Article Text

Download PDFPDF

Prognostic value of phosphorylated STAT3 in advanced rectal cancer: a study from 104 French patients included in the EORTC 22921 trial
  1. Franck Monnien1,2,
  2. Harouna Zaki1,
  3. Christophe Borg3,4,
  4. Christiane Mougin2,5,
  5. Jean-François Bosset2,7,
  6. Mariette Mercier2,6,
  7. Francine Arbez-Gindre1,
  8. Bernadette Kantelip1,2
  1. 1Department of Pathology, Besançon University Hospital, Besançon, France
  2. 2EA 3181, IFR133, University of Franche-Comté, Besançon, France
  3. 3Department of Oncology, Besançon University Hospital, CIC-BT506, Besançon, France
  4. 4INSERM U645, University of Franche-Comté, Besançon, France
  5. 5Department of Cell and Molecular Biology, Besançon University Hospital, Besançon, France
  6. 6Department of Clinical Research in Cancerology, Besançon University Hospital, Besançon, France
  7. 7Department of Radiotherapy, Besançon University Hospital, Besançon, France
  1. Correspondence to Franck Monnien, Service d'Anatomie et Cytologie Pathologiques, Centre Hospitalier Universitaire Jean Minjoz, 2 Boulevard Fleming, 25030 Besançon Cedex, France; fmonnien{at}


Background Signal transducer and activator of transcription 3 (STAT3) has been implicated as an oncogene in several neoplastic diseases. However, the biological effects of STAT3 have not been extensively studied in rectal carcinogenesis.

Aims To evaluate STAT3 activation in advanced rectal cancers and its association with clinicopathological variables and prognosis.

Methods Nuclear immunohistochemical expression of phosphorylated STAT3 (p-STAT3) was studied in 104 advanced rectal cancers (T3–T4). All patients were participating in the EORTC 22921 trial to assess whether preoperative chemoradiotherapy followed by postoperative chemotherapy improved overall and progression-free survival.

Results Nuclear p-STAT3 expression was detected in 37.5% of rectal cancer patients. No correlation was observed between p-STAT3 and any clinicopathological variables tested. However, patients with tumours positive for p-STAT3 had significantly improved overall survival.

Conclusion These results highlight an unexpected role for nuclear p-STAT3 expression in advanced rectal cancers and need further investigation to clarify this finding.

  • Rectal cancers
  • STAT3
  • immunohistochemistry
  • survival analysis
  • colorectal cancer
  • oncogenes

Statistics from


In France there are approximately 12 000 new cases of rectal cancers per year.1 2 Preoperative chemoradiotherapy, which is established as standard treatment for locally advanced disease,3–5 has been reported to result in a 5-year local recurrence rate of 8%, distant metastasis rate of 34% and overall 5-year survival of 65%.4 Many prognostic factors exist, some of which are based on the pathological examination of the resected specimens.6 7

In the setting of upfront chemoradiation, the identification of prognostic parameters before treatment is increasingly important for the appropriate selection of patients. The most important of these is the cTNM stage defined after digital rectal examination and the results of endorectal ultrasonography (EUS) and nuclear magnetic resonance. However, because the local recurrence rate has been dramatically decreased and distant metastasis has emerged as the major event, surrogate markers are required to more precisely identify subgroups of patients with high risk of distant metastasis. Recent studies have shown that the status of signal transducer and activator of transcription 3 (STAT3) protein may be a good candidate for this purpose.8–13

STATs are cytoplasmic transcription factors involved in epidermal growth factor and IL-6 signalling pathways.13 14 On tyrosine phosphorylation, STAT proteins usually form homodimers, translocate into the nucleus and bind to DNA response elements in the promoter of target genes.15–17

