Targeting TRPC channels for control of arthritis-induced bone erosion

Abstract

Arthritis leads to bone erosion due to an imbalance between osteoclast and osteoblast function. Our prior investigations revealed that the Ca²⁺-selective ion channel, Orai1, is critical for osteoclast maturation. Here, we show that the small-molecule ELP-004 preferentially inhibits transient receptor potential canonical (TRPC) channels. While ELP-004 minimally affected physiological RANKL-induced osteoclast maturation in murine bone marrow- and spleen-derived myeloid cells (BMSMCs) and human PBMC-derived cells, it potently interfered with osteoclast maturation driven by TNFα or LTB4. The contribution of TRPC channels to osteoclastogenesis was examined using BMSMCs derived from TRPC4^-/- or TRPC(1-7)^-/- mice, again revealing preferential interference with osteoclastogenesis driven by proinflammatory cytokines. ELP-004 also reduced bone erosion in a mouse model of rheumatoid arthritis. These investigations reveal TRPC channels as critical mediators of inflammatory bone erosion and provide insight into the major target of ELP-004, a drug currently in preclinical testing as a therapeutic for inflammatory arthritis.

Fig. 1.. ELP-004 is a TRPC inhibitor.

(A to D) TRPC3-, TRPC6-, and TRPC7-overexpressing HEK293 cells (n = 3 to 5) plated on coverslips were loaded with Fura2-acetoxymethyl ester (Fura2-AM) (2 μM) before treating with 100 μM OAG (1-oleoyl-2-acetyl-sn-glycerol), Ca²⁺ (1 mM), and ELP-004 (0-1 mM). Each representative trace represents 50 to 100 cells with SEM shown as a dotted line. (D) Ca²⁺ entry was measured, normalized to control and then plotted to calculate IC₅₀s by nonlinear regression. (E to H) TRPC4-, TRPC5-, and TRPC7-overexpressing HEK293 cells (n = 3 to 5) plated on coverslips were loaded with Fura2-AM (2 μM) before treating with Cch (100 nM), Sr²⁺ (3 mM), and ELP-004 (0 to 100 μM). (H) Sr²⁺ entry was measured, normalized to control and then plotted to calculate IC₅₀s by nonlinear regression. (I) qPCR for TRPC(1–7) and 18S RNA was performed in murine BMSMCs incubated for 14 days under the indicated conditions (n = 3). Data were analyzed by two-way ANOVA with Dunnett’s multiple comparisons [P(TRPC expression) < 0.0001, P(treatment) = 0.2479, P(interaction) = 0.0029]. Data below log2(x) = 1 not shown. (J) Representative traces of whole-cell currents at ±100 mV for HEK293 cells transiently expressing mouse TRPC4β. Cells in nominally Ca²⁺-free bath solution (30 s) were treated with 30 nM Englerin A (EA). At ~10 s after, the +100 mV current peak, EA (control), or EA plus ELP-004 (ELP-004, 30 nM) was applied; yellow-green points indicate washout. Currents were normalized to that at the peak. a, b, c, a′, b′, and c′ indicate the time points where current-voltage curves are displayed in fig. S2. (K) Summary of current decrease 45 s after the peak ±10, 30, 100, and 1000 nM ELP-004. Data are presented as means ± SEM. Cell numbers (n) are labeled above each column. Significant differences compared to the desensitization (0) group were assessed with one-way ANOVA followed by Dunnett’s multiple comparisons. *P < 0.05, **P < 0.01, and ***P < 0.001

Fig. 2.. ELP-004 inhibits basal calcium and related NFATc1 translocation in inflammatory cytokine-dependent osteoclastogenesis.

