J Pediatr Endocrinol Metab 2016; 29(4): 441–449
Mackenzie J. Bodo, Elizabeth Yakes Jimenez, Carole Conn, Alyssa Dye, Page Pomo, Deborah Kolkmeyer, Robert Orlando and Alberta S. Kong*
Association between circulating CCL2 levels and modifiable behaviors in overweight and obese adolescents: a cross-sectional pilot study DOI 10.1515/jpem-2015-0260 Received June 29, 2015; accepted October 27, 2015; previously published online December 16, 2015
Abstract Background: We evaluated the relationship between an early inflammatory biomarker, chemokine (C-C motif) ligand 2 (CCL2), and other clinical biomarkers and lifestyle behaviors, in overweight/obese adolescents at high risk of developing cardiometabolic derangements. Methods: We collected anthropometric measurements, clinical biomarkers, and three 24-h dietary recalls from 21 vocational high school students (91% male), 14–19 years, with body mass index (BMI) ≥ 25 kg/m2. Pearson’s or Spearman’s correlation coefficients were used to examine relationships. Results: Mean BMI was 33.2 kg/m2 (range 25.7–45.6) and 38% were prediabetic by fasting glucose. Mean CCL2 was 512.9 pg/mL (range 220–917) and positively correlated with triglycerides (r = 0.45; p = 0.04) and TNF-α (r = 0.57; p = 0.007) and marginally negatively correlated with fruit/ *Corresponding author: Dr. Alberta S. Kong, MSC 10 5590, 1 University of New Mexico, Albuquerque, NM, 87131, USA, E-mail:
[email protected]; Department of Pediatrics, Division of Adolescent Medicine, University of New Mexico (UNM) School of Medicine, Albuquerque, NM, USA; and Department of Family and Community Medicine, UNM School of Medicine, Albuquerque, NM, USA Mackenzie J. Bodo and Carole Conn: Nutrition and Dietetics Program, Department of Individual, Family and Community Education, UNM, Albuquerque, NM, USA Elizabeth Yakes Jimenez: Nutrition and Dietetics Program, Department of Individual, Family and Community Education, UNM, Albuquerque, NM, USA; Department of Family and Community Medicine, UNM School of Medicine, Albuquerque, NM, USA; and Pacific Institute for Research and Evaluation, Albuquerque, NM, USA Alyssa Dye and Page Pomo: UNM School of Medicine, Albuquerque, NM, USA Deborah Kolkmeyer: Southwest Endocrinology Associates, Albuquerque, NM, USA Robert Orlando: Department of Biochemistry and Molecular Biology, UNM School of Medicine, Albuquerque, NM, USA
vegetable intake (r = –0.42, p = 0.06) and omega-3 fatty acids (r = –0.41, p = 0.07). Conclusions: CCL2 was positively associated with proinflammatory biomarkers and negatively associated with some anti-inflammatory dietary factors. Keywords: chemokine CCL2; inflammation; nutrition assessment; obesity; pediatrics.
Introduction Obesity has more than tripled in adolescents (12–19 year olds) over the past 35 years in the US. Currently, one in five adolescents meet the clinical criteria for obesity [1]. As US adolescents transition to adulthood, they experience a significant jump in obesity prevalence, with prevalence increasing from 20.5% in 12–19 year olds to 30.3% in 20–39 year olds to 39.5% in 40–59 year olds [1, 2]. Obesity can have pervasive effects on children’s health, by increasing the risk of insulin resistance, gallstone development, sleep apnea, hypertension, and continued obesity during adulthood, which leads to additional age-dependent metabolic consequences [3, 4]. Obesity is associated with chronic, low-grade systemic inflammation, which is linked to insulin resistance, metabolic dysfunction, hypertension, cardiovascular disease, and increased levels of inflammatory adipokines [5–9]. Based on these observations, identification of obesity-dependent inflammatory markers may provide a diagnostic or prognostic indicator of inflammatory status and staging of pathological changes leading to insulin resistance and metabolic dysfunction. Although the association of certain inflammatory mediators with onset of insulin resistance has been established in animal, in-vitro and adult human studies [7, 8, 10–13], few of these mediators have been investigated in adolescents [14–16]. Identification of biological markers in adolescents may allow health professionals to better intervene to slow or prevent onset of obesity-dependent pathologies and potentially reduce
Unauthenticated Download Date | 7/7/16 10:20 PM
