Novel Adipokines CTRP1, CTRP9, and FGF21 in Pediatric Type 1 and Type 2 Diabetes: A Cross-Sectional Analysis

The prevalence of obesity has increased tremendously over the last several decades [1], with a concurrent increase in the incidence of pediatric diabetes over the last 2 decades [2]. Despite the long-held notion that children with type 1 diabetes (T1D) have a thin body habitus, there has also been an increase in obesity and insulin resistance in this population of children [3, 4], such that the specific type of diabetes at diagnosis can often be difficult to distinguish [5]. While type 2 diabetes (T2D) is directly related to obesity, and up to 80% of youth with T2D are obese [5], often with insulin resistance and signs of metabolic syndrome, the impact of overweight and obesity on youth with T1D is just starting to be explored [6, 7].

Our understanding of the role of adipokines in the cross-talk of the immune, endocrine and cardiovascular systems in the setting of obesity and insulin resistance is evolving in this young population. Previous studies have demonstrated higher leptin levels in obese individuals with T1D compared to lean T1D controls [8], similar to trends seen in other obese populations [9-12], and obese youth with T2D [13]. Further, lower adiponectin levels have been shown with increasing BMI in T1D youth [8], although this relationship is not consistent as other studies have demonstrated paradoxically high adiponectin levels in youth with T1D even in the setting of insulin resistance [14]. Common adipokines such as leptin and adiponectin have been well studied in obese youth with and without metabolic syndrome and T2D [8, 15-21], but given the increase in obesity even in the setting of T1D, it is also important to consider the role of other adipokines and inflammation in this cohort of youth as well.

Novel adipokines including the C1q/TNF-related proteins (CTRPs) family of proteins and the hepatokine, fibroblast growth factor 21 (FGF21) have been implicated in glucose and/or lipid metabolism in the adult population, but there is less data in pediatrics [22-28]. CTRP9 has been shown to be elevated in adults with obesity and T2D [12, 29, 30], but this same relationship has not been shown in the pediatric population [13]. CTRP1 is higher in adults with obesity and hyperglycemia [24, 31, 32], and has been shown to be higher in adolescents with obesity and hyperglycemia compared to normoglycemic obese youth [13]. The hepatokine, FGF21 has been shown to be higher in adults with obesity and T2D [27, 33-39]. Our group previously showed that obese youth with impaired glucose tolerance have higher FGF21 levels compared to obese youth with normal glucose tolerance [13], but that it may be due to a compensatory mechanism as longitudinal studies of children demonstrating weight loss had a decrease in FGF21 levels [40].

While most studies have focused on youth with obesity and T2D, with the increase in overweight and obesity affecting youth with T1D and the potential risk for cardiovascular complications as a result of the additive impact of obesity, it is also important to understand the inflammatory milieu in this population. We sought to determine the differences in adipokine levels between youth with T1D and T2D and how they relate to clinical characteristics.

This was a cross-sectional study utilizing samples from the Pediatric Diabetes Consortium (PDC) Registry in a secondary analysis. A cohort of youth with newly diagnosed T1D was enrolled in the PDC T1D NeOn registry by 7 US pediatric diabetes centers between 2009 and 2011, and a cohort of youth with T2D was enrolled by 8 pediatric diabetes centers between 2012 and 2014. A detailed description of the PDC T1D NeOn registry has been published [41, 42]. The PDC T2D registry included participants who were <21 years of age, and were diagnosed with T2D using the criteria of the American Diabetes Association including being negative for T1D associated autoantibodies (including insulin autoantibody, glutamic acid decarboxylase antibody, islet antigen 2 antibody, islet cell antibody, and zinc transporter 8 antibody), BMI >85th percentile for age and sex prior to diabetes-associated weight loss and at least one of the following: (i) HbA1c ≥6.5% (≥48 mmol/mol), (ii) random glucose >200 mg/dL (11.1 mmol/L), (iii) fasting glucose ≥126 mg/dL (7.0 mmol/L), or (iv) 2-h OGTT glucose ≥200 mg/dL (11.1 mmol/L) [43]. The PDC protocol was approved by the IRBs at each participating center. Written informed consent was obtained from parents with written assent in participants <18 years of age, and informed consent was obtained from participants ≥18 years of age. This study protocol was reviewed and approved by the Johns Hopkins IRB, approval number: IRB00219817. The source of the data is the PDC registry, but the analyses, content, and conclusions presented herein are solely the responsibility of the authors and have not been reviewed or approved by the PDC.

