High Expression of RECQL Protein in ER-Positive Breast Tumours Is Associated With a Better Survival

Background RECQL (also known as RECQ1 and RECQL1) is a gene of recent interest in breast cancer and an association between high levels of RECQL protein in breast cancer tumour cells and good survival of patients has been reported. Methods To validate this association, we measured the RECQL protein levels in tumours of 933 breast cancer patients using immunohistochemistry analysis and followed the patients for death from breast cancer. Results Women with a level of RECQL protein above the 75th percentile had better 15-year disease-specific survival among ER-positive patients (62.5% vs. 48.7%, HR= 0.72, 95%CI= 0.52-0.98, p-value = 0.04), but not among ER- patients (48.9% vs. 48.0%, HR= 1.07, 95%CI= 0.67-1.69, p-value= 0.79). Among the ER-negative patients, high RECQL protein levels were associated with better survival among women who received tamoxifen treatment (67.0% vs. 51.5%, HR= 0.64, 95%CI= 0.41-0.99, p-value= 0.04). Conclusion RECQL might be a new predictive marker for tamoxifen treatment among ER-positive patients.


INTRODUCTION
Altered expression levels of several genes in breast cancer predict patient prognosis. A correlation has been observed between poor breast cancer prognosis and lower levels of RECQL (also known as RECQ1 and RECQL1) mRNA or RECQL protein expression (1,2). RECQL is the smallest and most abundant member of the RecQ family of DNA helicases (3,4). It has two main domains; the core helicase domain, involved in ATP binding and hydrolysis, and the RecQ C-terminal domain (RQC) which play an essential role in the unwinding of DNA (4)(5)(6). RECQL performs its helicase activity in an ATP-dependent manner in a 3' to 5' direction (7). As a helicase, it has many essential functions in DNA replication, such as maintaining the DNA replication fork progression and restarting stalled replication forks (8)(9)(10)(11). It is also involved in maintaining genome stability, double-strand break repair, and telomere maintenance (4,6,(12)(13)(14). In addition to its roles as a DNA helicase, RECQL is involved in branch migration of Holiday junctions and strand annealing (7). The depletion of this helicase could impair normal cellular function and lead to increased DNA damage and compromised genome stability.
In addition to the cancer susceptibility role, RECQL was suggested to be a prognostic marker. In a study by Arora et al. low RECQL mRNA and protein levels in breast tumours were associated with poor breast cancer prognosis (1). This correlation was observed in a second study which showed low RECQL protein levels were correlated with poor survival (HR: 2.12, p-value: 0.015) (2). Furthermore, in a recent study, the cellular mechanism by which RECQL affects ER-positive cells was proposed, wherein RECQL in cooperation with FOXA1, directly regulates the expression of the ESR1 gene which encodes the ERa protein (28).
To investigate this association further, we measured the RECQL protein levels in tumours from 933 breast cancer patients by immunohistochemistry (IHC) and analyzed their 15-year survival.

Study Population
This study was conducted by analyzing patients from the Banting study (29). Total of 1,601 breast cancer patients were diagnosed and enrolled between 1987 to 1999 in the Banting study and their therapy reflected the commonly used treatments during that time period. Data files and tumour samples from 933 breast cancer patients were available and analyzed in this study. Patients were followed from the date of diagnosis to the date of death from breast cancer or the date of last follow up (if alive). The average follow-up was 12.1 years (range: 0 to 31.3 years). The age at diagnosis ranged from 24 to 93 (mean 55.4 years). 55.1% of subjects were post-menopausal at diagnosis and 77.0% were ER-positive. The majority of ER-positive patients received tamoxifen (60.6%). No other endocrine therapy was used.

Tissue Microarray and Immunohistochemistry
Tissue microarrays (TMA) were made from 0.6-mm cores sampled from the formalin-fixed paraffin-embedded tumour blocks. Each TMA contains three cores from each tumour block. Immunohistochemical staining for the RECQL protein was conducted based on a previously described method (30)(31)(32)(33) by using a combination of Thermo Scientific Shandon Sequenza chamber system (REF: 72110017), Novolink Max Polymer Detection System (RE7280-K: 1,250 tests), and the Leica Bond Primary Antibody Diluent (AR9352), according to the manufacturer's instructions (Leica Microsystems). The slides were dewaxed and dehydrated by Leica Autostainer XL machine. TMA sections were pretreated with sodium citrate buffer (pH 6.0) and heated for 20 minutes at 95°C in a microwave (Whirlpool JT359 Jet Chef 1000W) for antigen retrieval. Each set of slides was incubated with primary anti-RECQL antibody (Bethyl Laboratories, catalog no. A300-450A) at a dilution of 1:1,000 for 60 minutes. For each run, positive and negative controls were included. Negative control was utilized to reassure all the staining was because of specific antibody-antigen interaction. The negative control slide had only breast tissue without adding any antibody to it. The positive control slide was a liver tissue with known expression of RECQL to control for the reactivity of the RECQL antibody and also breast cancer tissue stained with an antibody for b-globulin to control for the reactivity of immunohistochemistry enzyme.

