![[Feedback]](/icons/banner/feedbackACT.gif){kind=icon}
![[For Subscribers]](/icons/banner/subscriberACT.gif){kind=icon}
![[Archive]](/icons/banner/archiveACT.gif){kind=icon}
![[Search]](/icons/banner/searchACT.gif){kind=icon}
![[Contents]](/icons/banner/tocACT.gif){kind=icon}
|
|
|
|
|
||
,
3
*
ESA, Inc., Chelmsford, MA 01824;
Dementia Research Service, Burke Medical Research Institute, White Plains, NY 10605;
**
Antioxidants Research Laboratory, Jean Mayer U.S. Department of Agriculture Human Nutrition Research Center on Aging at Tufts University, Boston, MA 02111 and
Departments of Biochemistry and Neuroscience, Cornell University Medical College, New York, NY 10021
3To whom correspondence should be addressed. E-mail: bkristal{at}burke.org
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
250
compounds in serum are sufficiently reliable, both analytically and
biologically, for potential use in building mathematical models of
serotype.
KEY WORDS: • dietary restriction • blood • high performance liquid chromatography • diet • coulometric array
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
, Kristal et al. 1998
). The
resultant ability to simultaneously screen a large set of metabolites
increases the chances of identifying markers that are altered by
caloric intake. Furthermore, such HPLC systems can reliably quantitate
redox-active, low-molecular-weight metabolites present in
concentrations as low as 10 fmol/125 µl or 80 pM (Milbury 1997
, Acworth et al. 1997
). This sensitivity
further increases the probability of identifying compounds of interest.
A disadvantage of building a serotype based on small molecules is that
the metabolites that comprise the serotype are initially identified as
peaks in a HPLC profile, rather than by complete biochemical identity.
This problem is solved at the end by isolating and biochemically
identifying (e.g., by mass spectroscopy) molecules determined to be of
interest. As with older more traditional methods of data collection, high throughput approaches to characterizing serotype shifts related to diet require attention to a series of issues related to analytical reliability and biological variability. Issues of analytical reliability and biological variability are often straightforward and adequately addressed by repetition and/or by the inclusion of control samples or control comparisons. High throughput studies, however, may generate data that overwhelm traditional statistical approaches and/or prevent analyte by analyte method validation. Studies in which data collection is expected to continue over years also must address drift that occurs in methods or in reagents. Likewise, issues related to signal strength and biologic variability also arise in studies of human or animal subjects. For example, dietary manipulations that do not generate signals that are relatively large in comparison to background noise generated by shifts in other dietary constituents or changes in the organisms (e.g., lab animals, human populations) over time will be technically difficult or impossible to study.
To help isolate the problem of analytical variability, we have begun
our investigation by focusing on a robust model of diet-related
physiological effects [DR (Kristal and Yu 1994
,
Weindruch and Walford 1988
)] and by using an analytical
method shown to be stable over years of constant analysis (HPLC
separations coupled with coulometric array detectors, Matson et al.,
unpublished data).
DR was chosen for our initial studies because it is the most potent and
reproducible known means of extending life span and reducing morbidity
in higher animals (Kristal and Yu 1994
, Weindruch and Walford 1988
). As such, dietary restriction stands as an
extreme example of the influence of diet on health. Animals maintained
on DR regimens are fed a fraction (usually 50–70%) of the amount
eaten by their ad libitum (AL) fed littermates, who have unlimited
access to food. DR has been extensively explored for over 70 y
because of its ability to extend both mean and maximum life span,
reduce age-related morbidity, and delay or prevent most
age-associated physiological dysfunction (Kristal and Yu 1994
, Weindruch and Walford 1988
). DR also
alters many basic physiological processes, including metabolism,
hormonal balance and the generation of, detoxification of and
resistance to reactive oxygen species (Yu 1996
). DR can
be implemented in multiple ways, ranging from harsh methods such as
McCay’s original protocols where animals were fed barely enough to
allow growth in a stair-step manner, to every-other-day feeding, to
controlled paradigms that decrease food intake from as little as 20%
to as much as 55% and more (McCay 1935
, Maeda et al. 1985
, Carlson and Hoelzel 1946
,
Goodrick et al. 1990
, Cheney et al. 1980
,
Ross and Bras 1970
, Weindruch and Walford 2000
, Nolen 1972
, Tannenbaum 1945
, McCay et al. 1935
). Restricting total
calories is more important than reducing intake of specific
constituents/nutrients (e.g., fat, proteins, vitamins and minerals,
etc.) (Iwasaki et al. 1988a
, Iwasaki et al. 1988b
). DR extends longevity in essentially all animals in
which it has been tested, including multiple mammalian species (rat,
mouse, guinea pig) (Kristal and Yu 1994
, McCay 1935
, Maeda et al. 1985
, Carlson and Hoelzel 1946
, Goodrick et al. 1990
,
Cheney et al. 1980
, Ross and Bras 1970
,
Weindruch and Walford 2000
, Nolen 1972
,
Tannenbaum 1945
, Stucklikova et al. 1975
). Together, these observations reveal that the
physiological effect of DR in mammals is large and suggest the
feasibility of identifying a metabolic serotype characteristic of this
nutritional intervention. Successful identification of a metabolic
serotype associated with DR should help define mechanisms for
evaluating serotype-based markers of nutrition and may be
intrinsically scientifically interesting.
