The islet-expressed Lhx1 transcription factor interacts with Islet-1 and contributes to glucose homeostasis

Abstract

The LIM-homeodomain (LIM-HD) transcription factor Islet-1 (Isl1) interacts with the LIM domain-binding protein 1 (Ldb1) coregulator to control expression of key pancreatic β-cell genes. However, Ldb1 also has Isl1-independent effects, supporting that another LIM-HD factor interacts with Ldb1 to impact β-cell development and/or function. LIM homeobox 1 (Lhx1) is an Isl1-related LIM-HD transcription factor that appears to be expressed in the developing mouse pancreas and in adult islets. However, roles for this factor in the pancreas are unknown. This study aimed to determine Lhx1 interactions and elucidate gene regulatory and physiological roles in the pancreas. Co-immunoprecipitation using β-cell extracts demonstrated an interaction between Lhx1 and Isl1, and thus we hypothesized that Lhx1 and Isl1 regulate similar target genes. To test this, we employed siRNA-mediated Lhx1 knockdown in β-cell lines and discovered reduced Glp1R mRNA. Chromatin immunoprecipitation revealed Lhx1 occupancy at a domain also known to be occupied by Isl1 and Ldb1. Through development of a pancreas-wide knockout mouse model (Lhx1^∆Panc), we demonstrate that aged Lhx1^∆Panc mice have elevated fasting blood glucose levels, altered intraperitoneal and oral glucose tolerance, and significantly upregulated glucagon, somatostatin, pancreatic polypeptide, MafB, and Arx islet mRNAs. Additionally, Lhx1^∆Panc mice exhibit significantly reduced Glp1R, an mRNA encoding the insulinotropic receptor for glucagon-like peptide 1 along with a concomitant dampened Glp1 response and mild glucose intolerance in mice challenged with oral glucose. These data are the first to reveal that the Lhx1 transcription factor contributes to normal glucose homeostasis and Glp1 responses.

INTRODUCTION

Diabetes mellitus (DM) results from a loss of functional insulin-producing pancreatic β-cells, via autoimmune destruction (type 1), or from β-cell dysfunction and failure (e.g., loss of cell identity and dedifferentiation, type 2) (19). DM has rapidly become a worldwide epidemic. Currently, there is no cure, with afflicted individuals requiring careful blood glucose monitoring along with anti-diabetic medications, which often includes exogenous insulin administration. Future treatments aim to restore functional β-cell mass, for example, through β-cell replacement strategies (19, 21, 41). Understanding β-cell development and function is crucial for the success of such therapies.

Transcription factors are critical effectors of pancreatic organogenesis and islet function (9, 40, 42, 49). In humans, maturity onset diabetes of the young (MODY), an autosomal dominant form of DM, illustrates the importance of transcription factors in β-cell function, as several genetic loci associated with MODY encode transcription factors (57). Numerous loss- and gain-of-function studies in mice have also demonstrated the critical nature of transcription factors during islet cell development and function. For example, global loss or disruption of Pdx1, a transcription factor gene expressed in embryonic multipotent pancreatic progenitors and then enriched in mature β-cells, results in pancreatic agenesis in mice and humans (22, 39, 51). Additionally, Neurogenin3 (Ngn3) null mice lack developing islet endocrine cells, yet appear to have a normally developed exocrine pancreas (15, 20). Other transcription factor deficiency models instead present with postnatal β-cell function defects. For example, MafA, is required for normal β-cell mass and glucose stimulated insulin secretion (16, 58).

Islet-1 (Isl1) is a LIM- (named after protein interaction domains identified in Lin11, Isl1, Mec3) homeodomain (HD) transcription factor expressed in the pancreatic epithelium and surrounding mesenchyme at about embryonic day (E)9.5 but is most notably maintained later in all adult islet cells (2, 11, 12, 54). Previous conditional knockout studies (using a Pdx1-Cre) demonstrated that Isl1 is required for the maturation, proliferation, and survival of the pancreatic endocrine cells. Additionally, Isl1-deficient mice had reduced insulin⁺, somatostatin⁺, and glucagon⁺ cells along with severe postnatal hyperglycemia (11). Key islet mRNAs such as MafA, Glp1R, and Arx were significantly reduced in Isl1-deficient mice (11).

LIM domain factors, like Isl1, interact with the LIM domain-binding protein 1 co-regulator (Ldb1) via the LIM-interaction domain (LID) to impact gene regulation in various tissues (1, 5, 23, 37). Ldb1 does not bind DNA directly but instead serves as a scaffold for LIM transcription factors to enable assembly of complexes at promoters/enhancers (33, 37). In prior reports, mice lacking pancreatic islet Ldb1 (using Pax6-Cre) had progressive postnatal hyperglycemia, reduced hormone⁺ cells, and impaired MafA, Glp1R, Arx, and Glut2 mRNA and protein levels (18). Additionally, Ldb1 and Isl1 were found to occupy 5′ control regions of MafA and Glp1R via chromatin immunoprecipitation (ChIP) (13, 18), suggesting a direct regulation of these gene targets. Strikingly, the Ldb1-deficient mice also exhibited phenotypes not observed in the Isl1 mouse model, for example, a reduction of Hb9/Mnx1, encoding a critical β-cell transcription factor (18). This suggested that 1) Ldb1 possesses Isl1-independent transcriptional activity and 2) other pancreatic LIM factors may interact with Ldb1 to regulate β-cell development and/or function.

Prior reports support that LIM homeobox 1 (Lhx1 or Lim1), a LIM-HD transcription factor related to Isl1, appears to be expressed in the developing and adult pancreas, albeit at a lower level than Isl1 (13, 18). Global methylation studies in sorted mouse β-cells and islets demonstrated that the Lhx1 locus was positive for active histone 3 lysine 4 trimethylation (H3K4me3) marks and negative for the repressive H3K27me3 mark, similar to Pdx1 (56). The enrichment pattern of these histone marks is indicative an active Lhx1 locus and highly supportive of β-cell Lhx1 expression. Lhx1 is required for head, nervous system, and kidney development in multiple species (e.g., mouse and zebrafish) (25, 28, 47, 52), with global Lhx1-null embryos lacking anterior head structures and having disrupted trunk and tail development (47). A head truncation phenotype was also observed in global Ldb1-null mice (38), implying a physical and functional interaction between Ldb1 and Lhx1 in co-expressing tissues. However, Lhx1 has not been characterized in the pancreas. Because of apparent Lhx1 expression and pancreatic Isl1-independent Ldb1 roles, we hypothesized that Lhx1 regulates islet development and function via interactions with Ldb1.

