Background
Colorectal cancer (CRC) is the world’s third most common cancer with over one million new cases and half a million deaths annually. Early detection, radical surgical and adjuvant chemotherapy are important for clinical outcome. The most crucial factor today for predicting patient outcome is stage of disease at diagnosis; roughly 40% has localised disease and another 40% regional disease [
1].
Over two decades ago, Bufill [
2] proposed that two different CRC entities exist according to location of the tumour either proximal to the splenic flexure (RHC = right hemicolon; caecum, ascending colon, hepatic flexure, and transverse colon) or distal to it (LHC = left hemicolon; splenic flexure, descending colon, sigmoid colon, and rectum). Cancers of the RHC (RHCC) and LHC (LHCC) exhibit different clinical and biological characteristics [
3‐
5].
Adjuvant therapy is today standard care for stage-III patients, giving an absolute 10% increase in 5-year overall survival, but for stage-II patients, the benefit of adjuvant therapy is still unclear. In stage-II patients, T4-stage, high histological grade, vascular invasion, tumour obstruction, bowel perforation, and inadequate lymph node resection favour the need for adjuvant therapy, even though limited prospective data support this [
6].
Podocalyxin-like 1 (PODXL) is an anti-adhesive transmembrane sialomucin expressed by normal vascular endothelia [
7], breast epithelial cells [
8], haematopoietic progenitors [
9], and renal podocytes [
10]. It is also a well-known stem cell marker [
11], and is closely related to stem cell marker CD34 and to endoglycan. It is thought to regulate cell morphology and adhesion through its connections to intracellular proteins and to extracellular ligands [
12‐
15]. Aberrant expression or allelic variation of PODXL or both occurs in many cancer forms, including renal cell carcinoma, breast, colorectal, testicular, prostate, and pancreatic cancer [
8,
13,
16‐
20]. In renal cell carcinoma, breast, and colorectal cancer it has also been an independent predictor of poor prognosis. The role of PODXL is not yet fully understood; though it evidence shows it to participate in epithelial-mesenchymal transition [
21] and it interacts with different mediators of metastasis [
13‐
15,
20,
22].
The aim of this study was to validate in a cohort of 840 CRC patients the role of PODXL expression as a marker of poor prognosis and to evaluate its association with clinicopathological variables by use of a novel monoclonal antibody. This new antibody HES9, produced against embryonic stem cells, is demonstrated to recognise PODXL.
Methods
Patients
The study population comprised 840 consecutive colorectal cancer patients surgically treated in 1983–2001 at the Department of Surgery, Helsinki University Central Hospital. Their median age was 66. The Finnish Population Register Centre provided the follow-up vital status data needed to compute survival statistics, and Statistics Finland provided cause of death for all those deceased. Median length of follow-up was 5.1 year (range 0–25.8), with a 5-year disease-specific overall survival rate of 58.9% (95% Cl 55.0–62.8%). The Surgical Ethics Committee of Helsinki University Central Hospital (Dnro HUS 226/E6/06, extension TMK02 §66 17.4.2013) and the National Supervisory Authority of Welfare and Health (Valvira Dnro 10041/06.01.03.01/2012) approved the study.
Preparation of tumour tissue microarrays
Formalin-fixed and paraffin-embedded tumour samples came from the archives of the Department of Pathology, University of Helsinki. An experienced pathologist marked representative areas of tumour samples on haematoxylin- and eosin- stained tumour slides. Three 1.0-mm-diameter punches taken from each sample were mounted on recipient paraffin block with a semiautomatic tissue microarray instrument (Beecher Instruments, Silver Spring, MD, USA) as described [
23].
