Marta Szabat, James D. Johnson
Hormones, growth factors, and cytokines coordinate metabolism between tissues and within tissues. The transforming growth factor superfamily signaling pathway encompasses members of the TGFβ, activin, bone morphogenetic protein, Nodal, and growth and differentiation factor subfamilies, in addition to numerous other ligands, receptors, coreceptors, downstream signaling effectors, and regulating molecules. The elaborate multifunctional effects of this superfamily are highly cell-type and context-dependent (1, 2). Many of these ligands are involved in organ specification, patterning, proliferation, and differentiation, including important roles in the pancreas (3, 4). Studies of transgenic and knockout mice highlight the importance of this signaling pathway in islet development (5), and more recent work has also implicated the TGFβ superfamily in adult β-cell function and maturity (6, 7). In this issue of Endocrinology, Boerner and colleagues (8) report a novel role for Nodal, a secreted TGFβ superfamily member known for its roles in early embryogenesis and its mitogenic signaling through Smad2/3 and Smad4. Detailed studies on the role of Nodal in adult β-cells had not been reported. Boerner et al (8) identified Nodel in adult islets and used complementary techniques, including flow cytometry, to demonstrate that Nodal stimulates human pancreatic β-cell proliferation. This contrasts with studies employing pancreatic cell lines wherein Nodal was reported to induce apoptosis and inhibit proliferation (9, 10) and highlights the need to perform proliferation studies in nontransformed cells. This novel role of Nodal adds to the numerous roles that this family appears to have on adult pancreatic cells. Here, we will review developments in the field of TGFβ signaling in islets and emphasize that delineating clear roles of this signaling pathway has been quite challenging. Genetically engineered mouse models have provided insight, as well as confusion, on the roles of TGFβ family members (Table 1). Life-long global knockouts can provide information on gene function in the whole organism over its entire lifetime. Distinguishing effects of TGFβ family members on adult β-cell physiology from effects on development is not possible without conditional loss-of-function or gain-of-function models, but these are relatively few in number (11–13). Despite these caveats, a picture of multiple effects of TGFβ signaling on adult β-cell homeostasis has begun to emerge. A recurring phenotype manifested in adult mice carrying mutations in various TGFβ superfamily genes includes decreased β-cell mass, impaired glucose tolerance, and in some cases diabetes. Table 1 provides a glimpse of the plethora of effects of this superfamily, and its downstream signaling molecules, on pancreatic islets. Deletion of the negative downstream effector of TGFβ signaling Smad7 (14) in Pdx1-expressing cells decreased hormone-positive cells in late development (5). On the other hand, conditional overexpression of Smad7 in adult Pdx1-expressing cells altered the β-cell gene signature and reduced pancreatic insulin production and release, leading to diabetes (11). Furthermore, Smad7 appears to be crucial for beta-cell proliferation after partial pancreatectomy (50). Thus, the same TGFβ effector can have opposing roles depending on the temporal context. Moreover, the TGFβ signaling effectors can have opposing roles, given that Smad3 appears to be a negative regulator of insulin secretion and glucose tolerance (13). Taken together, many lines of evidence clearly suggest a role of this signaling pathway in regulating β-cell mass and function. More inducible knockout mouse studies are needed to clarify the specific cell targets and mechanisms of actions of this family on pancreatic cells. Pancreatic Actions of the TGFβ Superfamily Abbreviations: ALK, activin receptor-like kinase, ActR, activin receptor; BMP, bone morphogenetic protein; GDF, growth and differentiation factor. Pancreatic Actions of the TGFβ Superfamily Abbreviations: ALK, activin receptor-like kinase, ActR, activin receptor; BMP, bone morphogenetic protein; GDF, growth and differentiation factor. The effects of the TGFβ superfamily ligands on adult pancreatic β-cells can be directly tested in vitro, bypassing the complexities of the in vivo environment. For example, acute activin A treatment increased glucose-stimulated insulin secretion from human (15) and rat (16) islets, whereas prolonged treatment decreased glucose-stimulated insulin secretion from mouse islets, an effect reversed by the endogenous antagonist follistatin (17). β-Cell proliferation was increased by activin A treatment in adult rat islets (18) and a mouse β-cell line (17). Interestingly, there appears to be a species difference in the islet expression and action of activin A, follistatin, and TGFβ1 (19). Although TGFβ1 acutely stimulates glucose-stimulated insulin release from rat islets (19), prolonged treatment of human islet cells with TGFβ1 had a negative impact on the expression of genes important for β-cell function, likely via the Smad3 pathway (13). Interestingly, Boerner et al (8) did not report evidence of β-cell dedifferentiation, although β-cell function remains to be rigorously investigated in this model. The range of (and sometimes opposing) effects of the same factors observed on pancreatic cells illustrates the complexity and context-dependence of this signaling pathway. Although studies using animals with genetic manipulations are crucial, results from human islets are vital to uncover any species differences and for future clinical translation. One of the most important contributions of the Boerner study was the use of islets from human donors (8). Studies with human islets, or at least >1-year-old mice (which are equally