Human Adenovirus Ad-36 Promotes Weight Gain in Male Rhesus and
Nikhil V. Dhurandhar,3Leah D. Whigham,* David H. Abbott,**‡Nancy J. Schultz-Darken,**
Barbara A. Israel,††Steven M. Bradley,* Joseph W. Kemnitz,**‡‡David B. Allison#
and Richard L. Atkinson*†
Department of Nutrition and Food Science and the Center for Molecular Medicine and Genetics, Wayne
State University, Detroit, MI; *Departments of Medicine and†Nutritional Sciences, **Wisconsin Regional
Primate Research Center,‡Department of Obstetrics and Gynecology,††Department of Pathobiological
Sciences,‡‡Department of Physiology of the University of Wisconsin, Madison, WI 53706;
and#Department of Biostatistics & Center for Research for Clinical Nutrition, University of Alabama
at Birmingham, Birmingham, AL
previously identified two viruses, SMAM-1, an avian adenovirus (Ad), and Ad-36, a human adenovirus, that produce
a syndrome of visceral obesity, with paradoxically decreased serum cholesterol and triglycerides in chickens and
mice. In the two studies presented in this paper, we used nonhuman primates to investigate the adiposity-
promoting potential of Ad-36. In study 1, we observed spontaneously occurring Ad-36 antibodies in 15 male rhesus
monkeys, and a significant longitudinal association of positive antibody status with weight gain and plasma
cholesterol lowering during the 18 mo after viral antibody appearance. In study 2, which was a randomized
controlled experiment, three male marmosets inoculated with Ad-36 had a threefold body weight gain, a greater
fat gain and lower serum cholesterol relative to baseline (P ?0.05) than three uninfected controls at 28 wk
postinoculation. These studies illustrate that the adiposity-promoting effect of Ad-36 occurs in two nonhuman
primate species and demonstrates the usefulness of nonhuman primates for further evaluation of Ad-36–induced
adiposity.J. Nutr. 132: 3155–3160, 2002.
Although obesity has multiple etiologies, an overlooked possibility is an infectious origin. We
● cholesterol ● adiposity ● obesity ● nonhuman primates ● infection
Obesity has multiple etiologies, but infectious agents have
been consistently overlooked as a possible origin of human
obesity. Three animal viruses have been reported to cause
obesity in nonprimate species, but have not been implicated in
initiating or maintaining obesity in humans (1–3). We have
now identified two additional viruses, SMAM-1, an avian
adenovirus (Ad), and Ad-36, a human adenovirus, that pro-
duce obesity in animals (4–7). In 6 separate experiments, we
have shown that these two adenoviruses produce a syndrome
of visceral obesity, along with paradoxically decreased serum
cholesterol and triglycerides in chickens and mice (4–8). A
capillary electrophoresis assay designed to detect Ad-36 DNA
(9) showed tropism of the virus for adipose tissue of the
infected animals (6,7). Ad-36–induced adiposity in animals
was hyperplastic and hypertrophic (10). Preliminary data
showed a marked up-regulation of adipocyte differentiation
induced by Ad-36 (11,12). In humans, serum antibodies to
both SMAM-1 and to Ad-36 are associated with obesity and
lower cholesterol and triglycerides levels (5,13). For ethical
reasons, the definitive experiment of injecting humans with
Ad-36 to determine a causal role for this virus in human
obesity is not possible. Nonhuman primates are the best sur-
rogates for human experiments. In this paper, we used two
disparate nonhuman primate species as models in which to
study the adiposity promoting potential of Ad-36.
MATERIALS AND METHODS
Study 1: spontaneously occurring antibodies to Ad-36 in rhesus
Adult male rhesus monkeys (Macaca mulatta) were
screened for the presence of spontaneously occurring antibodies to
Ad-36, to ascertain their association with longitudinal changes in
body weight and cholesterol. Frozen plasma samples from adult male
rhesus monkeys (n ? 15) were obtained from the Wisconsin Regional
1Presented in part at: 1) Experimental Biology ’99 April 1999, Washington,
DC [Dhurandhar, N. V., Bradley, Kemnitz, J. W. & Atkinson, R. L.
tibodies to human adenovirus Ad-36 are associated with body weight changes in
monkeys. FASEB J. 13: A 369 (abs.)]; 2) The European Congress on Obesity, May
2000, Vienna, Austria (Atkinson, R. L., Dhurandhar, N. V., Abbott, D. H. & Darken,
N.(2000)Weight gain and reduced serum lipids in non-human primates due to
a human virus. Int. J. Obes. (suppl. 1): S39 (abs.)]; and 3) Experimental Biology
’01, April 2001, Orlando, FL [Whigham, L. D., Dhurandhar, N. V., Abbott, D. H.,
Schultz-Darken, N., Israel, B. A., Kolesar, J., Strasheim, A. &Atkinson, R. L.
