Stage and tissue-specific prognostic impact of miR-182 in NSCLC

Between 1990 and 2004, 371 patients with pathological stage I to IIIA non-small cell lung cancer were diagnosed at the University Hospital of North Norway and Nordland Central Hospital and treated with curative intent. Resected tissues from the primary tumors in these patients were used in our retrospective study. Out of 371 patients, 36 were excluded from the study due to radiotherapy or chemotherapy prior to surgery (n = 10), other malignancy within 5 years before NSCLC diagnosis (n = 13) or inadequate paraffin-embedded fixed tissue blocks (n = 13). Adjuvant chemotherapy was not introduced in Norway during this period (1990 – 2004). Thus, 335 patients with complete demographic and clinicopathological data were eligible for this study. Of these, postoperative radiotherapy was offered to 55 patients with non-radical surgical margins or mediastinal lymph node disease (N2).

This report includes follow-up data as of January 10, 2011. The median follow-up time of survivors was 105 months (range 73–234). Formalin-fixed, paraffin-embedded tumor specimens were obtained from the archives of the Departments of Clinical Pathology at the University Hospital of North Norway and Nordland Central Hospital. The pathological data were revised according to the 7^th edition of UICC TNM classification of lung cancer [18]. If the morphological characteristics for adeno- and squamous cell carcinomas were easily recognizable, it was not always necessary to do further examinations (IHC) of the tumor samples. If the tumors were not well differentiated, IHC was necessary. CK7, TTF1, p63 and CK5/6 was the markers most frequently used. The National Data Inspection Board and the Regional Ethics Committee North (REC North) approved this study.

We used a 0.6 mm-diameter stylet to sample two cores with neoplastic tissue and two cores with tumor stroma from different areas of the primary tumors from each patient. Normal lung tissue localized distant from the tumor and lung tissue sample from 20 patients without cancer diagnosis were used as controls. The TMAs were assembled using a tissue-arraying instrument (Beecher Instruments, Silver Springs, Md). Eight tissue microarray blocks were made to include all the tissue samples. Multiple 4-μm-sections were cut with a Micron microtome (HM355S) and stained by specific antibodies for immunohistochemical analyses or stained by in situ hybridization. The detailed methodology has been previously reported [19].

In situ hybridization was performed following the protocol developed by Exiqon, Vedbaek, Denmark [20]. Digoxigenin (DIG) labelled locked nucleic acid (LNA) modified probes from Exiqon for miR-182 (hsa-miR-182), positive control (U6, hsa/mmu/rno) and negative control (scramble-miR) were used in this study. Some adjustments were done to get a specific and sensitive detection of miRNA in our sections from formalin-fixed paraffin-embedded (FFPE) TMA blocks.

We placed 4 μm sections of the TMA blocks in a heater at 59˚C over night to attach cores to Super Frost Plus slides. Sections were deparaffinised with xylene (3 × 5 min) and then rehydrated with ethanol solutions (99.9% - 96% - 70%) ending up in PBS, pH 7.4. Proteinase-K (20 μg/ml) (Exiqon, Vedbaek, Denmark) treatment was done in PK-buffer (5 mM Tris–HCl, pH 7.5, 1 mM EDTA, 1 mM NaCl, autoclaved) at 37˚C for 20 min in a HYBrite automated hybridizer (Abbot laboratories, IL, US). After a PBS wash the sections were dehydrated through increasing gradient of ethanol solutions and air-dried. The LNA-probes were denatured by heating to 90˚C for 4 min. Hybridization of the LNA-probe miR-182 (100nM) and scramble miR (50nM) control was carried out in the HYBrite automated hybridizer at 50˚C for 60 min. The positive control U6 (1nM) was hybridized at 55˚C for 60 min. Stringent washes was performed in pre-heated SSC buffers, 1 × 5 min in 5x SSC, 2 × 5 min in 1× SSC and 0,2× SSC. Sections were blocked against unspecific binding in blocking solution from DIG wash and Block Buffer set (Roche, Mannheim, Germany) for 15 min at room temperature (RT). Alkaline phosphatase (AP)-conjugated anti-DIG (Roche) 1:800 was incubated for 60 min at RT for immunologic detection. After PBS-T wash the substrate enzymatic reaction was carried out with NBT/BCIP (Roche) at 30˚C in the hybridizer for 120 min. The reaction was stopped with a 2 × 5 min wash in KTBT buffer (50 mM Tris–HCl, 150 mM NaCl, 10 mM KCl). Sections were counter stained with nuclear fast red (WALDECK, ZE-012-250) at RT for 1 min and then rinsed in tap water. Dehydration followed through increasing gradient of ethanol solutions and finally mounting with Histokitt mounting medium (Assistant-Histokitt, 1025/250).

We used data from previous publications with the following antibodies for correlation analyses: VEGF (−A, –C, -D, R-1, R-2, R-3), PDGF (−A, -B, –C, -D, R-α, R-β), FGF (−2, R-1), Notch (−1, -2), Jagged1, DLL4, Hif (−1α, -2α), GLUT-1, LDH5, CAIX, PHD (−1, -2, -3), FIH, Ang (−1, -2, -4), Tie-2 and MMP (−2, -7, -9). Detailed IHC procedures for the antibodies which correlated significantly with miR-182 (FGF2, Hif2α and MMP-7) have been previously published [21‐23].

