OR14-01 FSTL3 Neutralizing Antibodies Restore Function to Diabetic Mouse and Human Islets: A New Approach for Treating Diabetes

Activin, GDF11 and myostatin are structurally related members of the TGFbeta superfamily of growth factors with many biological roles in animal models and humans. Their actions are neutralized by extracellular proteins such as follistatin and follistatin like-3 (FSTL3). We have previously demonstrated that genetic inactivation of Fstl3 results in enlarged pancreatic islets containing increased numbers of beta cells that produce more insulin in response to glucose compared to wild type litter mates. We further discovered that at least some of these new beta cells arise via transdifferentiation from alpha cells. We also demonstrated that functional human islets from normal donors produce very high levels of activin. In contrast, activin biosynthesis is vastly reduced and FSTL3 synthesis is significantly increased in human islets from diabetic donors suggesting that activin is critical for normal insulin production. This was substantiated by direct treatment of human diabetic islets with activin which restored their response to glucose. These observations support the hypothesis that an FSTL3 neutralizing antibody would constitute a novel therapeutic approach to curing diabetes through restoring beta cell function as well as accelerating generation of new beta cells through transdifferentiation. To test this hypothesis, we produced a mouse monoclonal antibody that neutralized hFSTL3 (FP-101), thereby releasing bioactive activin, GDF11, and myostatin. We have now tested this antibody for biological activity in vitro on mouse and human islets. We used islets from high fat diet (HFD) treated mice to model diabetes-inducing effects of obesity as well as 24-hour incubation in hyperglycemic (33 mM glucose) medium to create human islets that lose responsiveness to high glucose as a model for human diabetes. In mouse islets we found that stimulation of normal (chow diet) islets by high glucose produced a stimulation index (SI) of 3.5 that was reduced to 2 in HFD islets. Treatment with activin, FP-101, or a commercial polyclonal antibody to mFSTL3 all increased response of HFD islets to elevated glucose and partially restored SI to normal levels. In human islets, hyperglycemia eliminated the normal (2.5 SI) response to high glucose while activin or FP-101 treatments dose-responsively restored this response. These results demonstrate that anti-FSTL3 therapy can restore function to compromised beta cells from mouse and human diabetes models. The observation that activin has the same action as anti-FSTL3 antibody indicates that FP-101 works through enhancing the activin signaling pathway. Finally, these results demonstrate that the FSTL3-activin pathway is an important regulator of beta cell function in humans as well as mice, supporting further development of this therapy as a diabetes treatment.