Among the STAT family members, STAT3 is primarily involved in growth control and cancer development. Indeed, STAT3 has been found to be constitutively activated in a wide variety of epithelial and haematopoietic malignancies.18–20 In colorectal cancer STAT3 promotes cell proliferation by deregulating expression of c-Myc, cyclin D1, p21Waf1/Cip1 and p27Kip1 involved in cell cycle progression, as well as expression of anti-apoptotic proteins Mcl-1, Bcl-2, Bcl-xL and survivin.21–24

However, the role of STAT3 in the pathogenesis of colorectal cancers has not been extensively evaluated. Kusaba et al reported in 2006 that the expression of phosphorylated STAT3 (p-STAT3) might be a factor of poor prognosis of colorectal cancer based on overall survival.25 In 2007 and 2008, two studies suggested that p-STAT3 was a factor of non-response to chemotherapy.26 27 Moreover it has been shown that nuclear expression of p-STAT3 is related to tumourous infiltration and lymphatic invasion.28 29 Targeting STAT3 signalling could represent an innovative approach for therapeutic intervention.30 31

However, the specific influence of STAT3 on rectal cancers was never investigated. The aim of this study was to explore the relation between the immunohistochemical expression of nuclear p-STAT3 in tumour cells and clinicopathological features, and to investigate the prognostic significance of p-STAT3 in patients with advanced rectal cancer.

Materials and methods

Patients and tissues

A total of 1011 patients with advanced rectal cancers were enrolled in the European Organisation for Research and Treatment of Cancer (EORTC) 22921 trial.4 Of these, 176 were recruited from the University Hospital Besançon, France. Our study therefore benefited from the clinicopathological and follow-up data collected for the EORTC trial. As previously described elsewhere,4 32 patients were randomly assigned by centre to four treatment arms: (1) preoperative radiotherapy (RT); (2) preoperative chemoradiotherapy (RT/CT); (3) preoperative radiotherapy plus postoperative chemotherapy (RT + adjuvant CT); and (4) preoperative chemoradiotherapy plus postoperative chemotherapy (RT/CT + adjuvant CT). Surgery was performed 3–10 weeks after the end of the preoperative treatment. Entry criteria were: T3 or resectable T4M0 adenocarcinoma of the rectum, according to the 1987 International Union Against Cancer (UICC) staging system33; WHO performance status of 0 or 1; age 80 years or less; and tumour located within 15 cm of the anal verge.32

Patients were followed at 6-month intervals until death or for at least 5 years. The 6-monthly evaluation included clinical examination, abdominal ultrasonography and chest radiography; colonoscopy was performed annually. Evaluation of late side effects included investigation of small bowel and rectal complications and the need for additional surgery. Recurrences were confirmed radiologically or by biopsy. Local recurrence was defined as tumour regrowth within the pelvis or perineum.32

Biopsies taken before preoperative treatment were assigned a histological type according to the WHO classification as follows: well differentiated adenocarcinoma, moderately differentiated adenocarcinoma, poorly differentiated adenocarcinoma, mucinous adenocarcinoma and signet ring type.34 Histological type was confirmed on routine haematoxylin–eosin–safran (HES) stained sections of the resection specimens on which lymphatic, perineural and venous invasions were scored (average number of nodes examined=11.1).


Archival formalin-fixed and paraffin-embedded pretherapeutic biopsies from patients were cut in 4 μm sections using standard techniques. The first section was assessed histopathologically by HES staining to confirm the presence of tumour cells.

Immunostaining for p-STAT3 (tyr705) was performed on the automated Ventana Benchmark (Ventana, Tucson, Arizona, USA). Sections were dewaxed in xylene and rehydrated using a graded series of alcohol. Antigen retrieval was carried out by incubating slides for 60 min in CC1 buffer. Sections were then incubated for 32 min with the primary rabbit polyclonal antibody p-STAT3 (sc-7993; dilution 1:200; Santa Cruz Biotechnology, Santa Cruz, California, USA) followed by biotin-labelled secondary anti-rabbit antibody for 8 min and finally a solution of streptavidin–peroxidase for 8 min. Tissues were then stained for 8 min with diaminobenzidine substrate, giving a brown nuclear precipitate. Tissues were finally counterstained with haematoxylin, dehydrated and mounted. Some serial sections were also incubated with antibody diluent only as negative controls.

