Intergenerational transmission of diet-induced obesity, metabolic imbalance, and osteoarthritis in mice

Abstract

Objective:

Obesity and osteoarthritis (OA) are two major public health issues affecting millions of people worldwide. While parental obesity affects the predisposition to diseases such as cancer or diabetes in children, transgenerational influences on musculoskeletal conditions such as OA are poorly understood.

Methods:

Metabolic phenotype and predisposition to OA of the first (F1) and second (F2) generation of offspring (n=10–16 per sex/per diet) from mice fed high-fat or control diets were investigated. OA was induced by destabilizing the medial meniscus. OA, synovitis, and adipose tissue inflammation were determined histologically, while bone changes were measured using microCT. Serum and synovial cytokines were measured by multiplex assay.

Results:

Parental high-fat feeding showed an intergenerational effect on the inheritance of increased weight gain (up to 19% in F1 and 9% in F2 generation), metabolic imbalance, and injury-induced OA in mice for at least two generations despite offspring being fed a low-fat diet. Strikingly, both F1 and F2 female mice showed increased predisposition to injury-induced OA (48% higher predisposition in F1 and 19% in F2 HFD Females) and bone microarchitectural changes due to parental and grandparental high-fat feeding.

Conclusion:

Our study reveals a detrimental role of parental high-fat diet and obesity on the musculoskeletal integrity of two generations of offspring and indicates the importance of the further investigation. An improved understanding of the mechanisms involved in the transmissibility of diet-induced changes through multiple generations may help in the development of future therapies targeting the effects of obesity on OA and related conditions.

INTRODUCTION

Osteoarthritis (OA) is the most common arthritic joint disease and the top indication for total joint replacement surgery, affecting more than 250 million people worldwide (1). One of the primary risk factors for OA is obesity and its associated metabolic syndrome (2). However, the mechanisms linking obesity and OA are multifactorial and are not fully understood. Whereas it was originally believed that the link between OA and obesity was simply due to increased “wear-and-tear” of excessive joint loading, it is now apparent that numerous additional factors associated with obesity, such as systemic inflammation or a high-fat diet, may also be critical mediators of OA (3). Additionally, a familial predisposition to OA in humans has been confirmed to possess strong polygenetic and environmental risk (4). Previous animal studies have proven that diet-induced obesity greatly affects the progression of OA. For instance, a diet rich in saturated fat with an excess of pro-inflammatory fatty acids (FAs) (i.e., omega-6 FAs), increases the predisposition to OA (5, 6).

In addition to the direct effects of obesity on OA, growing evidence is now showing that prenatal and postnatal development are affected by both nutrition and environmental stimuli, particularly in the context of obesity (7). The health of future offspring is thought to be influenced by diet and exercise aspects of parental lifestyle (8). Some diseases in children such as cancer (9) and type 2 diabetes mellitus (T2DM) have been shown to correlate with both maternal (10) and paternal obesity (11). Several mechanisms are postulated to play an important role in mediating these changes, including epigenetic modifications of reproductive cells (12), disturbance in hypothalamic hunger/satiety signaling (10), and finally the role of intrauterine (13) and maternal microbiome milieu (14) during development.

These previous findings suggest that parental obesity may be an important factor in the heritability of obesity and metabolic dysfunction, and therefore may influence the pathogenesis and susceptibility to musculoskeletal diseases such as OA. We studied the intergenerational effects on the metabolic and skeletal phenotype, systemic inflammation, and the predisposition to OA in two generations of offspring (F1 and F2) bred from mice fed either low- or high-fat diet (HFD) (Fig. 1A). We hypothesized that, despite a control low-fat diet post-weaning, the offspring would exhibit disturbed metabolism and “obese-like” phenotype, thus have an increased the predisposition to develop injury-induced OA.

Fig. 1. Parental HFD affects weight gain, adiposity, and metabolic profile of F1 generation offspring.

(A) Study design: F0 generation mice were fed either a control diet (10% fat by kcal) or a high-fat diet (60% fat by kcal). F1 and F2 generations were fed a control diet after weaning. Destabilization of medial meniscus (DMM) surgery was performed on the left knee at 16 weeks, and mice were microCT scanned at 26 weeks for body fat analysis. Animals were sacrificed at 28 weeks for further analysis. (B) Longitudinal body weight measurement. (C) Adipose tissue content assessment by microCT. (D) Representative histology (H&E) staining of subcutaneous adipose tissue (SAT), scale bar = 100μm. (E) Serum levels for adipokines (leptin, insulin, and glucagon). All data are presented as mean ± 95% CI, n=10–16 per sex, per parental diet group. One-Way Repeated Measures ANOVA (A) or Mann-Whitney U test (B, D), § p<0.001, ** p<0.01, * p<0.05.