(2001)Presence of obesity-associated human adenovirus-36 DNA in tissues of
marmosets and humans. FASEB J. 15: A300 (abs.)].
2Supported by funds from the Wisconsin Alumni Research Foundation’s
Beers-Murphy Clinical Nutrition Center, University of Wisconsin Medical School
Research Committee and the William Hardy Endowment for Obesity Research of
Wayne State University.
3To whom correspondence should be addressed.
0022-3166/02 $3.00 © 2002 American Society for Nutritional Sciences.
Manuscript received 2 May 2002. Initial review completed 19 May 2002. Revision accepted 20 June 2002.
by guest on December 31, 2011
Primate Research Center (WRPRC),4Madison, WI. Blood samples
and body weight measurements were collected every 6 mo for 90 mo.
Monkeys were offered once daily a purified diet (#85387, Teklad,
Madison, WI) supplemented daily with a piece of fresh fruit. The
composition of the diet is described in Table 1. Energy value of the
diet eaten was ?167 kJ [700 kcal/(monkey ? d)]. Monkeys consumed
water ad libitum. All monkeys were between 8 and 14 y of age when
the first plasma sample available for this study was drawn (baseline
sample). Plasma samples were used to determine total cholesterol as
well as antibodies to Ad-36, using the neutralization assay described
below. For each monkey, the time of first appearance of Ad-36
antibodies, as well as body weight and cholesterol were determined
for 18 mo before and after the first antibody appearance.
Study 2: infection of marmosets with Ad-36. In a randomized
experiment, adult male common marmosets (Callithrix jacchus) were
inoculated with Ad-36 to investigate prospectively the adiposity-
promoting effect of the virus. Ad-36 antibody-free adult male mar-
mosets (n ? 6, age range 2–6 y) were obtained from the WRPRC and
were divided into two weight- and age-matched groups (each n ? 3)
that were individually housed in two separate rooms. Marmosets were
fed once daily at 1230–1400 h ?20 g of a specialized diet (Table 2
Zu Preem Marmoset Diet, Premium Nutritional Products Topeka,
KS). Two small chunks of fruit (varied daily, ?10 g) were added to
provide variety and ?5 mL of yogurt was spread on top, providing
supplements of vitamin C, cholecalciferol, and calcium. Based on our
previous experience, 20 g of the diet is slightly more that the amount
consumed by marmosets in a day. To verify ad libitum consumption,
it was noted that some food remained in the cage each day. Monkeys
consumed water ad libitum. After 4 wk of acclimatization, marmosets
were anesthetized with ketamine-xylazine [10 and 0.5 mg/kg, respec-
tively, intramuscular (i.m.)] and inoculated intranasally (i.n.) with
tissue culture media containing either Ad-36 virus (5 ? 105plaque
forming units; Ad-36 group) or no virus (control group). Blood
samples were drawn before inoculation (baseline) and also 10, 17 and
28 wk postinoculation and used for antibody screening (using the
serum neutralization method described below) and for determination
of serum cholesterol and triglycerides. Using the stable isotope dilu-
tion method described below, total body fat was determined at base-
line and at 28 wk postinoculation, when the experiment was termi-
nated and the monkeys were killed by administration of an i.m.
injection of 10 mg/kg ketamine and 0.5mg/kg xylazine followed by an
intravenous (i.v.) injection of 100 mg/kg pentobarbital. Visceral fat
(intraperitoneal fat) was carefully removed from each carcass and
weighed. Samples (?1 g each) of the visceral fat, liver, skeletal
muscle, lung and brain of all monkeys were flash-frozen in liquid
nitrogen to screen for Ad-36 DNA using a nested polymerase chain
reaction (PCR) assay.
Animal care. Rhesus and marmoset procedures were approved by
the Institutional Animal Care and Use Committee (IACUC) of the
University of Wisconsin at Madison. Monkeys were housed individ-
ually in cages that allowed auditory, visual and olfactory contact. The
Ad-36 group and the uninfected control marmosets were housed in
two separate rooms under biosafety level 2 containment.