The use of HPLC-based analytical systems follows from the general
ability of these systems to reproducibly separate multiple different
types of low-molecular-weight species. The use of gradient
chromatography extends the capacity of a given system to separate
analytes of interest based on a physical property such as solubility or
binding capacity in a given solvent. Coulometric detectors allow
detection of essentially 100% of analytes of interest having a given
oxidation potential (Kristal et al. 1998
, Matson et al. 1984
, Milbury 1997
, Acworth and Gamache 1996
). Arrays of coulometric detectors increase the
ability of chromatographic systems to separate molecules of interest by
extending the dimension of time to the dimension of oxidation potential
(Kristal et al. 1998
, Matson et al. 1984
,
Milbury 1997
, Acworth and Gamache 1996
).
Furthermore, the use of coulometric arrays enables qualitative
identification of a compound (and determination of purity) based on
sequential partial reactivity (Kristal et al. 1998
,
Beal et al. 1990
, Matson et al. 1987
,
Matson et al. 1990
, LeWitt et al. 1992
,
Ogawa et al. 1992
, Beal et al. 1992
,
Matson et al. 1984
, Milbury 1997
,
Acworth and Gamache 1996
). Together, these properties
give these modern HPLC systems the analytical power and sensitivity
noted above.
In this report we present data indicating the analytical validity of the approach taken.
|
METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
Male and female Fischer 344 x Brown Norway
F1 rats were obtained monthly from the National
Institute on Aging colony at Harlan (Indianapolis, IN). Cohorts consist
of 5–20 rats each. Rats were fed National Institutes of Health
(NIH)-31 (AL rats only) or vitamin-mineral–fortified NIH-31
(DR rats only [see below for diet ingredients and Table 1
for diet compositions] Harlan, Indianapolis, IN). All animals were
individually housed and DR feeding regimens were implemented at 6 wk of
age. Food intake studies were carried out on all animals within 1 wk of
sacrifice. Rats were killed by decapitation to avoid known differential
effects of anesthesia on parameters of interest. Characteristics of the
rats used are reported in the Results section. All animal experiments
were performed under institutionally approved protocols and complied
with the Guide for the Care and Use of Laboratory Animals.
|
. HPLC methodology
HPLC separations and coulometric array detection were conducted
essentially as described previously using an ESA CoulArray system
[ESA, Chelmsford, MA (Kristal et al. 1998
,
Matson et al. 1984,
Milbury 1997
)].
Analytes were separated on two serial TosoHaas TSK-Gel
ODS-80TM, 4.6 ID x 250-mm columns (Montgomeryville, PA). The 16
channels in the coulometric array were set at 0–900 mV evenly
incremented at 60 mV. To compensate for chromatographic drift that
occurs over time due to environmental changes and/or column changes,
chromatograms were mathematically normalized using a propriety (ESA)
algorithm that scales the chromatograms based on user-identified
peaks. We identified 24 conserved, relatively evenly spaced peaks that
could be used as points of reference (benchmarks) for this scaling
procedure. The results of this normalization procedure are shown in
Figure 1
, which shows an example of a standard pool (A), a
nonnormalized, sample chromatogram (B) and the same sample
chromatogram after normalization using this procedure (C).
|
Statistical analysis. Data were analyzed using the programs Statview 5.0.1 (SAS Institute, Cary, NC), Sigma Plot 5.0 (SPSS, Inc Chicago, IL), NCSS 2000 (Number Cruncher Statistical Systems, Kaysville, UT), and Pirouette 2.7 (Infometrix, Woodinville, WA). Unless noted, data were analyzed by Student’s t test or by analysis of variance (ANOVA) followed by the statistically conservative Tukey/Kramer posthoc test (P < 0.05 and P < 0.01); values significant at P < 0.01 were further evaluated by Fisher’s protected least significant difference test.
