Tumor Biol. DOI 10.1007/s13277-013-0952-3

RESEARCH ARTICLE

Overexpression of CPE-ΔN predicts poor prognosis in colorectal cancer patients Kun Zhou & Hongyan Liang & Yang Liu & Chun Yang & Peijia Liu & Xiaofeng Jiang

Received: 29 April 2013 / Accepted: 17 June 2013 # International Society of Oncology and BioMarkers (ISOBM) 2013

Abstract Carboxypeptidase E (CPE) is one of the most important carboxypeptidases involved in biosynthesis of numerous peptide hormones and neurotransmitters and has an important role in endocrine regulation. A splice variant of CPE (CPE-ΔN) has been detected and the mechanism of CPE-ΔN action in tumorigenesis has been studied in many different cancers. The aim of this study was to examine CPE-ΔN expression in human colorectal cancer (CRC) and to evaluate its possible use as a potential prognostic marker. Two hundred nineteen primary colorectal tumors and corresponding normal tissues were included in the study. We have analyzed CPE-ΔN isoform expression by qRT-PCR and Western blot in 219 CRC patients. Correlations between CPE-ΔN mRNA expression and clinicopathological variables were determined with chi-square tests. Survival probabilities were determined using Kaplan–Meier analysis, and univariate and multivariate analyses of the prognostic factors were performed with a Cox regression model. Our results show that CPE-ΔN is overexpressed in colorectal tumor tissue and that high CPE-ΔN mRNA expression is closely correlated with

tumor differentiation, pT classification, pN classification, tumor recurrence, and lymph node metastasis (P=0.042, 0.036, 0.031, 0.006, and 0.008, respectively). However, no correlation was observed between CPE-ΔN expression and age, gender, tumor localization, gross features, and the tumor size. In addition, patients with high CPE-ΔN expression had a significantly shorter survival (P<0.001, logrank test). Tumor differentiation, gross feature, pT classification, pN classification, tumor recurrence, lymph node metastasis, and CPE-ΔN status were significantly associated with poor prognosis after performing a univariate Cox survival analysis. High CPE-ΔN expression was also identified as an independent prognostic factor using a multivariate analysis (P=0.011). Based on these results, we can conclude that CPE-ΔN expression might be a potential prognostic marker for colorectal cancer patients. Keywords Carboxypeptidase E . Recurrence . Metastasis . Colorectal cancer . Prognosis

Introduction Kun Zhou & Hongyan Liang contributed equally to this work. K. Zhou : H. Liang : Y. Liu : C. Yang : X. Jiang (*) Department of Clinical Biochemistry Laboratory, The 4th Affiliated Hospital of Harbin Medical University, 37 Yiyuan Street, Nangang, Harbin, People’s Republic of China e-mail: [email protected] K. Zhou (*) Department of Clinical Laboratory, Daqing Oilfield General Hospital Group, 9 Zhongkang Street, Saertu, Daqing, People’s Republic of China e-mail: [email protected] H. Liang : X. Jiang Heilongjiang Province Key Laboratory of Molecular Imaging, Harbin, People’s Republic of China P. Liu The 2nd Affiliated Hospital of Harbin Medical University, Harbin, People’s Republic of China

Recurrence and metastasis are the two main reasons leading to death of patients with colorectal cancer (CRC). The major cause of high recurrence and metastasis rate is mainly inadequate treatment after CRC surgery and presence of residual tumor cells prone to recurrence and metastasis. Therefore, identification of biomarkers that could accurately predict future recurrence and metastasis from the primary tumor could help clinicians improve treatment strategies. Carboxypeptidase E (CPE), also known as carboxypeptidase H or enkephalin convertase [1], is a member of the metallocarboxypeptidase gene family [2]. Although it is generally considered that the enzyme-active site binds Zn2+ ions, enzymatic activity is enhanced by millimolar concentrations of Co2+ with a pH optimum of 5.5 (pH range: 5.0–6.0) [3, 4]. CPE is one of the most important carboxypeptidases involved in biosynthesis of numerous peptide hormones

Tumor Biol.

