| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
* Institute of Pathology and
Animal Resource Center, Case Western Reserve University, Cleveland, Ohio 44106, USA
1Correspondence: Institute of Pathology, Case Western Reserve University, 2085 Adelbert Road, Cleveland, OH 44106, USA. E-mail: drs7{at}po.cwru.edu
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Key Words: aging • biomarker • glucose • oxidation • Maillard reaction
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
. Glucose and other sugars, or their
degradation products, have been implicated in the aging process by
formation of such irreversible molecules (2)
. In support
of this hypothesis, studies show that glucose tolerance progressively
declines with age (3
, 4)
whereas advanced glycation and
glycoxidation products increase with age in the extracellular matrix
(5
6
7)
. In contrast, the anti-aging intervention caloric
restriction, which is known to increase life span (8)
,
improves glucose tolerance in rats (9
, 10)
and monkeys
(11
12
13)
, and decreases glycoxidation in rodents (5
, 14)
. Conversely, diets high in sugars increase glycoxidation
(15)
and decrease longevities in rats (16)
and mice (17)
.
In the late 1980s, the National Institute on Aging (NIA) launched a
10-year program aimed at identifying biomarkers of aging, i.e.,
biological or biochemical markers that would predict the onset and
progression of age-related processes and longevity (18)
.
Rodents were chosen as models because their life spans are considerably
shorter than those of humans and with the hope of using the biomarkers
in future studies of aging processes and their interventions in humans
(18
, 19)
. However, the concept of the actual existence of
markers of aging has been controversial (20)
, with many
differences in opinions about the exact criteria and paradigm that make
good candidate markers (18
19
20
21)
. Some authors suggested
that because of the complexity of aging phenomena, not just one, but
many diverse markers, or a panel of markers, would be necessary in
order to achieve success if, indeed, such a paradigm could be
identified (20)
.
In this prospective, previous results from this (5
6
7)
and
other laboratories (14
, 22)
have shown that glycoxidation
markers correlate strongly with chronological age in cross-sectional
studies, suggesting that they have strong potential as candidate
biomarkers for practical application in longevity and intervention
studies. Subsequently, it was also shown that glycoxidation rate
determined by pentosidine in skin collagen was inversely related to
maximum life span across several mammalian species (5)
. In
this latter study, pentosidine (Fig. 1
) was related to longevity not by absolute level, but by the rate of
formation represented by slopes, i.e., its rate of change over time.
However, data in these previous studies were obtained cross-sectionally
such that inference to individual longevities could not be made. Thus,
in the present study we have investigated three markers of aging
longitudinally in mice (Fig. 1)
: furosine, i.e., a measure of
glycation; and N
-(carboxymethyl)lysine (CML)
and pentosidine, i.e., measures of glycoxidation. The hypothesis tested
is that longitudinal determination of glycation and glycoxidation rates
can predict individual longevities in ad libitum (AL) and
caloric restricted (CR) mice and that CR will retard this rate. We have
further investigated the relationship of these rates to major tissue
pathologies present at death. In short, these results show that these
rates modeled as linear equations were significantly inhibited by CR
vs. AL for all parameters, and in all cases were found to vary
inversely with longevity. Furosine, i.e., glycated lysine residues,
surprisingly emerged as the most important factor in explaining life
span of these mice.
|
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
. CR mice were fed a modified version of
this diet that was made nutritionally adequate in minerals and vitamins
by supplementation but deficient in calories by restricting the diet to
60% of the average calorie intake of AL mice, as described
(23)
. This consisted of pellets cut according to known
weight and daily feeding according to a predetermined pellet weight and
age schedule provided by the NCTR as described (24)
. Each
mouse was weighed approximately once a month until death on an
electronic balance. The histories of body weights of these mice prior
to shipment were obtained from the NCTR. The assessment of SPF status of these mice was done biannually by in-house disease surveillance monitoring using CF1 sentinel mice. This consisted of transferring dirty bedding from cages of experimental to those of the sentinel mice, which were later killed and examined for tissue serology, bacteriology, and parasitology by Charles River Laboratories (Wilmington, Mass.).
