Metabolic and Immunological Features of the Failing Islet-Transplanted Patient

Patients were retrospectively split into two groups: 17 patients with long-term functioning graft (long fx) (C-peptide >1.0 ng/ml for more than 12 months) and 20 patients with short-term functioning graft (short fx) (C-peptide levels <1.0 ng/ml within 12 months of islet transplantation). Please see the online appendix (available at http://dx.doi.org/10.2337/dc07-1831) for a full description of methods.

Blood samples were collected from these patients on a monthly basis after islet transplantation. We measured metabolic and immune peripheral markers obtained at baseline (before islet transplantation), from a second time point at the peak of transplanted islet function (defined by the highest C-peptide level: mean ± SE 4 ± 1 vs. 11 ± 2 months for patients with short fx and long fx, respectively), and at time of failure in the short fx group (9 ± 1 months) or at the latest time point available in the long fx group (24 ± 4 months).

The two groups of patients appeared similar regarding baseline characteristics (supplemental Table).

The long fx group showed higher C-peptide levels (P <0.01 at both time points), lower exogenous insulin requirement (P = 0.04 and P = 0.01 at the second and third time points, respectively), and lower glycated hemoglobin compared with that of the short fx group (Fig. 1A and B and data not shown).

The number of islets was similar in the two groups (equivalent number: short fx 542,776 ± 63,606 vs. long fx 585,600 ± 59,481; NS [supplemental Table]), excluding a potential bias due to a different number of transplanted islets. Furthermore, an l-arginine test performed 1 month after islet transplantation showed a similar insulin secretory reserve in the two groups (area under the curve of insulin release 2,674 ± 516 vs. 2,772 ± 338 for short fx vs. long fx; NS).

A higher proinsulin–to–C-peptide ratio was evident in the short fx group, particularly at the third sampling (failure vs. long-term function) (short fx 32.4 ± 11.9 vs. long fx 7.0 ± 1.6; P = 0.04), (Fig. 1C and D). These data suggest a disproportion between the amount of proinsulin produced and the amount processed. Levels of activin-A, which is considered a good marker of α-cell activity because it is cosecreted with glucagons (16,17), were slightly higher in the short fx group than in the long fx group (NS, data not shown). Amylin level as a marker of fibrillogenesis did not differ between the two groups (data not shown).

Homeostasis model assessment of insulin resistance (HOMA-IR) has been previously validated in kidney-transplanted patients as one of the most reliable methods to assess insulin resistance (18). HOMA-IR scores were similar in the two groups at the time of islet transplantation, indicating the groups' homogeneity. HOMA-IR was higher in the short fx group than in the long fx group at the peak of islet function (7.5 ± 2.3 vs. 3.5 ± 0.5, respectively, P = 0.02) (Fig. 1E and F), indicating that insulin resistance is already present in the short fx group long before failure of islets.

During follow-up, sIL-2r was higher in the long fx group, particularly at the third time point (4,901 ± 1,033 vs. 2,754 ± 490 pg/ml for long fx and short fx, respectively; P = 0.04) (Fig. 1G and H). sIL-2r is a marker of T-cell activation, released peripherally during any immunological process, but it has a role in modulating CD4⁺CD25⁺ regulatory T-cells' function. The persistence of high sIL-2r levels suggests a higher percentage of CD4⁺CD25⁺ cells with regulatory ability releasing the soluble form of the receptor that they expressed (CD25 is an interleukin-2 receptor). High sIL-2r levels could also be a sign of some activation/tolerogenic processes.

High levels of autoantibodies are associated with increased failure of the islet graft (12). A rise in autoantibodies to protein tyrosine phosphatase isoforms IA-2 was observed in the short fx group (short fx first time point 1.07 ± 0.62 arbitrary units [AU] vs. second time point 15.45 ± 13.94 AU, P <0.05; long fx first time point 2.52 ± 1.21 AU vs. second time point 2.37 ± 1.30 AU, NS) but not in the long fx group (data not shown). Antibodies to glutamic acid decarboxylase also increased in the short fx group, though without reaching statistical significance. It is not known whether members of the long fx group are simply less prone to the recurrence of autoimmunity or are more immunosuppressed, but it is generally accepted that the immunosuppressive regimens (including cyclosporine) currently used in islet transplantation are not specific for autoimmunity (19).

No patients showed detectable levels of peripheral annexin V (data not shown), but soluble FasL (sFasL) (a marker of apoptotic process) was higher in the short fx group at the earlier time point, when islet function was peaking.

An increase in sFasL levels was particularly evident in the short fx group (from 0.570 ± 0.422 ng/ml at baseline to 1.862 ± 1.639 ng/ml at islet function peak; P <0.05). This was not seen in the long fx group, where sFasL was more stable (from 0.356 ± 0.136 ng/ml at baseline to 0.706 ± 0.424 ng/ml at islet function peak). This increase was maintained even at the third time point of follow-up (short fx group 0.701 ± 0.452 ng/ml vs. long fx group 0.387 ± 0.161 ng/ml). There were no differences in sFasL levels between the two groups (data not shown).

In our study, increased insulin resistance and altered insulin processing are evident before failure of the graft (2,3,11,20,21). The early failing group showed high levels of sFasL, low levels of sIL-2r, and a rise in autoantibody titers (12,22), indicating that immunological phenomena also predict failure of islets.

Metabolic and immunological markers could help in identifying patients at high risk for early graft failure (23,24). Our data can help to define early markers that could be used as routine tests to identify or predict islet rejection.