Differential Effects of Linagliptin on the Function of Human Islets Isolated from Non-diabetic and Diabetic Donors

Linagliptin is a dipeptidyl Peptidase-4 (DPP-4) inhibitor that inhibits the degradation of glucagon-like peptide 1 (GLP-1), and has been approved for the treatment of type 2 diabetes (T2D) in clinic. Previous studies have shown linagliptin improves β cell function using animal models and isolated islets from normal subjects. Since β cell dysfunction occurs during diabetes development, it was not clear how human islets of T2D patients would respond to linagliptin treatment. Therefore, in this study we employed human islets isolated from donors with and without T2D and evaluated how they responded to linagliptin treatment. Our data showed that linagliptin significantly improved glucose-stimulated insulin secretion for both non-diabetic and diabetic human islets, but its effectiveness on T2D islets was lower than on normal islets. The differential effects were attributed to reduced GLP-1 receptor expression in diabetic islets. In addition, linagliptin treatment increased the relative GLP-1 vs glucagon production in both non-diabetic and diabetic islets, suggesting a positive role of linagliptin in modulating α cell function to restore normoglycemia. Our study indicated that, from the standpoint of islet cell function, linagliptin would be more effective in treating early-stage diabetic patients before they develop severe β cell dysfunction.

GLP-1 receptor (GLP-1R) agonists that mimic GLP-1 action but are resistant to DPP-4 degradation, and another is DPP-4 inhibitors that block GLP-1 degradation and thus prolong the activity of endogenously expressed native GLP-1.
Linagliptin (Tradjenta, Boehringer Ingelheim) is a specific DPP-4 inhibitor that is in clinic use for the treatment of T2D patients 18,19 . Studies have shown linagliptin increases GLP-1 concentrations in blood and improves β cell survival and function 20,21 . In addition, since GLP-1 is also produced locally within pancreatic islets, and its expression increases with the progress of diabetes development [22][23][24][25][26][27] , linagliptin may exert insulinotropic and glucagonostatic effects directly on the islets by elevating local GLP-1 concentrations. Indeed, Shah et al. have shown linagliptin significantly increases GLP-1 concentration in the culture media of human islets, and protects β cells from gluco-, lipo-, and cytokine-induced toxicity 28 . However, it is well documented that β cell dysfunction occurs with T2D development, in which β cells gradually lose their characteristic markers such as transcription factors and functional proteins, and become glucose insensitive 29,30 . Therefore, islets of diabetic patients may respond to linagliptin treatment differently from normal human islets. This study is thus designed to evaluate the direct effects of linagliptin on human islets with different diabetic conditions. To accomplish this, we employed human islets isolated from non-diabetic and diabetic donors, and examined whether linagliptin affects their hormone secretion including GLP-1, glucagon, and insulin.

Results
Expression of GLP-1 and DPP-4 in human islets supported a local effect of linagliptin on islet cell function. Classically deemed as a gut hormone, GLP-1 has also been shown to be produced locally within pancreatic islets as GLP-1 and glucagon share the same precursor proglucagon that is expressed by α cells [24][25][26][27] . Using a monoclonal antibody that specifically recognizes the cleaved and amidated C-terminal end of active GLP-1, we confirmed GLP-1 expression in the islets of human pancreatic tissue, in which GLP-1 and glucagon were co-detected in about 70% of α-cells (Fig. 1A). To determine whether linagliptin could affect local GLP-1 concentration, we examined whether DPP-4 was expressed in the islet cells. As shown in Fig. 1B, DPP-4 was detected in the lysates of human islets, which appeared as a highly specific band at expected molecular weight of 88 KDa. To determine its cellular localization, we performed immunofluorescence triple staining for DPP-4, insulin, and glucagon using human pancreatic slices. We found that DPP-4 was predominantly expressed in β cells, with much lower level expression in α cells as well as in exocrine cells (Fig. 1C). Intra-islet expression of GLP-1 and DPP-4 supported that linagliptin could have significant effects on islet cell function through protecting local GLP-1 from degradation.

