Obesity promotes lipid accumulation in mouse cartilage—a potential role of acetyl-CoA carboxylase (ACC) mediated chondrocyte de novo lipogenesis

Abstract

Objective:

Obesity promotes the development of osteoarthritis (OA). It is also well-established that obesity leads to excessive lipid deposition in non-adipose tissues, which often induces lipotoxicity. The objective of this study was to investigate changes in the levels of various lipids in mouse cartilage in the context of obesity and determine if chondrocyte de novo lipogenesis is altered.

Methods:

We used Oil Red O to determine the accumulation of lipid droplets in cartilage from mice fed high-fat diet (HFD) or low-fat diet (LFD). We further used mass spectrometry-based lipidomic analyses to quantify levels of different lipid species. Expression of genes involving in fatty acid (FA) uptake, synthesis, elongation, and desaturation were examined using quantitative PCR. To further study the potential mechanisms, we cultured primary mouse chondrocytes under high-glucose and high-insulin conditions to mimic the local microenvironment associated with obesity and subsequently examined the abundance of cellular lipid droplets. The acetyl-CoA carboxylase (ACC) inhibitor, ND-630, was added to the culture medium to examine the effect of inhibiting de novo lipogenesis on lipid accumulation in chondrocytes.

Results:

When compared to the mice receiving LFD, the HFD group displayed more chondrocytes with visible intracellular lipid droplets. Significantly higher amounts of total FAs were also detected in the HFD group. Five out of six significantly upregulated FAs were ω-6 FAs, while the two significantly down-regulated FAs were ω-3 FAs. Consequently, the HFD group displayed a significantly higher ω-6/ω-3 FA ratio. Ether linked phosphatidylcholine (PC) was also found to be higher in the HFD group. Fatty acid desaturase (Fad1-3), fatty acid binding protein 4 (Fabp4), and fatty acid synthase (Fasn) transcripts were not found to be different between the treatment groups and fatty acid elongase (Elovl1-7) transcripts were undetectable in cartilage. Ceramide synthase 2 (Cers-2), the only transcript found to be changed in these studies, was significantly upregulated in the HFD group. In vitro, chondrocytes upregulated de novo lipogenesis when cultured under high-glucose, high-insulin conditions, and this observation was associated with the activation of ACC, which was attenuated by the addition of ND-630.

Conclusion:

This study provides the first evidence that lipid deposition is increased in cartilage with obesity and that this is associated with the upregulation of ACC-mediated de novo lipogenesis. This was supported by our observation that ACC inhibition ameliorated lipid accumulation in chondrocytes, thereby suggesting that ACC could potentially be targeted to treat obesity-associated OA.

INTRODUCTION

Osteoarthritis (OA) is a leading cause of disability, affecting 32.5 million adults at an annual cost of $136.8 billion in the United States¹. The prevalence of OA has doubled since World War II, in parallel with the growing obesity epidemic². It was initially believed that obesity increased OA risk simply due to greater joint loading, however, several lines of evidence indicate this is not the case since obesity also increases OA risk in non-weight-bearing hand joints^3,4. It is well-known that obesity promotes excessive lipid deposition in non-adipose tissues, which is causal for lipotoxicity and tissue dysfunction⁵. However, it remains unknown if obesity promotes disproportionate lipid accumulation in cartilage and what role this may play in the development of OA.

Chondrocytes of various cartilages (e.g., costal, bronchial, and articular cartilage) naturally contain lipid droplets in their cytoplasm as well as their surrounding extracellular matrix^6,7. These lipids are traditionally not considered evidence of degeneration but as an energy source for chondrocytes given their avascular environment and limited access to macronutrients in the circulation⁸. However, recent evidence has shown that articular cartilage accumulates lipids during the aging process⁹, similarly to other tissues such as liver and muscle¹⁰. Moreover, lipid accumulation in cartilage was found to positively correlate with OA severity¹¹. Further profiling of the lipids in cartilage revealed that glycolipids account for the vast majority of increased lipids with OA¹¹. These findings stirred interest in studying the potential roles of lipids in cartilage degeneration. For example, Lippiello et al. suggested that high levels of arachidonic acid (20:4n6) in cartilage during OA could be related to elevated pro-inflammatory eicosanoid biosynthesis¹¹. Moreover, Kimmerling et al. recently found that decreasing the ω-6 to ω-3 fatty acid (FA) ratio by forced-expression of fat-1 to convert ω-6 to ω-3 FA in mice delayed injury-induced OA development¹². Additionally, lipid peroxidation has been linked to increased oxidative stress in OA pathogenesis^13-15. Despite these important findings, it remains unknown how cartilage accumulate lipids during obesity and if the accumulation plays a role in obesity-associated OA pathogenesis.

