To assess if osteoclast differentiation was associated with alterations in the expression of ECSIT, we started by evaluating ECSIT protein levels in total and mitochondrial-enriched fractions of osteoclast precursors. Specifically, bone marrow-derived macrophage (BMM) cultures were stimulated with RANKL in the presence or absence of E₂, for 6 h. This time point was chosen because of our previous findings that RANKL stimulates complex I activity and E₂ inhibits this effect after 6 h of exposure (10). ECSIT protein levels in total cell extracts were not changed in cultures with RANKL or E₂ (Figure 1A). However, quantification of ECSIT in mitochondrial-enriched fractions revealed a 2-fold increase after RANKL stimulation and treatment with E₂ abrogated this effect. In the absence of RANKL, E₂ had no effect on the levels of ECSIT in mitochondria. These finding are in line with previous evidence that the effects of E₂ on complex I and other measurements of mitochondria activity are seen in the presence, but not in the absence, of RANKL (10).

ECSIT is known to localize to both cytoplasm and mitochondria (12, 14). We next examined whether the accumulation of ECSIT in the mitochondria was due to the translocation of ECSIT from the cytosol to the mitochondria. Using isolated cytoplasmatic and mitochondrial enriched fractions, we found that the increase of ECSIT in the mitochondrial fraction in response to RANKL was associated with a decrease in the cytoplasmic fraction (Figure 1B). These actions of RANKL were attenuated by E₂.

In macrophages, inflammatory signals promote the association between TRAF6 and ECSIT (13). Here, we evaluated ECSIT and TRAF6 interaction by co-immunoprecipitation. RANKL promoted the interaction of ECSIT with TRAF6 while E₂ attenuated this effect (Figure 1C). Overall, these results suggest that RANKL enhances the interaction between ECSIT and TRAF6 and promotes the translocation of ECSIT to the mitochondria, while E₂ prevents these effects.

To examine the role of ECSIT on osteoclastogenesis, we used two different short hairpin RNA (shRNA) plasmids directed against exon 4 (target #1) and against exon 7 (target #2) to knockdown ECSIT in BMMs. Either plasmid was effective in decreasing the protein levels of ECSIT when compared to cells transduced with a control shRNA (Figure 2A). Silencing ECSIT decreased the number of TRAP-positive multinucleated osteoclasts formed after 5 d treatment with RANKL (Figure 2B). We then selected shRNA plasmids (#2) to further characterize the role of ECSIT on osteoclast differentiation. We first examined whether the decreased osteoclastogenesis with ECSIT silencing could be explained by a decrease in Tnfrsf11a which encodes for RANK. Similar to previous findings (15), RANKL increased Tnfrsf11a mRNA levels in control cells (Figure 2C). Silencing ECSIT increased Tnfrsf11a levels, perhaps as a compensatory mechanisms for the decrease in osteoblastogenesis. Therefore, the changes in Tnfrsf11a do not contribute to the decrease in osteoclastogenesis with ECSIT knock-down. As described before, E₂ decreased the number of osteoclasts formed in cultures of cells transduced with control shRNA (Figure 2D). In contrast, E₂ did not alter osteoclast number in ECSIT knock-down cells.

To examine the effects of ECSIT during early osteoclast differentiation, we evaluated the mRNA expression of osteoclast markers in cells cultured with or without RANKL and E₂ for 1 and 3 d. RANKL increased the levels of TRAP in cells transduced with control shRNA after 1d when compared to cells cultured with M-CSF alone (Figure 2E) and E₂ suppressed this effect. Likewise, silencing of ECSIT greatly attenuated the stimulation of TRAP by RANKL. E₂ had no effect in cells lacking ECSIT. As expected, progression of osteoclastogenesis led to a much higher expression of TRAP after 3 d (Figure 2F) when compared to 1 d in the presence of RANKL, in cells with or without ECSIT. Nonetheless, the levels of TRAP remain lower after 3 d in cells silenced for ECSIT as compared to control cells. As seen after 1 d, inhibition of TRAP expression by E₂ in control cells was of similar magnitude as the one seen with ECSIT silencing after 3 d; and, E₂ had no effect in cells lacking ECSIT. Very similar changes caused by RANKL, ECSIT silencing, and E₂ were observed when we examined the expression of cathepsin K, another marker of osteoclast differentiation. In contrast to the changes seen with Acp5 (TRAP) and cathepsin K expression, the RANKL stimulation of NFATC1 mRNA levels was not affected by ECSIT silencing or E₂ (Figure 2F). Together, these results show that knock-down of ECSIT mimics the suppressive effects of estrogens on osteoclast differentiation. Furthermore, these support the premise that ECSIT mediates the rapid effects of both RANKL and E₂ on osteoclast precursors.