442 Bodo et al.: Association between MCP-1 and modifiable behaviors in adolescents the incidence of insulin resistance as adolescents transition into adulthood. The inflammatory biomarker known as chemokine (C-C motif) ligand 2 (CCL2) is involved in the primary events of inflammation, leading to a positive feedback loop that results in chronic inflammation [8, 17]. Recent studies have shown that CCL2 is increased in obese adolescents when compared with lean adolescents [14–16, 18]. Importantly, CCL2 is also an independent predictor of insulin resistance in adults, although its predictive value in children and adolescents is less certain [8, 10–13]. Examination of the relationship between CCL2 and diet and exercise will aid in hypothesizing how to intervene to reduce CCL2 levels. In adults, excessive caloric intake, low physical activity, aging, and obesity are closely associated with initiating and establishing a chronic inflammatory response in adipose tissue [8]. In addition, numerous dietary factors are associated with increased or decreased inflammation. For example, the downstream signaling pathways of toll-like receptors (TLR) 2 and 4 are responsible for CCL2 production and pro-inflammatory cytokine activation, and are activated by dietary saturated fatty acids (SFAs) and inhibited by omega-3 polyunsaturated fatty acids (n-3 PUFAs) [19, 20]. Additional dietary variables shown to downregulate TLR4 include phytochemicals found in cinnamon (cinnamaldehyde), turmeric (curcumin) and vitamin D [21, 22]. Fruits, vegetables and their phytochemicals have also been associated with immunomodulatory and anti-inflammatory characteristics [23, 24]. Some studies have examined the impact of manipulating diet and physical activity on CCL2 levels. Studies in overweight/obese adult populations show that calcium intake of approximately 1400 mg/day for 12-weeks, 4 g/day of fish oil supplementation for 12 weeks, or high total intake of antioxidants for 9-months significantly lowered CCL2 levels [25–27]. Weight loss stabilization in adults has also been positively associated with reduced CCL2 levels [28]. Similarly, an aerobic training program in combination with a 21-day hospitalized controlled hypocaloric (1000 kcal/day) and macronutrient specific diet (55% carbohydrate, 25–30% fat with 7% saturated fat, and 54 g of animal protein) is capable of significantly lowering CCL2 levels [29]. To date, there have been only three studies addressing whether dietary intake [30], physical activity [15], or weight loss [18] are related to circulating CCL2 levels in adolescents. In this pilot study of overweight and obese adolescents, we broadly examined relationships between circulating levels of CCL2 and selected components of dietary intake and physical activity in overweight and
obese adolescents. In addition, we examined correlations between CCL2 and clinical biomarkers associated with inflammation. We aimed to identify associations of interest for future, larger studies. We hypothesized that biomarkers of insulin resistance, high dietary intake of SFAs and total energy would be positively associated with circulating CCL2 levels, while higher levels of physical activity and intake of fruits, vegetables, antioxidants and n-3 PUFAs would be negatively associated with CCL2 levels.
Materials and methods Study site and participants Students were recruited from a vocational training high school located in New Mexico, NM, USA. From April to May 2014, 28 students identified as overweight or obese through the school’s health clinic were referred to research staff. Students were eligible to participate in the pilot study if they were enrolled in the collaborating vocational high school, had a body mass index (BMI) ≥ 25 kg/m2, and a minimum remaining school-term of at least 1 month for follow-up. All students included in this analysis had a BMI percentile > 85th percentile [31]. Exclusion criteria included diagnosed diabetes, parental refusal (n = 2), stage 2 hypertension requiring pharmacotherapy, systemic corticosteroid medication use during the 2 weeks prior to the beginning of the study, current use of antipsychotic medications, inability to perform moderate to vigorous physical activity and pregnancy. All students provided written informed consent or assent, and parental consent was obtained for students between 14–17 years of age. The University of New Mexico Health Sciences Center Institutional Review Board approved the study protocol. A total of 21 male and female students, 14–19 years of age, completed all baseline data collection, allowing detection of a correlation of ≥ 0.52 (strong correlation) with a significance of 0.05 and 80% power. Measurements included anthropometrics, blood pressure, biomarkers, estimated dietary intake, self-reported physical activity level, health history, demographic information and eating behaviors. Five participants were either missing biochemical data or three complete days of dietary data, and were excluded from this analysis.
