' : ;:-,r-,.
£3.23" 53h

'le
‘ W“;

:61
a

‘ ‘ '~ 733'
.r ~ . . e"?
r I, n,- . "FA " . ,V . . , , i" fidtnl§§fig§
I '.[E:i‘2—4n.}f7;l h'” .' W, J ' . '77 Aflifr v . - . .
.:- gm‘rfix‘ - , 131132319, . M S
V '. . ' '1 ‘ . ~
‘ $35?ng a .

. 3.5 .3. ‘
« r n: . .

‘ 1 “:4‘63313‘1‘

N :fi-‘éfijfi‘

I; a > P17
2 ' $7»
a“:-

c’. .
1-;
\‘ ':

‘ l
I: :-
‘ ‘vw

 

' .3-
.”l ’ .
. . 3%»; _ 3
1 “A???“
“.thb‘wn

MiCHIGAN S NIVERSITY LIBRARIES

M WW\\\i\i\i\‘i\\t\1\\\\\\mM‘“W

 

 

i
H
3 1293 01399 5570

 

This is to certify that the

dissertation entitled

SEPARATION, IDENTIFICATION, AND QUANTITATION OF
MONOHYDROXYLATED POLYUNSATURATED FATTY ACIDS UTILIZING

GC/MS AND CHIRAL HPLC
presented by

Jennifer A. Johnson

has been accepted towards fulfillment

of the requirements for

Ph . D . degree in Chemistry

NW

Major professor

Date 5

MSU is an Affirmative Action/Equal Opportunity Institution 042771

 

LUERARY
M‘CMQan Etate
University

 

 

 

PLACE ll RETURN Boxwromavothhchockomm your round.
0 AVOID FINES rotum on or boforo doto duo.

1
DATE DUE DATE DUE DATE DUE

m i...

'T R 08LU¢JL

”if—7 ...T..
MT "713

r a.
L,-

MSU lo An Nfinnottvo Action/Equol Oppodunlty I
Wm1

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

SEPARATION, IDENTIFICATION, AND QUANTITATION OF
MONOHYDROXYLATED POLYUNSATURATED FATTY ACIDS UTILIZING
GC/MS AND CHIRAL HPLC

By

Jennifer A Johnson

A DISSERTATION

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

DOCTOR OF PHILOSOPHY

Department of Chemistry

1995

ABSTRACT

SEPARATION, IDENTIFICATION, AND QUANT IT ATION OF
MONOHYDROXYLATED POLYUNSATURATED FATTY ACIDS UTILIZING
GC/MS AND CHIRAL HPLC

By

Jennifer A. Johnson

While a number of mechanisms have been reported to explain the stimulatory
efl‘ect of high levels of dietary fat on mammary (breast) tumon'genic processes in
experimental animals, to date, none of these mechanisms has proven to be entirely
satisfactory. We are proposing that oxygenated metabolites (products) of linoleic acid, via
free radical and/or enzymatic processes, are the key products of fat hyperalimentation that
stimulate the mammary tumorigenic process. The oxygenated products of primary interest
are 9- and l3-hydroxyoctadecadienoic acid (9-HODE, l3-HODE) as these products have
been shown to have a growth stimulatory effect in experimental animals in. an array of
organ sites including the mammary gland.

We have developed an analytical methodology for the isolation, identification, and
quantitation by GC/MS and chiral HPLC of 9- and 13-HODE in mouse mammary gland
tissue. The mammary gland tissue is pulverized in liquid nitrogen, internal standards
consisting of "Oz-labeled analogs of the 9- and 13-HODEs are added, and the tissue is

extracted twice in ethanol. Non-lipid materials are removed in a chloroform:

methanolzwater extraction step. The lipid material remaining is methylated with
diazomethane, and much of the non-oxygenated fatty acids are removed via solid phase
extraction on silica columns. The sample is derivatized with BSTFA to form the TMS
ethers prior to quantitation by GC/MS. Those samples for HPLC analysis are
reconstituted in hexane and eluted isocratically with a hexanezmethanol mobile phase on a
Chiralcel OD chiral stationary phase.

The concentrations of 9- and 13-HODE were highest (P < 0.01) in the mammary
glands of mice fed high levels of polyunsaturated fatty acids (20% corn oil). Enantiomeric
distributions of the HODEs in the tissues show that l3(S)-HODE and9(R)-HODE are the
more abundant isomers in the tissues.

A positive correlation between the mammary gland concentrations of 9-HODE and
l3-HODE and the reported mammary gland tumorigenic activities of these diets in rodents
(20% corn oil > 5% corn oil > 20% beef tallow) is observed in these studies. These results
provide evidence that 9- and 13-HODE are produced in significat quantities in animals fed

diets rich in polyunsaturated fatty acids.

ACKNOWLEDGIVIENTS

I wish to thank my committee members, Dr. J. Throck Watson, Dr. Clifford
Welsch, Dr. Douglas Gage, and Dr. Stan Crouch, for their assistance and advice during
this project. Their willingness to meet with me was greatly appreciated. I also wish to
thank Dr. Art Bull, Oakland University, and Larry Nicholson, The DOW Chemical
Company, for their contributions to the project.

As a member of the Watson laband the MSU/NIH Mass Spectrometry Facility, I
have been associated with many wonderful people. I would like to thank Mel Micke and
Mike Davenport for their technical assistance over the years. I especially would like to
thank Melinda Berning, Bev Chamberlin, Kate Noon, Jian Wang, Ken Roth, Naxing Xu,
and Ben Gardner for their friendship.

I am grateful for the funding I have received from the Environmental Toxicology
training program (Public Health Service/NUT), as well as having the work supported by
A funds from the Biomedical Research Training Program of the National Center for

Research Resources of NIH.

iv

TABLE OF CONTENTS

List of Tables
' List of Figures
List of Abbreviations
Chapter One: Dietary Fat and Breast Cancer
I. Introduction
11. Fatty acids
111. Diet and breast cancer
A. Human population studies ‘
B. Animal studies
1. Amount of fat
2. Type of fat
3. Proposed mechanisms
IV. Lipid oxidation
V. Structural requirements for biological activity
VI. Goals of this research
References

Chapter 2: Development of Analytical Methodology

page number
viii

X

xiii

13
16

19

vi
1. Characterization of fatty acids by GC/MS

II. Characterization of hydroxylated fatty acids by GC/MS
III. Methods other than GC/MS for the analysis of fatty acids
IV. Initial attempts to extract mammary gland tissue

A Extraction

B. High-performance liquid chromatography

C. The use of GC/MS for detection
V. An extraction procedure for mammary glands

A. Recovery experiments

B. Internal standards

VI. Preliminary application of the extraction protocol to glands
from mice fed diets differing in the amount and type of fat

VII. Validation of the extraction methodology
A Method precision as a function of storage time
B. Chiral analysis of HODE enantiomers by HPLC

1. Chiral chromatography with a DNBPG
HPLC column

2. Chiral analysis of 13- and 9-HODE from
tissue samples

References

Chapter 3: Development of a New Methodology for Separation of HODE
Enantiomers

1. Selection of a different chiral column
11. Selection of a mobile phase

A. Background

19

21

22

23

23

24

27

31

32

35

37

44

44

47

52

S4

57

61

61

62

62

vii
B. Selection of a mobile phase

C. Alteration of the mobile phase

III. Application of the HPLC conditions to enantiomer separation
of other monohydroxylated polyunsaturated fatty acids

References
Chapter 4: Application of the Extraction Methodology to the Analysis of
Mammary Gland Tissue from Mice Fed Variously Controlled
Fat Diets
1. Experimental
11. Results
111. Data analysis and discussion
A Concentration of HODEs
B. Chiral‘analysis
IV. Summary

V. Conclusion

References

65

68

68

74

76

76

77

77

’77

8O

84

85

86

Table 1.1

Table 2.1

Table 2.2

Table 2.3
Table 2.4

Table 2.5

Table 2.6

Table 2.7

Table 2.8

Table 4.1
Table 4.2
Table 4.3
Table 4.4

Table 4.5

LIST OF TABLES

Important polyunsaturated fatty acids.

Percent recovery for 3H-13--HODE from rat mammary
tissue extract. '

Concentration Of 9- and 13-HODE in mammary gland tissue
from mice fed three different diets.

Analysis of variancefor 9-HODE.
Analysis of variance for 13-HODE.

Tukey’s test results for differences among the levels of 9-
and 13-HODE among the different diet groups.

ConcentratiOn of 9-HODE in aliquots of a common pool of
mammary gland tissue as a function of storage time and
temperature.

Concentration of lB-HODE in aliquots of a common pool of
mammary gland tissue as a function of storage time and
temperature.

Chiral analysis of tissues for autooxidation during sample
workup.

Concentration and enantiomeric ratios for 13-HODE.
Summary of concentration data for 13-HODE.
Concentration and enantiomeric ratons‘for 9-HODE.
Summary of concentration data for 9-HODE.

Summary of chiral data for both 9- and l3-HODE.

viii

page number
2

35
41
42
42
43

46

46

56

78
78
79
79

80

Table 4.6

Table 4.7

ix
Composition of the positional and optical isomers formed by
various lipoxygenases from linoleic acid (1).

Summary of the quantitative data from the diet study. All
data reported as mean :t standard error of the mean.
P < 0.05 for c vs. d, c vs. e, and d vs. e.

82

84

Figure 1.1

Figure 1.2

Figure 1.3
Figure 1.4

Figure 1.5

Figure 2.1a
Figure 2.1b
Figure 2.2

Figure 2.3a
Figure 2.3b

Figure 2.4

Figure 2.5

Figure 2.6

LIST OF FIGURES .

Trend in breast cancer incidence vs. dietary fat
consumption among countries (2).

Free radical oxidation of fatty acids.

Structures of representative biologically active oxidized
fatty acids. -

Structures of representative biologically inactive fatty
acids.

Lipoxygenase pathway of linoleic acid metabolism.

HPLC chromatogram of a lB-HODE ME standard,
A = 235 nm.

HPLC chromatogram of a methylated mammary gland
extract, 3. = 235 nm.

UV absorption spectrum of the peak at 25.97 min. in the
chromatogram from the mammary gland extract.

Reaction of 9-HODE with diazomethane to form 9-HODE
methyl ester.

Reaction of 9-HODE ME with BSTFA to for the methyl
ester/TMS ether.

GC/MS characteristics of 13- and 9-HODE ME/I‘MS.

RTIC and mass chromatograms of a mammary gland
extract.

Extraction procedure for mammary glands.

page number
4

11

12

15

25

25

26

28

28

29

3O

33

Figure 2.7

Figure 2.8

Figure 2.9

Figure 2.10

Figure 2.11

Figure 2.12

Figure 2.13a

Figure 2.13b

Figure 2.14

Figure 3.1

Figure 3.2

Figure 3.3

xi
RTIC and selected mass spectra recorded during the

MS analysis of a mammary gland extracted utilizing
the new procedure.

Calibration curves for 9-and 13-HODE ME/I‘MS.

SIM chromatogram of mammary gland extracts: m/z 382
represents the analyte, and m/z 386 represents the
oxygen-18 labeled internal standard.

9- and 13-HODE enantiomers.

Representation of three point interaction necessary for
chiral recognition.

Model used to interpret the stereoselectivity obtained with
charge-transfer based chrial selectors (51).

HPLC chromatogram of 13(S,R)- and 9(S,R)-HODE ME,
J.T. Baker Column, 71. = 234 nm.

HPLC chromatogram of 13(S,R)- and 9(S,R)-HODE ME
plus added S enantiomer of each, J.T. Baker Column,
A = 234 nm.

HPLC chromatogram of 13(S,R)- and 9—(S,R)-HODE
ME, Regis columns A = 234 nm.

Chromatogram resulting from chiral analysis of a mixture
of l3(PcS)cand 9(R,S)-HODE ME on a Chiralcel OD
column, hexanezlPA 100:1, flow = 1 mL/min.. The
enantiomers are not resolved under these conditions.

Chromatogram resulting from enantiomeric separation of
13(R,S)- and 9(R,S)-HODE methyl esters with a Chiralcel
OD chiral stationary phase, pentanezmethanol 180: 1, flow
= 1 mL/min.. The methyl esters of 13(S)- and 9(S)-HODE
have been added to determine the retention order of the
enantiomers.

Chromatograms resulting from enantiomeric separation of
13(R,S)- and 9(ILS)-HODE methyl esters, Chiralcel OD
CSP, pentanezmethanol 180:1, flow rate = 2 mL/min..

34

38

4o

48

49

50

53

53

55

63

66

67

Figure 3 .4

Figure 3.5

Figure 3.6

Figure 3.7

Figure 4.1

xii
l3(R,S)- and 9(R,S)-HODE ME, hexane:methanol 180: 1,
flow = 1 mL/min.

Chromatograms resulting from enantiomeric separation of
HEDE methyl esters (hexanezmethanol 180: 1,
flow = 1 mL/min.).

Chromatograms resulting from enantiomeric separation of
HETE methyl esters (hexanezmethanol 180:1,
flow = 1 mL/min.).

Published results for the analysis of I-IETE ME
enantiomers with a hexanezlPA mobile phase and a Pirkle
DNBPG chiral stationary phase (1). '

Chromatogram resulting from enantimeric separation of
13(R,S)- and 9-(R,S)-HODE methyl esters from a
mammary gland extract.

69

70

71

72

81

13-HODE

9-HODE

ANOVA

BSTFA

DNA

DNBPG

EGF

EI

GC

GC/MS

HEDE

HETE

HPLC

IPA

m/z

LIST OF ABBREVIATIONS

13-hydroxyoctadecadienoic acid
9-hydroxyoctadecadienoic acid

analysis of variance
bis(trimethysilyl)trifluoroacetamide
deoxyribonucleic acid

dinitrobenzoyl phenylglycine

epidermal growth factor

electron impact ionization

gas chromatography

gas chromatography/mass spectrometry
hydroxyeicosadienoic acid
hydroxyeicosatetraenoic acid
high-performance liquid chromatography
2-propanol

molecular ion

mass to charge ratio

methyl ester

xiii

xiv
MFJTMS methyl ester/trimethylsilyl

PUFA polyunsaturated fatty acid

RSD relative standard deviation

RTIC reconstructed total ion chromatogram
SHE ' Syrian hamster embryo

SIM selected ion monitoring

TFA trifluoroacetic acid

TLC thin-layer chromatography '

TMS trimethylsilyl

UV ultraviolet

Chapter One: Dietary Fat and Breast Cancer

L Introduction:

Human and animal studies have shown that there is a positive association between
mammary (breast) cancer and total fat intake. Animal experiments on dietary fat and
mammary cancer have shown that increasing the amount of dietary fat increases mammary
tumorigenesis, whether measured in terms of incidence or multiplicity of tumors. These
studies have also shown that polyunsaturated fats stimulate tumorigenesis whereas
saturated fats do not, unless the saturated fat is supplemented with a minimal amount of
linoleic acid. It has been postulated that oxidized unsaturated fatty acids may play a role in
stimulating the tumorigenic process. This dissertation describes the development of
analytical methodology to isolate, identify, and quantitate hydroxylated fatty acid
metabolites of linoleic acid in mouse mammary glands.

