HSulf‑1 and palbociclib exert synergistic antitumor effects on RB‑positive triple‑negative breast cancer.

Human sulfatase‑1 (HSulf‑1) is emerging as a novel prognostic biomarker in breast cancer. Previous studies demonstrated HSulf‑1 to function as a negative regulator of cyclin D1 in breast cancer. Accumulating preclinical evidence is supporting the efficacy of cyclin‑dependent kinase (CDK) 4/6 inhibitors against the luminal androgen receptor sub‑type of triple‑negative breast cancer (TNBC). It was therefore hypothesized that HSulf‑1 may cooperate with CDK4/6 inhibitors to control cell cycle progression in breast cancer cells. HSulf‑1 expression was found to be downregulated in TNBC tissues and cell lines compared with that in healthy tissues and non‑breast cancer cell lines, respectively. High levels of HSulf‑1 expression was also found to be associated with increased progression‑free survival and overall survival in patients with TNBC. Functionally, it was demonstrated that HSulf‑1 served as tumor suppressor in TNBC by inducing cell cycle arrest and apoptosis whilst inhibiting proliferation, epithelial‑mesenchymal transition, migration and invasion. Subsequent overexpression of HSulf‑1 coupled with treatment with the CDK4/6 inhibitor palbociclib exhibited a synergistic antitumor effect on retinoblastoma (RB)‑positive TNBC. Further studies revealed the mechanism underlying this cooperative antiproliferative effect involved to be due to the prohibitive effects of HSulf‑1 on the palbociclib‑induced accumulation of cyclin D1 through AKT/STAT3 and ERK1/2/STAT3 signaling. Taken together, findings from the present study not only suggest that HSulf‑1 may be a potential therapeutic target for TNBC, but also indicate that combinatorial treatment could be an alternative therapeutic option for RB‑positive TNBC, which may open novel perspectives.

Introduction

Triple-negative breast cancer (TNBC), which accounts for 10-20% of all types of breast cancers (1,2), is characterized by the lack of estrogen receptor (ER) and progesterone receptor expression in addition to the absence of human epidermal growth factor receptor 2 (HER2) amplification. Although it exhibits positive responses to chemotherapy, TNBC remain the most aggressive sub-type among younger patients since it is associated with a worse outcome and a higher risk of relapse within 5 years compared with other sub-types of breast cancer (3-7). Treatment of patients with TNBC has always been challenging due to the absence of effective therapeutic targets. Recently, olaparib, an inhibitor of poly (adenosine diphosphate-ribose) polymerase, has demonstrated promising antitumor activity in HER2-negative patients with a germline mutation in the BRCA gene (8). Although BRCA germline mutations have a relatively higher prevalence in TNBC compared with that in the non-TNBC sub-types, its prevalence in TNBC is only 12% (9,10), severely limiting the efficacy of this treatment strategy. Consequently, there is an urgent need to identify promising alternative treatment strategies for TNBC.

A hallmark of cancer is the dysregulation of cell division, ultimately favoring aberrant proliferation that fuels tumorigenesis and disease progression (11). The cyclin-dependent kinase (CDK)-retinoblastoma (RB)-E2F axis controls transition through the G1/S cell cycle checkpoint. CDK4/6 first forms a complex with cyclin D1 to promote the phosphorylation of the RB protein (12). Phosphorylated (p)-RB then allows the dissociation of transcription factor E2F from the pRb–E2F complex to promote G1/S transition (12). Inhibition of CDK4/6 results in RB dephosphorylation, leading to cell cycle arrest (12). Since CDK4/6 activity is frequently dysregulated and constitutively active in breast cancer cells, inhibition of CDK4/6 may prove to be an effective therapeutic intervention against breast cancer (2). Treatment using highly selective CDK4/6 inhibitors, including palbociclib, ribociclib and abemaciclib, have been previously reported to improve progression-free survival (PFS) in the first and later lines of ER+ metastatic breast cancer therapy (13-18). Since TNBC is a group of highly prolifera-tive tumors enriched in cell cycle genes, intrinsic resistance to CDK4/6 inhibitor monotherapy has been observed owing to the loss or mutation of RB (19-21). By contrast, it was also reported that the loss of RB in TNBC can result in improved response to chemotherapy (22-24). The Cancer Genome Atlas study revealed that since only 20% of patients with TNBC carry mutations in RB or are RB-negative, the use of CDK4/6 inhibitors may serve as a promising therapeutic strategy for RB-positive TNBC (2).

Human sulfatase-1 (HSulf-1) is an endosulfatase that selectively removes 6-O-sulfate from heparan sulfate proteoglycans (HSPGs) in the extracellular matrix (25). The sulfation of HSPG residues is required for their interaction with a number of heparin-binding growth factors (HBGFs), including fibroblast growth factor 2 (FGF-2), hepatocyte growth factor (HGF) and heparin-binding epidermal growth factor (HB-EGF), to promote their activation (25,26). In normal human tissues, HSulf-1 is stably expressed and serves important roles in cell proliferation and differentiation (27). However, HSulf-1 expression has been reported to be downregulated in various tumors, including ovarian, breast and gastric carcinomas (28,29). A number of studies previously revealed that the down-regulation of HSulf-1 in cancers is mediated by epigenetic modification, including CpG island methylation and histone acetylation (28,29) Reduced expression of HSulf-1 is one of the causes behind the increased sulfation of HSPGs, leading to excessive activation of the signal transduction pathways downstream of growth factor receptors. Ultimately, this results in cancer cell proliferation, increased metastatic activity, inhibition of apoptosis and reduced sensitivity to radio- and chemotherapy (30-37). Previously, several studies revealed that the overexpression of HSulf-1 inhibited tumor proliferation and angiogenesis in xenografts derived from breast cancer cell lines (36). Additionally, high levels of HSulf-1 have also been found to correlate with increased PFS and overall survival (OS) in patients with breast cancer (38,39). Mechanistically, it was reported that HSulf-1 inhibited autocrine cyclin D1 expression, leading to decreased S phase fractions whilst increasing G2-M fractions and increased cell death (39). However, the potential relationship between HSulf-1 and CDK4/6 inhibition remains unexplored, which serves as the subject of the present study.