Skin procurement by biopsy
Skin biopsies were obtained surgically from mice at age 605
days. Cages were moved to a location under a laminar flow/fume hood,
the surface of which was kept moist during surgery with Alcide sprayed
onto a layer of paper towels. Each mouse was anesthetized by Metofane
(Schering-Plough Animal Health Corp., Omaha, Nebr.). The continuous
maintenance of the anesthetic during surgery was provided by inserting
the nose of the mouse into the plugger end of a 10 cc syringe
containing gauze moisten with Metofane. An approximate 4 mm diameter
skin biopsy was taken from the dorsal region near the neck after
shaving the region with surgical clippers and prepping with iodine
solution. The incision was closed with a surgical clip applied by the
Autoclip Wound Closing System (Stoelting, Wood Dale, Ill.). Each biopsy
was washed in saline and stored frozen at -70°C. When the incision
healed, the clip was removed by an autoclip remover utensil
(Stoelting).
Postmortem handling of mice
During this study, mice were carefully examined daily for
vitality and sickness. At death, mice were weighed and evaluated at
necropsy for gross pathological diagnoses by the attending veterinarian
(Dr. N. Kleinman). Their skin was removed starting from the neck down
and stored frozen at -70°C.
Skin processing and analytical assays
Hair was removed from skin by scraping frozen tissue samples
with the blade of a scalpel as described previously (7)
.
Likewise, samples were homogenized and extracted (7)
. In
all analyses described below, specimens belonging to the same animal
were processed consecutively in order to minimize experimental error.
Pentosidine was determined in acid hydrolysates of processed skin
samples by high-performance liquid chromatography (HPLC). Acid
hydrolysates equivalent to 50 µg hydroxyproline (hyp) contained in
volumes of either 152–350 µl (tissue at biopsy) or 57–163 µl
(tissue at death) were pipetted into the bottom of 12 x 75 mm
borosilicate glass tubes, followed by addition of the internal
standard, 50 µl of 1 µg/ml pyridoxamine (Sigma Chem. Co., St.
Louis, Mo.). Contents of the tubes were then dried using a Speed-Vac
Rotary Centrifuge (Savant Instruments, Holbrook, N.Y.), followed by
reconstituting each tube with 200 µl of water containing 0.1 M
heptafluorobutyric acid. Totals of 160 µl equivalent to 40 µg
hyp—i.e., equivalent 256 µg collagen based on 14% hyp of collagen
by weight (5)
—were injected onto a HPLC containing a
4.6 x 250 mm 218TP54 C18 Vydac, 5 µ column equilibrated and
eluted as described (7)
. Pyridoxamine and pentosidine
eluted at ~11 and 33 min.
Lysine, furosine, and CML were determined in acid hydrolysates of
processed skin samples derivatized as their trifluoroacetyl methyl
esters by selected ion-monitoring gas chromatography/mass spectrometry
(SIM-GC/MS) using modified procedures of Dunn et al. (25)
and Glomb and Monnier (26)
. Acid hydrolysates equivalent
to amounts previously described for pentosidine were pipetted into the
bottom of 13 x 100 mm screw-type cap culture tubes, followed by
the addition of 100 µl of 4 nmol/ml deuterated CML
(d4-CML) as the internal standard. The contents of each tube
were then dried by Speed-Vac (Savant). One milliliter of anhydrous
methyl alcohol containing 0.2 M thionyl chloride (Aldrich, Milwaukee,
Wis.) was pipetted into the tube, followed by sealing with a Teflon
liner screw cap. The tube was placed into a 110°C oven for ~30 min,
followed by evaporating contents with a stream of nitrogen. A total of
0.5 ml anhydrous dichloromethane and 1 ml of trifluoroacetic anhydride
(Aldrich) was added and the tube was left to sit at room temperature
for 1 h. After evaporation of organic solvents (nitrogen), the
contents of each tube were transferred to 100 µl glass inserts to
autosampler vials (Hewlett-Packard, Wilmington, Del.) with three
separate 100 µl washes with dichloromethane. The solvent in the
insert vial was dried between washes with a stream of nitrogen. After
the last wash, the solvent was dried and the contents in the vial were
reconstituted with 25 µl of dichloromethane. For each sample, 2 µl
was injected onto a Hewlett-Packard GC/MS equipped with a 5890 Series
II Gas Chromatograph, 5971 Series Mass Selective Detector, and a 6890
Series Automatic Injector. Separations were made on a 25 m X 0.2
mm X 0.33 µM Ultra 2 column cross-linked with 5% PH ME siloxane
(Hewlett-Packard) using the temperature program described by Dunn et
al. (25)
. Lysine, D4-CML, CML, and furosine were monitored
at ions m/z = 320, 396, 392, and 110, which eluted at 23, 29.4,
29.4, and 36.2 min, respectively. Quantitation was by comparison to
calibration curves using standards prepared and run under similar
conditions.
