In total, 185 breast cancer patients with preoperative neoadjuvant chemotherapy (NAC) in Fujian Medical University Union Hospital between March 2012 to December 2019 were retrospectively selected in this study. All patients were histologically confirmed with TNBC and received at least six cycles of NAC with anthracycline and taxanes-based regimens. Surgical resection was performed followed by the completion of neoadjuvant chemotherapy. Patients with residual invasive tumor after NAC received oral capecitabine for six or eight cycles as recommended by the treating physician. Radiation therapy was performed at the discretion of the radiologist. Our study was approved by the Research Ethics Committee of Fujian Medical University Union Hospital (2022KY122) and written informed consent was obtained from each participant before inclusion in this study.

In this study, the cut-off value for ER and PR was less than 1% of positive tumor cells with nuclear staining. HER2 was considered as positive when immunohistochemistry (IHC) expression was 3 + or fluorescence in situ hybridization (FISH) was positive. Only tumors with negative expression of ER, PR, and HER2 were confirmed with TNBC. Pathological complete response (pCR) was defined as no residual invasive tumor in breast or lymph nodes. Patients only with ductal carcinoma in situ (DCIS) were also considered as pCR responders. Pathological evaluation was conducted to measure the continuous RCB index (wherein pathologic complete response has RCB 0; residual disease is categorized as RCB-I, RCB-II, and RCB-III) by two independent experienced pathologists [24]. Disease-free survival (DFS) was defined as the time of diagnosis to the date of disease relapse (with histopathology confirmation or radiological evidence of tumor recurrence). Overall survival (OS) was defined as the time of diagnosis until death from any cause. The last follow-up date was 1 May 2021.

Blood samples were obtained prior to the start of NAC (at baseline) and at surgery. All serum samples were centrifuged and stored at −80°C until use with no more than two freeze-thaw cycles allowed. Serum THBS2 (sTHBS2) level was measured by a commercially available enzyme-linked immunosorbent assay (ELISA) kit (Cat #DTSP20; R&D). Measurements were performed by strictly following the manufacturer’s instructions and repeated three times, with the final result marked as average level. All non-pCR patients were classified into different TNBC subtypes (LAR, IM, BLIS, MES, and US) with the surgical specimens based on the immunohistochemistry (IHC) staining procedures described previously [25]. Histological THBS2 (hTHBS2) expression for the surgical specimens in non-pCR patients were also evaluated by the IHC staining method. Slides were incubated with THBS2 (ab112543, abcam, 1:500) and a negative control was prepared by the substitution of primary antibody with phosphate-buffered saline (PBS, 5% BSA). The IHC staining scores were independently assessed by two pathologists according to the proportion of stained tumor cells and intensity of cellular staining. The proportion of stained tumor cells was scored from 1 to 4: 1, 0%–25%; 2, 26%–50%; 3, 51%–75% and 4, 75%–100%. The intensity of cellular staining was scored from 0 to 3: 0, no staining; 1, weak staining; 2, moderate staining; and 3, strong staining. The percentage of positive tumor cells and the staining intensity were multiplied to produce a weighted score for each patient. A score of 8–12 was defined as high expression level and a score of 0–7 was defined as low expression.

Human TNBC MDA-MB-231 and BT-549 cells were obtained from the Cell Bank of Type Culture Collection of The Chinese Academy of Sciences. TNBC cells were cultured in DMEM (HyClone; Cytiva) supplemented with 10% FBS (Gibco) and 1% penicillin and streptomycin solution. Cell culture was maintained in a 37°C incubator with 5% CO₂. Short hairpin RNA (shRNA) targeting THBS2 and its negative control were subcloned into GV493 lentiviral vector (GeneChem) and named sh-THBS2 and sh-Ctrl. Cells were transfected with lentivirus vectors for 48 h and further selected with 2 μg/ml puromycin. The efficiency of THBS2 knockdown was validated by quantitative real-time PCR (qRT-PCR) and western blot. The target sequences of shRNA were as follows:

