Islet Transplantation Stabilizes Hemostatic Abnormalities and Cerebral Metabolism in Individuals With Type 1 Diabetes

Type 1 diabetes (T1D) is associated with microvascular (e.g., retinopathy, nephropathy, and neuropathy) or macrovascular (e.g., cardiovascular, cerebrovascular, and peripheral vascular disease) complications, which cause an increase in morbidity and mortality and negatively affect patient quality of life (1,2). Poor glycemic control is the primary contributor to the onset of chronic complications (3).

By replenishing destroyed pancreatic islets, islet transplantation leads to improved glycometabolic control—if not to improved insulin independence—and reduces the frequency of hypoglycemic episodes, thus delaying diabetes complications and improving patient quality of life (4–6). However, islet transplantation requires the chronic administration of immunosuppressive agents (7), which increase cardiovascular disease (8), atherosclerosis (9), and dyslipidemia (10) and are toxic to β-cells, thus shortening the life span of the transplanted islets (11,12). In the islet after kidney transplantation group, patients who received a kidney transplantation received immunosuppressive therapy before transplantation of the islet graft, whereas individuals with T1D receiving an islet allograft need to begin an immunosuppression regimen at the time of the transplant. It is therefore necessary to demonstrate that the benefit obtained from improved glycometabolic control overcomes the deleterious effect of immunosuppression. Indeed, islet transplantation after kidney transplantation demonstrated an overall beneficial effect in halting the progression of several major diabetes complications (13) such as nephropathy (14,15) and cardiovascular disease (16), yet fewer data are available on the effect of islet transplantation alone (ITA) on diabetes complications (17,18).

In this study we examined the effect of ITA on hemostatic and cerebrovascular abnormalities in individuals with T1D. Hemostatic abnormalities are evident in T1D (19) and may contribute to the onset of accelerated atherosclerosis and microangiopathy of different vascular districts, including the cerebral vasculature, thus leading to morphological, metabolic, and functional cerebral abnormalities (20). Features of hemostatic abnormalities include enhanced activation of the clotting system and severe platelet function, which lead to a prothrombotic state, with vascular dysfunction/early atherosclerosis observed in several organs, particularly in the cerebral district, in individuals with T1D (20). The resulting diabetic encephalopathy is characterized by a reduction in the volume of the cerebral cortex; a significant loss of neocortical neurons (21); a decrease in mnemonic and abstract reasoning, problem-solving abilities, and hand-eye coordination; and an increased risk of Alzheimer disease (22). We hypothesized that improvement of glycometabolic control (if not full restoration of normoglycemia) obtained with islet transplantation may reestablish the physiology of the hemostatic system, thus halting the progression of cerebral abnormalities. The aim of this study was to investigate the relatively short-term effects of islet transplantation on hemostatic profile and platelet function as well as cerebral morphology (brain volume), metabolism (neuronal/axonal metabolism), and function (cognitive and neuropsychological function).

This nonrandomized pilot study included 22 individuals with T1D who were actively enrolled on the islet transplantation waiting list. Twelve of these individuals underwent ITA, while 10 remained on the waiting list (T1D-WL) (Supplementary Fig. 1). Individuals with T1D were retrospectively enrolled for a 24-month period according to the major criteria applied to enter the islet transplantation program, as published elsewhere (4,7). The baseline and demographic characteristics of the two groups of individuals with T1D were comparable (Supplementary Table 1) and were reexamined at 15 months of follow-up (15.5 ± 2.5 months). Glycometabolic parameters were determined at enrollment and during the follow-up using a single determination; thus they are only partially representative of the individual’s glycometabolic control. Ten healthy volunteers matched for age and sex (control group [CTRL]) were studied as well. Patients with a history of cerebrovascular disease (transient ischemic attack/stroke) and/or who were taking an oral anticoagulant agent were excluded. All subjects provided informed consent before enrollment. Studies not included in the routine clinical follow-up were approved as appropriate by an institutional review board. ITA is associated with different adverse events because of both the procedure and immunosuppressive treatments. On the basis of our experience, acute complications are represented by thrombosis of a peripheral branch of the portal vein (9%) and bleeding (36%), whereas chronic complications may include decreased renal function (15%), viral myocarditis (6%), and reactivation of cytomegalovirus infections (6%) (23).

