Purpose: In patients with colorectal cancer, a high-density local inflammatory infiltrate response is associated with improved survival, whereas elevated systemic inflammatory responses are associated with poor survival. One potential unifying mechanism is the IL6/JAK/STAT3 pathway. The present study examines the relationship between tumor total STAT3 and phosphorylated STAT3Tyr705 (pSTAT3) expression, host inflammatory responses, and survival in patients undergoing resection of stage I–III colorectal cancer.

Experimental Design: Immunohistochemical assessment of STAT3/pSTAT3 expression was performed using a tissue microarray and tumor cell expression divided into tertiles using the weighted histoscore. The relationship between STAT3/pSTAT3 expression and local inflammatory (CD3+, CD8+, CD45R0+, FOXP3+ T-cell density, and Klintrup–Mäkinen grade) and systemic inflammatory responses and cancer-specific survival were examined.

Results: A total of 196 patients were included in the analysis. Cytoplasmic and nuclear STAT3 expression strongly correlated (r = 0.363; P < 0.001); nuclear STAT3 and pSTAT3 expression weakly correlated (r = 0.130; P = 0.068). Cytoplasmic STAT3 was inversely associated with the density of CD3+ (P = 0.012), CD8+ (P = 0.003), and FOXP3+ T lymphocytes (P = 0.002) within the cancer cell nests and was associated with an elevated systemic inflammatory response as measured by modified Glasgow Prognostic Score (mGPS2: 19% vs. 4%, P = 0.004).

The combination of nuclear STAT3/pSTAT3 stratified 5-year survival from 81% to 62% (P = 0.012), however, was not associated with survival independent of venous invasion, tumor perforation, or tumor budding.

Conclusions: In patients undergoing colorectal cancer resection, STAT3 expression was associated with adverse host inflammatory responses and reduced survival. Upregulation of tumor STAT3 may be an important mechanism whereby the tumor deregulates local and systemic inflammatory responses. Clin Cancer Res; 23(7); 1698–709. ©2016 AACR.

Translational Relevance

The presence of a conspicuous local inflammatory reaction is associated with improved survival of patients with colorectal cancer independent of stage, whereas the presence of elevated systemic inflammatory responses, as measured by acute phase proteins, is associated with decreased survival. One potential candidate pathway linking these responses is the Janus kinase/signal transduction and activator of transcription-3 (JAK/STAT3) pathway, with increasing evidence that this is a potential therapeutic target in cancer. In the present study, increased expression and nuclear localization of STAT3 was associated with aberrant local and systemic inflammatory responses in patients undergoing resection of stage I–III colorectal cancer, and was associated with poorer survival. In addition to suggesting a role for JAK/STAT3 inhibitors in restoring host antitumor immune responses, the results of the present study further support the rationale for stratifying patients by host inflammatory responses in future trials of such agents.

Colorectal cancer is the second most common cause of cancer death in Europe. Despite improved outcomes over the past decades, survival still remains poor, with 5-year survival of around 50% across all disease stages (1). Indeed, it is clear that the present tumor–node–metastasis (TNM)-based staging of colorectal cancer is suboptimal, with a need to identify characteristics pertaining to both the tumor and the host which may not only guide prognosis, but also the need for existing and novel adjuvant therapies.

One such characteristic is tumor-associated inflammation, which is now undisputed as affecting both the development and progression of cancer (2). In patients with colorectal cancer, for example, host local and systemic inflammatory responses are important determinants of disease progression, and their assessment is now accepted as holding independent prognostic value. To date, over 100 studies have examined the role of the local inflammatory cell infiltrate in determining outcome in patients with colorectal cancer (3), with a consistent, stage-independent decrease in disease recurrence and increase in survival observed in association with the presence of a conspicuous inflammatory cell infiltrate (3, 4).

In contrast, an elevated systemic inflammatory response is associated with increased risk of recurrence and reduced survival across a number of cancers including colorectal cancer (5). Upregulation of the systemic inflammatory response, as characterized by dysregulation of circulating proinflammatory cytokines, acute phase proteins, such as C-reactive protein (CRP) and albumin (6), and myeloid cells (7, 8), propagates a systemic environment which favors tumor growth and metastasis (9). Furthermore, routine assessment of the systemic inflammatory response utilizing routinely available biomarkers, such as acute phase proteins and components of the differential white cell count, informs prognosis complimentary to current TNM-based staging (8, 10).

One potential mechanism linking the local and systemic inflammatory responses is activation of the Janus-activated kinase/signal transduction and activator of transcription-3 (JAK/STAT3) pathway by IL6. Circulating IL6 is commonly elevated in a number of cancers, including colorectal cancer (11–13), and is the predominant stimulus for the hepatic synthesis of acute phase proteins, including CRP (6). Cancer-associated fibroblasts and inflammatory cells contribute to high levels of IL6 within the tumor microenvironment (14, 15), with subsequent tumor cell activation of the soluble IL6 receptor/glycoprotein 130 complex (16). IL6 trans-signaling regulates JAK activity within the tumor cell to promote phosphorylation of the tyrosine 705 residue of STAT3. Phosphorylated STAT3 (pSTAT3) translocates to the nucleus where it is a key transcription factor for numerous Th2-type cytokines, including IL6 (12, 14), which promote a protumor, immunosuppressive environment and attenuate host antitumor immune responses (15, 17). Indeed, given its role in not only deregulation of the host antitumor immune response, but also in orchestrating numerous pro-oncogenic processes (15, 18, 19), it is not surprising that STAT3 expression and activation has previously been associated with reduced survival in a number of gastrointestinal cancers, including colorectal cancer (20).

We hypothesize that the host systemic and local inflammatory responses in patients with colorectal cancer may be linked by STAT3. As such, the aim of the present study was to examine the relationship between tumor cell STAT3 expression, host local and systemic inflammatory responses, and survival in a cohort of patients undergoing potentially curative, elective resection of stage I–III colorectal cancer.

Patients who between 1997 and 2007 had undergone elective, potentially curative resection for stage I–III colorectal cancer in a single surgical unit in Glasgow Royal Infirmary, and who were included in a previously constructed tissue microarray (TMA) were included. Resection was considered curative on the basis of preoperative CT and intraoperative findings. Patients were excluded on the following criteria: emergency resection, resection with palliative intent, resection for IBD-associated colorectal cancer, known familial cancer syndrome, neoadjuvant therapy, underlying inflammatory condition, or death within 30 days of surgery. Local ethical approval was obtained from the West of Scotland Research Ethics Committee, and tissue for analysis of STAT3 expression was obtained from the National Health Service Greater Glasgow & Clyde Tissue Biorepository.

Patient demographics were collected prospectively. Tumors were staged using AJCC/UICC-TNM 5th edition, consistent with current Royal College of Pathologist guidelines (21). The presence of venous invasion was assessed routinely using elastica staining. Following surgery, patients were discussed at local colorectal cancer multidisciplinary meetings, where patients with stage III or high-risk stage II disease were considered for 5-fluorouracil–based chemotherapy according to treatment protocols at the time. Patients were followed up for a minimum of 5 years according to local guidelines at the time. Date and cause of death were crosschecked with the cancer registration system and the Registrar General (Scotland). Death records were complete until March 31, 2014, that served as the censor date. Cancer-specific survival was measured from date of surgery until date of death from colorectal cancer.

Serum CRP, albumin, and differential white cell count were measured within 30 days prior to surgery and recorded prospectively. Preoperative systemic inflammatory responses were defined using the modified Glasgow Prognostic Score (mGPS), the neutrophil:lymphocyte ratio (NLR), and the neutrophil:platelet score (NPS). The mGPS was calculated as previously described (10); patients with CRP ≤ 10 mg/L were allocated a score of 0, patients with CRP >10 mg/L and albumin ≥35 g/L were allocated a score of 1, and patients with CRP >10 mg/L and albumin <35 g/L were allocated a score of 2. On the basis of previously published thresholds, NLR > 5 was considered elevated (7). The NPS was calculated as previously described (22); patients with a platelet count <400 × 109/L and neutrophil count < 7. 5 × 109/L were allocated a score of 0, either a neutrophil count > 7.5 × 109/L or platelet count > 400 × 109/L a score of 1, and those with both an elevated neutrophil and platelet count a score of 2.

Assessment of tumor microenvironment

The tumor-associated stroma, the generalized local inflammatory cell infiltrate, and tumor budding have previously been characterized in this cohort using full hematoxylin & eosin (H&E)–stained sections of the deepest point of invasion according to previously published methodology (23–26). The tumor-associated stroma was assessed using tumor stroma percentage (TSP) and graded as either low (≤50%) or high (>50%; ref. 25). The local inflammatory cell infiltrate was assessed using the Klintrup–Mäkinen (KM) grade and graded as either low grade (no increase or mild/patchy increase in inflammatory cells at invasive margin) or high grade (prominent band-like inflammatory reaction or florid cup–like inflammatory reaction at invasive margin with destruction of cancer cell islands; ref. 24). Tumor budding was examined using the 10-high powered field method as previously described (26). Budding was considered high grade if greater than 20 buds (tumor cells with less than five nuclei or single tumor cells) were identified per 10-high powered fields.

Immunohistochemistry for CD3+ (mature), CD8+ (cytotoxic), CD45R0+ (memory), and FOXP3+ (regulatory) T lymphocytes was performed using formalin-fixed paraffin-embedded full sections of the deepest point of invasion as previously described (24). Cellular epitopes were identified using the following antibodies: CD3 (Vector Labs; code VP-RM01, 1:100 dilution), CD8 (DakoCytomation; code M7103, 1:100 dilution), CD45R0 (DakoCytomation; code M0742, 1:150 dilution), and FOXP3 (Abcam; code 20034, 1:200 dilution). T-lymphocyte density at the invasive margin or within the cancer cell nests was semiquantitatively graded using a four-point scale (absent/low/moderate/high). For the purposes of statistical analysis, density was subsequently graded as low (absent/low) or high (moderate/high). Investigators blinded to clinicopathologic data and outcomes performed all assessments, with independent scoring of 10% of cases by two investigators to ensure consistency.

