Increased Plasma Amylin in Type 1 Diabetic Patients After Kidney and Pancreas Transplantation: A sign of impaired β-cell function?

Type 1 diabetic patients (n = 11) after successful combined PKT were matched for age, sex, and BMI with nine healthy control subjects and six nondiabetic patients after successful kidney transplantation (Table 1). Type 1 diabetic patients (duration of type 1 diabetes: 31 ± 2 years) had received a whole pancreas with systemic venous anastomosis to the iliac vein (2.4 ± 0.5 years [range 0.8–5.2]) before the study (12). By this method of systemic drainage, pancreatic endocrine secretions bypass the physiological first-pass clearance of the liver (9,12,20). At the time of examination, the immunosuppressive regimen in PKT patients had been free of glucocorticoids for at least 5 months and included tacrolimus (2.5–8 mg/day, plasma level on study day: 13.2 ± 2.8 ng/ml) combined with either mycophenolate mofetile (1–2 g/day, n = 9) or azathioprine (25–50 mg/day, n = 2). Until the beginning of the study, none of the PKT patients had a graft rejection. Glucocorticoids were included in the initial immunosuppressive protocol but were discontinued within 3 months after successful transplantation.

The kidney transplant patients underwent successful kidney transplantation 11.2 ± 0.5 years before the study. One of the patients had had an acute rejection of the first kidney graft 9 years ago, following a successful re-transplantation of a second kidney graft 1 year later. None of the other patients suffered from any graft rejection period. At the time of examination, the immunosupressive regimen in kidney transplant patients included prednisolon (2.5–10 mg/day) and tacrolimus (2 mg/day, n = 1, plasma level on study day: 10.1 ng/ml) combined with either mycophenolate mofetile (1–2 g/day, n = 4) or azathioprine (50–75 mg/day, n = 2) and/or cyclosporine-A (75–200 mg/day, n = 4, plasma level on study day: 68.0 ± 2.8 ng/ml).

Normal fasting plasma glucose, A1C <6.5%, and stable creatinine were required for participation in the study. No subject was using insulin or any other hypoglycemic agent to maintain glucose control. All participants gave their written informed consent to the protocol, which had been approved by the local ethics committee.

At Lainz hospital (Vienna, Austria), a 3-h OGTT was performed after a minimum 12-h overnight fast. A catheter (Venflon; Becton Dickinson, Stockholm, Sweden) was inserted into an antecubital vein of the forearm for blood sampling. The subjects were given the standard dose of 75 g glucose in H₂O solution (Glukodrink; Roche Diagnostics, Vienna, Austria) within 5 min and were maintained on bed rest. Venous blood samples for measurement of metabolites and hormones were drawn in the fasting state and at 10, 20, 30, 40, 60, 90, 120, 150, and 180 min after glucose load. A1C, plasma glucose, and serum creatinine were measured using routine laboratory methods (www.kimcl.at). Blood was rapidly centrifuged, and plasma and serum aliquots were stored at −70°C until further analysis.

Plasma FFAs were determined with a commercially available enzymatic colorimetric method (Wako Chemicals, Neuss, Germany). Plasma insulin was determined with a radioimmunoassay (RIA; BioChem Immunosystems, Freiburg, Germany) and serum C-peptide by dissociation-enhanced lanthanide flouroimmunoassay (Delfia, Wallac Oy, Finland). Plasma human total amylin was measured using an enzyme-linked immunosorbent assay (Linco Research, St. Charles, MO).

Total areas under the concentration curves of glucose (AUC_glucose), insulin (AUC_insulin), C-peptide (AUC_C-peptide), and amylin (AUC_amylin) were calculated with the trapezoidal rule. To better compare hormone release in response to the glucose challenge, suprabasal AUCs (ΔAUC), i.e., total AUC − basal concentration × 180 min, were used. Differences between groups were assessed by performing χ² tests for categorical variables. Continuous variables were analyzed with ANOVA following a Dunnet or Bonferroni post hoc test (as indicated). Linear correlations are based on Spearman’s correlations. Statistical analyses were performed using SPSS (SPSS, Chicago, IL), Statistica (StatSoft, Tulsa, OK), and/or MedCalc (MedCalc Software, Mariakerke, Belgium) computer software. Data are presented as means ± SE. Differences between groups at P ≤ 0.05 were considered to be statistically significant.

