How Piezo1 controls cartilage cell energy function and helps bone healing through β-catenin and LARS2 signals

Similarly, immunohistochemistry (IHC) experiments revealed an increased proportion of HTCs in the callus of Piezo1^Col2a1 mice compared with Piezo1^f/f mice. The expression of Piezo1 (Fig. 2i, j) and the osteogenic markers OPN (Fig. 2k, l) and RUNX2 (Fig. 2m, n) were significantly reduced in the chondrocytes. Furthermore, we performed micro-computed tomography (micro-CT) scans on both Piezo1^f/f and Piezo1^Col2a1 mice 14 days (Fig. 2o), 21 days (Fig. S1p), and 28 days (Fig. S1p) after femoral fracture. The results showed that the fractures in Piezo1^f/f mice had essentially healed, and the callus tissue had largely completed remodeling by 28 days post-fracture (Fig. S1p). The results also depicted a significant reduction in bony callus tissue in the Piezo1^Col2a1 mice. Since our focus was on the endochondral ossification process during fracture healing, we selected 14-day post-fracture mice for the study. Bone volume/total volume (BV/TV) (Fig. 2p), and trabecular thickness (Tb.Th) (Fig. 2q) values of callus tissue in Piezo1^Col2a1 mice were significantly lower compared with Piezo1^f/f mice 14 days after femoral fracture. These findings suggest that Piezo1 plays a role in regulating endochondral ossification and, consequently, affects fracture healing.

To further assess changes in mitochondrial function, we measured the OCR and extracellular acidification aate (ECAR) in both Piezo1^WT and Piezo1^–/– cells. The results revealed a significant decrease in the overall OCR values and a rise in the overall ECAR values in Piezo1^–/– ATDC5 cells compared with Piezo1^WT cells (Fig. 3e, i). The reduced basal respiration OCR value in Piezo1^–/– ATDC5 cells suggests a decrease in the energy demand of these cells under basal conditions (Fig. 3f). The reduced ATP production in Piezo1^–/– ATDC5 cells indicated a diminished capacity for cellular ATP synthesis (Fig. 3g). And the decrease in the OCR during maximal respiration indicates a reduced capacity for cellular respiration, suggesting a decline in the ability of cells to reach their maximum respiratory potential (Fig. 3h). The rise of glycolytic proton efflux rate (glycoPER) in Piezo1^–/– ATDC5 cells indicated that the anaerobic respiration of cells increases (Fig. 3j). The decreased ratio of OCR/ECAR indicated that the Piezo1^–/– ATDC5 cells shift from aerobic respiration to anaerobic respiration, and the cellular energy metabolism declines (Fig. 3k). We also measured ATP levels using an ATP Assay Kit. The results revealed a significant decrease in ATP concentration in Piezo1^–/– ATDC5 cells compared to Piezo1^WT cells (Fig. 3l).

What's more, we observed obvious morphological changes in Piezo1^–/– cells compared with Piezo1^WT cells through optical microscope (Fig. S1a). Then, we used transmission electron microscopy (TEM) to observe the ultrastructure of mitochondria. We found a significant reduction in mitochondrial cristae and enlargement of the mitochondrial matrix in Piezo1^–/– cells (Fig. 3m). And we found through the CCK-8 assay that the proliferation ability of Piezo1^–/– cells decreased significantly (Fig. S1b). Similarly, we found that the level of reactive oxygen species (ROS) in Piezo1^–/– cells increased significantly (Fig. 3n). Furthermore, qPCR and WB results demonstrated a marked changes in the mRNA and protein levels of the mitochondrial dynamics-related markers mitofusin-1 (Mfn1), mitofusin-2 (Mfn2), optic atrophy 1 (OPA1), and dynamin-related protein 1 (Drp1) following Piezo1 knockout (Fig. S1c-k). These results demonstrate that Piezo1 plays a critical role in maintaining the mitochondrial structure and function. In conclusion, our data suggest that Piezo1 deficiency leads to defects in the mitochondrial structure and function. When we used mitoquinone mesylate (MitoQ) to protect the mitochondrial function in Piezo1^–/– ATDC5 cells, ALP and alizarin red staining revealed that the osteogenic ability of the cells were enhanced (Fig. S1l-o). These results suggest that Piezo1 can regulate the transdifferentiation of chondrocytes into osteoblasts by influencing mitochondrial function.