PODXL monoclonal antibody
For the novel monoclonal antibody (mAb) HES9 used here, immunization of mice was with the undifferentiated human embryonic (hES) stem cell line SA167 (Cellartis, Gothenburg, Sweden,
http://www.cellectis-bioresearch.com); and by conventional hybridoma technology [
24] we established hybridoma cell lines producing mAbs against hES cells. Mimotope analysis, immunoprecipitation, and mass-spectrometry identified the target antigen as PODXL. The mimotope sequence corresponds to amino acid residues 189 to 192 in the PODXL protein sequence (NCBI Reference Sequence: NP_001018121.1). For a detailed description see Additional file
1.
Immunohistochemistry
Tumour tissue microarray blocks were freshly cut into 4-μm sections. After deparaffinization in xylene and rehydration through a gradually decreasing concentration of ethanol to distilled water, slides were treated in a PreTreatment module (Lab Vision Corp., Fremont, CA, USA) in Tris–HCl (pH 8.5) buffer for 20 min at 98°C for antigen retrieval. Staining of sections was performed in an Autostainer 480 (Lab Vision) by the Dako REAL EnVision Detection system, Peroxidase/DAB+, Rabbit/Mouse (Dako, Glostrup, Denmark). Tissues were incubated with the mouse mAb HES9, at dilution of 1:500 (=5 μg/ml) for one hour at room temperature. A sample of renal tissue served as a positive control in each staining series.
Scoring of samples
HES9 expression in tumour cells was mainly cytoplasmic, evenly distributed, and often granular. Membranous positivity was seen only in cells with strong cytoplasmic staining. Positivity in tumour cells was uniform, with no nuclear expression. Cytoplasmic HES9 immunoreactivity was scored independently by T.K. and J.H., who were blinded to clinical data and outcome. Negative cytoplasmic staining was scored as 0, weakly positive as 1, moderately positive as 2, and strongly positive as 3. The highest score of the triplicates of each sample was considered representative for analysis. Differences in scoring were discussed until consensus.
Statistical analyses
For statistical purposes, categories of PODXL expression were dichotomised into low (0–2) and high (3). Evaluation of the association between PODXL expression and clinicopathological parameters was done with the Fisher exact-test or linear-by-linear association test for ordered parameters. The effect of laterality on PODXL expression was confirmed by binary logistic regression adjusted for differentiation, age, gender, and Dukes classification. Disease-specific overall survival was counted from date of surgery to date of death from colorectal cancer, or until end of follow-up. Survival analysis was done by the Kaplan-Meier method and compared by the log rank test. The Cox regression proportional hazard model served for uni- and multivariable survival analysis, adjusted for sex, age, Dukes classification, and differentiation. Testing of the Cox model assumption of constant hazard ratios over time involved including a time-dependent covariate separately for each testable variable. Hazard ratios of differentiation and Dukes class D were analyzed in two periods (0 to 1.25 and 1.25 to 5 years) in order to meet the assumptions of the Cox model, and the time-dependent COX model was used. Interaction terms were considered, but no significant interactions found. All test were two-sided. A p-value of 0.05 was considered significant. All statistical analyses were done with SPSS version 20.0 (IBM SPSS Statistics, version 20.0 for Mac; SPSS, Inc., Chicago, IL, USA, an IBM Company).
Discussion
Here we used a novel mAb HES9 produced by hybridoma technology against human embryonic stem cells to show that this new mAb is specific for PODXL, known to be a stem cell marker [
11]. By immuhistochemical staining, we show that in CRC, PODXL is an independent prognostic factor. CRC has been analyzed as one disease, but here we analyzed separately tumours from the colon and rectum as well as right and left hemicolon. To our knowledge these are the first results to show that PODXL is an independent prognostic factor in subgroups of LHCC and rectal cancers. We also show a difference in PODXL expression depending on tumour location.
