Abstract Background This study aimed to delineate the cell heterogeneity in the bone-implant interface and investigate the fibroblast responses to implant-associated S. aureus infection. Methods Single-cell RNA sequencing of human periprosthetic tissues from patients with periprosthetic joint infection (PJI, n = 3) and patients with aseptic loosening (AL, n = 2) was performed. Cell type identities and gene expression profiles were analyzed to depict the single-cell landscape in the periprosthetic environment. In addition, 11 publicly available human scRNA-seq datasets were downloaded from GSE datasets and integrated with the in-house sequencing data to identify disease-specific fibroblast subtypes. Furthermore, fibroblast pseudotime trajectory analysis and Single-cell regulatory network inference and clustering (SCENIC) analysis were combined to identify transcription regulators responsible for fibroblast differentiation. Immunofluorescence was performed on the sequenced samples to validate the protein expression of the differentially expressed transcription regulators. Results Eight major cell types were identified in the human bone-implant interface by analyzing 36,466 cells. Meta-analysis of fibroblasts scRNA-seq data found fibroblasts in the bone-implant interface express a high level of CTHRC1. We also found fibroblasts could differentiate into pro-inflammatory and matrix-producing phenotypes, each primarily presented in the PJI and AL groups, respectively. Furthermore, NPAS2 and TFEC which are activated in PJI samples were suggested to induce pro-inflammatory polarization in fibroblasts, whereas HMX1, SOX5, SOX9, ZIC1, ETS2, and FOXO1 are matrix-producing regulators. Meanwhile, we conducted a CMap analysis and identified forskolin as a potential regulator for fibroblast differentiation toward matrix-producing phenotypes. Conclusions In this study, we discovered the existence of CTHRC1 + fibroblast in the bone-implant interface. Moreover, we revealed a bipolar mode of fibroblast differentiation and put forward the hypothesis that infection could modulate fibroblast toward a pro-inflammatory phenotype through NPAS2 and TFEC.