Based on current knowledge, we surmise that cartilaginous lipids may come from the diffusion of lipids from the surrounding environment, or de novo lipogenesis locally in chondrocytes. It is well known that the majority of lipids found in blood are associated with lipoproteins or bound to carrier proteins. However, due to the avascular nature of articular cartilage, it is unknown what proportion of cartilaginous lipids come directly from the circulation. It also remains debatable that circulating lipids have any direct links with the development of OA. For example, a recent report suggests that serum triglycerides and cholesterol are positively associated with cartilage degeneration and the incidence of bone marrow lesions¹⁶. However, another longitudinal study reported that higher serum HDL cholesterol levels appeared to protect against radiographic hand OA, while no relationship was found with total or LDL cholesterol¹⁷. Other studies have also found that various types of serum lipids, including sphingomyelins¹⁸, phosphatidylcholines, and lysophosphatidylcholines¹⁹, changed in abundance in serum samples from OA patients. Surprisingly, few studies to date have examined the role of chondrocyte de novo lipogenesis.

The goal of the current study was to examine changes in the levels of various lipids in mouse cartilage in the context of obesity and determine if chondrocyte de novo lipogenesis is altered. We hypothesized that cartilage accumulates more lipids during obesity, and this increase is associated with increased de novo lipogenesis in chondrocytes. To test this hypothesis, we first stained cartilage neutral lipids with Oil Red O in mouse cartilage tissue from mice fed HFD or LFD. We then quantified various lipid species in the cartilage using mass spectrometry-based lipidomic analyses. In an effort to better understand the potential mechanisms by which lipids accumulate in cartilage, we analyzed the expression of genes involved in FA uptake, synthesis, elongation, and desaturation. We also evaluated de novo lipogenesis in primary mouse chondrocytes subjected to high-glucose, high-insulin media, which mimics the local microenvironment in vivo in obese mammals.

MATERIALS AND METHODS

Animals and diet treatment:

All mice (C57BL/6, male) were group housed (5/cage) on corncob bedding with cardboard enrichment tubes and nestlets at 22 ± 0.5°C on a 12:12-hour light-dark cycle and maintained on low-fat diet (TestDiet 58YP; 66.6% CHO, 20.4% PRO, 13.0% FAT) until 12 weeks of age. At 12 weeks of age, mice were randomized into low-fat diet (LFD; TestDiet 58YP) or high-fat diet (HFD; TestDiet 58V8; 35.5% CHO, 18.3% PRO, 45.7% FAT) treatment groups. The mice were provided ad libitum access to food and water throughout the 30-week dietary treatment period. At the conclusion of each treatment period mice were euthanized with isoflurane in the fasted state (5-6 hours) and tissues were collected as described below. All animal procedures were reviewed and approved by the Institutional Animal Care and Use Committees at the Ohio University.

Mouse joint histology:

Isolated knee joints were fixed in 4% formaldehyde for 48 hours at 4°C and decalcified in 10% EDTA for 14 days at 4°C. The samples were then dehydrated and embedded in paraffin for coronal sectioning using standard histology procedures. Paraffin tissue blocks were then sectioned at a thickness of 8 μm. Joint sections were then deparaffinized, rehydrated, and stained with Hematoxylin, Fast Green, and Safranin-O for histological grading. Two experienced graders (H.L. and S.Z.) evaluated the sections (n=7/group). Slides were randomized and assigned a temporary identification code to blind graders to diet treatment. OARSI mouse OA grading scores²⁰ were assigned separately for the lateral femur, lateral tibia, medial femur, and medial tibia independently by each grader. Scores that differed by >2 between graders were re-evaluated for consensus scoring. Scores were then averaged for both graders to obtain a final score per section and location.

