Serotonin Regulates Adult b-Cell Mass by StimulatingPerinatal b-Cell ProliferationJoon Ho Moon1 Yeong Gi Kim1 Kyuho Kim1 Sho Osonoi2 Shuang Wang3 Diane C Saunders4

Juehu Wang5 Katherine Yang5 Hyeongseok Kim16 Junguee Lee7 Ji-Seon Jeong8 Ronadip R Banerjee9

Seung K Kim10 Yingjie Wu311 Hiroki Mizukami2 Alvin C Powers412 Michael S German513 andHail Kim114

Diabetes 202069205ndash214 | httpsdoiorg102337db19-0546

A sufficient b-cell mass is crucial for preventing diabe-tes and perinatal b-cell proliferation is important in de-termining the adult b-cell mass However it is not yetknown how perinatal b-cell proliferation is regulatedHere we report that serotonin regulates b-cell prolifer-ation through serotonin receptor 2B (HTR2B) in an auto-crineparacrine manner during the perinatal period Inb-cellndashspecific Tph1 knockout (Tph1 bKO) mice perina-tal b-cell proliferation was reduced along with the loss ofserotonin production in b-cells Adult Tph1 bKO miceexhibited glucose intolerance with decreased b-cellmass Disruption of Htr2b in b-cells also resulted in de-creased perinatal b-cell proliferation and glucose intoler-ance in adulthood Growth hormone (GH) was found toinduce serotonin production in b-cells through activationof STAT5 during the perinatal period Thus our resultsindicate that GH-GH receptor-STAT5-serotonin-HTR2Bsignaling plays a critical role in determining the b-cellmass by regulating perinatal b-cell proliferation and

defects in this pathway affect metabolic phenotypesin adults

Diabetes develops when b-cells fail to meet insulin de-mand a sufficient b-cell mass is crucial for preventingdiabetes and people with decreased b-cell mass haveincreased susceptibility to type 2 diabetes (12) Despitethe importance of the adult b-cell mass it is not yet clearhow a proper amount of b-cell mass is achieved

During fetal development insulin-producing b-cellsdifferentiate from endocrine progenitor cells and thenewly developed b-cells actively proliferate during theperinatal period to achieve the adult b-cell mass (3) Afterthe perinatal period b-cells maintain a low proliferationrate unless sustained metabolic stress increases systemicinsulin demand Therefore perinatal b-cell proliferation isimportant in determining the adult b-cell mass and sus-ceptibility to type 2 diabetes However we know little

1Graduate School of Medical Science and Engineering Korea Advanced Institute ofScience and Technology Daejeon Korea2Department of Pathology and Molecular Medicine Hirosaki University GraduateSchool of Medicine Hirosaki Japan3Institute of Genome Engineered Animal Models for Human Disease and NationalCenter of Genetically Engineered Animal Models for International Research DalianMedical University Dalian Liaoning China4Division of Diabetes Endocrinology and Metabolism Department of MedicineVanderbilt University Medical Center Nashville TN5Eli and Edythe Broad Center of Regeneration Medicine and Stem Cell ResearchDiabetes Center and Hormone Research Institute University of California SanFrancisco San Francisco CA6Department of Biochemistry Chungnam National University School of MedicineDaejeon Korea7Department of Pathology Daejeon St Maryrsquos Hospital College of Medicine TheCatholic University of Korea Daejeon Korea8Center for Bioanalysis Division of Chemical and Medical Metrology KoreaResearch Institute of Standards and Science Daejeon Korea9Division of Endocrinology Diabetes and Metabolism Department of MedicineUniversity of Alabama School of Medicine Birmingham AL

10Department of Developmental Biology Stanford University School of MedicinePalo Alto CA11Division of Endocrinology Diabetes and Bone Disease Department of MedicineIcahn Mount Sinai School of Medicine New York NY12VA Tennessee Valley Healthcare System Nashville TN13Department of Medicine University of California San Francisco San Francisco CA14KAIST Institute for the BioCentury Korea Advanced Institute of Science andTechnology Daejeon Korea

Corresponding authors Hail Kim hailkimkaistedu and Michael S Germanmgermanucsfedu

Received 1 June 2019 and accepted 14 November 2019

This article contains Supplementary Data online at httpdiabetesdiabetesjournalsorglookupsuppldoi102337db19-0546-DC1

JHM and YGK contributed equally to this work

copy 2019 by the American Diabetes Association Readers may use this article aslong as the work is properly cited the use is educational and not for profit and thework is not altered More information is available at httpwwwdiabetesjournalsorgcontentlicense

Diabetes Volume 69 February 2020 205

ISLETSTUDIES

about how b-cell proliferation is physiologically regulatedduring the perinatal period

Serotonin (5-hydroxytryptamine [5-HT]) is a neuro-transmitter that is synthesized from tryptophan by thesequential actions of tryptophan hydroxylase (TPH) andaromatic amino acid decarboxylase 5-HT exerts its bi-ological action by binding to 5-HT receptor (HTR) (45)Two isoforms of TPH are expressed in a mutually exclusivemanner TPH1 in peripheral nonneuronal tissues andTPH2 in neurons of the central and enteric nervoussystems (45) In pancreatic b-cells TPH1 expression and5-HT production are simultaneously increased duringpregnancy 5-HT released from b-cells acts locally to in-crease b-cell proliferation and insulin secretion throughHTR2B and HTR3 respectively (67) Intriguingly 5-HTproduction in b-cells also increases robustly during theperinatal period (68) These data prompted us to explorethe physiological role of 5-HT in perinatal b-cell expan-sion

Here we show that 5-HT regulates the adult b-cell massby stimulating perinatal b-cell proliferation throughHTR2B in an autocrineparacrine manner Growth hor-mone (GH) but not prolactin (PRL) induces 5-HT pro-duction in b-cells during the perinatal period Humanb-cells also produce 5-HT during the perinatal periodsuggesting that 5-HT may contribute to determining thehuman b-cell mass and susceptibility to diabetes

RESEARCH DESIGN AND METHODS

Animal ExperimentsAll mice used in this study were C57BL6J backgroundstrain Mice were housed in a specific-pathogenndashfreefacility at the Korea Advanced Institute of Science andTechnology with 12-h light and dark cycles under iso-thermal and isohumid conditions Normal chow diet (SCDD10001 Research Diets) and water were given ad libitumTo generate b-cellndashspecific knockout (bKO) mice of Tph1Htr2b Prlr and Stat5 Ins2-Cre (MGI 2387657) mice weremated with mice harboring a Tph1 floxed allele (MGI6271993) an Htr2b floxed allele (MGI 3837400) a Prlrfloxed allele and a Stat5 floxed allele (MGI 3055318)respectively (9ndash14) To generate Ghr bKO mice Ins2-Cremice were mated with mice harboring a Ghr floxed allele(MGI 5056123) (1516) All Cre mice used in this studylacked the GH minigene which has been shown to induceectopic 5-HT production (1718) Htr1btm11(KOMP)Vlcg