In the present study, HSulf-1 expression level was first measured in TNBC tissues and cell lines, following which the function of HSulf-1 in TNBC was next demonstrated. Secondly, the effects of HSulf-1 overexpression in combination with treatment with the CDK4/6 inhibitor palbociclib on RB-positive TNBC was investigated. Finally, further studies were performed to explore the mechanism underlying this effect. Based on the present study, it could be speculated that HSulf-1 may not only serve as a potential therapeutic target but also a reliable efficacy predictor for CDK4/6 inhibitors. Therefore, in clinical practice, patients with TNBC and elevated HSulf-1 expression can potentially be selected for palbociclib treatment, which may achieve superior therapeutic efficacy.

Materials and methods

Human tissue specimens

In total, 86 tissue TNBC specimens and paired adjacent normal breast tissues were collected from patients with TNBC (sex, all female; age range, 25-73 years; median age, 47) at the General Hospital of The Yangtze River Shipping (Wuhan, China), Wuhan Polytechnic University (Wuhan, China) and the Henan Provincial People's Hospital, Zhengzhou University (Zhengzhou, China) between January 2000 and January 2019. All specimens were collected by surgical resection and the distance between the tumour and adjacent normal tissue was >1 cm. Non-TNBC patients, patients receiving chemotherapy or radiotherapy prior to surgical treatment and pregnancy concomitant with the diagnosis were excluded from the present study. The follow-up time range was 6-210 months. Informed consent was obtained from all patients. This experiment was approved by the Ethics Committee of the General Hospital of The Yangtze River Shipping, Wuhan Polytechnic University (approval no. 2017IEC0003). All specimens were classified according to the 2003 World Health Organization Consensus Classification and the Eighth Edition AJCC Cancer Staging for breast cancer (40,41). Detailed clinicopathological features are shown in Table I.

Table I

Association between HSulf-1 expression and clinicopathologic characteristics of 86 TNBC patients.

Parameters	HSulf-1 expression				P-valuea
	Low (n=55)	%	High (n=31)	%
Age (years)					0.4238
≤50	35	67	17	33
>50	20	59	14	41
Grade					0.6237
Low	10	59	7	41
High	45	65	24	35
Tumor size (cm)					0.5089
≤5	41	63	24	37
>5	14	67	7	33
Lymph node metastasis					0.7535
Negative	30	64	17	36
Positive	25	64	14	36

χ2 test. HSulf-1, human sulfatase.

Immunohistochemistry

Tissue samples were first fixed in 4% formalin for 24 h at room temperature and then embedded in paraffin blocks. Tissue sections (thickness, 4 µm) were subsequently deparaffinized in a graded series of xylene followed by rehydration in a descending series of ethanol before antigen retrieval in citrate buffer (pH 9.0) at 100°C for 2 min. Then, the tissue sections were blocked for 30 min with 3% BSA (Sigma-Aldrich; Merck KGaA) at room temperature. After incubation with primary anti-HSulf-1 (1:250; cat. no. ab32763; Abcam) at 4°C overnight and secondary antibodies (biotin-conjugated goat anti-rabbit IgG, 1:400 dilution, cat. no. SA00004-2; Proteintech Group, Inc.) at room temperature for 30 min, the sections were then incubated with avidin-biotin complex-horseradish peroxidase (HRP; Vectastain® ABC kit, Maravai LifeSciences) at 37°C for 20 min followed by incubation with 3'3-diaminobenzidine (Sigma-Aldrich; Merck KGaA) and counterstained with hematoxylin (Sigma-Aldrich; Merck KGaA) for 5 min at room temperature. The percentage of positively stained cells was scored on a scale of 0 to 4 as follows: i) 0, ≤5%; ii) 1, 6-25%); iii) 2, 26-50%); iv) 3, 50-75%; and v) 4, 76-100%. The staining intensity was scored from 0 to 3 as follows: i) 0, negative); ii) 1, weak; iii) 2, moderate; and iv) 3, strong). The scores for the percentages of positive cells and staining intensities were then multiplied to generate an immunoreactivity score (IS) for each case. The IS ranged from 0-12 (0, 1, 2, 3, 4, 6, 8, 9 and 12). Cut-off values for this scoring system were assigned as follows: i) High HSulf-1 expression, IS ≥4 (4, 6, 8, 9 and 12); ii) low HSulf-1 expression, IS <4 (0, 1, 2 and 3). Tissue slides were visualized under a light microscope (magnification, x400; Olympus FSX100; Olympus Corporation). Five high-power fields were examined randomly for each section. Quantifications of the proteins of interest were performed using the Image Pro Plus 6.0 software (Media Cybernetics, Inc.).

Cell lines and culture

The breast cancer cell lines (Hs578T and MDA-MB-231) human immortalized breast epithelial cell line (MCF10A) and 293T cells were purchased from American Type Culture Collection (ATCC). Hs578T MDA-MB-231 and 293T cells were cultured in DMEM (Gibco; Thermo Fisher Scientific, Inc.) with 10% FBS (Gibco; Thermo Fisher Scientific, Inc.) in a humidified atmosphere at 37°C in 5% CO2. MCF10A cells were cultured in DMEM/F12 (Procell Life Science & Technology Co., Ltd.) containing 5% horse serum, 20 ng/ml EGF, 0.5 µg/ml hydrocortisone, 10 µg/ml insulin and 1% NEAA (Procell Life Science & Technology Co., Ltd.) in a humidified atmosphere at 37°C in 5% CO2. Palbociclib (Sigma-Aldrich; Merck KGaA) was dissolved in DMSO, following which the cells were treated with 500 nM palboci-clib for 72 h at 37°C. Palbociclib (500 nM) was used for the majority of the experiments as previously described (42,43). An equivalent concentration of DMSO was used as negative control treatment in parallel with palbociclib treatment.