Lipidomic Analyses:

Cartilage tissue (~8 mg) from left joints of mice (n=3) was collected and subjected to lipidomic analysis. Total lipids were extracted following the method of Bligh and Dyer with modifications²¹. Both 15:0 and 17:0 were added as internal standards to each extract. Extracted samples underwent acid hydrolysis/methanolysis to generate fatty acid methyl esters (FAMEs)²². FAMEs were quantified using an Agilent Technologies 6890N gas chromatograph with flame ionization detector (GC-FID)²³. Data was obtained as relative molar concentration for each fatty acid and presented as relative values to the LFD group. For phospholipid analysis, the method has been described previously²⁴. Briefly, total lipid extracts were diluted 1:40 with 2-propanol/methanol/chloroform (4:2:1 v/v/v) containing 20 mM ammonium formate and 1.00 μM PC 14:0/14:0 as internal standard. Samples were introduced into a triple quadrupole mass spectrometer (TSQ Ultra, Thermo Scientific) using a chip-based nano-ESI source (Advion NanoMate) operating in infusion mode. PC lipids were measured using precursor ion scanning of m/z 184. Quantification of lipid molecular species was performed using the Lipid Mass Spectrum Analysis (LIMSA) software’s peak model fit algorithm. Data are represented as relative percent of each measured species within each class (i.e. PC) ± standard deviation.

Mouse primary chondrocyte isolation and culture:

Primary mouse chondrocytes were isolated from 7 day old male and female wild-type (WT) mice according to a previously published protocol²⁵. Briefly, both tibial and femoral cartilage tissue were dissected from the knee joint. The cartilage was then incubated in 3 mg/mL Collagenase D (Roche) in DMEM solution for two 45 minute periods and transferred to 0.5 mg/mL Collagenase D in DMEM solution supplemented with 3% Liberase TL (Sigma) overnight. The tissue was then homogenized by pipetting up and down to release and suspend the cells, and the homogenate was filtered through a 40 μm strainer to remove large debris. Passage 0 cells were cultured and expanded in 6-well plates in complete DMEM media (Life Technology, 10567014) supplemented with 10% fetal bovine serum (FBS) and 1% penicillin/streptomycin (P/S) at 37 °C and 5% CO2. Chondrocytes reached confluency after 3-4 days in culture. Chondrocytes were then trypsinized, counted, and reseeded as passage 1 cells for specific experiments described in detail in the following sections.

RESULTS

Obesity promotes lipid accumulation in mouse cartilage.

The average body weight of mice with high-fat diet (HFD) feeding almost doubled at the time of euthanasia, which is significantly higher than the body weight of low-fat diet (LFD) fed mice (Figure 1A). We further analyzed the joint histopathology of mice from both groups. Cartilage in mice fed LFD displayed only slight undulating surface in medial tibia (Figure 1B, left panel). In contrast, small lesions across the cartilage surface were present in mice subjected to HFD feeding (Figure 1B, right panel). Consistently, OARSI grading showed greater overall OA pathology in medial tibia in the HFD group (Figure 1C). To evaluate whether obesity promotes lipid accumulation in cartilage, we used Oil Red O staining on cryosections of joint tissues from both LFD and HFD groups. Consistent with previous findings in other species, mouse articular cartilage naturally contained oil droplets inside chondrocytes in the LFD group (Figure 1D). The droplets tended to form around nuclei suggesting their cytoplasm localization (Figure 1D). Interestingly, the majority of the oil droplets seemed to be inside chondrocytes in the middle to deep zone of cartilage. Compared with the LFD group, the cellular oil droplets in the HFD group were visually larger and appeared more abundant (Figure 1D). Quantification of the ratio of chondrocytes containing oil droplets showed that significantly higher ratio of oil droplet-containing chondrocytes among total chondrocytes in the HFD group (Figure 1E). And this difference is not caused by the change of the total number of chondrocytes (Figure 1F). These results suggested that HFD-induced obesity promoted cartilage degeneration and cellular lipid deposition in mouse chondrocytes.

