Intracellular iron uptake is favored in Hfe-KO mouse primary chondrocytes mimicking an osteoarthritis-related phenotype

HFE-hemochromatosis is a disease characterized by a systemic iron overload phenotype mainly associated with mutations in the HFE protein ( HFE ) gene. Osteoarthritis (OA) has been reported as one of the most prevalent complications in HFE-hemochromatosis patients but the mechanisms associated with its onset and progression remain incompletely understood. In this study, we have characterized the response to high iron concentrations of a primary culture of articular chondrocytes isolated from newborn Hfe-KO mice and compared the results to a similar experiment developed in cells from C57BL/6 (WT) mice. Our data provide evidences that both WT and Hfe-KO derived chondrocytes, when exposed to 50 µM of iron, develop characteristics of an OA-related phenotype, such as an increased expression of metalloproteases, a decreased extracellular matrix production and a lower level of expression of aggrecan. In addition, Hfe-KO cells also showed an increased expression of Mmp3 , indicating an increased susceptibility to intracellular iron accumulation. Accordingly, upon treatment with 50 µM of iron, these chondrocytes were found to preferentially differentiate towards hypertrophy with increased expression of collagen I, transferrin and downregulation of SRY (Sex-Determining Region Y)-box containing gene 9 ( Sox9 ). In conclusion, high iron exposure can compromise chondrocyte metabolism which, when simultaneously affected by an Hfe loss of function, appear to be more susceptible to the establishment of an OA-related phenotype. or 50 μM ferric citrate, or in vehicle ([Na-Cit]); Results are from n=5 pools of three cell culture wells (9.6cm 2 ) for each genotype (WT and Hfe-KO).Each pool corresponded to one independent experiment. All results were normalized by Gapdh and statistical analysis was done by Two-way ANOVA followed by Tukey multiple comparison test with p<0.05; Statistical significance: (a)- compared to control [Na-Cit]; (b)- compared to [Fe-Cit 10 µM]); For Bglap2 expression WT groups did not show a normal distribution, therefore multiple comparison between treatments WT chondrocytes was done by Kristal-Wallis method and multiple comparison with Dunn’s post-hoc test. For Hfe-KO chondrocytes One-way ANOVA was used to compare the impact of treatments with control, with Tukey test for post-hoc multiple comparison with p<0.05. and matrix metallopeptidase 3 (MMP-3) in articular cartilage.- Immunohistochemistry for knee articular cartilage in tibia of control (WT MNX) and Hfe-KO (Hfe-KO MNX) mice at 8 weeks after MNX surgery. ADAMTS5 and MMP3 detection was done with DAB Kit SK4100 and correspond to brown staining. Counterstaining was done with 1% methyl green, which stains hyaline cartilage tissues. The comparison between WT and Hfe-KO showed higher degree of ADAMTS5 and MMP3 detection in Hfe-KO mice relative to WT.


1-Introduction
Articular cartilage is a highly specialized tissue, avascular and hypocellular, composed primarily of water and extracellular matrix produced by chondrocytes, the only cell type present in this tissue [1]. Chondrocytes are specialized cells responsible for the production of extracellular cartilage matrix (ECM) composed by several macromolecules including aggrecan and type II, IX and XI collagens, which are important for the maintenance of the cartilage properties including tensile strength and flexibility [1,2].
Osteoarthritis (OA) is the most frequent adult joint disease and is characterized by structural and molecular modifications in subchondral bone, synovial membrane, muscle, ligament and cartilage, that ultimately lead to joint destruction [3][4][5]. Several factors contribute to OA development including mechanical overload, ageing, genetic background, obesity, metabolic syndromes and the presence of calcium-containing microcrystals in the joint. However its underlying physiopathology is still incompletely understood [3][4][5].
Hemochromatosis is an inherited metabolic disorder characterized by systemic iron overload and is associated to mutations in genes involved in iron metabolism [6,7].
The most common gene affected is the HFE, encoding a protein that functions as an important sensor for systemic iron levels, the most prevalent HFE pathogenic mutation being the C282Y (p.Cys282Tyr) which results in the substitution of an adenine (A) for a cytosine (C), (NG_008720.2:g.10633G4A; NM_000410.3:c.845G4A), and individuals homozygotes for this mutation have an increased risk of developing iron overload. The association between OA and HFE-hemochromatosis was first described by Schumacher in 1964, who showed an increased prevalence of arthritis in those patients [8]. Since then, several studies have confirmed this observation [9][10][11][12][13][14], which suggested an association between elevated iron concentrations, HFE loss of function and articular cartilage integrity, however it was not clear which could be the main factor affecting cartilage metabolism. Recently, we have shown that Hfe-KO mice develop more severe knee OA than wild type (WT) mice after joint destabilization induced by partial meniscectomy [14], but the underlying molecular mechanisms remain unclear. Therefore, we have developed an in vitro approach to further investigate the molecular events occurring in chondrocytes exposed to high levels of ferric citrate. Primary cultures of articular chondrocytes derived from Hfe-KO mice were exposed to 50 µM of ferric citrate in order to identify molecular factors and pathways involved in chondrocyte metabolism which

