Histopathology 2016, 68, 191–198. DOI: 10.1111/his.12726

Loss of RNA-binding motif protein 3 expression is associated with right-sided localization and poor prognosis in colorectal cancer Nathaniel Melling, Ronald Simon,1 Martina Mirlacher,1 Jakob R Izbicki, Philip Stahl,1 Luigi M Terracciano,2 Carsten Bokemeyer,3 Guido Sauter1 & Andreas H Marx1 Department of Surgery, University Medical Centre Hamburg-Eppendorf, Hamburg, 1Institute of Pathology, University Medical Centre Hamburg-Eppendorf, Hamburg, Germany, 2Institute of Pathology, University Hospital Basel, Basel, Switzerland, and 3Department of Oncology, Haematology, BMT with section Pneumology, Hubertus Wald Cancer Centre, University Medical Centre Hamburg-Eppendorf, Hamburg, Germany Date of submission 12 January 2015 Accepted for publication 24 April 2015 Published online Article Accepted 29 April 2015

Melling N, Simon R, Mirlacher M, Izbicki J R, Stahl P, Terracciano L M, Bokemeyer C, Sauter G & Marx A H (2016) Histopathology 68, 191–198. DOI: 10.1111/his.12726

Loss of RNA-binding motif protein 3 expression is associated with right-sided localization and poor prognosis in colorectal cancer Aims: RNA-binding motif protein 3 (RBM3) has recently been suggested as a prognostic biomarker in an array of human cancers. This study aimed to examine its effects in colorectal cancers. Methods and results: RBM3 expression was analysed by immunohistochemistry on a tissue microarray containing 1800 colorectal cancers (CRCs). Nuclear RBM3 immunohistochemical staining was found in 95.9% of all interpretable CRCs. Loss of RBM3 expression was linked to advanced tumour stage (P < 0.0001),

right-sided tumour localization (P < 0.0001), and poor prognosis (P = 0.0003). In a multivariable analysis including RBM3 staining, tumour grade, tumour stage, and nodal status, only tumour stage and nodal status proved to be independent prognostic markers (P < 0.0001 each), whereas the prognostic impact of RBM3 staining was not significant (P = 0.2655). Conclusions: Our observations indicate that loss of RBM3 expression is an unfavourable prognostic marker in CRC, and is linked to right-sided tumour localization.

Keywords: colorectal cancer, RBM3, tissue microarray

Introduction Colorectal cancer (CRC) is the fourth most common malignant disease, with >1 million new cases each year worldwide.1 Despite recent advances in the management of the disease, CRC remains the second leading cancer-related cause of death in western countries.1 Therefore, elaboration and analysis of molecular features that may help to predict tumour behaviour and enable personalized therapy in individual CRC patients are urgently needed. Address for correspondence: A H Marx, MD, Institute for Pathology, University Medical Centre Hamburg-Eppendorf, Martinistr. 52, 20246 Hamburg, Germany. e-mail: [email protected] © 2015 John Wiley & Sons Ltd.

RNA-binding motif protein 3 (RBM3) is a glycinerich protein that is transcriptionally induced by low temperature2 and hypoxia.3,4 It has been suggested to play a key role in carcinogenesis and proto-oncogene function.5 RBM3 overexpression is related to decreased cell death6 and increased cell proliferation.7 Immunohistochemical studies have shown that high nuclear RBM3 expression is associated with improved prognosis in an array of human malignancies, including melanoma, and breast, ovarian, oesophageal, gastric and urothelial bladder cancer.8–15 In prostate cancer, different prognostic impacts of RBM3 immunostaining results have been reported. Whereas one study suggested high RBM3 expression to be a good prognostic sign,16 another recent large study on

192

N Melling et al.

operated prostate cancer showed that high RBM3 expression is an independent prognostic marker of early biochemical recurrence, and is tightly linked to ERG activation and PTEN deletions.17 It might well be possible that RBM3 acts differently in various tumour types. To further expand our knowledge of the relevance of RBM3 expression in CRC, we analysed a series of 1800 cancers with follow-up data on a set of tissue microarrays (TMAs). Our data demonstrate that loss of RBM3 expression in CRC is a predictor of poor prognosis, and is linked to right-sided tumour localization.

the transverse colon, and left-sided colon cancer (descending and sigmoid colon, and rectum). The utilization of tissues and clinical data was according to the Hamburger Krankenhaus Gesetz (§12 HmbKHG), and was approved by our local Ethical Committee. Table 1. Clinicopathological features of the cohort Clinical/pathological features Gender Female

640

Male

896

Materials and methods

Mean age: 69 years (29–96 years)