Tissue culture techniques. A549 cells (human lung carcinoma
cells) obtained from American Type Culture Collection (ATCC,
Manassas, VA) were used to grow Ad-36. Minimum essential medium
Eagle (MEM) (Cat # M-0643, Sigma Chemical, St. Louis, MO) with
nonessential amino acids, Earle’s salts, L-glutamine, 10% fetal bovine
serum and 2.9% NaHCO3(v/v), pH 7.4, was used for growing the
Virus growth. Ad-36 was obtained from the ATCC. The virus
was plaque purified as previously described (6,7) and grown in A549
cells. A working stock of virus was prepared as previously described
(6,7) and was titrated on A549 cells, divided into aliquots and stored
Serum neutralization test for detecting antibodies.
plasma and marmoset serum were tested for the presence of Ad-36
antibodies. The assay used A549 cells and was conducted as a con-
stant-virus-decreasing-serum method, as previously described (6,7).
The absence of cytopathic effect (CPE) of the virus on A549 cells in
the presence of the test serum is considered an indication of effective
neutralization of the virus and the serum is considered to have
antibodies against the virus. Samples were considered antibody pos-
itive if the serum titer was ?1:8. A few rhesus samples demonstrated
cell-toxicity up to 1:16 dilutions. For these samples, a more stringent
criterion of titer (?1:32) was used to denote antibody positivity.
Fasting total cholesterol was determined in
duplicate with a cholesterol-oxidase-peroxidase method (Cat # 352–
500P; Sigma Chemical) using 10 ?L of serum.
Development of a nested PCR assay for detection of Ad-36
DNA. Four primers were designed to unique regions of the Ad-36
fiber protein gene for use in nested PCR detection of viral DNA.
Sequences of primers were as follows: outer forward primer (5?-
GTCTGGAAAACTGAGTGTGGATA), outer reverse primer (5?-
ATCCAAAATCAAATGTAATAGAGT), inner forward primer
(5?-TTAACTGGAAAAGGAATAGGTA), inner reverse primer
(5?- GGTGTTGTTGGTTGGCTTAGGATA). DNA was isolated
using a QIAamp Tissue Kit (Cat #29304; Qiagen, Valencia, CA).
Negative PCR controls were water and DNA from uninfected A-549
cells. Positive PCR control was DNA from Ad-36 infected A-549
cells. DNA was denatured for 2 min at 95°C and subjected to 35
cycles of PCR (94°C for 1 min, 55°C for 1 min, 72°C for 2 min)
followed by incubation at 72°C for 5 min. PCR products were
visualized on a 1% agarose gel with a size marker. Nested PCR
products from positive control (Ad-36 DNA) and infected marmoset
brain tissue were sequenced to confirm amplification of targeted
region of the gene.
Adenovirus infection in marmosets was con-
firmed by collecting fecal samples (wk 4 and 9 after Ad-36 inocula-
tion) and growing the virus on cell cultures as described earlier (7).
4Abbreviations used: Ad, adenovirus; ATCC, American Type Culture Collec-
tion; CPE, cytopathic effect; IACUC, Institutional Animal Care and Use Commit-
tee; i.m., intramuscular; i.n., intranasal; i.v., intravenous; MEM, minimum essential
media Eagle; PCR, polymerase chain reaction; WRPRC, Wisconsin Regional
Primate Research Center.
Composition of diet offered to rhesus monkeys1,2
Energy, kJ (kcal)
Crude fat, g
Crude fiber, g
1Minerals: contains calcium, 7.9 g; phosphorus, 4.9 g; potassium,
3.6 g; magnesium, 1.2 g; sodium, 2.1 g; chloride 3.2 g; iodine, 2 mg;
copper 10 mg; selenium, 1 mg; chromium, 2 mg; iron, 276 mg; zinc, 31
mg; manganese, 62 mg.
2Vitamins: contains riboflavin, 22 mg; niacin, 99 mg; pantothenic
acid, 55 mg; folic acid, 2 mg; thiamin, 19 mg; biotin, 4 mg; choline, 1259
mg; cholicalciferol, 34 mg; vitamin A, 40 mg; vitamin K, 50 mg; vitamin
E, 227 mg; vitamin B-12, 30 ?g; vitamin C, 1017 mg.
Composition of the canned diet offered to marmosets1,2
1Minerals: contains calcium, 3.3 g; phosphorus, 2.4 g; sodium,
2.2 g; potassium, 3.3 g; magnesium, 0.5 g; iron, 54 mg; zinc, 56 mg;
copper, 5.3 mg; iodine, 0.8 mg; manganese, 7.4 mg.
