Neurturin and a GLP-1 Analogue Act Synergistically to Alleviate Diabetes in Zucker Diabetic Fatty Rats

Type 2 diabetes results from several pathophysiologies leading to impaired glucose homeostasis (1). Most diabetes therapies do not target this underlying cause of diabetes and eventually lose their effectiveness (2–4). Thus, developing combination therapies of agents with distinct mechanisms of action may represent a superior therapeutic strategy.

Efforts are under way to identify new treatments for diabetes (5,6), and strategies exploiting neuronal communication among brain, gut, liver, and pancreas in regulating metabolic homeostasis are being developed (7–10). Several groups have examined the role of neurotrophic factors like nerve growth factor, brain-derived neurotrophic factor, and glial-derived neurotrophic factor (GDNF) that regulate growth, survival, and differentiation of mature and developing neurons in regulating diabetes and its complications (11–14). GDNF also regulates β-cell differentiation and proliferation, enhances β-cell mass, improves glucose tolerance, and restricts development of obesity (12–14).

In this study, we describe identification of the GDNF family member neurturin (NRTN) as a secreted factor expressed in embryonic pancreas. NRTN and GDNF bind to distinct receptors, GDNF family receptor (GFR) a2 and GFRa1, respectively, and then recruit the RET coreceptor to mediate their action (15–17). NRTN and GFRa2 are critical for the development of enteric and parasympathetic innervation of several target tissues (15). In this study, we demonstrate a novel function for NRTN in preventing development of diabetes. NRTN and the glucagon-like peptide 1 (GLP-1) receptor agonist (GLP-1RA) liraglutide use distinct mechanisms to improve diabetes, and the combination of NRTN and liraglutide better restored glucose homeostasis and metabolic parameters in hyperglycemic rats than either molecule alone.

Animal studies were approved by the Institutional Animal Care and Use Committee at MedImmune (Gaithersburg, MD), or Gubra (Horsholm, Denmark) under licenses issued by the Danish Committee for Animal Research. All studies used male Zucker diabetic fatty (ZDF) rats on Purina 5008 chow (LabDiet, St. Louis, MO) or Zucker lean (ZL; fa/?) control rats (Charles River Laboratories, Wilmington, MA) and were administered vehicle (15% captisol and 10 mmol/L sodium phosphate [pH 7.4]), recombinant human NRTN, or liraglutide (0.4 mg/kg; Novo Nordisk, Copenhagen, Denmark) daily subcutaneously (1 or 2 mL/kg as indicated).

Male ZDF rats (7 to 8 weeks), were subjected to an overnight fast for analysis of blood glucose (Breeze2 glucometer; Bayer, Pittsburgh, PA), insulin (MesoScale Discovery, Rockville, MD), and percentage HbA_1c (%HbA_1c; Cobas c-111; Roche Diagnostics, Indianapolis, IN) and allocated to treatment groups: vehicle (n = 15), liraglutide (n = 10), or NRTN (1, 3, or 10 mg/kg; n = 15). Doses of NRTN were based on pharmacokinetic data demonstrating that 3 mg/kg achieved 24-h exposure of 12 nmol/L, projected to be the minimal efficacious exposure (data not shown). Body weight, food weight, and nonfasted glucose levels were measured three times per week, and drug was dosed for 30 days. Acid-ethanol extraction of the pancreas followed by insulin measurement determined pancreatic insulin content. Serum concentrations of NRTN were determined using ELISA (Eurofins, Abington, U.K.).

Male ZDF rats (7 weeks; n = 6/group) were administered a single dose of vehicle, NRTN (3 mg/kg), or liraglutide. Rats were euthanized 4 h later and c-Fos in brain regions assessed as described (18).