Individuals who underwent islet transplantation received five doses of intravenous daclizumab 1 mg/kg (Zenapax, Roche) for 2 weeks as induction therapy. As maintenance, tacrolimus (Prograf, Fujisawa; serum levels of 3–6 ng/mL) and sirolimus (Rapamune, Wyeth-Ayerst; ranging from 12–15 ng/mL for 90 days and 7–10 ng/mL thereafter) were used. Four individuals were switched to a different immunosuppressive regimen because of side effects: either tacrolimus plus mycophenolate mofetil (MMF; n = 2), sirolimus plus MMF (n = 1), or cyclosporine plus MMF (n = 1). Four of 12 individuals in the ITA group and 4 of 10 individuals with T1D received ACE inhibitors. One patient in the ITA group received statins during follow-up. No drugs were administered to controls.

To understand whether immunosuppression represents a potential bias in our study and whether it affects coagulatory/inflammatory markers, we considered an additional 10 individuals with end-stage renal disease (ESRD) and 10 individuals with ESRD who subsequently received a kidney transplant; these patients were studied at 15 months of follow-up and treated with thymoglobulin as induction therapy and tacrolimus plus sirolimus (or MMF in few cases) as maintenance therapy. Individuals with ESRD were all receiving hemodialysis treatment at our hospital; patients received ACE inhibitors (7 of 10), β-blockade agents (3 of 10), calcium-phosphate binders (10 of 10), and/or 1-OH vitamin D (8 of 10) as concomitant therapy. Finally, with the same purpose of understanding whether immunosuppression affects coagulatory/inflammatory markers in a pure T1D population, we considered five individuals with T1D receiving sirolimus monotherapy (0.1 mg/kg) for 37–197 days (target levels 8–10 ng/m) and atorvastatin 10 mg for 44–279 days as preconditioning for islet transplantation (at least 30 days).

Platelet-poor plasma was obtained by centrifugation at room temperature (10 min at 1,500 g). Determinations of prothrombin time, activated partial thromboplastin time, fibrinogen (Fg), antithrombin, fasting homocysteine d-dimer fragments (d-dimer), levels of prothrombin fragments 1+2 (F1+2), protein C, and protein S were analyzed using fresh plasma samples (24).

Platelet-rich plasma was obtained by centrifugation of blood with 1 mL of anticoagulant citrate dextrose (ACD) solution and collection of supernatant. Platelet intracellular calcium ([Ca²⁺]i) was evaluated under resting conditions as previously described (25).

Aliquots of platelet-rich plasma were fixed, and ultrastructural evaluation of four cases from each group included platelet size, morphology, and granule content, and platelet areas were measured using the Measure Arbitrary Area tool of Analysis Image Processing 3.0 software (Soft Imaging System, Münster, Germany).

Magnetic resonance imaging (MRI) and magnetic resonance spectroscopy (MRS) of protons (¹H-MRS) were performed using a 1.5-T scanner (Vision; Siemens, Erlangen, Germany) with a head-dedicated coil for scanning. Acquired images (T₂, T₁, axial fluid attenuated inversion recovery, and coronal spin) were analyzed to assess micro- and macroangiopathy during the clinical course of diabetes, as previously reported (26). A diagnosis of cerebrovascular disease was formulated when lesions were hyperintense in T₂-weighted and fluid attenuated inversion recovery images and hypointense on T₁-weighted images. T₁-weighted axial images of each participant were transferred to an offline workstation (Sun Sparcstation; Sun Microsystems, Mountain View, CA) for brain volume assessment (27). Single-voxel MRS was performed at least three times for each individual (28). The peaks of choline (Cho), creatine (Cr), and N-acetylaspartate (NAA) were calculated, and their integral function was solved by automatic and manual measurement. Results were expressed as ratios: NAA to Cr, Cho to Cr, and NAA to Cho. Cr was considered an internal standard because its levels are relatively uniform even in pathological conditions (29). All measurements were performed by the P.V.; whole-brain NAA MRS was analyzed as previously described (28).

