SEARCH

SEARCH BY CITATION

Abstract

  1. Top of page
  2. Abstract
  3. Materials and Methods
  4. Results
  5. Discussion
  6. Acknowledgment
  7. Disclosure Statement
  8. References

A downstream target of the Wnt pathway, neurone glial-related cell adhesion molecule (Nr-CAM) has recently been implicated in human cancer development. However, its role in colorectal cancer (CRC) pathobiology and clinical relevance remains unknown. In this study, we examined the clinical significance of Nr-CAM protein expression in a retrospective series of 428 CRCs using immunohistochemistry and tissue microarrays. Cox proportional hazards regression was used to calculate hazard ratios (HR) of mortality according to various clinicopathological features and molecular markers. All CRC samples were immunoreactive for Nr-CAM protein expression, compared to 10/245 (4%) matched normal tissue (< 0.0001). Of 428 CRC samples, 97 (23%) showed Nr-CAM overexpression, which was significantly associated with nodal (= 0.012) and distant (P = 0.039) metastasis, but not with extent of local invasion or tumor size. Additionally, Nr-CAM overexpression was associated with vascular invasion (P = 0.0029), p53 expression (P = 0.036), and peritoneal metastasis at diagnosis (P = 0.013). In a multivariate model adjusted for other clinicopathological predictors of survival, Nr-CAM overexpression correlated with a significant increase in disease-specific (HR 1.66; 95% confidence interval 1.11–2.47; P = 0.014) and overall mortality (HR 1.57; 95% confidence interval 1.07–2.30; P = 0.023) in advanced but not early stage disease. Notably, 5-fluorouracil-based chemotherapy conferred significant survival benefit to patients with tumors negative for Nr-CAM overexpression but not to those with Nr-CAM overexpressed tumors. In conclusion, Nr-CAM protein expression is upregulated in CRC tissues. Nr-CAM overexpression is an independent marker of poor prognosis among advanced CRC patients, and is a possible predictive marker for non-beneficence to 5-fluorouracil-based chemotherapy. (Cancer Sci 2011; 102: 1855–1861)

Colorectal carcinoma (CRC) is one of the most common human cancers in the world,(1) and is frequently diagnosed at advanced stages that require a multimodality approach for treatment. In addition to surgery and radiotherapy, contemporary treatment options for advanced CRC include multidrug fluoropyrimidine-based chemotherapeutic regimens as well as novel targeted therapies.(2) However, despite substantial progress in the understanding of CRC pathobiology in recent years, it remains a challenge in translating this knowledge to allow the identification of patients at risk for treatment failure and poor prognosis. As such, much effort has been instilled in the field of biomarker discovery to advance towards effective personalized therapy and improved clinical outcome.(3)

Recently, cell adhesion molecules (CAMs) have caught the attention of cancer researchers, due to their counter-intuitive roles in cancer invasion and metastasis. Cell adhesion molecules are known to aid intercellular adherence, yet they have also been shown to promote cell motility, invasion, and migration.(4) Activation of the canonical Wnt signaling pathway has been strongly suggested to hold a direct causal relationship to CRC initiation and progression.(5) Interestingly, several CAMs associated with the Wnt pathway have been implicated in human cancer development and progression, and their overexpression has been linked with features of aggressive disease and worse prognosis.(6–8) In particular, two members belonging to the L1 family of CAMs within the immunoglobulin-like superfamily, L1-CAM and neurone glial-related (Nr)-CAM, have been identified as targets of the β-catenin–Tcf complex. L1-CAM has been shown to be exclusively detected at the invasive front of colon cancer tissue, and its expression promoted tumorigenesis and conferred metastasis in colon cancer cells.(9) L1-CAM expression has also been identified as a marker for poor prognosis in patients with CRC.(10) Although the roles of L1-CAM in CRC biology have been well-characterized, that of Nr-CAM remain relatively obscure.