Male ZDF rats (n = 16/group; 7 to 8 weeks) were administered vehicle, liraglutide, or NRTN (3 mg/kg); ZL rats (n = 14) served as controls. Body weight and food intake were measured daily. Blood was collected in the nonfasted state for glucose and insulin measurements (AlphaLisa; PerkinElmer, Waltham, MA). An oral glucose tolerance test (OGTT; 1 g/kg) was performed on days 3 and 16 and an insulin tolerance test (ITT; Humulin R U-100, 1 unit/kg; Eli Lilly and Company, Indianapolis, IN) on day 20, following a 4-h fast. In a second study, male ZDF rats (n = 10/group; ∼8 weeks) were administered NRTN (3 or 10 mg/kg, s.c.), rosiglitazone (3 mg/kg, by mouth, in 0.5% hydroxypropyl methylcellulose in 0.1% Tween; Sigma-Aldrich, St. Louis, MO), or vehicle (both subcutaneously and by mouth) once daily. On days 6 and 17, an ITT was performed (0.75 units/kg Humulin R U-100) after a 4-h fast. Body weight and nonfasted glucose (Breeze2; Bayer) was measured three times per week.

Male diabetic ZDF rats (n = 10/group; ∼11 weeks) were dosed for 27 days once daily with two simultaneous injections (n = 10/group): vehicle/vehicle, liraglutide/vehicle, NRTN (10 mg/kg)/vehicle, or liraglutide/NRTN (10 mg/kg); ZL rats (n = 8) were used as controls. Body weight and food intake were measured daily. Nonfasted blood glucose was measured as indicated. An OGTT (2 g/kg) was performed on day 16 after overnight fast. On day 27, blood was collected after a 4-h fast and rats euthanized on day 28 in nonfasted conditions. Liver lipid content from one lobe was assessed (EchoMRI, Houston, TX).

Human NRTN protein was synthesized by cloning full-length coding sequence in a bacterial expression system. Lysates from large-scale cultures were purified using column chromatography to obtain ≥98% pure NRTN. NRTN in 15% captisol-phosphate buffer (pH 7.4) was used for both in vivo and in vitro studies. Recombinant human GDNF was purchased from PeproTech (Rocky Hill, NJ).

Rat insulinoma INS1 832/3 cells (provided by Christopher Newgard, Duke University, Durham, NC) were as described (19). Human neuroblastoma cell line TGW (Japanese Collection of Research Bioresources, Osaka, Japan) was cultured as described (20). EndoC cells were acquired from Dr. Raphael Scharfmann (INSERM, Paris, France) and used under a license agreement with AstraZeneca.

Pancreata were bisected longitudinally, weighed, fixed in 10% neutral buffered formalin for 24 h, and paraffin embedded. β-Cell mass was assessed using guinea pig anti-insulin, rabbit antiglucagon, anti-rabbit horseradish peroxidase, anti-mouse horseradish peroxidase, ChromoMap DAB, and DS Discovery Purple (Ventana Medical Systems, Inc., Tucson, AZ). Sections were scanned and analyzed using Definiens Developer XD64 (Definiens AG, Munich, Germany) to measure relative β-cell area.

TGW and INS1 cells were serum-starved for 3 h and then treated with indicated stimuli for 15 min. Cells were lysed, and phosphorylated Akt (p-Akt; Ser⁴⁷³) and total Akt were determined using ELISA kits (MesoScale Discovery) to calculate p-Akt/tAkt ratio.

INS1 cells washed in basal-KRB medium (2.7 mmol/L glucose) for 1 h and incubated with indicated glucose, NRTN, and GLP-1 concentrations for an additional hour. NRTN was maintained during washing and treatment in the overnight group. Pancreatic islets were isolated from 2-month-old adult male Wistar rats (Charles River Laboratories) by collagenase digestion. Twenty-five islets were added to 200 µL Krebs-Ringer bicarbonate HEPES buffer (KRBH; 20 mmol/L HEPES and 0.1% BSA [pH 7.4]) at 2.8 mmol/L glucose and allowed to equilibrate for 2 h at 37°C and 5% CO₂ with KRBH media replaced after 1 h. Islets were incubated for 1 h at 37°C and 5% CO₂, tap spun and the media collected. Then, 200 µL KRBH at 16.8 mmol/L glucose was added alone, with 10 µg/mL NRTN, or with 100 nmol/L GLP-1 and incubated for 1 h at 37°C and 5% CO₂. Media was removed and insulin measured (MesoScale Discovery).

