Purpose: A high frequency of head and neck squamous cell cancers (HNSCC) contain constitutively activated signal transducer and activator of transcription 3 (STAT3). To further elucidate the prognostic role of STAT3 in HNSCC, the expression pattern of STAT3 was correlated with outcome in two independent data sets.

Experimental Design: STAT3 protein expression analysis was done on a test cohort of 102 patients with HNSCC recruited between 1992 and 2005. Automated quantitative analysis was used to assess STAT3 protein expression. We evaluated associations with clinicopathologic parameters and survival prognosis. Associations were validated in a second, independent cohort of 58 patients with confirmed HNSCC enrolled in the Early Detection Research Network–sponsored study who underwent surgical resection with curative intent at the University of Pittsburgh Medical Center between 2000 and 2004.

Results: STAT3 displayed mixed nuclear and cytoplasmic staining. Survival analysis showed that high nuclear STAT3 expression (top tertile versus the rest) was associated with longer progression-free survival (n = 70, mean survival of 88.9 versus 46.7 months, P = 0.012 for the first cohort; n = 37, mean survival of 60.3 versus 33.0 months, P = 0.009 for the second cohort). After best model selection in the multivariable analysis context, only STAT3 was significant, revealing a lower risk of progression and death for patients with high nuclear STAT3-expressing tumors (hazard ratio, 0.28; 95% confidence interval, 0.10-0.82; P = 0.019; and hazard ratio, 0.23; 95% confidence interval, 0.07-0.76; P = 0.016, respectively).

Conclusions: Our results indicate that high nuclear STAT3 expression levels by automated quantitative analysis are associated with favorable outcome in HNSCC. Clin Cancer Res; 16(8); 2427–34. ©2010 AACR.

Translational Relevance

Signal transducers and activators of transcription 3 (STAT3) is activated by a variety of signals that all converge into nuclear translocation of STAT3 to the nucleus, where it is involved in the regulation of genes that promote the malignant phenotype. Therefore, nuclear STAT3 may represent a sensitive marker for detecting the activation of oncogenic signaling pathways (i.e., epidermal growth factor receptor) and aid in the selection of treatment modalities that may result in more efficient eradication of disease. Although the role of constitutive activation of STAT3 in head and neck carcinogenesis has been well documented, its prognostic value for patients with head and neck squamous cell cancer (HNSCC) has not been rigorously studied. Here, using AQUA, we correlated nuclear STAT3 protein levels with patient outcome in two independent cohorts of patients with HNSCC. We found that high nuclear STAT3 expression levels by AQUA are associated with favorable outcome in HNSCC.

Head and neck squamous cell cancer (HNSCC) is the sixth leading cause of cancer-related deaths worldwide. Despite advances in diagnosis and treatment, morbidity and mortality rates of patients with HNSCC remain high (1). Advances in molecular biology have led to the identification of molecules that promote the malignant properties of tumor cells and induce resistance to chemotherapy and radiotherapy. Targeting these molecules has become a major goal in cancer therapy and may eventually increase cure rates.

The signal transducer and activator of transcription (STAT) proteins are transcription factors that have been implicated in signaling by oncoproteins, cytokine receptor–associated kinases, growth factor receptor tyrosine kinases, and nonreceptor tyrosine kinases, originally defined as the signaling mechanism for IFN receptors. After activation, STAT proteins dimerize and translocate to the nucleus, where they bind to DNA-response elements and alter the transcription of genes controlling normal cell functions such as growth, apoptosis, and differentiation (2). Among the seven STAT members, STAT3 is particularly interesting due to its constitutive phosphorylation in several human cancers (3), its essential role in cell transformation by the Src oncoprotein (4), and its ability to induce malignant transformation (57).

Constitutive activation of STAT3 has been described in HNSCC and is involved in deregulation of cell cycle, increased proliferation, and inhibition of apoptosis (8, 9). STAT3 is activated frequently in matched tumor/normal mucosa samples from HNSCC patients compared with normal mucosa from patients without cancer (8). Constitutive activation of STATs was detected in extracts from HNSCC cells and was dependent on transforming growth factor–induced activation of the epidermal growth factor receptor (EGFR) tyrosine kinase (9).

Despite the diversity in STAT3 activation signals, all result in the translocation of STAT3 to the nucleus, where it is involved in the regulation of genes that promote the malignant phenotype. These findings suggest that activated STAT3 may provide a sensitive marker for detecting the activation of oncogenic signaling pathways (i.e., EGFR) and aid in the selection of treatment modalities that may result in more efficient eradication of disease.

Although the role of constitutive activation of STAT3 in head and neck carcinogenesis has been well documented, its prognostic value for patients with HNSCC has not been rigorously studied. Studies (10, 11) evaluating the prognostic significance of STAT3 are limited by the technical difficulties inherent in assessing pSTAT3 with conventional immunohistochemistry such as variability in immunohistochemical techniques, different methods of pathologist-based scoring, and the semiquantitative nature of the assay. To overcome this problem, a method of in situ automated quantitative analysis (AQUA) has been developed (12), which allows measurements of protein expression within subcellular compartments that results in several directly proportional to the number of molecules expressed per unit area. Thus, we avoid biases introduced from the arbitrary cutoff points used in conventional immunohistochemistry studies while at the same time preserving spatial and morphologic information that techniques such as Western blotting lose.