Neurone glial-related cell adhesion molecule is a 200–220 kDa transmembrane protein composed of six Ig-like domains and five FnIII repeats in the extracellular region plus a highly conserved cytoplasmic tail, and is known to be involved in the development and function of the mammalian nervous system.(11) First identified in neural tissue, Nr-CAM functions as an adhesion molecule and mediates axonal outgrowth and guidance along with other aspects of neuronal tissue development.(12–16) Nr-CAM was initially believed to be present exclusively in the nervous system,(11) but subsequent studies revealed that it is expressed in a variety of normal non-neural tissues and cells including those of the pancreas, adrenal gland, placenta, testis, thyroid, endothelium, and lens fibre.(17–21) More significantly, Nr-CAM overexpression has been confirmed in various neural(22–24) as well as non-neural cancers such as pancreatic adenocarcinoma,(19) renal cell carcinoma, papillary thyroid carcinoma,(25) malignant melanoma, and colonic carcinoma.(6) Furthermore, Nr-CAM overexpression has been shown to enhance cancer cell motility and invasiveness in vitro,(6) and is associated with aggressive clinical phenotypes.(23,24) Given the potential significance of Nr-CAM in cancer pathobiology, its clinical relevance in human CRC deserves to be investigated. Therefore, the aim of this study was to examine the protein expression of Nr-CAM in a well-characterized retrospective series of primary human CRC and matched normal tissues, as well as to evaluate Nr-CAM expression as a possible biomarker.

Materials and Methods

  1. Top of page
  2. Abstract
  3. Materials and Methods
  4. Results
  5. Discussion
  6. Acknowledgment
  7. Disclosure Statement
  8. References

Patient cohort and clinicopathological data.  The study cohort consisted of 428 consecutive CRC cases who had undergone surgical resection for histologically-proven primary CRC at the National University Hospital of Singapore (Singapore) between 1 January 1990 and 31 December 1999. Clinicopathological information available included gender, age, tumor size, tumor stage, histological grade, vascular invasion, perineural invasion, lymphatic invasion, cancer-specific mortality, and overall mortality. These data were obtained at the time of diagnosis and at subsequent follow-up. Staging was based on pathological findings according to the American Joint Committee on Cancer classification.(26) A total of 146 patients received chemotherapy (35 stage II, 75 stage III, and 36 stage IV). Cases treated with chemotherapy received intravenous 5-fluorouracil (5-FU) based on the Mayo regimen. One complete cycle of treatment involved dosage ranging from 500 to 900 mg/m2 per day for 3–5 consecutive days. Each cycle was repeated monthly for 6 months or until progression of disease, patient refusal, or adverse reactions to the treatment. Tumor response to treatment was assessed according to the standard World Health Organization response criteria.(27) This work was done under approval from the ethics committee of the National University of Singapore (NUS IRB 05-017 and DSRB Domain B/09/284).

Tissue microarray (TMA) samples.  Tissue microarray blocks containing cores from 428 CRC and 245 matched normal tissue samples were constructed from the study cohort. The TMA blocks were constructed as described previously.(28,29) Briefly, a 0.6-mm diameter needle was used to punch a donor core from morphologically representative areas of a donor paraffin-embedded tissue block. The core was then inserted into a recipient paraffin block using an ATA-100 tissue arrayer (Chemicon International, Temecula, CA, USA). Three cores were taken from the center of the tumor tissue and a single core was taken from histologically normal colorectal epithelium of matched cases where available. Consecutive TMA sections of 5-μm thickness were cut and placed on slides for immunohistochemical analyses.

