Prognostic Value of LISCH7 mRNA in Plasma and Tumor of Colon Cancer Patients

Tumor metastases are the final consequence of tumor progression and the most common cause of death in cancer patients. The first step in the development of metastasis is invasion of the adjacent tissue; then, in the second step, intravasation, the neoplastic cells enter the blood or lymph vessels, and some survive to the next step, extravasation at distal sites, where a subset of such metastasizing cells will form a tumor metastasis (1).

Experimental animal models have been successfully used to identify molecular elements during metastasis in melanoma and breast cancer (2, 3). Recently, Yang et al. (4) have studied a mouse mammary tumor model, in which a set of otherwise isogenic tumor cell populations completed distinct steps of metastasis when implanted into the mammary glands of BALB/c mice. One of the most up-regulated genes related to development of visible metastasis was LISCH7, liver-specific bHLH-Zip transcription factor, which is involved in the regulation of transcriptional processes. In addition, Kannan et al. (5) have proposed that LISCH7 could be a primary target of p53, although transactivation of LISCH7 by stress conditions is not discarded.

The search for new methods to detect metastatic or recurrent disease at preclinical or presymptomatic stages may contribute to improved patient survival. Cancer patients have higher concentrations of circulating nucleic acids in plasma and serum, and patients with metastasis have still higher levels (6, 7). Moreover, tumor alterations in DNA and RNA have been detected in plasma of patients with various types of cancer (8–18) and are associated with poor prognosis (14–17, 19–22). Nevertheless, mRNA-based amplification methods can offer greater specificity and sensitivity than DNA alteration detection (23). These data suggest the possibility of developing a noninvasive diagnostic and prognostic method by measurement of tumor RNA in plasma, which would be useful during the follow-up of patients.

Based on such evidence, we studied the presence of the LISCH7 mRNA in plasma from colorectal patients and healthy blood donors to examine its specificity for the disease and to search its possible prognosis value. Finally, the expression levels of this gene in primary tumors were also studied to infer a possible relationship with survival or the behavior of the disease, as well as its cellular origin.

Patients, samples, and RNA extraction. The study was approved by the research ethics board of our hospital and informed consent was obtained from all participants.

Between January 2005 and June 2006, blood samples (20 mL) were taken from 83 patients with colorectal carcinoma by venipuncture before intervention on the day of surgery. Blood samples from 20 healthy blood donors were also obtained at the hematology unit of our hospital. Plasma was prepared by centrifugation of peripheral blood at 2,500 rpm for 25 min and snap frozen at −80°C until processing. Tissue samples were also obtained from 20 of the colorectal carcinoma patients immediately after surgery, immersed in RNA later (Ambion, Inc.), snap frozen in liquid nitrogen, and stored at −80°C until processing. Normal tissues were taken from a noncancerous region of the colon mucosa at least 3 cm from the outer tumor margin.

Between June 1999 and December 2000, we recruited 95 patients diagnosed with colorectal cancer. They were included in a prospective study to evaluate overall survival (OS) and disease-free survival (DFS). Tissue samples from these patients were obtained, as mentioned above.

RNA of plasma samples was obtained from 1 mL of plasma by the Dynabeads mRNA DIRECT kit (Dynal Biotech ASA). RNA from tissue samples was extracted from ∼30 mg of tumor and normal samples, using RNeasy Mini kit (Qiagen, Inc.).

Epithelial cell selection/purification. Normal and tumor fresh samples from five patients were disaggregated in a Medimachine instrument (DAKOCytomation). Cell viability was evaluated by trypan blue exclusion, all (>95%) of the cells in this fraction being viable. Epithelial cells were immunomagnetically purified from 2 × 10⁶ total cells using superparamagnetic polystyrene beads coated with the Ber-EP4 antibody (Dynabeads Epithelial Enrich, Dynal Biotech ASA). mRNA was extracted from the epithelial cell–enriched fraction (>97%) and from the supernatant containing nonepithelial cells. Cellular nuclei were labeled with 4′,6-diamidino-2-phenylindole (Sigma). Monoclonal antibodies against cytokeratin 8, 18, 19, conjugated with FITC, were used to label and to visualize epithelial cells specifically (DAKOCytomation) under the fluorescence microscope.

