Transplantation

With the introduction of the glucocorticoid-free immunosuppressive regimen, there was a considerable increase of the success rate of clinical islet transplantation (¹). However, within some 5 years most patients had to resume exogenous insulin therapy, demonstrating a progressive deterioration of islet graft function over time (²). The reasons for this failure are not known, but there are obviously other reasons than just pure allogeneic graft rejection. To obtain a more complete understanding of this process, a detailed morphological characterization of intraportally grafted islets in autopsy material would be of great interest. In a recent report, we described the presence of wide spread amyloid depositions, consisting of the β-cell islet amyloid polypeptide (IAPP), in human pancreatic islets intraportally grafted into a diabetic patient 60 and 6 months before his death (³). In previous studies, we and others had described amyloid deposits in human islets or mouse islets, transgenic for human IAPP, transplanted into nude mice (^{4, 5}). This may be of great interest because aggregated IAPP, as amyloid fibrils or as oligomeric assemblies, is believed to be of importance for the loss of β cells in type 2 diabetes (^{6, 7}). Consequently, deposition of IAPP amyloid in transplanted islets may indicate a mechanism for loss of β cells. Our first report was based on the findings in one individual only. The present study was therefore carried out to extend our knowledge as regards the formation of amyloid in clinically transplanted human islets. Because there is evidence to suggest that there is a prohormone convertase 2 (PC2) deficiency in experimentally grafted human islets (⁸), we also examined the presence of PCs in these clinically grafted islets.

RESULTS

Liver material from four deceased islet-bearing recipients has been made available for the demonstration of amyloid depositions in grafted pancreatic islets. Neither of the recipients was obese and they had all therapy-controlled hypertension. Excised pancreas was preserved in the two-layer cold storage solution or in University of Wisconsin solution and the ischemic period lasted 2 to 14 hr. Two of the recipients (identity acronyms 001, 002) were described by the Edmonton team (⁹), one (003) by the Milan group (¹⁰), and the final one (004) by the Nordic Network for Clinical Islet Transplantation (³).

Patient 001

This patient (⁹) died almost 2 years after transplant from a methadone overdose. He was using a small dose of insulin at death and had slightly elevated serum HbA1C levels (Table 1). Fasting serum C-peptide concentrations were fairly low. The implant was a mixture from three donors, two with body mass index (BMI) more than 30 kg/m² and age varying from 32 to 52 years. A total of 28 islets were recovered in the liver blocks available. These islets were possible to follow by means of serial sectioning. In seven of them, amyloid deposits were found in grades 1+ to 3+ (Table 2). The amyloid was spread in the affected islets and appeared around capillaries and at the outer border of the islets (Fig. 1A, B). More nodular deposits were sometimes seen. Intracellular amyloid was not possible to identify using light microscope. As expected, the material was labeled with antibodies against IAPP (not shown).

Patient 002

This patient (⁹) was off insulin 17 months after transplant when he died of a myocardial infarct. He had the lowest HbA1C value of the four patients examined and the second highest fasting C-peptide value (Table 1). Islets were from two male donors aged 50 and 53 years and one with a BMI more than 30 kg/m². Only 14 islets were available for assessment and none of them were found to contain amyloid deposits, neither after Congo red staining viewed in cross polarized light or with fluorescence microscope (Table 2).

Patient 003

This patient (¹⁰) died of a myocardial infarct more than 4 years after transplant. She was off insulin, had slightly elevated serum HbA1C levels and the highest serum C-peptide concentration of the four investigated patients (Table 1). This patient was the only one on glucocorticoid immunosuppressive regimen. Implanted islets were isolated from two male donors both with BMI less than 28 kg/m². In the sections still available for amyloid diagnosis, nine islets were identified. In four of them, amyloid deposits were recognized, two of them with pronounced deposits (Table 2; Fig. 1D, E). The amyloid had typical affinity for Congo red and showed bright green birefringence.