This secondary analysis of participants from the PDC registries included a total of 80 randomly selected participants: 40 with T1D and 40 with T2D, who were pubertal males and females, sex and age matched between the two groups, with a range of BMI and hemoglobin A1c values. Data on enrollment into the Registry were collected via the medical record and interviews with participants and/or parents. Data included clinical (age, race/ethnicity, duration of diabetes), anthropometric variables (BMI, BMI%, and BMI Z-score), and metabolic variables (glucose, C-peptide, Hemoglobin A1c [HbA1c], and vitamin D). BMI was measured by the health care provider, and BMI percentile for age and gender was calculated using the 2000 CDC population growth chart data.

Human CTRP9, CTRP1, and FGF21 were measured by ELISA (Biovendor, Czech Republic) per the manufacturer: CTRP9 ELISA intra-assay variation 5.5%, inter-assay variation 7.9%; CTRP1 ELISA intra-assay variation 2.7% inter-assay variation 8.5%; FGF21 ELISA intra-assay variation 2% inter-assay variation 3.3%. Laboratory methods for metabolic variables are previously described [41-43] and are presented here in brief: Hemoglobin A1c levels were measured at each center by the DCA Vantage (Siemens, New York, NY, USA). Glucose levels were nonfasting random levels in 37 participants in both T1D and T2D groups, and fasting in 3 participants in both T1D and T2D groups.

Initial comparisons between T1D and T2D groups were analyzed using Student’s t test for normally distributed continuous variables. Variables that were non-normally distributed were either log transformed to fit a normal distribution or compared using the Wilcoxon rank-sum test. A logistic regression was used for sex and a Fischer’s exact test was used for race/ethnicity. Multivariable regression was used to correlate variables across all 80 participants controlling for age, sex, BMI, diabetes duration when C-peptide was in the model, and glucose when FGF21 was in the model. For each group, the novel adipokines CTRP1, CTRP9, and FGF21 were correlated with each of the metabolic factors using a univariate and multivariable linear regression models. Variables significant in univariate regression models were controlled for in multivariable regression models additionally controlling for age, sex, BMI, diabetes duration when C-peptide was in the model, and glucose when FGF21 was in the model. Data are reported as either mean ± SD if normally distributed or log transformed to fit a normal distribution or median with interquartile range if skewed. Statistical analyses were performed using the R statistical software package. p values <0.05 were considered statistically significant across all analyses.

A total of 80 participants were included, 40 with T1D and 40 with T2D as shown in Table 1. By design, each group had 20 males and 20 females, with an average age of 14.9 ± 2 years. Patients with T1D had a mean duration of diabetes of 3.3 ± 0.7 years, and mean HbA1c was 9.2 ± 2.4%. T2D participants had a lower mean duration of diabetes of 2.3 ± 1.7 years (p = 0.0009) and lower mean HbA1c of 8.1 ± 2.9% (p = 0.02) when compared to T1D participants. Weight (61.9 ± 13.1 vs. 93.6 ± 24.8 kg, p < 0.0001), BMI (22.7 ± 3.6 vs. 33.9 ± 7.4 kg/m², p < 0.0001), BMI% (78.0 [61.0, 91.0] vs. 98.2 [96.6, 99.4] %, p < 0.0001), and C-peptide levels (0.1 [0.03, 0.3] vs. 4.2 [2.6, 6.4] ng/mL, p < 0.0001) were lower in participants with T1D compared to T2D, but there was no difference in BMI Z-scores. Glucose levels were higher in participants with T1D than T2D (198.8 ± 115.9 vs. 158.2 ± 113.7 mL/dL, p = 0.04).