We next examined the contribution of ECSIT to the effects of both RANKL and E₂ on mitochondria. Because ECSIT contributes to the assembly of complex I in macrophages (12) (13, 16), we first examined complex I activity. As expected, knock down of ECSIT in BMMs, cultured in the presence of M-CSF alone, decreased complex I activity (Figure 3A). Moreover, as we have previously shown (10), E₂ prevented the stimulatory effect of RANKL on complex I in control cells. In cells silenced for ECSIT, the effects of RANKL and E₂ were abrogated. We next examined key parameters of mitochondria function by performing a Mito Stress test to measure oxygen consumption rate (OCR). Basal, ATP-linked, maximal respiration, and spare respiratory capacity all followed a similar pattern to the one seen with complex I activity (Figures 3B–E). Specifically, the stimulatory effect of RANKL and inhibition by E₂ were prevented in ECSIT silenced cells. In contrast, non-mitochondrial respiration, and proton leak – basal respiration not coupled to energy production – were not affect by RANKL or E₂ in control or ECSIT silenced cells (Figures 3F, G).

Oxidation of NADH at complex I supplies the electron transport chain with electrons for the production of ATP and also regenerates the NAD⁺ pool. In line with the changes in complex I activity and OCR, ECSIT silencing decreased NAD⁺/NADH ratio (Figure 3H). RANKL increased NAD⁺/NADH ratio in control cells, an effects that was attenuated by E₂. ECSIT silencing prevented the effects of both RANKL and E₂. ATP levels, in turn, were not altered by ECSIT knockdown in macrophages because of a compensatory increase in glycolysis (Figures 3J, K), as previously described by Carneiro et al. (16). The stimulation of ATP levels by RANKL was attenuated by E₂ in control cells (Figure 3I). The effect of RANKL on ATP levels was attenuated in the absence of ECSIT while the effect of E₂ was indistinguishable from the one of RANKL. Silencing of ECSIT increased lactate production, a marker of glycolysis (Figure 3J). In control cells, RANKL or E₂ did not impact glycolysis, while in the absence of ECSIT, E₂ modestly decreased lactate levels. Very similar findings were obtained by analyzing the extracellular acidification rate (ECAR) except that no effect of E2 was detected with this assay (Figure 3K). While the lactate in the medium reflects the cumulative changes in glycolysis throughout the experimental period, ECAR reflects the secretion of lactate in real time. Overall, these findings support the notion that ECSIT mediates the early stimulatory effects of RANKL on mitochondria and their abrogation by estrogens.

ECSIT deletion in macrophages leads to mitochondria dysfunction and increased mitochondrial ROS production (13, 16). Because RANKL stimulates mitochondrial ROS which contribute to osteoclastogenesis (17), we examined the role of ECSIT in this process. In the absence of RANKL, mitochondrial ROS production was increased by 2-fold in ECSIT-deleted cells (Figure 4A). Mitochondrial ROS production was stimulated by RANKL in control cell, but this effect was abrogated in ECSIT-deleted cells. E₂ attenuated RANKL-induced ROS production in control, but not in ECSIT lacking cells. Other sources of cellular ROS are the NADPH oxidases (NOXs) present at the cell membrane. To examine whether the effects of ECSIT were specific to mitochondrial ROS, we also evaluated intracellular ROS using DCFDA. ECSIT deletion modestly increased intracellular ROS in the absence of RANKL (Figure 4B). Nonetheless, RANKL increased ROS in both control and ECSIT silenced cells and the inhibitory effects of E₂ were seen in the presence and absence of ECSIT. These results suggest that ECSIT contributes specifically to the generation of mitochondrial ROS in response to RANKL and E₂.