Anthropometrics and blood pressure Anthropometric measurements included height, weight, and waist circumference. Height was measured to the nearest 0.1 cm using a vertical measuring rod (Seca Model 213, Chino, CA, USA) and weight was measured to the nearest 0.1 kg using a portable strain-gauge digital scale (Seca Model 770, Chino, CA, USA), according to the methods described by Lohman and Martorell [32] and the National Health and Nutrition Examination Survey [33]. Waist circumference was measured to the nearest 0.1 cm using a retractable steel tape at the top of the iliac crest, according to the methods described by NHANES [33]. All measurements were taken in duplicate, with a third measurement taken if height measurements differed by > 1.0 cm, if weight measurements differed by > 0.5 kg, and if waist circumference
Unauthenticated Download Date | 7/7/16 10:20 PM
Bodo et al.: Association between MCP-1 and modifiable behaviors in adolescents 443 measurements differed by > 0.5 cm [33]. If three measurements were taken, the average of the closest two measurements was used for analysis. BMI was calculated as kilograms per meters squared, and Z-scores and corresponding BMI percentiles were generated from the participants’ measured BMI using the Center for Disease Control and Prevention’s equation and age- and sex-specific LMS parameters for the BMI-for-age charts, ages 2–20 years [34]. Teens were considered overweight if they had a BMI percentile from 85 to 94.9, and obese if their BMI percentile was ≥ 95. Waist circumference was converted to age- and sex- specific percentiles using LMS tables shared by Cook et al. [35]. Blood pressure was measured on the right arm in a seated position with an aneroid sphygmomanometer. Prior to taking the measurements, participant mid-upper arm circumference was measured to determine the appropriate cuff size to use for each individual [36]. Participants were seated at rest for at least 5 min prior to taking the measurements. Measurements were repeated in triplicate, with at least 1 min between measurements, and the average of the second two measurements was used for analysis. Age- and sex-specific blood pressure percentiles were calculated following the statistical procedures specified in the Fourth Task Force Report on high blood pressure in children and adolescents [37].
Biomarkers Biomarkers measured included glucose, insulin, hemoglobin A1c (HbA1c), lipids, and pro-inflammatory markers (high sensitivity C-reactive protein, hsCRP; tumor necrosis factor-α, TNF-α; monocyte chemoattractant protein-1, CCL2). Glucose was measured twice, about 4 weeks apart, to assess fasting glucose 100–125 mg/dL in the prediabetes range [38]. Blood samples were drawn after a 12-h fast and serum samples were prepared by incubating whole blood at room temperature for 30 min, followed by refrigerated centrifugation at 1000 × g for 10 min. Supernatants (serum) were transferred to clean tubes and stored at 4 °C until assayed. All samples were centrifuged on-site and transported in an ice chest with freezer packs to the Clinical and Translational Science Center (CTSC), University of New Mexico, Health Sciences Center laboratory for processing. Glucose: Glucose concentration was measure by a standard oxidoreduction assay using glucose-6-phosphate dehydrogenase and quantifying NADH formation. Samples were processed using an ACE Alera® Clinical Chemistry System (Alfa Wassermann Diagnostic Technologies, West Caldwell, NJ, USA). Insulin: Insulin concentration was determined by capture ELISA using a Siemens Immulite 1000 Immunoassay System (Siemens Healthcare, Malvern, PA, USA). The homeostatic model assessment insulin resistance index (HOMA-IR) was calculated as [(fasting glucose in mmol/L × fasting insulin in μU/mL)/22.5 [39]. HOMA-IR increases with increasing insulin resistance, and has been shown to have reliable sensitivity and specificity in assessing insulin resistance [40, 41]. HbA1c: HbA1c levels were measured using a competitive agglutination assay and quantified using a DCA Vantage Analyzer® (Siemens Healthcare, Malvern, PA, USA).
Pro-inflammatory cytokines: hsCRP, TNF-α and CCL2 were measured by a Bio-Plex® Precision Pro™ Human Cytokine Assay System (BioRad, Hercules, CA, USA). The assay uses fluorochrome-labeled magnetic bead technology in a capture immunoassay format coupled with fluorescent detection antibodies for precise quantification of cytokine levels.
