Nerve growth factor is associated with islet graft failure following intraportal transplantation

Abstract

Nerve growth factor (NGF) has recently been recognized as an angiogenic factor with an important regulatory role in pancreatic β-cell function. We previously showed that treatment of pancreatic islets with NGF improved their quality and viability. Revascularization and survival of islets transplanted under the kidney capsule were improved by NGF. However, the usefulness of NGF in intraportal islet transplantation was not previously tested. To resolve this problem, we transplanted syngeneic islets (360 islet equivalents per recipient) cultured with or without NGF into the portal vein of streptozotocin-induced diabetic BALB/c mice. Analysis revealed that 44.4% (4/9) of control and 12.5% (1/8) of NGF-treated mice attained normoglycemia (≤ 200 mg/dL) (p = 0.195). NGF-treated islets led to worse graft function (area under the curve of intraperitoneal glucose tolerance tests (IPGTT) on post-operative day (POD) 30, control; 35,800 ± 3,960 min*mg/dl, NGF-treated; 47,900 ± 3,220 min*mg/dl: *p = 0.0348). NGF treatment of islets was also associated with increased graft failure [the percentage of TdT-mediated dUTP-biotin nick-end labeling (TUNEL)-positive and necrotic transplanted islets on POD 5, control; 23.8% (5/21), NGF-treated; 52.9% (9/17): p = 0.0650] following intraportal islet transplantation. Nonviable (TUNEL-positive and necrotic) islets in both groups expressed vascular endothelial growth factor (VEGF) and hypoxia-inducible factor-1α (HIF-1α). On the other hand, viable (TUNEL-negative and not necrotic) islets in both groups did not express VEGF and HIF-1α. In the present study, pre-transplant NGF treatment was associated with impaired survival and angiogenesis of intraportal islet grafts. The effect of NGF on islet transplantation may significantly vary according to the transplant site.

Keywords: apoptosis; diabetes; intraportal; islet transplantation; nerve growth factor.

Figure 1. Effect of NGF treatment for 24 h in culture on in vitro islet viability and function test. (A) Islet viability was tested by fluorescence microscopy using SYTO green (green for viable area) and ethidium bromide (red for dead area). The ratio of the green area relative to total stained (green + red) area was calculated. It was improved by NGF treatment in a dose-dependent manner (control; 84.0 ± 2.41%, NGF 20 ng/mL; 86.5 ± 2.51%, NGF 100 ng/mL; 91.7 ± 1.55%*, NGF 500 ng/mL; 95.5 ± 0.852%**: *p < 0.05, **p < 0.01 vs. NGF 0 ng/mL, ANOVA p = 1.06 × 10⁻⁵). (B) Islet function was tested by glucose-stimulated insulin secretion assay. Stimulation Index was calculated as the ratio of the insulin secretion in the high glucose relative to that in the low glucose. There were no significant differences between these groups (Stimulation Index, control; 1.78 ± 0.345, NGF 20 ng/mL; 1.92 ± 0.347, NGF 100 ng/mL; 1.72 ± 0.203, NGF 500 ng/mL; 1.45 ± 0.279: ANOVA p = 0.735). Data are reported as the mean ± standard error of the mean.

Figure 2. Effect of pre-transplant NGF treatment on glucose control. Streptozotocin-induced diabetic BALB/c mice received syngeneic islets (360 islet equivalents per recipient) cultured for 24 h with or without 2.5 S mouse NGF (100ng/ml) via the portal vein. Blood glucose levels in the control (A) and NGF-treated (B) groups after transplantation. Bold lines represent parameters of normal glycemic mice in both groups. (C) Achievement of normal glycemia up to POD 28 was defined as a non-fasting blood glucose level consistently maintained ≤ 11 mmol/L (200 mg/dL) after transplantation. During the observation period, 44.4% (4/9) of control and 12.5% (1/8) of NGF-treated mice attained normoglycemia (p = 0.195).

Figure 3. Results of intraperitoneal glucose tolerance tests (IPGTT) on POD 30. (A) IPGTT were performed by overnight fasting for 12 h and then injecting mice with 2.0 g/kg body weight of glucose solution followed by tail vein blood samples at 0, 15, 30, 60, 90 and 120 min after injection. (B) Area under the curve of IPGTT in the control group was significantly lower than that of the NGF-treated group (control; 35,800 ± 3,960 min*mg/dl, NGF-treated; 47,900 ± 3,220 min*mg/dl: p = 0.0348). Data are reported as the mean ± standard error of the mean.

Figure 4. Histological findings 2 h after transplantation. Insulin staining (A), NGF staining (B), TUNEL (C), and H&E staining (D) were performed in the control group. Insulin staining (E), NGF staining (F), TUNEL (G), and H&E staining (H) were performed in the NGF-treated group. The same islet was used for (A–D) and (E–H). Most of the transplanted islets were TUNEL-negative (not apoptotic) (control; 20/22, NGF-treated; 46/49: p = 0.651) (C and G), and all of them were not necrotic in both groups (D and H) (see Table 1). NGF-positive liver regions were observed in both groups, and most parts of those regions corresponded with TUNEL-positive regions. The ratio of transplanted islets that were in contact with NGF and TUNEL-positive liver regions was 59.1% (13/22) in the control group, and 85.7% (42/49) in the NGF-treated group (*p = 0.0130) (B, C, F and G) (see Table 1). (I) The ratio of intra-islet NGF positive area to each transplanted islet area was 55.9 ± 6.83% in the control, and that was 77.4 ± 3.73% in the NGF-treated group (*p = 0.0125). Pre-transplant co-culture with NGF significantly increased intra-islet detection of NGF 2 h after transplantation. Calibration bar = 100 μm. Data are reported as the mean ± standard error of the mean.

Figure 5. Histological findings on POD 5. Insulin staining (A and F), TUNEL (B and G), HIF-1α staining (C and H), VEGF staining (D and I), and H&E staining (E and J) were performed in the control group. Insulin staining (K and P), TUNEL (L and Q), HIF-1α staining (M and R), VEGF staining (N and S), and H&E staining (O and T) were performed in the NGF-treated group. The same islet was used for (A–E), (F–J), (K–O) and (P–T). Nonviable (TUNEL-positive and necrotic in H&E staining) islets in both groups (A–E, K–O) expressed VEGF and HIF-1α in all islet area and were in contact with nonviable and severely hypoxic (HIF-1α positive) liver regions. The percentage of nonviable islets was 23.8% (5/21) in the control (A–E) and 52.9% (9/17) in the NGF-treated group (K–O). Viable (TUNEL-negative and not necrotic in H&E staining) islets in both groups (F–J, P–T) did not express VEGF and HIF-1α at all, and the liver parenchyma in the immediate vicinity appeared normal. The percentage of viable islets was 76.2% (16/21) in the control (F–J) and 47.1% (8/17) in the NGF-treated group (P–T) (p = 0.0650). Cellular infiltration was observed around the dead islets and liver regions in both groups. Arrows indicate cellular infiltration (E and O). Calibration bar = 100 μm.

Figure 6. Blood vessel numbers of viable grafts on POD 5. Histological staining for CD31 was performed to count the vessel numbers of viable grafts. Islet viability had been confirmed by TUNEL method and H&E staining before vascular assessment. In the control group, 16 transplanted islets were viable from five mice (A). In the NGF-treated group, eight transplanted islets were viable from five mice (B). The blood vessel numbers in viable grafts were equal between these groups (control: 4.17 ± 0.834 × 10⁻⁴/µm²; hepatectomy: 4.34 ± 0.718 × 10⁻⁴/µm², p = 0.886) (C). Arrows indicate typical blood vessel morphology in high magnification (A and B). The dotted line is drawn along the margin of the transplanted islets. Calibration bar = 100 μm (low magnification) and 20 μm (high magnification).