(MGI 5463939) and Htr1dtm11(KOMP)Vlcg (MGI 5529062)mice were purchased from KOMP For experiments involvingadult mice 9-week-old male mice were used For high-fatdiet (HFD) studies the diet (60 kcal fat D12492Research Diets) was fed to 12-week-old male mice for8 weeks For the S-961 (insulin receptor antagonist NovoNordisk) study 100 nmolkg mouse was given twice a dayintraperitoneally for 7 days Given that glucose levels arecomparable in Cre mice and floxed control mice Cre

littermates of the experimental (KO) mice were used asthe control unless otherwise stated

Human Pancreatic TissuesHuman pancreas autopsy specimens were obtained fromthe archives of the Department of Pathology and Molec-ular Medicine Hirosaki University All specimens wereformalin-fixed and paraffin-embedded for further stainingConventional hematoxylin and eosin (HampE) staining wasperformed in all cases Fetal cases were from a stillbornbaby with no apparentmalformations on the outer surfaceAn 8-day-old sample was obtained from a male newbornbaby whose cause of death was cardiopulmonary insuffi-ciency secondary to osteogenesis imperfecta A 3-month-old sample was obtained from a female infant whose causeof death was sudden infant death syndrome Pancreassamples from a 35-year-old pregnant woman (gestationalage 37 weeks cause of death pulmonary embolism) anda 72-year-old man (cause of death myocardial infarction)were used as controls The use of paraffin blocks and thestudy design were approved by the ethics committee of theHirosaki University School of Medicine (approval number2014-269) and the study conformed to the provisions ofthe Declaration of Helsinki

Immunostaining and b-Cell Area MeasurementsMouse pancreas tissues were harvested and fixed in 10neutral-buffered formalin solution (Sigma-Aldrich) for 2 h(perinatal) or 4 h (adult) at room temperature and washedwith deionized water for 30 min After washing the tissueswere processed with an automatic tissue processor(TP1020 Leica Biosystems) and embedded in moltenparaffin wax Paraffin-embedded tissues were sliced at4-mm thickness and mounted on adhesive glass slides(081000 Paul Marienfeld) Sections were manually depar-affinized in xylene and rehydrated in ethanol for furtherstaining Antigen retrieval was performed by incubatingthe slides in sodium citrate buffer (10 mmolL sodiumcitrate pH 60) with a pressure cooker for 15 min Forimmunofluorescent staining samples were blocked with2 donkey serum in PBS for 1 h at room temperatureSlides treated with the following primary antibodies wereincubated at 4degC overnight anti-5-HT antibody (rabbitdilution factor 11000 ImmunoStar) anti-glucagon anti-body (11000 mouse Sigma) anti-insulin antibody(11000 guinea pig Dako) anti-Nkx61 antibody (1100mouse Developmental Studies Hybridoma Bank) anti-Pdx1 antibody (1100 mouse Developmental StudiesHybridoma Bank) anti-Ki-67 (11000 rabbit Abcam) andanti-phospho-Histone H3 antibody (11000 rabbit Milli-pore) After incubation and washing the following sec-ondary antibodies were treated and incubated for 2 h atroom temperature Alexa 594ndashconjugated donkey anti-mouse IgG (11000 Jackson ImmunoResearch) Alexa647ndashconjugated donkey anti-rabbit IgG (11000 JacksonImmunoResearch) and FITC-conjugated donkey anti-guinea pig IgG (11000 Jackson ImmunoResearch) In

206 Serotonin Regulates Perinatal b-Cell Proliferation Diabetes Volume 69 February 2020

Situ Cell Death Detection Kit Fluorescein (TUNEL Sigma-Aldrich) was used to measure apoptosis Samples weretreated with DAPI (12000 Invitrogen) for 5 min andmounted with fluorescence mounting medium (Dako)Pictures were obtained using a confocal microscope (LSM780 Carl Zeiss) or a fluorescence microscope with a DS-Ri2camera (Nikon) Immunofluorescence intensity was mea-sured with region-of-interest statistics calculated us-ing Nikon NIS-Elements version 43 software (referencerange 0ndash256) 5-HT immunofluorescence intensity wasmeasured in b-cells positive for immunofluorescencestaining of insulin The background intensity was mea-sured in exocrine pancreas to evaluate whether theobtained images were suitable for comparison

Immunohistochemistry was conducted with a biotin-free polymer detection system for 5-HT (MACH 3 BiocareMedical) After the application of secondary antibody forrabbit IgG the immunoreaction products were colorizedwith diaminobenzidine (Thermo Fisher Scientific) Speci-ficity was confirmed by the lack of positive staining afteromission of the first antibody or replacement with non-immune serum The sections were observed and captured bybright-field microscopy (Axio Imager A1 or M1 Carl Zeiss)

For the b-cell mass and proliferation measurementsslides of fresh-frozen paraffin-embedded pancreas were cho-sen with an 80-mm interval These slides were immunos-tained with insulin and whole slides were scanned with JuLIStage (NanoEntek) b-Cell mass was measured by insulinimmunoreactive area divided by whole-pancreas area (morethan eight slides per mouse) b-Cell proliferation was mea-sured by the percentage of insulin and phospho-HistoneH3 copositive cells in all b-cells (2000 b-cells per mouse)

Glucose DynamicsFor intraperitoneal glucose tolerance test and in vivoglucose-stimulated insulin secretion (GSIS) test mice werefasted for 16 h and D-glucose (2 gkg body weight) wasinjected intraperitoneally Blood glucose level was mea-sured from tail vein using a glucometer (GlucoDrPlusAllmedicus) For insulin measurement blood was collectedat 0 15 and 30min after glucose injection in a heparinizedtube which was immediately centrifuged in 1500g for10 min at 4degC Plasma insulin concentration was measuredusing an ultrasensitive insulin ELISA kit (80-INSMSU-E01ALPCO) For insulin tolerance test mice were fasted for6 h before an intraperitoneal Humulin R (Eli Lilly) injec-tion (075 unitskg body weight)

For ex vivo GSIS mouse islets were isolated usingcollagenase The major process of the technique was de-scribed previously (19) First islets were incubated at 37degCin humidified gas of 5 CO2 and 95 air in RPMI medium(Thermo Fisher Scientific) with 100 unitsmL penicillin-streptomycin (Thermo Fisher Scientific) and 10 FBS(Thermo Fisher Scientific) After 4 h of incubation isletswere moved to Krebs-Ringer HEPES (KRH) buffer (pH 74115 mmolL NaCl 25 mmolL HEPES 24 mmolLNaHCO3 5 mmolL KCl 1 mmolL MgCl2 25 mmolL