and neurotransmitters [2]. Soluble CPE in secretory granules can cleave C-terminal basic amino-acid residues to generate mature neuropeptides and hormones [5, 6]. Another form of CPE is membrane-bound CPE, which acts as a sorting receptor in the trans-Golgi network [7, 8]. High CPE expression is present in pancreatic islets, pituitary glands, and adrenal glands [3], while medium and low CPE expression is present in the brain and normal neuroendocrine tissues [9]. However, CPE expression has not been detected in some tissue types such as hepatocyte cells [10] and cervical tissues [11]. A CPE splice variant (CPE-ΔN) lacking the N-terminus has been described in previous studies [12]. Interestingly, in some cancer types, only the CPE-ΔN splice variant is expressed. Moreover, CPE-ΔN expression is frequently observed in highly metastatic cancers as opposed to benign tumors or normal tissues [12], and CPE-ΔN mRNA expression correlated with tumor growth and invasiveness in some tumor types [11]. In the CPE-ΔN isoform, 98 nucleotides from the first exon are removed by alternative splicing via noncanonical alternative 3′ and 5′ splice sites, leading to a translational “advantage” during tumor metastasis [13]. Based on the data in the expressed sequence tag (EST) database, this variant codes for a peptide that might be secreted into the extracellular matrix and function as a signaling molecule or ligand [13]. It has been proposed that CPE-ΔN may participate in cancer development by inducing both Wnt signaling activity [14] and NEDD9 gene expression, leading to tumor development and progression [12]. In the first stages of CRC development, the expression of CPE is lost and CPE-ΔN begins to be expressed as the disease progresses [14]. Lee et al. [12] demonstrated that CPE-ΔN was a potentially powerful biomarker for predicting future metastasis and recurrence in hepatocelullar carcinoma (HCC) and pheochromocytoma/paraganglioma (PHEO/PGL) patients. The aim of this study was to examine CPE-ΔN mRNA and protein expression in CRC tissues and corresponding normal tissues by qRT-PCR and Western blot and to evaluate the association between CPE-ΔN expression and clinicopathological features as well as survival of CRC patients.

Patients and methods Clinical samples Specimens from surgical resection of 250 CRC patients from the Fourth Clinical Medical College of Harbin Medical University or Affiliated Tumor Hospital of Harbin Medical University were included in this study between January 2008 and March 2009. Following the resection, tissue

samples were immediately frozen in liquid nitrogen until further use. All specimens were confirmed by a pathologist. Patients included in this study did not receive any chemotherapy or radiotherapy prior to surgery. All patients were followed until March 2012 (range 36–50 months), except for those who died before the end of the follow-up period. Since 31 patients did not have usable normal tissues, 219 primary colorectal tumors and corresponding normal tissues (≥5 cm away from the tumor) were included in the study. The age range of patients included in this study was from 36 to 82 years at the time of diagnosis and 61.6 % of the patients were men. All tumors were staged according to the seventh edition of the American Joint Committee on Cancer (AJCC) staging for CRC. According to the AJCC classification, there were 36 stage I patients, 95 stage II patients, 79 stage III patients, and nine stage IV patients. Thirty patients were diagnosed with T1 tumors, 48 patients had T2, 104 patients had T3, and 37 patients had T4. In addition, 131 patients were classified as N0, 60 as N1, and 28 as N2. The diagnosis of recurrence and metastasis were determined by high levels of carcinoembryonic antigen (CEA) and verified by CT or MRI scans. The study was approved by the ethics committees of institutions included in this study, and written consent from all patients was obtained. Quantitative real-time PCR For the quantitative real-time PCR (qRT-PCR) analysis, RNA was extracted from the tumors and surrounding nontumor tissue using the TRIzol reagent (Invitrogen, USA). Next, complementary DNAs (cDNAs) were reverse transcribed from 1 μg of total RNA. The primers were designed according to the reported gene sequences of human CPE (GenBank accession no. NM_001873.2). 18S RNA was used as a control gene. The primer sequences were as follows: CPE-ΔN forward, 5′-ATGGCCGGGCATGAGGCGGC-3′ and reverse 5′GCTGCGCCCCACCGTGT AAA-3′ designed according to the Lee et al. [12]. 18S RNA forward, 5′-CTCTTAGCTGA GTGTCCCGC-3′ and reverse 5′-CTGATCGTCTTC GAAC CTCC-3′. Amplification of 18S RNA and CPE-ΔN transcripts were performed in 25-μl reaction volume using the SYBR Green Master Mix PCR kit (Applied Biosystems, USA) with a DA7600 PCR amplifier (DAAN, China), respectively. The reaction mixture was denatured at 95 °C for 3 min and subjected to 35 cycles of 95 °C for 30 s, 63 °C for 30 s and 72 °C for 30 s, followed by a final extension at 72°C for 10 min. All the experiments were performed two times and averaged to obtain the data for each specimen. The relative levels of CPE-ΔN mRNA transcripts were normalized to 18S RNA, which served as an endogenous control. Relative threshold cycles (ΔCt) were calculated according to Livak and