2Vitamins: contains vitamin A, 42 mg; cholicalciferol, 0.235 mg;
vitamin E, 94 mg; thiamine, 40.0 mg; riboflavin, 8.0 mg; pyridoxine, 5.2
mg; niacin, 55.0 mg; pantothenic acid, 25.0 mg; biotin, 0.28 mg; folic
acid, 0.2 mg; choline, 587 mg.
DHURANDHAR ET AL.
by guest on December 31, 2011
Body composition analysis by stable isotope dilution method.
Blood (1 mL) was drawn followed by i.v. administration of 0.1 g/kg
deuterium oxide. The dose was determined by weighing the syringe
before and after administration. A second blood sample was obtained
2 h after the dose. Equilibration time for the stable isotope is ?60–75
min for rhesus monkeys; therefore, 2 h was considered an adequate
equilibration period for the marmosets. Isotope dilution spaces were
determined by forcing the serum through a 100,000-Da exclusion
filter and distilling and reducing over zinc for mass spectrometric
determination of the isotopic abundance. Dilution space was calcu-
lated from the isotopic enrichment of the 2-h sample relative to the
predose sample. Fat-free mass was calculated from total body water,
assuming a hydration ratio of 0.732. Body fat was calculated by
subtracting fat-free mass from the total body mass.
Study 1. Multivariate (O’Brien) test. Before conducting univar-
iate tests and to increase power, a multivariate test, in which the
dependent variable Y, was defined as Y ? (ZWeight ? ZCholesterol)
was used (14). When Y was regressed on Ad-36 status, monkey, and
the polynomials of age in the mixed effects model, the effect of Ad-36
status was significant (F ? 8.14; df ? 1,96; P ? 0.005). Thus, there
was a clear effect of Ad-36 status on combination of weight and
cholesterol. We then examined each dependent variable separately.
Univariate tests. Univariate tests were conducted only when the
multivariate test was significant. Each dependent variable (weight
and cholesterol) was regressed on Ad-36 antibody status, monkey and
polynomials of age in a mixed-effect model.
Study 2. Groups were compared using Student’s t test. Values in
the text are means ? SD.
Study 1: spontaneously occurring antibodies to Ad-36 in
rhesus monkeys. During the 90-mo period, all 15 monkeys
showed Ad-36 antibodies at some point in time. Because blood
samples were obtained at 6-mo intervals, the exact date of
seroconversion during the preceding 6 mo of the first positive
sample is unknown. The appearance of antibodies in the
plasma did not demonstrate an annual pattern. Of 15 mon-
keys, 8 were seronegative at baseline. Only these monkeys
were included in the statistical analysis to compare the body
weights and cholesterol levels before and after the first appear-
ance of Ad-36 antibodies. The remaining 7 monkeys were
seropositive for Ad-36 antibodies at baseline and were ex-
cluded from statistical analysis.
Before separate univariate tests were conducted, a multi-
variate test (see Methods) was conducted that indicated a
clear effect of Ad-36 status on the combination of weight and
cholesterol (P ? 0.005).
The monkeys were completely free of antibodies from ?18
to ?6 mo before the first positive sample; this was designated
as the “baseline” period, whereas the period from ?6 to ?18
mo after seroconversion was designated as the “postinfection
period.” Body weight and plasma cholesterol level changes for
18 mo before and 18 mo after the first appearance of Ad-36
antibody are presented in Figures 1 and 2. Time point “0 mo”
denotes the first antibody positive serum sample for each
Body weight changed little during the baseline
period, decreasing by ?0.3% (0.04 ? 1.5 kg; P ? 0.87). In
contrast, the body weight increased by ?10% in 6 mo and by
?15% (1.7 ? 0.8 kg, P ? 0.03) at 18 mo during the “postin-
fection” period (Fig. 1). In the full statistical model, the effect
of Ad-36 antibody status was significantly associated with
weight gain (F ? 5.47; df ? 1,96; P ? 0.021). The parameter
estimate for the effect of Ad-36 antibody status was 0.81 kg,
indicating that the generation of Ad-36 antibodies was asso-
ciated with an increase in body weight of 0.81 kg.