Maintenance of mitochondrial membrane potential (Δψm) is essential for mitochondrial function and alterations in Δψm are associated with mitochondrial damage. As described previously (13, 15), ECSIT deleted macrophages had lower Δψm when compared to control cells. RANKL is known to increase Δψm by stimulating OXPHOS (18, 19). Accordingly, RANKL increased Δψm in control cells while this response was abrogated in cells lacking ECSIT (Figure 4C). Moreover, the inhibitory effect of E₂ on RANKL-induced mitochondrial potential was reduced in ECSIT-deleted cells.

We have previously shown that E₂ stimulates the mitochondrial death pathway in osteoclast progenitor cells, an effect that is dependent on the presence of RANKL (10). Here, we examined whether the pro-apoptotic effects of E₂ are dependent on ECSIT. In control cells, RANKL alone stimulated caspase 3 activity, in line with previous findings showing that RANKL can promote both pro-survival and pro-apoptotic signals simultaneously (20, 21). As expected, E₂ further increased caspase 3 activity (Figure 5A). Silencing ECSIT did not alter BMM apoptosis in the absence of RANKL. Strikingly, in the absence of ECSIT, RANKL increased apoptosis by about 3-fold when compared to the effects in control cells. E₂ further increased apoptosis in the absence of ECSIT, however this effect was much smaller than the one seen in the presence of ECSIT. Similar results were seen when measuring Caspase 9, although the overall magnitude of the effects was attenuated (Figure 5B). These findings suggest that dysregulation of mitochondria due to suppression of ECSIT alters the effects of RANKL on cell survival; and that ECSIT contributes to the pro-apoptotic actions of E₂.

Bone marrow-derived macrophages (BMMs; used as osteoclast precursors) were obtained, as described previously (17, 38), from 3-month-old female C57BL/6 mice. Briefly, whole bone marrow cells were flushed from the tibiae and femora, depleted of red blood cells with ACK buffer, and plated in α-MEM (11900-024, Thermo Fisher Scientific, Hampton, NH, USA), supplemented with 10% FBS and 1% penicillin/streptomycin with macrophage-colony stimulating factor (M-CSF; 10 ng/ml; 416-ML, R&D systems, Minneapolis, MN, USA). Twenty-four hours later, non-adherent cells were plated in Petri dishes with M-CSF (30 ng/ml) for 4 days to obtain BMMs. BMMs were then re-plated and cultured in α-MEM complete medium with M-CSF (30 ng/ml) and receptor activator of nuclear factor kappa B ligand (RANKL; 30 ng/ml; 462-TR, R&D systems), in the presence or absence of E₂ (E1024, Sigma-Aldrich, St. Louis, MO, USA) to perform the assays described below. In all experiments, E₂ was used at a dose of 10^-8 M. To generate mature osteoclasts, BMMs were cultured for 4–5 days with M-CSF and RANKL, with or without E₂ (10^-8 M). Osteoclasts were fixed with 10% neutral buffered formalin for 10 minutes and stained for tartrate-resistant acid phosphatase (TRAP), using the Leukocyte Acid Phosphatase Assay kit, following the manufacturer’s instructions (387A, Sigma-Aldrich). Osteoclasts were defined as multinucleated (>3 nuclei) TRAP-positive cells.

Mitochondria and cytosolic fractions from BMMs were obtained using the Mitochondria isolation kit (89874, Thermo Fisher Scientific) according to the manufacturer’s instructions using 2x10⁷ cells. The pellets were resuspended to a final protein concentration of 2 mg/mL.