Mathematical models were used to assess β-cell function, secretion, and clearance of C-peptide, insulin, and amylin from OGTT data (21,22). In PKT patients, insulin and C-peptide are delivered directly into the systemic circulation, whereas in control subjects, they undergo a first pass through the liver. Because the liver does not play any significant role in C-peptide dynamics, we analyzed the kinetics of C-peptide (23,24) instead of insulin. The secretion model described (21) and thoroughly validated elsewhere (22) reconstructs the patterns per unit volume of C-peptide release (basal secretion rate and total stimulated secretion rate) and its systemic appearance. Therefore, hepatic insulin extraction (as percentage of the secreted hormone) was simply approximated by [1 − (AUC_insulin/AUC_C-peptide)]. Given the different drainage in PKT patients, this formula evaluates extraction in every circumstance; in addition, the calculated value refers to the whole extraction of the hormone during the entire OGTT period.

Amylin basal secretion, total secretion, clearance, and amylin/insulin cosecretion factor were estimated using a mathematical model described (21,25) and thoroughly validated with ad hoc kinetic studies (26,27) and applied in several studies (28). The amylin/insulin cosecretion factor may be regarded as a physiological parameter that relates amylin delivery to β-cell insulin secretion rate under the assumption of parallel secretion (25,27,28).

The pattern of FFAs was assessed by calculating the slope of the curve between zero time and 60 min, thus mapping the course of FFA suppression during OGTT. The basal level and the difference between nadir and fasting FFAs were also determined.

Insulin sensitivity during an OGTT was assessed by the oral glucose insulin sensitivity index (OGIS), which takes into account known relationships between glucose disappearance and insulin (29). OGIS, which describes glucose clearance, has been validated against and applied instead of the corresponding value obtained from the gold standard (namely the euglycemic-hyperinsulinemic clamp test) in healthy, obese, and diabetic subjects (29–31).

The adaptation index describes the capacity of the β-cell to release increasing amounts of insulin to compensate for the increasing insulin resistance (32). The adaptation index was calculated as OGIS × ΔAUC_C-peptide. β-Cell sensitivity to glucose was calculated as total stimulated secretion rate over ΔAUC_glucose. For comparison of β-cell dynamics between the three groups, the different ways of pancreatic venous drainage were taken into account by using C-peptide modeling (23,24) instead of insulin kinetics, since C-peptide is eliminated preliminary by the kidney. With respect to reduced renal function, thus possibly influencing C-peptide elimination kinetics, we examined a group of kidney transplant patients matched for creatinine clearance.

The clinical characteristics of PKT patients, kidney transplant patients, and control subjects are shown in Table 1. The groups were matched for age, sex, and BMI. When compared with the control subjects, PKT and kidney transplant patients had similar blood pressure values but a lower creatinine clearance (P < 0.001, PKT and kidney transplant patients vs. control subjects; NS, PKT vs. kidney transplant patients). PKT patients had higher A1C values than control subjects and kidney transplant patients (P < 0.03).

Data from the OGTT for all measured compounds are shown in Fig. 1 and Table 1.

Fasting plasma glucose concentration was not different in PKT patients, kidney transplant patients, and control subjects (Table 1 and Fig. 1A). While OGTT plasma glucose of kidney transplant patients did not differ from that in control subjects or PKT patients, the PKT group had higher plasma glucose values between 90 and 150 min by 9–31% (Fig. 1A; each P < 0.05).

No parameter of FFA concentrations, kinetics, or modeling (data not shown) differed between the three groups (Fig. 1B).

Plasma insulin concentrations were 90% higher at fasting in PKT patients (P < 0.01 vs. control subjects) and increased by 138% at 90 min in kidney transplant patients (P < 0.04 vs. control subjects). Total and suprabasal AUC_insulin levels were higher by 82–97% in kidney transplant patients (P < 0.05), but similar in PKT patients when compared with control subjects (Fig. 1C and Table 1). C-peptide was similar between PKT patients and control subjects, but markedly increased in kidney transplant patients both at fasting (48%, P < 0.007 vs. control subjects) and throughout the entire course of the OGTT (Fig. 1D and Table 1; each P < 0.05 vs. control subjects and/or PKT patients). Total and suprabasal ΔAUC_C-peptide levels were elevated by ∼50–70% in kidney transplant patients (Table 1; each P < 0.05 vs. control subjects and PKT patients) but were comparable between PKT patients and control subjects.

