Bone phenotyping of murine hemochromatosis models with deficiencies of Hjv, Alk2, or Alk3: The influence of sex and the bone compartment

Abstract

Osteopenia is frequently observed in patients with iron overload, especially in those with HFE‐dependent hereditary hemochromatosis (HH). Interestingly, not all mouse models of HH show bone loss, suggesting that iron overload alone may not suffice to induce bone loss. In this study, the bone phenotypes of Hjv ^−/− and hepatocyte‐specific Alk2‐ and Alk3‐deficient mice as additional mouse models of HH were investigated to further clarify, how high iron levels lead to bone loss and which signaling mechanisms are operational. Neither male nor female 12‐week‐old Hjv ^−/− mice had an altered trabecular or cortical bone mass or bone turnover, despite severe iron overload. Male 12‐month‐old Hjv ^−/− mice even presented with a higher femoral trabecular bone volume compared to wildtype mice. Similarly, female mice with hepatocyte‐specific Alk2 or Alk3 deficiency did not show an altered bone phenotype at 3, 6, and 12 months of age. Male hepatocyte‐specific Alk3‐deficient mice also had a normal trabecular bone mass at all ages analyzed, despite showing increased bone resorption and decreased bone formation parameters. Interestingly, hepatocyte‐specific Alk2‐deficient mice showed reduced femoral trabecular bone at 6 months of age due to suppressed bone formation. Cortical thickness at the femur was reduced in both, 6‐month‐old male hepatocyte‐specific Alk2‐ and Alk3‐deficient mice. Raising hepatocyte‐specific Alk2‐deficient male mice on an iron‐deficient diet rescued the bone phenotype. Taken together, despite iron overload, trabecular bone microarchitecture was not altered in mice deficient of Hjv or Alk3. Only male hepatocyte‐specific Alk2‐deficient mice showed site‐specific lower trabecular and cortical bone mass at the femur, which was dependent on iron. Thus, bone loss does not correlate with the extent of iron overload in these mouse models, but may relate to the amount of iron‐loaded macrophages, as precursors of osteoclasts, in the bone marrow.

Iron overload is most severe in Hjv−/− and Alk3fl/fl; Alb‐cre mice, followed by Alk2fl/fl; Alb‐Cre mice and WT mice. Severe hepcidin suppression in Hjv−/− and Alk3fl/fl; Alb‐cre mice leads to lower numbers of iron‐loaded macrophages in the bone marrow. Bone loss is largest in Alk2fl/fl; Alb‐Cre mice, which have the weakest hepcidin suppression and highest number of iron‐loaded cells in the bone marrow.

Abbreviations

Alk2

BMP type I receptor (Acvr1)

Alk3

BMP type I receptor (Bmpr1a)

ANOVA

analysis of variance

BFR/BS

bone formation rate/bone surface

BMP

bone morphogenetic protein

BV/TV

bone volume/total volume

CTX

cross‐linked C‐telopeptide of type I collagen

Hamp

Hepcidin

HFE

High Fe

HJV

Hemojuvelin

kVp

kilo electron voltage

MAR

mineral apposition rate

MS/BS

mineralizing surface/bone surface

µM

micro molar

ms

milliseconds

P1NP

pro‐collagen type I amino‐terminal propeptide

PPM

Parts per million

SD

standard deviation

TRAP5B

tartrate resistant acid phosphatase 5b

TFR2

tranferrin receptor 2

1. INTRODUCTION

Iron is essential for several physiological processes. As such, well‐balanced iron levels are indispensable for health and both, iron deficiency as well as iron overload result in major health problems. While iron deficiency causes anemia, weakness, and shortness of breath, iron overload can lead to failure of the heart, pancreas, and liver due to the accumulation of iron into these organs. ¹ , ² In addition to cardiomyopathy, diabetes, and liver failure, osteoporosis and an increased rate of fractures have also been observed in patients with hemochromatosis. ³ , ⁴ , ⁵ , ⁶ , ⁷ , ⁸ , ⁹ In HFE‐dependent hereditary hemochromatosis (HH), osteoporosis is present in up to 34% of the patients, ¹⁰ , ¹¹ , ¹² with men being more susceptible to an increased risk for osteoporosis. ¹³ Besides osteoporosis, also other skeletal manifestations such as osteonecrosis and bone marrow edema have been reported in patients with HFE‐HH. ¹⁴ Although the underlying disease mechanisms are not fully understood, similar to other hemochromatosis‐related pathologies (e.g., liver disease, cardiomyopathy), the occurance of osteoporosis is associated with the severity of iron overload. ¹⁰ , ¹⁵ , ¹⁶ , ¹⁷ , ¹⁸ , ¹⁹ , ²⁰