Immunohistochemistry.  Neurone glial-related cell adhesion molecule rabbit polyclonal antibody was used to detect the immunohistochemical expression of Nr-CAM protein (Abcam, Cambridge, MA, USA). After deparaffinization and rehydration, tissue arrays were subjected to high temperature-induced epitope retrieval by briefly steaming them in target retrieval solution (10-mM citrate buffer, pH 6.0) (Dako, Glostrup, Denmark) in a MicroMED TT Microwave Processor (Milestone, Sorisole, Italy) for 5 min at 120°C. Subsequently, the sections were treated with serum-free protein blocking solution (Dako, Oslo, Norway) and incubated overnight with 200 μL Nr-CAM rabbit polyclonal antibody (1:500). A peroxidase-3, 3′-diaminobenzidine-based detection system (LSAB2; Dako, Norway) was used to detect the immunoreaction in the sections in accordance with the manufacturer’s specifications, and counterstained with haematoxylin. Nr-CAM staining intensity of colonic mucosa was graded using a semiquantitative score comparable to that used in previous published reports.(19,25) Staining intensity was graded as absent/weak (1+), moderate (2+), or strong (3+). Unequivocal intense staining of neural tissue was used as the internal built-in positive control for strong staining (3+) (Fig. 1d). Negative controls consisted of the omission of primary antibody without any other changes to subsequent procedures (Fig. 1e). Absent/weak (1+) staining was defined as an intensity comparable to negative controls (Fig. 1a). Moderate (2+) staining was defined as clearly positive but not intense (Fig. 1b). Nr-CAM overexpression was defined as strong (3+) immunostaining intensity of the sample, comparable to that of neural tissue (Fig. 1c). In all runs of immunohistochemistry, appropriate positive and negative controls were included. The scoring was carried out by the principle author following consultation with an expert colorectal pathologist (MST) and in the absence of information on patient outcome or tumor pathology. β-Catenin expression status was obtained for 356 cases and classified using criteria defined by Jass et al.(30) Scoring of β-catenin was based upon the distribution within the cell membrane (0 to 1+), cytoplasm (0 to 2+), and nuclei (0 to 2+). Immunohistochemical expression of CK7, CK20, Cox-2, Ki-67, p27, and p53 were carried out as detailed previously.(31)

image

Figure 1.  Expression of neurone glial-related cell adhesion molecule (Nr-CAM) in colorectal cancer. (a) Absent/weak (1+) Nr-CAM expression in normal colonic mucosa. (b) Moderate (2+) expression and (c) strong (3+) expression in colorectal cancer. (d) Intense staining of neural tissue as internal positive control (arrowheads). (e) Negative control by omission of primary antibody.

Download figure to PowerPoint

KRAS and BRAF mutational analysis.  In the stage III/IV cases, KRAS and BRAF mutations were determined by an allelic discrimination assay using the ABI 7900HT Sequence Detection System (Applied Biosystems, Foster City, CA, USA). Briefly, genomic DNA was extracted from the formalin-fixed paraffin-embedded tissues using the Gentra Puregene kit (Puregene, Gentra Systems Inc, Minneapolis, MN, USA) followed by PCR. The PCR reaction mix contained 5 μL comprising 10 ng DNA, 1 μL TaqMan Universal PCR Master Mix (Applied Biosystems), and 5 μL primers. The primer sequences for KRAS (exon 2, codons 12 and 13) were 5′-AAGGTACTGGTGGAGTATTTG-3′ (forward) and 5′-GTACTCATGAAAATGGTCAGAG-3′ (reverse). For BRAF (V600E), the sequences were 5′-CTACTGTTTTCCTTTACTTACTACACCTCAGA3-3′ (forward) and 5′-ATCCAGACAACTGTTCAAACTGATG-3′ (reverse). DNA was then submitted to the following cycle conditions: 95°C for 15 min; 40 cycles, 95°C for 15 s; and 60°C for 1 min. Data were analyzed with SDS 2.0 software (Applied Biosystems). Each mutation detected by allelic discrimination was verified by direct sequencing.

Microsatellite instability (MSI) analysis.  In the stage III/IV cases, BAT26 was used as a surrogate marker for MSI.(32,33) Briefly, the fluorescent-labeled PCR products were mixed with a size standard (GeneScan 400HD ROX; Applera, Foster City, CA, USA), ran on an automatic ABI3130 DNA analyzer, and evaluated with the GeneScan software (Applera). A variation in number and size of peaks of a marker in tumor DNA was interpreted as generalized instability for that marker.