1]. Fatty acids

The fatty acids of plant, animal, and microbial origin generally contain even
numbers of carbon atoms in straight chains, with a carboxyl group at one end and with
double bonds in the cis configuration. In animal tissues, the common fatty acids vary in
chain-length from 14 to 22 carbons, and may have one to six double bonds. The most
abundant saturated fatty acids (no double bonds) are straight-chain compounds with 14,
16, and 18 carbon atoms. The fatty acid with 18 carbon atoms and the structural formula, ,
CH3(CH2)15COOH, is systematically named octadecanoic acid, having the common name

stearic acid. This acid may also be termed a C13 fatty acid, or more specifically an 18:0

2
fatty acid where the first number refers to the number of carbon atoms in the chain. The

number following the colon depicts the number of double bonds.

Polyunsaturated fatty acids (PUFA), those fatty acids with more than one double
bond, can be subdivided into families according to their derivation from specific
biosynthetic precursors. Each of the families contain from two up to a maximum of six cis-
double bonds, separated by single methylene groups (methylene-interrupted unsaturation).
Some of the more important PUFA are listed in Table 1.1.

Table 1.1 Important polyunsaturated fatty acids.

 

 

Systematic name Common name Shorthand notation
9,12-octadecadienoic acid linoleic acid 18:2 '
6,9,12:-octadecatrienoic acid y-linolenic acid 18:3
8,11,14-eicosatrienoic acid homo-y-linolenic acid 20:3

' 5,8,11,14-eicosatetraenoic acid arachidonic acid 20:4
9,12,15-octadecatrienoic acid or-linolenic acid . g , 18:3

 

Linoleic acid is the most wide-spread PUFA. It is found in most animal and plant
tissues. Linoleic acid is an essential fatty acid in animal diets, as it cannot be synthesized in
animal tissues yet is required for normal growth, reproduction, and healthy development.
This is because enzymes in animals are only able to insert new double bonds between an
existing double bond and the carboxyl group. Therefore, linoleic acid serves as a precursor

to a family of fatty acids that are formed by desaturation and chain elongation.

III. Diet and breast cancer
A. Human population studies

Women in the United States develop cancer of the breast more frequently than at
any other anatomical site (1). A tremendous variability in breast cancer incidence is seen
among countries (see Figure 1.1 (2)). Those countries with the highest incidence are ones
with a “Western” culture such as northern Europe and North America. Figure 1.1
demonstrates that there is a more than fivefold range between those countries with. the
highest and lowest rates of breast cancer. A wide international variation in per capita fat
consumption also exists. The fat consumption-breast cancer relationship in international
population studies is direct and linear. As can be seen, those countries with high fat
consumption rates have high rates of breast cancer, those with low fat consumption have
low breast cancer rates. Prentice et al. (3) have reported that the strong international
correlation with breast cancer holds for total calories from fat but not from nonfat Calories.

Cohort pOpulation studies have also shown that. there is a positive association
between fat consumption and breast cancer. Howe et al. (4) reported in a Canadian
National Screening study consisting of a cohort of 56,837 women there was evidence of a
positive association between breast cancer and total fat intake. An Italian case-control
study reported that multivariate analyses of the data showed that the women in the high-
fat group (46% of ' calories from fat) had a greater breast cancer risk than those in the
control group (26% of calories from fat) (5). Studies of population migration from
countries with low breast cancer incidence to countries with high breast cancer incidence

have provided evidence implicating lifestyle, especially diet. For those individuals that

.5 83:98 mecca 53.5.8.8 «ah—80% .m> 8:365 30:8 635 E 25¢. u: onE

 

 

 

 

 

 

 

 

can. see 88.8 .2
2m 6366350 8“.
iv . any
09 amp 0: our cow on on o
r p p n n n O
829.3 m o
8360 o

cgfl. .W
. 0032 O . &
02m 9.26 0 «36900 o «Siam—Ed «5 r n W:
...... m
r m
m a
I Op 0 m
m m
m m

n
r 9 m m
m- ll
u m
m...
U.
gnaw; 1 cm W
mug-w vase: I pnthflwflo= .09». O . , m ru xmmm gum—um .0

ocaaoN 30223.0 o «350 BE< 538 o
«9335-2 2. o gases .

 

5
migrated fi'om Poland or Italy to either the United States, the United Kingdom, or

Australia it was reported that their age adjusted breast mortality rates were comparable to
those of the host country afier 10-30 years (3).
B. Animal studies:

1. Amount of fat

‘ Because it is difficult to control many variables in human studies, the use of animal

studies to model human breast cancer has allowed dietary factors to be controlled
independently. In 1942, Tannenbaum reported a high incidence of mammary carcinomas in
mice fed high-fat diets (6). Since 1942 numerous laboratories have confirmed and
extended this observation (7). It has been seen that increased amounts of ingested fat were
reported to increase the development of mammary tumors in experimental animal models.
Caloric consumption also increased mammary tumor development. An as yet unresolved
question is whether the enhancement of mammary tumorigenesis by hyperalimentation
with fat results from specific metabolic. activity of the fat, per se, or is due to excessive
energy intake (8).

2. Type of fat

It has been seen that not only the amount of fat but also the type of fat influences
the incidence of mammary tumors. High levels of saturated fatty acids from animals (beef
tallow) or plants (coconut oil or palm oil) suppress the development of mammary tumors
in experimental animals compared to high levels of polyunsaturated fatty acids from
vegetables (reviewed in 8). If saturated fatty acid diets, which are low in essential fatty

acids (i.e., linoleic acid, linolenic acid, and arachidonic acid), are supplemented with a

6
small amount of unsaturated fatty acid (linoleic acid), the inhibitory effect of the saturated

fat diet is often reversed (9, 10). Mammary tumor development is as high in rats fed a diet
containing 20% sunflower seed oil (high linoleic acid content) as it is in rats fed a diet
containing 17% beef tallow and 3% sunflower seed oil (9) or 17% coconut oil and
3% ethyl linoleate (10). These experiments show that the amount and the type of fat in the
diet is important in influencing mammary tumorigenesis in experimental animal models.
3. Proposed mechanisms
The mechanism for the stimulatory effect of the dietary fat on mammary -
tumorigenesis has not as yet been determined. Some mechanisms that have been proposed
to explain the effect include (11):
0 changes in the immune system
0 alterations of the endocrine system
0 modifications of cell membrane lipids
0 changes in prostaglandin activity
0 direct effects on tumor cell metabolism
0 changes in the activity of carcinogen-metabolizing enzymes
0 production of oxidized products from unsaturated fats
IV. Lipid oxidation
In recent years, interest in the role of lipid peroxidation and the production of
oxidized products in tissues has increased. Lipid hydroperoxides and their breakdown
products display a variety of toxic effects such as impairment of membrane function,

decreased cell membrane fluidity, inactivation of membrane bound receptors and enzymes,

7
and increased non-specific permeability of ions such as Ca2+. These lipid products have

also been proposed to contribute to inflammation, reoxygenation injury, and parasitic
infections, as well as contributing to degenerative diseases such as cancer, atherosclerosis,
rheumatoid arthritis, and aging (12). The oxidation of polyunsaturated fatty acids in vitro
has been well described, and both the products and the kinetics are well understood (13,
14).

Naturally occuning polyunsaturated fatty acids contain two to six methylene-
interrupted double bonds within the hydrocarbon. portion of the molecule. The greater the
number of double bonds the greater the oxidizability. The peroxidation reaction is
triggered by the abstraction of a wealdy bonded allylic hydrogen by a strong oxidant. The
alkyl radical of the fatty acid is stabilized by resonating to a conjugated diene system. The
addition of molecular oxygen to an alkyl radical produces a conjugated peroxy radical
which, upon abstracting a hydrogen from another polyunsaturated fatty acid, produces a
hydroperoxide and an alkyl radical (see Figure 1.2). Several mechanisms involving one—
electron transfer from a metal ion, radiation, photolysis, active oxygens, chlorinated
hydrocarbons, and radicals generated from enzymatic processes are thought to be
responsible for initiation of the radical reaction in biological systems (15).

There has been no direct evidence to support the idea that lipid peroxidation is
influencing the incidence of mammary cancer in mice fed a high-fat diet. Slaga et. al., (16)
however, reported that benzoyl peroxide, a free radical generating compound, is an in viva
promoter in the dimethylbenzanthracene-induced mouse skin model of tumorigenesis. Bull
et. al. (17) provided evidence that oxidized fatty acids containing a site of unsaturation

adjacent to the oxygen-containing functionality are mitogenic to the rat colon

Initiation \L Removal of H-
f '
\f/M
\L Molecular rearrangement

0
Major reaction \J/ 02

02 .
Lipid peroxyl radical i H- abstraction from

adjacent lipid
02H

Lipid hydroperoxide

Figure 1.2. Free radical oxidation of fatty acids.

9
in viva at the same concentrations as tumor-promoting bile acids. The mitogenic

properties of these compounds suggests a role in the enhancement of tumorigenesis by
high levels of dietary fat.

Antioxidants can act as effective scavengers of lipid peroxy radicals and/or oxygen
radicals. A number of laboratories have treated rats bearing mammary tumors with an
array of antioxidants during the ingestion of diets rich in unsaturated fatty acids and found
a decrease in tumorigenesis (18, 19). There also exists evidence to the contrary (20). It is
possible that the effects seen by antioxidants are the result of altered carcinogen
metabolism. King et al. (20) found that an antioxidant administered with a non-
metabolized carcinogen did not reduce the promotional effect of the high-fat diet. Even
though there exists conflicting evidence for the effect of lipid peroxidation on mammary
tumor promotion, the idea remains an attractive concept.

V. Structural requirements for biological activity

Bull et a1. (17) reported that primary autooxidation products of major dietary
polyunsaturated fatty acids, arachidonic and linoleic acids, were capable of stimulating rat
colonic mucosa] deoxyribonucleic acid (DNA) synthesis and omithine decarboxylase
activity in viva, two early events in the activation of mitogenesis. This group carried the
work firrther to determine the structural features of the fatty acid molecule that are
responsible for the biological activity. A site of unsaturation adjacent to an oxidized
moiety was found to be the minimal requirement. No significant differences in biological
activity were seen if the oxidized functionality was a hydroperoxide, hydroxide, or

carbonyl (21). Loss of mitogenic activity was seen with saturation of the hydrocarbon

. 10
backbone, separation of the double band from the oxidized moiety by a methylene group,

or the absence of an oxidized group. Structures of the compounds found to be active are
shown in Figure 1.3. Structures of the inactive compounds are shown in Figure 1.4.

Recently, Glasgow et al. (22) determined that enzymatically derived linoleate
metabolites were observed to enhance [3H]thymidine incorporation in epidermal growth
factor (EGF)-activated Syrian hamster embryo (SI-IE) cells. Autoradiographic studies
confirmed the mitogenic potential of the linoleate derivatives. Investigations into
structure-activity relationships of analogous lipid compounds found that metabolites of
arachidonic acid and linolenic acid were much less active than linoleic acid metabolites,
and the lack of activity by ricinoleic acid suggested either a requirement for an allylic
oxygenated functionality or a requirement for a diene. The structure-activity revealed that
the primary derivative of linoleic acid, 13-hydroperoxyoctadecadienoic acid, was the most
efi‘ective lipid compound in stimulating EGF-regulated DNA synthesis in SHE cells. The
carbon chain length, degree of unsaturation, type and position of oxygenated firnctionality,
double-bond geometry, and chirality are all factors that modulated the mitogenic activity
of related compounds (23).

Welsch (24) assessed the potential of 9-hydroperoxyoctadecadienoic acid and 13-
hydroperoxyoctadecadienoic acid for their ability to stimulate proliferation of primary
mouse (Balb/c) mammary gland epithelial cells in a collagen gel cell culture system. 13-
hydroperoxyoctadecadienoic acid stimulated cell proliferation more than did the parent
fatty acid, linoleic acid. In contrast, 9-hydroperoxyoctadecadienoic acid inhibited

proliferation of these cells. It was concluded that specific oxidation products of linoleic

11

OOH (OH)

CH3(CH2)4 (CH2)2000H

15-hydr0peroxyeicosatetraenoic acid
15-hydroxyeicosatetraenoic acid

OOH (OH)

W

CH3(CH2)4 (CH2)6000H

13-hydroperoxyoctadecadienoic acid
13-hydroxyoctadecacfienoic acid

OOH (OH)

/\/\

CH3(CH2)7 (CH2)6CX)OH

10-hydroperoxyoctadecenoic acid
10-hydroxyoctadecenoic acid

0
M
CH3(CH2)7 (CH2)GCOOH

10-ketooctadecenoic acid

Figure 1.3: Structures of representative biologically active oxidized fatty acids.

CH3(CH2)4 (CH2)COOH
Arachidonic acid

m

CH3(CH2)4 (CH2)7COOH

CH3(CH2)7 (CH2)7COOH
Oleic acid
(CH2)7COOH

CH3(CH2)8
Eladic acid

OCH (CH)

CH3(CH2)5 (CH2)10COOH

12-hydroperoxyoctadecanolc acid
124iydroxyoctadecanoic acid

M
CH3(CH2)5 (CH2>7C°°t[
Ricinoleic acid

Figure 1.4: Structures of representative biologically inactive fatty acids.

13
acid could provide critical mitogenic stimuli to normal and neOplastic mammary gland

epithelial cells, particularly in animals fed a high-fat, high linoleic acid containing diet.

