SEARCH

SEARCH BY CITATION

Keywords:

  • RAD51;
  • DNA repair;
  • EGFR;
  • colorectal adenocarcinoma;
  • prognostic marker

Abstract

  1. Top of page
  2. Abstract
  3. Patients and Methods
  4. Results
  5. Discussion
  6. Acknowledgements
  7. References
  8. Supporting Information

RAD51 is the central protein in the homologous recombination pathway and is therefore of great relevance in terms of both therapy resistance as well as genomic stability. By using a tissue microarray analysis of 1,213 biopsies taken from colorectal adenocarcinomas (CRCs), we investigated whether RAD51 expression can be used as a prognostic marker as well as potential associations between this and the expression of other proteins known to be related to CRC. Strong RAD51 expression was observed in 1% of CRC, moderate in 11%, weak in 34% and no expression in 44%. No correlation was found between RAD51 expression and clinicopathological parameters. RAD51 expression correlated significantly (p = 0.001) with overall survival, with a median survival of 11 months for patients with strong, 46 with moderate, 76 with weak and 68 with negative expression. Multivariate analyses revealed that in addition to tumor stage (p < 0.0001) and nodal status (p < 0.0001), RAD51 expression is also an independent prognostic parameter (p = 0.011). Strong RAD51 expression was found to be associated with the loss of the two DNA mismatch repair proteins MSH (p = 0.0003), MLH (p = 0.002) and β-catenin (p = 0.012) as well as with elevated p21 (p = 0.003) and EGFR expression (p = 0.0001). However, a correlation with overall survival could only be found for EGFR expression (p = 0.008), although no added benefit in risk stratification could be determined when evaluated together with RAD51. Overexpression of RAD51 is a predictor of poor outcome in CRC. This finding indicated the promise of future studies using RAD51 as a prognostic marker and therapeutic target.

Colorectal cancer (CRC) is one of the most frequent cancers, with over 1 million new cases worldwide each year and a disease-specific mortality of ∼33–40%.1 Although treatment has advanced in recent years, the disease-specific mortality remains unchanged. Therefore, identifying CRC patients who could potentially benefit from a given therapy continues to remain an important research goal. Numerous biomarkers have already been tested to this end.2 About 15% of all CRC3 are characterized by microsatellite instability (MSI), indicating a defect in mismatch repair (MMR). These patients benefit notably when treated by chemotherapeutic drugs.2 At present, the only other robust indicator of CRC prognosis justifying routine clinical assessment is KRAS mutational analysis, which can be used to select CRC patients for an EGFR-targeting therapy.1, 4, 5

Hence, clinical and pathologic staging information are the only data currently used to determine prognosis and to select CRC patients to receive adjuvant chemotherapy. However, the considerable stage-independent variability in patient outcomes indicated the need for prognostic markers identifying high-risk patients who may benefit from adjuvant chemotherapy and beyond this, from new targeted therapies. The discovery of new targeted therapies has offered dramatic improvements in therapeutic oncology over the last several decades. Particularly, with regard to DNA repair, PARP inhibitors show promise as a powerful therapeutic tool, especially in the management of tumors deficient in homologous recombination (HR).6 Consequentially, other proteins involved in the regulation and execution of DNA repair might constitute additional new potential prognostic markers.

RAD51 plays a central role in the repair of DNA double-strand breaks (DSBs) performed by HR.7 After the induction of specific DNA damage, RAD51 localizes to nuclear foci that represent sites of DNA repair. As part of HR, RAD51 facilitates strand transfer between interrupted sequences and their undamaged homologies.8 Several studies have shown that the level of RAD51 protein expression is elevated in immortalized cells as well as in a wide variety of human cancer cell lines.9, 10 It is generally suggested that RAD51 overexpression results in an increased cellular resistance to radiation and some chemotherapeutic drugs such as topoisomerase inhibitors or crosslinking agents.10–12

RAD51 overexpression has also been investigated in numerous tumor tissues using immunohistochemistry (IHC). Clear overexpression has been observed for breast, pancreatic, head and neck, lung and esophageal squamous cell carcinomas as well as for gliomas.13–18 Additionally, RAD51 overexpression was associated with poor prognosis in most of these studies,13, 15, 19, 20 with only two studies showing the opposite effect.17, 18

Our study offers the first immunohistochemical analysis of RAD51 expression in CRC. The study was performed using a tissue microarray (TMA) containing specimens resected from 1,210 CRCs. We also investigated the extent to which RAD51 expression is associated with that of other proteins suggested to be of relevance for CRC. This analysis was performed for MSH and MLH, both required for DNA MMR, as well as for p53, p21 and c-myc, all involved in cell cycle regulation; BCL2, relevant for apoptosis; β-catenin as part of the Wnt-pathway, and HER2 and EGFR, both involved in growth factor signaling.21 The overall aim of our study was to test whether RAD51 expression can be used as a prognostic parameter for CRC and how this prognostic sensitivity can be improved when combined with the expression of the other proteins. Such a study might help not only to establish further prognostic markers for CRC but also to identify new potential therapeutic targets.