Cultured cells were washed twice with ice-cold PBS and lysed with a buffer containing 20 mM Tris-HCL, 150 mM NaCl, 1% Triton X-100, and protease inhibitor mixture, and phosphatase inhibitor cocktail (P8340, Sigma-Aldrich) on ice for 30 minutes. The cell lysates were centrifuged at 13,200 rpm for 15 min at 4°C, and the supernatants were collected in new 1.5 ml microcentrifuge tube. The protein concentration of cell lysates was determined using a DC Protein Assay kit (Bio-Rad, Hercules, CA, USA). The extracted protein (30-40 μg per sample) was subjected to 10%–12% SDS-PAGE gels and transferred electrophoretically onto polyvinyl difluoride membranes (MilliporeSigma). The membranes were blocked in 5% fat-free milk/Tris-buffered saline for 90 min and incubated with each primary antibody overnight, followed by secondary antibodies conjugated with horseradish peroxidase. The following monoclonal antibodies were used: ECSIT (Abcam, Cambridge, UK; ab21288; 1:1000), TRAF6 (Santa Cruz Biotechnology, Santa Cruz, CA; sc-8409; 1:1000), VDAC (Cell Signaling, MA, USA; D73D12; 1:1000), β-actin (Sigma-Aldrich; A5316; 1:5000). Blots were stripped and reprobed with anti–β-actin or anti–VDAC antibodies. Bound antibodies were detected with ECL reagents (MilliporeSigma) and were imaged and quantified with a VersaDoc imaging system (Bio-Rad).

Cells were washed twice with ice-cold PBS and lysed on ice in a buffer containing 20 mM Tris, pH 7.5, 50 mM NaCl, 0.1% NP-40, 2 mM EDTA, and protease/phosphatase inhibitors, incubated on ice for 30 minutes and then cleared by centrifugation at 13,200 rpm for 15 min at 4°C. A total of 1 mg of protein was incubated with 2 ug of the primary antibody overnight at 4°C with rotation. Then the samples were incubated with 40 μl of beads (20421, Thermo Fisher Scientific) for 2 h at 4°C. Immunoprecipitates were centrifuged at 10,000 rpm, 3 min at 4°C, washed 3x with RIPA buffer and resuspended in 2x Laemmli buffer. Samples were further fractioned by SDS-PAGE and analyzed by Western blotting as described above.

Total RNA was purified from cultured BMMs using TRIzol reagent (Invitrogen, Carlsbad, CA). After extraction, RNA was quantified using a Nanodrop instrument (Thermo Fisher Scientific), and 2 μg of RNA was then used to synthesize cDNA using a High-Capacity cDNA Archive Kit (Applied Biosystems, Grand Island, NY) according to the manufacturer’s instructions. Transcript abundance in the cDNA was measured by qPCR using Taqman Universal PCR Master Mix (Thermo Fisher Scientific). The primers and probes for murine Acp5 (Mm 00432448_m1), Cathepsin K (Mm 00484039_m1), Tnfrsf11a (Mm00437132_m1), and NFATC1 (Mm00479445_m1) were manufactured by the TaqMan^® Gene Expression Assays service (Applied Biosystems). Relative mRNA expression levels were normalized to the house-keeping gene ribosomal protein S2 (Mm00475528_m1) using the ΔCt method.

Lentiviral particles, pLKO.1 with short hairpin RNAs (shRNAs), containing ECSIT-specific sequences (GenBank accession No. NM_012029; GTGTCTACTATCACATCCTAA and GCTTCGTAATAAGTGT GTCTA) were purchased from Sigma-Aldrich. Non-targeted shRNA lentiviral particles (pLKO.1-empty vector) were used as shRNA control. Whole bone marrow cells were obtained from 3-month-old female C57BL/6 mice, as described above. Twenty-four hours later, non-adherent cells were submitted to a Ficoll-Hypaque gradient, and cells at the interface were cultured in the presence of M-CSF (30 ng/ml), polybrene (8 µg/ml; Santa Cruz Biotechnology), and the lentiviral particles for 16 h. We further cultured the infected BMMs with M-CSF for 24 h and then added puromycin (2 µg/ml; Santa Cruz Biotechnology) for 48 h to remove uninfected cells.