Lars2, a gene associated with mitochondrial function, plays a crucial role in maintaining mitochondrial function by regulating the translation of mitochondrial genes that are essential for cell differentiation. Through the scRNA-seq analysis mentioned earlier, we observed an increased expression of Lars2 in chondrocytes from Piezo1^Col2a1 mice. To test this hypothesis, we performed a series of in vivo and in vitro experiments. qPCR and WB analyses revealed that both the mRNA and protein levels of Lars2 were significantly elevated in Piezo1^–/– ATDC5 cells compared with the Piezo1^WT group (Fig. 4e–h). Moreover, immunofluorescence (IF) staining further confirmed the markedly increased expression of Lars2 in Piezo1^–/– ATDC5 cells (Fig. 4i). These results are consistent with in vivo observations. In the callus tissue of Piezo1^Col2a1 fracture mice, Lars2 protein levels were also significantly higher than those in Piezo1^f/f fracture mice (Fig. 4j, l). Furthermore, IHC and IF experiments revealed significantly increased Lars2 expression in hypertrophic chondrocytes of callus tissue from Piezo1^Col2a1 mice (Fig. 4m–o). Collectively, these findings suggest that Piezo1 knockout leads to a significant increase in Lars2 levels both in vitro and in vivo.

To investigate whether Lars2 acts as a key downstream regulator of Piezo1, we constructed Lars2 interference plasmids (shRNAs) and a negative control plasmid (NC) for in vitro experiments. Three interference sequences targeting Lars2 were designed (Table S1) and incorporated into the U6-MCS-CAG-EGFP vector to construct corresponding plasmids (shLars2-1, shLars2-2, and shLars2-3). These plasmids were then transfected into both Piezo1^WT and Piezo1^–/– ATDC5 cells. The results indicated that shLars2-1 plasmid demonstrated the most effective interference (Fig. S2a, b). Therefore, shLars2-1 interference plasmid was selected for subsequent studies.

WB analysis was used to assess Piezo1 and osteogenic markers, OPN and RUNX2, in callus tissue 14 days after femoral fracture (Fig. 6d–g). The results reported that shLars2-AAV significantly increased the expression of OPN and RUNX2 in Piezo1^Col2a1 mice (Fig. 6f, g). Moreover, the expression of Piezo1, Lars2, and the osteogenic markers OPN and RUNX2 in femoral fractures was examined in each treated group by IHC staining (Fig. 6h–o). The findings revealed that shLars2-AAV treatment accelerated fracture healing in Piezo1^Col2a1 mice. These results suggest that inhibiting Lars2 reverses the osteogenic impairment induced by Piezo1 deletion in chondrocytes and consequently restores endochondral ossification.

Finally, we performed micro-CT scans of the mouse specimens 14 days after the femoral fracture in each group, providing representative 3D reconstructions of the callus and coronal section images. The results demonstrated that compared with the Piezo1^f/f + AAV-NC group, the Piezo1^Col2a1 + AAV-NC group exhibited significantly reduced bone callus, clear fracture ends, and less woven bone coverage (Fig. 6p). Moreover, a significant decrease in BV/TV and Tb.Th of the callus was observed, indicating that a specific knockout of Piezo1 in chondrocytes inhibits fracture healing (Fig. 6q, r). Compared with Piezo1^Col2a1 + AAV-NC group, Piezo1^Col2a1 + AAV-shLars2 group exhibited increased bone callus with trabecular coverage at the fracture ends (Fig. 6p), along with an increase in BV/TV and Tb.Th (Fig. 6q, r). These results suggest that the inhibition of Lars2 can reverse the delay in fracture healing caused by Piezo1 knockout, highlighting that Piezo1 affects endochondral ossification in fracture healing by regulating Lars2 in chondrocytes.

