Research ArticleBone BiologyMetabolism Open Access | 10.1172/JCI159928

Marta Martinez-Calle,1 Guillaume Courbon,1 Bridget Hunt-Tobey,1 Connor Francis,1 Jadeah Spindler,1 Xueyan Wang,1 Luciene M. dos Reis,2 Carolina S.W. Martins,2 Isidro B. Salusky,3 Hartmut Malluche,4 Thomas L. Nickolas,5 Rosa M.A. Moyses,2 Aline Martin,1 and Valentin David1

1Division of Nephrology and Hypertension, Department of Medicine, and Center for Translational Metabolism and Health, Institute for Public Health and Medicine, Northwestern University Feinberg School of Medicine, Chicago, Illinois, USA.

2LIM 16, Nephrology Department, Hospital das Clínicas da Faculdade de Medicina da USP (HCFMUSP), Universidade de São Paulo, São Paulo, Brazil.

3Department of Pediatrics, David Geffen School of Medicine at UCLA, Los Angeles, California, USA.

4Division of Nephrology, Bone and Mineral Metabolism, Department of Internal Medicine, University of Kentucky, Lexington, Kentucky, USA.

5Department of Medicine, Columbia Irving University Medical Center, New York, New York, USA.

Address correspondence to: Valentin David, Northwestern University Feinberg School of Medicine, Division of Nephrology and Hypertension & Center for Translational Metabolism and Health, 320 East Superior Street, Searle Building, Suite 10-517, Chicago, Illinois 60611, USA. Phone: 312.503.4159; Email: valentin.david@northwestern.edu.

Find articles by Martinez-Calle, M. in: JCI | PubMed | Google Scholar

1Division of Nephrology and Hypertension, Department of Medicine, and Center for Translational Metabolism and Health, Institute for Public Health and Medicine, Northwestern University Feinberg School of Medicine, Chicago, Illinois, USA.

2LIM 16, Nephrology Department, Hospital das Clínicas da Faculdade de Medicina da USP (HCFMUSP), Universidade de São Paulo, São Paulo, Brazil.

3Department of Pediatrics, David Geffen School of Medicine at UCLA, Los Angeles, California, USA.

4Division of Nephrology, Bone and Mineral Metabolism, Department of Internal Medicine, University of Kentucky, Lexington, Kentucky, USA.

5Department of Medicine, Columbia Irving University Medical Center, New York, New York, USA.

Address correspondence to: Valentin David, Northwestern University Feinberg School of Medicine, Division of Nephrology and Hypertension & Center for Translational Metabolism and Health, 320 East Superior Street, Searle Building, Suite 10-517, Chicago, Illinois 60611, USA. Phone: 312.503.4159; Email: valentin.david@northwestern.edu.

Find articles by Courbon, G. in: JCI | PubMed | Google Scholar |

1Division of Nephrology and Hypertension, Department of Medicine, and Center for Translational Metabolism and Health, Institute for Public Health and Medicine, Northwestern University Feinberg School of Medicine, Chicago, Illinois, USA.

2LIM 16, Nephrology Department, Hospital das Clínicas da Faculdade de Medicina da USP (HCFMUSP), Universidade de São Paulo, São Paulo, Brazil.

3Department of Pediatrics, David Geffen School of Medicine at UCLA, Los Angeles, California, USA.

4Division of Nephrology, Bone and Mineral Metabolism, Department of Internal Medicine, University of Kentucky, Lexington, Kentucky, USA.

5Department of Medicine, Columbia Irving University Medical Center, New York, New York, USA.

Address correspondence to: Valentin David, Northwestern University Feinberg School of Medicine, Division of Nephrology and Hypertension & Center for Translational Metabolism and Health, 320 East Superior Street, Searle Building, Suite 10-517, Chicago, Illinois 60611, USA. Phone: 312.503.4159; Email: valentin.david@northwestern.edu.

Find articles by Hunt-Tobey, B. in: JCI | PubMed | Google Scholar

1Division of Nephrology and Hypertension, Department of Medicine, and Center for Translational Metabolism and Health, Institute for Public Health and Medicine, Northwestern University Feinberg School of Medicine, Chicago, Illinois, USA.

2LIM 16, Nephrology Department, Hospital das Clínicas da Faculdade de Medicina da USP (HCFMUSP), Universidade de São Paulo, São Paulo, Brazil.

3Department of Pediatrics, David Geffen School of Medicine at UCLA, Los Angeles, California, USA.

4Division of Nephrology, Bone and Mineral Metabolism, Department of Internal Medicine, University of Kentucky, Lexington, Kentucky, USA.

5Department of Medicine, Columbia Irving University Medical Center, New York, New York, USA.

