix

Capillary column (DB-l, 30—m long, 0.314 mm i.d., 0.25
pm thick film) GC-FID chromatogram of 1.5% of
hydrogenated extract from new isolation procedure with
elution times of some FAMEs designated GC conditions:
He flow rate 1 mL/min; temperature program-180°, 4°lmin
to 325°; attentuation=2.

Two problems caused by inadequate spectral generation
rates in GC-MS. Top: True chromatographic profile is
poorly represented due to an insufficient number of TIC
points. Bottom: Mass spectral peak intensities are skewed
due to the change in partial pressure of the analytc which
gours during the scan acquisition time. Source: Reference

 

Calibration curve and equation for the determination of 12-
OTMS 18:0 ME by GC-EI/MS with SIM at m/z 187.

Selected ion current profile at m/z 173 from GC-EI/MS
analysis with SIM of the ten ions listed in Table 3-2,
representing an tit-cleavage ion of 13-OTMS 18:0 ME. The
signal represents injection of 11 pg of the analytc, or 10 ng
analyte/g tissue (10 ppb).

Calibration curve for determination of 12-OTMS 18:0 ME
by GC-FID. The average area response of replicate
injections (N=2, 3, or 4) is plotted. Error bars represent 21:1
standard deviation.

Reconstructed total ion current chromatogram and mass
chromatograms at m/z 301 and m/z 187 from GC-EI/MS
(scanning m/z 50-500) analysis of tissue spiked with 12-
OH 18:0 and extracted by the new isolation procedure.
Structure indicates the origin of these a-clcavage fragment
ions.

Selected ion-current profiles at m/z 229.2 & 259.2 (9-
OTMS 18:0 ME). m/z 301.2 & 187 (12-OTMS 18:0 ME),
m/z 313.2 & 175 (ii-OTMS 18:0 ME) and m/z 315.2 & 173
(13-OTMS 18:0 ME),representing the a-cleavage fragment
pairs, and m/z 339.2 [M-47]"' from GC-EI/MS with SIM of
these nine ions.

Selected ion-current profiles at m/z 315.2 (top) and m/z
173 (bottom) from GC-EI/MS with SIM of the nine ions
displayed in Figure 4—1. The arrow indicates elution time
of endogenous 13-OTMS 18:0 ME.

Calibration curves and linear regression equations for the
determination of 12-OTMS 18:0 ME (top) and of B-OTMS
18:0 ME (bottom), using 2.00 ng 2-OTMS 12:0 ME as an
internal standard.

45

48

51

54

57

59

67

68

71

Figure 4-4:

X

Reconstructed total ion chromatogram and mass
chromatograms at m/z 259, 229, 301, 187, 313, 175, 315,
173, representing the OTMS 18:0 ME isomers listed in
Figure 4-1, from GC-EI/ scanning MS (m/z 170-320).
Injection of 2.0% of the hydrogenated portion of tissue B-5,
results of quantitation are listed in Table 4-3.

74

DMBA
BHT
ODC
DNA
BHA
FAMEs
BF3
HCl

tBDMS
ECD
OTMS
EI/MS

PGs
NICI/MS
'I'Xs
HPLC
PFB
KI-12PO4
NaOH

LIST OF ABBREVIATIONS

dimethylbenzanthracene

butylated hydroxytolucne

ornithine decarboxylasc
deoxyribonucleic acid I

butylated hydroxyanisole

fatty acid methyl esters

boron trifluoride

hydrochloric acid

trimethylsilyl

t-butyl dimethyl silyl

electron capture detection
trimethylsilyl ester

electron ionization mass spectrometry
mass to charge ratio

atomic mass units

prostaglandins

negative ion chemical ionization mass spectrometry
thromboxanes

high performance liquid chromatography
pcntafluorylbcnzyl

potassium phosphate, dibasic

sodium hydroxide

(1. H20
ODS
CHC13
MeOH

v/v
TBA-H804
BSTFA
ECD

i.d.

SPE

EtOH

rpm
OH-FAMEs

yi
YT
S/N
TIC
RSD

distilled water

octadecylsilyl

chloroform

methanol
volume/volume

tetrabutylammonium hydrogen sulfate
bis(trimethylsilyl)trifluoroacetamide
electron capture detection

inner diameter

solid phase extraction

ethanol
revolutions per minute

hydroxylated fatty aicd methyl esters
selected ion monitoring

signal of the analyte at its limit of detection
the signal from the blank

standard deviation of the slope
concentration at the limit of detection
slope of the calibration curve
calculated y-interccpt

quantity injected
quantity at the limit of detection
abundance of ion at m/z i
total ion signal

signal to noise ratio
total ion current

relative standard deviation

dpm

MS/MS

retention factor
disintegrations per minute
parts per million

tandem mass spectrometry
methyl ester

minutes

1. CHAPTER ONE: BIOCHEMICAL SIGNIFICANCE or OXYGENATED FATTY ACIDS
A. Introduction

The role of dietary polyunsaturated fatty acids in increasing the incidence of
tumors, particularly in mammary glands, has been well documented in numerous animal
and epidemiological studies. Several hypotheses have been proposed to explain the
mechanism of this tumorigcnic process (1). One of these theories proposes activation by
oxidation at the sites of unsaturation in the fatty acid, and has been supported by reports
from some laboratories that administration of antioxidants decreases or eliminates the
enhancement in tumorigenesis usually realized with a diet high in polyunsaturated fatty
acids. Yet no direct evidence of oxygenated metabolites in glands of animals fed diets
high in polyunsaturated fatty acids has been reported This thesis describes the
development of analytical methodologies to extract and analyze one type of oxygenated
metabolite, hydroxylated fatty acids, from mammary tissue.
B. Effects of Dietary Polyunsaturated Fat on Tumorigenesis

As early as 1942, Tannenbaum (2) reported an increase in the growth of
spontaneous mammary tumors in mice fed a diet high in fat with respect to those mice
fed a low-fat diet. This same effect of high fat diet on the incidence of mammary tumors
in rats was observed by Engel and Copeland (3) when the tumors were chemically
induced. Results from numerous laboratories have demonstrated a similar correlation of
high dietary fat and an increase in mammary tumorigenesis in rats or mice when other
carcinogens were employed (4-6). I

Epidemiological evidence has been presented that suggests a similar correlation
between the level of dietary fat and the incidence of breast cancer (as well as prostate,
colon and other cancers) (5, 7). Although the trends are not strictly defined, a
relationship has been observed between the age-adjusted mortality due to breast cancer

and the amount of dietary fat available in numerous countries (8, 9). The

integrity of these assumptions and conclusions of these epidemiological studies has been
disputed (10).

Further studies demonstrated that the degree of unsaturation in the fat is positively
related to mammary tumorigenesis in experimental animals (11, 12). Those rodents fed a
diet high in polyunsaturated fat had a higher rate of tumorigenesis than those rodents fed
a similar diet high in saturated fat, which in turn exhibited a higher rate of tumorigenesis
than rodents fed a same-calorie low-fat diet.

Animal models for the study of cancer in various organs have been developed.
To investigate mammary tumorigenesis in experimental animals, rats and mice are often
treated with a chemical carcinogen. Commonly used are lipophilic carcinogens such as
7,12-dimethylbenzanthraccne (DMBA), but hydrophobic carcinogens such as N-nitroso-
N-mcthylurea have also been utilized, with the same positive correlation of level of
unsaturated fat to mammary carcinogenesis (13). These studies have focussed mostly on
the effect of dietary fat on the number of tumors and rate of tumor development.

Supporting evidence of the role of dietary fat in mammary tumorigenesis also has
been provided by studies of transplantablc tumors (14). The rate of growth of
transplantable mammary tumors was highest when the animals were fed a diet high in
linoleic or linolenic acid, as compared to those animals fed diets high in oleic or other
monounsaturated and saturated fats.

More recent studies with polyunsaturated fatty acids differing in the position of
the double bonds have demonstrated that an increase in tumorigenesis is dependent upon
the position of the double bonds (15). Those polyunsaturated fatty acids with double
A bonds starting at the sixth carbon atom from the methyl group (denoted n-6 or to-6), such
as linoleic and arachidonic acid, have been shown to cause an increase in tumorigenesis.
Yet those animal fed diets high in n-3 polyunsaturated fatty acids, the so-called omega-3
fish oils, demonstrated no increase, or even a decrease, in mammary tumorigenesis, when

compared to rodents fed diets high in n-6 fats (Figure 1-1).

\/\/\/\/\/\/\/\/‘ COOH stearic 18:0
H
W COO oleic: n-9—1821
/\/\/=\/‘=\/\/\/\/‘ coon linoleic: n—e—iazz
COOH '
W iinoicm'c: n-3—18z3
COOH

WW eioosapeqtaonoic: n-3-20:S
Figure 1-1: Structures of the difl’erent families of unsaturated fatty '
acids, demonstrating the different positional arrangements
of the conjugated double bonds, counting from the methyl
terminus. Source: Reference 16.
C. Proposed Mechanisms for the Effect of Dietary Fat on Mammary Tumorigenesis
l. Endocrine System
Numerous studies have been performed in the past twenty years to attempt to
discover the mechanism(s) of increased mammary tumorigenesis by dietary fat. Several
possible types of factors may be responsible for the carcinogenic pathway. One of the
most researched areas involves the endocrine system. Initial studies which indicated an
increase in secretion of prolactin and ovarian steroids in rats fed diets high in fat (17),
were refuted by subsequent, more sensitive analyses of these hormones in the blood of
animals fed high-fat diets (18). No increase in secretion of these hormones was observed
for the animals fed high-fat diets when compared to those animals fed diets of moderate ,
fat levels (19). The observance of increased tumorigenesis of organs not affected by the
' pituitary-ovarian system also casts serious doubts on possible mechanisms related to
endocrine function.
2. Interceilular Considerations
It also has been proposed that hi gh-fat diets may affect hormone and/or growth
factor receptors (20), especially via a lipid-dependent enzyme system. The. activation of

protein kinase C and extracellular signal transduction seems to be enhanced by

unsaturated diacyl glycerides with respect to saturated diacyl glycerides.

The findings that polyunsaturated fatty acids and prostaglandins can restrict some
immune functions has suggested that the immune system is responsible for the increase
in mammary tumorigenesis caused by high levels of dietary fat. Several studies have
indicated that diets high in unsaturated fatty acids decreases lymphocyte levels or
responsiveness ( 18). But the increase in mammary tumor cell growth in vitro observed
when unsaturated fatty acids are added to the culture medium indicates a mechanism
other than the immune is (also) involved.

Enhanced mammary tumorigenesis by high dietary levels of polyunsaturated fat
could result in changes in the structure of the mammary gland or its sub-cellular
components (14). A dietary increase in unsaturated fatty acids has been shown to
increase the fluidity of cell membranes, which may increase cell mitosis (18). Cell
membrane fluidity has been demonstrated to be increased in mammary tumors by certain
hormonal agents which are known to promote mammary tumor growth. This increase in
membrane fluidity may affect intercellular communication via a decrease in metabolic
cooperation (18). Unsaturated medium-chain fatty acids have been shown to decrease
metabolic cooperation, while their saturated analogues did not.

3. Dietary Factors

Another theory attempting to explain the positive relationship between mammary
tumorigenesis and dietery fat involves the level of caloric consumption. The repeated
observance that caloric restriction decreases mammary tumorigenesis (and tumorigenesis
in other organs) suggests a mechanism relating caloric intake and the tumorigenic process
(21, 22). But, many well controlled-diet studies in rodents utilizing isocaloric diets have
not supported this theory (18). As an example, mice fed isocaloric high fat diets, which
differed only in the amount of saturation or unsaturation exhibited different rates of
mammary tumor development. The incidence of mammary tumors was significantly
higher in the mice fed the unsaturated fat diet. These latter findings do not support the

concept that increased caloric intake is solely responsible for the increase in mammary
tumor development documented in animals fed high-fat diets.

An alternative possible cause of decreased mammary tumorigenesis when a low-
calorie diet is administered ad libitum (and animals consume increased amounts of food)
is increased consumption of particular nutrients (22). Numerous studies have
demonstrated an inverse relationship between the level of several nutrients and incidence
of mammary tumors (23). Increased dietary levels of protein have been shown to
suppress mammary tumorigenesis in rodents, possibly by increased activation of hepatic
mixed function oxidases which metabolize, thus detoxify, chemical carcinogens (21).
When administered in very high doses, vitamin A and other retinoids have prevented, or
significantly reduced, chemically-induced mammary tumorigenesis in rats and mammary
tumor cell growth in vitro (1, 17). Other nutrients have been shown to inhibit mammary
tumor growth when fed at high levels, e.g., selenium, vitamin E, and a combination of
these two compounds (1, 23). Several mechanisms have been proposed for the anti-
tumorigenic activity of these latter three chemicals including their roles as antioxidants,
but such a mechanism has yet to be supported in a compelling manner.

D. Antioxidant Suppression of Fat-Enhanced Tumorigenesis

Besides the above-mentioned potential antioxidants, butylated hydroxytolucne
(BHT) and propyl gallate - two more antioxidants, have been shown to inhibit DMBA-
induced mammary tumorigenesis in rats fed high-fat diets (18, 24). These findings
support the theory that oxygenated metabolites of fatty acids may be involved in this
carcinogenic process. This mechanism is particularly attractive since diets high in
unsaturated fatty acids enhance mammary tumorigenesis more so than do diets high in
saturated fatty acids. Unsaturated fatty acids are susceptible to oxidation, possibly
explaining the observed difference in tumor growth between the different levels of
double bond character in the two high-fat diets (17).

The products and mechanisms of lipid oxidation, particularly of unsaturated fatty
acids, have been extensively studied (25-27). Oxidation of fatty acids can be activated by
enzymes (lipoxygcnase reactions), free radicals (autoxidation and peroxidation), or light
(photooxidation). Hydroperoxides are the initial products of lipid oxidation (25), but are
unstable and would most likely be reduced to other oxygenated forms such as hydroxyl,
carbonyl, or epoxy moieties.

To test this theory that oxygenated fatty acids play a role in tumorigenesis, Bull
and coworkers have used rat colon tumorigenesis as a model (28). This group has shown
that hydroperoxidc and hydroxide derivatives of linoleic and arachidonic acids increase
two indicators of tumor promotion - ornithine decarboxylase (ODC) and
deoxyribonucleic acid (DNA) synthesis - in colonic mucosa. Subsequent analysis of the
effect of structurally-similar fatty acids indicates specifically that an allylic oxygenated
functional group is necessary for induction of CDC and stimulation of DNA sythesis,
both indicators of mitogenic activity (29).

Additional studies investigating the roles of high-fat (saturated and unsaturated)
diets and antioxidants in rodent mammary tumorigenesis support the theory that an
oxygenated intermediate of unsaturated fatty acids is critical in the mechanism. One such
study is the work by McCay era]. (30) in which rats were fed purified diets of high
polyunsaturated fatty acid content, high saturated fatty acid content, or low fat content.
The ability of several antioxidants as inhibitors in DMBA-induced mammary
tumorigenesis in these rats was assessed. Butylated hydroxytolucne was shown to be a
strong inhibitor of induced mammary tumorigenesis in all three diet groups, but was
more effective in the high saturated fat diet group than in the group fed high levels of
unsaturated fat. But no inhibitory effect on mammary tumorigenesis was observed for
those animals fed butylated hydroxyanisole (BHA), nor tocopheral acetate. These latter
findings contradict other studies demonstrating an inhibitory effect of BHA and

tocopheral on DMBA-induced mammary tumors in rats fed commercial rat diets. It is

suggested that the purified diets lack certain cofactor(s) which make these two
antioxidants effective in the commercial diets, or that the purified diets contain a factor or
factors which inactivate BHA and tocopheral acetate, but not BHT.

Another laboratory (31) later reported selective inhibition of DMBA-induced
mammary tumorigenesis in rats by different antioxidants. Mammary carcinomas were
inhibited by BHA or ethoxyquin, while development of mammary fibroadenomas was
reduced by BHA, ethoxyquin, BHT and diaminodiphenylmethanc (separately). In this
study, tocopherol did not appear to inhibit mammary tumorigenesis.