shTHBS2-1: 5′-CTG​CGA​CCT​CAT​AGA​CAG​CTT-3′

shTHBS2-2: 5′-CCG​CTT​CGT​GCG​CTT​TGA​CTA-3′

shTHBS2-3: 5′-TTG​CTT​CAG​AAC​GTC​CAC​CTA-3′

shNC: 5′-TTC​TCC​GAA​CGT​GTC​ACG​T-3′

Total RNA in MDA-MB-231 and BT-549 cells was extracted with TRIzol reagent following the manufacturer’s instructions (Invitrogen). Complementary DNA was synthesized using PrimeScript RT Master Mix (Takara Bio, Inc.) and qRT-PCR was subsequently performed on a model 7500 Real-Time PCR System (Applied Biosystems) with SYBR Green kit (Takara Bio, Inc.). β-actin gene was detected for normalization of data. The respective forward and reverse primers were listed in Supplementary Table S1. Fold changes of gene expression were calculated by the 2−ΔΔct method; three independent replicates of all biological samples were assessed.

Cell proliferation was detected by cell count kit-8 (CCK8) assay and colony formation assay. For the CCK8 assay, 2000 cells were seeded per well in quadruplicate for 96 well plates. 10ul CCK-8 reagent (DOJINDO) was added to each well and the plates were maintained in a 37°C incubator for 2 h. The absorbance at 450 nm was measured by a microplate reader at indicated time points. For the colony formation assay, cells were plated into 6-well plates at a density of 500 cells per dish and cultured for 2 weeks. Colonies were fixed in methanol for 30 min at room temperature and stained with 1000 ul of crystal violet solution (Sangon Bio, Inc.). A colony was defined as >50 cells and the colony formation rate was assessed by colony number/seeding number.

MDA-MB-231 and BT-549 cells were spread at the bottom of 96-well plates and cultured overnight. A line wound was made by scraping 10 ul tips across the confluent cell layer. The floating cells were washed three times with PBS and the serum-free medium was used to maintain the cells. The images were captured at 0 and 24 h with an inverted light microscope.

Transwell assay was performed to assess the invasion. Matrigel (Corning, Inc.) was mixed with serum-free medium and added onto the upper surface of the chambers (Corning, Inc.; 8.0-µm filter), while the lower chambers were filled with medium with 30% FBS. After incubation at 37°C for 24 h, the tumor cells attached on the upper surface were removed with cotton swabs. The cells on the lower surface of the membrane were stained with Giemsa (Sigma-Aldrich) after fixation with methanol. The images of invasive cells were collected with a light microscope (magnification, ×100).

Correlations among sTHBS2 expression, clinicopathological characteristics, and NAC treatment response were compared by student’s t-test (for continuous variables) and chi-squared (χ2) test (for categorical variables). Survival analyses were conducted to explore the relationship between hTHBS2 expression, clinicopathological factors, and survival of non-pCR responders. Logistic regression analysis was also applied to identify independent predictors of treatment response. Survival curves were calculated by the Kaplan-Meier method and analyzed by log-rank test. Cox proportional hazard regression model was used for univariate and multivariate survival analysis. Each experiment was repeated 3 times and presented as mean ± standard deviation (SD). Student’s t-test and one-way analysis of variance were conducted for comparisons among the groups. Statistical analyses were performed with Statistical Package for the Social Sciences (SPSS, version 24.0) for Windows (Chicago, United States), with a two-sided p value of less than 0.05 considered statistically significant.

Of the 185 patients, 48 (25.9%) achieved pathological complete response (pCR). The detailed clinicopathological characteristics are shown in Table 1. The included patients had a median age of 49 years (range, 24–72 years). All patients were diagnosed with stage II and stage III disease, among whom 63 (34.1%) were T3/T4 and 151 (81.6%) had lymph node involvement. Most patients (89.7%) were pathologically diagnosed with invasive ductal carcinoma and 104 (56.2%) had a higher histology grade (Grade III). The Ki67 expression of the biopsy specimen was also evaluated, with 120 patients (64.9%) having Ki67 more than 50%.