Assessment of STAT3 expression

Immunohistochemical assessment of tumor cell STAT3 activity was performed using a previously constructed colorectal cancer TMA consisting of four 0.6 mm cores per patient (27). The TMA was constructed from formalin-fixed paraffin-embedded tissue blocks corresponding to the full sections utilized for assessment of the tumor microenvironment. In addition to activation of STAT3 by phosphorylation of the tyrosine 705 residue by IL6/JAK activation, activation may also occur through phosphorylation of the serine 727 residue in response to MAPK activation (28). As the present study hypothesized an association between the systemic and local inflammatory response via IL6/JAK/STAT3 activation, only total STAT3 expression and phosphorylated STAT3Tyr705 (pSTAT3) expression was measured. Sections were dewaxed in xylene before being rehydrated using graded alcohols. Antigen retrieval was performed using a citrate buffer at 96°C for 20 minutes for STAT3, and using a Tris-EDTA buffer at high pressure in a microwave for 5 minutes for pSTAT3. Endogenous peroxidase activity was blocked using 3% hydrogen peroxide for 10 minutes before rinsing in water. Casein and 5% horse serum in TBS were applied for 20 minutes at room temperature as a blocking agent for STAT3 and pSTAT3, respectively. Sections were then incubated overnight at 4°C with the primary antibody (STAT3: product code 9132; Cell Signaling Technologies; pSTAT3: product code 9131; Cell Signaling Technologies) at a concentration of 1:100 and 1:50 for STAT3 and pSTAT3, respectively, before washing in TBS for 10 minutes. Envision (Dako) was then added to the sections for 30 minutes at room temperature before washing in TBS for 10 minutes. DAB substrate was added for 5 minutes until color developed before washing in running water for 10 minutes. Slides were counterstained in hematoxylin for 60 seconds and blued with Scotts' tap water before dehydration through graded alcohols. Cover slips were applied using distrene, plasticizer, and xylene.

Sections were scanned using a Hamamatsu NanoZoomer at x20 magnification, and visualization was carried out using Slidepath Digital Image Hub (Slidepath, Leica Biosystems). Assessment of STAT3 and pSTAT3 expression within the cancer cell cytoplasm and nucleus was performed at x20 magnification by a single examiner (J.H. Park) blinded to clinical data using the weighted histoscore (29). To ensure reproducibility of scoring, 15% of tumors were co-scored by a second investigator (J. Clark); the intraclass correlation coefficient was 0.826 and 0.837, respectively. For the purposes of the present study, cytoplasmic STAT3 expression was considered representative of total STAT3 expression, whereas nuclear STAT3 and pSTAT3 expressions were considered representative of STAT3 transcriptional activity.

Assessment of mismatch repair status

Mismatch repair (MMR) protein deficiency was determined using the TMA utilized for STAT3 assessment. Sections were stained for MLH1, MSH6, MSH2, and PMS2 (product codes: M3640, M3646, M3639, and M3647, respectively; Dako UK Ltd.) as described previously (30). MMR status was determined according to UK NEQAS guidelines (31), using appendix and normal colon as positive controls and intratumoral lymphocytes as internal positive controls. Tumors were considered MMR competent if tumor cell nuclear expression was positive with positive immune cell expression, and MMR deficient if tumor nuclear expression was absent with normal immune cell expression.

Statistical analysis

For the purpose of statistical analysis, patients were divided into tertiles (low/moderate/high) on the basis of cytoplasmic and nuclear STAT3 and pSTAT3 expression as measured by h-score. The relationship between clinicopathologic characteristics and cytoplasmic and nuclear STAT3 expression was examined using the χ2 test for linear trend. The relationship between STAT3 expression and 5-year cancer-specific survival was examined using the Kaplan–Meier log-rank analysis and displayed as percentage surviving (standard error). The relationship between STAT3 expression, clinicopathologic characteristics, and cancer-specific survival was examined using Cox proportional hazards regression; variables with a P ≤ 0.1 on univariate analysis were entered into a multivariate model using a backwards conditional model to calculate HRs and 95% confidence intervals (CI). Given the number of comparisons performed, a P value ≤0.01 was considered statistically significant for χ2 analysis, with a P value ≤0.05 considered statistically significant for survival analysis. All analyses were performed using SPSS version 22.0 (IBM SPSS).

A total of 196 patients who underwent elective, potentially curative resection of stage I–III colorectal cancer were included. Clinicopathologic characteristics are displayed in Table 1. Almost two thirds of patients were older than 65 at time of surgery, and 52% were male. Pathologic assessment confirmed Stage I disease in 16 patients (8%), stage II disease in 94 patients (48%), and stage III disease in 86 patients (44%). Fifty-four patients (28%) received adjuvant therapy; 1 patient with stage I disease, 14 patients with stage II disease, and 39 patients with stage III disease received adjuvant therapy. MMR deficiency was identified in 27 patients (14%).

Table 1.

The relationship between tumor cell STAT3 and pSTAT3 expression and clinicopathologic characteristics of patients undergoing elective, potentially curative resection of stage I–III colorectal cancer

