476 Diabetes Volume 68, March 2019

Targeting the Brain to Cure Type 2 Diabetes

Bernard Thorens

Diabetes 2019;68:476–478 | https://doi.org/10.2337/dbi18-0051

Present pharmacological treatments of type 2 diabetes aim induce a short-term normalization of glycemia in diabetic mainly at increasing insulin secretion by pancreatic b-cells, rats and mice; this effect is probably through activation of at improving insulin action on liver, fat, or muscle, and, arcuate nucleus neurons and the regulation of insulin and more recently, at favoring renal glucose elimination. Al- counterregulatory hormones secretion (4–8). though efficacious, these treatments do not have the rapid In their previous study, Scarlett et al. (2) showed that and long-term efficacy of bariatric surgery on correcting i.c.v. FGF1 normalized glycemia in ob/ob, db/db, and high- diabetic hyperglycemia, suggesting that alternate thera- fat diet–fed mice by a mechanism that could not be peutic targets can still be identified. Neuronal circuits in ascribed to increased insulin sensitivity, insulin secretion, the brain, in particular in the hypothalamus and brain or insulin-independent glucose disposal, referred to as stem, can exert strong control on glucose homeostasis. The glucose effectiveness. In the current study (1), the basis activity of these circuits is controlled by a variety of cell for the hypoglycemic effect of i.c.v. FGF1 was investigated types—neurons, glial cells, tanycytes—that are sensitive in Zucker diabetic fatty (ZDF) rats. Diabetes development to interoceptive signals such as glucose and various hor- in these rats follows a well-described pattern, with an mones, including insulin, leptin, or ghrelin. When acti- increased in insulinemia preceding the appearance of vated, these neuronal circuits can regulate, by controlling hyperglycemia. The onset of diabetes is triggered by a fail- autonomic nervous system activity, the secretion of insu- ure of the b-cells to further increase insulin secretion and lin and glucagon by pancreatic endocrine cells as well as by glucolipotoxicity-induced apoptosis, which reduces glucose metabolic pathways in liver, fat, and muscle. b-cell mass (9,10). Here, Scarlett et al. (1) showed that Targeting these neuronal circuits could lead to innovative a single i.c.v. injection of FGF1 in ZDF rats reversibly

COMMENTARY therapies for diabetes. In the report by Scarlett et al. (1) in normalized their glycemia for approximately 30 days. In this issue of Diabetes and in a previous publication (2), the addition, when injected in prediabetic rats, development of Schwartz group demonstrated that a single intracerebro- hyperglycemia was delayed by 3–4 weeks, an effect asso- ventricular (i.c.v.) injection of fibroblast growth factor ciated with a transient increase in b-cell mass at 3 weeks 1 (FGF1) in diabetic mice and rats leads to a long-term posttreatment, which was nevertheless reduced within (up to 18 weeks) correction of diabetic hyperglycemia. 4 weeks to the low level found in vehicle-treated rats. FGF1 belongs to the fibroblast growth factor (FGF) Analysis of the components of glucoregulation showed, family, which contains 18 members (3). Most of the similarly to the previous mouse study, no difference in FGFs act as paracrine regulators because they bind to insulin secretion, insulin sensitivity, or glucose effective- heparan sulfate proteoglycans present in the extracellular ness. In contrast, the authors found an increase in liver Gck space, preventing their diffusion beyond the local envi- mRNA expression, Gck activity, and increased index of ronment where they are secreted. The exceptions to this glycolytic flux as determined by increased plasma lactate rule are FGF15 (or the human ortholog FGF19) and following glucose injection. They conclude (1) that the FGF21, which lack the heparan sulfate proteoglycan bind- main effect of i.c.v. FGF1 is to increase hepatic glucose ing site and, thus, act as classic endocrine hormones. FGF1 uptake, probably through a sympathetic regulation. binds to all FGF receptors (FGFRs) (FGFR1 to FGFR4) and The results of this and the preceding study (1,2) are their splice variants; high-affinity binding for FGFs quite spectacular, further supporting the central nervous requires receptor dimerization and association with hep- system as an important target for diabetes treatment. Of aran sulfates (3). In contrast, FGF15/19 and FGF21 bind to importance, i.c.v. FGF1 does not induce hypoglycemia in FGFRs in obligatory association with the coreceptor diabetic animals and has no impact on glycemia in regular b-klotho. Intracerebroventricular injections of FGF19 chow–fed mice. In addition, the mechanisms involved do