We also show that immunostaining of PODXL by our new mAb gives prognostic results similar to those achieved by the commercial polyclonal antibody (HPA 2110, Atlas Antibodies, Stockholm, Sweden) [
17,
18]. The antigenic determinant of the polyclonal PODXL antibody is amino acid residues 278–415, whereas our antibody reacts with amino acid residues 189–192. One proposal is that distinct membranous PODXL expression, but not cytoplasmic expression, associates with poor prognosis in different cancers [
17,
25]. Interestingly, staining by our monoclonal antibody was mainly cytoplasmic in cancer cells, with no distinct membranous immunopositivity. The reason for the difference in cancer cells compared to benign cells is not known. The difference in staining pattern may reflect different PODXL function in cancer compared to normal tissue. It is also possible that the HES9 mAb in cancer cells recognices splice variants not expressed in cytoplasm of normal cells. Similar change in expression from membranous to cytoplasmic is also seen for instance in some Toll-like receptor (TLR) stainings [
26]. The proportion of tumours with high PODXL expression was relatively small compared to the proportion in previous studies (5.7% vs. 7.9–13.4%) ([
17,
18], which could be explained by differences in antibody, patient series, staining methods, and staining evaluation/cut- off points. Similar results to ours has been reported in uterine endometroid carcinoma, where the ectopic apical expression of PODXL in benign uterine endometroid tissue was transformed into cytoplasmic expression in carcinoma [
27]. Further studies compare the expression pattern of these two antibodies in the same patient series.
High expression of PODXL was an independent marker of poor prognosis in colorectal cancer, but no difference emerged between moderate, low, or negative expression. These results are similar to earlier ones on colorectal [
17,
18] and breast cancer [
8].
Over two decades ago Bufill’s [
2] suggestion of subdivision of CRC by tumour location was not based solely on anatomical site. It is based also on developmental differences, because the RHC is derived from midgut and perfused by the superior mesenteric artery with a multilayered capillary network, whereas the LHC is derived from hindgut and perfused by the inferior mesenteric artery with a single layer capillary network.
Here, cancers of the RHC were more poorly differentiated, had higher PODXL expression, and were in older patients, more often female, although differences in age and gender were not statistically significant. Results are in concordance with earlier ones, because cancers of the right hemicolon tend to be less differentiated and more locally advanced, the patients tend to be older, and more often female [
4,
5,
28]. On the other hand, a recent study by Yamauchi [
29] suggests no discrete transition point at the splenic flexure, but a gradual change in histological and molecular characteristics from ascending colon to rectum.
The higher expression of PODXL in the RHCC was expected, because these cancers are more poorly differentiated and PODXL expression correlates with differentiation. On the other hand, we show that the difference in PODXL expression between right and left sides is independent of other predictors of survival. Interestingly, only in the LHCC group was high PODXL expression a sign of poor prognosis, not in the RHCC group. This is unlikely due to inadequate statistical power in our study, as there were 237 cases of RHCC, suggesting a different role for PODXL in tumours of the RHC compared to the LHC.
By the TMA technique, only a small proportion of the tumours are evaluated compared to studies of whole tissue sections. Moreover, for technical reasons, up to 9% of the specimens were lost in the TMA-production and -staining process. On the other hand, TMA allows analysis of larger patient cohorts. The strength of this study is a large, well-characterised CRC-patient cohort with long follow-up time that permitted subgroup analyses of patients with various tumour locations.
Acknowledgements
We thank Päivi Peltokangas, Gynel Arifdshan, and Elina Aspiala for their excellent technical assistance. This study was supported by grants from Finska Läkaresällskapet Kurt och Doris Palander Foundation, the Sigrid Jusélius Foundation, and Medicinska understödsföreningen Liv och Hälsa, and a special governmental subsidy for research and training.
Competing interests
No authors have any competing interests.
Authors’ contributions
TK performed the statistical analyses, participated in data collection, and drafted the manuscript. CF and ON were responsible for production and characterisation of mAb HES9. JH was responsible for scoring of HES9 staining. HM was responsible for statistical analyses. CB participated in data collection and figure design. CH planned the study, was responsible for the immunohistochemical methods, and helped to draft the manuscript. All authors read and approved the final manuscript.