Cytoplasmic STAT3 h-scoreNuclear STAT3 h-scoreNuclear pSTAT3 h-score
AllLow (0–20)Mod (21–65)High (66–168)Low (0–15)Mod (16–30)High (30–130)Low (5–80)Mod (81–105)High (106–205)
n = 196 (%)n = 76 (%)n = 56 (%)n = 64 (%)Pn = 75 (%)n = 66 (%)Pn = 72 (%)n = 61 (%)n = 63 (%)P
Host characteristics 
 Age     0.571    0.199    0.026 
  <65 72 (37) 23 (30) 22 (39) 27 (42)  29 (39) 25 (37) 18 (33)  19 (26) 26 (43) 27 (43)  
  65–74 61 (31) 29 (38) 18 (32) 14 (22)  25 (33) 22 (33) 14 (26)  23 (32) 19 (31) 19 (30)  
  >75 63 (32) 24 (32) 16 (29) 23 (36)  21 (28) 10 (29) 23 (41)  30 (42) 16 (26) 17 (27)  
 Sex     0.833    0.647    0.906 
  Female 94 (48) 40 (53) 21 (37) 33 (52)  35 (47) 31 (47) 28 (51)  38 (53) 23 (38) 33 (52)  
  Male 102 (52) 36 (47) 35 (63) 31 (48)  40 (53) 35 (53) 27 (49)  34 (47) 38 (62) 30 (48)  
 Adjuvant therapy     0.532    0.038    0.389 
  No 142 (72) 55 (72) 44 (79) 43 (67)  48 (64) 50 (76) 44 (80)  56 (78) 41 (67) 45 (71)  
  Yes 54 (28) 21 (28) 12 (21) 21 (33)  27 (36) 16 (24) 11 (20)  16 (22) 20 (33) 18 (29)  
Tumor characteristics 
 Tumor location     0.375    0.242    0.860 
  Colon 130 (66) 48 (63) 37 (66) 45 (70)  47 (63) 43 (65) 40 (73)  49 (68) 37 (61) 44 (70)  
  Rectum 66 (34) 28 (37) 19 (34) 19 (30)  28 (37) 23 (35) 15 (27)  23 (32) 24 (39) 19 (30)  
 T stage     0.288    0.480    0.694 
  1–2 25 (13) 10 (13) 9 (16) 6 (9)  10 (13) 10 (15) 5 (9)  10 (14) 8 (13) 7 (11)  
  3 121 (61) 49 (65) 34 (61) 38 (59)  46 (61) 41 (62) 34 (62)  43 (60) 39 (64) 39 (62)  
  4 50 (26) 17 (22) 13 (23) 20 (31)  19 (25) 15 (23) 16 (29)  19 (26) 14 (23) 17 (27)  
 N stage     0.183    0.470    0.039 
  0 110 (56) 47 (61) 30 (53) 33 (51)  34 (45) 46 (70) 30 (54)  47 (65) 33 (54) 30 (48)  
  1 68 (35) 24 (32) 20 (36) 24 (38)  34 (45) 16 (24) 18 (33)  21 (29) 21 (34) 26 (41)  
  2 18 (9) 5 (7) 6 (11) 7 (11)  7 (10) 4 (6) 7 (13)  4 (6) 7 (12) 7 (11)  
 TNM stage     0.211    0.494    0.051 
  I 16 (8) 6 (8) 7 (13) 3 (5)  6 (8) 8 (12) 2 (4)  7 (10) 5 (8) 4 (6)  
  II 94 (48) 41 (54) 23 (41) 30 (47)  28 (37) 38 (58) 28 (50)  40 (55) 28 (46) 26 (41)  
  III 86 (44) 29 (38) 26 (46) 31 (48)  41 (55) 20 (30) 25 (46)  25 (35) 28 (46) 33 (52)  
 Tumor differentiation     0.530    0.108    0.174 
  Mod/well 174 (89) 69 (91) 49 (87) 56 (87)  63 (84) 60 (91) 51 (93)  60 (83) 57 (93) 57 (91)  
  Poor 22 (11) 7 (9) 7 (13) 8 (13)  12 (16) 6 (9) 4 (7)  12 (17) 4 (7) 6 (10)  
 Venous invasion     0.465    0.337    0.271 
  No 129 (66) 51 (67) 39 (70) 39 (61)  46 (61) 45 (68) 38 (69)  51 (71) 39 (64) 39 (62)  
  Yes 67 (34) 25 (33) 17 (30) 25 (39)  29 (39) 21 (32) 17 (31)  21 (29) 22 (36) 24 (38)  
 Margin involvement     0.856    0.649    0.562 
  No 187 (95) 72 (95) 54 (96) 61 (95)  70 (93) 65 (98) 52 (94)  70 (97) 57 (93) 60 (95)  
  Yes 9 (5) 4 (5) 2 (4) 3 (5)  5 (7) 1 (2) 3 (6)  2 (3) 4 (7) 3 (5)  
 Peritoneal involvement     0.423    0.794    0.787 
  No 144 (3) 57 (75) 43 (77) 44 (69)  55 (73) 50 (76) 39 (71)  53 (74) 46 (75) 45 (71)  
  Yes 52 (27) 19 (25) 13 (23) 20 (31)  20 (27) 16 (24) 16 (29)  19 (26) 15 (25) 18 (29)  
 Tumor perforation     0.652    0.799    0.087 
  No 192 (98) 74 (97) 55 (98) 63 (98)  73 (97) 66 (100) 53 (96)  69 (96) 60 (98) 63 (100)  
  Yes 4 (2) 2 (3) 1 (2) 1 (2)  2 (3) 0 (0) 2 (4)  3 (4) 1 (2) 0 (0)  
 MMR status     0.741    0.406    0.491 
  Competent 169 (86) 65 (85) 48 (86) 56 (87)  62 (83) 59 (89) 48 (87)  61 (85) 52 (85) 56 (89)  
  Deficient 27 (14) 11 (15) 8 (14) 8 (13)  13 (17) 7 (11) 7 (13)  11 (15) 9 (15) 7 (11)  
Cytoplasmic STAT3 h-scoreNuclear STAT3 h-scoreNuclear pSTAT3 h-score
AllLow (0–20)Mod (21–65)High (66–168)Low (0–15)Mod (16–30)High (30–130)Low (5–80)Mod (81–105)High (106–205)
n = 196 (%)n = 76 (%)n = 56 (%)n = 64 (%)Pn = 75 (%)n = 66 (%)Pn = 72 (%)n = 61 (%)n = 63 (%)P
Host characteristics 
 Age     0.571    0.199    0.026 
  <65 72 (37) 23 (30) 22 (39) 27 (42)  29 (39) 25 (37) 18 (33)  19 (26) 26 (43) 27 (43)  
  65–74 61 (31) 29 (38) 18 (32) 14 (22)  25 (33) 22 (33) 14 (26)  23 (32) 19 (31) 19 (30)  
  >75 63 (32) 24 (32) 16 (29) 23 (36)  21 (28) 10 (29) 23 (41)  30 (42) 16 (26) 17 (27)  
 Sex     0.833    0.647    0.906 
  Female 94 (48) 40 (53) 21 (37) 33 (52)  35 (47) 31 (47) 28 (51)  38 (53) 23 (38) 33 (52)  
  Male 102 (52) 36 (47) 35 (63) 31 (48)  40 (53) 35 (53) 27 (49)  34 (47) 38 (62) 30 (48)  
 Adjuvant therapy     0.532    0.038    0.389 
  No 142 (72) 55 (72) 44 (79) 43 (67)  48 (64) 50 (76) 44 (80)  56 (78) 41 (67) 45 (71)  
  Yes 54 (28) 21 (28) 12 (21) 21 (33)  27 (36) 16 (24) 11 (20)  16 (22) 20 (33) 18 (29)  
Tumor characteristics 
 Tumor location     0.375    0.242    0.860 
  Colon 130 (66) 48 (63) 37 (66) 45 (70)  47 (63) 43 (65) 40 (73)  49 (68) 37 (61) 44 (70)  
  Rectum 66 (34) 28 (37) 19 (34) 19 (30)  28 (37) 23 (35) 15 (27)  23 (32) 24 (39) 19 (30)  
 T stage     0.288    0.480    0.694 
  1–2 25 (13) 10 (13) 9 (16) 6 (9)  10 (13) 10 (15) 5 (9)  10 (14) 8 (13) 7 (11)  
  3 121 (61) 49 (65) 34 (61) 38 (59)  46 (61) 41 (62) 34 (62)  43 (60) 39 (64) 39 (62)  
  4 50 (26) 17 (22) 13 (23) 20 (31)  19 (25) 15 (23) 16 (29)  19 (26) 14 (23) 17 (27)  
 N stage     0.183    0.470    0.039 
  0 110 (56) 47 (61) 30 (53) 33 (51)  34 (45) 46 (70) 30 (54)  47 (65) 33 (54) 30 (48)  
  1 68 (35) 24 (32) 20 (36) 24 (38)  34 (45) 16 (24) 18 (33)  21 (29) 21 (34) 26 (41)  
  2 18 (9) 5 (7) 6 (11) 7 (11)  7 (10) 4 (6) 7 (13)  4 (6) 7 (12) 7 (11)  
 TNM stage     0.211    0.494    0.051 
  I 16 (8) 6 (8) 7 (13) 3 (5)  6 (8) 8 (12) 2 (4)  7 (10) 5 (8) 4 (6)  
  II 94 (48) 41 (54) 23 (41) 30 (47)  28 (37) 38 (58) 28 (50)  40 (55) 28 (46) 26 (41)  
  III 86 (44) 29 (38) 26 (46) 31 (48)  41 (55) 20 (30) 25 (46)  25 (35) 28 (46) 33 (52)  
 Tumor differentiation     0.530    0.108    0.174 
  Mod/well 174 (89) 69 (91) 49 (87) 56 (87)  63 (84) 60 (91) 51 (93)  60 (83) 57 (93) 57 (91)  
  Poor 22 (11) 7 (9) 7 (13) 8 (13)  12 (16) 6 (9) 4 (7)  12 (17) 4 (7) 6 (10)  
 Venous invasion     0.465    0.337    0.271 
  No 129 (66) 51 (67) 39 (70) 39 (61)  46 (61) 45 (68) 38 (69)  51 (71) 39 (64) 39 (62)  
  Yes 67 (34) 25 (33) 17 (30) 25 (39)  29 (39) 21 (32) 17 (31)  21 (29) 22 (36) 24 (38)  
 Margin involvement     0.856    0.649    0.562 
  No 187 (95) 72 (95) 54 (96) 61 (95)  70 (93) 65 (98) 52 (94)  70 (97) 57 (93) 60 (95)  
  Yes 9 (5) 4 (5) 2 (4) 3 (5)  5 (7) 1 (2) 3 (6)  2 (3) 4 (7) 3 (5)  
 Peritoneal involvement     0.423    0.794    0.787 
  No 144 (3) 57 (75) 43 (77) 44 (69)  55 (73) 50 (76) 39 (71)  53 (74) 46 (75) 45 (71)  
  Yes 52 (27) 19 (25) 13 (23) 20 (31)  20 (27) 16 (24) 16 (29)  19 (26) 15 (25) 18 (29)  
 Tumor perforation     0.652    0.799    0.087 
  No 192 (98) 74 (97) 55 (98) 63 (98)  73 (97) 66 (100) 53 (96)  69 (96) 60 (98) 63 (100)  
  Yes 4 (2) 2 (3) 1 (2) 1 (2)  2 (3) 0 (0) 2 (4)  3 (4) 1 (2) 0 (0)  
 MMR status     0.741    0.406    0.491 
  Competent 169 (86) 65 (85) 48 (86) 56 (87)  62 (83) 59 (89) 48 (87)  61 (85) 52 (85) 56 (89)  
  Deficient 27 (14) 11 (15) 8 (14) 8 (13)  13 (17) 7 (11) 7 (13)  11 (15) 9 (15) 7 (11)  

Expression of STAT3 was observed in both the cytoplasm and nucleus, whereas pSTAT3 expression was only observed in the nucleus. The h-score range for cytoplasmic and nuclear STAT3 expression ranged from 0 to 168 and from 0 to 130, respectively. Nuclear pSTAT3 h-score for the cohort ranged from 5 to 205. Cytoplasmic expression of STAT3 was associated with nuclear expression of STAT3 (Spearman r = 0.363, P < 0.001) but not pSTAT3 (r = 0.111, P = 0.121). Nuclear STAT3 expression was not significantly associated with nuclear pSTAT3 expression (r = 0.130, P = 0.068). Normal, noncancer epithelium expression of STAT3 was available for 10 patients. Although this precluded meaningful statistical analysis, it was of interest that 7 patients showed similar or higher expression of cytoplasmic STAT3, nuclear STAT3, and nuclear pSTAT3 in normal tissue compared with cancer tissue. The remaining 3 patients showed heterogeneous expression of each of the studied markers.

The relationship between STAT3 and pSTAT3 expression tertiles and clinicopathologic characteristics is displayed in Table 1. Cytoplasmic expression of STAT3 was not associated with any clinicopathologic characteristics. Although failing to reach statistical significance, nuclear STAT3 expression showed an inverse association with use of adjuvant chemotherapy (P = 0.038), whereas pSTAT3 expression was associated with younger age (P = 0.026) and an increased prevalence of lymph node positive disease (low pSTAT3 expression—35% vs. high pSTAT3 expression—52%, P = 0.039).

The relationship between STAT3 and pSTAT3 expression and components of the tumor microenvironment is displayed in Table 2. Cytoplasmic STAT3 expression was inversely associated with the cancer cell nest density of CD8+ and FOXP3+ (both P < 0.01) T lymphocytes and showed a trend toward a similar relationship with CD3+ density (P = 0.012) but was not associated with TSP, tumor budding, or the local inflammatory cell density at the invasive margin as measured by KM grade or T-lymphocyte density. Nuclear expression of STAT3 showed no statistically significant association with characteristics of the tumor microenvironment; however, a lower density of CD8+ (P = 0.039) and CD3+ (P = 0.055) T lymphocytes was identified in patients with nuclear STAT3 expression. There were no statistically significant associations between nuclear pSTAT3 expression and tumor microenvironment characteristics; patients with high nuclear pSTAT3 expression however were observed to have a lower density of CD45R0+ T lymphocytes (P = 0.037) and more frequent high-grade tumor budding (P = 0.022).

Table 2.

The relationship between tumor cell STAT3 and pSTAT3 expression and tumor microenvironment of patients undergoing elective, potentially curative resection of stage I–III colorectal cancer