Murine BMSMCs were differentiated to form osteoclasts using mCSF (50 ng/ml) and RANKL (100 ng/ml) (A and C) or mCSF (50 ng/ml), RANKL (1 ng/ml), and TNFα (2 ng/ml) (B and D). (A and B) Cells plated on coverslips were loaded with Fura2-AM (2 μM) and treated as indicated in the presence or absence of ELP-004 (100 nM) before measuring basal Ca²⁺ concentration at days 1,7,14, and 21. ELP-004 had no effect on basal Ca²⁺ levels during RANKL-induced osteoclastogenesis (two-way ANOVA, P > 0.05), but significantly inhibited basal Ca²⁺ for TNFα-induced [TNFα: two-way ANOVA, P(ELP-004) < 0.0001, P(time) < 0.0001, P(interaction) = 0.0260]. (C and D) At the same time points, NFATc1 nuclear translocation was measured by immunocytochemistry; 7-aminoactinomycin D (7-AAD) was used to stain the nucleus. Average Pearson’s correlation of NFATc1 colocalization with nuclear signal was calculated for 50 to 100 cells; representative images of each condition are displayed below. ELP-004 had no significant effect on NFATc1 colocalization in RANKL-induced osteoclasts as determined by two-way ANOVA, although increased nuclear translocation was observed throughout the incubation [P(time) < 0.0001; (C)]. ELP-004 significantly inhibited NFATc1 colocalization for TNFα-induced osteoclasts [two-way ANOVA P(ELP-004) < 0.0001, P(time) < 0.0001, P(interaction) = 0.0002; (D)]. *P < 0.05, **P < 0.01, ***P < 0.001, and ****P < 0.0001.

Fig. 3.. ELP-004 has minimal effect on RANKL-induced osteoclastogenesis.

Murine BMSMCs were grown for up to 21 days in media supplemented with mCSF (50 ng/ml) and RANKL (100 ng/ml) in the presence or absence of ELP-004 (100 nM). Each experiment was repeated a minimum of three times. Cells were fixed and stained for tartrate-resistant acid phosphatase (TRAP) activity at days 1, 7, 14, and 21. (A) Number (top) and intensity (bottom) of TRAP-positive osteoclast-like cells (of 100 cells) were counted for each time point; representative examples of TRAP-stained cells are depicted to the right. No significant changes in TRAP signal were observed. (B) At each time point, cells were also stained with 7-AAD to facilitate counting the number of nuclei per cell; multinucleation increased from day 1 to day 21 with no apparent effect of ELP-004. (C) RANKL-induced osteoclasts were grown on osteo-assay surface plates in the presence or absence of ELP-004 (100 nM). Pits formed by the resorptive activity of these osteoclasts were analyzed at days 7, 14, and 21, and the average pit sizes of 10 to 30 pits are shown along with representative images. Significant ELP-004–induced changes in pit sizes in RANKL-induced osteoclasts were observed at day 21 only [two-way ANOVA P(Interaction) = 0.0005]. **P < 0.01. (D) BMSMCs were differentiated for 14 days in the presence of RANKL before extracting RNA at days 7 and 14 to measure DC-STAMP and OC-STAMP mRNA expression by qPCR. ELP-004 significantly inhibited increase in DC-STAMP expression at day 14; *P < 0.05. Each experiment was repeated a minimum of three times with two technical replicates at each time point. ns, not significant.

Fig. 4.. ELP-004 blocks TNFα-induced osteoclast fusion.