Statistical analyses.  Data were analyzed using spss version 14.0 for Windows (SPSS, Chicago, IL, USA). Cox proportional hazards regression was used to calculate hazard ratios (HR) of mortality according to various clinicopathological features and protein markers. Multivariate Cox regression model through a stepwise selection procedure was used to test for independence of significant factors identified in univariate analyses. For analyses of CRC-specific mortality, survival was measured from the date of surgery till the date of death from CRC, or was censored till the date of the last follow-up for non-CRC related deaths and survivors. Median follow-up time for the patient cohort was 3.41 years. Kaplan–Meier survival curves were plotted to describe the distribution of CRC-specific and overall survival times, and compared using the log-rank test. Comparisons of the frequencies of Nr-CAM expression with categorical variables were carried out using the Pearson’s Chi-square test or Fisher’s exact test, as appropriate. A two-sided P < 0.05 was considered to be of statistical significance.

Results

  1. Top of page
  2. Abstract
  3. Materials and Methods
  4. Results
  5. Discussion
  6. Acknowledgment
  7. Disclosure Statement
  8. References

Neurone glial-related cell adhesion molecule protein expression in CRC.  Neurone glial-related cell adhesion molecule protein expression was evaluated by immunohistochemistry in TMA containing CRC and matched normal tissues. Representative slides of Nr-CAM immunostaining are shown in Figure 1. Peripheral nerves displayed intense Nr-CAM reactivity and were used as internal positive controls for strong staining (3+) (Fig. 1d). Nr-CAM staining was predominantly localized to the cell membrane of the mucosal epithelial cells. All of the 428 CRC samples analyzed showed at least moderate (2+) Nr-CAM expression, compared to only 10/245 (4%) adjacent matched normal colorectal tissue (< 0.0001). Furthermore, 97/428 (23%) of CRC samples showed Nr-CAM overexpression (3+). None of the normal tissue displayed strong Nr-CAM expression. Notably, Nr-CAM overexpression in CRC was significantly associated with nodal (N stage) and distant (M stage) metastasis (P = 0.012 and P = 0.039, respectively) but not with extent of local invasion (T stage) or tumor size. In addition, Nr-CAM overexpression was associated with vascular invasion (P = 0.0029), p53 expression (P = 0.036), as well as with peritoneal metastasis (P = 0.013). Membranous β-catenin expression was lost in 83 (23.3%) CRC cases and strong cytoplasmic/nuclear (2+) expression was observed in 130 (36.5%) cases. Normal colonic mucosa showed only isolated cell membrane staining without cytoplasmic/nuclear staining. These results are consistent with previously published reports.(34,35) Loss of membranous β-catenin, but not cytoplasmic/nuclear expression, was positively correlated with Nr-CAM overexpression (= 0.027). No correlation was observed with gender, age at diagnosis, ethnicity, tumor site, histological grade, lymphatic or perineural invasion, nor with p27, CK7, CK20, Ki-67, and Cox-2 expression status, MSI, KRAS and BRAF mutations. Tables 1 and 2 summarize the clinicopathological and molecular features of the cancer samples according to Nr-CAM overexpression status.

Table 1.   Clinicopathological features and neurone glial-related cell adhesion molecule (Nr-CAM) expression in 428 patients with colorectal cancer
Characteristic (n)Nr-CAM overexpression (%)P
NegativePositive
  1. AJCC, American Joint Committee on Cancer.

