Article Text

other Versions

Rsf-1 expression in rectal cancer: with special emphasis on the independent prognostic value after neoadjuvant chemoradiation
  1. Ching-Yih Lin1,2,
  2. Yu-Feng Tian3,4,
  3. Li-Ching Wu5,
  4. Li-Tzong Chen6,
  5. Li-Ching Lin7,
  6. Chung-Hsi Hsing8,
  7. Sung-Wei Lee9,
  8. Ming-Jen Sheu1,
  9. Hao-Hsien Lee10,
  10. Yu-Hui Wang11,
  11. Yow-Ling Shiue12,
  12. Wen-Ren Wu12,
  13. Hsuan-Ying Huang13,
  14. Han-Ping Hsu14,
  15. Chien-Feng Li5,6,12,15,16,
  16. Shang-Hung Chen6,17,18
  1. 1Division of Gastroenterology and Hepatology, Department of Internal Medicine, Chi-Mei Medical Center, Tainan, Taiwan
  2. 2Departments of Nursing and Nutrition & Health Science, Chang Jung Christian University, Tainan, Taiwan
  3. 3Division of General Surgery, Department of Surgery, Chi-Mei Medical Center, Tainan, Taiwan
  4. 4Department of Health and Nutrition, Chia Nan University of Pharmacy and Science
  5. 5Department of Pathology, Chi-Mei Medical Center, Tainan, Taiwan
  6. 6National Institute of Cancer Research, National Health Research Institutes, Tainan, Taiwan
  7. 7Department of Radiation Oncology, Chi-Mei Medical Center, Tainan, Taiwan
  8. 8Department of Anesthesiology, Chi-Mei Medical Center, Tainan, Taiwan
  9. 9Department of Radiation Oncology, Chi-Mei Medical Center, Tainan, Taiwan
  10. 10Department of Surgery, Chi-Mei Medical Center, Liouying, Tainan, Taiwan
  11. 11Institute of Biosignal Transduction, National Cheng Kung University, Tainan, Taiwan
  12. 12Institute of Biomedical Science, National Sun Yat-Sen University, Kaohsiung, Taiwan
  13. 13Department of Pathology, Kaohsiung Chang Gung Memorial Hospital and Chang Gung University College of Medicine, Kaohsiung, Taiwan
  14. 14College of Medicine, China Medical University, Taichung, Taiwan
  15. 15Department of Biotechnology, Southern Taiwan University, Tainan, Taiwan
  16. 16Department of Medical Technology, Chung Hwa University of Medical Technology, Tainan, Taiwan
  17. 17Division of Hematology and Oncology, Department of Internal Medicine, Chi-Mei Medical Center, Liouying, Tainan, Taiwan
  18. 18Institute of Clinical Pharmacy and Pharmaceutical Sciences, National Cheng Kung University, Tainan, Taiwan
  1. Correspondence to Dr Shang-Hung Chen, Division of Hematology and Oncology, Department of Internal Medicine, Chi-Mei Medical Center, Liouying, Tainan, Taiwan; bryanchen0615{at}


Aims Neoadjuvant chemoradiation therapy (CRT) is an increasingly used therapeutic strategy for rectal cancer. Clinically, it remains a major challenge to predict therapeutic response and patient outcome after CRT. Rsf-1 (HBXAP), a novel nuclear protein with histone chaperon function, mediates ATPase-dependent chromatin remodelling and confers tumour aggressiveness and predicts therapeutic response in certain carcinomas. However, the expression of Rsf-1 has never been reported in rectal cancer. This study examined the predictive and prognostic impacts of Rsf-1 expression in patients with rectal cancer following neoadjuvant CRT.

Methods Rsf-1 immunoexpression was retrospectively assessed for pre-treatment biopsies of 172 rectal cancer patients without initial distant metastasis. All of them were treated with neoadjuvant CRT followed by surgery. The results were correlated with the clinicopathological features, therapeutic response, tumour regression grade and metastasis-free survival (MeFS), local recurrent-free survival and disease-specific survival.