Address correspondence to: Valentin David, Northwestern University Feinberg School of Medicine, Division of Nephrology and Hypertension & Center for Translational Metabolism and Health, 320 East Superior Street, Searle Building, Suite 10-517, Chicago, Illinois 60611, USA. Phone: 312.503.4159; Email: valentin.david@northwestern.edu.

Find articles by Francis, C. in: JCI | PubMed | Google Scholar

1Division of Nephrology and Hypertension, Department of Medicine, and Center for Translational Metabolism and Health, Institute for Public Health and Medicine, Northwestern University Feinberg School of Medicine, Chicago, Illinois, USA.

2LIM 16, Nephrology Department, Hospital das Clínicas da Faculdade de Medicina da USP (HCFMUSP), Universidade de São Paulo, São Paulo, Brazil.

3Department of Pediatrics, David Geffen School of Medicine at UCLA, Los Angeles, California, USA.

4Division of Nephrology, Bone and Mineral Metabolism, Department of Internal Medicine, University of Kentucky, Lexington, Kentucky, USA.

5Department of Medicine, Columbia Irving University Medical Center, New York, New York, USA.

Address correspondence to: Valentin David, Northwestern University Feinberg School of Medicine, Division of Nephrology and Hypertension & Center for Translational Metabolism and Health, 320 East Superior Street, Searle Building, Suite 10-517, Chicago, Illinois 60611, USA. Phone: 312.503.4159; Email: valentin.david@northwestern.edu.

Find articles by Spindler, J. in: JCI | PubMed | Google Scholar

1Division of Nephrology and Hypertension, Department of Medicine, and Center for Translational Metabolism and Health, Institute for Public Health and Medicine, Northwestern University Feinberg School of Medicine, Chicago, Illinois, USA.

2LIM 16, Nephrology Department, Hospital das Clínicas da Faculdade de Medicina da USP (HCFMUSP), Universidade de São Paulo, São Paulo, Brazil.

3Department of Pediatrics, David Geffen School of Medicine at UCLA, Los Angeles, California, USA.

4Division of Nephrology, Bone and Mineral Metabolism, Department of Internal Medicine, University of Kentucky, Lexington, Kentucky, USA.

5Department of Medicine, Columbia Irving University Medical Center, New York, New York, USA.

Address correspondence to: Valentin David, Northwestern University Feinberg School of Medicine, Division of Nephrology and Hypertension & Center for Translational Metabolism and Health, 320 East Superior Street, Searle Building, Suite 10-517, Chicago, Illinois 60611, USA. Phone: 312.503.4159; Email: valentin.david@northwestern.edu.

Find articles by Wang, X. in: JCI | PubMed | Google Scholar

1Division of Nephrology and Hypertension, Department of Medicine, and Center for Translational Metabolism and Health, Institute for Public Health and Medicine, Northwestern University Feinberg School of Medicine, Chicago, Illinois, USA.

2LIM 16, Nephrology Department, Hospital das Clínicas da Faculdade de Medicina da USP (HCFMUSP), Universidade de São Paulo, São Paulo, Brazil.

3Department of Pediatrics, David Geffen School of Medicine at UCLA, Los Angeles, California, USA.

4Division of Nephrology, Bone and Mineral Metabolism, Department of Internal Medicine, University of Kentucky, Lexington, Kentucky, USA.

5Department of Medicine, Columbia Irving University Medical Center, New York, New York, USA.

Address correspondence to: Valentin David, Northwestern University Feinberg School of Medicine, Division of Nephrology and Hypertension & Center for Translational Metabolism and Health, 320 East Superior Street, Searle Building, Suite 10-517, Chicago, Illinois 60611, USA. Phone: 312.503.4159; Email: valentin.david@northwestern.edu.

Find articles by dos Reis, L. in: JCI | PubMed | Google Scholar

1Division of Nephrology and Hypertension, Department of Medicine, and Center for Translational Metabolism and Health, Institute for Public Health and Medicine, Northwestern University Feinberg School of Medicine, Chicago, Illinois, USA.

2LIM 16, Nephrology Department, Hospital das Clínicas da Faculdade de Medicina da USP (HCFMUSP), Universidade de São Paulo, São Paulo, Brazil.

3Department of Pediatrics, David Geffen School of Medicine at UCLA, Los Angeles, California, USA.

4Division of Nephrology, Bone and Mineral Metabolism, Department of Internal Medicine, University of Kentucky, Lexington, Kentucky, USA.

5Department of Medicine, Columbia Irving University Medical Center, New York, New York, USA.

Address correspondence to: Valentin David, Northwestern University Feinberg School of Medicine, Division of Nephrology and Hypertension & Center for Translational Metabolism and Health, 320 East Superior Street, Searle Building, Suite 10-517, Chicago, Illinois 60611, USA. Phone: 312.503.4159; Email: valentin.david@northwestern.edu.