MATERIALS AND METHODS

Animal Studies and Experimental Design

All experimental procedures were approved by and conducted in accordance with the Institutional Animal Care and Use Committee guidelines of Washington University in St. Louis. C57BL/6J mice were purchased from Jackson Lab (Bar Harbor, Maine). Mice were housed in 12h light/12h dark cycle with ad libitum water and food access. At 4 weeks of age, mice were allocated either on Control Diet (Research Diets #D11120103) or HFD (Research Diets #D11120105) for 11–15 weeks. The content of this diet (see Supplement Table 1 (Table S1)) has been reported previously (5). Two groups of breeding pairs (F0 breeders) were formed: F0 Control (F0 mice receiving Control diet) and F0 HFD (F0 mice receiving HFD) (n=6 per each breeding group). The F1 generation of offspring (both sexes) was collected forming F1 Control and F1 HFD groups. A subset of F1 mice (n=4 per each breeding group) was also used to create the F2 generation (both sexes) forming F2 Control and F2 HFD groups. The F1 breeder mice were siblings from the same litter to the F1 offspring used for the main experiments. One litter per each breeding pair was used and n=1–5 offspring per litter per sex was collected. F1 and F2 groups were allocated on the Control Diet throughout the study (Fig. 1A). Offspring were housed separately from the breeders after weaning, and siblings from the same litter were housed together. F1 and F2 offspring groups underwent destabilization of medial meniscus (DMM) surgery at 16 weeks of age, and animals were sacrificed at 28 weeks of age. At sacrifice, the knee joints, visceral (VAT) and subcutaneous (SAT) adipose tissue, serum, and synovial fluid were collected.

Body weight and composition

Mice were weighed bi-weekly. The body fat content was measured at 26 weeks of age using the microCT scanner (microCT, SkyScan 1176, Bruker) with a 35 μm isotropic voxel resolution (357 μA, 500 ms integration time, 1 frame averaging). A 1 mm aluminum filter was used to reduce the effects of beam hardening. Images were reconstructed using NRecon software (with 30% beam hardening and 20 ring artifact correction). For in vivo scans, mice were anesthetized by 2–3% isoflurane inhalation. Scans were reconstructed for the abdominal region (between the proximal end of L1 and the distal end of L6, as shown in Supplement Figure 1 (Fig. S1) to quantify representative AT content in each animal. Adipose tissue percentage was measured using a modified script provided by CTan software (SkyScan 1176, Bruker) which delineated AT threshold from the other tissues as previously published (15). Briefly, the algorithm separated the mouse body from the background to provide the total tissue volume (TV), while AT threshold delineated total fat mass from lean mass and bone (Supplement Fig. S1), forming AT percentage values for each scanned mouse.

OA and synovitis assessment

At 16 weeks of age, all mice underwent surgery to destabilize the medial meniscus (DMM) to induce knee OA in the left hind limb as previously described. To evaluate the effects of parental diet on OA development, mice were sacrificed at 28 weeks of age. Briefly, both left and right hind-limbs were harvested and fixed in 10% neutral buffered formalin (NBF). Limbs were then decalcified in Cal-Ex solution (Fisher Scientific, Pittsburgh, PA, USA), dehydrated and embedded in paraffin. The joint was sectioned in the coronal plane at a thickness of 8 μm. Joint sections were stained with hematoxylin, fast green, and Safranin-O to determine OA and osteophytes formation. For synovitis, sections were stained with hematoxylin and eosin (H&E) to reveal infiltrated cells and synovial structure. Three independent, blinded graders scored joint sections for OA using a modified Mankin scoring system (16), osteophytes and synovitis using previously established protocol (17). Scores were averaged among graders for the whole joint (Total Mankin Score, synovitis) or medial part of the joint (Total Medial Osteophytes).