Center for Integrative Genomics, University of Lausanne, Lausanne, Switzerland © 2019 by the American Diabetes Association. Readers may use this article as fi Corresponding author: Bernard Thorens, [email protected] long as the work is properly cited, the use is educational and not for pro t, and the work is not altered. More information is available at http://www.diabetesjournals .org/content/license. See accompanying article, p. 654. diabetes.diabetesjournals.org Thorens 477 not involve changes in food intake or body weight. Then (12,13). The anatomical organization of the tanycytes– what are the mechanisms responsible for these long-term arcuate neurons barrier and the fact that FGF1 is a poorly effects? A number of specific questions could be asked: diffusible protein suggests that its sustained action may be what is the primary site of action of FGF1 in the brain? due to its retention in the immediate vicinity of tanycytes, How is the signal generated and why does it last so long? thereby providing constant stimulation of FGFRs. Alter- How is it transmitted to the periphery and what is the natively, or in addition, FGF1 may induce the reorganiza- primary peripheral target? Why is it observed only in tion of neuronal circuits or tanycyte-neuron connections diabetic animals and does not induce hypoglycemia? that may have been hampered by the diabetic milieu. The When FGF1 is injected i.c.v., a strong induction of c-Fos stability of these repaired structures could then explain the immunostaining is observed in tanycytes lining the third long-lasting effect of FGF1 and the fact that it has no ventricle (2). These astroglial cells, which line the baso- hypoglycemic effect in regular chow-fed mice. Using FGF1 lateral part of the third ventricle, are exposed to the variants that lack the heparan sulfate proteoglycan bind- cerebrospinal fluid and form intricate contacts with neu- ing site may help resolve these questions. rons from the dorsomedial, ventromedial, or arcuate nuclei Scarlett et al. (1) speculate that the action of i.c.v. FGF1 of the hypothalamus; some also extend projections to the on hepatic glucose uptake could be directed by an increase blood capillaries of the median eminence (11) (Fig. 1). in sympathetic activity. However, the firing activities of Their role is not fully understood, but they influence the the sympathetic or vagal nerves have not been directly activity of the arcuate neurons that control feeding and recorded. It would be important to determine which glucose homeostasis, possibly in response to variations in branch of the autonomic nervous system is activated. cerebrospinal fluid and/or blood nutrients or hormones Indeed, instead of invoking an increase in sympathetic

