Association of Adenovirus 36 Infection With Adiposity and Inflammatory-Related Markers in Children

Abstract

Context:

Although animal studies suggest that adenovirus 36 (Ad36) infection is linked to obesity and systemic inflammation, human data are scant and equivocal.

Objective:

Associations of Ad36 infection with total body adiposity and inflammatory-related markers were determined in 291 children aged 9–13 years (50% female, 49% black).

Design:

Fasting blood samples were measured for presence of Ad36-specific antibodies and TNF-α, IL-6, vascular endothelial growth factor (VEGF), and monocyte chemoattractant protein-1 (MCP-1). Fat mass and fat-free soft tissue mass were measured by dual-energy X-ray absorptiometry.

Results:

The overall prevalence of Ad36 seropositivity [Ad36(+)] was 42%. There was a higher percentage of Ad36(+) children in the highest tertiles of TNF-α and IL-6 compared with their respective middle and lowest tertiles (both P < .03). There was also a trend toward a higher prevalence of Ad36(+) children in the highest tertile of VEGF compared with tertiles 1 and 2 (P = .05). Multinomial logistic regression, adjusting for age, race, sex, and fat-free soft tissue mass, revealed that compared with children with the lowest TNF-α, IL-6, and VEGF levels (tertile 1), the adjusted odds ratios for Ad36(+) were 2.2 [95% confidence interval (CI) 1.2–4.0], 2.4 (95% CI 1.4–4.0), and 1.8 (95% CI 1.0–3.3), respectively, for those in the highest TNF-α, IL-6, and VEGF levels (tertile 3). No association was observed between Ad36(+) and greater levels of fat mass or MCP-1 (all P > .05).

Conclusions:

In children, our data suggest that Ad36(+) may be associated with biomarkers implicated in inflammation but not with greater levels of fat mass.

As pediatric obesity continues to be an emergent concern, research efforts to disentangle its etiology have considered novel influences. One such influence that may trigger obesity development is infection with human adenovirus 36 (Ad36) (1–5), a common upper respiratory infection transmitted through person-to-person contact (6) that elicits an antibody response ranging from days to years after exposure. In vitro, Ad36 has been shown to enhance adipocyte differentiation and lipid accumulation in human-derived cells (7–9), which prompted studies in children and adults (9–12). Indeed, children seropositive for Ad36 antibodies [Ad36(+)] have been shown to exhibit higher body mass index (BMI)-for-age percentile and waist circumference vs those who are seronegative [Ad36(−)] (10), although not all studies show consistent results (4, 12).

The discrepancies in the aforementioned studies can be attributed in part to differences in the populations examined and the study designs and instruments used. However, it is possible that systemic inflammation may explain the Ad36-adiposity link because it is an underlying characteristic of both viral infection and obesity (13–15). In vitro, Ad36 has been reported to induce cellular activation of an inflammatory pathway that signals the production of TNF-α, IL-6, vascular endothelial growth factor (VEGF), and monocyte chemoattractant protein-1 (MCP-1) (9, 16). In vivo, these inflammatory-related markers were recently shown to stimulate adipocyte and macrophage migration into adipose tissue of infected mice, a hallmark of obesity development (9, 17, 18). In addition, they have been implicated in the pathogenesis of cardiovascular diseases, diabetes mellitus, osteoporosis, and a host of adverse health outcomes (13).

To our knowledge, only one human study has investigated the association of Ad36 infection with adiposity and inflammatory-related markers. Na and Nam (9) found that Ad36(+) adults had higher serum MCP-1 vs those who were Ad36(−) (2.5-fold higher in the total sample and 3.5-fold higher in obese subjects). It is important to elucidate the Ad36-adiposity-inflammation link in youth because measures may be introduced during childhood to prevent obesity and related disorders over the long term. The primary objective of this study was to determine associations of the human Ad36 infection with total body adiposity and inflammatory-related markers in a cohort of otherwise healthy children.

Materials and Methods

Participants

The study included baseline data from 291 children who were part of a multisite vitamin D supplementation trial conducted at the University of Georgia (UGA; Athens, Georgia), Purdue University (PU; West Lafayette, Indiana) and Indiana University (IU; Indianapolis, Indiana) (19). At each testing site, two cohorts of children were enrolled in the study, the first from October through December 2009 and the second from October through December 2010. Inclusion criteria were as follows: 1) aged 9–12 years for females and aged 10–13 years for males; 2) white or black/African American race; and 3) sexual maturation stages 2 and 3, estimated using self-administered questionnaires for genitalia or breast development (20). Children were excluded if they were taking medications or had any medical condition known to affect growth, maturation, nutritional status, metabolism, or inflammatory response. The Institutional Review Board for Human Subjects at each testing site approved the study procedures. All participants and their parent/guardian provided informed assent and consent, respectively.

