Article Text


Associations of Rsf-1 overexpression with poor therapeutic response and worse survival in patients with nasopharyngeal carcinoma
  1. Hui-Chun Tai1,2,
  2. Hsuan-Ying Huang3,
  3. Sung-Wei Lee4,
  4. Ching-Yih Lin5,
  5. Ming-Jen Sheu5,
  6. Shih-Lun Chang6,
  7. Li-Ching Wu7,
  8. Yow-Ling Shiue8,
  9. Wen-Ren Wu8,
  10. Chun-Mao Lin9,
  11. Chien-Feng Li7,8,10,11
  1. 1Department of Pathology, Changhua Christian Hospital, Changhua, Taiwan
  2. 2Institute of Clinical Medicine, Kaohsiung Medical University
  3. 3Department of Pathology, Kaohsiung Chang Gung Memorial Hospital and Chang Gung University College of Medicine, Kaohsiung, Taiwan
  4. 4Department of Radiation Oncology, Chi-Mei Medical Center, Liouying, Tainan, Taiwan
  5. 5Division of Gastroenterology and Hepatology, Department of Internal Medicine, Chi-Mei Medical Center, Tainan, Taiwan
  6. 6Department of Otolaryngology, Chi-Mei Medical Center, Tainan, Taiwan
  7. 7Department of Pathology, Chi-Mei Medical Center, Tainan, Taiwan
  8. 8Institute of Biomedical Science, National Sun Yat-Sen University, Kaohsiung, Taiwan
  9. 9College of Medicine, Taipei Medical University, Taipei, Taiwan
  10. 10National Institute of Cancer Research, National Health Research Institutes, Tainan, Taiwan
  11. 11Department of Biotechnology, Southern Taiwan University, Tainan, Taiwan
  1. Correspondence to Chien-Feng Li, Department of Pathology, Chi-Mei Foundation Medical Center, 901 Chunghwa Road, City, Tainan County 710, Yung Kang 710, Tainan, Taiwan; angelo.p{at}


Aims Deregulated chromatin remodelling often leads to aberrant gene expression in cells, thereby implicating tumour development and progression. As a subunit of remodelling and spacing factor complex, Rsf-1 (HBXAP), a novel nuclear protein with histone chaperon function, mediates ATPase-dependent chromatin remodelling and confers tumour aggressiveness in common carcinomas. However, the expression of Rsf-1 has never been reported in nasopharyngeal carcinoma (NPC). This study aimed at evaluating the expression status, associations with clincopathological variables and prognostic implications of Rsf-1 in a well-defined cohort of NPC.

Methods Rsf-1 immunoexpression was retrospectively assessed in biopsies of 108 consecutive NPC patients without initial distant metastasis and treated with consistent guidelines. The results were correlated with the clinicopathological features, therapeutic response, local recurrence-free survival (LRFS), distant metastasis-free survival (DMFS) and disease-specific survival (DSS).

Results Present in 49 cases (45%), Rsf-1 overexpression was associated with N2,3 status (p=0.016), American Joint Committee on Cancer stage 3, 4 (p=0.004), and incomplete therapeutic response (p=0.041). In multivariate analyses, Rsf-1 overexpression not only emerged as the sole independent adverse prognosticator for LRFS (p=0.0002, RR 5.287) but also independently predicted worse DMFS (p=0.0011, RR 3.185) and DSS (p<0.0001, RR 4.442), along with T3,4 (p=0.0454) and N2,3 (p=0.0319), respectively.

Conclusion Rsf-1 overexpression is common and is associated with adverse prognosticators and therapeutic response, which confers tumour aggressiveness through chromatin remodelling, and represents a potential prognostic biomarker in NPC.

  • AIDS
  • autoimmunity
  • bacteriology
  • bone pathology
  • breast pathology
  • cancer
  • carcinoma
  • chromatin remodelling
  • molecular genetics
  • molecular oncology
  • molecular pathology
  • nasopharynx
  • oncogenes
  • Rsf-1

Statistics from

Nasopharyngeal carcinoma (NPC) is an endemic head and neck epithelial malignancy in southeast Asia and southern China.1 2 In Taiwan, the incidence of NPC is eight to 10 per 100 000 person-years, similar to that of Singapore (database from Bureau of Health Promotion, Taiwan).3 NPC is strongly linked with Epstein–Barr virus, especially in the subtypes of differentiated and undifferentiated non-keratinising carcinoma.1 2 4 Apart from Epstein–Barr virus infection required for tumour initiation, several lines of evidence indicate that genetic susceptibility and environmental factors may also contribute to deregulation of multiple oncogenes and tumour suppressor genes that induce progression from pre-neoplastic lesions to frank malignancies.1 5 The advances in diagnostic imaging, radiation therapy and adjuvant chemotherapy of NPC have achieved better locoregional control.1 6 However, the current challenge of the management of patients with NPC appears less satisfactory in final treatment outcomes.1 6 Despite being an important parameter, tumour–node–metastasis (TNM) staging still has room for improvement in terms of providing the optimal prognostication to patients. In this context, it is desirable to identify an effective biomarker that better correlates with tumour growth and/or treatment outcomes in patients with NPC, so as to aid in risk stratification and perhaps the development of therapeutic targets.

