Abstract
Background
Impaired insulin sensitivity (IS) predicts complications and mortality in type 1 diabetes (T1D). IS improves shortly after islet cell transplant for type 1 diabetes, yet long-term changes in IS, and associated factors such as patient characteristics, transplant factors, clinical management, and IS-related biomarkers are unknown.
Methods
Up to nine years (mean four) of longitudinal data were available on 22 adults (18 female) with T1D who received 1–3 transplants in Phase 1/2 or 3 clinical trials (2004–2014). Metabolic testing post-transplant estimated IS by the Homeostasis Model Assessment for Insulin Resistance (HOMA-IR; 111 observations) and the Simple Index of Insulin Sensitivity (SIis; 95 observations).
Results
SIis significantly increased the first year post-transplant (p=0.02), then stabilized (p=0.39); HOMA-IR remained stable post-transplant (p=0.92). Adjusting for age and BMI, higher SIis was associated with lower HbA1c following transplant (p=0.03). Greater IS as measured by lower HOMA-IR and higher SIis was associated with lower fasting C-peptide (both p≤0.04), and also with higher exenatide dose (both p≤0.01). More islets transplanted was associated with higher SIis (p<0.0001). Lower leptin at transplant predicted lower HOMA-IR and higher SIis after transplant, and lower bone marker RANKL predicted lower HOMA-IR (all p≤0.01).
Conclusions
IS measured by SIis was improved several years following transplant, while IS measured by HOMA-IR did not worsen. Higher exenatide dose, more islets transplanted, and diet and exercise (lowering leptin and RANKL) may improve IS, which may enhance glycemic control and lower metabolic demand on transplanted islets. Long-term clamp studies are needed to confirm these results.
Keywords: type 1 diabetes, islet cell transplantation, insulin sensitivity, leptin, receptor activator of nuclear factor kappa-B ligand (RANKL), exenatide
Introduction
Impaired insulin sensitivity (IS) is common not only in type 2 diabetes but also type 1 diabetes (T1D)1. Tissue uptake of blood glucose is lower than normal, resulting in hyperglycemia despite higher endogenous or exogenous insulin levels, which may lead to complications such as retinopathy, nephropathy, and cardiovascular disease2,3. IS can be quantified using the gold-standard hyperinsulinemic-euglycemic (HIEG) clamp4, but the procedure is labor- and time-intensive. Insulin resistance or IS can also be estimated using validated mathematical models such as the Homeostasis Model Assessment for Insulin Resistance (HOMA-IR)5 and the Simple Index of Insulin Sensitivity from the oral glucose tolerance test (SIisOGTT)6, both highly correlated with HIEG clamp results in nondiabetic individuals5–7. However, as neither HOMA-IR nor SIisOGTT are valid measures in T1D due to the lack of endogenous insulin, an alternative model of IS called the estimated glucose disposal rate (eGDR) was previously derived specifically for patients with T1D8. Impaired IS calculated using eGDR has been shown to be clinically significant in T1D as it is strongly associated with longer T1D duration9 and predicts diabetes complications and mortality10,11.
To attenuate the burden of hyperglycemia and complications in T1D, islet cell transplantation is a minimally invasive therapy currently being tested in clinical trials as a potential durable treatment for T1D12,13. Briefly, islets are isolated from a deceased human donor pancreas, then transplanted into a T1D patient’s liver via the portal vein13. The goal is to provide a sufficient number of islets to restore adequate endogenous insulin to control blood glucose and eliminate severe hypoglycemic episodes. However, not all recipients maintain long-term insulin independence, and some experience graft failure14. Therefore, optimizing recipients’ insulin sensitivity could potentially lead to more durable islet graft function by placing less metabolic demand on the transplanted beta-cells. Indeed, research has shown that islet transplant improves IS as measured by the HIEG clamp within 6–7 months following transplant15.
However, to the best of our knowledge, longitudinal data are lacking describing both long-term changes in IS, and predictors of IS, in patients with T1D who have undergone long-term successful islet transplantation. A better understanding of clinical, modifiable, and/or physiological factors associated with recipients’ improved IS could inform future research on interventions aimed at providing more durable islet graft function. Therefore, using longitudinal granular data from patients enrolled in islet transplant clinical trials13, the current study: 1) investigated changes in IS using HOMA-IR and SIisOGTT, up to a mean of four years following islet transplant; and 2) determined the longitudinal associations of IS with patient characteristics, transplant factors, clinical management strategies, and select biomarkers known to be associated with IS in nondiabetic individuals (leptin and adiponectin16, insulin-like growth factor [IGF-1]17, and bone-derived markers: receptor activator of nuclear factor kappa-B ligand [RANKL]18 and osteocalcin [OC]19).
Materials and Methods
Study Design & Participants
The current study is a secondary analysis of existing data extracted from the medical records of 22 patients with T1D who: 1) were enrolled in the University of Illinois at Chicago’s (UIC) islet transplant Phase 1/2 (n=10) or Phase 3 (n=12) clinical trial (NCT00679042) between 2004–2014; and 2) had archived serum tested for biomarkers as part of a prior ancillary study in 2013. The clinical trials have previously been described13. Briefly, eligibility criteria included 18–70 years old with T1D for more than five years, experiencing severe hypoglycemic episodes with hypoglycemic unawareness despite optimal therapy. Exclusion criteria included ongoing cardiovascular disease, history of non-skin cancer, other untreated clinically significant comorbidities, previous organ transplant, active smoking, body mass index (BMI) >26 (Phase 1/2) or >27 (Phase 3), insulin requirement ≥0.7 U kg−1 day−1, or HbA1c ≥12% (108 mmol/mol). Patients received up to three islet transplants to achieve insulin independence. In sum, nine patients received one, six patients received two, and seven patients received three transplants. The first four patients received the Edmonton Protocol for immunosuppression12; the remaining 18 received the UIC protocol13, which is the Edmonton Protocol plus etanercept and exenatide. Mycophenolate mofetil was substituted for sirolimus if patients presented with intolerable side effects. Management of blood pressure and lipids adhered to the concurrent American Diabetes Association guidelines20,21.
