Abstract Osseointegration for implants, especially bioinert implants, poses significant clinical challenges. Overcoming fibrotic encapsulation and promoting osseointegration at the implant interface are critical for successful bone repair, which highly expected biomaterials with osteoblast over fibroblast selectivity. However, few materials possess the function. β-amino acid polymers have demonstrated cell adhesion property, easy preparation, and robust stability to resist proteolysis as emerging biomaterials. Here, we develop amphiphilic β-amino acid polymers that demonstrate exceptional osteoblast vs fibroblast selectivity, outperforming the natural osteoblast-selective KRSR peptide. The optimal polymer selectively supports osteoblast adhesion by manipulating the adsorption of serum proteins and the presentation of RGD motifs on polymer-modified surfaces. In vivo study using polymer-modified titanium-implants in female rat maxillary bone reveals that the optimal polymer substantially promotes osseointegration of titanium-implants compared to uncoated titanium-implants, which tend to develop fibrous encapsulation. This study demonstrates the effectiveness of our strategy in designing osteoblast-selective biomaterials and implies the promising application of β-amino acid polymer as emerging osteoblast-selective biomaterials to promote osseointegration. Subject terms: Biomaterials - cells, Implants, Polymers __________________________________________________________________ Biomaterials exerting osteoblast over fibroblast selectivity are promising for overcoming fibrotic encapsulation and promoting osseointegration at the implant interface, but remain underdeveloped. Here, the authors report amphiphilic β-amino acid polymers that exhibit selectivity for osteoblasts over fibroblasts, outperforming the natural osteoblast-selective peptide. Introduction Osseointegration plays a vital role in bone regeneration and realizing the function of permanent implants, including orthopedic and dental implants^[44]1,[45]2. The process is primarily determined by the initial osteoblast adhesion to the implant surface, as these cells are responsible for synthesizing and depositing the extracellular matrix (ECM), crucial for bone remodeling and mineralization^[46]3. However, during the process of implant placement, the wound healing response is caused by the foreign body reaction, or when the local tissue healing microenvironment is disrupted due to individual patient differences, fibroblasts may be recruited to the area near the implant. This leads to the formation of a fibrous capsule, which inhibits the colonization of osteoblasts on the implant surface, thereby increasing the likelihood of osseointegration failure^[47]4,[48]5 (Fig. [49]1a). Moreover, compromised osseointegration and fibrous capsule formation can result in implant loosening and elevate the risk of bacterial biofilm-related infections at the implant-tissue interface, ultimately causing implant failure^[50]6. Fig. 1. High-throughput screening of β-amino acid polymers for osteoblast-selective adhesion. [51]Fig. 1 [52]Open in a new tab a Schematic illustration of the competition between osteoblasts and fibroblasts at the interface between implants and surrounding bones after dental implantation. b Thiol-terminated β-amino acid polymers composed of a cationic subunit (NM, MM, DM, or βHLys) and a hydrophobic subunit (CP, CH, or CO), x + y = 1, x = 0.4 − 1. c Osteoblasts (MC-3T3-E1) and fibroblasts (NIH-3T3) were cultured on polymer or peptide-modified surfaces for 24 h and then stained with Calcein-AM. Heatmap depicts the ratio of osteoblasts to fibroblasts for cell coverage area. d, e Representative fluorescent images of cell morphology and quantification of cell area for human primary osteoblasts (d) and human primary fibroblasts (e) on RGD, KRSR, and MM[50]CH[50]-modified surfaces, respectively, after cell seeding for 3, 12, and 24 h (n = 10). The red fluorescence represents the F-actin that was stained by rhodamine-phalloidin. Scale bar: 50 μm. Data were presented as mean ± SD Statistical analysis: one-way ANOVA with Tukey post-test. f, g Schematic diagrams (summarizing the findings in d, e) showing the distribution of F-actin for osteoblasts and fibroblasts. Currently, two strategies are widely employed to promote osteoblast adhesion and osseointegration: altering the implant surface topography and modifying the surface with bioactive molecules such as the ECM-derived cell adhesion peptide Lys-Arg-Ser-Arg (KRSR) and Arg-Gly-Asp (RGD)^[53]7–[54]13. However, these strategies fail to achieve the desired osteoblast-selective adhesion and optimal osseointegration. Most existing strategies promote osteoblast adhesion but lack selectivity between osteoblast and fibroblast. While the KRSR peptide promotes selective osteoblast adhesion, it suffers from the common shortcomings of natural peptides, such as susceptibility to proteolysis, difficulty in large-scale synthesis, and expensiveness^[55]14. Therefore, developing biomaterials and implant coatings that can selectively promote osteoblast adhesion over fibroblast is crucial for improved osseointegration. Osteoblast and fibroblast exhibit different adhesion behaviors due to their distinct cell types and functions^[56]15,[57]16, which implies the feasibility to develop osteoblast-selective bioactive molecule by adjusting surface chemistry composition and protein adsorption. β-amino acid polymers are promising for biomaterial applications because of that their backbone is inherently protein-mimetic, and that they overcome the shortcomings of ECM-derived peptides by offering excellent stability against proteolysis, easy large-scale synthesis, and low cost^[58]17–[59]20. In this work, we explore the selectivity between osteoblasts and fibroblasts using a library of amphiphilic β-amino acid polymers with varying compositions, tunable positive charges, and amphiphilicity (Fig. [60]1b). Notably, we observe tunable protein adsorption among these polymers and identified an optimal polymer, MM[50]CH[50], which demonstrate desired osteoblast-selective adhesion in vitro and potent in vivo osseointegration in a rat oral implant model, outperforming the gold standard of cell adhesion RGD peptide and the osteoblast-selective KRSR peptide (Fig. [61]1c). Compared to the general cell adhesion function, achieving selective osteoblast adhesion using polymeric materials is considered impossible before the achievement in this work. These advantages and the structural diversity of osteoblast-selective β-amino acid polymers make our design a promising and effective strategy for developing biomaterials that enhance the osseointegration of implants in dental or orthopedic surgery at the early stage of osseointegration and prevent aseptic loosening. Results High-throughput screening of β-amino acid polymers for osteoblast-selective adhesion We investigated the selectivity between osteoblasts and fibroblasts on a library of cationic and amphiphilic β-amino acid polymers. A total of 76 β-amino acid polymers were prepared in about 20 residues (degree of polymerization (DP) of ~20) by ring-opening polymerization on a mixture of two β-lactams, one bearing the cationic side chain group and the other bearing the hydrophobic side chain group. These polymers were composed of variable ratios