Total (428)33197
Gender
 Male (202)161 (49)41 (42)0.32
 Female (226)170 (51)56 (58)
Age (years)
 >70 (264)200 (60)64 (66)0.38
 ≤70 (164)131 (40)33 (34)
Ethnicity
 Chinese (375)292 (88)83 (86)0.60
 Non-Chinese (53)39 (12)14 (14)
Tumor site
 Colon (331)253 (76)78 (80)0.49
 Rectum (97)78 (24)19 (20)
Tumor size
 ≥5 cm (173)133 (40)40 (41)0.65
 <5 cm (236)187 (60)49 (59)
AJCC stage
 I (50)42 (13)8 (8)0.048
 II (159)130 (40)29 (31)
 III (115)88 (27)27 (28)
 IV (97)66 (20)31 (33)
T stage
 1 (9)9 (3)0 (0)0.59
 2 (49)38 (11)11 (11)
 3 (340)259 (78)81 (84)
 4 (30)25 (8)5 (5)
N stage
 0 (244)199 (60)45 (46)0.012
 1 (114)84 (25)30 (31)
 2 (70)48 (15)22 (23)
M stage
 0 (324)260 (79)64 (66)0.039
 1 (97)66 (20)31 (32)
 x (7)5 (1)2 (2)
Tumor grade
 Low/Moderate (377)295 (90)82 (85)0.22
 High (45)31 (10)14 (15)
Vascular invasion
 Present (43)25 (8)18 (19)0.0029
 Absent (385)306 (92)79 (81)
Lymphatic invasion
 Present (30)19 (6)11 (11)0.094
 Absent (398)312 (94)86 (89)
Perineural invasion
 Present (15)9 (3)6 (6)0.19
 Absent (413)322 (97)91 (94)
Table 2.   Molecular features and neurone glial-related cell adhesion molecule (Nr-CAM) expression in 428 colorectal cancer samples
Characteristic (n)Nr-CAM overexpression (%)P
NegativePositive
  1. †Microsatellite instability (MSI) analysis carried out for stage III/IV cases only.

Membranous β-catenin
 Loss (83)57 (21)26 (33)0.027
 Present (273)221 (79)52 (67)
Cytoplasmic/nuclear β-catenin
 Positive (130)108 (39)22 (28)0.11
 Negative (226)170 (61)56 (72)
p53
 Positive (172)122 (47)50 (61)0.036
 Negative (170)138 (53)32 (39)
p27
 Positive (130)104 (40)26 (32)0.22
 Negative (212)156 (60)56 (68)
CK7
 Positive (40)35 (13)5 (6)0.11
 Negative (302)225 (87)77 (94)
CK20
 Positive (195)156 (60)39 (48)0.064
 Negative (147)104 (40)43 (52)
Ki-67
 Positive (248)191 (73)57 (70)0.58
 Negative (94)69 (27)25 (30)
Cox-2
 Positive (253)187 (72)66 (80)0.16
 Negative (89)73 (28)16 (20)
MSI†
 Positive (9)7 (6)2 (4)1.0
 Negative (148)104 (94)44 (96)
KRAS mutation†
 Positive (46)37 (33)9 (20)0.13
 Negative (111)74 (67)37 (80)
BRAF mutation†
 Positive (19)14 (13)5 (11)0.97
 Negative (138)97 (87)41 (89)

Neurone glial-related cell adhesion molecule overexpression and patient survival in CRC.  Among the 428 patients included in the study, there were 207 CRC-specific deaths among a total of 284 deaths. We assessed the prognostic significance of clinicopathological features and protein expression in these patients. Univariate analysis revealed that older patient age, poor histological grade, advanced stage, and the presence of vascular, lymphatic, and perineural invasion were all associated with worse survival. p53 and CK20 expression were also associated with poor prognosis, whereas p27 expression was a marker of good prognosis. Consistent with our result that Nr-CAM overexpression was associated with nodal and distant metastasis, these patients (n = 97) had significantly higher disease-specific mortality compared to those without Nr-CAM overexpression (n = 331) (log-rank, P = 0.0069) (Table 3). Subgroup analysis showed that Nr-CAM overexpression was associated with increase in both disease-specific as well as overall death in advanced (P = 0.017 and P = 0.034, respectively) but not early stage disease (data not shown). In a multivariate model adjusted for other clinicopathological predictors of survival, Nr-CAM overexpression was associated with a significant increase in both disease-specific mortality (multivariate HR 1.66; 95% confidence interval [CI] 1.11–2.47; P = 0.014), and overall mortality (multivariate HR 1.57; 95% CI 1.07–2.30; P = 0.023) in advanced disease (Table 4).

