Serum Cytokines and Growth Factors in Subjects with Type 1 Diabetes: Associations with Time in Ranges and Glucose Variability

The detrimental effect of hyperglycemia and glucose variability (GV) on target organs in diabetes can be implemented through a wide network of regulatory peptides. In this study, we assessed a broad panel of serum cytokines and growth factors in subjects with type 1 diabetes (T1D) and estimated associations between concentrations of these molecules with time in ranges (TIRs) and GV. One hundred and thirty subjects with T1D and twenty-seven individuals with normal glucose tolerance (control) were included. Serum levels of 44 cytokines and growth factors were measured using a multiplex bead array assay. TIRs and GV parameters were derived from continuous glucose monitoring. Subjects with T1D compared to control demonstrated an increase in concentrations of IL-1β, IL-1Ra, IL-2Rα, IL-3, IL-6, IL-7, IL-12 p40, IL-16, IL-17A, LIF, M-CSF, IFN-α2, IFN-γ, MCP-1, MCP-3, and TNF-α. Patients with TIR ≤ 70% had higher levels of IL-1α, IL-1β, IL-6, IL-12 p70, IL-16, LIF, M-CSF, MCP-1, MCP-3, RANTES, TNF-α, TNF-β, and b-NGF, and lower levels of IL-1α, IL-4, IL-10, GM-CSF, and MIF than those with TIR > 70%. Serum IL-1β, IL-10, IL-12 p70, MCP-1, MCP-3, RANTES, SCF, and TNF-α correlated with TIR and time above range. IL-1β, IL-8, IL-10, IL-12 p70, MCP-1, RANTES, MIF, and SDF-1α were related to at least one amplitude-dependent GV metric. In logistic regression models, IL-1β, IL-4, IL-10, IL-12 p70, GM-CSF, HGF, MCP-3, and TNF-α were associated with TIR ≤ 70%, and MIF and PDGF-BB demonstrated associations with coefficient of variation values ≥ 36%. These results provide further insight into the pathophysiological effects of hyperglycemia and GV in people with diabetes.