Tumor Biol. Fig. 1 Western blot analysis on 12 % SDS-PAGE gels showing CPE-ΔN protein expression in CRC tissues and surrounding normal tissues from 10 patients. T, primary CRC, N, corresponding normal tissue

Schmittgen [15]. The amount of target=2−ΔΔCt, where ΔΔCt =(Ct(CPE-ΔN) −Ct(18S RNA))test −(Ct(CPE-ΔN) −Ct(18S RNA))control. Western blot For the Western blot analysis, proteins from 50 patients were extracted from their tumors and surrounding normal tissue specimens using a 2× loading buffer [250 mM tris–HCl, pH 6.8; 10 % sodium dodecyl sulfate (SDS); 50 % glycerol; 0.5 % bromophenol blue; 5 % β-mercaptoethanol; protease inhibitor cocktail; 1 mM phenylmethanesulphonyl fluoride (PMSF); and DNase I]. Equal amounts of protein were run

Fig. 2 a Melting curve analysis for 18S RNA in CRC primary tumors. b Melting curve analysis for CPE-ΔN in CRC primary tumors

on 12 % SDS-PAGE gels and transferred onto PVDF membranes (Millipore, USA). CPE-ΔN was detected using a purified mouse anti-CPE monoclonal antibody (BD Biosciences, USA) at a 1:3,000 dilution. Membranes were then incubated with a secondary anti-mouse horseradish peroxydase-conjugated antibody (Bio-Rad, Hercules, CA, USA) diluted at 1:2,000. Statistical analysis All statistical data were analyzed using SPSS 16.0 software (SPSS, Chicago, IL, USA). Correlations between CPE-ΔN

Tumor Biol. Table 1 The association between CPE-ΔN expression and clinicopathological features of CRC patients

* P value is considered significant if p<0.05 a

Average of CPE-ΔN/18S RNA mRNA expression in CRC tissues

b P value was analyzed in highlevel CPE-ΔN (CPE-ΔN-H) versus the low-level CPE-ΔN (CPE-ΔN-L) expression by chisquare test

Mean±SDa

χ2

P valueb

28 (34.6) 48 (34.8)

1.26±0.05 1.26±0.06

0.001

0.974

Male 135 Female 84 Tumor localization Colon 125 Rectum 94 Tumor differentiation Well/moderate 154

44 (32.6) 32 (38.1)

1.27±0.04 1.26±0.06

0.692

0.406

48 (38.4) 28 (29.8)

1.26±0.04 1.27±0.06

1.756

0.185

60 (39.0)

1.29±0.05

4.151

0.042*

Poor Gross features Non-infiltrating Infiltrating Tumor size (cm) ≤5.0 >5.0 pT classification pT1/2 pT3/4 pN classification

65

16 (24.6)

1.25±0.05

45 174

18 (40.0) 58 (33.3)

1.27±0.05 1.23±0.02

0.701

0.402

98 121

40 (40.8) 36 (29.8)

1.27±0.05 1.26±0.05

2.925

0.087

78 141

20 (25.6) 56 (39.7)

1.28±0.04 1.21±0.05

4.391

0.036*

pN0 131 pN1–2 88 Tumor recurrence No 147 Yes 72 Lymph node metastasis No 130 Yes 89

38 (29.0) 38 (43.2)

1.29±0.04 1.24±0.05

4.667

0.031*

42 (28.6) 34 (47.2)

1.29±0.04 1.25±0.05

7.419

0.006*

36 (27.7) 40 (44.9)

1.28±0.04 1.22±0.06

6.939

0.008*

Variables

No. of cases

CPE-ΔN high expression cases (%)

All cases Age (years) ≤60 >60 Gender

219

76 (34.7)

81 138

mRNA expression and clinicopathological variables were determined with chi-square tests. The survival probabilities were determined using a Kaplan–Meier analysis, and