Table 3.   Univariate mortality analysis, all cases of colorectal cancer assessed in this study (n = 428)
CharacteristicDisease-specific mortalityOverall mortality
HR (95% CI)PHR (95% CI)P
  1. CI, confidence interval; HR, hazards ratio; Nr-CAM, neurone glial-related cell adhesion molecule.

Gender (female vs male)1.07 (0.85–1.35)0.540.97 (0.79–1.18)0.75
Age (>70 vs≤70 years)1.61 (1.31–2.15)<0.00012.07 (1.83–2.84)<0.0001
Ethnicity (Chinese vs non-Chinese)0.99 (0.69–1.42)0.951.01 (0.73–1.39)0.95
Grade (high vs low/moderate)1.73 (1.32–2.99)0.0011.43 (1.07–2.18)0.020
Stage (III/IV vs I/II)3.12 (2.47–3.97)<0.00011.98 (1.67–2.52)<0.0001
Site (rectum vs colon)1.22 (0.93–1.65)0.141.20 (0.94–1.56)0.13
Tumor size (≥5 vs <5 cm)1.05 (0.82–1.33)0.710.93 (0.76–1.15)0.51
Distant metastasis5.65 (16.91–35.82)<0.00014.54 (12.10–24.61)<0.0001
Vascular invasion1.95 (1.59–3.73)<0.00011.69 (1.32–2.82)0.0006
Lymphatic invasion2.13 (1.74–4.86)<0.00011.88 (1.46–3.70)0.0004
Perineural invasion2.50 (2.12–8.42)<0.00012.06 (1.50–5.28)0.0013
β-Catenin (cytoplasmic/nuclear)1.01 (0.74–1.37)0.970.92 (0.70–1.20)0.52
β-Catenin (membranous)0.90 (0.63–1.28)0.560.92 (0.68–1.25)0.58
p532.53 (2.00–3.43)<0.00011.95 (1.62–2.58)<0.0001
p270.84 (0.65–1.11)0.220.78 (0.62–0.98)0.034
CK71.24 (0.82–1.84)0.291.05 (0.73–1.51)0.78
CK201.34 (1.03–1.74)0.0301.10 (0.88–1.38)0.39
Ki-671.13 (0.85–1.49)0.411.04 (0.82–1.32)0.73
Cox-20.88 (0.65–1.19)0.390.87 (0.67–1.13)0.30
Nr-CAM1.51 (1.14–2.22)0.00691.25 (0.95–1.69)0.11
Table 4.   Multivariate mortality analysis according to staging of colorectal cancer (n = 428)
CharacteristicDisease-specific mortalityOverall mortality
HR (95% CI)PHR (95% CI)P
  1. –, non-significant; CI, confidence interval; HR, hazards ratio; Nr-CAM, neurone glial-related cell adhesion molecule.

Stages I/II
 Age (>70 vs≤70 years)1.68 (1.05–2.70)0.0331.61 (1.00–2.58)0.049
 p531.88 (1.17–3.02)0.0095
 CK202.05 (1.25–3.36)0.0047
 Cox-20.51 (0.31–0.84)0.0082
Stages III/IV
 Age (>70 vs≤70 years)1.69 (1.15–2.49)0.0075
 Grade (high vs low-moderate)1.81 (1.07–3.06)0.028
 Distant metastasis6.74 (4.39–10.34)<0.00016.25 (4.11–9.51)<0.0001
 Nr-CAM1.66 (1.11–2.47)0.0141.57 (1.07–2.30)0.023

Neurone glial-related cell adhesion molecule overexpression and survival benefit from chemotherapy.  The survival of patients with CRC (stage II–IV) treated with or without chemotherapy was compared using Kaplan–Meier analysis, and classified according to tumoral Nr-CAM protein expression. 5-FU-based chemotherapy conferred significant disease-specific survival (HR 1.43; 95% CI 1.01–2.00; P = 0.044) and overall survival (HR 1.94; 95% CI 1.40–2.52; < 0.0001) benefit to patients with Nr-CAM (2+) tumors. However, patients with Nr-CAM overexpressed (3+) tumors did not benefit in terms of either disease-specific survival (HR 1.38; 95% CI 0.81–2.35; P = 0.24) or overall survival (HR 1.39; 95% CI 0.85–2.23; < 0.19) (Fig. 2).