Amplification of the chromosomal region 11q13 is frequently detected in several types of human cancer, including high-grade ovarian serous carcinoma, squamous cell carcinoma of head and neck and NPC, etc.7–10 There are several established and putative oncogenes harboured in 11q13, including MEN1, CCND1, FGF3, EMS1, GARP, PAK1 and RSF1, etc.9 Of these, CCND1 (encoding cyclin D1), FGF3, EMS1 and PAK1 genes have been reported to be amplified in NPC.9 Recently, Rsf-1 protein encoded by the amplified RSF1 gene was characterised in a particularly aggressive subset of ovarian high-grade serous carcinoma,10 and Rsf-1 protein overexpression was also subsequently detected in other histotypes of cancers, including bladder, oesophageal and head and neck carcinomas. Those studies reinforced the oncogenic property of Rsf-1 in carcinogenesis. For instance, Rsf-1 was shown to promote cellular proliferation and inhibit cellular apoptosis in oral squamous cell carcinomas.8 In the setting of TP53 mutation, Rsf-1 overexpression may contribute to genomic instability in ovarian high-grade serous carcinomas.11 12 To our knowledge, the expression status and clinical implication of Rsf-1 have never been explored in NPC. This prompted us to assess systematically the immunoexpression of this nuclear protein, a subunit of remodelling and spacing factor (RSF) complex that mediates ATPase-dependent chromatin remodelling.10 13

Materials and methods

Patients and tumour specimens

The institutional review board had approved the procurement of formalin-fixed tissue of NPC for this study (IRB100-09-003). Available paraffin-embedded tissue blocks were retrieved from 108 NPC patients who underwent biopsy between January 1998 and December 2002. These patients were all free of distant metastasis at initial presentation. Two pathologists (HCT and HYH) reappraised the histological subtypes according to the current WHO classification. The tumour staging was reappraised by the 7th American Joint Committee on Cancer (AJCC) system.14

Immunohistochemical staining and assessment of Rsf-1 expression

Tissue sections were cut from paraffin-embedded blocks at 3-μm thickness. Slides were deparaffinised with xylene, rehydrated with ethanol, heated by microwave treatment in a 10 mM citrate buffer (pH 6) for 7 min. Endogenous peroxidase was quenched by 3% hydrogen peroxide treatment. Slides were washed with Tris buffered saline for 15 min and then incubated with a primary monoclonal antibody targeting Rsf-1 (1:2000; Upstate, 5H2/E4, Lake Placid, New York, USA) for 1 h, the specificity of which has previously been reported.10 Primary antibodies were detected using the ChemMate EnVision kit (DAKO, K5001, Carpinteria, California, USA). The slides were incubated with the secondary antibody for 30 min, developed with 3,3-diaminobenzidine for 5 min, and then counterstained with Gill's haematoxylin. An ovarian high-grade serous adenocarcinoma known to express RSf-1 was used as a positive control. Rabbit serum IgG was used to replace primary antibody as a negative control. Rsf-1 immunoexpression was scored by two pathologists (HCT and HYH) using a multiheaded microscope to reach a consensus for each case without knowledge of clinical and follow-up information. The percentage of tumour cells with Rsf-1 immunoexpression was recorded for each NPC specimen and 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. For statistical analysis, Rsf-1 overexpression was defined as cases featuring 3+ or 4+ immunoreactivity.

Treatment and follow-up

All 108 patients were regularly monitored after radiotherapy until death or their last appointment, with the mean follow-up duration being 64.3 months (range 4–117). Nevertheless, treated in the earlier phase of this cohort, seven patients were not available to obtain instantaneous image studies after therapy as the baseline for treatment response evaluation. Another 101 patients received a complete course of radiotherapy (total dose ≧7000 cGy) and also cisplatin-based chemotherapy in the case of stage II–IV disease, based on the published protocol.15 There were 94 complete and seven partial responders according to the methodology reported by Lin et al,16 who adopted WHO criteria with elaborations of clinical examinations and imaging modalities to assess treatment response.

Statistical analysis

Statistical analyses were performed using SPSS V.14 software package. The comparisons of Rsf-1 expression status between the subgroups of various clincopathological parameters were evaluated by χ2 or Fisher's exact test. The endpoints analysed were local recurrence-free survival (LRFS), distant metastasis-free survival (DMFS) and disease-specific survival (DSS), calculated from the starting date of radiotherapy to the date of the event. Patients lost to follow-up were censored on the latest follow-up date. Survival curves were plotted using the Kaplan–Meier method, and log-rank tests were performed to evaluate prognostic differences between groups. Multivariate analysis was performed using the Cox proportional hazards model. For all analyses, two-sided tests of significance were used with p<0.05 considered significant.