Find articles by Martins, C. in: JCI | PubMed | Google Scholar

1Division of Nephrology and Hypertension, Department of Medicine, and Center for Translational Metabolism and Health, Institute for Public Health and Medicine, Northwestern University Feinberg School of Medicine, Chicago, Illinois, USA.

2LIM 16, Nephrology Department, Hospital das Clínicas da Faculdade de Medicina da USP (HCFMUSP), Universidade de São Paulo, São Paulo, Brazil.

3Department of Pediatrics, David Geffen School of Medicine at UCLA, Los Angeles, California, USA.

4Division of Nephrology, Bone and Mineral Metabolism, Department of Internal Medicine, University of Kentucky, Lexington, Kentucky, USA.

5Department of Medicine, Columbia Irving University Medical Center, New York, New York, USA.

Address correspondence to: Valentin David, Northwestern University Feinberg School of Medicine, Division of Nephrology and Hypertension & Center for Translational Metabolism and Health, 320 East Superior Street, Searle Building, Suite 10-517, Chicago, Illinois 60611, USA. Phone: 312.503.4159; Email: valentin.david@northwestern.edu.

Find articles by Salusky, I. in: JCI | PubMed | Google Scholar

1Division of Nephrology and Hypertension, Department of Medicine, and Center for Translational Metabolism and Health, Institute for Public Health and Medicine, Northwestern University Feinberg School of Medicine, Chicago, Illinois, USA.

2LIM 16, Nephrology Department, Hospital das Clínicas da Faculdade de Medicina da USP (HCFMUSP), Universidade de São Paulo, São Paulo, Brazil.

3Department of Pediatrics, David Geffen School of Medicine at UCLA, Los Angeles, California, USA.

4Division of Nephrology, Bone and Mineral Metabolism, Department of Internal Medicine, University of Kentucky, Lexington, Kentucky, USA.

5Department of Medicine, Columbia Irving University Medical Center, New York, New York, USA.

Address correspondence to: Valentin David, Northwestern University Feinberg School of Medicine, Division of Nephrology and Hypertension & Center for Translational Metabolism and Health, 320 East Superior Street, Searle Building, Suite 10-517, Chicago, Illinois 60611, USA. Phone: 312.503.4159; Email: valentin.david@northwestern.edu.

Find articles by Malluche, H. in: JCI | PubMed | Google Scholar

1Division of Nephrology and Hypertension, Department of Medicine, and Center for Translational Metabolism and Health, Institute for Public Health and Medicine, Northwestern University Feinberg School of Medicine, Chicago, Illinois, USA.

2LIM 16, Nephrology Department, Hospital das Clínicas da Faculdade de Medicina da USP (HCFMUSP), Universidade de São Paulo, São Paulo, Brazil.

3Department of Pediatrics, David Geffen School of Medicine at UCLA, Los Angeles, California, USA.

4Division of Nephrology, Bone and Mineral Metabolism, Department of Internal Medicine, University of Kentucky, Lexington, Kentucky, USA.

5Department of Medicine, Columbia Irving University Medical Center, New York, New York, USA.

Address correspondence to: Valentin David, Northwestern University Feinberg School of Medicine, Division of Nephrology and Hypertension & Center for Translational Metabolism and Health, 320 East Superior Street, Searle Building, Suite 10-517, Chicago, Illinois 60611, USA. Phone: 312.503.4159; Email: valentin.david@northwestern.edu.

Find articles by Nickolas, T. in: JCI | PubMed | Google Scholar |

1Division of Nephrology and Hypertension, Department of Medicine, and Center for Translational Metabolism and Health, Institute for Public Health and Medicine, Northwestern University Feinberg School of Medicine, Chicago, Illinois, USA.

2LIM 16, Nephrology Department, Hospital das Clínicas da Faculdade de Medicina da USP (HCFMUSP), Universidade de São Paulo, São Paulo, Brazil.

3Department of Pediatrics, David Geffen School of Medicine at UCLA, Los Angeles, California, USA.

4Division of Nephrology, Bone and Mineral Metabolism, Department of Internal Medicine, University of Kentucky, Lexington, Kentucky, USA.

5Department of Medicine, Columbia Irving University Medical Center, New York, New York, USA.

Address correspondence to: Valentin David, Northwestern University Feinberg School of Medicine, Division of Nephrology and Hypertension & Center for Translational Metabolism and Health, 320 East Superior Street, Searle Building, Suite 10-517, Chicago, Illinois 60611, USA. Phone: 312.503.4159; Email: valentin.david@northwestern.edu.

Find articles by Moyses, R. in: JCI | PubMed | Google Scholar |

1Division of Nephrology and Hypertension, Department of Medicine, and Center for Translational Metabolism and Health, Institute for Public Health and Medicine, Northwestern University Feinberg School of Medicine, Chicago, Illinois, USA.

