Ashley Untereiner1, Jie Xu1,2, Alpana Bhattacharjee3, Over Cabrera4, Cheng Hu2,5, Feihan F. Dai1,3,†, Michael B. Wheeler1,3,†
ABSTRACT
Aims:GABA was shown to increase murine pancreatic β -cell mass and enhance β -cell proliferation in both human and mouse islets. In the present study, we sought to address GABA’s mechanism of action on β-cell proliferation, and ifLy49 (a novel GABAA-R PAM) co-treatment
further amplifies this effect.Methods: Human or mouse islets were co-treated for 4-5 days with GABA and selected receptor and cell signaling pathway modulators. Immunofluorescence was used to determine protein colocalization, cell number or proliferation, and islet size. Osmotic minipumps were surgically implanted in mice to assess Ly49 effects on pancreatic β -cells.Results: Amplification of GABAA-R signaling, but not GABAB-R, enhanced GABA-stimulated β-cell proliferation in cultured mouse islets. Co-treatment of GABA with an inhibitor specific for PI3K, mTORC1/2, or p70S6K abolished GABA-stimulated β -cell proliferation in mouse and human islets. Nuclear p-AktSer473 and p-p70S6KThr421/Ser424 expression in pancreatic β -cells were increased in GABA-treated mice compared to vehicle-treated mice; an effect further augmented with GABA+Ly49 co-treatment. Mice co-treated with GABA and Ly49 exhibited enhanced β – cell area and proliferation compared to GABA-treated mice. Furthermore, S961 injection (an insulin receptor antagonist) resulted in enhanced plasma insulin in GABA and Ly49 co-treated mice compared to GABA-treated mice. Importantly, GABA co-treated with Ly49 increased β -cell proliferation in human islets providing a potential application for human subjects.Conclusions: We demonstrate that GABA stimulates β-cell proliferation via the PI3K/mTORC1/p70S6K pathway in both mouse and human islets. Furthermore, we demonstrate that Ly49 enhances the β-cell regenerative effects of GABA, showing promising potential in the intervention of diabetes.
KEYWORDS:GABA; GABAA receptor; mTORC1; p70S6K; β-cell; proliferation
1.INTRODUCTION
The development of medications to safely promote β -cell replication would be ground-breaking for the treatment of all forms of diabetes1. It has recently become appreciated that activation of type A or B γ-aminobutyric acid receptors (GABAA/B-Rs) on pancreatic β-cells promotes cell
survival and replication2-7.GABA is endogenously produced in β-cells and interacts with and activates the ionotropic receptor GABAA (a Cl− ion channel)8- 11 and the metabotropic G-protein-coupled receptor GABAB2,12 on neighboring islet cells in an autocrine and paracrine fashion. Previously, GABA administration was shown to promote β-cell replication and inhibit β-cell apoptosis in a mouse model of streptozotocin-induced diabetes and in human islets3,4,6 . Although controversial13,14, GABA15 or artemisinins (a GABAA-R agonist)16 may promote the transdifferentiation of α-cells into β -like cells through a GABAA-R-dependent process. Nonetheless, GABA administrated in these studies led to the reversal of type 1 diabetes in mice15, rats16, and zebrafish16 We7 and others6,15 previously reported that oral GABA administration expanded murine pancreatic β -cell mass and enhanced β -cell proliferation in both human and mouse islets. GABA was shown to stimulate β-cell proliferation via activating Akt3,6 ; however, Akt signaling diverges into many pathways, making the proliferative nature of GABA unclear. In our previous study, we found that GABA stimulated β -cell proliferation via GABAA-R, as bicuculline (a GABAA-R antagonist) abrogated the proliferative effects of GABA7. However, others have reported that GABAB-R4,17 activation stimulates β-cell proliferation, thereby calling into question the identityof the main GABA-R through which GABA induces β-cell replication.