Pre-transplant data used in the current analysis were those closest to the initial transplant. Follow-up after initial islet transplant ranged up to 9.4 years. At the end of 2014, ten of the 22 patients were no longer active in the clinical trials due to: graft failure (n=6), adverse events (n=1), malignancy (n=1), death (n=1), or withdrawal (n=1). However, all of their available data pre- and post-transplant were included in the current analysis. UIC’s Institutional Review Board approved the clinical trials and the ancillary study, and written informed consent was obtained.
HOMA-IR and SIisOGTT – Dependent Variables
To calculate HOMA-IR after transplant, fasting glucose and insulin levels were obtained prior to multiple metabolic tests (OGTT, mixed meal test [MMT], intravenous glucose tolerance test [IVGTT], and glucagon stimulation test [GST])5. To calculate SIisOGTT after transplant, both fasting and stimulated glucose and insulin levels obtained only from the OGTT were analyzed6. Glucose was measured by the glucose oxidase method (UIC Pathology Laboratory, Chicago, IL) and insulin by quantitative chemiluminescent immunoassay (ARUP Laboratories, Salt Lake City, UT). As neither HOMA-IR nor SIisOGTT are valid in T1D, neither was calculated pre-transplant.
OGTT
Glucose and insulin were measured in blood samples taken while fasting, then 30, 60, 90, and 120 minutes after drinking 75gm of glucose (in 300mL solution) within five minutes. This test was scheduled 6, 24, and 52 weeks after each islet transplant, then yearly. Fasting and stimulated levels from 95 OGTTs were used to calculate SIisOGTT, and fasting levels from 47 to calculate HOMA-IR.
MMT, IVGTT, GST
These tests were scheduled 6 (except MMT), 24 and 52 weeks after each islet transplant and every year thereafter. Fasting levels from 18 MMTs, 20 IVGTTs, and 26 GSTs were used to calculate additional HOMA-IR values. The few cases when metabolic testing was also conducted in the days following transplant provided additional data.
Patient Characteristics, Transplant and Clinical Factors – Independent Variables
Patient characteristics included age and duration of T1D at initial transplant, sex, and menopausal status across follow-up. Transplant variables included clinical trial phase, number of transplants, cumulative number of islet equivalents transplanted per kg body weight (IEQs/kg), and immunosuppressant and exenatide doses from each visit. Clinical variables from each visit included BMI, blood pressure after five minutes at rest, insulin dose, and antihypertensive and statin use. Waist and hip circumferences were not part of the clinical trial protocol, so were not available. A comprehensive metabolic profile, which included fasting glucose, was performed at each visit (UIC Pathology Laboratory). Fasting C-peptide was also measured at each visit using quantitative chemiluminescent immunoassay (ARUP Laboratories). Beta-cell function was calculated at the same time points as SIisOGTT as the area under the curve of fasting and stimulated (30, 60, 90, and 120 minute) C-peptide values. Complete C-peptide data were only available in a subgroup of 70 OGTTs. HbA1c was measured every three months by high performance liquid chromatography (UIC Pathology Laboratory). Estimated glomerular filtration rate (eGFR) was calculated using serum creatinine (part of the comprehensive metabolic profile) from each visit (Jaffe reaction, UIC Pathology Laboratory) using the CKD-EPI equation22. An albumin-to-creatinine ratio (Beckman LX20 standard chemistry method, UIC Pathology Laboratory) was obtained every six months. Lipids and inflammatory markers were measured in fasting morning blood samples at pre-transplant, 1, 20, and 52 weeks after each islet transplant, and yearly thereafter (Clinical Reference Laboratory, Lenexa, KS). The lipid panel included LDL, HDL, total cholesterol, triglycerides, and non-esterified fatty acids (NEFAs). Inflammatory markers included plasminogen activator inhibitor-1 (PAI-1) antigen and activity, vascular cell adhesion molecule-1 (VCAM-1), intercellular adhesion molecule-1 (ICAM-1), matrix metalloproteinase-9 (MMP-9), monocyte chemoattractant protein-1 (MCP-1), and high-sensitivity C-reactive protein (CRP). T1D autoantibodies were measured pre-transplant, three times in the first year after initial transplant, and yearly thereafter, and included anti-islet cell (by semi-quantitative indirect fluorescent antibody assay), glutamic acid decarboxylase (by semi-quantitative enzyme-linked immunosorbent assay), IA2 and anti-insulin (both by quantitative radioimmunoassay) (ARUP Laboratories).
Biomarkers – Independent Variables
As part of the clinical trials, serum was archived at −80°C within one hour of the morning fasting blood draw, at three time points: pre-transplant, and 1 and 20 weeks after each islet transplant. As part of an ancillary study in 2013, the archived serum samples from the first transplants only (baseline) were analyzed. Eleven patients had serum available from all three time points, eight from two, and three from one. There were 18 pre-transplant samples (mean [standard deviation (SD)]: 28 [24] weeks prior to first transplant), 20 1-week post-transplant samples (5 [2] days), and 14 20-week post-transplant samples (20 [2] weeks). The following biomarkers were measured: undercarboxylated OC and carboxylated OC by enzyme immunoassay (EIA, Takara Bio, Clontech Laboratories, Mountain View, CA); soluble RANKL by Human sRANKL (TOTAL) ELISA (BioVendor LM, Asheville, NC); IGF-1 by Quantine ELISA Human IGF-1 Immunoassay (R&D Systems, Minneapolis, MN); and leptin and adiponectin by high-sensitivity EIA (R&D Systems). Total OC was calculated by summing the levels of undercarboxylated and carboxylated OC. Patients’ mean baseline biomarker levels were calculated using all data available from the three time points.