In PKT patients, fasting plasma amylin levels were approximately threefold higher (Fig. 1E and Table 1; P < 0.01 vs. control subjects). In PKT patients, amylin levels were increased at 10 and 40 min as well as between 120 and 180 min during the course of OGTT, resulting in an approximately twofold elevation of total AUC_amylin compared with control subjects (Fig. 1E and Table 1; each P < 0.05). In kidney transplant patients, plasma amylin concentrations and suprabasal and total AUC_amylin levels tended to be higher than those of control subjects, but did not reach statistical significance, which was due to the extremely broad range in kidney transplant patients (AUC_amylin in control subjects: 1,899–6,776 nmol/l per min; in PKT patients: 3,065–14,896 nmol/l per min; and in kidney transplant patients: 2,849–28,684 nmol/l per min).

Insulin sensitivity, calculated using the OGIS model, was not statistically different in all groups (Table 1).

At baseline, the insulin secretion rate was 49 and 115% higher in PKT and kidney transplant patients, respectively (Table 1; each P < 0.05 vs. control subjects). Total insulin secretion was elevated in kidney transplant patients by 82% (Table 1; P < 0.02 vs. control subjects). The hepatic insulin extraction rate during the test was ∼10% (P < 0.05) lower in PKT patients than in control subjects (Table 1).

In PKT patients, β-cell sensitivity to glucose during OGTT was reduced to one-third (85.5 ± 8.6, P < 0.009), while that of kidney transplant patients (136.7 ± 19.3) was not different when compared with control subjects (235.7 ± 54.3). The adaptation index, a measure of β-cell adaptation to ambient insulin sensitivity, was higher by 29 and 49% in kidney transplant patients when compared with control subjects and PKT patients, respectively (Table 1; each P < 0.05).

Basal amylin secretion was statistically unchanged in kidney transplant patients, but was threefold higher in PKT patients, resulting in a doubled amylin/insulin cosecretion factor in PKT patients (Table 1; P < 0.05 each). Total amylin secretion was not statistically different in all groups. Amylin clearance was similar in PKT patients, kidney transplant patients, and control subjects (Table 1).

In all subjects, both higher basal amylin (r = −0.543, P < 0.004) and higher AUC_amylin levels (r = −0.479, P < 0.01) correlated with lower β-cell sensitivity to glucose. A comparison of circulating amylin and C-peptide in PKT patients, kidney transplant patients, and control subjects at all time points during OGTT revealed a very close correlation (each P < 10⁻⁶) (all participants: r = 0.701; PKT patients: r = 0.575; kidney transplant patients: r = 0.737; control subjects: r = 0.837). In all participants, OGIS was inversely associated with basal plasma amylin (r = −0.437, P < 0.03), as well as with suprabasal (r = −0.557, P < 0.003) and total AUC_amylin (r = −0.560, P < 0.003). A direct relationship was registered between basal amylin and basal insulin (r = 0.560, P < 0.003) in all subjects.

The main finding of this study was that there was a rise in amylin, both in the fasting state and after a glucose challenge, in type 1 diabetic patients who had undergone successful PKT. Plasma amylin concentrations in PKT patients, kidney transplant patients, and control subjects were inversely associated with β-cell function and OGIS. Despite systemic drainage, which resulted in peripheral hyperinsulinemia, after PKT, type 1 diabetic patients on a glucocorticoid-free regimen showed a degree of insulin sensitivity comparable to that in nondiabetic individuals.

Plasma amylin is commonly elevated in insulin-resistant conditions that go along with hyperinsulinemia, such as impaired glucose tolerance, the early stage of type 2 diabetes, and obesity. However, because of β-cell destruction, amylin is completely absent in absolutely insulinopenic type 1 diabetic patients (20).

We showed that successful PKT restored and even increased plasma amylin, which, to our knowledge, has not been reported previously.