Specific positive allosteric modulators (PAMs) for GABAA-R offer increased specificity and are currently in clinical use (e.g., the benzodiazepine Valium). GABAA-R PAMs are not agonists; they enhance GABAA-R conductance and activity when GABA (or another GABAA-R agonist) is present18. Repurposing GABAA-R-PAMs to treat diabetes is theoretically appealing due to their potential to enhance the ability of GABA to promote β-cell proliferation. It is clearly of interest to understand the mechanism of GABA action and whether GABA-induced β -cell
proliferation can be enhanced via GABAA-R PAM co-treatment.Could GABA-induced β-cell proliferation be amplified, and what would be its mechanism of proliferation? Therefore, we aimed to address: 1) the direct involvement of GABAA- and GABAB-R in β-cell proliferation; 2) the mechanism of GABA action in β-cell proliferation in both human and mouse islets; and 3) whether or not co-treatment of GABA and a GABAA-R PAM can further enhance the anti-diabetic potential of GABA in vivo. In the current study, we report new mechanistic insight into the proliferative nature of GABA in both mouse and human β -cells, along with the promising potential of a novel GABAA-R PAM (i.e., Ly49, a metabolite of the benzodiazepine Oxazepam) to enhance the anti-diabetic effects of GABA in vivo.
2. MATERIALS AND METHODS
2.1 Animals and treatment regimens
Eight-week-old male FVB mice were purchased from Charles River Laboratories (Laval, Canada). To treat mice with a continuous fusion of the brain-restricted GABAA-R PAM, Ly49 [also known as LSN497448; Supplement Tables 1 & 2] (Eli Lilly and Company, Indianapolis, USA), osmotic minipumps (Alzet Model 2006; DURECT Corporation, Cupertino, USA) filled with either Ly49 (0.02 mg/kg/day) or vehicle (85% PEG/15% PBS) were implanted in a subcutaneous pocket on the back of each mouse. Following surgery, mice were administered GABA-supplemented drinking water (6 mg/mL; Sigma-Aldrich, St Louis, USA) or standard water for the duration of the 6-week treatment, and maintained on standard rodent diet, as described previously7. The University of Toronto animal care committee approved all animal
experiments and methods (Protocol# 20011236).
2.2 Mouse islet isolation, human islets, and cytospin
Mouse islets were isolated from 8-week old male FVB mice, as described previously19,20 . Human islets isolated from healthy donors were obtained from the Alberta Islet Distribution Program (University of Alberta,Canada) and the University Health Network Human Islet Isolation Program (UHN, Toronto, Canada) (Supplement Table 3). A detailed description of islet isolation and treatment can be found in the Supplementary Appendix. After treatment, intact islets were dispersed into single cells with TrypLE (Invitrogen, Burlington, Canada) treatment. Dispersed cells were loaded onto the slides using ShandonTM Single Cytofunnel(ThermoFisher, Mississauga, Canada) for immunostaining.
2.3 Immunofluorescence and confocal microscopy
The whole pancreata from each mouse were fixed, as described previously7. Sections on two different levels were stained with anti-insulin (DAKO,Carpinteria, USA), anti-glucagon (Abcam, Toronto, Canada), anti-p-AKTSer473 (Abcam), anti-p-p70S6KThr421/Ser4254 (New England Biolabs, Ltd, Whitby, Canada), or anti-Ki67 (Abcam). Section images were acquired using the Zeiss Axioscan Slide Scanner (Zeiss, Toronto, Canada).At least 50-80 islets per slide were analyzed.For confocal, dispersed single islet cells were fixed with 4% paraformaldehyde and probed with anti-insulin, anti-glucagon, anti-Ki67, or EdU (Click-iTTM EdU Cell Proliferation Kit for Imaging; ThermoFisher). Secondary anti-guinea pig Alexa488 (ThermoFisher), antimouse Alexa555 (Abcam), and anti-rabbit Cy5 (ThermoFisher) were used to detect the protein of interest. Images were obtained using the Zeiss LSM700 confocal microscopy (Zeiss) or the Zeiss Axioscan Slide Scanner.All quantifications were performed by HALO version 2.0.1145.14 (Indica Labs, Corrales,USA). A more detailed description of the methods can be found in the Supplementary Appendix.
2.4 Liquid chromatography-tandem mass spectrometry (LC-MS/MS)
LC-MS/MS methods are detailed in the Supplementary Appendix.
