Abstract
Background/Aim: The tumor microenvironment (TME) balances tumor growth and suppression through humoral factors and cell-cell interactions. In oral squamous cell carcinoma (OSCC), TMEs have been associated with prognosis of cancer patients and are evaluated by microscopy; however, these methods of evaluation vary among studies. Materials and Methods: To evaluate the TME, borderline microenvironment fibrosis (bMF) was evaluated histologically in 236 OSCC cases and used to determine the clinicopathological status. Results: bMF was observed in 47% (110 in 236 cases) of OSCC cases and associated with higher T category, N category, stage, histological grade and mode of invasion. bMF-positive was related to overall survival (OS) and progression-free survival (PFS). Multivariate analysis revealed that bMF-positive was an independent factor for OS in all cases [n=226; HR=1.683 (1.018–2.781); p=0.042], especially in T1+T2 cases [n=186; HR=1.926 (1.079–3.440); p=0.024], and PFS in all cases [n=226; HR=2.254 (1.397–3.637); p=0.001]. Conclusion: bMF may act as a novel biomarker for OSCC.
Carcinoma of the oral cavity and pharynx is among the most common types of cancer worldwide, with an estimated 53260 new cases in the United States in 2020 (1). Oral squamous cell carcinoma (OSCC) is a major type of carcinoma of the oral cavity and pharynx, and its incidence has increased by 50% over the past decade. Although the disease can be diagnosed at an early stage by visual inspection and biopsy, it is aggressive even in its early stages, with a lymph node metastasis rate of 30%, and high mortality despite appropriate treatment (2).
OSCCs are genetically unstable, and several oncogene and tumor suppressor gene mutations have been associated with the disease. Histologically, OSCC contains heterogenous types of SCC, ranging from well- to poorly-differentiated. Previous studies have demonstrated that histological grading based on differentiation of SCCs correlates with poor prognosis in OSCC (3, 4). In addition, patterns of invasion correlate with sensitivity to chemotherapy, lymph node metastasis and poor prognosis (3, 4). Perineural invasion, lymph vascular invasion and tumor budding are also independent poor prognostic markers for OSCC (4, 5). Together, these findings indicate that morphological information about cancer cells can play a significant role in the prediction of several factors.
In addition to the tumor cells themselves, the environment surrounding the tumor, known as the tumor microenvironment (TME), has also been suggested to play a role in tumor progression. Thus the TME might provide novel biomarkers for OSCC (6). TMEs comprise populations of cells such as cancer-associated fibroblasts (CAFs), tumor-associated macrophages (TAMs), myeloid derived suppressor cells (MDSCs), and several lymphocyte subsets, including regulatory T cells (Tregs) and cytotoxic T lymphocytes (CTLs). These cell populations can be identified by their molecular characteristics using techniques such as immunohistochemical staining (IHC). One previous study attempted to identify TMEs by using only Hematoxylin and Eosin (HE) staining (3); however, most studies use IHC (7-9), which makes routine evaluation of the TME difficult. These TMEs have been evaluated in a relatively wide range of cancer invasive fronts using microscopy. However, like cancer cells, TMEs are heterogenous and methods for evaluating them vary among researchers.
Because interaction between cancer cells and the TME depends on humoral factors and cell-cell interaction, we hypothesized that proximity between cancer cells and TME is important for the biological function of TMEs. We therefore evaluated the TME in the invasive borderline area of OSCCs by HE staining and identified borderline microenvironment fibrosis (bMF) as a novel independent poor prognostic marker of OSCC.
Patients and Methods
Patients. OSCC patients diagnosed and treated in the Sapporo Medical University Hospital between January 2004 and May 2015 were enrolled in this study. Patients treated with arterial infusion chemotherapy and neoadjuvant chemotherapy were excluded. This study was approved by the Institutional Review Board (IRB) of Sapporo Medical University Hospital (#292-1044). Written informed consent was obtained from all patients in accordance with the guidelines of the Declaration of Helsinki.