Cholesterol. Plasma cholesterol levels were stable during
the baseline period, but decreased by ?23% (P ? 0.06) during
the 6 mo immediately after the appearance of Ad-36 antibod-
ies and remained low for at least 18 mo during the postinfec-
tion period (Fig. 2). In the full statistical model, the effect of
Ad-36 antibody status was significant (F ? 5.33; df ? 1,96; P
? 0.023). The parameter estimate for the effect of Ad-36
antibody status was 0.49 mmol/L, indicating that the genera-
tion of Ad-36 antibodies was associated with a decrease in
weight change in rhesus monkeys. Values are mean ? SD; n ? 8.
Plasma samples from adult male rhesus monkeys were collected every
6 mo for 90 mo and tested for antibodies to Ad-36. Body weights are
plotted for the 18 mo before the onset of Ad-36 antibodies (?18, ?12
and ?6 mo) and after the monkeys became antibody positive (0, 6, 12
and 18 mo). *Different from the body weight at ?6 mo (the last anti-
body-free value), P ? 0.05.
Adenovirus (Ad)-36 antibody appearance and body
cholesterol change in rhesus monkeys. Values are mean ? SD, n ? 8.
Plasma samples from adult male rhesus monkeys were collected every
6 mo for 90 mo and tested for antibodies to Ad-36. Plasma cholesterol
values are plotted for the 18 mo before the onset of Ad-36 antibodies
(?18, ?12 and ?6 mo) and after the monkeys became antibody pos-
itive (0, 6, 12 and 18 mo). *Different from the body weight at ?6 mo (the
last antibody-free value), P ? 0.05.
Adenovirus (Ad)-36 antibody appearance and plasma
WEIGHT GAIN IN MONKEYS DUE TO A HUMAN VIRUS
by guest on December 31, 2011
plasma levels of cholesterol by 0.49 mmol/L. Thus, when
controlled for age, and after the first appearance of Ad-36
antibody, the rhesus monkeys had greater body weights and
lower cholesterol levels.
Study 2: infection of marmosets with Ad-36. Body weights
of both groups did not differ at the time of inoculation (336.4
? 19.8 g vs. 346.3 ? 27.8 g, P ? 0.65; for the Control and the
Ad-36–infected groups, respectively). The serum neutraliza-
tion assay showed an absence of Ad-36 antibodies in the
control group at all times during the experiment. Two of the
three Ad-36–inoculated marmosets were antibody positive at
both 10 and 17 wk postinoculation. Ad-36 could not be
isolated from the fecal samples of the control group at any time
during the study. However, the infective virus was isolated at
both 4 and 9 wk postinoculation from the fecal samples of the
two Ad-36 inoculated marmosets that had detectable Ad-36
antibodies. The nested PCR assay, however, detected Ad-36
DNA in the adipose tissue, liver, skeletal muscle, lung and the
brain samples of all three Ad-36 inoculated marmosets, but
not in any monkey in the control group. The nested PCR assay
results for the adipose tissue of the two groups are presented in
Figure 3. Isolation of Ad-36 virus from fecal samples and the
presence of viral DNA in the tissue of all infected males demon-
strated that marmosets were readily infected with Ad-36.
At 28 wk after inoculation, the Ad-36 group had greater
body weight gain (41.4 ? 11.2 g vs. 10.8 ? 13.4 g, P ? 0.039;
Fig. 4) and total body fat gain (36.3 ? 6.1 g vs. 23.0 ? 3.0 g,
P ? 0.013) than controls. The mean weight gain of 41 g
represented a 12% increase in the Ad-36 group compared with
a 3.2% increase in the control group, a threefold difference.
The Ad-36 group tended to have more visceral fat than
controls (7.3 ? 2.5 g vs. 4.4 ? 0.9 g, P ? 0.089). Despite the
gain in fat mass and body weight, serum cholesterol levels were
lower in the Ad-36 group than in controls (change from
baseline: ?0.79 ? 0.28 mmol/L vs. 0.10 ? 0.08 mmol/L, P
? 0.006; Fig. 5). The change in serum triglycerides from
baseline tended to be greater in infected marmosets that in
controls (?0.52 ? 0.23 mmol/L vs. 0.14 ? 0.53 mmol/L, P
Although 50 types of human adenoviruses are deposited
with the ATCC, Ad-36 is the first reported to cause obesity in
animals (6,7). Ad-36 is serologically different from at least 47
of the other 49 human adenoviruses (15–22); it was first
isolated in 1978 in Germany from the feces of a diabetic girl
suffering from enteritis (17). Relative antigenic uniqueness
was one of the main reasons for selecting Ad-36 to test for the
The concept of virus-induced obesity is of greater impor-
tance if shown to be relevant to human obesity. Although we
have demonstrated the adipogenic and hypolipidemic effects
of a human virus in animals such as chickens and mice, we can
not conclusively extrapolate the results, without verification,
to human obesity. Differences in lipid metabolism between
lower animals and primates preclude such a direct extrapola-
tion. For instance, the energy metabolism of chickens is based
on free fatty acids, not glucose, and insulin is of minor impor-
tance. Such metabolic differences clearly warranted the use of
a higher model to establish the relevance of the findings to
human obesity. In addition, because of ethical reasons, hu-
mans can not be infected experimentally with Ad-36 to verify
its adipogenic effect directly. Therefore, we decided to use
nonhuman primate species as the best way of determining the
relevance of Ad-36 in human obesity.