Statistical Analysis
All analyses were performed using SAS (version 9.4, Cary, NC). A p-value <0.05 was considered statistically significant. The distribution of continuous variables was assessed and transformed if skewed, such as the log transformation of HOMA-IR and SIisOGTT. Frequencies and percentages were calculated for categorical variables, means and SDs for continuous variables. Histograms and scatter plots identified the following extreme influential outliers that were excluded from the analysis: two log-transformed SIisOGTT values, and one baseline mean value each for adiponectin, undercarboxylated OC, and carboxylated OC. For one patient, a single leptin value was excluded and a new mean leptin was calculated.
Longitudinal changes (i.e., the slope) during follow-up in biomarkers, HOMA-IR, SIisOGTT, and other metabolic outcomes were determined using mixed-effects linear regression modeling of repeated measures. Mixed-effects modeling was used to account for the non-independence of repeated measures within each patient. The autoregressive variance matrix was specified to account for the diminishing within-person correlation of observations that are farther apart in time. To determine the longitudinal association between HOMA-IR and SIisOGTT, each were standardized by dividing by their respective SD then regressed using the same procedure.
Multivariable mixed-effects linear regression modeling (both forwards and backwards stepwise) was then used to assess for significant associations of the slope of change in HOMA-IR or SIisOGTT with the independent variables; both final models were adjusted for age and BMI. Non-linear (i.e., quadratic) relationships of HOMA-IR or SIisOGTT with significant biomarkers and continuous independent variables were determined and retained if significant. Effect modification of the associations by patient sex, age, and BMI was also tested and retained if significant.
Results
Patient Characteristics
The majority were female (n=18), and all but one were Non-Hispanic white. The mean (SD) age and duration of T1D at the time of initial transplant were 45.3 (11.9) years and 27.9 (11.8) years, respectively (Table 1). Both pre-transplant and post-transplant, patients were on average normotensive and had normal lipid levels, though many were using anti-hypertensives (n=12 pre/14 post) and/or statins (n=12 pre/17 post). BMI pre- and post-transplant was within the normal range. eGFR levels were, on average, above 80 mL min−1 1.73(m2) −1 pre- and post-transplant. HbA1c and insulin dose were higher pre-transplant (7.2%, 55.7 mmol/mol; 0.50 U kg−1 day−1) compared to post-transplant (6.2%, 43.8 mmol/mol; 0.10 U kg−1 day−1). The median (25th; 75th percentile) number of IEQs/kg transplanted was 14,309 (7,900; 25,931).The mean (range) follow-up was 4.2 (0–9.4) years.
Table 1.
Characteristics of Study Sample (n=22)
| Characteristic | Pre-Transplant | Post-Transplant | ||
|---|---|---|---|---|
| Age (years), mean (SD) | 45.3 | (11.9) | ||
| Female sex, n (%) | 18 | (82) | ||
| Pre-menopausal, n (%) | 10 | (56) | 7 | (39) |
| BMI (kg/m2), mean (SD) | 23.3 | (1.9) | 22.1 | (1.9) |
| Duration of T1D (years), mean (SD) | 27.9 | (11.8) | ||
| HbA1c, mean (SD) a | ||||
| % | 7.2 | (0.9) | 6.2 | (0.7) |
| mmol/mol | 55.7 | (9.7) | 43.8 | (7.5) |
| Insulin dose (units/kg), mean (SD) | 0.50 | (0.16) | 0.10 | (0.16) |
| Blood pressure (mmHg), mean (SD) | ||||
| Systolic | 115.8 | (12.8) | 125.8 | (15.8) |
| Diastolic | 67.3 | (8.0) | 72.8 | (9.3) |
| Antihypertensive use, n (%) | 12 | (55) | 14 | (64) |
| Cholesterol (mg/dL), mean (SD) | ||||
| Total | 164.8 | (34.7) | 182.1 | (43.4) |
| HDL | 64.5 | (19.7) | 64.7 | (26.1) |
| LDL | 85.5 | (27.5) | 98.1 | (32.1) |
| Triglycerides | 76.5 | (60.5) | 94.5 | (45.4) |
| Statin use, n (%) | 12 | (55) | 17 | (77) |
| eGFR (mL min−1 1.73(m2)−1), mean (SD) | 82.2 | (20.1) | 81.1 | (22.1) |
| Number of transplants received, n (%) | ||||
| 1 | 9 | (41) | ||
| 2 | 6 | (27) | ||
| 3 | 7 | (32) | ||
| Total IEQs/kg, median (25th; 75th percentile) | 14,309 | (7,900; 25,931) | ||
| Protocol, n (%) | ||||
| Edmonton | 4 | (18) | ||
| UIC | 18 | (82) | ||
| Phase, n (%) | ||||
| 1/2 | 10 | (45) | ||
| 3 | 12 | (55) | ||
| Follow-up time (years), mean (SD) | 4.2 | (2.9) | ||
Pre=calculated using closest value prior to first transplant; Post=calculated using all data available post-first transplant
HbA1c pre-transplant, n = 21
Biomarkers
Table 2 presents the average levels of each biomarker at the three time points and mean changes over time. There were no significant changes in undercarboxylated OC, carboxylated OC, total OC, or IGF-1. There were trends towards a decrease in RANKL (p=0.09) and an increase in adiponectin (p=0.06). There was a significant decrease in leptin (p<0.0001), specifically from pre- to 1-week post-transplant.
Table 2.