Figure 1—Proposed model for the control of hyperglycemia by i.c.v. FGF1. Intracerebroventricular injection of FGF1 activates tanycytes lining the basolateral region of the third ventricle (3V). At least four tanycyte populations have been described, a1, a2, b1, and b2, located as shown in the figure. These cells send projections and form intricate contacts with neurons in the dorsomedial (DMH), ventromedial (VMH), and arcuate (AN) nuclei of the hypothalamus, and the b tanycytes also send projections to the median eminence (ME) where they contact blood vessels. Injection of FGF1 is proposed by Scarlett et al. (1) to activate the sympathetic innervation of the liver, thereby increasing glucokinase (Gck) expression and activity. This leads to increased glucose uptake and glycolysis, as revealed by increased production of lactate and a reduction of glycemia. Interesting questions involve: which tanycytes are activated by FGF1? Which hypothalamic neurons are regulated by the activated tanycytes? Does this activation lead to stable change in tanycytes–hypothalamic neurons organization to sustain the long-term effect of the treatment? Is the signal transmitted dependent only on sympathetic activity and hepatic glucose uptake or is activation of parasympathetic nerves also involved to, for instance, protect b-cell mass and insulin secretion capacity? 478 Commentary Diabetes Volume 68, March 2019 activity in the control of hepatic glucose uptake, an acti- Pharmaceutical Industries and Associations, and the Swiss State Secretariat vation of the vagal nerve could explain the transient for Education, Research and Innovation under contract number 16.0097. increase in b-cell mass and insulinemia in the prediabetic Duality of Interest. No potential conflicts of interest relevant to this article ZDF rats. Indeed, increasing vagal activity, for instance by were reported. lesion of the ventromedial hypothalamus, strongly in- References b creases -cell proliferation (14), and the observed hyper- 1. Scarlett JM, Muta K, Brown JM, et al. Peripheral mechanisms mediating the insulinemia can also explain the increased expression of sustained antidiabetic action of FGF1 in the brain. Diabetes 2019;68:654–664 hepatic Gck mRNA and Gck activity. Thus, whether glyce- 2. Scarlett JM, Rojas JM, Matsen ME, et al. Central injection of fibroblast growth mia normalization is primarily due to a b-cell protective factor 1 induces sustained remission of diabetic hyperglycemia in rodents. Nat effect or to an increase in hepatic glucose uptake is difficult Med 2016;22:800–806 to determine based on the presented data. More generally, 3. Beenken A, Mohammadi M. The FGF family: biology, pathophysiology and because rats are studied several weeks after FGF1 treat- therapy. Nat Rev Drug Discov 2009;8:235–253 ment, distinguishing between the primary events that 4. Ge H, Baribault H, Vonderfecht S, et al. Characterization of a FGF19 variant fi trigger glycemia normalization versus secondary events with altered receptor speci city revealed a central role for FGFR1c in the regulation is not straightforward. of glucose metabolism. PLoS One 2012;7:e33603 5. Marcelin G, Jo YH, Li X, et al. Central action of FGF19 reduces hypothalamic Nevertheless, this study is a new illustration of the AGRP/NPY neuron activity and improves glucose metabolism. Mol Metab 2013;3:19–28 importance of the central nervous system in glucoregula- 6. Morton GJ, Matsen ME, Bracy DP, et al. FGF19 action in the brain induces tion. It also indicates that understanding the pathogenesis insulin-independent glucose lowering. J Clin Invest 2013;123:4799–4808 of type 2 diabetes requires understanding precisely what 7. Perry RJ, Lee S, Ma L, Zhang D, Schlessinger J, Shulman GI. FGF1 and FGF19 these brain control mechanisms are, how they become reverse diabetes by suppression of the hypothalamic-pituitary-adrenal axis. Nat dysfunctional, and whether they precede and cause deregu- Commun 2015;6:6980 lated insulin secretion and insulin action. Also, earlier 8. Ryan KK, Kohli R, Gutierrez-Aguilar R, Gaitonde SG, Woods SC, Seeley RJ. studies showed that peripheral injection of FGF1 normal- Fibroblast growth factor-19 action in the brain reduces food intake and body weight izes glycemia in diabetic mice (15), but the effect was and improves glucose tolerance in male rats. Endocrinology 2013;154:9–15 short-lived and dependent on the adipocyte FGFR1. Thus, 9. Unger RH, Orci L. Diseases of liporegulation: new perspective on obesity and – the beneficial effect of FGF1 as presented in this study related disorders. FASEB J 2001;15:312 321 10. Topp BG, Atkinson LL, Finegood DT. Dynamics of insulin sensitivity, b-cell depends on its delivery to the central nervous system. This function, and b-cell mass during the development of diabetes in fa/fa rats. Am J still poses a major challenge because of the need for these Physiol Endocrinol Metab 2007;293:E1730–E1735 drugs to cross the blood-brain barrier. Targeting the basal 11. Langlet F. Tanycytes: a gateway to the metabolic hypothalamus. J Neu- hypothalamus may be more feasible because of the leakiness roendocrinol 2014;26:753–760 of the blood-brain barrier, but if delivery needs to specif- 12. Langlet F, Levin BE, Luquet S, et al. Tanycytic VEGF-A boosts blood-hy- ically target tanycytes, this may represent an additional pothalamus barrier plasticity and access of metabolic signals to the arcuate difficulty. nucleus in response to fasting. Cell Metab 2013;17:607–617 13. Coppola A, Liu ZW, Andrews ZB, et al. A central thermogenic-like mechanism in feeding regulation: an interplay between arcuate nucleus T3 and UCP2. Cell Funding. The work in B.T.’s laboratory is supported by a grant from the Swiss Metab 2007;5:21–33 National Science Foundation (310030B_163458), a European Research Council 14. Kiba T, Tanaka K, Numata K, Hoshino M, Misugi K, Inoue S. Ventromedial Advanced Grant (694798-INTEGRATE), and a grant from the Innovative Medicines hypothalamic lesion-induced vagal hyperactivity stimulates rat pancreatic cell Initiative 2 Joint Undertaking under grant agreement no. 115881 (RHAPSODY). proliferation. Gastroenterology 1996;110:885–893 This Joint Undertaking receives support from the European Union’s Horizon 15. Suh JM, Jonker JW, Ahmadian M, et al. Endocrinization of FGF1 produces 2020 research and innovation program, the European Federation of a neomorphic and potent insulin sensitizer. Nature 2014;513:436–439