Patients and Methods

  1. Top of page
  2. Abstract
  3. Patients and Methods
  4. Results
  5. Discussion
  6. Acknowledgements
  7. References
  8. Supporting Information

Patients, clinical samples and microarray construction

Two different TMAs with a total of 1,800 CRC samples, one manufactured from resection specimens from 1,420 patients at the University Hospital of Basel and one created using specimens from 380 CRCs at the University Medical Center Hamburg-Eppendorf, were investigated. Of these 1,800 tumor biopsies, only the 1,213 adenocarcinomas were chosen for survival analysis.

TMA construction was completed as described in Ref.22. Patient information and clinical data such as age, sex, pTNM stage and carcinoma grade were retrospectively retrieved from clinical and pathological databases (Table 1). Follow-up data were obtained from local cancer register boards or through the attending physicians.

Table 1. Relation between RAD51 expression and clinicopathological parameters
inline image

Immunohistochemistry

Freshly cut TMA sections were used for all immunostains. Four-micrometer sections were dewaxed in xylene and dehydrated in a descending ethanol series. Antigen retrieval was performed in an autoclave in TRIS-EDTA-citrate buffer adjusted to pH 7.8. The primary antibodies were diluted as follows: RAD51 (US Biological, Swampscott, MA) 1:150; MLH1 (Zytomed, Berlin, Germany) 1:20; MSH2 (Life Technologies, Carlsbad, CA) 1:50; β-catenin (Life Technologies, Carsbad, CA) 1:1,000; HER2 (DAKO, Hamburg, Germany) 1:100; EGFR (DAKO, Hamburg, Germany) 1:100; BCL2 (Dako) 1:250; p53 (Calbiochem, Darmstadt, Germany) 1:2,000, and incubated for 2 hr at room temperature. Diaminobenzidine was used as a chromogen. The primary antibody was omitted for the negative controls. A single pathologist evaluated all the IHC experiments. Staining intensity was estimated by visual inspection and documented using a four-step scale (0, 1, 2 and 3) and the fraction of positive stained tumor cells from 0 to 100%. A final IHC score was composed using these two parameters (intensity and % positive tumor cells) according to the following criteria: tumors without any detectable staining were rated as “negative”; tumors with 1+ staining intensity in ≤70% of tumor cells, or tumors with 2+ staining intensity in ≤30% of tumor cells were rated as “weak”; tumors with 1+ staining intensity in >70% of tumor cells, or 2+ staining intensity in >30% but ≤70% tumor cells, or tumors with 3+ staining intensity in ≤30% of tumor cells were rated as “moderate” and tumors with 2+ staining intensity in >70% of tumor cells, or tumors with 3+ staining intensity in >30% tumor cells were rated as “strong.”

Statistical analysis

Contingency table analyses and the chi-square (likelihood) test were used to study the relationship between the IHC results and the morphological parameters. The Kaplan–Meier method was used for survival analysis, whereas the log-rank test was used to test for significant associations between stratified survival functions. Multivariable Cox regression analysis was used to show prognostic relevance of RAD51 expression after adjusting for morphological parameter or other biomarker. A p-value of <0.05 was regarded as significant. JMP 8.0 software (SAS Institute, Cary, NC) was used for data analysis. Correlations between markers were computed by means of Spearman's test.

Results

  1. Top of page
  2. Abstract
  3. Patients and Methods
  4. Results
  5. Discussion
  6. Acknowledgements
  7. References
  8. Supporting Information

RAD51 immunohistochemistry

A total of 1,213 tumor biopsies from patients suffering from colorectal adenocarcinomas (CRCs) were informative in determining the patterns and intensities of RAD51 staining by IHC. Representative examples of negative, weak, moderate or strong staining are shown in Figures 1a1d. Overall, 54% of the samples were classified as negative for RAD51 (n = 651), 34% showed a weak staining (411), 11% a moderate (139) and only 1% a strong staining (12) (Fig. 1e).

thumbnail image

Figure 1. Representative examples (ad) and the distribution (e) of RAD51 protein expression in 1,210 biopsies from patients with CRCs. Biopsies exhibiting negative (a), weak (b), moderate (c) and strong (d) RAD51 staining were classified and plotted according to RAD51 expression (e). Biopsies were counterstained with hematoxylin.