Systemic iron homeostasis is mainly controlled by the hepatic hormone hepcidin, which regulates iron availability in the blood and thus iron distribution in the body by blocking the sole iron exporter ferroportin. Several other iron regulatory proteins have been characterized, with many of them being involved in the regulation of the signaling pathway involved in hepcidin expression, the bone morphogenetic protein‐son of mother against decapentaplegic homolog (BMP‐SMAD) signaling pathway. Deficiency of these iron regulators leads to abnormally low hepcidin levels and iron overload of various grades. As reviewed by Bartnikas et al., hepatocyte‐specific deficiency of the BMP type I receptor Alk2 and Hfe in mice leads to moderate iron overload compared to wildtype littermates, ²¹ while in contrast, global or hepatocyte‐specific deficiency of Alk3, Hjv (hemojuvelin), and Hamp (hepcidin) are characterized by severe iron overload with up to five‐fold elevated liver iron content compared to controls due to severe depletion of hepcidin levels. ²¹ , ²² , ²³ , ²⁴ , ²⁵ Of note, also deficiency of the BMP ligands Bmp2 and Bmp6 in liver sinusoidal endothelial cells leads to severe iron overload.

In an effort to better understand the underlying mechanisms of iron‐induced bone loss, various mouse models of hemochromatosis have been employed. Initial studies of Hfe ^−/− mice showed low bone mass associated with reduced bone formation and increased numbers of osteoclasts. ²⁶ , ²⁷ However, no bone phenotype was observed in subsequent studies using Hfe ^−/− , liver‐specific Hfe as well as osteoblast‐ and osteoclast‐specific Hfe knockout mice of various ages. ²⁸ , ²⁹ Interestingly, the lack of bone loss is not only evident in Hfe‐deficient mice, but also in mice lacking the iron sensor transferrin receptor 2 (Tfr2). Unexpectedly, these mice display a high bone mass phenotype that is independent of iron overload, being present also in animals lacking Tfr2 in osteoblasts, demonstrating a cell type‐specific role of TFR2. ³⁰ Deficiency of hepcidin or the use of hepcidin‐resistant ferroportin transgenic mice was shown to cause inhibition of bone formation and low bone mass in mice and zebrafish. ³¹ , ³² , ³³ , ³⁴ At the same time, mice overexpressing hepcidin in hepatocytes or osteoblasts also lead to bone loss due to increased osteoclast activity. ³⁵ Finally, interference with the BMP signaling pathway led to various bone outcomes: blockade of BMP signaling with dorsomorphin resulted in iron overload and suppression of bone formation. ³⁶ Global deficiency of Bmp6 resulted either in low ³⁷ or normal ³⁸ trabecular bone mass, and finally, deficiency of Bmp2 in liver sinusoidal endothelial cells showed no gross skeletal abnormalities. ³⁹ Overall, these studies do not yet provide a clear picture about the underlying mechanisms of iron‐induced bone loss.

Thus, in this study, we aimed to investigate the bone phenotype of three additional mouse models of genetic iron overload: Hjv ^−/− mice, which are amongst the most severely iron overloaded mice, as well as mice with a hepatocyte‐specific deficiency of Alk2 (moderate iron overload) and of Alk3 (severe iron overload comparable to Hjv ^−/− mice).

2. MATERIALS AND METHODS

2.1. Mouse cohorts

Hepatocyte‐specific Alk2‐ or Alk3‐deficient mice: Mouse breeding and animal procedures were approved and conducted in compliance with the guidelines of the institutional animal care committee and the Regierungspräsidium Darmstadt (permit number FK/2028). Male and female hepatocyte‐specific Alk2 (Alk2 ^fl/fl ; Alb‐Cre) or Alk3 (Alk3 ^fl/fl ; Alb‐Cre) conditional knockout mice and Cre‐ littermates on a C57BL/6J background were fed a standard rodent diet (188 ppm iron) with water ad libitum. ⁴⁰ A cohort of male hepatocyte‐specific Alk2 (Alk2 ^fl/fl ; Alb‐Cre) conditional knockout mice also received an iron‐deficient diet (<10 ppm iron, Ssniff) from weaning (around 21 days old) until sacrifice (6 months old). Mice were exposed to a 12 h light/dark cycle and an air‐conditioned SPF room at 23°C (no specific pathogen‐free room). Enrichment was provided in the form of cardboard houses and bedding material. Mice were euthanized at the age of 12 weeks, at 6 or 12 months under deep anesthesia and blood, organs and bones were collected for further analysis.

Hjv global knockout mice: Male and female, 12‐week‐old and 12‐month‐old Hjv ^−/− mice and WT mice on an inbred 129S6/SvEvTac background were used and housed under a standard 12‐hour light/dark cycle with water and chow ad libitum in a pathogen‐free animal facility of the San Raffaele Scientific Institute in accordance with the European Union guidelines (IACUC n° 1026). ⁴¹ Mice were fed a standard rodent diet (Safe diets, Essingen, Germany, irradiated complete universal vegetal diet, 150 SP‐25 with 188 ppm iron) with water ad libitum and analyzed at 12 weeks of age.

2.2. Iron measurements

Non‐heme iron content in the livers was measured using the bathophenanthroline colorimetric method as previously described. ³⁰ In brief, 100 mg of liver tissue was dried for 3 days at 37°C, and afterwards the samples were digested, followed by the reaction with 0.01% bathophenanthrolinedisulfonic acid. Values were recorded spectrophotometrically at 535/540 nm. Non‐heme iron content is reported as μg iron/g dry tissue weight.