Diagnosis and management of dermatitis
Mice were assessed for absence or presence of dermatitis. Cases
were categorized into being mild, moderate, or severe. A case was
designated mild if scratching occasionally occurred but was cured by a
single clip of toenails. Moderate cases were diagnosed when scratching
was chronic but held in check by multiple clips of toenails over time
(i.e., 
two clips). More severe cases involving episodes of both
scratching and chewing, causing skin lesion ulcerations, were held in
check by multiple toenail clipping and application of a mouse
Elizabethan collar (Stoelting) until healing.
Evaluation at necropsy for gross pathology and major diagnosis at
death
Each mouse was evaluated at death for major pathological
diagnosis. The thorax cavity was examined first. Cardiomegaly was
diagnosed upon observing unilateral or bilateral heart enlargement.
Cardiopathy/cardiomyopathy was diagnosed if inflammation or other
apparent abnormalities such as chamber dilations, blood clots, masses,
or valve nodules were present. Lungs were examined for collapse,
atelectasis, congestion, effusions, edema, and masses. In severe cases
of cardiomegaly and cardiopathy, congestive heart failure was
diagnosed.
The abdominal cavity was examined next. The internal contents of the stomach and the remainder of the intestines were examined for the presence of melena. If melena was present, it was diagnosed as ulcerative gastroenteritis. In some cases, melena existed only in the stomach; however, most frequently it extended to portions of the small intestines down to the level of the cecum and colon.
The liver was examined for size, color, architecture, infiltrations, and growths. Liver enlargement, or hepatomegaly, was frequently accompanied by infiltrative hepatopathy involving discoloration, loss of normal architecture, and lobular accentuation. However, in other cases, especially in CR mice, infiltrates observed as white or red foci on lobes were not necessarily accompanied by increased liver size. In contrast, in some mice, especially those of the CR cohort, the liver appeared to be smaller than normal and frequently had a yellowish discolored appearance. Mice that showed chronic weight loss and emaciation were diagnosed with hepatic degeneration/insufficiency. However, whether this diagnosis actually caused death or was secondary to other pathology was not investigated.
The entire spleen or one of its portions was diagnosed upon enlargement as splenomegaly. Enlargement was frequently homogenous whereas others showed dark or white colored foci. Other spleens showed accentuation of white pulp, growths, cysts, or blood-filled hemangiomas. Similar to the liver, many CR mice had enlarged spleens as well; however, others had very small spleens, appearing below normal in size.
Renal pathology was diagnosed upon observing unilateral or bilateral enlargement of kidneys, tumor-appearing masses, growths, or nodules on the kidney surface, and dilations of renal tract parts (i.e., dilations in renal pelvis).
Seminal vesiculitis was diagnosed and graded for severity as either mild, moderate, severe, or extremely severe. Mild cases involved unilateral or bilateral enlargement (i.e., hypertrophy or dilatation) of the seminal vesicles with opacity. Moderate cases, which were most numerous, involved greater enlargement, yellow discoloration, opacity, and turgidity. Severe cases were similar to moderate ones, but also showed unilateral or bilateral internal hemorrhaging and/or suppuration. Extreme conditions further showed red and black discoloration, torsion, ischemia, necrosis, and, in one case, rupture of the vesicle contents into the peritoneal cavity.
Statistical methods
Statistical methods have been described elsewhere
(5
6
7)
and thus are only summarized here. Data were tested
for normality by the W test and homogeneity of variance by either the
Burr-Foster Q test or the Levene’s test by SPSS software (SPSS, Inc.,
Chicago, Ill.). Pearson correlation tables, regression analyses, and
ANOVA were done by SPSS. Age adjustment of data, outlier test, and
comparison of regression equations for slopes were done according to
methods previously described (5
6
7)
. The Kruskal-Wallis
one-way ANOVA was computed by SPSS.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
). The first mouse in the AL cohort to die was 629 days old, which also
coincided with the rapid rate phase of death in the survival curve
(Fig. 2)
. Likewise, the first death in the CR cohort occurred at the
age of 638 days; however, mortality acceleration did not occur until
the death of the third mouse at age 791 days (Fig. 2)
. The median and
maximum life spans for cohorts were as follows: AL, 842 and 951 days;
CR, 1040 and 1442 days. The tailing effect frequently observed in
survival curves due to the delay of death in the oldest living
individuals of the population was not observed for the AL cohort. The
tail observed for CR mice was essentially due to a single mouse that
lived to age 1442 days, superseding the death of the previous CR mouse
at age 1195 days (Fig. 2)
.