2LIM 16, Nephrology Department, Hospital das Clínicas da Faculdade de Medicina da USP (HCFMUSP), Universidade de São Paulo, São Paulo, Brazil.

3Department of Pediatrics, David Geffen School of Medicine at UCLA, Los Angeles, California, USA.

4Division of Nephrology, Bone and Mineral Metabolism, Department of Internal Medicine, University of Kentucky, Lexington, Kentucky, USA.

5Department of Medicine, Columbia Irving University Medical Center, New York, New York, USA.

Address correspondence to: Valentin David, Northwestern University Feinberg School of Medicine, Division of Nephrology and Hypertension & Center for Translational Metabolism and Health, 320 East Superior Street, Searle Building, Suite 10-517, Chicago, Illinois 60611, USA. Phone: 312.503.4159; Email: valentin.david@northwestern.edu.

Find articles by Martin, A. in: JCI | PubMed | Google Scholar |

1Division of Nephrology and Hypertension, Department of Medicine, and Center for Translational Metabolism and Health, Institute for Public Health and Medicine, Northwestern University Feinberg School of Medicine, Chicago, Illinois, USA.

2LIM 16, Nephrology Department, Hospital das Clínicas da Faculdade de Medicina da USP (HCFMUSP), Universidade de São Paulo, São Paulo, Brazil.

3Department of Pediatrics, David Geffen School of Medicine at UCLA, Los Angeles, California, USA.

4Division of Nephrology, Bone and Mineral Metabolism, Department of Internal Medicine, University of Kentucky, Lexington, Kentucky, USA.

5Department of Medicine, Columbia Irving University Medical Center, New York, New York, USA.

Address correspondence to: Valentin David, Northwestern University Feinberg School of Medicine, Division of Nephrology and Hypertension & Center for Translational Metabolism and Health, 320 East Superior Street, Searle Building, Suite 10-517, Chicago, Illinois 60611, USA. Phone: 312.503.4159; Email: valentin.david@northwestern.edu.

Find articles by David, V. in: JCI | PubMed | Google Scholar |

Published April 20, 2023 - More info

Published in Volume 133, Issue 11 on June 1, 2023
J Clin Invest. 2023;133(11):e159928. https://doi.org/10.1172/JCI159928.
© 2023 Martinez-Calle et al. This work is licensed under the Creative Commons Attribution 4.0 International License. To view a copy of this license, visit http://creativecommons.org/licenses/by/4.0/. Published April 20, 2023 - Version history
Received: March 4, 2022; Accepted: April 17, 2023 View PDF Abstract

Renal osteodystrophy (ROD) is a disorder of bone metabolism that affects virtually all patients with chronic kidney disease (CKD) and is associated with adverse clinical outcomes including fractures, cardiovascular events, and death. In this study, we showed that hepatocyte nuclear factor 4α (HNF4α), a transcription factor mostly expressed in the liver, is also expressed in bone, and that osseous HNF4α expression was dramatically reduced in patients and mice with ROD. Osteoblast-specific deletion of Hnf4α resulted in impaired osteogenesis in cells and mice. Using multi-omics analyses of bones and cells lacking or overexpressing Hnf4α1 and Hnf4α2, we showed that HNF4α2 is the main osseous Hnf4α isoform that regulates osteogenesis, cell metabolism, and cell death. As a result, osteoblast-specific overexpression of Hnf4α2 prevented bone loss in mice with CKD. Our results showed that HNF4α2 is a transcriptional regulator of osteogenesis, implicated in the development of ROD.

Graphical Abstract Introduction

Chronic kidney disease (CKD) is a costly public health burden that increases the risk of mortality (1). Disordered bone and mineral metabolism is a nearly universal complication of CKD, collectively termed CKD–mineral and bone disorder (CKD-MBD), that begins early and worsens progressively as kidney function declines (2–4).

Renal osteodystrophy (ROD) is the bone disease associated with CKD. ROD is a disorder of bone cell function and metabolism that leads to abnormal structure and compromised bone strength. Loss of bone quantity and quality due to high and low bone turnover and further onset of bone lesions are strongly associated with progressive impairment of kidney function. The exact pathogenesis of ROD is poorly understood, but it is often described as a particular subset of metabolic bone disease. Although disturbances in circulating factors, such as calcitriol, parathyroid hormone, and fibroblast growth factor 23, and the resulting impact on phosphate and calcium levels, have major skeletal effects, in recent years it became clear that intrinsic osseous mechanisms might contribute to the onset and progression of ROD (5–7). Indeed, skeletal abnormalities persist despite therapy with different active vitamin D sterols and phosphate binders (8–10), and bone deformities, fractures, and growth retardation remain the long-term consequences of CKD for the growing skeleton (11–13). To date, the molecular mechanisms of bone loss in ROD remain to be determined.