Psychological assessment evaluated emotional patterns and quality of life (QoL) of patients. The Profile of Mood State (POMS) was used to measure present mood states and disturbance symptoms, dissecting six dimensions of mood: tension-anxiety, depression-dejection, anger-hostility, vigor-activity, fatigue-inertia, and confusion-bewilderment (30). To assess the patients’ QoL, the LEIPAD questionnaire was administered, as reported by our group (28,31). A neuropsychologist performed a neuropsychological evaluation in one session by analyzing childhood intellectual failures, global cognitive efficiency, language, attention, reasoning, and verbal/visual-spatial memory, as previously reported (28). The Mini-Mental State Examination was used to examine patients’ global (intellectual) cognitive abilities. Language functioning (oral comprehension and access to internal lexicon) was evaluated using the token test and phonemic and semantic word fluency test (32). Attention abilities were analyzed using the Stroop test (28) and the paced auditory serial addition test, which evaluated sustained attention and information processing speed (33). Verbal and visual short-term memory function was evaluated with the verbal span and Corsi span tests, whereas verbal long-term memory was assessed with short story recall. Finally, abstract reasoning was evaluated through visuospatial stimuli using the Wisconsin card sorting test (48-card form) (34). Evaluations were performed three times after extensive internal validation.

Continuous variables are presented as mean ± SD and compared with two-tail Student t test for unpaired data (two groups) or the one-way ANOVA (three or more groups); the Mann-Whitney U test (two groups) or the Kruskal-Wallis test (three or more groups) were used if no normal distribution was evident. Categorical variables are presented as proportions with 95% CIs and were compared using the χ² test if five or more observations are in each cell, or the Fisher exact test if there are fewer than five observations. A P value < 0.05 was considered statistically significant. CIs were estimated using the binomial exact. Analysis was performed using Stata software, version 11.1 (StataCorp, LP, College Station, TX).

Individuals with T1D and healthy controls did not differ in any major demographic characteristics at baseline, when considering individuals with T1D who remained on the waiting list (T1D-WL) or those who underwent ITA (T1D-ITA) (Supplementary Table 1). The mean age, duration of diabetes, creatinine level, total cholesterol, HDL cholesterol, exogenous insulin requirement, and mean arterial pressure were similar in the three groups of subjects examined (Supplementary Table 1). Significant differences were observed in the major parameters accounting for glycometabolic control between individuals with T1D (T1D-WL and T1D-ITA) and controls, including glycated hemoglobin (HbA_1c), serum glucose (S-glucose) and C-peptide, whereas no differences were found when comparing the values of the T1D-WL versus the T1D-ITA groups (Supplementary Table 1). All individuals with T1D (T1D-WL and T1D-ITA) were homogenous for platelet function and ultrastructural characteristics, levels of prothrombotic factors, and cerebral morphology/metabolism/function at baseline, with consistent results in all the parameters analyzed.