In this study, we demonstrate that Lhx1 is indeed expressed in the developing and adult pancreas. We observed that Lhx1 interacts with Ldb1 and, surprisingly, Isl1 and regulates β-cell targets, including Glp1R. To gain insight into in vivo function, we conditionally removed Lhx1 from the mouse pancreas. Aged Lhx1-deficient mice exhibited mild glucose intolerance, impaired insulin secretion, and a significant elevation of islet α-cell mRNAs, including MafB, Arx, and glucagon. Loss of Lhx1 in the mouse pancreas also led to a significant reduction of Glp1R mRNA, resulting in a dampened β-cell response to the incretin glucagon like peptide 1 (Glp1). Taken together, these results support that Lhx1 participates in complexes with Ldb1 and Isl1 and is a novel regulator of islet cell function.

MATERIALS AND METHODS

Animal models.

Lhx1^LoxP (29), Lhx1^LacZ (48), and Pdx1-Cre (17, 35) mouse lines have been described previously; mice were maintained on a C57BL/6 background. The Lhx1^LoxP mice were crossed into the Pdx1-Cre line to create a pancreas-wide Lhx1 deletion (Lhx1^∆Panc). Male Lhx1^∆Panc and control (Lhx1^F/F and/or Lhx1^F/+) littermates (indicated as CTL) were housed up to five per cage under a light-dark cycle of 12 h, with ad libitum access to food and water, except when noted. We initially examined Cre⁺ controls (i.e., Lhx1^F/+; Pdx1-Cre) and found no effect on glucose tolerance under these conditions (data not shown), and therefore, we used Lhx1^F/F and/or Lhx1^F/+ littermates as controls moving forward. The University of Alabama Institutional Animal Care and Use Committee approved all animal experiments.

X-gal staining.

X-gal staining for β-gal activity was performed as described (48) with minor modifications. Embryos were collected at E14.5 and briefly fixed for 15 min in 4% paraformaldehyde (PFA) in PBS at 4°C. Tissues were then washed in X-gal rinse solution (PBS containing 0.02% NP-40 and 0.01% deoxycholate) three times for 15 min each. The wash solution was removed, and tissues were stained with X-gal: 5 mM K₃Fe(CN)₆, 5 mM K₄Fe(CN)₆, 0.02% NP-40, 0.01% deoxycholate, 2 mM MgCl₂, 5 mM EGTA, and 2 mg/ml X-gal solution dissolved in DMSO in PBS. Tissues were incubated at 37°C for 12–18 h until a blue signal was observed. Tissues were washed three times in PBS and then postfixed in 2% PFA-PBS for 20 min. Fixed tissues were cleared in glycerol-PBS (1:1 ratio) overnight at 4°C and photographed.

Immunofluorescence and immunohistochemistry analyses.

Pancreata were dissected from adult mice, fixed overnight in 4% formaldehyde diluted in 1× PBS at 4°C, and then embedded in paraffin or optimal cutting temperature (OCT; Tissue-Tek 4583). Sections were cut to 6 (paraffin) or 10 µm (OCT) and then blocked using 5% normal donkey serum in 1% bovine serum albumin in 1× PBS for 1 h at room temperature. Sections were then incubated with primary antibodies overnight at 4°C: goat α-Ldb1, 1:1,000 (sc-11198X; Santa Cruz Biotechnology); mouse α-Isl1, 1:1,000 (DSHB, 39.4D5; Developmental Studies Hybridoma Bank); rabbit α-Lhx1, 1:500 (Novus NB 110-12933); guinea pig α-insulin, 1:1,000 (Dako 0564); mouse α-glucagon, 1:4,000 (G2654; Sigma). Cy-2-, Cy-3-, and Cy-5-conjugated α-rabbit, α-goat, α-guinea pig, or α-mouse IgG secondary antibodies 1:500 (Jackson ImmunoResearch Laboratories) were used to detect indirect immunofluorescence. For immunohistochemistry, signal was detected by using a Vectastain Elite ABC Kit and a DAB Peroxidase Substrate Kit (SK-4100; Vector). Tissues were then subjected to eosin counterstaining. Slides were imaged using an Olympus IX81 fluorescence or Zeiss LSM710 confocal microscope, and the images were processed by CellSens Dimension version 1.12 (Olympus) or Zen software (Zeiss).

Quantitative RT-PCR.

RNA was collected from collagenase-digested and hand-picked mouse islets (14, 18) or βTC3 and Min6 cells using the RNeasy Mini Plus kit (no. 74134; Qiagen, Valencia, CA) and cDNA was synthesized using the iScript cDNA Synthesis kit (no. 170-8840; Bio-Rad, Hercules, CA). Single gene quantitative (q)PCR was performed using iTaq SYBR Green (no. 172-5124; Bio-Rad, Hercules, CA) in duplicate using the gene primers listed in Table 1. qPCR reactions were amplified using a LightCycler 480 II (Roche), normalized to Gapdh and analyzed using the 2^ΔΔCt method. Primer sequences are provided in Table 1.

Table 1.

Quantitative RT-PCR primer pairs for mRNA analysis

Target	Forward	Reverse
Lhx1	GACCTACCCTTTGTGCCATCA	CCACCATTGACCGACAGAGAT
Isl1	GCAACCCAACGACAAAACTAA	CCATCATGTCTCTCCGGACT
MafA	CCTGTAGAGGAAGCCGAGGAA	CCTCCCCCAGTCGAGTATAGC
Pdx1	CGGCTGAGCAAGCTAAGGTT	TGGAAGAAGCGCTCTCTTTGA
Ldb1	ACTCATGTGGATGCCTGTGTG	CCCCAACATTTAGCCCCTAAG
Glut2	CAGTTCGGCTATGACATCGGT	GTTAATGGCAGCTTTCCGGTC
Kcnj11	TGTGCAGAATATCGTCGGGCTGAT	GCATGCTTGCTGAAGATGAGGGTT
Glp1R	TGAACCTGTTTGCATCCTTCA	ACTTGGCAAGCCTGCATTTGA
Ins1	CCAGCTATAATCAGAGACCA	CCAGGTGGGGACCACAAAGA
Ins2	GGCTTCTTCTACACACCCAT	CCAAGGTCTGAAGGTCACCT
Sst	AACGCAAAGCTGGCTGCAAGAA	TCAGAGGTCTGGCTAGGACAACAA
Gcg	TTAGCGCAGACAGAGCTACCGAAA	ATACTCTTTACACTCCCACCCTCG
PP	TTGCAGCCTCTCTTGTCTTCA	TAGTTTGCAAGGGAGCAGGTT
Nkx6.1	CCTCTGGACCCGAACTCTGA	GCTGCCACCGCTCGATT
Arx	TCCGGATACCCCACTTAGCTT	GACGCCCCTTTCCTTTAAGTG
Gck	CTGTTAGCAGGATGGCAGCTT	TTTCCTGGAGAGATGCTGTGG
Sur1	TCAACTTGTCTGGTGGTCAGC	GAGCTGAGAAAGGGTCATCCA
GcgR	ACATCATTCACCTTCTTGTGG	GCCAGCGAGGTCTCCATA
Gapdh	CCTGGAGAAACCTGCCAAGTA	TGGAAGAGTGGGAGTTGCTGT

Glp1R, glucagon-like peptide 1 receptor; Glut2, glucose transporter 2; Ldb1, LIM domain-binding protein 1 co-regulator; Lhx1, LIM homeobox 1.

