The breast cancer cell line MCF7 was cultured in DMEM/high glucose (SH30243.01B, HyClone, Logan, UT) with 10% fetal bovine serum (FBS) (04-001-1ACS, Biological Industries, Kibbutz Beth Haemek, Israel) and 1% penicillin/streptomycin (p/s) (C0222, Beyotime Institute of Biotechnology, Jiangsu, China). SKBR3, HCC1806 and 786-0 cells were cultured in RPMI-1640 (SH30809.01B, HyClone) containing 10% FBS and 1% p/s. MDA-MB-231 cells were cultured in DMEM/F12 (SH30023.01B, HyClone), supplemented with 10% FBS and 1% p/s. The immortalized human breast epithelial cell line MCF10A was maintained in DMEM/F12, supplemented with 5% horse serum (16050-130, Gibco, New Zealand), 20 ng/mL EGF (PHG0311, Invitrogen, Carlsbad, CA), 0.5 mg/mL hydrocortisone (MB1567, Meilunbio, Dalian, China), 100 ng/mL cholera toxin (C8052, Sigma-Aldrich, St. Louis, MO), 10 μg/mL insulin (Wanbang Biopharmaceuticals, Xuzhou, China), and 1% p/s. The cells were purchased from Conservation Genetics CAS Kunming Cell Bank. Cell lines were tested to be mycoplasma-free by PCR (16).

For siRNA experiments, cells were transfected with two TICRR-specific siRNA and control siRNA (RiboBio, Guangzhou, China) at a final concentration of 50 nM using riboFECT™ CP Transfection Kit (C10511-1, RiboBio) according to the manufacturer's instructions. The sequences of the siRNA were listed in Table S1. For generation of stable cell population, the shRNA targeting TICRR was cloned into pLKO.1 lentiviral vector. The lentiviruses were generated from HEK-293T cells and collected at 48 h and 72 h after transfection. Then cells were infected with lentiviruses, and selected in the presence of puromycin for three generations.

Total RNA were extracted using TRIzol reagent (11667165001, Invitrogen), followed by treatment of DNase I (EN0521, Thermo scientific, MA, USA). Reverse transcription was performed with oligo (dT) primers using GoScript^TM reverse transcription system (A5001, Promega, Madison, WI) according to the manufacturer's protocol. Quantitative real-time PCR with gene-specific primers was performed using GoTaq® qPCR Master Mix (A6002, Promega). The comparative C_T method was applied for quantification of gene expression, and values were normalized to beta actin (ACTB). The primers used in the study were shown in Table S2.

Cells were lysed with RIPA Lysis Buffer with PMSF (ST506, Beyotime Institute of Biotechnology). Proteins were quantified by BCA Protein Assay Kit (P0010, Beyotime Institute of Biotechnology). 20–80 μg of total protein was loaded onto SDS-PAGE and subsequently transferred to PVDF membranes (162-0177, Bio-Rad, Richmond, CA). After blocking with 5% nonfat milk or bovine serum albumin (BSA) in PBST, membranes were incubated with the following primary antibodies at the suggested dilutions: anti-CDKN1A/p21 (A1483, ABclonal, Cambridge, MA), anti-TP53/p53 (sc-6243, Santa Cruz Biotechnology, Santa Cruz, CA), anti-ERK1/2 (A0229, ABclonal), anti-p-ERK1/2 (AP0472, ABclonal), anti-p-Histone H2A.X (S139) (AP0099, ABclonal), anti-p-ATM (10H11.E12) (sc-47739, Santa Cruz Biotechnology), anti-p-CHK1 (Ser345) (sc-17922, Santa Cruz Biotechnology), anti-p-CHK2 (Thr68) (sc-16297-R, Santa Cruz Biotechnology), anti-PUMA (sc-374223, Santa Cruz Biotechnology), anti-TICRR (NBP2-41283, Novus Biologicals, Littleton, CO), anti-cyclin D1 (2922S, Cell Signaling Technology, Danver, MA) and anti-ACTB (AA128, Beyotime Institute of Biotechnology). Primary antibodies were detected with HRP-labeled goat anti-rabbit (A0208, Beyotime Institute of Biotechnology) or anti-mouse IgG (H+L) (A0216, Beyotime Institute of Biotechnology).