Data points in univariate comparisons were predetermined to be
considered outliers only if they failed the Dixon test (Dixon 1955
) or if they were more than four standard deviations from
the mean. No data points failed either test. Sample data for
multivariate comparisons were similarly examined for outliers using
Pirouette as described in detail in the Results section.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
Animal characteristics
This report focuses predominantly on the use of two cohorts of female Fischer 344 x Brown Norway F1 rats to demonstrate technical feasibility of the approach used to determine serotypes (Kristal et al., unpublished results). One cohort composed of eight AL and eight DR rats was used as the primary dataset, and a second, essentially identical, cohort was used to confirm results and/or to determine the relative robustness of the data (as described below). The basic animal husbandry for the animals used followed National Institute on Aging/NIH guidelines as implemented by Harlan. DR (40% less food than eaten by AL rats) was initiated at 6 wk of age.
Rats were imported from Harlan at 5 mo of age and were killed
1 mo later. In the cohort used as the primary dataset, female animal
weights at entry were 205 ± 10 g (AL) or 165 ± 5 g (DR). Harlan’s protocol called for the DR rats to receive 8 g
of diet/d, but we determined empirically that DR rats fed over 7.5 g left food and rapidly gained weight. These data suggested the
possibility that they were being (relatively) overfed. In addition,
food intake studies suggested that AL rats were eating 11.8 g of
diet/d (Fig. 2A
), suggesting that a 40% restriction should give DR rats
7.1 g of diet/d. Rather than go too far below the 8 g/d initially
recommended, however, we chose a compromise value of 7.5 g/d (36%
restriction vs. 40%). Food intake was relatively unchanged in the
second cohort (11.8 vs. 11.1 g/d, not significant by Tukey/Kramer), but
dropped subsequently (Fig. 2A
). Therefore, food intake in
the DR rats was dropped to 6.5 g as well in these subsequent
cohorts (not reported here). Despite the relative change in intake,
weight remained constant in both the two cohorts reported here and in
the subsequent groups. DR animals were consistently 25% lighter
than AL controls (Fig. 2B
). Intake and weight had no obvious
relationship in the AL group (by regression analysis in SigmaPlot).
|
indicate that the cohorts presented are
relatively representative of the larger population being studied.
However, our AL animals do eat less than do those maintained at Harlan
or those maintained at the National Center for Toxicological Research
(Little Rock, AR, 12.8 g/d) (Turturro et al. 1999
). Body
weights are, however, essentially identical to those reported at the
National Center for Toxicological Research (Turturro et al. 1999
). Analytical sensitivity
The first analytical criterion that must be met is one of sensitivity. That is, can the HPLC-based approach being used identify any differences between the serums of AL and DR rats? In this context, the term sensitivity refers to a combination of analytical sensitivity (i.e., limits of detection, signal vs. noise) and biological sensitivity (i.e., are the analytical limits of detection sufficient to detect interanimal differences).
Comparison of the complete chromatograms of serums samples from
an AL and a DR rat (Fig. 3
) reveals two important points. First, visual inspection reveals that
nearly all major peaks (metabolites) are present and essentially
identical between the two samples. Nonetheless, closer inspection
reveals that some differences between these two chromatograms are
apparent (Fig. 3
, arrows). In Figure 4
, the region surrounding the peak of interest at 63.25 min (arrow in DR
portion of Fig. 3
) is expanded (threefold vertically, 33-fold
horizontally) to show detail. At this magnification, the difference is
most clearly seen on channel 8, because channel 10, the dominant
channel, is off-scale. This magnified view makes it clear that the
identity of this peak can be confirmed by matching dominant channel and
ratio of reactivity (Milbury 1997
, Acworth and Gamache 1996
, Matson et al. 1987
, Matson et al. 1990
, Kristal et al. 1998
). In addition,
the peak is surrounded by three peaks that appear elevated in AL rats
(marked with arrows in the top portion of Fig. 4
). Note that instrument
sensitivity is 600-fold greater than the scale shown in Figure 4
(i.e., peaks present at 0.5% of the total scale can be readily
qualitatively identified and quantitatively analyzed). Thus, the
CoulArray system clearly has sufficient sensitivity to approach the
problem of defining serotype.