In the present study, we hypothesized that assessment of nuclear STAT3 levels by AQUA would be a more reliable indicator of STAT3 activation status. We therefore determined the effect of nuclear STAT3 protein levels on patient outcome in two independent cohorts of patients with HNSCC.

Patient population

For cohort 1, inclusion criteria were histologically confirmed primary squamous cell carcinomas of the head and neck treated at Yale-New Haven Hospital and Aristotle University Hospital between 1992 and 2005, and therapy with either external beam radiotherapy or gross total surgical resection and postoperative radiotherapy (n = 102). Exclusion criteria included presentation with metastatic or recurrent disease or failure to receive a full course of radiation therapy. Patients with incomplete clinical-pathologic data or those lost to follow-up were also excluded. For cohort 2, patients with confirmed HNSCC (oral cavity, oropharynx, hypopharyx, or larynx) who were enrolled in an Early Detection Research Network–sponsored study and who underwent surgical resection with curative intent at the University of Pittsburgh Medical Center between 2000 and 2004 were considered for inclusion in this study (n = 161). Tumor specimens from 58 of these HNSCC patients, 50 with paired adjacent histologically normal adjacent mucosa, were used to construct a tissue microarray (TMA), using triplicate tissue cores. HNSCC tumors included in cohort 2 TMA did not differ significantly by tumor site, clinical stage, or patient gender or age from those HNSCC tumors not incorporated into the TMA. Tissues were collected under the auspices of a tissue bank protocol approved by the University of Pittsburgh Institutional Review Board.

TMA construction

Two TMAs were constructed: (a) a TMA consisting of tumors from patient cohort I was constructed at the Yale University Tissue Microarray Facility including 102 cases and (b) a TMA consisting of HNSCC tumors from patient cohort II including 58 cases. Following Institutional Review Board approval, the TMAs were constructed as previously described (13). Tissue cores 0.6 mm in size were obtained from paraffin-embedded, formalin-fixed tissue blocks from the archives of the Yale University and Aristotle University of Thessaloniki Department of Pathology or from the University of Pittsburgh Head and Neck Department of Pathology. H&E-stained slides from all blocks were first reviewed by a pathologist to select representative areas of invasive tumor to be cored. The cores were placed on the recipient microarray block using a Tissue Microarrayer (Beecher Instrument). All tumors were represented with at least 2-fold redundancy. Previous studies (14) have shown that the use of TMAs containing one to two histospots provides a sufficiently representative sample for analysis by immunohistochemistry. Addition of a duplicate histospot, although not necessary, does provide marginally improved reliability (14). Cores from human papillomavirus 16–positive SiHa cell lines fixed in formalin and embedded in paraffin were selected for positive controls and included in the cohort 1 array.

Quantitative immunofluorescent staining

TMAs were deparaffinized with xylene and stained as previously described (15). Briefly, slides were rehydrated and antigen retrieval was achieved by pressure cooking for 20 min in citrate buffer (pH = 6). Slides were preincubated with 0.3% bovine serum albumin in 0.1 mol/L of TBS (pH 8) for 30 min at room temperature. Slides were then incubated with primary antibody to STAT3 (1:1,000, clone 124H6; Cell Signaling Technology) and a wide spectrum rabbit anti-cow cytokeratin antibody (1:100, Z0622; DAKO) at 4°C overnight. The STAT3 antibody has been extensively validated in previous studies using immunohistochemistry and Western blot analysis of neoplastic tissue and tumor cell lines. Subsequently, slides were incubated with Cy3-conjugated goat anti-rabbit secondary antibody (A11010; Molecular Probes) diluted 1:100 in mouse EnVision reagent (K4001; DAKO Corp.) for 1 h at room temperature. Cy5 directly conjugated to tyramide (FP117; Perkin-Elmer) at a dilution was used as the fluorescent chromagen for STAT3 detection. Cy5 (red) was used because it is well outside the green-orange spectrum of tissue autofluorescence. Prolong mounting medium (ProLong Gold, P36931; Molecular Probes) containing 4′,6-diamidino-2-phenylindole was used to identify nuclei.

Automated image acquisition and analysis

Automated image acquisition and analysis using automated in situ quantitative measurement of protein analysis (AQUA) has been previously described (12). In brief, monochromatic, high-resolution (1,024 × 1,024 pixel; 0.5 μm) images were obtained of each histospot. We distinguished areas of tumor from stromal elements by creating a mask from the cytokeratin signal. 4′,6-Diamidino-2-phenylindole signal was used to identify nuclei and the cytokeratin signal was used to define cytoplasm. Overlapping pixels [to a 99% confidence interval (CI)] were excluded from both compartments. The signal (AQUA score) was scored on a normalized scale of 0 to 255 expressed as pixel intensity divided by the target area. AQUA scores for each subcellular compartment (nuclear and cytoplasmic STAT3) as well as the tumor mask were recorded. AQUA scores for duplicate tissue cores were averaged to obtain a mean score for each tumor.