Immunohistochemical expression of Rsf-1 and associations with clinicopathological variables in NPC samples

As shown in the table 1, these 108 cases of NPC consisted of five keratinising squamous cell carcinoma (figure 1A), 47 non-keratinising differentiated carcinomas and 56 non-keratinising, undifferentiated carcinomas (figure 1B). There were 82 men and 26 women with a mean age of 48 years (range 20–77). Seven cases were classified as stage I, 18 as stage II, 52 as stage III and 31 as stage IV. Rsf-1 immunoexpression was seen and successfully scored in all 108 cases, with a wide range of positively stained tumorous nuclei, varying from 10% to 90% (median 45%). Fifty-nine cases showed 1∼2+ (<50%, figure 1C) staining, while 49 cases were 3∼4+ and therefore classified as the Rsf-1-overexpressing group (figure 1D). Rsf-1 overexpression was present in 49 (45.4%) of 108 cases and was significantly associated with NPC cases featuring N2,3 status (p=0.016), AJCC stage 3, 4 (p=0.004) and incomplete therapeutic response (p=0.041) (table 2). However, there was no association between Rsf-1 expression and other clinicopathological factors.

Table 1

Clinical pathological features of 108 NPC

Figure 1

Photomicrographs of one each representative keratising (A) and non-keratinising undifferentiated (B) nasopharyngeal carcinoma. Rsf-1 immunohistochemistry in representative nasopharyngeal carcinoma cases that showed a varying extent of nuclear labelling interpreted as low (C) and overexpression (D), respectively.

Table 2

Associations between Rsf1 expression with other important clinicopathological variables

Prognostic impact of Rsf-1 expression in NPC

As shown in table 3, patients with NPC more frequently presented with N2,3 status (p=0.0019), AJCC stage 3, 4 (p=0.0211) and only partial response to radiotherapy. The median durations to the development of local recurrence and distal metastasis were 66 and 70 months, respectively (figure 2A–C). Rsf-1 overexpression (p<0.0001) correlated with a more aggressive clinical course, with a significantly shorter DSS in patients with NPC (figure 2D). In addition, it was also strongly predictive of poor outcomes for DMFS (figure 2E) and LRFS (figure 2F) (p<0.0001 for both), together with N2,3 status (p=0.0369 for DMFS; p=0.0212 for LRFS). Other univariately significant prognosticators included T3,4 status (p=0.0054) and AJCC stage 3, 4 (p=0.0112) for worse DMFS and LRFS, respectively. In multivariate analysis, Rsf-1 overexpression still remained as the most powerful prognosticator for all three endpoints evaluated, and independently portended inferior LRFS (p=0.0002, HR 5.287) alone, worse DSS (p<0.0001, HR 4.442) along with N2,3 status (p=0.0454) and inferior DMFS (p=0.0011, HR 3.185) together with T3,4 status (p=0.0319) (table 4).

Table 3

Univariate log-rank analyses

Figure 2

For the entire cohort, survival curves were shown for various endpoints (A–C). In addition, disease-specific survival (D), distant-metastasis-free survival (E) and local recurrence-free survival (F) were plotted using the Kaplan–Meier method according to Rsf-1 immunoexpression.

Table 4

Multivariate survival analyses


With an increasing need for patient-tailored strategy to manage high-risk NPC, it is desirable to identify biomarkers independently correlated with tumour aggressiveness to facilitate appropriate allocation of adjuvant therapy.1 4 6 Approximately 15–20% of NPC patients receiving radiotherapy and/or adjuvant chemotherapy still experienced distant metastases, constituting the major cause of mortality.1 6 Apart from tumour staging, it is therefore highly desirable to identify prognostic biomarkers independently correlated with tumour aggressiveness. Non-random molecular alterations resulting from DNA copy number gains or losses are known to implicate the development and progression of NPC. In candidate tumour suppressor loci associated with NPC, multiple regions of deletions were mapped to 3p14-24.2, 9p21.3, 11q21-23, 13q12-14, 13q31-32, 14q24-32 and 16q22-23.5 9 17 However, the understanding of activated oncogenes in NPC driven by gene amplification as a result of chromosomal gains is relatively rare. Recently, an array comparative genomic hybridization (CGH) study on NPC reported frequent chromosomal gains involving 1q24.3-32.1, 1q42-ter, 3q27.3-28, 8q21-24, 11q13, 12p12.1-13.3 and 12q13.13-13.2.9 Among these, 11q13 amplicon showed the highest prevalence in well above 60% of NPC samples, and harboured several established or putative oncogenes, including amplified CCND1 gene at 11q13.9