|
|
The second analytical criterion that must be met is one of
analytical reproducibility. That is, can the HPLC-based approach
being used consistently identify peaks of interest in the serums of AL
and DR rats? This issue was addressed by analyzing 8 aliquots of a set
of pooled samples spaced out over 1 wk. One sample pool was then
chosen arbitrarily, and all peaks present at > 0.3 nA (10
fmol/125 µL serums) were automatically identified by computer. Peaks
that occurred before 5 min (end of void) or after 107 min (beginning of
wash phase) are not analytically reproducible and, therefore, were
removed from further consideration, leaving 1138 peaks (Table 2A
). Both chromatographic and software reliability were
simultaneously assessed by determining the ability to consistently
identify and quantitate these same peaks from the total set of 8 sample
pools. Overall, the software was able to find 297 peaks (26%) at
62.5% frequency, with a mean ± 50% and a coefficient of
variation of < 50% (Table 2A
). Although not
directly addressed in this study, the relative inefficiency of the
software-based peak recognition seems to result from the
instability of the ratio of the dominant and subdominant channels in
minor peaks (e.g., < 1 nA) and in peaks that have naturally low ratios
(Kristal et al. 1998
, Kristal et al. 1999
, Matson et al. 1984
, Milbury 1997
).
|
) reveals that
peaks found more frequently in the test pools also passed sequential
tests (mean, coefficient of variation) more frequently, with 88% of
peaks found in all pools passing the variation tests, compared with
only 31% of peaks identified in only 5/8 samples. This is consistent
with expectations that peaks that are more consistently identified by
combinations of ratios and retention times are more likely to be
correctly assigned. Outlier analysis
An additional analysis that should be carried out before
attempting to identify serotypes reflective of caloric intake is to
determine whether all members of the primary training set reflect the
population as a whole. Specifically, it is important to know whether
there are samples in the dataset that are statistical outliers. This
process is analogous to that used before determining a mean and
standard deviation in an experiment involving univariate statistical
calculations. In a univariate dataset, outliers may be removed by
analysis of their z-score relative to a standard population
or by using formal outlier analysis such as the Dixon test
(Dixon 1955
). For multivariate studies, such as the one
contemplated in our proposed experiments, we use a multivariate
equivalent of these tests. In the analysis shown in Figure 5
, data from approximately 100 variables of interest were used to test
the 16 rat samples in the training set for outliers (preset value:
expected to occur < 5% of the time by chance). This analysis was
conducted using the principal components analysis portion of Pirouette.
Sample data are examined using two different criteria to determine
outliers. Sample residual analysis, shown on the y-axis,
determines whether a sample is an outlier with respect to the sum of
the residuals of that sample (i.e., the residuals of a variable or
variables of the sample are greater than would be predicted by chance).
This is to say that this test determines whether the variability that
is not captured by the model exceeds expectations. Outliers, if
present, would fall above the heavy horizontal line at the top of the
Figure 5
. The other analysis, shown on the x-axis, uses the
Mahalanobis distance to determine whether a sample is an outlier with
respect to the mathematical factors of that sample. This is to say that
this test determines whether the variability that is captured by the
model exceeds statistical expectations. Outliers, if present, would
fall to the right of the heavy vertical line at the right of the Figure 5
. As can be seen in Figure 5
, no samples in the current study were
statistical outliers. Note that because AL and DR rats are predicted to
differ from each other, these tests would be expected to also recognize
an animal on an intermediate diet as belonging to the sample
population. Therefore, these tests are used to define as outliers those
rats that would fall outside of either the AL or DR populations (e.g.,
because of disease).
|
The final aspect of analytical validation pursued in the series of
experiments reported in this manuscript was to determine whether the
peaks that appeared sufficiently analytically well defined to be used
to test for serotype development were consistently found in female AL
and DR rats (validation for male animals will be carried out
separately). More specifically, the data shown in Table 2
reveal
analytical reproducibility—that we can reproducibly find 300 peaks
in a given serum sample chromatogram without optimizing the system. The
data shown in Figures 3
and 4
demonstrate that it is possible to find a
set of peaks that differ between a given AL and a given DR animal. The
data shown in Figure 5
indicate that a series of such peaks identify
the cohort chosen as being free of obvious outliers when examined using
a series of such metabolites as variables. For these
metabolites/markers to be useful, however, they must be consistently
observed across different cohorts as well as across individual animals.