Mitochondrial respiration was measured, as described previously (10, 38). Briefly, BMMs were plated and cultured with RANKL (30 ng/ml) for 6 h with or without E₂. The culture media was replaced with XF assay media (Agilent), and the plate was kept in a non-CO2 incubator for 1 h at 37°C. Oxygen consumption rate (OCR) was quantified using the Mitostress Test and Seahorse XF96 Extracellular Flux Analyzer (Agilent) according to the manufacturer’s instructions. Oligomycin (3 µM), FCCP (2 µM), and a mix of Rotenone/Antimycin A (2 µM) were injected sequentially to assess the different mitochondrial respiration parameters and the extracellular acidification rate (ECAR).

Cultured BMMs were collected in 300 µl of phosphate buffer and dissociated by 30 passages through a 27-gauge needle, followed by three cycles of freeze/thawing in liquid nitrogen and another round of 30 passages through a 27-gauge needle after addition of 1 ml of 10 mM ice-cold hypotonic Tris buffer (pH 7.6). The mitochondrial-enriched fractions were obtained by adding 200 µl of 1.5 M sucrose solution to the cell homogenate followed by differential centrifugations (600g for 10 min at 2°C and 14,000g for 10 min at 2°C). The mitochondrial pellet was resuspended in 0.3 ml of 10 mM hypotonic Tris buffer (pH 7.6) and the protein content was quantified using the DC Protein Assay kit. Complex I activity was determined using 10 µg of protein according to a protocol developed by Spinazzi et al. (45). Briefly, a mixture containing potassium phosphate buffer, fatty acid–free BSA, KCN and NADH was added to the samples. In separate wells the same mixture was prepared but with the addition of rotenone. After mixing, the baseline was read at 340 nm for 2 min. The reaction was started by adding ubiquinone. After mixing the decrease of absorbance at 340 nm was follow for 2 min. Activity of complex I is calculated by subtracting complex I activity (without rotenone) and rotenone-resistant activity (with rotenone). The enzymatic activities of complex I was further normalized to the activity of citrate synthase, a mitochondrial matrix enzyme, used as a marker of the abundance of mitochondria within a cell. Briefly a mixture containing Tris with Triton X-100, DTNB, Ac CoA, and the sample were mixed and baseline activity was read at 412 nm for 2 min. Reaction was started by adding oxaloacetic acid and the increase in absorbance at 412 nm monitored for 3 min.

NAD⁺ and NADH levels were measured by using the EnzyFluoTM NAD⁺/NADH Assay kit (Bioassay Systems). Briefly, BMMs were plated in a 96-well black-wall tissue culture plate and cultured with RANKL and E₂ for 6 h. NAD and NADH extraction buffers were then added to the respective wells according to the manufacturer’s instructions. The fluorescent signal (_λex/em =530/585 nm) was quantified in a Cytation™ 5 microplate reader (BioTek Instruments, Winooski, VT, USA).

Intracellular ATP levels were measured by a luciferin-luciferase based assay using CellTiter- Glo^® Luminescent Cell Viability Assay (G7570, Promega), according to the manufacturer’s protocol. Briefly, BMMs were cultured in a 96-well white-wall tissue culture plate with RANKL and E₂ for 6 h. Culture media was replaced with 100 µl of assay reagent (CellTiter-Glo Buffer and CellTiter-Glo Substrate). Each extracted samples were mixed for 2 min on an orbital shaker to promote cell lysis, followed by 10 min incubation at room temperature. The luminescence signal was monitored in a Cytation™ 5 microplate reader.