Hepatocyte nuclear factor 4α (HNF4α) is a highly conserved transcription factor, a member of the nuclear receptor (NR) family, which regulates gene transcription by binding DNA as a dimer. In contrast to other types of NR, HNF4α is constitutively localized in the nucleus and does not require binding of a ligand to homodimerize and interact with the response elements of its target genes (14). HNF4α can function as an activator or repressor of genes involved in cell metabolic activity, transport, glucose and lipid homeostasis, and detoxification of xenobiotics (15–20). HNF4α was initially discovered as a regulator of liver-specific gene expression. However, HNF4α expression has also been described in multiple other organs, including pancreas, kidney, stomach, small intestine, and colon (19, 21–23). Mutations of HNF4α and HNF4α response elements cause maturity-onset diabetes of the young 1 (MODY1), a rare disease; certain types of hemophilia; and hepatitis B viral infections. In addition, HDL-cholesterol (24), metabolic dyslipidemia (25), and type 2 diabetes mellitus (26, 27) have been associated with the HNF4α locus by genome-wide associations studies. Importantly, HNF4α is also associated with coronary artery calcification in the Chronic Renal Insufficiency Cohort (28) and with osteoporosis in the Framingham Osteoporosis Study (29), which was mostly attributed to HNF4α function in liver and kidney. However, the direct role of HNF4α in bone has never been investigated despite the clear associations between HNF4α and disturbances in bone and mineral metabolism.

In the present study, we report the expression of 2 main isoforms of HNF4α in bone, HNF4α1 and HNF4α2, and we investigated the role of osseous HNF4α in the pathology of ROD in human and experimental models. First, we show that HNF4α expression is nearly completely suppressed in bone from patients and mice with CKD. We report that HNF4α2 is a major regulator of osteogenesis using genetics and multi-omics approaches in vitro and in vivo. Finally, we show the key impact of restoring osteoblastic HNF4α2 expression on bone mass in mice with CKD. These results establish the direct role of osseous HNF4α2 in the regulation of osteogenesis, suggest that osseous HNF4α2 deficiency contributes to the pathogenesis of ROD, and propose a mechanism to explain intrinsic bone defects in patients with CKD.

Results

HNF4α is expressed in bone and its expression is reduced in patients and animals with CKD. We performed RNA-Seq on bone biopsies collected from patients (Supplemental Table 1; supplemental material available online with this article; https://doi.org/10.1172/JCI159928DS1) with or without ROD (GEO GSE194056), showing either low or high bone turnover to identify ROD-specific transcriptomic profiles (Figure 1A). We mainly identified alterations in expression of genes involved in osteogenesis, metabolism, and cell death (Figure 1, B–D). Among the metabolic genes, we identified HNF4A, a gene not known for its osseous expression, as a gene suppressed in all ROD patient groups compared with non-ROD patients, irrespectively of their bone remodeling status. In humans and mice, HNF4A encodes 12 annotated isoforms with distinct molecular functions and varying transcriptional regulatory potentials (Figure 1E). Accordingly, 12 distinct HNF4α proteins have been annotated in both humans and mice (30, 31). HNF4A isoforms are generated under the control of 2 alternative promoters, P1 and P2, which results in over 60 potential HNF4A homo- or heterodimer isoforms with different impacts on gene expression regulation (30). In adult mice, total Hnf4α mRNA expression was highest in liver. In comparison, the expression of total Hnf4α mRNA was only 40% lower in osteoblast- and osteocyte-enriched bone fraction alone (Figure 1F). Analysis of mRNA expression of the 12 annotated Hnf4α isoforms in mouse bones showed that isoforms 1–3 were the most represented of all Hnf4α isoforms, as in adult liver and kidney (32), and further analysis identified Hnf4α1 and Hnf4α2 as the predominant osseous isoforms (Figure 1, G and H). We next used the Col4a3KO mouse model of progressive CKD, which recapitulates most of the typical features of human CKD, including ROD (4, 33, 34), to investigate changes in Hnf4α expression. As in patients with ROD, we found that expression of Hnf4α1/2 was nearly completely suppressed in the bone of the Col4a3KO mice (Figure 1I).