Also supportive of the oxygenated polyunsaturated fatty acid mechanism is a
more biochemically-oriented study by King's group (32) in which rats fed diets (iso-
nutrient and iso-caloric) differing only in the amount of fatty acid or its degree of
saturation were given DMBA. Select animals of each diet group were fed non-toxic
levels of one of various antioxidants for different lengths of time before and after DMBA
administration. Drug metabolism in the mammary gland and some hepatic microsomal
parameters were analyzed. It was shown that those antioxidants which had the greatest
inhibition of mammary tumorigenesis, as measured by incidence of tumors and by
biochemical measures of mitogenic activity, were those which were incorporated into the
mammary gland for the longest time. These results strongly support a mechanism for
DMBA-induced mammary tumorigenesis involving an oxygenated intermediate of
unsaturated fatty acids derived from dietary fat.

Reports from other laboratories have contradicted the oxidized metabolite theory
(18, 24); one study found that a high-fat diet caused only a decrease in the latency period
of mammary tumor appearance in mice, not an increase in the total number of tumors at
long-term necropsy (33). Although there have been some studies which contradict the

oxidized metabolite theory, this mechanism is still a promising, yet unproven, hypothesis.

E. Lack of Direct Evidence for Oxygenated Intermediate

Although conclusions from the above studies suggest that an oxygenated
intermediate may play a role in a mechanism of the effect of dietary fatty acids on
mammary tumorigenesis, no direct evidence has been reported in the literature. There
are no known studies in which mammary glands were analyzed for possible oxygenated
intermediates which may be involved in the tumorigenic process. It would be of great
interest in attempting to test the proposed oxygenated intermediate theory to examine
mammary glands or mammary tumors for any oxygenated intermediates, particularly as a
function of the amount and type of dietary fat fed to the animal.

It is anticipated that any such oxygenated intermediate would be present in only
trace quantifies in the tissue, thus requiring a very sensitive detection technique. The
exact nature of the intermediate(s) is unknown, but it is suspected that the most stable
intermediates would be those which had been reduced to hydroxyl, ketone, and epoxide
moieties. Since the exact oxidized fatty acids are unknown, the detection method must
also be a fairly general one which could also provide structural identification.

Those oxygenated fatty acids proposed to be involved in this possible mechanism
of tumorigenesis would most likely be metabolites of dietary fatty acids. This would
indicate that the origin of the possible intermediates would have carbon chain lengths of
16 or 18, since palmitic, oleic, linoleic, and linolenic acids are the major fatty acid
components of the common dietary fats and oils. These dietary fatty acids could be
metabolized to form oxygenated intermediates of slightly different carbon length.

II. CHAPTER Two: ADAPTATION or EXISTING ANALYTICAL TECHNIQUES

Fatty acids, particularly methyl esters of fatty acids, have been well characterized
by gas chromatographic techniques. There have been additional techniques for the
separation and identification of oxygenated, particularly hydroxylated, fatty acids. After
derivatization of the hydroxy fatty acids, separation is easily performed by gas
chromatography with flame ionization or mass spectrometry as a detection method.

Several laboratories have utilized these gas chromatography techniques, as well
as a few others, to identify oxygenated metabolites of fatty acids in body fluids and
tissues. But, no reports have been found in the literature which describe analysis of
mammary tissue for hydroxylated fatty acid content. Application of some techniques for
extracting and analyzing oxygenated fatty acids in tissue for determination of
hydroxylated fatty acids in mammary tissue was attempted, with unsatisfactory results.
A. Identification of Hydroxylated Fatty Acids

Analysis of fatty acids as their methyl esters has been and still is one of the most
widely-used techniques. Traditionally, gas chromatography with flame ionization
detection utilized packed, deactivated glass columns. The advent of fused silica open
tubular columns improved the stability, resolution, and reproducibility of separation of
fatty acid methyl esters (FAMES) in numerous laboratories (34).

l. Esterification of Fatty Acids

Fatty acids are derivatized to make them more volatile and non-polar, thus more
amenable to gas chromatography (35). The derivatized compound is more easily
vaporized into the gas (mobile) phase and chromatographically behaves more ideally. As
. mentioned above, methyl esters are by far the most commonly used derivatives of fatty
acids. They are easily prepared from the free acids by reaction with ethereal
diazomethane in the presence of catalytic methanol (Figure 2-1). The esterification is
rapid and the reactants are quickly evaporated. Fatty acid methyl esters also can be

prepared by other methods such as with methanol-BF3 or methanol-HCl (36).

9

10

+ +
o -
RC OH —--- RC OH > RC OH ——> RC=O
N2 -H+ .

 

Figure 2- 1: Acid catalyzed methylation of carboxylic acid by diazomethane.

Fatty acid methyl esters are resolved by polar and non-polar stationary phases
according to their molecular weights (carbon number) and, if the stationary phase is
correctly chosen, the number and position of the double bond(s) in the compound can be
determined. The retention indices of the methyl esters of the nattually-occurring fatty
acids are well characterized. Standards of many FAMES are commercially available to
characterize the individual gas chromatographic system. When the compound is an
unknown, mass spectrometry is often used for identification.

Free fatty acids may also be derivatized into other esters, such as the
trimethylsilyl (TMS) or t-butyl dimethyl silyl (tB DMS) esters. These derivatives are
made when higher molecular weights are desired. This is often the case for Short-chain
fatty acids, where the methyl ester would have a shorter retention volume and resolution
may be poor (36). Pentafluorobenzyl esters can be prepared to make the fatty acids
suitable for electron capture detection (ECD). Pyrrolidine esters and picolinyl esters can
be prepared if mass spectral analysis will be performed, as they have distinctive
fragmentation patterns (35).

2. Trimethylsilyl Ethers from Hydroxylated FAMES

Methyl esters of hydroxylated fatty acids are still polar and somewhat involatile
Since they contain the hydroxyl moiety. Thus, if a real fatty acid mixture is methylated, it
is desirable to further derivatize the sample to convert any possible hydroxylated
components to others. Preparation of the TMS ether (OTMS) of the fatty acid methyl
esters provides very volatile and chromato graphically well-behaved compounds (37).

11

Standards of methyl ester, TMS ethers of hydroxylated fatty acids are not
commercially available, but are easily prepared from the free hydroxylated fatty acids or
their methyl esters, some of which are available. Retention indices of the OTMS fatty
acid methyl ester can then be determined experimentally, as these are not catalogued in
the literature. On non-polar Stationary phases, these derivatized compounds are separated
by chain length, number of double bonds, number of OTMS moieties, and position of
these functional groups. Positional isomers of OTMS-FAMES will have Similar retention
indices, but are resolvable by the hi gh-resolution capillary columns now so commonly
used. Therefore, the approximate retention time of a TMS ether, FAME which is
difficult to purchase or synthesize can be determined from a positional isomer which is
more easily obtained. .

If electron ionization mass spectrometry (El/MS) is to be used, this TMS ether
derivatization provides additional an benefit in its abundant fragment at m/z 73 from
McLafferty rearrangement, which can be a marker to indicate the presence of derivatized
hydroxylated fatty acids (Figure 2-2). Other characteristic fragment ions include minor
peaks corresponding to [M-CH3]"' and [M-CH3OHT", allowing determination of the
carbon chain length. As a further aid in identification of positional isomers, a-cleavage
about the TMS ether provides two characteristic fragment ions, the mass sum of which
equals the molecular weight plus 102 it.

B. Isolation Techniques for Hydroxylated Fatty Acids from Tissue

1. General Lipid Methods
Before identifying the hydroxylated fatty acids in a real sample, such as a bodily fluid or
tissue, the interfering components of the matrix must be removed. This is usually done
by extraction of the fatty acids, and subsequent cleanup to obtain the desired compounds
to be analyzed. The objective of this research is to isolate and identify any hydroxylated
fatty acids in mammary glands using the mouse as a model.

l2

 

100

 

 

 

 

 

 

 

I73 301
8"-..
m - ' 187
I
a..(ae,e-a-(ae.ee-Me
E L-sor
)5
z m
E
‘ C
3 4o -
3
C
8
“~31
m . “-1
366
o an All]. A A ' u 1 A 1 I ,
so 1“ 15¢ zoo 250 one see ‘

Figure 2-2: The electron ionization mass spectrum of the TMS ether
derivative of the methyl caster of 12-hydroxy stearic acid.
Source: Reference 37.

No literature procedures were found for the isolation of oxygenated fatty acids,
that is, hydroxy-, keto-, or epoxy- fatty acids, from mammary tissue, yet there are many
laboratories which have analyzed body fluids for oxygenated fatty acids. Pace-Asciak
and Micallef used negative ion chemical ionization to analyze rat blood for
prostaglandins (PGs), their keto catabolites, and their keto, hydroxy catabolites (38).
Negative ion chemical ionization mass spectrometry (NICI/MS) was also used to analyze
saliva, urine, cerebrospinal fluid and platelets for PGS, thromboxanes (TXS) and hydroxy
C20 and C17 fatty acids (39), and to determine hydroxylated C20 (40). After
derivatization to a fluorescent compound, high performance liquid chromatography
(HPLC) was used to detect PGs, TXs, keto metabolites of prostaglandins, and hydroxy
C20 fatty acids in rat plasma (41). These oxygenated fatty acid determinations were
from fluid matrices, which differ greatly from the mammary gland that is to be examined

in this application. Therefore, these extraction techniques were not considered applicable

to the dissimilar matrix of the mammary gland.

13

Some laboratories report the isolation and analysis of oxygenated fatty acids,
mostly arachidonic acid metabolites, from tissues. Electron ionization mass spectrometry
was used to analyze TMS ether, methyl ester derivatives of hydroxy arachidonic
metabolites from monkey seminal vesicles (42); TMS ether, methyl ester derivatives of
hydroxylated lipoxygenase metabolites of C20 and C13 fatty acids from rabbit peritoneal
tissue (43); TMS or tBDMS ether, methyl ester derivatives of hydroxylated metabolites
of C20 and C22 fatty acid in trout gill tissue (44); TMS ether, methyl ester derivatives of
PGS and hydroxylated arachidonic and linoleic fatty acids from rat and rabbit aortae
(45,46); and hydroxylated C20 fatty acids and PGS in brain cortical tissue (47). In these
studies the fragmentation provided by EI/MS allowed more definitive identification of
oxygenated metabolites than the high performance liquid chromatography with
spectrophotometric or radiological analysis concurrently performed.

The pentafluorylbenzyl (PFB) ester, trimethylsilyl ether derivatives of these
oxygenated fatty acid metabolites have been prepared for negative ion chemical
ionization, sometimes in addition to the methyl ether, TMS ethers for EI/MS analysis.
The detection limits with NICI/MS can be lower than those with EI/MS by as much as
1000 fold (48). One reason for this is the ions formed by chemical ionization are
primarily the characteristic anion [M-PFBT (Figure 2-3), and the subsequent loss of the
ether and hydrogen [M-PFB-HOTMS]' (49).

Gleispach and co-workers (48) utilized both EI/MS and positive ion chemical
ionization of methyl ester, TMS ethers and NICI/MS of PFB ester, TMS ethers of
hydroxy fatty acids, TXS, and PGS to determine the arachidonic acid metabolites in skin
fibroblast and saliva. Qualitative data were provided by the fragmentation of these
compounds, while quantitative determinations were made by the more sensitive chemical
ionization techniques and isotope dilution. Murphy has used the PFB ester, TMS ethers
with NICI-MS to analyze lipoxygenase products (50) particularly hydroxy arachidonic

acids and leukotrienes, as well as the methyl ester, TMS ethers with EI/MS in lung tissue

14

(49). Kcto PGS and keto, hydroxy catabolites of PGS in rat kidney were determined by
NICI/MS of the PFB ester, methoxime, TMS ester derivatives (51).

 

 

 

 

1001
101
- NlCi (CH4)
" )Ot
]-T“SON
7:5 o ”‘ f f 3"

. _ «-.l..,©.
). _ F f
.— -— l
177
Z
w -t
.—
Z
.5
Ad” 0 v fl #1 l1 ijijijij

DO 200 300 400 m/z

Figure 2-3: Negative ion chemical ionization mass spectrum of the
TMS ether, pentafluorobenzyl ester of 5-HETE using
methane as the auxiliary gas. Source: Reference 49.

So it was from these procedures, and general lipid extraction methods described
in the literature, that preliminary attempts to extract hydroxylated fatty acids from mouse
mammary tissue were based. In the fifties, early extraction techniques for the extraction
of lipids from tissue were published. Chloroform-methanol was the conventional organic
solvent system for homogenization of tissues, followed by an aqueous phase extraction
(52, 53). These early extraction procedures are still widely used today, as they are rapid
and fairly quantitative (54, 55). Further sample cleanup is usually achieved with
chromatographic methods (56).

The samples used for testing and further development of the existing analytical
methodology to isolate and identify hydroxylated fatty acids were obtained from three
month old female Balb/c mice under the care of Dr. C.W. WelSch's laboratory. The
inquinal (number four and five) mammary glands were removed from the animals by

trained laboratory personnel with particular care to excise only mammary tissue.

15

2. Pace-Asciak Isolation of Keto-FAS from Kidney

The procedure for extraction of keto- and keto, hydroxy- prostaglandins from rat
kidney used by Pace-Asciak and Micallef (38) was applied to the mouse mammary
tissues. This procedure was selected due to the complex nature of both matrices, for
which the sample cleanup procedure may prove to be appropriate. After extraction, the
Canadian workers prepared the PFB ester, TMS ether derivatives of the oxygenated PG
metabolites for analysis by GC-ECD and NICI/MS. Since GC-FID and EI/MS are'the
preferred analysis methods for the unknown oxygenated fatty acids possibly present in
the mammary glands, the methyl ester, TMS ether derivatives were also prepared in this
study.

The extraction procedure is outlined in Figure 2-4. Each mammary gland was
approximately 0.1 g. The tissue was rinsed with 0.05 M KH2PO4 - NaOH buffer (pH
7.4), then homogenized with 10 volumes of buffer (10 mL) in a silanized homogenizer
with a Teflon pestle. Five volumes of ethanol (5 mL) was added and the mixture was
filtered with suction on No. 42 Whatrnan paper. These latter steps were to stop the
incubation that Pace-Asciak and Micallef desired, and to filter precipitated protein,
respectively.

The filtrate was evaporated in vacuo and 1 mL of distilled water (d. H20)
acidified to pH 3 with HCl. This mixture was passed through an octadecylsilyl (ODS)
C18 Sep Pak solid phase extraction column (Waters) which had been pretreated with 20
mL each of methanol and d. H20. The column was washed sequentially with 8 mL of d.
H20, hexane and methanol. The methanol fraction was collected and dried by rotary
evaporation. The residue was dissolved in 200 ILL chloroform and applied to a 325 mesh
silicic acid (Sigma) column in a small pasteur pipet. This column was washed with 8 mL
chloroform, and 8 mL CHCl3-MeOH (9:1, v/v). The elucnt was dried in vacuo and
reconstituted in 2.00 mL methylene chloride. The sample was divided into 1.00 mL

Figure 2-4:

16

Rinse tissue in 0.05M KH2P04-NaOH buffer (pH 7.4)
Homogenize in 10 volumes of buffer
Add 5 volumes EtOH
Filter from precipitated protein on Whatrnan No. 42 paper
Evaporatc filtrate in vacuo
Dissolve in 1 mL d. H20 (pH 3)
Pass through C18 Sep-Pak
Wash Sep-Pak with 8 mL each of:
d.H20
hexane
McOH
Collect MeOH eluent
Dry in vacuo
Dissolve in 200 uL CHC13
Deposit on 325 mesh silicic acid column
Wash column with 8 mL each of:
CHC13
CHCl3zMeOH (9:1)
Collect CHCL3: MeOH eluent
Dry in vacuo
Prepare methyl ester (CHZNZ), then TMS ether (BSTFA)

Pace-Asciak procedure for the extraction of oxygenated
fatty acids from rat kidney.

1?

portions, each of which was placed in silanized vials. From these two aliquots the PFB
ester, TMS ether or methyl ester, TMS ether was prepared.