|
Mean body weights as shown in Fig. 3
were significantly (P<0.0001) affected by diet
(r=0.83) and age (r=0.42). Body
weights for AL mice increased during growth and maturation, reached a
peak between the ages of 364 and 504 days, and thereafter progressively
declined into senescence (Fig. 3)
. Body weights for CR mice were
consistently lower than AL mice of similar age and showed considerably
less variation (Fig. 3)
. The maximum body weights attained for AL and
CR mice were 41 ± 5 and 28 ± 2 g, which occurred at
504 days of age. Thereafter, the progressive decline observed in body
weights was considerably more rapid for AL vs. CR mice (Fig. 3)
.
|
Results for survival of AL and CR mice, including median and maximum
life spans and the degree of curve triangulation (Fig. 2)
, are similar
to previous results published for this strain by the NIA/NCTR, i.e., AL
vs. CR: median, 840 vs. 966 days; maximum, 1050 vs. 1246 days (8
, 28)
. The observed lack of a distinct tail in survival curves
(Fig. 2)
is probably due to fewer mice represented in the present study
compared with the previous survival data published for these mice
(8
, 28)
. Similarly, the patterns in mean body weights for
this strain over life span (Fig. 3)
, including maximums achieved and
the observed middle age peaking for AL mice prior to late age decline,
also agree with previously reported data (8
, 28)
. The
further observations of a lack of a peaking effect in body weights for
CR mice, but with greater weight uniformity compared with AL mice, have
also been previously observed (8
, 28)
.
The effects of diet, age, and longevity on lysine content of mouse
skin collagen
The biochemical relationships among markers assayed in this study
are shown in Figs. 4
5
6
7
. Mean content of lysine residues, a common
denominator used in many studies of the Maillard reaction in
vivo (14
, 25
, 27
, 29)
, was found to be significantly
(P<0.0001) greater at biopsy than at the time of
death: 773 vs. 280 nmol/mg collagen (compare scale of y axis
for Fig. 4A
vs. Fig. 4B
). The results at biopsy were significantly
(P=0.001) less for AL vs. CR, 598 vs. 942 nmol/mg collagen.
In contrast, lysine content at death was significantly
(P<0.0001) but moderately greater for AL vs. CR, 293
vs. 266 nmol/mg collagen.
|
|
|
|
Additional relationships were noted among the variables lysine,
sampling time, and longevity when data from mice were pooled across
diets for regression analyses (i.e., AL+CR, n=63). At the
time of biopsy (Fig. 4A
), lysine correlated weakly but
positively (r=0.29) with longevity
(P=0.023). In contrast, a significant
(P<0.002) inverse relationship
(r=-0.39) was noted at the time of death between
lysine and longevity (Fig. 4B
). In these analyses, the
effect of diet on lysine level of skin was significant at biopsy
(r=0.43, P=0.0004) and at death
(r=0.43, P=0.0005). However, none
of these aforementioned relationships were significant
(P>0.05) when data were analyzed separately within
diets, i.e., AL (n=31) vs. CR (n=32) (Fig. 4A, B
). In further analyses, lysine levels of skin collagen
at biopsy were correlated with levels at death (Fig. 4C
).
When mice were analyzed across diets (n=63), a significant
(P<0.002) inverse relationship
(r=-0.38) was noted between levels at biopsy vs.
death. When analyzed within diets, however, the relationship was
significant for CR (r=-0.40,
P<0.025, n=32) but not AL
(r=-0.10, P>0.05,
n=31). Overall, these results suggest that the fewer the
lysines modified by the aging process, the greater the longevity of the
animal.