Real-time PCR. Because LISCH7 mRNA was not detected in plasma of healthy donors, or in normal tissue from patients, two methods for expression analysis were used in the study. First, LISCH7 mRNA expression was measured only in plasma and tumor tissue samples in a relative quantification in samples in which its detection was possible. When LISCH7 expression was not detected in tumor tissues, half the minimum value detected in the series was assigned (this value was 0.01). Second, as there were some patients in whom LISCH7 mRNA in plasma was not detected, LISCH7 expression in plasma was also evaluated as present or absent.

LISCH7 mRNA expression in each sample was measured as a ratio against the geometric average of three reference housekeeping genes, succinate dehydrogenase complex subunit A, TATA binding protein, and ubiquitin C (24). The relative concentrations of the target and the reference genes were calculated by interpolation using a standard curve of each gene plotted from the same serial dilution of cDNA from tumor tissue. The quantitative mRNA analysis was done in duplicate. Primers are shown in Supplementary Table (Supplementary Materials and Methods). For the synthesis of cDNA, RNA was retrotranscribed using the Gold RNA PCR Core kit (PE Biosystems).

Real-time PCR was done in a LightCycler apparatus (Roche Diagnostics) using the LightCycler FastStart^PLUS DNA Master SYBR Green I kit (Roche Diagnostics).

Clinicopathologic variables. Clinicopathologic variables, listed in Table 1, were obtained from the medical records of the 178 patients. Pathologic stage was assessed using the tumor-node-metastases classification. Presence of lymph node metastases (LNM) was evaluated by optical microscopy.

Patient follow-up. Clinical follow-up after surgery and diagnosis was based on periodic visits and biochemical and imaging techniques (chest X-ray, endoscopy, bone scan, and other areas as clinically indicated). Ultrasonic study was done when liver function was impaired. OS and DFS were defined as the period from time of diagnosis until death and the interval between diagnosis and first recurrence, respectively.

Statistical analysis. Association between presence or absence of LISCH7 mRNA in plasma and presence or absence in healthy donors and clinicopathologic variables of the tumor or LISCH7 expression in tumors were contrasted by the χ² test.

LISCH7 mRNA data in plasma was not normally distributed (Kolmogorov-Smirnov test, Lilliefors correction). We standardized data distribution by using log₁₀ for statistical analysis and we also used the geometric average to describe expression gene data and association between LISCH7 mRNA in plasma and clinicopathologic characteristics were contrasted by the one-way ANOVA test.

LISCH7 expression data in tumors were not normally distributed (Kolmogorov-Smirnov test), although log₁₀ was used, so the geometric average was used to describe the gene expression, and the Kruskal-Wallis test was applied to contrast expression data with clinicopathologic variables.

To study OS and DFS, the expression data of LISCH7 were divided by tertiles. The expression levels defining the three groups were 0.05 (33%) and 2.24 (66%). The relationship between the cumulative probability of OS and DFS, as well as analyzed predictors, was calculated following the Kaplan-Meier method (25), whereas significant differences between curves were evaluated with Mantel's log-rank test (26). To identify factors that might be of independent significance in influencing OS and DFS, multivariate analysis (Cox proportional risk regression model) was applied (27). ‘Confuser’ and interacting variables were analyzed. The model's basic assumptions (proportional hazards) were evaluated.

In all statistical tests, two-tailed P ≤ 0.05 was considered statistically significant. Statistical analysis was done using version 11.0 of the SPSS.

LISCH7 expression in plasma and tumor samples. Of the 83 plasma samples analyzed from colorectal cancer patients, 25 (30.1%) showed presence of LISCH7 mRNA. However, it was not detected in any of the 20 plasma samples from healthy donors (P = 0.002). The quantification of LISCH7 mRNA in plasma showed a median of 1.45 (minimum, 0.05; maximum, 9.55).