image

Figure 2.  Kaplan–Meier survival analysis for colorectal cancer patients according to neurone glial-related cell adhesion molecule (Nr-CAM) overexpression status. The survival benefit from chemotherapy is lost in patients positive for Nr-CAM overexpression. Continuous lines, patients treated with chemotherapy; broken lines, patients not treated with chemotherapy.

Download figure to PowerPoint

Discussion

  1. Top of page
  2. Abstract
  3. Materials and Methods
  4. Results
  5. Discussion
  6. Acknowledgment
  7. Disclosure Statement
  8. References

In this study, we examined the protein expression of Nr-CAM in a cohort of primary human CRC samples and evaluated its utility as a possible biomarker. We showed that Nr-CAM expression occurred in all CRC cases but was rarely observed in normal tissue. This is in keeping with a previous study that detected significant Nr-CAM RNA expression in a small number of CRC cell lines and patient samples, but not in normal colonic tissue.(6) More significantly, we noted that a subset of CRC samples displayed a greater degree of Nr-CAM expression, which correlated with adverse clinicopathological features, including the presence of vascular invasion as well as nodal and distant metastasis. Our results suggest that Nr-CAM overexpression is a strong prognostic biomarker in advanced CRC, independent of several clinicopathological and molecular variables.

It is known that Nr-CAM is a downstream target of canonical Wnt signaling.(6) However, in our study, we observed that 72% of Nr-CAM overexpressed cases did not show cytoplasmic/nuclear β-catenin expression, suggesting that other upstream factors, apart from oncogenic β-catenin activation, might be responsible for regulating Nr-CAM expression. In line with this thinking, we found a novel relationship between Nr-CAM and p53 expression, suggesting that normal functional p53 signaling might directly or indirectly suppress Nr-CAM expression. In colorectal cancers, early oncogenic activation of β-catenin is followed by a sharp increase in p53 mutation frequency.(36) p53 downregulates β-catenin signaling by enhancing β-catenin degradation through the ubiquitin–proteasome system.(37) Of note, inactivating mutations of p53 have been associated with activation of β-catenin.(35) It may be hypothesized that p53 mutation leads to oncogenic activation of β-catenin signaling, and consequently results in upregulation of Nr-CAM gene expression. Interestingly, we also observed that loss of membranous β-catenin staining was positively correlated with Nr-CAM overexpression. It has been shown previously that L1-CAM disrupts the E-cadherin/β-catenin complex at adherens junctions, resulting in β-catenin activation and a positive feedback loop.(38) It remains to be investigated if such a signaling loop exists for Nr-CAM as well.

There is considerable evidence implicating Nr-CAM expression in cancer initiation and progression. Forced expression of Nr-CAM in NIH3T3 fibroblasts through retroviral transduction stimulated cell growth, increased cell motility, proliferation, and induced rapid tumorigenesis in nude mice. Similarly, human melanoma cells of more advanced stages express high levels of Nr-CAM and form tumors in mice, whereas those lacking Nr-CAM do not.(6) Clinically, Nr-CAM gene overexpression was observed in highly proliferative ependymoma, which conferred a poor prognosis.(23) Additionally, high-risk neuroblastomas negative for MYCN amplification showed upregulation of the Nr-CAM gene in comparison to the high-risk MYCN-amplified or intermediate-risk counterparts.(24) Notably, an apparently contrasting role of Nr-CAM may surface depending on tumor type. This is illustrated in pancreatic carcinoma, whereby poorly differentiated, invasive, and metastatic tumors showed a loss of cell surface Nr-CAM expression, compared to their early stage counterparts showing Nr-CAM overexpression.(19) Nonetheless, such a counter-intuitive finding remains consistent with the role of Nr-CAM as an adhesion molecule, as its overexpression could inhibit tumor metastasis yet promote cell–cell interactions that facilitate carcinogenesis during early stages. Another interesting observation in our study was that Nr-CAM overexpression correlated with peritoneal metastasis at diagnosis, supporting a previous study suggesting that CAMs are involved in organ selectivity during metastasis.(39) Taken together, aberrant Nr-CAM expression occurs in several human cancers and is associated with features of aggressive disease, however, the exact function might be tumor-specific and remains to be explored.