Further, iron and transferrin saturation were measured in the serum using the total iron binding capacity kit from Randox as previously described. ³⁰

2.3. μCT analysis of bone microarchitecture

The distal femur and the fourth lumbar vertebrae were excised, and their bone microarchitecture was measured using a vivaCT40 (ScancoMedical, Switzerland). Images were taken at an isotropic voxel‐size of 10.5 μm with X‐ray energy of 70 kVp/114 μA and 200 ms of integration time. For femora, half the femur was scanned, and 100 slices below the growth plate of the distal femur was evaluated for trabecular bone, and 150 slices in the mid‐diaphysis were evaluated for cortical bone. For the vertebral bone, the entire 4th lumbar vertebrae was measured, and 100 slices in the middle of the bone were measured. Trabecular and cortical bone parameters were assessed using standard protocols from Scanco Medical. μCT parameters are reported according to international guidelines.

2.4. Bone histomorphometry

Hepatocyte‐specific Alk2;Alb‐Cre‐ and Alk3‐Alb‐Cre‐deficient mice, Hjv ^−/− mice, and control animals received two intraperitoneal injections with 20 mg/kg calcein (Sigma) five and 2 days before sacrifice. For dynamic bone histomorphometry, the third and fourth lumbar vertebrae were fixed in 4% PBS‐buffered paraformaldehyde and dehydrated in an ascending ethanol series. Subsequently, bones were embedded in methacrylate and cut into 7 μm sections to assess the fluorescent calcein labels. Sections were analyzed using fluorescence microscopy to determine the mineralized surface/bone surface (MS/BS), the mineral apposition rate (MAR), and the bone formation rate/bone surface (BFR/BS).

To determine numbers of osteoclasts, the fifth lumbar vertebra was decalcified for 1 week using Osteosoft (Merck), dehydrated, and embedded into paraffin. Tartrate‐resistant acid phosphatase (TRAP) staining was used to identify osteoclasts. These sections were also used to quantify the number of osteoblasts based on their location and morphology. Bone sections were analyzed using the Osteomeasure software (Osteometrics, USA).

Iron was stained on paraffin‐embedded bone sections using Perls' Prussian blue staining. ⁴²

2.5. Serum analysis of bone turnover markers

Concentrations of pro‐collagen type I amino‐terminal propeptide (P1NP) and tartrate resistant acid phosphatase 5b (TRAP5b) were quantified in the serum using enzyme‐linked immunosorbent assays from Immundiagnostik, Bensheim, Germany.

2.6. Statistical analysis

Data are presented as mean ± standard deviation (SD). Two‐way ANOVA for comparison of age and genotype or a two‐sided unpaired Student's t‐tests were used to compare two groups. Calculations were performed using GraphPad Prism 8 (GraphPad Software Inc., San Diego, CA, USA). p‐values <.05 were considered statistically significant.

3. RESULTS

3.1. Hjv knockout mice do not display bone loss despite iron overload

To investigate whether Hjv ^−/− mice, which are severely iron overloaded, exhibit bone loss, we conducted a microarchitectural, histological, and serological characterization of bone. Analysis of the liver iron content confirmed that both, 12‐week‐old male and female Hjv ^−/− mice had exceedingly high levels of iron stored in the liver (Figure 1A). Perl's Prussian Blue staining revealed iron‐loaded macrophages in the bone marrow of male wildtype mice, but much fewer in male Hjv ^−/− mice (wildtype: 48 ± 57/mm², Hjv ^−/− : 2.50 ± 1.75/mm², p = .0042) (Figure 1B), as expected with hepcidin being strongly reduced and ferroportin stabilized in Hjv ^−/− mice. No staining or iron was found adjacent to the bone surface (Figure 1B). Neither male nor female Hjv ^−/− mice displayed alterations in the femoral cortical thickness (Figure 1C). At the trabecular bone compartment, no changes in the bone volume fraction were observed in the distal femur (Figure 1D). However, trabecular number was decreased in male Hjv ^−/− mice, whereas trabecular separation was increased and trabecular thickness was unchanged compared to wild‐type littermates (Figure 1E–G). No changes in trabecular bone structure were observed in the fourth vertebral body (Figure 1H–K).

FIGURE 1.

Bone phenotype of Hjv ^−/− mice. Twelve‐week old male and female Hjv^−/− and wildtype mice were analyzed. (A) Liver iron content, (B) iron staining of the bone/bone marrow, (C) cortical thickness (Ct.Th) of the femoral mid‐diaphysis, (D) trabecular bone volume/tissue volume (BV/TV), (E) trabecular number (Tb.N), (F) trabecular thickness (Tb.Th), (G) trabecular separation (Tb.Sp) of the distal femur. (H) Trabecular bone volume/tissue volume (BV/TV), (I) trabecular number (Tb.N), (J) trabecular thickness (Tb.Th), (K) trabecular separation (Tb.Sp) of the fifth lumbar vertebra. Individual dots represent individual mice. Mean and SD are indicated as horizontal lines. A two‐sided t‐test was used for statistical analysis. *p < .05; ***p < .001.