Plasmids, lentivirus production and cell transfection

The plasmids (pcDNA3.0-vector, pcDNA3.0-HSulf-1, pcDNA3.0-cyclin D1, LV105-HSulf-1 and LV105-EGFP) and the vector psi-H1 used for cloning the HSulf-1 short hairpin RNA (shRNA) were purchased from iGene Biotechnology Co., Ltd. Hs578T and MDA-MB-231 cells (5x105 cells/well) were plated into six-well plates 24 h prior to transfection. Plasmids (2.5 µg/well) were transfected into cells using Lipofectamine® 3000 (Invitrogen; Thermo Fisher Scientific, Inc.) according to the manufacturer's protocol for in vitro assays. Cells were incubated for 48 h at 37°C prior to further experimentation. The lentivirus particles were produced by transfecting 293T cells (ATCC) with pEZ-Lv105 lentiviral vectors encoding HSulf-1 (LV105-HSulf-1) and the control empty vector (LV105-EGFP) using the LentiPac™ Expression packaging kit (GeneCopoeia, Inc.) according to the manufacturer's protocols. The lentivirus-containing supernatants were harvested 72 h following transfection and were filtered through 0.45-µm PVDF filters (EMD Millipore). The supernatant was then concentrated by ultracentrifugation at 100,000 x g for 2 h at room temperature. MDA-MB-231 cells (4x105 cells/well; multiplicity of infection, 10) were infected with the lentiviral particles (2.03x108 TU/ml) where the stable cell lines were established by treatment with puromycin (2.5 µg/ml) for 2 weeks at 37°C for in vivo study. Transfection efficiency was determined by reverse transcription-quantitative PCR (RT-qPCR) and western blot analysis. The target sequences used for shRNA were as follows: ShHSulf-1 1, 5'-CCC AAA TAT GAA CGG GTC AAA-3' and shHSulf-1 2, 5'-CCA AGA CCT AAG AAT CTT GAT-3'. The plasmid shHSulf-11 was chosen for further study based on its superior silencing effect.

RT-qPCR

Total RNA was extracted from MDA-MB-231 cells transfected with the HSulf-1 overexpression or vector plasmid using RN07-EASYspin kit (Aidlab Biotechnologies Co., Ltd) according to manufacturer's protocols. cDNA was then synthe-sized using the PrimeScript™ RT Master Mix (Takara Bio, Inc.) from 1 µg RNA according to manufacturer's protocols The following temperature protocol was used for the reverse transcription reaction: 37°C for 15 min, followed by reverse transcriptase inactivation reaction: 85°C for 5 sec. qPCR reactions were performed using SYBR® Premix Ex Taq™ (Takara Bio, Inc.) according to manufacturer's protocols. The thermo-cycling conditions were as follows: Initial denaturation at 95°C for 30 sec, followed by 40 cycles of 95°C for 5 sec and 60°C for 30 sec. Relative expression was calculated using the 2-∆∆Cq method (44). GAPDH was used as an internal control. The sequences of the primers were as follows: Cyclin D1 forward, 5'-CCC ACT CCT ACG ATA CGC-3' and reverse, 5'-AGC CTC CCA AAC ACC C-3'; GAPDH forward, 5'-GGA GCG AGA TCC CTC CAA AAT-3' and reverse, 5'-GGC TGT TGT CAT ACT TCT CAT GG-3'.

Western blotting

Protein extracts were prepared using RIPA buffer (Thermo Fisher Scientific, Inc.) supplemented with protease and phosphatase inhibitors. Protein concentrations were determined using a bicinchoninic acid protein assay kit (Thermo Fisher Scientific, Inc.). A total of 20 µg total protein was loaded per lane and separated by SDS-PAGE (10 or 12% gels) before transferal to polyvinylidene fluoride membranes (EMD Millipore). The membranes were blocked in 5% skimmed milk diluted with Tris-buffered saline/Tween-20 (0.1%) (TBS-T) at room temperature for 1 h and subsequently incubated overnight at 4°C with the following primary antibodies: Anti-RB (1:1,000, cat. no. 9309; Cell Signaling Technology, Inc.), anti-p-RB (1:1,000, Ser780; cat. no. 9307; Cell Signaling Technology, Inc.), anti-HSulf-1 (1:1,000, cat. no. ab32763; Abcam), anti-E-cadherin (1:500, cat. no. ab15148; Abcam), anti-vimentin (1:1,000, cat. no. ab92547; Abcam), anti-N-cadherin (1:2,000, cat. no. ab76011; Abcam), anti-cyclin D1 (1:200, cat. no. ab16663; Abcam), anti-STAT3 (1:1,000, cat. no. 30835; Cell Signaling Technology, Inc.), anti-p-STAT3 (Y705; 1:2,000, cat. no. 9145; Cell Signaling Technology, Inc.), anti-JAK2 (1:1,000, cat. no. 74987; Cell Signaling Technology, Inc.), anti-p-JAK2 (Tyr1007; 1:1,000, cat. no. 4406; Cell Signaling Technology, Inc.), anti-AKT (1:1,000, cat. no. 9272; Cell Signaling Technology, Inc.), anti-p-AKT (Ser473; 1:1,000, cat. no. 9271; Cell Signaling Technology, Inc.), anti-ERK1/2 (1:1,000, cat. no. 4695; Cell Signaling Technology, Inc.), anti-pERK1/2 (Thr202/Tyr204; 1:2,000, cat. no. 4370; Cell Signaling Technology, Inc.), anti-GAPDH (1:5,000, cat. no. 60004-1-Ig; Proteintech Group, Inc.), and anti-β-actin (cat. no. 60008-1-Ig; Proteintech Group, Inc.). The next day, the membranes were washed with TBS-T and then incubated with secondary antibodies including horseradish peroxidase-conjugated goat anti-mouse (1:10,000, cat. no. SA00001-1; Proteintech Group, Inc.) and goat anti-rabbit (1:10,000, cat. no. SA00001-2; Proteintech Group, Inc.) at 37°C for 1 h. Immunoreactive bands were visualized using ECL reagent (Advansta, Inc.) and detected using the ChemiDoc XRS+ system (Bio-Rad Laboratories, Inc.). Gray values were calculated using ImageJ software (version 1.52u; National Institutes of Health).

Cell counting kit-8 (CCK-8) assay

Hs578T and MDA-MB-231 cells (3,000 cells/well) were first seeded into 96-well plates in triplicate, allowed to attach for 24 h at 37°C and treated with the indicated concentrations of palbociclib for 0, 24, 48 and 72 h at 37°C. Subsequently, 10 µl CCK-8 solution (MedChemExpress) was added into each well before the plates were subsequently incubated for 4 h at 37°C. The optical density was measured at 450 nm using the SpectraMax M5 microplate reader (Molecular Devices, LLC).

