Rac deletion in osteoclasts causes severe osteopetrosis

Abstract

Cdc42 mediates bone resorption principally by stimulating osteoclastogenesis. Whether its sister GTPase, Rac, meaningfully impacts upon the osteoclast and, if so, by what means, is unclear. We find that whereas deletion of Rac1 or Rac2 alone has no effect, variable reduction of Rac1 in osteoclastic cells of Rac2(-/-) mice causes severe osteopetrosis. Osteoclasts lacking Rac1 and Rac2 in combination (Rac double-knockout, RacDKO), fail to effectively resorb bone. By contrast, osteoclasts are abundant in RacDKO osteopetrotic mice and, unlike those deficient in Cdc42, express the maturation markers of the cells normally. Hence, the osteopetrotic lesion of RacDKO mice largely reflects impaired function, and not arrested differentiation, of the resorptive polykaryon. The dysfunction of RacDKO osteoclasts represents failed cytoskeleton organization as evidenced by reduced motility of the cells and their inability to spread or generate the key resorptive organelles (i.e. actin rings and ruffled borders), which is accompanied by abnormal Arp3 distribution. The cytoskeleton-organizing capacity of Rac1 is mediated through its 20-amino-acid effector domain. Thus, Rac1 and Rac2 are mutually compensatory. Unlike Cdc42 deficiency, their combined absence does not impact upon differentiation but promotes severe osteopetrosis by dysregulating the osteoclast cytoskeleton.

Fig. 1.

LysMRacDKO mice are osteopetrotic. (A) Micro-CT image of distal tibia of 7-week-old WT (i.e. Rac1^f/f), LysMRac1^ΔOCΔOC, Rac2^−/− and LysMRacDKO mice. (B) Radiograph of tibia and coccygeal vertebrae of WT and LysMRacDKO mice. (C) Toluidine Blue staining demonstrating osteopetrotic cartilaginous bars (purple reaction product) in tibia of LysMDKO mice. The black box indicates the region of magnification shown in the right-most panel. (D) TRAP staining (red reaction product) of same bones illustrated in C. Black boxes indicate the region of magnification shown in the lower panels. Scale bars: 500 μm (C, left two panels; D, upper panels); 100 μm (C, right-most panel); 50 μm (D, lower panels).

Fig. 2.

LysMRacDKO osteoclasts fail to effectively resorb bone. Equal numbers of WT, LysMRac1^ΔOCΔOC, Rac2^−/− and LysMRacDKO BMMs were cultured on bone slices for 7 days in the presence of M-CSF and RANKL. (A) The cells were removed and the resorption pits visualized by wheat germ agglutinin-lectin staining and outlined. Scale bars: 20 μm. (B) Culture medium was analyzed for content of CTx (CTX), a marker of global bone resorption. ***P<0.001 compared with WT.

Fig. 3.

LysMRacDKO osteoclast number is reduced in vitro. (A) WT and LysMRacDKO BMMs were cultured for 6 days in M-CSF and RANKL. The cells were stained for TRAP activity and the number of osteoclasts counted. (B) WT, LysMRac1^ΔOCΔOC (Rac1KO) and LysMDKO BMMs were cultured in M-CSF and RANKL with time. β3 integrin subunit, Src (c-Src) and cathepsin K (CatK) were measured, by immunoblotting, as a function of osteoclast differentiation. β-actin serves as a loading control. (C) The percentage of CD11B- and F4/80-expressing cells in marrow of WT and LysMRacDKO mice was determined by FACS analysis. (D) WT and LysMRacDKO BMMs were cultured for 60 hours in M-CSF and RANKL at which time apoptosis was measured by ELISA. (E) WT and LysMRacDKO BMMs were maintained in M-CSF and RANKL for 3 days. The cells were starved of cytokine and serum for 3 hours and exposed to 100ng/ml RANKL, with time. Lysates were immunoblotted for phosphorylated Akt S473 and Akt T308. Total Akt (tAKT) serves as loading control. **P<0.01; ***P<0.001.

Fig. 4.