2.5 Induction of transient insulin resistance
Non-fasted mice were given a single i.p. injection of 30 µM/kg S961 (Novo Nordisk, Princeton, USA). Blood glucose was determined via One-touch glucometer. Blood samples were collected to assess plasma insulin level via Ultrasensitive Insulin ELISA kit (ALPCO, Salem, USA).
2.6 Statistics
Shapiro-Wilk normality test was used to determine data normality. Mann-Whitney U, unpaired student-t test, one- or two-way ANOVA were applied to determine statistical significance where applicable. P-values less than 0.05 were regarded as statistically significant
3.RESULTS
3.1 Functional distinctions between GABAA- and GABAB-R medical libraries activity in murine pancreatic β-cells
The general consensus is that GABA induces β-cell proliferation via GABAA- and GABAB- R7,21,22 . However, it has yet to be determined which receptor GABA predominantly exerts its proliferative effects; therefore, we sought to address this issue using a pharmacological approach. Cell proliferation was assessed via the cytospin-immunofluorescent method (Fig. 1A). Harmine was used as a positive control for both β and α-cell proliferation23(Fig.1B). Treatment with muscimol (a GABAA-R agonist) or Ly49 (a GABAA-R PAM) increased β -cell proliferation by 27 ± 10% or 64 ± 13%, respectively (Fig. 1C, middle graph). The effect of β -cell proliferation in both treatments was effectively inhibited by bicuculline co-treatment (a GABAA- R antagonist). No α-cell proliferation was observed upon treatment with GABA or a GABAA-R- specific agent (Fig. 1C, right graph). Treatment with baclofen (a GABAB-R agonist) induced a 47 ± 16% increase in β-cell proliferation, which was effectively blocked with saclofen (a GABAB-R antagonist) co-treatment (Fig. 1D, middle graph). Although GABA treatment alone did not induce α-cell proliferation, surprisingly, baclofen treatment significantly enhanced α-cell proliferation by 72 ± 21% (Fig. 1D, right graph); which was also effectively blocked by saclofen co-treatment (Fig. 1D, right graph).
Next, we sought to determine if the mitogenic effects of GABA on β -cell proliferation can be enhanced via increasing GABAA- or GABAB-R activity.Co-treatment of GABA with either baclofen or Ly49 significantly increased β -cell proliferation by 125 ± 12% or 192 ± 22%, respectively, compared to the control group (Fig. 1E, middle graph). Intriguingly, only the Ly49 and GABA co-treatment induced a significantly greater response in β -cell proliferation compared to either the GABA-treated or GABA and baclofen co-treated group (Fig. 1E, middle graph). Similar to the effects seen in Fig. 1D, co-treatment of baclofen with either Ly49 or GABA led to a significant induction of α-cell proliferation (73 ± 21% or 116 ± 22%, respectively; Fig. 1E, right graph). Overall, these observations suggest that 1) GABA stimulates β-cell proliferation through both GABAA- and GABAB-R, and 2) enhancing GABAA-R activity(via GABAA-R PAM) further increased GABA-mediated induction of β -cell proliferation.
3.2 GABA stimulates β-cell proliferation via the PI3K/mTORC1/p70S6K pathway in primary mouse and human islets
Two observations were considered when investigating the underlying mechanism of GABA- induced β -cell proliferation: 1) GABA stimulated β-cells proliferation via Akt pathway in mouse and human islets3,6, and 2) GABA stimulated developing neuron proliferation via mTORC1 activation24 ; whereby mTORC1 was also demonstrated to induce β-cell proliferation25-27. Therefore, we sought to determine if mTORC1, a downstream target of Akt, was crucial for GABA-mediated β -cell proliferation. Mouse or human islets were treated with various inhibitors for 4-5 days (Fig. 2A; Fig. S1). In mouse islets, GABA treatment significantly increased co- staining of insulin+Ki67+ or insulin+EdU+ by 65 ± 22% or 55 ± 14%, respectively, compared to their corresponding control groups (Fig. 2B & C, respectively). In accordance with previous observations3,6, we found wortmannin co-treatment(a PI3K [phosphoinositide 3-kinase] antagonist; an upstream target of Akt) significantly inhibited GABA-stimulated murine β -cell proliferation (Fig. 2B & C). Interestingly, inhibition of mTORC1/2 (via rapamycin) and p70S6K (a downstream target of mTORC1; via PF-4708671) significantly blocked GABA-induced β -cell replication in mouse islets; effects observed using both the Ki67 and EdU stains (Fig. 2B & C,respectively).