Download figure to PowerPoint

The relationship between RAD51 expression, tumor phenotype and the various clinical parameters was analyzed (Table 1). Demographic, clinical and histopathological data (including age, sex, TNM stage and tumor grade) were collected retrospectively. The median age of the patients was 67.9 years (range from 30 to 90 years) at time of diagnosis; 52% of the patients were female and 48% were male, and the mean survival time was 51.3 months. Most of the biopsies were moderately graded (G2) T3 tumors without regional node involvement (N0). No significant correlation was observed between RAD51 expression and age (p = 0.649), tumor stage (p = 0.1155) or nodal status (p = 0.7363). However, a nearly significant correlation with tumor grade could be identified (p = 0.051) (Table 1).

RAD51 expression and clinical outcome

RAD51 expression was analyzed with regard to the overall survival time of the patients (Fig. 2). Patients whose tumors displayed high-level RAD51 expression showed a shorter median survival time of 11 month compared to patients with moderate (46 months), weak (76 months) or no (68 months) RAD51 expression. Univariate survival analyses (log-rank test) demonstrated a significant association between overall survival and RAD51 status (p = 0.0013), with p = 0.0049 for strong and p = 0.0188 for moderate RAD51 expression compared to no expression.

thumbnail image

Figure 2. Kaplan–Meier survival analysis of 1,209 patients with CRCs according to RAD51 expression (log-rank test). Overall survival is plotted against time after therapy.

Download figure to PowerPoint

As variables found to have prognostic influence in univariate analyses might covariate, all statistically significant variables from the univariate analysis were included in the multivariate regression analysis to identify independent prognostic factors. RAD51 expression proved to be an independent prognostic factor (p = 0.011), as did tumor stage and nodal status (p < 0.0001) (Table 2), whereas grading did not.

Table 2. Cox regression multivariate analysis
inline image

RAD51 expression and other biomarkers

We also tested the association between RAD51 expression and that of other proteins known to be relevant for the outcome of CRCs: the two DNA MMR proteins MSH and MLH were used as markers for MSI, p53 and c-myc as indicators of stress or oncogene-induced transcriptional changes, p21 as a marker of cell cycle arrest, BCL2 to monitor the regulation of apoptosis, β-catenin as a marker for the Wnt pathway, an integral pathway for colorectal tumorigenesis, as well as the two tyrosine growth factor receptors HER2 and EGFR (Supporting Information Table 1 and Fig. 3).23, 24 Immunohistochemical staining of MSH, MLH, p53, c-myc, p21, BCL2, β-catenin, HER2 and EGFR was informative for 808, 808, 806, 800, 804, 806, 771, 809 or 775 cases, respectively.

thumbnail image

Figure 3. Relationship between RAD51 expression and other protein markers. Samples were classified as either MSH (a), MLH (b), β-catenin (c), p21 (d) and EGFR (e) negative (open bars) or positive (black bars) and plotted against negative, weak, moderate or strong RAD51 expression. Correlations between markers were computed by means of Spearman's test.

Download figure to PowerPoint

RAD51 expression showed no correlation with the expression of p53, c-myc, BCL2 or HER2 (Supporting Information Table 1). These markers failed to reveal an association with the overall survival of CRC patients (Supporting Information Figs. S1A–D). In contrast, a significant correlation was observed between RAD51 expression and the loss of the two DNA MMR proteins MSH, MLH and β-catenin, as well as with the rise in expression of p21 and EGFR (Supporting Information Table 1 and Figs. 3a3e). Overall, these data indicate that strong RAD51 expression appears to be associated with the downregulation of the two DNA MMR proteins MSH, MLH and β-catenin as well as with the upregulation of p21 and EGFR, the latter of which is especially pronounced in CRCs. Surprisingly, for the majority of these markers, no significant association was observed with the overall survival of CRC patients as obtained from the Kaplan–Meier curve analysis coupled with the log-rank test [Supporting Information Figs. S2A–D; MSH (p = 0.5451), MLH (p = 0.8283), p21 (p = 0.3183) and β-catenin (p = 0.5375)].