Changes in Biomarker Levels During Follow-up (22 patients with 52 observations)
| Biomarker, mean (SD) | Pre-Transplant n=18 | 1 week Post n=20 | 20 weeks Post n=14 | β per 10 weeks (SE) | p-value | ||||
|---|---|---|---|---|---|---|---|---|---|
| Undercarboxylated OC (ng/mL) | 6.1 | (5.0) | 5.5 | (2.9) | 3.9 | (6.0) | −0.10 | (0.14) | 0.50 |
| Carboxylated OC (ng/mL) | 23.5 | (14.0) | 21.9 | (14.8) | 24.1 | (12.6) | 0.16 | (0.39) | 0.69 |
| Total OC (ng/mL) | 29.6 | (14.4) | 27.4 | (14.8) | 28.0 | (13.3) | 0.06 | (0.41) | 0.88 |
| RANKL (pmol/L) | 4.0 | (5.2) | 4.1 | (4.4) | 3.4 | (3.4) | −0.24 | (0.13) | 0.09 |
| IGF-1 (ng/mL) | 98.6 | (38.4) | 95.2 | (33.9) | 96.1 | (34.8) | −0.40 | (1.20) | 0.74 |
| Leptin (ng/mL) | 43.2 | (23.3) | 19.3 | (10.4) | 19.4 | (10.5) | −4.34 | (0.83) | < 0.0001 |
| Adiponectin (ng/mL) | 7464 | (3700) | 8483 | (3370) | 9012 | (3140) | 254 | (127) | 0.06 |
Metabolic Outcomes
Table 3 presents changes in metabolic outcomes: (1) pre-transplant throughout all follow-up, and (2) post-transplant follow-up only. Overall post-transplant, no changes were observed in HOMA-IR or SIisOGTT. HOMA-IR (log) decreased but not significantly within one year after initial transplant (β=−0.12 per 10 weeks, p=0.33; Figure 1a). In contrast, within the first year after initial transplant, there was a significant increase in SIisOGTT (log) (β=0.013 per 10 weeks, p=0.02), which remained stable at this higher level throughout the remainder of follow-up (β=0.001 per 10 weeks, p=0.39; Figure 1b). There was only a moderate, though significant, negative association between HOMA-IR and SIisOGTT (β=−0.33, p=0.002). Fasting glucose and HbA1c decreased significantly from pre- to post-transplant, but only HbA1c decreased significantly during post-transplant only. Fasting C-peptide increased significantly from pre- to post-transplant, but did not change during post-transplant only. No significant changes were observed in fasting insulin or BMI.
Table 3.
Changes in IS and Other Metabolic Outcomes During Follow-up (22 patients)
| Outcome | Pre- thru Post-Transplant | Post-Transplant Only | ||||||
|---|---|---|---|---|---|---|---|---|
| β per 10 weeks (SE) | n | p-value | β per 10 weeks (SE) | n | p-value | |||
| HOMA-IR (log) | – | – | – | 0.0009 | (0.009) | 111 | 0.92 | |
| SIisOGTT (log) | – | – | – | 0.0003 | (0.001) | 95 | 0.71 | |
| Fasting glucose (mg/dL; log) | −0.006 | (0.002) | 151 | 0.01 | −0.002 | (0.002) | 130 | 0.48 |
| Fasting C-peptide (ng/mL) | 0.017 | (0.006) | 151 | 0.003 | 0.004 | (0.006) | 130 | 0.47 |
| Fasting insulin (μU/mL; log) | 0.0004 | (0.008) | 117 | 0.96 | 0.001 | (0.008) | 111 | 0.92 |
| HbA1c | ||||||||
| % | −0.03 | (0.01) | 138 | 0.0006 | −0.01 | (0.01) | 117 | 0.04 |
| mmol/mol | −0.27 | (0.08) | 138 | 0.0006 | −0.15 | (0.07) | 117 | 0.04 |
| BMI (kg/m2) | −0.01 | (0.02) | 147 | 0.43 | 0.01 | (0.02) | 126 | 0.78 |
n=number of longitudinal observations
Figure 1.
Individual patient change in a) HOMA-IR (inset log-transformed), and b) SIisOGTT (inset log-transformed) during follow-up post-transplant. Dashed line indicates 1 year follow-up.
Associations with HOMA-IR
Table 4 presents factors associated with HOMA-IR, adjusted for age and BMI. With regard to biomarkers, HOMA-IR after transplant was positively associated with mean leptin at baseline. In addition, mean RANKL at baseline demonstrated a quadratic association with HOMA-IR after transplant (p=0.001; Figure 2a), such that HOMA-IR was slightly negatively associated with RANKL at the lowest levels of RANKL, while positively associated with RANKL at higher levels of RANKL. Fasting C-peptide also had a quadratic association with HOMA-IR (p=0.01; Figure 3a), with a stronger positive association at higher fasting C-peptide levels. Exenatide dose was negatively associated with HOMA-IR in males, but had a null association in females (sex interaction p=0.01). However, female sex had a negative association with HOMA-IR (i.e., they were more insulin sensitive than males, p=0.001). Further adjusting for the type of metabolic test used to calculate HOMA-IR did not change the results.
Table 4.
Factors Associated with HOMA-IR and SIisOGTT (adjusted for age and BMI; n=22)
| Independent Variable | HOMA-IR (log) 107 observations | SIisOGTT (log) 85 observations | ||||
|---|---|---|---|---|---|---|
| β | (SE) | p-value | β | (SE) | p-value | |
| Intercept | 3.30 | (0.89) | 0.002 | −1.30 | (0.09) | <0.0001 |
| Mean leptin (ng/mL) | 0.049 | (0.008) | <0.0001 | −0.009 | (0.002) | 0.001 |
| Mean leptin x mean leptin | 0.0001 | (0.00004) | 0.01 | |||
| Mean RANKL (pmol/L) | −0.34 | (0.09) | 0.002 | |||
| Mean RANKL x mean RANKL | 0.025 | (0.006) | 0.001 | |||
| Sex (male=0, female=1) | −1.14 | (0.29) | 0.001 | |||
| Fasting C-peptide (ng/mL) | −0.77 | (0.37) | 0.04 | 0.04 | (0.03) | 0.23 |
| Fasting C-peptide x fasting C-peptide | 0.35 | (0.13) | 0.01 | −0.02 | (0.01) | 0.04 |
| HbA1c (%) | −0.02 | (0.01) | 0.03 | |||
| Exenatide dose (μg) | −0.12 | (0.04) | 0.01 | 0.004 | (0.001) | 0.0001 |
| Exenatide dose x sex | 0.12 | (0.05) | 0.01 | |||
| Cumulative IEQs/kg (per 1000) | 0.003 | (0.001) | <0.0001 | |||
| Pre-menopausal women vs. men | 0.05 | (0.02) | 0.03 | |||
| Post-menopausal women vs. men | −0.02 | (0.02) | 0.43 | |||
Figure 2.