Anthropometry

Standing height was measured using a wall-mounted stadiometer to the nearest 0.1 cm. Body weight was measured using an electronic scale to the nearest 0.1 kg (21). BMI z-scores and BMI-for-age percentiles were calculated (22), with obesity defined as having a BMI-for-age greater than or equal to the 95th percentile.

Body composition

Fat mass, fat-free soft tissue (FFST) mass, and body fat percentage were assessed using dual-energy X-ray absorptiometry [DXA; Delphi-A; Hologic Inc (UGA); Lunar iDXA; GE Medical Instruments (PU); and Hologic Discovery-W (IU)]. The same technician at each site conducted scans and performed analyses using instrument-specific software and protocols. Cross-calibration was performed to facilitate comparability of data from each study site/scanner, as described by Warden et al (23). Briefly, the UGA and PU study sites were cross-calibrated by scanning 26 individuals on the Hologic Delphi-A scanner and a Lunar iDXA scanner, whereas the IU and PU study sites were cross-calibrated by scanning ten individuals on the Hologic Discovery-W and Lunar iDXA scanners. Regression formulae between data obtained at UGA and PU and IU and PU were derived for each variable and used to adjust data obtained at UGA and IU to PU values (23).

Biochemical analyses

Blood samples were collected from fasting participants for assessment of Ad36 seropositivity and to determine the concentrations of TNF-α, IL-6, VEGF, and MCP-1 in sera. An indirect ELISA was used to determine Ad36 seropositivity [ie, Ad36(+) or Ad36(−)] as described by Laing et al (24). For the indirect ELISA assay, 10 μg Ad36 in 100 μL PBS was added to each well and incubated at 37°C for 1 hour. Cells were washed and blotted, and then primary antibody sera (1:10 dilution) or control (PBS) was added to the wells and incubated at 37°C for 1 hour. Secondary antibody (goat antihuman Ig (HþL) whole molecule-alkaline phosphatase conjugated) diluted 1:500 was added to wells (100 mL/well) and incubated at 37°C for 1 hour. Wells were washed and 100 mL paranitrophenyl phosphate substrate per well was added and incubated at 37°C for 10–15 minutes. Absorbance was measured at an OD of 405/495 nm (24).

The Luminex xMAP system, a high-throughput, microsphere-based suspension array, was used with a MILLIPLEX MAP human cytokine/chemokine immunoassay (Millipore,) for the quantification of serum inflammatory-related markers, TNF-α, IL-6, VEGF, and MCP-1, according to the manufacturer's guidelines. Biochemical measures were conducted in a single laboratory at UGA in batch analysis. The assay was analyzed on a Luminex 200 instrument (Luminex Corp) using Luminex xPONENT version 3.1 software. An additional analysis was performed using MILLIPLEX Analyst (Millipore). The intra- and interassay coefficients of variation were 2.6% and 13.0% for TNF-α, 2.0% and 18.3% for IL-6, 3.7% and 10.4% for VEGF, and 1.5% and 7.9% for MCP-1, respectively.

Statistical analyses

Data were checked for outliers and normality with histograms and tests of skewness and kurtosis for normality. Because TNF-α, IL-6, VEGF, and MCP-1 had skewed distributions, they were log transformed (ie, TNF-α, VEGF, and MCP-1) or square root transformed (ie, IL-6) for analyses but backtransformed when we present the results in Table 1 for ease of interpretation. We used an ANOVA to compare age, sexual maturation stage, anthropometric, body composition, and the inflammatory-related variables between the Ad36(+) and Ad36(−) groups. Group differences in categorical variables were tested using χ² tests. Multinomial logistic regression was used to estimate odds ratios (OR) and 95% confidence intervals (CI) for the presence of Ad36(+) according to tertiles of fat mass and the inflammatory-related variables (ie, TNF-α, IL-6, VEGF, and MCP-1) after adjusting for age, sex, race, and FFST mass. Subsequent models for TNF-α, IL-6, VEGF, and MCP-1 were adjusted additionally for fat mass. All statistical analyses were conducted using SPSS software (version 21; IBM SPSS Statistics), and statistical significance was set at a value of P < .05.