IHC Evaluation
After scoring whole-field inspection of the cores, intensities of the nuclear stainings were classified (0 = no staining, 1 = weak staining, 2 = moderate staining, and 3 = strong staining; Supplementary Figure 1). For each intensity classification, the percentage of the stained nuclei was estimated. An H-index (range 0-300) was calculated by multiplying the staining intensity score by the percentage of stained nuclei for each core. For each tumour sample, the median H-index of the three core replicates was used for the data analysis. The distribution of the median H-index for the entire cohort was shown in Supplementary Table 1. The mean and median of the coefficient of variation (CV) of H-index for the three cores across all samples were 10.4% and 6.9%, respectively.

Statistical Analysis
Subjects were divided into two groups based on their RECQL protein levels ( Table 1). We considered a high RECQL level to be one in the highest quartile (n = 205) and a medium/low RECQL level to be one in the bottom three quartiles (n = 728). This classification had the best performance in terms of distinguishing patients with poor from good survival compared to using the first quartile or median as the cut-off point. Student t tests and Fisher exact tests were used as appropriate. Estimation of cumulative survival probabilities was conducted by the Kaplan-Meier  Table 2). Multivariate survival analysis was conducted separately for ERpositive and ER-negative patients. Among ER-positive patients a higher 15-year survival rate was observed for those with a high level of RECQL protein (62.5% vs. 48.7%, HR = 0.72, 95%CI = 0.52-0.98, p-value = 0.04) ( Figure 1B; Table 2). Among the ER-negative patients no difference was seen (48.9% vs. 48.0%, HR = 1.07, 95%CI = 0.67-1.69, p-value = 0.79) ( Figure 1C; Table 2).
In terms of 15-year survival, the benefit of tamoxifen treatment was greater in ER-positive patients with a high   (Figure 3; Table 3).