Nonlinear associations between a) HOMA-IR post-transplant and RANKL at baseline, and b) SIisOGTT post-transplant and leptin at baseline
Figure 3.
Nonlinear associations between a) HOMA-IR and fasting C-peptide post-transplant, and b) SIisOGTT and fasting C-peptide post-transplant
Associations with SIisOGTT
Adjusting for age and BMI (Table 4), mean leptin at baseline demonstrated a quadratic association with SIisOGTT post-transplant (p=0.01; Figure 2b), with the negative association being stronger at lower leptin levels, and the association plateauing at higher leptin levels. Fasting C-peptide also demonstrated a quadratic association with SIisOGTT, but in opposite directions from the HOMA-IR association (p=0.04; Figure 3b), with a stronger negative association at higher fasting C-peptide. In a smaller number of observations (n=59), the C-peptide area under the curve from the OGTT replaced fasting C-peptide and demonstrated a negative, but linear, association with SIisOGTT (log) (β=−0.004, p=0.003). SIisOGTT was positively associated with exenatide dose and cumulative IEQs/kg transplanted, while negatively associated with HbA1c. Cumulative IEQs unadjusted for recipient weight was interchangeable with IEQs/kg in the model. Pre-menopausal women were more sensitive to insulin than men, while post-menopausal women and men had similar SIisOGTT.
Variables not associated with either HOMA-IR or SIisOGTT were: duration of T1D, blood pressure, lipids and NEFAs, inflammatory markers, statin and antihypertensive use, insulin dose post-transplant, immunosuppression, eGFR, albumin-to-creatinine ratio, T1D auto-antibodies, OC, adiponectin, and IGF-1. Further, adjusting all the final models for the minimal insulin dose patients may have still required immediately post-transplant did not substantially change the results for either HOMA-IR or SIisOGTT.
Discussion
To the best of our knowledge, this is the first study to investigate long-term changes in IS using HOMA-IR and SIisOGTT in patients with T1D who have received islet cell transplantation (even if requiring small doses of exogenous insulin immediately after the transplants); and to determine clinical, modifiable, and/or physiological factors associated with long-term changes in IS following islet transplant in these patients. In summary, IS as measured by SIisOGTT improved within one year following first transplant, and remained stable at this improved level for up to a mean of four years, whereas IS as measured by HOMA-IR remained stable (i.e., did not worsen) over the entire follow-up. This is particularly important given that decreasing IS is associated with longer T1D duration9. IS after transplant was associated with both potentially modifiable factors and with transplant outcomes. These results need to be confirmed using HIEG clamp studies in the decade following islet transplant, but indicate potential clinical benefits of optimizing IS surrounding transplant.
Our study is in agreement with prior research that also demonstrated IS (peripheral, hepatic, and total) improved immediately post-islet transplant, as quantified by the FSIVGT23 and the HIEG clamp15, however both of these studies were limited to at most one year follow-up. Research in a more heterogeneous sample, where 4 of the 10 participants were islet-after-kidney recipients, similarly found improvements in IS measured by FSIVGT within 12 months following islet transplant24. The beneficial effect that islet transplant has on IS has been hypothesized to be a result of restoring endogenous insulin, specifically the resolution of hyperglycemia and normalization of insulin’s suppression of NEFA production23. The current study interestingly found a significant improvement in IS as measured by SIisOGTT, but no change in IS as measured by HOMA-IR. Further mechanistic research is needed to determine if islet transplant has differential effects on parameters quantified by SIisOGTT’s stimulated metabolic outcomes, specifically muscle and adipose tissue glucose metabolism, versus parameters captured by HOMA-IR, specifically fasting hepatic glucose production6,25. Or perhaps SIisOGTT may be a more sensitive measure of insulin action in this population.
With respect to clinical outcomes, as expected, HbA1c and fasting C-peptide improved after transplant26. Among individuals with T1D, previous research has shown that higher IS (HIEG clamp) was associated with lower glycosylated hemoglobin27. The current study similarly found that for patients undergoing islet transplant for T1D, enhanced IS following islet transplant was associated with improved glycemic control. Lower IS was significantly associated with greater fasting C-peptide, and in a smaller subgroup, with greater stimulated C-peptide as well. In this case, C-peptide, a marker of endogenous insulin secretion, might therefore also be interpreted as a marker for beta-cell demand or stress, since it is generally accepted that impaired IS requires higher levels of insulin, increasing demand on beta-cells28. The quadratic association between fasting C-peptide and IS as measured by both HOMA-IR and SIisOGTT (Figure 3) demonstrates that the positive and negative associations, respectively, are stronger above the normal value of fasting C-peptide, about 1.0 ng/mL, such that it is the patients who are less sensitive to insulin that are placing more demand on their islet grafts’ beta-cells. Therefore, improving patients’ IS surrounding islet transplant could potentially lower the demand on beta-cells, extending islet graft function and survival.