PATIENTS AND TMA CONSTRUCTION

Tumour grade

Two different TMAs with a total of 1800 CRC samples were included in this study. The first TMA was manufactured from resection specimens of 1420 CRC patients at the Institute of Pathology of the University Hospital of Basel. Raw survival data were obtained from the responsible physicians for all of the 1420 patients. The median follow-up period was 46 months (range 1–152 months). The second TMA included samples from 380 CRC patients whose tumour resection specimens were examined at the Institute of Pathology of the University Medical Centre Hamburg-Eppendorf. For this TMA too, overall survival data were available for all of the 380 patients, with a median follow-up period of 36 months (range 1–179 months). TMA construction was performed as previously described.18 In brief, haematoxylin and eosin-stained sections were made from each block to define representative tumour regions. One tissue cylinder with a diameter of 0.6 mm was then punched from the tumour on the ‘donor’ tissue block with a homemade semi-automated precision instrument, and placed in empty recipient paraffin blocks. Four-micrometre sections of the resulting TMA blocks were transferred to an adhesive-coated slide system (Instrumedics, Hackensack, NJ, USA). Patient information and clinical data, such as age, sex, localization and type of tumour, pTNM stage, and carcinoma grade, were retrospectively retrieved from clinical and pathological databases (Table 1). All tumours were reclassified by two pathologists (P.S. and A.M.). Follow-up data were obtained from local cancer register boards or via attending physicians. For statistical analyses, tumour localizations were grouped as follows: rightsided cancer (caecum and ascending colon), cancer of

No. available

G1

23

G2

1144

G3

158

Tumour stage pT1

60

pT2

222

pT3

845

pT4

203

Nodal status pN0

708

pN1

333

pN2/N3*

271

Tumour type Tubular carcinoma

947

Mucinous carcinoma

72

Others

10

Localization Right-sided colon

264

Transverse colon

88

Left-sided colon

330

Rectum

337

Total number of patients

1536

*Only one case with pN3. © 2015 John Wiley & Sons Ltd, Histopathology, 68, 191–198.

RBM3 in colorectal cancer

IMMUNOHISTOCHEMISTRY

Freshly cut TMA sections were analysed on 1 day and in one experiment. Slides were deparaffinized, and exposed to heat-induced antigen retrieval for 5 min in an autoclave at 121°C in pH 7.8 Tris– EDTA–citrate buffer. A primary antibody specific for RBM3 (polyclonal rabbit, HPA003624; 1:150 dilution; Sigma-Aldrich, St. Louis, MO, USA) was applied at 37°C for 60 min. Bound antibody was then visualized with the EnVision Kit (Dako, Glostrup, Denmark), according to the manufacturer’s instructions. This polyclonal antibody was also used in a different study, where it was validated against a monoclonal antibody.10 RBM3 staining was analyzed by one person (N.M.) experienced in immunohistochemistry. As nuclear staining was typically paralleled by similar or slightly lower cytoplasmic staining, only nuclear staining was considered. For statistical analyses, the staining results were categorized into three groups as previously described.19 Tumours without any staining were considered to be RBM3-negative. Tumours with 1 + or 2 + staining in up to 50% of cells or 3 + staining in up to 20% of cells were considered to be weakly positive. Tumours with 2 + staining in >50% of cells or 3 + staining in >20% of cells were considered to be strongly positive.

STATISTICS

Statistical calculations were performed with JMP 10.0.2 software (SAS Institute, Cary, NC, USA). Contingency tables and the chi-square test were performed to search

A

Figure 1. Representative pictures of (A) negative, (B) weak and (C) strong RNAbinding motif protein 3 (RBM3) staining in colorectal cancer. © 2015 John Wiley & Sons Ltd, Histopathology, 68, 191–198.

C

193

for associations between molecular parameters and tumour phenotype. Survival curves were calculated with the Kaplan–Meier method. The log-rank test was applied to detect significant survival differences between groups. Cox proportional hazards regression analysis was performed to test the statistical independence and significance between pathological, molecular and clinical variables.

Results RBM3 IMMUNOHISTOCHEMISTRY

A total of 264 of 1800 tissue samples (14.7%) were non-informative, owing to the absence of unequivocal cancer tissue or complete loss of tissue. RBM3 immunostaining was seen in 1473 of the remaining 1536 CRCs (95.9%). Immunostaining was typically nuclear. RBM3 staining was considered to be weak in 224 cases (14.6%) and strong in 1249 cases (81.3%). Representative examples of RBM3-stained cancers are shown in Figure 1. The relationship between RBM3 staining, tumour phenotype and clinical parameters is shown in Table 2. Decreased RBM3 expression was significantly associated with high tumour stage (P < 0.0001), but there were no correlations with tumour grade (P = 0.0508), nodal status (P = 0.4444), peritumoral lymphocytes (PTLs) (P = 0.4690), or vascular invasion (P = 0.1717). RBM3 staining was more frequent in mucinous carcinomas than in the common tubular type or other subtypes (P = 0.0041). Moreover, high RBM3 expression levels were significantly related to left-sided tumour localization (P < 0.0001). Strong

B

194

N Melling et al.

Table 2. Association between RNA-binding motif protein 3 (RBM3) immunohistochemistry (IHC) and clinicopathological features RBM3 IHC No. available All cancers

Negative (%)

Weak (%)

Strong (%)

1536

4.1

14.6

81.3

G1

23

4.3

13.0

82.6

G2

1144

3.0

13.4

83.7

G3

158

5.7

20.9

73.4

6.7

93.3

P-value*

Tumour grade 0.0508

Tumour stage pT1

60

0

pT2

222

1.4

10.8

87.8

pT3

845

3.0

13.8

83.2

pT4

203

8.9

20.7

70.4

pN0

708

2.4

13.8

83.8

pN1

333

4.5

15.0

80.5

pN2/N3†

271

4.8

14.4

80.7

947

3.3

13.9

82.8