Results Present in 82 cases (47.7%), high-expression of Rsf-1 was associated with advanced pre-treatment tumour status (T3, T4, p=0.020), advanced post-treatment tumour status (T3, T4, p<0.001) and inferior tumour regression grade (p=0.028). Of note, high-expression of Rsf-1 emerged as an adverse prognosticator for diseases-specific survival (p=0.0092) and significantly predicted worse MeFS (p=0.0006). Moreover, high-expression of Rsf-1 also remained prognostic independent for worse MeFS (HR 2.834; p=0.0214).

Conclusions High-expression of Rsf-1 is associated with poor therapeutic response and adverse outcome in rectal cancer patients treated with neoadjuvant CRT, which confers tumour aggressiveness and therapeutic resistance through chromatin remodelling and represents a potential prognostic biomarker in rectal cancer.

Statistics from


Colorectal cancer has long been considered as a Western disease, but its incidence has gradually increased in Taiwan (database from Bureau of Health Promotion, Department of Health, R.O.C., Although surgery is the cornerstone of curative therapy, randomised phase III trials of resectable rectal cancer show the concurrent addition of 5-fluorouracil to neoadjuvant radiation increases the pathological complete response rate over radiotherapy alone1 and improves locoregional control.2 3 Recent studies have indicated that downstaging of the tumour after neoadjuvant chemoradiation therapy (CRT) and consequently R0 resection is correlated with improved survival.4 5 Currently, neoadjuvant CRT is generally considered for locally advanced rectal cancer featuring lymph node metastases or perirectal extension (T3, T4 or N+) with the reported response rate approximating 45%.4 Otherwise, the other advantage of neoadjuvant CRT might allow some patients to undergo sphincter-preserving low anterior resection rather than an abdominoperineal resection. The variation of sphincter preservation incidence in such patients also exists, ranging from 39% to 94%.5–7 Determining the effective biomarkers to predict the response of neoadjuvant CRT in rectal cancer patients is important for further aid risk stratification and also the development of novel therapeutic targets.

RSF1, also known as hepatitis B X-antigen associated protein (HBXAP), is a member of ATP-dependent chromatin remodelling factors that interact with sucrose non-fermenting protein 2 homologue to form a complex belonging to the ISWI chromatin remodelling family.8 9 Recently, it has been demonstrated to be an oncoprotein encoded by Rsf-1 gene located at 11p13.5-14 and frequently amplified in human ovarian high-grade serous carcinomas.10 Its overexpression is also associated with high-grade serous carcinoma, and shorter overall survival.11 Additionally, Rsf-1 protein overexpression has been detected in other histotypes of cancers, including, oesophageal, gallbladder, head and neck and nasopharynx.12–14 To the best of our knowledge, the expression status and clinical implication of Rsf-1 have never been studied in rectal cancer patients and has not linked with response to CRT. In the current study, we focused on determining the biological significance of Rsf-1 immunoexpression in a well-characterised cohort of rectal patients treated with neoadjuvant CRT.

Patients and methods


The institutional review board had approved procurement of formalin-fixed tissue of locally advanced rectal cancer patients for this study (IRB 10009-L04). Between 1998 and 2004, a total of 172 patients with rectal cancer were enrolled in this study (table 1). Those patients were confirmed as having an adenocarcinoma of the rectum by using a colonoscopic biopsy. Detailed information is provided in the online supplementary data.

Table 1

Associations and comparisons between Rsf-1 expression and clinicopathological factors in 172 rectal cancer patients receiving neoadjuvant chemoradiation therapy

Treatment plan and follow-up

Preoperatively, radiation therapy was given at a total dose of 45 Gy in 25 fractions over a period of 5 weeks with 24-h continuous infusion of 5-fluorouracil concurrently. Adjuvant systemic chemotherapy was administered if the pre-treatment (Pre-Tx) or post-treatment (Post-Tx) stage of the tumour was beyond T3 or N1. All patients were regularly monitored after diagnosis until death or their last appointment at our hospital. Detailed information is provided in the online supplementary data.

Histopathological evaluation

Following surgery, pathological analyses of the tumour specimens were performed by two pathologists (HCT and H-YH) who had not been informed of the patients' clinical information. Post-Tx T and N stages of all patients were documented according to the 7th American Joint Committee on Cancer TNM staging system.15 H&E stained sections were reviewed, and proximal, distal and circumferential resection margins were evaluated. Tumour regression grade (TRG), used as end points for evaluation of tumour response after neoadjuvant CRT, were documented as described by Dworak et al.16 Detailed information is provided in the online supplementary data.