To further explore the mechanism by which Piezo1 regulates Lars2 expression, we used the Piezo1 agonist Yoda1 and β-catenin signaling inhibitor XAV939. qPCR and WB analyses revealed that in the Yoda1-treated group of ATDC5 cells, the expression of both Piezo1 and β-catenin was significantly increased, while Lars2 expression was significantly decreased compared with the control group. This suggests that the activation of Piezo1 promotes β-catenin expression and inhibits Lars2 expression (Fig. 7i–m). In contrast, in the Yoda1 + XAV939 co-treatment group, Piezo1 expression remained unchanged, β-catenin protein levels were reduced, and Lars2 expression significantly increased (Fig. 7i–m). These results indicate that Piezo1 regulates Lars2 expression through the β-catenin signaling pathway.

To investigate whether Piezo1 regulates cellular mitochondrial function through the β-catenin/Lars2 pathway, we examined the mitochondrial parameters in ATDC5 cells. Compared with the control group, the Yoda1-treated group exhibited increased mitochondrial membrane potential and activity along with decreased superoxide generation (Fig. 7p–s). Conversely, inhibition of β-catenin signaling with XAV939 resulted in decreased mitochondrial membrane potential and activity and increased superoxide production in ATDC5 cells (Fig. 7p–s). Additionally, we found that activation of Piezo1 regulates the expression of mitochondrial dynamics-related markers Mfn1, Mfn2, OPA1, and Drp1 through the β-catenin signaling pathway (Fig. S3h–p). These findings suggest that Piezo1 regulates Lars2 expression through the β-catenin signaling pathway, thereby influencing mitochondrial function.

To verify the molecular mechanism through which Piezo1 affects the transdifferentiation of chondrocytes into osteoblasts by regulating Lars2 expression by β-catenin signaling, we conducted a series of correlation experiments. qPCR and WB results revealed that in the Yoda1-treated group, Piezo1 expression was significantly higher than that in the control group, and the expression of osteogenic markers OPN, RUNX2, Col1, ALP, and OCN were also markedly elevated, indicating that Piezo1 activation promotes endochondral ossification (Fig. 7i–o; Fig. S3a–g). However, in the Yoda1 + XAV939 group, Piezo1 expression remained unchanged, β-catenin expression was reduced, Lars2 expression was increased, and the levels of osteogenic markers were significantly decreased. These results suggest that Piezo1 regulates Lars2 expression through β-catenin signaling, thereby affecting endochondral ossification. Furthermore, ALP and alizarin red staining confirmed these findings (Fig. 7e–h). Taken together, these data demonstrate that activation of Piezo1 inhibits Lars2 expression through β-catenin signaling, promoting endochondral ossification in vitro.

To further investigate the molecular mechanisms involved, we conducted IHC and tissue IF experiments. Compared with the control group, activation of Piezo1 using Yoda1 significantly increased Piezo1 expression in chondrocytes within the callus (Fig. 8d, e). Moreover, the β-catenin levels were significantly elevated (Fig. 8f, g), whereas Lars2 expression was significantly reduced (Fig. 8h, i). Moreover, osteogenic markers such as OPN and RUNX2 demonstrated notable increases (Fig. 8j). Conversely, when the Wnt/β-catenin signaling pathway was inhibited by XAV939, there were no significant changes in Piezo1 expression in chondrocytes of the callus (Fig. 8d, e). However, β-catenin levels were significantly reduced (Fig. 8f, g), Lars2 expression was significantly increased (Fig. 8h, i), and the osteogenic markers, OPN and RUNX2, were significantly reduced (Fig. 8j). These results confirmed that Piezo1 promotes endochondral ossification by regulating Lars2 expression through the β-catenin signaling pathway.