For electron capture detection, the PFB ester, TMS ether was prepared by adding
another 1 mL of methylene chloride and 2 mL of 0.1 M TBA-H804
(tetrabutylammonium hydrogen sulfate) in 0.2 M NaOH to the vial. Twenty-five
microliters of pentafluorobenzyl bromide was added. The vial was vortcxed at five
minute intervals for 30 minutes. The methylene chloride layer (top) was placed in
another silanized vial. The aqueous layer was washed two times with methylene
chloride. Solvent was removed by nitrogen evaporation from the combined organic
fractions.

To form the TMS ether, 25 ul of dry pyridine and 100 |.iJ
bis(trimethylsilyl)trifluoroacetamide (BSTFA) were added to the dried residue. The vial
was tightly capped and incubated at 80 °C for one hour. The reactants were dried by
nitrogen evaporation and the sample was dissolved in an exact volume (usually 200 pl) of
hexane-ethyl acetate (1:1, v/v). _

The methylene chloride from the other aliquot of silicic acid column eluent was
dried by nitrogen evaporation. About 200 “L and an excess of ethereal diazomethane
(ca. 4 mL) were added to the vial and mixed. The solution was allowed to sit for 20
minutes at room temperatrne, then dried by nitrogen evaporation. The TMS ether of the
methylated mixture was prepared and reconstituted as described above.

The samples were analyzed by gas chromatography with electron capture
detection (ECD) (V arian 3740) or flame ionization detection (FID) (HP5890). For the
ECD analysis, a 15-m megabore (0.52 mm i.d., 1.0 um thick film) fused silica column
(DB-1, J&W Scientific) was used; a 60-m capillary (0.25 mm i.d., 0.25 um thick film)
column of the same stationary phase and manufacturer was utilized. Results of these
analyses are presented in Figures 2-5 and 2-6. Aliquots of the PFB ester, TMS ether
derivatives produced chromatograms with at most a few major peaks, corresponding to

18

.38985

3835 3 Cocoa 803 came man 93 one come mum out:
SE m .003 - 09: 5828 SE a: 3 :3 as S9
80% 08 Tmo E 2 .253 EEaE Lo 88:8 0302
.85 Co 550 mg :28 mum Lo 83338020 Gum

“TN 8:me

T 95:: 08c

mum oém mum one“

9 e

 

 

 

 

 

 

 

 

 

AA-j!
‘ l .

 

 

 

 

 

 

 

(— osuodsor

19

t
16.919

response -->

 

 

fin
(AxlA. .

 

 

 

 

 

 

 

 

 

b LLL‘L“ "LipivALL-w-e M

HC 16 HC 20

time (mins) —->

Figure 2-6: FID chromatogram of methyl ester, TMS ether of extract of
mammary tissue processed by Pace-Asciak method. 60 m
DB-l capillary (0.25 mm i.d., 0.25 um film) column, 180° -
280°, 4°lmin. Hydrocarbons HC 10, HC 12, HC 14, HC
16, HC 18 and HC 20 were added for retention indexing.

20

residue from the derivatizing agents as determined by blanks. The other, minor
components were not well resolved by the megabore column. Flame ionization detection
of the methyl ester, TMS ether aliquots produced chromatograms with only one
consistent major peak, which was identified as plasticizer by GC-EI/MS.

Several different trials using this extraction protocol were attempted, but the
methyl ester, TMS ether derivative showed only plasticizer in its GC-FID
chromatograms. Gas chromatography with flame ionization and electron capture
detection of the PFB ester, TMS ether of several trials showed no biological components.
(The plasticizer impurity was reduced by removal of a tygon tubing connection in the
column set-up.)

Not only did this extraction procedure not extract any detectable levels of
biologically-significant compounds, but it was very time consuming. Elution through the
silicic acid column was extremely slow, due to the need to tightly pack the glass wool
plug so the fine stationary material would not fall through the column. Thus, it was
desired to investigate another extraction technique for this application.

3. Use of Funk and Powell Extraction Methodology

Funk and Powell isolated and analyzed prostaglandins and hydroxylated linoleic
and arachidonic acid metabolites from rat and rabbit aortae (45). Their isolation
procedure employed classical chloroform-methanol extraction and sample clean-up on
ODS cartridge columns as outlined in Figure 2-7. All glassware used in these
experiments was silanized prior to use. The sample was sliced into several pieces and
then hand-homogenized in a glass homogenizer with a Teflon pestle in a dry ice-acetone
bath in CHCl3-McOH (2:1) containing an internal Standard and 0.05% butylated hydroxy

toluene (BHT) as an antioxidant.

21

Cut up tissue on ice batlr
Homogenize @ -20° inCHCl3MeOH (2:1), 0.05% BHT, int. std.
Filter rapidly through sintered glass

Concentrate in vacuo

Divide into two equal aliquots
Free Lipid Fraction Free + Esterified Lipid Fraction
[21851113] E . 5 'fi .
Evaporatc in vacuo Evaporatc under Ar
Add 15% aq. MeOH Dissolve in 0.28N KOH in 95% MeOH
Apply to C18 column Incubate @ 55°for 45 min under Ar
Wash column with: Follow extraction as for Free Lipid Fraction
(20 mL each)
15% MeOH
Distilled H20
petroleum ether
pet. ether:CHCl3 (65:35)
Collect pet. ether:CHCl3 eluent
Prepare methyl ester (CH2N7)
Make TMS ether (BSTFA)

Figure 2-7:

Funk and Powell procedure for the extraction of
oxygenated fatty acids from rat and rabbit aortae, as
modified in this laboratory.

22

The homogenate was rapidly filtered through scintered glass under vacuum. The
filtrate was dried by rotary evaporation, then 6.00 mL CHCl3-McOH was added. One-
half of the sample was used to determine the free fatty acids, while the other aliquot was
hydrolyzed to provide analysis of total (free and esterified) lipids.

The solvent in the free fatty acid fraction was removed by rotary evaporation and
reconstituted in 15% aqueous ethanol. An ethanolic solvent was used in the literature
procedure (45, 57), but was changed to 15% aqueous methanol after the first trials to
prevent any ethyl ester formation. Once cthylated, fatty esters will not be converted to
methyl esters during diazomethane methylation. The diluted homogenate portion was
applied to a C18 Sep-Pak solid phase extraction (SPE) cartridge (Waters) prewashed as
described in the preceding section. The column was washed with 20 mL each of 15%
EtOH (or MeOH), d. H20, petroleum ether, and petroleum ether-chloroform (65:35).
The water removes alcohol from the column; the petroleum ether wash dries the column
of the water. The petroleum ether-CHC13 fraction contains fatty acids and monohydroxy
fatty acids, and was collected and dried in vacuo. These mobile phases were pushed
through the SPE cartridge with a syringe to achieve a flow rate of about 1-2 drops per
second.

The aliquot designated for saponification was dried by nitrogen evaporation. The
residue was dissolved in several milliliters of degassed 0.28 M NaOH in 95% ethanol (or
methanol) and the vial was capped under nitrogen. The reaction mixture was incubated
at 55 °C for 45 minutes. It was determined that polypropylene vials were the most
appropriate vessels for the hydrogenation as the basicity of the solution desilylates and
etches glassware, thus causing retention of the polar hydroxylated fatty acids (determined
by retention of radio-labeled hydroxy fatty acid). The alcohol in the solution will leach
phthalates from plastic vials, so these are unacceptable for hydrogenation. Upon
completion of the incubation time, the sample was acidified slightly with HCl and dried

under nitrogen. This step was time-consumin g since the vial contained some water, but

23

rotary evaporation would require yet another transfer of the sample. The residue
contained considerable particulate matter, presumably salts, and was dissolved in 34 mL
of 15% aqueous ethanol (or methanol). The hydrolyzed aliquot was extracted by the
same ODS silica column procedure described above.

The methyl ester of each fraction of the isolated mammary gland was prepared as
described above. Dcrivatization to the TMS ether was achieved by addition of 50 uL dry
pyridine and 100 uL BSTFA to vial containing the dried methyl esters. This reaction
mixture was incubated for 45 minutes at 80 °C. The reactants were evaporated under
nitrogen and the sample was dissolved in an exact amount of degassed toluene containing
0.05% BHT. These extracts were analyzed by GC-FID and by GC-EI/MS.

The complexity of the extract required use of 60-m long, capillary (0.25 mm i.d.,
0.25 um thick film) non-polar (DB-l, J &W Scientific) columns. Initial F11) and EI/MS
(LKB 2091 mass Spectrometer) analyses with similar megabore (DB-1, 0.52 mm i.d., 1.0
um thick film) columns demonstrated that the numerous components were not resolved
(Figure 2-8). Capillary column GC-FID (HP 5890) chromatograms revealed
approximately 120 components in the free fatty acid fraction (Figure 2-9) and over 150
components in the total lipid extract (Figure 2-10).

By comparison of retention indices and fragmentation from EI/MS with reference
compounds, several of the major components of these extracts were identified as fatty
acid methyl esters. Table 2-1 lists the FAMES and their approximate amount ill the free
fatty acid fraction. Methyl esters of the even-carbon saturated fatty acids C14 through
C24 and/or their unsaturated analogues were the most abundant components. Lower
- levels of odd-carbon saturated fatty acid methyl esters (C15, C17, C19) were detected in
the extracts. Also present were phosphate esters and cholesterol. The identities of these

compounds were confirmed by their mass spectra.

24

Flame Ionization Detector

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

.x'! .’
m i if:
«1'
T v:
N
C
0
04
V3 0“
8 A r K! --
34 c: :3
.2 ..= «4
a - - ., E 3?
b I t V'
- 5“: . A‘s . B stir-4 the em
3' mg ': U.$ (.1; V; '3 “nu. . g
.4: N .i
r— 0+
Si‘l
.—
c”i 18:0 ME
time(mins) —)
Mass Spectrometer
: rear: 737344. Tu:
..N.... .._Y , A NJ- ._/\__
I I I "'I'IT‘I‘"‘1‘f"l"'vlvvvfiI-vwvl1..v .
a 25 so 75 100 125 150 175 200 225 250 275
fcan
18:0 ME

Figure 2-8: FID chromatogram (top) and reconstructed total ion
chromatogram (bottom) of the methyl ester, TMS ether of a
free fatty acid extract processed by the modified Funk and
Powell method. Both analyses - 15-m DB-l fused silica
megabore (0.52 mm i.d., 1.0 pm thick film) column .

25

. T823358“

58>... «mam- as: .828 EE 8:: a:

mo :3 88 3.8 .5:an 7mm 8.8 .3508 zesom
28 x85 BEBE 05 .3 33308 585 Sum 98.“

88 e co .850 mg £23 3508 05 .8 89803885 OE

m2 ouom m2 ouwfl m2 OAS

 

88'”

 

83'Si

9t'SE

 

 

93'9i

 

 

 

 

 

 

 

R
‘___‘

-i

'0
h

'II
3.1....
’0...
can.

“.1
65
E'Q'Ei

”em 2:5

T 988v 08:

m2 ONE

 

 

#91
i8]

 

 

 

 

6.74:.

It.)
23.

 

 

<— asuodser

26

 

.mueoueaeoza .88\ev .emum- cow“ £828 86 328 81
S .8 88 3.8 88:88 KB 8-8 .888 838
98 x85 3588 05 .3 38808 88x0 Ba: :29

a co .850 m5 .828 3508 05 Lo 88.898885 GE 57m 88E

m2 is

 

-,..' b
"'3 [i

 

 

 

an.

 

 

 

 

 

tiE'Si
tr9‘9i
Lri'Si
art-fir“
6c
(Cf'fi

If“

 

 

 

 

 

n0
n3

 

 

 

 

 

 

 

 

 

 

 

g
l_‘
L__.

I '8 II. I 't I 6
I. r2 CA... or. 19.... on. ./c
.u... n 03... 36 (hr. on. .T
P? St river P 097. st) 0:
03 to it. I; E 0%»... 0L

té'S 2;;

9i! 8

82 on:
c

j

(

 

T 958v 08.5

 

8 S
698'1
GS 32

w I
C

 

Q9

 

 

 

 

_ :4 a.
.flu
.J
9
S
d
9. 0
TC w
E 9
g. E
O J
.hf
6 .
.. L
01.. rhe-
." .0...-
/L n. ..
r" 1 D1:

Table 2-1:

EM

14:0

14:19

15:0

16:0

16:19

17:0

18:0

18: 19

18:29-12

19:0

20:0

20;38,1 1,14

2031 1.14.17
20:45.8,1 1,14

22:0

22:19
22:64.7,10,13,l6,19
24:0

24: 1

unidentified components

27

Fatty Acid Composition of Free Fatty Acid Extract

Amount
ins)

420
87

5200
230
1010
9120
700
510
700
330
880

350
430
410
360
170
330

0.2
0.04
0.2

0.1
0.4

0.4
0.3
0.4
0.2
0.5
0.2
0.2
0.2
0.2
0.2
0.1
0.2
88%

28

There were numerous other peaks in the FID chromatograms of the extracts
which did not have retention indices corresponding to a derivative of a common fatty
acid. Mass Spectral analysis did not allow identification of numerous components, either
due to inability to interpret the mass spectrum (no characteristic fragmentation) or the
inability of the detector to accurately record the true mass spectra of the clucnts.

High levels of unsubstituted saturated and unsaturated fatty acids in the mammary
gland could provide interferences in the detection of any oxygenated fatty acids possibly
present. Any oxygenated fatty acid would be expected to be a minor component, in the
parts per million or less range. For a 100 mg tissue, this would correspond to 100 ng in
the combined extracts (assuming 100% recovery). Thus, if there were any overlap in the
elution of a the methyl ester, TMS ether of an oxygenated fatty acid and the methyl ester
of one of the unsubstituted fatty acids so abundant in the mammary gland, the FID or ion
signals from the oxygenated component would be dwarfed by the signals from the latter.
This overlap is very possible, as illustrated by the retention times of the methyl ester of
arachidic acid and the methyl ester, TMS ether of 13-OH 18:0 (Figure 2-11).

4. Hydrogenation of Extracts

One way of correcting for the possible chromatographic overlap is to Simplify the
extract. Hydrogenation of the extract will saturate the unsaturated fatty acids, decreasing
the possibility of chromatographic overlap of the high-level unsubstituted fatty acids with
any oxygenated fatty acids. After reduction, any unsaturated hydroxylated fatty acids of
the same carbon number and same hydroxy position will coelute, merging the signals.
This would possibly increase the signals from any low-level oxygenated fatty acids.

The extracts were hydrogenated after the methyl ester was prepared by a classical,
platinum-catalyzed reaction (58). The methylated extracts were dissolved in methanol in
a silanized vessel. Initially, a side arm flask fitted with a ground glass connection to a
gas buret was utilized. Later a gas manifold was constructed to permit simultaneous

reaction in four screw cap vials. Hydrogen from the manifold was leaked into the

29

. . INIO'I'I'.\O' ’0'--.IIII("..I. .Y'Bel3'! loo -I'e'lll-I

A

Jfi‘itrl‘lu-ll‘l)lljf{e‘i".ne elli.enillel(.o..le III...

W‘II‘IDO'.‘.DCQ)O ‘J‘x’l-"I...II’ .T.‘ oi} '1‘- '.1l ....... .' $0.."

.I v -
irififérrzrwbxe f.c__l}\__\rmx.lr<t/5%....5.19“,— _—.....x)_.’_\</ % xx xxx-3% \ far. 31!. H £115....5x...\L’(../.}€..c ‘JM/qu
m m.

11...... .l..“f..e.;‘le’.l(o o... 0.1.- 'Ill1'.Oc-'.I ,. .clei 8"}

.838 3508 as 88888: 28 0:. 05 E coon 850
.28 we 08 82 one” 88 28 8 o: 82 3: 82.8-2
Co 9826 03888988020 Rue-ken MES-52: .Aoon-om
88 8858 8.8860 88 m. 88 m: in .2: .sm

£8 8 8888988020 $88 83 com 39 Buosbmeooom ; Tm 083m

m2 8: $8.08

838:: 58 «
emu“ game .