Rsf-1 immunohistochemistry and scoring

Tissue sections from Pre-Tx rectal tumour biopsies were cut onto precoated slides from paraffin-embedded tissue blocks at 3 μm thickness. The detail experimental protocols are provided in the online supplementary data. Immunoexpression of Rsf-1, using a multiheaded microscope to reach a consensus for each case, was scored by two pathologists (HCT and H-YH) without prior knowledge of clinical results. The percentage of tumour cells with Rsf-1 immunoexpression for each specimen were classified into five groups of various expression levels from 0 to 4+, denoting none, 1%∼24%, 25%∼49%, 50%∼74% and 75%∼100% of tumour cells with moderate to strong nuclear reactivity, respectively.

Statistical analysis

High expression of Rsf-1 was defined as cases featuring 3+ or 4+ immunoreactivity for statistical analysis. The comparisons of Rsf-1 expression status between the subgroups of various clincopathological parameters were evaluated by χ2 test. The other end points analysed in this study were metastasis-free survival (MeFS), local recurrent-free survival (LRFS) and disease-specific survival (DSS), calculated from the date of operation to the date of event. Detailed information is provided in the online supplementary data.


Clinicopathological features and Rsf-1 expression in association with Pre-Tx tumour features

Patient characteristics are summarised in table 1. Detected in the cellular nuclei, Rsf-1 immunoexpression was seen and successfully scored in all 172 cases with a wide range of positively stained tumourural nuclei, varying from 5% to 90% (median, 45%). Low-expression of Rsf-1, stained less than 50% of tumour cells, was detected in 90 (53.3%) cases (figure 1A). Rsf-1 was high-expressed in 82 (47.7%) cases that displayed moderate or strong nuclear immunostaining in less than 50% of tumour cells (figure 1B). As shown in table 1, Rsf-1 high-expression was correlated with advanced Pre-Tx tumour status (p=0.020), indicating a potential role of Rsf-1 expression in promoting aggressive phenotypes in rectal cancers. Otherwise, Rsf-1 expression status was not related to other clincopathological variables statistically.

Figure 1

Representative immunostainings of Rsf-1 low expression (A) and high expression (B) cases from pre-treatment rectal cancer specimens, which demonstrate a remarkable tumour regression (C) and low tumour regression grade (D) after chemoradiation, respectively.

Rsf-1 expression in association with Post-Tx tumour features

As shown in table 1, high-expression of Rsf-1 in patients with rectal cancers was correlated with advanced Post-Tx tumour status (p<0.001) but also associated with lower-degree TRG (p=0.028, figure 1C,D).

Prognostic impact of Rsf-1 expression in rectal cancer

By univariate analysis (table 2), a number of clinicopathological parameters included Pre-Tx tumour status (p=0.0484), Pre-Tx nodal status (p=0.0059), Post-Tx tumour status (p=0.0014), vascular invasion (p=0.0123) and TRG (p=0.0037, figure 2A) were predictive for DSS. For LRFS, Pre-Tx nodal status (p=0.0025), Post-Tx tumour status (p=0.0056), vascular invasion (p=0.0023), perineurial invasion (p=0.0083) and TRG (p=0.0021) were correlated to poor prognosis. In addition, Post-Tx tumour status (p=0.0123) and TRG (p=0.0008, figure 2C) also conferred as significant prognosticator for MeFS. Except LRFS, high-expression of Rsf-1 was associated with shorter duration of DSS and MeFS (p=0.0092 and p=0.0006, respectively, figure 2B,D). More importantly, Rsf-1 high-expression (p=0.0214, HR 2.834) remained to be prognostically significant for MeFS in multivariate analysis (table 3). Besides, TRG was found to have independent prognostic impact in all survival analyses (DSS, p=0.0123, HR 2.41; LRFS, p=0.0019, HR 3.861; MeFS, p=0.0122, HR 2.426).