Serum and Synovial Fluid cytokine levels

Serum and synovial fluid (SF) from the DMM and contralateral control limbs were collected as described previously (5). For cytokine and adipokine levels in the serum and SF fluid, a multiplexed bead assay (Discovery Luminex 31-plex, Eve Technologies, Calgary, AB, Canada) was used to determine the concentration of Eotaxin, G-CSF, GM-CSF, IFNγ, IL-1α, IL-1β, IL-2, IL-3, IL-4, IL-5, IL-6, IL-7, IL-9, IL-10, IL-12 (p40), IL-12 (p70), IL-13, IL-15, IL-17A, IP-10, KC, LIF, LIX, MCP-1, M-CSF, MIG, MIP- 1α, MIP-1β, MIP-2, RANTES, TNF-α, and VEGF. A Mouse Metabolic Array (MRDMET, Eve Technologies, Calgary, AB, Canada) was used to measure levels for Amylin, C-Peptide, ghrelin, glucagon, insulin, leptin, PP, PYY, and resistin in serum.

Bone microstructural analysis

Bone structural and morphological changes were measured in intact hindlimbs by microcomputed tomography (microCT, SkyScan 1176, Bruker) with an 8.75 μm isotropic voxel resolution (500 μA, 500 ms integration time, 4 frame averaging). A 1 mm aluminum filter was used to reduce the effects of beam hardening. Images were reconstructed using NRecon software (with 30% beam hardening and 20 ring artifact correction). Subchondral/trabecular regions were segmented using CtAn automatic thresholding software. Tibial and femoral epiphysis were selected using the subchondral plate and growth plate as references. Tibial metaphysis was defined as the 1mm region directly below the growth plate. Hydroxyapatite calibration phantoms were used to calibrate bone density values (mg/cm³). MicroCT limb images were analyzed for bone volume/total volume (BV/TV), Trabecular Bone Number (Tb.N), Trabecular Bone Thickness (Tb.Th), Trabecular Bone Separation (Tb.Sp), and Bone Mineral Density (BMD).

Immunofluorescent detection of adipose tissue macrophages

Visceral adipose tissues (VAT) were collected, flash-frozen in OCT, and cryosectioned at 5μm thickness. Tissue slides were then acetone-fixed followed by incubation with Fc-receptor blocking in 2.5% goat serum (Vector Laboratories) and incubation with primary antibodies cocktail containing anti-CD11b: AlexaFluor488 and CD11c:PE (Biolegend). Nuclei were stained with DAPI. Samples were imaged using fluorescence microscopy (VS120, Olympus).

Adipose Tissue Histology

Visceral adipose tissues were fixed in 10% neutral buffered formalin, paraffin-embedded and cut into 5 μm sections. Sections were deparaffinized, rehydrated and stained with standard H&E method. Samples were imaged using fluorescence microscopy (VS120, Olympus).

Real-Time PCR

RNA from VAT samples (n = 5–9 samples per group) was obtained using Total RNA Purification Kit (Norgen Biotek, Canada). RNA purity was assessed by NanoDrop absorbance measurement with 260 nm/280 nm and 260 nm/230 nm absorbance ratios of >1.9. One microgram of RNA was reverse transcribed using iSCRIPT (Bio‐Rad, USA) in accordance with the instructions of the suppliers. Real‐time PCR was conducted on a QuantStudio (ThermoFisher, USA). Ten nanograms of cDNA was analyzed. The primer concentration was used at 10 μM. The reactions were performed in duplicate for each analyzed gene. Reactions using SYBR Green chemistry were also subjected to melting curve analysis. The amplification curves and efficiency of amplification of each gene were validated, and the efficiency values ranged from 95% to 105%. Values for target gene expression were normalized to average of three housekeeping genes Rlp32 (ribosomal protein 32), peptidylprolyl isomerase A (Ppia), and glycerylaldehyde-3-phosphate dehydrogenase (Gapdh). Relative messenger RNA (mRNA) expression was assessed using the $2^{-\Delta \Delta {\text{C}}_{\text{t}}}$ method, with the lowest C_t value of the group serving as a calibrator (according to the method described by Schmittgen and Livak (18). The primer sequences used for assessment of gene expression were as follows: Lep (Forward: GGGCTTCACCCCATTCTGAG and Reverse: AAGGCCAGCAGATGGAGGAG, Lepr (Forward: CTGCAGTCTTCGGGGATGTG and Reverse: TGGGCTGCAGTGACATCAGA, Cd36 (Forward: GACGTGGCAAAGAACAGCAG and Reverse: GGCTCCATTGGGCTGTACAA), AdipoQ (Forward: CCCTCCACCCAAGGGAACTT and Reverse: TCCAGGAGTGCCATCTCTGC, AdipoR1 (Forward: TGCCTCAGGAAGAGGAGGAG and Reverse: TTTCAGCCAGTCAGGAAGCA, AdipoR2 (Forward: GCTCAGAAAAGGGCACCAAC and Reverse: ATCATGACACTCGGGCTCCT, Pparg (Forward: GCCCTGGAACTGCAGCTAAA and Reverse: GTGCTCTGTGACGATCTGCCT, Fabp4 (Forward: TGTGATGCCTTTGTGGGAAC and Reverse: ATGCCTGCCACTTTCCTTGT,), Rpl32 (Forward: GGTGAAGCCCAAGATCGTCA and Reverse: CTCCGCACCCTGTTGTCAAT, Ppia (Forward: AGGATTCATGTGCCAGCGTG and Reverse: TTGCCATGGACAAGATGCCA), Gapdh (Forward: GGCAAATTCAACGGCACAGT and Reverse: GTCTCGCTCCTGGAAGATGG).