Colony formation assay

Cells were seeded in 6-well plates at a density of 1,000 cells per well. After 24 h of culture at 37°C, the Hs578T and MDA-MB-231 cells were transfected with the HSulf-1 overexpression or vector plasmid and then treated with 500 nM palbociclib for 14 days at 37°C. After 14 days, cell colonies were fixed with 4% paraformaldehyde and stained with 0.1% crystal violet, each for 30 min at room temperature. Cell colonies containing >50 cells were counted using the ImageJ software (version 1.52u; National Institutes of Health) and images were taken with a camera (Canon, Inc.).

Cell cycle assay

Hs578T and MDA-MB-231 cells were harvested and fixed with 70% cold ethanol at 4°C overnight, following which the cells (8x105/tube) were washed with PBS twice, mixed with 400 µl propidium iodide (PI) (Beyotime Institute of Biotechnology) and incubated for 30 min in the dark at room temperature and then analyzed by flow cytometry (Cytoflex, Beckman Coulter, Inc.) with a CytExpert software (version 2.3; Beckman Coulter, Inc.). Each sample was tested in triplicate.

Apoptosis analysis

Hs578T and MDA-MB-231 cell apoptosis was analyzed by flow cytometry (CytoFLEX; Beckman Coulter, Inc.) as per the FITC-conjugated Annexin V/PI method (Annexin V-FITC/PI Apoptosis Kit; Beyotime Institute of Biotechnology). After 72 h of the indicated treatment, adherent and suspended cells (3x105/tube) were both harvested by centrifugation and resuspended in 195 µl binding buffer, following which Annexin V-FITC (5 µl) and PI (10 µl) were added to each sample and mixed. Cells were incubated for 15 min in the dark at 4°C and then analyzed by flow cytometry (Cytoflex; Beckman Coulter, Inc.) using a CytExpert software (version 2.3; Beckman Coulter, Inc.). Each sample was tested in triplicate.

Wound-healing assay

Hs578T and MDA-MB-231 cells were seeded into six-well plates at a density of 5x105 cells per well and allowed to attach for 24 h at 37°C. Plasmids (pcDNA3.0-vector, pcDNA3.0-HSulf-1, shNC and shHSulf-1) were transfected into cells and then treated with 500 nM palbociclib for 72 h 37°C. Cells were harvested and reseeded into six-well plates (1x106 cells/well) and cultured to 90% confluence. A 200 µl sterile pipette tip was then used to scratch a wound (time 0 h), following which fresh low-serum (2% FBS) DMEM was immediately replaced after three times of washing with PBS. Cell migration was imaged under a light microscope (magnification, x100; Olympus Corporation) at 0 and 24 h after the injury. The wound width was evaluated by measuring the distance between the two edges of the scratch at five sites in each image. Cell migration was determined using the following formula: Percentage of wound healing (%) = [(wound width at the 0 h time point-wound width at the 24 h time point)/wound width at the 0 h time point] x100.

Transwell assay

Transwell chambers (Corning, Inc.) were coated with 50 µl of DMEM-diluted Matrigel (1:8 dilution; BD Biosciences) which was precooled at 4°C for 12 h and then incubated overnight at 37°C. Cells were suspended in serum-free culture medium (5x104 cells/well). A total of 200 µl cell suspension was transferred into the upper chambers whilst 600 µl DMEM supplemented with 10% FBS was added to the lower chambers. After 24 h incubation, the cells that migrated to the lower chambers were fixed with 4% paraformaldehyde and stained with 0.1% crystal violet each for 30 min at room temperature. The number of migratory cells were counted from three randomly chosen fields per chamber under a light microscope (magnification, x100; Olympus Corporation).

Tumor xenografts

Animal experiments were performed with the approval of the Animal Care and Use Committee of the General Hospital of The Yangtze River Shipping, Wuhan Polytechnic University (approval no. 2017IEC0003; Wuhan, China). A total of 12 four-week-old female BALB/c nude mice were maintained under standard conditions (room temperature, 22±2°C; relative humidity, 55±10%) on a 12-h light/dark cycle (lights on at 6:00 a.m.), where they had ad libitum access to food and water. A daily observation was preformed to prevent the animals from anger, restlessness, fear, anxiety, pain or damage to keep them at a normal status. These mice (age, 4 weeks) were randomly divided into two groups (n=6 mice/group). MDA-MB-231 cells (2x107) stably transfected with LV105-HSulf-1 plasmids or the control vector were suspended in 100 ml PBS and subcutaneously injected into the nude mice under the axilla of the left forelimb. Once the tumors of the control vector group reached ~300 mm3, mice with tumors of similar size were distributed into the treatment cohorts (n=3/group). Mice were orally treated with 50 mg/kg palbociclib dissolved in 0.5% methylcellulose or vehicle (0.5% methylcellulose) daily for 21 days. Tumor sizes were monitored every 3 days for 3 weeks. At the completion of 3 weeks of palbociclib treatment, all 12 mice were euthanized by cervical dislocation. Death was confirmed by the cessation of breathing, muscle relaxation and the absence of nerve reflex. The experiment lasted a total of 6 weeks. No accidental death was observed before these mice were euthanized. The tumors were then excised and measured. The length (l) and width (w) of each tumor were measured using digital calipers, where tumor volumes were calculated according to the following formula: Tumor volume (V)=lxw2x0.5.

Statistical analysis

Statistical analyses were performed using GraphPad Prism 8.0 software (Graphpad Software, Inc.). Data are presented as the means ± standard deviation. Data were analyzed using Student's t-test or one-way ANOVA followed by either Dunnett's or Tukey's post hoc test. Wilcoxon-signed rank sum test and the χ2 test was used to analyze the IHC data, whilst the log-rank test was used to assess the statistical signifi-cance of the Kaplan-Meier plots and Cox proportional-hazards model was used to determine the hazard ratio and 95% CI. P<0.05 were considered to indicate a statistically significant difference.