Mean duration of follow-up was 15.5 ± 2.5 months. Renal function, blood pressure, and lipid metabolism measurements (creatinine: T1D-WL 0.8 ± 0.1 vs. ITA 1.0 ± 0.2 mg/dL; total cholesterol: T1D-WL 172.5 ± 11.0 vs. ITA 183.5 ± 12.5 mg/dL; HDL cholesterol: T1D-WL 54.3 ± 10.5 vs. ITA 58.9 ± 9.9 mg/dL; triglycerides: T1D-WL 97.0 ± 37.7 vs. ITA 64.1 ± 18.8 mg/dL; and mean arterial pressure: T1D-WL 98.6 ± 8.1 vs. ITA 103.8 ± 1.8 mmHg) were similar at 15 months of follow-up in individuals with T1D. Patients did not show any sign of cardiovascular disease (absence of both any electrocardiographic signs of myocardial ischemia and any echocardiography signs of systolic/diastolic dysfunction). A significant improvement in glycometabolic control was evident in patients in the ITA group compared with individuals in the T1D-WL group. S-glucose, C-peptide, HbA_1c, and exogenous insulin requirement values were all improved in patients in the ITA group (Fig. 1A–D). Islet transplantation restored glycometabolic control to near normal.

We examined an average of 10 platelets per patient to evaluate platelet area, calcium platelet homeostasis, and number of granules, and we examined 10 granules to evaluate granular area (Fig. 1E and F). Platelet size at 15 months of follow-up was higher in individuals with T1D compared with controls and ITA patients (T1D-WL 3.860 ± 0.288 × 10⁶, CTRL 3.076 ± 0.197 × 10⁶, and ITA 3.199 ± 0.287 × 10⁶ nm²; P = 0.02, T1D-WL vs. CTRL), whereas ITA patients and controls showed similar platelet sizes (P = not significant). Moreover, platelets in individuals with T1D showed the tendency to aggregate, exhibited a more electron-dense cytoplasm, and displayed more numerous and larger granules than platelets in controls and the ITA patients (Fig. 1G). This suggests that platelet morphology is altered in individuals with T1D and that islet transplantation is associated with near-normalization of platelet size and aggregation and platelet granule size and number. The levels of resting [Ca²⁺]i, analyzed at 15 months of follow-up, were higher in individuals with T1D, but not in ITA patients, compared with controls (CTRL 72.2 ± 15.0, T1D-WL 107.7 ± 40.0, ITA 87.6 ± 18.8 nmol; P = 0.01) (Fig. 1F). This suggests that T1D is associated with aberrant calcium platelet homeostasis and that islet transplantation promotes near-normalization of both platelet signaling and the cytosolic calcium pathway.

Analysis of hemostatic profile demonstrated high levels of plasmatic Fg, F1+2, and d-dimer in individuals with T1D. Lower levels of Fg (CTRL 262.5 ± 43.0, T1D-WL 328.5 ± 35.0, ITA 367.0 ± 26.0 mg/dL; P < 0.005, CTRL vs. all) and F1+2 (CTRL 63 ± 17.8, T1D-WL 113 ± 35.7, ITA 96 ± 45.2 mmol/L; P = 0.005, CTRL vs. T1D-WL) were evident in controls compared with ITA patients and individuals with T1D (Fig. 2A and B). Finally, d-dimer levels were higher in individuals with T1D compared with controls and ITA patients (CTRL 0.19 ± 0.02, T1D-WL 1.07 ± 0.80, ITA 0.24 ± 0.02 µg/mL; P = 0.01, CTRL vs. all) (Fig. 2C). No statistically significant differences were found in fasting homocysteine levels among the three groups of subjects (Fig. 2D). Analysis of prothrombin time, activated partial thromboplastin time, and antithrombin III did not reveal differences among the three groups at 15 months of follow-up. Significantly lower levels of protein S and protein C were found in individuals with T1D compared with controls, whereas in ITA patients levels were almost normalized at the 15-month follow-up (Figs. 2E and F). A prothrombotic state is evident in individuals with T1D, and ITA is associated with near-normalization of hemostatic abnormalities.