Dietary intake and physical activity Selected questions from the New Mexico Youth Risk and Resiliency Survey (NMYRRS) were used to assess physical activity levels and general eating behaviors [42]. Physical activity level was estimated based on one question asking about how many days per week the participant completed at least 60 min of physical activity that increased his/her heart rate and made him/her breathe hard, based on the U.S. Department of Health and Human Service’s recommendation that adolescents complete at least 60 min of moderate to vigorous physical a day [43]. Those who reported 4 days a week or more were considered to be physically active adolescents. Dietary intake was assessed via an in-person interview by one of four researchers who were trained in collecting 24-h recalls using the multiple-pass method [44]. This method uses verbal probes for portion size description, food preparation, and commonly forgotten foods, and has been validated for estimating energy intake against the doubly labeled water method in small groups of normal weight men and women and younger children (ages 3–12) [44–46]. An illustrated portion size booklet was used to help the participants estimate their daily intake. The three 24-h diet-recalls included at least 1 weekday and 1 weekend day. The Nutrition Data System for Research software (NDSR), version 2014, developed by the Nutrition Coordinating Center (University of Minnesota, Minneapolis, MN, USA), was used to estimate dietary intake of nutrients and food groups and to tally meal intake. Dietary intake was assessed related to guidelines for a nutritious diet such as the Dietary Reference Intakes’ average macronutrient distribution range (AMDR) and estimated average requirement (EAR) for specific nutrients.
Statistics Study data were managed using REDCap electronic data capture tools hosted at the University of New Mexico [47]. REDCap (Research Electronic Data Capture) is a secure, web-based application designed to support data capture for research studies. Stata/ SE12 (StataCorp LP, College Station, TX, USA) was used for statistical analysis. Descriptive statistics were calculated for all variables of interest, including mean, standard deviation and range for continuous variables and frequency for categorical variables. Normality of variables was examined using the Shapiro-Wilk test. Pearson and Spearman bivariate correlation coefficients and χ2 tests were used, as appropriate based on the assumptions of each test, to test relationships between CCL2 and 35 individual pro- or anti-inflammatory dietary variables, including dietary fats, antioxidants and food groups such as fruits and vegetables, and anthropometric, clinical biomarker and physical activity variables. Results were considered significant if the p-value was ≤ 0.05, unless otherwise noted.
Unauthenticated Download Date | 7/7/16 10:20 PM
444 Bodo et al.: Association between MCP-1 and modifiable behaviors in adolescents
Results Demographics and baseline characteristics of the participants (n = 21), who were predominantly male (91%), are presented in Table 1. The population comprised both obese (76%) and overweight (24%) adolescents ages 14–19. Eight students (38%) had two consistent fasting blood glucose measurements in the prediabetic range (100–125 mg/dL). Fourteen percent had high-density lipoprotein (HDL), and about one third had low HDL and high triglycerides. Total average reported energy intake was 1688 kcal, with a range from 918 to 3014 kcal, and an average macronutrient distribution of 36% fat, 49% carbohydrate and 15% protein. Based on the proportion of participants with intake below the estimated average requirements (EAR), intake of several vitamins and minerals was estimated to be low. It was estimated that 100% (vitamins E and D,
and magnesium), 86% (calcium), 76% (vitamin A), 71% (vitamin C), 67% (phosphorus), 38% (vitamin B6), 33% (riboflavin) and 29% (folate) of participants had estimated intake below the EAR. Table 2 highlights estimated intake for nutrients and food groups with pro-inflammatory and anti-inflammatory properties, compared to recommended intake for adolescents. The relationship between CCL2 and select b iomarkers, select pro- and anti-inflammatory nutrients, and selfreported physical activity is presented in Table 3. C linical variables that had significant, moderate positive correlations to CCL2 were triglycerides (r = 0.45, p = 0.04) and TNF-α (r = 0.57, p = 0.007). CCL2 displayed a moderate negative trending correlation to total fruit and vegetable intake (r = –0.42, p = 0.06) and to total linolenic acid, eicosapentaenoic acid (EPA), and docosahexaenoic acid (DHA) intake (r = –0.41, p = 0.07).