of cationic and hydrophobic residues, with the cationic content ranging from 40% to 100%). The cationic residues were selected from NM (non-methyl), MM (monomethyl), DM (dimethyl), and βHLys (β-homo-lysine), while the hydrophobic residues were selected from CP (cyclopentyl), CH (cyclohexyl), and CO (cyclooctyl) (Fig. [62]1b). To quickly evaluate the selectivity of β-amino acid polymers for osteoblast vs. fibroblast adhesion, these thiol-terminated polymers were covalently modified to a glass surface via the maleimide-functionalized octaethylene glycol (OEG[8]) as an anti-fouling layer^[63]21. The universal cell-adhesion RGD peptide and osteoblast-specific KRSR peptide served as controls for non-selective and osteoblast-selective adhesion, respectively. Additionally, OEG[8] was used as a negative control for cell adhesion. The cell adhesion selectivity of these β-amino acid polymers was evaluated by measuring the surface coverage of osteoblasts and fibroblasts on each polymer-modified surface after 24 h of cell culture and calculating the selective adhesion ratio (Fig. [64]1c). Some polymers, such as DM[50]CO[50], can promote cell adhesion for both osteoblast and fibroblast but without cell-selectivity (Supplementary Fig. [65]1). Among these polymers, both MM[40]CH[60] and MM[50]CH[50] exhibited the highest selective adhesion ratio in the series polymer. However, based on more detailed observations, we found that fibroblasts showed better adhesion regarding cell spreading on the MM[40]CH[60] surface compared to the MM[50]CH[50] surface (Supplementary Fig. [66]2). Therefore, to achieve the best osteoblasts vs. fibroblasts selectivity, MM[50]CH[50] was chosen as the optimal polymer and further studied for its selectivity and potential to enhance osseointegration. (^1H NMR spectra and gel permeation chromatography characterization of MM:CH polymers in Supplementary Figs. [67]3–[68]10 and Table [69]S1, respectively). The morphology of osteoblasts and fibroblasts on the MM[50]CH[50] surface To assess the osteoblast-selectivity of MM[50]CH[50], this thiol-terminated polymer was modified to the OEG[8]-covered glass slide, and the resulting surface was characterized by X-ray photoelectron spectroscopy (XPS) and water contact angle (WCA). XPS characterization showed apparent C1s and N1s peaks following the amination process, and the subsequent change of C:N element ratio, indicating a successful surface amination and the further modification of OEG[8] and MM[50]CH[50] (Supplementary Fig. [70]11 and Table [71]S2). The WCA of the surface increased from 55° to 69° after the modification with MM[50]CH[50], further indicating successful modification of the glass surface (Supplementary Fig. [72]12). Human primary osteoblast and human primary fibroblast adhesion were evaluated for cell morphology on the RGD, KRSR, and MM[50]CH[50]-modified surfaces, staining F-actin in red and the nucleus in blue (Fig. [73]1d, e). On the RGD surface, both cell types exhibited normal spreading with well-defined actin stress fibers (Fig. [74]1d, e and Supplementary Fig. [75]13). For MM[50]CH[50] and KRSR-modified surfaces, osteoblast displayed polygonal morphology with multiple lamellipodia after 3 h of incubation. For cell culture over 12 h, osteoblasts on the MM[50]CH[50]-modified surface showed significant spreading and formed an organized network of the actin stress fibers (Fig. [76]1d). In contrast, fibroblasts on MM[50]CH[50] or KRSR-modified surfaces exhibited a round shape and poor adhesion after 3 h of culture. With the extension of cell culture time, fibroblasts showed minimal spreading with little stress fibers (Fig. [77]1e). These observations of cell spreading were described with diagrams that closely follow the contours of the cytoskeletal network to effectively visualize the individual fibers and fiber bundles (Fig. [78]1f, g). It’s worth mentioning that even on the non-selective RGD surface, obvious differences in cytoskeletal morphology were observed: fibroblasts displayed long and spindle-shaped extensions aligned with cell movement, while osteoblasts had short and multidirectional F-actin bundles. This suggests different spreading behaviors and adhesion mechanisms between the two cell types, possibly relating to the osteoblast-selectivity of the β-amino acid polymers. The difference of osteoblast vs. fibroblast adhesion To unveil the cell-selective adhesion mechanism and gain insight into the differences in adhesion behavior and intracellular events between osteoblasts and fibroblasts under a non-selective condition, RNA sequencing (RNA-seq) was performed on cells that were cultured on the RGD surface. The two types of cells were incubated for 24 h to allow adhesion maturation. Three replicates of osteoblast and fibroblast adhered on the RGD surface were sequenced, respectively. The analysis identified 8008 differentially expressed genes (DEGs) between fibroblasts and osteoblasts (adjust Q value ≤0.05), with 3854 upregulated DEGs and 4154 downregulated DEGs (Supplementary Fig. [79]14a). Then, gene ontology (GO) enrichment analysis was performed for cellular components categories to identify the key biology processes that are different between osteoblasts and fibroblasts. The results showed that DEGs are mainly ascribed to cellular components such as cytoplasm, nucleus, cytosol, and cytoskeleton (Fig. [80]2a and Supplementary Fig. [81]14b). Among these, the GO term related to the cytoskeleton was most correlated with cell adhesion. To investigate the biological functions of these 665 DEGs in the cytoskeleton, a Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway enrichment analysis was further performed. The top 15 pathways sorted by Q-values were presented, among which the “regulation of actin cytoskeleton pathway”, a dominant pathway related to cell adhesion, existed as the top 1 pathway, indicating a difference in cytoskeleton formation between the two types of cells when they adhered to the RGD surface (Fig. [82]2b). Fig. 2. RNA-seq analysis of osteoblasts vs. fibroblasts on the RGD-modified surface. [83]Fig. 2 [84]Open in a new tab a Venn diagram of four key gene ontology (GO) terms (cytoplasm, cytosol, cytoskeleton, and nucleus) from cellular components categories. b Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway enrichment analysis of the cytoskeleton in GO cellular components categories. The top 15 pathways were shown. The y-axis label represents the KEGG pathway, and the x-axis represents the rich ratio. The size of the dots represents the number of gene enrichment. c–g The level of gene expression by osteoblasts and fibroblasts with respect to: Cdc42 to Arp2/3 pathway (c), Rac to Arp2/3 pathway (d), ROCK downstream pathways toward the activation of myosin II (e), Integrin downstream pathways (f), and focal adhesion (g). Genes expressed to a higher level in osteoblasts and in fibroblasts are blue and red, respectively. h Working model for osteoblast and fibroblast adhesion on the RGD-modified surface. To evaluate the activation degree of “regulation of actin cytoskeleton pathway” in osteoblasts and fibroblasts, the regulation of actin cytoskeleton-related genes among each group were mapped by hierarchical clustering analysis (Supplementary Fig. [85]15). Several genes (e.g., Diaph3, Baiap2, Enah, Wasl, Arpc4, Arpc5, Arpc5l, Cyfip2, Brk1, and