CaCl2 and 1 mgmL BSA) containing 28 mmolL glucosefor 30 min Islets were handpicked (n 5 10 size-matchedislets per individual sample) to a 12-well noncoated dishTo measure basal insulin secretion samples were incu-bated with 28 mmolL glucose containing KRH buffer for15 min and 50 mL of supernatants were obtained forinsulin measurements For high-glucose stimulation sam-ples were incubated with 168 mmolL glucose containingKRH buffer for 15 min and supernatants were obtainedNext the islets were sonicated and incubated in acid-ethanol (15 HCl in 100 mL of 70 ethanol) for 18 h at4degC for intracellular insulin extraction The same volume of1 molL Tris Cl buffer (pH 80) was added for neutralizationAll supernatants were immediately snap frozen with liquidnitrogen and kept frozen at280degC until ELISA experimentsInsulin secretion was calculated by dividing the secretedinsulin by the total insulin content extracted from the islets

GH Studies and 5-HT MeasurementsFor experiments with GH islets were isolated frommice at3 to 4 weeks of age and incubated in RPMI medium bTC3and MIN6 cells were maintained in DMEM containing25 mmolL glucose and supplemented with 10 or 15FBS 100 unitsmL penicillin 01 mgmL streptomycinand where applicable (ie for MIN6 cells) 715 mmolL2-mercaptoethanol Cells were incubated at 37degC in a hu-midified incubator containing 5 CO2 For GH treatmentcells were incubated in serum-free medium for 12 h andthen recombinant mouse GH (National Hormone andPeptide Program) was added to the medium RNA wasextracted 24 h after the treatment

For measurement of 5-HT secretion neonatal pancreaswas harvested at postnatal day 0 (P0) and washed withPBS The P0 pancreas was incubated in 1 mL of KRH buffersupplemented with protease inhibitor cocktail for 10 minin a CO2 incubator Supernatants were obtained and thesecreted 5-HT concentration was measured using ELISA(LDN Nordhorn Germany) This value was normalized tothe protein content which was measured using thebicinchoninic acid technique (Thermo Fisher Scientific)

Quantitative RT-PCRTotal RNA was extracted using TRIzol (Invitrogen) follow-ing the manufacturerrsquos protocol One microgram of totalRNA was used to generate cDNA High-Capacity cDNAReverse Transcription Kit (Applied Biosystems) Quanti-tative RT-PCR (qRT-PCR) was performed with Fast SYBRGreen Master Mix (Applied Biosystems) using a Viia7 Real-Time PCR System (Applied Biosystems) Gene ex-pression was calculated by relative gene expression anal-ysis using the DCt (threshold cycle) method Actb was usedas an internal control Primer sequences are listed inSupplementary Table 1

Statistical AnalysisAll data are presented as the mean 6 SEM Statisticalsignificance was obtained by Student t test (two-tailed) or

diabetesdiabetesjournalsorg Moon and Associates 207

Situ Cell Death Detection Kit Fluorescein (TUNEL Sigma-Aldrich) was used to measure apoptosis Samples weretreated with DAPI (12000 Invitrogen) for 5 min andmounted with fluorescence mounting medium (Dako)Pictures were obtained using a confocal microscope (LSM780 Carl Zeiss) or a fluorescence microscope with a DS-Ri2camera (Nikon) Immunofluorescence intensity was mea-sured with region-of-interest statistics calculated us-ing Nikon NIS-Elements version 43 software (referencerange 0ndash256) 5-HT immunofluorescence intensity wasmeasured in b-cells positive for immunofluorescencestaining of insulin The background intensity was mea-sured in exocrine pancreas to evaluate whether theobtained images were suitable for comparison

Immunohistochemistry was conducted with a biotin-free polymer detection system for 5-HT (MACH 3 BiocareMedical) After the application of secondary antibody forrabbit IgG the immunoreaction products were colorizedwith diaminobenzidine (Thermo Fisher Scientific) Speci-ficity was confirmed by the lack of positive staining afteromission of the first antibody or replacement with non-immune serum The sections were observed and captured bybright-field microscopy (Axio Imager A1 or M1 Carl Zeiss)

For the b-cell mass and proliferation measurementsslides of fresh-frozen paraffin-embedded pancreas were cho-sen with an 80-mm interval These slides were immunos-tained with insulin and whole slides were scanned with JuLIStage (NanoEntek) b-Cell mass was measured by insulinimmunoreactive area divided by whole-pancreas area (morethan eight slides per mouse) b-Cell proliferation was mea-sured by the percentage of insulin and phospho-HistoneH3 copositive cells in all b-cells (2000 b-cells per mouse)

Glucose DynamicsFor intraperitoneal glucose tolerance test and in vivoglucose-stimulated insulin secretion (GSIS) test mice werefasted for 16 h and D-glucose (2 gkg body weight) wasinjected intraperitoneally Blood glucose level was mea-sured from tail vein using a glucometer (GlucoDrPlusAllmedicus) For insulin measurement blood was collectedat 0 15 and 30min after glucose injection in a heparinizedtube which was immediately centrifuged in 1500g for10 min at 4degC Plasma insulin concentration was measuredusing an ultrasensitive insulin ELISA kit (80-INSMSU-E01ALPCO) For insulin tolerance test mice were fasted for6 h before an intraperitoneal Humulin R (Eli Lilly) injec-tion (075 unitskg body weight)

For ex vivo GSIS mouse islets were isolated usingcollagenase The major process of the technique was de-scribed previously (19) First islets were incubated at 37degCin humidified gas of 5 CO2 and 95 air in RPMI medium(Thermo Fisher Scientific) with 100 unitsmL penicillin-streptomycin (Thermo Fisher Scientific) and 10 FBS(Thermo Fisher Scientific) After 4 h of incubation isletswere moved to Krebs-Ringer HEPES (KRH) buffer (pH 74115 mmolL NaCl 25 mmolL HEPES 24 mmolLNaHCO3 5 mmolL KCl 1 mmolL MgCl2 25 mmolL

CaCl2 and 1 mgmL BSA) containing 28 mmolL glucosefor 30 min Islets were handpicked (n 5 10 size-matchedislets per individual sample) to a 12-well noncoated dishTo measure basal insulin secretion samples were incu-bated with 28 mmolL glucose containing KRH buffer for15 min and 50 mL of supernatants were obtained forinsulin measurements For high-glucose stimulation sam-ples were incubated with 168 mmolL glucose containingKRH buffer for 15 min and supernatants were obtainedNext the islets were sonicated and incubated in acid-ethanol (15 HCl in 100 mL of 70 ethanol) for 18 h at4degC for intracellular insulin extraction The same volume of1 molL Tris Cl buffer (pH 80) was added for neutralizationAll supernatants were immediately snap frozen with liquidnitrogen and kept frozen at280degC until ELISA experimentsInsulin secretion was calculated by dividing the secretedinsulin by the total insulin content extracted from the islets