In line with the bone microarchitectural data, serum levels of the bone turnover markers CTX (indicating bone resorption) and P1NP (indicating bone formation) in males and females, as well as numbers of osteoclasts and osteoblasts at the femur and spine as assessed using histology in male bones, were not different between the genotypes (Figure 2A–F).

FIGURE 2.

Bone turnover in Hjv ^−/− mice. (A) Serum C‐terminal telopeptide of type I collagen (CTX) levels and (B) serum procollagen type I N‐terminal peptide (P1NP) levels in male and female 12‐week‐old Hjv^−/− and wildtype littermate controls. (C–F) The number of osteoblasts per bone perimeter (N.Ob/B.Pm) and the number of osteoclasts per bone perimeter (N.Oc/B.Pm) was assessed at the fifth lumbar vertebra and distal femur of male mice. Individual dots represent individual mice. Mean and SD are indicated as horizontal lines. A two‐sided t‐test was used for statistical analysis to compare genotypes of the same sex.

To investigate if iron overload induces bone loss over time, 12‐month‐old Hjv ^−/− mice were analyzed as well. As shown in Figure S1, despite increased liver iron levels in male and female Hjv ^−/− mice, they did not show trabecular or cortical bone loss. In fact, males even presented with a higher trabecular bone volume at the distal femur with no significant alterations in trabecular parameters, compared to age‐matched WT controls (Figure S1). The fourth lumbar vertebrae did not show major alterations (Figure S1D). The serum bone turnover makers CTX and P1NP were unchanged in aged male and female Hjv ^−/− mice (Figure S2A,B). Likewise, osteoclast parameters at the femur and spine were normal in Hjv ^−/− mice; however, the bone formation rate at the femur was increased in Hjv ^−/− mice, suggesting that increased bone formation may be causing the higher femoral bone mass of Hjv ^−/− mice (Figure S2C–F). Taken together, despite severe iron overload, Hjv ^−/− mice do not have a low bone mass phenotype.

3.2. Hepatocyte‐specific Alk3‐deficient male mice present cortical bone loss associated with higher number of osteoclasts and lower bone formation

Next, we examined the bone phenotype in mice with hepatocyte‐specific Alk3 deficiency, which exhibit iron‐overload comparable to the genetic hemochromatosis model Hjv ^−/− mice. As previously published, the liver iron content was higher amongst the male Alk3 ^fl/fl ; Alb‐Cre mice across all ages compared to Alk3 ^fl/fl controls (Figure 3A). Moreover, serum iron levels (incl. ferrous and ferric iron) and transferrin saturation were higher in male Alk3 ^fl/fl ; Alb‐Cre mice at 6 months of age (Table 1). Similar to Hjv ^−/− mice, Alk3 ^fl/fl ; Alb‐Cre mice showed a reduced number of iron‐loaded macrophages in the bone marrow (Figure 3B). Trabecular bone volume and structural parameters at the lumbar spine and the distal femur were not changed in male Alk3 ^fl/fl ; Alb‐Cre mice at any age investigated, although the femoral bone volume showed tendencies to be increased (Figure 3C–F, H–K). Interestingly, cortical thickness was decreased in Alk3 ^fl/fl ; Alb‐Cre mice at 6 months of age and tended to be decreased at 12 months of age (Figure 3G).

FIGURE 3.

Bone microarchitecture of Alk3 ^fl/fl ; Alb‐Cre mice assessed with μCT. Male and female Alk3 ^fl/fl ; Alb‐Cre positive and negative mice of different ages were analyzed. (A) Liver iron content and (B) iron‐stained cells in the bone marrow with representative images. Scale bar: 100 μm. (C) Trabecular bone volume/tissue volume (BV/TV), (D–F) trabecular number, thickness and separation at the distal femur, and (G) cortical thickenss at the femoral midshaft, (H) trabecular bone volume, (I–K) trabecular number, thickness and separation at the fourth vertebral body. Individual dots represent individual mice. Mean and SD are indicated as horizontal lines. A two‐sided t‐test was used for statistical analysis to compare genotypes of the same age. *p < .05, ***p < .001.

TABLE 1.

Serum iron levels and transferrin saturation in male 12‐week‐old Hjv^−/− mice and male 6‐month‐old hepatocyte‐specific Alk2‐ and Alk3‐deficient mice.

	WT	Hjv −/−	Alk3 fl/fl	Alk3 fl/fl ; Alb‐Cre	Alk2 fl/fl	Alk2 fl/fl ; Alb‐Cre
Serum total iron (μM)	79.9 ± 17.7	130.0 ± 42.9**	50.8 ± 16.5	84.0 ± 9.5*	52.6 ± 8.4	59.9 ± 8.8
Serum ferrous iron (μM)	35.8 ± 9.4	51.4 ± 26.9	24.4 ± 2.7	43.5 ± 7.6**	21.7 ± 1.1	29.8 ± 3.4**
Serum ferric iron (μM)	44.2 ± 10.2	78.6 ± 17.9***	26.4 ± 18.2	40.5 ± 13.5*	30.9 ± 7.8	30.2 ± 6.9
Transferrin saturation (%)	n.a.	n.a.	33.3 ± 8.2	84.8 ± 12.5***	39.2 ± 12.3	64.4 ± 19.4*

^* p < .05;

^** p < .01;

^*** p < .001.