Tmsb4x) expressed to a higher level in osteoblasts than in fibroblasts, and all of them were involved in Cdc42 to Arp2/3 pathway and Rac to Arp2/3 pathway that are associated with the formation of filopodia and lamellipodia^[86]22, respectively (Fig. [87]2c, d). Gene-gene interaction networks also showed the relationship of these genes (Supplementary Fig. [88]16). Moreover, the formation of actomyosin stress fiber in osteoblasts was also promoted owning to the activation of myosin light chain (MLC) (Fig. [89]2e). For fibroblast, several genes (PtK2, Bcar1, Crk, Src, Vcl, Actn1, Actn4, Pikfyve, Pip5k1c, and Pip5k1a) expressed to a higher level than that in osteoblast, and all of them were located in the pathway related to the formation of focal adhesion and F-actin (Fig. [90]2f, g). Among them, some genes were selected as representative genes. We examined their protein level expression and found that the results were consistent with that of the gene analysis (Supplementary Fig. [91]17). Therefore, we speculated that more filopodia and lamellipodia were formed to promote osteoblast adhesion and spreading, while fibroblasts achieve mature adhesion maybe depending on more focal adhesion and stress fibers. Lamellipodia and filopodia often occur on the stage of cell nascent adhesion or migration, however, the formation of focal adhesion depends on integrin clustering^[92]23. We further examined the expression of Lamin A/C, a core nuclear lamina component responsive to mechanical cues^[93]24, and integrin β1, involved in cell adhesion and mechanotransduction, where its assembly indicates increased binding to RGD^[94]25. The result indicated that fibroblasts possess stronger intracellular force compared to osteoblasts, likely due to fibroblasts’ higher number of stress fibers^[95]26, and that they rely on more RGD binding sites to balance the contractile force of these fiber bundles (Supplementary Fig. [96]18 and Fig. [97]2h). Mechanism analysis of osteoblast-selected adhesion on the MM[50]CH[50] surface Ethylenediaminetetraacetic acid (EDTA), a divalent cation chelator, was also used to block integrin-RGD binding^[98]27,[99]28. In a serum-containing environment, the presence of EDTA substantially diminished cell spreading on both RGD and MM[50]CH[50] surfaces (Fig. [100]3a, b), which indicates that cell attachment onto the β-amino acid polymer surface relies on the integrin-binding RGD motif of surface absorbed cell-adhesion proteins. Hence, the type, amount, and conformation of these adsorbed proteins could regulate the density and position of cell adhesion ligands and influence their cell adhesion behaviors^[101]29,[102]30. To investigate whether the osteoblast-selective mechanism of MM[50]CH[50] is related to the exposure of the RGD motif, the β-amino acid polymer DM[50]CO[50], which lacks cell-selective function, was used for comparison. Fibronectin (FN) and vitronectin (VN), known as critical ECM proteins for mediating cell adhesion, were employed to study cell-adhesion protein adsorption on polymer-modified surfaces. The amount of adsorbed FN and VN on polymer surfaces was measured by immunofluorescence analysis, but indicated no significant difference between the surface of DM[50]CO[50] and MM[50]CH[50] (Fig. [103]3c). Fig. 3. Mechanism analysis of osteoblast-selective adhesion on MM[50]CH[50]-modified surface. [104]Fig. 3 [105]Open in a new tab a, b Representative fluorescent micrographs on cell adhesion (a) and quantitative analysis on cell area (b) of Calcein-AM stained osteoblasts (MC-3T3-E1) adhered to MM[50]CH[50] and RGD-modified surfaces in serum-containing media for 3 h with or without EDTA treatment (n = 6). Data were presented as mean ± SD. Statistical analysis: two-tailed t-test. c The amount of surface-adsorbed fibronectin and vitronectin on MM[50]CH[50] and DM[50]CO[50]-modified surfaces (n = 3). Data were presented as mean ± SD. Statistical analysis: two-tailed t-test. d Quantification of the available integrin-binding RGD motif by antibody HFN7.1, from III9-10 of fibronectin that was adsorbed onto MM[50]CH[50] and DM[50]CO[50]-modified surfaces (n = 3). Data were presented as mean ± SD. Statistical analysis: two-tailed t-test. e Schematic representation of fibroblast adhesion on DM[50]CO[50]-modified surfaces bearing high density of RGD motif and MM[50]CH[50]-modified surfaces bearing low density of RGD motif. f Quantification of BSA adsorption on MM[50]CH[50] and DM[50]CO[50]-modified surfaces (n = 3). Data were presented as mean ± SD. Statistical analysis: two-tailed t-test. g, h Representative fluorescent images of cell morphology (g) and quantified average cell area (h) of fibroblasts (NIH-3T3) and osteoblasts (MC-3T3-E1) on MM[50]CH[50] and RGD-modified surfaces in serum-containing or serum-free medium after 2 h of incubation (n = 9), Data were presented as mean ± SD. Statistical analysis: two-tailed t-test. i, j Representative fluorescent images of cell morphology (i) and quantified cell area (j) of fibroblasts (NIH-3T3) and osteoblasts (MC-3T3-E1) in a serum-free environment (as a control), supplemented with BSA (BSA medium), or plated on BSA-pretreated surfaces (BSA surface) after 12 h of cell culture (n = 3). Data were presented as mean ± SD. Statistical analysis: one-way ANOVA with Tukey post-test. k Schematic representation of osteoblast and fibroblast adhesion in BSA-containing environment, summering the potential adhesion mechanism of two types of cells when BSA blocks part of the exposed RGD motif. To explore the conformational differences of FN adsorbed onto different surfaces, the exposed amount of RGD motif of surface-adsorbed FN was measured using the monoclonal antibody HFN7.1. The result showed that the amount of RGD motif on the non-selective DM[50]CO[50] surface was higher than that on the osteoblast-selective MM[50]CH[50] surface, which is consistent with the above hypothesis that inadequate RGD motif on the MM[50]CH[50] surface is unable to sustain stress fiber formation, resulting in poor fibroblast adhesion (Fig. [106]3d, e). In addition to the impact of the cell adhesion proteins, the predominant serum protein in the medium, bovine serum albumin (BSA), might also affect the surface density of the RGD motif and even block the RGD motif, therefore affecting cell adhesion behavior. Surface adsorption analysis of BSA on the MM[50]CH[50] and DM[50]CO[50] surfaces showed that osteoblast-selective MM[50]CH[50] obviously promoted BSA adsorption compared to the non-selective DM[50]CO[50], suggesting that BSA adsorption plays a crucial role in the osteoblast-selective function of MM[50]CH[50] (Fig. [107]3f). It is noteworthy that MM[50]CH[50] was almost deprived of osteoblast-selective function when cells were cultured in a serum-free environment without BSA (Fig. [108]3g, h). To better understand the role of BSA in the osteoblast-selectivity of MM[50]CH[50], cells were cultured in a serum-free environment either supplemented with BSA or on BSA-pretreated surfaces. In both cases, fibroblast adhesion was largely inhibited on the MM[50]CH[50] surface, while osteoblast adhesion remained unaffected, indicating that the decrease of RGD motifs due to BSA adsorption contributed substantially to the osteoblast-selectivity of MM[50]CH[50] (Fig. [109]3i, j). Altogether, the above analysis indicates that fibroblasts and osteoblasts