EGFR expression and clinical outcome

In addition to RAD51, EGFR expression appears to influence CRC prognosis, showing a significant association with the overall survival of CRC patients (p = 0.0078) (Fig. 4a). Expression of EGFR correlated with a shorter median survival time compared to negative expression (50 vs. 78 months), and was in this way comparable to RAD51 expression (Fig. 2). These data demonstrate that of all markers tested, only the expression of EGFR and RAD51 appears to be associated with the overall survival of CRC patients. Given the strong association between RAD51 and EGFR (Fig. 3e), cases were grouped based on their expression of these markers. For this combination, the RAD51 and EGFR data for 775 patients were grouped as follows: patients negative for either proteins or weak RAD51 expression (n = 520, 67%), patients positive for one protein (n = 216, 28%) and patients with both markers strongly positive (n = 39, 5%). A significant association with overall survival (Fig. 4b) was observed; cases that were positive for both markers showed the shortest median overall survival of 42 months compared to 52 months for patients positive for one and 84 months for those negative for both marker or weak RAD51 expression (p = 0.0026).

thumbnail image

Figure 4. Kaplan–Meier survival analysis according to EGFR alone (a) and EGFR together with RAD51 (b). Samples were classified as either EGFR negative or positive (a), or as negative for both EGFR and RAD51, positive for one of these two proteins or positive for both proteins (b). Overall survival is plotted against time after therapy. Significance is indicated by log-rank.

Download figure to PowerPoint

Again, multivariate regression analyses were performed to identify independent prognostic factors. Both RAD51 and EGFR proved to be independent prognostic factors (p = 0.0126 and 0.0155) (Supporting Information Table 2). By far the strongest increase in risk ratio from 0.66 (0.51–0.9 95% CI) to 3.39 (1.4–6.46 CI) was found for negative vs. strong RAD51 expression compared to EGFR with 0.87 (0.79–0.97 CI) to 1.14 (1.03–1.27 CI). Therefore, RAD51 expression alone remains the strongest parameter for survival in CRC, indicating no added benefit in risk stratification in this series when evaluating double negativity (RAD51 and EGFR) compared to single negativity and effect on outcome. In fact, it appears that the less favorable prognosis observed for cases of strong RAD51 expression becomes negligible when combined with EGFR (for comparison, see Fig. 2).

Impact of EGFR and RAD51 expression on tumor therapy

RAD51 and EGFR expression was analyzed with regard to tumor therapy. First, the impact of both RAD51 and EGFR overexpression on the subgroup of rectal carcinomas (ReCa), which are predominantly treated by radiotherapy, was analyzed (Figs. 5a and 5b). For this subgroup, high RAD51 expression was found to be of similar relevance as for all CRCs (Fig. 5a, p = 0.0022 vs. Fig. 2, p = 0.0013), whereas a slightly greater significance was detected for EGFR expression (Fig. 5b, p = 0.0022 vs. Fig. 4a, p = 0.0078). Second, the hypothesis that RAD51 overexpression can affect tumor prognosis both through increased therapy resistance as well as the elevated risk of metastasis was proven. Patients were grouped according to predominantly treated with surgery alone (pT1-3, pN0) (Fig. 5c) or patients treated with surgery plus chemotherapy (pT1-3, pN1-3 and T4 pN0-3) (Fig. 5d). RAD51 expression was found not to have significant effect on prognosis for lymph node-negative patients treated by surgery alone (p = 0.24), in contrast to lymph node-positive CRCs for which the risk of metastasis is high; RAD51 expression was found to have a rather strong effect on prognosis (p < 0.0001).

thumbnail image

Figure 5. Kaplan–Meier survival analysis according to RAD51 (a, c, and d) or EGFR (b) expression. Rectal carcinomas were classified according to RAD51 (a) or EGFR (b) expression or CRCs were grouped in pT1-3, pN0 tumors (c), pT1-3, pN1-3 and pT4 N0-3 tumors (d), and plotted against overall survival. Significance is indicated by log-rank.

