Impact of Nuclear Interleukin-1 Alpha and EGFR Expression on Recurrence and Survival Outcomes in Oral Squamous Cell Carcinomas

Interleukin-1 alpha (IL-1α) is a pleiotropic cytokine involved in inflammation and immune response and is upregulated in many solid tumors including head and neck squamous cell carcinomas. Although IL-1α expression is generally associated with poor prognosis, the implications of the subcellular localization of IL-1α expression in patient outcomes are poorly understood. This study is aimed at investigating the prognostic value of nuclear and cytoplasmic immunohistochemical IL-1α expression in oral squamous cell carcinomas (OSCCs). Tissue microarrays containing 146 OSCCs were analyzed for IL-1α and epidermal growth factor receptor (EGFR) expression by immunohistochemistry. IL-1α and EGFR expression scores were correlated with clinicopathological parameters and survival outcomes. IL-1α expression was observed in the nuclear and/or cytoplasmic compartments in 98% of evaluable tumors and 78% of tumors expressed IL-1α in both compartments. There were no differences observed in overall survival or progression-free survival between high, moderate, or negative IL-1α nuclear/cytoplasmic expression scores. When IL-1α nuclear/cytoplasmic expression scores were stratified by positive or negative EGFR expression, tumors with a combined EGFR-positive and high nuclear IL-1α expression profile were significantly more likely to possess perineural invasion and were significantly associated with a high risk of tumor recurrence and worse progression-free survival compared to all other EGFR and combined IL-1α/EGFR expression profiles. Altogether, nuclear IL-1α expression may enhance the prognostic value of EGFR in OSCC and warrants further study as a prognostic biomarker for recurrence.


Introduction
The IL-1 pathway plays a plays a critical role in the regulation of immune and inflammatory responses to infections and sterile insults [1]. IL-1 signaling is also frequently upregulated in many solid tumor types including head and neck squamous cell carcinomas (HNSCCs) [2]. Although increased IL-1 signaling is typically associated with poor prognosis in cancer patients [3][4][5], the role of IL-1 signaling in cancer is controversial since IL-1 plays roles in both tumor promotion via the expression of genes involved in tumor survival, angiogenesis, metastasis, and immune cell recruitment [6][7][8] and tumor suppression via increased natural killer (NK) cell activity and enhanced Th1-mediated immunity [9, 10].

Tissue Microarrays (TMAs).
Formalin-fixed paraffinembedded tumor samples of patients were obtained from the archives of the Department of Pathology at the University of Iowa Hospitals and Clinics. TMAs were constructed using 3-6 morphologically representative tumor regions (1 mm) chosen from 146 carcinomas from the oral cavity. The 146 cases were surgical resection specimens and chosen selectively to ensure a mixture of patients with and without recurrence, with and without node metastases or positive margins, young and old ages, and smokers and nonsmokers from a population of 421 patients spanning 10 years of time (2005)(2006)(2007)(2008)(2009)(2010)(2011)(2012)(2013)(2014). After generating a list of 266 patients proposed for inclusion in the TMA, only 146 were included in the TMA block construction due to paraffin-embedded formalin-fixed tissue block availability and quality and resource constraints. Chi-square analysis of the T stage distribution of the patients in the larger database (T1/T2/T3&4: 31/29/39%) showed no statistically significant difference from the patient cohort included in the study (29/29/40%). None of the patients included in the TMA underwent prior radiation or chemotherapy. Four m sections were obtained from the TMAs on poly-L-lysinecoated glass slides. Routine hematoxylin and eosin (H&E) sections were reviewed to confirm the original diagnosis.

Immunohistochemistry (IHC).
Antigen retrieval was performed on freshly cut sections in a decloaking chamber for 5 min at 125 ∘ C in TRIS buffer (pH 9.0). Endogenous peroxidase was blocked by incubation with 3% peroxide at room temperature for 2 min. For IL-1 staining, human specific IL-1 antibody (ab9614, Abcam) was applied at 1:250 in Dako diluent for 2 h at room temperature. The peptide sequence used to raise this antibody is SAPFSFLSNVKYNFMRIIKYEFIL-NDALNQSIIRANDQYLTAAALHNLDEAVKFDMGAYK-SSKDDAKITVILRISKTQLYVTAQDEDQPVLLKEMPEIP-KTITGSETNLLFFWETHGTKNYFTSVAHPNLFIATK-QDYWVCLAGGPPSITDFQILENQA (amino acids 113-271) and therefore recognizes both full length/pro-IL-1 (amino acids 1-271) and mature IL-1 (amino acids 113-271). EGFR immunostaining was performed with antibody (H11, Dako) at 1:200 dilution. Bound antibody was detected using Envison6 + HRP, rabbit (Dako North America) for 30 min at room temperature followed by incubation with diaminobenzidine substrate (DAB) for 5 min at room temperature. HPV status was determined by p16 expression [19,20]. After completion of IHC, slides were stored at room temperature and a virtual scanned copy of the TMA slides was kept for further reference.