have different adhesion mechanisms. Fibroblast, having more stress fibers, exhibit stronger intracellular forces than osteoblasts, which leads to differing responses to the density of RGD motifs on a surface, i.e., fibroblast adhesion and spreading require a higher density of RGD motifs compared to osteoblasts. This point was echoed by the observation that expression of Lamin A/C and integrin β1 in fibroblast was significantly lower on the MM[50]CH[50] surface than on the RGD surface (Supplementary Fig. [110]18). The fact that the osteoblast-selective MM[50]CH[50] surface manipulates adsorbed serum protein, including cell-adhesion proteins (such as FN) and BSA, to present fewer exposed RGD motifs than the non-selective DM[50]CO[50] surface, enables MM[50]CH[50] favorable osteoblast adhesion but not fibroblast adhesion. So, both cell-adhesion proteins and BSA significantly contribute to the osteoblast-selective adhesion of MM[50]CH[50] (Fig. [111]3k). RNA-seq analysis of osteoblast and fibroblast on the MM[50]CH[50] surface To further investigate the mechanism behind the osteoblast-selective adhesion of MM[50]CH[50], RNA-seq was employed to explore the differences in cellular adhesion between osteoblasts and fibroblasts on the MM[50]CH[50] surface by analyzing broad changes in the cellular transcriptome. We identified 9064 DEGs between fibroblasts and osteoblasts (adjust Q value ≤0.05) on the MM[50]CH[50] surface, including 4413 upregulated and 4651 downregulated DEGs (Supplementary Fig. [112]19a). DEGs related to cytoskeleton via cellular components categories of GO enrichment analysis were obtained and analyzed by KEGG pathway enrichment analysis (Fig. [113]4a and Supplementary Fig. [114]19b). We found that the “regulation of actin cytoskeleton” and “focal adhesion” pathway, relevant to cell adhesion behavior, were both presented in top 15 pathways. According to the two pathways, several genes expressed to a higher level in fibroblast than in osteoblast were used to perform hierarchical clustering analysis (Fig. [115]4b). In line with our prior results of RNA-seq preformed on the RGD surface, most of these genes were involved in stress fiber formation and focal adhesion. For fibroblast, while Crk, Dock, Rac3, and Iqgap1 were significantly upregulated, Vcl, Actn1, Actn4, and Actg1 were downregulated on the MM[50]CH[50] surface relative to those on the RGD surface (Fig. [116]4c, d). These genes were examined on their protein level expression and we found that the results were consistent with the gene analysis (Supplementary Fig. [117]20). Fig. 4. RNA-seq analysis of osteoblast and fibroblast on MM[50]CH[50] surface. [118]Fig. 4 [119]Open in a new tab a KEGG pathway enrichment analysis of the cytoskeleton in GO cellular components categories. The y-axis label represents the KEGG pathway, and the x-axis represents the rich ratio. The size of the dots represents the number of gene enrichment. b Heatmap of selected genes related to the “regulation of actin cytoskeleton” and “focal adhesion” pathways in each cell type. c, d The level of gene expression by osteoblasts and fibroblasts with respect to: integrin downstream pathways toward the activation of adherens junction (c), ROCK pathway (d). Genes expressed to a higher level in osteoblasts and in fibroblasts are blue and red, respectively. e Bright field microscopy images of fibroblast adhered to MM[50]CH[50] and RGD-modified surfaces after cell seeding for 24 h. Scale bar: 500 μm. f Heatmap of selected genes related to the formation of filopodia and lamellipodia in each cell type. g Volcano plot obtained from RNA-seq analysis of osteoblasts vs. fibroblasts on MM[50]CH[50]-modified surface. According to the analysis for pathways relevant to these genes, we found fibroblasts tend to form stress fibers and focal adhesion when they are cultured on the RGD surface. In contrast, cell–cell interaction is a dominant pathway when they are cultured on the MM[50]CH[50] surface. The “adherens junction”, “tight junction”, and “gap junction” pathways, all related to cell–cell interaction, were among the top 20 pathways in the KEGG enrichment analysis. This RNA-seq analysis results were consistent with bright field microscopy observations, which showed that fibroblasts on the MM[50]CH[50] surface tended to cluster after 24 h, unlike the well-dispersed morphology on the RGD surface (Fig. [120]4e). This result echoed the earlier conclusion that the density of exposed RGD motifs on the MM[50]CH[50] surface is not enough to enable fibroblast adhesion, leading to cell–cell interaction and eventually fibroblast detachment from the surface. Further, several genes related to filopodia and lamellipodia formation, which are more highly expressed in osteoblast than in fibroblast, were analyzed. The results from hierarchical clustering analysis and volcano plot showed that genes relevant to drive filopodia and lamellipodia formation (e.g., Wasl, Diaph3, Enah, and Baiap2) in osteoblast were upregulated on the MM[50]CH[50] surface relative to those on the RGD surface (Fig. [121]4f, g). To assess whether lamellipodia and filopodia promote osteoblast adhesion, small-molecule inhibitors targeting actin nucleators, including NSC23766 (Rac inhibitor)^[122]31, CK666 (Arp2/3 inhibitor)^[123]32, and SMIFH2 (Formin inhibitor)^[124]33, were used to treat human primary osteoblasts (Fig. [125]5a). In agreement with the RNA-seq analysis, inhibition of either lamellipodia or filopodia formation causes a decrease of the cell spreading area on the MM[50]CH[50] surface (Fig. [126]5b), indicating that the pseudopodia-based adhesion may be a dominant pathway to promote osteoblast adhesion and spreading under a relatively low density of exposed RGD motifs on surface-adsorbed serum proteins. Fig. 5. Cell adhesion behavior analysis on human primary osteoblasts and human primary fibroblasts using small-molecule inhibitors. [127]Fig. 5 [128]Open in a new tab a Representative fluorescent images of human primary osteoblasts were captured after 12 h of incubation with the small-molecule inhibitors CK666, NSC23766, and SMIFH2, respectively. F-actin was stained by rhodamine-phalloidin (red) and nucleus were stained by DAPI (blue). Non-treated cells were used as the control. b Cell area quantification on cells adhered to MM[50]CH[50] and RGD-modified surfaces after treatment with inhibitors (n = 10). c–f Representative fluorescent images of human primary osteoblasts (c) and human primary fibroblasts (e) were captured after 12 h of incubation with the small-molecule inhibitors Blebbistatin, Calyculin A, and Y-27632, respectively. F-actin were stained by rhodamine-phalloidin (red) and nucleus was stained by DAPI (blue). Non-treated cells were used as the control. Quantification of cell area for osteoblasts (d) and fibroblasts (f) adhered to MM[50]CH[50] and RGD-modified surfaces after treatment with inhibitors (n = 10). Data were presented as mean ± SD. Statistical analysis: one-way ANOVA with Tukey post-test. In order to a comprehensive understanding of osteoblast-selectivity mechanisms, we continued to explore the impact of intracellular force and stress fibers on cell adhesion and spreading using small-molecule inhibitors. These inhibitors were utilized to modulate the actomyosin-based contractility in human primary osteoblasts and human primary