Table 2

Univariate log-rank analysis for important clinicopathological variables and Rsf-1 expression

Figure 2

Kaplan–Meier survival curves are plotted to predict disease-specific survival in accordance with tumour regression grade (TRG) (A) and Rsf-1 immunoexpression (B). The associations between TRG (C) and Rsf-1 immunoexpression (D) and metastasis-free survival are also illustrated.

Table 3

Multivariate analysis


Major challenges encountered in the management of rectal cancer are local tumour control and preservation of the anal sphincter. Neoadjuvant CRT is an increasingly used strategy for patients with rectal cancer, because it has been shown to improve local control and survival.1–3 For patients with distal rectal cancers, converting the surgical procedure from an abdominoperineal resection to a sphincter-preserving operation such as low anterior resection with coloanal anastomosis might be achieved after neoadjuvant CRT.5–7 However, the response rate of neoadjuvant CRT varies in rectal cancers and this multimodal treatment implies a rather high risk of serious toxicity. Therefore, there is an urgent need for new predictive biomarkers with a perspective of individualised treatment for rectal cancers.

Chromatin remodelling is a fundamental process in several biological activities, such as nucleotide synthesis, transcription regulation, DNA repair, methylation and recombination.17 Because chromatin remodelling factors play the crucial roles in biology, defects in, or aberrant expression of, chromatin remodelling proteins are associated with various developmental disorders and cancer.18 19 Rsf-1 forming the RSF protein complex by partnering with human sucrose non-fermenting protein 2 homologue, is one of the components of various ATP-dependent chromatin remodelling complexes.8–10 Rsf-1 protein also contains a plant homeodomain zinc finger motif, and mediates protein-protein interaction and transcriptional regulation in response to various growth and environmental signals.9 20–22 Recently, oncogenic properties of Rsf-1 have been demonstrated in human ovarian high-grade serous carcinomas and oral squamous cell carcinoma.10 23 Patients who had Rsf-1 amplification in high-grade serous carcinomas have been proven a shorter overall survival compared with those without amplification.10

In the present study, high-expression of Rsf-1 immunostaining was associated with advanced Pre-Tx tumour status, and Post-Tx tumour status. These characters indicate its potential role in promoting tumour progression. In univariate analysis, high-expression of Rsf-1 was demonstrated a prognostic factor for poor DSS and MeFS. These survival analyses were similar to what was previously reported in ovarian high-grade serous carcinomas and oral squamous cell carcinomas.10 23 In multivariate analysis, Rsf-1 expression level was not correlated with LRFS statistically and this could attenuate its impact on DSS. However, in MeFS, high expression of Rsf-1 was an independent prognosticator, suggesting the metastatic talent of Rsf-1 high-expression rectal cancers. Certainly, local recurrence after surgery confers serious distress to patients with rectal cancers. Fortunately, with advances in treatment, such as total-mesorectal excision and preoperative CRT, the local recurrence rate in rectal cancers has decreased recently.24 25 Conversely, the incidence of distant metastasis is still higher than local recurrence at 28–38%, despite adjuvant treatment in rectal cancer.26 27 More importantly, distant metastases equal to incurable disease and inevitably death while there exists some possibilities for local recurrence to be cured. Therefore, we suggest evaluating Rsf-1 expression in tumour biopsy tissues before treatment may be useful for predicting distant metastasis after neoadjuvant CRT and surgery in rectal cancers, albeit further validation is necessary.

Recently, postoperative nodal status has been reported to be one of the most important survival predictors in rectal cancers with neoadjuvant CRT, and is even more significant for N2 patients.28 However, postoperative nodal status was not a significant prognosticator in the present series. We speculated that is because there are much fewer N2 patients in our study than previous work (5.8% and 11.6%, respectively),28 leading an insignificant result in the prognostic evaluation of postoperative node status in our study. In addition to postoperative node status, a standardised 5-point TRG, used for tumour responses after neoadjuvant CRT in rectal cancers, is also considered a prognosticator for survival. Higher TRG, meaning better response after CRT for rectal cancer, has been proved closely related with better survival and low local recurrence.16 28 In our study, the correlation of higher TRG and better survival was confirmed again. With respect to all survival endpoints, higher TRG was significantly correlated with better survival in univariate and multivariate analyses. Tumour regression after preoperative CRT is multifactorial, including smaller, less aggressive disease, and the molecular tumour profile regulating treatment response. In expectation, high-expression of Rsf-1 was corresponding to lower TRG, meaning poor response after neoadjuvant CRT. This finding also suggests Rsf-1 properties of aggressiveness, resistance to radiotherapy or chemotherapy and poor prognosticator in rectal cancers.

