Evidence of bisphosphonate-conjugated sitafloxacin eradication of established methicillin-resistant S. aureus infection with osseointegration in murine models of implant-associated osteomyelitis

Abstract

Eradication of MRSA osteomyelitis requires elimination of distinct biofilms. To overcome this, we developed bisphosphonate-conjugated sitafloxacin (BCS, BV600072) and hydroxybisphosphonate-conjugate sitafloxacin (HBCS, BV63072), which achieve "target-and-release" drug delivery proximal to the bone infection and have prophylactic efficacy against MRSA static biofilm in vitro and in vivo. Here we evaluated their therapeutic efficacy in a murine 1-stage exchange femoral plate model with bioluminescent MRSA (USA300LAC::lux). Osteomyelitis was confirmed by CFU on the explants and longitudinal bioluminescent imaging (BLI) after debridement and implant exchange surgery on day 7, and mice were randomized into seven groups: 1) Baseline (harvested at day 7, no treatment); 2) HPBP (bisphosphonate control for BCS) + vancomycin; 3) HPHBP (hydroxybisphosphonate control for HBCS) + vancomycin; 4) vancomycin; 5) sitafloxacin; 6) BCS + vancomycin; and 7) HBCS + vancomycin. BLI confirmed infection persisted in all groups except for mice treated with BCS or HBCS + vancomycin. Radiology revealed catastrophic femur fractures in all groups except mice treated with BCS or HBCS + vancomycin, which also displayed decreases in peri-implant bone loss, osteoclast numbers, and biofilm. To confirm this, we assessed the efficacy of vancomycin, sitafloxacin, and HBCS monotherapy in a transtibial implant model. The results showed complete lack of vancomycin efficacy while all mice treated with HBCS had evidence of infection control, and some had evidence of osseous integrated septic implants, suggestive of biofilm eradication. Taken together these studies demonstrate that HBCS adjuvant with standard of care debridement and vancomycin therapy has the potential to eradicate MRSA osteomyelitis.

Fig. 1

Schematic illustration of murine 1-stage revision model for MRSA implant associated osteomyelitis with outcomes and antibiotic treatments. a Osteomyelitis is initiated via surgical implantation of a 6-hole femoral plate that is secured by a sterile screw in hole 3, and a contaminated screw containing ~10⁵ CFU of USA300LAC:Lux in hole 4. On day 7, all implants are removed with debridement and irrigation, a sterile plate is implanted secured by sterile screws in holes 1 and 6 with initiation of antibiotic therapy, and the mice are euthanized on day 21. b Timeline of the outcome measures. c Seven groups of mice (n = 4) were studied. “Baseline” mice were euthanized on day 7 as a control for the infection at the time of revision surgery. The 6 different treatment groups received the indicated i.p. dosing regimen. d Molecular structures of the two experimental drugs 1) hydroxybisphosphonate-conjugated sitafloxacin (HBCS, BV63072) and 2) bisphosphonate-conjugated sitafloxacin (BCS, BV600072), 3) parent sitafloxacin to assess effects of non-bone targeted antibiotic, and the bone targeting moieties 4) hydroxyphenylethanehydroxybisphosphonate (HPHBP, BV6301) and 5) hydroxyphenylethanebisphosphonate (HPBP, BV6001) to control for non-nitrogen containing bisphosphonate effects on bone

Fig. 2

Confirmation of implant-associated osteomyelitis and post-op body weight recovery. a–c The original screws and femoral plate were removed from the mice on day 7 post-implantation, and the sonicate from each was assessed for bacterial load via CFU assay. Data for each implant are presented with the mean ± SD. No differences between any groups were detected. The body weight for each mouse was determined weekly and reported as the percentage of pre-op body weight for the group (d mean over time), and for each mouse with mean ± SD for the group (e *P < 0.05 2-way ANOVA)

Fig. 3

Combination of bisphosphonate-conjugated sitafloxacin and vancomycin therapy eradicate planktonic MRSA following 1-stage revision surgery. Longitudinal in vivo BLI was performed on the indicated day following the original septic implant surgery on day 0, and representative BLI images are shown (a), with the longitudinal data presented as the percentage of BLI normalized to day 1 for each group (b, mean over time). Unfortunately, several mice were lost to anesthesia and the group sizes for statistical analysis at day 14 are indicated. c BLI data for each mouse on day 7 and 14 are presented with mean ± SD for the group (*P < 0.05 Wilcoxon matched-pairs signed rank test)

Fig. 4

Combination of bisphosphonate-conjugated sitafloxacin and vancomycin therapy prevents catastrophic fractures and peri-implant osteolysis following 1-stage revision surgery. Longitudinal X-rays were obtained on the indicated day following the original septic implant surgery on day 0, and micro-CT was performed on the femurs after harvest on day 21. a Representative in vivo X-rays are shown to illustrate the progression of peri-implant osteolysis with Boolean assessment of catastrophic fractures indicated by yellow arrows on day 21 (b; *P < 0.05 vs. Baseline via Fishers exact test). c The implants were removed for ex vivo volumetric micro-CT of the femurs on day 21, and representative 3D renderings are shown to illustrate the osteolysis area at each screw hole and catastrophic fractures (yellow arrows) observed in the groups that did not receive bisphosphonate-conjugated sitafloxacin. d–g The volume of each screw hole for each mouse are presented with mean ± SD for the group (*P < 0.05 1-way ANOVA)