Figure 1. Dietary-induced obesity promotes cartilage degeneration and lipid intracellular deposition in mice.

A. body weight (g) of mice before and after 30 weeks of LFD or HFD feeding. n=7. B. representative Safranin-O staining images of medial tibia from C57/BL6 mice fed with LFD or HFD. Scale bar, 100 μm; C. OARSI scores of medial tibias from mice fed with LFD or HFD. n=7 for both LFD and HFD groups; D. representative Oil Red O staining images of medial tibia in LFD and HFD groups. Scale bar, 20 μm. Black dashed box indicates the area that is magnified and shown on the right; E. ratio of chondrocytes containing oil droplets to total chondrocytes. n=4 for LFD and n=5 for HFD. F. Total number of chondrocytes used for calculating the lipid-containing chondrocytes ratio. *p<0.05, **p<0.01, ***p<0.001.

Lipidomics analysis revealed that obesity increased both total and pro-inflammatory ω-6 fatty acids (FAs) in cartilage in mice.

To further analyze the lipids in mouse cartilage tissue, we performed lipidomic analyses to profile individual FA and lipid species. In both the LFD and HFD groups, the predominant FAs were palmitic (16:0, oleic (18:1) and linoleic (18:2) acids, which collectively represented more than 80% of the total FA content (Supplementary table 2). The total content of FA (nmol/mg tissue weight) increased by nearly 50% with HFD feeding (Figure 2B). Through the comparison of the relative abundance of individual FAs, we found that five of the six significantly increased FAs were ω-6 FAs [linoleic acid (18:2n6), eicosadienoic acid (20:2n6), arachidonic acid (20:4n6), adrenic acid (22:4n6), and docosapentaenoic acid (22:5n6)] (Figure 2A). Additionally, the two significantly decreased FAs were both ω-3 FAs (Figure 2A). As a result, the ω-6 to ω-3 FA ratio was significantly higher in the HFD group (Figure 2C). We also examined the abundance of glycerophospholipid composition in mouse cartilage tissue. Consistent with a previous study in human cartilage⁹, we detected a large amount of phosphatidylcholine (PC) in mouse cartilage from both LFD and HFD groups. Interestingly, we found that there was a significantly higher percentage of ether-linked PC in the HFD group (12.2% vs 15.9%) (Figure 2D). To understand the potential mechanism that contributes to differential FAs abundance in cartilage during HFD feeding, we examined the expression level of genes involved in FA uptake (fatty acid binding protein, Fabp4), synthesis (fatty acid synthase, Fasn and Ceramide synthase 2, Cers2), elongation (elongation of very long chain fatty acids protein 1-7, Elovl1-7), and desaturation (fatty acid desaturase 1-3, Fads1-3). Surprisingly, we did not see any changes in Fads1-3, Fabp4, or Fasn expression between HFD and LFD groups (Suppl Figure 1A-E). Cers2 was the only gene that was found to be significantly upregulated in the HFD group (Suppl Figure 1F). We were not able to detect expression of Elovl1-7 in cartilage in either the HFD or LFD group (data not shown).

Figure 2. Lipidomic analyses reveals that obesity increases both fatty acid (FA) and lipid levels in mouse cartilage.

A. relative abundance of individual FAs in the cartilage tissue of mice fed with LFD or HFD; B. total nanomoles of FA per mg wet tissue weight in LFD and HFD groups; C. ratio of ω-6 to ω-3 FA in LFD and HFD groups; D. sum of percent ether-phosphatidylcholine (PC) species of total detected phosphatidylcholines. *p<0.05, **p<0.01, ***p<0.001. n=3 for both LFD and HFD.

High concentrations of insulin and glucose in culture conditions promote lipid deposition in chondrocytes.