Interestingly, RSF1 at 11p13.5-14, encoding the chromatin-remodelling nuclear protein, was also spanned by 11q13 and has been substantiated to be a driver-amplified oncogene in ovarian high-grade serous carcinoma and head and neck squamous cell carcinoma.7 8 10 12 The chromatin structure is highly dynamic and its conformation is reorganised by chromatin remodelling factors when genes are activated or repressed.18 19 Several lines of evidence have highlighted that altered levels of these chromatin remodelling factors are linked to cancer progression, because aberrations of this kind have been shown to result in incorrect gene activation or improper gene silencing.18–22

Rsf-1 was originally identified as an interacting factor for hepatitis B virus X antigen in host cells, hence the synonym hepatitis B virus x-associated protein (HBXAP).13 23 Rsf-1 forms the RSF protein complex by partnering with human sucrose non-fermenting protein 2 homologue (hSNF2H, encoded by SMARCA5 gene at 4q31), one of the components of various ATP-dependent chromatin remodelling complexes, such as ACF/BAZ-like, WICH, NuRD and RSF complexes.10 24 In the lattermost complex, Rsf-1 functions as a histone chaperone to modulate DNA binding activity, whereas hSNF2H possesses nucleosome-dependent ATPase and helicase activities for DNA unwinding.24 In addition, Rsf-1 protein contains a plant homology domain (PHD) zinc finger domain to mediate protein–protein interaction and transcriptional regulation in response to a variety of growth signals and environmental cues.7 24 The changes in the homeostasis of Rsf-1 and hSNF2H proteins have been linked to the clinical aggressiveness of Rsf1-overexpressing cancers.24

For the first time, the expression status of Rsf-1 immunostaining in NPC was documented in our study. Present in 45% of cases, Rsf-1 protein overexpression was found to be significantly associated with more advanced nodal status, AJCC stages, and incomplete therapeutic response, indicating its potential role in promoting tumour progression, aggressiveness and resistance to radiotherapy of NPC. Despite not being entirely relevant, Rsf-1 overexpression may change the expression status of several genes and activate certain signalling pathways, contributing to paclitaxel resistance in ovarian cancer.7 With respect to all three survival endpoints, Rsf-1 ovexpression was further identified as an independent adverse prognosticator of NPC in addition to the TNM staging system. These findings were essentially in keeping with what was previously reported in ovarian high-grade serous carcinomas, clear cell carcinoma and head and neck squamous cell carcinomas,8 10 25 26 suggesting the potential utility of Rsf-1 immunostaining as a prognostic biomarker of NPC. However, Rsf-1 gene amplification does not entirely account for human carcinomas with Rsf-1 protein overexpression, as reported in ovarian clear cell carcinomas.25 Given that 11q13 gain was frequent in NPC,9 it is conceivable to explore further whether RSF-1 amplification was indeed present in NPC with Rsf-1 immunohistochemical overexpression and aggressive behaviour.

In conclusion, Rsf-1 overexpression is present in 45% of NPC cases and significantly correlates with incomplete response to radiotherapy and more advanced nodal status and AJCC stages. Our data suggest that Rsf-1 overexpression may contribute to disease progression in NPC, which warrants further elucidation of the molecular mechanisms underlying its protein expression. More importantly, Rsf-1 overexpression is an independent adverse prognosticator of patient survival in this study, justifying prospective prognostic analysis in a future large-scale study.

Take-home messages

  • The chromatin structure is highly dynamic and its conformation is reorganised by chromatin remodelling factors when genes are activated or repressed. Deregulated chromatin remodelling may perturb the regulation of gene expression in tumour cells, thereby implicating tumour development and progression.

  • Compared with improved locoregional control, the current challenge of management of patients with NPC appears less satisfactory in final treatment outcomes, necessitating the identification of an effective prognostic marker to facilitate appropriate allocation of adjuvant therapy.

  • RSF1 encodes the chromatin-remodelling nuclear protein to mediate ATPase-dependent chromatin remodelling and is known to be a driver-amplified oncogene in common carcinomas.

  • In NPC, Rsf-1 overexpression is present in a substantial proportion of cases and significantly correlates with incomplete response to radiotherapy and more advanced nodal status and AJCC stages. Moreover, RSF-1 overexpression represents an independent adverse prognosticator of patient survival.


View Abstract


  • HCT and HYH contributed equally to this work.

  • Funding This work was supported in part by grants from Chi-Mei Medical Center (99-CM-TMU-03) and Department of Health, Taiwan (DOH99-TD-C-111-004).

  • Competing interests None.

  • Ethics approval Ethics approval was provided by the institutional review board, who had approved the procurement of formalin-fixed tissue and further immunohistochemical study of nasopharyngeal carcinoma for this paper (IRB100-09-003).

  • 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.