Statistical analysis

Histospots containing <5% tumor as assessed by mask area (automated) were excluded from further analysis. The AQUA STAT3 score was assessed by tertiles. Progression-free survival (PFS) and overall survival (OS) were assessed by Kaplan-Meier analysis with log-rank score for determining statistical significance. Comparisons of STAT3 expression with gender, tumor-node-metastasis (TNM) stage, histologic grade (only available in cohort 1), and tumor site were made by χ2 test or Fisher's exact test. Relative risk was assessed by Cox proportional hazards regression by multivariable analysis for these variables. In cohort 1, backward selection of the best model was used with removal criterion P > 0.10. The prognostic ability of this model was evaluated in cohort 2. All calculations and statistical analyses were done by SPSS 15.0 for Windows (SPSS, Inc.).

Clinical and pathologic variable analysis

In cohort 1, there were 70 patients with primary squamous cell carcinoma of the head and neck who met inclusion criteria and for whom we had complete STAT3 expression data. We excluded from the analysis 32 cases (among 102) missing STAT3 expression information. These cases did not differ from the ones included in the analysis with respect to patient gender, tumor site, TNM stage, histologic grade, PFS and OS as assessed by Fisher's exact test, and log-rank test, respectively. For cohort 2, 37 of the 58 arrayed HNSCC tumors were successfully evaluated for nuclear STAT3 levels using AQUA. Unsuccessful assessment was due to tumor core loss; cases with lost cores did not differ from successfully evaluated tumors with respect to all parameters. Of those tumors successfully evaluated for nuclear STAT3 protein expression, tumors with high versus intermediate/low nuclear STAT3 levels did not differ by patient gender, tumor TNM stage, or tumor site for cohort 1 or cohort 2 (Table 1).

Table 1.

Demographic, clinical, and pathologic data by nuclear STAT3 status (cohorts 1 and 2)

Nuclear STAT3 expression status
Cohort 1Cohort 2
nLow (n = 47)%High (n = 23)%PnLow (n = 25)%High (n = 12)%P
Gender 
    Male 58 80.9 87.0 0.738 29 76.0 83.3 1.00 
    Female 12 19.1 13.0  24.0 16.7  
TNM stage 
    I 6.8 8.7 0.696 5.0 8.3 0.492 
    II 15.9 8.7  20.0 16.7  
    III 18 29.5 21.7  20.0 0.0  
    IV 35 47.7 60.9  20 55.0 75.0  
    Unknown       
Tumor site 
    Oral cavity 18.2 0.100 14 39.1 41.7 0.950 
    Larynx 31 40.9 59.1  12 34.8 33.3  
    Oropharynx 25 38.6 36.4  17.4 25.0  
    Hypopharynx 2.3 4.5  8.7 0.0  
    Unknown       
Tumor grade 
    Well differentiated 10 12.8 23.8 0.407     
    Moderately differentiated 32 59.0 42.9      
    Poorly differentiated 18 28.2 33.3      
    Unknown 10        
Nuclear STAT3 expression status
Cohort 1Cohort 2
nLow (n = 47)%High (n = 23)%PnLow (n = 25)%High (n = 12)%P
Gender 
    Male 58 80.9 87.0 0.738 29 76.0 83.3 1.00 
    Female 12 19.1 13.0  24.0 16.7  
TNM stage 
    I 6.8 8.7 0.696 5.0 8.3 0.492 
    II 15.9 8.7  20.0 16.7  
    III 18 29.5 21.7  20.0 0.0  
    IV 35 47.7 60.9  20 55.0 75.0  
    Unknown       
Tumor site 
    Oral cavity 18.2 0.100 14 39.1 41.7 0.950 
    Larynx 31 40.9 59.1  12 34.8 33.3  
    Oropharynx 25 38.6 36.4  17.4 25.0  
    Hypopharynx 2.3 4.5  8.7 0.0  
    Unknown       
Tumor grade 
    Well differentiated 10 12.8 23.8 0.407     
    Moderately differentiated 32 59.0 42.9      
    Poorly differentiated 18 28.2 33.3      
    Unknown 10        

Quantitative immunohistochemistry for STAT3 protein expression

As visualized by fluorescent immunohistochemistry, STAT3 displayed mixed nuclear and cytoplasmic staining (Fig. 1). Normalized AQUA scores for STAT3 expression in tumors ranged from 1 to 100. Nuclear, cytoplasmic, and tumor STAT3 expression followed a skewed distribution as expected for a cancer tissue biomarker (Supplementary Fig. S1). To assess for intratumor heterogeneity of STAT3 expression and control for reproducibility of the assay, we compared AQUA scores from redundant tumor cores and observed significant correlation (Supplementary Fig. S2; R = 0.89). AQUA scores in the nuclear compartment were averaged between the two histospots and final scores ranging from 3.92 to 97.92 were obtained for 70 patients.

Fig. 1.