Co-Immunoprecipitation and Western blotting.

Antibodies raised against Ldb1 (sc-11198X; Santa Cruz Biotechnology), Isl1 (39.4D5; DSHB), Lhx1 (AB3200; Millipore), rabbit, and goat IgG (Bethyl Laboratories) were covalently bound to Protein-G Dynabeads (10004D; Life Technologies) using 20 mM dimethyl pimelimidate crosslinker (21667; Life Technologies) and then incubated with βTC3, Min6 nuclear extract (45), or WT mouse islet whole cell extract diluted with a PBS-protease inhibitor cocktail (PIC; Thermo Scientific) for 3 h at 4°C. The beads were washed five times with PBS-PIC, and the bound proteins were eluted using radioimmunoprecipitation assay (RIPA) buffer [50 mmol/l Tris (pH 7.7), 150 mmol/l NaCl, 1% Nonidet P-40, 0.5% deoxycholic acid, and 0.1% SDS] at 37°C for 15 min. Elutions were separated on a 10% SDS-PAGE (Bio-Rad) gel and then transferred to a polyvinylidene fluoride (no. 162-077; PVDF Bio-Rad) membrane. The membrane was blocked using 5% nonfat dry milk in 1% PBS-Tween for 1 h at room temperature, followed by overnight 4°C incubation with α-Lhx1 (sc-515631, 1:1,000; Santa Cruz Biotechnology or 4F2, 1:1,000; DSHB), α-Isl1 (39.4D5, 1:1,000; DSHB), and α-Ldb1 (sc-11198X,1:1,000; Santa Cruz Biotechnology). The membrane was washed three times using PBS-Tween and incubated with species-matched HRP-conjugated secondary antibodies (α-rabbit, Promega; α-mouse and α-goat, Santa Cruz Biotechnology), followed by detection using Luminata Forte substrate (Millipore, Billerica, MA), and imaged using Chemidoc XRS+ Imager (Bio-Rad).

siRNA knockdown.

Min6 cells were seeded in 1 × 10⁶ cells/well in six-well tissue culture plates (Corning) and then transfected with 50 nM On-Target Plus Smart Pool Lhx1 targeted (siLhx1, L-042643-01; GE/Dharmacon) or nontargeted scrambled control (siScr, D-001810-10-05; GE/Dharmacon) using RNAiMAX (13778030; Life Technologies). mRNA analyses were performed 48 h after transfection. Each experiment was performed in duplicate on at least three independent occasions.

Chromatin immunoprecipitation.

Chromatin immunoprecipitation (ChIP) assays were performed as described (14, 18). Briefly, βTC3 cells were cross-linked for 10 min in 1% formaldehyde diluted in 1× PBS. Chromatin was precleared using Protein-G Dynabeads (10004D; Invitrogen,) for 2 h at 4°C and then incubated with α-Lhx1 (AB3200; Millipore) or rabbit IgG (P120-101; Bethyl Laboratories) at 4°C overnight. Bound complexes were then eluted, and cross-links were reversed. qPCR was performed on the Lhx1 or IgG-immunoprecipitated DNA using SYBR Green PCR master mix (Bio-Rad) on MafA and Glp1R regulatory sequences (14, 18). The enrichment of target control sequences in the ChIP DNAs was normalized to inactive albumin control sequences and then calculated relative to rabbit IgG enrichment set as onefold.

Glucose and insulin tolerance testing.

Male mice (age matched over the study time course: ∼16–23 wk) were fasted ≤16 h, and glucose (2 g or 2.5g/kg body wt) was administered by oral gavage or via intraperitoneal (ip) injection. For ITT, regular human insulin [0.50 U/kg (Lilly Humalog) in saline] was administered intraperitoneally. Blood was drawn from the tail vein at 0, 15, 30, 60, and 120 min after administration, and glucose levels measured using a Breeze2 glucometer (Bayer HealthCare). During ITT, glucose disappearance rate (k_g) was defined as Δblood glucose per minute.

β-Cell mass.

To quantify staining, slides were digitally scanned and analyzed with Olympus CellSens Dimensions software. β-Cell mass was calculated by averaging the percentage of insulin-stained tissue area over three sections that were taken at 100 μm apart. The fraction of insulin/eosin-stained tissue was then multiplied by the wet pancreas mass measured at tissue harvest.

Synthesis of a novel glucagon-like peptide 1 receptor agonist (IUB282).

Boc Thr (Bzl; 0.2 mmol) was coupled to an MBHA resin and served as the initial synthetic substrate upon which the target peptide was assembled through repetitive DEPBT/diisopropylethylamine (DIEA)-activated single couplings on a CSBio336 peptide synthesizer (10). The assembled peptide resin was treated with 20% piperidine/dimethylformamide (DMF) at room temperature for 10 min to remove the side chain Fmoc protection. The resin was filtered and washed with DMF, and an activated solution of Fmoc Glu-αOBzl was added to acylate the Lys10 side chain [previously prepared by dissolving FmocGlu-αBzl (Aapptec) in 0.5 mol/l DEPBT/DMF and subsequently adding DIEA]. After 1 h, the resin was filtered and washed with DMF. The above process was repeated using an equivalent aliquot of Fmoc Glu-αOBzl and once more with 2.0 mmol palmitic acid (Sigma, St. Louis, MO). The completed peptide resin was filtered, washed, treated with 50% trifluoroacetic acid-dichloromethane, and neutralized with 5% DIEA-dichloromethane. The peptide was cleaved from the resin in liquid hydrogen fluid in the presence of p-cresol at 0°C for 1 h. The peptide was solubilized in 50% aqueous acetonitrile, diluted with dilute ammonium bicarbonate, and purified by Amberchrom XT20 chromatography with a linear gradient of aqueous acetonitrile in 0.025 mol/l ammonium bicarbonate. The pure peptide was identified by analytical high-performance liquid chromatography and electrospray ionization mass spectral analysis. The sequence is HXEGTFTSDZSKYLDXRAAQDFVQWLMDT-amide, where X = AIB (amino, isobutyric acid) and Z = K-γEγEC16 (a γ-glutamic acid dipeptide with a C16-saturated fatty acid acylated to the epsilon amine of lysine).

In vitro receptor-mediated cAMP production.