GH Studies and 5-HT MeasurementsFor experiments with GH islets were isolated frommice at3 to 4 weeks of age and incubated in RPMI medium bTC3and MIN6 cells were maintained in DMEM containing25 mmolL glucose and supplemented with 10 or 15FBS 100 unitsmL penicillin 01 mgmL streptomycinand where applicable (ie for MIN6 cells) 715 mmolL2-mercaptoethanol Cells were incubated at 37degC in a hu-midified incubator containing 5 CO2 For GH treatmentcells were incubated in serum-free medium for 12 h andthen recombinant mouse GH (National Hormone andPeptide Program) was added to the medium RNA wasextracted 24 h after the treatment

For measurement of 5-HT secretion neonatal pancreaswas harvested at postnatal day 0 (P0) and washed withPBS The P0 pancreas was incubated in 1 mL of KRH buffersupplemented with protease inhibitor cocktail for 10 minin a CO2 incubator Supernatants were obtained and thesecreted 5-HT concentration was measured using ELISA(LDN Nordhorn Germany) This value was normalized tothe protein content which was measured using thebicinchoninic acid technique (Thermo Fisher Scientific)

Quantitative RT-PCRTotal RNA was extracted using TRIzol (Invitrogen) follow-ing the manufacturerrsquos protocol One microgram of totalRNA was used to generate cDNA High-Capacity cDNAReverse Transcription Kit (Applied Biosystems) Quanti-tative RT-PCR (qRT-PCR) was performed with Fast SYBRGreen Master Mix (Applied Biosystems) using a Viia7 Real-Time PCR System (Applied Biosystems) Gene ex-pression was calculated by relative gene expression anal-ysis using the DCt (threshold cycle) method Actb was usedas an internal control Primer sequences are listed inSupplementary Table 1

Statistical AnalysisAll data are presented as the mean 6 SEM Statisticalsignificance was obtained by Student t test (two-tailed) or

diabetesdiabetesjournalsorg Moon and Associates 207

ANOVA to compare differences among three or moregroups Tukey honestly significant difference method wasused for post hoc analysis Statistical analyses were con-ducted using SPSS version 22 software (IBM Corporation)The levels of statistical significance were P 005 P 001 and P 0001

Data and Resource AvailabilityAll data that support the findings of this study are availablefrom the authors upon reasonable request

RESULTS

5-HT Production in b-Cells During the Perinatal PeriodWe detected the presence of 5-HT in human pancreaticislets during the perinatal period ranging from a gesta-tional age of 24 weeks to a postnatal age of 19 months (Fig

1A and Supplementary Fig 1A and B) 5-HTndashpositive cellswere more frequently observed during the perinatal pe-riod and 5-HT staining was confined to endocrine cellsSimilarly 5-HT staining was observed in mouse b-cells atP0ndashP5 and the expression pattern of the Tph1 mRNA inpancreas paralleled the change in 5-HT immunostaining(Fig 1B and Supplementary Fig 1CndashE) In humans themRNA expression of TPH1 was higher than that of TPH2in b-cells and TPH1 mRNA expression was higher in fetalb-cells than in adult b-cells (20) (Supplementary Table 2)These data suggest that 5-HT may have a common func-tion in human and rodent b-cells during the perinatalperiod and that 5-HT production mainly depends on TPH1than on TPH2

To elucidate the physiological role of 5-HT in b-cells wegenerated Tph1 bKO mice (12) Tph1 bKO was confirmed

Figure 1mdash5-HT regulates b-cell proliferation during the perinatal periodA Immunohistochemical staining (5-HT in brown) and HampE stainingin perinatal human pancreata Fetal (gestational age 24 weeks) postnatal 8-day-old and 3-month-old pancreata were stained Pancreassamples from a 35-year-old pregnant woman (gestational age 37 weeks) and a 72-year-old man were used as a positive and negativecontrol respectively B Immunofluorescent (IF) staining of perinatal pancreas from C57BL6J mice labeled insulin (green) and 5-HT (red)CndashF Pancreata of control (Cre) and Tph1 bKO mice were assessed at P0 C Representative IF staining labeled insulin (green) and amylase(AMYL) (red) at P0 Images of a whole-slide section were obtained using a slide scannerD b-Cell mass was quantified as the percentage ofinsulin-positive area over whole pancreas area at P0 (n 5 4 micegroup) E IF staining labeled insulin (green) and phospho-Histone H3(PHH3) (red) Arrows indicate insulin and PHH3 copositive cells b-Cell proliferation was quantified by counting the percentage of insulin-positive cells labeled with PHH3 at P0 (n5 4 micegroup) F mRNA levels of cyclin families were measured by qRT-PCR in RNAs extractedfrom pancreata at P0 (n 5 4ndash5 micegroup) Scale bars 5 50 mm Data are mean 6 SEM P 005 P 001 P 0001 INS insulin

208 Serotonin Regulates Perinatal b-Cell Proliferation Diabetes Volume 69 February 2020

by decreased Tph1mRNA expression in the pancreas at P0and the reduction of 5-HT production to a level undetect-able by immunofluorescence staining in b-cells during theperinatal period and pregnancy (Supplementary Fig 2Aand B) Next we evaluated 5-HT secretion from b-cells inthe neonatal P0 pancreas (Supplementary Fig 2C) Glucosedid not stimulate 5-HT secretion in the P0 pancreasbecause b-cells were not mature enough to sense theincrease in glucose levels Tolbutamide treatment in-creased 5-HT secretion in the P0 pancreas but this re-sponse was perturbed in the Tph1 bKO pancreas Theseresults suggest that 5-HT is produced and secreted in theb-cell during the perinatal period and can activate HTR inan autocrineparacrine manner

5-HT Stimulates Perinatal b-Cell ProliferationWe further explored the physiological role of 5-HT in Tph1bKO mice The loss of 5-HT production did not affectb-cell differentiation as evidenced by normal expressionof PDX1 and NKX61 (Supplementary Fig 2D) Intrigu-ingly the b-cell mass was remarkably reduced in Tph1 bKOmice whereas the size of the pancreas was unchanged (Fig1C and D and Supplementary Fig 2E) Immunofluores-cence staining of proliferation markers (phospho-Histone