Fluorescent immunohistochemistry for automated analysis (AQUA). A, cytokeratin was used to identify tumor. B, binary gating of cytokeratin expression created the tumor mask (white). C, pseudocolored colocalization image showing compartment assignment. Cytokeratin-Cy3 (green) was used to define the nonnuclear compartment; 4',6-diamidino-2-phenylindole (blue) was used to define the nuclear compartment. D, Cy5 (red) was used to identify STAT3. Original magnification, ×20.

Fig. 1.

Fluorescent immunohistochemistry for automated analysis (AQUA). A, cytokeratin was used to identify tumor. B, binary gating of cytokeratin expression created the tumor mask (white). C, pseudocolored colocalization image showing compartment assignment. Cytokeratin-Cy3 (green) was used to define the nonnuclear compartment; 4',6-diamidino-2-phenylindole (blue) was used to define the nuclear compartment. D, Cy5 (red) was used to identify STAT3. Original magnification, ×20.

Close modal

Univariate survival analysis

Progression-free survival

The status of nuclear STAT3 expression was evaluated for association with PFS using the log-rank statistic to determine significance. High nuclear STAT3 expression is associated with superior PFS. Patients in cohort 1 with high nuclear STAT3 had a higher cumulative survival at 5 years compared with patients with intermediate or low STAT3. The cumulative PFS at 5 years was 75.2% for high nuclear STAT3 expressors versus 35.4% for others. The median PFS has not been reached and the mean PFS is 88.9 months for high expressors, compared with a median of 21.3 and mean of 46.7 months for those with intermediate or low nuclear STAT3 expression by AQUA (Fig. 2A). As shown in Table 2, this difference in PFS was significant (P = 0.012; log-rank test). For cohort 2, elevated nuclear expression of STAT3 was also associated with improved PFS (high STAT3: median, 81.0; mean, 60.3 months; low STAT3: median, 17.2; mean, 33.0 months; P = 0.035). A greater proportion of patients in cohort 2 with high nuclear STAT3 levels remained PFS compared with patients with intermediate or low nuclear STAT3 levels (Table 2).

Fig. 2.

A, Kaplan-Meier survival curve comparing PFS estimation between low and high expressing nuclear STAT3 groups. Patients with high nuclear STAT3 expression exhibit a higher probability of PFS (cumulative PFS at 5 y, 75.2% versus 35.4%). B, Kaplan-Meier survival curve comparing OS estimation between low- and high-expressing nuclear STAT3 groups. Patients with high nuclear STAT3 expression exhibit a higher probability of OS (cumulative OS at 5 y, 72.4% versus 38.3%).

Fig. 2.

A, Kaplan-Meier survival curve comparing PFS estimation between low and high expressing nuclear STAT3 groups. Patients with high nuclear STAT3 expression exhibit a higher probability of PFS (cumulative PFS at 5 y, 75.2% versus 35.4%). B, Kaplan-Meier survival curve comparing OS estimation between low- and high-expressing nuclear STAT3 groups. Patients with high nuclear STAT3 expression exhibit a higher probability of OS (cumulative OS at 5 y, 72.4% versus 38.3%).

Close modal
Table 2.

Univariate survival analysis (Kaplan-Meier log-rank; cohorts 1 and 2)

Mean survival (95% CI; mo)Median survival (95% CI; mo)Cumulative survival or recurrence (no. of events) at 5 yP
Cohort 1 
PFS    0.012 
    High nuclear STAT3 (n = 23) 88.91 (67.33-110.49) n.e (n.e-n.e) 75.2% (4)  
    Low nuclear STAT3 (n = 47) 46.65 (32.09-61.21) 21.30 (12.95-29.65) 35.4% (24)  
OS 
    High nuclear STAT3 (n = 23) 119.92 (86.13-153.71) n.e (n.e-n.e) 72.4% (3) 0.009 
    Low nuclear STAT3 (n = 47) 57.30 (38.48-76.11) 32.3 (18.24-46.36) 38.3% (23)  
Cohort 2 
PFS    0.035 
    High nuclear STAT3 (n = 12) 60.3 (41.2-79. 5) 81.0 (n.e-n.e) 75.0% (3)  
    Low nuclear STAT3 (n = 25) 33.0 (20.2-45.8) 17.2 (4.4-30.0) 39.3% (15)  
OS 
    High nuclear STAT3 (n = 12) 63.1 (47.1-79.1) 81.0 (n.e-n.e) 68.8% (3) 0.244 
    Low nuclear STAT3 (n = 25) 45.3 (31.2-59.4) 62.7 (0-135.9) 51.4% (12)  
Mean survival (95% CI; mo)Median survival (95% CI; mo)Cumulative survival or recurrence (no. of events) at 5 yP
Cohort 1 
PFS    0.012 
    High nuclear STAT3 (n = 23) 88.91 (67.33-110.49) n.e (n.e-n.e) 75.2% (4)  
    Low nuclear STAT3 (n = 47) 46.65 (32.09-61.21) 21.30 (12.95-29.65) 35.4% (24)  
OS 
    High nuclear STAT3 (n = 23) 119.92 (86.13-153.71) n.e (n.e-n.e) 72.4% (3) 0.009 
    Low nuclear STAT3 (n = 47) 57.30 (38.48-76.11) 32.3 (18.24-46.36) 38.3% (23)  
Cohort 2 
PFS    0.035 
    High nuclear STAT3 (n = 12) 60.3 (41.2-79. 5) 81.0 (n.e-n.e) 75.0% (3)  
    Low nuclear STAT3 (n = 25) 33.0 (20.2-45.8) 17.2 (4.4-30.0) 39.3% (15)  
OS 
    High nuclear STAT3 (n = 12) 63.1 (47.1-79.1) 81.0 (n.e-n.e) 68.8% (3) 0.244 
    Low nuclear STAT3 (n = 25) 45.3 (31.2-59.4) 62.7 (0-135.9) 51.4% (12)  