refractory to β-cell proliferation), are crucial for clinical translation of the limited basic knowledge around primary β-cell proliferation mechanisms (20). The ultimate goal of this research is to provide the basis for a diabetes cure. Diabetes results from the immune-mediated destruction (type 1 diabetes) or functional failure (type 2 diabetes) of pancreatic insulin-producing β-cells (21). There are theoretically numerous approaches to treating or reversing diabetes using cell-based therapy, including in vivo regeneration via induction of replication of remaining β-cells, transdifferentiation of other adult cell types (in vivo and in vitro) and differentiation of embryonic (or induced) pluripotent stem cells into functional β-cells (22). Significant preclinical obstacles remain for each of these approaches. At this point, islet transplantation from cadaveric donors remains the only viable alternative to insulin therapy for treating type 1 diabetes (23). Nevertheless, the lack of islet donor tissue prevents widespread patient access to this therapy. Hence, the expansion of primary human islets is of great interest to the diabetes research community. Various studies have shown efficient induction of young rodent β-cell proliferation, yet the same approaches have either not been tested or failed to produce significant results on adult human β-cells (20, 24). Failure to induce proliferation in human islets could be in part a species difference and the fact that the human islet donor age is generally much higher in relative comparison with the young adolescent mice being used for studies (25). As has been elegantly shown, β-cell proliferation in mice and humans ceases to nearly zero after 1 and 30 years of age, respectively (26, 27). Hence, any study showing that a ligand can overcome this resistance of adult human β-cells to reenter the cell cycle, and stimulate proliferation even modestly, is an important contribution to the field. Boerner et al (8) found an ∼1.5-fold increase in human β-cell proliferation with Nodal treatment without affecting viability. Because it is estimated that only 1% of type 1 diabetes patients can be treated with the current scarce availability of donor islets (28) and the fact that a single patient requires at least 103 islet equivalents (from ∼2 pancreas donors) per kilogram of body weight (29), the expansion of islets from a single human donor pancreas would require at least an ∼200-fold increase for a widespread application of human islet transplantation. Therefore, more extensive research in human β-cell expansion is needed. Boerner et al (8) echoed statements by many in the field when they proposed in their paper that the ultimate therapeutic approach may require simultaneously harnessing multiple local signaling pathways. Indeed, genomic advances have revealed that islets produce hundreds of soluble ligands and express hundreds of receptors for ligands produced locally and in other tissues (30). It is clear that using high-throughput approaches will be required to make breakthroughs in this area. Pancreatic β-cell proliferation is exceedingly hard to measure in samples from humans and older adult rodents, given the rarity of events and the potential for false-positives with semiautomated counting. This means that there must be exceptional rigor. Bulk measurements of DNA synthesis are clearly unacceptable in heterogeneous human islet cell cultures where β-cells usually make up a minority of endocrine cells, to mention nothing of the fibroblast-like cells that dominate these cultures after a few days. Imaging must be of sufficient X, Y, and Z resolution such that highly proliferative fibroblasts, which overlay virtually all human islet cultures, are not mistaken for proliferating β-cells. In this regard, the use of flow cytometry by Boerner et al (8) to detect human β-cell proliferation is an important technical proof-of-concept. Clearly, the bar for measuring β-cell proliferation needs to continue to get higher to prevent costly false-positives in the field. Even considering the technical limitations, only a handful of studies have shown significant in vitro stimulation of human β-cell proliferation. These include cocktails of 1) hepatocyte growth factor, fibroblast growth factor 4, nicotinamide and fibrin gel (∼2-fold) (31), the L-type calcium channel agonist Bay K8644 (∼1.5-fold) (32), a diarylamide compound WS6 (∼5-fold) (33), and glycogen synthase kinase-3 inhibitors (∼2-fold) (34). It remains unclear whether any ligand alone will be able to safely increase human β-cell proliferation to therapeutic levels in vivo. It is not clear at this point whether TGFβ superfamily ligands are useful therapeutic targets for manipulating β-cell mass or function in vivo, because they may have off-target systemic effects. Most of the TGFβ family ligands share the same receptors, downstream effectors and antagonists (1) and are expressed ubiquitously in various tissues, including pancreatic islets (8, 12, 17, 19, 30, 35). Thus, the local concentration and length of exposure of the ligands as well as the combination of receptors, effectors and antagonists locally expressed will ultimately provide the final concerted effect on the β-cell, both in vivo and in vitro. Nodal and other TGFβ family members may be more promising factors for in vitro differentiation, expansion, and maturation of pancreatic β-cells. However, for human in vitro β-cell expansion to become clinically relevant, much more potent mitogens and/or combinations of factors working synergistically to retain β-cell maturity, viability, and function need to be evaluated. It will take the collaborative efforts between disciplines from basic β-cell biologists to physicians and engineers (36) to accomplish such an endeavor. Disclosure Summary: The authors have nothing to disclose.