In vivo data were analyzed by either one- or two-way repeated-measures ANOVA with Tukey post hoc test or one-way Kruskal-Wallis nonparametric test with Dunn’s post hoc assessment when appropriate. Graphs and analyses were generated using Prism 6.0 (GraphPad Software Inc., La Jolla, CA). For in vitro studies, data were analyzed using the Student t test. For β-cell mass, ANOVA followed by Tukey-Kramer pairwise testing was performed on log10-transformed data. Multiplicity-adjusted two-sided P values are reported. A P value <0.05 was considered significant.

A chicken embryonic pancreatic cDNA library was examined to identify factors that may improve β-cell function in diabetes. Clones that gave positive in situ hybridization signals on embryonic pancreatic sections were sequenced to determine if they represented secreted factors. NRTN was identified as one such secreted factor, and its expression in the embryonic pancreas was reconfirmed by in situ hybridization (Supplementary Fig. 1). NRTN and its receptor GFRa2 are expressed in the pancreas during embryonic development and postnatally along the neuronal network (21–23). Furthermore, mice lacking the Gfra2 gene showed a deficit in parasympathetic innervation of pancreatic islets and impaired vagal tone, but retained normal homeostatic responses to exogenous glucose (22). This suggested a role for NRTN signaling in pancreatic development and innervation, but also indicated that such effects of NRTN are distinct from a putative functional role in regulating glucose homeostasis. Given the role of GDNF in regulating pancreatic function and glucose tolerance, we explored the potential of NRTN to elicit similar pharmacological effects.

The potential of NRTN in preventing development of diabetes was evaluated in a ZDF rat model (24,25). Eight-week-old male ZDF rats received daily subcutaneous injections of vehicle, NRTN, or liraglutide (as positive control) for 4 weeks. Liraglutide reduced both body weight and food intake relative to vehicle-treated rats, whereas NRTN had no effect on body weight or food intake (Fig. 1A and B). Nonfasted glucose levels of vehicle-treated rats increased from 174 ± 17 to 387 ± 38 mg/dL by the end of the study (Fig. 1C). Liraglutide prevented the increase in glycemia (146 ± 18 mg/dL). NRTN had a dose-dependent effect on the development of hyperglycemia, with terminal glucose values ranging from 321 ± 44 to 228 ± 29 mg/dL for 1- and 3-mg/kg NRTN dose, respectively, whereas 10 mg/kg prevented the development of hyperglycemia (144 ± 19 mg/dL) (Fig. 1C).

At the end of the study, plasma from overnight-fasted animals was analyzed. NRTN (10 mg/kg) and liraglutide significantly reduced fasting glucose relative to controls (Fig. 1D). High-dose NRTN significantly reduced HOMA of insulin resistance and tended to reduce fasting insulin compared with liraglutide (Fig. 1E and F). There was a significant reduction in the %HbA_1c levels in liraglutide and 3 and 10 mg/kg NRTN-treated animals (Fig. 1G) that was significant when assessed as a change in %HbA_1c levels from baseline (P < 0.05) (Fig. 1H). Total plasma triglyceride levels trended lower in the high-dose NRTN and liraglutide groups, and plasma cholesterol levels were significantly lower upon liraglutide and mid- and high-dose NRTN administration (Fig. 1I and J). Terminal plasma concentrations of NRTN were 7.2 ± 0.4, 17.9 ± 1.0, or 47.3 ± 2.8 ng/mL for the 1-, 3-, or 10-mg/kg groups, respectively. These observations suggest that high-dose NRTN could prevent development of diabetes and deterioration of metabolic phenotype as effectively as liraglutide.

