BMPs in periprosthetic tissues around aseptically loosened total hip implants

Abstract

Background and purpose: Primary and dynamically maintained periprosthetic bone formation is essential for osseointegration of hip implants to host bone. Bone morphogenetic proteins (BMPs) play a role in osteoinductive bone formation. We hypothesized that there is an increased local synthesis of BMPs in the synovial membrane-like interface around aseptically loosened total hip replacement (THR) implants, as body attempts to generate or maintain implant fixation.

Patients and methods: We compared synovial membrane-like interface tissue from revised total hip replacements (rTHR, n = 9) to osteoarthritic control synovial membrane samples (OA, n = 11. Avidin-biotin-peroxidase complex staining and grading of BMP-2, BMP-4, BMP-6, and BMP-7 was done. Immunofluorescence staining was used to study BMP proteins produced by mesenchymal stromal/stem cells (MSCs) and osteoblasts.

Results and interpretation: All BMPs studied were present in the synovial lining or lining-like layer, fibroblast-like stromal cells, interstitial macrophage-like cells, and endothelial cells. In OA and rTHR samples, BMP-6 positivity in cells, inducible by the proinflammatory cytokines tumor necrosis factor-alpha and interleukin-1beta, predominated over expression of other BMPs. Macrophage-like cells positive for BMP-4, inducible in macrophages by stimulation with particles, were more frequent around loosened implants than in control OA samples, but apparently not enough to prevent loosening. MSCs contained BMP-2, BMP-4, BMP-6, and BMP-7, but this staining diminished during osteogenesis, suggesting that BMPs are produced by progenitor cells in particular, probably for storage in the bone matrix.

Figure 1.

Immunolocalization of BMP-2 (A, E), BMP-4 (B, F), BMP-6 (C, G), and BMP-7 (D, H) in consecutive sections of synovial membrane from the hip joint of one control osteoarthritis patient (A–D) and in consecutive sections of interface tissue surrounding aseptically loosened total hip replacement component from the hip joint of one revision-operated patient (E–H). The synovial lining is present at the right-hand side of osteoarthritis samples (A–D). A similar synovial lining-like layer present in revision total hip replacement samples (E–H) is marked with “s”. Some small arteries and veins are marked with an asterisk. No counterstaining; original magnification ×400. Scale bars (panels A and E only) represent 10 mm.

Figure 2.

Cell density scores for BMP-2, BMP-4, BMP-6, and BMP-7 in synovial membrane in osteoarthtritis (white bars) compared to synovial membrane-like lining membrane in revision total hip replacement (gray bars). Medians and 95% CIs for medians are presented. The difference in BMP-4 expression between osteoarthritis and aseptic loosening is marked with an asterisk (p = 0.04).

Figure 3.

Cell density scores for BMPs (BMP-6, BMP-2, BMP-4, and BMP-7) in synovial membrane in osteoarthritis (upper row, white bars) and in synovial membrane-like lining membrane in revision total hip replacement (lower row, gray bars). Medians and 95% CIs for medians are presented. Statistically significant differences (p < 0.05) between BMP-6 and other BMP types are marked with an asterisk, and statistically significant differences (p < 0.05) between BMP-4 and BMP-7 are marked with #.

Figure 4.

Immunofluorescence staining of undifferentiated bone marrow-derived human mesenchymal stromal/stem cells (A, C, E, G, and I) and after differentiation for 14 days in osteogenic medium (B, D, F, H, and J). Staining for BMP-2 (A, B), BMP-4 (C, D), BMP-6 (E, F), and BMP-7 (G, H). Negative control is staining with normal nonimmune goat IgG (I, J). Nuclear staining with DAPI is shown to demonstrate the presence of cells in the photographed sample fields.