The main objective of the present study was to assess p-STAT3 status before any treatment. Therefore this work was done on tumour biopsies taken during endoscopies. These small fragments were taken on the superficial part of the tumours.

The evaluation of p-STAT3 expression was performed at high magnification power (×400). Five carcinomatous fields were selected to count the number of tumour cells that exhibited nuclear staining. When there were multiple fragments, at least one field was selected on each. When the size of the fragments was too small the account was made on the entire biopsy. The intensity of nuclear staining varied among specimens and could even be heterogeneous within the same sample. However the intensity of STAT3 staining had not been taken into account.

As previously described in colorectal carcinoma, biopsies were considered positive for p-STAT3 if nuclear staining was detected in more than 15% of tumour cells; they were judged negative if ≤15% of tumour cells showed nuclear staining.25 Cytoplasmic staining was observed, but not taken into account.

Analysis of the immunohistochemical staining was interpreted by two independent pathologists who were blinded to patient identity and clinical status. In case of discrepancies, the pathologists reviewed the slides together and reached a consensus.

Statistical analyses

Quantitative variables were recoded into categorical variables. All categorical data were compared using Fisher's exact test and the χ2 test. The primary endpoint was progression-free survival (PFS) that was defined as the time between the date of treatment randomisation and the first recurrence (local recurrence or metastasis). The endpoint date was 31 December 2008. Assuming a total of 40 events, our sample of 104 patients ensured 80% statistical power to detect a HR higher than 1.6 or lower than 0.6 using a log rank test.35 PFS and overall survival (OS) were calculated with the Kaplan–Meier method.36 Bivariate analysis was performed with the log-rank test. Multivariate analysis was performed to identify independent prognostic factors for survival using a Cox proportional hazards model with forward stepwise selection. The significance levels for entry and for stay were 0.2 and 0.1, respectively. Statistical significance for all analyses in the present report is considered at the two-sided 5% level. The data analysis was generated using SAS Version 9.1.3 for Windows.


Clinicopathological characteristics

Archival pretherapeutic biopsies were available from 169 of the 176 (96%) patients included in the EORTC trial in the University Hospital of Besançon. Of these biopsies, 104 were formalin-fixed and used for the present study while 65 were excluded because of Bouin fixation or insufficient material. To control for potential bias, we assessed whether the clinicopathological characteristics of the study patients were significantly different from those of the excluded patients. There was no significant difference between groups in terms of patient characteristics and treatment (data not shown).

Therefore our study included specimens from 104 patients. Table 1 presents the sociodemographic and clinicopathological characteristics.

Table 1

Relationship between expression of nuclear phosphorylated signal transducer and activator of transcription 3 (p-STAT3) and clinicopathological parameters among 104 patients with advanced rectal cancer

Relation between p-STAT3 expression and clinicopathological characteristics

We identified immunostaining in an expected nuclear pattern in 39 of 104 (37.5%) biopsies of advanced rectal cancer samples. The κ value for concordance between both pathologists was 0.88, corresponding to a good concordance. Figure 1 shows positive and negative p-STAT3 staining. Nuclear p-STAT3 staining was also observed in peritumour rectal tissue and at the base of crypts of the normal mucosa.

Figure 1

Representative immunohistochemical stainings of tyr705-phosphorylated signal transducer and activator of transcription 3 protein. (A) Positive nuclear staining in tumourous cells. (B) Negative nuclear staining in tumourous cells.

There was no significant relationship between p-STAT3 expression and the clinicopathological characteristics of the patients, including age, sex, anatomical location, treatment modalities, tumour differentiation, lymph node metastasis, and venous, perineural and lymphatic invasions (table 1).