IUB282 potency at the Glp1 receptor was determined by the ability of the peptide to induce cAMP and was measured in a firefly luciferase-based reporter gene assay (Fig. 5E). In brief, human embryonic kidney-293 cells co-transfected with the GLP-1 receptor and the luciferase gene linked to the cAMP-responsive element were serum deprived by culturing for 16 h in Dulbecco’s modified Eagle’s medium (Fisher Scientific, Carlsbad, CA) supplemented with 0.25% bovine growth serum (HyClone, Logan, UT) and then incubated with serial dilutions of either glucagon, GLP-1, or IUB282 for 5 h at 37°C, 5% CO₂, and 90% humidity in 96-well poly-d-lysine-coated Biocoat plates (BD Biosciences, San Jose, CA). At the end of the incubation, 100 µl of Steady-Lite luminescence substrate reagent (PerkinElmer, Wellesley, MA) was added to each well. The plate was shaken briefly and incubated for 10 min in the dark, and luminescence signal was measured on an Ensight multimode plate reader (PerkinElmer). Signal versus concentration plots were generated, and EC₅₀ values were calculated by using three-parameter regression in GraphPad Prism 7 software (GraphPad, La Jolla, CA).

Fig. 5.

LIM homeobox 1 (Lhx1) contributes to the β-cell incretin effect and insulin secretion. A: response to an oral glucose load (2.5 g/kg) between control and Lhx1^∆Panc mice. B: assessment of insulin sensitivity of Lhx1^∆Panc and control mice by insulin tolerance test (0.50 U/kg) at 15 wk of age. There was no significant difference across genotypes at any time point or in the rate of change in glucose levels from time 0 to the nadir of 30 min (k_g30; bottom inset). Plasma glucagon levels were measured by ELISA at time 0 and 30 min during the insulin tolerance test (ITT), with a significant elevation in glucagon observed at 30 min (top inset). C: intraperitoneal glucose tolerance test modified with an injection of exendin-4 (Ex-4; 10 ng·mol⁻¹·kg⁻¹ ip at −10 min) or PBS using age- and sex-matched Lhx1^∆Panc and control mice at 23 wk. Glucose (2.5 g/kg ip) was injected at 0 min, and blood glucose was measured at the designated time points (2-way ANOVA). *P < 0.05 comparing the two Ex-4 treated genotypes; **P < 0.01 comparing Ex-4 vs. PBS-treated CTL mice. D: plasma glucagon-like peptide 1 (Glp1) was quantified from control and Lhx1^∆Panc mice at 0 and 15 min after a mixed meal challenge (2.0 g/kg body wt of Ensure) at 15 wk. No significant difference was observed. E: to characterize the novel glucagon-like peptide 1 receptor (Glp1R) agonist IUB282, signal vs. concentration plots of luminescence signal and EC₅₀ values were calculated by using 3-parameter regression. Human embryonic kidney-293 cells co-transfected with the Glp1 receptor and luciferase gene linked to the cAMP-responsive element and then incubated with serial dilutions of glucagon, Glp1, or IUB282. F: glucose tolerance tests were conducted after injection of the Glp1r agonist IUB282 (10 ng·mol⁻¹·kg⁻¹ ip at −10 min) or PBS using age-matched Lhx1^∆Panc and control mice at 22 wk. Glucose (2.5 g/kg ip) was injected at 0 min, and blood glucose was measured at the designated time points. #P < 0.05, comparing the 2 IUB282-injected groups (dashed lines) at 15 min; **P < 0.01 comparing the PBS- or IUB282-treated control (CTL) groups at 15 min; ****P < 0.0001 comparing the PBS- or IUB282-treated Lhx1^∆Panc mice at 30 and 60 min. G: insulin secretion (expressed as fold over basal) from islets isolated from 20-wk-old Lhx1^∆Panc and control mice in response to low glucose (3 mM), 11 mM stimulatory glucose with or without Ex-4, and high glucose (16 mM). H: insulin content was quantified from acid-ethanol extracted islet fractions in the 11 mM + Ex-4 glucose-stimulated insulin secretion (GSIS) groups. No significant differences were noted. I: in vivo GSIS performed with Lhx1^∆Panc and control mice at 17 wk. Plasma was collected at 5 and 15 min post-glucose injection (t = 0), and insulin was quantified by ELISA. All data are shown as means ± SE; n = 3–7 mice/group. NS, not significant. In G and I: *P < 0.05; **P < 0.01; ***P < 0.001; ****P < 0.0001.

Treatment with Glp1R agonists.

Lhx1^∆Panc and littermate controls were treated for 10 min before intraperitoneal glucose tolerance test (IPGTT) with vehicle (PBS pH 7.4, no. 10010-023; Life Technologies), IUB282 Glp1R agonist (10 nm/kg body weight), or exendin-4 (Ex-4; 10 nm/kg body wt, no. 5-42489; Chi Scientific).

Static insulin secretion.

Intact control and Lhx1^∆Panc mouse islets were isolated using collagenase digestion (43). Static insulin secretion assays were performed on preparations consisting of three islets incubated with various secretagogues (n = 4/genotype) (7). Basal-normalized islet insulin secretion was determined by enzyme-linked immunosorbent assay (ELISA).

In vivo insulin secretion.

Fasted control and Lhx1^∆Panc mice were subject to an ip glucose challenge (2.5g/kg body wt) after baseline plasma was isolated at time = 0. At 5 and 15 min after injection, plasma was isolated from tail vein blood and subjected to insulin ELISA (see below).

Plasma hormone analyses.

Plasma insulin (no. 90800; Crystal Chem), glucagon (no. 10-1281-0; Mercodia), or Glp-1 (no. 80-GLP1A-CH01; Alpco) was quantified by ELISA according to the manufacturer’s instructions. Analytical variation of assays was assumed as reported by the manufacturer.

Statistical analysis.

All data are represented as means ± SE. Statistical significance was assessed using either Student’s t-test or a two-way ANOVA, followed by post hoc analysis using GraphPad Prism version 7.0 statistical software. P < 0.05 was considered significant.

RESULTS

Lhx1 is expressed in β-cell lines and in the developing and adult mouse pancreas.

To characterize pancreatic Lhx1 expression, we employed the use of Lhx1 knock-in LacZ reporter (24) mouse pancreata for histological analysis. LacZ reporter staining of E14.5 whole embryos demonstrated Lhx1 mRNA expression in the central nervous system, as expected (34, 59) (see whole mount stained embryos Fig. 1A, inset). In tissue sections, LacZ reporter signal was also observed within the developing pancreatic epithelium (Fig. 1A) in a pattern appearing similar to developing endocrine cells (i.e., trunk domain; see Ref. 49). Immunostaining was also performed with wild-type adult mouse pancreas to determine the spatial Lhx1 protein expression. In tissue from adult (5-mo-old) mice, Lhx1 was expressed in insulin⁺ cells (Fig. 1B). Although Lhx1 appeared primarily in insulin⁺ cells, there were a few Lhx1⁺/insulin⁻ cells within the islet. However, for technical reasons, we could not identify, for example, the Lhx1⁺/Ins⁻ cells as α- or δ-cells. Notably, we also observed nuclear Lhx1 co-expression with Isl1 and Ldb1 in adult mouse islets (Fig. 1C). Using cDNA isolated from rodent β-cell lines and mouse islets, we compared Lhx1 and Isl1 expression. Lhx1 mRNA was expressed similarly to Isl1 in βTC3 and Min6 cell lines, but at a lower level in mouse islets, as observed prior (Fig. 1D) (13, 18). Similar trends in expression were observed for Lhx1 and Isl1 protein levels (Fig. 1D).