Quantification of IL-1 and EGFR
Staining. IL-1 immunoreactivity was evaluated by KGC using an Olympus BX53 microscope with an Olympus DP72 camera. Human kidney and spleen were used as positive and negative controls for IL-1 expression, respectively. Given that IL-1 expression varied in both the nuclei and cytoplasm, IL-1 expression was scored separately for the nuclei (N) and cytoplasm (C) on a scale from 0 to 2, with 0 representing no staining, 1 low/weak staining, and 2 strong/intense staining. The percentage of tumor cells with positive staining was determined by scoring 10 microscopic fields of 100 tumor cells each. Quantitative evaluation of EGFR staining was performed by AR according to Gamboa-Domingez et al. and Modern Pathology, 2004 [21], using a semiquantitative score (0-3+) where 0 represents no staining or membranous positivity in <10% neoplastic cells (negative) and 1, 2, and 3 represent weak, moderate, and strong membranous immunopositivity in >10% neoplastic cells.

Statistical Analysis.
The associations between IL-1 and/or EGFR expression with clinicopathological features such as sex, age, smoking history, tumor site, pathological TNM classification (UICC 7th), differentiation (well/moderately/poor), perineural, lymphovascular, and bone invasion were tested using the generalized linear modeling (GLM) framework and the Kruskal-Wallis test. Differences in survival outcomes (overall survival [OS] and progression-free survival [PFS]) were plotted using the Kaplan-Meier method while estimates for the group hazard ratios were obtained using Cox proportional hazards (PH) modeling. OS is defined as the length of time from the date of diagnosis that the patients remain alive. PFS is defined as the time from diagnosis to disease progression or death from any cause. All testing was performed on the univariate level and unadjusted for multiple comparisons. Differences between survival curves were compared using the log-rank test. A p-value below 0.05 was considered statistically significant. All analyses were completed using SAS 9.4.

Patient Characteristics.
A total of 146 patient samples were included in the OSCC TMA, of which 141 showed interpretable IL-1 and EGFR immunostaining. Reasons for lost samples included loss/absence of tissue in the TMA section during IHC processing. The baseline characteristics for these patients are summarized in Table 1. Of the 141 patients, 82 patients were males with an average age of diagnosis of 58 years, and 59 patients were female with an average age of diagnosis of 66 years (Table 1). Smoking histories were reported in most patients with 56 (40%) active smokers, 13 (9%) that had quit smoking for less than 10 years, 17 (12%) that quit smoking for more than 10 years, and 6 (4.3%) tobacco chewers (Table 1). Males made up the majority (78%) of the active smokers and females made up the majority of the patients with no smoking history (68%, (data not shown)). Majority of tumor sites represented in the TMA were from the oral tongue, N0 (52%), and moderately differentiated (64%) ( Table 1). Approximately 78 (55%) patients received adjuvant radiotherapy and 22 (16%) patients received chemotherapy following surgery (Table 1). Only 5 of the evaluable tumors were HPV-positive as detected by p16 expression (data not shown).  (Table 2).

Correlation of IL-1 Expression with Clinicopathologic
Parameters. There were no differences in nuclear or cytoplasmic IL-1 expression scores based on sex or age (Table 1). However significant differences in both nuclear and cytoplasmic IL-1 expression cores were observed based on smoking history, tumor site, T stage, N stage (cytoplasmic only), differentiation, perineural and lymphovascular invasion (cytoplasmic only), bone invasion, and the number of patients that received radiotherapy or chemotherapy (Table 1). There were not enough HPV-positive tumors to assess differences in IL-1 expression. Since the IL-1 expression profiles 1/1, 2/1, and 1/0 represented the majority of tumors in the TMA (Table 2), we further analyzed the association of these expression profiles with clinicopathological features of the represented OSCC patients. We found no differences in overall survival (p=0.27) or progression-free survival (p=0.29) with respect to these 3 major IL-1 expression profiles (Figures 2(a) and 2(b)).

EGFR Expression.
We next examined the role of EGFR expression in survival outcomes in OSCC patients. Examples of EGFR expression scores are shown in Figure 3. EGFR expression was observed in 61% of the tumors with strong (score of 3) expression in 34%, moderate (score of 2) expression in 16%, low (score of 1) expression in 11%, and no (score of 0) expression in of 39% of tumors (Table 3). There were no differences observed in overall survival (p=0.69) according to EGFR scores (Figure 4(a)); however significant differences were observed in progression-free survival (p=0.04) with higher EGFR expression being associated with worse progression-free survival (Figure 4(b)). A comparison of combined strong (3)+moderate (2) EGFR (designated as EGFR+) expression and combined low (1)+no (0) EGFR (designated as EGFR-) expression also showed a significant difference in progression-free survival (p=0.02) (Figure 4(d)) but not overall survival (p=0.22) (Figure 4(c)). These results support prior reports that EGFR expression is a strong predictor of progression-free survival in HNSCC patients [16,22].