fibroblasts, including Blebbistatin (a myosin II inhibitor)^[129]34, Calyculin A (a phosphatase inhibitor that promotes myosin contraction)^[130]35, and Y-27632 (a ROCK inhibitor that prevents stress fiber formation)^[131]36. The results showed that Blebbistatin treatment did not obviously change the spreading area of the osteoblasts (Fig. [132]5c, d), but it significantly increased the spreading area of fibroblasts on the MM[50]CH[50] surface, supporting the hypothesis that reduced intracellular force in fibroblast will promote their adhesion on the polymer surface with fewer exposed RGD motifs (Fig. [133]5e, f). Calyculin A treatment, which increases myosin II activity and intracellular force, resulted in obvious decreases in the spreading area for both osteoblast and fibroblast on MM[50]CH[50] surface (Fig. [134]5c–f), which echoed the aforementioned conclusion that cell spreading is impaired due to insufficient exposed RGD motif of adsorbed proteins on the MM[50]CH[50] surface to bind integrins and counterbalance the increase of intracellular force. A further study utilizing the ROCK inhibitor Y-27632 to block stress fiber-based adhesion on the MM[50]CH[50] surface showed little effect on osteoblast adhesion, but a significant increase in fibroblast spreading area (Fig. [135]5c–f). This suggests that fibroblast adhesion requires fewer exposed RGD motifs when the contractive force of stress fiber bundles is reduced by inhibitors. In normal conditions, fibroblast adhesion requires a relatively high density of exposed RGD motifs on surface-adsorbed proteins. These results echoed the aforementioned observation that osteoblasts are less sensitive to and require a lower density of exposed RGD motifs on surface-adsorbed serum proteins compared to fibroblasts. This explains how β-amino acid polymer MM[50]CH[50] achieves osteoblast-selectivity by manipulating surface-adsorbed serum proteins and their exposed RGD motifs. In vitro evaluation of osteoblast-selective MM[50]CH[50] To evaluate the function of osteoblast-selective β-amino acid polymer MM[50]CH[50], we studied the growth and migration behaviors of osteoblasts on the MM[50]CH[50] surface. Cell proliferation study showed that MM[50]CH[50] promotes osteoblast proliferation, comparable to the RGD peptide and surpassing the KRSR peptide after culture for 5 days (Fig. [136]6a, b). An assay of scratch-induced migration with 24 h showed that MM[50]CH[50] promotes osteoblast migration as effectively as the RGD peptide and significantly better than the KRSR peptide (Fig. [137]6c, d). A further co-culture study was performed to observe the competitive growth behavior and evaluate the selectivity of MM[50]CH[50] for osteoblasts over fibroblasts. Osteoblasts and fibroblasts were prelabeled with cell-tracker dyes (Green CMFDA for osteoblast and orange CMTMR for fibroblast) and seeded at a 1:1 ratio on MM[50]CH[50], RGD, and KRSR surfaces. After 6 h, both MM[50]CH[50] and KRSR significantly promoted osteoblast adhesion and spreading, but not fibroblast adhesion; in contrast, the RGD surface promoted adhesion for both cell types (Fig. [138]6e, f). After 12 h, the RGD peptide facilitated excellent adhesion for both osteoblasts and fibroblasts, lacking osteoblast-selectivity. MM[50]CH[50] maintained distinct selectivity for osteoblast adhesion, outperforming the osteoblast-specific KRSR peptide (Fig. [139]6e, g). This outstanding osteoblast vs. fibroblast selectivity of MM[50]CH[50] implied its significant potential for application in the osseointegration of implants. Fig. 6. In vitro evaluation of osteoblast-selective MM[50]CH[50]. [140]Fig. 6 [141]Open in a new tab a Representative fluorescent images of Calcein-AM stained osteoblasts that were seeded on MM[50]CH[50], RGD, KRSR, and OEG[8]-modified surfaces after culture for 1, 3, and 5 days. b Quantification of cell density by Alamar Blue assay. c Scratch-induced osteoblast migration assay performed on MM[50]CH[50], RGD, and KRSR-modified surfaces. Red dashed lines indicate the gap at different points in time (0, 8, 16, and 24 h) (n = 3). Data were presented as mean ± SD. Statistical analysis: one-way ANOVA with Tukey post-test. d The coverage of the scratch area in the scratch-induced osteoblast migration assay (n = 6). Data were presented as mean ± SD. Statistical analysis: two-tailed t-test. Statistical analysis: one-way ANOVA with Tukey post-test. e Competitive adhesion of osteoblast and fibroblast under co-culture conditions on different surfaces after cell seeding over 6 and 12 h. Osteoblasts and fibroblasts were prelabeled with CMFDA (green) and CMTMR (red), respectively. f, g Cell area of osteoblast and fibroblast after co-culture for 6 h (f) and 12 h (g) (n = 4). Data were presented as mean ± SD. Statistical analysis: two-tailed t-test. In vivo evaluation of osteoblast-selective MM[50]CH[50] Since dental implants are located in the oral cavity, the early osseointegration of the implants and the prevention of aseptic loosening are of great importance. Therefore, we have chosen an SD rat oral implant model as a proof-of-concept demonstration to further explore the potential application of the osteoblast-selective β-amino acid polymer MM[50]CH[50]. Titanium (Ti) has been widely used for dental implants, owning to its excellent biocompatibility and corrosion resistance^[142]37–[143]39. However, a challenging and unresolved issue in dental implants lies in the poor osseointegration between bioinert Ti implants and host bone tissue, due to the formation of fibroblast-resulted fibrous encapsulation at the interface^[144]38,[145]40,[146]41. In this study, MM[50]CH[50]-modified Ti-screws were inserted into rat maxillary bones to validate whether MM[50]CH[50] could enhance osseointegration of the implant, using RGD, KRSR, and DM[50]CO[50]-modified Ti-screws and uncoated Ti-screws, for comparison (Fig. [147]7a and Supplementary Fig. [148]21). Before this, human primary osteoblasts and fibroblasts were seeded onto the modified Ti surfaces, and the results were consistent with experiments conducted on glass using cell lines (Supplementary Fig. [149]22). Additionally, we proved the existence of peptides and polymers pre- and post-implantations by observing no difference on the fluorescence intensity of Ti surface-modified peptides and polymers tethered with Rhodamine-B, and by characterizing the fluorine signal of the fluorinated polymers through XPS (Supplementary Figs. [150]23, [151]24). Fig. 7. In vivo evaluation of osteoblast-selective MM[50]CH[50] to promote osseointegration. [152]Fig. 7 [153]Open in a new tab a Schematic representation and experimental protocol of the rat oral Ti-implant model: maxillary after the incision, flap elevation, creation of the defect (I), and implantation of Ti-screw in the defect region (II and III). b Histological morphologies of methylene blue/acid fuchsin staining showing the bone (red) forming around the Ti implants (black) in five groups after 8 weeks of implantation. c Quantitative analysis of histomorphometrical analyses of BIC%. d Micro-CT images of RGD, KRSR, DM[50]CO[50], or MM[50]CH[50]-modified Ti-screw and the uncoated Ti-screw, respectively. e–i Quantitative analysis of bone volume/tissue volume (BV/TV, %) (e), trabecular thickness (Tb.Th, mm) (f), trabecular number (Tb.N) (g), bone surface/bone volume (BS/BV) (h), and