Introduction
The burden of type 1 diabetes (T1D) is enormous and is expected to increase.In 2021, there were about 8.4 million individuals worldwide living with T1D.The remaining life expectancy of a 10-year-old diagnosed with T1D ranges from a mean of 13 years in low-income countries to 65 years in high-income ones [1].Vascular diabetes complications remain an important determinant of mortality in subjects with T1D [2][3][4].The Diabetes Control and Complications Trial (DCCT) and its longitudinal observational follow-up study, the Epidemiology of Diabetes Interventions and Complications (EDIC) study, clearly showed a reduction in the incidence of diabetic retinopathy, diabetic nephropathy, diabetic neuropathy, and cardiovascular disease (CVD) outcomes, following the optimization of glycemic control in people with T1D [5,6].However, the benefits of intensive insulin treatment in terms of the risk of major adverse cardiovascular events vary substantially between individuals with T1D [7].Age, diabetes duration, and glucose level do not completely explain the associations between complications and mortality [4].This underlines the necessity for a better understanding of the drivers of diabetic complications.
Clinicians noticed a long time ago that many patients with excessive glucose fluctuations and unstable glycemic control rapidly develop complications.One of the first experimental studies testing the deteriorating effect of glucose fluctuations on the kidneys was published in 1957: it showed the rapid development of glomerulosclerosis in rats when excessive blood glucose fluctuations were induced by the intermittent administration of glucose and insulin [8].In recent years, the role of glucose variability (GV) as a trigger of diabetic vascular complications has attracted increasing attention [9,10].In the observational Finnish Diabetic Nephropathy (FinnDiane) study, glycated hemoglobin A1c (HbA1c) variability predicts incident cardiovascular events, microalbuminuria, and overt diabetic nephropathy in subjects with T1D [11].A recent analysis of 14 studies with 254,017 patients with diabetes revealed that the highest HbA1c variability is associated with increased risks of CVD.The CVD risk associated with HbA1c variability might be even higher among patients with T1D [12].However, the role of short-term GV as a trigger of diabetic complications remains to be clarified.
Continuous glucose monitoring (CGM) opened the way for the comprehensive assessment of short-term GV in patients with diabetes.Two principal dimensions of GV, amplitude and time, can be assessed and visualized from CGM data [9].Time in range (TIR), time above range (TAR), and time below range (TBR) became standardized CGM metrics for clinical care [13].Accumulating evidence suggests a negative association between time in range (TIR) and microvascular complications in subjects with T1D [14,15].In addition, associations of coefficient of variation (CV) and the mean amplitude of glycemic excursions (MAGE) with diabetic microvascular complications [16] and cardiovascular autonomic neuropathy [17] were reported.
The mechanisms of the harmful effects of excessive GV on target organs in diabetes have been intensively studied in recent years.The cardiovascular system, pancreas, adipose and muscle tissues, gastrointestinal tract, and kidney have been recognized as the loci with the highest expression of GV-related genes [18].Current data indicate that GV effects can be realized through oxidative stress, non-enzymatic glycation, chronic low-grade inflammation, endothelial dysfunction, platelet activation, impaired angiogenesis, and renal fibrosis.At the molecular level, these pathophysiological processes may be mediated through shifts in the production of cytokines and growth factors playing an important role in intercellular interactions [19].At present, the relationships between GV and changes in the production of these regulators are not well understood.
The large number of secreted cytokines and their interactions make it difficult to assess changes in the cytokine response during stress and disease.Moreover, measuring individual regulators can lead to a one-sided or even incorrect interpretation of the response.In such situations, multiplex platforms, widely used to measure multiple biomarkers from a single assay, may have advantages over single assay-based detection methods [20,21].In recent years, a multiplex bead assay has become increasingly popular in studies of cytokine panels in various diseases, including diabetes [22][23][24].
Therefore, we aimed to assess a broad panel of circulating cytokines and growth factors in subjects with T1D via a multiplex bead array assay and to determine the associations of these regulators with CGM-derived TIR and GV parameters.

Design
We performed a cross-sectional observational single-center comparative study.Caucasian male and female patients with T1D aged 18 to 70 years were included.Acute infections within three months prior to the study, pregnancy, malignant neoplasm, chronic inflammatory or autoimmune diseases, current diabetic ketoacidosis or hyperglycemic hyperosmolar state, end-stage renal disease, and diabetic foot syndrome were established as the principal exclusion criteria.Subjects without the above-mentioned diseases and conditions who had normal glucose tolerance (NGT) verified by the results of the oral glucose tolerance test and HbA1c measurement were included in the control group.
All study participants underwent a detailed clinical examination with real-time CGM.Digital CGM data were used for the calculation of TIRs and GV parameters.The panel of serum cytokines and growth factors was assessed with a multiplex bead array assay.
The study design is shown in Figure 1.
Caucasian male and female patients with T1D aged 18 to 70 years were included.Acute infections within three months prior to the study, pregnancy, malignant neoplasm, chronic inflammatory or autoimmune diseases, current diabetic ketoacidosis or hyperglycemic hyperosmolar state, end-stage renal disease, and diabetic foot syndrome were established as the principal exclusion criteria.Subjects without the above-mentioned diseases and conditions who had normal glucose tolerance (NGT) verified by the results of the oral glucose tolerance test and HbA1c measurement were included in the control group.
All study participants underwent a detailed clinical examination with real-time CGM.Digital CGM data were used for the calculation of TIRs and GV parameters.The panel of serum cytokines and growth factors was assessed with a multiplex bead array assay.
The study design is shown in Figure 1.