We next performed serological and histological bone analyses of the bones of 6‐month‐old Alk3 ^fl/fl ; Alb‐Cre mice. Despite showing no alterations in bone microarchitecture, the bone resorption marker CTX and the number of osteoclasts in the femur and spine were both increased by 35–45% in Alk3 ^fl/fl ; Alb‐Cre mice (Figure 4A–D). While the bone formation marker P1NP was not changed in Alk3 ^fl/fl ; Alb‐Cre mice, the number of osteoblasts at the spine (but not the femur) decreased by half in Alk3 ^fl/fl ; Alb‐Cre mice (Figure 4E–G). In line with these findings, dynamic bone histomorphometry showed fewer calcein labels in the fourth vertebral body of Alk3 ^fl/fl ; Alb‐Cre mice compared to controls (Figure 4H). Quantification of the labels showed that the mineralizing surface per bone surface and the mineral apposition rate tended to be decreased, resulting in a significantly dereased bone formation rate (Figure 4I–K).

FIGURE 4.

Histological and serological analysis of bone turnover of male Alk3 ^fl/fl ; Alb‐Cre mice. Bone turnover parameters were analyzed in the serum and at the lumbar vertebrae and femur of male 24‐week‐old Alk3 ^fl/fl ; Alb‐Cre mice. (A) Serum C‐terminal telopeptide of type I collagen (CTX) levels, (B) number of osteoclasts per bone perimeter (N.Oc/B.Pm) in the fourth lumbar vertebral body, (C) representative images of TRAP‐stained vertebral sections, scale bar: 100 μm, and (D) number of osteoclasts per bone perimeter in the femoral bone. (E) Serum procollagen type I N‐terminal peptide (P1NP) levels, (F–G) number of osteoblasts per bone perimeter (N.Ob/B.Pm) in spine and femur, and (H) representative images of calcein vertebral sections, scale bar: 100 μm. (I) Mineralizing surface per bone surface (MS/BS), (J) mineral apposition rate (MAR), and (K) bone formation rate per bone surface (BFR/BS) at the fourth vertebral body. Individual dots represent individual mice. Mean and SD are indicated as horizontal lines. A two‐sided t‐test was used for statistical analysis. *p < .05.

Taken together, hepatocyte‐specific deficiency of Alk3 does not only lead to iron overload in blood and organs, but also altered bone remodeling. Interestingly, this did not lead to changes in trabecular bone microarchitecture, but to femoral cortical bone loss at 6 months of age.

3.3. Hepatocyte‐specific Alk2‐deficient male mice show femoral bone loss

We next analyzed the bone phenotype of Alk2 ^fl/fl ; Alb‐Cre mice, which have a milder iron overload phenotype compared to Hjv ^−/− and Alk3 ^fl/fl ; Alb‐Cre mice, as displayed by higher liver iron levels (Figure 5A) as well as higher serum iron and transferrin saturation levels (Table 1) as compared to their Alk2 ^fl/fl littermate controls. In contrast to Hjv ^−/− and Alk3 ^fl/fl ; Alb‐Cre mice, Alk2 ^fl/fl ; Alb‐Cre mice showed a higher number of iron‐stained cells in the bone marrow (Figure 5B). Male 6‐month‐old Alk2 ^fl/fl ; Alb‐Cre mice showed a 50% reduction in trabecular bone volume, along with a reduction in trabecular thickness, while trabecular number and separation were unchanged (Figure 5C–F). Cortical thickness at the femoral midshaft was reduced by 10% (Figure 5G). No changes in femoral bone microarchitecture were observed in 3‐month‐old male Alk2 ^fl/fl ; Alb‐Cre mice (Figure 5C–G). Besides reduced trabecular thickness, no alterations in trabecular bone structure were observed in the fourth lumbar vertebrae of 3‐ or 6‐month‐old Alk2 ^fl/fl ; Alb‐Cre male mice (Figure 5H–K). Female Alk2 ^fl/fl ; Alb‐Cre mice showed no changes in their trabecular or cortical bone mass, neither at the femur nor at the spine (Figure S3).

FIGURE 5.

Bone microarchitecture of Alk2; Alb‐cre cKO mice assessed with μCT. Male Alk2 ^fl/fl controls and Alk2 ^fl/fl ; Alb‐Cre mice of different ages were analyzed. (A) Liver iron content and (B) iron‐stained cells in the bone marrow with representative images. Scale bar: 100 μm. (C) Trabecular bone volume/tissue volume (BV/TV), (D–F) trabecular number, thickness and separation at the distal femur, and (G) cortical thickness at the femoral midshaft, (H) trabecular bone volume, (I–K) trabecular number, thickness and separation at the fourth vertebral body. Indiviual dots represent individual mice. Mean and SD are indicated as horizontal lines. A two‐sided t‐test was used for statistical analysis to compare genotypes of the same age and sex. *p < .05, **p < .01.