Insulin and glucagon costained pancreatic sections from vehicle, liraglutide, and high-dose NRTN groups were examined to determine changes in pancreatic islet morphology (Fig. 2). As expected, ∼75% of pancreatic islets in the vehicle group, although enlarged, appeared to lose some architectural integrity. Although pancreata from both liraglutide- and NRTN-treated rats also contained some larger islets, the proportion of islets with normal architecture was higher (Fig. 2). Quantification of the insulin⁺ area showed a nearly twofold increase in β-cell mass in the NRTN group and a trend toward an increase in β-cell mass in the liraglutide group (Fig. 2F). Pancreatic insulin content showed an increasing trend in all treatment groups and reached significance in ZDF rats receiving liraglutide and high-dose NRTN (Fig. 2E). Pancreatic α-cell mass was not altered, and analysis of proliferating insulin⁺ cells (Ki67⁺/insulin⁺) relative to overall number of insulin⁺ cells was unchanged (Supplementary Fig. 2). These observations demonstrate that high-dose NRTN and liraglutide treatments were effective in maintaining islet organization and pancreatic insulin content.

Whether NRTN-dependent improvements in islet phenotype are mediated by direct action of NRTN on β-cells was investigated. The expression of Gfra2 and Ret in the TGW neuroblastoma cell line (20) was much higher relative to expression of these genes in key metabolic tissues and β-cell lines (Supplementary Fig. 3). The ability of NRTN and GDNF to activate Akt in TGW and INS1 cells (Fig. 3A and B) and the potential for cross-talk between NRTN and insulin signaling were examined. Insulin increased Akt phosphorylation in a concentration-dependent manner in TGW and INS1 cells. Treatment with NRTN or GDNF also enhanced Akt phosphorylation in TGW cells, which was further enhanced by insulin (Fig. 3A). However, in INS1 cells, there was no effect of NRTN or GDNF treatment on Akt phosphorylation or potentiation of insulin signaling (Fig. 3B). NRTN was examined as an insulin secretagogue. In both INS1 cells (Fig. 3C) and isolated rat islets (Fig. 3D), incubation with NRTN concomitant with high glucose or with GLP-1 in INS1 cells did not affect an insulin secretion response. The expression of Gfra2 in rat islets was relatively low (Supplementary Fig. 3). Furthermore, GFRa2 and RET protein was predominant in parasympathetic nerve bundles in the pancreas, but not within islets (Supplementary Fig. 4). Consistent with the lack of NRTN receptors on β-cells, NRTN did not enhance proliferation of INS-1 cells (data not shown). Collectively, these data indicate that NRTN is unlikely to act directly on β-cells.

GLP-1’s ability to activate several key brain centers was assessed using an indicator of neuronal activation-induced c-Fos expression (26,27). We evaluated whether acute NRTN treatment, like GLP-1, activates neurons in brain of ZDF rats 4 h after a single subcutaneous injection of vehicle, liraglutide, or NRTN (3 mg/kg). As reported (26,27), following liraglutide treatment, c-Fos expression was increased in key hypothalamic and brainstem/medullary centers relevant for metabolic control including the paraventricular nucleus (Fig. 4A and B), as well as the area postrema and the nucleus of the solitary tract (Fig. 4C–E), and the central nucleus of amygdala and the lateral parabrachial nucleus (data not shown). However, c-Fos expression was not enhanced in these (Fig. 4) or any other central nervous system regions (data not shown) following NRTN administration.

The effect of middose NRTN was evaluated in an OGTT after 3 and 16 days of administration, and insulin sensitivity was examined after 20 days. NRTN (3 mg/kg) did not affect body weight and partially reduced hyperglycemia (Fig. 5A and B). After 3 days, OGTTs were not significantly different between vehicle and NRTN groups, whereas liraglutide-treated rats exhibited lower 4-h fasting glucose and improved glucose tolerance (Fig. 5C). Plasma insulin excursions were also reduced in the liraglutide group, but were comparable in vehicle and NRTN groups, except that the 4-h fasting insulin levels were significantly lower after 3 days of NRTN administration (Fig. 5D). OGTT after 16 days of treatment showed a significant reduction in fasting glucose and glucose excursions in NRTN- and liraglutide-treated ZDF rats compared with vehicle controls (Fig. 5E). Consistent with this improvement in glucose tolerance, the NRTN and liraglutide groups had enhanced insulin secretion following glucose administration relative to the vehicle group (Fig. 5F). During ITTs, there was a more rapid reduction in glucose in the NRTN-treated group compared with vehicle controls. The rate of glucose clearance for the first 60 min after insulin injection was significantly greater in the NRTN-treated group versus vehicle controls (Fig. 5G and H), indicating a potential improvement in insulin sensitivity.