Abbreviation: n.e, nonevaluable.

Overall survival

The expression status of STAT3 was also evaluated for association with OS. Kaplan-Meier analysis showed that there was a significant correlation between high nuclear STAT3 expression and improved OS. Patients in cohort 1 with high nuclear STAT3 levels had significantly better OS (median not reached; mean, 119.9 months), whereas patients with intermediate or low nuclear STAT3 levels had a shorter OS (median, 32.3; mean, 57.3 months; P = 0.009; log-rank test). Patients with high nuclear STAT3 expression had an OS at 5 years of 72.4% compared with 38.3% for patients with intermediate or low STAT3 (Fig. 2B). A nonsignificant trend was observed for cohort 2: patients with high nuclear STAT3 levels had a higher cumulative survival at 5 years than patients with intermediate or low STAT3 levels, 68.8% and 51.4%, respectively.

Multivariable survival analysis

Using the Cox proportional hazards model, we carried out multivariable analysis to assess the prognostic value of nuclear STAT3 expression by AQUA for both PFS and OS in cohort 1. We included the following prognostic variables in the regression model: gender, TNM stage, and tumor grade and tumor site. After backward selection, the model for PFS including only STAT3 is chosen. A 77% decrease in risk of a progression-defining event was estimated for patients with STAT3 protein expression in the upper tertile [hazard ratio (HR), 0.23; 95% CI, 0.06-0.91; P = 0.037], which for OS corresponds to an 87% decrease in risk (HR, 0.13; 95% CI, 0.03-0.67; P = 0.014; Table 3). For cohort 2, although the number of events is small, the significant association is sustained for PFS (HR, 0.31; P = 0.044) whereas it is not for OS (HR, 0.54; P = not significant).

Table 3.

Multivariable PFS and OS analysis, by Cox regression

VariableHR95% CIP
PFS 
    Male gender    
    Female gender 3.76 1.09-12.98 0.036 
    TNM stage I    
    TNM stage II 1.32 0.11-15.46 0.827 
    TNM stage III 0.96 0.10-8.90 0.973 
    TNM stage IV 2.59 0.30-22.52 0.388 
    Histology, well differentiated    
    Histology,moderately differentiated 0.39 0.10-1.62 0.196 
    Histology, poorly differentiated 0.24 0.05-1.23 0.086 
    Tumor site, oral cavity    
    Tumor site, larynx 1.52 0.31-7.46 0.604 
    Tumor site, oropharynx 1.95 0.48-8.01 0.352 
    Nuclear STAT3 AQUA score 0.23 0.06-0.91 0.037 
OS 
    Male gender    
    Female gender 2.53 0.78-8.23 0.123 
    TNM stage I    
    TNM stage II 1.05 0.06-17.89 0.973 
    TNM stage III 0.80 0.07-8.93 0.849 
    TNM stage IV 3.99 0.41-39.11 0.234 
    Histology, well differentiated    
    Histology,moderately differentiated 0.17 0.03-0.84 0.030 
    Histology, poorly differentiated 0.14 0.02-0.85 0.032 
    Tumor site, oral cavity 0.79 0.17-3.62 0.762 
    Tumor site, larynx 1.33 0.38-4.65 0.658 
    Tumor site, oropharynx    
    Nuclear STAT3 AQUA score 0.13 0.03-0.67 0.014 
VariableHR95% CIP
PFS 
    Male gender    
    Female gender 3.76 1.09-12.98 0.036 
    TNM stage I    
    TNM stage II 1.32 0.11-15.46 0.827 
    TNM stage III 0.96 0.10-8.90 0.973 
    TNM stage IV 2.59 0.30-22.52 0.388 
    Histology, well differentiated    
    Histology,moderately differentiated 0.39 0.10-1.62 0.196 
    Histology, poorly differentiated 0.24 0.05-1.23 0.086 
    Tumor site, oral cavity    
    Tumor site, larynx 1.52 0.31-7.46 0.604 
    Tumor site, oropharynx 1.95 0.48-8.01 0.352 
    Nuclear STAT3 AQUA score 0.23 0.06-0.91 0.037 
OS 
    Male gender    
    Female gender 2.53 0.78-8.23 0.123 
    TNM stage I    
    TNM stage II 1.05 0.06-17.89 0.973 
    TNM stage III 0.80 0.07-8.93 0.849 
    TNM stage IV 3.99 0.41-39.11 0.234 
    Histology, well differentiated    
    Histology,moderately differentiated 0.17 0.03-0.84 0.030 
    Histology, poorly differentiated 0.14 0.02-0.85 0.032 
    Tumor site, oral cavity 0.79 0.17-3.62 0.762 
    Tumor site, larynx 1.33 0.38-4.65 0.658 
    Tumor site, oropharynx    
    Nuclear STAT3 AQUA score 0.13 0.03-0.67 0.014 