Results

HSulf-1 is downregulated in TNBC tissues and cell lines and high HSulf-1 levels associate with improved prognosis in patients with TNBC

It was demonstrated previously that HSulf-1 expression was downregulated in ~60% of breast cancers, which was in turn associated with poor patient prognosis (38,39). Therefore, the expression levels of HSulf-1 in 86 pairs of TNBC and adjacent normal mammary tissues were first measured, where it was found that HSulf-1 expression was significantly downregulated in TNBC tissues compared with that in the paired normal tissues (Fig. 1A and B). In addition, higher expression levels of HSulf-1 was found to significantly associate with prolonged PFS (HR, 1.943; 95% CI, 1.054–3.581; P=0.025) and OS (HR, 2.159; 95% CI, 1.041–4.478; P=0.032) in TNBC (Fig. 1C and D). However, HSulf-1 levels showed no significant associations with the clinicopathological indicators, including age, tumor grade, tumor size and lymph node status (Table I). A number of preclinical studies previously suggested that CDK4/6 inhibitors exerted inhibitory effects in RB-positive TNBC cell lines (43,45-47). Therefore, two RB-positive cell lines Hs578T and MDA-MB-231 were chosen to examine the expression levels of HSulf-1 and RB phosphorylation in these two cell lines and one human immortalized breast epithelial cell line MCF10A (43,47). HSulf-1 expression also was also found to be significantly downregulated in Hs578T and MDA-MB-231 compared with that in MCF10A (Fig. 1E). Collectively, these data suggested that HSulf-1 expression was downregulated in TNBC tissues and cells, where reduced HSulf-1 expression was highly predictive of poor patient prognosis.

Figure 1

Detection of HSulf-1 expression in TNBC and its association with patient survival. (A) Representative immunohistochemistry images for HSulf-1 staining in paraffin-embedded tumor and adjacent normal mammary tissues from 86 patients with TNBC. (B) The relative protein expression of HSulf-1 was detected using immunohistochemistry in 86 pairs of TNBC tissues and paired adjacent normal breast tissues (C) Association between HSulf-1 expression levels and progression-free survival in the 86 patients with TNBC. (D) Association between HSulf-1 expression levels and overall survival in 86 patients with TNBC. (E) HSulf-1 expression, RB phosphorylation and total RB expression were measured by western blotting in Hs578T, MDA-MB-231 and MCF10A cells. β-actin was used as an internal control. The results were derived from three independent experiments. Data are presented as the mean ± SD. *P<0.05. TNBC, triple-negative breast cancer; HSulf-1, human sulfatase-1; RB, retinoblastoma; p-, phosphorylated; t-, total; HR, hazard ratio; CI, confidence interval.

Overexpression of HSulf-1 inhibits cell proliferation and exhibits a synergistic antiproliferative effect with palbociclib on RB-positive TNBC cells, both in vitro and in vivo

To explore the effects of HSulf-1 and/or palbociclib treatment on TNBC cell proliferation, an HSulf-1 overexpression plasmid was transfected into Hs578T and MDA-MB-231 cells. CCK-8 assay was performed to measure cell viability following treatment with palbociclib for various periods of time (0, 24, 48 and 72 h). Overexpression of HSulf-1 was found to significantly reduce the viability of TNBC cells compared with those transfected with the control vector (Fig. 2A). In addition, combined HSulf-1 overexpression and palbociclib treatment synergistically potentiated the reduction of cell viability compared with that following either intervention alone (Fig. 2A). Colony formation assays yielded similar trends of results (Fig. 2B). By contrast, downregulation of HSulf-1 expression using shRNAs was found to significantly increase proliferation compared with cells transfected with the control vector (Fig. S1A-C).

Figure 2

HSulf-1 inhibits cell proliferation and when combined with palbociclib, exerts antiproliferative effects on RB-positive TNBC cells in a synergistic manner both in vitro and in vivo. (A) Hs578T and MDA-MB-231 cells were transiently transfected with the control vector or HSulf-1 overexpression plasmids and incubated for 6 h at 37°C then incubated with fresh complete medium cantaining 500 nM palbociclib or DMSO for 0, 24, 48 and 72 h. Subsequently, Cell Counting Kit-8 assays were performed to assess cell viability. (B) Hs578T and MDA-MB-231 cells were transiently transfected with the control vector or HSulf-1 overexpression plasmids and incubated for 6 h at 37°C then incubated with fresh complete medium containing 500 nM palbociclib or DMSO for up to 14 days for colony formation assay. (C) Representative images of the mouse xenograft models and tumors extracted from the four treatment groups after euthanasia. The results were derived from three independent experiments. Data are presented as the mean ± SD. *P<0.05. TNBC, triple-negative breast cancer; HSulf-1, human sulfatase-1; RB, retinoblastoma; EGFP, enhanced green fluorescent protein.

The effects of HSulf-1 overexpression and/or palbociclib treatment on tumor proliferation was next investigated in vivo. MDA-MB-231 cells stably expressing either the control vector or HSulf-1 overexpression plasmid were injected into female nude mice to form xenografts. The xenograft models were treated with 50 mg/kg palbociclib dissolved in 0.5% methylcellulose or vehicle daily for 21 consecutive days. Treatment with palbociclib or overexpression of HSulf-1 alone was found to significantly hinder tumor growth compared with tumors in the Lv105-EGFP + vehicle group. (Fig. 2C). Combined treatment with palbociclib and HSulf-1 overexpres-sion potentiated the reduction in tumor volume compared with that following either interventions alone (Fig. 2C). Collectively, these results suggested that the combination of HSulf-1 over-expression and palbociclib treatment co-operatively inhibited the proliferation of RB-positive TNBC cells both in vitro and in vivo.

HSulf-1 and palbociclib exhibit a cooperative antitumor effect by inducing G1/S cell cycle arrest, inhibiting migration, invasion and epithelial-to-mesenchymal transition (EMT) in RB-positive TNBC in vitro

To investigate the biological effects of HSulf-1 and palbociclib in RB-positive TNBC cells further, the following experiments were performed. Cell cycle analysis indicated that palbociclib treatment significantly inhibited S phase entry, whilst HSulf-1 overexpression increased the % of cells in the G1 and G 2phases of the cell cycle (Fig. 3A). The combination of HSulf-1 overexpression and palbociclib treatment enhanced G1 arrest in RB-positive TNBC cells further compared with either treatments alone (Fig. 3A). Apoptosis analysis revealed that palbociclib treatment had no impact on the apoptosis of TNBC cells (Fig. 3B). By contrast, significantly increased apoptosis was observed in the HSulf-1-overexpressing TNBC cells compared with those transfected with the control vector (Fig. 3B). These results suggested that the cooperative antiproliferative effect of the combinatorial treatment was dependent on the induction of G1/S cell cycle arrest instead of the apoptosis of RB-positive TNBC cells.