Visual inspection of the entire serum chromatograms generated from
animals in cohorts 1 and 2 shows that, as expected, the majority of the
chromatograms are qualitatively conserved (Fig. 6
). That is to say that neither the analytical nor the animal husbandry
aspects of our study have greatly altered the serum components.
|
20-min
portion of the channel 9 chromatogram at 20-fold vertical magnification
shows nearly perfect qualitative and quantitative conservation of all
peaks in 12 rats from the two cohorts (Fig. 7A
). In contrast, the 20-min portion of channel 4 shown in
Figure 7B
shows that the same 12 rats are nearly completely
biochemically distinct in this region of the chromatogram. The solid
arrows in Figure 7B
mark conserved peaks and serve to
demonstrate that the chromatography was reliable. The open arrow marks
a triplet where the peaks are conserved qualitatively, and some but not
all of the peaks are conserved quantitatively. The CoulArray software
can generally correctly analyze the data from all three such peaks.
|
The above data suggest that we have a valid analytical approach to
address the problem proposed (i.e., serotypes) and that serums samples
are sufficiently conserved to attempt this analysis. The final step of
validation examined the computer’s ability to identify the 297 peaks
of interest in 8 AL and 8 DR samples. Analysis reveals that 40–50% of
the peaks of interest were found in all samples automatically, with
80% found in at least 5/8 samples. This suggests that a pool of
240 variables exists that are legitimate targets for automated
metabolic serotype analysis in female rats. Note that peaks that are
found less frequently may represent male-specific peaks (and hence
absent in the female samples) or peaks with as yet undetermined high
biochemical variability. Further optimization of search criteria were
previously shown to improve this success rate to 95% when analytes
are present (Kristal et al. 1998
).
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
The results reported reflected the need to validate some aspects of the
technology that we are using to characterize metabolic serotypes. The
basic theory and application of technology underlying HPLC separations
coupled with coulometric array detectors have been validated over
several years and do not require further validation (Milbury et al. 1998
, Kristal et al. 1998
, Kristal et al. 1999
, Beal et al. 1990
, Matson et al. 1987
, Matson et al. 1990
, LeWitt et al. 1992
, Ogawa et al. 1992
, Beal et al. 1992
, Matson et al. 1984
, Milbury 1997
, Acworth and Gamache 1996
). Consistent with
this, there were no problems in carrying out the basic analysis of
serum metabolites using this experimental protocol (Fig. 1)
. Analysis
of redox-active species present at subnanomolar levels, however,
has not yet been described in the literature. Validation of this
methodology has three components, analytical reliability, sensitivity
and biological variability. In other words, validation studies ask the
following questions: 1) Which, if any, metabolites present
in rat serums are sufficiently stable and analytically reproducible to
be correctly identified by computer-based searching algorithms?;
2) Can differences between two chromatograms of interest be
observed?; and 3) Which, if any, metabolites present in rat
serums are sufficiently consistent from rat to rat both within and
between cohorts to be useful for classification analysis?
The current study suggests that the system chosen meets these
requirements. Given the criteria presented in the Results section, we
estimate that 300 peaks (of 1150 total) have sufficient
analytical reproducibility to be screened for their potential use as
components of serotype (Table 2)
. Other peaks will require additional
approaches (work in progress). Of the peaks that appear analytically
valid, 250 appear appropriate for use in studies of the effects of
DR in young female animals (Table 3
). Sensitivity studies show that even relatively crude comparison of AL
and DR samples identified potential serums differences (e.g., Fig. 3
, 4
). The results of the studies presented in Figure 5
validate the
cohort chosen to serve as a baseline used to determine the consistency
of these differences. Furthermore, we found that the majority of peaks
appeared relatively consistent between rats within a given cohort (Fig. 6
, 7)
. These studies of biological variability suggest that most, but
not all, peaks are sufficiently biologically conserved to serve as
potential markers (Fig. 6
, 7)
of differences between the diet groups.