Patient 004

The finding of amyloid in this patient was reported earlier (³). This patient also died from a myocardial infarct 60 months after the first transplant. He had slightly increased HbA1C levels, fairly low serum C-peptide concentrations, and required a low dose of insulin. The first transplant was isolated from two female donors whereas the second and third implants were from single male donors, given 4 and 54 months after the first implant, respectively. These four donors had the highest age (54–64 years) and one of them had a BMI more than 30 kg/m². High numbers of islets had been recovered because the availability of liver tissue was essentially unlimited. More than 40% of the examined islets contained amyloid deposits in greater or smaller amounts as assessed by means of Congo red staining (ordinary and polarized light) (Fig. 1C) and immunolabeling at both the light and electron microscopical level.

Immunolabeling With Antibodies Against PC1/3 and PC2

In double immunolabeling experiments, monoclonal antibodies against PC1/3 and PC2 antibodies were used together with antiinsulin antibodies. PC2 labeling was detected in insulin containing cells and also in noninsulin reactive cells in all three investigated grafts (Fig. 2A, B). PC1/3 reactivity was restricted to insulin reactive cells (not shown). The fairly strong and even labeling with antibodies against the processing enzymes did not differ from that seen in islets in normal human pancreas.

DISCUSSION

We reported earlier the wide-spread occurrence of amyloid in islets transplanted into the liver of an individual with type 1 diabetes and suggested that this may be a common event of importance for the long-term loss of β-cell function in such cases. Indeed, when human islets or islets from transgenic mice expressing human IAPP are transplanted into nude mice, amyloid develops rapidly in the transplanted islets (^{4, 5, 11, 12}). The present study extends and strengthens our previous observation that development of IAPP amyloid deposits in transplanted islets is common also in the clinical situation when islets are infused into the portal system of the liver.

In their previous histological assessment of clinical islet grafts, the Edmonton group reported that the two autopsy specimens available were both negative for amyloid. We can now demonstrate that at least in one of the two patients (001), there were amyloid depositions as demonstrated by Congo red staining with polarization microscopy (Fig. 1A,B). The reasons for this discrepant reporting can simply be that the islets investigated in the first report were, indeed, islets lacking amyloid. In the present study, as many as three of four islets were amyloid negative, and, in the Nordic network patient (³), approximately 6 of 10 islets were amyloid negative. It should, however, be pointed out that the diagnosis of amyloid in pancreatic islets may be difficult and requires considerable experience. By the same token, it might be argued that the fact that we were unable to demonstrate islet amyloid in the second Edmonton patient (002) was because just 14 islets have been scrutinized. It might well be that with a greater material some amyloid positive islets might have shown up here as well. It should be pointed out that in our first published patient (³), sufficient amounts of fixed tissue were present to make electron microscope examinations possible. In the three newly added patients, this was not the case.

Although our clinical material is still very small, it is nevertheless intriguing to notice that the patient in whom we found no amyloid depositions in the grafted islets (002) had the shortest observation time. The two patients with most depositions (004 and 003) had survived for more than 5 and 4 years, respectively, after the first islet transplantation. Amyloid formation in the clinical setting may be a slow process and its importance for the functional impairment of grafted islets is still not known. There is, however, strong support for the effect of islet amyloid on islet function in a primate diabetes model (¹³).

The mechanisms behind the formation of amyloid in the grafted islets are not fully understood. Insufficiencies in the revascularization of the grafted islets have been one intriguing reason for the long-term failure of grafted (¹⁴). It is therefore of great interest that it was recently shown that intramuscular islet transplantation might offer advantages, because it induces an almost normal blood supply to the grafted islets (¹⁵). Against that background, it will be important to analyze such islets for amyloid deposition. It should be noted though that there was no difference in amyloid development in human islets transplanted into the liver, subcapsularly in the kidney or in the spleen of nude mice (⁵). Another important issue is the species-dependent difference in amyloid formation. The fibril formation propensity varies strongly between species depending on the primary structure of IAPP (¹⁶). Xenotransplantation with porcine islets—although it involves considerable problems—may therefore offer an interesting alternative because porcine IAPP does not aggregate into amyloid fibrils (¹⁷).