2.1-Ethical statement
All personnel involved in animal handling and experimentation received proper training

2.2-Biological models used in this study
Animals were maintained under pathogen-free conditions in individually

Page 4 of 34 BioFactors
The method for isolation of primary chondrocytes was adapted from Gosset and colleagues (2008) [18]. In short, newborn mice (P6-day 6 after birth) were euthanized according with the recommendations of FELASA (Federation of European Laboratory Animal Science Associations). For isolation of articular cartilage, mouse anterior legs were dissected by removing the skin and soft tissues in order to isolate femoral heads and tibial plateau using a modular stereomicroscope (Leica MZ6). The articular cartilage pieces were submitted twice to 45 minutes incubation with 2mg/ml collagenase D (Gibco) solution followed by an overnight digestion with 0.5mg/ml of collagenase D (Gibco). All collagenase solutions were diluted in culture medium (Dulbecco's modified Eagle's medium (DMEM), (Gibco), supplemented with 1% Penicillin/Streptomycin (250 μg/ml, Gibco) and 1% Fungizone (250 μg/ml, Gibco). Articular cartilage pieces were retrieved and passed sequentially through 25-, 10-, 5-and 2-ml pipettes and filtered through a 40μm cell strainer (Fisherbrand, nylon mesh) followed by 2000g centrifugation for 10 minutes at room temperature. Cell pellets were resuspended in 10ml of culture medium supplemented with 10% Fetal Bovine Serum (FBS, Gibco). The chondrocyte suspension was counted in a hemocytometer and cell viability was determined by a quick trypan blue exclusion test (0.4%, Sigma). The isolated chondrocytes were seeded at 25,000 cells/cm 2 /well onto a 12 well culture dish and cultures (Sarstedt) allowed to reach confluence. From this point onward FBS supplemented medium was changed every two days.

2.4-Preparation of ferric citrate treatments
To determine the ferric citrate concentrations to be applied in primary culture of chondrocytes, a preliminary assay with several ferric citrate concentrations was performed (data not shown) and two concentrations were chosen: 10µM, which induced significant changes in iron metabolism markers, and 50µM, which induced not only iron metabolism changes but also chondrocyte metabolism alterations. In addition, the iron concentrations chosen were within the physiological range identified for Hfe-KO mice, which, in the presence of iron overload, showed total iron concentrations in plasma ranging between 30 and 60µM [19,20]. A negative control for iron presence was used by substituting the iron chloride solution by purified water (Sigma) [Na-Cit]. ferric citrate treatments were applied by adding it at each medium change, which occurred every two days, and evaluation for the effect of high ferric iron concentrations was done at days 7 (1 day after last treatment) and 10 (3 days after last treatment) of culture.

2.5-Qualitative evaluation of extracellular cartilage matrix (ECM) production
The evaluation of ECM production was done by staining with alcian blue the primary chondrocyte cultures, which had been proliferating for 10  for 30 min at -20ºC. Then cells were rinsed 2x with PBS and incubated for at least 1hour with 1ml of 0.5% alcian blue (8GX-Sigma-Aldrich) dissolved in HCl (1M-Sigma-Aldrich). After staining, each well was rinsed 2x with PBS and left in 1ml of PBS for analysis under a light-contrast microscope (Axiovert 25). Cell plates were photo-scanned and separated into three color channels (blue, green and red) for detection of pixel intensity of alcian blue staining using Image J. software v.151h (https://imagej.net/, RRID:SCR_003070) [21]. For each condition, three independent wells were evaluated and measured.