Figure 1

HNF4A is expressed in bone and is reduced in humans and mice with CKD. (A) Number of differentially regulated genes identified by RNA-Seq of bone biopsies from CKD patients with low–bone remodeling (LR) and high–bone remodeling (HR) renal osteodystrophy (ROD) versus healthy volunteers. (B–D) Heatmap-represented expression of genes identified in the topmost differentially regulated pathways in LR-ROD and HR-ROD bone biopsies versus healthy volunteers. n = 9 (Healthy and LR-ROD) and 11 (HR-ROD); corrected P < 0.05. Statistical analysis was performed with an ANOVA test followed by unpaired Student’s t test and corrected by the FDR. (E) Schematic representation of Hnf4α gene and different promoter P1– and P2–driven Hnf4α isoforms. (F–H) Comparative analysis of total Hnf4α mRNA in liver, kidney, and bone (F), Hnf4α isoforms 1 to 12 mRNA in bone (G), and Hnf4α isoforms 1 to 3 mRNA in bone of WT mice (H). (I) mRNA expression of Hnf4α1/2 in bone of WT and Col4a3KO mice with CKD. Values are expressed as the mean ± SEM. N = 5 per group. Corrected P < 0.05 versus aliver, bkidney, cHnf4α1–3, dHnf4α1, eHnf4α2, and *WT. Statistical analysis was performed with an unpaired Student’s t tests (I) or with an ANOVA followed by post hoc t tests to determine statistical differences and multiple-testing correction using the Holm-Bonferroni method (F–H).

HNF4α2 regulates osteoblastogenesis and osteoblast metabolism. Hnf4α1/2 isoforms were expressed in bone marrow stromal cells (BMSCs) and primary osteoblasts, and to a lesser extent in the MC3T3-E1 osteoblast cell line, cultured for 21 days in osteogenic medium (Figure 2A). To identify the specific role of HNF4α1 and HNF4α2 in osteoblast differentiation, we overexpressed Hnf4α1 and Hnf4α2 in MC3T3-E1 osteoblasts (Hnf4α1Tg and Hnf4α2Tg, respectively). Compared with empty vector–transfected (Ctr) MC3T3-E1 osteoblasts, Hnf4α expression was about 500 times higher in both transgenic cell lines (Figure 2B). Interestingly, overexpression of Hnf4α2, but not Hnf4α1, increased expression of osteoblastic markers such as Runx2 and Sp7, suggesting a major role for Hnf4α2 in osteoblastogenesis (Figure 2, C and D). RNA-Seq (GSE190315) and subsequent pathway analyses of MC3T3-E1 Ctr, Hnf4α1Tg, and Hnf4α2Tg cells showed that overexpression of each isoform modified the expression of known HNF4α targets (Supplemental Figure 1). In addition, Hnf4α2Tg cells displayed increased cell cytoskeleton remodeling pathways, osteogenesis, and metabolic signaling and reduced cAMP/PKA signaling, cell death, calcium/NFAT, and nitric oxide pathways compared with Ctr cells. However, overexpression of Hnf4α1 showed either a milder or an opposite effect on these pathways (Figure 2, E and F), suggesting that HNF4α1 and HNF4α2 functions are non-redundant. Importantly, Hnf4α2Tg cells showed highly modified gene expression profiles of osteogenic and metabolic markers, compared with Ctr and Hnf4α1Tg cells (Figure 2, G and H), supporting a specific role of HNF4α2 in osteoblastogenesis.

Figure 2

HNF4α2 is a major regulator of osteogenesis and metabolism in osteoblasts. (A and B) Hnf4α1/2 mRNA expression in differentiated primary bone marrow stromal cells (BMSCs), mature osteoblasts (OBs), and MC3T3-E1 osteoblasts (A), and in MC3T3-E1 osteoblasts transfected with an empty vector (Ctr), Hnf4α1 (Hnf4α1Tg), and Hnf4α2 (Hnf4α2Tg) expression transgene (B). (C and D) mRNA expression of markers of osteoblast differentiation Runx2 and Sp7. Values are expressed as the mean ± SEM. n ≥ 3 per group of a representative experiment performed at least 3 times; corrected P < 0.05 versus *BMSC or Ctr. Statistical analysis was performed with an ANOVA test followed by post hoc t tests to determine statistical differences and multiple-testing correction using the Holm-Bonferroni method. (E) Number of differentially regulated genes identified by RNA-Seq in Hnf4α1Tg and Hnf4α2Tg osteoblasts versus Ctr. (F) Canonical pathway analysis and prediction of pathway activation of differentially regulated genes identified by RNA-Seq of Ctr, Hnf4α1Tg, and Hnf4α2Tg osteoblasts. (G and H) Heatmap-represented expression of genes modified and involved in osteogenesis and metabolism pathways in Ctr, Hnf4α1Tg, and Hnf4α2Tg osteoblasts. Corrected P < 0.05; n = 3 per group. Statistical analysis was performed with an unpaired Student’s t test and corrected by the FDR.