Statistical analysis

Statistical analyses were performed using IBM SPSS, with significance reported at the 95% confidence level. All data are presented as mean ± 95% confidence interval, Regular Two-Way ANOVA, Mann-Whitney U test, or Repeated Measures Two-Way ANOVA, followed by Bonferroni test were used as indicated in each figure. For Two-Way ANOVA and Mann-Whitney U test, statistical significance was indicated by § p<0.001, ** p<0.01, * p<0.05. For Repeated Measures Two-Way ANOVA, different letters indicate significantly different outcomes, p<0.05. Right (non-DMM) contralateral leg was used as control. All analyses were performed within separate sex groups. Spearman bivariate regression was used to evaluate the association between DMM-operated (left) joint OA Score and other outcomes. The Spearman’s rank-order correlation was conducted between left OA score and 52 separate measurements, including weight, serum cytokines, and adipokines, as well as synovial fluid cytokines.

RESULTS

F1 offspring weight gain, adiposity, and metabolic profile

F1 HFD Female and F1 HFD Male were significantly heavier than their corresponding control diet group (Fig. 1B). F1 HFD Female had a significantly higher percentage of adipose tissue than did the F1 Control Female, and there was a trend (p=0.055) towards increased abdominal fat in F1 HFD Male compared to F1 Control Male (Fig. 1C). We observed increased adipocyte hypertrophy in SAT in both F1 HFD Female and Male compared to their control diet counterparts (Fig. 1D, and Supplement Figure Fig. S2A). Serum leptin level was significantly higher, insulin trended towards higher, while glucagon towards lower in F1 HFD Female compared to F1 Control Female (Fig. 1E). No significant differences were found in leptin, insulin, or glucagon serum adipokines between F1 Male diet groups. Other serum adipokine levels did not differ between diet groups (Supplement Fig. S2).

F1 offspring systemic and local inflammation

Serum IL-1β trended higher in F1 HFD Female compared to F1 Control Female, while there was no change in serum TNFα and IL-6 levels. Serum IL-1α, IL-10, and IP-10 were significantly lower in F1 HFD compared to F1 Control Female. There were trends toward higher IL-17 in F1 HFD compared to F1 Control Female (Fig. 2A). Other cytokines did not differ in analyzed groups (Table S2). Visceral adipose tissue (VAT) from F1 HFD offspring had a higher content of fibrotic areas when compared to their sex-controls (Fig. 2B, and Supplement Fig. S3B). CD11b⁺CD11c⁺ labeling indicated an increased number of M1-like macrophages in F1 HFD Male compared to F1 Control Male in VAT. The number of M1-like macrophages was comparable in the VAT of both Female diet groups (Fig. 2C, and Supplement Fig. S3E). Gene expression analysis of VAT showed significantly higher leptin (Lep) levels and trends towards higher Cd36 levels in F1 HFD Male, but significantly lower Lepr levels in F1 HFD Female compared to their controls. Similar trends in higher Lep expression and significant Cd36 expression was found in SAT in F1 HFD Males compared to F1 Control Males (Fig. 2D).

Fig. 2. Parental HFD affects immune profile of F1 generation offspring.

(A) Serum levels for cytokines (IL-1β, TNFα, IL-6, IL-1α, IL-10, IP-10, Eotaxin, IL-13, and IL-17). (B) Representative histology (H&E) staining of visceral adipose tissue (VAT), scale bar = 100μm, black arrows indicate fibrosis. (C) Immunohistochemistry staining of VAT showing CD11b (green), CD11c (red), and nuclei (blue) of M1 (pro-inflammatory) macrophages, scale bar = 100μm, white arrows indicate double positive cells. (D) Leptin (Lep), Leptin Receptor (Lepr), and Cd36 gene expression in VAT and SAT. All data are presented as mean ± 95% CI, n=6–16 per sex, per parental diet group. Mann-Whitney U test, ** p<0.01, * p<0.05.