Fig. 1.

LIM homeobox 1 (Lhx1) is expressed in the pancreas. A: embryonic day (E)14.5 Lhx1-LacZ reporter mouse pancreas tissue staining demonstrates an active Lhx1 locus in the developing pancreas. Tissues were stained with X-gal (blue) and eosin (pink). Bottom: ×40 magnification of E14.5 Lhx1-LacZ reporter activity in mouse pancreas, highlighted by black arrowheads B: immunofluorescence from 5-mo-old adult mouse pancreas tissue demonstrating Lhx1 (red) in insulin (green) co-expressing cells. C: co-immunofluorescence with 5-mo-old wild-type mouse pancreata indicates that Lhx1 (blue), Islet-1 (Isl1; red), and LIM domain-binding protein 1 co-regulator (Ldb1; green) are co-expressed in islets. D: Lhx1 and Isl1 expression relative to Gapdh was assessed with βTC3, Min6, and adult mouse islet cDNA (top); Western blotting analysis of relative Lhx1 protein levels in βTC3, Min6, and adult mouse islet nuclear extracts as compared with Actin loading control (bottom). Arrowhead points to the Lhx1-specific signal, based on apparent size. Data are representative of at least 3 independent experiments. Pa, pancreas; St, stomach. ***P < 0.001.

Lhx1 interacts with Ldb1 and Isl1 in β-cell lines and islets.

The observed co-expression of Lhx1, Isl1, and Ldb1 in adult mouse islets (Fig. 1C) led us to investigate whether these factors interact in β-cells. To test for endogenous Lhx1 interactions, we conducted co-immunoprecipitation (co-IP) experiments using βTC3, Min6, and islet extracts. βTC3 IP of Ldb1, or Isl1 recovered Lhx1 protein as compared with species-matched IgG control IP (Fig. 2A, top). Reciprocal co-IPs with Lhx1-specific antibodies or control IgG confirmed the endogenous interaction with Ldb1 and Isl1 (Fig. 2A, bottom), and similar results were observed with Min6 cell extracts (data not shown). Co-IPs with primary wild-type mouse islets confirmed endogenous Lhx1 interaction with Isl1 (Fig. 2B), whereas a preliminary result revealed Lhx1 and Ldb1 interacting in human islets (data not shown). However, the primary mouse and human co-IPs studies were conducted using extracts prepared from whole islets, and thus we cannot conclude that Lhx1 interactions with Ldb1 or Isl1 are occurring only in β-cells (Fig. 2B). Additionally, Lhx1 IP did not recover other islet-enriched transcription factors from βTC3 cells, including Pdx1, MafA, Nkx6.1, and Pax6 (Fig. 2C), suggesting specificity of the Lhx1 interaction with Ldb1 and Isl1.

Fig. 2.

LIM homeobox 1 (Lhx1) interacts with LIM domain-binding protein 1 co-regulator (Ldb1) and Islet-1 (Isl1) in β-cell lines. A: co-immunoprecipitation (IP) using βTC3 nuclear extracts. Endogenous Ldb1 (top left) or Isl1 (top right) IP recovers Lhx1 as compared with input positive control and species-matched IgG negative control. Reciprocal Lhx1 IP is also enriched for Ldb1 (bottom left) and Isl1 (bottom right); n = 3. B: co-IP using primary mouse islet extracts (n = 3). Endogenous Isl1 IP recovers Lhx1. C: Lhx1 does not enrich for Pdx1, MafA, Nkx6.1, or Pax6 in βTC3 cells, suggesting binding specificity with Ldb1 and Isl1 (n = 3). WB, Western blot.

Glp1R expression is regulated by Lhx1 in β-cell lines.

The interaction between Lhx1, Ldb1, and Isl1 led us to assess whether Lhx1 is localized to Isl1 and Ldb1-bound Glp1R or MafA control regions in β-cells (13, 18). Using a βTC3 chromatin immunoprecipitation (ChIP), we observed significant Lhx1 enrichment for known Ldb1-Isl1-bound 5′ regulatory regions of the MafA (i.e., region 3) and Glp1R loci (Fig. 3, A and B).

Fig. 3.

LIM homeobox 1 (Lhx1) directly regulates glucagon-like peptide 1 receptor (Glp1R) in β-cell lines. A: schematic of conserved MafA 5′ control domains (left) and chromatin immunoprecipitation (ChIP) analysis of Lhx1 enriching for the MafA region 3 domain (right). B: schematic of Glp1R 5′ domains that specify known Islet-1 (Isl1)-bound ChIP peaks (left) and ChIP analysis demonstrating Lhx1 enrichment of the Glp1R 5′ region (right). Values are presented as fold enrichment over normal mouse IgG (set to 1-fold) after normalizing to the inactive albumin promoter. C: Western blot for Lhx1 or actin loading control using Min6 cell extracts transfected with a scrambled control siRNA (siSCR) or Lhx1 siRNA (siLhx1) (top); densitometry of Lhx1 protein band intensity normalized to actin (bottom). D: relative quantification of key β-cell mRNA levels after siLhx1 knockdown demonstrating significant reduction of Glp1R in Min6 β-cells. mRNA levels were normalized to Gapdh and compared with scrambled siRNA control. All data are shown as means ± SE; n = 4–8 independent experiments. *P < 0.05; **P < 0.01; ****P < 0.0001.

To test the importance of Lhx1 for β-cell gene expression, we conducted siRNA-mediated knockdown of Lhx1 in a β-cell line. Because of similar Lhx1 protein and mRNA levels as primary mouse islets, Min6 β-cells were transfected with siRNAs targeting Lhx1 or scrambled control to significantly reduce Lhx1 protein (Fig. 3C) and mRNA levels (Fig. 3D). Consequently, Lhx1 knockdown imparted significant reduction of Glp1R (Fig. 3D), yet we did not observe significant changes in expression of MafA (despite occupation by Lhx1, Fig. 3A) or other islet-enriched transcription factor mRNAs, including Pdx1 and Isl1 (Fig. 3A). This implies that Lhx1 is dispensable for MafA activation or that the remaining Lhx1 (i.e., 40%) in siLhx1-treated cells is sufficient for MafA expression. Collectively, these data support that Lhx1 is required for normal expression of the Ldb1/Isl1 target gene Glp1R.

Lhx1 deficiency causes mild glucose intolerance and altered islet gene expression.