Because of the differences noted in lysine levels with tissue sampling
time and longevity, skin parameters of glycation (i.e., furosine) and
glycoxidation (i.e., CML and pentosidine) were expressed separately
both per milligram collagen and per µmoles lysine. Pearson
correlation analyses in Table 1
show that two methods for expression of parameters were highly
correlated with each other at the time of death
(r=0.90 to 0.97, P<0.0001), but
correlated considerably less at the time of biopsy
(r=0.13 to 0.67, Table 1
). However, overall composite
relationships (n=126) between the two methods of expression
were significantly correlated for all parameters as follows: furosine
(r=0.72, P<0.0001), CML
(r=0.38, P=0.0001), and pentosidine
(r=0.72, P<0.0001). Except where
noted, the results below express measured parameters on the basis of
µmoles lysine.
|
Relationship between longevity and markers of glycation and
glycoxidation
Pearson correlation analyses of levels of furosine, CML, and
pentosidine at the time of biopsy and at death vs. longevity are
summarized in Table 2
. Levels at biopsy when expressed per µmoles lysine correlated weakly
with longevity within and across diets. Only in one of these
comparisons did the results reach significance
(P<0.05): pentosidine in CR mice vs. longevity,
r=-0.36 (Table 2)
. These correlations improved
somewhat when levels at death were analyzed (Table 2)
, and were
especially significant and positively correlated with longevity
(r=0.33, 0.44, 0.81) when results were pooled across
diets for furosine (P<0.035), CML
(P<0.002), and pentosidine
(P<0.0001), respectively. There was a significant
(P<0.02) trend for these parameters measured at
biopsy in AL mice to correlate inversely with longevity when results
were expressed as per milligram of collagen. As shown in Table 2
and
plotted in Fig. 5
, the results reached significance for furosine
(r=-0.41, P=0.021) and CML
(r=-0.50, P=0.005), but not
pentosidine (r=-0.19, P=0.39).
Further multifactor analysis consisting of levels at biopsy vs.
longevity in mice showed that prediction (r=0.55)
reached significance (P=0.019) with the equation:
longevity = 1032 -1.28 (furosine) - 0.18 (CML) - 6.95
(pentosidine).
|
In another study, levels at the time of biopsy and death were plotted
vs. age (Fig. 6
). Initial analyses showed no significant (P>0.05)
effect of AL vs. CR on parameters at the time of biopsy: mean furosine,
548 vs. 481 pmol/µmol lysine (P=0.14); CML, 274 vs.
225 (P=0.42); pentosidine, 2.2 vs. 2.9 (P=0.39)
(data not shown). Further correlations across sampling times and diets
(n=126, Table 2
) showed significant (P<0.0001)
relationships with age for furosine (r=0.75), CML
(r=0.46) and pentosidine (r=0.73)
(Table 2)
. As shown in Fig. , 6a
line was drawn connecting each
mouse’s level at the time of biopsy to its level at the time of death.
Thus, each line represented a slope or rate of change of the parameter
(dP/dt) from the period starting at biopsy (i.e., age 605
days) to the time of death. These slopes were in turn plotted vs.
individual longevities, as shown in Fig. 7
.
Figures 6
and 7
show that the slopes are negative for some mice in the
study (Fig. 7)
, which is due to the level at biopsy being greater than
the level at the time of death. Second, some mice showed slopes that
deviated from those of the main group. As proved by a statistical test,
such deviations were found to be outliers that occurred particularly in
mice dying soon after biopsy. These outliers, as listed in Table 3
, were not included in the further statistical analyses. Third, mice
with early demise, especially those of the AL cohort, showed a rapid
increase in rate over the time period. This is most pronounced for
furosine levels of AL mice shown in Fig. 6A
. This figure
also shows an age-related, accelerated spreading of furosine levels,
which is elevated in AL vs. CR mice.
|
An inverse relationship between longevity and rate of change for each
parameter is evidenced by the regression lines drawn for AL and CR
cohorts of Fig. 7
. Regression equations for these lines are shown in
Table 3
. For AL mice, a highly significant
(P<0.0001) inverse relationship is noted between
longevity of mice and rates of change in furosine
(r=-0.70) and pentosidine
(r=-0.70), whereas that of CML
(r=-0.33) approached significance
(P=0.067). The same inverse relationship noted for AL
mice holds true for CR mice. However, the relationship is attenuated by
CR as evidenced by the lack of decline of these lines shown in Fig. 7
and equated in Table 3
, being significant for furosine
(r-0.39, P=0.028) and CML
(r-0.56, P=0.0015), but
approaching significance for pentosidine (r-0.31,
P=0.09). A comparison of these equations for slopes
shows that for each parameter the overall rate of change is
significantly (P<0.001) less for CR vs. AL mice
(i.e., slopes for furosine, 0.006 vs. 0.016; CML, 0.001 vs. 0.005;
pentosidine, 2x10-5 vs.