Finally, we performed micro-CT scans of the mouse specimens 14 days after femoral fracture in each group, which provided representative 3D reconstructions of the callus and coronal section images. The results demonstrated that compared with the control group, activation of Piezo1 using Yoda1 significantly increased the bone callus in mice, blurred the fracture ends, and enhanced the coverage of newly woven bone (Fig. 8k). Additionally, there was a significant increase in BV/TV and callus Tb. Th (Fig. 8l, m), indicating that activation of Piezo1 accelerated the fracture healing process. In contrast, in the Yoda1 + XAV939 group, we observed a decrease in the bone callus, clearer fracture ends, and significantly reduced coverage of the newly woven bone (Fig. 8k), indicating that blocking the Wnt/β-catenin signaling pathway inhibited the osteogenic activity of Piezo1. In summary, this study suggests that Piezo1 regulates Lars2 through the β-catenin signaling pathway, promoting endochondral ossification and accelerating fracture healing in vivo.

All in vivo experiments were conducted on 12-week-old male mice. Wild-type C57BL/6J mice were obtained from the SiPeiFu Biotechnology Co. Ltd. (Beijing, China). Col2a1-CreERT2 mice were obtained from Cyagen Biosciences Inc. (Guangzhou, China). Col2a1-CreERT2; R26-tdTomato gene mice were obtained from the Shulaibao Biotechnology Co. Ltd. (Wuhan, China). The Piezo1^f/f mice were provided by Professor Weiguo Zou (Chinese Academy of Sciences, University of Chinese Academy of Sciences, Shanghai, China). To generate Piezo1^Col2a1 mice, Piezo1^f/f mice were mated with Col2a1-CreERT2 mice, and filial-generation 1 (F1) littermates were mated with each other. The genotypes of Piezo1^Col2a1 mice were determined by genotyping all filial-generation 2 (F2) offspring mice.⁴⁵ To produce Col2a1-CreERT2; Piezo1^f/f; R26-tdTomato mice, Piezo1^f/f mice were mated with Col2a1-CreERT2; R26-tdTomato mice and F1 littermates were mated with each other. The genotypes of Col2a1-CreERT2; Piezo1f/f; R26-tdTomato mice were genotyped in all F2 offspring. The primers used for each genotype are listed in Table S2. All animal experiments were approved by the Ethics Committee of the Hebei Medical University (IACUC-Hebmu-P2022017).

The surgical procedure for the middle femoral fracture model was as follows: After isoflurane gas anesthesia (3% for induction and 1% for maintenance), the right lower limb was shaved using an electric shaver. After disinfection with alcohol, an incision was made in the anterolateral side of the right femur. The muscles were separated by sharp dissection to expose the middle femur and the patella. The femur was then cut using a number 10 scalpel blade, and a 23G stainless steel intramedullary pin was inserted retrograde through the intercondylar notch into the medullary cavity, maintaining parallel alignment with the femoral mechanical axis. The tail of the needle was carefully trimmed to prevent the movement of the patellofemoral joint. The periosteum remained intact, and the soft tissue was well protected. The wound was closed after fracture stabilization. The mice were revived under anesthesia and returned to their housing cages for free movement. For the gene mice, tamoxifen (75 mg/kg) was dissolved in corn oil and administered subcutaneously for 5 consecutive days to induce the deletion of Piezo1 in chondrocytes. The first injection was performed on the 5th day after surgery, as chondrocytes in the callus first appeared on postoperative day 5.¹⁴ The mice were euthanised, the skin and muscles were dissected, and the entire femur was exposed. The hip and knee joints were detached, the intramedullary needle removed, and complete femoral samples collected.