Given our in vitro observations suggesting the expression, interactions, and roles for Lhx1, we next focused on determining in vivo Lhx1 roles in the mouse pancreas. To this end, we crossed transgenic Pdx1-Cre (17, 35) and Lhx1 floxed mice (29), creating a pancreas-wide Lhx1 deficiency model termed Lhx1^∆Panc (Fig. 4A). Mice were born at the expected Mendelian ratios, and Lhx1 was significantly reduced at the mRNA and protein levels in adult mouse islets (Fig. 4B). At postnatal day 1 (P1), Lhx1^∆Panc mice had no differences in body weight, pancreas mass index, or ad libitum blood glucose when compared with littermate controls (data not shown and Fig. 4C). Islet morphology and hormone immunoreactivity appeared similar between Lhx1^∆Panc and control neonatal tissue (Fig. 4D). P1 plasma insulin in ad libitum fed neonates was unchanged across the genotypes; however, plasma glucagon was significantly elevated in Lhx1^∆Panc mice (Fig. 4E).

Fig. 4.

In vivo LIM homeobox 1 (Lhx1) deficiency alters glucose tolerance and islet gene expression. A: schematic of LoxP sites flanking the Lhx1 locus (29). In the presence of Pdx1-driven Cre recombinase, the targeted allele is recombined in the pancreas, a model termed Lhx1^∆Panc. B: Lhx1 mRNA levels were significantly reduced in islets isolated from 17-wk-old Lhx1^∆Panc mice relative to Gapdh and compared with control (CTL) littermates (top); Western blotting demonstrates that Lhx1 protein is reduced in Lhx1^∆Panc islet extracts as compared with littermate controls (bottom). Actin serves as loading control. C: there was no significant (NS) change in P1 pancreas mass index (ratio of pancreas to body mass; top), or ad libitum blood glucose levels (bottom) compared with CTL littermates. D: postnatal day 1 (P1) tissue immunofluorescence of insulin (green), glucagon (red), and somatostatin (blue) indicate no overt deficiency in neonatal islet architecture or differentiation. E: P1 ad libitum plasma insulin levels were unchanged, whereas there was a significant increase in glucagon, as measured by ELISA. F: fasting blood glucose levels (FBG; left) and body weight (right) were not significantly different in Lhx1^∆Panc mice compared with controls at 16 wk; n = 12–13. G: in vivo response to an intraperitoneal (ip) glucose load (2 g/kg intraperitoneal glucose tolerance test) in Lhx1^∆Panc mice at 18 wk was significantly different from controls at the 30-, 60-, and 120-min time points (2-way ANOVA). H and I: confocal imaging of islet hormone immunofluorescence of 20-wk-old pancreas tissue from control (H) and 2 independent Lhx1^∆Panc mice (I), indicating somatostatin (red), insulin (green), and glucagon (blue) expression. J, top: β-cell mass was not significantly different in Lhx1^∆Panc mice from controls. J, bottom: representative insulin immunohistochemistry of CTL and Lhx1^∆Panc islets counterstained with eosin. K: islet hormone mRNA levels were quantified relative to Gapdh in Lhx1^∆Panc and control mice, indicating significantly elevated somatostatin, glucagon, and pancreatic polypeptide. Islet transcription factor mRNAs were quantified from CTL and Lhx1^∆Panc mouse islet cDNA, with a significant elevation in Arx and MafB. Glucose-stimulated insulin secretion (GSIS)-related mRNAs were also quantified relative to Gapdh, with a significant decrement in Glp1R mRNA and elevation in Glut2. All data are shown as means ± SE; n = 3–7 mice/group; *P < 0.05; **P < 0.01; ***P < 0.001; ****P < 0.0001.

Aged Lhx1^∆Panc mice had no difference in body weight or fasting blood glucose levels when compared with age-matched controls at 16 wk (Fig. 4F). An intraperitoneal glucose tolerance test (IPGTT) revealed that Lhx1^∆Panc mice were glucose intolerant (Fig. 4G), suggesting that Lhx1 may be required for glucose homeostasis. As described in materials and methods, an initial age-matched 16-wk-old cohort included assessment of IPGTT with Lhx1^∆Panc, floxed Lhx1, and Lhx1^F/+;Pdx1-Cre (i.e., Lhx1 heterozygous) mice. Indeed, the floxed control and Lhx1^F/+;Pdx1-Cre mice had a normal glucose excursion, whereas only the Lhx1^∆Panc mice were glucose intolerant (data not shown). This supports that the Cre recombinase does not impact glucose homeostasis in this model. As a consequence, we employed Lhx1^F/F and/or Lhx1^F/+ littermate controls throughout.

Qualitative islet hormone immunofluorescence analyses demonstrated that Lhx1^∆Panc islets appear to have greater glucagon⁺ and somatostatin⁺ immunoreactivity (Fig. 4, H and I), with somatostatin⁺ or glucagon⁺ cells often found throughout the islet core, as opposed to the typical mantle location (Fig. 4I). These overt islet changes led us to assess β-cell mass; however, no significant difference was exhibited in Lhx1^∆Panc islets when compared with control animals, although there was a trending reduction (Fig. 4J). Strikingly, the observed increased glucagon⁺ and somatostatin⁺ immunoreactivity in Lhx1^∆Panc pancreata (Fig. 4, H and I) was supported by a significant increase in islet-expressed glucagon, somatostatin, and pancreatic polypeptide mRNA, whereas there was no significant impact on insulin1 or insulin2 (Fig. 4K).

To gain additional mechanistic insight into the impaired glucose tolerance exhibited in Lhx1^∆Panc animals, islet-enriched transcription factor mRNAs along with transcripts encoding critical glucose-stimulated insulin secretion (GSIS) factors were evaluated in 16-wk-old control and Lhx1^∆Panc islets. Expression of α-cell-specific mRNAs Arx and MafB was significantly upregulated in Lhx1^∆Panc islets (Fig. 4K). However, the levels of key β-cell transcription factor mRNAs Pdx1, MafA, and Nkx6.1 were unaffected by Lhx1 loss. Assessment of GSIS genes revealed that, similarly to the Lhx1 knockdown experiments (Fig. 3D), mRNA levels of Glp1R were significantly reduced in Lhx1^∆Panc islets (Fig. 4K). Of particular note is that loss of Lhx1 in pancreatic islets also led to increased Glut2 mRNA. Collectively, these data support that Lhx1 is required for normal glucose homeostasis and islet gene expression in aged mice.

Lhx1 contributes to the β-cell incretin effect.

Because islet mRNA obtained from aged Lhx1^∆Panc animals had significant reductions of Glp1R and an increase of non-β-cell mRNAs (Fig. 4K), we tested for additional impacts on glucose homeostasis. An oral glucose tolerance test revealed that Lhx1^∆Panc animals exhibited impaired oral glucose tolerance as compared with littermate controls (Fig. 5A). To examine whether the elevated α-cell mRNAs were associated with altered glucagon levels, we measured plasma glucagon during an insulin tolerance test (ITT), when the insulin-induced reduction of plasma glucose is expected to impart counterregulatory glucagon secretion. Although there was no significant difference in insulin action between the experimental groups during the ITT (Fig. 5B), significantly elevated plasma glucagon levels were observed in Lhx1^∆Panc mice 30 min after insulin challenge (Fig. 5B, inset), suggesting that Lhx1^∆Panc mice may have dysregulated glucagon secretion, or altered α-cell numbers.