Biological functions of Rsf-1 during tumour development have been further clarified and some evidences may support why Rsf-1 plays a role in prognosticator of cancer treatment. First, DNA damage response (DDR), which cells have evolved to combat threats posed by DNA damage, have diverse impact on a wide range of cellular events. Rsf-1 overexpression is recently shown to promote chromosomal instability in cancer cells though activating ataxia-telangiectasia mutated (ATM)-dependent DDR pathway in the presence of mutated p53 signalling.21 Further genetic alterations and oncogene activation during cancer development could be allowed. This may explain the aggressive phenotype of tumour cells with Rsf-1 high-expression, implying its poor nature prognosis. Second, the associations of DDR and radioresistance have been demonstrated in glioma stem cells.29 30 Cancer stem cells contribute to glioma radioresistance through preferential activation of the DNA damage checkpoint response and an increase in DNA repair capacity. Although it needs to be further validated, DDR pathway activated by higher level of Rsf-1 may also be related to radioresistance and explain poor prognosis of Rsf-1 high-expression in rectal cancers with neoadjuvant CRT. Finally, Rsf-1 is proved to enhance chemotherapeutic resistance to paclitaxel in ovarian carcinomas.20 The potential resistance to chemotherapy also increases difficulties to overcome the cancers with Rsf-1 high-expression.

In conclusion, this is the first time that Rsf-1 is shown to be highly expressed in a substantial proportion of rectal cancers and significantly correlated with advanced tumour status and lower-grade TRG after neoadjuvant CRT. More importantly, high-expression of Rsf-1 represents a significant prognosticator of worse prognosis, especially MeFS, in rectal cancer after neoadjuvant CRT. Our data suggest that high-expression of Rsf-1 contributes to disease progression and resistance of CRT in rectal cancers. Further investigation of molecular mechanisms underlying this protein expression and prospective prognostic analysis in large-scale study is warranted.

Take-home messages

  • Aberrant expression of chromatin remodelling proteins, a fundamental process in several biological activities, is associated with various developmental disorders and cancer.

  • Rsf-1 forms the RSF protein complex, one of the components of various ATP-dependent chromatin remodelling complexes, and oncogenic properties of Rsf-1 have been demonstrated in certain human cancers.

  • The response rate of neoadjuvant chemoradiation varies in patients with rectal cancer, necessitating the identification of effective prognostic and therapeutic markers.

  • In rectal cancers, high-expression of Rsf-1 was present in a substantial proportion of cases treated with neoadjuvant chemoradiation and was significantly correlated with advanced Pre- and Post-Tx tumour status. Moreover, high-expression of Rsf-1 represents an independent adverse prognosticator of patient survival and also confers a potential therapeutic target.


The authors are grateful to Translational Research Laboratory of Human Cancers of Chi-Mei Medical Center for providing critical technical assistance.

View Abstract
  • Supplementary Data

    This web only file has been produced by the BMJ Publishing Group from an electronic file supplied by the author(s) and has not been edited for content.

    Files in this Data Supplement:


  • CYL and YFT contributed equally to this work.

  • Funding This work was supported in part by grants from Chi-Mei Medical Center (CMFHR10044) and Department of Health, Taiwan (DOH99-TD-C-111-004).

  • Competing interests None.

  • Patient consent The current study used tumour materials from the Biobank of Chi-Mei Medical Center. As a rule, all tumour samples are collected only when patient consent is completed. Once being enrolled, the samples are disconnected with identifiable private information and thus no more patient consent is needed.

  • Ethics approval The institutional review board had approved procurement of formalin-fixed tissue of 172 LARC patients for this study (IRB 10009-L04).

  • Provenance and peer review Not commissioned; externally peer reviewed.

Request permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.