Effects of linagliptin treatment on basal insulin and glucagon secretion of human islets varied between diabetic and non-diabetic donors. Previous studies have shown beneficial effects of linagliptin
on β cell survival and function in islets isolated from normal human donors 28 . Islets of diabetic patients, however, might respond to linagliptin differently. To test this, we employed human islets isolated from non-diabetic (ND, n = 5) and diabetic (T2D, n = 5) donors (Table 1), and incubated them with 0 (untreated), 5 nM, 30 nM or 50 nM linagliptin for 3 days with daily media refreshing. The culture media was collected and used to evaluate their basal hormone secretion under normal glucose concentration of 5.5 mM.
As shown in the representative sets of experiment (Fig. 2), linagliptin treatment significantly increased GLP-1 concentration in the culture media of both ND and T2D islets in a dose-dependent manner, with the highest concentrations reaching ~5 times of that in untreated groups ( Fig. 2A,B). The dramatic elevation of GLP-1 in the islet culture had a tendency to stimulate basal insulin secretion (Fig. 2C,D) and to inhibit basal glucagon secretion (Fig. 2E,F), but the significance varied between normal and T2D islets. Specifically, normal islets did not significantly alter their basal insulin (Fig. 2C) or glucagon (Fig. 2E) secretion in response to linagliptin-mediated GLP-1 elevation. This was not surprising because the islets were cultured at normal glucose of 5.5 mM and studies have shown incretin effect on islet hormone secretion is glucose-dependent [31][32][33][34] . In T2D islets, however, basal insulin secretion was significantly stimulated and glucagon secretion significantly inhibited by linagliptin when its concentrations reached 30 nM − 50 nM (Fig. 2D,F). These data suggested that the glucose-dependent incretin responses diminished in T2D islets.
Linagliptin improved β cell function with different efficacies in normal and diabetic islets. Next we examined whether linagliptin affected glucose-stimulated insulin secretion (GSIS), a key feature of β cell function, in human islets of different diabetic conditions. The islets were incubated at either low glucose (2.5 mM) or high glucose (25 mM) in the presence of various concentrations of linagliptin, and insulin concentration in the supernatants was measured. Our data showed that, without linagliptin, non-diabetic human islets (n = 5) often had stimulation indices of 1.5-3.0, whereas T2D (n = 5) islets had stimulation indices of 0.7-1.5. As shown in the representative sets of experiment (Fig. 3), linagliptin was highly efficient in improving GSIS of ND islets, in which all linagliptin dosages significantly improved their stimulation indices ( Fig. 3A and B). In contrast, the T2D islets only showed significant GSIS improvement at 50 nM ( Fig. 3C and D), suggesting linagliptin was less effective in improving β cell function of T2D islets than that of normal islets.
Because linagliptin treatment resulted in similar GLP-1 elevation between ND and T2D islets ( Fig. 2A,B), the reduced efficiency of linagliptin in T2D islets might be attributable to the change of GLP-1 receptor (GLP-1R) expression in the islets. Indeed, previous studies have shown GLP-1R expression in T2D human islets was significantly lower than that of normal islets 29,30 . To confirm whether this was true in the human islets that were used in our study, we performed western blotting assays. Our data showed that GLP-1R expression was indeed significantly reduced in T2D islets compared to that in ND islets ( Fig. 4A and B). DPP-4 expression also decreased in T2D islets ( Fig. 4A and C), consistent with a previous study 35 . Linagliptin treatment led to increased GLP-1 versus glucagon production in human islets. In addition to evaluating linagliptin effects on β cell function, we also investigated whether linagliptin affected α cell hormone production. It is known that proglucagon can be cleaved to produce either glucagon or GLP-1 via distinct post-translational processes in pancreatic islets, specifically by α cells 22 . Note: the term α cells herein was used to describe islet cells expressing proglucagon (instead of glucagon). We examined whether linagliptin had an effect on GLP-1 vs glucagon production in human islets of ND (n = 5) and T2D (n = 5) patients. To accomplish this, the human islets were treated with various concentrations of linagliptin for 3 days. Then the islet lysates were collected and processed for GLP-1, glucagon and total protein measurements. In ND human islets, without linagliptin treatment, GLP-1 production was ~30 pg/ug islet protein, whereas glucagon was produced at ~2000 pg/ug protein. This is consistent with previous studies showing proglucagon in pancreatic islets is processed to Shown are representative data obtained from the 3 rd day culture. All data were normalized with the total protein of islet lysates that were collected at the end of the experiments (per ug protein), and expressed as Mean ± SEM. *p < 0.05; **p < 0.01 when compared to untreated group.