All of the active compounds studied above can arise fiom autooxidation of the
major dietary unsaturated fatty acids. The primary products of fatty acid autooxidation are
allylic hydroperoxides containing trans double bonds (oleic acid derived) or a cis-trans
conjugated diene system (linoleic and arachidonic acid derived) (14). The hydroxy fatty
acids can arise from reduction of the hydroperoxides by either tissue or bacterial
peroxidase activity. These types of hydroperoxide reductions are well precedented both in
viva and in vitra (25, 26, 27)., Fatty acid hydroperoxides may also be produced by in situ
generation in various tissues. Funk and Powell (28) have shown that linoleic acid can be
metabolized, via a hydroperoxide intermediate, to 9-hydroxyoctadecadienoic acid (9-
HODE) and 13-hydroxyoctadecadienoic acid (1 3-HODE) by the prostaglandin H synthase
activity present in particulate fractions of the aorta. Similar reactions catalyzed by
prostaglandin synthase or lipoxygenases have been reported in VX2 carcinoma cells (29),
rabbit peritoneal tissue (30), human (31) and porcine leukocytes (32), and SHE fibroblast
cells (33). The lipoxygenase pathway of linoleic metabolism may be seen in Figure 1.5.

VI. Goals of this research

As stated earlier in this chapter, linoleic acid seems to play an important role in the
enhancement of mammary gland tumorigenesis in experimental animals. Unsaturated fat
diets that are high in linoleic acid stimulate tumorigenesis. Furthermore, the addition of
linoleic acid to a saturated fat diet returns the stimulatory effect. It is possible that the

hydroxylated fatty acid metabolites, 9- and 13-HODE, are playing some role in the

14
tumorigenic process. The goals of this research are 1) to develop a methodology for the

separation, isolation, and detection of 9- and 13-HODE in mammary gland tissue, and 2)
to apply this methodology to quantitating the levels of 9- and 13-HODE from mammary

glands in experimental animals fed diets that differ in the amount and the type of fat.

15
W

 

 

 

   

Linoleic Acid
COOH
/ — \
OOH i OOH
V \V
9-Hydroxyoctadecadienoic acid 13-Hydroxyoctadecadienoic acid

9—HODE 13-HODE

Figure 1.5: Lipoxygenase pathway of linoleic acid metabolism.

References

10.

11.

12.

13.

14.

15.

16.

Silverberg, E., CA. A Cancer Journalfar Clinicians 32, 1982, 15-31.

Schatzkin, A., Greenwald, P., Byar, D.P., Clifford, C.K., Journal of the American
Medical Association 261, 1989, 3284-3287.

Prentice, RL., Kakar, F., Hurstion, S., Sheppard, L., Klein, R., and Kushi,iL.H.,
Journal of the National Cancer Institute 80, 1988, 802-814.

Howe, GR, Freidenreich, C.M., Jain, M., and Miller, AB., Journal of the

_ National Cancer Institute 83, 1991, 336.

Toniolo, P., Riboli, E., Protta, F., Chanel, M., and Cappa. AP.M., Journal of the

National Cancer Institute 81, 1989, 278.

Tannenbaum, A, Cancer Research 2, 1942, 49-53. '
Welsch, C.W., and Alysworth, or, In Dietary Fats andHealth, E.G. Perkins and
WJ. Visek, eds.', Monograph No. 10, Charnpaign, IL, American Oil Chemists
Society 1982. '

Welsch, C.W., Cancer Research(Suppl.) 52, 1992, 20403-20485.

Hopkins, GI. and Carroll, K.K., Journal of the National Cancer Institute 62,
1979, 1009-1012.

Hopkins, G.J., Kennedy, TO, and Carroll, K.K., Journal of the National Cancer
Institute 66, 1981, 517-522.

Clinten, SK. and Visek, W.J., In Xenobiotic MetabolisrmNutritianal Efl'ects,
“Nutrition and Experimental Breast Cancer: The Effects of Dietary Fat and
Protein” 1985, The American Chemical Society.

Frei, B., Analytical Biochemistry 175, 1988, 120-130.

Porter, N.A., Accounts of Chemical Research 19, 1986, 262-268.

Porter, N.A., Journal of the American Chemical Society 103, 1981, 6447-6455.
Pryor, W.A., Lipids 1976, 370-379.

Slaga, T.J., Klein-Szanto, A.J.P., Triplett, L.L., Yotti, LP, and Trosko, J E
Science 213, 1981, 1023-1025.

16

17.

18.

19.

20.
21.

22.

23,

24.

25.

26.

27.

28.

29.

30.

31.

32.

17

Bull, AW., Nigro, N.D., Golembieski, W.A., Crissman, J.D., and Mamett, L.J.,
Cancer Research 44, 1984, 4924-4928.

Horvath, RM. and Ip, C., Cancer Research 43, 1983, 5335-5341.

Alysworth, C.F., Cullum, M.E., Zile, M.H., and Welsch, C.W., Journal of the
National Cancer Institute 76, 1986, 339-345.

King, M.M, McCay, RB, and Kosnake, S.D., Cancer Letters 14, 1981, 219-226.
Bull, AW., Nigro, N.D., Mamett, L.J., Cancer Research 48, 1988, 1771-1776.

Glasgow, W.C., Afshari, C.A., Barrett, J.C.,and Eling, T.E., Journal of
Biological Chemistry 267,1992, 10771-10779.

Glasgow, W. C. and Eling, T. E. ,Archives of Biochemistry and Biophysics 311,

. 1994, 286-292.

Welsch, C. W. Free Radical Biologr & Medicine 18, 1995, 757-773.

Wendel, A, In: Enzymatic Basis of Detoxtcatron W. B. J akoby (Ed. ), Vol. 1, New
York, Academic Press, 1980.

Forstrom, J. W, Stults, F. H., and Tappel, A. L. ,Archives of Biochemistry and
Biophysics 193, 1979, 51-55.

Prohaska, JR and Ganther, H.E., Biochemical and BiOphysical Research
Communications 76, 1979, 437-445.

Funk, CD. and Powell, W.S., Journal of Biological Chemistry 260, 1985, 7481-
7488.

Hubbard, W.C., Hough, A.J., Brash, AR., Watson, J .T., and Oates, J.A,
Prostaglandins 20, 1980, 431-447.

Claeys, M., Coene, M.C., Herman, AG., Jouvenaz, G.H., and Nugteren, D.H.,
Biochimica et Biophysica Acta 713, 1982, 160-169.

Claeys, M., Coene, M.C., Herman, AG., Jouvenaz, G.H., and Nugteren, D.H.,
Biochimica et Biophysica Acta 837, 1987, 35-51.

Reinaud, O., Delaforge, M., Boucher, J.L., Rocchiccioli, F., Mansuy, D.,
Biochemical and Biophysical Research Communications 161, 1989, 883-891.

18

33. Glasgow, WC. and Eling, T.E., Molecular Pharmacology 38, 1990, 503-510.

Chapter 2: Development of Analytical Methodology

Oxygenated products (hydroxy- (1, 2), keto- (3, 4), epoxy- (5), and aldehydes (6))
have been isolated from many different body fluids and tissues. In general, each tissue
studied requires its own methodology for extraction and/or removal of matrix interferents.
There also exists a wide array of detection methods utilized to analyze the tissue extracts,
yet none have been specifically developed for mammary gland tissue. Since it is the goal of
this research to analyze hydroxylated fatty acids fiom mammary gland tissue, a method for
the quantitative extraction and isolation of the analytes of interest had to be developed.

L Characterization of fatty acids by GC/MS

Gas chromatography (GC) is a widely used and popular technique for the analysis
of fatty acids after they have been converted to esters in most cases. However, fatty acids
may only tentatively be identified by' gas chromatographic retention times alone, but GC
combined with derivatization and chemical degraditive or spectroscopic procedures,
especially mass spectrometry, is an extremely powerful means of characterization of fatty
acids (7).

Mass spectrometry has become a valuable tool for the identification of lipids and
fatty acids. There are many ionization techniques that may be utilized for the analyses of
lipids, but one that is most common is electron impact (E1) ionization. The principle of this
technique is that organic molecules in the gas phase are bombarded by a beam of electrons
in an ion source of a mass spectrometer producing positively-charged, radical ions. The

positively-charged intact molecule is known as the molecular ion (M“). This ion may then

19

20
fragment to form other smaller positively-charged ions. All of the ions are then accelerated

out of the ion source, separated according to their mass to charge (tn/z) ratios, and
detected. A histogram plot of abundance vs. W2 is a mass spectrum. Often it is possible to
deduce the structure of a fatty acid by its mass spectrum, i.e. by its fragmentation pattern.
The combination of mass spectral and gas chromatographic retention time data may aid in
eliminating alternative molecular structures.

Before gas chromatography/mass spectrometry (GC/MS) may be utilized, the fatty
acids must be vaporized. Free fatty acids contain a polar functional group and typically are
derivatized to make them more volatile and non-polar, improving peak shape and
resolution simultaneously. A popular and widely used derivative for the analysis of fatty
acids is the formation of the methyl esters. Methyl esters are easily prepared by reaction
with ethereal diazomethane with a catalytic amount of methanol, or by reaction with
methanol-BF; or methanol-HCl (8). The methyl ester derivatives of long-chain saturated
fatty acids are easily characterized by electron ionization mass spectrometry. Their spectra
are characterized by a prominent molecular ion and other significant ions consisting of m/z
= M”- 31 and M”— 43. There also exists a series of ions of the general formula
[CH3OOC(CH2),,]+ with peaks located at m/z = 87, 143, and 199 (9). The base peak, m/z =
74, represents the “McLafferty rearrangement ion”, which is formed after hydrogen
rearrangement and cleavage of the parent molecule beta to the carboxyl group (10).

Free fatty acids may also be derivatized into other esters to increase their volatility,
or to improve their chromatographic or mass spectrometric characteristics. When a higher

molecular weight is desired, trimethylsilyl (TMS) or t-butyldimethylsilyl (t-BDMS) esters

, 21
may be formed. Pentafluorobenzyl esters make fatty acids suitable for electron capture

detection or negative ion mass spectrometry. The determination of the position of
functionalities such as double and triple bonds, branches, or cyclopropane rings have been
described using pyrrolidides (11), picolinyl (12), piperidyl and morpholinyl esters (13), 2-
alkenylbenzoxazoles (14), and 4,4-dimethyloxazoline derivatives (15).

II. Characterization of hydroxylated fatty acids by GC/MS

Because of the hydroxyl functionality, the esters of hydroxylated fatty acids are
still quite polar and somewhat involatile. It is, therefore, desirable to convert the hydroxyl
group to an ether to improve the gas chromatographic and mass spectrometric
characteristics of the compound. The choice of the derivative will depend on the
compound, and it may be necessary to prepare several types of derivatives to confirm
identification. Some common hydroxyl group derivatizations include acetylation (16),
trifluoroacetylation (17), and trimethylsilylation (18, 19, 20).

Trimethylsilyl (TMS) ether derivatives are especially useful with EI mass
spectrometry. The fragmentation of these molecules can be diagnostic for locating the
position of the hydroxyl functionality in the carbon chain as major fragmentation occurs
alpha to the carbon containing the functional group. A discernible peak for the molecular
ion is not usually present, but an [M - 15]* ion is formed allowing molecular weight and
carbon chain length to be determined.

The positions of hydroperoxy groups formed during the oxidative deterioration or
degradation of lipids are most commonly determined by reduction of the hydroperoxide to

a hydroxide and hydrogenation of any double bonds, followed by methylation and

22
conversion to the TMS ether prior to analysis by GC/MS. This approach has been of use

in the identification of eicosanoids derived from arachidonic acid (21, 22, 23, 24).

If the hydroxylated fatty acid is unsaturated, it is essential to prepare the methyl
ester TMS ether if the mass spectra are to be interpretable (25, 26). The position of the
double bond(s) relative to the TMS ether affects the fragmentation pattern. With esters in
which the hydroxyl group is allylic to the double bond(s), no fragmentation occurs
between them. Cleavage occurs on either side of the hydroxyl and double bonds. When
there is one methylene group between the double bond and the carbon linked to the TMS
ether, the ions caused by fi'agmentation alpha to the TMS carbon on the side of the double
bond are most abundant. Ifthe functional groups are separated by two methylene groups,
the two peaks in the mass spectrum representing alpha cleavage on either side of the TMS
carbon are of equal intensity (7).

III. Methods other than GC/MS for the analysis of fatty acids

Thin-layer chromatography (TLC) has been used in lipid analysis for many years
on both an analytical and preparative scale. TLC procedures with silica gel G (with a
calcium sulfate binder) have been employed most frequently for lipid class separations. A
common solvent elution system consists of hexane-diethyl ether-formic acid (80:20:2 by
volume). Bands on the TLC plate may be detected by spraying with an 0.1% (w/v)
solution of 2',7'-dichlorofluorescein in 95% methanol and viewing under ultraviolet (UV)
light. The lipids appear as yellow spots against a dark background, and the lipid fractions

may be recovered from the plate by elution with solvents for further analysis.