Methods
General clinical tests.The levels of HbA1c, serum biochemical parameters, and urinary albumin were assessed in patients with T1D with the use of a AU480 Chemical Analyzer (Beckman Coulter, Brea, CA, USA) and commercially available cartridges.A complete blood count was performed on a hematology analyzer (Analyticon Biotechnologies AG, Lichtenfels, Germany).The fasting C-peptide was determined using chemiluminescent immunoassay with an Immulite 2000 XPi immunological analyzer (Siemens Healthineers, Erlangen, Germany).

Methods
General clinical tests.The levels of HbA1c, serum biochemical parameters, and urinary albumin were assessed in patients with T1D with the use of a AU480 Chemical Analyzer (Beckman Coulter, Brea, CA, USA) and commercially available cartridges.A complete blood count was performed on a hematology analyzer (Analyticon Biotechnologies AG, Lichtenfels, Germany).The fasting C-peptide was determined using chemiluminescent immunoassay with an Immulite 2000 XPi immunological analyzer (Siemens Healthineers, Erlangen, Germany).
A standard oral glucose tolerance test with 75 g of glucose was performed in nondiabetic subjects.Blood samples for the measurement of glucose were taken from the cubital vein at 0 and 120 min of the test.
CGM parameters.Patients with T1D underwent real-time CGM within 3-13 days (median 5.4 days) with MMT-722 or MMT-754 monitoring systems and CareLink ® Pro software 2.5.524.0 (version 2.5A, Medtronic, Minneapolis, MN, USA).Study participants were instructed to calibrate the system at least three times a day.The results of the CGM of each subject were reviewed individually and recording defects were eliminated.
Serum cytokines and growth factors.Serum samples for the assay of cytokines and growth factors were obtained from the fasting blood and stored at −80 • C until the analysis.Repeated freeze-thaw cycles were avoided.
A multiplex bead array assay was performed according to the manufacturer's instructions.Serum samples were centrifuged at 10,000× g for 10 min at 4 • C, diluted 1:4 with Bio-Plex Sample Diluent, and incubated with antibody-coupled beads, detection antibody, and streptavidin for 30 min, 30 min, and 10 min, respectively.After each coating with antigen, the plate was washed with a Bio-Plex Handheld Magnetic Washer, resuspended, and vortexed.Fluorescence was measured on a two-beam laser automated analyzer Bio-Plex ® 200 system.Data were acquired with Bio-Plex Manager Software 4.0.The values below the detection limit were set to zero.

Statistical Analysis
Statistics 12.0 software package (Dell, Round Rock, TX, USA) was used for analysis.The outliers were excluded with Dixon's Q test.Quantitative data are presented as medians (lower quartiles; upper quartiles); frequencies are presented as percentages (%).The Kolmogorov-Smirnov (KS) test was applied to test the normality of data distribution.As most of the studied parameters were not distributed normally, a non-parametric Mann-Whitney U-test was used for group comparisons.Spearman's rank correlation analysis and logistic regression analysis were applied to test the associations between studied parameters.p-values less than 0.05 were considered significant.

Clinical Characteristics of the Study Participants
One hundred and thirty subjects with T1D, fifty-five men and seventy-five women, aged from 18 to 70 years (median 33 years), with diabetes duration from 0.5 to 55 years (median 15 years), were included in the study.Eighty eight patients had normal body mass index (BMI), twenty-four were overweight, and eighteen individuals had obesity.The level of fasting C-peptide in most patients was below the sensitivity limit (<0.1 ng/mL); however, 19 subjects had detectable C-peptide (range: 0.102-1.3ng/mL).Eighty-two individuals received multiple daily injections of insulin and forty-eight were on continuous subcutaneous insulin infusion.Clinical characteristics are presented in Table 1.Twenty-seven non-obese individuals with NGT, twelve men and fifteen women, from 19 to 62 years of age (median 32 years), were included in the control group.
Mean monitoring glucose, CGM-derived TIRs, and GV parameters in subjects with T1D are presented in Table 2.
In addition, there were trends towards an increase in the levels of G-CSF, SCF, TRAIL, and HGF.On the other hand, concentrations of IL-1α, IL-4, and GM-CSF were decreased and MIF demonstrated a tendency to decrease.Other molecules showed no significant changes (Figure 2).filtration rate; HbA1c, glycated hemoglobin A1c; HDL, high density lipoprotein; hsCRP, high sensitivity C-reactive protein; LDL, low-density lipoprotein; RBC, red blood cells; UACR, urinary albumin-to-creatinine ratio; WBC, white blood cells.
Twenty-seven non-obese individuals with NGT, twelve men and fifteen women, from 19 to 62 years of age (median 32 years), were included in the control group.
Mean monitoring glucose, CGM-derived TIRs, and GV parameters in subjects with T1D are presented in Table 2.