Figure 3

HSulf-1 in combination with palbociclib exerts synergistic antitumor effects by inducing G1/S stage cell cycle arrest and by inhibiting migration, invasion and epithelial-mesenchymal transition in vitro. (A-E) Hs578T and MDA-MB-231 cells following transient transfection with the control vector or HSulf-1 overexpression plasmids and incubation with 500 nM palbociclib or DMSO for 72 h. (A) Cell cycle analysis. G1 and S phases of pcDNA3.0-Vector+Vehicle vs. pcDNA3.0-Vector+Palbociclib; G1 and G2 phases of pcDNA3.0-Vector+Vehicle vs. pcDNA3.0-HSulf-1+Vehicle; G1 phase of pcDNA3.0-Vector+Palbociclib vs. pcDNA3.0-HSulf-1+Palbociclib; G1 phase of pcDNA3.0-HSulf-1+Vehicle vs. pcDNA3.0-HSulf-1+Palbociclib. (B) Apoptosis analysis. (C) Wound healing assays. (D) Transwell assays. Magnification, x100. (E) Western blotting was performed to measure E-cadherin, vimentin and N-cadherin expression, with β-actin used as an internal control. The results were derived from three independent experiments. Data are presented as the mean ± SD. *P<0.05. HSulf-1, human sulfatase-1; vec, vector.

The anti-metastatic effect of palbociclib on TNBC remains controversial (48-51). As a result of these contradictory reports, the potential anti-migratory and anti-invasive properties of palbociclib in the two RB-positive TNBC cell lines were investigated. Overexpression of HSulf-1 or palbociclib treatment alone significantly impaired the migration and invasion of Hs578T and MDA-MB-231 cells compared with those treated with vehicle and transfected with the control vector (Fig. 3C and D). By contrast, when HSulf-1 expression was downregulated using shRNAs, it was shown that the migratory and invasive abilities of Hs578T and MDA-MB-231 cells were significantly increased compared with those transfected with the control vector (Fig. S1D and E). Additionally, palbociclib treatment combined with HSulf-1 overexpression significantly enhanced the anti-migratory and invasive effects further compared with those following either treatment alone (Fig. 3C and D). Since EMT is typically the first crucial process in cancer metastasis (52-55), potential changes in the expression of EMT markers was examined. Hs578T and MDA-MB-231 cells treated with palbociclib or transfected with the HSulf-1 overexpression plasmid showed significantly increased expression levels of E-cadherin, along with signifi-cant reductions in those of vimentin and N-cadherin compared with cells in the control vector+vehicle group (Fig. 3E). When the two interventions were combined, a significant synergistic effect was observed (Fig. 3E). Collectively, these data demonstrated the cooperative antitumor effects of combining palbociclib with HSulf-1 overexpression was mediated via the induction of G1/S cell cycle arrest and the inhibition of migration, invasion and EMT in vitro.

HSulf-1 inhibits cell cycle progression by suppressing the expression of cyclin D1 via non-canonical AKT/STAT3 and ERK1/2/STAT3 signaling

Previous studies reported that HSulf-1 downregulated cyclin D1 expression by inhibiting the STAT3 signaling pathway in hepatocellular carcinoma cells (56). In addition, it was reported that constitutive STAT3 activation occurs in >40% of breast cancers, especially in the triple-negative sub-type (31,57). Therefore, to ascertain if HSulf-1 regulated cyclin D1 expression in TNBC cells via the STAT3 pathway, HSulf-1 was overexpressed in TNBC cells, following which the expression levels of the associated critical signaling components were evaluated in vitro. Overexpression of HSulf-1 significantly reduced cyclin D1 expression and STAT3 phosphorylation without affecting total STAT3 expression in TNBC cells compared with those transfected with the control vector (Fig. 4A). Additionally, it was hypothesized that HSulf-1 could inhibit the expression of cyclin D1 through the canonical JAK2/STAT3 pathway. Subsequent measurement of JAK2 phosphorylation and total JAK2 expression in parental and HSulf-1-overexpressing cells showed that neither were significantly unaffected by HSulf-1 overexpression (Fig. 4B).

Figure 4

HSulf-1 inhibits cell cycle progression by suppressing the expression of cyclin D1 via non-canonical AKT/STAT3 and ERK1/2/STAT3 signaling. (A) Western blotting was performed to measure the levels of HSulf-1 and cyclin D1 expression in addition to STAT3 phosphorylation in Hs578T and MDA-MB-231 cells following transient transfection with the control vector or HSulf-1 overexpression plasmids. β-Actin was used as an internal control. (B) Western blotting was performed to measure HSulf-1 expression and JAK2 phosphorylation in MDA-MB-231 cells following transient transfection with the control vector or HSulf-1 overexpression plasmids. GAPDH was used as an internal control. (C) Western blotting was performed to measure the expression of HSulf-1 and the degree of phosphorylated ERK1/2, AKT and STAT3 phosphorylation in MDA-MB-231 cells following transient transfection with the control vector or HSulf-1 overexpression plasmids. GAPDH was used as an internal control. The results were derived from three independent experiments. Data are presented as the mean ± SD. *P<0.05. HSulf-1, human sulfatase-1; JAK, janus kinase; vec, vector; ctrl, untransfected control.

Previous studies have also reported that HSulf-1 inhibited the AKT and ERK signaling pathways, leading to cell cycle arrest, apoptosis, suppression of EMT and suppression of metastasis (34,39,58-61). Additionally, it has been demonstrated that ERK1/2 can directly phosphorylate Ser727 at the C-terminus of STAT3 (62). Malanga et al previously showed that suppression of AKT reduces the phosphorylation and transcriptional activity of STAT3, whereas the opposite effects were observed following AKT activation in non-small cell lung cancer cells (63). Based on these previous observations, it was hypothesized that HSulf-1 may suppress STAT3 activity via the ERK1/2 and AKT signaling pathways. Overexpression of HSulf-1 was found to significantly attenuate ERK1/2, AKT and STAT3 phosphorylation in MDB-MB-231 cells compared with those transfected with the control vector (Fig. 4C). In conclusion, these observations suggest that HSulf-1 suppressed cyclin D1 expression via non-canonical AKT/STAT3 and ERK1/2/STAT3 signaling but not the canonical JAK2/STAT3 pathway, leading to cell cycle arrest and the inhibition of proliferation of TNBC cells both in vitro and in vivo.