Correlation of Combined EGFR and IL-1 with Patient
Outcomes. We next evaluated if differences in nuclear/ cytoplasmic IL-1 expression scores altered the predictive value of EGFR. There were no significant differences (p=0.31) in EGFR expression in tumors with 1/0, 1/1, and  2/1 IL-1 expression profiles ( Figure 5(a)). Tumors with a high nuclear/moderate cytoplasmic IL-1 expression profile (2/1) combined with a moderate/strong (2/3) EGFR expression score, which was designated as 2/1/EGFR+, were significantly (p=0.0058) more likely to experience tumor recurrence compared to the other IL-1 /EGFR expression profiles ( Figure 5(b)). We also found a significant interaction (p=0.02) between the 2/1 IL-1 profile score and EGFR+ expression ( Figure 5(b)) suggesting that this particular 2/1/EGFR+ expression profile could be considered a predictor of tumor recurrence. The high rates of recurrences observed in 2/1/EGFR+ tumors encompassed local, regional, and distant sites alone and in combination ( Figure 5(c)). Additionally, 2/1/EGFR+ tumors displayed significantly higher rates of perineural invasion (p<0.0001) compared to the other IL-1 /EGFR expression profiles ( Figure 6(a)). There were no differences among the IL-1 /EGFR expression profiles with respect to lymphovascular invasion (Figure 6(b)) and bone invasion ( Figure 6(c)). Lastly, we found that patients with 2/1/EGFR+ tumors trended toward worse overall survival compared to 2/1/EGFR-tumors although this association did not reach statistically significance (p=0.06) (Figure 7(a)). However, 2/1/EGFR+ tumors were significantly associated with worse progression-free survival compared to 2/1/EGFR-tumors (p<0.0001) (Figure 7(b)). EGFR expression did not affect survival outcomes in 1/1 (Figures 7(c) and 7(d)) or 1/0 tumors (Figures 7(e) and 7(f)). Compared to patients with an overall EGFR-expression profile (median survival = 61.4 months, Figure 4 survival outcome (median survival=not reached) out of all of the other profiles (Figure 7(b)). Furthermore, we confirmed in Table 4 that adjuvant chemotherapy and/or radiation treatment did not influence the differences in progressionfree survival observed with 2/1/EGFR+ versus 2/1/EGFRtumors (Figure 7(b)). Altogether, these data suggest that high nuclear IL-1 expression in combination with moderate/high EGFR expression may be associated with worse outcomes in OSCC patients.