H3 and Ki-67) and TUNEL assays revealed that b-cellproliferation was reduced by 75 in Tph1 bKO micewithout obvious induction of b-cell death (Fig 1E andSupplementary Fig 2F and G) The expression levels ofcyclin family members (ie D1 D2 D3 A2 B1) werereduced in the pancreas of Tph1 bKO mice at P0 (Fig 1F)Collectively these data suggest that 5-HT increases theb-cell mass by promoting perinatal b-cell proliferation

To examine the long-term impacts of the reducedperinatal b-cell proliferation seen in Tph1 bKO mice weevaluated the metabolic phenotypes of Tph1 bKO mice at9 weeks of age Tph1 bKOmice showed normal growth andinsulin sensitivity (Supplementary Fig 3A and B) how-ever their glucose tolerance was impaired (Fig 2A) Plasmainsulin levels at fasting and 15 min after glucose admin-istration were lower and the insulin secretory capacity waslower in Tph1 bKO islets than in control islets (Fig 2B andC) In adult Tph1 bKO mice the b-cell proliferation andpancreas size were comparable to those of control micebut the b-cell mass was reduced by more than half (to40) compared with that of control mice (Fig 2DndashF) Toinvestigate the mechanism underlying the reduction in theb-cell mass of adult Tph1 bKO mice we evaluated

Figure 2mdash5-HT regulates adult b-cell mass and is necessary to maintain glucose homeostasis Metabolic phenotypes were evaluated in9-week-old male control and Tph1 bKOmice A and B Blood glucose and plasma insulin levels were measured after intraperitoneal glucoseinjection (2 gkg) after a 16-h fast Both Cre-only and floxed-only mice were used as controls for the glucose tolerance test Because glucoselevels were comparable between the two groups Cre-only mice were used as the control in further experiments (n 5 4ndash8 micegroup) CSecreted insulin wasmeasuredwith ELISA after a 15-min incubation of islets isolated from 9-week-old control (Cre) and Tph1bKOmice in 28or 168 mmolL glucose Insulin secretion was normalized to the total insulin content extracted from the islets (n 5 5group) D b-Cell masswas quantified as the percentage of insulin-positive area over whole-pancreas area of 9-week-old mice (n 5 4 micegroup) EImmunofluorescent staining labeled insulin (green) and Ki-67 (red) with arrows indicating insulin and Ki-67 copositive cells b-Cellproliferation was quantified by counting the percentage of insulin-positive cells labeled with Ki-67 in 9-week-old mice (n 5 4 micegroup)F Pancreas weight wasmeasured after harvest in 9-week-oldmice (n5 4micegroup) Scale bar5 50mm Data aremean6 SEM P 005P 001 P 0001 control (Cre) vs KO P 001 control (floxed) vs KO INS insulin

diabetesdiabetesjournalsorg Moon and Associates 209

apoptosis and endoplasmic reticulum stress Howeverwe did not observe any significant difference in theTUNEL-positive cell numbers or endoplasmic reticulumstress gene expression levels in islets of Tph1 bKO miceand mature b-cell markers were appropriately expressed(Supplementary Fig 3CndashF) These data indicate thatthe reduced b-cell mass seen in adult Tph1 bKO micecan be attributed to the reduction in perinatal b-cellproliferation Thus 5-HT stimulation of perinatal b-cell

proliferation is necessary to achieve the adult b-cell massneeded to accomplish normoglycemia

5-HT Signaling Through HTR2B Regulates b-CellProliferationAmong the 14 HTRs in rodents 6 are expressed inpancreatic islets (6) To determine the downstream HTRsinvolved in perinatal b-cell proliferation we examined theexpression levels of these 6 (HTR1B 1D 2A 2B 3A and

Figure 3mdashHTR2B is the downstream target of 5-HT in perinatalb-cellsA mRNAexpression levels of indicatedHTRswere assessed by qRT-PCR in RNAs extracted from pancreata during embryonic development (n5 4micegroup) BndashE Pancreata of control (Cre) andHtr2b bKO atP0 were evaluated B HampE staining was performed in control (Cre) and Htr2b bKO pancreas at P0 C b-Cell mass was quantified as thepercentage of insulin-positive area over whole-pancreas area at P0 (n5 4micegroup)D Pancreas weight wasmeasured after harvest at P0(n 5 4ndash5 micegroup) E Immunofluorescent staining labeled insulin (green) and phospho-Histone H3 (PHH3) (red) with arrows indicatinginsulin andPHH3 copositive cells b-Cell proliferationwas quantified by counting the percentage of insulin-positive cells labeledwith PHH3 atP0 (n5 4 micegroup) FndashJ Metabolic phenotypes were evaluated in 9-week-old male control and Htr2b bKOmice F and G Blood glucoseand plasma insulin levels were measured after intraperitoneal glucose injection (2 gkg) after a 16-h fast Both Cre-only and floxed-only micewere used as controls for glucose tolerance tests Because glucose levels were comparable between the two Cre-only mice were used asthe control in further experiments (n5 4micegroup)H Secreted insulin wasmeasuredwith ELISA after a 15-min incubation of islets isolatedfrom 9-week-old control (Cre) andHtr2b bKOmice in 28 or 168mmolL glucose Insulin secretion was normalized to the total insulin contentextracted from the islets (n5 3group) I b-Cell mass was quantified as the percentage of insulin-positive area over whole-pancreas area of9-week-old mice (n 5 4 micegroup) J b-Cell proliferation was quantified by counting the percentage of insulin-positive cells labeled withKi-67 in 9-week-old mice (n5 4 micegroup) K Pancreas weight was measured after harvest in 9-week-old mice (n5 4 micegroup) Scalebar 5 50 mm Data are mean 6 SEM P 005 P 001 control (Cre) vs KO P 005 control (floxed) vs KO INS insulin

210 Serotonin Regulates Perinatal b-Cell Proliferation Diabetes Volume 69 February 2020

3B) in embryonic pancreas The expression levels of Htr1bHtr2b and Htr3amRNAs showed increasing trends duringpancreatic development (Fig 3A) Among them HTR3 isknown to be involved in insulin secretion but not in b-cellproliferation (7) and the loss of HTR1B or HTR1D did notdecrease b-cell proliferation at P0 (Supplementary Fig 4)In contrast the b-cell mass was reduced by55 inHtr2bbKOmice at P0 while the weight of the pancreas remainedunchanged (Fig 3BndashD) The b-cell proliferation was re-duced in Htr2b bKO mice without any obvious increase inb-cell death (Fig 3E and Supplementary Fig 2G) Thusreduced perinatal b-cell proliferation was responsible forthe reduction of b-cell mass in Htr2b bKO mice during theperinatal period