Murine BMSMCs were grown for up to 21 days in media supplemented with mCSF (50 ng/ml), RANKL (100 ng/ml), and TNFα (2 ng/ml) in the presence or absence of ELP-004 (100 nM). Each experiment was repeated a minimum of three times. Cells were fixed and stained for TRAP activity at days 1, 7, 14, and 21. (A) Number (top) and intensity (bottom) of TRAP-positive osteoclast-like cells (of 100 cells) were counted for each time point; representative examples of TRAP-stained cells are depicted to the right. ELP-004 had no effect on the number of TRAP-positive cells but significantly inhibited TRAP intensity [two-way ANOVA P(ELP-004) < 0.0001, P(time) < 0.0001, and P(interaction) < 0.0001]. (B) At each time point, cells were also stained with 7-AAD to facilitate counting the number of nuclei/cell; multinucleation increased from day 1 to day 21 in untreated, but not ELP-004–treated cells, as determined by two-way ANOVA with multiple comparisons; ***P < 0.001. (C) TNFα-induced osteoclasts were grown on osteo-assay surface plates in the presence or absence of ELP-004 (100 nM). Pits formed by the resorptive activity of these osteoclasts were analyzed at days 7, 14, and 21, and the average pit sizes of 10 to 30 pits are shown along with representative images. ELP-004 significantly inhibited pit formation for TNFα-induced osteoclasts [two-way ANOVA P(interaction) < 0.0001]. (D) BMSMCs were differentiated for 14 days in the presence of RANKL/TNFα before extracting RNA at days 7 and 14 to measure DC-STAMP and OC-STAMP mRNA expression via qPCR. ELP-004 significantly inhibited increase in both OC-STAMP [two-way ANOVA P(interaction) = 0.0051] and DC-STAMP [two-way ANOVA P(interaction) = 0.0017] expression. Each experiment was repeated a minimum of three times with two technical replicates at each time point. *P < 0.05, **P < 0.01, and ****P < 0.0001.

Fig. 5.. TRPC4 is required for osteoclast maturation.

BMSMCs collected from C57, TRPC4^−/−, outbred, and TRPC(1–7)^−/− mice were differentiated for 14 days using mCSF (50 ng/ml) and RANKL (100 ng/ml) or mCSF (50 ng/ml), RANKL (1 ng/ml), and TNFα (2 ng/ml). C57 was control for TRPC4^−/− samples, whereas outbred was control for TRPC(1–7)^−/− samples. Data represent at least three biological replicates. (A and B) Cells were fixed and stained for TRAP activity; nuclei were marked with 7-AAD. The average number of nuclei per 100 cells is shown under each condition [unpaired t test comparing mononuclear and multinuclear cells; P(TRPC(1–7)^−/−, RANKL) = 0.0061; P(TRPC4^−/−, TNFα) = 0.0210; P(TRPC(1–7)^−/−, TNFα) = 0.0017]. (C) Frequency of TRAP-positive cells was plotted against TRAP signal intensity. The frequency was unaffected by loss of TRPC4 but was significantly lowered by loss of all seven TRPC channels when osteoclastogenesis was induced by TNFα only [two-way ANOVA P(interaction<0.0001)]. TRAP signal intensity was also significantly lower in TRPC4^−/− and TRPC(1–7)^−/− than controls when osteoclastogenesis was induced by both RANKL alone or TNFα [two-way ANOVA P(interaction) = 0.0076]. (D) Lack of TRPC4, similar to lack of all seven TRPC channels, significantly reduced cell sizes in both RANKL-and TNFα-driven osteoclasts [C57 versus TRPC4^−/− two-way ANOVA P(cytokine treatment) < 0.0001; outbred versus TRPC(1–7)^−/− two-way ANOVA P(interaction) = 0.0067]. (E and F) BMSMCs were plated on coverslips and loaded with Fura2 to measure basal Ca²⁺ concentration and oscillations. Average basal Ca²⁺ was measured for 30 to 100 cells in the same buffer for 3 to 5 min and plotted against Ca²⁺ oscillations alongside representative calcium traces for each condition. Both average basal Ca²⁺ and Ca²⁺ oscillations for both TRPC4^−/− and TRPC(1–7)^−/− cells were significantly lower than C57 or outbred cells, respectively, when driven by both (E) RANKL and (F) TNFα. *P < 0.05, ***P < 0.001, and ****P < 0.0001.

Fig. 6.. Daily oral treatment with ELP-004–LNE reduces arthritis phenotypes in SKG mice.