Rac organizes the osteoclast cytoskeleton through its effector domain. (A) WT, LysMRac1^ΔOCΔOC, Rac2^−/− or LysMRacDKO BMMs, on plastic, were differentiated into osteoclasts by 5 days of exposure to M-CSF and RANKL. The cells were stained for TRAP activity. (B,C) WT and mutated Rac1 constructs were retrovirally transduced into LysMRacDKO BMMs. (B) Expression of each transductant was determined by immunoblotting. Empty vector (pmx) transduced into WT and LysMRacDKO BMMs served as a control. (C) Left panels: osteoclasts, generated on plastic were stained for TRAP activity (purple reaction product). Right panels: osteoclasts generated on bone were removed and resorption pits, visualized by wheat-germ-agglutinin–lectin staining, were outlined. (D) WT or LysMDKO BMMs were maintained in M-CSF and RANKL for 3 days to generate pre-fusion osteoclasts. The cells were lifted and placed in the upper chamber of transwell dishes. The lower chamber contained unaltered medium or vitronectin-coated filter surface with or without recombinant h-CSF and RANKL. After 8 hours, the filters were stained with crystal violet and the number of prefusion osteoclasts present on the lower filter surface determined. ***P<0.001 compared with WT. Scale bars: 200 μm (A,C, right-hand panels); 100 μm (C, left-hand panels).

Fig. 5.

Cytoskeletal organization of LysMRacDKO osteoclasts is deranged. (A) WT, LysMRac1^ΔOCΔOC, Rac2^−/− and LysMRacDKO BMMs were cultured on bone slices for 7 days in the presence of M-CSF and RANKL. The actin cytoskeleton was visualized by Alexa-Fluor-488–phalloidin staining. Actin rings are present in WT, LysMRac1^ΔOCΔOC and Rac2^−/− osteoclasts, and podosomes are dot-like structures in LysMRacDKO osteoclasts. (B,C) WT and LysMDK osteoclasts were generated on bone slices as detailed in A. The specimens were plastic embedded. (B) Semi-thin sections were prepared. WT osteoclasts (left panel) are juxtaposed to bone in resorption lacunae. By contrast, LysMDKO osteoclasts are thin, elongated and poorly apposed to bone, which they fail to resorb. (C) Transmission electron microscopy demonstrates actin rings (asterisk) encompassing ruffled border (RB) in WT osteoclasts (10,000× magnification) and absence of both structures in LysMRacDKO cells (20,000× magnification). Scale bars: 100 μm.

Fig. 6.

Rac regulates Arp3 distribution in osteoclasts. WT and LysMRacDKO osteoclasts, generated on bone, were starved of serum and cytokine and then exposed to M-CSF for 5 minutes. The cells were fixed, and Arp3 was visualized by immunofluorescence. Scale bar: 20 μm.

Fig. 7.

CatKRacDKO osteoclasts resorb bone in vitro. (A) WT (Cre-) and CatKRacDKO BMMs were cultured with M-CSF and RANKL with time. Lysates were subjected to Rac1 immunoblots daily for 5 days to assess Rac1 deletion in the total culture. β-actin serves as the loading control. (B) WT and CatKRacDKO BMMs, cultured with M-CSF and RANKL for 5 days, on plastic, were stained for TRAP activity (scalebar). (C,D) WT and CatKRacDKO BMMs were cultured with M-CSF and RANKL on bone slices. (C) After 7 days the cells were stained with Alexa-Fluor-488–phalloidin to visualize the actin cytoskeleto. (D) Following removal of the cells, the resorption pits were visualized by wheat-germ-agglutinin–lectin staining and outlined. Scale bars: 200 μm (B); 100 μm (C,D).

Fig. 8.

CatKRacDKO mice are osteopetrotic. (A) Sequential micro-CT images of tibia of WT (Cre-) and CatKDKO mice at 4 and 7 weeks of age. (B) Histological sections of tibias of 4-week-old WT and CatKRacDKO mice stained with TRAP (red reaction product). (C) Toluidine-Blue-stained histological section of CatKRacDKO tibia demonstrating growth plate and persistence of cartilaginous bars (scalebar). (D) Higher magnification of the bones illustrated in B demonstrating CatKRacDKO osteoclasts are attenuated and poorly attached to bone. (E) WT and CatKRacDKO mice demonstrating failure of tooth eruption in the latter. Scale bars: 500 μm (B); 100 μm (C,D).