The up to twofold higher plasma amylin levels in PKT patients could be due to its increased release and/or reduced clearance. Kidney function in both groups of transplanted patients was comparable, but worse than that in control subjects. However, amylin clearance was similar in all groups, which could be expected since amylin is predominantly cleared by the kidney. Indeed, the tight correlation of the time curves for plasma amylin and C-peptide levels (the latter is exclusively eliminated by the kidneys) revealed that the two hormones were closely associated with each other. This also substantiates the fact that amylin is primarily eliminated by the kidney, in line with previous reports (26). Finally, since hepatic amylin extraction was shown to be negligible during OGTT (27), the systemic versus portal drainage in PKT patients and control subjects, respectively, cannot be held responsible for the higher amylin concentrations in PKT patients. Therefore, the similar amylin clearance in PKT patients and control subjects indicates that elevated concentrations of plasma amylin are most likely because of enhanced secretion. To further investigate this hypothesis, we also studied another group of matched nondiabetic patients after successful transplantation of a kidney, only with normal pancreatic situs and drainage. However, because of the current guidelines on treatment of kidney transplant patients (33), regular glucocorticoid therapy had not been tapered off in our kidney transplant patients. Amylin’s plasma concentration and secretion in kidney transplant patients was not significantly different from that in control subjects, but tended to be increased. In this context, it is of relevance that 5-day dexamethasone intake in healthy humans did not only increase amylin, but also insulin release (34). Our kidney transplant patients were also clearly hyperinsulinemic, most likely because of the glucocorticoid treatment. In a healthy individual, amylin is cosecreted with insulin in quite a stable proportion (35). When compared with control subjects, our kinetic analysis of amylin revealed an approximately twofold higher amylin/insulin cosecretion factor in PKT patients, but not kidney transplant patients, indicating that the β-cell releases markedly larger amounts of amylin compared with C-peptide or insulin in PKT patients only. From this it can be suggested that the trend of higher plasma amylin in kidney transplant patients after glucose challenge is rather because of increased stimulation of the β-cell that cosecretes more of both insulin and amylin. In contrast, in PKT patients, post–glucose load insulin secretion was similar to that in control subjects, whereas plasma amylin was significantly higher, which might be an early sign of β-cell damage. The lower β-cell sensitivity to glucose in these patients seems to confirm this hypothesis.

Of note, smaller amounts of amylin are also secreted from δ-cells in the pancreas and the antral and fundic mucosa (36,37). In the stomach, amylin stimulates somatostatin secretion and thereby inhibits gastrointestinal motility (36), which could result in gastroparesis and/or delayed absorption due to reduced intestinal circulation. Plasma glucose concentrations during the first 60 min of the OGTT (corresponding to the major glucose absorption period) were similar in all groups, and the diabetic patients did not show any clinical signs of gastroparesis. On the other hand, somatostatin is known to inhibit β-cell secretion. Thus, it cannot be ruled out that amylin release from the stomach stimulated somatostatin secretion, which also reduced insulin secretion and thereby worsened β-cell function in PKT patients.

We studied insulin secretion through that of C-peptide instead of insulin because of the different pancreatic drainage in the examined groups. Since C-peptide is not cleared by the liver (38), we used a mathematical model for C-peptide, which allows estimation of β-cell function on the basis of peripheral hormone levels. Thus, in the PKT patients, β-cell sensitivity to glucose is markedly reduced when compared with control subjects. These findings are in line with several studies that also found defects of β-cell function in type 1 diabetic patients after PKT (10,11,13). In contrast, in kidney transplant patients, β-cell sensitivity to glucose was comparable to that in control subjects, and the adaptation index, a measure of releasing adequate increasing quantities of insulin and C-peptide to compensate for the ambient insulin resistance, was even higher than that in control subjects or PKT patients, suggesting an improved β-cell function.

It has been known for a long time that augmented plasma amylin levels precede β-cell failure and that its amyloid is deposited around the pancreatic islets of type 2 diabetic patients (39,40). In the present study, plasma amylin was inversely correlated not only with β-cell sensitivity to glucose, but also with OGIS. Thus, elevated plasma amylin levels are a sign of both impaired β-cell function and insulin resistance and might well serve as a simple peripheral measure of β-cell function.

Impaired β-cell function in type 1 diabetes PKT might be due to 1) previous minor lesions during transplantation, 2) the use of specific immunosuppressive agents, and/or 3) the effects of increased circulating amylin itself. Major graft damage seems unlikely because all PKT patients had a nondiabetic glucose tolerance test and no longer require insulin treatment. However, the PKT patients showed slightly higher A1C levels, which is also reflected by the elevated OGTT plasma glucose concentrations.