Cells were seeded in 96-well plates at a density of 3 × 10³ cells. Then, cell viability were quantified by CellTiter 96® AQueous One Solution Cell Proliferation Assay (G3580, Promega) with 490 nm plate reading at indicated time points according to the manufacturer's protocol. Cells were seeded into 96-well plates and DNA synthesis was measured using Cell-Light™ EdU Apollo®488 in vitro Imaging Kit (100T) (C10310-3, RiboBio) following the manufacturer's protocols. In brief, cells were labeled with 50 μM EdU for 2 h, then fixed in 4% paraformaldehyde for 30 min, then stained with Apollo®488 and Hoechst 33342. Cells were imaged by Nikon eclipse Ti inverted microscope.

For cell cycle analysis, cells were collected after transfection, then washed and fixed with cold 75% alcohol overnight. After wash with PBS, cells were labeled with propidium iodide (PI) (P4170-10, Sigma-Aldrich) and incubated at room temperature in the dark for 30 min. Cells were then filtered through a nylon mesh filter and subjected to flow cytometry (BD Biosciences). For cell apoptosis analysis, cells were harvested at 48 h after transfection, and stained using the FITC-Annexin V apoptosis detection kit and PI staining solution (88-8005-72, eBioscience, San Diego, CA) according to manufacturer's protocol. FACS (fluorescence activated cell sorter) analysis was performed within 4 h and the results were analyzed by FlowJo software (Version 7.6.1).

MCF7 cells were transfected with TICRR-specific siRNA and control siRNA for 48 h, then labeled with 50 μM IdU (I7125, Sigma-Aldrich) for 20 min. DNA fiber spreads were performed as previously reported (17, 18). Briefly, cell were harvested and re-suspended with cold PBS. 2.5 μL of the cell suspension was spotted onto a glass slide and mixed with 7.5 μL lysis solution (0.5% SDS, 50 mM EDTA, 200 mM Tris-HCl). Slides were then tilted to 15° to allow the fibers to spread. Fibers were air-dried and fixed in methanol and acetic acid (3:1) and subsequently acid treated with hydrochloric acid (2.5 N) to denature the DNA fibers. Later, slides were stained with immunofluorescent anti-BrdU (347580, BD Biosciences, San Jose, CA). Slides were imaged at 60 × using an Olympus FluoView™ FV1000. Fiber length was analyzed using Image-Pro Plus software (Version 6.0). Table S3 shows the number of fibers and independent experiments performed under each condition. The median and mean of replication tract length and p-values derived from the Welch's two-tailed t-test were calculated using R software.

For colony formation assay, cells after transfection were seeded at 500 cells per well in a 6-well plate and incubated for 15 d. Cells were fixed with fixative (methyl alcohol: glacial acetic acid = 3:1) for 15 min and then stained with 0.1% crystal violet for 20 min. The colony formation rate was calculated as colony number/cell number seeded.

Cell migration was evaluated by wound healing and transwell assays. For the wound healing assay, transfected cells were seeded in 6-well plates and cultured to 90% confluence. Cells were scraped with a 200 μL tip and washed with PBS for 3 times, and then cultured in fresh medium containing 2% serum for 24 h. Photographs were took at 0 h and 24 h. The width was measured with Image-Pro Plus software. The relative migration in TICRR knockdown cells was normalized to the control cells. For transwell migration assay, 24-well polycarbonate inserts were used. After transfection, cells were cultured on the top chamber of 24-well transwell plate (3422, Corning, Glendale, AZ) in 2% FBS medium and medium with 20% FBS was added into the bottom chambers. After 24 h, the cells on the surface of top chamber membrane were removed with a cotton swab. The migrated cells on the bottom surface of chamber membrane were fixed with 4% paraformaldehyde for 20 min, stained with 0.1% crystal violet for 20 min, washed with PBS and air dried. The crystal violet was dissolved with 500 μL 33% acetic acid, and the OD570 nm value was recorded.