DISCUSSION
We analyzed the association between RECQL protein levels and breast cancer survival among 933 breast cancer patients A B  diagnosed from 1987 to 1999 in Toronto, Canada. Among unselected breast cancer patients medium/low levels of RECQL protein were associated with inferior disease-specific survival. Additional analysis revealed that this association was seen only among ER-positive patients, in particular among those who received endocrine therapy (Tamoxifen). Together, these data show that RECQL protein level is a prognostic factor for realizing a benefit from endocrine therapy.
Three other studies have investigated the association between RECQL protein or mRNA levels and breast cancer survival (1,2,34). The first study showed that in separate cohorts of 848 and 1977 breast cancer patients, lower protein and mRNA levels of RECQL were associated with worse prognosis; further analysis revealed that this association only held among the ER-positive patients (1). The second study also observed that among 774 breast cancer patients, lower mRNA levels of RECQL were associated with poor distant recurrencefree survival (HR: 2.77, p-value <0.001) and disease-specific survival (HR: 3.10, p-value <0.001). In a cohort of 322 breast cancer patients, low RECQL protein levels correlated with poor survival (HR: 2.12, p-value: 0.015); however, the authors did not compare this association among ER-positive versus ERindividuals (2). These studies also showed that lower RECQL protein levels were associated with poor clinicopathological characteristics (1,2). We also observed that medium/low RECQL levels were associated with poor clinicopathological characteristics, such as a higher proportion of lymph nodepositive tumours (56.1% vs. 47.9%, p-value= 0.05) and larger tumour sizes (27.3 vs 24.2, p-value= 0.01). However, other clinicopathological characteristics were better among the medium/low RECQL patients of our cohort; medium/low RECQL levels were associated with a larger proportion of ERpositive tumours (79.6% vs. 69.0%) and a smaller proportion of triple-negative tumour (12.3% vs. 17.9%), while there was no difference between the medium/low versus high RECQL patients in HER2 status, tumour stage, tumour grade, and tumour histology. Lastly, the third study was only focused on RECQL mRNA levels (34). They observed that a higher expression of RECQL mRNA was correlated with shorter relapse-free survival (RFS) (HR: 1.28, p-value <0.001, n= 3955) and post-progression survival (PPS) (HR: 1.32, p-value: 0.027, n= 414) in all breast cancers. However, higher expression of RECQL mRNA did not affect overall survival (OS) (HR: 1.04, p-value: 0.74, n= 1402) or distant metastasis-free survival (DMFS) (HR: 1.06, p-value: 0.57, n= 1747). The current body of evidence suggests that higher RECQL protein levels are associated with higher survival rates among breast cancer patients with ER-positive tumours.  The cellular mechanism of RECQL effect on the prognosis of ER-positive tumours might be explained by a recent study that explored the role RECQL plays in regulating ERa expression. According to this newly published study, in a helicase dependant manner, RECQL in cooperation with FOXA1 increases the chromatin accessibility at the regulatory site of the ESR1 gene (the gene encoding ERa), and a group of other ERa target genes (28). Therefore, higher levels of RECQL protein would increase ERa expression and its downstream effect. This is an important observation in breast cancer biology for two main reasons. First, higher ERa levels are associated with a better prognosis because ERa inhibits tumour invasiveness (35)(36)(37)(38). Second, higher levels of ERa would enable a more desirable response to endocrine therapy (38)(39)(40)(41). Therefore, higher RECQL level directly increases ERa level, and in turn could reduce tumour invasiveness and improve the response to endocrine therapy. Unfortunately, we did not have ER expression values in our cohort to compare with the RECQL data to confirm their positive correlation. The relationship between RECQL levels and ER expression, tumour invasiveness as well as endocrine treatment efficiency could be the subject of further research. More importantly, future studies should investigate how to induce RECQL expression in ERpositive patients to improve their prognosis, especially in patients with lower RECQL protein levels. The findings of such studies could have crucial clinical implications, especially considering that more than 70% of breast cancer cases are ER-positive (35,42).
One question that remains to be answered is why patients with lower RECQL levels had lower survival, while lower RECQL levels result in lower levels of ERa, and as a result the mitogenic effects of estrogen and ERa should be reduced. Three factors should be considered in answering this question. First, lower ERa levels would result in higher invasiveness, reducing the survival rate (35,36). Second, we observed a larger tumour size in patients with lower levels of RECQL protein, so either lower RECQL levels do not cause enough reduction in ERa levels to dampen the mitogenic effects of ERa, or there are other factors involved that not only compensated for the reduced mitogenic effects of ERa due to its reduction, but caused increased mitogenic effects and a larger tumour size. Third, it is likely that RECQL also impacts breast cancer patient survival through non-ERa dependent effects and through its role in maintaining the chromosomal stability and DNA damage response. Additional studies should be conducted to investigate if such factors exist, and if they do, what the mechanism by which they act as mitogens is, and what implications do they have for breast cancer and possibly other neoplasms.

CONCLUSION
We have shown that higher RECQL protein levels are associated with improved ER-positive breast cancer-specific survival and better response to endocrine therapy with tamoxifen. Therefore, RECQL could be a prognostic and predictive candidate biomarker in ER-positive breast cancer patients responding to tamoxifen's endocrine treatment.

DATA AVAILABILITY STATEMENT
The original contributions presented in the study are included in the article/Supplementary Material. Further inquiries can be directed to the corresponding author.

ETHICS STATEMENT
This study has been approved by the ethics board committee at Women's College Hospital. The patients/participants provided their written informed consent to participate in this study.

AUTHOR CONTRIBUTIONS
AM wrote the manuscript draft, AS did the laboratory assays, PS and VG did the statistical analysis, CC, SN-M, SS, SM, SN, and MA worked together in conceptualizing the study idea, contributing laboratory and biospecimen resources required for the study and writing the manuscript, AAS reviewed the IHC slides and helped with the revision of the manuscript. All authors contributed to the article and approved the submitted version.

FUNDING
The study was funded by Canadian Institute for Health Research, grant# 152939.

ACKNOWLEDGMENTS
We acknowledge the pathology lab at Sunnybrook Health Science Centre, Toronto, Canada for making the tissue microarrays used in this study.