Assessment of OA severity, osteophyte formation, and synovitis in F1 generation offspring

A modified Mankin scoring system was used to analyze the severity of DMM-induced OA (5). Two-way ANOVA analysis within each sex revealed that there was a significantly higher predisposition to injury-induced OA in F1 HFD Female, with a similar trend (p=0.058) in F1 HFD Male compared to their F1 Controls (Fig. 3A and C). Osteophyte formation was detected in DMM-joints (Fig. 3A), with Total Medial Osteophyte Score significantly higher in F1 HFD Female compared to F1 Controls (Fig. 3D). Furthermore, although there was no significant interaction, we found that a significant parental diet effect on total synovitis score in F1 HFD compared to F1 Control Female (Fig. 3B and E). There were no significant differences in synovial fluid (SF) cytokine levels in the analyzed F1 offspring group (Fig. 3F). Additionally, we did not observe significant differences in synovial F4/80+ macrophages influx in analyzed groups (Supplement Fig. S4A). Total OA Mankin scores from DMM-operated joint correlated positively with weight and serum leptin in F1 Female, but not in F1 Male. Additionally, serum IP-10 correlated negatively with F1 Female OA score, while serum IL-6 showed a positive correlation with F1 Male OA scores (Fig. 3G and H).

Fig. 3. Parental HFD affects musculoskeletal integrity of F1 generation offspring.

(A) Representative Safranin-O/Fast Green staining of DMM-operated joints, scale bar = 100 μm. (B) Representative H&E staining of the medial femoral condyles of DMM-operated joints, scale bar = 100 μm. (C) Total joint Modified Mankin scores. (D) Total Medial Osteophyte scores. (E) Total Synovitis scores. (F) Synovial fluid (SF) cytokine levels (IP-10, IL-10, IL-1α) from non-operated control (right) and DMM-operated (left) joints. G) Spearman bivariate correlation between DMM-operated (left) joint OA Score and weight, serum leptin levels, and serum IP-10 levels in F1 Female group. (H) Spearman bivariate correlation between DMM-operated (left) joint OA Score and weight, serum leptin levels, and serum IL-6 levels in F1 Male group. (I) Representative microCT reconstructed images of the tibial metaphysis. (J) MicroCT analysis of DMM-operated (left) joint tibial metaphysis showing trabecular bone fraction bone volume/total volume (BV/TV), trabecular bone number (Tb.N), and bone mineral density (BMD). All data are presented as mean ± 95% CI, n=10–16 per sex, per parental diet group. (C-F) Two-way ANOVA with Bonferroni post hoc analysis. Groups not sharing the same letter are significantly different within sex, p < 0.05. Main effect of diet reported as #. Main effect of DMM surgery reported as *. (J) Mann-Whitney U test, * p<0.05.

Analysis of bone microstructure changes in F1 offspring

MicroCT imaging revealed differences in bone microstructure (Fig. 3I). Bone volume/total volume (BV/TV), bone mineral density (BMD), and trabecular bone number (Tb.N) of tibial metaphysis were significantly lower in F1 HFD Female compared to F1 Control Female (Fig. 3J). Additionally, F1 HFD Female displayed significantly lower BV/TV and Trabecular Bone Thickness (Tb.Th) of trabecular bone in the medial femoral condyle, but no major differences were detected in the tibial plateau epiphysis (Supplement Fig. S5). We did not observe major significant changes in bone microstructure between F1 Male whose parents received different diets.

F2 offspring weight gain, adiposity, and metabolic profile

F2 Female HFD and Control groups did not differ in weight, while F2 HFD Male were significantly heavier than F2 Male Control (Fig. 4A). F2 HFD Male had also a significantly higher percentage of adipose tissue compared to F2 Male Control (Fig. 4B). There was no noticeable difference in SAT adipocyte size in F2 Female diet groups, while a trend toward enhanced adipocyte hypertrophy was present in F2 HFD Male compared to their Controls (Fig. 4C, and Supplement Fig. S2A). Additionally, we observed trends towards lower glucagon level in F2 HFD Female and toward higher serum insulin and leptin, as well as lower resistin levels in F2 HFD Male compared to F2 Controls (Fig. 4D and Supplement Fig. S6). Other serum adipokines did not differ in analyzed groups (Supplement Fig. S6).