23
High-performance liquid chromatography (HPLC) is increasingly being utilized for

separation and detection of fatty acids and other lipid classes (27). Most lipids lack
chromophores for spectrophotometric detection, but the absorbance of double bonds in
the low ultraviolet (UV) range can be used successfully if care is taken in the choice of
solvents for the mobile phase. Only a few {solvents are transparent (methanol, hexane,
acetonitrile, isopropanol, and water) in the low UV range, and the molar extinction
coefficients are low so that impurities in the solvent can interfere with the signal of the
compounds of interest. However, many different separations have been reported using
both normal and reverse phase, and well as with isocratic and gradient elution (27).
IV. Initial attempts to extract mammary gland tissue

The high fatty acid content of the mammary gland extract was shown to produce a
high chemical background when published isolation procedures for other tissues were
applied to the mammary gland (28). As a starting point to learn practical methodology, the
initial attempts at extraction of 9- and 13-HODE from mammary gland tissue were begun
at the laboratory of Dr. Arthur Bull, Assistant Professor of Chemistry at Oaldand
University, Rochester, Michigan. Dr. Bull’s laboratory has developed an extraction
procedure to isolate 13-HODE from rat colon tissue using HPLC with UV detection as a
means of analysis.
A. Extraction

A one-gram manunary gland tissue sample was stirred with 10 mL of a 2:1
chloroformzmethanol (29) mixture, the supernatant was saved and the extraction was

repeated with 5 mL of solvent; the two extracts were combined and the final volume

24
determined prior to the addition of 1/4 volume of water for an aqueous phase extraction

(30, 31), resulting in a biphasic system. The upper phase was removed with a pipet and
discarded. Solvent from the lower phase was removed in vacuo, prior to methylation with
diazomethane and methanol for 30 minutes at room temperature. The sample was dried
under nitrogen, reconstituted in 100 uL of 20 % acetonitrile, 0.1 % trifluoroacetic acid
(TF A) in water for analysis by HPLC.
B. High-performance liquid chromatography

The entire 100 uL was injected onto a 25-cm octadecasilica column. Detection of
the analytes was performed with a Perkin-Elmer LC-235 diode array detector at a
wavelength of 235 nm to detect the absorption of the conjugated diene in the 13-HODE
methyl ester (ME). The chromatographic conditions were as follows: solvent A was 20 %
acetonitrile, 0.1 % TFA in water; solvent B was 100 % acetonitrile with 0.1 % TFA The
HPLC run consisted of a linear gradient from 50% A to 100% B over 30 minutes with a
20-minute hold. The flow rate was 1 mL/minute. Figure 2.1a is a chromatogram of a
standard of 13-HODE methyl ester which elutes at 25.45 minutes. Figure 2.1b is a
chromatogram of the methylated mammary gland extract. A spectrum of the peak eluting
at 25.97 min. (Figure 2.2) shows possible conjugated diene features, but the absorbance at
235 nm is too intense to give a definitive result, and the elution time is not the same as that
for the standard. From this work it was concluded that the extraction procedure may be

applicable to mammary gland tissue, but a detection method with greater selectivity was

necessary.

ooaocnomc>

 

25

 

 

 

 

 

 

13-HODE Standard Mammary Gland Extract
. . pea!“ @ .
Retention Tune 25.34 and 25.97 mm.
25.45 min.
A
b
. T
o
r
b
a
n
c
r V e
Time Time
Figure 2.1a: HPLC chromatogram of a Figure 2.1b: HPLC chromatogram of a
13-HODE ME standard methylated mammary gland extract
k=235 nm. A=235 nm.

26

.25-9? Min.

 

 

 

 

c.5111-
nu
_e 1822 I I 9 l l l I I J l l 1 1 I l l l l l 1_1 J j
' l l '1 l
195 229 245 273 295

Figure 2.2: UV absorption spectrum of the peak at 25.97 min. in the chromatogram from

the mammary gland extract.

27

C. The use of GC/MS for detection

Because the use of UV detection was too general for this application, gas
chromatography/mass spectrometry was investigated as a method of detection. As
previously discussed, GC/MS has a long history in the analysis of lipids. This method of
detection is advantageous in this case because the analytes may be identified by retention
times and mass spectra.

Since the analytes of interest are both hydroxylated and contain points of
unsaturation, the free fatty acids were derivatized to methyl ester/TMS (MFJTMS) ethers.
Figure 2.3a shows the reaction of 9-HODE with diazomethane to form the methyl ester
and Figure 2.3b shows the subsequent reaction with bis(trimethylsilyl)trifluoracetamide
(BSTFA) to form the TMS ether. This reaction was carried out in pyridine at 60° for 30
minutes. This reaction mixture may be injected directly into the gas chromatograph.

The gas chromatographic mass spectrometric characteristics of 9- and 13-HODE
MFJTMS are shown in Figure 2.4. The fragmentation of these compounds was first
characterized by Hubbard et al. (32). The mass spectrum of 13-HODE ME/TMS has
characteristic peaks at m/z 382 (M“), 311 [M — 71]’, and 225 [M - 157T. The mass
spectrum of 9-HODE ME/TMS also has characteristic peaks at m/z 382 (MT), 311 [M —
HT, and 225 [M - 157]+. The major difference between the mass spectra of these two
isomers is the relative intensities of the peaks.

A reconstructed total ion chromatogram (RTIC) with mass chromatograms
representing the analysis of a mammary gland extract with GC/MS detection is shown in

Figure 2.5. There is evidence of the presence of both 9- and 13-HODE ME/TMS,

28

see menace... Ens 2: Ea a 5:8 as. m2 moons .6 seem 52 use...

:3 E8 _
22558;... a z 01.4.0 n : issaacgfiaemsfisaza 2232;
. _ . f0 . I o
o . _
I \ filmlzuw tom. + \
, All ,
m I o o o 0 f0 w a I o o o o
glml of
£0 _

use ease moons Ea a ogzoaoae as. moons co 888m a2 saw

us. moo—.3 moozé

:0 IO
.l \ . . I. \

29

 

 

 

 

5 ..g 1344005 9-HODE
5 ..,: rag/ms \ / ME/TMS
i

 

 

 

 

 

 

 

 

 

 

 

 

: e
'- i
l l
i l
E t
1 .. :
ua

 

 

 

311

 

Figure 2.4: GC/MS characteristics of 13- and 9-HODE MEI’I'MS.

30

 

rind

mkflS

003.3357)

mhfll

 

 

Mann

 

 

 

 

 

thE

if“?!

ooa~o=cc>

 

 

 

mama

 

 

 

 

 

1550

165° Seen Nmaeelm

Figure 2.5: RTIC and mass chromatograms of a mammary gland extract.

31 .
however, interferences also exist so that peaks representing these compounds are barely

detectable in the RTIC making it impossible to quantitate these compounds. From this
result it was decided that this extraction procedure was not directly applicable to
mammary gland tissue.
V. An extraction procedure for mammary glands

The first step taken to develop an extraction procedure for mammary glands was
to use ethanol as an extraction solvent. Ethanol was tried for two reasons: 1) all of the
hydroxylated fatty acid standards purchased were dissolved in ethanol, and 2) Lehmann et
al. (33) utilized ethanol as an extraction solvent for the extraction of HODEs and
hydroxyeicosatetraenoic acids (HETEs) from epidermal tissue. This proved to give similar
results to the chloroformzmethanol extraction, i.e. much interference. A silica solid-phase
extraction step was added after methylation to remove interferences (34).

The 200 mg silica columns are preconditioned by washing with 2 mL hexane, 1 mL
50% ether/hexane, and 2 mL hexane. The methylated extract is reconstituted in 1 mL
hexane and applied to the silica column which is then washed with 2 mL 5% ether/hexane
and eluted with 5 mL 50% ether/hexane and 2 mL ethyl acetate. Results with either the
ethanol or the chloroformzmethanol extractions with the silica column cleanup were still
not satisfactory.

We observed that mammary glands extracted with ethanol would produce a pellet
upon centrifugation, whereas the chloroformzmethanolzwater extractions resulted in a
suspension where the densities of the tissue and the solvent were similar so the tissue

would not form a pellet making it difficult to obtain a clean extract.

32
The use of the two solvent extractions was attempted; ethanol was used first to

extract the lipids and to pellet out the tissue residue. Afier evaporation of the ethanol in
vacuo, a chloroform:methanol:water extraction was used to remove any non-lipid material
remaining. A clean extract resulted following silica gel cartridge cleanup.

The extraction procedure in its final format may be seen in Figure 2.6. A
mammary gland tissue sample (100-200 mg) is pulverized in liquid nitrogen prior to
extractions in 10 mL, then 5 mL of ethanol. The ethanol extracts are combined and are
centrifuged, the supernatant transferred to, a pear-shaped flask, and the ethanol is
evaporated in vacuo. A second extraction in dichloromethanezmethanol (2:1) and water
removes the remaining non-lipid material. The switch was made to dicholoromethane fi'om
chloroform as it has been reported that dichloromethane is less likely to promote oxidation
of the lipids in the sample (35, 36). The mixture of dichloromethane:methanol:water
separates into two phases; the upper phase is pipetted off and discarded. The remaining
solvent is evaporated in vacuo, and the sample is methylated with diazomethane and
methanol. After 30 minutes the solvent is evaporated under a stream of nitrogen, the
residue is reconstituted in 1 mL of hexane and applied to silica columns according to the
above procedure. Figure 2.7 shows the RTIC and selected mass spectra obtained during
GC/MS analysis of a mammary gland extract processed according to the new procedure.
As can be seen, the interferences have been removed.

A. Recovery experiments
Recovery experiments were performed in the laboratory of Dr. Arthur Bull at

Oakland University. Tritium-labeled 13-HODE was prepared by soybean lipoxygenase

33

Tissue (100-200 mg)

Pulverize the tissue in liquid nitrogen

Extract in ethanol and centrifirge

Extract in CHzClzzMeOH (2: 1)
Water '
to form a biphasic system

Methylate with diazomethane and methanol

Solid phase extraction on silica columns
dissolve sample in hexane and apply sample to column
wash with hexane
5% ether/hexane
elute with 50% etherzhexane
ethyl acetate

l
Derivatize with BSTFA

l
Analyze by GC/MS

Figure 2.6: Extraction procedure for mammary glands.

34

 

 

 

 

 

 

 

5095 ”DE W 73 94-005 m P
73 ‘ I-
' ' ' 225

l i

1 r
130 ' P

25 130
171 311
mr'T m . 311 m r

 

 

 

 

 

v v I I I I I v I 1 V v v v I ‘ Wj Y I v 1 I v I Y ‘ I 1 1 V
2000 2100 2200 {mu-1915‘
trme(mrn.)

30.00 31.49 32.99 34.49

Figure 2.7: RTIC and selected mass spectra recorded during the GC/MS analysis of a
mammary gland extracted utilizing the new procedure.

35
catalyzed oxygenation of 3H-labeled linoleic acid followed by sodium borohydride

reduction of the resulting 13-hydroper0xyoctadecadienoic acid (3 7, 38, 39). The labeled
HODE was dissolved in ethanol, and the standard purity was determined by HPLC. The
concentration of the standard solution was 3.1 mM with a specific activity of 2.68
uCi/umol.

The recovery experiments were performed by adding the labeled standard to 150
mg rat mammary gland tissue sample prior to extraction with. ethanol. An aliquot of a
known volume of solvent was removed in duplicate after the initial ethanol extraction,
following methylation, and after elution from silica columns to monitor losses throughout
the extraction procedure. Radioactivity was measured by scintillation counting. The
average recovery was 72% with a 5% relative standard deviation. (RSD). Recoveries are
summarized in Table 2.1.

Table 2.1: Percent recovery for 3H-13-HODE from rat mammary gland tissue extract.

 

 

 

Recovery (%) mean
(RSD)
Extract #1 Extract #2 Extract #3
after ethanol extraction 100 100 97 99 (1.7)
after methylation 74 79 70 73 (4.9)
final 72 76 69 72 (4.8)

 

 

B. Internal standards

Standards are essential for calibration of the instrument and for quantitation of the
analyte. The addition of an internal standard increases the accuracy of the experiment
especially if the internal standard has the same or similar chemical and physical-

characteristics as the analyte(s) of interest. If the internal standard of similar chemical and

36
physical properties is added at the beginning of the assay, it will undergo the same losses

during chemical reactions (derivatizations) and physical manipulations such as extractions
and transfers fiom flask to flask. Because the internal standard experiences the same losses
as the analyte, the quantity of the internal standard relative to the analyte accounts for
losses of the analyte. Quantification of the analyte can be canied out with the use of
standard curves.

An ideal internal standard for ms is a stable isotope-labeled analog of the
analyte of interest. Biochemical synthesis from deuterium-labeled arachidonate has been
described n.- isotopically labeled leukotrienes and HETEs, but dilution with arachidonic
acid endogenous to the biological system reduces the isotopic purity of the labeled
products (40). Chemical synthesis of deuterated lipoxygenase products has also been
described (41). An alternative for eicosanoid internal standards is to incorporate stable
isotope oxygen-18 [“0] atoms into the carboxyl moiety of the fatty acid. The exchange of
the carboxyl oxygen atoms with oxygen-18 water can be accomplished with acid catalysis,
basic ester hydrolysis, 0r enzymatic means (42, 43).

For the preparation of oxygen-18 labeled HODEs as internal standards in this
project, the procedure of cholinesterase-catalyzed carboxyl oxygen exchange was utilized
(44, 45). Oxygen-18 labeled HODEs were prepared by the addition of 50 mg of 9- or 13-
HODE dissolved in 20 uL methanol to 500 1.1L H2‘80 (97-98% l"0) containing 30 units of
butyrylcholine esterase. After incubation for 24 hours at 37°, the reaction mixture was
extracted four times with ethyl acetate, combined, dried under nitrogen, and reconstituted

in ethanol. The concentration of the solution was determined by measuring the absorbance

37
at a wavelength of 234 nm and using a value of 23000 moi L‘l cm" for the molar

absorptivity constant (8). Standards were stored at -20°.

Oxygen-18 labeled HODEs were selected for internal standards for this research
because of the simplicity of preparation and a four mass unit increase in molecular weight.
The internal standards are added to the sample prior to extraction with ethanol. The mass
spectrum of the labeled m02-HODE has a peak at m/z 386 for M“, and a peak at m/z 315
representing the fi'agment ion containing the label. The use of either selected ion
monitoring (SIM) 0r repetitive scanning (mass chromatograms) to monitor the ion current
at m/z 386 and m/z 315 gives a quantitative index of the internal standard, and monitoring
the ion current at m/z 382 and m/z 311 gives a quantitative index of the amount of analyte ‘
introduced into the mass spectrometer. A plot of the ratios of the peak areas of the
internal standard to the analyte vs. the analyte concentration is linear as can be seen in
Figure 2.8.