To further confirm these observations, we assessed the effect of 3 or 10 mg/kg NRTN on an ITT in ZDF rats following 6 or 17 days of treatment compared with rats treated with rosiglitazone. As previously observed, NRTN had no effect on body weight and dose-dependently inhibited hyperglycemia in ZDF rats (Fig. 6A and B). Rosiglitazone exerted its known effect on increasing body weight (Fig. 6A) and significantly prevented hyperglycemia (Fig. 6B). On day 6, all rats responded to exogenous insulin by reducing glucose (Fig. 6C). After 17 days, however, vehicle-treated rats exhibited some insulin resistance, with minimal glucose-lowering the first hour following insulin administration (Fig. 6D). Both NRTN and rosiglitazone treatments were associated with improved glucose-lowering in response to insulin, with a slightly enhanced effect in the low-dose NRTN group (Fig. 6D).

We examined whether NRTN (with or without liraglutide) could reverse overt hyperglycemia in ZDF rats. Ten-week-old ZDF rats (blood glucose ∼330 mg/dL) were assigned to one of four treatment groups: vehicle, NRTN (10 mg/kg), liraglutide (0.4 mg/kg), or a group receiving both NRTN and liraglutide daily for 28 days. ZL rats were used as normal controls. Liraglutide reduced body weight for the first 8 days (Fig. 7A). NRTN alone did not alter body weight; however, body weight in rats coadministered NRTN and liraglutide was significantly lower than all other ZDF groups from days 9 to 22 (Fig. 7A). Whereas NRTN had no effect on hyperglycemia, liraglutide improved glycemia (Fig. 7B), which was further lowered to normal levels of lean control ZL rats in the NRTN plus liraglutide group (Fig. 7B). After 16 days of treatment, an OGTT indicated that NRTN administration was associated with a significant reduction in overnight-fasted glucose at t = 0 (Fig. 7C). Fasting glucose levels in liraglutide and combination groups were identical, and they had reduced glucose excursion compared with the vehicle or NRTN groups. Liraglutide-treated rats required a higher amount of insulin to maintain the similar glucose levels as seen in the combination group (Fig. 7D). Terminal pancreatic weight, β-cell mass, or α-cell mass was not altered by NRTN alone or in combination with liraglutide (Supplementary Fig. 5). Liraglutide alone, however, was associated with increased β-cell mass (Supplementary Fig. 5).

The effect of these treatments on terminal metabolic parameters was also measured (Table 1). A significant reduction in glucose and %HbA_1c in the liraglutide and combination group was observed, whereas the NRTN group showed a tendency for reduction in glucose levels. Likewise, the change in blood glucose and %HbA_1c levels from the baseline was lower in NRTN alone compared with the vehicle control. The combination group showed a further significant reduction in these parameters compared with the liraglutide group. Plasma insulin levels were higher in the liraglutide group, but not different in other treatment groups. The reversal of hyperglycemia in the liraglutide and combination groups was accompanied with a reduction in the circulating total cholesterol and triglyceride levels. Liver mass and liver lipid were not altered by treatment. Plasma alanine aminotransferase levels were also increased in ZDF rats and significantly reduced in NRTN and combination groups (Table 1). Together, these results indicate that a combination of NRTN and liraglutide can be more effective in reversing diabetes compared with liraglutide alone.

Identification of novel therapeutic mechanisms that can prevent or reverse the progression of diabetes are much needed to successfully stem the tide of the ongoing diabetes epidemic and its associated complications. In this study, we describe a new function for a neurotrophic factor, NRTN, in inhibiting the development of diabetes by a mechanism not yet fully defined, but distinct from GLP-1RAs. We demonstrated that NRTN can prevent the development of hyperglycemia, reduce fasting glucose, and improve metabolic parameters independent of weight loss and reduced food intake. Coadministration of NRTN with liraglutide synergistically reduced weight and reversed hyperglycemia in diabetic ZDF rats. Our results highlight the potential of novel combinatorial approaches to prevent or reverse diabetes progression.