Download figure to PowerPoint

Discussion

  1. Top of page
  2. Abstract
  3. Patients and Methods
  4. Results
  5. Discussion
  6. Acknowledgements
  7. References
  8. Supporting Information

In our study, we investigated the impact of RAD51 expression on the outcome of CRCs. We observed that 46% of CRCs showed an enhanced expression of RAD51. Similar percentages were found for other tumor entities.13–16, 18, 20 The large number of tumor samples represented in our TMA (n = 1202) also allowed for a differential analysis of RAD51 expression, indicating that the majority (34%) of CRCs are characterized by a weak, 11% by a moderate and only 1% by strong staining. A high level of RAD51 expression appears to correlate with advanced histological grading in CRCs (Table 1), a finding that has previously been reported for breast cancer.14, 17

The most important result reported here is the fact that RAD51 expression is clearly an independent prognostic marker for CRC prognosis (Fig. 2). CRC patients with strong RAD51 expression showed a median survival time, which was significantly shorter than that of patients with moderate, low or no RAD51 expression. Similar observations have previously been made for breast, pancreatic, lung, head and neck cancers as well as for esophageal squamous cell carcinomas.13, 15, 19, 20 Only for glioblastomas and in one study also for breast cancer could a positive impact of high RAD51 expression on tumor prognosis be determined.17, 18 In the case of the latter study, it might be important to note that our study was performed with premenopausal patients with early-onset breast cancer. It is likely that genetic factors determining the early onset of this cancer negate the impact of RAD51 overexpression. Overall, these data demonstrate that high RAD51 expression is generally a negative prognostic marker for tumor outcome.

For tumor patients, negative outcomes arise either from failed local tumor control or from the occurrence of local or distant metastases. One study of breast cancer patients treated by combined radiochemotherapy demonstrates that the poor overall survival seen for tumors with high RAD51 expression may result from decreased local tumor control.19 Such an association may also exist for the CRCs analyzed in our study. This tumor entity is often treated by chemotherapy using either FOLFIRI (5-FU, Leucovorin and Irinotecan) or FOLFOX (5-FU, Leucovorin and Oxaliplatin).25 Both chemotherapy combinations induce S-phase-dependent DNA damage. If this damage interferes with DNA replication, new DSBs may be formed which are then repaired predominantly by HR with RAD51 as the key player.26 Several in vitro studies have shown that an increase in RAD51 expression stimulates HR, resulting in greater cellular resistance to treatment with crosslinking agents, etoposide or irradiation.10–12 High numbers of RAD51 foci in tumor biopsies were also positively associated with greater chemoresistance in breast cancer patients.27

The stimulation of HR by RAD51 overexpression has also been reported to result in an increased genomic instability as detected either by a larger number of chromosomal translocations28 or genomic rearrangements.29 These observations can be considered to explain the accelerated tumor progression detected for myelomas or in Barrett's syndrome overexpressing RAD51.30, 31 The elevated genomic instability is thought to result from disturbed replication caused by RAD51.29 As genomic instability is also known to result in therapy resistance, these data suggest that RAD51 overexpression may result in therapy resistance because of both better repair as outlined above as well as through its effect on genomic instability.

However, it should also be noted that RAD51 overexpression is also associated with poor prognosis, even when cancer patients are treated solely by surgery, as reported for esophageal tumors.20 These data strongly suggest that RAD51 expression also has an impact on the risk of metastasis, although direct data are lacking thus far. As it is generally accepted that chromosomal/genetic instability results in an enhanced risk of metastasis,32 RAD51 overexpression may increase this risk by causing greater chromosomal instability.

Overall, these data indicate that RAD51 overexpression can affect tumor prognosis both through increased therapy resistance as well as the elevated risk of metastasis, a notion supported by the data obtained here (Figs. 5c and 5d). For CRCs treated predominantly with surgery alone and for which the risk of metastasis is low, RAD51 expression was found not to have significant effect on prognosis, in contrast to CRCs treated primarily using chemotherapy and for which the risk of metastasis is high, where RAD51 expression was found to have a rather strong effect on prognosis.

The data obtained here also raise questions concerning the pathological role of RAD51 in tumorigenesis and tumor progression. RAD51 is essential for genetic recombination via HR during embryogenesis, with the elimination of RAD51 leading to early embryonic lethality.33 RAD51 expression is typically not observed in most normal differentiated tissues as reported for esophagus, pancreas, breast as well as head and neck tissues.13, 14, 16, 19, 20 Only in actively proliferating tissues such as testis, thymus and stem cells was active RAD51 expression detected.20, 34 This might partially explain the elevated RAD51 protein expression seen in tumors, which arise mostly from proliferating cells. However, the manner in which some tumors are able to develop an extreme upregulation of RAD51 remains to be elucidated. It should be noted that this RAD51 overexpression is not a simple surrogate of enhanced proliferation, because this correlated neither with Ki-67 nor with PCNA expression.14, 16, 18 Similarly, high RAD51 expression in a specific tumor sample is generally seen in nearly all the tumor cells and not merely in a subgroup of cells (Fig. 1d).