To our knowledge, this is the first study to perform a quantitative assessment of nuclear STAT3 in head and neck cancer in association with patient prognosis. We used AQUA, a method that allows quantitative measurement of protein expression within subcellular compartments. Because activation of STAT3 signaling involves translocation to the nucleus, we hypothesized that assessment of nuclear STAT3 expression could act as a marker of activation of a STAT3-mediated signaling pathway. Our findings show higher nuclear STAT3 level is predictive of favorable clinical outcome. Furthermore, nuclear STAT3 was the only identified prognostic factor when adjusting for other commonly used prognostic markers and pathologic parameters.

We previously reported the prognostic effect of tyrosine-phosphorylated STAT3 immunoexpression on survival in two independent cohorts of patients with HNSCC and evaluated pSTAT3, transforming growth factor-α, EGFR, and gastrin-releasing peptide receptor expression in matched tumor and lymph node metastases in one of these cohorts (11). Immunoexpression of pSTAT3 alone did not correlate with clinical outcome in either cohort. Masuda et al. (10) analyzed the immunohistochemistry the levels of pSTAT3 in 51 primary squamous cell carcinomas of the tongue and found that p-STAT3 was an independent predictor of poor prognosis. P-STAT3 levels were also correlated with cyclin D1 levels, nodal metastasis, and clinical stage. However, in those studies, immunohistochemistry was done on archived surgical specimens that were not processed using a specific protocol for evaluation of phospho proteins. Routine formalin-fixed tissue from surgical resections does not provide the degree of preservation required to analyze functional proteins, particularly such transient events as phosphorylation. In addition, these studies are limited by the technical difficulties inherent in assessing pSTAT3 with conventional immunohistochemistry. Shah et al. (16) analyzed 135 oral squamous cell carcinomas for cytoplasmic or nuclear STAT3 expression by immunohistochemistry. The authors found that nuclear STAT3 was an independent predictor for poor outcome in early-stage patients and concluded that STAT3 activation is an early event in head and neck carcinogenesis.

Our finding that high levels of nuclear STAT3 are associated with better outcome in HNSCC is subject to numerous interpretations. Tumors that activate these pathways appear less aggressive than tumors that progress even in the absence of STAT3 activation, perhaps because the later rely on alternate pathways more tightly related to invasiveness or treatment resistance. Activated STAT3 in early-stage HNSCC may play a role in tumor development by sparing cells from apoptotic death or through cooperation with other oncogenes. In our cohorts, the majority of tumors were of advanced stage. Thus, at advanced stage, tumors probably become independent of STAT3-mediated signaling. An alternative explanation suggests that STAT3 functions as a tumor suppressor protein. STAT3 plays a pleomorphic role in signal transduction and it is possible that it functions as an oncoprotein or a tumor suppressor protein depending on the setting. In mouse mammary gland, STAT3 is activated both during the apoptotic phase and during the highly proliferative state of early pregnancy (17). There are also several examples in which activated STAT3 seems to play a role in differentiation and promoting apoptosis (1820). Conditional knockout studies in murine models have shown that STAT3 function is essential in facilitating mammary gland involution by inducing extensive epithelial cell apoptosis through upregulating IGFBP-5 (18, 21, 22). De La Iglesia et al. (23) reported that STAT3 plays a pro-oncogenic or tumor suppressive role in glioblastoma depending on the mutational profile of the tumor. Deficiency of PTEN triggers a cascade that inhibits STAT3 signaling in murine astrocytes and human glioblastoma tumors. Accordingly, PTEN knockdown induces efficient malignant transformation of astrocytes upon knockdown of the STAT3 gene. In contrast to the tumor suppressive function of STAT3 in the PTEN pathway, STAT3 forms a physical complex with the pro-oncogenic protein EGFRVIII in the nucleus and mediates EGFRVIII-induced glial transformation. STAT3 may play a dual role in head and neck carcinogenesis depending on the genetic background of the tumor similar to glioblastoma.