Other islet transplant trials have shown that in patients receiving more than 600,000 IEQs in total, 75–80% reached insulin independence, compared with 55% who received fewer29. The current study found that having more IEQs/kg transplanted was also associated with improved IS. This may be a function of a greater number of IEQs providing a more rapid resolution of hyperglycemia and suppression of NEFA production as previously discussed23. Also demonstrated was a sex difference in the relationship of exenetide with IS as calculated by HOMA-IR, such that exenatide was associated with improved IS among men, but not in women. In comparison, exenatide was associated with improved IS as calculated by SIisOGTT in both sexes. Exenatide has been previously shown to improve IS (HIEG clamp) by 40% in patients with T1D30. The lesser impact of exenatide specifically on IS in women following islet transplant may be a function of their already significantly higher IS compared to the men in the present sample.
The current study is innovative in that it investigated the association of IS post-transplant with IS-related biomarkers measured at baseline. The demonstrated association between HOMA-IR and RANKL (specifically at higher levels of RANKL) is similar to the findings of an experimental study whereby the blocking or deletion of RANKL signaling specifically in the liver in mice led to significantly lower HOMA-IR compared to wild type mice on a high fat diet18. In an epidemiological study, higher baseline circulating levels of RANKL were found to be associated with greater odds of developing type 2 diabetes over a 15 year period18. Further research is needed to determine if the quadratic association between HOMA-IR and RANKL (Figure 2a) reflects various thresholds where there is an optimal circulating range of RANKL, above and below which IS decreases. Regarding leptin, similar to the current study, previous literature found a positive association between leptin and HOMA-IR among nondiabetic individuals16. The current study also found a negative, quadratic association between leptin and SIisOGTT (Figure 2b). Again, further research is needed to clarify if the association is capturing a threshold effect where below a certain leptin concentration, the negative association of leptin with IS is stronger.
Based on our findings, islet transplant outcomes may be optimized by considering both transplant and patient modifiable factors that enhance patients’ IS. Maximizing the number of IEQs transplanted in the patient’s first transplant may result in better IS, which then may lead to lower fasting C-peptide and better glycemic control as found in this study, as well as potentially longer lasting graft function. Prescribing and titrating exenatide to the maximum tolerable dose may be an option for improving IS after transplant, and potentially pre-transplant. Another option for enhancing patients’ long-term IS after transplant may be lowering leptin and RANKL levels surrounding first transplant by prescription of healthy diet31 and exercise31,32.
Although this study does not utilize the gold standard for measuring IS (HIEG was not part of the clinical trial protocols), the employed metrics, HOMA-IR and SIisOGTT, have been documented to be highly correlated with HIEG clamp measures in nondiabetic, non-obese individuals5,7. The trends in, and associations with, HOMA-IR and SIisOGTT remained robust even after adjusting for the minimal exogenous insulin therapy patients were requiring shortly after the islet transplants, supporting the use of these mathematical models in this patient population after T1D was reversed. Just as HOMA and SIisOGTT were not calculated pre-transplant as they are not valid in the presence of T1D, eGDR was not calculated post-transplant as it was specifically derived and validated for use in those with T1D8. As a standard IVGTT (not a FSIVGT) was part of the clinical trial protocols, with fewer blood samples and without a bolus of insulin, we were not able to calculate the Insulin Sensitivity Index using minimal modeling. As is the case for all islet transplant clinical trials, a control group was not available, because it is unethical to randomize patients to either islet transplant or an alternative treatment for hypoglycemic unawareness given that in order to be eligible for islet transplantation, the patients would have already failed such existing treatments. Furthermore, patients cannot serve as their own control because neither HOMA-IR nor SIisOGTT can be calculated pre-transplant as these measures are not valid in patients with T1D. However, looking at existing controls in the literature, it has been demonstrated that IS decreases with longer duration of diabetes in those living with T1D9, indicating the improvement in SIisOGTT and the stabilization of HOMA-IR after transplant may be clinically important. The sample had limited diversity, with mostly Non-Hispanic white females, limiting the generalizability of these findings. Because the primary endpoints for islet transplant clinical trials are safety and efficacy, diet and exercise data were not obtained. A strength of this study was its use of longitudinal data, with a mean of four years of follow-up. With the smaller sample size, all available data were able to be analyzed despite unequal follow-up times using the mixed-effects modeling of repeated measures.
In summary, IS as measured by SIisOGTT improved within one year of functionally curing T1D via islet transplant and remained improved, while IS as measured by HOMA-IR did not worsen, over several years of follow-up. Furthermore, improved IS was associated with better glycemic control and lower fasting C-peptide. However, these results need to be confirmed using HIEG clamp studies over the decade following islet cell transplant. With the long-term goal being to lessen the requirement of islets transplanted along with the need for subsequent islet transplants, utilizing clinical and behavioral modifications pre- and post-transplant to improve patients’ IS may be considered. Future research is needed to confirm the clinical significance of such interventions on subsequent clinical outcomes in this patient population.
Acknowledgments
We thank Jessica M. Madrigal, MS, at UIC for her assistance with data management. Parts of this study were presented orally at the 77th Scientific Sessions of the American Diabetes Association, San Diego, CA.
Funding
The Chicago Diabetes Project; UIC’s Campus Research Board Pilot Research Program; UIC’s Center for Clinical and Translational Science (CCTS), UL1RR029879 & UL1TR002003 from the National Center for Research Resources; American Diabetes Association’s Innovative Clinical or Translational Science Award 1-16-ICTS-022; and UIC’s Building Interdisciplinary Research Careers in Women’s Health, K12HD055892 from the National Institute of Child Health and Human Development and the National Institutes of Health Office of Research on Women’s Health.
Footnotes
Disclosure Summary
The authors have nothing to disclose.