To determine the translational value of our mouse studies, human islets obtained from nine donors were treated with the PI3K/mTORC1/p70S6K pathway inhibitors. Firstly, we observed a significant increase in human β -cell proliferation by 58 ± 10% with GABA treatment compared to their respective control groups (Fig. 2D). Similar to mouse pancreatic β -cells, GABA-stimulated β -cell proliferation was also inhibited with co-treatment of wortmannin, rapamycin, and PF-4708671 in human islets (Fig. 2D). Importantly, Ly49 and GABA co-treated significantly amplified GABA- or Ly49-mediated β -cell proliferation in human islets (Fig. 2E). Lastly, co-treatment with rapamycin or PF-4708671 abolished Ly49-mediated induction of β -cell proliferation in Donor 1 islets (Fig. S2). These observations suggest that GABA-mediated β -cell proliferation is dependent upon a functional PI3K/mTORC1/p70S6K pathway in both human and mouse islets.
3.3 GABA or GABA plus Ly49 treatment upregulates nuclear p-Akt and p-p70S6K protein expression in murine pancreatic β-cells in vivo
As GABA-induced β -cell proliferation was abolished upon co-treatment with a PI3K or p70S6Kinhibitor, we sought to determine if GABA or GABA+Ly49 treatment stimulates p-Akt and p-p70S6K colocalization in the pancreatic β -cell nuclei in vivo. Mice were treated either vehicle (85% PEG/15% PBS) or Ly49 (0.02 mg/kg/d) and were administered standard or GABA (6 mg/mL)-supplemented drinking water for 6 weeks (Fig. 3A). As expected, a corresponding increase in plasma GABA or Ly49 were observed in the appropriate treatment groups (Fig. 3Ai & Aii). Interestingly, we found a significant 1.2- or 2.2-fold increase in nuclear p-AktSer473 expression in pancreatic β -cells in mice treated with GABA or GABA+Ly49, respectively, compared to the control mice (Fig. 3B & C). Moreover, p-AktSer473 expression in the β-cell nuclei was further enhanced by 49 ± 17% in mice co-treated with GABA and Ly49 compared to mice biomedical detection treated with only GABA. Additionally, we observed a significant 1.4- or 2-fold increase in p-p70S6KThr421/Ser424 expression in the β-cell nuclei in GABA-treated or GABA and Ly49 co- treated mice, respectively, compared to the vehicle-treated mice (Fig. 3B & D). Similar to p- AktSer473, colocalization of p-p70S6KThr421/Ser424 in the β -cell nuclei was further enhanced by 63 ± 16% in mice co-treated with GABA and Ly49 compared to mice treated with GABA. This data supports our in vitro observations that GABA activates the Akt/mTORC1/p70S6K pathway in murine pancreatic β -cells.
3.4 Co-treatment with Ly49 enhances GABA-stimulated β-cell proliferation and islet expansion and improves glucose homeostasis in transient insulin resistance mice Mice given GABA water (6 mg/mL; 6 weeks) displayed a significant increase in average islet size, along with an increase in β -cell area, number, and proliferation compared to mice given standard drinking water (Fig. S3). Next, we sought to determine if Ly49 amplifies GABA- mediated induction of β -cell proliferation in vivo (mouse treatment regimen shown in Fig. 4A). A substantial increase in averaged islet size was observed (Fig. 4B-D), along with significant increases in β -cell area and number in mice co-treated with GABA and Ly49 compared to GABA-treated mice (Fig. 4E & F). Interestingly, α-cell area, but not α-cell number, was significantly increased in mice co-treated with GABA and Ly49 compared to GABA-treated mice (Fig. 4G & H, respectively). Additionally, β-cell proliferation (via Ki67) was significantly increased in the pancreata from mice given both GABA and Ly49 compared to mice given only GABA (Fig. 4I & J). No changes were observed in α-cell proliferation (Fig. 4I & K), nor in the oral glucose tolerance test (OGTT), plasma insulin, or insulin tolerance test (ITT) (Figs. S3 & S4).