Furthermore, we examined GLP-1 and glucagon expression by immunohistochemistry following linagliptin treatment of ND and T2D human islets. GLP-1 and glucagon were co-expressed and readily detected in most α cells with or without linagliptin treatment (Fig. 6A). Quantification of GLP-1 positive cells (GLP-1 + ) and glucagon positive (glucagon + ) cells showed increase in the ratio of GLP-1 + : glucagon + cells with increasing concentrations of linagliptin, in both ND and T2D islets (Fig. 6B). It should be noted that the number of GLP-1 + or glucagon + cells did not reflect the amount of GLP-1 or glucagon produced in these cells due to the nature of the assays. For example, a cell would become GLP-1 positive in immunohistochemistry as long as its expression reached the detectable level. The increased number of GLP-1 + cells indicated that linagliptin treatment resulted in more α cells to express detectable GLP-1, which was in agreement with the increased GLP-1 production in these islets (Fig. 5)

Discussion
DPP-4 inhibitors including linagliptin, sitagliptin, and vildagliptin have been shown to enhance β cell function of T2D patients as assessed by physiological parameters, and the effect is mainly attributed to upregulation of circulating GLP-1 [38][39][40] . Little information was available regarding whether and how a DPP-4 inhibitor could directly affect islet cell function in diabetic patients. In this study, we examined this matter using human islets isolated from diabetic donors and compared it with those from non-diabetic donors. Our data showed that linagliptin significantly improved β cell function, but the effectiveness reduced in T2D islets compared to normal islets. The differential effects were attributable to reduced GLP-1R expression in diabetic islets, which agreed with β cell dysfunction during diabetes progression 29,30 . Furthermore, we examined whether linagliptin affected α cell function, and found that linagliptin significantly up-regulated the relative production of GLP-1 vs glucagon in α cells of both non-diabetic and diabetic human islets, which would be beneficial for restoring normoglycemia in diabetes treatment.
DPP-4 is a widely expressed protease. Omar et al. have recently shown DPP-4 is expressed in human islets, and its expression decreases with T2D development 35 , which is in agreement with our own discoveries. Nonetheless, there is some discrepancy: their study showed DPP-4 expression in α cells of human islets (and in β cells of mouse islets), while our data showed that DPP-4 was predominantly expressed in β cells of human islets. The DPP-4 antibody used in our study was commercially available and our western blotting assay confirmed that it was highly specific -it recognized one band at the expected molecular weight of DPP-4 in the full lane of whole islet extracts with little background staining ( Fig. 1B and Fig. S2). The discrepancy with Omar et al. study is most likely due to the use of different antibodies and/or different staining protocols.
Linagliptin's effects on normal human islets have been examined previously by Shah et al., in which the investigators have found that linagliptin restores β cell function through GLP-1 stabilization 28 . In addition, they have shown linagliptin has protective effects on human islets against cytokines-induced toxicity as well as glucoand lipotoxicity that are often associated with diabetes development 28 . Our results on non-diabetic islets supported their discoveries. Furthermore, we obtained informative data regarding how the islets of diabetic patients responded to linagliptin directly. We found that linagliptin significantly increased GLP-1 concentration to the similar extent in both normal and T2D islet cultures despite different DPP-4 expression levels, but its effectiveness on improving GSIS differed. The phenomenon could be readily explained by the expression of GLP-1R in islets of patients at different stages of their diabetes. In some severe cases, β cells might be so dysfunctional that they wouldn't respond to linagliptin at all. In fact, we did encounter such a case, in which even 200 nM linagliptin didn't significantly improve β cell function in response to high glucose (supplemental Fig. S1). Together with β cell protection effect of linagliptin 28 , these results indicate linagliptin would be more beneficial and more effective when used in clinic for patients at early stage of T2D. Another interesting discovery in the study is that linagliptin did not significantly affect basal insulin and glucagon secretion of ND islets, but did in T2D islets at normal culture condition, although GLP-1 levels were elevated to similar levels (Fig. 2). It is well documented that GLP-1 stimulates insulin secretion and suppresses glucagon secretion in a glucose-dependent manner [31][32][33][34] . Therefore, it is not surprising that basal insulin and glucagon secretion in ND islets was not altered by linagliptin-mediated GLP-1 elevation. In contrast, T2D islets might have developed certain degree of β-cell dysfunction so that the glucose-dependence of GLP-1 action was compromised in these islets. On the other hand, it should be noted that this is beneficial because increased insulin and decreased glucagon is good for lowering blood glucose in T2D patients.