3 ." "Ib‘l‘ ‘- O 'L'Ii‘ "0... "0 JP

. b i s P . b h .

3mm
b an. .0-..I‘IhTO-IP'II

\..’ xii-£11.... .\-!')<IT’S“' .Q'rltiifive 0" n’]- It]? ‘-Il-’\“O\I.OA

< u

. elite-1.0.3.40; 5‘31.'11t01}(\.0.1033..l01/ \_\lu’a\u.tltlll\4\l\.l
..

[/3 m2

m2 3: 92.5-2 li- __ _ m:
_ h

‘Otc‘i0.eT V’-.-IO ‘lee‘tio‘I'T 0 '0':‘\ .u'zeo'0.'. el_“ .DII- ft ‘T'I-e-c“l . do. ".OO'¢TI'¢T.OII| -OO""1.'0 O

I. \\ mzcuom

J _

1 i ..

fiam

\II....CI 0" 'l "’ Col

 

.84. Rafi. __._.__._ <

U:-

 

 

4

60 n... E..-

fi) CID?— ...... JI'IOeuon)?ll\"-"--oe‘0"00.00 ' \D'“. 9' .1 l‘f't... f";.el v

el.'90 .0 l‘ltlf 0......Iile'lu'1'

v

I

v

 

. .
m: l - - m; l v.8 t i .
A825 2:: 5:58

;Mfltll

lIxrzc-udcon

H mmm . .m xe:

30

reaction vial by a needle through a Teflon-lined septum. A stirbar, a very small amount
of platinum dioxide (Aldrich), and one drop of acidified methanol was added to the
vessel before it was sealed. The vessel was evacuated and magnetically stirred. The
vessel was purged with hydrogen and evacuated, and the cycle was repeated. If the
reaction vessel atmosphere is hydrogen and the oxygen has been replaced, the platinum
oxide reduces to platinum black. A positive hydrogen pressure was left on the vessel for
20 minutes. The reaction mixture was filtered on Whatman No. 1 paper, and rinsed with
methanol three times. Care was taken to keep the paper wet, as the platinum could spark.
After rinsing was complete, the paper was doused with water. The filtrate was dried
under nitrogen evaporation, and the TMS ether was prepared as described above.

The FID chromatograms (Figure 2-12) of hydrogenated and non-hydrogenated
portions of a total lipid extract show the extent of reduction. The chromatograms of the
extracts were greatly simplified. Noting the shift in retention time of certain peaks
signals the presence of unsaturated components.

The efficiency of the hydrogenation was verified with standards of high levels (20
and 5 mg) of unsubstituted fatty acid methyl ester and 10 and 0.5 ppm, respectively,
hydroxylated FAME (Table 2-2). Analysis by GC-FID showed that under these
conditions of sample level, hydrogenation of these fatty acid methyl esters was complete.
It should be noted that these conditions were not robust enough to reduce all of the
unsaturated fatty acids in some of the extracts. In particular, methyl esters of oleic and
linoleic acid were detected in some of these extracts. This is due to the high amounts of
unsubstituted fatty acids in the mammary gland, a fatty tissue. It is this high content of
. unsubstituted fatty acid that made analysis of hydroxylated fatty acids with these
extraction procedures very difficult. There were no mass spectral data that indicated any

derivatized hydroxylated fatty acids were present in the fatty acid extracts.

31

4 . 3 3:
N u zt<
eo<mkxm
322303»: ._ _
F u 3.2 A _
_»o<mexm
omEzmoomoEéoz L : m m g

.5953 .owmm

. .9: 528 Asa 8: 3 :3 as 2.9 bangs 35

E om .Smbxo Ba: :29 .«o 30:33 A8988 Engaged»:
98 30c couscomofimaéoc mo 9.8838885 DE "NTN 8sz

T 08:

ms. 06—. us. on:

a «

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

<— osuodsar

32

Table 2-2: Efficiency of Hydrogenation Reaction

(ms) Etfirim ma Efiifiifinfl
1 20 99.6% 100 100.0%
2 20 100.0% 100 100.0%
3 20 100.0% 100 100.0%
4 5 100.0% 50 100.0%
5 5 100.0% 50 100.0%

111. CHAPTER THREE: DEVELOPMENT or NEW ISOLATION PROCEDURE

The high fatty acid content of the mammary gland provided a high chemical
background in the determination of any possible trace hydroxylated fatty acids when the
isolation procedures described above were employed. These isolation procedures were
developed in other laboratories for the analysis of tissues with much lower lipid content,
and have been found in this study to be unsuitable for the analysis of mammary tissue.
After isolation by the above-mentioned techniques, the extract has found to contain many
components, many of which coeluted, even with high-resolution chromatography. Thus
a new extraction technique needed to be developed to separate any possible hydroxylated
fatty acids, which would be expected to be present only in trace amounts, from the
unsubstituted fatty acids, which constitute a large portion of the mammary gland.

The original intent of the new extraction procedure was to perform a class
separation of the oxygenated fatty acids in the mammary tissue from the non-lipids and
the unsubstituted fatty acids. Thus, the extract would become greatly simplified and
some potentially coeluting compounds would be eliminated. This goal was restricted
somewhat by the use of only hydroxylated fatty acids as model compounds during the
development and assessment of the new isolation procedure. This decision was made
because some standards for these hydroxylated compounds are readily available and
relatively temperature- and light-insensitive (with respect to the analogue keto— and
epoxy- fatty acids).

A. Development and Description of New Isolation Procedure

1. Solvent Extractions

The first steps of the mammary gland extraction are to homogenize the tissue and
to separate the lipids from the non-lipid matter. To achieve this latter goal an aqueous
phase extraction (52, 53) was performed after the initial chloroform-methanol (2:1)
solvent extraction (45). (All glassware was silanized before use). For a tissue of 0.1

gram, two milliliters of CHCl3zMeOH (2:1) was used during homogenization in a dry

33

34

ice:isopropanol bath. After the homogenization, which was mechanically driven by a
drill motor, the sample was quickly filtered through sintered glass under vacuum and an
additional 0.5 mL of the solvent mixture was used to quantitatively rinse the glassware.
The filtrate was collected into an inert vessel, and 20% by volume (0.5 mL) deionized
water was added.

The mixture was capped and vortexed, then separated in a bench-top centrifuge at
about 500 rpm for one minute. A polypropylene test tube was found to be the inert vessel
easiest to use and prepare . No phthalates were found by gas chromatography-flame
ionization detection (GC-FID) in blank solvent systems (CHC13, MeOH, and water)
which were allowed to sit in the polypropylene test tubes for 15 minutes. The test tubes
readily fit into the centrifuge and into the side arm flask used for vacuum filtration. The
polypropylene test tubes required no preparation, unlike glass test tubes, which must be
cleaned and silanized before each use. There was visible material in the aqueous wash
which separated from the organic layer. It was determined by GC-FID that no lipids
were extracted into the aqueous layer, even when the mammary gland was spiked with a
mixture of unsubstituted and hydroxylated fatty acids just before homogenization.

2. Methylation

The organic layer was transferred to a one dram, screw-top vial and dried under
nitrogen. The lipids were methylated with diazomethane and methanol at room
temperature for 20 minutes, as described in the previous chapter. At this point the extract
contained methyl esters of all of the lipids originally present in the homogenate (tissue
plus any internal standard added). As described in the previous chapter, a portion of the
extract can be removed at this time for acidic, platinum-catalyzed hydrogenation.

3. Hydrogenation

This hydrogenation step was originally performed in individual vials and was
redesigned to a simultaneous and less labor-intensive process. A gas manifOld was

designed in this laboratory and crafted in the Chemistry Department Glassblowing Shop

35

with four outlet ports, each with an individual Teflon vacuum st0pcock (Figure 3-1).
Each outlet was connected to an 18- gauge needle which easily introduced hydrogen into
the vial through a Teflon-lined septum in the screw cap. To rid the vial of oxygen, a
vacuum connected to the manifold could be used or the vial could be purged via a 22-
gauge needle through the septum during the initial three to five minutes. Contents of the
vial were stirred by magnetic stirrers below the vials. A small positive hydrogen pressure
was applied to the manifold to replenish any hydrogen consumed or leaked. After 20
minutes, the hydrogenated extracts were filtered and dried as previously described.

4. Normal-Phase Chromatography

A column chromatography step was developed to separate the methyl esters of the
unsubstituted fatty acids from the extract, leaving the hydroxylated (and possibly other
oxygenated) fatty acid methyl esters (FAMES). Since separation by published methods
with octadecylsilyl C18 (non-polar) columns did not appear to fulfill this goal, normal
phase chromatography with silica (polar) columns was investigated.

In the 19505, early work with silica columns utilized different proportions of
diethyl ether-petroleum ether as solvents for elution of lipid classes. But Homing and
coworkers (59) utilized benzene-hexane mobile phases to elute various lipid classes from
silicic acid columns. The latter mobile phases were easier to use, partially due to its
lower volatility. The increase in greater differences in percent composition with the
benzene-hexane versus the diethyl ether-petroleum ether also makes the benzene-hexane
system more desirable.

Initial studies in this laboratory were conducted with a standard mith of
unsubstituted and hydroxylated FAMES, both alone and in the presence of mammary
tissue extract to determine the percentage of benzene in hexane needed to remove the
unsubstituted FAMES from the column, while keeping the hydroxylated FAMES (OH-
FAMES) adsorbed to the stationary phase for subsequent elution. The first trials were
performed with silicic acid (Sigma 325 mesh) in a 2.5 mm i.d. column, slurry (hexane)

36

53088386 3383 58 22%83:
9 com: 28:38 new 2: mo couanmoEB ogEofim "Tm 0.8me

   
  

2002
323.1 I. l. I. l.
338-8233 A“ w Am g

 

Ensow> 0a

 

 

37

packed by gravity to a bed height of about 40 cm. After dilution in hexane, the standard
or spiked tissue was applied to the column, which was rinsed with one column volume of
hexane. Mobile phases of 6%, 20%, 30%, 40%, 50%, and 60% benzene in hexane;
CHCl3zMeOH (1:1); and 100% MeOH were prepared and applied to the column
sequentially. Fractions of eluent from each mobile phase composition were collected,
dried under rotary evaporation, and analyzed by GC-FID (DB-l, 15-m, 0.52 mm i.d., 1.5
um thick film). The results of these analyses for a mammary gland spiked with the
standard mixture are given in Figure 3-2. The unsubstituted FAMES began to elute when
the 6% benzene in hexane mobile phase was applied, and 20% benzene in hexane
appeared to be the mobile phase in which the last of the unsubstituted FAMES were
removed from the stationary phase. The hydroxylated FAMES began to elute with the
50% benzene in hexane and were not completely removed from the column until
CHCl3zMeOH (1:1) was used as a mobile phase.

This large gravity-fed column probably contained an excess of theoretical plates
and made-the procedure very time-consuming, even when a slight head pressure of argon
was applied to the column to achieve a flow rate of about 2 drops per second. The
process was further improved by the use of solid phase extraction columns and a vacuum
manifold with Teflon stopcocks (Burdick & Jackson). Silica SPE cartridges (Burdick &
Jackson, 500 mg) were pretreated by rinsing with two to three column volumes of
CHCl3zMe0H (1:1) and then with enough hexane to fully saturate the stationary phase.
The vacuum manifold provided simultaneous pretreatment of twelve columns and elution
of up to three columns with a constant, controllable flow. With the addition of adapters
. and 50 mL reservoirs, the mobile phase could be loaded onto the apparatus, requiring
much less attention from the operator. The claims of inertness of adapters and reservoirs
from different manufacturers were tested with washes of CHC13 and MeOH and

megabore GC-FID, and the Burdick & Jackson parts were found to be the best.

 

 

 

response

 

TR

figme32:

 

 

 

 

 

 

38

 

 

{V 6% BENZENE IN HEXANE
g
100% BENZENE

LC

 

tune

 

D

 

Megabore column (DB-l, 15-m long, 0.52 mm i.d., 1.5 pm
thick film) GC-FID chromatograms resulting from analysis

of spiked mammary gland extract eluents from selected

mobile phases through silica SPE columns. The bottom
chromatogram is 12-OH 18:1 ME and 12-OH 18:0 ME in

the eluent from 100% benzene (1.2% of the eluent injected; - '
attenuation=3). The top chromatogram represents a ’
mixture of 14:0, 16:0, 18:1, and 18:0 MES (1.7% of the

eluent injected; attentuation=8). GC conditions: N2 flow

rate 15.0 mL/min; temperature program-160° for 4 min,

then 8°lmin to 280°.

39

The new silica SPE system was implemented with the following mobile phases:
a) application of sample dissolved in ca. 2 mL hexane
b) 7 mL hexane to rinse SPE column
c) 50 mL 35% benzene in hexane to elute the unsubstituted FAMES
d) 50 mL CHC13 to elute the OH-FAMEs
e) 50 mL CHCl3zMeOH (1:1) to determine if all OH-FAMEs eluted.

The last column washing step was eliminated after several trials of mammary
gland spiked with the standard fatty acid mixture demonstrated that no hydroxylated
FAMES remained on the column after elution with 100% chloroform. The volumes of
the benzenezhexane and chloroform mobile phases were reduced to 20 mL, with no loss
in efficiency of elution of the desired type of fatty acid methyl ester.

5. Trimethylsilylation of Alcohols

The eluent from the 100% chloroform mobile phase was collected in a round
bottom flask and dried under rotary evaporation. The residue was quantitatively
transferred to a conical reaction vial with methanol and dried under nitrogen. The
trimethylsilyl ether was prepared with bis(trimethylsilyl)trifluoroacetamide as described
in the previous chapter. The derivatized extract was reconstituted in an exact amount
(usually 100 pl.) of dry hexane:ethyl acetate (1:1) containing 0.05% BHT and 5%
BSTFA. The new isolation procedure is summarized in Figure 3-3.

6. Reduction of Contaminants

Throughout the development of the new isolation procedure, care was taken to
minimize the amount of blank interferences from the apparatus and chemicals. The
selection of Teflon and inert plastic (polypropylene) parts for the vacuum manifold and
the choice of polypropylene test tubes for centrifugation were described earlier in this
chapter. All solvents used were analytical or I-IPLC grade. High boiling impurities were
discovered by GC-FID in I-IPLC grade chloroform. After unsuccessful attempts to dry or

chromatographically clean the chloroform, distillation removed the impurities. To

l. Mechanically Homogenize Tissue in CHCl3/MeOH (2:1) with 0.05% BHT (2 mL).
2. Filter through Sintered glass.
3. Aqueous Wash (0.2 Volumes).
4. Separate Organic Layer and Remove solvents with Nitrogen.
5. Methylate with ethereal CH2N2/MeOH.
6. Remove excess reagent and solvent with Nitrogen.
7. (Optional Hydrogenation - Pt catalyst, positive H2 pressure).
8. Silica Column Clean-Up:
a. Apply Sample in Hexane.
b. Wash with 35% Benzene in Hexane to Elute Non-Substituted Fatty Acids.
c. Wash with 100% CHCl3 and Collect Eluate of Oxygenated Fatty Acids.
9. Remove solvents by rotary evaporation.
10. (Optional Conversion of Ketones to Methoximes.)
11. Trimethylsilylate Hydroxyl Groups with BSTFA/Pridine
12. Dry excess reagents with Nitrogen.

13. Add 50 or 100 [AL 0.05% BHT in Solvent/BSFTA (95:5), Solvent is Hexane:EtOAc
or Toluene.

Figure 3-3: Summary of new procedure to isolate hydroxylated fatty
acids from mammary glands. See text for detailed
description.

41

minimize chlorine formation, 0.05% ethanol was added to the pure CHC13;fortunately,
this did not appear to cause reintroduction of the high boiling impurities.