At a cellular level, male Alk2 ^fl/fl ; Alb‐Cre mice showed lower serum CTX levels, while no changes in osteoclast parameters were found histologically at the femur or spine (Figure 6A–D). Likewise, the number of osteoblasts and the mineralizing surface were not different between the genotypes at the spine or femur, but serum levels of P1NP, the mineral apposition rate, and the bone formation rate were significantly decreased in Alk2 ^fl/fl ; Alb‐Cre mice (Figure 6E–K).

FIGURE 6.

Histological and serological analysis of bone turnover of male Alk2 ^fl/fl ; Alb‐Cre mice. Bone turnover parameters were analyzed in the serum and at the lumbar vertebrae of male 24‐week‐old Alk2 ^fl/fl ; Alb‐Cre mice. (A) Serum C‐terminal telopeptide of type I collagen (CTX) levels, (B) number of osteoclasts per bone perimeter (N.Oc/B.Pm) in the fourth lumbar vertebral body, (C) representative images of TRAP‐stained vertebral sections, scale bar: 100 μm, and (D) number of osteoclasts per bone perimeter in the femoral bone. (E) Serum procollagen type I N‐terminal peptide (P1NP) levels, (F and G) number of osteoblasts per bone perimeter (N.Ob/B.Pm) in spine and femur, and (H) representative images of calcein vertebral sections, scale bar: 100 μm. (I) Mineralizing surface per bone surface (MS/BS), (J) mineral apposition rate (MAR), and (K) bone formation rate per bone surface (BFR/BS) at the fourth vertebral body. Individual dots represent individual mice. Mean and SD are indicated as horizontal lines. A two‐sided t‐test was used for statistical analysis. *p < .05.

Taken together, Alk2 ^fl/fl ; Alb‐Cre male mice show reduced trabecular bone parameters at the age of 6 months, likely due to a suppression of bone formation.

3.4. Low iron diet rescues femoral bone loss in hepatocyte‐specific Alk2‐deficient mice

Given the decreased femoral trabecular volume and cortical thickness and the higher number of iron‐stained cells in the bone marrow of 6‐month‐old Alk2 ^fl/fl ; Alb‐Cre‐deficient male mice, we wondered whether femoral bone loss was due to their higher systemic iron levels. Therefore, we fed Alk2 ^fl/fl ; Alb‐Cre male mice a low iron diet after weaning until 6 months of age. This diet prevented the increase in liver iron levels (Figure 7A) as well as the femoral trabecular and cortical bone loss in Alk2 ^fl/fl ; Alb‐Cre mice (Figure 7B–D). Trabecular bone volume and trabecular thickness of the fourth lumbar vertebrae also remained unchanged in Alk2 ^fl/fl ; Alb‐Cre mice on the low iron diet (Figure 7E,F). This is consistent with the measurement of bone turnover markers, which showed no difference between genotypes (Figure 7G,H). Thus, the iron‐deficient diet normalized the femoral bone phenotype of the Alk2 ^fl/fl ; Alb‐Cre mice, suggesting that the higher iron load drives femoral bone loss.

FIGURE 7.

Iron‐deficient diet rescues the femoral bone phenotype of Alk2 ^fl/fl ; Alb‐Cre mice. Male Alk2 ^fl/fl controls and Alk2 ^fl/fl ; Alb‐Cre mice were put on an iron‐deficient diet (IDD) after weaning. At the age of 24 weeks, mice were analyzed regarding their liver iron content (A) and bone phenotype (B–H). (B) Trabecular bone volume/tissue volume (BV/TV) and (C) trabecular thickness at the distal femur. (D) Cortical thickness (Ct.Th) of the mid‐diaphysis. (E) Trabecular bone volume/tissue volume (BV/TV) and (F) trabecular thickness at the fourth lumbar vertebrae. (G) CTX and (H) P1NP serum levels. Individual dots represent individual mice. Mean and SD are indicated as horizontal lines. A two‐sided t‐test was used for statistical analysis.

4. DISCUSSION

Bone loss has been reported frequently in HH patients as well as in some mouse models of iron overload, including Hfe ^−/− mice, Bmp6 ^−/− mice, and hepcidin‐deficient mice, ³ , ¹⁰ , ²⁷ , ³² , ³⁴ , ³⁷ , ⁴³ although with controvertial results. In contrast, Tfr2 ^−/− mice and mice with a deletion of Bmp2 in liver sinusoidal endothelial cells showed no obvious bone loss, ²⁸ , ²⁹ , ³⁰ , ³⁹ , ⁴⁴ despite hepcidin deficiency and body iron overload.