Fig. 4. Grandparental HFD affects weight gain, adiposity, and metabolic profile of F2 generation offspring.

(A) Longitudinal body weight measurement. (B) Adipose tissue content assessment by microCT. (C) Representative histology (H&E) staining of subcutaneous adipose tissue (SAT), scale bar = 100μm. (D) Serum levels for adipokines (leptin, insulin, and glucagon). All data are presented as mean ± 95% CI, n=10–16 per sex, per parental diet group. One Way ANOVA (A) or Mann-Whitney U test (B, D), § p<0.001, * p<0.05.

F2 offspring systemic and local inflammation

Similar to the F1 generation, we found that serum IL-1β trended towards higher, while IP-10 was significantly lower, in F2 HFD Female compared to their Controls (Fig. 5A). No significant differences were observed in other cytokines, including IL-10, between F2 Female diet groups (Table S3). We also did not find any significant difference in serum cytokine levels between F2 Male groups. Additionally, we did not observe remarkable differences in fibrosis among analyzed groups, although there were trends towards a higher number of M1-like macrophages in F2 HFD Male compare to their Controls (Fig. 5B and C, and Supplement Fig. S3E). Gene expression analysis of VAT and SAT from F2 offspring groups showed no significant differences in Lep, Lepr, or Cd36 gene expression (Fig. 5D).

Fig. 5. Grandparental HFD affects immune profile of F2 generation offspring.

(A) Serum levels for cytokines (IL-1β, TNFα, IL-6, IL-1α, IL-10, IP-10, Eotaxin, IL-13, and IL-17). (B) Representative histology (H&E) staining of visceral adipose tissue (VAT), scale bar = 100μm. (C) Immunohistochemistry staining of VAT showing CD11b (green), CD11c (red), and nuclei (blue) of M1 (pro-inflammatory) macrophages, scale bar = 100μm, white arrows indicate double positive cells. (D) Leptin (Lep), Leptin Receptor (Lepr), and Cd36 gene expression in VAT and SAT. All data are presented as mean ± 95% CI, n=5–16 per sex, per parental diet group. Mann-Whitney U test, ** p<0.01, * p<0.05.

Assessment of OA severity, osteophyte formation, and synovitis in F2 generation offspring

We found that F2 Female from HFD grandparents had a significantly higher predisposition to develop OA, while there was no difference in F2 Male between diet groups (Fig. 6A and C). We also found that osteophyte scores were not affected by grandparental diet in both analyzed F2 groups (Fig. 6D). Interestingly, there was a significantly higher total synovitis score in F2 HFD Male compared to Control (Fig. 6B and E). We did not find differences in SF cytokines levels in analyzed F2 offspring (Fig. 6F). Additionally, we did not observe significant differences in the presence of synovial F4/80+ macrophages in analyzed groups (Supplement Fig. S4B). Total OA Mankin score from DMM-operated joint did not correlate with weight and serum leptin in F2 Female and Male (Fig. 6G and H). However, serum C-peptide correlated positively with F2 Female OA score, while serum insulin showed a positive correlation with F2 Male OA scores.

Fig. 6. Grandparental HFD affects musculoskeletal integrity of F2 generation offspring.

(A) Representative Safranin-O/Fast Green staining of DMM-operated joints, scale bar = 100 μm. (B) Representative H&E staining of the medial femoral condyles of DMM-operated joints, scale bar =100 μm. (C) Total joint Modified Mankin scores. (D) Total Medial Osteophyte scores. (E) Total Synovitis scores. (F) Synovial Fluid (SF) Cytokine levels (IP-10, IL-10, IL-1α) from non-operated control (right) and DMM-operated (left) joints. (G) Spearman bivariate correlation between DMM-operated (left) joint OA Score and weight, serum leptin levels, and serum C-peptide levels in F1 Female group. (H) Spearman bivariate correlation between DMM-operated (left) joint OA Score and weight, serum leptin levels, and serum insulin levels in F1 Male group. (I) Representative microCT reconstructed images of tibial metaphysis. (J) MicroCT analysis of DMM-operated (left) joint tibial metaphysis showing trabecular bone fraction bone volume/total volume (BV/TV), trabecular bone number (Tb.N), and Bone Mineral Density (BMD). All data are presented as mean ± 95% CI, n=10–16 per sex, per parental diet group. (C-F) Two-way ANOVA with Bonferroni post hoc analysis. Groups not sharing the same letter are significantly different within sex, p < 0.05. Main effect of DMM surgery reported as *. (J) Mann-Whitney U test, § p<0.001.