There were 115 tumor and normal counterpart tissue samples. LISCH7 expression was not detected in any normal samples. However, 80 (69.6%) adenocarcinomas showed LISCH7 expression. The quantification of LISCH7 expression in tumor tissue showed a median expression of 2.46 (minimum, 0.02; maximum, 28.00).

There were 20 colorectal cancer patients in whom tumor and plasma samples were available. Among these 20 patients, only 1 was a stage I patient and both plasma and tumor tissue were LISCH7 mRNA negative; 10 were patients at stage II, in 7 of whom LISCH7 mRNA was detected in tissues samples and 2 plasmas of these showed positive presence of the target gene; 3 cases of the 8 patients at stage III had LISCH7 mRNA in tumors and 2 of them also had it in plasma. Finally, there was 1 patient at stage IV, and LISCH7 mRNA was only detected in the plasma sample. In other words, of the 5 patients with LISCH7 mRNA in plasma, 4 also showed it in tumor tissues.

LISCH7 cellular expression origin. We tested the possibility that the expression of LISCH7 could correspond to stromal cells rather than carcinoma cells in the biopsies. Nonepithelial cytokeratin-negative cells constituted 5% to 15% of the whole tumor sample disaggregates but were absent in epithelial-enriched fractions following separation with paramagnetic beads (Fig. 1). LISCH7 mRNA was detected specifically in epithelial fraction from three of the five tumoral samples (data not shown).

LISCH7 mRNA and pathologic data in plasma samples. Presence of LISCH7 mRNA in plasma was associated with stage, LNM, and vascular invasion (Table 1). A difference toward statistical significance was observed for the tumor side because the presence of LISCH7 mRNA was greater in patients with a rectum cancer.

In the plasma samples in which LISCH7 mRNA was detected, the amounts of it were contrasted with pathologic variables, and some associations were found. Thus, it was associated with presence of LNM (Fig. 2C; Table 1). However, no association was found with vascular invasion. For the stage, a difference close to statistical significance was found (P = 0.056), but this difference was clearly significant when the stages were grouped in early (I + II) and advanced (III + IV) stages (P = 0.005; Fig. 2A and B; Table 1).

LISCH7 expression and pathologic data in tumor samples. In tumors, an association between up-regulation of LISCH7 expression and vascular invasion was observed (Fig. 2D; Table 1). Although no association was found between LISCH7 expression and stage, a similar scenario to those for plasma LISCH7 mRNA was observed. Therefore, there were 7 (58%) patients with LISCH7 expression at stage I, 39 (71%) with it at stage II, 29 (69%) with it at stage III, and, again, the highest proportion of patients with LISCH7 expression was at stage IV, 5 (83%).

Overall survival. The 5-year OS for patients was 63% [95% confidence interval (95% CI), 52-74%].

Patients were divided into tertiles based on LISCH7 expression. When OS was analyzed for LISCH7 expression, a significant difference was observed (P = 0.018; Fig. 3I, A). The Kaplan-Meier graph revealed similar behavior of the medium- and high-level tertiles. Thus, these patients were grouped as a high expression group and additional analysis of LISCH7 expression was carried out with only two categories, low- and high-expression levels of LISCH7. In this analysis, the significant difference is clearer than in the previous one (P = 0.008; Fig. 3I, B).

The univariate analysis showed that variables that could be regarded as statistically supported factors in OS prediction were sex of patients, vascular invasion, LNM, stage (III and IV versus I and II), and LISCH7 expression. Multivariate analysis showed that sex, stage, and LISCH7 expression [hazard ratio (HR), 3.45; 95% CI, 1.19-9.98] had independent relations to OS (Table 2A).

The correlation between overall survival and LISCH7 expression was analyzed separately in stages I + II and in III + IV. A significant difference was found for LISCH7 expression at stages III + IV (P = 0.043), but not at stages I + II (Fig. 3II, A and B).

The multivariate analysis with the series stratified according to stage confirmed that LISCH7 expression had clearer independent relations to OS at stages III + IV (HR, 3.63; 95% CI, 0.83-15.89; Table 3A) than at stages I + II.

Disease-free survival. During follow-up, 32 (33.7%) relapses were observed.