Tumor xenografts were performed by injecting shControl-HCC1806 cells and shTICRR-HCC1806 cells (1.5 × 10⁶ cells per 100 μL DMEM with 30% BD matrigel) into subcutaneous of 5-week-old female BALB/c nude mice (Vital River, Beijing, China). Once tumors were detectable, the mice were monitored and the tumor volumes (V) were measured twice a week by determining length (L) and width (W) using Vernier calipers and calculated using the formula: V = L × W²/2. One month later, mice were sacrificed, and the tumors were excised for mass measurement and imaging. The mouse experiment was approved by the animal ethics committee of the Kunming Institute of Zoology, Chinese Academy of Sciences.

Gene expression data and clinical data of 15 cancer types were downloaded from The Cancer Genome Atlas (TCGA, https://tcga.xenahubs.net). Overall survival (OS) analysis was performed in 15 cancer types using a web-tool OncoLnc (http://www.oncolnc.org) (19), and the disease free survival (DFS) was analyzed using R software. The samples were grouped by the median of TICRR expression. For survival analyses, log-rank tests were used to determine the statistical significance.

Total RNA was isolated using TRIzol reagent and RNA sequencing was completed by Novogene Company. The process of data analysis was described previously (20).

Data were expressed as means ± SEM. Statistical analyses were performed using GraphPad Prism 7.0 (GraphPad Software) or R software (version 3.3.2). P < 0.05 were considered statistically significant.

In agreement with our previous observation based on a newly developed algorithm (15), re-analysis of the TICRR expression data extracted from TCGA database did show that this gene mRNA level was strikingly up-regulated in all of the 15 cancer types compared to the normal tissues (all p < 1.00 × 10⁻³) (Figure 1). Based on the expression profile, we tested TICRR expression using quantitative real-time PCR assay (RT-qPCR) in several cell lines, including the immortalized human breast epithelial cell line (MCF10A), four breast cancer cell lines (MCF7, SKBR3, HCC1806, and MDA-MB-231), and one kidney cancer cell line (786-0). The results showed that TICRR displayed the highest expression in MCF7 cells (p = 3.30 × 10⁻³, compared to MCF10A), followed by HCC1806 (p = 0.017) (Figure S1A). At the protein level, TICRR expression in MCF7 cells was also higher than that in MCF10A cells (Figure S1B).

To investigate the role of TICRR in cancer cell growth, shRNA and siRNA mediated TICRR knockdown were performed in MCF7 cells (Figure 2A). Knock down efficiency was validated in other cell lines (Figure S2A). TICRR silencing caused 20–30% inhibition of proliferation in MCF7 cells compared to the control cells at 48 h after transfection with siRNA (all P < 0.05) (Figure 2B). Cell growth curves of MCF7 and HCC1806 (Figures 2C,D), SKBR3, MCF10A, and 786-0 cells (Figures S2B,C) consistently showed the inhibitory effect of TICRR knockdown on cell growth, with 30~65% inhibition at 96 h (all P < 0.05). We further examined cell apoptosis to test whether the reduced cell viability by TICRR knockdown is attributed to cell death. Results showed that the percentage of apoptotic cells was significantly increased in TICRR-knockdown cells (p = 8.70 × 10⁻³, Figure S3), suggesting enhancement of cell death through apoptosis. We further examined the ability of TICRR to regulate colony formation. Foci number in TICRR-knockdown cells was significantly reduced in MCF7 (p = 0.02) (Figure 2E), which was validated in HCC1806 cells (p = 0.04) (Figure 2F). Furthermore, we found that TICRR depletion significantly inhibited cancer cell migration in vitro as revealed by the wound healing (Figure S4A) and transwell assays (Figure S4B).

To explore the role of TICRR in affecting tumor growth in vivo, we injected shTICRR- and shControl-HCC1806 cells into BALB/c nude mice and found that the growth of xenograft tumors was strikingly slower in mice injected with shTICRR-HCC1806 cells (tumor volume on the last day: control = 538.4 ± 49.68 mm³, TICRR-depletion = 119.4 ± 14.56 mm³, p = 2.12 × 10⁻⁶) (Figures 3A,B). Consistently, the tumor weight in the shTICRR group (0.13 ± 0.02 g) was dramatically reduced than that in the control (0.54 ± 0.06 g) (p = 1.43 × 10⁻⁵) (Figure 3C). The TICRR protein was still lower in the shTICRR tumors compared to the control (Figure 3D). These results collectively support a critical role of TICRR in tumorigenesis in vivo.