For the ScRNA-seq experiments in this study, we used callus tissues from Piezo1^Col2a1 (n = 3) and Piezo1^f/f (n = 3) mice 14 days post-surgery. Enzymatic digestion was used to dissociate all the callus samples into suspensions of individual cells.⁴⁶ Next, complementary DNA (cDNA) amplification, chromium library construction, and library sequencing were performed using Lianchuan BioTechnology (Hangzhou, China). scRNA-seq data analysis was conducted using the Seurat R package (version 3.1.4) as previously described.⁴⁶ The scRNA-seq results were grouped and visualized using uniform manifold approximation and projection dimensionality reduction technology. DEGs were identified by setting the adjusted P-value threshold at 0.05 to define the DEGs. All scRNA-seq data were uploaded to the National Center for Biotechnology Information (NCBI) Gene Expression Omnibus database (accession number: GSE266774).

The removed fresh femoral specimens were scanned using a SkyScan 1176 micro-CT instrument (Micro-CT; 50 kV, 500 μA, and 9 μm/pixel). Micro-CT scan images were 3D reconstructed using the NRecon software (version 1.6, SkyScan; Microphotonics Inc., Allentown, PA, USA). Quantitative analysis of the fracture callus tissue images was performed using CTAn software (version 1.9, SkyScan). The parameters for micro-CT analysis were bone volume (BV), total volume (TV), bone volume fraction (BV/TV), and trabecular thickness (Tb. Th.).

Femoral samples were decalcified for 21 days using an EDTA decalcifying solution (EDTA, Solarbio Science & Technology, Beijing, China). After decalcification, samples were dehydrated and embedded in paraffin. A series of 4 μm consecutive paraffin-embedded sections were cut from each sample for HE, SO/FG, IHC, and IF staining. For Lineage tracing staining, we dehydrated the samples using 30% sucrose solution and performed frozen sectioning.

HE staining was performed using a commercial kit (G1120; Solarbio, Beijing, China) following standardized histopathological protocols. Tissue sections underwent sequential processing: (1) Deparaffinization through two xylene immersions (5 min each); (2) Rehydration in graded ethanol series (100%, 95%, 85%, 75%; 5 min per concentration); (3) Nuclear staining with Mayer's hematoxylin (2 min) followed by differentiated bluing in running tap water (5 min); (4) Cytoplasmic counterstaining with eosin Y (1 min); (5) Rapid dehydration through reverse ethanol gradients; (6) Final clearing in xylene and mounting with neutral gum. Stained sections were imaged under a microscope.

A modified Safranin O/Fast Green cartilage staining kit (G1371; Solarbio, Beijing, China) was used. Paraffin sections were dewaxed and rehydrated in water. A freshly prepared Weigert dye solution was applied for 5 min and then washed with water. The acidic differentiation solution was used for 15 s, followed by washing with distilled water for 10 min. The sections were then stained with a Solid Green staining solution for 5 min. Subsequently, the sections were quickly washed with a weak acid solution for 15 s to remove the residual green solid and dried. The sections were stained with a safranin staining solution for 5 min. Finally, the sections were dehydrated in 95% ethanol for 2–3 s, in absolute ethanol for 2–3 s, and then in absolute ethanol for 1 min. Xylene was used for transparency, followed by neutral gum sealing, and the slides were observed under a microscope.

Paraffin sections were routinely dewaxed, rehydrated, immersed in EDTA antigen retrieval solution in a 55 °C water bath for 10 h. An appropriate amount of endogenous peroxidase blocker was added dropwise and incubated for 10 min at room temperature. Goat serum was added, and the mixture was blocked for 1 h at room temperature. Diluted primary antibody was added and incubated at 4 °C overnight. The following day, enzyme-conjugated goat anti-rabbit/mouse IgG polymer added dropwise. The sections were incubated at 37 °C for 20 min. Appropriate amount of freshly prepared DAB solution was added and incubated for 5-8 min at room temperature. Finally, the samples were counterstained with haematoxylin for 1 min until the nuclei appeared blue, and the staining was immediately stopped. The sections were washed with tap water, dehydrated using an ethanol gradient, cleared with xylene, sealed with neutral gum, and observed under a microscope. The primary antibodies used were Piezo1 (1:100, 15939-1-AP, Proteintech, USA), Lars2 (1:200, 17097-1-AP, Proteintech, USA), β-catenin (1:200, 51067-2-AP, Proteintech, USA), RUNX2 (1:200, sc-390351, Santa Cruz Biotechnology, USA), and Osteopontin (1:200, 22952-1-AP, Proteintech, USA).