Table 1: Characteristics of vocational high school students (n = 21). Demographics Age in years, mean±SD (range) Sex, n (%) male Race/Ethnicity, n (%) Hispanic/Latino American Indian/Alaska Native White Anthropometrics Height, cm [mean±SD (range)] Weight, kg [mean±SD (range)] Waist circumference, cm [mean±SD (range)] Waist circumference percentile, mean±SD (range) Body mass index (BMI), kg/m2 [mean±SD (range)] BMI percentile, mean±SD (range) BMI percentile category Overweight (85th – 94.9th percentile), n (%) Obese (≥95th percentile), n (%) Biomarkers Hemoglobin A1c, % [mean±SD (range)] Homeostatic model assessment insulin resistance index (HOMA-IR), mean±SD (range) Glucose, mg/dL, [mean±SD (range)] Systolic blood pressure, mm Hg [mean±SD (range)] Diastolic blood pressure, mm Hg [mean±SD (range)] Cholesterol, mg/dL [mean±SD (range)] Low density lipoprotein (LDL), mg/dL [mean±SD (range)] High density lipoprotein (HDL), mg/dL [mean±SD (range)] Triglycerides, mg/dL [mean±SD (range)] Chemokine (C-C motif) ligand 2 (CCL2), pg/mL [mean±SD (range)] TNF-α, pg/mL [mean±SD (range)] High sensitivity C-reactive protein, mg/L [mean±SD (range)] Family history and lifestyle Family history of type 2 diabetes, n (%) Family history of heart attack before age 56, n (%) Actively trying to lose weight, n (%) Participates in 60 min of physical activity, four or more days a week, n (%)
Unauthenticated Download Date | 7/7/16 10:20 PM
16.2±1.4 (14–19) 19 (91) 15 (71) 1 (5) 5 (24) 173.3±6.0 (157.4–181.2) 100.3±22.3 (65.5–149.0) 106.2±15.6 (85.6–140.5) 93.6±5.6 (79.4–99.4) 33.2±6.0 (25.7–45.6) 96.6±4.0 (85.2–99.9) 5 (24) 16 (76) 5.3±0.2 (4.9–5.9) 14.0±18.2 (2.0–80.3) 102.0±10.8 (88–136) 121.9±8.1 (104.5–134) 68.1±6.6 (56–79) 158.7±31.0 (117–230) 89.8±25.4 (47–146) 43.3±8.0 (31–65) 136.8±70.3 (63–308) 512.9±184.4 (220–917) 16.9±4.8 (9–32) 2.2±2.3 (0.2–7.4) 9 (43%) 2 (10%) 17 (81%) 7 (33%)
Bodo et al.: Association between MCP-1 and modifiable behaviors in adolescents 445 Table 2: Adolescent intake of pro-inflammatory and anti-inflammatory nutrients and food groups, compared to recommended intake, by sex. Pro-inflammatory Total sugars, g
% kcal from trans-fatty acids % kcal from saturated fatty acids Linoleic acid, g Arachidonic acid, g Sodium, g Anti-inflammatory Alpha-linolenic acid, g Vitamin C, mg Vitamin E, mg Food groups Total fruits (cups) Total vegetables (cups) Total whole grains (oz) Total dairy (cups)
Average intake: Mean±SD (range)
Male (M): 90.3±66.0 (9.7–273.3) Female (F): 97.6±49.1 (62.9–132.3) Total: 1.5±0.7 (0.8–3.6) Total: 13.0±2.9 (9.0–18.4) M: 11.6±5.6 (3.9–25.6) F: 15.2±15.9 (4.0–26.5) Combined M/F average macronutrient distribution range (AMDR): 6.3±2.3% (3.5–10.7) M: 0.10±0.06 (0.03–0.24) F: 0.05±0.031 (0.03–0.07) M: 3.1±1.4 (1.7–6.7) F: 3.8±3.5 (1.3–6.3) Male: 1.3±0.7 (0.45–3.14) Female: 1.9±2.1 (0.5–3.5) Combined M/F AMDR: 0.7±0.3% (0.4–1.3) M: 68.6±66.5 (12.2–227.0) F: 70.8±22.7 (54.8–86.9) M: 5.27±2.64 (1.51–11.10) F: 6.97±5.47 (3.10–10.83) M: 0.59±0.70 (0–2.84) F: 0.48±0.62 (0.04–0.92) M: 0.92±0.60 (0.07–2.2) F: 1.91±1.16 (1.09–2.73) M: 0.62±0.59 (0–2.03) F: 2.27±0.34 (2.03–2.52) M: 1.40±0.94 (0.06–4.33) F: 1.20±0.16 (1.08–1.32)
Discussion In this pilot study, we found that overweight and obese adolescents had a substantial range of circulating CCL2 levels, and that high CCL2 levels were positively correlated with high triglycerides, a component of metabolic syndrome, and TNF-α, another pro-inflammatory mediator. Students had high estimated intake of pro-inflammatory nutrients and low estimated intake of anti-inflammatory nutrients and foods. Students with higher intake of antiinflammatory fruits and vegetables and omega-3 fatty acids (total linolenic acid, EPA and DHA) had lower CCL2 that trended towards significance. Given the importance of interrupting the progression from obesity and
Recommended intake (14–18 years of age) Not determined (ND) American Heart Association (AHA):