Selection of the best commercially available SPE silica column was based upon a
determination of which generated the lower amount of background contaminants as
analyzed by GC-FID analysis of column washes with distilled chloroform. Two SOO-mg
silica SPE columns from each of Burdick & Jackson and J.T. Baker were pretreated with
CHC13:MeOH (1 :1) and hexane as previously described. Then 40 mL of distilled CHC13
were run through each column and collected in individual round bottom flasks. The
flasks were dried in vacuo and the residue was transferred with methanol to a one-dram
vial. The contents were dried under nitrogen, and 200 uL of hexane:ethyl acetate were
added. The reconstituted eluents were analyzed by megabore column (DB-1, 15-m long,
0.52 mm i.d., 1.5 um thick film) GC-FID. The gas chromatogram of the eluents from the
SPE column manufactured by Burdick & Jackson had fewer impurities, especially in the
area of retention of the higher boiling FAMES (Figure 3-4).

B. Detection Techniques for Analysis of Extracts

A summary of historical and current procedures for gas chromatography with
flame ionization detection or electron ionization mass spectrometry is presented in
Chapter Two. In order to resolve the large number of components in the mixture of
methyl ester, TMS ethers from the mammary gland extract that was processed by
adaptation of the Funk and Powell method described in Chapter Two, a very long (60-m)
high-resolution capillary column (DB-l, 0.25 mm i.d., 0.1 pm thick film) was needed.
The final derivatized mammary gland extract from the new isolation procedure described
above contained many fewer components, particularly in the region of elution of the
methyl ester, TMS ethers of the C18 hydroxylated fatty acids, due to the absence of the
unsubstituted FAMES (Figure 3—5). The GC:FID chromatograms of extracts from the
new isolation procedure did not contain any peaks which were thought to represent any

of the unsubstituted FAMES which were detected in the extracts processed by the Funk

42

 

 

 

 

 

 

m
‘1’ m
-- can
a a“
«S WC:
on
A"
37
a, it
\D
‘1‘:
9:: 25.
.n
or? Z
"" o
I O
. 5
’RL
d.)
8
8.
a)
a mm:
“on
“as

3%

 

DII-lKl

 

18:0 ME 20:0 ME

time-9

Figure 3—4: Megabore column (DB-1, 15-m long, 0.52 mm i.d., 1.5 tun
thick film) GC-FID chromatograms of chloroform solvent
blanks through silica SPE columns from Burdick &
Jackson (bottom) and J .T. Baker (top) with elution times of
some FAMES designated. GC conditions: N2 flow rate
15.0 mL/min; temperature program- 140° for 4 min, then
8°/min to 280°; 2 mL injection volumes. Attenuation=5.
The marked peak has an area equivalent to the area of the
signal from 0.1 pg 18:0 methyl ester.

43

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

.33 .3
" 5°. a
% a
‘37 lfflh'; . i g
:! ufii .- 3
T E I 'm '
o 1’
(I)
r:
0
CL
(I)
9?.
sea a gage a ea era '2
cacao; a; 2:222 swag: r.
r z... '
F $11“
$2.; 5
. $3; _
z 2 M
1
16:0 ME 18:0 ME 20:0 ME
time —9
Figure 3-5: Capillary column (DB-l, 60-m long, 0.25 mm i.d., 0.1 pm

thick film) GC-FID chromatograms of 4.0% of extract from
new isolation procedure (tOp) and 0.5% of total lipid
extract from modified Funk and Powell Method (bottom)
with elution times of some FAMES designated. GC
conditions: He flow rate 1 mL/min; temperature program-
180°, 4°lmin to 320°; attentuation=3.

44

and Powell method (Table 2-1). Mass spectral analysis confirmed that no C16 nor C18
unsubstituted FAMES, nor phosphate esters, were present in the extracts processed by the
new isolation method. The TMS ether of cholesterol was identified by GC-FID and GC-
EI/MS to be a very late eluting component of extracts from both procedures. Many of
the extracts processed by the new isolation procedure contained minor components which
were confirmed by GC-EI/MS to be TMS ether, methyl esters of C18 hydroxylated fatty
acids (see Chapter Four for details). These OTMS C18 methyl esters (nor any other
OTMS -FAMEs) were not observed by the same mass spectral or flame ionization
detection methods in the extracts from the Funk and Powell extraction method, possibly
due to the high level of other lipids, particularly unsubstituted FAMES, in the region in
which the former compounds elute. Thus, only a 30-m capillary column (0.25 or 0.314
mm i.d., 0.25 um film) was needed to resolve the components near the retention indices
of the C13- and C20-hydroxylated FAMES (Figure 3-6).

Flame ionization detection provides a sensitive and continuous method of analysis
of the column eluent, but its capacity to identify components is limited, since retention
indices must be known from standards, which are not always available. Mass
spectrometry, EI/MS, can provide additional structural information to assist in the
identification of compounds which are not resolved chromatographically or are not
readily identified by their retention indices. Specific fragmentation characteristics of the
methyl ester, TMS ethers of hydroxylated fatty acids were described in the previous
chapter.

1. Electron Ionization/Scanning Mass Spectrometry

The methyl ester, TMS ethers of the monohydroxylated C18 and C20 fatty acids
have molecular weights in the 380-420 u range. Although a molecular ion [M]"'° is not
always observed with EI/MS, loss of a methyl group [M-15]+ is a common fragment ion,
strongly suggesting that the mass range of the mass spectrometer be extended to at least

m/z 430. Characteristic tit-cleavage fragments give intense peaks in the mass spectra of

45

l

299'
£3§a 951

.mucoamacozm ”ommm 9

58>...» .oOmTEEmBq 8380922 “£2348 a 8m.“ >5: om
”202980 00 .cozmcwmmoe mm2<m 088 .«o 88m 3320
SS, 2:385 comm—8m Bo: 89a zombxo Essence»:

he s? we sameneoao Bade Ea fie a:

mac :3 as 38 .meg 8-8 .38 5:28 5:36

MU“ [I I . I ole
. so, ’ 'I I I o I
a... . . [tr-Pt . . 0 ”Dub
.7 909” 79/57 t. .. .5). 9 019.1. to _I.
CL 839. 8 but i 8 &O . . 9
a x. 9 3m... .9. My. .: n m1: m.
9 CS.” £9
92 I.
.L
I

19V”

822‘8

 

[299

”mm 29mm

836'9

 

239's

T 95.5 new.

 

 

(9Y9

 

 

 

 

 

 

natal.
Outd-
bah".

  

(— osuodser

a...
m»
can...

46

these compounds and they provide important structural information, therefore it is critical
to monitor these ions. A list of these a-cleavage ions from the methyl ester, TMS ethers
of the possible monohydroxy isomers of stearic acid is given in Table 3-1. The range of
these ions is large (m/z 103-327) and the instrument's mass range must be extended to
encompass all of the fragments. At the lower m/z values, other characteristic fragments,
especially those for the trimethylsilyl fragment (m/z 73) and the McLafferty
rearrangement fragment at m/z 74, are observed. Thus, it is ideal to analyze all of the
ions in the range m/z 70-430 in order to record all of the ions which may give structural
information and help identify the methyl ester, TMS ethers of hydroxylated fatty acids.

With sector mass spectrometers, the range of ions to be analyzed is scanned
repetitively in a finite time. The time of ion sampling and the rest time can restrict the
instrument to only one to a few scans of a large mass range, such as m/z 50—500, per
second. This causes possible distortions in the mass spectra for compounds that are
eluting from the high-resolution column in a very short time span (60). If the duration of
elution of a component is only two to four seconds, then the partial pressure of that
compound would change significantly during a typical 0.9 second scan time. The
relative peak intensities would be distorted from those observed when the compound is at
a constant pressure in the mass spectrometer source (Figure 3-7). When the repetitive
scan rate is low, another problem encountered is that the reconstructed mass
chromatogram is composed of only a relatively few data acquisition points during the
short elution time of the compound, thus distorting the apparent elution profile.

The highest scan repetition rate possible for the particular instrument is desired
when high resolution chromatography is utilized. However, conventional magnetic
sector instruments are severely limited by the rest time needed to correct for the
hysteresis of the magnetic field, limiting scan rates for the m/z 50-500 range to less than
two sweeps per second. On the instrument available for this project, the greatest scan

speed requires 0.9 second per scan of m/z 50-500.

47

Table 3-1: List of m/z Values at which Peaks for a-cleavage Ions
Would Occur for Isomers of OTMS-18:0 ME

15ng W

ml: ml;

(a)2-OTMS 161 327
(BB—OTMS 175 313
4—OTMS 189 299
S-OTMS 203 285
6-OTMS 217 271
7-OTMS 231 257
8-OTMS 245 243
9-OTMS 259 229
lO—OTMS 273 215
ll-OTMS 287 201
12-OTMS 301 187
13-OTMS 315 173
14-OTMS 329 159
lS-OTMS 343 145
16-OTMS 357 131
17-OTMS 371 117

lS-OTMS 385 103

48

 

True
Chromatogram

Reconstructed
Chromatogram
l l I I l I

& Spear-a

(a) Acquisitions

 

 

gmMmgraphrc Mass Spectrum

Aw- ”M Ideall—Ll—‘LLJ-

Acquisition
(b ) \ «mean

 

 

 

Figure 3-7: Two problems caused by inadequate spectral generation
rates in GC-MS. Top: True chromatographic profile is
poorly represented due to an insufficient number of TIC
points. Bottom: Mass spectral peak intensities are skewed
due to the change in partial pressure of the analyte which
occurs during the scan acquisition time. Source: Reference
61.

49

A disadvantage to such rapid scanning is the decrease in ion counting statistics,
lowering the signal to noise ratio. This signal to noise ratio will be improved only if the
time allowed to detect a particular ion current is increased (Nyquist theory). Therefore,
limiting the mass range scanned will allow a higher scanning rate since the magnetic field
will not have to change as much and the hysteresis will not be as great, allowing a
shortened rest time. With a shorter rest time and decreased scanning time, the scan
repetition rate will be increased for a limited mass range. This will tend to correct for
some of the inaccuracies in recorded relative peak intensity and simultaneously allow for
more scans while a given compound elutes, thus improving ion counting statistics and
increasing the signal to noise ratio.

2. EI/MS with Selected Ion Monitoring

Selected ion monitoring (SIM), where only a few ions are analyzed and recorded,
has been used quite often when specific ions of interest are known and a more sensitive
technique is desired. But the limitations of this technique for analysis of numerous
possible compounds are great; usually only ten or fewer ions can be monitored during a
single chromatographic run. If the abundant, characteristic tat-cleavage ions are selected
to identify the methyl ester, TMS ethers of hydroxylated fatty acids, then only five or
fewer compounds may be analyzed during a given interval following each injection.
Since these chromatographic runs last 40-45 minutes each, searching for several isomers
of different compounds could be very time-consuming and require several injections of
the extract, which is available only in limited quantity. Another disadvantage found with
SIM is that should the selected ions have m/z values which are too widespread, the scan
repetition rate is set by the instrument to be the same or lower than during scanning over
the same range.

An alternative, compromising method is to scan rapidly over a narrow mass
range, particularly a higher mass range. The sweep of the magnetic field in most mass

spectrometers is not linear, but is an exponential function, scanning at an increasingly

50

faster rate as the magnetic field, and thus the mass/charge ratio being measured,
increases. Thus, the mass range m/z 200-300 is scanned more quickly than the range m/z
50-150. With the magnetic sector instrument used in this study, the range m/z 170-330
can be scanned in 0.6 seconds per cycle. The signals observed for a-cleavage ions of
standard ME, TMS ether hydroxy fatty acids under the limited scanning method were
comparable in intensity to those for the same ions in SIM mode when only ten ions in the
range m/z 173-315 were observed. The 0.6 second per cycle scan frequency over the m/z
170-320 range allows more frequent scans than does SIM of these ten ions (1.0 second
per scan file). The short duration of elution of the compound means that the additional
scan(s) during analyte elution can provide peak enhancement approaching that achieved
with the limited ion sampling SIM provides.

3. Determination of Limit of Detection

The limit of detection of an analyte in a real sample can be determined from the
linear expression for its calibration curve (62,63). The signal of the analyte at its limit of
detection, yd, is equal to at least the sum of the signal from the blank, yb, plus three times
the variability in the blank's signal, which can be approximated by the standard deviation
of the slope, s5. (yd = Yb + 35s) The value of three as the factor for the variability is
liberal, giving only a 7% probability of false positives. Since this is an acceptable level,
this study used three as the variability factor. The concentration at the limit of detection,
Cd, is derived from the slope, A, of the calibration curve (Figure 3-8). Thus, Cd 2 (yd -
b)/A z (yb + 3ss - b)/A, where b is the calculated y—intercept. When no sample blank is
available, the intercept is an acceptable estimate of Yb and the equation simplifies to Cd
2 (yb + 338 - yb)/A = 3ss/A. Since it is not known if these hydroxylated fatty acids exist
in these tissues, no blanks were possible. Two isomers, 12- and B-, of OTMS 18:0
methyl ester was chosen as the model compound on which to base the estimate of the
limit of detection. In this study, the calibration curves were plotted as quantity injected,

Qx, vs. signal, not concentration vs. signal, but the equation is easily adapted to

51

CALIBRATION OF 12-OTMS 18:0 ME

 

8.00

6.00 "

AREA

4.00

2.00

 

   

y = 1.9669e-2 + 0.68448x R"2 = 0.991

  

 

   

 

Figure 3-8:

 

‘I

I 1

‘ .
8.00 10.00

V

' I I
2.00 4.00 6.00
AMT. INJECTED (NG)

Calibration curve and equation for the determination of 12-
OTMS 18:0 ME by GC-EI/MS with SIM at m/z 187.

 

12.00

52

Qd 2 3531A. A sample calibration curve is shown in Figure 3-8 and the results for this
definition of limit of detection for two isomers (12- and B-) of OTMS 18:0 ME under the
different mass spectral scanning modes described above are given in Table 3-2. A Qd of
0.10 ng corresponds to 67 parts per billion (ppb). based on a typical 2 [.11 injection of 100
pl total extract from a tissue weighing 75 mg.

These estimates of the limits of detection are a function of the ratio of the
abundance of the ion used for detection to the analyte's total ion signal, that is Yi/YT-
Thus even if the ion of highest abundance (base peak) is chosen for detection for each of
two analytes, the limits of detection may not be similar if the abundance of one base peak
is of a different proportion to the sum of all of the ions from that analyte. It is seen that
the lower limit of detection must be evaluated separately for each analyte and its most
abundant ion(s), unless these ions have the same Yi/YT

Other factors which greatly influence the estimation of limit of detection are the
slope of the calibration curve (the sensitivity) and the variability in the slope. A larger
slope for the calibration line (a more sensitive method), will lower the limit of detection.
This is usually caused by a higher analyte signal and has been achieved with the use of an
internal standard for the calibration curves while scanning from mlz 170-320. The
variability of the slope, that is the precision of the method, is a function of the number of
points used to calculate the regression line. As the number of points increases, the
method should become more precise, decreasing the standard deviation of the slope ( and
the signal of the blank), therefore lowering the limit of detection.

These statistical estimates of the limit of detection can be compared to another,
but less rigorous, method of determining the minimum signal that can be considered a
positive measure of the analyte. That method is to calculate the signal to noise ratio
(SIN) for the suspected analyte response, with three being the lowest value generally
accepted. Figure 3-9 shows a sample selected ion current profile in which SIM of ten
ions in the range mlz 173-315 were analyzed and the S/N was calculated to be five. This

53

NE 92 mEO-u 658:5» 358.: u 2N ~68 .~.2m .N._om 4an £an dag .mdmm K2 .n: .22 a): .. BEA—SB «=2 5.:

ag8n
53d

2:8a
a3

owed

and

nESN
he...

2 we 8m
> 3

an m: 8."
$85

26
and

5nd

:6

2

2

2|

Rod $3
085 E..—
23 Red
and 8.“
God Em
39¢ 2:
23 some
e85 366
Sod 22.
«to m:
25 mom
a... name.
a?

823:3 .o :88? co Beam e38— m2 3: £5 5.. 5%er co :5:

836-

836

mused

nnd-

256.

886

886
unwed
N2 .o

82.

2mm-

nag

96

n2 mac."—
96

22 92.0.2
52 32.0.2
woe—«Eco: .96

52 22.5.2
52 3.5.0.2
52 2290.2
n: . mEQn
>2 22.0.2
n: 34.5.”.
m2 2290-2
52 2290.2

33

m3 «Sufi

sensed
82: S.