NRTN was identified as a survival factor for sympathetic neurons (28) and is implicated in the development of the enteric nervous system and neurological disorders (29,30). Mice lacking the NRTN receptor Gfra2 exhibit a profound reduction in the parasympathetic innervation of pancreatic islets and impaired vagal tone, but retain normal glucose tolerance (22). NRTN and its family member GDNF have also been reported to improve hind limb innervation in streptozotocin diabetic mice (31). The lack of a direct effect of NRTN on β-cells or islets in vitro and the inability of 3-day NRTN administration to reduce hyperglycemia or enhance insulin secretion suggest that NRTN’s ability to prevent diabetes is not mediated by acute modulation of islet innervation to enhance islet function. It is possible that long-term NRTN may overcome diabetes by controlling innervation defects in a yet unidentified metabolic tissue, which would be consistent with NRTN alleviating insulin resistance only after 2 weeks of dosing as observed in this study.

In addition to regulating survival of many types of central and peripheral neurons (15), GDNF is a potential regulator of metabolic homeostasis. GDNF enhances differentiation and proliferation of pancreatic β-cells during development in mice (14). Transgenic Gdnf overexpression in mice enhanced islet area, pancreatic insulin content, and glucose tolerance (12). We likewise observed increased β-cell mass in prediabetic ZDF rats treated with NRTN, but this was not associated with increased β-cell proliferation. Furthermore, the amount of apoptotic β-cells in the pancreas from vehicle and treatment groups was too low to provide any meaningful quantification (data not shown), suggesting that prevention of cell death may not account for the increased β-cell mass. Additional studies are required to determine the exact temporal relationship between the reduction in glucose and alleviation of glucotoxicity with NRTN treatment and the preservation of islet morphology and β-cell mass. In overtly diabetic ZDF rats, however, NRTN alone or in combination with liraglutide failed to significantly increase β-cell mass or retain islet morphology. Of interest, NRTN did not alter glucose levels in this model, whereas in combination with liraglutide glycemic status was normalized. Whether the lack of impact on islets/β-cells is because of differences in model or analytical technique is not clear. Certainly, enhanced glycemic control with NRTN in diabetic ZDF rats in combination with liraglutide does not require enhanced β-cell mass.

The mechanisms by which GLP-1 regulates metabolism are well appreciated (32,33). In contrast to GLP-1, acute administration of NRTN was unable to stimulate either c-Fos expression in brain areas that are activated by GLP-1RA (26,27) or insulin secretion from β-cells. Consistent with these observations, several days of NRTN administration were required before any in vivo changes in the glucose and insulin levels were observed, whereas such changes were apparent immediately after liraglutide administration. Furthermore, unlike GLP-1RA, NRTN prevented the development of hyperglycemia without affecting body weight or food intake. NRTN-treated animals required lower insulin levels to maintain similar glucose levels despite comparable glucose levels to liraglutide-administered animals. Comparing pancreatic insulin content with plasma insulin levels for individual animals revealed that NRTN-treated animals secreted a relatively lower proportion of insulin to maintain glycemia than liraglutide treatment (Supplementary Fig. 6). The putative insulin sensitizing action of NRTN was supported by the observation that glucose was reduced similarly by insulin in NRTN- (3 or 10 mg/kg) and rosiglitazone-treated rats. The weight-reducing and insulinotropic functions of liraglutide contrast with the weight-independent and potential insulin-sensitizing actions of NRTN. However, additional studies in the appropriate animal model will be necessary to better delineate the specific contribution of insulin sensitization to NRTN's ability to prevent hyperglycemia in ZDF rats.

Remarkably, coadministration of NRTN and liraglutide in hyperglycemic ZDF rats synergistically improved glycemic control and enhanced metabolic parameters compared with liraglutide. Although NRTN alone did not reduce random glucose levels, a significant reduction in fasting glucose and %HbA_1c levels were observed, indicating that NRTN was functionally active but ineffective at reducing glucose and inducing insulin secretion. This suggests that for NRTN to be effective at reducing hyperglycemia in diabetic states, it may require a degree of functional insulin signaling in metabolic tissues and/or residual β-cell function.