References
- 1.DeFronzo RA, Hendler R, Simonson D. Insulin resistance is a prominent feature of insulin-dependent diabetes. Diabetes. 1982;31:795–801. doi: 10.2337/diab.31.9.795. [DOI] [PubMed] [Google Scholar]
- 2.The DCCT Research Group. The effect of intensive treatment of diabetes on the development and progression of long-term complications in insulin-dependent diabetes mellitus. N Engl J Med. 1993;329:977–986. doi: 10.1056/NEJM199309303291401. [DOI] [PubMed] [Google Scholar]
- 3.Lachin JM, Orchard TJ, Nathan DM. Update on cardiovascular outcomes at 30 years of the Diabetes Control and Complications Trial/Epidemiology of Diabetes Interventions and Complications study. Diabetes Care. 2014;37:39–43. doi: 10.2337/dc13-2116. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.DeFronzo RA, Tobin JD, Andres R. Glucose clamp technique: a method for quantifying insulin secretion and resistance. Am J Physiol. 1979;237:E214–223. doi: 10.1152/ajpendo.1979.237.3.E214. [DOI] [PubMed] [Google Scholar]
- 5.Matthews DR, Hosker JP, Rudenski AS, Naylor BA, Treacher DF, Turner RC. Homeostasis model assessment: insulin resistance and beta-cell function from fasting plasma glucose and insulin concentrations in man. Diabetologia. 1985;28:412–419. doi: 10.1007/BF00280883. [DOI] [PubMed] [Google Scholar]
- 6.Bastard JP, Vandernotte JM, Faraj M, Karelis AD, Messier L, Malita FM, Garrel D, Prud'homme D, Rabasa-Lhoret R. Relationship between the hyperinsulinemic-euglycaemic clamp and a new simple index assessing insulin sensitivity in overweight and obese postmenopausal women. Diabetes Metab. 2007;33:261–268. doi: 10.1016/j.diabet.2007.02.004. [DOI] [PubMed] [Google Scholar]
- 7.Antuna-Puente B, Disse E, Lavoie ME, Rabasa-Lhoret R, Laville M, Bastard JP. Evaluation of the SIisOGTT index in a healthy population with normal glucose tolerance. Diabetes Metab. 2009;35:236–237. doi: 10.1016/j.diabet.2009.02.004. [DOI] [PubMed] [Google Scholar]
- 8.Williams KV, Erbey JR, Becker D, Arslanian S, Orchard TJ. Can clinical factors estimate insulin resistance in type 1 diabetes? Diabetes. 2000;49:626–632. doi: 10.2337/diabetes.49.4.626. [DOI] [PubMed] [Google Scholar]
- 9.Teixeira MM, Diniz Mde F, Reis JS, Abreu Ferrari TC, Bravim de Castro MG, Teixeira BP, da Silva Anantes IC, Bicalho DM, Foscolo RB. Insulin resistance and associated factors in patients with Type 1 Diabetes. Diabetol Metab Syndr. 2014;6:131. doi: 10.1186/1758-5996-6-131. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Pop A, Clenciu D, Anghel M, Radu S, Socea B, Mota E, Mota M, Panduru NM. Insulin resistance is associated with all chronic complications in type 1 diabetes. J Diabetes. 2016;8:220–228. doi: 10.1111/1753-0407.12283. [DOI] [PubMed] [Google Scholar]
- 11.Olson JC, Erbey JR, Williams KV, Becker DJ, Edmundowicz D, Kelsey SF, Tyrell KS, Orchard TJ. Subclinical atherosclerosis and estimated glucose disposal rate as predictors of mortality in type 1 diabetes. Ann Epidemiol. 2002;12:331–337. doi: 10.1016/s1047-2797(01)00269-1. [DOI] [PubMed] [Google Scholar]
- 12.Shapiro AM, Lakey JR, Ryan EA, Korbutt GS, Toth E, Warnock GL, Kneteman NM, Rajotte RV. Islet transplantation in seven patients with type 1 diabetes mellitus using a glucocorticoid-free immunosuppressive regimen. N Engl J Med. 2000;343:230–238. doi: 10.1056/NEJM200007273430401. [DOI] [PubMed] [Google Scholar]
- 13.Gangemi A, Salehi P, Hatipoglu B, Martellotto J, Barbaro B, Kuechle JB, Qi M, Wang Y, Pallan P, Owens C, Bui J, West D, Kaplan B, Benedetti E, Oberholzer J. Islet transplantation for brittle type 1 diabetes: the UIC protocol. Am J Transplant. 2008;8:1250–1261. doi: 10.1111/j.1600-6143.2008.02234.x. [DOI] [PubMed] [Google Scholar]
- 14.Barton FB, Rickels MR, Alejandro R, Hering BJ, Wease S, Naziruddin B, Oberholzer J, Odorico JS, Garfinkel MR, Levy M, Pattou F, Berney T, Secchi A, Messinger S, Senior PA, Maffi P, Posselt A, Stock PG, Kaufman DB, Luo X, Kandeel F, Cagliero E, Turgeo NA, Witkowski P, Naji A, O’Connell PJ, Greenbaum C, Kudva YC, Brayman KL, Aull MJ, Larsen C, Kay TW, Fernandez LA, Vantyghem M-C, Bellin M, Shapiro AM. Improvement in outcomes of clinical islet transplantation: 1999–2010. Diabetes Care. 2012;35:1436–1445. doi: 10.2337/dc12-0063. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Rickels MR, Kong SM, Fuller C, Dalto-Bakes C, Ferguson JF, Reilly MP, Teff KL, Naji A. Improvement in insulin sensitivity after human islet transplantation for type 1 diabetes. J Clin Endocrinol Metab. 2013;98:E1780–1785. doi: 10.1210/jc.2013-1764. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.Rasmussen-Torvik LJ, Wassel CL, Ding J, Carr J, Cushman M, Jenny N, Allison MA. Associations of body mass index and insulin resistance with leptin, adiponectin, and the leptin-to-adiponectin ratio across ethnic groups: the Multi-Ethnic Study of Atherosclerosis (MESA) Ann Epidemiol. 2012;22:705–709. doi: 10.1016/j.annepidem.2012.07.011. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17.Friedrich N, Thuesen B, Jorgensen T, Juul A, Spielhagen C, Wallaschofski H, Linneberg A. The association between IGF-I and insulin resistance: a general population study in Danish adults. Diabetes Care. 2012;35:768–773. doi: 10.2337/dc11-1833. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18.Kiechl S, Wittmann J, Giaccari A, Knoflach M, Willeit P, Bozec A, Moschen AR, Muscogiuru G, Sorice GP, Kireva T, Summerer M, Wirtz S, Luther J, Mielenz D, Billmeier U, Egger G, Mayr A, Oberhollenzer F, Kronenberg F, Orthofer M, Penninger JM, Meigs JB, Bonora E, Tilg H, Willeit J, Schett G. Blockade of receptor activator of nuclear factor-kappaB (RANKL) signaling improves hepatic insulin resistance and prevents development of diabetes mellitus. Nat Med. 2013;19:358–363. doi: 10.1038/nm.3084. [DOI] [PubMed] [Google Scholar]
- 19.Lee NK, Sowa H, Hinoi E, Ferron M, Ahn JD, Confavreux C, Dacquin R, Mee PJ, McKee MD, Jung DY, Zhang Z, Kim JK. Endocrine regulation of energy metabolism by the skeleton. Cell. 2007;130:456–469. doi: 10.1016/j.cell.2007.05.047. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20.American Diabetes Association. Standards of medical care in diabetes. Diabetes Care. 2004;27(Supplement 1) [Google Scholar]
- 21.American Diabetes Association. Standards of medical care in diabetes. Diabetes Care. 2015;38(Supplement 1) doi: 10.2337/dc14-2142. [DOI] [PubMed] [Google Scholar]
- 22.Levey AS, Stevens LA. Estimating GFR using the CKD Epidemiology Collaboration (CKD-EPI) creatinine equation: more accurate GFR estimates, lower CKD prevalence estimates, and better risk predictions. Am J Kid Dis. 2010;55:622–627. doi: 10.1053/j.ajkd.2010.02.337. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23.Rickels MR, Naji A, Teff KL. Insulin sensitivity, glucose effectiveness, and free fatty acid dynamics after human islet transplantation for type 1 diabetes. J Clin Endocrinol Metab. 2006;91:2138–2144. doi: 10.1210/jc.2005-2519. [DOI] [PubMed] [Google Scholar]
- 24.Hirsch D, Odorico J, Radke N, Hanson M, Danobeitia JS, Hullett D, Alejandro R, Ricordi C, Fernandez LA. Correction of insulin sensitivity and glucose disposal after pancreatic islet transplantation: Preliminary results. Diabetes Obes Metab. 2010;12:994–1003. doi: 10.1111/j.1463-1326.2010.01290.x. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25.Abdul-Ghani MA, Jenkinson CP, Richardson DK, Tripathy D, DeFronzo RA. Insulin secretion and action in subjects with impaired fasting glucose and impaired glucose tolerance: Results from the veterans administration genetic epidemiology study. Diabetes. 2006;55:1430–1435. doi: 10.2337/db05-1200. [DOI] [PubMed] [Google Scholar]
- 26.The Clinical Islet Transplantation Consortium. Phase 3 trial of transplantation of human islets in type 1 diabetes complicated by severe hypoglycemia. Diabetes Care. 2016;39:1230–1240. doi: 10.2337/dc15-1988. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27.Yki-Jarvinen H, Koivisto VA. Natural course of insulin resistance in type I diabetes. N Engl J Med. 1986;315:224–230. doi: 10.1056/NEJM198607243150404. [DOI] [PubMed] [Google Scholar]
- 28.Vilsbøll T. The effects of glucagon-like peptide-1 on the beta cell. Diabetes Obes Metab. 2009;11:11–18. doi: 10.1111/j.1463-1326.2009.01073.x. [DOI] [PubMed] [Google Scholar]
- 29.Balamurugan AN, Naziruddin B, Lockridge A, Tiwari M, Loganathan G, Takita M, Matsumoto S, Papas K, Trieger M, Rainis H, Kin T, Kay TW, Wease S, Messinger S, Ricordi C, Alejandro R, Markmann J, Kerr-Conti J, Rickels MR, Lin C, Zhange X, Witkowski P, Posselt A, Maffi P, Secchi A, Berney T, O’Connell PJ, Hering BJ, Barton FB. Islet product characteristics and factors related to successful human islet transplantation from the collaborative islet transplant registry (CITR) 1999-2010. Am J Transplant. 2014;14:2595–2606. doi: 10.1111/ajt.12872. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 30.Sarkar G, Alattar M, Brown RJ, Quon MJ, Harlan DM, Rother KI. Exenatide treatment for 6 months improves insulin sensitivity in adults with type 1 diabetes. Diabetes Care. 2014;37:666–670. doi: 10.2337/dc13-1473. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 31.Abbenhardt C, McTiernan A, Alfano CM, Wener MH, Campbell KL, Duggan C, Foster-Schubert KE, Kong A, Toriola AT, Potter JD, Mason C, Xiao L, Blackburn GL, Bain C, Ulrich CM. Effects of individual and combined dietary weight loss and exercise interventions in postmenopausal women on adiponectin and leptin levels. J Intern Med. 2013;274:163–175. doi: 10.1111/joim.12062. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 32.Pichler K, Loreto C, Leonardi R, Reuber T, Weinberg AM, Musumeci G. RANKL is downregulated in bone cells by physical activity (treadmill and vibration stimulation training) in rat with glucocorticoid-induced osteoporosis. Histol Histopathol. 2013;28:1185–1196. doi: 10.14670/HH-28.1185. [DOI] [PubMed] [Google Scholar]