To assess β -cell function in vivo, mice were given a single injection of S961 (a specific insulin receptor antagonist; 30 µM/kg). A single injection of S961 was shown to induce transient insulin resistance by preventing insulin uptake into the peripheral tissues in rodents28. Accordingly, we observed a significant rise in plasma insulin level at 2- and 3-hour post-S961 injection in mice treated only with GABA water compared to mice given standard drinking water (Fig. 5A-C). Interestingly, co-treatment of Ly49 with GABA led to a significant improvement in glucose homeostasis along with increased plasma insulin at 2-hour post-S961 injection when compared to mice given GABA water (Fig. 5D-F). Overall, these results demonstrate the add-on capabilities of Ly49 to further enhance GABA-mediated increase in β -cell area, number, and function in vivo.
4.DISCUSSION
Repurposing GABAA-R PAMs to treat diabetes is theoretically appealing due to their safety profile and potential to enhance the ability of GABA to promote β -cell replication. The current study provides new mechanistic insight in the proliferative nature of GABA in both mouse and human β -cells. We also show in vitro and in vivo evidence of the repurposing potential of a novel GABAA-R PAM to enhance GABA-mediated β -cell proliferation.PAMs require the presence of a GABAA-R agonist to be effective. Ly49 administration in vitro induced β -cell proliferation equal to that of GABA itself (Fig. 1), suggesting Ly49 was working through endogenous GABA to enhance GABAA-R activity. Under the conditions studied, we found muscimol induced a marginal 27% increase in β -cell proliferation, whereas Ly49 treatment promoted β-cell replication as effectively as GABA (which activates both GABAA- and GABAB-Rs). This could be due to more favorable pharmacokinetics ofLy49. As a GABAA-R PAM, Ly49 likely enhanced GABA-mediated β -cell proliferation via amplifying the conductance of the channel without increasing its desensitization29,30 . In regards to the role ofGABAB-R in cell proliferation, we found that baclofen treatment increased α- and β -cell proliferation in cultured FVB islets. Interestingly, baclofen was shown to activate CXCR4 receptor31, which is also expressed in both α- and β -cells32. Although not significant, baclofen co-treated with Ly49 or GABA elicited a higher response in β -cell proliferation compared to the GABA-treated islets. These data suggest that GABA converges on both GABAA- and GABAB-R to induce β -cell proliferation, and that enhancing GABAA-R activity (via Ly49) significantly increased GABA-mediated induction of β -cell proliferation in mouse islets.
Previously, we showed that GABA treatment induced β -cell proliferation and mass, and enhanced glucose clearance in CD1 but not C57 mice7. We also found that GABA treatment stimulated α-cell proliferation in CD1 islets7. In the current study, we found GABA treatment induced β -cell but not α-cell proliferation in FVB islets (determined via Ki67 [Fig. 1] and EdU [data not shown] staining). We assume this difference could be due to a substrain effect (e.g., differential expression of GABAA-R subunits, intracellular trafficking proteins, etc.), as we
previously observed murine substrain effects to GABA treatment7.In developing neurons, GABA acts as an excitatory transmitter and induces neuron proliferation24,33-35 via mTORC1 signaling24. mTORC1 modulates the synthesis and stability of cyclin D2 and D3, which controls cell cycle progression in β-cells27 by phosphorylating p70S6K and eukaryote initiation factor 4E-binding protein36,37 . GABA was shown to induce β -cell proliferation by activating Akt/CREB3,6 ; however, Akt signaling diverges into many pathways, making the proliferative nature of GABA unclear.In the current study, we found mTORC1/p70S6K signaling played a crucial role in GABA-stimulated β -cell replication in primary mouse and human islets (Fig. 2). Through pharmacological inhibition, GABA-induced β-cell replication required a functional PI3K/mTORC1/p70S6K pathway. Prolonged treatment with rapamycin (e.g., >48 hours) was shown to inhibit not only mTORC1 but also mTORC2 in β-cells38. As such, we verify the reliance of GABA on mTORC1 activity as inhibition of its downstream target, p70S6K (via PF-4708671), prevented GABA-induced β -cell replication in mouse and human islets. Moreover, assessment of mTORC1 activity revealed a significant increase in p-p70S6KThr421/Ser424 expression in the nucleus of pancreatic β-cells in mice treated with GABA for six weeks; whereby, the colocalization of p-p70S6KThr421/Ser424 in the β -cell nuclei was further enhanced via GABA and Ly49 co-treatment (Fig. 3).