Fig. 3 a Kaplan–Meier analysis of CRC patients according to CPE-ΔN expression. The 4-year survival rate was 63.5 % for patients with a low-level of CPE-ΔN expression (CPEΔN-L) (n=143) and 40.4 % for patients with a high-level of CPE-ΔN expression (CPEΔN-H) (n=76). b Kaplan– Meier analysis of recurrencefree survival according to CPEΔN expression

survival differences were analyzed using log-rank tests based on the CPE-ΔN expression status. Univariate and multivariate analyses of the prognostic factors were performed with a

Tumor Biol. Table 2 Univariate analysis of prognostic markers in 219 patients with colorectal cancer

CI confidence interval *

p<0.05 (statistical significance)

β

Variable

χ2

P

HR

95 % CI Lower

Upper

Age (year)

0.180

0.677

0.411

1.197

0.780

1.839

Gender Tumor differentiation Gross features pT classification pN classification Tumor recurrence Lymph node metastasis CPE-ΔN expression

0.069 0.643 0.938 1.711 1.187 1.670 1.867 0.775

0.111 9.401 7.850 30.429 31.579 57.180 72.704 14.165

0.739 0.002* 0.005* <0.001* <0.001* <0.001* <0.001* <0.001*

1.071 1.903 2.555 5.534 3.276 5.314 6.468 2.172

0.714 1.261 1.326 3.013 2.166 3.454 4.211 1.450

1.607 2.870 4.925 10.163 4.955 8.174 9.935 3.252

Cox regression model. For continuous variables, the data are expressed as mean±standard deviation (SD). P<0.05 was considered statistically significant.

Results Association of CPE-ΔN expression with clinicopathological variables In order to evaluate CPE-ΔN protein expression in 50 CRC patients, Western blot was performed (Fig. 1). The results showed that CPE-ΔN protein expression was significantly higher in colorectal tumors than in the surrounding normal tissue. For the CPE-ΔN mRNA expression analysis in 219 CRC patients, qRT-PCR was performed. The melting curve of PCR products showed a single peak indicating good reaction specificity (Fig. 2a and b). In addition, we examined the association of CPE-ΔN mRNA expression and clinicopathological features of patients and their tumors (Table 1). The CPE-ΔN/18S RNA ratio was used to reflect the relative CPE-ΔN mRNA expression level. The average CPE-ΔN/18S RNA ratio in 219 in colorectal tumors was 1.41±0.05. Patients with a ratio higher than 1.41±0.05 and a CPE-ΔN Ct value higher than 20 cycles were classified as patients in the low expression CPE-ΔN-L group (n=143), while patients with a ratio less than 1.41±0.05 and a

CPE-ΔN Ct value lower than 20 cycles were classified as the high expression CPE-ΔN-H group (n=76). High CPE-ΔN mRNA expression closely correlated with tumor differentiation, pT classification, pN classification, tumor recurrence, and lymph node metastasis (P=0.042,0.036, 0.031, 0.006, and 0.008, respectively). There was no significant association between CPE-ΔN mRNA expression and age, gender, tumor localization, gross features, and tumor size (P=0.974, 0.406, 0.185, 0.402, and 0.087, respectively) (Table 1). Correlation of CPE-ΔN expression with colorectal cancer prognosis The results of the survival analysis using the Kaplan–Meier method showed that the CPE-ΔN-H had a significantly shorter overall survival and recurrence-free 4-year survival period compared with those included in the CPE-ΔN-L group (log-rank test P<0.001 and P=0.001, respectively) (Fig. 3a and b). Univariate Cox survival analysis In order to evaluate the possible associations between prognosis and different clinicopathologic features, including age, gender, tumor differentiation, gross features, pT classification, pN classification, tumor recurrence, lymph node metastasis, and

Table 3 Multivariate Cox regression analysis of prognostic markers in 219 patients with colorectal cancer (backward stepwise model) Variable

pN classification Lymph node metastasis Tumor recurrence CPE-ΔN expression CI confidence interval *

p<0.05 (statistical significance)

β

0.582 2.526 0.716 0.542

χ2

6.935 69.008 9.309 6.542

P

0.008* <0.001* 0.002* 0.011*

HR

1.789 12.498 2.047 1.719

95 % CI Lower

Upper

1.160 6.887 1.292 1.135

2.759 22.679 3.243 2.603

Tumor Biol. Fig. 4 a and b Kaplan–Meier survival curves of CRC patients in relation to CPE-ΔN expression in tumors with different pT and pN classifications. c Kaplan–Meier survival curves in relation to CPE-ΔN expression in patients with infiltrating and non-infiltrating tumors. d Kaplan–Meier survival curves in relation to CPE-ΔN expression in tumors based on differentiation status