Discussion
The results from this study imply that nuclear IL-1 expression may enhance the prognostic value of EGFR with respect to progression-free survival. EGFR is already well known as a prognostic indicator in HNSCCs; therefore the question remains of what role does nuclear IL-1 play in EGFR signaling? IL-1 has been reported to be associated with poor prognosis in a wide range of cancers [    Combined IL-1/EGFR Scores Combined IL-1/EGFR Scores     [5][6][7]. On the other hand IL-1 is involved in antitumor immunity via increased natural killer (NK) cell activity, dendritic cell activity, and enhanced Th1mediated immunity [9,[27][28][29]. These 2 opposing properties create a controversy surrounding the potential long-term clinical effectiveness of IL-1 inhibitors for cancer therapy. The additional presence of nuclear IL-1 further complicates this field since the activity of nuclear IL-1 is independent of the IL-1R1. For these reasons, little/no attention has been placed on the in situ analysis of IL-1 protein expression and the prognostic implications of IL-1 subcellular location.
In this study we have found that IL-1 is expressed in the vast majority of OSCC tumors and that IL-1 is expressed in both the nuclear and cytoplasmic components (Figure 1). We observed predominant nuclear expression of IL-1 since most of the IL-1 expressing tumors had either nuclear only or nuclear+cytoplasmic expression (Table 2). Only 4 tumors had cytoplasmic only staining highlighting the importance of nuclear IL-1 (i.e., pro-IL-1 and/or IL-1 propiece (ppIL-1 )) in tumor cells. In support of this observation, prior studies have also shown that pro-IL-1 and ppIL-1 appear to be predominantly intranuclear in IL-1 -expressing or IL-1 -transfected cells [30][31][32][33]. Based on the particular IL-1 antibody (ab9614, Abcam) used in these studies, the nuclear IL-1 detected is likely pro-IL-1 and not ppIL-1 . To date it is not clear what role nuclear IL-1 plays since some studies have reported that nuclear IL-1 inhibits cell proliferation [30,31] and triggers apoptosis [32]; and in other studies nuclear IL-1 promotes cell proliferation [34]. It is possible that differences in experimental techniques and cell models may explain these contradicting reports. Nevertheless, we find no prognostic value in IL-1 subcellular localization (Figures 2(a) and 2(b)) in this cohort of OSCC patients. However, when we take into consideration EGFR expression we found unexpected but interesting results. When separating patients from the 2/1, 1/1, and 1/0 IL-1 -expressing groups into EGFR+ and EGFR-subgroups, we uncover a subset of tumors with a 2/1/EGFR+ profile that demonstrates significantly increased rates of perineural invasion compared to all other IL-1 /EGFR expression profiles (Figure 6(a)). The presence of perineural invasion is a strong and independent predictor of local and regional failure in OSCC patients [35][36][37][38]. Therefore it is no surprise that patients bearing tumors with this 2/1/EGFR+ expression profile were significantly more likely to recur compared to all of the other IL-1 /EGFR expression profiles ( Figure 5(b)). Patients with 2/1/EGFR+ tumors also demonstrated worse progression-free survival compared to 2/1/EGFR-tumors (Figures 7(a) and 7(b)). This difference with EGFR expression was not observed in 1/1   (Figures 7(e) and 7(f)) tumors suggesting a possible interaction between high nuclear IL-1 activity and EGFR signaling which is supported by the significant interaction (p=0.02) found in Figure 5(b) between the 2/1 IL-1 profile score and EGFR+ expression. Of note, we found that 61% of OSCC patients with EGFR+ (score of 2 and 3) tumors experienced tumor recurrence with a median progression-free survival of 13 months (Figure 4(d)). However, if only patients with 2/1/EGFR+ tumors are taken into account, 94% of these patients experienced tumor recurrence with a median survival of 7 months (Figure 7(b)). This suggests that 2/1/EGFR-positivity may be a stronger and more accurate indicator of recurrence and progression-free survival than EGFR expression alone. At this time, we are unclear as to the role of nuclear IL-1 activity in EGFR signaling. Prior reports have shown that nuclear IL-1 is involved in transcriptional control by interacting with the histone acetyltransferases p300, PCAF, and Gcn5 [39][40][41]. Transcription factors important in cancer and proinflammation pathways such as NFkB, Elk-1, C/EBP , or AP-1 are activated by both IL-1 ligands and EGF [42]. EGFR activation has been shown to increase IL-1 ligand expression via increased NFkB activity in breast cancer cells resulting in increased growth and invasion [43]. Also, IL-1 ligands have been reported to transactivate EGFR through a CXCL1-CXCR2 axis [17] and ADAM17 [18] suggesting crosstalk between the EGFR and IL-1 pathways. These reports all support the synergistic interaction between the EGFR and IL-1 pathways but does not fully explain the preferential interaction of EGFR expression with nuclear IL-1 . Recently nuclear IL-1 has been shown to increase cell proliferation in T-lymphocytic leukemia cells by binding to the promoter region of sp1 leading to increased sp1 expression and activity [15]. Sp1 is a transcription factor involved in cell growth, immune responses, and chromatin remodeling [44]. Sp1 is also involved in the regulation of numerous genes involved in invasion and metastasis [45][46][47]. EGFR promoter activation requires sp1 and multiple binding sites for sp1 have been discovered [48][49][50]. This suggests that high nuclear IL-1 expression may promote EGFR signaling via sp1 activity which would explain the high recurrence rates observed in patients with 2/1/EGFR+ tumors ( Figure 5(b)). Moreover, the IL-1 gene has been reported to be regulated by sp1 [51,52] suggesting a feed-forward relationship between IL-1 and sp1 which would ultimately promote EGFR signaling and tumor progression. Further mechanistic studies in this area are necessary to investigate this theory.

Conclusions
Altogether, we believe we have identified a combined high nuclear IL-1 /EGFR+ tumor expression profile as a strong prognostic biomarker for progression-free survival in OSCC patients which warrants further study in other HNSCCs and other EGFR-expressing tumors.

Data Availability
The datasets generated during and/or analyzed during the current study are not publically available due to privacy and other restrictions but are available from the relevant authors on reasonable request.

Ethical Approval
All procedures performed in studies involving human participants were in accordance with the ethical standards of the institutional and/or national research committee and with the 1964 Helsinki Declaration and its later amendments or comparable ethical standards. For this type of study formal consent is not required. This article does not contain any studies with animals performed by any of the authors.

Disclosure
An earlier version of this study was presented as an abstract in the Experimental Biology 2019 Meeting (Abstract # 250.2) and at the AACR 2019 Annual Meeting (Abstract #5531).