2x10-4). In contrast, prediction of these
slopes by testing against life span and diet as the independent
variables during regression analyses (data not shown) shows life span
to be significant for all three parameters: furosine
(P=0.0001), CML (P=0.03), and
pentosidine (P=0.0002). However, the effect of diet
was marginal: furosine (P=0.38), CML
(P<0.10), and pentosidine
(P<0.05).
In further evaluation, slopes of Figs. 6
and 7
were regressed on
longevities by using stepwise multiple regression analysis in order to
indicate the most important factors in determining the life spans of
mice in this study. Diet (P<0.0001) and furosine
(P=0.0002) as dFur/dt were selected with
the equation (r=0.75): longevity = 748 +
169(diet) - 34(dFur/dt). Further proof of the importance of
furosine in explaining life spans was by age adjustment of levels for
all parameters at the time of death (data not shown). Statistical
comparison of AL vs. CR data by the Kruskal-Wallis one-way ANOVA for
levels age-adjusted to 950 days showed significance for furosine
(P=0.003), but not CML (P=0.65) or
pentosidine (P=0.49).
Relationship between diet and tissue pathology at death
Specific pathological changes were found either decreased,
increased, or unaffected by CR vs. AL (Fig. 8
). As previously reported (30
31
32
33)
, many pathological
changes were decreased by CR vs. AL including the percent incidences of
seminal vesiculitis (22 vs. 83%, r=0.65,
P<0.0001), ulcerative dermatitis (3 vs. 58%,
r=0.60, P<0.0001), splenomegaly
(22 vs. 42%, r=0.34, P=0.006) and
enlarged mesenteric lymph nodes (6 vs. 26%, r=0.27,
P=0.034). Furthermore, there was also a
nonsignificant (P>0.05) trend for CR to reduce both
heart (r=0.15) and renal (r=0.11)
pathologies, although the incidence of the latter pathology was minor
in this study (6 vs. 13%). In contrast, several pathologies were
increased by CR, which included hepatic degeneration/insufficiency (41
vs. 13%, r=0.31, P=0.013) and
minor problems with tooth loss/malocclusion observed only in CR mice
(13% vs. ND, r=0.26, P=0.043).
Finally, a large incidence of ulcerative gastroenteritis was observed
in both CR vs. AL mice (78 vs. 87%, P>0.05).
Incidences of infiltrative hepatopathy (38 vs. 48%) and subcutaneous
(s.c.) edema (22 vs. 16%) were not significantly
(P>0.05) different between dietary cohorts (Fig. 8)
.
|
Correlations between pathologies and associations with longevity
Because postmortem pathology varied widely between AL and CR
cohorts, further analyses were conducted within diets with regard to
correlations between pathologies and their relationship to age. To
summarize, seminal vesiculitis was the only pathology that correlated
with age, but only in AL mice. Incidences (r=0.54,
P=0.002) and severities (r=0.54,
P=0.002) were positively correlated with longevities:
older AL mice tended to have more severe cases. Furthermore, seminal
vesiculitis mildly but negatively correlated with slopes representing
rates of glycation and glycoxidation as follows: furosine
(r-0.29, P=0.11); CML
(r-0.38, P=0.038); and
pentosidine (r-0.45, P=0.012).
Finally, the most noteworthy and consistent observation in both AL and
CR cohorts was the interrelationships between s.c. edema and
pathologies involving both the spleen and liver. In short, infiltrative
hepatopathy was positively correlated with hepatomegaly (AL,
r=0.71, P<0.0001; CR,
r=0.58, P=0.001), splenomegaly (AL,
r=0.44, P=0.013; CR,
r=0.49, P=0.004) and s.c. edema
(AL, r=0.45, P=0.011; CR,
r=0.68, P<0.0001). These
observations are most likely related to metastatic neoplasia causing
sequential hepatic dysfunction, hypoproteinemia, and edema.
Major diagnoses as probable cause of death
From necropsy examinations and previous pathological analyses,
major diagnoses were evaluated as to the most probable cause of death
for each mouse. These results are summarized in Table 4
. As observed, the causes of death were not determined in 13 CR and 9 AL
mice. Neoplasia, including both hepatoma (n=10) and lymphoma
(n=9), was the major diagnosis made in AL mice. Even though
hepatoma (n=10) was major in CR mice, the incidence of
lymphoma (n=3) was suppressed in this cohort.
|
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
. The significance of this impaired tolerance is
important and has profound significance for the elderly population. The
impairment can either contribute, cause, or catalyze a wide range of
diseases such as diabetes and various cardiovascular anomalies
including hypertension, atherosclerosis, and coronary artery disease
(34
, 35)
. A further concern is for nondiabetic or
borderline diabetic individuals who are unaware or apathetic to its
diagnosis, risks, and ramifications (34
35
36)
. In rodents,
however, experimental data concerning impairment of glucose tolerance
with advancing age have depended on species, strain, and gender
examined (6)
. However, it has generally been concluded
that an insulin-resistant state exists in aging rodents similar to that
observed in humans (3)
.
