The OS cell lines (143B (RRID : CVCL_2270), KHOS (RRID : CVCL_2546), MG-63 (RRID : CVCL_0426), U2OS (RRID : CVCL_0042)) were purchased from the American Type Culture Collection (ATCC). All the cell lines have been authenticated within the last three years and were mycoplasma-free. 143B cells were cultured in DMEM (Gibco), KHOS, MG63, U2OS cells were cultured in RPMI-1640 (Gibco), with 10% fetal bovine serum (FBS, Wisent) and 1% penicillin/streptomycin (PS, Gibco). All cells were cultured in the incubator at 37 °C in humidified air with 5% CO² and stored in CELLSAVING solution (New Cell & Molecular Biotech) at −80°C. Anlotinib (S8726) and Tocilizumab (anti-human IL-6R, A2012) were both purchased from Selleckchem. All products were stored and used according to manufacturer’s instructions.

To establish anlotinib-resistant cells, we treated the KHOS, 143B, MG63 and U2OS cells with gradually increasing anlotinib concentrations from 0.5 to 5 μM over 5 months. The culture medium was changed and anlotinib was added daily. Once the cells maintained stable proliferationin the existing anlotinib concentration, the anlotinib concentration was increased in with an 0.5 μM increments. Half maximal inhibitory concentrations (IC₅₀) were tested by CCK8 assay to detect cellular sensitivity to anlotinib-resistant cells (KHOS-R, 143B-R).

The OS cells and anlotinib-resistant cells were seeded into 96-well plates at a density of 3× 10³ cells/well. The cells were treated with different anlotinib concentrations and 100ng/ml tocilizumab. After 24-hour treatment, Cell Counting Kit-8 (CCK8, Dojindo) were added into culture medium following the manufacturer’s instructions. The cells were incubated for 1 hour in incubator, and the optical density (OD) value was measured at 450 nm using a microplate reader (Bio- Rad). IC₅₀ of OS cells to anlotinib were calculated using GraphPad Prism 9.0.

The cells were seeded into six-well plates at a concentration of 800 cells/well, incubated overnight, subsequently treated with 3μM anlotinib and 100ng/ml tocilizumab. The culture medium was changed every three days. After one week of treatment, the cells were fixed with 4% fixative solution (P1110, Solarbio) and stained with 0.1% crystal violet stain solution (G1063, Solarbio).

The OS cells were seeded into six-well plates at a concentration of 5×10⁵ cells/well, incubated overnight, and treated with different drugs at the mentioned concentrations for 24 h. Then the cells were stained with the FITC Annexin V Apoptosis Detection Kit I (556547, BD Biosciences) following the manufacturer’s instructions, and examined by flow cytometer (C6 Plus, BD Biosciences).

RNA was isolated from 4 kinds of OS cells and corresponding resistant OS cells, RNA sequencing (RNA-Seq) was performed by Biomarker Technologies Corporation (Beijing, China). The cDNA library was constructed using the NEBNext Ultra RNA Library Prep Kit for Illumina (NEB, E7530) and NEBNext Multiplex Oligos for Illumina (NEB, E7500). Gene expression levels were estimated using fragments per kilobase of exon per million fragments mapped (FPKM) values using the Cufflinks software. Differentially expressed genes (sensitive vs. resistant) were identified as log2 (fold change) > 2 and adjusted P value <0.05.

Total RNA was isolated using TRIzol reagent (15596018, Invitrogen) following the manufacturer’s protocol. Complementary DNA (cDNA) was synthesized using PrimeScript RT Master Mix (TaKaRa Biotechnology). qRT-PCR was performed to measure the relative target gene expressions using SYBR Green premix Ex Taq (TaKaRa, RR420A). Data were normalized to the controls and measured using the comparative CT method. The sequence primers were listed in Supplementary Table 1.

The cell culture supernatants of OS cells and anlotinib-resistant OS cells were collected to be used immediately or stored at -80°C. ELISA was performed by following the manufacturer’s instructions using enzyme‐linked immunosorbent assay kit (70-EK106, Multiscience).

Total protein was extracted from cell lysates using cell lysis buffer (Cell Signaling Technology). The protein concentration was measured using the BCA Protein Assay kit (Solarbio). Then, protein was loaded into SDS- PAGE gels and transferred onto PVDF membranes, and blocked with rapid Closure Solution (WB1601, Biotides) for 20min. The membranes were incubated with primary antibodies including anti-IL6 (77888-1-Ig, Proteintech), anti-N-cadherin (22018-1-AP, Proteintech), anti-Vimentin (60330-1-Ig, Proteintech), anti-STAT3 (10253-2-AP, Proteintech), anti-GAPDH (60004-1-Ig, Proteintech) and anti-β-actin (ab8226, Abcam) at 4 °C overnight. The next day, the secondary antibodies were added onto the membranes at 37 °C for 1 h. Protein bands were visualized using the Image Lab Software (Bio-Rad).