Discussion
In this study, we assessed serum levels of 44 cytokines and growth factors that are discussed as mediators of diabetic complications in subjects with T1D depending on CGM-derived TIRs and GV metrics.To determine the circulating regulators, we used multiplex analysis, or multiplex bead array assay.This method makes it possible to simultaneously determine a large number of biomarkers in one biological sample [20,21].Moreover, we performed a comprehensive analysis of the GV [10,27] that included an assessment of time-dependent parameters (TIR, TAR L1, TAR TBR L1, TBR L2), the dispersion of glucose values (CV), the amplitude of glucose fluctuations (MAGE), and the rate of glucose changes (MAG).
When interpreting the levels of cytokines in subjects with T1D, the autoimmune nature of the disease should be taken into account.Some studies have matched the levels of circulating cytokines with markers of the autoimmune process in patients with T1D.It was reported that children with one or two diabetes-related antibodies (IA-2 and/or GAD65) have significantly higher levels of IL-1β and IL-2 and lower levels of IL-4 than children with diabetes who were negative for these markers [40].In patients with recentonset T1D, an increase in IL-18 and decrease in MIF and MCP-1 levels were associated with IA-2 and GAD65 antibody positivity [42].Another study showed a dependency of accelerated autoimmunity and beta cell destruction on increased IFN-γ, IL-12, and IL-17 and decreased IL-4, IL-6, and IL-13 in pediatric patients with T1D [37].The onset of T1D in children was characterized by the upregulation of GM-CSF, IL-1β, IL-7, IL-8, IL-10, IL-17F, IL-21, IL-23, and IL-27, but not IL-6 or TNF-α; the presence of autoantibodies (anti-IA-2 and -ZnT8) influenced the blood cytokine levels [43].In this study, we did not test patients for autoimmune markers of diabetes but assessed fasting C-peptide levels.Though we found negative correlations between fasting C-peptide and IL-1β, IL-2Rα, and IL-16 concentrations, it should be emphasized that the majority of patients in our study had long-term diabetes with undetectable beta-cell function.Therefore, the autoimmune process hardly played a significant role in changing the cytokine profile in our cohort.
The greatest changes in the levels of the studied molecules were revealed in patients with TIR values < 70%.Compared to individuals with target TIR, these patients showed higher concentrations of IL-1β, IL-1Ra, IL-2Rα, IL-3, IL-6, IL-7, IL-12 p70, IL-16, IL-17A, G-CSF, M-CSF, IFN-γ, MCP-1, MCP-3, TNF-α, TNF-β, SCF, and HGF, and lower levels of IL-1α, IL-4, IL-8, IL-10, GM-CSF, and MIF.Associations of IL-1β, IL-4, IL-10, IL-12 (p70), GM-CSF, HGF, MCP-3, and TNF-α with target TIR were confirmed in a multivariate logistic analysis.It is interesting to note that many of the studied molecules showed associations with TIR, but not with HbA1c.Naturally, TIR reflects glucose levels over a much shorter period than HbA1c.Therefore, it can be hypothesized that recent glucose fluctuations may have a greater impact on cytokine production than long-term glycemic control.Our data on the associations of TIR with the levels of important biological regulators represent another argument for the widespread use of TIR in assessing glycemic control in people with diabetes.