Relation between p-STAT3 expression and clinical outcomes

Patients with a positive expression of p-STAT3 had a better OS than those with a negative expression (p=0.04; 5-year survival rate: 71.8% and 65.5%, respectively). Regarding PFS, there was no difference between patients with positive versus negative expression of p-STAT3 (p=0.52; 5-year survival rate: 63.9% and 62.6%, respectively).

Among the other common prognostic factors of PFS and OS in rectal cancer, lymph node status, perineural invasion and lymphatic invasion were associated with PFS and OS, but venous invasion was not (table 2). Figure 2 shows the Kaplan–Meier curves for p-STAT3 status.

Table 2

Bivariate analysis between clinicopathological characteristics and clinical outcomes

Figure 2

Kaplan–Meier curves for (A) progression-free survival, and (B) overall survival, according to p-STAT3 expression.

Multivariate analysis

Table 3 presents the results of multivariate Cox regression. Perineural invasion and lymphatic invasion were not included in the multivariate analysis because of a high degree of collinearity with lymph node status. Regarding PFS, the only significant unfavourable prognostic factor was the presence of lymph node metastasis (HR 2.9, 95% CI 1.5 to 5.6). Lymph node metastasis and p-STAT3 status were independent significant prognostic factors for OS. Indeed, patients with lymph node metastasis had a poor overall survival (HR 2.3, 95% CI 1.3 to 4.6), while p-STAT3 expression was correlated to a better survival (HR 0.4, 95% CI 0.2 to 0.9).

Table 3

Multivariate analysis for progression-free and overall survival


In this study we explored the relation between p-STAT3 expression and clinicopathological features, as well as its prognostic value in a cohort of patients with advanced rectal cancer included in a randomised trial (EORTC 22921).

In the EORTC 22921 clinical trial, the cumulative 5-year overall survival rate was 62.5%. In the present cohort of patients, lymph node and perineural invasions are prognostic factors for a short-term survival (table 2). The precise role of STAT3 in rectal cancer prognosis was never studied.

Here, we showed that STAT3 phosphorylation in biopsies of rectal cancers was not associated with the clinicopathological staging of these carcinomas. Moreover, STAT3 phosphorylation in T3–T4 rectal cancers is not a negative prognostic factor for progression free and overall survival. Interestingly, our results show a correlation between constitutively activated STAT3 and better overall survival.

Many studies pointed out the presence of a constitutive activated form of STAT3 in different types of cancers such as cervical,37 prostate,38 laryngeal,39 gastric40 and breast cancers.41 These studies suggested that activated STAT3 may play an important role in tumourigenesis, and showed a correlation with pathological T stage and lymph node involvement, as well as identifying activated STAT3 as a risk factor for outcomes. Specifically, in colorectal cancer, Kusaba et al showed that activated STAT3 was correlated with the depth of tumour invasion28 and worse prognosis.25

Overall, the results of these pioneering studies suggest that tumours with constitutive activation of STAT3 may be more aggressive. This is clearly in contrast with the results provided by Dolled-Filhart et al,42 Torres-Roca et al43 and Stewart et al,44 respectively in breast and prostate cancer, and diffuse large B-cell lymphoma (DLBCL). Of note, these latter studies included precise inclusion criteria for patient selection, leading to a higher potential to clarify the prognostic value of STAT3. In line with these results, our study failed to identify a negative impact of nuclear p-STAT3 in rectal cancer outcome since p-STAT3 expression did not influence the PFS.