Analysis of bone microstructure changes in F2 offspring

Bones were imaged with microCT (Fig. 6I). BV/TV and Tb.N but not BMD of the tibial metaphysis were significantly lower in F2 HFD Female compared to Controls (Fig. 6J). Additionally, F2 HFD Female displayed significantly lower trabecular bone BV/TV in the epiphysis of the lateral femoral condyle, while F2 Male had significantly lower BV/TV in the lateral tibial plateau (Supplement Fig. S7).

DISCUSSION

Our findings show that high-fat feeding has an intergenerational effect on the inheritance of increased risk of weight gain, metabolic imbalance, and OA in mice for at least two generations of offspring. Both F1 and F2 offspring groups, despite being allocated on the standard control diet, demonstrated dysregulated systemic metabolism, weight gain, and disturbed musculoskeletal integrity. Strikingly, both F1 and F2 Female mice showed an increased predisposition to injury-induced OA and severe bone microarchitectural changes due to parental and grandparental high-fat feeding. These findings indicate that in addition to well-characterized genetic factors (19), the inheritability of OA risk may involve epigenetic factors secondary to diet-induced obesity (20) that may be passed through multiple generations of offspring.

Significantly higher weight and adipose tissue content in F1 (Male and Female) and F2 Male offspring of HFD-fed parents, in comparison to offspring of control diet-fed parents, is consistent with the results of previous studies showing that combined parental diet had a cumulative effect on the weight gain and adiposity in multiple generations of offspring. For instance, it has been reported that maternal fructose consumption during pregnancy increases metabolic syndrome and adipose tissue content in F1 females (21), while maternal high-fat feeding increases adiposity in adult male rats (22). Paternal hyperglycemia was also shown to affect obesity predisposition in offspring (23), and maternal obesity leads to an increased weight gain across multiple generations (24). While the mechanisms involved in this process remain to be determined, it has been postulated that reduced quality of oocyte and sperm cells due to a high-fat diet, epigenetic modifications, dysregulated hypothalamic signaling, the effect of the maternal microbiome (14), and intrauterine nutrient transfer (7) may play an essential role.

Leptin and insulin signaling have been postulated as key contributors in fetal programming. Dysregulated leptin hypothalamic signaling in leptin-sensitive neurons of overfeeding mothers has been shown to be passed on to male offspring, and in fact, to mediate an “obese” phenotype in future generations (25). Several previous studies have also shown that both maternal hyperglycemia (26) and paternal obesity lead to a modification in insulin sensitivity and glucose tolerance (27). Here we show that there are significant sex-dependent differences in circulating leptin, insulin, and glucagon, as well as LEPR gene expression, within both F1 and F2 HFD offspring as compared to their controls. Serum leptin level significantly correlated with OA scores in F1 Female but did not in any other analyzed groups, while serum insulin correlated significantly with OA score only in the F2 Male group. Interestingly, leptin is postulated to be a key mediator in the pathogenesis of OA (28), while the loss of leptin signaling is protective against obesity-induced OA in mice (29). Additionally, circulating leptin levels differ between male and female mice, and leptin has been shown to affect males and females differently (30). Although we did not investigate fasting metabolite levels in our animal cohort, the data presented here implies that both leptin and glucose metabolism may be significantly affected in offspring in a sex-dependent manner.

Obesity-induced systemic and local inflammation plays an important role in the progression of OA damage. Our and other previous studies have shown that immune cells colocalize within the knee joint during obesity-induced OA (31–33). Additionally, the systemic pro-inflammatory milieu has been shown to be important in OA pathogenesis, potentially through interactions with local injury-induced inflammation (17). Before we addressed the OA ramifications in our study, we showed that obesity-related local and systemic immune changes are affected by parental and grandparental HFD. Local adipose tissue inflammation, indicated by adipose tissue fibrosis and CD11b+CD11c+ macrophage infiltration, was mainly observed in F1 HFD and F2 HFD Male, although no serum proinflammatory cytokines were significantly affected in these groups. On the other hand, we found that serum levels of pro-inflammatory IL-1β trended higher in F1 HFD Female while anti-inflammatory IL-10 was significantly lower in F1 and F2 HFD Female, compared to their Controls. Studies suggest that immune cells play an important role in mediating the transgenerational effect of obesity in offspring. For instance, recent studies showed that monocytes and macrophages obtained from cord blood of obese mothers had significantly changed levels of IL-1β and IL-10, as well as significant differences in DNA methylation of IL-10 promotor in those cells (34). Furthermore, children born to obese mothers were shown to have a significantly lower number of eosinophils and CD4⁺ T helper cells and reduced monocyte response to inflammation (35). Finally, our study shows unexpectedly that serum IP-10 was significantly lower in both F1 and F2 HFD Female compared to their Controls and was negatively correlated with OA score in the F1 Female group. In conclusion, the maternal diet has been proposed as a regulator of the fate of hematopoietic cells (36). Our findings and those of others are consistent with this notion and indicate that the immunometabolism in offspring of mice on HFD is compromised, potentially leading to observed changes in body composition and musculoskeletal integrity.