rus (Ad)-36 DNA detection in the marmoset adipose tissue (n ? 6). Key:
St: DNA ladder, 1: Ad-36 DNA positive control, 2–4: adipose tissue
DNA from the uninfected control marmosets, 5–7: adipose tissue DNA
from the Ad-36 infected marmosets.
Nested polymerase chain reaction assay for adenovi-
inoculation with adenovirus (Ad)-36. Values are means ? SD, n ? 3 per
group. *Different from control, P ? 0.05.
Cumulative body weight gains in marmosets after
sets after inoculation with Ad-36. Values are means ? SD, n ? 3 per
group. *Different from control, P ? 0.05.
Cumulative changes in serum cholesterol in marmo-
DHURANDHAR ET AL.
by guest on December 31, 2011
In our rhesus monkey study, the fact that all of the 15
monkeys had Ad-36 antibodies at some time during the 90-mo
period suggests a covert epidemic of Ad-36 infection in the
rhesus monkey colony at WRPRC. The source of Ad-36
infection in monkeys is not yet known. However, transmission
from humans is a possibility. Although most adenoviruses are
species specific in their replication cycle, we have recovered
infectious Ad-36 virus from Ad-36–infected chickens (6,7)
and have detected naturally occurring antibodies to Ad-36 in
chickens as well as rats (unpublished data). The ability of
Ad-36 to infect and replicate in widely disparate vertebrate
species is in itself unusual. Therefore, the presence of naturally
occurring antibodies in rhesus monkeys to a human adenovirus
is not very surprising.
Serum neutralization is considered the “gold standard” to
specifically detect neutralizing antibodies. In addition to the
previously reported minimal antigenic cross-reactivity of
Ad-36 (16,–22), we confirmed that Ad-36 does not cross-react
with other human adenoviruses such as Ad-2, Ad-31 or Ad-37
(unpublished data). Therefore, it is unlikely that the antibod-
ies detected in the rhesus monkey plasma that neutralized
Ad-36 were antibodies to other human adenoviruses. Anti-
bodies to simian adenovirus(es) that may cross-react with
Ad-36 are a possibility, although cross-reactivity between a
simian adenovirus and Ad-36 has not been reported.
In rhesus monkeys, a positive correlation of age with body
weight and serum cholesterol would have been expected. The
increase in body weight and drop in cholesterol during the
postinfection period that persisted even after controlling for
age supports our hypothesis that these changes were associated
with the appearance of Ad-36 antibodies.
A comparison of body weight and plasma cholesterol
changes between antibody positive and antibody-free monkeys
would have been optimal to eliminate the effect of age on
these variables. However, all 15 rhesus monkeys developed
Ad-36 antibodies at different times during the study period,
thus precluding such a comparison. Nevertheless, the follow-
ing measures ensure that the changes in body weight and
cholesterol observed were not age related.
The age of the monkeys ranged from 8 to 14 y. The age of
the first appearance of antibodies varied among the monkeys.
Regardless of the age of seroconversion, there was little change
in body weights or plasma cholesterol in the 18 mo before the
first appearance of antibodies. Thus, although the period be-
fore the onset of antibodies represents different ages for indi-
vidual monkeys, their body weights were stable during this
period. Such a stabilization of body weight is expected in adult
rhesus monkeys who have completed their growth. Also, an
age-related decline in plasma cholesterol is not expected in
freely fed rhesus monkeys. In effect, each monkey acted as his
own control for body weight and cholesterol comparisons
before and after the seroconversion. As stated earlier, we
further ascertained the age effect by controlling statistically for
age, and found a clear effect of seroconversion on body weight
and cholesterol that was independent of age. These measures
demonstrate that the observed changes were not age related.