CPE-ΔN expression (CPE-ΔN-H versus CPE-ΔN-L), we performed a univariate Cox survival analysis. Among the parameters, tumor differentiation (P=0.002), gross feature (P=0.005), pT classification (P<0.001), pN classification (P<0.001), tumor recurrence (P<0.001), lymph node metastasis (P<0.001), and CPE-ΔN status (P<0.001) were significantly associated with poor prognosis (Table 2). The variables found to have prognostic influence may represent covariates; therefore, all significant variables from the univariate analysis were included in the multivariate regression analysis in order to identify independent prognostic factors (Table 3). CPE-ΔN expression was identified as an

Table 4 Correlation of CPE-ΔN mRNA expression with recurrence in CRC patients Variable

CPE-ΔN expression

Tumor recurrence

Number

T/N≤2

T/N>2

P value

Yes No

62 83

8 63

54 20

<0.001

T/N Tumor versus Non-tumor

Tumor Biol. Table 5 Correlation of CPE-ΔN mRNA expression with lymph node metastasis in CRC patients Variable

CPE-ΔN expression

Lymph node metastasis

Number

T/N≤2

T/N>2

P value

Yes

59

7

52

<0.001

No

86

69

17

T/N Tumor versus Non-tumor

independent prognostic factor (P=0.011), as were pN classification (P=0.008), tumor recurrence (P=0.002), and lymph node metastasis (P<0.001). We next examined correlations between CPE-ΔN expression (CPE-ΔN-H versus CPE-ΔN-L) and prognosis using the Kaplan–Meier method. Survival analysis was performed with regard to CPE-ΔN mRNA expression levels in patients with tumors of different pT and pN classifications. The results indicated that CPE-ΔN-H was a CRC prognostic factor regardless of the pN status (pN0, P=0.007; pN1–2, P=0.005). However, CPE-ΔN expression could not differentiate outcomes for pT1/2 CRC patients (P=0.163). Kaplan–Meier survival curves in relation to CPE-ΔN expression in different pathological pT and pN classifications of CRC patients are shown in Fig. 4a and b. In addition, the CPE-ΔN-H group had shorter survival than the CPE-ΔN-L group among patients with infiltrating tumors (P<0.001), but not among noninfiltrating tumors (P=0.195) (Fig. 4c). Patients in the CPE-ΔN-H group had a poorer prognosis regardless of tumor differentiation (well/moderate, P=0.002; poor, P<0.001) (Fig. 4d). CPE-ΔN mRNA expression in the corresponding normal tissue for all 219 patients was examined by qRT-PCR. CPE-ΔN mRNA tumor expression versus nontumor expression was presented as a ratio (T/N). First, 74 patients were analyzed in order to establish the cutoff value for mRNA expression in normal tissue. Future recurrence and metastasis were predicted by a T/N value >2. Next, 145 CRC patients were used to verify this ratio. In 63/83 patients without recurrence, the CPE-ΔN mRNA T/N ratio was 2 or less, whereas in 54/62 patients with Fig. 5 Kaplan–Meier survival curves of 4-year recurrence-free survival and 4-year metastasisfree survival for 145 CRC patients

recurrence, the T/N ratio was greater than 2. Therefore, the CPE-ΔN mRNA T/N ratio greater than 2 could predict recurrence within 4 years, with a sensitivity of 87 % and a specificity of 75.9 % (Table 4). In addition, the T/N ratio predicted future metastasis with a sensitivity of 88.1 % and a specificity of 80.2 % (Table 5). Kaplan–Meier curves for 4-year recurrencefree survival and 4-year metastasis-free survival (Fig. 5) indicated a shorter survival time (P<0.001) for patients with a CPE-ΔN mRNA T/N ratio higher than 2 when compared with patients with a T/N ratio of 2 or less.