In previous cross-sectional studies with C57BL/6 mice, it was found
that glycohemoglobin levels increased until ~12 months of age and
declined thereafter, thereby reflecting the body weight data in Fig. 2
.
This suggests that glycohemoglobin measurement is not a reliable
indicator of cumulative glycemia at late age in this strain
(6)
. Cross-sectional comparisons of CR vs. AL subsequently
showed no difference in pentosidine levels in skin and tendon collagens
at ages 6, 12, 18, or 24 months (7)
. Conversely, CR vs. AL
was found to significantly inhibit furosine formation in skin collagen
at 24 months, but not at earlier ages (6)
. Taken together,
these results suggest that impairment of glucose tolerance probably
starts some time late in the life span of these mice—soon after the
age of 20 months.
In three separate studies of the accumulation of AGEs and intervention
paradigms on Type I diabetes in humans it was found that the degree of
glycemic control strongly modulated the levels of glycation and AGE
formation in skin collagen (27
, 39
, 40)
. In general, these
results showed that the more intensive the therapy, the slower rate of
the accumulation. Most noteworthy was that glycation (Fig. 1)
was
reversed with improved glycemic control in all studies (27
, 39
, 40)
. However, the hypothesis of whether AGE formation can be
reversed with diabetic intervention remains questionable (27
, 39)
. In one study comparing long-term intensive vs. conventional
therapy in participants to the Diabetes Control and Complications
Trial, collagen glycation as furosine emerged as the parameter most
consistently associated with diabetic complications in subjects not
under intensive glycemic control (40)
even when all
parameters (including glycated hemoglobin and measures of collagen
cross-linking, glycation, and glycoxidation) were evaluated. In this
study, it was observed that despite the reversibility of glycation
(Fig. 1)
, glycated residues in collagen surprisingly reflected
cumulative glycemia over 12 months and more. Thus, steady-state levels
of glycemia tend to change little in diabetic individuals over long
periods of time (40)
. A similar conclusion was reached in
the studies of age-related glycemic patterns in healthy individuals
(41)
.
In the present investigation, the finding that furosine formation rates
(dFur/dt) are better predictors of individual longevities in
mice than those of CML and pentosidine (Figs. 6
and 7)
may best be
explained by the biochemical origins of these products (Fig. 1)
.
Whereas furosine can originate only from glucose (42)
and
reflect cumulative glycemia (40
, 41)
, CML and pentosidine
can originate from multiple sugars and are influenced by oxidative
events that are catalyzed by transition metals and inhibited by
antioxidative factors intrinsic and extrinsic to tissues
(43)
. Furthermore, CML has multiple origins, including
serine myeloperoxidase (44)
and lipid peroxidation
(45)
reactions. In contrast, pentosidine is specific for
the Maillard reaction (6)
. Indeed, superimposed on these
events is the fact that collagen turnover decreases with age
(1)
. Undoubtedly, the reversibility of the glycation
reaction shown in Fig. 1
may be minimized by this factor.
Previous studies by Masoro et al. (46)
showed that CR
maintains lower plasma glucose and insulin levels than AL control rats.
Such results suggest a primary role of an altered carbohydrate
metabolism by CR in influencing aging processes and enhanced life span.
Based on these findings (46)
, it was subsequently
postulated by Parr (47)
that insulin exposure and
sensitivity control the rate of mammalian aging. The enhancement by CR
was speculated to be mediated through insulin-induced activation of
tyrosine kinase receptor and subsequent signaling pathways. Studies
with the nematode model of aging, C. elegans, showed that
mutations in the daf-2 gene that encodes an insulin
receptor-like molecule, DAF-2, increases life span (48)
.