We further analyzed the adult phenotypes of Htr2bbKO mice Htr2b bKO mice became glucose intolerant inadulthood while exhibiting normal body weight and insulinsensitivity (Fig 3F and Supplementary Fig 5A and B)Plasma insulin levels at fasting and 15 min after glucoseadministration were lower in Htr2b bKO mice (Fig 3G)However GSIS from isolated Htr2b bKO islets was notsignificantly impaired (Fig 3H) Instead the b-cell masswas persistently reduced in adult Htr2b bKO mice to40 of control mice (Fig 3I) The b-cell proliferation andpancreas size were similar between control and Htr2b bKOmice and b-cell markers were appropriately expressed inHtr2b bKO mice (Fig 3J and K and Supplementary Fig5C) Thus Htr2b bKO mice phenocopied Tph1 bKO miceexcept for the ex vivo GSIS These data support theconclusion that 5-HT stimulates b-cell proliferationthrough HTR2B during the perinatal period Furthermoredefects in 5-HTndashdependent perinatal b-cell proliferationare associated with decreases in the adult b-cell massindicating that 5-HT plays an important role in determin-ing the adult b-cell mass

5-HT Is Not Involved in HFD-Induced b-CellProliferationHFD is known to induce b-cell proliferation to compensatefor the increased insulin demand To further assesswhether 5-HT is necessary for the ability of b-cells to

endure metabolic stress we fed mice an HFD for 8 weeksbeginning at 12 weeks of age Neither Tph1 expression nor5-HT production was induced in b-cells by HFD (21)(Supplementary Fig 6A) Tph1 bKO mice showed a similarbody weight gain and insulin sensitivity compared withcontrol mice (Supplementary Fig 6B and C) Howeverglucose intolerance was more rapidly aggravated uponHFD in Tph1 bKO mice compared with control micewhereas b-cell proliferation was comparable (Fig 4A andB) These data suggest that 5-HT is not involved in theinduction of compensatory b-cell proliferation in responseto metabolic stress This conclusion was further supportedby measuring b-cell proliferation in mice with insulinresistance induced by the insulin receptor antagonistS-961 Induction of b-cell proliferation was comparablebetween control and Tph1 bKO mice upon S-961 admin-istration (Fig 4C and Supplementary Fig 6D) Thus 5-HTstimulates b-cell proliferation not in response to metabolicstress but rather in response to physiological stimuliduring the perinatal period This notion prompted us toexplore the upstream signals that regulate 5-HT produc-tion in b-cells during the perinatal period

GH Induces 5-HT Production in b-Cells During thePerinatal PeriodDuring pregnancy placental lactogen stimulates Tph1expression through the PRL receptor (PRLR)-STAT5 path-way (622) and plasma PRL levels remain high during theperinatal period (23) To determine whether PRL inducesTph1 expression in perinatal b-cells through the PRLR-STAT5 pathway we investigated b-cellndashspecific Prlr KO(Prlr bKO) and Stat5 KO (Stat5 bKO) mice Similar toobservations made during pregnancy 5-HT was not de-tectable in the b-cells of Stat5 bKOmice but in contrast topregnancy the b-cells in Prlr bKOmice still produced 5-HTat P0 (Fig 5A) These findings suggest that differentupstream signals induce Tph1 expression during the peri-natal period

Similar to PRL signaling through PRLR GH has beenshown to act through the GH receptor (GHR) on b-cells tostimulate STAT5 phosphorylation and b-cell proliferation

Figure 4mdash5-HT is not involved in b-cell proliferation induced by metabolic stress HFD was fed to 12-week-old mice for 8 weeks A Bloodglucose levels weremeasured after intraperitoneal glucose injection (2 gkg) after a 16-h fast (n5 4ndash5micegroup)B b-Cell proliferation wasquantified by counting the percentage of insulin-positive cells labeled with Ki-67 (n 5 4 micegroup) C b-Cell proliferation upon S-961administration (100 nmolkg mouse twice a day) for 7 days was quantified by counting the percentage of insulin-positive cells labeled withKi-67 (n 5 4 micegroup) Data are mean 6 SEM P 005 P 001

diabetesdiabetesjournalsorg Moon and Associates 211

(1624) Disruption of Ghr in b-cells results in a decrease inthe b-cell mass and insulin secretory function upon HFD(16) GHR expression is increased in human fetal b-cellswhereas PRLR expression is not (Supplementary Table 2)Therefore we generated b-cellndashspecific Ghr KO (Ghr bKO)mice and found that they failed to produce 5-HT in b-cellsat P0 (Fig 5A) Consistently Tph1 mRNA expression wasdecreased in pancreas at P0 in both Stat5 bKO mice andGhr bKO mice but not in Prlr bKO mice (SupplementaryFig 6EndashG) The loss of 5-HT production was associatedwith a reduction in b-cell proliferation in both Ghr bKO

mice and Stat5 bKO mice but not in Prlr bKO mice (Fig5BndashE) Consistent with these reductions in b-cell prolif-eration the b-cell mass was reduced in Ghr bKO mice andStat5 bKO mice but not in Prlr bKO mice (Fig 5FndashH)

Next we examined whether GH can directly induce5-HT production in b-cells during the perinatal periodPlasma GH levels peaked at embryonic day 175 (E175)and gradually declined after birth (Supplementary Fig 6H)GH treatment directly increased the expression of Tph1and Mki67 in wild-type islets and this response wasperturbed in Tph1 bKO islets (Fig 5I and J) GH also

Figure 5mdashGH regulates 5-HT production in b-cells during the perinatal period Pancreas of control (Cre) Prlr bKO Stat5 bKO andGhr bKOmice at P0 were examined A 5-HT production in b-cells at P0 were assessed by immunofluorescent (IF) staining which labeled 5-HT (red)and insulin (green)B Representative HampE staining and IF staining labeled insulin (green) and phospho-HistoneH3 (PHH3) (red) at P0 Arrowsindicate insulin and PHH3 copositive cells CndashE b-Cell proliferation of Prlr bKO Stat5 bKO and Ghr bKO mice at P0 was quantified bycounting the percentage of insulin-positive cells labeled with PHH3 (n5 4ndash6micegroup) FndashH b-Cell mass was quantified as the percentageof insulin-positive area over whole-pancreas area in Prlr bKO Stat5 bKO andGhr bKOmice at P0 (n5 4ndash6 micegroup) I and J Induction ofTph1 or Mki67 in islets isolated from 3-week-old control (Cre) and Tph1 bKO mice was measured by qRT-PCR Islets were incubated withrecombinant mouse GH (1000 ngmL) for 24 h (n 5 3ndash5group) K Schematic describing the GHR-STAT5-TPH1-HTR2B axis for b-cellproliferation during the perinatal period Scale bars 5 50 mm Data are mean 6 SEM P 005 P 001 P 0001 INS insulin