Table 1.

Participant Characteristics

	Total Sample	Ad36(+) Group	Ad36(−) Group	P Valuea
n	291	122	169
Age, y	11.3 ± 1.2	11.4 ± 1.2	11.3 ± 1.3	.23
Sexual maturation stage	2.4 ± 0.5	2.4 ± 0.5	2.4 ± 0.5	.84
Female, %b	50	53	47	.27
Black, %b	49	41	54	.02
BMIc	20.7 ± 4.3	20.1 ± 4.0	21.3 ± 4.5	.02
Z-score	0.7 ± 1.0	0.5 ± 1.0	0.8 ± 1.1	.02
Percentile	69.1 ± 28.6	64.8 ± 27.4	72.1 ± 29.1	.03
Obesity, %b	21	13	27	.01
FFST mass, kg	30.7 ± 6.9	30.3 ± 7.2	30.9 ± 6.7	.50
Fat mass, kg	14.7 ± 7.3	14.0 ± 7.2	15.2 ± 7.4	.17
Body fat, %	30.9 ± 9.3	30.2 ± 9.2	31.4 ± 9.4	.28
TNF-α, pg/mL	9.9 ± 1.7	10.9 ± 1.8	9.1 ± 1.6	<.01
IL-6, pg/mL	2.8 ± 2.0	3.5 ± 2.3	2.3 ± 1.8	.03
VEGF, pg/mL	190.6 ± 2.6	239.8 ± 2.5	160.8 ± 2.6	<.01
MCP-1, pg/mL	428.4 ± 1.7	437.0 ± 1.7	419.9 ± 1.7	.47

Values are means ± SD or percentage.

^aTests of significance between groups were based on ANOVA.

^bTests of significance between groups were based on the χ² test of goodness of fit.

^cBMI is calculated as weight in kilograms divided by height in meters squared. Obesity indicates a BMI-for-age greater than or equal to the 95th percentile.

Results

Participant characteristics are presented in Table 1. The sample was composed of 291 children, aged 9–13 years, with 50% female, 49% black, 21% obese, and 42% Ad36(+). There were no differences in the prevalence of Ad36(+) by testing site (P = .17). No significant differences were observed between groups [Ad36(+) vs Ad36(−)] in age, sexual maturation stage, sex, FFST mass, fat mass, or body fat percentage. In the Ad36(+) group, there was a greater proportion of whites (59%) vs blacks (41%), and obese (13%) vs nonobese (87%) than in the Ad36(−) group (both P < .05). The Ad36(+) group had significantly lower BMI, BMI z-score, and BMI-for-age percentile and significantly higher serum concentrations of TNF-α, IL-6, and VEGF than the Ad36(−) group (all P < .05).

The prevalence and adjusted odds ratios for Ad36 seropositivity across tertiles of total body fat mass and serum concentrations of TNF-α, IL-6, VEGF, and MCP-1 are presented in Table 2. There were a higher percentage of children with Ad36 seropositivity in the highest tertiles of serum TNF-α and serum IL-6 compared with their respective middle and lowest tertiles (both P < .03). There was also a trend for a higher percentage of children with Ad36 seropositivity in the highest tertile of serum VEGF compared with the middle and lowest tertiles (P = .05). Multinomial logistic regression, adjusting for age, race, sex, and FFST mass, revealed that compared with children with the lowest serum TNF-α, IL-6, and VEGF levels (tertile 1), the adjusted ORs for Ad36(+) were 2.2 (95% CI 1.2–4.0), 2.4 (95% CI 1.4–4.0) and 1.8 (95% CI 1.0–3.3), respectively, for those in the highest serum TNF-α, IL-6, and VEGF levels (tertile 3). When fat mass was added as a covariate, these significant adjusted ORs for Ad36(+) persisted for TNF-α (OR 2.2; 95% CI 1.2–4.0), IL-6 (OR 2.4; 95% CI 1.4–4.1), and VEGF (OR 1.9; 95% CI 1.0–3.3)]. No association was observed between Ad36 seropositivity and greater levels of fat mass or increased serum MCP-1 concentrations (all P > .05).

Table 2.