IF staining was performed according to standard experimental protocols.⁴⁷ The sections were dewaxed and rehydrated for antigen retrieval. Membranes were then incubated with the primary antibody overnight at 4 °C. The next day, the sections were rewarmed, washed with TBST, and incubated with fluorescent secondary antibodies for 1 h at room temperature. After washing the samples with TBST, the slides were sealed with DAPI anti-fluorescence mounting medium and observed under a fluorescence microscope. The detection markers in the IF staining experiment were Lars2 (1:200, 17097-1-AP, Proteintech, USA), RUNX2 (1:200, sc-390351, Santa Cruz Biotechnology, USA), and Osteopontin (1:200, 22952-1-AP, Proteintech, USA).

After restoring the frozen sections to room temperature, wash them three times with TBST. Goat serum was added and blocked for 1 h at room temperature. After removing the blocking solution, add the diluted primary antibody Collagen I (Col1) (1:100, Ab270993, abcam, USA) 4 °C to spend the night. The next day, restore the slices to to room temperature, wash them three times in TBST, drip an appropriate amount of Alexa Fluor 488 secondary antibody (1:200, A0423, Beyotime, China), and incubate at room temperature for 1 h. After washing the samples with TBST, the slides were sealed with DAPI anti-fluorescence mounting medium and observed under a fluorescence microscope.

Piezo1^WT and Piezo1^–/– ATDC5 cells were obtained from Cyagen Biosciences. Osteogenic differentiation of Piezo1^WT and Piezo1^–/– ATDC5 cells was performed following the protocol reported by Hendrickx et al.¹¹ After 7 days of chondrogenic differentiation, the chondrogenic induction medium of Piezo1^WT and Piezo1^–/– ATDC5 cells was replaced with osteoblast induction medium. After 7 days of osteogenic induction, ALP staining was performed using a BCIP/NBT alkaline phosphatase chromogen kit (Beyotime). After 21 days of osteogenic induction, calcium deposition was visualized using Alizarin Red staining (Solarbio Science & Technology). RNA and protein samples were extracted from the cells and collected after 14 days of osteogenic induction.

Total RNA was extracted from ATDC5 cells using the RNeasy RNA extraction kit (Thermo Fisher Scientific, CA, USA). Reverse transcription was performed using GoScript™ Reverse Transcription Mix and Oligo dT primers (Promega Corporation, Madison, WI, USA). The mRNA expression levels were analyzed by RT-PCR using a One Step RT-qPCR Kit (Sangon Biotech, Shanghai, China). Primers used for qPCR are listed in Table S3. The 2^−ΔΔCt method was used for data analysis.