There is evidence that pro-IAPP may be even more amyloidogenic than mature IAPP and that the first, intracellular amyloid is formed from the propeptide. The report of reduced β-cell PC2 immunolabeling and hyperproinsulinemia after transplantation of human islets to nude mice was therefore interesting (⁸). However, in this study, we detect PC2 labeling in β-cells from all three patients examined. Therefore, it is possible that the disappearance of PC2 is a transient phenomenon, which still may be of importance in the initiation of islet amyloid formation. This issue deserves more studies.

Our finding that amyloid often develops in transplanted islets should increase our efforts to unravel the effect of aggregation of IAPP on the function of transplanted human islets. Experimentally, it has been shown that aggregated human IAPP added to isolated islets in vitro exerts a toxic effect to the β cell. Several different mechanisms have been suggested, including formation of pathological cell membrane pores. It should be mentioned, however, that early islet amyloid formation takes place intracellularly (^{4, 18, 19}), indicating that interference with intracellular events is of relevance. The site of the early aggregation events may be important for the development of future therapeutical alternatives.

MATERIALS AND METHODS

Patient Characteristics

Relevant information on glucose homeostasis, donor numbers, and time posttransplantation is given in Table 1 and on islet amyloid is given in Table 2. The information has been extracted from earlier publications (^{3, 9, 10}).

Staining Methods

From three of the patients (001, 002, and 004), several blocks of formalin-fixed and paraffin-embedded liver autopsies were available. Material from the Italian case (003) was in the form of unstained sections from seven blocks, containing 1-2 formalin-fixed and paraffin-embedded specimens. Islets were localized in the sections by means of immunostaining for insulin and glucagon. All specimens were stained for amyloid with Congo red and examined under ordinary and polarized light (²⁰). Other specimens were immunostained with rabbit antibodies against IAPP (A110; [²¹]) and visualized by goat anti-rabbit antibodies labeled with Alexa-488 (Invitrogen, Stockholm, Sweden). The amount of amyloid deposition in each single islet based on the viewing of one or several Congo red-stained sections was estimated on a 5-graded scale (0 to 4+).

To investigate whether PC2 deficiency remains in clinically transplanted islets and whether PC1/3 expression is influenced, we performed double immunolabeling of sections from 004, 001, and 002 with well-characterized mouse monoclonal antibodies against PC2 and PC1/3 (²¹) together with antiinsulin antibodies produced in guinea pig. Sections used for PC2 labeling were subjected to pretreatment with hot 0.02 M sodium citrate buffer for 20 min before overnight incubation at 4°C with PC2 antibodies, diluted in tris HCl buffer, pH 7.4, containing 0.15 M NaCl, and 0.001% Tween. Sections were incubated with PC1/3 antibodies, diluted in tris HCl buffer, pH 7.4, containing 0.15 M NaCl, overnight at room temperature. PC2 and PC1/3 reactivity was detected by horseradish peroxidase-labeled goat anti-mouse antibodies (Dako, Glostrup, Denmark) and visualized with 3,3′-diaminobenzidine tetrahydrochloride. After extensive rinses in water and tris HCl buffer, pH 7.4, with 0.15 M NaCl, sections were incubated with guinea pig antiinsulin antibodies (Dako) overnight followed by Alexa 488 labeled goat anti-guinea pig antibodies (Invitrogen). To facilitate islet detection, sections were incubated with rabbit antiglucagon antibodies (Dako) overnight and the reactivity was detected with horseradish peroxidase-labeled goat anti-rabbit antibodies (Dako) and visualized with 3,3′-diaminobenzidine tetrahydrochloride before Congo staining.

REFERENCES