Prevalence and Adjusted ORs (95% CI) for Ad36 Seropositivity Across Tertiles of Total Body Fat Mass and Serum Concentrations of TNF-α, IL-6, VEGF, and MCP-1 in Children Aged 9–13 Years

	Total Body Fat Massa			P Value
	Tertile 1 7.8 kg (<10.4 kg)	Tertile 2 13.3 kg (10.4–16.1 kg)	Tertile 3 22.0 kg (>16.1 kg)
n	98	97	96
Ad36 seropositivity
Prevalence, %	45	49	32	.06
Adjusted OR (95% CI)b	1.0	1.2 (0.6, 2.1)	0.6 (0.3, 1.2)c	.15

	Serum TNF-αa			P Value
	Tertile 1 6.2 pg/mL (<8.2 pg/mL)	Tertile 2 9.6 pg/mL (8.2–11.9 pg/mL)	Tertile 3 14.9 pg/mL (≥12.0 pg/mL)
n	96	96	99
Ad36 seropositivity
Prevalence, %	31	44	51	.02
Adjusted OR (95% CI)b	1.0	1.7 (0.9, 3.2)	2.2 (1.2, 4.0)c	.01

	Serum IL-6a			P Value
	Tertile 1 0.9 pg/ml (<1.0 pg/ml)	Tertile 2 1.3 pg/ml (1.1–1.5 pg/ml)	Tertile 3 6.6 pg/ml (≥1.6 pg/ml)
n	97	95	99
Ad36 seropositivity
Prevalence, %	35	50	55	<.01
Adjusted OR (95% CI)b	1.0	1.6 (0.2, 12.1)	2.4 (1.4, 4.0)c	<.01

	Serum VEGFa			P Value
	Tertile 1 82 pg/ml (<136 pg/ml)	Tertile 2 195 pg/ml (137–313 pg/ml)	Tertile 3 463 pg/ml (≥313 pg/ml)
n	98	94	99
Ad36 seropositivity
Prevalence, %	33	44	49	.05
Adjusted OR (95% CI)b	1.0	1.6 (0.9, 2.8)	1.8 (1.0, 3.3)c	.04

	Serum MCP-1a			P Value
	Tertile 1 264 pg/ml (<332 pg/ml)	Tertile 2 419 pg/ml (334–536 pg/ml)	Tertile 3 694 pg/ml (≥536 pg/ml)
n	95	97	99
Ad36 seropositivity
Prevalence, %	39	41	45	.65
Adjusted OR (95% CI)b	1.0	0.8 (0.4, 1.5)	0.8 (0.4, 1.6)	.53

^aValues are median (range) in a given tertile.

^bData are adjusted for age, race, sex, and FFST mass.

^cSignificantly different from the reference group, tertile 1 (P < .05).

Discussion

To our knowledge, this is the first pediatric study to investigate associations of the human Ad36 infection with a robust measurement of adiposity and with serum markers implicated in inflammation. We did not observe a relationship between Ad36 infection and total body fat mass after controlling for age, sex, race, and FFST mass. On the other hand, Ad36 infection was associated with several biomarkers implicated in inflammation, including TNF-α, IL-6, and VEGF. These relationships were independent of potential confounding factors, such as age, sex, race, FFST mass, and fat mass.

The overall prevalence of Ad36(+) in our sample was 42%, which is similar to what has been reported in other pediatric studies (2, 10, 11). We found a higher prevalence of Ad36(+) in white vs black children, and although earlier adenovirus research has also shown differences by racial/ethnic group, findings are inconsistent. For example, although Gabbert et al (10) reported that the prevalence of Ad36(+) was not different by race in 124 children, Broderick et al (25) showed that Ad36(+) was significantly higher in black vs white adults. Although determining the association of Ad36(+) with adiposity and inflammatory-related markers by race was not an objective of this study, future work in this area is warranted.