Fig. 5

Combination of bisphosphonate-conjugated sitafloxacin and vancomycin therapy reduces peri-implant osteoclast numbers at the site of the screw holes following 1-stage revision surgery. The infected femurs (n = 3 or 4) described in Fig. 4 were processed for TRAP-stained histology, and semi-automated histomorphometry was performed to quantify the TRAP⁺ area proximal to the revision screw hole #1, and the original MRSA-infected screw hole #4. Representative x1 images (a–g) with 40x region of interest (h–n) are shown to illustrate the number of osteoclasts at the site of bone infection. The TRAP-stained tissue area proximal to screw hole #1 (o) and #4 (p) was quantified with ImageJ, and the data for each femur is presented with the mean ± SD for the group (*P < 0.05 1-way ANOVA)

Fig. 6

Combination bisphosphonate-conjugated sitafloxacin and vancomycin therapy reduces SACs and Gram⁺ biofilm following 1-stage revision surgery. The infected femurs (n = 3 or 4) described in Fig. 4 were processed for Brown and Brenn-stained histology, and semi-automated histomorphometry was performed to quantify the Gram⁺ area proximal to screw holes #1 and #4. Representative 1× images (a–g) with 5x region of interest (h–n) are shown to illustrate the SACs (yellow arrows) and bacterial biofilm on bone (red arrows) at the site of bone infection. The Gram-stained tissue area proximal to screw hole #1 (o) and #4 (p) was quantified with ImageJ, and the data for each femur is presented with the mean ± SD for the group (*P < 0.05 1-way ANOVA)

Fig. 7

Schematic of transtibial implant model of S. aureus implant-associated osteomyelitis with antibiotic therapy. L-shaped wires contaminated with ~10⁵ CFU of USA300LAC:Lux were transcortically implanted through the right tibia (a) of 30 mice, which remained untreated for 1 week to established chronic osteomyelitis. On day 7, the mice were randomized into three treatment groups (n = 10) : Group 1 Vancomycin (110 mg·kg⁻¹ per 24 h); Group 2 Sitafloxacin (2.5 mg·kg⁻¹ per 24 h); and Group 3 BV63072 (3 mg·kg⁻¹ per 24 h). The infected mice were treated for 14 days, euthanized via perfusion with 4% paraformaldehyde/2.5% glutaraldehyde on day 21 post-op, and the infected tibiae were processed for histology and TEM (b)

Fig. 8

Radiographic and histologic evidence of HBCS-mediated MRSA eradication with implant osseous integration. Representative X-rays, H&E and TRAP-stained histology of the MRSA infected tibiae treated with antibiotics described in Fig. 7 and the sterile pin control are shown to illustrate the radiographic (a–d) and histologic outcomes of H&E (e–h), TRAP (i–l) after 2 weeks of therapy. Note the pin dislocation in the X-rays, and the marrow necrosis with peri-implant osteolysis adjacent to the pinhole (*), observed in 1x images of H&E-stained section from most of the Vancomycin (e, i) and Sitafloxacin (f, j) treated tibiae (Table 1), which is consistent with septic implant loosening and chronic osteomyelitis in this model. In contrast, the implant appears to be fixed in the HBCS-treated tibia (c, g, k), there is also healthy bone marrow and new bone formation completely surrounding the flat wire-shaped pinhole (*), which is consistent with the eradication of the infection and osseous integration of the implant, that compared with sterile pin implant (d, h, l)

Fig. 9

Sitafloxacin kills MRSA within SACs with evidence of fibrin capsule disintegration. To assess antibiotic effects on intramedullary SAC eradication and effects on bacteria morphology, Brown and Brenn-stained histology with subsequent TEM was performed on the infected tibiae described in Fig. 7. Representative images from the Vancomycin (a–e), Sitafloxacin (f–j) and HBCS (k–o) treatment groups are shown. Gram-positive SACs identified at 1x (boxed regions in a, f and k) with corresponding higher power images (b, c, g, h, l and m), and x3 000 TEM (d, i and n) with an enlarged region of interest at x6 000 (e, j and o), are shown. Note the complete lack of vancomycin efficacy, as evidenced by the thick intact fibrin ring (blue arrow in c and red arrows in e) surrounding the TEM electron-dense bacteria within the SAC. In contrast, sitafloxacin treated SACs contain a necrotic core (*) and a degraded fibrin capsule. These bactericidal effects were more profound in the HBCS treatment Group, as the fibrin capsule was disintegrated, the necrotic core (#) extended all the way to the edge of the SAC, and very few TEM electron-dense bacteria remain (yellow arrow in o)

Fig. 10

Decreased incidence of MRSA colonization of OLCN in necrotic bone fragments in mice treated with HBCS versus vancomycin treatment control. To assess antibiotic effects on the incidence of MRSA colonization of OLCN and effects on bacteria morphology, histology was analyzed from the mice that received MRSA contaminated transtibial pins described in Fig. 7. Representative 1x images from the Vancomycin (a–d) and HBCS (e–h) treatment groups are shown. Gram positive necrotic bone fragments (boxed regions in (a, e) with corresponding higher power images in (b, f) highlighting the biofilm (red arrows) were identified in 8 out of 8 tibiae in the Vancomycin treatment group, but only 2 out of 10 tibiae in the HBCS treatment group (Table 1). TEM assessment of this biofilm at x5 000 confirmed MRSA colonization of the OLCN (c, g), and x15 000 high power images of the boxed regions (d, h) failed to detect any remarkable differences in bacteria morphology between groups. However, while the marrow necrosis (#) with peri-implant osteolysis as typically seen in this model is observed around the pinhole in the Vancomycin treated tibiae (* in a), the pinhole in the HBCS treated tibiae is surrounded by new bone (* in e), which may suggest infection control despite the infected necrotic bone fragment in the adjacent marrow space