CTRP9 was significantly higher in T1D compared to T2D participants (13,903.6 vs. 3,608.5 pg/mL, p = 0.04), FGF21 was significantly higher in T2D compared to T1D (113.1 vs. 70.6 pg/mL, p = 0.03), but there were no differences observed in CTRP1 between these two groups in unadjusted analysis (Fig. 1). Adjusting for age, sex, duration of diabetes, ethnicity, and BMI showed BMI to be the only confounder to the observed differences in CTRP9 and FGF21 levels between both groups, as BMI is collinear with diabetes type.

As shown in Table 2, CTRP1 was not observed to be associated with any clinical variables in T1D. CTRP9 was positively associated with C-peptide levels in univariate analysis (p = 0.005) and when adjusting for age, sex, BMI, and duration of diabetes (p = 0.006). FGF21 was higher in females (152.9 vs. 72.1 pg/mL, p = 0.009) than males, and positively associated with HbA1c (p = 0.04) even when adjusting for age, sex, BMI, and glucose levels.

In participants with T2D, CTRP1 was positively associated with HbA1c (p = 0.0002), glucose (p = 0.004), and FGF21 (p < 0.0001) even when adjusting for age, sex, and BMI (Table 3). CTRP9 was inversely associated with C-peptide (p = 0.03) and FGF21 (p = 0.005), and remained significantly associated with FGF21 on multivariable analysis (p = 0.01) adjusting for age, sex, BMI, glucose, and duration of diabetes. FGF21 was positively associated with C-peptide levels (p = 0.007), which remained significant (p = 0.02) when adjusting for age, sex, BMI, diabetes duration, and other significant characteristics. In the multivariable analysis model, FGF21 was positively associated with CTRP1 (p = 0.002), but not with CTRP9.

As the obesity epidemic continues [44], and the prevalence of T1D and T2D increases [2, 45, 46], it is important to understand the connection of adipokines and their relationship to metabolic factors in youth with diabetes. To our knowledge, this is the first comparison of the circulating novel adipokines CTRP1, CTRP9, and FGF21 in pediatric patients with T1D and T2D. We demonstrate that CTRP9 is higher in youth with T1D compared to T2D, and the hepatokine FGF21 is higher in youth with T2D compared to T1D. Further, we show that in youth with T2D, the novel adipokine CTRP1 is positively associated with HbA1c and glucose. The associations identified in this study of youth with T1D and T2D, while noting the absence of a control group, may help us begin to elucidate early alterations in adipokines and identify future research areas that may be able to inform future therapeutics and prevention of diabetes-related complications.

Similar to our prior study of CTRP9 in obese adolescents with and without hyperglycemia, we did not see an association of CTRP9 with obesity or measures of glycemia in the T2D cohort [13]. This is different than results from adult T2D studies demonstrating elevated CTRP9 levels in the setting of obesity and diabetes [12, 30, 47]. Based on these data, one could hypothesize that CTRP9 levels are not yet elevated in the pediatric or adolescent T2D population and suggest that CTRP9 levels may be changing in the course of T2D disease progression, but future longitudinal studies would be required to further evaluate this. Prior research suggests that CTRP9 may have protective effects on diabetes-related complications. In vivo studies in diabetic mice demonstrate that administration of recombinant CTRP9 attenuates development of diabetic nephropathy and development of kidney fibrosis [48], and reduces cardiac injury and infarct size [49]. In vitro studies also show that CTRP9 overexpression attenuates hyperglycemia-induced oxidative stress in human retinal pigment epithelial ARPE-19 cells [50]. Thus, future research is needed to investigate CTRP9, as a potential biomarker for those at risk of developing diabetes, and what its role may be as a therapy to mitigate diabetes-associated complications. To our knowledge, this is the first study evaluating CTRP9 levels in youth with T1D. CTRP9 levels in youth with T1D were associated with C-peptide levels, even after adjusting for age, sex, BMI and diabetes duration, and suggests that CTRP9 levels may be changing in the course of T1D disease progression, but future longitudinal studies would be required to further evaluate this.