The favourable prognostic value of STAT3 was already established in node negative breast cancer patients. Nuclear staining of p-STAT3 was related to better overall survival in a cohort of 346 node-negative women (HR 2.35, 95% CI 1.01 to 5.46 for negative nuclear staining).42 Also, regarding PFS, a correlation was observed between a high level of activated STAT3 and a lower rate of distant metastasis in a cohort of 62 patients with locally advanced prostate cancer (T2–T3) uniformly treated by RT or RT plus hormonal therapy in a randomised, prospective, phase III trial of the Radiation Therapy Oncology Group (RTOG 86-10) (HR 0.79, 95% CI 0.63 to 0.99 for unit increase of 10% in STAT3).43 More recently, Stewart et al have shown on 33 poor prognosis patients with DLBCL that the detection of p-STAT3 expression was associated with improved 5-year PFS (93% vs 47%, p=0.006).44 Furthermore, preclinical models have suggested that STAT3 signalling might be a negative regulator of intestinal tumour progression in APCMin mice.45

In line with these studies, we did not observe any significant relation between nuclear p-STAT3 expression and tumour differentiation, lymph node metastasis, and venous, perineural or lymphatic invasions in locally advanced rectal cancers.

Regarding the significant difference observed in OS, we hypothesised that increased activity of STAT3 is associated with a better response to the treatment. However, we failed to observe any correlation of p-STAT3 status and progression free survival, suggesting that STAT3 does not influence the clinical benefit of preoperative treatments of rectal cancers. In addition, the literature gives several examples of studies that shown an association between p-STAT3 expression and CT resistance in breast cancer, melanoma, and head and neck squamous cell carcinoma.46–48 However, the study of Stewart et al45 in DLBCL is contradictory.

On the other hand, as previously described,25 the interpretation of the immunohistochemical staining was based on the proportion of stained tumour cells in the field, with 0–15% positive tumour cells considered negative for STAT3; it was considered positive when there were >15% positive cells. However, this threshold of 15% is not standardised and others used different thresholds, such as 10% or 20%.40 49 A composite score taking into account the intensity of staining and the proportion of stained tumour cells has also been used by Gong et al50 in gastric cancer. On tissue microarray analysis (TMA), only the intensity of staining was used by Dolled-Filhart et al42 in breast cancer. Therefore, there is a degree of uncertainty concerning patients whose tumour expressed 15% of positive cells for p-STAT3, as this can depend on the choice of the threshold. Our results were assessed in a double-blind manner in order to take this difficulty into account.

Finally, the influence of p-STAT3 expression on OS in this study deserves further investigation. Particularly, the recent reports establishing STAT3 as a pivotal component of chronic inflammatory diseases prompted us to investigate the role of this signalling pathway in the tumour micro-environment.51

Take-home messages

  • Phosphorylation (activation) of signal transducer and activator of transcription 3 (p-STAT3) was observed in 39 of 104 pretherapeutic biopsies of T3–T4 resectable rectal cancers (37.5%).

  • The expression of p-STAT3 was not associated with the clinicopathological staging of these rectal cancers and was not a negative prognostic factor for progression free and overall survival.

  • Multivariate Cox regression analysis showed that p-STAT3 expression was a significant prognosis factor of a better overall survival (p=0.01, HR 0.3, 95% CI 0.1 to 0.8).

  • The role of p-STAT3 in the tumour micro-environment should be investigated.


Ligue contre le cancer (Doubs, Alsace, Lorraine, Haute-Marne).

We are also grateful to technicians from the department of Pathology for their excellent technical assistance with rectal tumour section preparation and immunohistochemical analysis. We also thank Dr Fiona Ecarnot and Ms Charline De Vettor for their contribution to this article.



  • Competing interests None.

  • Patient consent Obtained.

  • Ethics approval The trial was approved by the medical ethics committees of all participating centres of the EORTC 22921 trial. Written informed consent was obtained from all patients before randomisation. For the Besançon University Hospital, the trial and this ancillary study were approved by the Protection Person Center (CPP) – Est II, Centre Hospitalier Universitaire, Hôpital Saint-Jacques, 2 place Saint-Jacques, 25030 Besançon Cedex, France.

  • Provenance and peer review Not commissioned; externally peer reviewed.

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.