Given the key role of TICRR in DNA replication initiation as well as the observation that alteration of TICRR level changed DNA initiation (12, 13, 21), we proposed that the reduced viability and tumor growth induced by TICRR knockdown are attributable to the deficiency of DNA replication. Here we found that TICRR-depletion led to significantly reduced percentage of MCF7 cells labeled with EdU (p = 2.61 × 10⁻⁴) (Figure 4A), a molecule combining to DNA during S phase of cell cycle, suggesting DNA synthesis likely to be inhibited. Then, we directly labeled replicating DNA with green fluorescent Iodo-deoxyuridine (IdU) (Figure 4B) and found that the inter-origin distance (IODs) in TICRR-depleted cells (mean: 39.63 μm) was significantly lengthened compared with the control (mean: 31.14 μm; p = 2.95 × 10⁻¹⁰) (Figure 4C; Table S3), indicating that TICRR depletion suppressed the initiation of DNA replication origins.

Interestingly, we also observed that the median of IdU tract length in TICRR knockdown cells (mean: 4.01 μm) was significantly shortened than that in the control (mean: 4.47 μm; p = 1.87 × 10⁻¹¹) (Figure 4D; Table S3), indicating a stalled replication fork progression after TICRR depletion. These results collectively suggest that silencing TICRR affects not only the initiation of DNA replication but also the progression of replication fork, thus significantly impairing the DNA synthesis.

Since TICRR functions to activate the origins from G1/S transition to S-phase (22) and regulate S-phase duration (12), cell cycle analysis was performed in TICRR knockdown cells. We found that TICRR silencing induced significant arrest of cell cycle at G0/G1-phase in MCF7 cells with increased percentage of G0/G1-phase cells (p = 1 × 10⁻⁴) (Figure 5A; Figure S5A). The results were validated in SKBR3 (p = 6.9 × 10⁻³) and MCF10A cells (p = 7.8 × 10⁻³) (Figures S5A,B). We also found that TICRR depletion decreased the expression of Cyclin D1, which controls cell cycle progression through the G1-S checkpoint, in MCF7 (Figure 5B), SKBR3 cells (Figure S5C), HCC1806 cells and tumors harvested from mice (Figure S5D). Moreover, the level of phosphorylated ERK1/2, a protein involved in the regulation of G1/S transition and DNA replication (23), also decreased in TICRR knockdown MCF7 cells (Figure 5C), indicating that cell cycle was interrupted at S phase entry.

To disclose genes and/or pathways involved in the arrestment, we obtained the transcriptomes of MCF7 cells transfected with TICRR-specific siRNA and the control after 48 h via RNA-sequencing. For the obtained gene expression data, mRNA levels of several randomly selected genes were estimated and confirmed to be replicable via RT-qPCR (Figure S6A). Our analysis showed that 53 genes were up-regulated and 225 genes down-regulated with fold change >2 (Figure 6; Table S4). As expected, gene ontology (GO) analysis revealed that the DEGs were significantly enriched in cell cycle and its regulation processes, such as cell division, DNA replication, G1/S and G2/M transition of mitotic cell cycle (Figure 7A; Table S5). KEGG pathway analysis also showed cell cycle and DNA replication as the most significant enrichments (Figure 7B; Table S6).

We further analyzed the mRNA levels of genes that directly interact with TICRR in DNA replication initiation process and found that MCM2, MCM5, and TOPBP1 expression decreased significantly after TICRR silencing (all P < 0.05), while CDC45 and MCM7 were not changed in MCF7 cells (Figure S6B). However, the genes CDC45, MCM2, MCM5, and MCM7, were found to be consistently down-regulated in SKBR3 cells and HCC1806 cells (all P < 0.05) (Figure S6C). These results altogether showed that the depletion of TICRR greatly affected the expression of cell cycle-related genes, especially those involved in DNA replication, and thus led to cell cycle arrest and subsequently inhibition of cell proliferation.