KHOS-R cells was added into 12 well plates and incubated overnight. Then, the cells were treated with 3μM anlotinib or 100ng/ml tocilizumab for 24h. EdU assay was performed using BeyoClick™ EdU Cell Proliferation Kit with Alexa Fluor 488 (C0071S, Beyotime Biotechnology) following the manufacturers’ protocol. Stained cells were photographed using a fluorescence microscope (Leica).

The cells were plated into a six-well plate and incubated until reached 80% confluence. A sterile pipette tip was used to scratch a line on the cells. Cells were washed twice with PBS and then treated with 3μM anlotinib or 100ng/ml tocilizumab for 24h.The images were taken at 0 hour and 24 hours under the microscope, the cell migration area was measured using Image J software.

Transwell inserts (3422, Corning, NY, USA) were used for our transwell assay. Cell complete medium with 3μM anlotinib was added into the bottom chamber of transwell plate, 1.2×10⁴ cells were seeded into the upper chamber and incubated with 100ng/ml tocilizumab for 24 hours. The migrating cells on the underside of the inserts were fixed with 4% fixative solution and stained with 0.1% crystal violet for 15 min. The cells remained on the upper chamber membrane were removed gently with a cotton swab. Images were acquired by the microscope, and the number of migrating cells was measured using Image J software.

The anlotinib-resistant cells were added into 24 well plates and incubated overnight. Then, the cells were treated with 3μM anlotinib or 100ng/ml tocilizumab for 24h, and fixed with 4% fixative solution. TRITC Phalloidin (CA1610, Solarbio) was used for cytoskeletal stain assay and cell nuclei were stained with DAPI (Solarbio) following the manufacturers’ protocol. Images were photographed using a fluorescence microscope.

We collected 104 osteosarcoma tissues after surgery, which were approved by the Ethics Committee of the Peking University People’s Hospital (2021PHE030), and made into 3 tissue arrays. Tumor tissues excised from the mice were fixed in 4% paraformaldehyde immediately, embedded in paraffin, and cut into 4-μm sections. The methods of Immunohistochemistry (IHC) and immunoreactive score (IRS) evaluation were performed as previously described (5). The Paraffin sections were reacted with the anti-IL-6 antibody.

The orthotopic osteosarcoma mouse models were performed using BALB/c-nude mice (female, 6-week-old). Animal experiments were approved by the ethics committee of Peking University People’s Hospital. To establishing the xenograft tumor model, 10μL 10⁶ anlotinib-resistant 143B-R cells were implanted into mice fibular marrow cavity and then randomly divided into four groups (4 mice per group): NC, anlotinib, tocilizumab, and combination therapy groups. Anlotinib was gastric gavage administration at a concentration of 2mg/kg once per day for 14 consecutive days, tocilizumab was injected at a concentration of 5mg/kg every one week intraperitoneally. The mice were treated once the obvious tumor formation, after 14-day treatment, the mice were sacrificed, the lungs and legs with tumor were removed, tumors and lungs were stored in 4% paraformaldehyde for IHC and HE staining analysis.

Data was expressed as mean ± S.D. (standard deviation). Unpaired two-tailed Student t test was used for analysis of two samples. The One- way ANOVA was used among different groups. All statistical analysis were performed with the GraphPad Prism 9.0. The level of significance was chosen as P < 0.05.

In order to construct anlotinib-resistant osteosarcoma cell lines, we gradually increased the anlotinib concentration added in the culture medium of 143B, KHOS, MG63 and U2OS cells over 4 months. To verify the establishment of the anlotinib-resistant cells, we first compared the IC_50s of OS cells with the anlotinib-resistant OS cells for 24 h by CCK8 assay. We found that the IC_50s of 143B and KHOS were 4.758μM and 13.91μM respectively, while the IC_50s of 143B-R and KHOS-R were increased to 15.25 μM and 26.38μM respectively ( Figures 1A, B ). In addition, the clone formation results showed that the proliferation of OS cells were obviously inhibited in response to 3 μM anlotinib, while the 143B-R and KHOS-R cells were not sensitive to anlotinib ( Figures 1C, D ). The results of apoptosis also showed that anlotinib-resistant OS cells significantly down-regulated their sensitivity to anlotinib ( Figure 1E ). All these results indicated the anlotinib-resistant OS cell lines were successfully constructed.