Furthermore, although hyperglucagonemia is commonly observed in T2D patients, we did not detect any significant change in basal glucagon secretion in cultured T2D islets compared to normal human islets (Fig. 2). It should be noted that the mechanisms underlying glucagon dysregulation in diabetes are not completely understood, but appear to be attributable to many factors that are affected by insulin resistance and hyperglycemia. In particular, studies have shown that impaired paracrine insulin inhibition (due to insulin resistance) [1][2][3]41 , impaired blood glucose-sensing in the α-cells 42 , and autonomic dysfunction in the nervous system 43,44 , are involved in this process. In this study, the cultured T2D human islets did not show increase in basal glucagon secretion compared to normal human islets (Fig. 2). This could mainly because: (1) the islets were cultured at normal glucose concentrations, thus not affected by hyperglycemia; and (2) there was no effect from autonomic nervous system in isolated islets culture.
The effect of linagliptin on α cell function, specifically hormone production and secretion, was also investigated in this study. It has become increasingly clear that α cells produce not only glucagon, but also GLP-1, although the latter was produced at much lower level than the former [22][23][24][25][26][27] . In addition, studies have shown GLP-1 expression in α cells increases with T2D development, suggesting they play an integral role in the attempt to regain normoglycemia 22,23,26,27 . Our results on untreated ND and T2D islets aligned well with these studies as we showed that GLP-1 was produced at ~1.5% of glucagon level in ND islets, whereas it was ~3.5% in T2D islets. Moreover, our data demonstrated that linagliptin significantly increased GLP-1 vs glucagon production in both ND and T2D islets. The number of α cells expressing detectable level of GLP-1 was also significantly upregulated by linagliptin in both ND and T2D islets. These data argue for a positive effect of linagliptin on α cells with regard to hyperglycemia alleviation.
In summary, our study provided important information regarding how linagliptin would affect islet cell function in diabetic patients when used in clinic. From the standpoint of islet cell function, linagliptin would be more effective in treating early-stage diabetes.
Human islets culture and treatment. Freshly isolated human islets were obtained from the Integrated Islet Distribution Program (IIDP) at City of Hope that was funded by National Institute of Diabetes and Digestive and Kidney Diseases (NIDDK, 1UC4DK098085). The donor and human islet information was listed in Table 1. The study was designated as "Not Human Subject Research" by the Institutional Review Board of Tulane University. Upon receiving the islet shipment, the islets were washed with fresh culture media (CMRL-1066 supplemented with 0.1 g/L L-glutamine, 10% FBS, 10 mM HEPES, 0.2% Sodium Bicarbonate, and 0.1 mg/ml ciprofloxacin), allocated into 48-well plates at a density of 150 Islet Equivalent Quantities (IEQ) per 0.5 ml per well, and cultured at 37 °C in a humidified incubator containing 5% CO 2 .
For linagliptin treatment, linagliptin (1 mg/ml stock solution) was diluted in the islet culture media and added into each well to achieve 5 nM, 30 nM or 50 nM final concentrations. Untreated islets (that is, 0 nM linagliptin) were included as control. The culture media was collected every day and replaced with fresh media containing the same concentrations of linagliptin, and cultured for up to 4 days.