HNF4α2 is a direct transcriptional regulator of osteoblastic genes. To identify genes directly regulated by HNF4α, we performed 3 different sets of chromatin immunoprecipitation sequencing (ChIP-Seq) analyses using 3 separate antibodies (GSE190314). We first performed HNF4α immunoprecipitation in Ctr, Hnf4α1Tg, and Hnf4α2Tg cell extracts using 2 different polyclonal anti-HNF4α antibodies purchased from Aviva Systems Biology and Abcam, respectively. In parallel, we generated 2 stable cell lines overexpressing Hnf4α2 coupled with a carboxy-terminal (Hnf4α2C-Halo-Tg) or amino-terminal (Hnf4α2N-Halo-Tg) Halo tag and used an anti-Halo antibody to immunoprecipitate HNF4α. Peaks were called in each separate experiment and consolidated as follows: common HNF4α1/2 peaks resulting from the intersection of samples overexpressing either Hnf4α1 or Hnf4α2; HNF4α1 peaks resulting from the intersection of 2 or more experiments overexpressing Hnf4α1 and/or Ctr cells; HNF4α2 peaks resulting from the intersection of 2 or more experiments overexpressing Hnf4α2 and/or Ctr cells (Figure 3A). For all chromatin immunoprecipitations, several HNF4α motifs were identified as the primary target (Figure 3B). HNF4α2 peaks were the most abundant in osteoblasts, and a majority of peaks (60%–70%) showed the expected HNF4α motif. However, a relatively large number of peaks remained without a match to the consensus motif (Figure 3C), as previously shown (35), suggesting that HNF4α binds DNA either through other motifs or by interacting with other cofactors. Consistent with prior reports, both HNF4α1 and HNF4α2 showed a preferential binding to intronic (~30%) and distal intergenic regions (~50%), with a small proportion (~10%) at gene promoters (17) (Figure 3D). Similar to results obtained in RNA-Seq analyses of Hnf4α1Tg and Hnf4α2Tg osteoblasts (Figure 2), downstream analyses of gene targets identified by ChIP-Seq showed enrichment of cell cytoskeleton remodeling, cAMP/PKA signaling, osteogenesis, metabolic, cell death, calcium/NFAT, and nitric oxide pathways (Figure 4A). Therefore, to determine whether the genes dysregulated in Hnf4α1Tg and Hnf4α2Tg osteoblasts are directly regulated by HNF4α binding to DNA, we intersected the HNF4α cistrome with the transcriptomic analyses performed in Figure 2. We found that about 2,500 genes were directly regulated by HNF4α1 and about 5,000 by HNF4α2 (Figure 4B). Downstream pathway analysis of these genes showed an enrichment in cAMP/PKA signaling, osteogenesis, metabolic, cell death, calcium/NFAT, and nitric oxide signaling pathways, supporting the important finding that HNF4α2 directly controls a large part of the osteoblast metabolic activity, differentiation, and death (Figure 4C).

Figure 3

HNF4α-specific ChIP sequencing analysis of HNF4α targets in MC3T3-E1 osteoblasts. (A) Representative illustration of final peak calls based on overlapping naive peaks found in MC3T3-E1 osteoblasts overexpressing an empty vector (Ctr), Hnf4α1 (Hnf4α1Tg), or Hnf4α2 (Hnf4α2Tg). (B) Enriched HNF4α motif sequences found in final peaks from position frequency matrices using MEME Suite (https://meme-suite.org/meme/tools/meme-chip) compared with the curated HNF4α consensus motif. (C and D) Number (C) and distribution across genomic regions (D) of HNF4α1, HNF4α2, or common HNF4α1/2 final peaks. n = 3 biological replicates per experimentally used antibody.

Figure 4

HNF4α2 is a direct transcriptional regulator of osteogenesis and metabolism in osteoblasts. (A) Canonical pathway analysis of HNF4α targets identified by ChIP sequencing of Ctr, Hnf4α1Tg, and Hnf4α2Tg osteoblasts. n = 3 biological replicates per experimentally used antibody. (B) Number of genes differentially regulated in Hnf4α1Tg and Hnf4α2Tg osteoblasts versus Ctr and directly regulated by HNF4α, obtained from the intersection between genes identified by RNA-Seq in Figure 2 and genes identified by ChIP sequencing in Figure 3. (C) Canonical pathway analysis and prediction of pathway activation of direct HNF4α targets identified in A.