In our study, many molecules were associated with time-dependent parameters reflecting hyperglycemic fluctuations (TAR L1 and TAR L2), and some molecules (IL-1β, IL-8, IL-12 p70, MCP-1, RANTES, MIF, SDF-1α) demonstrated correlation(s) with at least one amplitude-dependent GV metric (CV, MAGE, and/or MAG indices).It has previously been shown that intermittently high glucose enhances the secretion of IL-6 and TNF-α by activated monocytes [53].Moreover, it has been demonstrated that glucose oscillations stimulate IL-6 production in endothelial cells to a greater extent than persistently high glucose [54,55].Similarly, acute glucose fluctuations were a more potent inducer of IL-1β and TNF-α in rat podocytes than constantly high glucose [56].In adipocytes, oscillating glucose induced a greater increase in MCP-1 secretion than constantly high glucose [57].In rats, blood glucose fluctuations induced by intermittent glucose infusions increased the expression of IL-6 and TNF-α in vascular endothelial cells [58].Therefore, the effect of hyperglycemia on cytokine synthesis may be exacerbated by that of GV.
Hypoglycemia may act as an additional trigger of inflammation in individuals with high GV.In cultured macrophages, intermittent episodes of hypoglycemia and hyperglycemia promote M1 polarization and enhance IL-1β, TNF-α, IL-6, and MCP-1 secretion [59].In healthy individuals and those with T1D, hypoglycemia promotes leukocyte mobilization into the bloodstream and induces proinflammatory changes in these cells [60].In individuals with T1D, an episode of two-hour hypoglycemia was followed by an increase in the levels of IL-6 [61].High blood glucose, replacing hypoglycemia, caused a further increase in the concentrations of IL-6 [62].Contrary to these data, we found weak negative correlations between TBR and some studied cytokines.This can be explained by associations of these molecules with TIR and TAR, which are in reciprocal relationships in TBR.It should also be noted that all hypoglycemic episodes in our study were non-severe and may not have affected cytokine levels.
When discussing possible confounders affecting cytokine levels in patients with T1D, BMI should be taken into account [63].The prevalence of obesity among people with T1D is increasing [64] and nearly a quarter of T1D patients are affected by metabolic syndrome [65].In our study, patients with excessive BMI had higher levels of IL-1β, IL-12 p70, IP-10, LIF, MCP-3, β-NGF, and TNF-α, and lower levels of IL-10.However, associations of IL-1β, IL-10, IL-12 p70, MCP-3, and TNF-α with TIR remained significant after adjustment to BMI, as well as other possible confounders (age, sex, diabetes duration, and eGFR).This demonstrates an independent association of excessive glucose oscillations with the changes in cytokine levels.
Our study is not without limitations.The recruitment of patients in one center could lead to some bias in the sample selection.A cross-sectional design does not prove the causality.The CGM duration was rather short.However, to the best of our knowledge, this is the first study that examined a panel of circulating cytokines and growth factors depending on GV parameters in T1D patients.The strengths of the study are a comprehensive analysis of GV with time-dependent and amplitude-dependent metrics and a determination of the relationship of these parameters with the levels of a large number of cytokines and growth factors.As a result, we have identified the regulators most associated with

Table 1 .
Clinical characteristics of the study participants with T1D.

Table 2 .
CGM-derived TIRs and GV parameters in subjects with T1D.

Table 2 .
CGM-derived TIRs and GV parameters in subjects with T1D.

Table 3 .
Serum concentrations (pg/mL) of cytokines and growth factors in T1D subjects depending on TIR.

Table 4 .
Serum concentrations (pg/mL) of cytokines and growth factors in T1D subjects depending on CV.

Table 5 .
Serum cytokines and growth factors associated with TIR ≤ 70% and CV ≥ 36% in subjects with T1D.