Hormone measurements. The concentrations of GLP-1, insulin and glucagon in the islet culture media were measured using corresponding ELISA kits from ALPCO Diagnostics (Salem, NH). Specifically, active GLP-1 was measured with GLP-1 (active 7-36) ELISA kit (cat# 43-GP1HU-E01), glucagon with glucagon (human, mouse, rat) ELISA kit (cat# 48-GLUHU-E01), and insulin with human insulin ELISA kit (cat# 80-INSHU-E01.1). All measurements were performed according to the manufacturer's protocols. To determine the hormone expression (production) in the human islets after linagliptin treatment, the islets were washed with PBS, and lysed in 100 ul/well RIPA buffer containing a cocktail of protease inhibitors and phosphatase inhibitors. Following a freeze/ thaw cycle, and 15 minutes' incubation on ice, the samples were centrifuged at 4 °C, 13000 rpm for 10 minutes. The supernatants (cell lysates) were processed for hormone measurements as described above. In addition, total protein concentrations of the cell lysates were measured using Pierce BCA Protein Assay kit (Thermo Fisher Scientific, Waltham, MA).
SCieNTifiC RepoRts | 7: 7964 | DOI:10.1038/s41598-017-08271-9 Glucose-stimulated insulin secretion. GSIS was performed essentially as described previously 45 except that various concentrations of linagliptin were included in Krebs-Ringer Bicarbonate (KRB) solutions that contained either 2.5 mM glucose (KRB-low) or 25 mM glucose (KRB-high). Specifically, following overnight culture after receiving the islets from IIDP centers, the human islets plated in 48-well plates (150 IEQ/well/0.5 ml) were pre-incubated in KRB-low solution for 30 minutes at 37 °C. Then the supernatants were replaced with either KRB-low with various concentrations of linagliptin (0, 5, 30, 50 nM), or KRB-high with various linagliptin concentrations (0, 5, 30, 50 nM), n = 4 wells per treatment group. Of note, 0 nM linagliptin depicts the islets without linagliptin treatment. After the plate was incubated for 1 hour at 37 °C in humidified 5% CO 2 incubator, the supernatants were collected, and insulin concentrations measured as described above.
Immunofluorescence staining. For isolated human islets, the islets were cultured and treated with linagliptin in collagen-coated 24-well plates for 4 days as described above. The islets were then fixed with 4% paraformaldehyde in PBS for 30 minutes, permeabilized with 0.25% Triton-X for 7 minutes, and blocked in blocking solution (2% Glycine, 2% BSA, 5% FBS, 50 mM NH 4 Cl in PBS) for 1 hour. The cells were then incubated with primary antibodies and fluorescence-conjugated secondary antibodies following standard protocols. For the paraffin-embedded human pancreatic slices, which were obtained from the network for Pancreatic Organ Donors (nPOD), de-paraffinization was performed prior to antibody staining using standard protocols 27 . Quantification of GLP-1 + and Glucagon + cells. The numbers of GLP-1-expressing cells (GLP-1 + ) and glucagon-expressing cells (Glucagon + ) in each islet was performed essentially as described before 27 . Briefly, the GLP-1 + and Glucagon + cells were manually counted from 25-30 islets in each group. The ratio of GLP-1 + cells (including GLP-1 + only and GLP-1 + Glucagon + double positive) to Glucagon + cells (including Glucagon + only and GLP-1 + Glucagon + double positive) was calculated for each islet, and averaged for each treatment group.
Statistical analysis. The statistical analysis was performed using SAS software. One-way ANOVA was used to compare the differences among different treatment groups. In the cases involving only 2 groups, student's t-test was used to determine whether the difference was statistically significant. The data were expressed as Mean ± SEM (standard error of the mean). P < 0.05 was considered statistically significant.
Western blotting assay. Human islets (5000 IEQ) were lysed in 300 ul of RIPA buffer containing a cocktail of protease inhibitors and phosphatase inhibitors. After ultrasonic homogenization and a freeze/thaw cycle, the samples were centrifuged at 4 °C, 13000 rpm, for 10 minutes. The supernatants (protein extracts) were then transferred to fresh microcentrifuge tubes, boiled, and processed for SDS-PAGE and western blotting assays following standard procedures 46 .
Data availability statement. The datasets generated during and/or analyzed during the current study are available from the corresponding author on reasonable request.