Osteoblast-specific deletion of Hnf4α reduces peak bone mass in mice. Next, to determine the physiological importance of HNF4α in bone, we deleted HNF4α in osteoblasts and osteocytes (Hnf4αOc-cKO). These mice showed an approximately 80% reduction in osseous Hnf4α expression (Supplemental Figure 2A). Hnf4αOc-cKO neonates were smaller and hypomineralized (Figure 5A) compared with their WT littermates, and showed an approximately 30% reduction in whole-body (Figure 5B) or femur (Figure 5, C and D) mineralized volume. Young and adult Hnf4αOc-cKO male mice displayed a reduction in body weight (Supplemental Figure 2B) and femur (Figure 5, E and F), tibia, and limb lengths (Supplemental Figure 2, C and D) and did not show modifications of the femur microarchitecture in cortical bone at 6 or 12 weeks of age (Supplemental Figure 2, E, D, and J). However, osteoblast-specific deletion of Hnf4α resulted in an approximately 50% loss of trabecular peak bone mass in 12-week-old male mice, as shown by reduced trabecular bone volume, number, and thickness and reduced trabecular bone mineral density (Figure 5, E and G–J). We observed similar changes in Hnf4αOc-cKO female mice (Supplemental Figure 3, A–E), but female mice also showed a reduction in cortical thickness and cortical area at 6 and 12 weeks of age (Supplemental Figure 3, F–J). Both male and female mice showed a reduced osteoid apposition as measured on Goldner Trichrome–stained nondecalcified bone sections and a lower bone formation rate as assessed by the reduced number of alizarin red–stained mineral seams and distance between the seams, coupled with an increase in osteoclastogenesis as shown by an increase in TRAcP-positive cells (Figure 5K and Supplemental Table 2). Notably, deletion of Hnf4α earlier in the osteoblastic lineage, using an Osterix-Cre–mediated deletion, exacerbated these changes in 12-week-old animals, in both male and female mice (Supplemental Figure 4).

Figure 5

Bone-specific deletion of Hnf4α leads to low bone mass and impaired bone growth. (A–J) 3D microtomography analysis of whole-body skeleton (A and B) and entire femur (C and D; bottom panel of D shows a longitudinal section) of WT and Hnf4αOc-cKO neonates, and of entire femur and femur metaphysis of young (6 weeks) and adult (12 weeks) WT and Hnf4αOc-cKO mice (E–J). BMD, bone mineral density; BV, bone volume; TV, total volume; Tb, trabecular; N, number; Th, thickness. (K) Microscopy analysis of modified Goldner Trichrome staining (left), alizarin red S staining (middle), and TRAcP staining (right) of femur trabecular bone from 6- and 12-week-old WT and Hnf4αOc-cKO mice. Values are expressed as the mean ± SEM. n ≥ 5 per group; P < 0.05 versus *age-matched WT. Statistical analysis was performed with unpaired Student’s t tests.

To determine the impact of osteoblast-specific deletion of Hnf4α on the expression of bone transcripts, we performed RNA-Seq on femora isolated from 6-week-old WT and Hnf4αOc-cKO male littermates (GSE190313). First, we show that reduction of Hnf4α in osteoblasts affected the expression of gene targets of HNF4α previously established in other tissues (Supplemental Figure 5). In addition, bones from Hnf4αOc-cKO mice showed impaired expression of genes from the major pathways identified in cultured osteoblasts (Figure 6A), leading to a defect in osteogenesis, metabolic, and cell death transcripts (Figure 6, B–D), consistent with profiles observed in patients with ROD. Interestingly, deletion of Hnf4α in osteoblasts increased the proinflammatory signaling in the bone, leading to activation of major cytokine signaling and prototypical NF-κB signaling (Figure 6A). Intersection of significantly altered genes in the bone of Hnf4αOc-cKO mice with transcripts directly regulated by either HNF4α1 or HNF4α2 (Figure 4B) in MC3T3-E1 osteoblast cultures identified 579 and 819 genes directly regulated by HNF4α1 and HNF4α2, respectively, in mouse bones (Figure 6E). Subsequent pathway analyses of these transcripts showed that HNF4α2 controlled cytoskeleton remodeling, metabolic, and proinflammatory signaling in bone, whereas HNF4α1 had a milder effect on these pathways, consistent with a different metabolic role (Figure 6F). In aggregate, these data demonstrate the critical role of HNF4α in bone development and structure, mediated mainly by the regulatory effects of HNF4α2 on the transcription of osteogenic, metabolic, and apoptotic gene targets.

Figure 6

Low bone mass is associated with altered osteogenesis and impaired bone metabolism in Hnf4αOc-cKO mice. (A) Canonical pathway analysis of differentially regulated genes identified by RNA-Seq of bone from 6-week-old Hnf4αOc-cKO mice versus WT. (B–D) Heatmap-represented, log-normalized expression of genes identified in the topmost differentially regulated pathways in Hnf4αOc-cKO bone versus WT. (E) Number of genes differentially regulated in bone in Hnf4αOc-cKO versus WT identified by RNA-Seq and directly regulated by HNF4α1 or HNF4α2, obtained from the intersection with previously identified direct HNF4α targets in osteoblast ChIP sequencing in Figure 4B. (F) Canonical pathway analysis of direct HNF4α1 and HNF4α2 gene targets in bone identified in E. In A and F, prediction