Discussion Microarray data from the Gene Expression Omnibus (GEO) profile database demonstrated that significant overexpression of CPE mRNA was detected in many different cancer types. CPE mRNA expression was present not only in neuroendocrine tumors, such as PHEO/PGL [11], pulmonary neuroendocrine tumors [16], insulinomas [17], and various types of astrocytomas and oligodendrogliomas, but also in many metastatic nonendocrine cancers, including liver cancer [10], breast cancer [18], cervical cancer, colorectal cancer, clear cell sarcoma, and Ewing sarcomas [11]. Recently, Lee et al. [12] detected a CPE isoform lacking the N-terminus (CPE-ΔN). The CPE-ΔN isoform has been detected in breast, colon, and head, and neck tumor cell lines by qRT-PCR, and CPE-ΔN mRNA expression predicted recurrence and metastasis in HCC and PHEO/PGL patients [12]. In addition, CPE-ΔN expression has been associated with highly metastatic cell lines from various human cancers [12]. The CPE expression is present in neuroendocrine tumors in which CPE acts as an autocrine growth factor [19]. In a study by Jeffrey et al. [20], CPE was shown to be involved in the regulation of apoptosis. In addition, He et al. [16] showed that high CPE expression predicted good prognosis in neuroendocrine tumors, small-cell lung cancers, and large-cell neuroendocrine carcinomas. Contrary to these findings, Krajnik et al. [21] showed that CPE overexpression in lung cancer was an indicator of poor patient prognosis. In glioma

Tumor Biol.

cells, CPE overexpression reduces the migratory and invasive potential of tumor cells [22], and CPE expression is reduced in gliomas and is associated with shorter survival in glioblastoma (GBM) patients. Furthermore, CPE enhances cell growth independent of the presence of CPE-ΔN [22]. Lee et al. detected the CPE-ΔN isoform in CRC cell lines [12]. However, the relationship between CPE-ΔN and clinicopathological features, as well as the prognostic significance of CPE-ΔN in CRC, has not been assessed. In this study, we evaluated CPE-ΔN expression at the mRNA and protein levels in colorectal tumors and surrounding normal tissues and investigated the potential use of CPE-ΔN as an independent prognostic factor for CRC patients. Our results showed that CPE-ΔN expression was significantly higher in CRC tissues than in the surrounding normal tissues. In addition, high CPE-ΔN expression closely correlated with poor tumor differentiation, advanced pT stage, advanced pN stage, tumor recurrence, and lymph node metastasis. Therefore, CPE-ΔN might have important roles in the progression, invasion, and metastasis of CRC. Furthermore, our results indicated that CPE-ΔN was an important prognostic factor for pT3/4 patients, but not in pT1/2 patients (Fig. 4a). CPE-ΔN expression also correlated with pN, regardless of the status in CRC patients, and CPE-ΔN expression was an independent prognostic factor for overall survival. In patients with high CPE-ΔN expression, recurrence was more frequent. In addition, we analyzed CPE-ΔN protein expression in 50 primary CRC tumors and the surrounding normal tissue. Patients diagnosed with lymph node metastasis had a significantly higher CPE-ΔN T/N ratio than those without metastatic disease (data not shown). Lee et al. analyzed the expression of CPE-ΔN in colorectal cancer. It is showed that 93.5 % of patients without lymph node metastasis had CPE-ΔN mRNA T/N ratios of 2 or less, while 86.5 % of patients diagnosed with lymph node metastasis within 1 year after surgery had tumor CPE-ΔN mRNA T/N ratios above 2 [13]. In our study, the prediction of future metastasis using the CPE-ΔN mRNA T/N ratio had a sensitivity of 88.1 % and a specificity of 80.2 %. In conclusion, our study showed that the expression of CPE-ΔN was significantly higher in colorectal tumors than in the surrounding normal tissue. Overexpression of CPE-ΔN might be used as an independent prognostic marker for CRC patients. In addition, CPE-ΔN overexpression might have important roles during the CRC tumorigenesis. Detection of CPE-ΔN expression in primary colorectal tumors and corresponding normal tissues and prediction of future recurrence or metastasis from the primary tumor may reduce or prevent tumor recurrence and metastasis and potentially improve treatment strategies for patients after surgery. Thus, the evaluation of CPE-ΔN expression status could be helpful in the identification of CRC patients in need of a more intensive treatment.

Acknowledgments We thank Medjaden Bioscience Limited for assisting in the preparation of this manuscript.

Conflict of interest None.