We previously discovered that supplementation of 20 nM insulin and 20 mM glucose (GI) into chondrocyte culture medium could mimic in vivo obesity associated conditions and promote protein post-translational malonylation²⁶. We then tested whether this ‘obesity-mimicking’ condition would also promote lipid accumulation in chondrocytes. We cultured primary mouse chondrocytes in a control growth medium or a growth medium supplemented with 20 nM insulin and 20 mM glucose (GI) for ten days. We used Oil Red O staining to visualize lipids. Interestingly, we found that under normal culture condition (2.5 mM glucose without insulin), some chondrocytes (~5%) contained lipid droplets (Figure 3A). Meanwhile, a large number of chondrocytes (~10%) were filled with large lipid droplets in the media supplemented with GI (Figure 3A). Absorbance measurement (at 492 nm) of the extracted stain solution showed that the GI culture condition significantly increased lipid content in chondrocytes (Figure 3B). We also used ELISA to measure the abundance of malonyl-CoA, the building block for lipids. It was found that compared to normal culture condition, GI upregulated malonyl-CoA level by approximately two-fold (Figure 3C). This increase did not appear to be in response to a greater cell number in the GI group, as the total cell numbers were not significantly different between the control and GI condition after ten days of culture (Figure 3D).

Figure 3. High concentrations of insulin and glucose in culture media promote lipid deposition in chondrocytes.

A. Oil Red O staining of mouse primary chondrocytes cultured in control (CTL) growth medium or CTL medium supplemented with 20 nM insulin and 20 mM glucose (GI) for seven days. B. absorbance measurement (at 492 nm) of the extracted stain solution in CTL or GI groups. C. ELISA measurement of abundance of metabolite malonyl-CoA in CTL or GI treated chondrocytes. D. cell number of chondrocytes cultured in CTL or GI medium. n=4 for both CTL and GI. *p<0.05, **p<0.01.

Acetyl-CoA carboxylase (ACC) is activated in chondrocytes under obesity associated conditions.

Since we observed a higher level of cellular malonyl-CoA in chondrocytes under GI conditions, we wanted to understand the potential underlying mechanisms. It is known that malonyl-CoA is formed by carboxylating acetyl-CoA through the enzyme acetyl-CoA carboxylase (ACC), which is the first step committed to lipogenesis. Hence, we first examined the expression of both ACC1 and ACC2 in cartilage in both LFD and HFD groups using immunofluorescent staining. Our results showed that both ACC1 and ACC2 were expressed at a low level in cartilage of mice with LFD feeding (Figure 4A). However, HFD feeding substantially increased ACC1 expression in chondrocytes across the full thickness of mouse cartilage tissue (Figure 4A). The positive staining signal of ACC1 was in a circular shape surrounding the nuclei inside chondrocytes indicating a cytosolic localization of ACC1 protein. This observation is consistent with previous studies investigating localization of ACC1 and ACC2²³. In contrast, the expression of ACC2 was at a similarly low level in both LFD and HFD group (Figure 4A). We further investigated whether the in vitro ‘obesity-mimicking’ GI condition would activate ACC. Interestingly, insulin treatment decreased phospho-ACC (S79) level (Figure 4B), indicating insulin upregulated ACC activity. Since specific antibodies against phospho-ACC1 or phospho-ACC2 are not currently available, we were only able to detect pan phospho-ACC levels, therefore it remains unclear which isoform is responsible for our observations. Moreover, 20 mM glucose treatment increased the expression of total ACC1 (Figure 4B). These results suggested that ACC, in particular ACC1, was activated in chondrocytes under obesity associated conditions. It is also worth noting that we were not able to detect total ACC2 using Western Blot in any of the groups of chondrocytes (data not shown). We next wanted to test whether inhibition of ACC would abolish lipid accumulation in chondrocytes in vitro. We found that ‘obesity-mimicking’ GI condition induced intracellular accumulation of numerous large oil droplets inside chondrocytes, which was blocked by the addition of ACC inhibitor, ND-630 (Figure 4C). Consistently, ELISA measurement of the metabolite malonyl-CoA showed that GI induced upregulation of malonyl-CoA is normalized by ND-630 (Figure 4D). These results suggested that chondrocytes increased lipogenesis under obesity-associated conditions and this process likely involved the enzyme ACC.

Figure 4. Acetyl-CoA carboxylase is activated in chondrocytes following exposure to GI culture condition, thereby contributing to chondrocyte de novo lipogenesis.