While there were no differences in CTRP1 levels between the T1D and T2D groups, CTRP1 was positively associated with glucose and HbA1c in adolescents with T2D, consistent with our previous study [13]. In our recent study of CTRP1 in obese adolescent youth with and without hyperglycemia, we found CTRP1 to be positively associated with glucose levels, and higher in obese hyperglycemic youth compared to obese normoglycemic adolescents, consistent with adult studies showing elevated CTRP1 levels in T2D and the metabolic syndrome [29, 31, 32, 51, 52]. We also showed CTRP1 to be positively associated with FGF21, consistent with prior research in adults [52]. Both CTRP1 and FGF21 are known to be elevated in adult subjects with T2D, which is posited to be a compensatory mechanism since administration of recombinant CTRP1 lowers blood glucose in mice, and overexpression of CTRP1 in transgenic mice improved insulin sensitivity [22]. Studies have shown that administration of FGF21 also reduces plasma glucose, improves insulin sensitivity, and preserves β-cell function in diabetic animal models [32, 53].

Childhood obesity is associated with increased serum FGF21 levels [36, 40, 54], consistent with our recent results illustrating increased FGF21 levels in obese adolescents with hyperglycemia [13]. We also found T2D participants to have elevated FGF21 levels when compared to T1D. This is consistent with a study by Xiao et al. [55] that found similar changes in serum FGF21 levels in adults with different types of diabetes, including T1D, Latent Autoimmune Diabetes in Adults and T2D. These authors found a positive association between FGF21 and C-peptide in adult subjects with all types of diabetes, which is consistent with our findings [55]. Interestingly, we found that FGF21 levels were positively associated with HbA1c in youth with T1D, suggesting that FGF21 levels are higher with worse glycemic control. As other studies have shown in T1D, this may be explained by an early-stage compensatory mechanism where hyperglycemia initially activates FGF21 secretion by the liver or adipose tissue but is not sustained long term with subsequently declining levels of FGF21 [55, 56]. Interestingly, an inverse relationship has been reported with studies suggesting that lower FGF21 levels are associated with worse diabetes-associated complications. An in vivo study of FGF21 deletion in diabetic mice demonstrated aggravated aortic remodeling and rescue with FGF21 attenuated diabetes-induced aortic damage [57]. Further, another study in humans showed that patients with T1D and retinopathy were shown to have lower FGF21 levels than healthy controls [58], whereas other clinical studies suggest that FGF21 administration may attenuate diabetes-related kidney disease [59]. Future longitudinal studies are needed to further investigate the compensatory mechanism and changing FGF21 levels throughout the course of disease progression in diabetes.

Although these data are well characterized from an established registry, we acknowledge several limitations. This study is limited by the modest sample size and demographic variability between groups. Puberty staging was not available so it could not be adjusted for in analysis. Additionally, the lack of a control group in this cross-sectional design limits the generalization of results. Despite limitations, the uniqueness of this study still provides important results to serve as a benchmark for further research.

With the concurrent increase in obesity and diabetes in youth, it is important to understand the relationship between adipokines and the inflammatory milieu in this young population. Our results suggest that CTRP9 levels are elevated in youth with T1D with intact C-peptide levels but longitudinal studies will need to determine how this changes over time and if CTRP9 levels decrease as seen in the adult population. In youth with T2D, we found elevated FGF21 levels similar to trends seen in adult T2D. While CTRP1 was not different between youth with T1D and T2D, this novel adipokine’s association with glucose and HbA1c that persists across studies and cohorts suggests it may play a role in the metabolome of youth with T2D. Future studies with larger cohorts and longitudinal follow-up may help further elucidate the relationship of adipokines and obesity in pediatric diabetes.

The PDC protocol was approved by the Institutional Review Board at each participating center, according to the Declaration of Helsinki. Written informed consent was obtained from parents with written assent in participants <18 years of age, and informed consent was obtained from participants ≥18 years of age. This study protocol was reviewed and approved by the Johns Hopkins IRB, approval number: IRB00219817.

The authors have nothing to disclose.

There was no funding for this project. The authors report the following funding: NIDDK (DK084171 to GWW), NICHD (HD071981 to SNM), and NEI (EY033233 to RMW).

R.M.W., S.N.M., and G.W.W. designed the study. R.M.W., D.C.S., and G.W.W. performed the laboratory work. A.A., R.M.W., S.N.M., and D.C.S. analyzed the data. A.A. and R.M.W. wrote the manuscript. All authors critically reviewed and revised the manuscript.

Data are available upon request from the PDC.