Cytoplasmic STAT3 h-scoreNuclear STAT3 h-scoreNuclear pSTAT3 h-score
Low (0–20)Mod (21–65)High (66–168)Low (0–15)Mod (16–30)High (30–130)Low (5–80)Mod (81–105)High (106–205)
n = 76 (%)n = 56 (%)n = 64 (%)Pn = 75 (%)n = 66 (%)n = 55 (%)Pn = 72 (%)n = 61 (%)n = 63 (%)P
KM grade    0.208    0.657    0.582 
 Weak 28 (37) 20 (36) 17 (27)  25 (33) 24 (36) 16 (29)  26 (36) 19 (31) 20 (32)  
 Strong 48 (63) 36 (64) 47 (73)  50 (67) 42 (64) 30 (71)  46 (64) 42 (69) 43 (68)  
Tumor stroma percentage (195)    0.241    0.794    0.090 
 Low 59 (78) 43 (78) 44 (69)  56 (75) 51 (77) 39 (72)  55 (78) 50 (82) 40 (64)  
 High 17 (22) 12 (22) 20 (31)  19 (25) 15 (23) 15 (28)  16 (22) 11 (18) 22 (36)  
Tumor budding (182)    0.834    0.822    0.022 
 Low 45 (64) 40 (74) 38 (65)  45 (63) 44 (76) 34 (64)  49 (75) 41 (71) 33 (56)  
 High 25 (36) 14 (26) 20 (35)  26 (37) 14 (24) 19 (36)  16 (25) 17 (29) 26 (44)  
CD3+ margin density (184)    0.332    0.055    0.648 
 Low 36 (49) 30 (60) 35 (57)  37 (51) 28 (46) 36 (71)  34 (54) 31 (52) 36 (58)  
 High 37 (51) 20 (40) 26 (43)  35 (49) 33 (54) 15 (29)  29 (46) 28 (48) 26 (42)  
CD3+ cancer cell nest density (192)    0.012    0.262    0.150 
 Low 38 (51) 42 (79) 45 (70)  47 (64) 38 (59) 40 (74)  43 (62) 35 (58) 47 (75)  
 High 37 (49) 11 (21) 19 (30)  27 (37) 26 (41) 14 (26)  26 (38) 25 (42) 16 (25)  
CD8+ margin density (184)    0.630    0.177    0.806 
 Low 41 (59) 34 (64) 33 (54)  38 (53) 37 (61) 33 (65)  38 (59) 33 (55) 37 (62)  
 High 29 (41) 19 (36) 28 (46)  34 (47) 25 (39) 18 (35)  26 (41) 27 (45) 23 (38)  
CD8+ cancer cell nest density (190)    0.003    0.039    0.730 
 Low 41 (57) 45 (83) 51 (80)  47 (63) 47 (76) 43 (80)  50 (72) 41 (68) 46 (75)  
 High 31 (43) 9 (17) 13 (20)  27 (37) 15 (24) 11 (20)  19 (28) 19 (32) 15 (25)  
CD45R0+ margin density (186)    0.960    0.089    0.282 
 Low 38 (52) 27 (51) 31 (52)  33 (47) 31 (48) 32 (63)  32 (48) 29 (50) 37 (57)  
 High 35 (48) 26 (49) 29 (48)  38 (54) 33 (52) 19 (37)  38 (52) 29 (50) 26 (43)  
CD45R0+ cancer cell density (192)    0.408    0.268    0.037 
 Low 48 (64) 43 (80) 44 (70)  48 (67) 46 (70) 41 (76)  46 (64) 39 (67) 50 (81)  
 High 27 (36) 11 (20) 19 (30)  24 (33) 20 (30) 13 (24)  26 (36) 19 (33) 12 (19)  
FOXP3+ margin density (186)    0.180    0.373    0.466 
 Low 37 (51) 29 (56) 38 (62)  39 (53) 34 (54) 31 (62)  40 (60) 32 (54) 32 (53)  
 High 36 (49) 23 (44) 23 (38)  34 (47) 29 (46) 19 (38)  27 (40) 27 (46) 28 (47)  
FOXP3+ cancer cell nest density (188)    0.002    0.807    0.181 
 Low 26 (36) 26 (49) 39 (63)  39 (53) 25 (39) 27 (53)  38 (56) 26 (44) 27 (44)  
 High 47 (64) 27 (51) 23 (37)  34 (47) 39 (61) 24 (47)  30 (44) 33 (56) 34 (56)  
Cytoplasmic STAT3 h-scoreNuclear STAT3 h-scoreNuclear pSTAT3 h-score
Low (0–20)Mod (21–65)High (66–168)Low (0–15)Mod (16–30)High (30–130)Low (5–80)Mod (81–105)High (106–205)
n = 76 (%)n = 56 (%)n = 64 (%)Pn = 75 (%)n = 66 (%)n = 55 (%)Pn = 72 (%)n = 61 (%)n = 63 (%)P
KM grade    0.208    0.657    0.582 
 Weak 28 (37) 20 (36) 17 (27)  25 (33) 24 (36) 16 (29)  26 (36) 19 (31) 20 (32)  
 Strong 48 (63) 36 (64) 47 (73)  50 (67) 42 (64) 30 (71)  46 (64) 42 (69) 43 (68)  
Tumor stroma percentage (195)    0.241    0.794    0.090 
 Low 59 (78) 43 (78) 44 (69)  56 (75) 51 (77) 39 (72)  55 (78) 50 (82) 40 (64)  
 High 17 (22) 12 (22) 20 (31)  19 (25) 15 (23) 15 (28)  16 (22) 11 (18) 22 (36)  
Tumor budding (182)    0.834    0.822    0.022 
 Low 45 (64) 40 (74) 38 (65)  45 (63) 44 (76) 34 (64)  49 (75) 41 (71) 33 (56)  
 High 25 (36) 14 (26) 20 (35)  26 (37) 14 (24) 19 (36)  16 (25) 17 (29) 26 (44)  
CD3+ margin density (184)    0.332    0.055    0.648 
 Low 36 (49) 30 (60) 35 (57)  37 (51) 28 (46) 36 (71)  34 (54) 31 (52) 36 (58)  
 High 37 (51) 20 (40) 26 (43)  35 (49) 33 (54) 15 (29)  29 (46) 28 (48) 26 (42)  
CD3+ cancer cell nest density (192)    0.012    0.262    0.150 
 Low 38 (51) 42 (79) 45 (70)  47 (64) 38 (59) 40 (74)  43 (62) 35 (58) 47 (75)  
 High 37 (49) 11 (21) 19 (30)  27 (37) 26 (41) 14 (26)  26 (38) 25 (42) 16 (25)  
CD8+ margin density (184)    0.630    0.177    0.806 
 Low 41 (59) 34 (64) 33 (54)  38 (53) 37 (61) 33 (65)  38 (59) 33 (55) 37 (62)  
 High 29 (41) 19 (36) 28 (46)  34 (47) 25 (39) 18 (35)  26 (41) 27 (45) 23 (38)  
CD8+ cancer cell nest density (190)    0.003    0.039    0.730 
 Low 41 (57) 45 (83) 51 (80)  47 (63) 47 (76) 43 (80)  50 (72) 41 (68) 46 (75)  
 High 31 (43) 9 (17) 13 (20)  27 (37) 15 (24) 11 (20)  19 (28) 19 (32) 15 (25)  
CD45R0+ margin density (186)    0.960    0.089    0.282 
 Low 38 (52) 27 (51) 31 (52)  33 (47) 31 (48) 32 (63)  32 (48) 29 (50) 37 (57)  
 High 35 (48) 26 (49) 29 (48)  38 (54) 33 (52) 19 (37)  38 (52) 29 (50) 26 (43)  
CD45R0+ cancer cell density (192)    0.408    0.268    0.037 
 Low 48 (64) 43 (80) 44 (70)  48 (67) 46 (70) 41 (76)  46 (64) 39 (67) 50 (81)  
 High 27 (36) 11 (20) 19 (30)  24 (33) 20 (30) 13 (24)  26 (36) 19 (33) 12 (19)  
FOXP3+ margin density (186)    0.180    0.373    0.466 
 Low 37 (51) 29 (56) 38 (62)  39 (53) 34 (54) 31 (62)  40 (60) 32 (54) 32 (53)  
 High 36 (49) 23 (44) 23 (38)  34 (47) 29 (46) 19 (38)  27 (40) 27 (46) 28 (47)  
FOXP3+ cancer cell nest density (188)    0.002    0.807    0.181 
 Low 26 (36) 26 (49) 39 (63)  39 (53) 25 (39) 27 (53)  38 (56) 26 (44) 27 (44)  
 High 47 (64) 27 (51) 23 (37)  34 (47) 39 (61) 24 (47)  30 (44) 33 (56) 34 (56)  

When analysis was restricted to patients with MMR-competent colorectal cancer (Supplementary Table S1), the observed relationship between cytoplasmic STAT3 and cancer cell nest density of CD3+ (P = 0.061), CD8+ (P < 0.05), and FOXP3+ (P < 0.01) T lymphocytes remained. Nuclear STAT3 was no longer associated with CD8+ density within cancer cell nests but was associated with CD3+ density within the invasive margin (P < 0.05). Nuclear pSTAT3 expression again showed a nonsignificant trend toward low cancer cell nest density of CD45R0+ T lymphocytes. Although the small number of patients limited statistical power, when analysis was restricted to patients with MMR-deficient colorectal cancer, the relationship between cytoplasmic STAT3 expression and cancer cell nest density of CD3+ (P < 0.05) and CD8+ (P < 0.01) T cells, and nuclear STAT3 expression, and cancer cell nest density of CD8+ T cells (P < 0.05) remained. Nuclear pSTAT3 expression, however, was not associated with T-lymphocyte density of patients with MMR-deficient colorectal cancer.

The relationship between STAT3 and pSTAT3 expression and systemic inflammatory responses is displayed in Table 3. Cytoplasmic STAT3 expression was associated with the systemic inflammatory response as measured by mGPS; this was predominantly due to an increase in the number of patients with mGPS = 2 (high expression—19% vs. low expression 4%, P = 0.004). Neither cytoplasmic nor nuclear STAT3 expressions were associated with the systemic inflammatory response as measured by circulating platelets or components of the differential white cell count. Nuclear pSTAT3 expression was not associated with any measures of the systemic inflammatory response.

Table 3.

The relationship between tumor cell STAT3 and pSTAT3 expression and systemic inflammatory responses of patients undergoing elective, potentially curative resection of stage I–III colorectal cancer