Evaluation of vasculopathy with conventional MRI revealed that 12 of 22 individuals with T1D had cerebrovascular disease at baseline, without differences between the two groups (T1D-ITA vs. T1D-WL; P = not significant), compared with 2 of 9 controls (T1D 54.5% [95% CI 36–69%] vs. CTRL 22.2% [95% CI 3–6]; P = 0.04). No major changes in cerebral morphology were evident at 15 months of follow-up. Mean cerebral volume was similar in individuals with T1D and in ITA patients, whereas it was significantly lower when compared with controls at 15 months of follow-up (CTRL 1.197 ± 98.3, ITA 1.027 ± 90.3, T1D-WL 1.105 ± 67.8; P < 0.03, CTRL vs. all) (Fig. 3A). Islet transplantation did not modify structural changes in cerebral morphology, which was evident in individuals with T1D at 15 months of follow-up.

The mean absolute NAA, a neuronal marker that indicates neuronal/axonal loss when decreased (29,35), showed similar values in the three groups of patients at 15 months of follow-up. Absolute NAA values, corrected for brain volume to adjust for interindividual variation (NAA concentration), failed to reveal any effect in the ITA group (Fig. 3B). Likewise, mean whole-brain NAA content did not differ among the three groups (CTRL 14.5 ± 1.5; T1D 14.4 ± 1.8; ITA 13.2 ± 1.6 arbitrary units [Aus]; P = not significant) (Fig. 3C).

The NAA-to-Cho ratio (an index of neuroaxonal loss/impairment associated with diabetes-related gliosis [36]) was similar in ITA patients and controls, but it was lower in individuals with T1D, suggesting major cerebral tissue damage (CTRL 2.0 ± 0.2, T1D-WL 1.7 ± 0.2, ITA 1.9 ± 0.2 AU; P = 0.02, T1D-WL vs. ITA; P = 0.003, CTRL vs. T1D-WL), (Fig. 3D). The NAA-to-Cr ratio, an index of neuronal loss/damage when decreased (37), did not show any significant difference in the three groups analyzed, although higher values were found in controls compared with others (CTRL 2.1 ± 0.2, T1D-WL 2.0 ± 0.1, ITA 2.0 ± 0.2 AU; P = not significant) (Fig. 3E). Conversely, we documented an increased Cho-to-Cr ratio (which has been associated with diabetes gliosis [36] when increased) in individuals with T1D, particularly when compared with ITA patients, indicating a near recovery of the neuronal tissue (T1D-WL 1.1 ± 0.1 and ITA 0.98 ± 0.1; P = 0.01) (Fig. 3F). Finally, an inverse linear correlation was found between HbA_1c and the NAA-to-Cho ratio (r = −0.61; P = 0.004) (data not shown), thus suggesting that poor glycemic control in individuals with T1D is important in promoting degenerative phenomena and altering cerebral metabolism.

Psychological evaluation was performed by analyzing patients’ emotional axis and QoL at 15 months’ follow-up using the POMS test and LEIPAD questionnaire, respectively (30,31). The POMS test, examining six different dimensions of mood (see Research Design and Methods), showed a significantly higher score in the POMS depression-dejection category among individuals with T1D compared with controls and ITA patients (Table 1). Moreover, the POMS fatigue-inertia category was significantly different among controls, individuals with T1D, and ITA patients, with the latter showing less fatigue-inertia (Table 1). Results of the POMS confusion-bewilderment category showed higher values in individuals with T1D compared with ITA patients, whereas the POMS anger-hostility category revealed a significant difference in controls compared with both individuals with T1D and ITA patients (Table 1). No differences were found in the POMS tension-anxiety and vigor-activity categories (Table 1). The analysis of the LEIPAD test did not show any difference in physical function and self care among the three groups, but it revealed a significantly higher score for depression and anxiety in individuals with T1D compared with ITA patients (Table 1). In addition, life satisfaction scores were significantly lower (in which lower score demonstrates improvement) in controls and ITA patients compared with individuals with T1D (Table 1). Considering both the POMS test and LEIPAD questionnaire, despite the fact that some values did not achieve statistical significance, an overall trend demonstrating improvement was evident in ITA patients compared with individuals with T1D. This was confirmed by neuropsychological assessment, in which ITA patients demonstrated a substantial benefit from transplantation compared with individuals with T1D in most of the tests (Table 1). Phonemic and semantic fluency also showed higher scores in controls and ITA patients compared with individuals with T1D. Furthermore, the paced auditory serial addition test revealed significantly higher scores among ITA patients compared with individuals with T1D and controls (Table 1). The color-naming and the word-color interference variants of the Stroop test were higher (indicating a worsened state) in T1D-WL and ITA patients compared with controls (Table 1). The color-color interference Stroop test revealed lower scores (i.e., a worse state) in T1D-WL and ITA patients compared with controls, although this difference did not reach statistical significance. The number of errors made by the patients during the three sections of the Stroop test was significantly lower in controls than in T1D-WL and ITA patients (Table 1). ITA is associated with a near-normalization of psychological and neuropsychological tests.