trabecular separation (Tb.Sp) (i). Data were mean ± SD of biological replicates (n = 6 rats per group for micro-CT and four rats were randomly selected for histological morphology analysis); Statistical analysis: one-way ANOVA with Tukey post-test. Histological staining and microcomputed tomography (micro-CT) analysis were performed to evaluate the new bone formation and osseointegration at the bone-implant interfaces. In the methylene blue/acid fuchsin images, the bones and implants are indicated in red and black, respectively. At 2 weeks after surgery, no obvious necrosis or inflammation occurred in any group, but the produced new bone matrix was only in some small area adjacent to the bone-implant interface. Bone-to-implant contact (BIC%), as the percentage of the bone directly contacting the implant surface, was used to evaluate the degree of osseointegration. The results showed that the BIC% of postoperative week 2 was not significantly different among all groups, but an increase in bone volume/total volume (BV/TV) of MM[50]CH[50]-modified screws than the uncoated Ti-screws was observed (Supplementary Figs. [154]25, [155]26). Although the histological features of postoperative week 4 were similar to those of week 2 in all groups, bone formation progressed obviously around the MM[50]CH[50]-modified implant compared to another group (Supplementary Figs. [156]27,[157]28). Eight weeks after implantation, the interface of MM[50]CH[50]-modified Ti-screws showed direct contact and excellent integration with the surrounding bone tissues (having the BIC% of 85.5 ± 7.1%); whereas, the fibrous connective tissue is interposed between the new bone and the interface of uncoated Ti-screws (having the BIC% of 32.8 ± 10.5%), which impeded osseointegration. The BIC% of DM[50]CO[50], RGD, and KRSR-modified Ti-screws were 60.5 ± 7.8%, 53.8 ± 10.5%, and 42.3 ± 7.0%, respectively, which fall between MM[50]CH[50]-modified Ti-screws and uncoated groups (Fig. [158]7b). In other studies, the BIC% of the optimal group is usually between 50 and 70%^[159]42,[160]43. 3D Micro-CT images reflected the same result as histological staining analysis (Fig. [161]7c). The bone volume/total volume (BV/TV) values, trabecular number (Tb.N), and trabecular thickness (Tb.Th) around the MM[50]CH[50]-modified Ti implants were generally higher than that of uncoated Ti-screws; whereas, bone surface/bone volume (BS/BV) and trabecular separation (Tb.Sp) around the MM[50]CH[50]-modified Ti implants was lower than that of uncoated Ti-screws (Fig. [162]7d–g). These observations indicated that MM[50]CH[50]-modified Ti implants benefited from the osteoblast vs. fibroblast selectivity of MM[50]CH[50] to have favorable integration with surrounding bones in vivo, which is consistent with the in vitro performance of osteoblast-selectivity. Discussion The poor osseointegration of bioinert implants with surrounding bone tissues often leads to implant loosening and even implant failure^[163]37. Ensuring direct contact between the implant interface and surrounding bones is crucial for effective osseointegration and bone regeneration^[164]44. To achieve this, it is highly expected to develop bioactive materials that selectively promote osteoblast adhesion while inhibiting fibroblast adhesion at the implant interface, because fibroblasts tend to adhere excessively to the implant interface, hindering osteoblast colonization^[165]45. Moreover, fibroblasts may also activate local inflammation and impair osteoblast function by secreting certain cytokines^[166]46. In this study, we designed a series of β-amino acid polymers and demonstrated that the optimal polymer, MM[50]CH[50], exhibits distinct selectivity for osteoblasts over fibroblasts. This finding is dramatically different from our previous work, where β-amino acid polymer was designed to mimic the RGD peptide to obtain general cell adhesion without exploring selective cell adhesion to promote osteoblast vs. fibroblast cell adhesion like in this study. Compared to the general cell adhesion function, achieving selective osteoblast adhesion using polymeric materials is way more difficult and considered impossible before the achievement in this work. The osteoblast-selectivity of MM[50]CH[50] surpasses that of both RGD peptides and the osteoblast-specific KRSR peptide. To elucidate the mechanism behind the selective osteoblast adhesion of MM[50]CH[50], we conducted research from the perspectives of genes, proteins, and cellular behavior. We found that osteoblasts require fewer adhesion sites (such as RGD) compared to fibroblasts on the polymer-modified surface, due to the activation of the Arp2/3 pathway which promotes the formation of more lamellipodia. Additionally, we found that the density of RGD motifs on MM[50]CH[50] surface is lower compared to the non-selectivity surface (DM[50]CO[50]), due to the conformation effect of cell adhesion protein (e.g., FN) and blocking of non-adhesion protein (e.g., BSA). Our study advances the understanding of the adhesion behaviors of osteoblasts and fibroblasts, highlighting the crucial role of protein regulation and cell adhesion site density in promoting selective cell adhesion. For an in vivo study, a rat oral implant model was used as a proof-of-concept demonstration. We find that MM[50]CH[50]-modified Ti implants have favorable osseointegration with surrounding bones in vivo by avoiding the interfaces fibrous capsule, compared to the cell adhesion gold standard RGD peptide and osteoblast-selective KRSR peptide. In addition, β-amino acid polymers are designed to address the innate limitation of natural ECM peptides to have easy and scalable synthesis via polymerization, cheap price, resistance to proteolysis, and stable function in vivo. These advantages over cell adhesive ECM peptides and functional validation of osteoblast-selectivity altogether imply the potent application of β-amino acid polymers as emerging osteoblast-selective biomaterials to promote osseointegration of implants. However, the anatomical differences between rats and humans, such as the smaller and thinner alveolar bone in rats, are inherent limitations. Therefore, it’s worth investigation and validation of our findings in larger animal models and even in human clinical trials in future studies. Methods All animal research was performed in accordance with the Guidelines for Care and Use of Laboratory Animals of East China University of Science and Technology. All procedures in this study were approved by the Animal Research Bioethics Committee, East China University of Science and Technology (ECUST-2023-031). Rats were raised at 20–26 °C and 40–70% humidity, with a dark–light cycle of 12 h. Synthesis of β-amino acid polymers All monomers and polymers were prepared using previously reported methods^[167]19,[168]47–[169]49. Briefly, taking MM[50]CH[50] as an example, the β-lactam MM (42.9 mg, 0.2 mmol), the cyclohexyl substituted β-lactam CH (25 mg, 0.2 mmol) and the co-initiator 2-(tritylthio) acetic acid succinimidoester (8.6 mg, 0.02 mmol) were dissolved in anhydrous tetrahydrofuran (THF, 2 mL), followed by addition of the lithium hexamethyldisilazide (LiHMDs, 8.4 mg, 0.05 mmol) in THF. The reaction was stirred at room temperature for 10 h, and then two drops of methanol were added to the reaction to terminate the polymerization. Cold petroleum