2.6-Evaluation of expression of molecular markers
RNA extractions of WT and Hfe-KO chondrocytes were done with High Pure RNA Isolation Kit (Roche) according with the manufacture instructions, which includes a DNase treatment during the procedure. For each condition tested, cells from three individual wells (Nucleon Delta Surface-Thermo scientific; 9.6cm 2 ) were pooled and used for each RNA extraction. RNA samples were quantified, assessed for puritity using a Nanodrop 1000 (Thermo Scientific) and for integrity using the Experion system (Bio-Rad). Then, 1μg of RNA from each condition tested was reverse transcribed using 1µL of oligo(dT) adapter primer (10 µM) and 1µL of M-MLV Reverse Transcriptase (1U/µL, Invitrogen) following manufacturer's instructions. The qPCR reaction (20μl) was performed using as template 2μl of 1:10 cDNA dilution (0.01μg) for each condition tested, 10µl of SsoFast Evagreen Taq supermix (BioRad), 0.6 µl of either target gene primers (300 nM) or housekeeping gene primers (300 nM), and the final volume completed with purified RNase-free water (Sigma). Amplification by qPCR was performed in a Step One Plus Realtime PCR System (Applied Biosystems). Results are from five different experiments. To determine the levels of gene expression, all results were normalized against Gapdh levels of expression and relative expression was determined by the ΔΔCt method [21] using as control group for basal expression cells without treatment, unless stated otherwise. All primer sequences are described in the supplementary data (Table A1).

2.7-Determination of intracellular iron
The evaluation of intracellular iron concentrations was only done at 10 days of culture when extracellular matrix production was significantly altered, in order to characterize the corresponding intracellular iron status. incubated for 2h at 60ºC. After this, 30µL of freshly prepared iron detection solution (Ferrozine (6.5mM), Cuproine (6.5mM), Acetate ammonium (2.5M) and Ascorbic acid (1M); all from Sigma-Aldrich) was added to each sample and 30 minutes later 280µL of the mix was removed and absorbance measured in a microplate reader (Synergy4) at 550nm. The same procedure was done for calibration curve samples diluted in HCl (10mM). To normalize iron quantification the corresponding protein concentrations in cell lysate were determined using the Bio-Rad Protein assay kit (Bio-Rad) according with manufacture recommendations.

2.8-Evaluation of articular cartilage integrity by OARSI score
WT and Hfe-KO mice (male and female) knee articulations were collected after euthanasia and fixed for 24h in 4% paraformaldehyde in PBS pH7.4, followed by decalcification with 0.5M Ethylenedinitrilotetraacetic acid (EDTA) in PBS pH 7.4 for 3 weeks. Samples were then dehydrated with graded alcohols, cleared with Xylol and included in paraffin, followed by 5μm sagittal sectioning in a rotary microtome (Microm HM340E, Germany) and stained with Safranin-O/Fast Green/Meyer's Hematoxylin. The evaluation was done in two sections per level and in three levels separated by 100µm each, starting at menisci separation. Two separate observers, blinded to the strain and type of intervention, graded the cartilage lesions using the semi-quantitative scoring system proposed by Glasson et al. 2010, also known as the OARSI scoring system for murine osteoarthritis. This system is based on the cartilage destruction extent, graded 0-6 depending on the depth of the lesion and on the percentage of the articular surface affected. The score for each knee was the sum of the scores from each of the three levels.

2.11-Statistical analysis
Unless stated otherwise, results are expressed as the mean and standard deviation (SD). Each group was evaluated for normality with Kolmogorov-Smirnov test and where normality was confirmed, the One-way ANOVA was used with Tukey test for post-hoc multiple comparison with p<0.05.

Hfe-KO primary chondrocyte cultures morphology and molecular phenotype
Primary chondrocytes were isolated from the articular cartilage of the knee joint from Hfe-KO and WT newborn mice and its morphological, histological and molecular profiles analyzed. In terms of morphology, no significant differences were observed, with

3.4-Inflammatory and antioxidant response to increased iron concentrations
Because inflammatory pathways have been associated to onset of OA phenotype, the expression of interleukin-1-beta (Il-1b) and interleukin-6 (Il-6) were evaluated in the    [29][30][31]. In addition, HFE loss of function can lead to increased intracellular iron absorption [32,33]. The fact that articular cartilage is an avascular tissue [1] suggests that HFE can only contribute to OA progression if articular cartilage is exposed to blood, where it is possible to find both iron bound (TBI) and non-bound (NTBI) to transferrin. This is expected in HFEhemochromatosis patients following a trauma event and/or lesion in the articulation associated with bleeding and the corresponding increase in iron exposure.