In the present study, we provide data to suggest that poor outcome conferred by Nr-CAM overexpression could be attributed to abrogation of survival benefit from 5-FU-based chemotherapy. This is a major finding, which is supported with several lines of corroborative in vitro evidence for a plausible underlying mechanism. Previously, Nr-CAM was shown to promote tumorigenesis in mice through metalloprotease-mediated shedding of its ectodomain, resulting in apoptotic resistance through constitutive activation of MAPK and AKT survival pathways, accompanied by enhanced cell motility and proliferation.(40) In addition, our results show that Nr-CAM overexpression was associated with p53 expression, which is known to be associated with apoptotic resistance in colon cancer cells.(41) Indeed, it may be speculated that Nr-CAM overexpression confers resistance to 5-FU-evoked apoptosis in CRC cells. Nonetheless, as Nr-CAM overexpression is correlated with disease stage, it remains to be examined in larger studies if there exists a differential response to 5-FU-based chemotherapy at various stages of disease progression. Future work should investigate whether other CAMs, such as L1-CAM, have predictive value for chemoresistance in CRC and in other tumor types as well. It may be that combinations of such markers will show the strongest predictive value for response to chemotherapy.

Cell adhesion molecules have been identified as potential therapeutic targets for cancer. In particular, there is emerging evidence that targeted antibody-based treatment against CAMs may delay the process of tumor progression.(4) Inoculation of antisense Nr-CAM in nude mice resulted in slower growth of glioblastoma(42) and melanoma cells.(40) Likewise, injecting antibodies and/or antisense RNA to L1-CAM into nude mice carrying human ovarian tumors(43) as well as intrahepatic cholangiocarcinoma(44) resulted in a significant reduction of tumor growth. Given that Nr-CAM overexpression may predict poor response to chemotherapy, patient stratification by Nr-CAM status may provide a more personalized approach to colon cancer therapy. Nr-CAM is an attractive protein for the development of novel agents in CRC targeted therapy, especially in those tumors expressing high levels of Nr-CAM.

The limitations of our study were the relatively small sample size, and the incomplete inclusion of known biomarkers such as status of CpG island methylator phenotype. In addition, the selection of patients for chemotherapy was not randomized, introducing a possible selection bias to our results. The advantages of our study, however, include the adequacy of follow-up and extensive data on disease characteristics, details of treatment, as well as other commonly studied tumoral molecular events, including MSI, KRAS mutation, and BRAF mutation. Thus, this allowed us to establish an effect of Nr-CAM on patient survival independent of several clinicopathological and other molecular prognostic markers.

In conclusion, this study provides the first evidence to suggest that tumoral Nr-CAM overexpression is an independent poor prognostic marker in advanced CRC. In addition, survival benefit from 5-FU-based chemotherapy may be abrogated by Nr-CAM overexpression. Our findings have several clinical implications, given the emerging roles of CAMs and related pathways as potential biomarkers and targets for cancer therapy.

Acknowledgment

  1. Top of page
  2. Abstract
  3. Materials and Methods
  4. Results
  5. Discussion
  6. Acknowledgment
  7. Disclosure Statement
  8. References

This study was funded by Singapore Cancer Syndicate, Agency for Science, Technology and Research, Singapore.

References

  1. Top of page
  2. Abstract
  3. Materials and Methods
  4. Results
  5. Discussion
  6. Acknowledgment
  7. Disclosure Statement
  8. References