Increased phosphorylation of p70S6K in pancreatic β-cell nuclei may reflect increased mTORC1 activity due to changes in protein translation39,40 . Therefore, we suggest GABA stimulates β -cell proliferation by activating both GABAA- and GABAB-R, whereby both receptors converge on the PI3K/mTORC1/p70S6K pathway to induce β -cell proliferation (Fig. S5). However, the complete lack of mTORC2 involvement cannot be ruled out in GABA’s mechanism of action to induce β -cell proliferation, as mTORC2 was also shown to promote β-cell survival38. Future in vivo studies will determine if rapamycin treatment abrogates the islet and β-cell enhancing effects of GABA.Constant β -cell depolarization was shown to lead to excitotoxicity, dedifferentiation, and eventual reductions in β -cell mass41. GABA was demonstrated to increase intracellular Ca2+ and enhance β -cell depolarization3,42 . On the other hand, we show that GABA activates PI3K/Akt signaling, which was demonstrated to promote β -cell survival43,44 . As GABA activates the P13K/Akt/p70S6K pathway leading to β -cell proliferation, we assume PI3K/Akt activation also protects the cells from persistent depolarization-toxicity, thereby preventing apoptosis (Fig. S5).
In the current study, Ly49 was chosen to study the enhancement effects of GABA- mediated β -cell proliferation in vivo because: 1) co-treatment with GABA significantly increased β-cell proliferation; 2) although controversial13,14, GABAA-R activation may further enhance islet expansion via α-cell-to-β -cell transdifferentiation15,16 ; and 3) Ly49 treatment did not elicit α-cell proliferation. Ly49 was continuously infused (via osmotic minipump) for six weeks, thereby avoiding unnecessary animal handling stress while providing stable and continuous drug delivery. Recently, Tian et al.,21 demonstrated that daily injections of alprazolam (a GABAA-R PAM) for two weeks promoted β-cell survival and replication in human islets after implantation into non-obese diabetic/scid mice. Although promising, we are confident that our delivery method was more reliable in assessing the pharmacodynamic properties of Ly49 over a longer (e.g., 3X) treatment period. Indeed, we found that co-treatment with GABA and Ly49 significantly increased averaged read more islet size as well as β-cell area, number, and proliferation compared to mice treated with only GABA (Fig. 4). Additionally, GABA and Ly49 co-treatment significantly increased β-cell area compared to Ly49-treated mice (Fig. S7).
Interestingly, Ly49 treatment alone did not significantly increase β-cell area compared to the vehicle group, however, the data was trending towards significance (e.g., P=0.0852) (Fig. S6). Lastly, the induction of transient insulin resistance (via S961 injection) revealed elevated plasma insulin in mice treated with both GABA and Ly49 compared to GABA-treated mice (Fig. 5); further supporting our observation that GABA and Ly49 co-treatment significantly increased pancreatic β-cell area and number in mice. No changes in glucose tolerance nor insulin secretion (both via OGTT) were observed in mice co-treated with GABA and Ly49 compared to GABA-treated mice (Fig. S4). We assume the unchanged insulin state between the two treatment groups could be due to islets responding appropriately to blood glucose level and not wastefully secreting excess insulin. Taken together, we show that GABA induces β -cell proliferation via the PI3K/mTORC1/p70S6K pathway, an effect amplified by Ly49 (a novel GABAA-R PAM) co- treatment in both human and mouse islets. We also provide evidence of the translational relevance of our mouse studies to humans, whereby the majority of our donors showed a higher rate of β -cell proliferation upon Ly49 treatment compared to GABA. This paper demonstrates the promising potential of using GABAA-R PAMs for the intervention of diabetes.