Cytoplasmic STAT3 h-scoreNuclear STAT3 h-scoreNuclear pSTAT3 h-score
Low (0–20)Mod (21–65)High (66–168)Low (0–15)Mod (16–30)High (30–130)Low (5–80)Mod (81–105)High (106–205)
n = 76 (%)n = 56 (%)n = 64 (%)Pn = 75 (%)n = 66 (%)n = 55 (%)Pn = 72 (%)n = 61 (%)n = 63 (%)P
Modified Glasgow Prognostic Score    0.004    0.244    0.651 
 0 53 (70) 33 (59) 33 (51)  46 (61) 42 (64) 31 (56)  44 (61) 36 (59) 39 (62)  
 1 20 (26) 18 (32) 19 (30)  23 (31) 20 (30) 14 (26)  20 (28) 17 (28) 20 (32)  
 2 3 (4) 5 (9) 12 (19)  6 (8) 4 (6) 10 (18)  8 (11) 8 (13) 4 (6)  
Neutrophil count (195)    0.515    0.676    0.470 
 ≤7.5 × 109/L 67 (88) 47 (85) 54 (84)  63 (85) 60 (91) 45 (82)  60 (85) 52 (85) 56 (89)  
 >7.5 × 109/L 9 (12) 8 (15) 10 (16)  11 (15) 6 (9) 10 (18)  11 (16) 9 (15) 7 (11)  
Lymphocyte count (195)    0.942    0.174    0.209 
 >4 × 109/L 76 (100) 54 (98) 64 (100)  74 (100) 66 (100) 54 (98)  71 (100) 61 (100) 62 (98)  
 ≤4 × 109/L 0 (0) 1 (2) 0 (0)  0 (0) 0 (0) 1 (2)  0 (0) 0 (0) 1 (2)  
Platelet count (176)    0.557    0.587    0.895 
 ≤400 × 109/L 58 (87) 44 (86) 48 (83)  55 (85) 49 (83) 46 (88)  57 (85) 44 (85) 49 (86)  
 >400 × 109/L 9 (13) 7 (14) 10 (17)  10 (15) 10 (17) 6 (12)  10 (15) 8 (15) 8 (14)  
NLR (195)    0.350    0.131    0.352 
 ≤5 62 (82) 45 (82) 48 (75)  61 (82) 55 (83) 39 (71)  56 (79) 45 (74) 54 (86)  
 >5 14 (18) 10 (18) 16 (25)  13 (18) 11 (17) 16 (29)  15 (21) 16 (26) 9 (14)  
NPS (176)    0.441    0.831    0.602 
 0 52 (78) 40 (78) 40 (69)  47 (72) 46 (78) 39 (75)  49 (73) 39 (75) 44 (77)  
 1 13 (19) 7 (14) 17 (29)  16 (25) 10 (17) 11 (21)  15 (22) 11 (21) 11 (19)  
 2 2 (3) 5 (8) 1 (2)  2 (3) 3 (5) 2 (4)  3 (5) 2 (4) 2 (4)  
Cytoplasmic STAT3 h-scoreNuclear STAT3 h-scoreNuclear pSTAT3 h-score
Low (0–20)Mod (21–65)High (66–168)Low (0–15)Mod (16–30)High (30–130)Low (5–80)Mod (81–105)High (106–205)
n = 76 (%)n = 56 (%)n = 64 (%)Pn = 75 (%)n = 66 (%)n = 55 (%)Pn = 72 (%)n = 61 (%)n = 63 (%)P
Modified Glasgow Prognostic Score    0.004    0.244    0.651 
 0 53 (70) 33 (59) 33 (51)  46 (61) 42 (64) 31 (56)  44 (61) 36 (59) 39 (62)  
 1 20 (26) 18 (32) 19 (30)  23 (31) 20 (30) 14 (26)  20 (28) 17 (28) 20 (32)  
 2 3 (4) 5 (9) 12 (19)  6 (8) 4 (6) 10 (18)  8 (11) 8 (13) 4 (6)  
Neutrophil count (195)    0.515    0.676    0.470 
 ≤7.5 × 109/L 67 (88) 47 (85) 54 (84)  63 (85) 60 (91) 45 (82)  60 (85) 52 (85) 56 (89)  
 >7.5 × 109/L 9 (12) 8 (15) 10 (16)  11 (15) 6 (9) 10 (18)  11 (16) 9 (15) 7 (11)  
Lymphocyte count (195)    0.942    0.174    0.209 
 >4 × 109/L 76 (100) 54 (98) 64 (100)  74 (100) 66 (100) 54 (98)  71 (100) 61 (100) 62 (98)  
 ≤4 × 109/L 0 (0) 1 (2) 0 (0)  0 (0) 0 (0) 1 (2)  0 (0) 0 (0) 1 (2)  
Platelet count (176)    0.557    0.587    0.895 
 ≤400 × 109/L 58 (87) 44 (86) 48 (83)  55 (85) 49 (83) 46 (88)  57 (85) 44 (85) 49 (86)  
 >400 × 109/L 9 (13) 7 (14) 10 (17)  10 (15) 10 (17) 6 (12)  10 (15) 8 (15) 8 (14)  
NLR (195)    0.350    0.131    0.352 
 ≤5 62 (82) 45 (82) 48 (75)  61 (82) 55 (83) 39 (71)  56 (79) 45 (74) 54 (86)  
 >5 14 (18) 10 (18) 16 (25)  13 (18) 11 (17) 16 (29)  15 (21) 16 (26) 9 (14)  
NPS (176)    0.441    0.831    0.602 
 0 52 (78) 40 (78) 40 (69)  47 (72) 46 (78) 39 (75)  49 (73) 39 (75) 44 (77)  
 1 13 (19) 7 (14) 17 (29)  16 (25) 10 (17) 11 (21)  15 (22) 11 (21) 11 (19)  
 2 2 (3) 5 (8) 1 (2)  2 (3) 3 (5) 2 (4)  3 (5) 2 (4) 2 (4)  

The median follow-up of survivors was 143 months (range, 101–204) with 57 cancer-associated deaths and 64 noncancer deaths. For the purposes of survival analysis, low and moderate expression of each marker was combined to form one group (low expression). The relationship between cytoplasmic STAT3, nuclear STAT3 and nuclear pSTAT3 and cancer-specific survival is displayed in Fig. 1 and in Table 4. High nuclear STAT3 expression was associated with poorer cancer-specific survival (P < 0.05). High expression of both cytoplasmic STAT3 expression and nuclear pSTAT3 expression showed a nonsignificant trend toward poorer survival (P = 0.068 and P = 0.116, respectively).

Figure 1.

The relationship between tumor cell STAT3 expression and cancer-specific survival of patients undergoing elective, potentially curative resection of stage I–III colorectal cancer (Kaplan–Meier log-rank analysis): (A) cytoplasmic STAT3 expression (P = 0.068), (B) nuclear STAT3 expression (P = 0.012), (C) nuclear pSTAT3 expression (P = 0.116), and (D) combined nuclear STAT3/pSTAT3 expression (P = 0.012).

Figure 1.

The relationship between tumor cell STAT3 expression and cancer-specific survival of patients undergoing elective, potentially curative resection of stage I–III colorectal cancer (Kaplan–Meier log-rank analysis): (A) cytoplasmic STAT3 expression (P = 0.068), (B) nuclear STAT3 expression (P = 0.012), (C) nuclear pSTAT3 expression (P = 0.116), and (D) combined nuclear STAT3/pSTAT3 expression (P = 0.012).

Close modal
Table 4.

Relationship between tumor cell STAT3 and pSTAT3 expression and cancer-specific survival of patients undergoing elective, potentially curative resection of stage I–III colorectal cancer

N5-year CSS % (SE)Univariate HR (95% CI)PMultivariate HR (95% CI)P
Cytoplasmic STAT3    0.072  – 
 Low-mod expression 132 81 (3) –  –  
 High expression 64 67 (6) 1.62 (0.96–2.65)  –  
Nuclear STAT3    0.018  – 
 Low-mod expression 141 78 (4) –  –  
 High expression 55 70 (6) 1.89 (1.12–3.22)  –  
Nuclear pSTAT3    0.119  – 
 Low-mod expression 133 80 (4) –  –  
 High expression 63 69 (6) 1.52 (0.90–2.57)  –  
Combined cytoplasmic STAT3/nuclear STAT3 (Model 1)    0.009  0.221 
 Both low-mod 106 81 (4)     
 One high 61 73 (6) 1.56 (1.20–2.17)  –  
 Both high 29 63 (9)     
Combined cytoplasmic STAT3/nuclear pSTAT3 (Model 2)    0.024  0.526 
 Both low-mod 95 80 (4)     
 One high 75 79 (5) 1.50 (1.06–2.13)  –  
 Both high 26 54 (10)     
Combined nuclear STAT3/nuclear pSTAT3 (Model 3)    0.008  0.008 
 Both low-mod 100 81 (4)     
 One high 74 74 (5) 1.63 (1.14–2.34)  1.63 (1.14–2.34)  
 Both high 22 62 (11)     
N5-year CSS % (SE)Univariate HR (95% CI)PMultivariate HR (95% CI)P
Cytoplasmic STAT3    0.072  – 
 Low-mod expression 132 81 (3) –  –  
 High expression 64 67 (6) 1.62 (0.96–2.65)  –  
Nuclear STAT3    0.018  – 
 Low-mod expression 141 78 (4) –  –  
 High expression 55 70 (6) 1.89 (1.12–3.22)  –  
Nuclear pSTAT3    0.119  – 
 Low-mod expression 133 80 (4) –  –  
 High expression 63 69 (6) 1.52 (0.90–2.57)  –  
Combined cytoplasmic STAT3/nuclear STAT3 (Model 1)    0.009  0.221 
 Both low-mod 106 81 (4)     
 One high 61 73 (6) 1.56 (1.20–2.17)  –  
 Both high 29 63 (9)     
Combined cytoplasmic STAT3/nuclear pSTAT3 (Model 2)    0.024  0.526 
 Both low-mod 95 80 (4)     
 One high 75 79 (5) 1.50 (1.06–2.13)  –  
 Both high 26 54 (10)     
Combined nuclear STAT3/nuclear pSTAT3 (Model 3)    0.008  0.008 
 Both low-mod 100 81 (4)     
 One high 74 74 (5) 1.63 (1.14–2.34)  1.63 (1.14–2.34)  
 Both high 22 62 (11)     

To examine the relationship between expression and activation of STAT3 and survival, the cumulative prognostic value of cytoplasmic STAT3, nuclear STAT3, and nuclear pSTAT3 was examined with respect to 5-year cancer-specific survival (Table 4). Three models were examined: model 1 (cytoplasmic STAT3/nuclear STAT3) stratified survival from 81% (low expression of both) to 63% (high expression of both; P = 0.022), model 2 (cytoplasmic STAT3/nuclear pSTAT3) stratified survival from 81% to 54% (P = 0.018), and model 3 (nuclear STAT3/nuclear pSTAT3) stratified survival from 81% to 62% (P = 0.012). When the three models were entered into a multivariate model using a backwards conditional method, only model 3 (nuclear STAT3/nuclear pSTAT3) remained independently associated with cancer-specific survival (HR, 1.63; 95% CI, 1.14–2.34; P = 0.008; Fig. 1).

The relationship between this prognostic model and cancer-specific survival was examined on multivariate analysis. As the prognostic value of the KM grade has previously been shown to be similar to assessment of individual T-lymphocyte subsets (24), only KM grade was entered into the multivariable model. On multivariate survival analysis (Table 5), combined nuclear STAT3/pSTAT3 expression was not associated with cancer-specific survival (P = 0.220), whereas venous invasion (HR, 2.89; P = 0.001), tumor perforation (HR, 8.30; P < 0.01), NPS (HR, 1.69; P < 0.05), and tumor budding (HR, 4.12; P < 0.00) were all independently associated with survival. Low KM grade (HR, 2.14; P = 0.060) and elevated mGPS (HR, 1.52; P = 0.060) showed a trend toward poorer survival however failed to reach statistical significance.