ether (45 mL) was added to the mixture to precipitate out the obtained β-amino acid polymer. The polymer was collected by centrifugation (at 2000 × g for 3 min) and then subjected to two more cycles of dissolution/precipitation using THF and petroleum ether. The collected polymer was dried under vacuum to give the NHBoc-protected MM[50]CH[50] polymer. Deprotection of the polymer was achieved by treating the polymer under gentle shaking for 2 h in trifluoroacetic acid (3 mL) supplemented with triethylsilane (0.15 mL). Then, the resulting solution was concentrated by blowing N[2]and the resulting oil was dispersed in methanol (1 mL). Cold methyl tert-butyl ether (40 mL) was added to the polymer solution in methanol to precipitate out a white solid. The obtained solid was subjected to the dissolution/precipitation cycle three times using methanol and methyl tert-butyl ether to give the fully deprotected MM[50]CH[50] polymer as a white powder, then characterized with GPC using dimethylformamide as the mobile phase at a flow rate of 1 mL/min. The polymer was dissolved in Milli-Q water (6 mL) and filtered through a 0.45 μm polyether sulfone membrane, followed by lyophilization to give a white floccule. Other poly-β-peptides were synthesized and purified using the same procedure. Preparation of ECM peptide and polymer-immobilized surfaces Materials surfaces modified with peptides and polymers were prepared using our recently developed method^[170]21. Briefly, glass coverslips (25 × 76 mm) or medically used titanium nails (length 3 mm, diameter 0.8 mm) were irradiated by a UV-ozone for 25 min, then immersed in anhydrous toluene solution containing 2% (v/v) (3-aminopropyl) triethoxysilane (APTES) overnight. After washing alternately with ethanol and Milli-Q water, APTES-modified glass slides were annealed in vacuum condition at 80 °C for 2 h. The APTES-modified surfaces were covered with 50-well or 8-well coverslips (103,380 or 103,350, Sigma-Aldrich), followed by incubation with MAL-OEG[8]-NHS solution (Biomatrik Inc, 10 mM in PBS) for 2 h. Then, the OEG[8]-modified surfaces were incubated with RGDSPC (Synpeptide, China), KRSRGYC (Synpeptide, China), or β-amino acid polymer solution (0.6 mM in PBS, pH 7.4) overnight to prepare the peptide or polymer-modified surfaces. A solution of thioglycerol (Adamas-beta®, 100 mM in PBS) was added to the surface and incubated for 40 min to block unreacted maleimide groups. Finally, the slides were rinsed with ethanol and Milli-Q water and dried with N[2] for further use. Characterization of material surface modification The water contact angles (WCA) were used to evaluate surface hydrophilicity using a contact angle meter (JC2000D2, Shanghai Zhongchen Digital Technology Co.Ltd, China). X-ray photoelectron spectroscopy (XPS) of modified surfaces were detected on a Thermo Scientific K-Alpha spectrometer (Thermo Fisher) that is equipped with a monochromatic Al Kα (1486.6 eV) X-ray source operated at 12 kV and a pressure of 8 × 10^−10 Pa^[171]50. Cell culture MC-3T3-E1 osteoblasts (ATCC) were cultured in α modification of Eagle’s Minimum Essential Medium (α-MEM) medium supplemented with 10% fetal bovine serum (FBS) (Gibco), 1% (v/v) Penicillin/Streptomycin (P/S), and 2 mM l-glutamine (Thermo Fisher) at 37 °C in a 5% CO[2] environment. NIH-3T3 fibroblasts (ATCC) were cultured in Dulbecco’s modified Eagle’s medium (DMEM) (Gibco) in the same condition. Human primary osteoblasts (iCell, s022) were cultured in primary osteoblast medium (PriMed-iCell-019) supplemented with 5% (v/v) fetal bovine serum, 1% (v/v) osteoblast growth supplement, and 1% (v/v) P/S. Human primary gingival fibroblasts (iCell, m005) cultured in primary fibroblast medium (PriMed-iCell-003) supplemented with 5% (v/v) fetal bovine serum, 1% (v/v) fibroblast growth supplement, and 1% (v/v) P/S. Cells were seeded on a series of β-amino acid polymers, RGD, KRSR, or OEG[8]-modified surfaces at a density of 1.4 × 10^4 cells/cm^2. Glass slides were placed in a petri dish and incubated for 2 h at 37 °C for initial cell attachment. Then fresh medium was added into the dish to immerse the entire slide for continuous cell culture. Immunofluorescent analysis of cell attachment and morphology Cells were thoroughly washed with PBS between all steps unless stated otherwise. 4% paraformaldehyde was added and incubated for 20 min to fix cells. Subsequently, cells were permeabilized with 0.4% triton X-100 in PBS for 5 min, blocked with 5% BSA in PBS for 30 min. For the cell morphology staining, cells were incubated with Alexa Fluor-555-conjugated anti-vinculin antibody (Yeasen Biotech, diluted 1:300 in PBS) overnight at 4 °C, then incubated with FITC-phalloidin (Yeasen Biotech, diluted 1:200 in PBS) for 2 h and 4′-6-diamidino-2-phenylindole (DAPI) (Yeasen Biotech, 5 μg/mL) for 10 min in sequence. Fluorescent images were collected on a confocal microscope (Leica TCS SP8) equipped with Leica Application Suite X software. Microscopy images for each experimental group were obtained from three to five independent experiments, with photographs taken at three to ten randomly selected locations. The fluorescence intensity of adherent cells was measured by a Typhoon TRIO variable mode imager with Typhoon Scanner Control software. To quantify protein expressed by cells, cells were incubated overnight with the primary antibodies against Crk p38, Dock, Rac3, Iqgap1, Vcl, Actn1/2/3/4, Arpc4, Arpc5l, Tmsb4x, and Actg1 (Affinity Bioscience, diluted 1:200 in PBS), then washed with PBS thrice and incubated with FITC-conjugated goat anti-rabbit IgG secondary antibody (diluted 1:400 dilution). The fluorescence intensity of protein secreted by cells was measured using a ChemiDoc MP Imaging system from Bio-Rad and quantitatively analyzed using ImageJ software. RNA-seq and data analysis MC-3T3-E1 and NIH-3T3 cells were cultured on MM[50]CH[50], RGD, and KRSR-modified surfaces. After 24 h incubation, total RNA was isolated using Trizol Reagent (Servicebio). RNA samples were quantified using a NanoDrop and Agilent 2100 bioanalyzer (Thermo Fisher Scientific, MA, USA). The RNA sequencing libraries were constructed and sequenced on the MGISEQ2000 platforms. Protein adsorption assay The protein adsorption assay was performed on MM[50]CH[50] and DM[50]CO[50]-modified surfaces, after incubating these surfaces for 2 h with a fibronectin solution (Sigma, 0.1 µg/mL, 10 μL/well), a vitronectin solution (Sigma, 0.1 µg/mL, 10 μL/well), or a collagen solution (Sigma, 0.1 µg/mL, 10 μL/well). Then the slides were incubated with anti-FN antibody (Abcam, 1:200 dilution), anti-VN (Abcam, 1:100 dilution), respectively. FITC-conjugated goat anti-rabbit IgG secondary antibody (Abcam, 1:250 dilution) was used to incubate these slides for 2 h. The fluorescence values of the secondary antibody were measured by a Typhoon TRIO variable mode imager with Typhoon Scanner Control software. For the bovine serum albumin (BSA) adsorption assay, these surfaces were incubated with a FITC-conjugated BSA solution (Solarbio, 0.01 mg/mL, 10 μL/well), and then measured the fluorescence values by a Typhoon TRIO variable mode imager with Typhoon Scanner Control software. Conformation study for surface-adsorbed FN The conformation of FN adsorbed on polymer-modified surfaces was studied by immunofluorescence staining. The slides were