HSulf-1 cooperates with palbociclib to induce antiproliferative effects by reducing palbociclib-induced cyclin D1 accumulation

A recently published preclinical study reported that a novel CDK4/6 inhibitor, SHR6390, increased the expression of cyclin D1 in RB-positive MCF7 cells without affecting the RB-negative TNBC cell line MDA-MB-468 (64). To determine if palbociclib can exert a similar effect on RB-positive TNBC cells, cyclin D1 expression was next measured following palbociclib treatment. It was found that following palbociclib treatment, the levels of RB expression and phosphorylation were significantly reduced, whilst that of cyclin D1 expression was significantly increased compared with those treated with vehicle (Fig. 5A). In addition, palbociclib-induced cyclin D1 accumulation was found to be significantly reduced following the overexpression of HSulf-1, leading to significantly lower levels of total RB expression and RB phosphorylation compared with cells transfected with the control vector (Fig. 5A). Finally, the possibility that the aforementioned synergistic effects may be due to a reduction in cyclin D1 expression caused by HSulf-1 overexpression was next considered. Therefore, a rescue experiment was performed on MDA-MB-231 cells, where the overexpression of cyclin D1 abolished the synergistic effects of HSulf-1 and palbociclib treatment on cell proliferation (Fig. 5C and D). Consequently, it was concluded that the mechanism by which HSulf-1 cooperated with palbociclib to exert antiproliferative effects was due to the reversal of palbociclib-induced cyclin D1 accumulation.

Figure 5

HSulf-1 exhibit synergy with palbociclib to exert antiproliferative effects on breast cancer cells by reducing palbociclib-induced cyclin D1 accumulation. (A) Western blotting was performed to measure the protein levels of cyclin D1, p-RB and total RB in Hs578T and MDA-MB-231 cells following transient transfection with the control vector or HSulf-1 overexpression plasmids and incubation with 500 nM palbociclib or DMSO. β-Actin was used as an internal control. (B) Reverse transcription-quantitative PCR was performed to measure mRNA level of cyclin D1 in MDA-MB-231 cells following transient transfection with the control vector or cyclin D1 overexpression plasmids. GAPDH was used as an internal control. (C) Cell Counting Kit-8 and (D) colony formation assays were performed to evaluate cell viability in MDA-MB-231 cells following incubation with 500 nM palbociclib for 14 days, overexpression of HSulf-1, overexpression of HSulf-1 followed by incubation with 500 nM palbociclib for 14 days or co-transfection of plasmids encoding HSulf-1 and cyclin D1 followed by exposure to palbociclib for 14 days. The results were derived from three independent experiments. Data are presented as the mean ± SD. *P<0.05. HSulf-1, human sulfatase-1; RB, retinoblastoma; p-, phosphorylated; t-, total; vec, vector.

Discussion

TNBC is the most heterogeneous sub-type of breast cancer, where gene expression profiling identified six distinct molecular subgroups: Luminal androgen receptor (LAR), mesenchymal stem-like, mesenchymal, two basal-likes and immunomodula-tory (21,65). Due to this heterogeneity, patients with TNBC frequently undergo a variety of clinical courses, where a and diverse set of therapeutic responses can occur (21). Therefore, specific therapeutic strategies should be identified based on these molecular sub-types and clinicopathological features. A number of preclinical studies have reported that the RB-positive LAR subgroup of TNBC is sensitive to CDK4/6 inhibitors (43,47,66). However, the anti-metastatic effect of palbociclib on TNBC remains controversial. Lamb et al (48) revealed that treatment with palbociclib increased migration and mammosphere formation and had the potential to increase the migratory capacity of TNBC cells, which could in turn increase metastasis and cancer recurrence. Conversely, other studies have previously demonstrated that palbociclib reduced the migration and invasion of TNBC cells in addition to suppressing metastasis in a xenograft metastasis model derived from TNBC cells (49-51). In the present study, RB-positive TNBC cell lines were found to be sensitive to palbociclib, which reduced tumor proliferation both in vitro and in vivo. Furthermore, palbociclib was found to inhibit TNBC cell migration, invasion and EMT by upregulating E-cadherin expression whilst downregulating that of vimentin and N-cadherin, thus further validating the antitumor effect of palbociclib on RB-positive TNBC.

In recent years, the roles of HSulf-1 in human cancers have attracted attention. HSulf-1 expression was reported to be downregulated in breast cancers, which is considered an early event in the tumorigenesis of breast cancer (37). Additionally, previous studies have demonstrated that HSulf-1 functions as a tumor suppressor in breast cancer; where it can be applied in predicting clinical outcomes (36,38,39). However, these previous studies aforementioned focused primarily on breast cancer without considering the subtype. It is well documented that breast cancer is a highly heterogeneous disease with a complex mechanism of pathogenesis that exhibits a wide range of clinical behaviours and treatment responses. Therefore, it becomes necessary to validate the role of HSulf-1 in TNBC. The present study was specifically performed on the RB-positive TNBC subtype both in vitro and in vivo. Accordingly, it was found that HSulf-1 expression was downregulated in TNBC tissues and cells compared with their non-cancerous counterparts. Notably, reduced HSulf-1 expression was highly predictive of poor PFS and OS in patients with TNBC. However, no association was observed between HSulf-1 expression and the other clinicopathological indicators tested, including age, tumor grade, tumor size and lymph nodal status in TNBC tissues. Collectively, these results indicate that HSulf-1 is likely to serve as a reliable prognostic biomarker of patient outcomes for TNBC.

Previous studies have demonstrated that HSulf-1 induces cell cycle arrest and inhibits tumorigenesis and angiogenesis in TNBC cells in vivo (36,39). However, the potential effects of HSulf-1 on other physiological processes in TNBC, including apoptosis, migration and invasion, remained unclear. In the present study, it was found that HSulf-1 overexpression inhibited tumor proliferation by inducing G1/S and G2/M cell cycle arrest and apoptosis, in accordance with previous reports (19,45,47). In addition, HSulf-1 was also found to suppress migration, invasion and EMT. Collectively, these results demonstrated that HSulf-1 serves as a tumor suppressor in TNBC. Following the discovery that HSulf-1 reduced cyclin D1 expression to induce G1/S arrest, the hypothesis that HSulf-1 overexpression combined with palbociclib treatment may exhibit synergistic antitumor effects on TNBC cells was investigated. HSulf-1 and palbociclib in combination exerted additive antitumor effects on the induction of G1/S cell cycle arrest, inhibition of migration, invasion and EMT in vitro. Accordingly, this combination may serve as a good alternative therapeutic option for RB-positive TNBC. However, in clinical practice, a suitable vector, including the likes of polyamidoamine dendrimers, transferrin-polyethylene glycol-polyethylenimine and adeno-associated viral vectors, may be required for the effective delivery of the HSulf-1 gene into tumors for patients with TNBC in the future (67-71), which require further clinical validation.