4-Discussion
Because it is known that HFE-hemochromatosis leads to a progressive iron  [37,38]. In Hfe-KO chondrocytes, levels of intracellular iron were significantly higher than in WT cells when subject to 50 µM of ferric citrate (Figure 1-B). This suggests a deregulation in the mechanism of iron import in those cells, probably related to the absence of a functional HFE protein, known to play an essential role in the regulation of iron transport into the cells by competing with TRF for TFRC binding [23]. If TRF binds to TFRC, it promotes the internalization of TFRC/TRF complex along with solute carrier family 11 member 2 (SLC11A2, also known as DMT-1) leading to intracellular iron absorption [39]. In this study, Tfrc was not significantly downregulated in Hfe-KO chondrocytes, in contrast to results in WT cells (Figure 1-A). Because SLC11A2 is a key player in intracellular iron uptake following TFRC/TRF internalization, the absence of a functional HFE could favour intracellular iron accumulation since it was previously shown that HFE/β2microglobin complex is able to colocalize and bind with SLC11A2 and inhibit iron absorption [40]. However, regarding NTBI cellular uptake it was showed that SLC11A2 was not able to impair NTBI absorption in the liver enterocytes only transferrin bound iron (TBI) [41], contrarily to what was observed in enterocytes [36]. SLC11A2 is a ferrous iron transporter [39], and the treatment was done with ferric citrate, which suggests that the iron uptake may be done through TFRC/TRF internalization with the influence of HFE, unless pH and/or oxygen levels decrease. In addition, we have analyse the expression of Slc39a14, known NTBI intracellular transporter [42], was possible to observe that Slc39a14 was upregulated when cells were subject to 10 µM of ferric citrate, however upon 50 µM treatment Slc39a14 return to basal expression (Figure 1-C), suggesting a response to inhibit iron uptake and without the influence of HFE.
Since hepcidin expression was only detected at residual levels (data not shown), this led us to hypothesize that the significant differences observed in response to iron metabolism markers in Hfe-KO chondrocytes relative to WT should be primarily connected with the absence of HFE protein, responsible for regulating the entry of iron inside the cells [36,43], and not with the regulation of ferroportin (SLC40A1) degradation by hepcidin [34,44]. The absence of an increased expression of inflammatory markers in both WT and Hfe-KO cells in this study was also previously observed during OA progression in Hfe-KO mice [14], and may be connected with primary response to stress, as also described previously in a blood induced joint damage study [45]. In that study, upregulation of Il-1b and Il-6 occurred only in the first hours after bleeding stimuli, and the levels of inflammatory markers decreased to basal levels after 24h, which could explain why we did not observe significant changes in the expression of pro-inflammatory markers, given the fact that we only analysed gene expression 24h and 72h after the last treatment. It will therefore be important in the future to study the inflammatory response in chondrocyte cultures in the first hours after ferric citrate treatment to understand the impact of inflammation in this context. Increased ROS concentrations have been associated with OA phenotype [46,47] which led us to evaluate indirect evidences of oxidative stress by analysing the expression of molecular markers associated with antioxidant response, like Cat, which showed to be upregulated in both WT and Hfe-KO cells (Figure 3-A). These results suggested that iron overload promoted an antioxidant response in chondrocytes [48,49]. Another marker known to be sensitive to iron availability is Epas1, which is associated with hypoxia [50,51] and one important catabolic regulator of cartilage destruction [52]. Our results show significant upregulation of Epas1 in response to the higher ferric citrate treatment (50 µM) (Figure 3-B) suggesting that Epas1 should be contributing for the establishment of OA phenotype on both WT and Hfe-KO chondrocytes.
Bmp6, besides being one of the main regulators of hepcidin and of iron metabolism [53][54][55], has also been shown to be an inducer of chondrocyte proliferation and ECM production [56,57]. The results for Bmp6 expression (supplementary figure   S3) suggested that, after an initial response to iron increase (10µM) in WT and Hfe-KO chondrocytes, and upon higher iron concentrations, Bmp6 expression was restrained. This was indicative of a decrease in the SMAD 1, 5 and 8 pathway [53,58], and reported to donwregulate Col10a1 and Sox9 and increase col1a1 and bglap2 [59], which is in agreement with what we observed ( In OA, the expression of Col10a1 has been mostly shown to be increased [14, 60,62], however downregulation of Col10a1 in human OA-associated articular chondrocytes has also been reported [61], similar to the gene expression pattern found in our study for Hfe-KO chondrocytes (Figure 4-A). Accordingly, upregulation of Bglap2 and downregulation of Sox9 in Hfe-KO chondrocytes (Figure 4-B) are in agreement with data reported previoulsy and shown to