References 1. Fricker LD, Snyder SH. Enkephalin convertase: purification and characterization of a specific enkephalin-synthesizing carboxypeptidase localized to adrenal chromaffin granules. Proc Natl Acad Sci USA. 1982;79:3886–90. 2. Chen H, Jawahar S, Qian Y, et al. Missense polymorphism in the human carboxypeptidase E gene alters enzymatic activity. Hum Mutat. 2001;18(2):120–31. 3. Fricker LD, Carboxypeptidase E. Annu Rev Physiol. 1988;50:309– 21. 4. Fricker LD. Activation and membrane binding of Carboxypeptidase E[J]. J Cell Biochem. 1988;38(4):279–89. 5. Fricker LD, Snyder SH. Purification and characterization of enkephalin convertase, an enkephalin-synthesizing carboxypeptidase. J Biol Chem. 1983;258:10950–5. 6. Hook VY, Loh YP. Carboxypeptidase B-like converting enzyme activity in secretory granules of rat pituitary. Proc Natl Acad Sci USA. 1984;81(9):2776–80. 7. Cool DR, Normant E, Shen F, Chen HC, Pannell L, Zhang Y, et al. Carboxypeptidase E is a regulated secretory pathway sorting receptor: genetic obliteration leads to endocrine disorders in Cpe(fat) mice. Cell. 1997;88:73–83. 8. Dhanvantari S, Shen FS, Adams T, Snell CR, Zhang C, Mackin RB, et al. Disruption of a receptor-mediated mechanism for intracellular sorting of proinsulin in familial hyperproinsulinemia. Mol Endocrinol. 2003;17:1856–67. 9. Song L, Fricker LD. Purification and characterization of carboxypeptidase D, a novel carboxypeptidase E-like enzyme, from bovine pituitary. J Biol Chem. 1995;270(42):25007–13. 10. Ge X, Yamamoto S, Tsutsumi S, Midorikawa Y, Ihara S, Wang SM. Aburatani H Interpreting expression profiles of cancers by genomewide survey of breadth of expression in normal tissues. Genomics. 2005;86:127–41. 11. Murthy SR, Pacak K, Loh YP. Carboxypeptidase E: elevated expression correlated with tumor growth and metastasis in pheochromocytomas and other cancers. Cell Mol Neurobiol. 2010;30(8):1377– 81. 12. Lee TK, Murthy SR, Cawley NX, et al. An N-terminal truncated carboxypeptidase E splice isoform induces tumor growth and is a biomarker for predicting future metastasis in human cancers. J Clin Invest. 2011;121(3):880–92. 13. Cawley NX, Wetsel WC, Murthy SR, Park JJ, Pacak K, Loh YP. New roles of carboxypeptidase E in endocrine and neural function and cancer. Endocr Rev. 2012;33:216–53. 14. Skalka N, Caspi M, Caspi E, Loh YP, Rosin-Arbesfeld R. Carboxypeptidase E: a negative regulator of the canonical Wnt signaling pathway. Oncogene. 2012; Jul 23. doi: 10.1038/onc.2012.308. 15. Livak KJ, Schmittgen TD. Analysis of relative gene expression data using real-time quantitative PCR and the 2−ΔΔCT method. Methods. 2001;25(4):402–8. 16. He P et al. Identification of carboxypeptidase E and gammaglutamyl hydrolase as biomarkers for pulmonary neuroendocrine tumors by cDNA microarray. Hum Pathol. 2004;35(10):1196–209. 17. Guest PC, Ravazzola M, Davidson HW, Orci L, Hutton JC. Molecular heterogeneity and cellular localization of carboxypeptidase H in the islets of Langerhans. Endocrinology. 1991;129(2):734–40.

Tumor Biol. 18. Du J, Keegan BP, North WG. Key peptide processing enzymes are expressed by breast cancer cells. Cancer Lett. 2001;165: 211–8. 19. Rozengurt E. Neuropeptides as growth factors for normal and cancerous cells. Trends Endocrinol Metab. 2002;13(3):128–34. 20. Jeffrey KD, Alejandro EU, Luciani DS, Kalynyak TB, Hu X, Li H, et al. Carboxypeptidase E mediates palmitate-induced beta-cell ER stress and apoptosis. Proc Natl Acad Sci USA. 2008;105:8452–7.

21. Krajnik M, Schafer M, Sobanski P, et al. Enkephalin, its precursor, processing enzymes, and receptor as part of a local opioid network throughout the respiratory system of lung cancer patients. Hum Pathol. 2010;41:632–42. 22. Horing E, Harter PN, Seznec J, Schittenhelm J, Buhring HJ, Bhattacharyya S, et al. The ‘go or grow’ potential of gliomas is linked to the neuropeptide processing enzyme carboxypeptidase E and mediated by metabolic stress. Acta Neuropathol. 2012;124:83–97.