29836
ommét QE

26
:8 2: 2a

sees?
82“ as

2933
82: S:

ensued
83: NE

sensed
88: as

:5 a: 26
use. S 26
~22 2 25

sesame
82“ as

«Halal: £58

”N4“ 03$.

54

GEEV 0E: 52:82

.33 as 033 28328

ma 2 B .2223 05 .6 ma 2 .6 :38? 9:829: :83...

BF .22 92 mSFO-2 00 com 03.63085 wcccomeme

.N-m 032. E was: 82 :2 05 00 Sum 523 $9323
$65.00 :80 n5 £8 8 20qu 88:8 com 382% ”Qm 23.5

0.00« 00.00“ vmv0.0 0.00“ 00.00« wmmm.0 2H mvm vmm 0mm ..mm.m« «

 

 

 

 

 

 

 

 

 

x flmm ucofimx x flmm mmc< ma>k 0cm unmum ao._. mgr—Hm .02
«\a moon «Olwmmmmmm.m 59.? HmuOF
« *Hmccmco « 0030.5 000.95." 0 man:
00.0 >4 ”.moava A..mm.m$ samba on ..m«.v« .E 8090mm 0» 0mm ficoom
“3:5: 58
amm svm mmm mom 0mm 8 m
.T'lz...rnl. pl! L. h p ..al..-..,.-..!:.l_ . llr ._|.I.I.L . pl _ LII}.-. .ol-alLo:ll.l*m
10
m: N? mom
T
wow
#00
0
"tom
.7
_.
TI-I' . alulllldl .. . q 0‘ 4 q q 1d 41 0 4 1| 111... Incl-l4... In. .l'.-nlllo1.'JlJ .vvndCi latillnHGGH
.3 H 9

55

peak was determined to be 13-OTMS 18:0 ME from its retention index and the
observance of a coeluting peak at m/z 301, the mass of the other tat-cleavage ion. This
signal represents injection of 0.011 ng corresponding to 9.9 ng/g tissue or 9.9 ppb, based
on methods described in the next chapter. It can be seen that the statistically calculated
criterion is more rigorous than the S/N criterion for determination of the limit of
detection.

C. Recovery of Hydroxylated Fatty Acids

l. lZ-OH 18:0 as the Model Compound

In order to determine the recovery of hydroxylated fatty acids in the tissue, a
known amount of 12-hydroxy stearic acid, the standard selected as the model
hydroxylated fatty acid, was added in the homogenizing tube just before tissue
homogenization. After sample work-up, the extract was analyzed for the derivatized
standard. In the initial stages of development, a large amount (10.0 pg) of the standard
was added to the tissue and excellent recovery was obtained (Table 3-3). With this
indication of the capability of the procedure to isolate hydroxylated fatty acids, a smaller
amount (80.0 or 60.0 ng) of the standard, closer to the amount expected to be present
naturally in the tissue, was added to fresh tissue before homogenization. With this lower
quantity of the polar fatty acid, a more accurate assessment of the recovery was possible,
since some standard was expected to be lost due to adherence, etc.

The high sensitivity, low limit of detection, and excellent resolution provided by
capillary GC-FID would make this detection technique ideal for detection of the internal
standard. But there were severe reproducibility problems with the instrument at the time
of these analyses. The deviation of replicate injections was unusually high and a typical
standard calibration curve from GC-FID illustrates the high imprecision (RSD = 6%-
46%) in replicate analyses of standards (Figure 3-10). The signal for later injections of
similar volumes of the same extract would usually be considerably higher than the initial

response. It was thought that the sample itself may be contaminating the column or

56

maucaou

 

eves: w w 83:58 Emmaim g V
88: as
e8 H .22 2 w 3:59. 92 mo-~ m: 32
82: as
«.8 H em... a _ 3:58 one mg w: 82
898 as
$0.. a .soom v v 3.28. 3: moé w: 98 3 0.8
e: a ea c v a: as 25 3: moé m: ode a as
$5 . .3:
832% w v «.2: S. 25 3: moé w: odes 98
.3 H saw v v 89% as
as. a as: n v «55.8. 3: moé m: 2.... s 0.8
89% as
can H E v a Esau. 3: moé m: 2:
.nm «:5 «295% 2.3:. 2: cm can
560% .oz .02 8522 8828 can «fine s25 .22.

88¢. 03m 8335.632 .8 bo>o8m ”Wm 052.

57

CALIBRATION OF 12-OTMS 18:0 ME BY GC-FID

 

    

3,1 200000
2
2
:3 150000 .. y=98.683+8528.5x R"2=0.986
a:
fi
E 100000 -
< .
o
3 50000 -
5 J

 
 

 

 

 

i

0.00 5.00 10.00 15.00
AVG AMT 12-OTMS 18:0 ME INJ (MG)

1 V V I r fi' ' l I I

Figure 3-10: Calibration cm've for determination of 12-OTMS 18:0 ME
by GC-FID. The average area response of replicate
injections (N=2, 3, or 4) is plotted. Error bars represent 11
standard deviation.

58

detector, thus causing fluctuations in response. Some, but not all, randomly-spaced
injections of blank solvent did indicate a high and variable background.

One possible way to eliminate these inconsistencies is to utilize the mass
discrimination provided by reconstructed mass chromatograms with GC-EI/MS. Figure
3-11 demonstrates how the selection of major fragment ions (in this case, a-cleavage
ions), can produce a cleaner chromatogram than the total ion current (TIC)
chromatogram. Some other component, either one in this spiked extract or a "ghost"
compound from the column, eluted in the region of the derivatized standard, making
quantification from the TIC impossible. Therefore, quantitation was henceforth
calculated from the area obtained from reconstructed mass chromatograms of
characteristic fragment ions.

The recovery of 6080 ng of 12-OH 18:0 was calculated from GC-scanning
EI/MS with a new column was good (78%), but the standard deviation was fair, possibly
due to the small number (four) of spiked extracts analyzed (Table 3-3). A saturated fatty
acid waa selected as the model hydroxylated fatty acid so that the efficiency of
hydrogenation would not influence the recovery estimation. The column appeared to be
clean of contaminants, as indicated by a very low signal for the characteristic fragment
ions after injections of solvent. In an effort to improve the sensitivity and reproducibility
of the analysis, SIM of ten or only one ion was performed.

Subsequent analysis of the same extracts by GC-EI/MS with SIM and the same
column just ten days later, gave recovery results which increased dramatically with each
subsequent run. Recovery was calculated to be 119% from the first injection, and 376%
from the sixth extract injection. With the next two injections, the response did tend to
level off. Even duplicates of the same extract gave quite different responses (RSD = 11-
15%). Immediately after these extract analyses, another standard injection was
performed, and its value agreed with the calibration line.

59

retention time (mins)

 

 

 

 

--.“E5---L--A--m -
TIC
A
b
“ 301
n
d
8
n
C
° 187
59° 700 710 720 730 740 750
scan number
12-OTMS 18:0 ME I 301
Eoms o
: . ‘ /
/W: :'\/\/\/\/\/C\
5 E OCH:
Figure 3-11: Reconstructed total ion current chromatogram and mass

chromatograms at m/z 301 and mlz 187 from GC-EI/MS
(scanning mlz 50-500) analysis of tissue Spiked with 12-
OH 18:0 and extracted by the new isolation procedure.
Structure indicates the origin of these (at—cleavage fragment
ions. .

60

After baking out the column again, these same extracts were re-analyzed by GC-
EI/MS with SIM set only at mlz 187. Recovery values for the tissues spiked with 60.0 or
80.0 ng 12-OH 18:0 were more reproducible, having relative standard deviations of 5%
or less for duplicate analyses. The average recovery from this day's analyses was 93%
(Table 3-3), but the deviation of the recovery estimates between the samples was very
large (RSD = 33% for six data points). The results from the two days which produced
reasonably precise values for duplicate injections indicate the recovery of the
hydroxylated C18 fatty acid was about 86%, an acceptable value for such a complex
matrix.

2. 2-OI-I 12:0 as the Model Compound

It was decided that one possible contributor to the high recovery was endogenous
12-0H 18:0 or precursors which produce that compound. Thus, a shorter carbon-chain
length compound, 2-OH 12:0, was used as an internal standard in a manner similar to that
described for 12-OH 18:0. The earlier elution time of the shorter chain-length compound
may reduce some of the possible interferences in the extract or from the column. Using
the peak intensity at mlz 243, representing a major fragment ion for quantification, a
preliminary estimation of the recovery of 100.0 ng from a mammary tissue with duplicate
analyses of one extract was 93% :I: 22%. Subsequent analysis of the recovery of
additional spiked tissues under the same scanning conditions displayed a similar
imprecision (150% :l: 80%). The irreproducibility in the latter data set can be partly
attributed to the scatter of the calibration curve, particularly in the region corresponding
to the peak intensities observed for the internal standard in the spiked extracts. But
again, the RSDs of duplicate analyses of the same extract (7.0% to 52%) indicated severe
irreproducibility.

3. 3H-s-HETE as the Model Compound

In order to separate the imprecision of the analytical method of detection from the

imprecision of the exuaction technique, an independent detection method was needed.

61

The sensitivity of radiation counting makes it an attractive quantitation technique. 3'H-S-
hydroxyeicosatetraeneoic acid (3H-5-HETE) (New England Nuclear) was purified by
thin layer chromatography on silica plates (10 cm x 10 cm, Merck) with an ethyl
acetate:isooctane:acetic acidzwater mobile phase (64). Parallel elution of 1-2 pg of cold
5-I-IETE (Cayman Chemicals) and visualization by iodine vapor was utilized to
determine its chromatographic mobility, which would be the same as any pure 3H-5-
I-IETE. The experimentally determined retention factor for the 5-HETE standard, Rf =
0.53, was close to the retention pictured in the literature (64). The corresponding 3H-5-
HETE band was quickly marked and scraped with suction through a plugged (silanized
glass wool) pasteur pipet. For the purpose of determining the purity of the original 3H-5-
HETE as it existed after -4 °C storage, the entire band was placed in a scintillation vial
with cocktail. The rest of the 3H-5-HETE vertical channel was likewise scraped in 12
fractions, collected, and counted (LKB Wallac, com'tesy of Dr. McConnell‘s research
laboratory). Quenching corrections are automatically performed by the instrument. The
band parallel to the cold 5-I-IETE contained 28% (82000 dpm of 291200 dpm) of the
radioactivity. To prepare a working solution of 3H—5-HETE, another silica plate was
spotted with the original 3H-5-HETB solution and with cold S-HETE, and developed
Only the band parallel to the cold 5-HETE, as determined by iodine vapor visualization,
was scraped and collected. The 3H-5-HETE was removed from the silica stationary
phase with CHCl3zMeOH (1:1) rinses in a tightly-plugged pasteur pipet. The eluent was
collected, solvents were removed by nitrogen evaporation, and an exact amount of
ethanol was added. This entire procedure was carried out as quickly as possible to
minimize degradation of the light- and air-sensitive compound.

Each of four tissues was spiked with 50.0 [LL of the working 3H-5-I-IE-ZTE solution
(200 dpm/ltL). Cold 5-HETE (50 “L, 5.0 pg) was added to each tissue before

homogenization to act as a carrier for the low amount of labeled compound. Each sample

62

and a blank were extracted according to the above procedures, including hydrogenation
of one-half of each tissue preparation.

A one-quarter portion of each extract was analyzed by a scintillation counting .
Recoveries of 3H-5-HETE from the spiked tissues, based on the ratio of the radioactivity
in the final extracts to the radioactivity of another 50 uL aliquot of the working 3H—5-
HETE solution and corrected for the average dpm of scintillation cocktail blanks, were
quite low and imprecise - 31% i 4% (13% RSD) for the non-hydrogenated extracts and
24% :t 7% (27% RSD) for the hydrogenated extracts.

These results were considered quite suspect when the radioactivity of the filter
paper used after hydrogenation and the discarded aqueous wash were examined. A most
striking feature is the high amount of radioactivity in the filter paper, which had been
rinsed three times with methanol (in which 5-HETE is very soluble). Each filter
contained 23-34% of the counts in one-half of the volume of 3H-5-HETE added to each
tissue. There was also an unexpectedly high amount of radioactivity (5% - 10% of the
total radioactivity) in the discarded phase from the aqueous wash. Earlier experiments
with GC—FID in this laboratory supported literature statements that an aqueous wash does
not extract lipids. Thus it is suspected that the "purified" 3H-5-HETE was not very pure.

One possible explanation is that the 3H-5-I-IE'I'l-3 had not been completely
unbound from the silica stationary particles after filtration through tightly-plugged glass
wool. If this were the case, 3H-5-I-IE'I'E irreversibly bound to the silica particles would
be too large to pass through the filter paper. Such bound 3H-5-HETE would also behave
uncharacteristically in other separation processes, such as the aqueous wash. The
imprecision in the calculated recovery can also be possibly explained by this theory. If
some (about one quarter) of the 3H-S-I-IE’I‘E was bound to silica particles, its solubility in
ethanol, the working solution solvent, could be lowered If so, not all of the 3H-5—I-IETE
was in solution, and some 3H-5-I-IE'I'E-silica particles were undissolved, making the
aliquots of the working 3H-S-HETE solution added to each tissue unequal in their tritium

63

content. It was concluded that these tritium-based recovery results are not reliable.
These recovery results based on 3H-5-HETE were also suspect because they were so low
compared to the estimates obtained with GC-MS, when the measurements are
reproducible.
D. Conclusions

The preceding chapter described the development of a new procedure to isolate
hydroxylated fatty acids from mouse mammary glands. Estimates for the recovery of
hydroxylated fatty acids by this technique were highly variable, probably due to chemical
interferences in the case of analysis by GC-FID and GC-EI/MS and probably due to an
impure radioactively-labeled standard in the case of radioactive counting. Nevertheless,
the extraction procedure did effectively remove the large amount of unsubstituted fatty
acids which possibly could interfere with the detection of any hydroxylated (or other
oxygenated) fatty acids. Thus, it was decided to utilize this new extraction procedure on
a preliminary number of mouse mammary glands to determine if they contain any
hydroxylated fatty acids.

IV. CHAPTER Form: ANALYSIS AND QUANTITATION or Hmnoxv FATTY ACIDS rN
MAMMARY TISSUES

The preceding chapter described the development of a new procedure to isolate
hydroxylated fatty acids from mouse mammary glands. This chapter describes the use of
this extraction procedure to analyze several mouse mammary tissues for a select group of
derivatized hydroxylated fatty acids. The purpose of these analyses was to provide a
preliminary determination of the presence, if any, and concentration of C18 mono-
hydroxylated fatty acids. Results and examples of two different methods of quantitation
of the OTMS C13 methyl esters are given.
A. Analysis of Mammary Tissues for C18 OH-Fatty Acids

This new technique for the isolation of hydroxylated fatty acids from mammary
glands was utilized on several tissues of mice which had been fed controlled-fat diets.
The diets were similar, except in their fat composition. The mice were fed diets of high
concentration in specific fats, either 20% corn oil, 20% butter, or 20% fish oil (Table 4-
1). ‘

The derivatized extracts from the tissues were analyzed by GC-EI/MS utilizing a
30-m, DB-l capillary column (0.314 mm i.d., 0.25 um thick film) at 10 psi head pressure,
1.5 mL/minute helium flow rate, and 260° injector temperature. The temperature
program was 50°-180° at 35°lminute, then 180°-290° at 4°lminute. All mass
spectrometric analyses were performed on a JEOL AX-505 magnetic sector instrument at
70 eV ionization voltage, 100 M ionization current, and - 10 kV post-acceleration
voltage.