Representative viable tissue sections were scored semi-quantitatively by light microscopy. The dominant staining intensity in tumor cells was scored as 0 = negative, 1 = weak, 2 = intermediate or 3 = strong (Figure 1). The TMA cores were scored anonymously and independently by one experienced pathologist and one oncologist. In case of disagreement, the slides were reexamined and consensus was reached by the observers.

Mean score for duplicate cores from each individual was calculated in tumor epithelial cells. We then categorized the staining into high and low expression. High expression in tumor cells was defined as score >0.

All statistical analyses were performed using the statistical package SPSS (Chicago, IL), version 19.0. The chi-square test and the Fisher exact test were used to examine the association between molecular marker expression and the clinicopathological markers. Correlations between markers were assessed using Spearman’s rank correlation. Univariate analyses were done using the Kaplan-Meier method, and statistical significance between survival curves was assessed by the log-rank test. Disease-specific survival (DSS) was defined as time from surgery to lung cancer death. Variables of significant value from the univariate analyses were entered into multivariate analysis using the backward stepwise Cox regression analysis. A P < 0.05 was considered statistically significant.

The National Data Inspection Board and the Regional Ethics Committee North (REC North) approved this study. Information and subsequent written consent from patients was considered, but as this was a retrospective study with more than half of patients deceased, the rest of the patients having to be reminded about the death rate of the disease and the possible raising of unrealistic hope for the individual, The Norwegian Data Inspection Board and REC North specifically waived the need for consent. All the patient data were anonymized after collecting the clinicopathological variables for each patient and before doing the statistical analyses.

Demographic, clinical and histopathological variables are listed in Table 1. The median patient age was 67 (range 28–85) and the majority were male (76%). Most patients (95%) were current or previous smokers. The NSCLC tumors comprised 191 squamous cell carcinomas (SCC), 113 adenocarcinomas (AC) including 18 bronchioloalveolar carcinomas (BAC) and 31 large-cell carcinomas (LCC).

miR-182 was homogenously expressed mainly in the cytoplasm of tumor cells. There was also some unspecific nuclear staining (Figure 1). The scoring was based on cytoplasmic staining. There was no staining of stromal cells, except for weak nuclear staining of some fibroblasts.

We tested correlations between miR-182 and angiogenic and hypoxia molecular markers. We found significant correlations between miR-182 and FGF2 (r = −0.147; P = 0.010), HIF2α (r = 0.115; P = 0.047) and MMP-7 (r = 0.172; P = 0.003).

As shown in Table 1, the clinicopathological variables performance status (P = 0.016), histology (P = 0.028), tumor differentiation (P < 0.001), surgical procedure (P = 0.007), pathological stage (P < 0.001), tumor status (P < 0.001), nodal status (P < 0.001) and vascular infiltration (P = 0.001) were significant prognostic indicators for DSS.

The results from the univariate analyses on miR-182 are presented in Table 2 and Figures 2 and 3. In the whole cohort, there was a tendency towards a better prognosis for those with tumors overexpressing miR-182 (P = 0.062, Figure 2). In subgroup analyses, patients with stage II disease had a significantly improved prognosis if they overexpressed miR-182 (P = 0.003, Figure 3E). In the histological subgroup SCC, high tumor cell miR-182 expression was associated with superior prognosis when compared to low expression (P = 0.042, Figure 3A), while for large cell carcinomas the trend was opposite (Figure 3C).

In the total cohort, performance status (P = 0.008), histology (P = 0.001), tumor differentiation (P = 0.007), tumor status (P = 0.007), nodal status (P = 0.022) and vascular infiltration (P = 0.004) all were independent prognostic factors.

Results of the multivariate analysis for miR-182 expression are presented in Table 2. Examining the total material, high miR-182 expression tended towards an independent association with a better prognosis (HR 0.73, CI 95% 0.50-1.06, P = 0.098). Among stage II patients, however, high tumor cell expression of miR-182 was an independent positive prognostic factor (HR 0.50, CI 95% 0.28-0.90, P = 0.020). Also in SCC, patients with a high miR-182 expression had an independent favorable outcome (HR 0.57, CI 95% 0.33-0.99, P = 0.048).

Among markers examined for correlations with miR-182, FGF2 and MMP-7 showed the strongest correlations. We assessed the co-expression combinations between miR-182 and FGF2 and MMP-7, respectively. The co-expression of low miR-182/high FGF2 was associated with poor survival (P = 0.017) as shown in Figure 4A. The combination showed an independently significant adverse prognosis compared to high miR-182/low FGF2 (HR 1.92, P = 0.015, Table 3). Patients expressing high miR-182/high MMP-7 had a better survival than other combinations (P = 0.036, Figure 4B). In the multivariate analyses, high miR-182/high MMP-7 showed an independently better prognosis than low miR-182/low MMP-7 (HR 0.49, P = 0.015, Table 3). In the SCC subgroup, we found an even bigger difference between these groups both in univariate and multivariate analyses (Figure 4C, Table 3).