Table 5.

Relationship between combined tumor cell nuclear STAT3/pSTAT3 expression, clinicopathologic characteristics, and cancer-specific survival of patients undergoing elective, potentially curative resection of stage I–III colorectal cancer

Cancer-specific survival
Clinicopathologic characteristicsUnivariate analysisPMultivariate analysisP
Age (<65/65–74/>75) 1.00 (0.73–1.37) 0.986 – – 
Sex (female/male) 1.43 (0.84–2.44) 0.188 – – 
Adjuvant therapy (no/yes) 1.43 (0.83–2.47) 0.196 – – 
Tumor site (colon/rectum) 0.99 (0.57–1.74) 0.983 – – 
TNM stage (I/II/III) 2.16 (1.35–3.48) 0.001 – 0.416 
Tumor differentiation (mod-well/poor) 1.18 (0.51–2.75) 0.700 – – 
Venous invasion (no/yes) 3.35 (1.97–5.70) <0.001 2.89 (1.59–5.28) 0.001 
Margin involvement (no/yes) 2.82 (1.12–7.09) 0.028 – 0.612 
Peritoneal involvement (no/yes) 2.45 (1.45–4.13) 0.001 – 0.650 
Tumor perforation (no/yes) 4.34 (1.04–18.11) 0.044 8.30 (1.84–37.43) 0.006 
Modified Glasgow Prognostic Score (0/1/2) 1.43 (0.99–2.08) 0.057 1.52 (0.98–2.35) 0.060 
NPS (0/1/2) 1.72 (1.13–2.62) 0.012 1.69 (1.07–2.67) 0.025 
NLR (<5/>5) 1.13 (0.60–2.13) 0.715 – – 
MMR status (competent/deficient) 1.37 (0.69–2.71) 0.370 – – 
KM grade (high/low) 2.33 (1.20–4.49) 0.012 2.14 (0.97–4.71) 0.060 
Tumor stroma percentage (low/high) 2.52 (1.48–4.30) 0.001 – 0.180 
Tumor budding (low/high) 3.92 (2.25–6.85) <0.001 4.12 (2.20–7.71) <0.001 
Nuclear STAT3/nuclear pSTAT3 (both low-mod/one high/both high) 1.63 (1.14–2.34) 0.008 1.28 (0.86–1.89) 0.220 
Cancer-specific survival
Clinicopathologic characteristicsUnivariate analysisPMultivariate analysisP
Age (<65/65–74/>75) 1.00 (0.73–1.37) 0.986 – – 
Sex (female/male) 1.43 (0.84–2.44) 0.188 – – 
Adjuvant therapy (no/yes) 1.43 (0.83–2.47) 0.196 – – 
Tumor site (colon/rectum) 0.99 (0.57–1.74) 0.983 – – 
TNM stage (I/II/III) 2.16 (1.35–3.48) 0.001 – 0.416 
Tumor differentiation (mod-well/poor) 1.18 (0.51–2.75) 0.700 – – 
Venous invasion (no/yes) 3.35 (1.97–5.70) <0.001 2.89 (1.59–5.28) 0.001 
Margin involvement (no/yes) 2.82 (1.12–7.09) 0.028 – 0.612 
Peritoneal involvement (no/yes) 2.45 (1.45–4.13) 0.001 – 0.650 
Tumor perforation (no/yes) 4.34 (1.04–18.11) 0.044 8.30 (1.84–37.43) 0.006 
Modified Glasgow Prognostic Score (0/1/2) 1.43 (0.99–2.08) 0.057 1.52 (0.98–2.35) 0.060 
NPS (0/1/2) 1.72 (1.13–2.62) 0.012 1.69 (1.07–2.67) 0.025 
NLR (<5/>5) 1.13 (0.60–2.13) 0.715 – – 
MMR status (competent/deficient) 1.37 (0.69–2.71) 0.370 – – 
KM grade (high/low) 2.33 (1.20–4.49) 0.012 2.14 (0.97–4.71) 0.060 
Tumor stroma percentage (low/high) 2.52 (1.48–4.30) 0.001 – 0.180 
Tumor budding (low/high) 3.92 (2.25–6.85) <0.001 4.12 (2.20–7.71) <0.001 
Nuclear STAT3/nuclear pSTAT3 (both low-mod/one high/both high) 1.63 (1.14–2.34) 0.008 1.28 (0.86–1.89) 0.220 

The prognostic value of combined nuclear STAT3/pSTAT3 expression as stratified by T stage and N stage was examined (Supplementary Fig. S1). Nuclear STAT3/pSTAT3 expression was associated with reduced survival of patients with T1-2 colorectal cancer (P < 0.001) but was not associated with survival of patients with T3-4 colorectal cancer (P = 0.192). Furthermore, nuclear STAT3/pSTAT3 expression stratified survival of patients with lymph node positive (P = 0.001) but not lymph node–negative disease (P = 0.516).

In the present study of patients undergoing elective, potentially curative colorectal cancer resection, STAT3 was not associated with clinicopathologic characteristics of the tumor but was associated with adverse host inflammatory responses. In particular, increased tumor cell STAT3 expression was associated with downregulation of the local inflammatory cell infiltrate.

Although in keeping with previous clinical studies of colorectal and pancreatic adenocarcinoma (32, 33), the present study is to our knowledge the first to examine the relationship between tumor STAT3 expression and the density of the local adaptive immune infiltrate as evidenced by T lymphocytes in the clinical context of patients with gastrointestinal cancer. Whereas previous studies found a decrease in the density of the generalized inflammatory cell infiltrate or tumor-infiltrating lymphocytes using H&E-based assessments (32, 33), the present study utilized immunohistochemistry and found a decrease in the density of tumor-associated T-lymphocyte populations. Indeed, this would suggest a direct effect of STAT3 activation on adaptive, T-lymphocyte–mediated anti-tumor immunity. Furthermore, the relationship between STAT3 expression and the local inflammatory cell infiltrate would appear to be independent of MMR status.

Although assessment of cytoplasmic STAT3 expression was significantly associated with the density of T lymphocytes, it was of interest that the K-M grade, an assessment of the generalized inflammatory cell infiltrate, did not differ with STAT3 expression. This may reflect the ability of STAT3 to simultaneously suppress antitumor immune responses whilst promoting protumor immunity (17, 34). Whereas antitumor, adaptive, Th1-polarized immune responses are down regulated (35, 36), STAT3-dependent transcription and release of Th2-type cytokines favors recruitment of tumor-promoting tumor-associated macrophages and myeloid-derived cells (17). Furthermore, STAT3 activation may additionally favor the differentiation of naïve T lymphocytes into tumor-promoting lymphocytic subsets (17). Consistent with such a hypothesis, Morikawa and colleagues found that although intratumoral lymphocyte density decreased, the density of the peritumoral inflammatory cell infiltrate increased with increasing STAT3 activity in a cohort of patients with stage I–IV colorectal cancer (32). Furthermore, it has been shown in some tumors, such as ependymomas, that STAT3 immunosuppression is mediated by upregulation of myeloid-derived cell activity, with a subsequent deleterious effect on T-lymphocytic, antitumor activity (37). As such, future studies of STAT3 activation in patients with gastrointestinal cancers should also consider the nature and density of local innate immune responses.

Of interest, pSTAT3 expression was associated with high-grade tumor budding. The presence of tumor buds is a phenotypic characteristic of epithelial–mesenchymal transitioning (EMT), a vital step in tumor cell dissemination which requires immune cell evasion (38). The present results may suggest that STAT3 activation is one mechanism by which tumor buds evade host antitumor immune responses. However, it is also recognized that STAT3 activation promotes tumor cell stemness and is an upstream activator of EMT (39, 40), which may explain the present associations.

Although failing to reach statistical significance, the density of tumor-associated stroma, as measured by TSP, appeared to be associated with pSTAT3 expression. Given that an increase in TSP primarily reflects an increased population of cancer-associated fibroblasts within the tumor microenvironment, this would further support the importance of IL6 secretion by fibroblasts in the activation of the JAK/STAT3 pathway in tumor cells (14, 15). Indeed, the present results suggest that the JAK/STAT3 pathway may be an important mechanism by which the tumor influences the composition of the tumor microenvironment and deregulates host antitumor immune responses.

The present study found that increased tumor STAT3 expression was associated with elevated systemic inflammatory responses as measured using the mGPS, a cumulative score based on serum CRP and albumin concentrations. Such routinely measured biomarkers of the systemic inflammatory response represent only “the tip of a far larger iceberg” of cancer-associated systemic inflammation, whereby circulating cytokines, growth factors, and myeloid-derived cells promote cancer progression and dissemination (9). One such cytokine, IL6, is commonly elevated in colorectal cancer (11, 13), and is the main determinant of hepatic synthesis of CRP and responsible for the acute phase reduction in hepatic albumin synthesis (6). Given the importance of IL6 as both an activator of the JAK/STAT3 pathway and as an end product of its activation, the present results are perhaps not surprising, and suggest that STAT3 activation may play a role in the systemic inflammatory response in colorectal cancer.

However, although STAT3 expression was associated with an elevated mGPS, it was not associated with components of the differential white cell count. This is in keeping with previous work from Guthrie and colleagues, whereby serum IL6 concentration correlated strongly with the mGPS but not the NLR in patients with colorectal cancer (13). However, other groups have found contradictory results, with a positive association between serum IL6 concentrations and the NLR in patients with colorectal cancer. This disparity may be explained by differences in the groups studied; whereas the patients in the present study and that of Guthrie and colleagues were undergoing potentially curative resection of stage I–III colorectal cancer, the groups studied by Kantola and Chen included patients with stage I–IV colorectal cancer at varying stages of treatment. Taken together, it would appear that, at least in patients with nonmetastatic colorectal cancer, the effects of the IL6/JAK/STAT3 pathway on the cancer-associated systemic inflammatory response may not be entirely modulated by an effect on circulating innate and adaptive immune cells.

Of interest, only total cytoplasmic STAT3 expression was associated with the systemic inflammatory response as measured by mGPS. The reason for this is not clear and, however, may represent the dynamic nature of JAK/STAT3 activation and translocation. Although activation of the IL6 receptor leads to rapid accumulation of STAT3, mechanistic studies have shown that less than 30% of total cytoplasmic STAT3 translocates to the nucleus on cytokine stimulation (41). Furthermore, STAT3 also exhibits transcription-independent activity within the cytoplasm without nuclear translocation (41, 42). Another plausible hypothesis is that rather than being directly causative, the presently observed associations between the mGPS and tumor cell STAT3 expression may represent separate downstream events of a common precursor, such as elevated systemic IL6 concentrations. Finally, given the lack of a consistent relationship across different measures of the systemic inflammatory response, the present results may simply represent a Type-I statistical error. Indeed, rather than the tumor itself, other end organs, such as liver or skeletal muscle, may be the predominant drivers of the systemic inflammatory response in such patients (43). As such, the present observations should be regarded as hypothesis-generating, and remain to be further investigated by mechanistic and clinical studies.