Standards were purchased or synthesized of the B- (Supelco), 9-, 12- (NuCheck
Prep), and 13- (Oxford Biomedical Research Inc., Oxford, MI) isomers of OTMS 18:0
methyl ester, and their retention times and fragmentation spectra under these laboratory
conditions were determined. As previously described, the methyl ester, TMS ethers of

hydroxylated fatty acids produce characteristic, relatively intense ions from or-cleavage

64

65

Table 4-1: Fatty Acid Composition of Oils and Butter

(Fish Oil)
Capric (10:0) -- —- 2.0%
Laurie (12:0) -- -- 2.3%
Myristic (14:0) -- 8.0% 8.2%
Palmitic (16:0) 10.1% 28.9% 21.3%
Palmitoleic (16:1) - 7.9 1.8%
Stearic(18:0) 1.6% 4.0% 9.8%
Oleic (18:1) 31.4% 13.4% 20.4%
Linoleic (18:2) 56.3% 1.1% 1.8%
Linolenic (18:3) -- 1.0% 1.2%
Eicosapentaenoic (20:5) -- 10.2% --
Docosahexaenoic (22:6) -- 12.8% -

Fatty acid concentrations of < 1% are not included.

Sources: Dr. C.W. Welsch and Reference 65.

66

about the carbon atom with the OTMS moiety (Table 3-1). Agreement with the
respective Standard of the ratio of the areas of ion current for each ion in the pair of major
a—cleavage ions, along with their retention indices, was used to identify these OTMS C18
FAMES in the extracts, as analyzed by capillary GC-EI/MS with SIM of nine ions (mlz
173, 175, 187, 229.2, 259.2, 301.2, 313.2, 315.2, 339.2) (Figure 41). Since the response
for many a-cleavage ions was near the limit of detection, especially for the ion with the
lower abundance of each pair (Figure 4—2), the window of acceptable values for the ratio
of peak intensities was liberal.

Quantitation of the 12- and B—O’IMS 18:0 ME compounds was based upon ratios
of the areas of mlz 187 and mlz 175, respectively, with 010 ng of analyte standard
injected. The available quantities of the 9- and 13- isomers were too low to prepare
nanogram concentrations for calibration curves based on the tat-cleavage fragments at mlz
229 and mlz 173, respectively. The ratios of characteristic ion current area to total ion
current area for these fragment ions were similar to the ratio of mlz 187 ion current area
to TIC area for 12-OTMS 18:0 ME. Therefore, the calibration curve of 12-OTMS 18:0
ME was used as a working curve for the 9- and 13- isomers. A list of the hydrogenated
and non-hydrogenated halves of the extracts analyzed by this method and the estimates of
quantitation are given in Table 4—2.

The small number of samples analyzed by this method and the uncertainty
inherent with calibration curves of only a few data points do not allow any definitive
conclusions to be made regarding the OTMS 18:0 methyl ester content of mammary
gland extracts. Some ion signal at the appropriate retention index was observed for each
isomer in the non-hydrogenated tissues from the animals fed a 20% butter diet. Very low
or zero concentrations of the 13-, 12-, or 9- isomers were detected in the tissues of mice
fed 20% corn oil or 20% fish oil. It does seem that, when present, the B-OTMS 18:0
methyl ester is present in higher concentration (parts per million) than the other isomers

(parts per billion).

67

mmom.v
Hamv.N
ommv.m
Nmmm.N
mm«v.«
vvmm.m
mHNm.m
mmmm.m

Nva.m

.ucznn

.82 05: 805

.3 25 a? mssmoo ace +5.5 Ndmm as ea .eaa
EoEwfim 033298 05 mcucomoaofimz ouwfi E29045 ,

m: a an: as e5 a: 3: msabé n: a. «a: a?

.32 cum“ m2hO-~C g cm. Ngom Q8 .32 on? mEHO
-ov down a. Ndmm Q8 3 3an Eobsoég 3823 “TV unswE

m: on: mzeoa
m2 3: 32.8-2 9

ME 93 $450-2 9 m2 onwfi mEHOé

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

aficsccfla 6
Quad owed t omen . amm.e. 0mm . arm . 0mm . 0mm , mum
m an all/l\\ll
o.m* oom.mmm
l< (all:
m.ma som.mm~
llI/zxxl,
oom.«om
o.N*
fi.efi* ll//t\\lll ooa.nmfi
ll/lllt
,l/xtxxlll oo~.mam
m.nm*. .
fill l5
oom.mafl
n.mm* w
lllll)///\\\lll/ll\ll oam.mam e
m.nm*. . M
ll//r\\\lllltl\ll som.mna e
v.mv* 3
llllllll;//\)//I;/\\\\l. oom.mmm u
. was $55 2:: 83:83 A" WE . .

mvnm.m« xo:

188

68

retention time (mins)

4 A A A A A A A A A A A 1 A I A A A A A A

 

ea-
60<
49-

291

 

948' ' '

m/z 315.2

1 I v v’ v

 

vss'a' . ' vséa' . 970 vsée' ' ' 'sée' ' lobe ' 18118 ' raza

 

 

 

 

 

m/z 173
949 ' res‘e' ' ' '969' ' 99:56 1 ' ‘9ee‘ ' '956' ' Widen ' ' ia'ta' ' ' Haze
scannumber
Figure 4-2: Selected ion-current profiles at mlz 315.2 (top) and m/z

173 (bottom) from GC-EI/MS with SIM of the nine ions
displayed in Figure 4-1. The arrow indicates elution time
of endogenous l3-OTMS 18:0 ME.

1"

69

.8958 25% 05 we :2 Basement»: 2E. - m.

 

cod odd cmdd awed odd odd cod cod dad ndo E Tm
2d mmd nmd Ned dd 3 Ed god wed ndo Tm
ddd odd omdd E d ddd cod ddd ddd dvd Sen m Ym
nod fid dud Rd v4 d4 «nod weed dvd Sen Ym
cod cod Sod 9d ed 3. awed Edd no.“ WE m Tm
odd 8d 8d odd ddd odd ddd odd cod QG Tm
Rod mEd cwdd nmdd ddd ood 8d 8d ovd mdw m 90
Bed Edd ddd ood 2d mud 8d 8d end mdw nO
:Vdd wood mad 2 d ddd ddd odd odd end mdo _: v-0
2d Ed 2d 5d 8d 8d cod ddd ow.” 08 DO
3&3 35 $38 85 3&3 $5 3&8 $5 REV
08% man: up: .~E< came mach: A5 .~E< 083 95m: unc— JE< 253 men: AS aE< @883 .5?
m2 dd“ mEOd m2 °an $290-3 m2 odd mSFOd m2 9E magm— Eoeom 033,—. 038%

Sam 53, 92>ch 3 Bug...“ 3 88.5mm 5 m2 cum" mSFO .«0 3.50:2 ”mé 035.

70

One curiosity of the data is the observance of some saturated OTMS C18 methyl
ester isomers in the non-hydrogenated portion of some tissue extracts, but no signal for
the same isomer on the hydrogenated portion of that same tissue extract. This is quite
unexpected, since any endogenous saturated hydroxylated C18 fatty acid should be
observed whether the extract was hydrogenated or not. The hydrogenated portion of the
extract is expected to have the same or an increased signal for each isomer of saturated
OTMS C18 methyl ester with respect to the non-hydrogenated portion, since any
unsaturated hydroxylated fatty acids of the same positional isomer possibly in the tissue
would become saturated during hydrogenation.

B. Use of Internal Standard for Quantitation

In order to provide more accurate determinations of these OTMS 18:0 methyl
esters in mammary tissue, a more rigorous calibration curve was needed. In addition, the
use of an internal standard to make corrections for variation in recovery of hydroxylated
fatty acids and fluctuations in analysis conditions and response (such as noise from
column background) would improve the analysis greatly. New mammary tissues were
spiked with 100.0 ng of 2-OH 12:0 (Matreya, Inc., Pleasant Gap, PA) just before
homogenization. The extracts were analyzed by capillary column GC-EI/MS under the
same conditions described above, but the mass range analyzed was only mlz 170-320 to
increase the scan cycle rate and lower the limit of detection. a-Cleavage ions and
retention indices were again used to determine the presence of an OTMS FAME.
Specifically, peaks at mlz 243 and m/z 287 were used to confirm the elution of the TMS
ether, methyl ester of the internal standard. For 12- and B— OTMS 18:0 ME, standard
curves of the ion current areas at mlz 187 or mlz 175, respectively, for the analyte ion
current area relative to the ion cm'rent area of the mlz 243 ion from the internal standard
versus the amount of analyte per 2.00 ng 2-O'IMS 12:0 ME were prepared (Figure 4-3).
The quantities of the OTMS 18:0 methyl ester positional isomers measured in non-

hydrogenated and hydrogenated portions of tissue extracts were calculated from these

71

 

 

AREA (187)/AREA (243)
to
I

   
 
 

y = 6.9146e-2 + 1.7878x 8‘2 = 0.992

El A187/A243

 

 

 

 

 

 

043 1 , .
0.00 1.00 2.00
NG 12-OTMS 18:0 ME/2.00 NG 2-OTMS 12:0 ME
5
8
E d
g 4- y=7.28GSe-2+0.46671x R“2=O.994
fi .
n: 3'
S d
3 2-
v . a
g 1- a A175/A243
a: .
<
O ' I f I 1 T T I ' I '
0.00 2.00 4.00 6.00 8.00 10.00 12.00
NG B-OTMS 18:0 IRE/2.00 NG 2-OTMS 12:0 ME
Figure 4-3: Calibration curves and linear regression equations for the

determination of 12-OTMS 18:0 ME (top) and of b-
OTMS 18:0 ME (bottom), using 2.00 ng 2-OTMS 12:0
ME as an internal standard.

72

curves and are given in Table 43. Representative reconstructed total ion and mass
chromatograms for the ions resulting from a-cleavage of selected OTMS 18:0 methyl
esters are given in Figure 4-4.

The limited number of tissues extracted and analyzed by the internal Standard
calibration method prohibits absolute conclusions to be made, but some generalizations
were noticed. Quantitation results from duplicate, same-day analyses of the
hydrogenated portion of two of the three extracts (samples B-S H and F-3 H) were in
good agreement, with relative standard deviations of 0% to 17%. This does not include
the one sample (B-5) in which no analyte signal was observed for 13-O'IMS 18:0 ME
during one trial analysis of the hydrogenated portion nor during the one trial of the non-
hydrogenated half of the extract. Results from duplicate analyses of sample F-II H are
suspect, however. The peak intensities, and thus the calculated OTMS 18:0 ME
concentrations, were consistantly lower for the second analysis trial than for the first.
This variation in calculated concentration was large, as evidenced by the high relative
standard deviation - 28% - 110%. Analysis of the hydrogenated portion of the extract
from sample F-II by GC-EI/MS scanning mlz 170-320 on another date supports the
results from the trial where 1.98 uL was injected, that is, the estimates at higher
concentrations of the OTMS 18:0 methyl esters.

As similarly observed in the tissues previously analyzed by GC-EI/MS with SIM,
the B« isomer of OTMS 18:0 methyl ester was determined to be present in low parts per
million (1-4 ppm) concentrations in all tissue extracts. The other isomers, 13-, 12-, and
9-OTMS 18:0 methyl esters, were detected at concentrations in the parts per billion range
- (2-140 ppb) in the hydrogenated portion of each extract. The sole exception is one of the
duplicate runs of the extract from the mouse fed the 20% butter diet. No 13-, 12-, or 9-
OTMS 18:0 methyl ester isomers were detectable in the non-hydrogenated portion of the

tissue extract from the mouse fed the 20% butter diet.

73

.8858 named on. .«o :2 Reasons—dz: 2E. - E

 

:od Rod omd Ed ood mud ood ood do; méw hi
2d 3 .o S. S .o _.m m.m E .o 2 .o mom m.mm m mi
3 .o m o .o omd _md mm m.m m“ .o E .o no; m.mn I m.&
mnod pod mood wood m._ m; mood wood 34 m.mn mi
mvod om mood mood md 3. wood omod mom o6: m =-m
omod E .o Sod ovod wees _ _ Sod 2 .o mo; oi o m oo-m
omod mwod ood ood m; ore. ood ood do; o5 o Fm
Sod 2d m_ .o w_ .o 3 o4 mood mood oo.m how I Wm
omod :d :d 2 .o Ed 2 ood ood oo.m 5.2. o: mm
ood ood ood ood vod v; ood ood oo.m 5.2. mm
Ede a weed eae a used Ede a 88d cage 2 mega 8.5
can: menu: .3 w: 253 wER: no: we can: 95»: no: we can: wash: up: w: 8.8%: 43
m2 oé mEO-o m2 oufi mSFOm— m2 odd macs... m2 od— med—.04: “cop—om 2.3:. oafiem

seesaw 3525 55 m2 «€535-00 3 Sign. a seesaw 2 m2 on: 34.5 as dessert "3. case.

74

.3. 035. 5 cam: 98 cowauoaasc do 23.2

.m-m 25% .«o comtoq 335mg? 05 do §o.m mo :26»?

ACNE: £5 m2 358... 5.8 sec .3. 05mm

5 was: «880$ 82 on? ago 2: manage? .m2

63 .m2 .mHm .an .Hom dam .omm #8 a mafiwofieezo
3&8 Ea gymoaaozo com :32 Boosbmcooom 31v oSwE

m2 3: mszbfi
m2 3: mszbé e

m2 oumo mEHOAu ,_‘ m2 oumH mEHOd

hon—ts: 58 F

8.3 - 8.3. - +3.3 8.3 8.3

 

D |D \

 

   

mnmm.m m.m*

mm«m.o m.m*. mmm .
mvvm.o m.m*. mmm .
mamm.o N.m*. _ «on .
vmuv.fi m.«*L um” .
vam.« a.«*. mam .
mumm.o m.m*¥ mud

mmmm . a N . m* L§§ir
MN

mmvmém 9:2,)» 1? 1 l1! i .

. :2...

 

 

 

 

 

 

 

CEJIDC‘OGCOO

.ucsno .mne . v«
95.5 08: con—.28 w E vam .« xcz

75

It should be noted that for these samples, and with this quantification method, the
hydrogenated portion of each extract was determined to possess a higher concentration of
each OTMS 18:0 methyl ester isomer than the concentration of that isomer in the non-
hydrogenated portion of the same tissue extract. Hydrogenated portions are expected to
contain the same or higher levels of saturated fatty acid than the non-hydrogenated
portion, as described above. This observation and the fact that the use of an internal
standard can correct for many variations in sample preparation and analysis indicates that
use of the internal standard for calibration is preferred

These few estimates of the concentrations of C 18 mono-hydroxylated fatty acids
are only preliminary, but indicate that those compounds probably exist in some
mammary glands at the parts per billion level. The results strongly suggest that this new
extraction procedure is appropriate to isolate hydroxylated fatty acids from mammary
glands and that there are possibly some isomers of mono-hydroxylated C18 fatty acids
which are detectable by GC—EI/MS in the parts per billion range.

Further analysis of a statistical number of tissues from mice fed diets of different
fatty acid content is needed to determine if there exists a relationship between the
amount, if any, of hydroxylated fatty acid in the mammary gland and the fat content of
the subject's diet. It is suggested that two or more glands from the same animal be pooled
to provide increased levels of any hydroxylated fatty acids present. With the same final
volume (100 pl.) for the extract of pooled glands, the ratios of weight of analyte/weight
of tissue will not be affected, but the amount of analyte introduced into the gas
chromatograph-mass spectrometer will be increased for each 2 11L injection (2% of the
extract). An alternative solution to increasing the amount of potential analyte in 100 p.L
total volume of extract is to use rats instead of mice as the study animal. The rat is a
much larger animal and each mammary gland can be extracted individually, providing

replicate trials for each subject.

76

The results from these GC-EI/MS analyses of mammary glands with SIM and
with small mass range scanning and an internal standard indicate that further use of the
isolation procedure is warranted. Some isomers of C 18 mono-hydroxylated fatty acids
do appear to be present in most of the tissues examined to date. B-I-Iydroxy stearic acid
was observed to be the isomer in highest concentration (low ppm), but other isomers (13-

, 12-, and 9-) are probably present in many tissues at the parts per billion level.

V. CHAPTER FIVE: DIRECTIONS FOR FUTURE WORK

The need for a new extraction procedure to isolate oxygenated fatty acids from
mouse mammary glands, to determine if there exists a relationship between dietary fat
and the oxygenated fatty acid content of the tissue, was explained earlier. This thesis has
described the development and preliminary evaluation of a new technique for the
isolation of hydroxylated fatty acids from mammary glands, with analysis by GC-EI/MS.
This chapter presents possible analytical directions for further application of the new
extraction technique to probe the biochemical problem under investigation.