Because those peaks that are not well conserved will fail the
statistical comparison between AL and DR and/or the secondary
cohort-cohort comparisons, we did not further pursue their
elimination in this initial study.
|
|
ACKNOWLEDGMENTS |
|---|
|
FOOTNOTES |
|---|
2 Supported by Grant R01-AG15354 from the National Institutes of Health National Institute on Aging (to B.S.K.), ESA, and the Winifred Masterson Burke Relief Foundation.
4 Abbreviations used: DR, dietary restricted; HPLC, high performance liquid chromatography; AL, ad libitum; NIH, National Institutes of Health; ANOVA, analysis of variance.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
1. Acworth I. N., Gamache P. H. The coulometric electrode array for use in HPLC analysis, part 1: theory. Amer. Lab. 1996;5:33-37
2. Acworth I. N., Naoi M., PArvez H., Parvez S. Coulometric electrode array detectors for the HPLC 1997;6 VSP International Science Publication Utrecht, The Netherlands.
3. Beal M. F., Matson W. R., Storey E., Milbury P. E., Ryan E. A., Ogawa T., Bird E. D. Kynurenic acid concentrations are reduced in Huntington’s disease cerebral cortex. J. Neurol. Sci. 1992;108:80-87[Medline]
4. Beal M. F., Matson W. R., Swartz K. J., Gamache P. H., Bird E. D. Kynurenine pathway measurements in Huntington’s disease striatum; evidence for reduced formation of kynurenic acid. J. Neurochem. 1990;55:1327-1339[Medline]
5. Carlson A. J., Hoelzel F. Apparent prolongation of the life span of rats by intermittent fasting. J. Nutr. 1946;31:363-375
6. Cheney K. E., Liu R. k., Smith G. S., Leung R. E., Mickey M. R., Walford R. L. Survival patterns in C57BL/6J mice subjected to undernutrition. Exp. Gerontol. 1980;15:237-258[Medline]
7. Dixon W. J. Analysis of extreme values. Ann. Math. Stat. 1955;21:488-506
8. Goodrick C. L., Ingram D. K., Reynolds M. A., Freeman J. R., Cider N. Effects of intermittent feeding upon body weight and lifespan in inbred mice: interaction of genotype and age. Mech. Age. Dev. 1990;55:69-87[Medline]
9. Iwasaki K., Gleister C. A., Masoro E. J., McMahan C. A., Seo E.-J., Yu B. P. Influence of the restriction of individual dietary components on longevity and age-related disease of Fischer Rats: the fat component and the mineral component. J. Gerontol. 1988b;43:B13-B21[Abstract]
10. Iwasaki K., Gleister C. A., Masoro E. J., McMahan C. A., Yu B. P. The influence of dietary protein source on longevity and age-related disease processes of Fischer rats. J. Gerontol. 1988a;43:B5-B12[Abstract]
11. Kristal B. S., Vigneau-Callahan K. E., Matson W. R. Simultaneous analysis of the majority of low-molecular-weight, redox-active compounds from mitochondria. Anal. Biochem. 1998;263:18-25[Medline]
12. Kristal B. S., Vigneau-Callahan K. E., Matson W. R. Purine catabolism: links to mitochondrial respiration and antioxidant defenses?. Arch. Biochem. Biophys. 1999;370:22-33[Medline]
13. Kristal B. S., Yu B. P. Aging and its modulation by dietary restriction. Yu B. P. eds. Modulation of Aging Processes by Dietary Restriction 1994:1-36 CRC Press Boca Raton, FL.
14.
LeWitt P. A., Galloway M. P., Matson W. R., Milbury P. M., McDermott M., Srivastava D. K. Markers of dopamine metabolism in parkinson’s disease: the Parkinson’s Study Group. Neurology 1992;42:2111-2117
15.
Maeda H., Gleister C. A., Masoro E. J., Murata I., McMahan C. A., Yu B. P. Nutritional influences on aging of Fischer 344 rats: II pathology. J. Gerontol. 1985;40:671-688
16. Matson W. R., Bouckoms A., Svendson C., Beal M. F., Bird E. D. Generating and controlling multiparameter databases for biochemical correlates of disorders. Basic, Clinical and Therapeutic Aspects of Alzheimer’s and Parkinson’s Diseases 1990;II:513-516 Plenum New York, NY.