Gel electrophoresis was performed using 20 µg protein samples per well at 150 V for 60 min. The PVDF membrane was operated at 400 mA for 30 min. The membranes were then blocked for 2 h with 5% nonfat milk powder. Immediately after blocking, the primary antibody was added and incubated overnight at 4 °C. After incubation, the primary antibody was removed, and the membranes were washed three times in TBST, followed by incubation with the secondary antibody. Finally, chromogenic experiments were performed using ECL chemiluminescence. The gray values of the bands were quantified using ImageJ software. The primary antibodies used were Piezo1 (1:1 000, 15939-1-AP, Proteintech, USA), Lars2 (1:1 000, 17097-1-AP, Proteintech, USA), RUNX2 (1:1 000, sc-390351, Santa Cruz Biotechnology, USA), Osteopontin (1:1 000, 22952-1-AP, Proteintech, USA), β-catenin (1:1 000, 51067-2-AP, Proteintech, USA), Collagen Type I (Col1) (1:1 000, 14695-1-AP, Proteintech, USA), ALP(1:1 000, PA5-106391, Thermo Fisher Scientific, USA), OCN (1:1 000, DF12303, Affinity Bioscience, USA), Mfn1 (1:1 000, 13798-1-AP, Proteintech, USA), Mfn2 (1:1 000, 12186-1-AP, Proteintech, USA), OPA1 (1:1 000, 27733-1-AP, Proteintech, USA), DRP1 (1:1 000, 12957-1-AP Proteintech, USA), β-Tubulin (1:1 000, ab7291, Abcam, USA), β-actin (1:1 000, 20536-1-AP, Proteintech, USA), and GAPDH (1:1 000, 10494-1-AP, Proteintech, USA). The secondary antibody used was (1:2 000, sc-2357, Santa Cruz Biotechnology, Santa Cruz, USA).

To examine the proliferation capacity of ATDC5 cells following Piezo1 knockout, we employed the Super-Enhanced Cell Counting Kit-8 (C0048S, Beyotime) for the assay. Briefly, Piezo1^WT and Piezo1^-/- ATDC5 cells (2 × 10³ cells/well) were resuspended and seeded into 96-well plates. After incubation for 48 h, the supernatants were removed and replaced with 100 μL of fresh medium containing 10 μL of CCK-8 solution. After incubation for 0 h, 1 d, 3 d, 5 d, 7 d, and 9 d at 37 °C in the dark, the absorbance was measured at 450 nm using a microplate reader.

ROS analysis was accomplished by staining with CM-H₂DCFDA (S0035S, Beyotime). The final concentration of CM-H₂DCFDA was 5 μmol/L. The ATDC5 cells were cultured as described above, then treated with CM-H₂DCFDA and Hoechst 33342 for ~30 min. Subsequently, the fluorescence of CM-H₂DCFDA was examined using flow cytometer (Sony ID7000). The data underwent processing with FlowJo version 10.8.1.

Cells cultured on slides were fixed with 4% PFA and blocked with 5% bovine serum albumin (BSA). The primary antibody was diluted in 5% BSA and incubated overnight at 4 °C. After rinsing, the cells were incubated with a fluorophore-conjugated secondary antibody for 1 h and counterstained with DAPI in the dark for 5 min. Images were acquired on a laser-scanning confocal microscope.

For immunoprecipitation, the lysates were precipitated with the indicated antibody or control IgG for 2 h. The immunoprecipitated complexes were collected and then washed four times with 1 mL of precooled IP lysate. Each wash was followed by centrifugation at 2 000 × g at 4 °C for 10 min, and the supernatant was carefully discarded. After the final washing step, as much of the supernatant as possible was discarded. Then, 80 μL of 1× reduced loading buffer was added, and the sample was boiled for 10 min in boiling water. The sample was centrifuged at 1 000 × g and 4 °C for 5 min, and the supernatant was aspirated. Finally, 10 μL of the supernatant was used for WB detection.

The SPR experiments were performed on a biacore T200 system at 25 °C. The β-catenin (HY-P74381, MCE, USA) (50 μg/mL) was immobilized onto a CM5 sensor chip (BR100530, Cytiva) by amine coupling, and the flow cell was equilibrated with the reaction buffer at a flow rate of 10 µL/min (1.0 × phosphate buffer, pH 7.4). Then, Lars2 protein at different concentration gradients (0.3125 μmol/L, 0.625 μmol/L, 1.25 μmol/L, 2.5 μmol/L, 5 μmol/L, 10 μmol/L) (Jinan Boshang Biotechnology Co., LTD) were injected over the chip, and the responses were recorded as resonance units (RU).⁴⁸