A. Determination of Other Hydroxylated Fatty Acids

The development and evaluation of the new isolation procedure used 12-hydroxy
stearic acid as the model oxygenated fatty acid. The mammary glands extracted by the
new method were analyzed only for a few isomers of mono-hydroxylated stearic acid.
The recovery and concentration of other hydroxylated fatty acids should be determined.
Those compounds of interest include unsaturated C13 hydroxylated fatty acids,
hydroxylated fatty acids of 20, 22, and 16 carbon-chain lengths, and dihydroxylated fatty
acids.

The non-hydrogenated portion of each extract can be analyzed for the presence of
derivatives of those unsaturated hydroxylated fatty acids which would be expected to be
oxidation products of dietary unsaturated fatty acids, such as 18:2, 18:3, 20:5 and 22:6
(65). There are other dietary and endogenous unsaturated fatty acids which may be
oxidized to hydroxylated compounds, including 16:1, 18:1, 20:1, 20:3, and 20:4. If
dihydroxylated fatty acids are present, their concentrations may not be sufficiently high
to be observed by the current methods. The isolation procedure should be evaluated for
recovery of these compounds, since they are more polar than their monohydroxy

analogues, and their retention with the silica column clean-up procedure is unknown.

77

78

B. Analysis for Other Oxygenated Fatty Acids

Hydroxylated fatty acids are not the only form an oxygenated fatty acid can have,
just one of the more stable, which is why it was selected as a model for the development
of the new isolation procedure. The initial oxidation products are highly unstable
hydroperoxides, which are quite easily reduced to alcohols, ketones or epoxides. The
carbonyl groups are fairly stable and should be detectable if present in the tissue.
Treatment of a ketone with BSTFA or other moderate to strong silylating reagents will
produce a different positional isolmers of OTMS fatty acid methyl esters. If
determination of the position of the carbonyl group is desired, methoximes can be
prepared from the FAMES just before the silylation. Then the extract can be analyzed by
GC—EI/MS for characteristic fragments of methoximes. Again, the recovery of keto-fatty
acids from tissue homogenate should be evaluated using the new isolation procedure.
Epoxides across a carbon-carbon double bond can also be a product of oxidation. The
extract can be analyzed by GC-EI/MS, where characteristic ions result from, tat-cleavage
on either side of the epoxide ring and B-cleavage on the far side of the ring with
hydrogen atom rearrangement. The recovery of a representative epoxide fatty acid
should be determined with the new extraction procedure.
C. Minimization of Background Contaminants

It is expected that any oxygenated fatty acids present will only exist in trace
quantities, and in this study some hydroxylated fatty acids were determined to be present
in the tissues at the parts per billion level. The analysis of such low levels of analytes
requires low background from the instrumentation used for analysis and from the
isolation procedure itself. During the development of the extraction procedure, care was
taken to attempt to eliminate contaminants, especially phthalates and other high-boiling
compounds. But not all background was eliminated, in particular, there was some
background noise from concentrated solvent which had passed through the silica SPE
column. This background could probably be reduced with the use of the recently-

79

available "inert" columns, which are made of silanized glass cartridges. But these
columns are quite expensive, and the background reduction must be great in order to
justify their cost.

Another source of background in the analysis of the extracts is from eluents from
the GC column. On certain days of analysis, there were many problems with a high
signal upon injection of solvent only, both with GC-FID and with GC-EI/MS using the
same GC column. The baseline signal was particularly high and/or variable after several
injections of extracts had been performed. Since this background was only observed on
certain days, and not after numerous extract injections on other days, there could be
another cause of the high background A change to a new, slightly more polar GC
column (DB-5, 5% phenylmethyl silicone) may reduce the amount of "ghost" peaks
resulting from previously-injected compounds which adhered to the column. This
change would require the redetermination of retention indices, so less dramatic methods,
such as changing the temperature programming of the GC oven, should be attempted
first. Using a lower final temperature (280°) for the GC oven after an overnight bake-out
at a higher temperature appeared to reduce the signal from injection of a solvent blank,
even after a dozen extract injections. Since it seemed to work, this possible solution was
not rigorously evaluated.

Corrections for background due to the matrix can be made with the discrimination
provided by tandem mass spectrometry (MS/MS) or triple quadrupole MS. The first
sector is set to allow ionized GC effluent with only a specified mlz value to pass to the
second sector. In triple quadrupole MS, the analyte reacts with an inert gas in this sector,
causing dissociations. The daughter ions pass into the next sector where they are
detected to give characteristic daughter spectra. The potential benefit of background
elimination with MS/MS warrants an examination of the feasibility of application of this

detection technique to the tissue extracts.

80

D. Separation of Extract by High Performance Liquid Chromatography

Recent developments in the use of HPLC-MS for analysis of low level
components (66) and for lipids (67) have offered a potential method of fatty acid
separation with on-line mass spectral detection which was not widely available at the
inception of this project. The picomole detection limit obtained by Hemling and
coworkers was for analysis of glycoproteins, not oxygenated fatty acids. Detection limits
for HPLC-MS are not as low for fatty acids as for proteins, which have been investigated
much more often by HPLC-MS.

The appropriate HPLC method for separation and the best interface to the mass
spectrometer could be investigated. Implementation of an HPLC-MS system could serve
to reduce some of the background noise, since the column in HPLC will have different
interactions with the extract's components in the GC column. The different stationary
phases and the different physical states of the extract for each separation method produce
dissimilar interactions.

The use of HPLC-MS/MS to analyze the oxygenated fatty acid methyl esters
could eliminate the use of a silica SPE clean-up column and a gas chromatographic
column for additional resolution of the resolution of the extract, combining these steps

into one process.

MP?!"

9‘

10.
11.
12.
13.

14.
15.
16.
17.

18.

LIST OF REFERENCES

Welsch, C.W., Cancer Res. 1985, 45, 3415-3443.

Tannenbaum, A. Cancer Res. 1942, 2, 49-53.

Engel, R.W.; Copeland, D.H. Cancer Res. 1951, 11, 180-183.

Gammel, E.B.; Carroll, K.K.; Plunkett, ER. Cancer Res. 1967, 27, 1737-1742.

Cgrroll, K.K.; Braden, L.M.; Bell, J.A.; Kalamegham, R. Cancer, 1986, 58, 1818-
1 25.

Carroll, K.K. Cancer Res. 1985, 45 , 3415-3443.

Reddy, B.S.; Cohen, L.A.; McCoy, GD.; Hill, P.; Weisenburger, J.H.; Wynder,
E.L. In Adv. Cancer Res.; Klein, 6.; Weinhouse, 8., Eds; Academic: New York,
1980; Vol.32, PP 295-325.

Carroll, K.K. J. Environ. Pathol. Toxicol. 1980, 3(4), 253-271.

Carroll, K.K.; Gammel, E.B.; Plunkett, E.R. Can. Med. Assoc. J. 1968, 98, 590-
594.

Klurfeld, D.M.; Kritchevsky, D. Proc. Soc. Exp. Biol. Med. 1986, 183, 287-292.
Carroll, K.K.; Khor, H.T. Lipids, 1971, 6, 415-420.

Sylvester, P.W.; Ip, C.; Ip, M.M. Cancer Res. 1986, 46, 763-769.

Kidwell, W.R.; Knaaek, R.A.; Vonderhaar, B.K.; Losonczy, I. In Molecular
Interrelations of Nutrition and Cancer; Amett, M.S.; van Eys, 1.; Wang, Y.-M.,
Eds.; Raven: New York, 1982; Vol. XVI, pp 219-236.

Erickson, K.L.; Thomas, I.K. Food Technol., 1986, 39(2), 69-73.

Jurkowski, J.; Cave, W.T. J. Nat. Cancer Inst. 1985, 74(5), 1145-1150.

Kinsella, J.E. Food Technology 1986, 39(2), 89-97.

Rogers, A.E.; Fernstrom, J.D.; Ge, K.; McConnell, R.G.; Leavitt, W.W.; Westel,
W.C.; Yang, S.O.; Camelio, EA. In Molecular Interrelations of Nutrition and
Cancer; Amett, M.S.; van Eys, J.; Yang, Y.-M., Eds. Raven: New York, 1982;
Vol. XVI, pp 381-399.

Welsch, C.W. Am. J. Clin. Nutr. 1987, 45 , 192-202.

81

19.

20.

21.

22.
23.
24.
25.
26.

27.

28.

29.
30.
31.

32.

33.

34.

35.

36.

82
Welsch, C.W.;DeHoog, J.V.; O'Connor, D.H.; Sheffield, LG. Cancer Res., 1985,
45, 6147-6154.
Carroll, K.K.; Davidson, MB. In Molecular Interrelations of Nutrition and
Cancer; Amett, M.S.; van Eys, J.; Yang, Y.-M., Eds. Raven: New York, 1982;
Vol. XVI. PP 237-245.
Clinton, S.K.; Visek, WJ. In Xenobion'c Metabolism: Nutritional Efiects; Finley,
J.W.; Schwass, D.E., Eds.; American Chemical Society: Washington, D.C., 1985;
Vol. IX, pp 310-325.
Kritchevsky, D.; Weber, M.M.; Klurfeld, D.M. Cancer Res. 1984, 44, 3174-3177.
Kline, K. Nutrition Today 1986, 21(3), 11-17.
Welsch, C.W. J. Natl. Cancer Inst. 1983, 70(2), 215-221.
Gunstone, F.D.; J. Am. Oil Chem. Soc. 1984, 61(2), 441-447.
Mead, J.F. In Free Radicals in Molecular Biology, Aging, and Disease;
Armstrong, D.; Sohal, R.S.; Cutler, R.G.; Slater, T.S. Eds.; Raven: New York,
1984; Vol. 27. PP 53-66.
Slater, T.F.; Cheeseman, K.H.; Proudfoot, K. In Free Radicals in Molecular
Biology, Aging, and Disease; Armstrong, D.; Sohal, R.S.; Cutler, R.G.; Slater,
T.S., Eds.; Raven: New York, 1984; Vol. 27, pp 293-305.

Bull, A.W.; Nigro, N.D.; Golembieski, W.A.; Crissman, J.D.; Mamett, LJ.
Cancer Res. 1984, 44 , 4924—4928.

Bull, A.W.; Nigro, N.D.; Mamett, L.J. Cancer Res. 1988, 48, 1771-1776.
McCay, P.B.; King, M.M.; Pitha, J.V. Cancer Res. 1981, 41, 3745-3748.

Hirose, M.; Masuda, A.; Inoue, T.; Fukushima, S.; Ito, N. Carcinogenesis 1986,
7(7), 1155-1159.

King, M.M.; Terao, J.; Brueggemann, G.; McCay, P.B.; Magarian, R.A. In
Xenobiotic Metabolism: Nutritional Efi'ects; Finley, J.W.; Schwass, D.E., Eds.;
American Chemical Society: Washington, D.C., 1985; Vol. IX, pp 131-150.

Lane, H.W.; Butel, J.S.; Howard, C.; Shepherd, F.; Halligan, R.; Hedina, D.
Carcinogenesis, 1985, 6(3), 403-407.

Traitler, H. In Progress in Lipid Research; Pergamon Journals Ltd: Oxford,1987;
Vol. 26, PP 257-280.

Christie, W.W. Gas Chromatography and Lipids; Oily: Ayr, Scotland, 1989; pp
64-84.

Kuksis, A. In Lipid Research Methodology, Story, J.A., Ed.; Alan R. Liss: New
York, 1984; pp 77-131.

37.

38.
39.

40.

41.
42.
43.

45.
46.
47.

48.

49.
50.

51.
52.
53.
54.

55.
56.

57.
58.

83
Christie, W.W. Gas Chromatography and Lipids; Oily: Ayr, Scotland, 1989; pp
180—182.
Pace-Asciak, C.R.; Micallef, S. J. Chrom. 1984, 310, 233-242.

Leis, H.J.; Hohenester, E.; Gleispach, H.; Malle, E,; Mayer, B. Biomed. Environ.
Mass Spec. 1987, 14, 617-621.

Brash, A. R.; Ingram, C.D. Prostaglandins, Leukotrienes, and Medicine 1986, 23,
149-154.

Kelly, R.A.; O'Hara, D.S.; Kelley. V. J. Chrom. 1987, 416, 247-254.
Oliw, E.H. J. Biol. Chem. 1989, 264(30), 17845-17853.

Claeys, M.; Coene, M.-C.; Herman, A.G.; Jouvenaz, G.H.; Nugteren, D.H.
Biochim. Biophys. Acta 1982, 713, 160-169.

German, J.B.; Bruckner, G.G.; Kinsella, J.E. Biochim. Biophys. Acta 1986, 875,
12-20.

Funk, C.D.; Powell, W.S. J. Biol. Chem. 1985, 260(12), 7418-7488.
Funk, C.D.; Powell, W.S. Biochim. Biophys. Acta 1983, 754, 57-71.

Sautebin, L.; Spagnuolo, C.; Galli, C.; Galli, G. Prostaglandins 1978, 16(6), 985-
988.

Glcispach, H.;Moser, R.; Mayer, B.; Esterbauer, H.; Skriletz, U.; Zierman, L.;
Leis, H]. J. Chrom. 1985, 344, 11-21.

Murphy, R.C. Prog. Biochem. Pharmacol. 1985, 20, 84-100

Strife, R.J.; Voelkel, N.F.; Murphy, R.C. Am. Rev. Respir. Dis. 1987, 135, S18-
S21.

Pace-Asciak, C.R.; Micellef, S. Can. J. Biochem. Cell Biol. 1984, 62, 709-714.
Bligh, E.G.; Dyer, W.J. Can. J. Biochem. Physiol. 1959, 37(8), 911-917.
Folch, J.; Lees, M.; Sloane Stanley, G.H. J. Biol. Chem. 1957, 226, 497-509.

Naito, H.K.; David, J.A. In Lipid Research Methodology, Story, J.A., Ed.; Alan
R. Liss: New York, 1984; pp 1-76.

Radin, N.S. Methods Enzymol. 1969, 14, 245-254.

Carroll, K.K. In Lipid Chromatographic Analysis; Marinetti, G.V., Ed; Marcel
Dekker: New York, 1976; Vol. 1, Second edition, pp 173-214.

Powell, W.S. Methods Enzymol. 1982, 86, 467-477.

Farquar, J.W.; Insull, W.; Rosen, W.; Stoffel, W.; Aherns, E.H. Nutrient Reviews
(Supplement), 8(II); Nutrition Foundation: New York, 1959; PP 29-30.

59.

61.

62.

63.

65.

67.

84

Homing, M.G.; Williams, E.A.; Homing, E.C. J. Lipid Research 1960, 1(5), 482-
485.

Holland, J.F.; Enke, C.G.; Allison, 1.; Stults, J.T.; Pinkston, J.D.; Newcome, B.;
Watson, J.T. Anal. Chem. 1983, 55, 997A-1012A.

Yefchak, G.; Schultz, G.; Tecklenburg, R.; Davenport, M.; McNitt, K.; Allison,
J.; Watson, J.T.; Enke, C.G.; Holland J.F. Annual Report 1989; Michigan State
University Mass Spectrometry Facility, East Lansing, MI.

Miller, J.C.; Miller, J .N. Statistics for Analytical Chemistry; Ellis Horwood
LimitedzChichester, 1984; Chapter 4.

ACS Committee on Environmental Improvement, Anal. Chem. 1980, 52, 2242-
2249.

Harris, 8.1.; Baenziger, N.L. Prostaglandins, 1983, 25(5), 733-739.

Composition of Foods: Fats and Oils Raw, Processed, Prepared; Science and
Education Administration, US. Department of Agriculture; Agriculture
Handbook No. 8-4, US Government Printing Office:Washington, DC, 1978.

Hemling, M.E.; et. a1. Proceedings of the 38th ASMS Conference on Mass
Spectrometry and Allied Topics, 1990, 1146-1147.

Hansen, E.B.; et.al. Proceedings of the 38th ASMS Conference on Mass
Spectrometry and Allied Topics, 1990, 1059-1060.