To better understand under which circumstances iron overload leads to bone loss, we explored the bone phenotypes of Hjv ^−/− as well as hepatocyte‐specific Alk2‐ and Alk3‐deficient mice as additional mouse models with HH. While Hjv ^−/− and hepatocyte‐specific Alk3‐deficient mice are severely iron overloaded and have almost undetectable hepcidin expression levels in the liver, hepatocyte‐specific Alk2‐deficient mice are only moderately iron‐overloaded with lower, but detectable levels of hepcidin mRNA in the liver. ²¹ , ⁴¹ At 12 weeks of age, none of the three models showed trabecular bone abnormalties. Suprisingly, hepatocyte‐specific Alk2‐deficient mice, despite having the lowest iron load, had the strongest bone phenotype at 6 months of age, showing reduced trabecular and cortical bone volume. Six month old male hepatocyte‐specific Alk3‐deficient mice also showed cortical bone loss. At 12 months of age, male Hjv ^−/− and to some extent male hepatocyte‐specific Alk3‐deficient mice even showed an increase in trabecular bone mass. These phenomena were present only in male mice, while female mice showed no bone abnormalities at any of the ages examined.

To date, most studies investigating the effects of iron overload on bone only examined trabecular bone parameters, while cortical bone parameters were often neglected. Focusing on trabecular bone, it was very suprising that no bone loss was observed in any of the three models examined here, despite the extent of iron overload. In fact, older Hjv ^−/− and hepatocyte‐specific Alk3‐deficient mice even had or tended to have higher trabecular bone mass at the femur, reminiscent of the bone phenotype of Tfr2 ^−/− mice, although here, increased bone volume was already present in younger mice and also at the spine. ³⁰ In all three cases (Tfr2 ^−/− , Hjv ^−/− and hepatocyte‐specific Alk3‐deficient mice), the increase in bone volume was higher in the femur than the vertebral bone, suggesting that potentially a factor in the bone marrow compartment may affect bone turnover in the femoral bone. However, this notion needs to be tested in future studies.

Besides the lack of a major trabecular bone phenotype in our models, it was striking to note that cortical bone loss was detected in male hepatocyte‐specific Alk2‐ and Alk3‐deficient mice at 6 months of age. Raising the Alk2 ^fl/fl ; Alb‐Cre mice on an iron‐deficient diet showed that the phenotype was iron‐dependent. So far, not too many studies have investigated the cortical bone compartment in hemochromatosis. Previously, studies in 30‐week‐old Hfe ^−/− mice and in humans with HFE‐HH showed that the cortical bone compartment was negatively affected, while the trabecular bone compartment was not. ¹⁴ , ²⁹ Explanations for the differential susceptibility of cortical vs. trabecular bone may stem from a differential expression of iron regulators in trabecular vs. cortical bone, from a different exposure to iron via the bone vasculature, or different machineries for redox detoxifying systems. None of these possibilities have yet been investigated in detail and remain the subject of future investigations. Finally, the cortical bone phenotype was only detected in males, but not in females. However, these sex differences are in line with previous studies that examined iron effects on bone, which show that males are more severely affected by iron overload than females, which may be due to the fact that females are better able to cope with iron fluctuations and that females are less iron overloaded per se. ¹³ , ⁴⁵

One explanation for the differential cortical bone phenotypes of our models may be the hepatocyte‐specific deletion of Alk2 and Alk3 vs. the global deficiency of Hjv. As for TFR2, which has been shown to have an osteoblast‐intrinsic role in bone remodeling, ³⁰ also HJV is expressed in bone cells at low levels ³⁴ and thus, may participate in the regulation of bone mass. Investigations of iron homeostasis in the liver have revealed that HJV is a BMP co‐receptor, and that HJV signaling is dependent on the BMP type I receptors ALK2 and ALK3. ⁴⁶ Neogenin, which is a multifunctional transmembrane receptor belonging to the immunoglobulin superfamily that forms the core of a signal transduction hub for HJV, has also been shown to play a role in endochondral bone formation by promoting lipid raft localization of repulsive guidance molecules and BMP receptors. ⁴⁷ Thus, with HJV being a repulsive guidance molecule, further studies should aim at investigating bone cell‐specific Hjv‐deficient mice to clarify a potential bone‐intrinsic role of HJV in the regulation of bone remodeling.