A. Immunofluorescent staining for ACC1 (green) and ACC2 (red) in cartilage of mice fed with LFD or HFD. Scale bar, 20 μm; B. Western blotting for p-ACC (S79), total-ACC1, p-AKT, total-AKT, and actin in chondrocytes treated with a combination of glucose (2.5 or 20 mM) and insulin (20 nM). 2.5 mM is the glucose concentration in chondrocyte culture medium, while 20 mM glucose is achieved through adding extra glucose solution to the culture medium. C. Oil Red O staining of mouse primary chondrocytes cultured in CTL or GI or GI supplemented with ACC inhibitor ND-630 (5 nM). D. ELISA measurement of abundance of metabolite malonyl-CoA in chondrocytes cultured in CTL or GI or GI supplemented with ND-630. n=3. **p<0.01.

DISCUSSION

Although it has been known for decades that cartilage tissue naturally contains a large amount of FA⁶, little is known about how FA quantity as well as quality change under different OA risk factors such as obesity. We hypothesized that, similar to other non-adipose tissues, obesity would promote excessive lipid accumulation in cartilage tissue and exert lipotoxic effects on chondrocytes. Using a combination of lipid staining and lipidomic analyses of mouse cartilage, we found that the total amount of FA is almost two times higher in obesity. Interestingly, this increase is largely attributed to the increased levels of pro-inflammatory ω-6 FAs. We further discovered that this increase is associated with the increased expression of ACC, the enzyme that commits the first step of FA biosynthesis. To test the role of FA biosynthesis, we inhibited ACC using ND-630 and found a significant decrease of lipid accumulation in chondrocytes in vitro.

Recent technical advances in metabolomics and its application in an attempt to identify OA biomarkers have found various classes of lipids in serum or joint tissues. For example, increased ratio of lysophosphatidylcholines (lysoPCs) to phosphatidylcholines (PCs)¹⁹, higher triglycerides level¹⁶, and lower high-density lipoprotein (HDL) cholesterol level¹⁷ have been reported in the serum samples of advanced OA patients. Additionally, analyses of synovial fluid have also discovered changes in various phospholipid species in OA patients^27,28. Besides, lipids are also found in chondrocytes and in the matrix of cartilage tissue^7,9. This has prompted a great deal of speculation on their significance in disease development. Bonner et al. first reported that polyunsaturated fatty acids (PUFA) increase in abundance in human cartilage with aging suggesting their potential involvement in cartilage degeneration⁹. However, this speculation is challenged by the observation that lipid deposition is also found in normal growing cartilage and in chondrocytes from elastic cartilage²⁹, a tissue that rarely undergoes degeneration even in old age. In 1991, Lipiello and colleagues discovered that there was a strong positive association of lipid, in particular arachidonic acid accumulation with OA joint histological severity¹¹.

Despite these reports, little is known about the direct effect of lipid species or FAs on cartilage and how chondrocytes accumulate lipid under normal or OA risk conditions. Our study is the first to report the profile of FAs and lipids in mouse articular cartilage under HFD-induced obesity conditions. We found a higher level of total and ω-6 FAs, as well as a lower level of ω-3 FAs in the cartilage of mice fed the HFD. These changes could potentially be associated with increased inflammation and greater cartilage degeneration in the HFD group.

While the mechanisms are not fully understood, ω-6 FAs are generally considered to be pro-inflammatory due to their role as precursors for prostanoid production, while ω-3 FAs can serve as precursors for anti-inflammatory mediators such as resolvins and protectins³⁰. Thus, the ratio of ω-6 to ω-3 FAs has been widely used as an indicator of inflammatory status. Our data shows that the ω-6/ω-3 ratio more than doubled in the HFD group compared to the LFD group. This is consistent with an increasing recognition of the role of inflammation in OA disease development, especially during obesity-mediated abnormal metabolic conditions (metabolic inflammation)³¹. Our lipidomics data showed that several precursors (including 18:2n6 linoleic acid, 20:4n6 arachidonic acid, and 22:4n6 adrenic acid) of prostaglandin E2 (PGE2) were significantly upregulated in the HFD group, suggesting that PGE2-mediated inflammation may be activated in joints with HFD feeding. Indeed, elevated levels of PGE2 have been reported in synovial fluid and cartilage from patients with OA^32-34. It has been demonstrated that PGE2 inhibited proteoglycan synthesis as well as promoted collagen degradation through upregulating MMP13 secretion^35,36. It is still unknown whether modulating synthesis of these FA precursors would inhibit PGE2 production and be beneficial for OA development. Additionally, future studies investigating the correlations between individual FA species and OA severity is warranted. It has the potential to identify novel biomarkers for obesity associated OA.