incubated with 0.05 µg/mL FN solution at 37 °C for 2 h. After being rinsed in PBS and blocked in 3% BSA at room temperature for 0.5 h, the FN-adsorbed slides were incubated with HFN7.1 monoclonal antibody (Novus Biologicals, 1:300 dilution) at 4 °C overnight. All slides were washed with PBS and incubated with FITC-conjugated goat anti-mouse IgG secondary antibody (Abcam, 1:250 dilution) for 2 h sequentially. The fluorescence values were obtained using a Typhoon TRIO variable mode imager. Cell adhesion in a serum-free environment In order to examine the impact of serum album, cells were centrifuged to remove the supernatant and resuspended in a serum-free culture medium. This step was repeated three times to remove proteins in the medium. For BSA preincubating assay, MM[50]CH[50] and RGD-immobilized surfaces were first incubated with BSA solution (Thermo Fisher, 30 mg/mL in PBS) at 37 °C for 2 h, followed by washing with PBS three times. Then, cells were seeded on these surfaces at a density of 1.4 × 10^4 cells/cm^2, followed by immersion in serum-free medium and incubation for 12 h. For a serum-free medium containing BSA, cells were seeded and immersed in a serum-free medium containing 0.9 mg/mL BSA for 12 h. Cell adhesion study under inhibitor treatment For the inhibition assay of EDTA^[172]27, cells were treated with 5 mM EDTA (leyan.com) at 37 °C for 10 min before seeding in serum-free or serum-containing medium. Cells were incubated for 3 h, and then stained with LIVE/DEAD Viability/Cytotoxicity Kits (Thermo Fisher) at room temperature for 15 min^[173]51. The fluorescent images were collected on a confocal microscope (Leica TCS SP8). The collected images were analysed using the ImageJ software. For small-molecule inhibitor experiments^[174]34, cells were plated on RGD or MM[50]CH[50]-modified surfaces at a density of 1.4 × 10^4 cells/cm^2 in the presence of one of the following reagents: blebbistatin (HY-13441, MedChemExpress, 20 μM), Calyculin A (HY-18983, MedChemExpress, 0.1 nM), Y-27632 (YH-10071, MedChemExpress, 20 μM), NSC23766 (HY-15723A, MedChemExpress, 50 μM), CK666 (HY-16926, MedChemExpress, 200 μM), or SMIFH2 (HY-16931, MedChemExpress, 10 µM). Inhibitors remained in the medium for the time window of the experiment for 12 h. Cell proliferation assay In cell proliferation experiments, 4000 cells/cm^2 were seeded on peptide and polymer-modified surfaces for 1, 3, and 5 days. Cells adhered to slides were stained by LIVE/DEAD Viability/Cytotoxicity Kits and imaged by a fluorescence microscope. After the addition of Alamar Blue (Thermo Fisher, 10%, v/v in medium) and incubation at 37 °C for 3 h, the solution in each well was transferred to a 384-well plate, and the fluorescence intensity was quantified using a microplate reader (λ[ex] = 560 nm; λ[em] = 590 nm) to evaluate the cell viability. Cell migration assay In scratch wound healing experiments, cells were seeded in the eight-well coverslip wells at a density 1.5 × 10^5 cells/cm^2. When the confluence of the adhered cells in each well reached 90%, a linear scratch was generated by a 200 μL pipette tip^[175]52. Wound healing images were recorded after continuous cell culture for 0, 8, 16, and 24 h, and the width of the scratches was measured by ImageJ software. Co-culture of osteoblasts and fibroblasts Osteoblast MC-3T3-E1 and fibroblast NIH-3T3 were prelabeled with CMFDA (green, C7025, Thermo Fisher) and CMTMR (red, C2927, Thermo Fisher) cell tracker, respectively^[176]53. After that, both of them were mixed at a ratio of 1:1 in α-MEM and DMEM mixed medium, followed by incubation on the peptide or polymer-modified glass slides at a density of 1.4 × 10^4 cells/cm^2[.] Fluorescent images of cells cultured on top of the slides were collected after culturing for 6 and 24 h. The cell area of the two cells was calculated using the ImageJ software. Characterization of the existence of modified layer for pre- and post-implantation Fluorophore Rhodamine-B were used to demonstrate that the polymers existed on the Ti nails for the pre- and post-implantation. Specifically, the polymer or peptide-modified Ti-screws were modified with the Rhodamine-B isothiocyanate ([177]M60789, Meryer) in a solution (5 mg/mL, DMSO/PBS = 4/1, pH = 8–9) for 12 h in the dark. Then the screws were alternately washed with Millipore water and 75% ethanol/water solution three times to remove the free Rhodamine-B fluorophores. The screws were dried with N[2] flow and implanted in the alveolar bone of the rats (n = 6). After the screw is screwed out again, it is washed successively with red blood cell lysis solution, PBST, and Millipore water. The fluorescence images of pre- and post-implantation nails were obtained with a fluorescence microscope. The fluorescence intensity of screws was obtained with a ChemiDoc MP Imaging system from Bio-Rad and quantitatively analyzed using ImageJ software. Fluorination modification: the screws modified with polymers were soaked in the N-succinimidyl trifluoroacetate solution (40 mg/ml, dissolved in dimethyl sulfoxide). After adding 2% V/V triethylamine and incubating at room temperature for 6 h, the screws were alternately washed with Millipore water and 75% ethanol/water solution. The elemental changes of these fluorinated nails before and after implantation were detected by X-ray photoelectron spectroscopy (XPS) characterization. Rat dental implant model for in vivo osseointegration study Female Sprague-Dawley rats (SD rats, eight-week old) with healthy oral cavities were randomly divided into two groups (n = 6 for each group). All rats were anesthetized with 50 mg/kg pentobarbital sodium in saline. Five groups of animals were implanted with RGD, KRSR, DM[50]CO[50], or MM[50]CH[50]-modified Ti-screws and uncoated Ti-screws into the alveolar bone, respectively. Animals were sacrificed after 2-, 4-, and 8-weeks post-operation, and the skulls were collected and fixed with 4% paraformaldehyde. The bone tissue in the implanted area was scanned with micro-CT (SkyScan 1172, SkyScan, Aartselaar, Belgium). The bone volume /tissue volume (BV/TV), trabecular number (Tb.N), trabecular thickness (Tb.Th), bone surface/bone volume (BS/BV), and trabecular separation (Tb.Sp) was calculated using VGstudio software. After micro-CT evaluation, tissue samples were subjected to histological analysis. Hard tissues were sectioned using the EXAKT system (Germany), then they were stained with methylene blue and acid fuchsin. Bone-to-implant contact (BIC%) was quantitatively analyzed using ImageJ software. The region of interest for BIC% statistical analysis was set to the area where the screw was inserted into the rat’s maxilla. Statistical analysis Statistical analysis was performed using Origin software. The respective numbers of data points, n, are indicated in the figure captions. Single comparison tests were performed by a two-tailed Student’s t-test, and multiple comparisons were performed using a one-way analysis of variance (ANOVA) with a Tukey post-test. Reporting summary Further information on research design is available in the [178]Nature Portfolio Reporting Summary linked to this article. Supplementary information [179]Supporting information^ (3.8MB, pdf) [180]Reporting Summary^ (261.5KB, pdf) [181]Transparent Peer Review file^ (752.5KB, pdf) Source data [182]Source Data^ (55.7KB, xlsx) Acknowledgements