Concentrations of unknowns may be calculated in the following manner: the peak
areas of the unknown analyte and the added internal standard are acquired from the mass
spectrometer response. The ratio of these two numbers gives the ratio of analyte to
intemal standard. Since the amount of internal standard added and the tissue sample size
are known, the concentration of analyte can be calculated by the following equation (46):

Area m / z 382 x amount of int. stand. added _ amount HODE
Area m / z 386 milligrams tissue mg. tissue

 

VI. Preliminary application of the extraction protocol to glands from mice fed diets
differing in the amount and type of fat

Female athymic nude mice (Balb/c) were divided into three groups to be fed three

different diets. The diets were isocaloric and chemically purified differing only in the

 

38

 

 

 

Standard Curve for 9-HODE
y = 0.0687x - 0.095
1.8 ~- R’ - 0.996
0 1.6 ..
3 1.4 er
5 1.2 T
1 ..
E 0.0 ~-
"‘ 0.6 ~~
g 0.4 ..
02 4.
0 4 i i a
0 5 10 15 20 25
ng/uL
Standard Curve for til-HODE
y - 0.0648x - 0.0295
1.6 -- R2 - 0.996
a 1.4 «~
8 1.2 ~~
g 1 ..
a 0.8 ~~
S 0.6 «~
g 0.4 ..
02 -
0 + i i 4
0 5 10 15 20 25
"9’“!-

Figure 2.8: Calibration curves for 9- and 13-HODE ME/T MS.

39
amount and type of fat. One diet group was fed a high-saturated-fat diet (20% beef tallow

by weight), one was fed a low-unsaturated-fat diet (5% corn oil by weight), and one was
fed a high-unsaturated-fat diet (20% corn oil by weight). Corn oil was used not only
because it is high in unsaturated fat, but also because it is high in linoleic acid, the
precursor to 9- and 13-HODE (see Figure 1.5). The mice in each group were fed these
diets beginning at three weeks of age. They were fed these diets for 6 weeks before being
sacrificed. Three pairs of mammary glands from each mouse were excised, placed in
polypropylene vials, and stored at -20° for approximately one month prior to being
extracted.

Mammary glands (100-200 mg) were extracted following the methodology
outlined above. Internal standards were added to the sample prior to the addition of
ethanol. One Sample from each of the diet groups was extracted per day. The amounts of
internal standards added to each sample were 862 ng of oxygen-18 labeled 9-HODE and
412 ng of oxygen-18 labeled 13-HODE. The extracts were analyzed on a JEOL AX505
magnetic sector mass spectrometer in the selected ion monitoring mode (SIM) acquiring
ion currents for W: 386, 382, 315, and 311. The extract was separated gas
chromatographically on a DB-SMS 30-m, 0.25-mm x 0.25-um, column. Five samples,
each representing a mammary gland sample from a different animal from each of the diet
groups, were extracted and analyzed except for the group on the beef tallow diet as one
sample was lost when a flask broke. An example of the chromatographic data may be seen
in Figure 2.9. Table 2.2 contains the calculated concentrations of 9- and 13- HODE in the

mouse mammary gland tissue.

40

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

- - - as ......... as ......... a! - RI'T
8 m/z 382 18-HODE errone-
a \ x '
'o A 20% Beef Tallow
g m/z 386 -
" 35o
Scan Number
- - --.- ......... a- ......... a: - 13:"
g m/z 382
5 M 5% Corn on
g m/z 386 A
war oaoo use 55:37— fibo
Scan Number
- - =9 ......... 3.0 ......... =7 .11ng

 

m/z 382 A '
~ 20% Corn Oil

 

 

 

 

 

 

 

 

8

5

'g .

<1 ~ .

' ado “'57: 755a robe
Scan Number

Figure 2.9: SIM chromatograms of mammary gland extracts: m/z 382 represents the
analyte, and m/z 386 represents the oxygen-18 labeled internal standards.

41

Table 2.2: Concentration of 9- and 13-HODE in mammary gland tissue from mice fed
three different diets.

Concentration (g HODE/mg tissue)

 

 

20% Beef Tallow 5% Com Oil 20% Corn Oil

13-HODE 0.018 0.19 1.0
mm 0.09 1.1

0.000 0.33 0.7

0.18 0.39 0.9

0.068 0.37 2.5

9-HODE 0.03 0.25 1.5
--- 0.14 1.4

0.00 0.46 0.8

0.21 _ 0.54 1.2

0.13 0.55 3.1

 

To test for significance a one-way analysis of variance (ANOVA) was performed.
The ANOVA values for 9-HODE are found in Table 2.3, and theANOVA values for 13-
HODE are found in Table 2.4. These data show that at a 95% confidence level there are
significant differences among the levels of both 9- and 13-HODE in mammary gland tissue
from mice fed isocaloric diets differing only in the amount and type of fat.

Analysis of variance tests for the main effects of independent variables and
potential interactions among them; it does not report which of the pairwise comparisons
are statistically different from each other. To find out which groups were different,
Tukey’s test for making paired comparisons was employed. This test allows for 100(1-
a)% confidence interval, where or = 0.05, for all comparisons. The method is based on a
studentized range distribution. The appropriate percentile point is a function of a, x, and

v, where or is the significance level, x is the number of treatments, and v is the number of

42

Table 2.3: Analysis of variance for 9-HODE

ANOVA: Single factor for 9-HODE

 

 

 

 

 

 

 

 

 

 

 

Summary

Groups Count Sum Average Variance

20% beef tallow 4 0.38 0.09 :1: 0.05 0.01

5% corn oil 5 1.9 0.39: 0.08 0.3

20% corn oil 5 7.9 1.5 :1: 0.4 0.8

ANOVA ,

Source of Variation SS df MS F P-value F crit
Between Groups 5.9 2 2.9 10.1 0.003 7.2
Within Groups 3.2 11 0.29

Total 9.1 13 '

Table 2.4: Analysis of variance for 13-HODE.

ANOVA: Single factor for l3-HODE

Summary

Groups Count Sum ‘ Average Variance

20% beef tallow 4 0.265 0.07 :l: 0.04 0.006

5% corn oil- 5 1.4 ‘ 0.27 :1: 0.06 0.02

20% corn oil 5 6.3 1.2 :l: 0.3 0.5

ANOVA

Source of Variation SS df MS F P-value F crit
Between Groups 3.8 2 1.9 10.5 0.003 7.2
Within Groups 2.0 1 1 O. 18

Total 5.8 13

43
degrees of freedom for 52. The method of paired comparisons by Tukey’s test involves

finding a significant difference between means i and j (iztj) if the difference in means
exceeds q[a, x, V]S"% , where n is the number of samples per group and q is a

tabulated value based on or, x, and v (47). The standard deviation (5) is obtained from the
ANOVA Table. The results of Tukey’s test for this diet study are shown in Table 2.5.
Table 2. 5: Tukey’ 5 test results for differences among the levels of 9- and 13-HODE

among the different diet groups.

13-HODE test statistic 9-HODE test statistic

 

Xm com _ sz beeftallow 0.21 q 0.73 0.30 ‘ 0.93
Xzoo/mm _ X20. /. beef Wow 1.19 0.73 1.50 0.93
X 20% com - X 5% cm 0.98 f 0.73. 1.20 0,93

 

The results of Tukey’s test indicate that the levels of 9- and 13-HODE in the mice
fed a high-unsaturated-fat diet (20% corn oil) are significantly different from those levels
in'the group fed the low-unsaturated-fat diet (5% corn oil) and the group fed the high-
saturated-fat diet (20% beef tallow) at a 95% confidence level. The groups fed the low-
unsaturated-fat diet and the high-saturated-fat diet were not significantly different fiom
each other at a 95% confidence level.

A closer look at the numbers show that the levels of 9-HODE in the tissues were
almost equal to the levels of 13-HODE. Enzymatic synthesis of 9- and 13-HODE would
favor the production of more of one or the other isomer, not production of equal amounts

of both. Baer et al. (48) states that “if the 13-HODE had been derived from non-enzymatic

V 44
(auto)oxidation, then the amount of 13-HODE would have equaled that of 9-HODE, and

the 13-HODE would have been racemic”. There was some concern that the results seen
here are artifactual (due to autooxidation during storage in the freezer or during sample
workup) rather than an indication of what had occurred biologically.
VII. Validation of the extraction methodology

’ Because there was concern that the results obtained may not reflect a biological
distribution of the HODEs, it was decided to approach answering the question of
autooxidation with two methods: 1) test the precision of the extraction methodology as a
function of storage time and 2) determine the enantiomeric distribution of the HODEs by
chiral Chromatography.
A. Method precision as a function of storage time .

A large variance from sample to sample (from the same animal) may indicate that
autooxidation is occurring during sample workup. To test this idea the precision of the
extraction methodology was tested using mammary glands from one animal instead of
multiple animals. Concurrently, precision was measured as a filnction of storage time. The
logic here is that if autooxidation is occurring only during sample workup, the variance
should be large in each case, but the levels of 9- and/or 13-HODE should be similar to
each other. If autooxidation is occuning during sample storage, the levels of 9- and/or 13-
HODE should increase, the levels of 9- and l3-HODE should start out being different
then become more equal, and the variance should increase the longer the tissue is stored.

For these experiments a male rat that had been fed a low-fat diet was sacrificed,

and two pairs of mammary glands were excised and immediately placed into liquid

45
nitrogen. The glands were pulverized in liquid nitrogen, thoroughly mixed, and divided

into replicate samples ranging from 100-200 mg. These samples were randomly assigned
to one of four test groups: one day (-80°), one week (-20°), one month (-20°), and one
month (-80°). Each group was. comprised of three Samples. The glands were extracted,
and data obtained are shown in tables 2.6 (9-HODE) and 2.7 (13-HODE).

There was one extra sample from the rat mammary glands used in this study. It
was stored in a -4° freezer for three weeks. When it was extracted and analyzed the level
of 13-HODE was 15.3 ng/mg of tissue. The level of 9-HODE was 14.8 rig/mg tissue.

Three points are evident from this data: 1) the concentration of 9-HODE is initially .
less than the concentration of l3-HODE, 2) the concentrations of 9- and 13-HODE are
elevated in the samples stored for one month at -20°, 3) the variance of the one month
(-20°) group has increased, 4) if samples are stored in an oxidizing environment (-4°) the
concentration of HODEs in the tissues are high and are equal to each other.

The relative standard deviation for the intra-animal tests range from 5-35% for
13-HODE with the highest deviation in the one month (-20°) group. Similar results for 9-
HODE were obtained with RSD ranging from 10-33%. The interanimal variation (fiom
the mouse diet study), for comparison, was greater than 50% for both 9- and 13-HODE in

each of the diet groups.

46

Table 2.6: Concentration of 9-I-IODE in aliquots of a common pool of mammary gland
tissue as a function of storage time and temperature.

Concentration (ng HODE/mg tissue)
One day (-80) One week (-20) One month (-20) One month (-80)

 

 

 

 

9-HODE
0.53 0.92 1.4 0.75
0.26 1.1 0.9 0.71
0.50 0.93 1.8 0.48

Summary

Groups Count Sum Average . Variance

one day (-80°) 3 1.3 0.43 i 0.08 0.02

one week (-20°) 3 2.9 0.98 :h 0.06 0.01

one month (-20°) 3 4.0 1.3 at: 0.3 0.2

one month (-80°) 3 1.9 0.64 :1: 0.08 0.02

 

Table 2.7: Concentration of 13-HODE in aliquots of a common pool of mammary gland
tissue as a filnction of storage time and temperature.

Concentration (ng HODE/mg tissue)
One day (-80) One week (-20) One month (-20) One month (-80)

 

 

 

13-HODE
1.19 1.12 , 1.6 0.99
0.95 1.22 1.4 1.11
1.29 1.24 2.6 0.74

Summary

Groups ' Count Sum Average Variance

one day (-80°) 3 3.43 1.1 :t 0.1 0.03

one week (-20°) 3 3 58 1.19 :l: 0.04 0.004

one month (-20°) 3 5.6 1.9 i 0.4 0.4

one month (-80°) 3 2. 85 0.9 i 0.1 0.03

 

47
The above data suggest that mammary gland tissue is oxidized during storage. The

levels of 9- and 13-HODE increase with increasing storage time, and the variance is
largest for a long storage time at the higher temperatures. The variance is consistent in the
one day, one week, and one month (-80°) groups, suggesting that autoOxidation is not
occurring during sample workup.

B. Chiral analysis of HODE enantiomers by HPLC

The most accurate test for autooxidation is to determine the enantiomeric
distribution of the HODEs in the mammary gland tissue. Each of the HODEs has a center
of chirality at the carbon containing the hydroxyl moiety. Each HODE therefore, has the
possibility of forming two enantiomers, an R form and an S form (see Figure 2.10).
Enzymatic oxidation of linoleic acid produces one enantiomeric form in mammalian-
systems. Autooxidation, however, produces racemic (1:1) mixtures. If autooxidation is
occurring, racemic mixtures of 9- and/or 13-HODE should be detected with chiral
chromatographic analysis of the mammary gland tissues.

Enantiomeric separation by HPLC has taken various approaches over the last 25
years, with the most recent progress focused on the design and manufacture of synthetic
chiral stationary phases. The three-point interactive rule of chiral recognition first
proposed by Dagliesh (see Figure 2.11)'for paper chromatographic separation (49), was
extended to HPLC and verified by Baczuk et al. (50). Pirkle and coworkers began the first
systematic approach to the design of chiral stationary phases for HPLC. They utilized
various optically active 7t-acids and n—bases (see Figure 2.12) and proved their utility (51,

52, 53, 54). Conformational rigidity was found to be important, and resolution is achieved

48

392.65

 

 

508355 maom.

30.152

80.125

\

2 98

IO
\
x

-o E .N edema

49

enantiomers

 

\\\\\\\\\

 

 

 

 

chiral selector

Figure 2.11: Representation of three point interaction necessary for chiral recognition.

 

 

\\\\\ \\\

/ widic sir: lt-basic site . a-basic site
j . acidic site
/ large small
‘ / \ - \ 4
fl // basic site \\ basic site Q
/ / with l
/ . . arge
stenc micron ate

 

Figure 2.12: Model used to interpret the stereoselectivity obtained with charge-transfer
based chiral selectors (51).

51
by a variety of diastereomeric interactions, including n-rt bonding, hydrogen bonding,

dipole-dipole interactions, and steric hinderence. Three sets of combinations of these
interactions are common. Each set of the interactions entails the approach of the analyte
enantiomers to the most accessible face of the immobilized chiral selector. The enantiomer
that forms the most stable complex with the stationary phase is retained on the column the
longest. In the more retained enantiomer, the complementary sites on the stationary phase
are arrayed more favorably for interaction than in the less retained enantiomer (55).