Serum THBS2 (sTHBS2) was measured by ELISA both prior to the start of NAC (at baseline) and at surgery. Of the patients, 128 (37 pCR and 89 non-pCR) were eligible for serum samples collected at two time points. The median level of sTHBS2 at baseline and surgery was 24.34 ng/ml (range: 8.65 to 47.34 ng/ml) and 17.45 ng/ml (range: 7.49 to 48.13 ng/ml), respectively. The change of sTHBS2 level was also calculated, with the median reduction at 5.19 ng/ml (range: −13.68 to 25.47 ng/ml). The cut-off values of sTHBS2 at baseline, surgery, or change were all categorized by its median level. In chi-squared test, it was demonstrated that lower clinical T stage, lower clinical N stage, higher expression of Ki67, sTHBS2 at surgery, and sTHBS2 change were correlated with higher possibility of achieving pCR. In multivariate logistic regression analysis, sTHBS2 at surgery and the change in sTHBS2 were indicated as independent predictors of pCR as continuous variables (OR = 0.88, 95%CI: 0.79–0.98, p = 0.020 and OR = 1.12, 95%CI: 1.02–1.23, p = 0.015, respectively). Clinical T sage and Ki67 expression at baseline were also identified to be independent predictors for achieving pCR (p = 0.035 and p = 0.002, respectively). The correlations among sTHBS2 at surgery, sTHBS2 change, and tumor response (according to the RCB system) are shown in Figure 1. Lower levels of sTHBS2 at surgery and higher levels of sTHBS2 change were significantly correlated with a better treatment response to NAC. The mean level of sTHBS2 at surgery and absolute sTHBS2 change for patients with RCB-I, RCB-II, and RCB-III was 16.38, 19.53, 22.95 ng/ml and 8.71, 5.15, and 3.87 ng/ml, respectively.

Of the 185 patients, 48 achieved pCR and 137 were non-pCR responders. Only two patients with pCR were found to have tumor relapse, while 56 (40.9%) cases of non-pCR developed disease recurrence or metastasis. Therefore, survival analyses were conducted to explore the relationship between hTHBS2 expression, clinicopathological factors, and survival of these 137 non-pCR responders. The distributions of hTHBS2 expression in each RCB grade (RCB0, RCB-I, RCB-II, and RCB-III) were displayed in Figure 2A. No statistically significant difference for hTHBS2 was observed within each group. Representative IHC images of strong and weak THBS2 staining in cytoplasms were shown in Figure 2B.

In univariate analysis (Table 2), clinical stage at baseline (p = 0.001; p = 0.003), residual tumor size (p < 0.001; p < 0.001), residual lymph nodes (p < 0.001; p < 0.001), and hTHBS2 expression (p < 0.001; p = 0.001) were identified to be significant predictors of both disease-free survival and overall survival, respectively. Kaplan-Meier curves of the DFS and OS for the low hTHBS2 and high hTHBS2 groups are presented in Figures 2C,D. Better survival was more frequently observed in patients with a lower expression of hTHBS2 in residual tumor. A multivariate analysis with the cox proportional hazards model was then performed and the relevant results are shown in Table 3. Residual lymph nodes involvement (p < 0.001) and hTHBS2 expression in residual tumor (HR = 2.21, 95%CI = 1.24–3.94, p = 0.007) were of independent prognostic value for disease-free survival. As for overall survival, residual tumor size (p = 0.004), residual lymph nodes involvement (p = 0.022), and hTHBS2 expression in residual tumor (HR = 2.07, 95%CI = 1.09–3.92, p = 0.026) were identified as independent prognostic factors for patients’ outcome.

To investigate the function role of THBS2 in TNBC cells, we selected MDA-MB-231 and BT-549 cell lines for further studies. Three specific shRNA (shTHBS2) were designed and synthesized to knock down THBS2 in both cell lines. After stable expression, the shRNA (shTHBS2-1) which exhibited the highest interference efficiency was selected for the following experiments (Figure 3A). Western blot was further conducted to validate the interference efficiency (Figure 3B). The expression of THBS2 in both cells was confirmed to be suppressed compared with the shCtrl group. CCK-8 and colony formation assays revealed that knockdown of THBS2 significantly inhibited the proliferation and colony formation ability for both cells (Figures 3C,D). For wound healing and transwell assays, we also observed a notable decrease migration and invasion potential for THBS2 knockdown cells (Figures 4A,B). Together, these results suggest that THBS2 could act as a tumor activator for the growth and metastasis of TNBC cells.