We also noticed that p53 signal pathway was significantly enriched in KEGG analysis (p = 7.25 × 10⁻⁶) (Figure 7B; Table S6), suggesting the p53 signal pathway likely be activated. The transcriptomic data revealed the downstream genes of p53, such as CDKN1A, CCNE2, CCNB1/2, CDK1, BID, and BBC3, were altered (Figure S7A), which may contribute to G1/G2 arrest and cell apoptosis (24). We then determined the protein levels of p53 and found that TICRR depletion enhanced the expression of p53 in MCF7 cells (Figure 8A). Indeed, at the translation level, CDKN1A/p21, the main regulator of cell cycle from G1 to S transition (25, 26), and the protein regulating cell apoptosis, BBC3 (27), were also up-regulated in TICRR-depleted MCF7 cells (Figure 8A). We also found the genes CCNE2, CCNB1/2, CDK1, and BID to be down-regulated and p21 and BBC3 to be up-regulated in HCC1806 which was reported as a p53-null cell line (Figure S7B and Figure 8B). In SKBR3, a p53 mutant cell line, we also found that knockdown of TICRR can increased p21 and BBC3 (Figure S7C), suggesting alternative way may function, such as p73 that can replace p53 to regulate cell proliferation and apoptosis in p53-null cells (28), which have yet to be determined.

As the impaired fork progression of DNA replication can lead to DNA damage (2, 29, 30), the known activator of p53 pathway (25), it is therefore possible that the activation of the p53 pathway in TICRR-depleted MCF7 cells is triggered by DNA damage response (DDR). Indeed, we found that the protein level of phosphorylated histone H2AX, a marker of DNA damage (31), was increased in TICRR depletion cells (Figure 8C). Further, one of the key kinases involved in DDR, ataxia telangiectasia mutated (ATM), was examined. Our results showed that the level of phosphorylated ATM was up-regulated in TICRR-depleted MCF7 cells (Figure 8D). Accordantly, the level of phosphorylated checkpoint kinases 2 (p-CHK2) was also increased (Figure 8D), indicating that both ATM and CHK2 were activated. The activated ATM-dependent p53 signaling induced by TICRR knockdown was validated in MCF10A cells (Figure S8A). We also determined but failed to detect the expression of the p-CHK1, the downstream protein of another DDR kinase ATR, no matter in TICRR-depletion or in the control cells (Figure S8B). Even under the treatment of cisplatin, a DNA damage inducer, p-CHK1 protein was observed to be increased in the control cells but not in TICRR knockdown cells (Figure S8B), suggesting that it is ATM/CHK2, rather than ATR/CHK1, which was activated in TICRR-depleted MCF7 cells.

The obtained evidence collectively lends support to a critical role of TICRR in tumorigenesis through regulating DNA replication in cancer cells, it is then possible that this gene would have great potential in clinical prognosis. We examined the association between TICRR expression and survival of cancer patients using the TCGA clinical data. The result revealed that the patients with higher TICRR expression (more than the median value) showing significantly poorer overall survival (OS) than those with lower expression in breast cancer (p = 0.042) (Figure 9A). We further determined the association between TICRR expression and distinct subtypes of breast cancer, classified by estrogen receptor (ER), progesterone receptor (PR), and human epidermal growth factor receptor 2 (HER2) expression patterns (32, 33). The results showed TICRR displayed the highest expression in triple-negative breast cancer (TNBC, ER/PR-negative, and HER2-negative), followed by HER2-positive breast cancer (HER2, ER/PR-negative, and HER2-positive) (Figure 9B), which are both associated with shorter survival time (32). Similar correlations were observed in the other cancer types, including KIRC, KIRP, LIHC, and LUAD (all P < 0.05) (Figure 9C). Besides, we also found that higher TICRR expression was associated with lower disease free survival (DFS) in LIHC, KIRP, UCEC, and PRAD (all P < 0.05) (Figure 10).

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.