The methodology that has been developed for the separation of HODE methyl
ester enantiomers uses a Pirkle concept chiral stationary phase consisting of a
dinitrobenzoyl phenylglycine (DNBPG) phase coupled either covalently or ionically over
amin0propyl residues on silica gel with a mobile phase composition of hexane:2-propanol
(IPA) in mixtures from 100:2 to 100:0.5 (hexanezIPA). Flow rates vary fiom 2 mL/min. to
0.5 mIJminute. '

Pirkle phase columns have been used for the analysis of polyenoic fatty acids
formed by the lipoxygenases of soybeans, reticulocytes, pea seeds, and tomato hits (56),
to the analysis of oxidative products of linoleic metabolism in human leukocytes (57), to
the analysis of the stereospecificity of the products of fatty acid oxygenases from psoriatic
scales (58, 59, 60), and to the analysis of structure-activity relationships in the potentiation
of mitogenesis by oxygenated metabolites of linoleic acid (61). In each case the
investigators used this approach to determine the enantiomeric composition in order to see

if autooxidation was occurring. The investigators also wanted to determine which one of

52
the enantiomers was in a greater abundance to be able to correlate the enantiomer to

biological activity.

1. Chiral chromatography with a DNBPG HPLC column.

Initial attempts at reproducing the chiral chromatographic analysis of the HODE
MES that had been reported in the literature was unsuccessfirl. A Pirkle column with a
covalently bonded dinitrobenzoyl phenylglycine to an aminopropyl silica stationary phase
was purchased from Supelco, Inc.. Mobile phase mixtures of hexane:isopropanol (IPA)
ranging from 100% hexane to a 100:5 hexanezlPA were attempted along with varying the
flow rate from 0.5 to 2 mUmin.. All samples are chromatographed as themethyl esters.
Resolution of the enantiomers was not obtained. It was suggested to try lengthening the
column to increase the number of theoretical plates. Increasing the number of theoretical
plates by using two columns in tandem resulted in no separation of the 13(R,S)-HODE
and only a slight separation of the 9(R,S)-HODE.

A J.T. Baker ionically bonded DNBPG column was purchased. With one column,
a mobile phase composition of hexanezIPA 100:1, and a flow rate of 0.5 mL/min., an
approximate 50% valley for both the 13(R,S)- and 9(R,S)-HODEs was obtained as seen in
Figure 2.13a. Figure 2.13b shows the addition of l3(S)-HODE to confirm the retention
times of the enantiomers. The R form of both the 13-HODE and the 9-HODE is more
retained on this type of column and elutes later. These results were reproducible until
some water was introduced on to the column destroying its selectivity. HPLC columns
with this type of stationary phase have large lot to lot variability, and a second column

from IT. Baker did not perform under any chromatographic conditions.

coupe-Homer).

 

 

 

 

__lL__l

53

GOflflc‘HOV’O'}

 

L.’

 

Time

Figure 2.13a: HPLC chromatogram of
13(S,R)- and 9-(S,R)-HODE ME, J.T.

Baker column 3. = 234 nm.

 

 

 

 

 

 

44w

 

Time

Figure 2.13b: HPLC chromatogram of
13(S,R)- and 9-(S,R)-HODE ME plus
added S enantiomer of each, J.T.
Baker column I = 234 nm.

S4

The use of two Regis ionically bonded DNBPG, 2S-cm. x 4.6-mm., columns, a
mobile phase composition of hexanezlPA 100: 1, and flow rate of 1 mL/min., gave results
similar to that of the previous J.T. Baker column as shown in Figure 2.14.

2. Chiral analysis of 13- and 9-HODE from tissue samples

Although the storage time experiments discussed above suggest that the tissues are
not being autooxidized during the extraction procedure, chiral analysis of an extract
provides more definitive results. Since an internal standard had been added to the samples
for the storage and precision study, they could not be analyzed by chiral chromatography.
Two mammary gland samples and one skin sample from the same animal as above were
extracted after one day at -80°. One of the mammary gland samples was extracted with
solvents containing butylated hydroxytoluene (an antioxidant). The samples were
extracted, and the methyl esters chromatographed on a Waters uBondapack reverse phase
C13 column 30-cm. x 3.9-mm. at a flow rate of 0.5 mL/min. with a mobile phase
composition of methanolzwater 80:20. Absorbance was measured at 234 nm. Under these
conditions the 9- and 13-HODE MES elute together and were collected for chiral
chromatography.

Two Regis chiral phase columns with a mobile phase of hexaneszA 100:1 and a
flow rate of 1 mIJmin. were used for the chiral analyses. Because the levels of the HODEs
in these samples were so low, the chiral analyses of these samples were carried out by
coinjection of racemic standards of 9- and 13-HODE ME. The 9-HODE standard had an
S:R ratio of 1.02:1; the 13-HODE standard had an S:R ratio of 1:1 based on peak heights.

The data from these experiments are found in Table 2.8.

55

ooawcnowcrgs

 

 

 

 

..,..l L J L...

Time

 

 

Figure 2.14: HPLC chromatogram of 13(S,R)- and 9-(S,R)-HODE ME, Regis columns
I. = 234 nm.

56
Table 2.8: Chiral analysis of tissues for autooxidation during sample workup.

 

Sample 13-HODE S:R 9-HODE S:R
Mammary gland 1.27 1.00
Mammary gland + BHT 1.16 ' 0.94
Skin 1.19 1.10

 

Ifautooxidation was occurring during sample workup, each of the values obtained
for these experiments should have remained 1:1. The S enantiomer of the 13-HODE
increased in each case supporting enzymatic formation rather than autooxidation. Also, the
level of 13-HODE is greater than that of 9-HODE. It is more difficult to draw any
conclusions for 9-HODE as the peaks were not detectable without coinjection of the
7 racemic standards. These chiral results support the quantitative data from GC/MS.

From these experiments it can be concluded that the tissues are not being
autooxidized during sample workup. However, the mammary gland samples will
autooxidize after prolonged storage. This raises questions concerning the biological

significance of the preliminary results.

References

10.

ll.

12.

13.

14.

15.

16.

17.

18.

Dix, T.A. and Mamett, L.J., Journal of Biological Chemistry 260, 1985, 5351-
5357.

Frankel, B.N., Neff, W.E., Rohwedder, W.K., Khambay, B.P.S., Garwood, RF,
and Weedon, B.C.L., Lipids 12, 1977, 901-907.

Plattner, R.D., Gardner, H.W., and Kleiman, R., Journal of the American Oil
Chemists Society 60, 1983, 1298- 1,303.

Weihrauch, J.L., Brewington, CR, and Schwartz, D.P., Lipids 9, 1974, 883-890.

' Knoche, H.W., Lipids 6, 1971, 581-583.

Su, KL. and Schmed, H.H.O., Lipids 9, 1974, 208-213.
Christie, W.W., In Gas Chromatography and Lipids Oilyszr, Scotland, 1989.

Kuksis, A, In Lipid Research Methodology Story, J .A., Ed.; Alan R. Liss; New
York, 1984.

Ryhage, R. and Stenhagen, E., Arla'vfar Kemi 13, 1959, 523-542.
McLafferty, F.W., Analytical Chemistry 31, 1959, 82-87.

Christie, W.W., Brechany, E.Y., Johnson, SB, and Holman, RT., Lipids 21,
1986, 657-662..

Harvey, D.J., Biomedical Mass Spectrometry 9, 1982, 33-3 8.
Jutila, M. and Jalonen, J., Biomedical Mass Spectrometry 17, 1988, 433-436.

Yu, Q.T., Zhang, J.Y., and Huang, Z.H., Biomedical Mass Spectrometry 13, 1986,
21 1-216.

Yu, Q.T., Lui, B.N., Zhang, J .Y., and Huang, Z.H., Lipids 23, 1988, 804-810.
Renkonen, o., Biochimica et Biophysica Acta 125, 1966, 288-309.
Kuksis, A, Canadian Journal of Biochemistry 49, 1971, 1245-1250.

Drozd, 1., Journal of Chromatography 113, 1975, 303-356.

57

19.

20.

21.

22.

, 23.
24.

25.

26.

27.

28.

29.

30.

31.

32.

33.

58

Poole, C.F., In Handbook of Derivatives for Chromatography, K. Blau and GS.
King, Eds, Heyden & Sons, London, 1978.

Poole, CF. and Zlatkis, A, Journal of Chromatographic Science 17, 1979, 115-
123.

Dix, TA. and Mamett, L.J., Journal of Biological Chemistry 260, 1985, 5351-
5357.

Mayer, B. M,oser, R., Leis, H-J., and Gleishpach, H. ,Joumal of Chromato-
graphy 378, 1986, 430-436.

Stan, H-J. and Scheutwinkel-Reich, M., Lipids 15, 1980, 1044-1054.
Woollard, RM. and Mallet, A.I., Journal of. Chromatography 306, 1984, 1-21.

Kleiman, R. and Spencer, G.F., Journal of the American Oil Chemists Society 50,
1973, 3 1-3 8.

lSessa,D..,J Gardner, I-I..,W KleimarLR, and Weisleder, D. ,Lipids 12, 1977, 613-

619.

Christie, W.W.,, In High-performance Liquid Chromatography and Lipids,

. Pergarnon Books, Oxford, 1987.

Yurek, E. M., A Selective Extraction Technique for H ydroxylated Fatty Acids in
Mammary Tissue Based on Analysis by Gas Chromatography-Mass Spectrometry,
A Masters’ Thesis, Michigan State University, East Lansing, MI, 1990.

Funk, CD. and Powell, W.S., Journal of Biological Chemistry 260, 1985, 7418-
7488.

Bligh, E.G., and Dyer, W.J., Canadian Journal of Biochemistry and Physiology
37, 1959, 911-917.

Folch, J., Lees, M., and Stanley, G.H., Journal of Biological Chemistry 226,
1957, 497-509.

Hubbard, W.C., Hough, A.J., Jr., AR. Brash, Watson, J.T., and Oates, J.A.,
Prostaglandins 20, 1980, 431-447.

Lehmann, W.D., Stephan, W., and Ffirstenberger, G., Analytical Biochemistry
204, 1992, 158-170.

34.

35.

36.

37.

38.

39.

40.

41.

42.

43.

44.

45.

46.

47.

48.

49.

59

Guido, D.M., McKenna, R, and Mathews, W.R, Analytical Biochemistry 209,
1993, 123-129.

van Kuijk, F.J.G., Thomas, D.W., Stephens, RJ., and Dratz, E.A, Journal of Free
Radicals in Biology and Medicine 1, 1985, 215-225. '

van Kuijk, F.J.G., Thomas, D.W., Stephens, R.J., and Dratz, E.A, Journal of Free
Radicals in Biology and Medicine 1, 1985, 387-393. '

Bull, AW., Nigro, N.D., and Mamett, L.J., Cancer Research 48, 1988, 1771-
1776.

Funk, M.O., Isaac, R, and Porter, N.A., Lipids 11, 1976,113-117.

Bull, AW., Nigro, N.D., Golembieski, W.A., Crissman, J.D., and Mamett, L.J.,
Cancer Research 44, 1984, 4924-4928.

Green, K. H,amberg, M, Samuelsson, B. ,,Smigel M., and Frolich, J. C., in
Advances in Prostaglandins and Ihromboxane Research, Vol. 5, ed. by J. C.
Frolich, Raven Press, New York, 197 8

. Hubbard, W.C., Phillips, M.A, and Taber, D.F., Prostaglandins 23, 1982, 61.

Pickett, WC. and Murphy, RC., Analytical Biochemistry 111, 1931, 115-121.

Murphy, RC. and Clay, K.L., inMethods of Enzymology Vol. 86, ed. by W.E.M.
Lands and W.L. Smith, Academic Press, New York, 1982.

Westcott, J. Y. C,lay, K. L., and Murphy, R. C., Biomedical Mass Spectrometry 12,
1985, 714-718.

Lehmann, W.D., Stephan, M., and Fflrstenburger, G., Analytical Biochemistry
204, 1992, 158-170.

Watson, J.T., in Introduction to Mass Spectrometry, 2nd edition, Raven Press,
New York, 1985.

Walpole, RE. and Myers, R.H., in Probability and Statistics for Engineers and
Scientists, 4th'ed. Macmillan Publishing Company, New York, 1989.

Baer, A.N., Costello, RB, and Green, F .A, Biochemical and Biophysical
Research Communications 180, 1991, 98-104.

Dagliesh, C., Journal of the Chemical Society 1952, 3940-3942.

50.
51.

52.

53.
54.

55.

56.
57.
58.
59.
60.

61.

6O

Baczuk, RJ., Landram, G.K., Dubois, RJ., and Dehm, H.C., Journal of
Chromatography 60, 1971, 351-361.

Pirkle, W.H., Huyn, M.H., and Bank, B., Journal of Chromatography 316, 1984,
585-604.

Pirkle, W.H. and Sikkenga, D.L., Journal of Organic Chemistry 40, 1975, 3430.

Pirkle, W.H. and Sikkenga, D.L., Journal of Chromatography 123, 1976, 400-
404.

Pirkle, W.H., Hause, D.W., and Finn, J.M., Journal of ChromatOgraphy 192,
1980, 143-158. .

Pirkle, W.H., Pochapsky, T.C., Burke, LA, and Deming, K.C., in Chiral .
Separations, edited by D. Stevenson and ID. Wilson, Plenum Press, New York,
1988.

Kuhn, H Wiesner, R., Lankin, V. Z. Nekrasov, A, Alder, L, and Schewe, T.,
Analytical Biochemistry 160, 1987, 24-34.

Reinaud, 0., Delaforge, M., Boucher, J.L., Rocchiccioli, F., and Mansuy, D.,
Biochemical and Biophysical Research Communications 161, 1989, 883 -89 l.

Baer, AN., Costello, P. B. ,and Green, F. A, Journal of Lipid Research 31,1990,
125-130.

Baer, AN., Costello, PB, and Green, F.A, Journal of Lipid Research 32, 1991,
341-347.

Baer, AN., Costello, RB, and Green, F .A, Biochemical and Biophysical
Research Communications 180, 1991, 98-104.

Glascow, WC. and Eling, T.E., Archives of Biochemistry and Biaphysics 311,
1994, 286-292.

Chapter 3: Development of a New Methodology for Separation of HODE
Enantiomers.