No clear differences in DFS for LISCH7 expression were observed (Fig. 3I, C).

The univariate analysis showed that sex, vascular invasion, LNM, and stage could be regarded as statistically supported factors in DFS. Multivariate analysis showed that sex and stage had independent relations with DFS (Table 2B).

Although the Kaplan-Meier test revealed no difference in DFS for LISCH7 expression in stage groups, the same tendency that LISCH7 expression showed better OS prognosis value in advanced than in early stages was seen (Fig. 3II, C and D). The multivariate analysis according to stage did not confirm this fact. However, the same tendency could be observed. Thus, LISCH7 expression had a HR of 0.95 (95% CI, 0.30-3.03; P = 0.921) at stages I + II and 2.42 (95% CI, 0.54-10.92; P = 0.251) at stage III (Table 3B).

The presence of LISCH7 mRNA in plasma from patients with colorectal cancer was studied to validate its disease-predictive value because it could then be used as a noninvasive tool in prognosis. An association between the presence of extracellular plasma RNA of specific genes and poor prognosis has been reported in some cancer types (14, 16, 17). Our results are consistent with these findings because LISCH7 mRNA in plasma was associated with presence of LNM, advanced stages, and vascular invasion, as well as a trend to tumor side in rectum, all of which suggest that it could be used as a prognosis factor.

LISCH7 mRNA was detected in 30.1% of plasmas from cancer patients. However, neither plasma from healthy control nor normal tissue samples showed the presence of LISCH7 mRNA. These results suggest that LISCH7 mRNA detected in plasma from patients with CC could have its origin in epithelial tumor cells, as it was observed with RNA isolated from the disaggregate tumor tissue and thus could be a specific tumor marker.

Interestingly, among patients with tumor and plasma samples, it was observed that LISCH7 mRNA was detected in plasma from a cancer patient in whom expression of LISCH7 was not detected in tumor tissue. We showed in a previous study that the presence of heterogeneous clones in the tumor samples could explain some nonmatched alterations between plasma and tumor (28). Additionally, in the current series, we observed that the detection of LISCH7 in plasma is increasingly related to the presence of LISCH7 in advanced-stage tumors. Thus, the LISCH7 mRNA detected in plasma may be related to metastases because the up-regulation of this gene has been documented in cells able to develop metastasis, at least in animal models (4).

The follow-up of plasma samples was too short to carry out a survival study. However, our results in a long series of human colon cancer support the previous results because expressions of LISCH7 were associated with a poor outcome of the disease and with more aggressive pathologic characteristics.

The finding that high levels of LISCH7 predicted OS but not DFS in patients suggests that this gene could specifically confer advantage on metastasis enlargement (4). In fact, LISCH7 expression levels are more clearly associated with poor prognosis for OS in advanced stages than in early stages, and the same tendency was observed for DFS. Therefore, for advanced stages (stage III in this case), 5-year DFS ran at 75% in patients with low LISCH7 expression levels versus 40% in patients with high levels. This marked difference between the two groups was not statistically significant, probably due to the insufficient number of patients included in the study.

An association between the presence of LISCH7 mRNA in plasma and advanced stages was found, but not clearly between the LISCH7 expression in tumors and advanced stages. This might be due to the distribution of the patients in the series, as the plasma series contained proportionally more patients at stage IV than the tumor series did. However, it might also be because LISCH7 mRNA in plasma is detected mainly at advanced stages. In this context, the detection of LISCH7 mRNA in plasma could be a better prognostic tool than tumor samples because plasma data could be the surrogate for the LISCH7 mRNA of primary tumor as well as of the metastatic tumor cells. Thus, hypothetically, it could have a possible clinical use as a serologic marker during the outcome of the patients.

Given the association between LISCH7 mRNA in plasma and several poor prognosis factors and because LISCH7 expression in tumors clearly influences disease outcome and given the high association between positive plasma and tumors, the value of LISCH7 mRNA in the plasma of colorectal cancer patients can be considered a marker of poor prognosis and one that can be obtained by a noninvasive method.

We thank M. Eaude for help with the English manuscript and I. Alonso for blood sample collection.