Enhanced cell migration and apoptosis resistance may underlie the association between high SERPINE1 expression and poor outcome in head and neck carcinoma patients.

High SERPINE1 expression is a common event in head and neck squamous cell carcinoma (HNSCC); however, whether it plays a role in determining clinical outcome remains still unknown. We studied SERPINE1 as a prognostic marker in two HNSCC patient cohorts. In a retrospective study (n = 80), high expression of SERPINE1 was associated with poor progression-free (p = 0.022) and cancer-specific (p = 0.040) survival. In a prospective study (n = 190), high SERPINE1 expression was associated with poor local recurrence-free (p = 0.022), progression-free (p = 0.002) and cancer-specific (p = 0.006) survival. SERPINE1 expression was identified as an independent risk factor for progression-free survival in patients treated with chemo-radiotherapy or radiotherapy (p = 0.043). In both patient cohorts, high SERPINE1 expression increased the risk of metastasis spread (p = 0.045; p = 0.029). The association between SERPINE1 expression and survival was confirmed using the HNSCC cohort included in The Cancer Genome Atlas project (n = 507). Once again, patients showing high expression had a poorer survival (p < 0.001). SERPINE1 over-expression in HNSCC cells reduced cell proliferation and enhanced migration. It also protected cells from cisplatin-induced apoptosis, which was accompanied by PI3K/AKT pathway activation. Downregulation of SERPINE1 expression had the opposite effect. We propose SERPINE1 expression as a prognostic marker that could be used to stratify HNSCC patients according to their risk of recurrence.

Introduction SERPINE1 (PAI-1) was selected as the candidate marker for this study. We studied the prognostic value of SERPINE1 (PAI-1) expression in head and neck squamous cell carcinoma (HNSCC) in a retrospective and a prospective studies in independent patient cohorts. Hypothesis and objectives are stated in introduction section (Pages 3-4)

Materials and Methods Patients
Patients with pathologically confirmed HNSCC were eligible for the study. The inclusion in the study did not modify patient treatment. Patients received a standard treatment according to their stage. Patient characteristics and treatment are described in table 1 and table 2. Inclusion and exclusion criteria are described in the methods section ("Patient characteristics, tissue samples and clinical follow-up") (Pages 16-17)

Specimen characteristics
The retrospective study was performed using paraffin-embedded pre-treatment tumor biopsies, whereas in the prospective study, we used pre-treatment fresh tumor biopsies. The collection, freezing and preservation of samples is detailed in the methods section (

"Patient characteristics, tissue samples and clinical follow-up") (Pages 16-17)
Assay methods SERPINE1 (PAI-1) expression was evaluated by Immunohistochemistry and real time RT-PCR. Details of sample processing and expression analysis are described in methods. Analysis were performed blinded to the study end point

Study design
Patients were recruited retrospectively and prospectively. A retrospective study (n = 80) was performed using pretreatment tumor biopsies from patients treated between 1995 and 2003 at Hospital de la Santa Creu i Sant Pau (HSCSP), Barcelona. A prospective study (n = 190) was run using fresh tumor biopsies obtained from patients treated at HSCSP and at Hospital Moises Broggi from 2002 to 2012. The median follow-up time was 68 months in the retrospective study and 37 months in the prospective study. The primary end points analyzed were local recurrence-free survival (LRFS), progression-free survival (PFS) and cancer-specific survival (CSS). Their description and secondary end points are defined in methods ("Patient characteristics, tissue samples and clinical follow-up") (Pages 16-17)

Statistical analysis methods
Classification and regression tree analysis was used to define patient subgroups based on SERPINE1(PAI-1) mRNA expression and their risk of disease relapse. The association between SERPINE1 (PAI-1) expression and clinicopathological variables was assessed using Mann Whitney and Kruskal Wallis tests. Kaplan-Meier analysis and Log-Rank test were used to analyze the association between the tumor subclassification and time to event outcome. Univariate and multivariate Cox models were used to test the association between clinical variables and SERPINE1 (PAI-1) expression with PFS and CSS ("Statistical analysis") (Pages 23-24).

Results. Data
The number of patients in each subgroup is detailed in the Kaplan Meier survival curves or tables. Basic demographic characteristics of the studied patients and standard prognostic variables are described in table 1 and 2.

Analysis and presentation
The association between SERPINE1 (PAI-1) expression and standard prognosis variables are detailed in Table 1 and 2. SERPINE1 (PAI-1) expression was associated with the appearance of local or metastatic recurrences after treatment. Kaplan-Meir and log rank test showed an association between SERPINE1 (PAI-1) expression and PFS and CSS (Figure 1-3). Multivariate analysis showed that PAI-1 expression, node stage and tumor size were associated with PFS (Table 3-4). Hazard ratios, confidence intervals and significance are detailed in tables, figures and described in results section (Pages 5-11).

Discussion
We showed that a high SERPINE1 (PAI-1) expression was associated with poor clinical outcome in patients with HNSCC. Larger multicenter studies and clinical trials are warranted to further validate the role of SERPINE1(PAI-1) in treatment decisions. The clinical implications and value of our findings are stated in the discussion section.