In the marmoset experiment, the persistence of Ad-36
DNA in various tissues of all marmosets 6 mo postinfection is
an important finding, demonstrating the spread of Ad-36 in
their bodies, even though one marmoset failed to generate
detectable levels of antibodies. In addition, the presence of
Ad-36 DNA in the adipose tissue is particularly intriguing
because our preliminary data showed Ad-36 induced up-regu-
lation of 3T3-L1 cell (rodent embryonic preadipocytes) differ-
entiation (11,12). Future experiments should investigate the
Ad-36–induced up-regulation of fat cell differentiation as a
possible mechanism of Ad-36–induced obesity in monkeys.
Similarly, Ad-36–induced hypothalamic damage is another
possible mechanism that should be investigated. Canine dis-
temper virus may promote obesity in mice by inducing hypo-
thalamic damage (1). Although we did not find histopatho-
logical hypothalamic lesions in Ad-36–infected mice in an
earlier study (6), the presence of Ad-36 DNA in the brains of
the infected group warrants such an investigation in marmo-
Considering the very small starting body weights of the
marmosets, the mean postinoculation weight gain of 41g rep-
resents a 12% increase in the Ad-36 group, which is approx-
imately threefold greater than the 3.2% mean weight gain in
the control group. Also, the Ad-36 group had an 11.5% total
fat gain vs. 6.8% in the control group and 66% more visceral
fat than the controls. Given the very small size of marmosets,
these findings are of major biological importance.
The data presented in this paper suggest that Ad-36 plays a
role in increasing body weight in nonhuman primates such as
rhesus monkeys and marmosets; either animal could be used as
a surrogate model with which to study the role of Ad-36 in
human obesity. However, as a model, marmosets appear to
have some advantages over rhesus monkeys because they are
smaller, easier to house and handle and less expensive. In
addition, in our initial screening of monkeys from the
WRPRC colony for Ad-36 antibodies, ? 4% of the marmosets
were antibody positive. This very low prevalence of Ad-36
antibodies in marmosets gives a much better chance to select
antibody-free marmosets for a prospective study (rhesus mon-
keys from the WRPRC had much greater and frequent contact
with humans, whereas the marmosets are housed in a more
controlled environment to prevent inadvertent infections).
Additionally, the fact that marmosets have a life-expectancy
of ?8 y (vs. 30–40 y for rhesus monkeys) means that in future
prospective studies, marmosets could be tracked relatively eas-
ily during their complete life-cycle to ascertain the long-term
consequences of Ad-36 infection.
A survey from three different states in the United States
showed a 30% prevalence of Ad-36 antibodies in obese but
only a 5% prevalence of the antibodies in nonobese subjects
(13). The antibody-positive obese subjects also had signifi-
cantly lower serum cholesterol levels (13). Body weight gain
and hypocholesterolemia in response to Ad-36 infection in
two divergent primate species support the hypothesis that the
virus may play a causative role in human adiposity and dem-
onstrate the suitability of the two animal models for further of
the phenomenon. In addition, the relationship of Ad-36 with
body composition and serum lipids in these models suggests
that in the future, investigators studying body weight, obesity,
obesity treatment or evaluation of serum lipids in rhesus mon-
keys or marmosets would be wise to take into account the
status of infection with Ad-36 because Ad-36 may add extra-
neous variance to the outcome.
We gratefully acknowledge the animal care staff of the University
of Wisconsin-Madison Medical School animal facility and WRPRC
for maintaining of the monkeys; and Darrel Florence, DVM, and the
veterinary staff of WRPRC for clinical care.
1. Lyons, M. J., Faust, I. M., Hemmes, R. B., Buskirk, D. R., Hirsch, J. &
Zabriskie, J. B.(1982) A virally induced obesity syndrome in mice. Science
(Wash., DC) 216: 82–85.
2. Carter, J. K., Ow, C. L. & Smith, R. E. (1983) Rous-associated virus
WEIGHT GAIN IN MONKEYS DUE TO A HUMAN VIRUS
by guest on December 31, 2011
type 7 induces a syndrome in chickens characterized by stunting and obesity.
Infect. Immun. 39: 410–422.
3. Gosztonyi, G. & Ludwig, H.(1995)
pathogenesis. Curr. Top. Microbiol. Immunol. 190: 39–73.
4. Dhurandhar, N. V., Kulkarni, P. R., Ajinkya, S. M. & Sherikar, A. A.
(1992) Effect of adenovirus infection on adiposity in chickens. Vet. Microbiol. 31:
5. Dhurandhar, N. V., Kulkarni, P. R., Ajinkya, S. M., Sherikar, A. A. &
Atkinson, R. L.(1997) Screening of human sera for antibody against avian
adenovirus. Obes. Res. 5: 464–469.