The fact that components of DAF-2 show strong structural and functional
homology to the mammalian insulin receptor, including a tyrosine kinase
domain, suggests involvement in the insulin signaling pathway that
controls both metabolic and diapause events in this worm. Thus, it was
concluded by Kimura et al. (48)
that mutations in the
daf-2 gene mimic the effects of CR on life extension
observed in rodents by a similar mechanism. However, a more recent
study of CR in C. elegans suggests this mechanism is
distinct from that of CR. Daf-2 mutations control diapause
events, leading to a distinct phenotypic dark appearance due to fat
accumulation in the intestine. In contrast, mutations in other genes
such as eat-2, which controls food intake, present a totally
different phenotype; namely, emaciation and a very pale appearance due
to decreased food intake and low intestinal stores of fat
(49)
.
Since the discovery of longevity genes that are associated with these glycation and glycoxidation reactions is of considerable interest for mammalian aging, collaborative efforts with Drs. Richard Miller and David Burke (University of Michigan) are in progress to map specific gene loci markers with measured parameters of this study in a population of genetically heterogeneous mice, i.e., progeny of (BALB/cJ X C57BL/6J) F1 mothers crossed to (C3H/HeJ X DBA/2J) F1 fathers. Preliminary results (unpublished observations) suggest genes on mouse chromosomes 1, 2, and 8 may have associations with CML, furosine, and pentosidine, respectively. Preliminary data, though not yet fully conclusive, suggest that a locus on chromosome 12 may influence both longevity and low levels of furosine in collagen.
In the present investigation, specific pathologies were followed for
major diagnoses at death in mice and associations with measured
parameters. Incidences of neoplasia, dermatitis, and seminal
vesiculitis (Fig. 8)
have been observed for this strain by other
investigators as well (32
, 50
51
52
53)
. This is in addition to
the noted delay/suppression by CR on these pathologies (30
31
32
, 54)
. Of particular interest is that in a previous study, CR was
able to retard the development of diethylnitrosamine-induced tumors in
the livers of mice (55)
. In this study, age-related
insulinaemia reflecting resistance of tissues to the effects of insulin
was much greater for AL vs. CR mice and correlated with the number of
liver foci. In consideration of the known mitogenic effects of insulin
on cultured hepatocytes, these investigators hypothesize that insulin
may contribute to the promotion of these tumors in vivo. In
light of the findings of Barzilai and Rossetti (3)
showing
the association between hepatic insulin resistance and age in rodents,
it is intriguing that numerous studies describing cell surface
molecules known to be involved in the signaling cascades for oncogene
activation are characterized as insulin receptors
(56
57
58)
. These discoveries are significant for three
reasons. First, the overexpression or mutation of oncogenes such as
Ras genes is a common cause of cancer especially liver
neoplasia in mice (57)
. Second, dysregulation of the
signaling pathways at the level of the receptor or other key elements
in the cascade has been shown to result in a variety of neoplastic
disorders such as hepatoma, lymphoma, and other carcinomas
(58)
. Third, recent epidemiological evidence points to a
link between non-insulin-dependent Type II diabetes and various cancers
in humans, suggesting an association between neoplasia and insulin
resistance (58)
.
In conclusion, the results of this longitudinal study show that
the skin collagen glycation rate measured as furosine could predict
early death in AL and CR mice (Table 3)
. Similarly, pentosidine and CML
when used as indices of glycoxidation could predict early death in AL
and CR mice, respectively (Table 3)
. However, further stepwise
regression analysis selected diet and rate of glycation of skin
collagen as the most important factors in explaining longevity of mice.
Hence, this latter result further supports the proposition made in a
previous study (6)
that glucose intolerance of yet unknown
etiology probably occurs at a late age in these mice and plays an
important role in catalyzing pathological lesions in tissues.
|
ACKNOWLEDGMENTS |
|---|
-(carboxymethyl)lysine. |
FOOTNOTES |
|---|
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
-N-(2-furoylmethyl)-L-lysine (furosine) levels in rodent skin collagen. The relationship to dietary restriction and glycoxidation. Mech. Ageing Dev. 95,81-99[Medline]
-(carboxymethyl)lysine and pentosidine, in rat skin collagen. J. Gerontol. Biol. Sci. 50A,B337-B341[Abstract]
-(carboxymethyl)lysine and N-(carboxymethyl)-hydroxylysine in human skin collagen. Biochemistry 30,1205-1210[Medline]
(carboxymethyl)lysine, is a product of both lipid peroxidation and glycoxidation reactions. J. Biol. Chem. 271,9982-9986This article has been cited by other articles:
