A static glucose-stimulated insulin secretion (sGSIS) assay that is significantly predictive of time to diabetes reversal in the human islet bioassay

Abstract

Introduction: Static incubation (static glucose-stimulated insulin secretion, sGSIS) is a measure of islet secretory function. The Stimulation Index (SI; insulin produced in high glucose/insulin produced in low glucose) is currently used as a product release criterion of islet transplant potency.

Research design and methods: Our hypothesis was that the Delta, insulin secreted in high glucose minus insulin secreted in low glucose, would be more predictive. To evaluate this hypothesis, sGSIS was performed on 32 consecutive human islet preparations, immobilizing the islets in a slurry of Sepharose beads to minimize mechanical perturbation. Simultaneous full-mass subrenal capsular transplants were performed in chemically induced diabetic immunodeficient mice. Logistic regression analysis was used to determine optimal cut-points for diabetes reversal time and the Fisher Exact Test was used to assess the ability of the Delta and the SI to accurately classify transplant outcomes. Receiver operating characteristic curve analysis was performed on cut-point grouped data, assessing the predictive power and optimal cut-point for each sGSIS potency metric. Finally, standard Kaplan-Meier-type survival analysis was conducted.

Results: In the case of the sGSIS the Delta provided a superior islet potency metric relative to the SI.ConclusionsThe sGSIS Delta value is predicitive of time to diabetes reversal in the full mass human islet transplant bioassay.

Keywords: Insulin Secretion; Islets of Langerhans Transplantation.

Figure 1

Grouping of measured static glucose-stimulated insulin secretion (sGSIS) Delta values by donor variables. Green dots represent potency values that resulted in rapid diabetes reversal (<5 days) and red dots represent those that resulted in delayed diabetes reversal (≥5 days). The distribution of Delta value was not correlated to any donor variable as can be seen from the nearly equivalent distribution of rapid and delayed diabetes reversal across the range of values on the X-axis. BMI, body mass index; CIT, cold ischemia time.

Figure 2

Work flow of the static glucose-stimulated insulin secretion (sGSIS) potency assay: islets were subjected to sGSIS loaded into the Sephadex bead slurry with a preincubation followed by three consecutive incubations/eluate collections in low, high and a second low-glucose exposure (A). Insulin content was measured by conventional ELISA (B), and the results analyzed and grouped through retrospective analyses into good (C1) or bad (C2) preparations. Islet transplants were performed in chemically induced diabetic immunodeficient mice for all preparations (D), and potency quantified as good (E1) or bad (E2) based on statistical analysis of time to diabetes reversal.

Figure 3

Final logistic regression plots for the Index (A) and the Delta (B) based on the recursive p value plot determination of the ideal cut-points for each metric (C for the Index and D for the Delta). The dashed lines in (C) and (D) indicate the diabetes reversal time cut-points determined by the statistical analysis.

Figure 4

Raw data for the Stimulation Index (SI) (A) and Delta (B) values versus time to diabetes reversal in transplanted chemically induced diabetic immunodeficient mice in potency assessment of 32 consecutive human islet preparations. The same data grouped according to receiver operating characteristic (ROC) analysis determined the cut-off points for both the SI (C) and the Delta (D).

Figure 5

(A, B) Receiver operating characteristic (ROC) curves for the Stimulation Index (SI) (left) and the Delta (right). The analysis demonstrates greater area under the curve (AUC) and higher sensitivity and specificity for the Delta relative to the SI confirming the superiority of the Delta analysis as a predictive metric of time to diabetes reversal of hyperglycemia in the chemically induced diabetic immunodeficient mouse bioassay. (C, D) Kaplan-Meier survival analysis of chemically induced diabetic immunodeficient mouse bioassay transplants grouped based on the cut-off points of each respective potency metric, SI (left) and the Delta (right). For each graph, the solid black line represents islet preparations with potency values greater than or equal to the cut-off point, and the solid red line represents preparations with potency values less than the cut-off point. The dashed lines represent the 95% CIs for each plot.

Figure 6

Diabetes reversal time groupings based on receiver operating characteristic (ROC) analysis for each metric with Mann-Whitney statistical analysis of differences between the two groups. The statistical grouping of both metrics resulted in significantly different mean times to diabetes reversal. The dashed line in each plot represents the mean diabetes reversal time. A depicts the grouping of the Stimulation Index values while B the grouping of the Delta values.

Figure 7

Representative comparison of human pancreata with similar Stimulation Index (SI) values but significantly different Delta values and transplant outcome. The Delta value was predictive of diabetes reversal/non-reversal in both HP while the SI, nearly identical in both preps, failed to distinguish transplant outcome differences. All seven mice transplanted with islets from HP 1 reverted diabetes at day 1, while the animal transplanted with islets from HP 26 failed to reverse diabetes.