Several studies have previously demonstrated that HBGFs, such as HGF binding to its corresponding receptors c-Met and HSPGs, form a ternary complex to activate downstream tyrosine kinases, resulting in the phosphorylation of substrate proteins, including PI3K-AKT and Ras-MAPK. Desulfation of HSPGs by HSulf-1 prevents the formation of these ternary complexes, followed by the inactivation of downstream signaling pathways (33,34,38,72,73). Therefore, in the context of HBGF signaling, HSPG serves as a core-ceptor to facilitate the interaction between the HBGFs and their cognate transducing receptors, activating the phosphorylation cascade. HSulf-1 inhibits the sulfation of cell-surface HSPGs and abrogates growth factor signaling. Mechanisms underlying the regulation of cyclin D1 by HSulf-1 have been proposed. Liu et al (56) previously documented that HSulf-1 downregulates cyclin D1 expression by suppressing STAT3 signaling in hepatocellular carcinoma. Notably, the CCND1 gene was demonstrated to be a target of STAT3, as STAT3-binding sites were identified in the CCND1 promoter (74). In the present study, HSulf-1 was identified to regulate cyclin D1 expression via non-canonical AKT/STAT3 and ERK1/2/STAT3 signaling instead of the canonical JAK2/STAT3 pathway. In addition, overexpres-sion of HSulf-1 prevented the apparent increase in cyclin D1 levels caused by palbociclib treatment. Therefore, data from the present study hinted at a potential crosstalk between the HSulf-1/ERK1/2/AKT/STAT3 and CDK-RB-E2F pathways to control cell cycle progression (Fig. 6).

Figure 6

Schematic diagram depicting the possible mechanism of HSulf-1-mediated inactivation of STAT3 via the PI3K/AKT and Ras/ERK1 pathways and crosstalk with CDK-RB-E2F signaling in TNBC cells. The initiation of HBGF signaling requires the binding of extracellular HBGF (FGF, EGF, HGF) and its cognate tyrosine kinase receptor to HSPGs to form a ternary ligand-HSPG-receptor complex at the cell surface. Ligand binding induces receptor dimerization and subsequent phosphorylation of the receptor tyrosine kinase, leading to a downstream phosphorylation cascade including the PI3K/AKT/STAT3 and Ras/ERK1/2/STAT3 pathways. Phosphorylated STAT3 forms a dimer and translocates into the nucleus, promoting cyclin D1 expression. Desulfation of HSPGs by HSulf-1 prevents formation of the ternary complex, leading to a reduction in HBGF signaling and cyclin D1 expression. Hsulf-1, human sulfatase 1; CDK, cyclin-dependent kinase; RB, retinoblastoma; YNBC, triple-negative breast cancer; HBGF, heparin-binding growth factors; HSPG, heparan sulfate proteoglycans; MCM7, minichromosome maintenance complex component 7; cdc6, cell division cycle 6; EGF, epidermal growth factor; FGF, fibroblast growth factor; HGF, hepatocyte growth factor.

A number of preclinical studies have reported that ER-positive breast cancer cells adapt rapidly to CDK4/6 inhibition to evade cytostasis, in part via non-canonical cyclin D1-CDK2-mediated S phase entry (75). This adaptation was found to be prevented by combining CDK4/6 inhibitors with PI3K/mTOR inhibitors, which reduced the levels of cyclin D1 and other G1-S phase cyclins (75,76). In the present study, it was also found that the levels of cyclin D1 were increased following palbociclib treatment, consistent with previously reported results (64). This may be due to an adaptive response to damaging stimuli or CDK4/6 inhibitors have a similar negative feedback effect of preventing CDK4/6 to form a complex with cyclin D1, which may increase cyclin D1 expression and further activate the crosstalk between the cyclin D1-CDK2 signaling pathway, which require further study (77). It was previously revealed that STAT3 phos-phorylation was significantly increased in palbociclib-resistant ER-positive breast cancer cells, where ER-positive metastatic breast cancer patients acquired resistance after treatment with Palbociclib (78). Furthermore, palbociclib-resistant cells appear to be more sensitive to the STAT3 inhibitor TTI-010, which directly targets the tyrosine-phosphorylated peptide-binding pocket within the STAT3 SH2 domain (78). Based on these observations, HSulf-1 may have the potential to prevent the early adaptation and prolong the durability of palbociclib efficacy in TNBC patients.

Although the present study demonstrated that HSulf-1 over-expression in conjunction with palbociclib has a synergistic antitumor effect in RB-positive TNBC cells. Only the synergistic anti-proliferative mechanism by which HSulf-1 reduced palbociclib-induced cyclin D1 accumulation was clarified. Additionally, although it was demonstrated that HSulf-1 inhibited cyclin D1 expression through the ERK1/2/STAT3 and AKT/STAT3 pathways, it remains unclear whether HSulf-1 could suppresses cyclin D1 expression through other pathways. For the wound healing experiments, it was found that the majority of cells died after 24 h incubation in medium without FBS following scratching a wound introduction. Therefore, medium containing 2% FBS was used, where little or no cell proliferation or cell death were observed. Accordingly, further studies are required to explore these possibilities.

In conclusion, the present study demonstrated that HSulf-1 expression is downregulated in TNBC tissues and cells, which serves an anti-oncogenic role by inducing cell cycle arrest and apoptosis in addition to preventing proliferation, EMT, migration and invasion in TNBC. In addition, it was found that downregulation of HSulf-1 was associated with poor clinical outcomes in patients with TNBC. HSulf-1 inhibited cyclin D1 expression via the non-canonical AKT/STAT3 and ERK1/2/STAT3 signaling pathways in TNBC. In particular, the combination of HSulf-1 overexpression and treatment with the CDK4/6 inhibitor palbociclib exerted additive, synergistic antitumor effects on RB-positive TNBC, which may serve as an effective alternative therapeutic option for patients with RB-positive TNBC.