To test the response to an exogenous Glp1r agonist, Lhx1^∆Panc mice were IP injected with exendin-4 (Ex-4) 10 min before glucose injection for an IPGTT. Lhx1^∆Panc mice treated with Ex-4 exhibited a trend toward a reduced ability to clear glucose when compared with control mice. However, Lhx1^∆Panc mice demonstrated a significant improvement in glucose clearance upon Ex-4 treatment when compared with PBS-injected Lhx1^∆Panc mice, suggesting that some fraction of Glp1 signaling remained intact (Fig. 5C). We also examined plasma Glp1 levels after a mixed meal challenge to determine whether a difference in the endogenous Glp1R ligand may contribute to the glucose intolerance exhibited in Lhx1^∆Panc mice; however, there was no difference (Fig. 5D). We also employed a novel Glp1R agonist, termed IUB282, with greater potency for Glp1R than native Glp1, as assessed by an in vitro cAMP reporter assay (Fig. 5E; also see materials and methods). Pretreatment with IUB282 10 min before IPGTT revealed that Lhx1^∆Panc mice displayed a significantly elevated glucose level at the 15 min time point as compared with IUB282-treated control mice (Fig. 5F). These observations support that Lhx1^∆Panc mice have dampened Glp1R responses, which is likely due in part to significant reductions in β-cell Glp1R.

To examine whether the Lhx1^∆Panc mice also have changes in insulin secretion, we performed static GSIS experiments with islets isolated from control and Lhx1^∆Panc mice. Lhx1^∆Panc islets secreted significantly less insulin than littermate controls when treated with 11 mM stimulatory glucose levels and Ex-4 (Fig. 5G), further supporting that Lhx1-deficient mice have impaired Glp1R-mediated potentiation of GSIS. This dampened insulin secretion in Lhx1^∆Panc islets upon Glp1R agonism was independent of insulin content, as levels were unchanged between Lhx1^∆Panc and control mice (Fig. 5H). Finally, we conducted an in vivo GSIS assay to assess insulin levels in vivo upon a glucose challenge. In Lhx1^∆Panc mice, insulin secretion was significantly reduced 15 min after glucose challenge (Fig. 5I). Furthermore, unlike littermate controls, the Lhx1^∆Panc animals failed to respond to the IP glucose challenge, as plasma insulin remained similar at the 0-, 5-, and 15-min time points. Collectively, these data support that Lhx1 is required for normal GSIS and has impacts on the incretin system, possibly via Glp1R regulation.

DISCUSSION

To determine the in vivo role of Lhx1 in islet function, we created and characterized mice with a conditional Lhx1 knockout in the pancreas (Lhx1^∆Panc). Adult male Lhx1^∆Panc mice exhibited mild glucose intolerance, an increase in mRNAs encoding for non-β-cell hormones and TF mRNAs critical for α-cell identity and function, an elevated counterregulatory glucagon response to IP insulin during an ITT, and a significant reduction in Glp1R. IPGTT accompanied with ex vivo GSIS analyses revealed that Lhx1^∆Panc animals have an impaired response to Glp1R agonism. Together, these data provide genetic evidence of a role for Lhx1 in the function of pancreatic islets.

Our in vitro observation that Lhx1 interacts with Ldb1 and Isl1 in β-cell lines and islets is noteworthy when considering the combinatorial nature of LIM domain transcription factor interactions (4, 6, 36). In spinal cord neuron subtype specification, the Ldb1 co-regulator interacts with the Lhx3 LIM-HD transcription factor to promote V2 interneuron development as a tetrameric complex (31, 37, 53). Thus, two Ldb1 molecules interact as an obligatory homodimer with two Lhx3 molecules to control Chx10 expression (31, 53), which may parallel a similar complex of Ldb1 and Isl1 impacting the MafA expression in β-cells (11, 18, 44). Our co-IP and co-immunofluorescence data (Figs. 1C and 2) suggested the possibility that, in β-cells or islets, a Ldb1-Isl1-Lhx1 complex also exists. This has also been described in neural specification, where Isl1 is expressed to participate with Lhx3 in a hexameric Ldb1-Isl1-Lhx3 complex. In this scenario, Isl1 acts as the direct Ldb1 binding partner, as the COOH terminus of Isl1 contains a binding domain for directly contacting Lhx3 (31, 53). This unique Ldb1-Isl1-Lhx3 hexameric complex (stoichiometry of 2:2:2) promotes motor neuron fates by activating transcription of Hb9/Mnx1 (31, 53). Although our data suggest the possibility of a hexameric Ldb1-Isl1-Lhx1 complex in β-cells, we observed that upon loss of Lhx1 in Lhx1^∆Panc islets, Ldb1-Isl1 interactions were largely unperturbed as tested by co-IP (data not shown), suggesting that an Isl1-Ldb1 islet complex is maintained upon Lhx1 loss or reduction.

Using ChIP, we observed Lhx1 occupancy at the Isl1-Ldb1 bound MafA region 3 control domain (44), along with a 5′ control domain of Glp1R. Although our experiments did not test for simultaneous binding, the data support that a Ldb1-Isl1-Lhx1 hexameric complex occupies the 5′ control domain of Glp1R, and all three factors are needed for activation of this target. However, despite the observed Lhx1 occupancy of MafA region 3, there was no change in MafA transcripts in the cell line knockdown or mouse islet knockout studies, suggesting that Lhx1 is dispensable for MafA activation. Moreover, our Lhx1^∆Panc islet mRNA data indicated that Lhx1 may also regulate Arx and/or MafB. A prior study established Isl1 as an activator of Arx transcription (32); however, in our study, the Isl1-interacting partner Lhx1 may act as a repressor, as Arx mRNAs were elevated in Lhx1-deficient islets. We do not yet know whether Lhx1 directly occupies the Arx or MafB loci in islets.

Previous reports revealed that Ldb1 and Lhx1 interact to regulate head development. Mice lacking Lhx1 or Ldb1 exhibit truncated head phenotypes, suggesting similar function (38, 47). However, we note that the Lhx1^∆Panc mice do not phenocopy previously reported Ldb1 and Isl1 conditional islet knockouts (11, 18). The islet Ldb1 and Isl1 knockout models exhibited severe reductions in hormone⁺ cell numbers, concomitant neonatal hyperglycemia, and early death. Conversely, Lhx1^∆Panc mice presented with a milder phenotype, exhibiting glucose intolerance, increased mRNAs encoding non-β-cell hormones and α-cell TFs, elevated plasma glucagon, and an altered incretin response. Thus, it is possible that the more highly expressed pancreatic Isl1 LIM-HD factor compensates for Lhx1 loss, preventing a severe pancreatic phenotype. This scenario is similar to another LIM-HD transcription factor, Lmx1b, which compensates for loss of Lhx2 and Lhx9 in dorsal limb patterning (55).