be associated with OA phenotypes [63][64][65], thus contributing to the significant impact of iron in chondrocyte differentiation towards hypertrophy. In addition, expression of the runt related transcription factor 2 (Runx2), an important transcription factor associated with chondrocytes hypertrophy [64], was not significantly changed by ferric citrate treatments (results not shown), whicht could explain the absence of Col10a1 upregulation in this context and suggests the existence of an alternative mechanism to induce an hypertrophy-like phenotype in WT and Hfe-KO chondrocytes subjet to 50 µM of ferric citrate. Indeed, Trf expression was significantly upregulated (Figure 4-B). Since this increase in expression has not been directly associated with an iron increase [66], it could be associated with the acceleration of chondrocyte differentiation process towards an hypertrophy-like phenotype, an hypothesis that remains to be further explored, however this response could favour intracellular iron uptake through TRF/TFRC complex in the absence of a functional HFE.  [67]. In addition, extracellular matrix was significantly affected throughout progressive iron overload exposure, both in Hfe-KO and WT chondrocytes ( Figure 5-A).
These results were also confirmed by the in vivo increase in MMP3 and ADAMTS5 ( Figure 6) and consequent decrease in methyl green staining (to detect hyaline cartilage [68]) in Hfe-KO mice subject to MNX relative to WT MNX animals, also observed in our previous work [14]. The significant upregulation of Mmp3 found in Hfe-KO chondrocytes suggests a possible synergistic effect between iron overload exposure and HFE loss of function. Altogether, these results are in agreement with published data since increase in MMP3 protein expression was previously associated to osteoarthritis phenotype and to extracellular matrix degradation [62,67]  The exposure of WT and Hfe-KO chondrocytes to ferric citrate loads favoured iron accumulation and the development of an OA-similar phenotype, which is in agreement to what was described previously with the identification of iron deposits in the articular cartilage and synovial membrane of HFE-hemochromatosis patients and cartilage associated degradation [69][70][71]. Iron deposits were also detected in thalassemia patients articulation [72], suggesting the putative toxic effect of iron overload inside articulations. Epidemiology studies report high prevalence of OA associated HFEhemochromatosis in patients averaging 60 years of age, with a strong association between transferrin saturation and serum ferritin concentrations [26,73]. Evidences showed that iron overload should play an important role on OA progression in HFE-hemochromatosis patients [14,26]. In addition, these patients showed higher susceptibility for the prevalence of OA when compared with non-HFE-hemochromatosis patients [74], which suggests additional susceptibility beyond systemic iron overload. Recently, it was reported a case were an HFE-hemochromatosis patient showed an early onset of knee OA (since he was 18 years old) [75]. The patient systemic iron parameters were considered normal even with constant OA related complains, however before the patient underwent bilateral knee replacement, significant iron deposits were observed in the knee synovial membrane [75], which may suggest a susceptibility associated with iron accumulation in synovium. Accordingly, the absence of HFE in cells cytoplasmatic membrane as consequence of pC282Y substitution [76] were previously shown to favor intracellular iron accumulation [33]. In addition, attending that HFE is a protein belonging to the major histocompatibility complex (MHC-I) family, we cannot exclude that OA associated HFEhemochromatosis could also be related with a possible immune response to the absence of pC282Y HFE in cytoplasmic membrane [77]. There is also a report in which hand osteoarthritis prevalence is increased in heterozygotes (pC282Y) [11], however a recent study showed no significand differences between the prevalence of OA in pC282Y heterozygotes compared to general population [78]. Therefore, further studies regarding HFE recognition by the immune system are required.
In conclusion, exposure of WT and Hfe-KO chondrocyte cultures to high iron concentrations promoted the establishment of an OA-like phenotype suggesting that high levels of iron can contribute to the progression of OA associated with HFEhemochromatosis. In addition, results suggest that Hfe loss of function can contribute to promote an increase in intracellular iron accumulation in primary chondrocytes.
However, this study cannot exclude the hypothesis that the onset of OA associated with HFE-hemochromatosis could be related with acute immune responses, even if the levels of inflammation are not comparable with those observed in chronic inflammation diseases like rheumatoid arthritis [70]. We propose that, upon articular lesion and/or trauma,       ADAMTS5 and MMP3 detection was done with DAB Kit SK4100 and correspond to brown staining. Counterstaining was done with 1% methyl green, which stains hyaline cartilage tissues. The comparison between WT and Hfe-KO showed higher degree of