We also tested the extent to which RAD51 is associated with other proteins known to be of relevance for the prognosis of CRC. No association was found for p53, c-myc, BCL2 or HER2 expression (Supporting Information Table 1). These proteins were also found not to be associated with prognosis (Supporting Information Fig. 1), findings that are in line with most of the other data obtained thus far.1, 35 For RAD51, significant associations were found with p21, the two DNA MMR proteins MSH, MLH, and β-catenin. However, none of these proteins was found to be associated with prognosis (Supporting Information Fig. 2), as seen in other reports for p21 and β-catenin.1, 35 In contrast, there are data for MSH and MLH1, 36 indicated that CRCs deficient in MMR, i.e., lacking MSH or MLH, have a better prognosis when treated by chemotherapy. As RAD51 expression correlates inversely with MSH and MLH expression (Fig. 3), it might be speculated that the lack of association seen here for MSH and MLH is because of compensation the better prognosis by the negative prognosis due to RAD51 overexpression. The compensatory upregulation of RAD51 expression in cells deficient in the DNA MMR pathway proteins MSH and MLH has also been observed in vitro.37–39

The strongest correlation with RAD51 expression was seen for EGFR (Supporting Information Table 2 and Fig. 3). EGFR was also the only other protein that showed a significant association with prognosis (Fig. 4a). There was, however, no further improvement in prognosis when the expression of RAD51 and EGFR was combined into a single score (Fig. 4 and Supporting Information Table 2), a finding compatible with the strong correlation seen for the two individual proteins.

A strong correlation of both RAD51 and EGFR overexpression was also observed for the subgroup of ReCa, which are predominantly treated by radiotherapy (Figs. 5a and 5b). For ReCa, high RAD51 expression was found to be of similar relevance as for all CRCs (p = 0.0022 vs. Fig. 2, p = 0.0013), whereas a slightly greater significance was detected for EGFR expression (p = 0.0022 vs. Fig. 4a, p = 0.0078). A significant association was also found by others when analyzing ReCa for EGFR expression.20, 40 These results indicate that RAD51 expression is of importance not only when tumors are treated by chemotherapy but also in terms of response to radiotherapy. In addition, the greater relevance of EGFR in the context of radiotherapy is in line with our recent data showing a strong increase in cellular radioresistance with increasing EGFR expression.41

The association seen here for RAD51 and EGFR has also been observed in other studies. When EGFR signaling is blocked, a clear reduction in both RAD51 protein expression as well as in RAD51-dependent HR can be noted.42–44 In addition, the persistent upregulation of EGFR signaling was also found to result in elevated HR activity in cells harboring EGFRvIII variants.45 The protein linking EGFR and RAD51 is likely PTEN, which is part of the intracellular signal cascade driven by receptor tyrosine kinases and is also active as a transcriptional regulator of the Rad51 gene.46

The observation made here that RAD51 expression appears to be a prognostic marker also suggests the possibility of new therapeutic concepts. The downregulation of RAD51 by RNAi or specific inhibitors may be used to sensitize tumors to radiation or chemotherapy as has already been demonstrated in cell culture.47 In addition, RAD51 expression may also be used to identify patients susceptible for new targeted therapies, particularly those through which secondary DSBs are induced or checkpoint signaling is inhibited as may be achieved by PARP1 or Chk1 inhibition.48, 49

In summary, our study introduces RAD51 expression as an independent prognostic factor in CRC, with elevated levels of RAD51 expression predicting poor outcome. Future studies are necessary to clarify whether this effect is due to enhanced therapy resistance or rather an elevated risk of metastasis.

Acknowledgements

  1. Top of page
  2. Abstract
  3. Patients and Methods
  4. Results
  5. Discussion
  6. Acknowledgements
  7. References
  8. Supporting Information

The authors thank all patients for their contribution to this project as well as Alexandra Zielinski for the excellent technical assistance.

References

  1. Top of page
  2. Abstract
  3. Patients and Methods
  4. Results
  5. Discussion
  6. Acknowledgements
  7. References
  8. Supporting Information

Supporting Information

  1. Top of page
  2. Abstract
  3. Patients and Methods
  4. Results
  5. Discussion
  6. Acknowledgements
  7. References
  8. Supporting Information

Additional Supporting Information may be found in the online version of this article.

FilenameFormatSizeDescription
IJC_27907_sm_SuppFigs.ppt1320KSupporting Information Figures
IJC_27907_sm_SuppTabs.ppt339KSupporting Information Tables

Please note: Wiley Blackwell is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.