Several studies have investigated the association of Ad36(+) with adiposity in children, yet the findings have also been disparate. Contrasting results may be related to differences in target populations and measures of body fatness, with findings in support of the adenovirus-adiposity link largely derived from obese cohorts and studies that rely on BMI-based outcomes (1, 10, 26). For example, in a cohort of 84 obese children, Atkinson et al (1) found that mean BMI z-score was higher in Ad36(+) vs Ad36(−) groups. In another sample of 157 children, Parra-Rojas et al (26) found no significant differences in mean BMI by Ad36 status but did observe a higher prevalence of Ad36(+) in obese vs nonobese participants (defined as BMI for age ≥95th percentile). Conversely, we observed a lower prevalence of Ad36(+) in obese vs nonobese children. Although it is unclear why participants with a BMI-for-age greater than or equal to the 95th percentile were less likely to be classified as Ad36(+), it is important to note that there were no significant differences in fat mass or body fat percentage by Ad36 status, and such robust measures of adiposity were not assessed in the previous studies. Our findings are in agreement with Na et al (11), who also reported no significant differences in body fat percentage between Ad36(+) and Ad36(−) groups in a cohort of 318 children who were obese and nonobese.

It has been postulated that systemic inflammation triggered in the early stages of Ad36 infection leads to the development of human obesity (9, 14). Given that there are no optimal ranges for serum concentrations of TNF-α, IL-6, VEGF, and MCP-1 for adults or children, we compared these inflammatory-related markers across tertile groups. In the present study, the prevalence of Ad36(+) was significantly associated with TNF-α, IL-6, and VEGF but not MCP-1. These inflammatory-related markers represent a variety of cytokines, chemokines and growth factors derived from a common cellular pathway that is activated by Ad36 infection (9, 27) and have been shown to act synergistically to induce obesity and related disorders (28, 29). Indeed, Na and Nam (9) found that Ad36 infection increased expression of MCP-1 and TNF-α, which induced macrophage migration and infiltration, respectively, and culminated in increased adiposity in mice. Similarly, Bouwman et al (16) reported that Ad36-infected human adipocytes increased production of IL-6, which has been shown to stimulate VEGF and contribute to vascularization preceding adipose tissue growth (30, 31). Although these inflammatory-related markers have also been implicated in the development of cardiovascular diseases (13) and diabetes mellitus (13, 32), our sample was relatively healthy with no history of metabolic disorders, and potential chronic effects were indeterminate at this time point.

Although it is unclear why Ad36(+) was not associated with serum MCP-1 in children when Na and Nam (9) found a positive relationship in Ad36(+) adults, it is possible that age and the relative duration of Ad36 infection has differential effects on inflammatory response and long-term health. The sample of children in our study entering the early stages of puberty was younger than those of other studies that support the Ad36-adiposity association (mean age of 11 vs 16 y) (1, 3, 10). Although the chronic effects of Ad36 infection and inflammatory-related markers cannot be determined from the present study, it is possible that the Ad36-adiposity relationship emerges with age after years of exposure. This may be plausible, given the proposed mechanism by which Ad36-induced inflammation triggers obesity development and the underlying actions of the aforementioned inflammatory-related markers in the pathogenesis of progressive diseases (9, 16, 30). The hypothesis that viral infection may lead to chronic health outcomes must be explored through long-term prospective studies.

Strengths of the present study include the ethnically diverse sample of young children and the use of both direct (fat mass and body fat percentage via DXA) and indirect (BMI z-score and BMI for age percentile) measures of body fatness for comparisons. We acknowledge some potential limitations. Although our study was conducted in a relatively large sample (1, 10), it is possible that our sample size was inadequate to detect differences in adiposity by Ad36 group, given findings from larger-scale studies of children (2, 3). Our study also used cross-sectional data, and therefore, we cannot determine how long Ad36(+) participants had been infected, nor can we be certain that Ad36(+) had a direct effect on serum inflammatory-related markers. We did not measure plasma C-reactive protein, a clinical marker of systemic inflammation, although it should be noted that C-reactive protein has been shown to highly correlate with TNF-α (33) and IL-6 (33, 34). Lastly, we assessed total body fatness rather than regional adiposity. Regional adiposity measures (eg, sc and visceral fat deposition) differ after infection with the virus (4). Therefore, location of adipose tissue accumulation may provide greater insight into the long-term effects of Ad36(+) on production of inflammatory-related markers and chronic disease.

In conclusion, Ad36 seropositivity in children was not associated with greater adiposity, assessed by total body fat mass, but was associated with several biomarkers implicated in inflammation. Research efforts in this area should focus on determining whether exposure to elevated serum inflammatory-related markers observed in Ad36(+) children has long-term effects on obesity and related outcomes. Moreover, future studies are warranted to examine the relationships between regional fat distribution, inflammation and risk factors for chronic disease in children classified as Ad36(+) vs Ad36(−).