6. Dhurandhar, N. V., Israel, B. A., Kolesar, J. M., Mayhew, G. F., Cook, M. E.
& Atkinson, R. L.(2000) Adiposity in animals due to a human virus. Int. J. Obes.
7. Dhurandhar, N. V., Israel, B. A., Kolesar, J. M., Mayhew, G. F., Cook, M. E.
& Atkinson, R. L. (2001) Transmissibility of adenovirus-induced adiposity in a
chicken model. Int. J. Obes. 25: 990–996.
8. Dhurandhar, N. V. & Atkinson, R. L.
Opin. Endocrinol. Diabetes 7: 247–251.
9. Kolesar, J. M., Miller, J. A., Dhurandhar, N. V. & Atkinson, R. L.
(2000) Direct quantification of AD-36 adenovirus DNA by capillary electrophore-
sis with laser-induced fluorescence. J. Chromatogr. B. 744: 1–8.
10. Atkinson, R. L. & Dhurandhar, N. V.
new thoughts on the etiology of obesity. In: Adipocyte Biology and Signaling.
(Ntambi, J. T., ed.), pp. 165–174. IOS Press, Amsterdam, The Netherlands.
11. Dhurandhar, N. V., Sheele, J., Israel, B. A. & Atkinson, R. L.
enovirus that induces adiposity in animals also influences in-vitro differentiation of
preadipocytes. Obes. Res 7: 39S.
12. Dhurandhar, N. V., Israel, B. A., Kolesar, J. M. & Atkinson, R. L.
(2000) Pathophysiology of the human adenovirus that induces adiposity in ani-
mals. FASEB J. 14: A732 (abs.).
Borna disease: neuropathology and
(2000) Viruses and obesity. Curr.
(2000) From cells to organisms:
13. Atkinson, R. L., Dhurandhar, N. V., Allison, D. B., Bowen, R. & Israel, B. A.
(1998) Evidence for an association of an obesity virus with human obesity at
three sites in the United States. Int. J. Obes. 22: S57 (abs.).
14. O’Brien, P. C.(1984) Procedures for comparing samples with multiple
endpoints. Biometrics 40: 1079–1087.
15. Pereira, H. G., Huebner, R. J., Ginsberg, H. S. & Van Der Veen, J.
(1963) A short description of the adenovirus group. Virology 20: 613–620.
16. Rubin, B. A.(1993) Clinical picture and epidemiology of adenovirus
infections (a review). Acta Microbiol. Hung. 40: 303–323. 20.
17. Wigand, R. H., Gelderblom, H. & Wadell, G.
novirus (candidate adenovirus 36), a novel member of subgroup D. Arch. Virol. 64:
18. De Jong, J. C., Bijlsma, K., Wermenbol, A. G., Verweij-Uijterwaal, M. W.,
Van Der Avoort,H. G., Wood, D.
A. D. (1993) Detection, typing and subtyping of enteric adenoviruses 40 and 41
from fecal samples and observation of changing incidences of infections with
types and subtypes. J. Clin. Microbiol. 31: 1562–1569.
19. De Jong, J. C., Wigand, R., Wadell, G., Keller, D., Muzerie, C. J., Wer-
menbol, A. G. & Schaap, G. J. (1981) Adenovirus 37: identification and charac-
terization of a medically important new adenovirus type of subgroup D. J. Med.
Virol. 7: 105–118.
20. De Jong, J. C., Wigand, R., Adrian, T. H, Hierholzer, J. C., Kapsenberg,
J. G., Muzerie, C. J. & Wermenbol, A. G.
adenovirus species of subgenus D. Intervirology 22: 164–169.
21. Wigand, R., Adrian, T. & Bricout, F.
subgenus D: candidate adenovirus type 42. Arch. Virol. 94: 283–286.
22. Hirholzer, J. C., Wigand, R., Anderson, L. J., Adrian, T. & Gold, J. W.
(1988) Adenoviruses from patients with AIDS: a plethora of serotypes and a
description of five new serotypes of subgenus D (Types 43–47). J. Infect. Dis.
(1980) New human ade-
J.,Bailey,A. S.& Osterhaus,
(1984) Adenovirus 38: a new human
(1987) A new human adenovirus of
DHURANDHAR ET AL.
by guest on December 31, 2011