We observed increased abundance of various ω-6 FAs in the HFD group, all of which are long-chain FAs. It is known that malonyl-CoA produced by acetyl-CoA carboxylase (ACC) fuels de novo synthesis of long-chain FAs and subsequent chain elongation. Consistently, we found that ACC1 was significantly upregulated in the HFD group. The rate of FA elongation is determined by the activity of the elongase³⁷. Seven elongase subtypes (ELOVL1-7) have been discovered thus far in mammals. Surprisingly, we found chondrocytes did not express Elovl1-7, suggesting other mechanisms may exist for the observed upregulation of the long-chain FAs in our study. Furthermore, considering the significantly changed FAs comparing HFD and LFD groups were all unsaturated FAs, we also examined the expression of the desaturases Fads1-3. However, no changes of Fads1-3 expression between HFD and LFD treatment groups were detected. This could be due to a transient activation of the genes in cartilage tissue during the course of HFD feeding. More longitudinal studies are warranted to test the temporal expression patterns of the genes.

A novel finding from this study was that chondrocytes upregulated de novo FA biosynthesis during obesity. Interestingly, the vast majority of previous studies have been focused on studying lipid uptake by chondrocytes while few studies have examined the role of de novo FA biosynthesis in cartilaginous lipid deposition. For example, there is evidence showing that the FA composition in cartilage is modulated directly by dietary fat intake³⁸, suggesting direct transport of lipids (lipoproteins) to cartilage from circulation or synovial fluid is happening. However, the dense network established by collagen and proteoglycans in cartilage matrix^39,40 restricts the transportation of large molecules such as lipids, which are usually bound to carrier plasma proteins in blood. Besides, synovial fluid is an ultrafiltrate of plasma which preventes a very large portion of lipids as in large complexes, from being transported into synovial fluid⁴¹. In the current study, we found that the enzyme ACC1, which commits the first step of de novo FA biosynthesis was significantly upregulated in cartilage tissue from mice with HFD-induced obesity. This result suggested that chondrocytes increased their FA biosynthetic capability in obesity. We consistently found that primary chondrocytes cultured in GI condition accumulated more intracellular lipids, and this was accompanied by activation of ACC. We further showed that pharmacological inhibition of ACC using ND-630 effectively blunted lipids intracellular accumulation. Two distinct isoforms of ACC exist: ACC1 and ACC2^42,43. Cytosolic ACC1 catalyzes the irreversible carboxylation of acetyl-CoA to produce malonyl-CoA. The malonyl-CoA produced by ACC1 is further utilized by FA synthase (FASN) as building materials for FA biosynthesis and subsequent chain elongation^44,45. In contrast, ACC2 is localized in the outer membrane of mitochondria. ACC2-generated malonyl-CoA functions as a potent inhibitor for carnitine/palmitoyl-transferase 1 (CPT1) activity and transfer of the fatty acyl group into the mitochondria for β-oxidation⁴⁶. Further in vivo studies are warranted using ACC1 cartilage-specific knockout mice to test if ACC1-mediated FA biosynthesis would play a role in cartilaginous lipid deposition and its contribution to obesity-associated OA pathogenesis.

In summary, this is the first study to evaluate the FAs and lipids profile in mouse cartilage tissue under obesity condition. The results extend our understanding of cartilaginous FA/lipid under different conditions. Our results also provide both in vitro and in vivo evidence that the modulation of de novo FA biosynthesis through regulating ACC activity could potentially correct the excessive accumulation of FAs in obesity.