Thiazolidinediones and liraglutide are less effective in reversing hyperglycemia in hyperglycemic ZDF rats (34–36). Consistent with these observations, liraglutide-treated animals reduced, but did not normalize, glucose levels or body weight. The addition of NRTN to liraglutide significantly improved these and other metabolic parameters. Furthermore, even for a similar fasting glucose, insulin levels were lower in the NRTN plus liraglutide group than in the liraglutide group, suggesting that either complete normalization of glucose levels or some other action of NRTN allowed liraglutide to sustain its normal ability to reduce body weight and glucose levels. At present, it is unclear how NRTN mediates its metabolic activity in the context of diabetes. Given its role in regulating the peripheral nervous system, it is conceivable that NRTN may restore some neuronal communication to improve insulin sensitivity altered during diabetes. Further studies will be required to test this hypothesis. In summary, we report a novel role of NRTN to act synergistically with a GLP-1 analog to ameliorate diabetes, thus illustrating a potential innovative therapeutic combinations for the treatment of diabetes.

Acknowledgments. The authors thank Drs. Jonah Rainey, Jo Goodman, Mark Berge, Thomas Linke, and Varnika Roy (MedImmune) and Robert Roth and Pia Davidsson (AstraZeneca) for generating human NRTN protein, bioanalysis, pharmacokinetic data, and scientific inputs to enable this study; Drs. Marcus Schindler (AstraZeneca), Lolke de Haan (MedImmune), and Fredrich Harder (Evotec) for scientific contributions and supporting the progression of this project; Anders Dahlstrand and Maria Nilsson at Offspring Biosciences AB (Södertälje, Sweden) who performed the GFRa2 and RET immunofluorescent imaging in ZL rats for able assistance; and the Laboratory Animal Research staff at MedImmune, Gaithersburg, MD, and Gubra, Hørsholm, Denmark, for assistance with animal husbandry, care, and experimental support.

Duality of Interest. J.L.T., H.J., S.A., J.L.L., S.O., N.B., B.B.B., J.C., Y.C., T.O., V.H., S.T., L.J., M.P.C., J.G., C.J.R., C.M.R., and A.S. are/were employees and shareholders of MedImmune. M.S.W. and D.M.S. are employees and shareholders of AstraZeneca. M.F. and P.B. are employees and shareholders of Gubra. C.R., K.S., M.A., U.A., and C.D. are/were employees and shareholders of Evotec. No other potential conflicts of interest relevant to this article were reported.

Author Contributions. J.L.T. designed, conducted, and analyzed studies; generated figures; and reviewed, edited, and wrote sections of the manuscript. C.B.S., H.J., S.A., J.L.L., N.B., V.H., M.F., and P.B. researched and analyzed data. S.O. researched data. B.B.B. performed experiments and analyzed data. J.C. researched and analyzed histological analyses, generated figures, wrote sections of the manuscript, and contributed to project progression. Y.C. contributed to production and supply of NRTN protein and contributed to project progression. T.O. performed statistical analysis and contributed to project progression. C.R. researched and analyzed data, contributed to project progression, and reviewed the manuscript. M.S.W. and D.M.S. researched and analyzed data and contributed to project progression. K.S. reviewed the manuscript. M.A. and C.D. conceived the project and reviewed the manuscript. U.A. analyzed data, contributed to study design and progression of the project, and reviewed the manuscript. S.T. contributed to study design and progression of the project. L.J., M.P.C., and J.G. analyzed data, contributed to study design and progression of the project, and reviewed the manuscript. C.J.R. contributed to discussion, study design, and project progression and edited and reviewed the manuscript. C.M.R. contributed to discussion and study design and reviewed the manuscript. A.S. designed, conducted, and analyzed studies, generated figures, and wrote the manuscript. J.L.T. is the guarantor of this work and, as such, had full access to all the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.

Prior Presentation. Parts of this study were presented in abstract form at the 76th Scientific Sessions of the American Diabetes Association, New Orleans, LA, 10–14 June 2016.