212 Serotonin Regulates Perinatal b-Cell Proliferation Diabetes Volume 69 February 2020

increased Tph1 expression in bTC3 and MIN6 cells (Sup-plementary Fig 6I) However no significant difference inHTR expression was observed upon GH treatment (Sup-plementary Fig 6J) These results suggest that GH signalsthrough GHR-STAT5 to stimulate TPH1 expression duringthe perinatal period which in turn induces 5-HT pro-duction in b-cells

Taken together we identified that 5-HT plays a key rolein determining the adult b-cell mass by regulating b-cellproliferation during the perinatal period (Fig 5K) GHbinds to GHR and activates STAT5 to stimulate 5-HTproduction by increasing TPH1 expression in perinatalb-cells In turn 5-HT stimulates b-cell proliferationthrough HTR2B in an autocrineparacrine manner Re-markably defects in this pathway lead to a reduction inadult b-cell mass and glucose intolerance which is exac-erbated by metabolic stress

DISCUSSION

Pancreatic b-cells undergo physiological mass expansionduring the perinatal period and pregnancy Intriguingly5-HT production increases robustly in b-cells during bothperiods and 5-HTHTR2B works as a common mechanismfor b-cell proliferation in response to different externalupstream signals (68) 5-HT production in human b-cellshas also been reported by multiple groups (625ndash27)although this has been argued as the expression of TPH1 islow and STAT5 functions differently in human b-cells(2829) Recent advances in b-cell transcriptomics supportour hypotheses that 5-HT is produced from perinatalb-cells to stimulate its proliferation through HTR2B (Sup-plementary Table 2) Blodgett et al (20) performed RNAsequencing analysis of sorted human b-cells whichshowed that TPH1 expression was enriched in fetal b-cellscompared with adult b-cells and GHR expression was10-fold higher in fetal b-cells HTR2B was among the mostabundant HTR transcripts in human fetal b-cells althoughpotential contributions by other HTRs should be furtherinvestigated Qiu et al (30) performed transcriptomicanalysis of mouse b-cells sorted from different perinatalperiods and found that Tph1 expression peaks at P0 whichis in accordance with our findings We also directly dem-onstrated the presence of 5-HT in fetal and perinatalhuman b-cells

GH production increases during fetal development andsurges at birth (10ndash50 ngmL) (3132) After birth the GHlevel falls within 1 month to 10 ngmL (3233) Tran-sient and robust 5-HT production in b-cells during theperinatal period parallels these dynamic changes in GHlevels in both humans and mice Our results showingdefects in b-cell proliferation in each bKO model of theGHR-STAT5-TPH1-HTR2B axis are in accordance withprevious reports showing that b-cell proliferation andcyclin D2 expression are altered in models of reduced GHsignaling (34ndash37) In humans children requiring GH re-placement have a six- to eightfold higher incidence ofdiabetes relative to age-matched untreated children

(3839) Inadequate GH stimulation at an earlier age inthese children may increase the risk of diabetes by re-ducing the perinatal expansion of the b-cell mass Inaddition the yield of islet clusters increases when humanfetal pancreas are cocultured with GH (40) These datasupport the potential implication of GH in the expansionof the b-cell mass

In summary we show that 5-HT plays a critical role inGH signaling to determine the b-cell mass by regulatingb-cell proliferation through HTR2B during the perinatalperiod Future studies of this pathway may provide newinsights into the genetics and pathophysiology of diabetesand could suggest newmethods for regenerating b-cells forpeople with diabetes

Acknowledgments The authors thank Hee-Saeng Jung Ju Eun Kim andHannah Jung (Korea Advanced Institute of Science and Technology) for technicalsupport Drs Gerold Grodsky and William Rutter (University of California SanFrancisco) for helpful discussions Dr Gerard Karsenty (Columbia University) forthe Htr2b floxed mice Dr Lothar Hennighausen (National Institute of Diabetes andDigestive and Kidney Diseases) for the Stat5 floxed mice Do Yon Kim and DrYong-Sam Kim (Korea Research Institute of Bioscience amp Biotechnology) for helpwith GHR studies Dr Lauge Schaffer (Novo Nordisk) for S-961 and all oflaboratory members for helpful discussionsFunding This work was supported by grants from the Health FellowshipFoundation (to JHM) National Research Foundation of Korea (grant numbersNRF-2017H1A2A1042095 to JHM NRF-2014H1A8A1022054 to YGK NRF-2013H1A8A1003985 and NRF-2018R1A6A3A01012333 to HyK and NRF-2015M3A9B3028218 and NRF-2018R1A2A3074646 to HaK) National Institutesof Health (P30-DK-063720 to MSG) Larry L Hillblom Foundation (2014-D-004-NET and 2018-D-006-NET to MSG) and JDRF (10-2010-553 to HaK) Some ofthis research was performed using resources andor funding provided by theNational Institute of Diabetes and Digestive and Kidney Diseasesndashsupported HIRN(RRIDSCR_014393 https hirnetworkorg UC4 DK104211 DK108120) by grantnumbers DK106755 and DK20593 and by grants from The Leona M and Harry BHelmsley Charitable Trust and the Department of Veterans Affairs (BX000666) toACPDuality of Interest No potential conflicts of interest relevant to this articlewere reportedAuthor Contributions JHM YGK KK SW JW KY HyK JLJ-SJ and YW performed the animal experiments and analyzed the obtaineddata SO DCS HM and ACP performed the human experiments andanalyzed the obtained data JHM YGK RRB SKK YW HM ACPMSG and HaK analyzed the data and reviewed the manuscript JHM YGKMSG and HaK generated the hypothesis designed the experiments and wrotethe manuscript HaK is the guarantor of this work and as such had full access toall the data in the study and takes responsibility for the integrity of the data and theaccuracy of the data analysis