Lactate production was measured using Glycolysis Cell-Based Assay (600450, Cayman), according to the manufacturer’s protocol. Briefly, BMMs were cultured in a 96-well tissue culture plate with RANKL and E₂ for 6 h. Lactate concentration was determined in the supernatant of the wells and the absorbance signal (490 nm) was recorded in a Cytation™ 5 microplate reader and normalized by cell number.

BMMs were seeded in a 96-well black microplate and treated with RANKL and E₂ for 6 h. Cells were incubated with assay medium (NaCl 120 mM, KCl 3.5 mM, NaHCO3 5 mM, NaSO4 1.2 mM, KH2PO4 0.4 mM, HEPES 20 mM, CaCl2 1.3 mM, MgCl2 1.2 mM and sodium pyruvate 10 mM, pH 7.4) containing 100 nM of Tetramethylrhodamine (TMRM) probe for 30 min at 37°C. The fluorescence signal (λ_ex/em =548/574 nm) was quantified in a Cytation™ 5 microplate reader and normalized by cell number.

Mitochondrial-specific probe MitoSOX Red Mitochondrial Superoxide Indicator (M36008, Thermo Fisher Scientific) was used to evaluate mitochondrial superoxide anion and cellular ROS levels were assessed using oxidative stress-sensitive report molecule CM-H2DCFDA (C6827, Life Technologies, Invitrogen). BMMs were seeded in a 96-well black microplate and treated with RANKL and E₂ for 6 h. Culture medium was then replaced by assay medium containing 3 µM of MitoSOX or 5 µM of CM-H₂DCFDA. For MitoSOX assay cells were incubated for 30 min at 37°C. A kinetic assay was then performed for 120 minute to evaluate the MitoSOX oxidation rate measured by fluorescence (λex/em =510/580 nm) using a Cytation™ 5 microplate reader. For CM-H₂DCFDA, cells were incubated for 15 min at 37°C. After incubation, cells were washed twice with 1x PBS, and the fluorescent signal (λ_ex/em =592/517 nm) was measured in a Cytation™ 5 microplate reader. Both measurements were normalized to cell number.

Caspase-9 activity was measured by the Caspase-Glo^® 9 Assay (G8210, Promega), according to the manufacturer’s instructions. Briefly, BMMs were cultured in a 96-well white-wall tissue culture plate with RANKL and E₂ for 24h. Cell culture media was replaced with 100 µl of Caspase-Glo^® 9 reagent. Contents were mixed for 30 seconds and incubated for 30 min at room temperature. The luminescence signal was measured in a Cytation™ 5 microplate reader.

Caspase-3 activity was determined by measuring the degradation of the fluorescent substrate DEVD-AFC (Biomol Research Labs, Plymouth, PA) and protein concentration was measured by a Bio-Rad detergent–compatible kit (Bio-Rad, Hercules, CA), as described previously (10). Briefly, cultured BMMs were lysed with 20 mM Tris-HCl (pH 7.5), 150 mM NaCl, 1 mM EDTA, 10 mM NaF, 1 mM sodium orthovanadate, 5 mg ml^-1 leupeptin, 0.14 U ml^-1 aprotinin, 1 mM phenylmethylsulfonylfluoride, and 1% Triton X-100. Cell lysates were then transferred to a new plate and incubated with 50 mM DEVD-AFC in 50 mM HEPES (pH 7.4), 100 mM NaCl, 0.1% CHAPS, 10 mM DTT, 1 mM EDTA, and 10% glycerol. The released fluorescent signal (λ_ex/em =400/510 nm) was measured kinetically in a Cytation™ 5 microplate reader.

All experiments were repeated at least twice. Depending on the assay, within each experiment we used 3-10 wells per treatment. The representative experiments are shown in the figures and depict individual data point, average, and SD. Statistical analysis and graphical design were performed in GraphPad Prism 8 software (Graphpad Software, CA, USA). Group mean values were compared, as appropriate, by one-way ANOVA or two-way ANOVA with Tukey’s test, after determining that the data were normally distributed and exhibited equivalent variances by D’Agostino & Pearson and Shapiro-Wilk tests.