To understand whether concomitant immunosuppressive treatment may have biased our results, we first explored the effect of immunosuppression (tacrolimus and sirolimus) in altering coagulatory/inflammatory markers in individuals with ESRD who had received a kidney transplant (n = 10). We finally evaluated whether immunosuppression (sirolimus) altered coagulatory/inflammatory markers in individuals with T1D (n = 5). Immunosuppression did not alter or change the peripheral levels of d-dimers, proteins S and C, homocysteine, Fg, or F1+2 in individuals who received a kidney transplant compared with individuals with ESRD (Supplementary Fig. 2A–F) or in patients with T1D (Supplementary Fig. 2G–N). These experiments confirm that immunosuppression has no effect on inflammatory/coagulatory markers, and thus it is not likely involved in the modifications of these markers observed in our study.

Insulin, the only treatment so far for individuals with T1D, can only partially prevent long-term diabetes complications, and it is still associated with fatal hypoglycemic episodes (4). Islet transplantation, a therapeutic option available for selected individuals with T1D, has been further examined in the past decade because it can potentially normalize glycometabolic control, at least in the short term, while providing partial islet function capable of halting the progression of diabetes complications (38,39). The safety, effectiveness, and minimal invasiveness of islet transplantation combined with kidney transplantation have been demonstrated by our group and others (4,5,40). Unfortunately, as with other allogeneic procedures, islet transplantation also requires administration of lifelong immunosuppressive treatment (7,41). Immunosuppressants (e.g., tacrolimus and sirolimus) increased the incidence of cardiovascular disease because of their detrimental effect on lipid metabolism and blood pressure (11), which may call into question the advantage of ITA in truly halting the progression of vascular complications associated with T1D. Our results confirmed that ITA is able to reestablish good glycemic control at 15 months of follow-up and maintain a stable lipid profile and acceptable blood pressure control, despite ongoing immunosuppressive treatment.

Furthermore, our analysis of the hemostasis/thrombosis system demonstrated that patients undergoing ITA had normal platelet volume/function and normal levels of prothrombotic factors. On the contrary, individuals with T1D who remained on the waiting list showed a persistent state of hypercoagulability, which represents a significant risk factor for vascular thrombosis and atherosclerosis in different systems, including the cerebral vascular system. In particular, activation of platelets strictly depends on cytosolic calcium signaling, which tends to be altered in T1D (24). Platelet activation, together with increased prothrombotic factors, leads to persistent activation of the clotting system and facilitates generation of thrombi and vessel occlusion.