References1 Prentki M Nolan CJ Islet beta cell failure in type 2 diabetes J Clin Invest20061161802ndash18122 Kendall DM Sutherland DE Najarian JS Goetz FC Robertson RP Effects ofhemipancreatectomy on insulin secretion and glucose tolerance in healthy hu-mans N Engl J Med 1990322898ndash9033 Gregg BE Moore PC Demozay D et al Formation of a human b-cellpopulation within pancreatic islets is set early in life J Clin Endocrinol Metab 2012973197ndash32064 Walther DJ Peter JU Bashammakh S et al Synthesis of serotonin bya second tryptophan hydroxylase isoform Science 200329976

diabetesdiabetesjournalsorg Moon and Associates 213

5 Zhang X Beaulieu JM Sotnikova TD Gainetdinov RR Caron MG Tryptophanhydroxylase-2 controls brain serotonin synthesis Science 20043052176 Kim H Toyofuku Y Lynn FC et al Serotonin regulates pancreatic beta cellmass during pregnancy Nat Med 201016804ndash8087 Ohara-Imaizumi M Kim H Yoshida M et al Serotonin regulates glucose-stimulated insulin secretion from pancreatic b cells during pregnancy Proc NatlAcad Sci U S A 201311019420ndash194258 Ohta Y Kosaka Y Kishimoto N et al Convergence of the insulin and se-rotonin programs in the pancreatic b-cell Diabetes 2011603208ndash32169 Herrera PL Orci L Vassalli JD Two transgenic approaches to define the celllineages in endocrine pancreas development Mol Cell Endocrinol 199814045ndash5010 Herrera PL Adult insulin- and glucagon-producing cells differentiate fromtwo independent cell lineages Development 20001272317ndash232211 Cui Y Riedlinger G Miyoshi K et al Inactivation of Stat5 in mouse mammaryepithelium during pregnancy reveals distinct functions in cell proliferation survivaland differentiation Mol Cell Biol 2004248037ndash804712 Kim H Kim YG Choi W et al Generation of a highly efficient and tissue-specific tryptophan hydroxylase 1 knockout mouse model Sci Rep 201881764213 Yadav VK Ryu JH Suda N et al Lrp5 controls bone formation by inhibitingserotonin synthesis in the duodenum Cell 2008135825ndash83714 Banerjee RR Cyphert HA Walker EM et al Gestational diabetes mellitus frominactivation of prolactin receptor and MafB in islet b-cells Diabetes 2016652331ndash234115 Li L Gao L Wang K et al Knockin of Cre gene at Ins2 locus reveals no Creactivity in mouse hypothalamic neurons Sci Rep 201662043816 Wu Y Liu C Sun H et al Growth hormone receptor regulates b cell hy-perplasia and glucose-stimulated insulin secretion in obese mice J Clin Invest20111212422ndash242617 Brouwers B de Faudeur G Osipovich AB et al Impaired islet function incommonly used transgenic mouse lines due to human growth hormone minigeneexpression Cell Metab 201420979ndash99018 Kim H Kim H Kim K GermanMS Kim H Ectopic serotonin production inb-cellspecific transgenic mice Biochem Biophys Res Commun 20184951986ndash199119 Szot GL Koudria P Bluestone JA Murine pancreatic islet isolation J Vis Exp2007(7)25520 Blodgett DM Nowosielska A Afik S et al Novel observations from next-generation RNA sequencing of highly purified human adult and fetal islet cellsubsets Diabetes 2015643172ndash318121 Kim K Oh CM Ohara-Imaizumi M et al Functional role of serotonin in insulinsecretion in a diet-induced insulin-resistant state Endocrinology 2015156444ndash45222 Schraenen A Lemaire K de Faudeur G et al Placental lactogens induceserotonin biosynthesis in a subset of mouse beta cells during pregnancy Dia-betologia 2010532589ndash259923 Freemark M Ontogenesis of prolactin receptors in the human fetus roles infetal development Biochem Soc Trans 20012938ndash4124 Houmlglund E Mattsson G Tyrberg B Andersson A Carlsson C Growth hormoneincreases beta-cell proliferation in transplanted human and fetal rat islets JOP200910242ndash248

25 Bennet H Balhuizen A Medina A et al Altered serotonin (5-HT) 1D and 2Areceptor expression may contribute to defective insulin and glucagon secretion inhuman type 2 diabetes Peptides 201571113ndash12026 Almaccedila J Molina J Menegaz D et al Human beta cells produce and releaseserotonin to inhibit glucagon secretion from alpha cells Cell Rep 2016173281ndash329127 Kim YG Moon JH Kim K et al b-cell serotonin production is associated withfemale sex old age and diabetes-free condition Biochem Biophys Res Commun20174931197ndash120328 Goyvaerts L Schraenen A Schuit F Serotonin competence of mouse betacells during pregnancy Diabetologia 2016591356ndash136329 Chen H Kleinberger JW Takane KK et al Augmented Stat5 signalingbypasses multiple impediments to lactogen-mediated proliferation in hu-man b-cells Diabetes 2015643784ndash379730 Qiu WL Zhang YW Feng Y Li LC Yang L Xu CR Deciphering pancreatic isletb cell and a cell maturation pathways and characteristic features at the single-cell level Cell Metab 20182770231 Kaplan SL Grumbach MM Shepard TH The ontogenesis of human fetalhormones I Growth hormone and insulin J Clin Invest 1972513080ndash309332 Hawkes CP Grimberg A Measuring growth hormone and insulin-like growthfactor-I in infants what is normal Pediatr Endocrinol Rev 201311126ndash14633 Zadik Z Chalew SA McCarter RJ Jr Meistas M Kowarski AA The influenceof age on the 24-hour integrated concentration of growth hormone in normalindividuals J Clin Endocrinol Metab 198560513ndash51634 Galsgaard ED Friedrichsen BN Nielsen JH Moslashldrup A Expression ofdominant-negative STAT5 inhibits growth hormone- and prolactin-induced pro-liferation of insulin-producing cells Diabetes 200150(Suppl 1)S40ndashS4135 Jackerott M Moslashldrup A Thams P et al STAT5 activity in pancreatic beta-cells influences the severity of diabetes in animal models of type 1 and 2 diabetesDiabetes 2006552705ndash271236 Friedrichsen BN Richter HE Hansen JA et al Signal transducer and activatorof transcription 5 activation is sufficient to drive transcriptional induction of cyclinD2 gene and proliferation of rat pancreatic beta-cells Mol Endocrinol 200317945ndash95837 Liu JL Coschigano KT Robertson K et al Disruption of growth hormonereceptor gene causes diminished pancreatic islet size and increased insulinsensitivity in mice Am J Physiol Endocrinol Metab 2004287E405ndashE41338 Cutfield WS Wilton P Bennmarker H et al Incidence of diabetes mellitus andimpaired glucose tolerance in children and adolescents receiving growth-hormonetreatment Lancet 2000355610ndash61339 Child CJ Zimmermann AG Scott RS Cutler GB Jr Battelino T Blum WFGeNeSIS International Advisory Board Prevalence and incidence ofdiabetes mellitus in GH-treated children and adolescents analysis fromthe GeNeSIS observational research program [published correction appearsin J Clin Endocrinol Metab 2011962624] J Clin Endocrinol Metab 201196E1025ndashE103440 Sandler S Andersson A Korsgren O et al Tissue culture of human fetalpancreas growth hormone stimulates the formation and insulin production of islet-like cell clusters J Clin Endocrinol Metab 1987651154ndash1158

214 Serotonin Regulates Perinatal b-Cell Proliferation Diabetes Volume 69 February 2020