(A) Schematic representation of experimental design. Arthritis was induced in male and female SKG mice with a single dose of 2 mg of zymosan. Beginning on day 3 after induction, mice were treated daily with ELP-004–LNE (180 mg/kg) or vehicle via oral gavage. The arthritic index (AI) was assessed weekly by scoring each paw individually on a scale of 0 to 4: 0, normal; 1, mild, but definite redness and swelling limited to individual digits; 2, two swollen joints/digits; 3, three or more swollen joints/digits; and 4, severe swelling of all joints/digits. All paws were evaluated; maximal AI per mouse was 16. The study was ended on day 56, which was 24 hours after final ELP-004–LNE dose. Blood, rear legs/paws, and spine were collected from each mouse for analyses. n = 14 to 15 per sex and treatment. (B) Total body weight on d56 was unaffected by treatment. (C) Concentration of ELP-004 was measured in the bone marrow to assess concentrations at the site of target engagement. Bone marrow was isolated from the femur and tibia 24 hours after the final treatment. (D) Weekly AI assessments were analyzed using a repeated-measures two-way ANOVA with Šídák’s correction for multiple comparisons. (E) The left hind paw of each mouse was assessed by microCT. Examples of bone erosion indicative of arthritis are indicated by white arrows. (F and G) Porosity and pore connectivity density of the tarsus (F) and calcaneus (G) (shown in white boxes) were quantified in each mouse. Cumulative erosion scores of metatarsals are quantified on the right (F). Vehicle- and ELP-004–treated animals were compared using a one-tailed, unpaired t test, or Mann-Whitney, when appropriate. *P < 0.05, **P < 0.01, ***P < 0.001, and ****P < 0.0001.

Fig. 7.. ELP-004–LNE improves bone structure parameters in the SKG mouse model of arthritis.

(A) Femurs and (C) L5 vertebrae in the spine were visualized using μCT. The trabecular bone analyzed in the region of interest (white box) is depicted below. Microarchitecture parameters were quantified in the femurs (B) and spines (D). Vehicle- and ELP-004–treated animals were compared using a one-tailed, unpaired t test, or Mann-Whitney, when appropriate. *P < 0.05.

Fig. 8.. ELP-004 inhibits human osteoclastogenesis.

Human PBMCs from four patients were incubated in media supplemented with (i) mCSF (50 ng/ml; control); (ii) mCSF (50 ng/ml) and RANKL (100 ng/ml); or (iii) mCSF (50 ng/ml), RANKL (1 ng/ml), and TNFα (2 ng/ml) ± ELP-004 (100 nM). (A and B) Human osteoclasts plated on coverslips were loaded with Fura2 after 14 days of growth. Average basal Ca²⁺ was measured for 30 to 50 cells in the same buffer for 3 to 5 min was plotted against variance for each condition. In RANKL alone, ELP-004 lowered average basal Ca²⁺ (unpaired t test, P = 0.0486) but had no effect on Ca²⁺ oscillations [(A)]. ELP-004 significantly lowered both average basal Ca²⁺ (unpaired t test, P = 0.0330) and Ca²⁺ oscillation generation (unpaired t test, P = 0.0189) cultured in TNFα/RANKL [(B)]. Asterisks on the x axis denote comparison of Ca²⁺ oscillation generation, whereas those on the y axis denote comparisons of average basal Ca²⁺. (C to F) Cells were fixed and stained for TRAP and 7-AAD after 14 days of growth. (C) Representative images. (D) Nuclei were stained with 7-AAD; multinucleation was inhibited by ELP-004 treatment in both RANKL- and TNFα-induced osteoclasts [unpaired t test, P(ELP-004, RANKL) = 0.0038; P(ELP-004, TNFα) = 0.0007]. (E) ELP-004 inhibited the number, but not intensity of TRAP⁺ cells driven by RANKL (unpaired t test, P = 0.0222); both the number (unpaired t test, P = 0.00050) and intensity (unpaired t test, P = 0.0001) of TRAP⁺ cells were inhibited in the presence of TNFα; asterisks on y axis denote comparison in TRAP intensity, whereas those in y axis denote comparison in TRAP⁺ cell count. (F) ELP-004 treatment significantly reduced cell size in TNFα-induced osteoclasts only (N = 3). (G) qPCR reveals TRPC1 expression only, with up-regulation in both RANKL- and TNFα-induced osteoclasts compared to macrophages. *P < 0.05, ***P < 0.001, and ****P < 0.0001.