Immunosuppressive agents are needed to prevent graft rejection, but may exert a toxic effect on β-cells and impair their function. All of the PKT and kidney transplant patients regularly ingested immunosuppressive agents. However, kidney transplant patients showed increased adaptation index when compared with PKT patients and control subjects. Apart from corticoids, the immunosuppressive treatment consisted of both an antimetabolite and a calcineurin inhibitor. The majority of both patient groups was treated with mycophenolate mofetil as an antimetabolite drug. Regarding calcineurin inhibitors, all PKT patients were under tacrolimus therapy, whereas all but one kidney transplant patient ingested cyclosporine-A. Previous in vitro and in vivo comparisons of both of these drugs on β-cell toxicity yielded conflicting results. While in cell culture, tacrolimus was described to be less (41) or as toxic to β-cells (42) as cyclosporine-A, a recent meta-analysis on diabetogenity in humans (43) and a study on biopsies of transplanted pancreata in PKT patients (44) reported a more harmful effect of tacrolimus than cyclosporine-A for β-cell function. Thus, it cannot be ruled out that tacrolimus therapy at least in part contributed to β-cell impairment in our PKT patients.

Increased plasma amylin levels might exert contrasting effects on glucose and insulin metabolism in different tissues. While a high plasma amylin level is suspected to facilitate the development of type 2 diabetes by β-cell failure (40) and indicates β-cell impairment in our patients, amylin analogs such as pramlintide have been developed for the treatment of type 1 diabetes (16–19). Pramlintide improved glycemic control in type 1 diabetic patients, leading to a reduction of daytime glucose excursions as well as an improvement of A1C levels (16–19). Thus, it appears that amylin might exert beneficial metabolic effects on tissues other than the β-cell. Regardless of the mechanism, plasma amylin appears to be a marker of overt β-cell impairment in humans and might also be harmful for β-cells.

Based on OGIS, insulin sensitivity in PKT and kidney transplant patients was not significantly lower than that in control subjects, although kidney transplant patients tended to show the lowest OGIS values due to a combination of corticoid and immunosuppressive treatment. The similar course of plasma FFAs during OGTT also indicates comparable insulin sensitivity in the adipose tissue of all groups. This indicates that both glucotoxicity and lipotoxicity, which had certainly been present in type 1 diabetic patients with former end-stage renal failure (45,46), can be reversed after successful PKT resulting in nondiabetic glycemia.

In contrast to our findings, several previous studies showed peripheral insulin resistance and elevated FFAs in patients after PKT (47,48). However, these studies were performed in patients taking glucocorticoids, which are known to deteriorate insulin sensitivity and to be diabetogenic (49,50). Nowadays, in most cases, pancreatic transplantation is followed by initial quadruple combination of immunosuppression, including antibody induction therapy and subsequent triple therapy for maintenance of immunosuppression with calcineurin inhibitors, antimetabolites, and glucocorticoids (4,49). If possible, prednisolone is tapered off during the first year of transplantation, thereby improving glucose tolerance (49). All of our PKT patients were on glucocorticoid-free immunosuppression. Considering the impaired β-cell function in PKT patients, our data also underline the necessity of a glucocorticoid-free regimen in these patients, because steroids would not only potentiate the β-cell toxicity of immunosuppressive agents, but also accelerate and favor the re-manifestation of diabetes (44).

In conclusion, type 1 diabetic patients after successful PKT on a glucocorticoid-free regimen exhibit nondiabetic glucose levels but impaired β-cell function. The normal FFA dynamics and insulin sensitivity in PKT patients confirm the benefits and necessity of glucocorticoid-free immunosuppressive treatment. Type 1 diabetic patients after PKT show increased fasting and post–glucose load amylin and increased amylin release in proportion to insulin, which appears to be associated with impaired β-cell function. Thus, amylin might serve as a marker of β-cell impairment in type 1 diabetic patients after combined PKT.

This study was supported by grants from the Austrian National Bank, awarded to R.P. (Project 10787), and the Genomics of Lipid-associated Disorders (GOLD) of the Austrian Genome Research Programme GEN-AU, awarded to F.K.

The authors gratefully acknowledge the technical assistance of H. Regal, PhD, G. Hofstetter, and E. Gauss, MA, from the laboratories of the 3rd Medical Department (Lainz Hospital, Vienna) and the Clinical Institute of Medical and Chemical Laboratory Diagnostics (Medical University, Vienna). We thank Szabo-Skandic (Vienna, Austria) for providing a Fluorescence-ELISA Reader (Bio-Tek FLx800). We thank S. Wagner for editing the manuscript.