The use of chiral chromatography in the analysis of mammary gland tissue is
important for two reasons: 1) the presence of racemic mixtures of HODEs in mammary
gland tissue is the most reliable indicator of autooxidation, and 2) biological activity may
be determined by the enantiomeric form of the hydroxylated fatty acids. Therefore it is
necessary to determine if the distribution of optical isomers favors the biologically active
enantiomer.

The published methods applied to the analysis of HODE enantiomers have resulted
in unresolved enantiomeric peaks with a 50% valley (1-6). This analysis has also suffered
fi'om poor reproducibility and large lot to lot variation in column performance. Because an
accurate assessment the distribution of optical isomers in a sample requires baseline
separation of enantiomers, a new protocol has been developed for the chiral analysis of
HODE enantiomers. This protocOl is applicable to other monohydroxylated fatty acid
enantiomers.

I. Selection of a different chiral column

The published procedures for the analysis of HODE enantiomers utilize a Pirkle
column with a stationary phase of DNBPG ionically or covalently bonded to
aminopropylsilica (1-6). Stationary phases comprised of cellulose derivatives on silica gel
have also been applied to chiral chromatography. These types of stationary phases
contains an assemblage of subunits acting together to achieve chiral recognition. The

chiral recognition processes have not been determined because of the complex nature of

61

62
the substrate (7). Some cellulose derivatives that have been applied to chiral stationary

phases include nitrates and aromatic esters such as benzoate, cinnamate, and phenyl
carbamate. The chiral discriminatory capabilities of these columns are enhanced by the
introduction of substituents on the phenyl groups of the chiral stationary phase (8, 9).

Capdevila et al. (10, 11) have reported the resolution of dihydroxyeicosanoates and
dihydroxyeicosatrienoates by chiral chromatography on Chiralcel OC (cellulose
tris(phenylcarbamate) on a silica-gel substrate) or OD (cellulose tris(3,5-dimethylphenyl
carbamate) on a silica-gel substrate) columns with hexanezlPA mobile phases. The
dihydroxy fatty acids were chromatographed as either methyl esters or pentafluorobenzyl
esters. Optimal resolution was achieved after extensive equilibration with the mobile
phase, and each compound studied required different chromatographic conditions (mobile
phase composition and flow rate). When the published conditions for the analysis of the
dihydroxy methyl esters on a Chiralcel OD column was applied to the analysis of HODE
methyl esters, no resolution of the HODE enantiomers was obtained as seen in Figure 3.1.
II. Selection of a mobile phase
A. Background

In I-H’LC, the mobile phase is a dynamic part of the system. In normal phase and
reversed phase chromatography, Optimization of the mobile phase is the most important
factor for improving separations (12). Optimization of the separation of racemates by
manipulation of the mobile phase rests on three principles: 1) solvent selectivity, 2) solvent
strength, and 3) adsorption of solvent on specific sites of the chiral stationary phase

followed by displacement of solvent by the solute (13). The selectivity of a nonpolar

63

 

 

 

 

 

 

13(RS)-HODE
I

A

b

S

O

5 9(R,S)-HODE
a

11

C

C

I I I l I I I I I I
30 40

time (min)

Figure 3.1. Chromatogram resulting from chiral analysis of a mixture of 13(R,S)- and
9(R,S)-HODE ME on a Chiralcel OD column, hexanezlPA 100: 1, flow = 1 mL/min.. The
enantiomers are not resolved under these conditions.

64
solvent like hexane can be modified by the addition of solvents that act as proton

acceptors, proton donors, or strong dipoles. The strength of the modified solvent is
affected by the quantity of the added modifier. A typical approach is to use a mobile phase
that is as inert as possible but possesses sufficient solubilizing ability to move the
enantiomers through the stationary phase. As has been discussed a typical solvent system
for the analysis of HODE enantiomers is hexane modified by a small amount of isopropyl
alcohol.

Snyder (14) has proposed that solute retention in liquid-solid chromatography on
polar solvents can be explained by solvent adsorption onto and displacement from specific
sites on the adsorbent by molecules of solute (14). On aminopropyl silica and silica, the
amino and silanol groups are firlly exposed sites for interaction with solvent. When chiral
groups are bonded via a spacer to silica, the amide linkages adjacent to the chiral carbon
are exposed for solvent interaction.

According to Snyder (14), the ease of displacement of solvent on the adsorbent by
solute will largely determine the retention time of the solute. A decrease in elution time
signifies that solvent molecules adsorbed onto active sites of the chiral phase are displaced
with greater difficulty by solute molecules.

Care must be taken when selecting solvent systems for cellulose based stationary
phases as these polymeric chiral phases are soluble in dichloromethane, chloroform, and
tetrahydrofuran. These stationary phases are limited to mobile phases such as hexane, 2-

propanol, methanol, and ethanol.

65
B. Selection of mobile phase

After conversations with Dr. Larry Nicholson, a scientist at The DOW Chemical
Company (Midland, M1), the utility of a pentane:methanol mobile phase with either
Chiralcel OD or OC chiral stationary phase was investigated. Alcohol modified pentane-
based mobile phases were described recently (15, 16). Use of a pentanezmethanol mobile
phase with a Chlialpak AD chiral stationary phase often yielded improved resolution on
enantiospecific separations that could not be obtained with conventional 2-propanol or
ethanol modified hexane mobile phases. The mechanism of this enhancement has not been
determined as yet. The racemic HODE methyl esters were not resolved on a Chiralcel OC
column with this mobile phase, but were baseline resolved on a Chiralcel OD column.
When racemic mixtures of the HODEs are spiked with the pure S enantiomer of each, the
elution order of the HODEs may be characterized (see Figure 3.2). This order is 13(R),
13(S), 9(R), and 9(S) indicating that the S enantiomer is more retained by the column
under these conditions (pentanermethanol 180: 1, flow rate = 1 mL/min.).

A flow rate of 1 mL/min. is used so that the 13(S)-HODE ME does not coelute
with the 9(R)-HODE ME. Total analysis time for this separation is less that 35 minutes.
Flow rates may be increased to shorten the analysis time if the 9- and 13-HODE are
analyzed separately as seen in Figure 3.3. An analysis time of 16 minutes for the 9(R,S)-
HODE ME results with a flow rate of 2 mL/min. with baseline separation of the

enantiomers. For comparison, the data reported in the literature with the DNBPG columns

66

13(S)-HODE

\n

9(S)-HODE

\1

9(R)-HODE

ll/

13(R)-HODE

1/

 

(canard-100:6}

 

 

 

 

 

 

 

 

 

WV

25 30

time (min)

Figure 3.2. Chromatogram resulting from enantiomeric separation of a mixture of
13(R,S)- and 9(R,S)-HODE methyl esters with a Chiralcel OD chiral stationary phase,
pentane2methanol 180: 1, flow = 1 mL/min.. The methyl esters of 13(S)- and 9(S)-HODE
have been added to determine the retention order of the enantiomers.

67
l 3 -HODE

l

 

 

 

00590'"ONO‘>

 

 

 

8 10 12 14'
time (min)

9-HODE

 

ooaua'fioucrg.

 

 

 

8 10 12 I4 16
time (min)
Figure 3.3. Chromatograms resulting from enantiomeric separation of 13(R,S)- and

9(R,S)-HODE methyl esters, Chiralcel OD CSP, pentane1methanol 180:1, flow rate = 2
mL/min..

 

 

68
and hexanezlPA mobile phases have analysis times of 40 to 60 minutes without baseline

separation of the racemates (1, 5).
C. Alteration of the mobile phase

Because pentane is a volatile solvent, its use as a mobile phase is plagued with
difficulties. The number one difficulty is the introduction of bubbles (even afier degassing
the solvents) into the HPLC system. In an attempt to circumvent this problem, the utility
of a mobile phase consisting of hexane:methanol was investigated. The chromatograms of
13(R,S)- and 9(RS)-HODE MES are shown in Figure 3.4 with hexanezmethanol 180:1 as
a mobile phase and a flow rate of 1 mL/min.. Baseline separation of the racemates with the
Chiralcel OD column is obtained. The analysis time is approximately thirty minutes, and
the 13(S)- is resolved from the 9(R)-HODE ME.

III. Application of the HPLC conditions to enantiomer separation of other
monohydroxylated polyunsaturated fatty acids

To determine the universality of this methodology to the general analysis of
monohydroxylated polyunsaturated fatty acid enantiomers, several hydroxyeicosadienoic
acid (HEDE) methyl esters; 8-HEDE, 11-HEDE, 15-HEDE (Figure 3.5), and several
hydroxyeicosatetraenoic acid (HETE) methyl esters; 5-HETE, 8-HETE, 12-I-IETE and
15-HETE (Figure 3.6), were successfully analyzed under these chromatographic
conditions (Chiralcel OD chiral stationary phase, hexanezmethanol 180:1, flow rate = 1
mL/min.). A chromatogram of the photooxidation products of arachidonic acid (Kuhn,
1987) is shown in Figure 3.7 as an example of published results utilizing a Pirkle DNBPG

column, hexanezlPA 100:0.5, flow = 0.5 mL/min..

69

HODE ME mixture

«0987-10-6)

 

 

 

 

 

v I v v 1' v v 1

20 Iimc'unin.) 30

l3-HODE ME

oosocnoac>

 

 

 

 

 

 

MUD

2'0 time‘(m‘in.)' 2'5
9-HODE M E

V

 

(scale-30106)

 

 

 

 

 

 

T V fl V ' V V V V V 1 V V Y 1 V 1 V V

Figure 3.4: 13(R,S)- and 9(R,S)-HODE ME, hexane:methanol 180: 1, flow = 1 mL/min.

70

Adi—DE _ 1 Bow 4 5E 328865085 come $52: mama? commando" 0108258”... Sou mew—82 gaoeanEo ”on edema

3.5 9.5

NM am an on am en
.._.T...._... _____-____ _______________

313...... a...

 

 

 

 

 

 

 

(amounacoo

 

 

 

 

 

 

 

 

' " it - - - " i- ‘1 v- -' t '1 iv -'

mam—Harm mag; _ mgrfl

71

68828 A 1 Bow .8 “SM 3588888886 8080 $508 amigo cough—om 0883580 82.“ wen—3.2 «88388980 .05 Emmi
Eaves: .

mm em VN om om . 3 on 3
1.1:: -_:__:___: :_:E:::: :__:___:_:___

J? 1437.11in 31.3

 

 

 

 

 

 

 

 

 

 

 

 

 

 

<nmor...oeur:oo

 

 

 

 

 

 

 

 

 

 

 

 

E E.» mhmmé— - mkmmrn M

72

 

 

 

 

 

 

 

 

VI ' w
m
«3 5 ‘3
B
m
u: w m UJ
a o— 0- :5 m '—
U m w I w
m F ¥ I I . f
2 $3 1‘ :2 0', 00 in
<

t—o———q——— -—9—-v-—o—o— N—O—O—o ¢—4—0—¢- “—0—.—
39 37 Ll 49 53

retention time (min)

Figure 3.7: Published results for the analysis of HETE ME enantiomers with a hexanezIPA
mobile phase and a Pirkle DNBPG chiral stationary phase (1).

73
This HPLC methodology for the analysis of the enantiomeric composition of

monohydroxylated fatty acids is an improvement over published procedures both in
enantiomeric resolution and in analysis time. Baseline separation of racemates result with
this procedure, providing confidence that the enantiomeric composition of 9- and 13-

HODE in mammary gland extracts will be reported accurately.

References

1.

10.

11.

12.

13.

14.

15.

Kuhn, H., Wiesner, R., Lankin, V.Z., Nekrasov, A., Alder, L., and Schewe, T.,
Analytical Biochemistry 160, 1987, 24-34.

Reinaud, 0., Delaforge, M., Boucher, J.L., Rocchiccioli, F., and Mansuy, D.,
Biochemical and Biophysical Research Communications 161, 1989, 883-891.

Baer, A.N., Costello, P.B., and Green, F.A., Journal of Lipid Research 31, 1990,
125-130.

Baer, AN., Costello, P.B., and Green, F.A., Journal of Lipid Research 32, 1991,
341-347.

Baer, AN., Costello, P.B., and Green, F.A., Biochemical and Biophysical
Research Communications 180, 1991, 98-104.

Glascow, WC. and Eling, T.E., Archives of Biochemistry and Biophysics 311,
1994, 286-292.

Finn, J.M., in Chromatographic Chiral Separations edited by M. Zief, and LI.
Crane, Marcel Dekker, Inc., New York, 1988.

Okamoto, Y., Kawashima, M., and Hatada, K., Journal of Chromatography 363,
1986, 173-179.

Ichida, A. and Shibata, T., in Chromatographic Chiral Separations, edited by M.
Zief, and LJ. Crane, Marcel Dekker, New York, 1988.

Capdevila, J.H., Wei, S. Kumar, A., Kobayashi, J ., Snapper, J.R., Zeldin, D.C.,
Bhatt, R.K., and Falck, J.R., Analytical Biochemistry 207, 1992, 236-240.

Karara, A., Wei, 8., Spady, D., Swift, L, Capdevila, J .H., and Falck, J .R.,
Biochemical and Biophysical Research Communicatons 182, 1992, 1320-1325.

Snyder, LR. and Kirkland, 1.1., in Introduction to Modern Liquid
Chromatography, 2nd. ed., John Wiley & Sons, Inc. New York, 1979.

Zief, M., in Chromatographic Chiral Separations, edited by M. Zief, and L].
Crane, Marcel Dekker, Inc., New York, 1988.

Snyder, L.R., Journal of Chromatography 255, 1983, 3-26.

Nicholson, L.W., PfeifTer, C.D., Goralski, C.T., Singaram, B., and Fisher, G.B.,
Journal of Chromatography A 687, 1994, 241-248.