OS cell lines (negative control, NC) and anlotinib-resistant cell lines (RE) were collected for second-generation RNA sequencing, and the differentially-expressed genes between NC and RE were shown in Figure 2A , we found cytokine-cytokine receptor interaction was highly activated by KEGG enrichment analysis ( Figure 2B ). Gene set enrichment analysis (GSEA) further validated the enrichment of cytokine activity in anlotinib-resistant OS cell lines ( Figure 2C ). We found that IL-6 expression was significantly increased in anlotinib-resistant OS cells ( Figures 2A, D ), which suggested that IL-6 might be a key cytokine promoting anlotinib resistance. qPCR, western blot and ELISA assays were performed to further validate the sequencing results. Unexpectedly, the results further showed IL-6 was all enhanced in mRNA, intracellular protein and secretory protein levels ( Figures 2E–G ). IL-6 acts directly at IL-6R located on tumor cells to induce the expression of STAT3 target genes, and then drive tumor proliferation, survival and drug resistance (28). According to the results of western blot, we found the STAT3 expression was up-regulated in anlotinib-resistant cells ( Figure 2H ). These results revealed that IL-6/STAT3 axis was the key pathway in OS anlotinib resistance.

The 24-hour IC_50s of resistant cells were first tested by CCK8 assay. The results showed the IC₅₀ values of anlotinib monotherapy for KHOS-R and 143B-R were 23.76μM and 16.91μM, combined with 100ng/ml tocilizumab (an IL-6R inhibitor) (29), the IC_50s were reverted to 12.70μM and 9.388μM respectively ( Figure 3A ). The cell proliferation viability results indicated that 100ng/ml tocilizumab affected proliferation of resistant cells lightly, but combination with anlotinib obviously diminished proliferation viability and regained sensitivity to anlotinib according to colony formation and EDU assays ( Figures 3B–E ). The effect of combination treatment on OS cells apoptosis was also tested, in KHOS-R cells, we found tocilizumab monotherapy was ineffective, while a significant apoptosis rate was showed in combination therapy ( Figure 3F ). As for 143B-R cells, no obvious apoptosis difference was shown between the NC and anlotinib groups, and low dose tocilizumab or combination therapy both upregulated apoptosis rates of 143B-R ( Figure 3G ).

Malignant cancer progression also encompasses the tumor migration and invasion, which closely related with tumor metastasis (30). The results of wound healing and transwell assays showed combination therapy was more effective to diminished tumor migrating and invasive ability ( Figures 4A–E ). Epithelial-mesenchymal transition (EMT) has been defined as a process which epithelial cells lose the epithelial features and acquire mesenchymal features. For a long time, EMT has been viewed associating with various tumor functions, including tumor cell migration, metastasis, and resistance to therapy (31). Combination with tocilizumab treatment significantly downregulated mesenchymal marker proteins including N-cadherin and Vimentin, which indicated combination with tocilizumab inhibited EMT of OS resistant cells ( Figure 4F ). Cytoskeletal structure is useful to maintain cell physiological activity and tissue homeostasis (32). F-actin staining was used to assess the effect of tocilizumab on the anlotinib resistant cell, and the results showed anlotinib combinated with tocilizumab obviously impaired the cytoskeletal structure ( Figure 4G ). We have found the STAT3 expression was upregulated in anlotinib-resistant cells as mentioned, whether IL-6/STAT3 pathway can be inhibited by tocilizumab was not clear. The results of western blot showed combination therapy decreased STAT3 expression in 143B-R cells. Meanwhile, 100ng/ml tocilizumab monotherapy and combination therapy were both able to inhibit IL-6/STAT3 pathway in KHOS-R cells ( Figure 4H ).

Orthotopic xenograft model of OS in nude mice was established employing the 143B-R cells. The obvious tumor formation was shown, and then treatments of different groups were performed. After 14-day treatment, we found combination therapy with tocilizumab significantly inhibited tumor growth, as shown by the results of the volume and weight of the tumor tissues ( Figures 5A–C ). The images of the lungs showed that tocilizumab monotherapy and combination therapy both decreased the number of pulmonary metastatic nodules. And combination therapy significantly inhibited the growth of osteosarcoma lung metastasis nodules compared with NC group and anlotinib treatment group by HE staining ( Figures 5D, E ). To detect whether STAT3 pathway was blocked in mouse models, STAT3 expression levels in different treatment groups were examined by IHC assay. And we found STAT3 expressions were all decreased in anlotinib group, anti-IL-6R group, and combination treatment group ( Figures 5F, G ).

Tissue microarrays were established using 104 tumor specimens, and were used to examine IL-6 expression in patients with osteosarcoma by IHC staining ( Figure 6A ). We found IL-6 was highly expressed in OS tissues, and we translated IRS scores into three levels: high, medium, and low groups ( Figure 6B ). Combined with patient clinical information, we found that IL-6 expression was correlated with poor prognosis of patients, the median survival of the patients in high, medium, and low IL-6 expression groups were 1217, 1508, and 2274 days respectively ( Figure 6C ). Figure 6D presented the hypothetic mechanisms of combination therapy with tocilizumab enhances sensitivity to anlotinib for anlotinib resistant OS cells.