Several studies in patients' samples in different sets of tumors have shown similar results to the present study. Dolled-Filhart et al. (22) reported that breast tumors with positive STAT3 nuclear expression had a significantly improved survival (P = 0.03) at 5 years of follow-up. Hsiao et al. (24) studied STAT3 and STAT5 expression patterns in relation to survival in nasopharyngeal cancer using immunohistochemistry. Strong intranuclear staining, or intranuclear and cytoplasmic staining, were noted in 31 of the 43 specimens classified as “positive” for significant constitutive STAT3 activation and in 32 of the 38 specimens classified as positive for significant constitutive STAT5 activation. The authors found that following radiotherapy, patients with constitutive STAT5 activation, or activation of both STAT3 and STAT5, had better PFS and OS than those without activated STAT5. Nuclear STAT3 staining was inversely correlated with the development of distant metastases in a cohort of prostate cancer patients treated with either radiotherapy alone or radiotherapy plus short-term androgen blockade (25). To the contrary, STAT3 expression has been associated with worse outcome in several studies (2628).

One limitation of our study is that our assay cannot differentiate between phosphorylated and unphosphorylated nuclear STAT3. It is known that STAT3 can enter the nucleus independently of its phosphorylation state. Unphosphorylated STAT3 enters the nucleus by binding to distinct importins (29). Although phosphorylation is not a prerequisite for STAT3 nuclear import, the requirement for phosphorylation for the transcriptional activity of STAT3 remains a controversial issue (5, 30, 31). Yang et al. (31) reported that a STAT3 mutated on Tyr705 retained its transcriptional activity but activated a different set of genes compared with wild-type STAT3. Nuclear phosphorylation of STAT3 by a nuclear tyrosine kinase has been reported in breast cancer and may represent an alternative STAT3 activation mechanism, contributing to its oncogenic potential (32).

To conclude, the present study shows that nuclear STAT3 protein expression in the upper tertile as assessed by AQUA is associated with better PFS in two independent cohorts of patients with HNSCC.

D. Rimm, founder, stockholder, consultant, HistoRx, licensee of AQUA Technology.

Grant Support: Yale School of Medicine Institutional startup funds (A. Psyrri) and the Virginia Alden Wright Fund (C. Sasaki).