Consistent with previous reports in patients with gastrointestinal cancers (20), the present study found that increased tumor cell STAT3 expression and activity was associated with reduced survival. The pleiotropic nature of STAT3 activation is reflected in the fact that combined assessment of total nuclear STAT3 and pSTAT3 held greater prognostic value than either measure alone. Whereas the present study investigated IL6/JAK-mediated activation of STAT3 by phosphorylation of tyrosine residue 705, MAPK-dependent activation results in phosphorylation of the serine 727 residue, with differing results on transcriptional activity (28). Furthermore, STAT3 may also undergo nuclear import without phosphorylation (44). In addition to its role in mediating host immune responses, STAT3 activation plays an integral role in many key tumor cell pathways, including proliferation, EMT, and promotion of cancer cell stemness (39). Indeed, the heterogeneity of upstream activation of STAT3 is reflected in the present results, whereby 74 patients showed discordant expression of nuclear STAT3 and pSTAT3. As such, further investigation of these other upstream pathways and their relationship to the present results is required. Furthermore, rather than targeting upstream activation of STAT3, future therapeutic strategies may benefit from targeting STAT3 itself and its subsequent activation.

Nuclear STAT3/pSTAT3 expression showed differential prognostic value according to T stage and N stage. Although potentially reflecting the limited statistical power of the present study for subgroup analysis, the present results could suggest that STAT3 activation and expression may have a differential effect dependent on disease stage and invasiveness. In addition, assessment of the local and systemic environment held greater prognostic value than nuclear STAT3/pSTAT3 expression. Rather than being defined by one mechanism such as the JAK/STAT3 pathway, characteristics within the tumor microenvironment and the systemic inflammatory response are likely to be multifactorial in origin. Therefore, it might be anticipated that such phenotypic characteristics would be of greater prognostic value than a single signal transduction pathway. Indeed, it would be of considerable interest to examine and compare other signal transduction pathways associated with inflammation, such as the NF-κB pathway (45, 46), in future studies.

The present study provides further clinical evidence of the role of the IL6/JAK/STAT3 pathway in the amelioration of host antitumor immune responses, and raises two interesting points that remain to be investigated. Firstly, it would suggest a role for inhibitors of the IL6/JAK/STAT3 pathway in restoring antitumor immune responses in patients with colorectal cancer (47, 48). Secondly, it would support the hypothesis that routine markers of the systemic inflammatory response, and in particular the mGPS, may aid in the identification and selection of patients likely to benefit from such targeted therapies (49). In keeping with such a scheme, one recent clinical trial of a JAK inhibitor in patients with metastatic pancreatic cancer found an increase in overall survival only in those patients with an elevated CRP or mGPS (50). Therefore, it is clear that markers of the host inflammatory response should be incorporated into future studies of agents targeting the IL6/JAK/STAT3 pathway in cancer.

Given the increasing appreciation of distinct molecular subtypes of colorectal cancer (51), the results of the present study are perhaps limited by the lack of molecular characterization of the tumors studied. Although not associated with MMR status in the present cohort, the relationship between STAT3 and other characteristics, such as KRAS and BRAF status, would be of interest. However, a previous comprehensive study by Morikawa and colleagues found no association between STAT3, a number of molecular characteristics, and survival in a cohort of over 700 patients (32). Furthermore, it has also been suggested that STAT3 may have a role in not only induction of KRAS-mutated tumors (52), but also in conferring chemoresistance in patients with KRAS wild-type tumors (53). Indeed, this would suggest that STAT3 is independent of such characteristics. A further limitation is the relatively small sample size, precluding meaningful subgroup analysis. Analysis was restricted to a previously constructed TMA, and only patients who had complete staining for both STAT3 and pSTAT3 were included. However, post-hoc power calculation shows that the present study has adequate power to examine the relationship between STAT3 and the local and systemic environment. For example, post-hoc analysis suggests that the present study holds 84% power to determine a difference in cancer cell nest CD8+ T-lymphocyte density between those with low and high cytoplasmic STAT3 expression.

In conclusion, the results of the present study suggest a relationship between tumor cell STAT3 expression and the host inflammatory response, and may be one potential mechanism whereby the tumor promotes a local and systemic environment amenable to tumor growth and dissemination. Further studies are required to confirm such a relationship, and whether therapeutic targeting of the IL6/JAK/STAT3 may be utilized in the treatment of patients with colorectal cancer and elevated systemic inflammatory responses.

No potential conflicts of interest were disclosed.

Conception and design: J.H. Park, D.C. McMillan, C.S.D. Roxburgh, P.G. Horgan, J. Edwards

Development of methodology: J.H. Park, C.S.D. Roxburgh, P.G. Horgan, J. Edwards

Acquisition of data (provided animals, acquired and managed patients, provided facilities, etc.): J.H. Park, D.C. McMillan, J. Clark, C.S.D. Roxburgh

Analysis and interpretation of data (e.g., statistical analysis, biostatistics, computational analysis): J.H. Park, D.C. McMillan, C.S.D. Roxburgh, J. Edwards

Writing, review, and/or revision of the manuscript: J.H. Park, D.C. McMillan, J. Quinn, C.S.D. Roxburgh, P.G. Horgan, J. Edwards

Administrative, technical, or material support (i.e., reporting or organizing data, constructing databases): J. Quinn, P.G. Horgan