17. Matson W. R., Gamache P. H., Beal M. F., Bird E. D. EC array sensor concepts and data. Life Sci 1987;41:905-908[Medline]
18.
Matson W. R., Langials P., Volicer L., Gamache P. H., Bird E. D., Mark K. A. n-electrode three dimensional liquid chromatography with electrochemical detection for determination of neurotransmitters. Clin. Chem. 1984;30:1477-1488
19. McCay C. M. Cellulose in the diets of rats and mice. J. Nutr. 1935;:435-447
20. McCay C. M., Crowell M. F., Maynard L. A. The effect of retarded growth upon the length of lifespan and upon the ultimate body size. J. Nutr. 1935;10:63-79
21. Milbury P. E. CEAS generation of large multiparameter databases for determining categorical process involvement of biomolecules. Coulometric Array Detectors for HPLC 1997:125-141 VSP International Science Publication Utrecht, The Netherlands.
22. Milbury P. E., Vaughan M. R., Farley S., Matula G. J., JR, Convertino V. A., Matson W. R. A comparative bear model for immobility-induced osteopenia. Ursus 1998;10:507-520[Medline]
23. Nolen GA Effect of various restricted dietary regimens on the growth, health, and longevity of albino rats. J. Nutr. 1972;102:1477-1494
24.
Ogawa T., Matson W. R., Beal M. F., Myers R. H., Bird E. D., Milbury P. E., Saso S. Kynurenine pathway abnormalities in Parkinson’s disease. Neurology 1992;42:1702-1706
25. Ross M. H., Bras G. Food preference and length of life. Science 1970;190:165-167
26. Stucklikova E., Juricova-Horakova E. M., Deyl Z. New aspects of the dietary effect of life prolongation in rodents: what is the role of obesity in aging. Exp. Gerontol. 1975;10:141-144[Medline]
27.
Tannenbaum A. The dependence of tumor formation on the composition of the calorie-restricted diet as well as on the degree of restriction. Cancer Res 1945;5:616-625
28. Turturro A., Witt W. W., Lewis S., Hass B. S., Lipman R. D., Hart R. W. Growth curves and survival characteristics of the animals used in the biomarkers of aging program. J. Gerontol. 1999;54A:B492-B501
29. Weindruch R., Walford R. The Retardation of Aging and Disease by Dietary Restriction 1988 Charles C. Thomas St. Louis, MO.
30. Weindruch R. H., Walford R. L. Dietary restriction in mice beginning at one year of age: effects on life-span and spontaneous cancer incidence. Science 2000;215:1415-1418
31. Yu B. P. Aging, oxidative stress: modulation by dietary restriction. Free Radic. Biol. Med. 1996;21:651-668[Medline]
This article has been cited by other articles:
|
|
 |
|
  A. L. Castle, O. Fiehn, R. Kaddurah-Daouk, and J. C. Lindon Metabolomics Standards Workshop and the development of international standards for reporting metabolomics experimental results Brief Bioinform, June 1, 2006; 7(2): 159 - 165. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 M. S. Sabatine, E. Liu, D. A. Morrow, E. Heller, R. McCarroll, R. Wiegand, G. F. Berriz, F. P. Roth, and R. E. Gerszten Metabolomic Identification of Novel Biomarkers of Myocardial Ischemia Circulation, December 20, 2005; 112(25): 3868 - 3875. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 H. Shi, K. E. Vigneau-Callahan, A. I. Shestopalov, P. E. Milbury, W. R. Matson, and B. S. Kristal Characterization of Diet-Dependent Metabolic Serotypes: Proof of Principle in Female and Male Rats J. Nutr., May 1, 2002; 132(5): 1031 - 1038. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
H. Shi, K. E. Vigneau-Callahan, A. I. Shestopalov, P. E. Milbury, W. R. Matson, and B. S. Kristal Characterization of Diet-Dependent Metabolic Serotypes: Primary Validation of Male and Female Serotypes in Independent Cohorts of Rats J. Nutr., May 1, 2002; 132(5): 1039 - 1046. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
B. C. Hansen Introduction J. Nutr., March 1, 2001; 131(3): 900S - 902. [Abstract] [Full Text]
|
|
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