Although OA was historically considered as a “wear-tear” disease, it is now commonly accepted that osteoarthritic-related changes are multifactorial processes affecting multiple joint tissues and correlate with the progression of systemic and local inflammation. We and other groups have previously shown that injury-induced OA is significantly affected by obesity (5, 17, 37). Additionally, previous human studies also indicate a strong correlation of the familial predisposition to OA (38–40). Here, we also show that there are significant changes in predisposition to injury-induced OA, particularly in F1 and F2 HFD Female, despite the fact that there were changes in body weight or composition in both male and female offspring. Additionally, we did not find significant differences in local synovial fluid cytokine levels or synovial F4/80+ macrophage content in the joints of the offspring, which may suggest that factors beyond injury-induced inflammation play an important role in the intergenerational inheritance of OA. The mechanisms modulating the inheritability of cartilage damage remain to be determined, but epigenetic regulation has become a central mechanism in the hypothesis of fetal developmental programming affected by parental obesity (41). For example, grandparental HFD-induced obesity has been shown to affect the skeletal muscle transcriptome transgenerationally (42). Additionally, maternal HFD has been shown to impair muscle mitochondria in female offspring (43).

Regardless of surgical induction, the structure and integrity of subchondral bone are postulated to play an important role in the progression of OA (44). Previous studies showed that maternal diabetes affects bone health and growth in offspring (45). Similarly, our study shows that HFD affects the bone structure of only the female offspring in F1 and F2 generations: trabecular bone BV/TV, Tb. N, and BMD are significantly impacted in F1 and F2 HFD Female offspring across the tibia metaphysis and femoral condyle of the knee joint. Previous studies have shown that maternal high-fat diet impairs offspring skeletal development by affecting osteoblast differentiation and promoting cell senescence (46, 47). Furthermore, intrauterine exposure to HFD was shown to negatively regulate bone microstructure (48). There are several factors which can possibly explain these observations. For instance, Liang et al. have pointed out that observed HFD-mediated reduction in bone quality may, in fact, mirror human epidemiological studies of the HFD-induced delayed skeletal development, which associates with bone fragility during aging (48). Furthermore, the detrimental effect of the microbiome on bone quality due to HFD has been postulated as an important mechanism in reducing bone microarchitectural changes (49).

It is important to note several points that may influence the interpretation of our findings. While we did not examine the role of the microbiome in this study, it has been reported that the obesity-induced microbiome may play an important role in musculoskeletal integrity (49). For example, obesity can significantly disrupt the gut-microbiome, which has a detrimental role in OA progression (32). Additionally, parental HFD has been shown to affect offspring microbiome, predisposing them to autoimmune diseases (50). Another potential limitation of the current study is the lack of direct measurement of the dysfunction in glucose metabolism and insulin resistance in offspring, which would allow further understanding of the potential role of metabolic syndrome in the observed outcomes. Furthermore, the described disturbance in leptin and its satiety/hunger role could possibly be explained by analysis of food intake in studied mice.

In conclusion, our results indicate that high-fat feeding plays a sex-dependent detrimental role in musculoskeletal integrity in at least two generations of offspring. We showed that offspring metabolism, adiposity, immune profile, cartilage and bone integrity are significantly affected by parental high-fat feeding. These findings have significant implications for the inheritability mechanism of obesity triggered-OA and should initiate an in-depth investigation of the role of epigenetic and environmental stimuli on OA progression. Future studies confirming the beneficial role of maternal and paternal diet, exercise, and microbiome interventions may benefit the offspring musculoskeletal integrity and perhaps mitigate the observed transgenerational inheritance of a predisposition to OA.