Study supervision: C.S.D. Roxburgh

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.
Oliphant
R
,
Nicholson
GA
,
Horgan
PG
,
Molloy
RG
,
McMillan
DC
,
Morrison
DS
, et al
Deprivation and colorectal cancer surgery: Longer-term survival inequalities are due to differential postoperative mortality between socioeconomic groups
.
Ann Surg Oncol
2013
;
20
:
2132
9
.
2.
Hanahan
D
,
Weinberg
RA
. 
Hallmarks of cancer: The next generation
.
Cell
2011
;
144
:
646
74
.
3.
Roxburgh
CSD
,
McMillan
DC
. 
The role of the in situ local inflammatory response in predicting recurrence and survival in patients with primary operable colorectal cancer
.
Cancer Treat Rev
2012
;
38
:
451
66
.
4.
Mei
Z
,
Liu
Y
,
Liu
C
,
Cui
A
,
Liang
Z
,
Wang
G
, et al
Tumour-infiltrating inflammation and prognosis in colorectal cancer: Systematic review and meta-analysis
.
Br J Cancer
2014
;
110
:
1595
605
.
5.
McMillan
DC
. 
The systemic inflammation-based Glasgow Prognostic Score: A decade of experience in patients with cancer
.
Cancer Treat Rev
2013
;
39
:
534
40
.
6.
Gabay
C
,
Kushner
I
. 
Acute-phase proteins and other systemic responses to inflammation
.
N Engl J Med
1999
;
340
:
448
54
.
7.
Guthrie
GJ
,
Charles
KA
,
Roxburgh
CS
,
Horgan
PG
,
McMillan
DC
,
Clarke
SJ
. 
The systemic inflammation-based neutrophil-lymphocyte ratio: Experience in patients with cancer
.
Crit Rev Oncol Hematol
2013
;
88
:
218
30
.
8.
Watt
DG
,
Martin
JC
,
Park
JH
,
Horgan
PG
,
McMillan
DC
. 
Neutrophil count is the most important prognostic component of the differential white cell count in patients undergoing elective surgery for colorectal cancer
.
Am J Surg
2015
;
210
:
24
30
.
9.
McAllister
SS
,
Weinberg
RA
. 
The tumour-induced systemic environment as a critical regulator of cancer progression and metastasis
.
Nat Cell Biol
2014
;
16
:
717
27
.
10.
Park
JH
,
Watt
DG
,
Roxburgh
CS
,
Horgan
PG
,
McMillan
DC
. 
Colorectal cancer, systemic inflammation, and outcome: Staging the tumor and staging the host
.
Ann Surg
2016
;
263
:
326
36
.
11.
Kantola
T
,
Klintrup
K
,
Vayrynen
JP
,
Vornanen
J
,
Bloigu
R
,
Karhu
T
, et al
Stage-dependent alterations of the serum cytokine pattern in colorectal carcinoma
.
Br J Cancer
2012
;
107
:
1729
36
.
12.
Guo
Y
,
Xu
F
,
Lu
T
,
Duan
Z
,
Zhang
Z
. 
Interleukin-6 signaling pathway in targeted therapy for cancer
.
Cancer Treat Rev
2012
;
38
:
904
10
.
13.
Guthrie
GJ
,
Roxburgh
CS
,
Richards
CH
,
Horgan
PG
,
McMillan
DC
. 
Circulating IL-6 concentrations link tumour necrosis and systemic and local inflammatory responses in patients undergoing resection for colorectal cancer
.
Br J Cancer
2013
;
109
:
131
7
.
14.
Nagasaki
T
,
Hara
M
,
Nakanishi
H
,
Takahashi
H
,
Sato
M
,
Takeyama
H
. 
Interleukin-6 released by colon cancer-associated fibroblasts is critical for tumour angiogenesis: Anti-interleukin-6 receptor antibody suppressed angiogenesis and inhibited tumour-stroma interaction
.
Br J Cancer
2014
;
110
:
469
78
.
15.
Huynh
PT
,
Beswick
EJ
,
Coronado
YA
,
Johnson
P
,
O'Connell
MR
,
Watts
T
, et al
CD90(+) stromal cells are the major source of IL-6, which supports cancer stem-like cells and inflammation in colorectal cancer
.
Int J Cancer
2016
;
138
:
1971
81
.
16.
Wang
SW
,
Sun
YM
. 
The IL-6/JAK/STAT3 pathway: Potential therapeutic strategies in treating colorectal cancer (Review)
.
Int J Oncol
2014
;
44
:
1032
40
.
17.
Yu
H
,
Pardoll
D
,
Jove
R
. 
STATs in cancer inflammation and immunity: A leading role for STAT3
.
Nat Rev Cancer
2009
;
9
:
798
809
.
18.
Yu
H
,
Kortylewski
M
,
Pardoll
D
. 
Crosstalk between cancer and immune cells: Role of STAT3 in the tumour microenvironment
.
Nat Rev Immunol
2007
;
7
:
41
51
.
19.
Taniguchi
K
,
Karin
M
. 
IL-6 and related cytokines as the critical lynchpins between inflammation and cancer
.
Semin Immunol
2014
;
26
:
54
74
.
20.
Li
MX
,
Bi
XY
,
Huang
Z
,
Zhao
JJ
,
Han
Y
,
Li
ZY
, et al
Prognostic role of phospho-STAT3 in patients with cancers of the digestive system: A systematic review and meta-analysis
.
PLoS One
2015
;
10
:
e0127356
.
21.
Loughrey
MB
,
Quirke
P
,
Shepherd
NA
. 
Dataset for colorectal cancer histopathology reports
. 3rd ed.
London, UK
:
The Royal College of Pathologists
; 
2014
.
22.
Watt
DG
,
Proctor
MJ
,
Park
JH
,
Horgan
PG
,
McMillan
DC
. 
The Neutrophil-Platelet Score (NPS) predicts survival in primary operable colorectal cancer and a variety of common cancers
.
PLoS One
2015
;
10
:
e0142159
.
23.
Richards
CH
,
Roxburgh
CS
,
Anderson
JH
,
McKee
RF
,
Foulis
AK
,
Horgan
PG
, et al
Prognostic value of tumour necrosis and host inflammatory responses in colorectal cancer
.
Br J Surg
2012
;
99
:
287
94
.
24.
Richards
CH
,
Roxburgh
CS
,
Powell
AG
,
Foulis
AK
,
Horgan
PG
,
McMillan
DC
. 
The clinical utility of the local inflammatory response in colorectal cancer
.
Eur J Cancer
2014
;
50
:
309
19
.
25.
Park
JH
,
Richards
CH
,
McMillan
DC
,
Horgan
PG
,
Roxburgh
CS
. 
The relationship between tumour stroma percentage, the tumour microenvironment and survival in patients with primary operable colorectal cancer
.
Ann Oncol
2014
;
25
:
644
51
.
26.
van Wyk
HC
,
Park
JH
,
Edwards
J
,
Horgan
PG
,
McMillan
DC
,
Going
JJ
. 
The relationship
between tumour budding, the tumour microenvironment and survival in patients with primary operable colorectal cancer
.
Br J Cancer
2016
;
115
:
156
63
.
27.
Roxburgh
CS
,
Richards
CH
,
Macdonald
AI
,
Powell
AG
,
McGlynn
LM
,
McMillan
DC
, et al
The in situ local immune response, tumour senescence and proliferation in colorectal cancer
.
Br J Cancer
2013
;
109
:
2207
16
.
28.
Lim
CP
,
Cao
X
. 
Serine phosphorylation and negative regulation of Stat3 by JNK
.
J Biol Chem
1999
;
274
:
31055
61
.
29.
Kirkegaard
T
,
Edwards
J
,
Tovey
S
,
McGlynn
LM
,
Krishna
SN
,
Mukherjee
R
, et al
Observer variation in immunohistochemical analysis of protein expression, time for a change?
Histopathology
2006
;
48
:
787
94
.
30.
Park
JH
,
Powell
AG
,
Roxburgh
CS
,
Horgan
PG
,
McMillan
DC
,
Edwards
J
. 
Mismatch repair status in patients with primary operable colorectal cancer: Associations with the local and systemic tumour environment
.
Br J Cancer
2016
;
114
:
562
70
.
31.
Arends
M
,
Ibrahim
M
,
Happerfield
L
,
Frayling
I
,
Miller
K
. 
Interpretation of immunohistochemical analysis of mismatch repair (MMR) protein expression in tissue sections for investigation of suspected lynch/hereditary non-polyposis colorectal cancer (HNPCC) syndrome
.
UK NEQAS ICC & ISH Recommendations
2008
;
1
.
32.
Morikawa
T
,
Baba
Y
,
Yamauchi
M
,
Kuchiba
A
,
Nosho
K
,
Shima
K
, et al
STAT3 expression, molecular features, inflammation patterns, and prognosis in a database of 724 colorectal cancers
.
Clin Cancer Res
2011
;
17
:
1452
62
.
33.
Denley
SM
,
Jamieson
NB
,
McCall
P
,
Oien
KA
,
Morton
JP
,
Carter
CR
, et al
Activation of the IL-6R/Jak/stat pathway is associated with a poor outcome in resected pancreatic ductal adenocarcinoma
.
J Gastrointest Surg
2013
;
17
:
887
98
.
34.
Nguyen
AV
,
Wu
YY
,
Liu
Q
,
Wang
D
,
Nguyen
S
,
Loh
R
, et al
STAT3 in epithelial cells regulates inflammation and tumor progression to malignant state in colon
.
Neoplasia
2013
;
15
:
998
1008
.
35.
O'Toole
A
,
Michielsen
AJ
,
Nolan
B
,
Tosetto
M
,
Sheahan
K
,
Mulcahy
HE
, et al
Tumour microenvironment of both early- and late-stage colorectal cancer is equally immunosuppressive
.
Br J Cancer
2014
;
111
:
927
32
.
36.
Yu
H
,
Liu
Y
,
McFarland
BC
,
Deshane
JS
,
Hurst
DR
,
Ponnazhagan
S
, et al
SOCS3 deficiency in myeloid cells promotes tumor development: Involvement of STAT3 activation and myeloid-derived suppressor cells
.
Cancer Immunol Res
2015
;
3
:
727
40
.
37.
Griesinger
AM
,
Josephson
RJ
,
Donson
AM
,
Mulcahy Levy
JM
,
Amani
V
,
Birks
DK
, et al
Interleukin-6/STAT3 pathway signaling drives an inflammatory phenotype in Group A ependymoma
.
Cancer Immunol Res
2015
;
3
:
1165
74
.
38.
Koelzer
VH
,
Dawson
H
,
Andersson
E
,
Karamitopoulou
E
,
Masucci
GV
,
Lugli
A
, et al
Active immunosurveillance in the tumor microenvironment of colorectal cancer is associated with low frequency tumor budding and improved outcome
.
Transl Res
2015
;
166
:
207
17
.
39.
Bak
Y
,
Kwon
T
,
Bak
IS
,
Hong
J
,
Yu
DY
,
Yoon
DY
. 
IL-32theta inhibits stemness and epithelial-mesenchymal transition of cancer stem cells via the STAT3 pathway in colon cancer
.
Oncotarget
2016
;
7
:
7307
17
.
40.
Rokavec
M
,
Oner
MG
,
Li
H
,
Jackstadt
R
,
Jiang
L
,
Lodygin
D
, et al
IL-6R/STAT3/miR-34a feedback loop promotes EMT-mediated colorectal cancer invasion and metastasis
.
J Clin Invest
2014
;
124
:
1853
67
.
41.
Sehgal
PB
. 
STAT-signalling through the cytoplasmic compartment: Consideration of a new paradigm
.
Cell Signal
2000
;
12
:
525
35
.
42.
Germain
D
,
Frank
DA
. 
Targeting the cytoplasmic and nuclear functions of signal transducers and activators of transcription 3 for cancer therapy
.
Clin Cancer Res
2007
;
13
:
5665
9
.
43.
Bonetto
A
,
Aydogdu
T
,
Kunzevitzky
N
,
Guttridge
DC
,
Khuri
S
,
Koniaris
LG
, et al
STAT3 activation in skeletal muscle links muscle wasting and the acute phase response in cancer cachexia
.
PLoS One
2011
;
6
:
e22538
.
44.
Liu
L
,
McBride
KM
,
Reich
NC
. 
STAT3 nuclear import is independent of tyrosine phosphorylation and mediated by importin-alpha3
.
Proc Natl Acad Sci U S A
2005
;
102
:
8150
5
.
45.
Wang
S
,
Liu
Z
,
Wang
L
,
Zhang
X
. 
NF-kappaB signaling pathway, inflammation and colorectal cancer
.
Cell Mol Immunol
2009
;
6
:
327
34
.
46.
De Simone
V
,
Franze
E
,
Ronchetti
G
,
Colantoni
A
,
Fantini
MC
,
Di Fusco
D
, et al
Th17-type cytokines, IL-6 and TNF-alpha synergistically activate STAT3 and NF-kB to promote colorectal cancer cell growth
.
Oncogene
2015
;
34
:
3493
503
.
47.
Sansone
P
,
Bromberg
J
. 
Targeting the interleukin-6/Jak/stat pathway in human malignancies
.
J Clin Oncol
2012
;
30
:
1005
14
.
48.
O'Shea
JJ
,
Schwartz
DM
,
Villarino
AV
,
Gadina
M
,
McInnes
IB
,
Laurence
A
. 
The JAK-STAT pathway: Impact on human disease and therapeutic intervention
.
Annu Rev Med
2015
;
66
:
311
28
.
49.
Diakos
CI
,
Charles
KA
,
McMillan
DC
,
Clarke
SJ
. 
Cancer-related inflammation and treatment effectiveness
.
Lancet Oncol
2014
;
15
:
e493
503
.
50.
Hurwitz
HI
,
Uppal
N
,
Wagner
SA
,
Bendell
JC
,
Beck
JT
,
Wade
SM
 3rd
, et al
Randomized, double-blind, phase II study of ruxolitinib or placebo in combination with capecitabine in patients with metastatic pancreatic cancer for whom therapy with gemcitabine has failed
.
J Clin Oncol
2015
;
33
:
4039
47
.
51.
Guinney
J
,
Dienstmann
R
,
Wang
X
,
de Reynies
A
,
Schlicker
A
,
Soneson
C
, et al
The consensus molecular subtypes of colorectal cancer
.
Nat Med
2015
;
21
:
1350
6
.
52.
Corcoran
RB
,
Contino
G
,
Deshpande
V
,
Tzatsos
A
,
Conrad
C
,
Benes
CH
, et al
STAT3 plays a critical role in KRAS-induced pancreatic tumorigenesis
.
Cancer Res
2011
;
71
:
5020
9
.
53.
Dobi
E
,
Monnien
F
,
Kim
S
,
Ivanaj
A
,
N'Guyen
T
,
Demarchi
M
, et al
Impact of STAT3 phosphorylation on the clinical effectiveness of anti-EGFR-based therapy in patients with metastatic colorectal cancer
.
Clin Colorectal Cancer
2013
;
12
:
28
36
.