Overall, it was surprising that the severely iron overloaded models (Hjv ^−/− and the hepatocyte‐specific Alk3 deficiency) did not show more pronounced bone loss. Spite the large extent of iron overload in the liver, iron did not accumulate in or along the bone tissue and was in fact depleted in macrophages found in the bone marrow of Hjv ^−/− and Alk3 ^fl/fl ; Alb‐Cre mice, in agreement with the stabilization of ferroportin, highly expressed in macrophages, due to hepcidin deficiency and thus less intracellular iron. This is in line with another previous finding in Hfe ^−/− and hepatocyte‐specific Hfe‐deficient mice ²⁹ and with our own previous observations in Hfe ^−/− mice, which also showed lower numbers of iron‐stained macrophages in the bone marrow and no iron staining of bone tissue. ³⁴ However, it is in contrast to previous publications showing iron staining along the trabecular bone in Hfe ^−/− mice. ²⁸ , ³⁷ Interestingly, these studies also show high iron levels in macrophages, which is rather surprising, given that the low expression of hepcidin in Hfe ^−/− mice should result in higher levels of ferroportin in macrophages. Along those lines, in ferroportin (Fpn ^C326S ) mutant mice, in which ferroportin is resistant to hepcidin‐mediated degradation, no iron‐stained macrophages were found at all in the bone marrow. ³⁴ The lack of trabecular bone phenotype in Hjv ^−/− and hepatocyte‐specific Alk3‐deficient mice may be related to the actions of hepcidin, or the deficiency here. Previously, it was hypothesized that the lack of hepcidin in HH models leads to increased release of iron from macrophages into the blood stream that thereafter also deposits in bone. In the case of iron dextran injections, hepcidin levels remain high and lead to limited release of iron into the circulation and no bone loss in this particular model of iron dextran injection. ³⁷ However, an alternative hypothesis may be that the low availability of hepcidin in HH leads to the stabilization of ferroportin on bone cells, allowing them to remove excess intracellular iron, which would otherwise be harmful due to the production of free radicals. This in fact would be supported by the bone phenotypes of hepatocyte‐specific Tfr2 ^−/− and Hfe ^−/− mice, which are iron overloaded, but show no bone loss, ²⁹ , ³⁰ , ³⁹ and the hepatocyte‐specific Alk3‐deficient mice, which also show no major trabecular bone loss. In contrast, administration of iron to animals would lead to high hepcidin levels, low ferroportin expression on bone cells, and thus, the accumulation of iron and its toxic effects in bone cells and eventually to bone loss. Besides the study of Robin et al. that did not show bone loss after iron dextran injection, several studies have shown bone loss in mice and rats treated with iron dextran, likely due to iron‐induced oxidative stress. ⁴⁸ , ⁴⁹ , ⁵⁰ , ⁵¹ One of the main differences between the studies showing bone loss after iron dextran administration, ⁴⁸ , ⁴⁹ , ⁵⁰ , ⁵¹ and the study by Robin et al. showing no bone loss, is the weekly injection of 100–500 mg/kg iron dextran over 12 weeks compared to a single injection of iron dextran (1000 mg/kg) with analyses 4 months later in the Robin study, ³⁷ which may lead to a different level of iron overload. However, all studies have in common that they show an increase in the number of iron‐stained cells (macrophages) in the bone marrow, indicating that iron accumulates in macrophages after iron treatment. ³⁷ , ⁵¹ While no study has investigated the iron load of bone cells in vivo directly, an increased expression of markers of oxidative stress was found in the bone tissue of iron dextran‐treated mice, which may stem from excess intracellular iron levels. ⁴⁸ Finally, one argument against our proposed working model is that hepcidin‐deficient mice show bone loss mostly due to reduced bone formation and that hepcidin treatment has been shown to protect against iron‐mediated suppression of bone formation and ovariectomy‐induced bone loss. ³² , ³³ , ⁴³ , ⁵² , ⁵³ However, again, opposing results have been published with hepcidin overexpression in hepatocytes or osteoblasts also leading to bone loss due to increased osteoclastogenesis. ³⁵ Currently, there is a lack of consistency in reporting bone phenotypes in iron overload models, such as e.g., detailed μCT analyses (e.g., analyses of at least two different skeletal sites and the differentiation between cortical and trabecular bone), provision of dynamic histomorphometry parameters and bone turnover markers, as well as bone iron concentrations, so that it remains challenging to produce a unifying working model about the effects of iron on bone at this point.

In summary, our data show that Hjv ^−/− mice with severe iron overload show no bone loss. In contrast, male, but not female, hepatocyte‐specific Alk2‐ or Alk3‐deficient mice show cortical bone loss with minor alterations in trabecular bone mass. Loss of bone mass in hepatocyte‐specific Alk2‐deficient mice is iron‐dependent. Whether the hepcidin deficiency protects trabecular bone from iron toxicity remains to be investigated in future studies.

AUTHOR CONTRIBUTIONS

Conceptualization: DYD, LS, MR, AUS; Data curation: DYD, IH, LS, UB, GS, MR, AUS; Formal analysis and methodology: DYD, IH, LS, UB, GS, MR, AUS; Funding acquisition: LCH, UB, MR, AUS; Supervision: MR and AUS; Writing—original draft: DYD, MR, AUS; Writing—review & editing: all authors.

FUNDING INFORMATION

This work was supported by grants from the DFG (FerrOs‐FOR 5146) to LCH, UB, GS, MR, and AUS.

DISCLOSURES

MR reports honoraria from UCB, Santhera, and Vifor for lectures and advisory boards. LH reports honoraria from Amgen, UCB, Ascendis, and Pharmacosmos for lectures and advisory boards. AUS reports a research grant from Pharmacosmos and from HemoClear, each for an investigator‐driven clinical trial. All other authors declare no conflicts of interest.

Supporting information

Figure S1.

Figure S2.

Figure S3.

Text S1.

Dogan DY, Hornung I, Pettinato M, et al. Bone phenotyping of murine hemochromatosis models with deficiencies of Hjv, Alk2, or Alk3: The influence of sex and the bone compartment. The FASEB Journal. 2024;38:e70179. doi: 10.1096/fj.202401015R

DATA AVAILABILITY STATEMENT

The data that support the findings of this study are available on request from the corresponding author (MR).