74

75

16. Nicholson, L..W, Pfeifier, C..,D Goralski, C..,T and Singaram, B. ,Journal of
Chromatography A in press.

Chapter 4: Application of the Extraction Methodology to the Analysis of Mammary
Gland Tissue from Mice Fed Variously Controlled Fat Diets

The goals of this research are twofold: 1) develop an improved analytical
methodology to isolate, identify, and quantitate 9- and 13-HODE in mammary gland
tissue, and 2) to be able to apply this methodology to help to answer biological questions;
specifically to determine if the levels of 9- and/or l3-HODE and their enantiomers are
higher in the glands from mice fed a diet that is high in linoleic acid. This chapter describes
the application of the analytical methodology to the analysis of mammary gland tissue
from mice fed diets differing only in the amount and/or type of fat. The tissues were
extracted and analyzed by GC/MS to quantitate the concentration of HODEs in the tissue.
In addition the enantiomeric distribution of the (R,S)-HODEs was determined by peak
area ratios.
L Experimental

Female Balb/c mice were divided into three groups at three weeks of age. Each
group was placed on one of three chemically purified isocaloric diets differing only in the
amount and type of fat in the diet. These mice were fed either a high-saturated-fat diet
(20% beef tallow by weight), a low-unsaturated-fat diet (5% corn oil by weight), or a
high-unsaturated-fat diet (20% corn oil by weight) for five weeks. At the end of the five
week period, one mouse from each of the diet groups was sacrificed on the day the tissue
was to be extracted to eliminate the storage time variable. Three pairs of mammary glands
were excised from the mouse and immediately frozen in liquid nitrogen. All of the glands

from each mouse were pulverized in liquid nitrogen and divided into two samples; one for

76

77

GC/MS quantitation which was extracted in the morning, the other for chiral analysis
which was extracted in the afiemoon. Tissues that were not being extracted were stored at
-80°.

11. Results

The results for this diet study are listed in Tables 4.1 and 4.2 for 13-HODE and
Tables 4.3 and 4.4 for 9-HODE. The concentration of the HODEs in the mammary gland
tissue are reported in ng HODE/mg. tissue. The chiral data are represented as a peak area
ratio of S/R for 13-HODE, but as the reciprocal R/S for 9-HODE because the R
enantiomer was in a greater abundance for most of the samples.

There was enough tissue from one sample from the high-unsaturated-fat diet to
extract it in triplicate. The relative standard deviation for 13-HODE was 13%; 9-HODE
was 24%.

III. Data analysis and discussion
A. Concentration of HODEs

An analysis of variance for the 13-HODE data from the second diet study shows
significance at a 95% confidence level (Furc = 6.4, Fab = 3.5). An ANOVA for the 9-
HODE data also indicates significance at a 95% confidence level (Fcalc = 8.7, Fm, = 3.5).
Tukey’s test on the data from both the 9- and 13-HODE finds no differences between the
HODE levels in the glands from mice fed the 20% beef tallow diet and those fed the 5%
corn oil diet. However, there are significant differences between levels of 9- and 13-
HODE in the 20% corn oil diet and the 20% beeftallow diet and between the 20% corn

oil diet and the 5% corn oil diet at a 95% confidence level.

78

Table 4.1: Concentration and enantiomeric ratios for l3-HODE.

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

13-HODE
20% BeefTallow diet I] 5% Corn on diet I] 20% Corn on diet
ngmg tissue area S/area R ngmg tissue area S/area R ngmg tissue area S/area R
0.73 2.23 1.5 4.63 1.6 6.13
0.98 1.46 0.53 2.31 0.54 1.83
1.6 3.32 0.64 4.57 5.9 11.18
0.52 1.15 1.4 3.88 5.4 4.39
0.68 1.15 0.53 3.19 1.6 4.48
1.4 0.64 0.75 0.95 0.54
0.66 1.40 0.45 1.79 2.5 1.54
0.35 1.16 0.79 4.28 3.3 5.25
2.15 2.01 1.93
1.01 13.22 1.32
Table 4.2: Summary of concentration data for 13-HODE.
Summary
Groups Count Sum Average
20% Beef Tallow Diet 8 6.9 0.9 $0.1
5% Corn Oil Diet 8 6.5 0.8 :t 0.1
20% Corn Oil Diet 8 21.6 2.7 :1: 0.7

 

Table 4.3: Concentration and enantiomeric ratios for 9-HODE.

 

 

 

 

 

 

 

 

 

 

 

 

 

9-HODE
20 % Beef Tallow diet 5% Corn Oil diet 20% Corn Oil diet
ng/mg tissue area S/area R ng/mg tissue area S/area R ng/mg tissue area S/area R
0.42 0.42 1.2 0.26 2.1 0.65
0.91 0.61 0.29 0.36 0.42 1.10
1.2 0.53 0.36 2.86 5.6 0.76
0.085 5.56 0.67 10 4.4 0.54
0.46 2.44 0.38 2 1.6 0.84
0.56 0.25 0.81 0.77 0.88
0.73 0.89 0.14 1.2 2.3 1.49
0.32 0.85 0.48 0.80 2.2 0.51
0.36 0.73 0.74
0.85 - 0.28 0.76
Table 4.4: Summary of concentration data for 9-HODE.
Summary
Groups Count Sum Average
20% Beef Tallow Diet 8 4.7 0.6 i 0.1
5% Corn Oil Diet 8 3.7 0.5 :1: 0.1
20% Corn Oil Diet 8 19.4 2.4 i 0.6

 

80

This study has shown that there are significant differences in the levels of 9- and
13-HODE in mammary gland tissue as a function of the amount and the type of fat in the
diet. These numbers may now show biological significance as well as mathematical
significance.

B. Chiral analysis

A representative chromatogram from the chiral analysis of the mammary gland
tissue extracts may be found in Figure 4.1. The small peaks that are unlabeled have not
been identified. The identity of the 9- and l3-HODE enantiomers has been confirmed by
coinjection of racemic standards of 9- and 13-HODE. Average values for the enantiomer

ratios for 13- and 9-HODE are listed in Table 4.5.

Table 4.5: Summary of chiral data for 9- and 13-HODE.

20% Beef Tallow diet 5% Corn Oil diet 20% Corn Oil diet

 

 

 

13-HODE (SIR)

average 1.6 :t 0.2 3 i 0.5 3 i 1
9-HODE (R/S)

average 0.9 i 0.2 1.0 i 0.3 0.82 :t 0.09

 

In calculating the averages for the 13-HODE (S/R) ratios for the 5% corn oil diet
and for the 20% corn oil diet and averages for 9-HODE (S/R) for 20% beef tallow diet
and the 5% corn oil diet, the largest value in each data set was removed by a Q-test so the

above numbers are based on n-1 data points.

81

13(S)-HODE

\l V...

J

 

13(R)-HODE 9(S)-HODE

/ /

anam6n0m6>

 

 

 

 

 

I I I I I 3" I I I I I I I‘M—

20 30

 

time (min)

Figure 4.1. Chromatogram resulting from enantiomeric separation of 13(R,S)- and 9-
(R,S)-HODE methyl esters from a mammary gland extract.

82

The data from the chiral analysis of the mammary gland extracts indicates that non-
racemic mixtures of the HODEs exist in these tissue samples. The 13(S)-HODE and the
9(R)-HODE are the more abundant enantiomers in these tissues.

Most of the studies reported in the literature have focused on in vitro studies of
enzymatic oxidation of linoleic acid by various lipoxygenase enzymes. Kuhn (1987)
investigated the positional and optical isomers formed by various lipoxygenases from
linoleic acid. This investigation found that different lipoxygenases formed different ratios
of 13:9 HODE as well as different S:R ratio. The results of this study are shown in
Table 4.6.

Table 4.6: Composition of the positional and optical isomers formed by various
lipoxygenases from linoleic acid (1).

 

Lipoxygenase 13/9 ratio (%) 13-HODE (SIR) 9-HODE (S:R)
Soybeans I 98:2 97:1 1:1
Reticulocytes 96:4 47: 1 1 :1
Wheat 15:85 2:1 41.521
Pea SeedsI 39:61 1.5:1 1.1:1

Pea Seeds II 89:11 44:1 1.2:1
Tomato fruit 15:85 6.521 84:1
Quasi-LOX of 50:50 1:1 1:1
hemoglobin

 

Reinaud et al. (2) reported that the major product of the biotransformation of
linoleic acid by human leukocytes is 13(S)-HODE. Glasgow and Eling reported that the
oxygenation of Syrian hamster embryo fibroblasts produces the pure (S) enantiomer of 13-
HODE with a minor product of 9-HODE (enantiomer not reported) in a 3:1 ratio (3).
They also demonstrated that it is the l3(S)-HODE that was active in stimulating DNA

synthesis.

83

Baer et al. has reported an in vivo experiment. In this study it was discovered that
the principal in viva oxygenase products of linoleic acid in psoriatic skin scales are 13-
HODE (SIR = 1.9) and 9-1-IODE (R/S = 2.4) (4). Further in vitro investigations with cell
cultures and radiolabeled linoleic acid produced an enantiopure l3(S)-["C]-
hydroxyoctadecadienoic acid. The difference between the in vivo and in vitro results was
explained as follows (5): the l3-HODE produced in vivo was not enantiopure, but was a
mixture of stereoisomers with an S/R ratio averaging 1.9. The discrepancy between the
pure (S) stereospecificity of the in vitro-produced 1 13-HODE and nonracemic
stereospecificity of the in vivo compound suggests that there is autooxidation of linoleic
acid in the epidermis resulting in a nonenzymatically derived racemic 13-HODE and an
enzymatically produced 13(S)-HODE.

Baer’s research on in vivo and in vitro production of HODEs in psoriatic skin may
help to explain the results seen in the in vivo study with mouse mammary gland tissue. The
data obtained with the mouse mammary glands follow the same trend as that for psoriatic
skin. 13(S)-HODE was produced in a greater abundance than was the l3(R)-HODE, and
the 9(R)-HODE is produced in a greater abundance than is the 9(S)-HODE in both the
mouse mammary gland tissue and the human psoriatic skin tissue. The mixtures of
enantiomers are non-racemic in the mammary gland tissue, but are not stereospecific for
13(S)-HODE suggesting that there may also be autooxidation in the cells of the mouse
mammary gland as was seen in the psoriatic skin.

A study by Kuhn (1990) on the oxygenation of biological membranes by

reticulocyte lipoxygenase provided some interesting results (6). At low substrate

84

concentrations, reticulocyte lipoxygenase enzymes led to pure 13(S)-HODE formation.
However, a high substrate concentration resulted in loss of the product enantiomeric
specificity. A similar decrease in product enantiomeric specificity has been reported for the
oxygenation of high concentrations of linoleic acid by the soybean lipoxygenase (7).
Dissociation of radical intermediates from the enzyme and then subsequent non-enzymatic
reactions with oxygen have been proposed as a mechanistic explanation of this lack of
specificity (7, 8, 9).
IV. Summary

This diet study using the validated methodology (including chiral chromatography)
has shown that mice fed a diet that is high in unsaturated fat show significantly higher
levels of 9- and 13-HODE in their mammary glands (Data are summarized in Table 4.7.)
Chiral chromatographic data indicate that 13(S)- and 9(R)-HODE are the more abundant
enantiomers in the tissues.

Table 4.7: Summary of the quantitative data from the diet study. bAll data reported as
mean :1: standard error of the mean. P < 0.05 for c vs. d, c vs. e, and (1 vs. e.

 

 

 

Diets Number of 9-HODE 1.3-HODE
Samples (ng/mg tissue)b (ng/mg tissue)”

20% beef tallow
(very low linoleic acid) 8 0.6 :t 0. 1° 0.9 :1: 02°
5% corn oil
(low linoleic acid) 8 0.5 d: 0.16 0.8 i 0.2d
20% corn oil
(high linoleic acid) 8 2.4 a: 06° 2.7 a: 07°

 

 

 

 

The results obtained here indicate enzymatic production of the HODEs on top of

autooxidation products as evidenced by the equivalent amounts of 9- and 13-HODE and

85

the non-racemic mixtures of the HODE enantiomers. This autooxidation most likely
occurs within cells primary prior to extraction.
V. Conclusion

An analytical methodology has been developed for the analysis of 9- and 13-
HODE in rodent mammary gland tissue. This methodology utilizes gas chromatography
mass spectrometry to quantitate the absolute amounts of these metabolites in the tissues.
Lipoxygenase oxidation of linoleic acid produces 9-hydroxyoctadecadienoic acid and 13-
hydroxyoctadecadienoic acid; these metabolites may also be formed by free radical
oxidation of linoleic acid. Analysis of the mammary gland tissue extracts by enantiomeric
separation with HPLC on a chiral stationary phase provides a qualitative assessment of the
degree of enzymatic vs. non-enzymatic production of the HODEs.

This methodology has been applied successfully to analyses of mouse mammary
gland‘tissues to assess the influence of diets consisting of different amounts and/or types
of fat on the production of HODEs. Significant differences were seen among the levels of
the HODEs in mammary gland tissue fi'om mice in different diet groups. The enantiomeric
distribution of 9(R,S)- and 13(KS)-HODE in mammary gland tissue as a function of diet
indicates that there is a combination of factors contributing to the significant difference

seen.

References
1. Kuhn, H., Wiesner, R., Lankin, V.Z., Nekrasov, A., Alder, L., and Schewe, T.,
Analytical Biochemistry 160, 1987, 24-34.

2. Reinaud, O., Delaforge, M., Boucher, J.L., Rocchiccioli, F ., and Mansuy, D.,
Biochemical and Biophysical Research Communications 161, 1989, 883-891.

3. Glascow, WC. and Eling, T.E., Archives of Biochemistry and Biophysics 311,
1994, 286-292.

4. Baer, AN., Costello, P.B., and Green, F.A., Journal of Lipid Research 31, 1990,
125-130.

5. Baer, AN., Costello, P.B., and Green, F.A., Journal of Lipid Research 32, 1991,
341-347.

6. Kuhn, H., Belkner J., Wiesner, R, and Brash, A.R., The Journal of Biological
Chemistry 265, 1990, 18351-18361.

7. Kemal, C., Yuan, LC, and Lorenzo, K., in Prostaglandins and Lipid Metabolism
in Radiation Injury, edited by Walden, TL. and Haywood, N.H., Plenum -
Publishing Co., New York, 1987.

8. van Os, C.P.A., Rijke-schilder, G.P.M., and Bliegenthart, J.F.G., Biochimica et
Biophysica Acta 575, 1979, 479-484. . .

9. Kuhn, H., Schewe, T., and Rapoport, S.M., Advances in Enzymology 58, 1986,
273-31 1. '

86

MICHIGAN sran UNIV. LIBRQRIES
llllllllllllllll“HilllllllllllWilli”lllllllllll
31293013995570