We then used ¹H-MRS to analyze brain metabolism and evaluate whether these hemostatic/platelet alterations resulted in or were associated with any modification of cerebral morphology, metabolism, and function (28). Amounts of total Cr (associated with energy metabolism and considered as an internal standard), compounds containing Cho (playing an important role in the turnover of cellular wall), and NAA (a marker of neuronal density and functioning) in the brain, the relative reductions of which have been correlated with neuronal/axonal defect in number and function (29,35), were measured. ¹H-MRS showed specific alterations in cerebral metabolism, which were otherwise undetectable with standard methods, including signs of atrophy, decreased brain volume, and reduced axonal metabolism in individuals with T1D compared with healthy controls. These metabolic abnormalities can be associated with early cognitive decline and possibly senile dementia. Conversely, we observed higher NAA-to-Cho and lower Cho-to-Cr ratio values in patients undergoing ITA, indicating a relative sparing of neuronal function from tissue degeneration and loss after ITA (36,37). Finally, we documented a significant improvement in mood profile, depression and anxiety, speed of information processing, and attention abilities, with greater maintenance of neuropsychological attitude in patients receiving ITA compared with individuals with T1D. The presence of a positive correlation between HbA_1c levels and the NAA-to-Cho ratio (a good marker of axonal/neuronal degeneration) demonstrated the importance of good metabolic control in patients receiving ITA (and overall in individuals with T1D) in influencing the preservation of neural structures and cognitive function/metabolism. We acknowledge that long-term immunosuppression is not an option for individuals with T1D, but our study is a proof of concept demonstrating that islet transplantation, despite immunosuppression, can prevent the occurrence of major diabetes complications.

Our study is not devoid of limitations. First, the sample size is relatively small, and a larger study is required to generalize conclusions. Second, although the two cohorts of individuals with T1D (T1D-WL and T1D-ITA) were homogeneous at baseline, we cannot completely exclude the presence of biases. In this context, we acknowledge the complexity of our whole clinical assessment, which will require further validation. Another bias may be represented by concomitant immunosuppression, which may affect coagulatory/inflammatory markers. Indeed, our data for individuals with ESRD who received a kidney transplant and for individuals with T1D treated with sirolimus demonstrated that immunosuppression has no effects on coagulatory/inflammatory markers—at least in our study—and thus should not be a confounding factor. Our results were obtained in a relative short period of time, underlining the relevance that ITA may have in halting diabetes complications in individuals with T1D (Supplementary Table 2). However, the field of islet transplantation needs randomized, well-designed, and powered studies to solidly confirm the benefits of this approach.In conclusion, the restoration of near-normal glucose metabolism with ITA improves glycometabolic control and hemostatic abnormalities, thereby leading to the near-normalization of cerebral disorders (Supplementary Table 2). Despite the disadvantageous effects of immunosuppressive therapy, islet transplantation may be an important option to limit the progression of major T1D complications but must to be validated in a larger, controlled clinical trial.

Funding. This work was supported by the Juvenile Diabetes Research Foundation (grant no. JT01Y01). F.D. has received an Italian Scientists and Scholars of North America Foundation–Fondazione Marche Fellowship. A.V. has received support from a National Institutes of Health Research Training Grant to Boston Children’s Hospital in Pediatric Nephrology (T32DK-007726-28). R.B. has received support from an American Diabetes Association (ADA) mentor-based fellowship to P.F. P.F. is supported by a Translational Research Program grant from Boston Children’s Hospital and has received a Juvenile Diabetes Research Foundation Career Development Award, an American Society of Nephrology Career Development Award, an ADA mentor-based fellowship, Istituto Superiore di Sanitá grant (“Staminali” RF-FSR-2008-1213704), and Italian Ministry of Health grants RF 2010-2314794 and RF 2010-233119.

Duality of Interest. No potential conflicts of interest relevant to this article were reported.

Author Contributions. F.D. designed the study, performed the research, analyzed data, and wrote the article. P.M., G.F., and S.L.R. performed research and analyzed data. P.V., A.V., A.M., R.B., R.N., M.F., E.C., A.D., I.F., and A.F. performed research. F.P. designed the research and wrote and edited the paper. G.M. and L.P. edited the paper. F.F. and A.S. designed the study and edited the article. P.F. designed the study and wrote and edited the article. P.F. is the guarantor of this work and, as such, had full access to all the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.

Prior Presentation. This study was presented in oral form at the 73rd Scientific Sessions of the American Diabetes Association, Chicago, Illinois, 21–25 June 2013.