Assessment of P1 Lhx1^∆Panc pups revealed no change in body weight, pancreas weight, ad libitum blood glucose, or plasma insulin levels. However, these animals exhibited significantly increased ad libitum plasma glucagon levels, suggesting that Lhx1^∆Panc neonates may have altered islet cell allocation or function. This elevation of plasma glucagon appeared to present with no consequence, for example, in elevating blood glucose levels, which could be due in part to an unobserved modulation of other counter-regulatory islet hormones that may dampen the glucagon effect.

The mild glucose homeostasis phenotype in aged Lhx1^∆Panc animals suggests that Lhx1 may play a role in β-cell compensation to metabolic stressors, such as age, pregnancy, or high-fat diet. Future studies will address this. Moreover, Lhx1^∆Panc islets had significantly elevated MafB, Arx, glucagon, somatostatin, and pancreatic polypeptide mRNAs. Our data also imply that Lhx1 or a yet identified downstream mediator may exert a suppressive effect on glucagon and/or pancreatic polypeptide expression. However, whether Lhx1 directly represses glucagon, somatostatin, and/or pancreatic polypeptide expression requires further investigation. A caveat of this study design involves use of a pancreas-wide Pdx1-Cre transgenic line. With this tool, it is difficult to conclude which islet cell types were directly impacted by Lhx1 loss or address whether the elevation of islet hormone mRNAs is due to transdifferentiation or a direct transcriptional effect. To begin addressing this, we conducted an Lhx1 siRNA-mediated knockdown experiment in the α-cell line αTC6 (data not shown) and observed a significant increase of glucagon, suggesting that the in vivo gene expression changes in the Lhx1^∆Panc mice were perhaps a consequence of lost repression by Lhx1. Future cell type-specific Lhx1 models (e.g., driven by Ins1-Cre or Glucagon-Cre) will shed light on Lhx1 function in β- and α-cells.

The elevation of glucagon mRNA possibly contributes to the glucose intolerance observed in the Lhx1^∆Panc mice. This follows that the impaired Glp1R response alone does not explain the glucose intolerance exhibited in aged Lhx1^∆Panc mice. Additionally, no significant difference was observed in the β-cell mass (Fig. 4J), insulin content (Fig. 5H), or basal insulin secretion (Fig. 5I) in Lhx1^∆Panc animals when compared with littermate controls. Although we did not observe changes in basal plasma glucagon in adult Lhx1^∆Panc mice, we did observe a significant glucagon elevation 30 min after an IP insulin injection during an ITT. This suggests that Lhx1^∆Panc mice may have defective glucagon secretion from α-cells. It has been demonstrated that insulin signaling in α-cells modulates glucagon secretion in vivo (26). Kawamori et al. (26) generated an α-cell insulin receptor-knockout mouse (αIRKO) and discovered a mild glucose intolerance and hyperglucagonemia when α-cells were met with stressors and/or stimuli, similar to the phenotype exhibited in the Lhx1^∆Panc mice.

Conversely, the islet paracrine actions of pancreatic polypeptide have not been fully elucidated. Some reports have suggested that pancreatic polypeptide inhibits both somatostatin and glucagon release from δ- and α-cells, respectively (3, 27). Thus, it is possible that the increased pancreatic polypeptide is an indirect compensatory action to restore normoglycemia in the Lhx1^∆Panc mice. Alternatively, perhaps Lhx1 participates in the direct transcriptional control of PP cells, with loss imparting elevated pancreatic polypeptide. However, our current study did not allow us to discern these possibilities.

The incretin system is an attractive therapeutic target for the treatment of type 2 diabetes (8), with Glp1R agonists such as exenatide and liraglutide acting to suppress postprandial glucagon release while increasing insulin sensitivity and secretion. Lhx1^∆Panc mice displayed oral glucose intolerance and an altered response upon Glp1R agonism. This suggests that the Glp1R reduction exhibited in the Lhx1^∆Panc model is sufficient to impact the incretin response. Distinct roles for Glp1R in glucose homeostasis have been established, as global Glp1R^−/− mice are glucose intolerant upon oral and intraperitoneal glucose challenge (46). However, it is less clear whether the β-cell expressed Glp1R is required for oral and intraperitoneal glucose tolerance. Our findings are in accord with those of Lamont et al. (30), which revealed the importance of β-cell Glp1R in both oral and intraperitoneal glucose tolerance through use of a transgenic mouse expressing human GLP1R in islets on a Glp1R^−/− background. Conversely, Smith et al. (50) employed an inducible model of Glp1R removal from β-cells and found that β-cell Glp1R is not necessary for oral glucose tolerance, but only for intraperitoneal glucose tolerance. However, there are differences in each of these studies (including ours) that may cloud direct comparisons. Because we generated a model of transcription factor loss, it is likely that multiple genes impacting GSIS are influenced, thus contributing to the phenotype. However, as already discussed, in addition to reduced Glp1R, Lhx1^∆Panc mice also had elevated plasma glucagon, along with mild glucose intolerance. Although the reductions in Glp1R resulted in mild physiological impacts, this alone does not account for the other glucose homeostasis phenotypes displayed in the Lhx1^∆Panc mice. We conclude that the Lhx1 transcription factor is a novel islet regulator contributing to normal glucose homeostasis and Glp1 signaling.

GRANTS

This work was supported by the National Institute of Diabetes and Digestive and Kidney Diseases (NIDDK) Grants R01-DK-111483 (to C. S. Hunter) and F31-DK-111181 (to M. Bethea), and American Diabetes Association Grant 1-16-JDF-044 (to C. S. Hunter). Core services in support of this research were provided by the University of Alabama at Birmingham Diabetes Research Center through NIDDK Grant P30-DK-079626.

DISCLOSURES

No conflicts of interest, financial or otherwise, are declared by the authors.

AUTHOR CONTRIBUTIONS

M.B., R.M., V.G., R.D., S.B., R.B., K.M.H., and C.S.H. conceived and designed research; M.B., Y.L., A.K.W., R.M., R.G., V.G., K.M.H., and C.S.H. performed experiments; M.B., Y.L., A.K.W., R.G., V.G., S.B., K.M.H., and C.S.H. analyzed data; M.B., R.D., S.B., R.B., K.M.H., and C.S.H. interpreted results of experiments; M.B. and C.S.H. drafted manuscript; M.B., R.M., R.D., S.B., K.M.H., and C.S.H. edited and revised manuscript; M.B. and C.S.H. approved final version of manuscript; C.S.H. prepared figures.