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

1
Al-Sarraf
M
. 
Treatment of locally advanced head and neck cancer: historical and critical review
.
Cancer Control
2002
;
9
:
387
99
.
2
Darnell
JE
 Jr.
,
Kerr
IM
,
Stark
GR
. 
Jak-STAT pathways and transcriptional activation in response to IFNs and other extracellular signaling proteins
.
Science
1994
;
264
:
1415
21
.
3
Alvarez
JV
,
Febbo
PG
,
Ramaswamy
S
,
Loda
M
,
Richardson
A
,
Frank
DA
. 
Identification of a genetic signature of activated signal transducer and activator of transcription 3 in human tumors
.
Cancer Res
2005
;
65
:
5054
62
.
4
Schlessinger
K
,
Levy
DE
. 
Malignant transformation but not normal cell growth depends on signal transducer and activator of transcription 3
.
Cancer Res
2005
;
65
:
5828
34
.
5
Bromberg
JF
,
Wrzeszczynska
MH
,
Devgan
G
, et al
. 
Stat3 as an oncogene
.
Cell
1999
;
98
:
295
303
.
6
Darnell
JE
 Jr
. 
Transcription factors as targets for cancer therapy
.
Nat Rev Cancer
2002
;
2
:
740
49
.
7
Yu
H
,
Jove
R
. 
The STATs of cancer-new molecular targets come of age
.
Nat Rev Cancer
2004
;
4
:
97
105
.
8
Grandis
JR
,
Drenning
SD
,
Zeng
Q
, et al
. 
Constitutive activation of Stat3 signaling abrogates apoptosis in squamous cell carcinogenesis in vivo
.
Proc Natl Acad Sci U S A
2000
;
97
:
4227
32
.
9
Grandis
JR
,
Drenning
SD
,
Chakraborty
A
, et al
. 
Requirement of Stat3 but not Stat1 activation for epidermal growth factor receptor-mediated cell growth in vitro
.
J Clin Invest
1998
;
102
:
1385
92
.
10
Masuda
M
,
Suzui
M
,
Yasumatu
R
, et al
. 
Constitutive activation of signal transducers and activators of transcription 3 correlates with cyclin D1 overexpression and may provide a novel prognostic marker in head and neck squamous cell carcinoma
.
Cancer Res
2002
;
62
:
3351
55
.
11
Seethala
RR
,
Gooding
WE
,
Handler
PN
, et al
. 
Immunohistochemical analysis of phosphotyrosine signal transducer and activator of transcription 3 and epidermal growth factor receptor autocrine signaling pathways in head and neck cancers and metastatic lymph nodes
.
Clin Cancer Res
2008
;
14
:
1303
9
.
12
Camp
RL
,
Chung
GG
,
Rimm
DL
. 
Automated subcellular localization and quantification of protein expression in tissue microarrays
.
Nat Med
2002
;
8
:
1323
27
.
13
Weinberger
PM
,
Yu
Z
,
Haffty
BG
, et al
. 
Prognostic significance of p16 protein levels in oropharyngeal squamous cell cancer
.
Clin Cancer Res
2004
;
10
:
5684
91
.
14
Rimm
DL
,
Camp
RL
,
Charette
LA
,
Costa
J
,
Olsen
DA
,
Reiss
M
. 
Tissue microarray: a new technology for amplification of tissue resources
.
Cancer J
2001
;
7
:
24
31
.
15
Yu
Z
,
Weinberger
PM
,
Haffty
BG
, et al
. 
Cyclin d1 is a valuable prognostic marker in oropharyngeal squamous cell carcinoma
.
Clin Cancer Res
2005
;
11
:
1160
6
.
16
Shah
NG
,
Trivedi
TI
,
Tankshali
RA
, et al
. 
Stat3 expression in oral squamous cell carcinoma: association with clinicopathological parameters and survival
.
Int J Biol Markers
2006
;
21
:
175
83
, (
in eng
).
17
Philp
JA
,
Burdon
TG
,
Watson
CJ
. 
Differential activation of STATs 3 and 5 during mammary gland development
.
FEBS Lett
1996
;
396
:
77
80
.
18
Chapman
RS
,
Lourenco
P
,
Tonner
E
, et al
. 
Suppression of epithelial apoptosis and delayed mammary gland involution in mice with a conditional knockout of Stat3
.
Genes Dev
1999
;
13
:
2604
16
.
19
Minami
M
,
Inoue
M
,
Wei
S
, et al
. 
STAT3 activation is a critical step in gp130-mediated terminal differentiation and growth arrest of a myeloid cell line
.
Proc Natl Acad Sci U S A
1999
;
93
:
3963
6
.
20
O'Farrell
AM
,
Liu
Y
,
Moore
KW
,
Mui
AL
. 
IL-10 inhibits macrophage activation and proliferation by distinct signaling mechanisms: evidence for Stat3-dependent and -independent pathways
.
EMBO J
1998
;
17
:
1006
18
.
21
Chapman
RS
,
Lourenco
P
,
Tonner
E
, et al
. 
The role of Stat3 in apoptosis and mammary gland involution. Conditional deletion of Stat3
.
Adv Exp Med Biol
2000
;
480
:
129
38
.
22
Dolled-Filhart
M
,
Camp
RL
,
Kowalski
DP
,
Smith
BL
,
Rimm
DL
. 
Tissue microarray analysis of signal transducers and activators of transcription 3 (Stat3) and phospho-Stat3 (Tyr705) in node-negative breast cancer shows nuclear localization is associated with a better prognosis
.
Clin Cancer Res
2003
;
9
:
594
600
.
23
de la Iglesia
N
,
Konopka
G
,
Puram
SV
, et al
. 
Identification of a PTEN-regulated STAT3 brain tumor suppressor pathway
.
Genes Dev
2008
;
22
:
449
62
.
24
Hsiao
JR
,
Jin
YT
,
Tsai
ST
,
Shiau
AL
,
Wu
CL
,
Su
WC
. 
Constitutive activation of STAT3 and STAT5 is present in the majority of nasopharyngeal carcinoma and correlates with better prognosis
.
Br J Cancer
2003
;
89
:
344
9
.
25
Torres-Roca
JF
,
DeSilvio
M
,
Mora
LB
, et al
. 
Activated STAT3 as a correlate of distant metastasis in prostate cancer: a secondary analysis of Radiation Therapy Oncology Group 86-10
.
Urology
2007
;
69
:
505
9
.
26
Abou-Ghazal
M
,
Yang
DS
,
Qiao
W
, et al
. 
The incidence, correlation with tumor-infiltrating inflammation, and prognosis of phosphorylated STAT3 expression in human gliomas
.
Clin Cancer Res
2008
;
14
:
8228
35
.
27
Kim
DY
,
Cha
ST
,
Ahn
DH
, et al
. 
STAT3 expression in gastric cancer indicates a poor prognosis
.
J Gastroenterol Hepatol
2009
;
24
:
646
51
.
28
Haller
F
,
Lobke
C
,
Ruschhaupt
M
, et al
. 
Increased KIT signalling with up-regulation of cyclin D correlates to accelerated proliferation and shorter disease-free survival in gastrointestinal stromal tumours (GISTs) with KIT exon 11 deletions
.
J Pathol
2008
;
216
:
225
35
.
29
Liu
L
,
McBride
KM
,
Reich
NC
. 
STAT3 nuclear import is independent of tyrosine phosphorylation and mediated by importin-α3
.
Proc Natl Acad Sci U S A
2005
;
102
:
8150
55
.
30
Liddle
FJ
,
Alvarez
JV
,
Poli
V
,
Frank
DA
. 
Tyrosine phosphorylation is required for functional activation of disulfide-containing constitutively active STAT mutants
.
Biochemistry
2006
;
45
:
5599
605
.
31
Yang
J
,
Chatterjee-Kishore
M
,
Staugaitis
SM
, et al
. 
Novel roles of unphosphorylated STAT3 in oncogenesis and transcriptional regulation
.
Cancer Res
2005
;
65
:
939
47
.
32
Liu
L
,
Gao
Y
,
Qiu
H
,
Miller
WT
,
Poli
V
,
Reich
NC
. 
Identification of STAT3 as a specific substrate of breast tumor kinase
.
Oncogene
2006
;
25
:
4904
12
.

Supplementary data