Histological analysis. Four-μm thick formalin-fixed paraffin-embedded sections of OSCC biopsy specimen or surgically resected specimen were HE stained and assessed. Three authors (KT, MN, and YH) independently assessed the histology. Borderline microenvironment fibrosis (bMF) was defined by the presence of fibrosis in the cancer invasion margin borderline area. The invasion margin was observed at 200× magnification using a microscope, and the stromal area within 100 μm from the invasive front of OSCC was evaluated. In this study, bMF-positive status was defined as an area occupied over 50% by fibrosis and/or fibroblasts. Below 50% was defined as bMF-negative. The mode of invasion at the invasive front was evaluated as described previously (10). Histological grading was defined as described previously (5).
Statistical analysis. The statistical package for social sciences (SPSS) version 23.0 for Windows (IBM, Chicago, IL, USA) was used for statistical analysis. Survival analyses were performed using the Kaplan–Meier method and compared using the log-rank test for each group. For all statistical analyses, p<0.05 was considered statistically significant.
Results
Definition of borderline microenvironment fibrosis (bMF). Previous studies have shown that OSCC grading based on differentiation and invasion mode and histological differentiation of invasion pattern are related to patient prognosis. Recently, TME has been associated with patient prognosis in several malignant diseases. We thus hypothesized that TME can provide additional prognostic biomarkers, independent of the OSCC invasion pattern, and so we analyzed the invasion pattern and surrounding TME.
The clinicopathological characteristics of each participant are summarized in Table I. Cases included histological Grade 1 (n=122), Grade 2 (n=92), and Grade 3 (n=15), and mode of invasion patterns Yamamoto Kohama Grade (YK)-1 (n=20), YK-2 (n=52), YK-3 (n=88), YK-4C (n=44), and YK-4D (n=10). TMEs are heterogenous and, in some cases, fibrosis-enriched regions and immune cell invasion-enriched regions existed in the same case (Figure 1A). However, we hypothesize that, because humoral factors and cell-cell interactions are important for TMEs, the borderline region of the tumor invasive margin might be important for TME assessment. We therefore analyzed interstitial lesions of OSCC within 100 μm of the invasive front and identified two patterns of interstitial lesion. One pattern was fibroblast/fibrosis enrichment, while the other contained less fibroblast/fibrosis and was more enriched with mononucleated lymphocytes (Figure 1B and C). We defined the former pattern as a borderline microenvironment fibrosis (bMF)-positive pattern and the latter pattern as bMF-negative. In total, 110 of 236 cases were bMF-positive.
Relationship between bMF-positive status and tumor size, lymph node metastasis, stage and mode of invasion. We analyzed the relationship between bMF-positive status and other clinicopathological factors. Cases with bMF-positive status were statistically significantly related to tumor size (T1+T2 vs. T3+T4) (p<0.001), lymph node metastasis (N category) (p=0.002), clinical stage (p<0.001), grade (p=0.037), and mode of invasion (p<0.001) (Table II). Furthermore, bMF, lymph node metastasis (N category), tumor size (T1+T2 vs. T3+T4), clinical stage, tumor grade, and mode of invasion were all related to overall survival (OS) and progression-free survival (PFS) (Figure 2).
Presence of bMF-positive status is related to poorer prognosis in OSCC. To further analyze the significance of bMF, we performed a subgroup analysis. In this analysis, bMF-positive status was correlated with poorer OS in T1+T2 cases (p=0.014) (Figure 3A), but not in T3+T4 cases (p=0.667) (Figure 3B). Furthermore, bMF-positive status was correlated with OS in T1+T2 and histological grade 2+3 cases (p=0.034), but not with T1+T2 and histological grade 1 cases (Figure 3C and D). Finally, we analyzed the significance of bMF using the Cox proportional-hazards model. Lymph node metastasis (p<0.001) and bMF (p=0.041) were correlated with OS across all cases (n=226), and histological grade (p=0.02) and bMF were correlated with OS in T1+T2 cases (n=186) (Table III). Tumor size (p=0.014), mode of invasion (p=0.022), and bMF (p=0.001) were all correlated with PFS across all cases (n=226) (Table IV). These results indicate that bMF is an independent prognostic factor for OSCC.
Discussion
In this study, we identified a novel morphological poor prognostic marker of OSCC, bMF, which is defined only by the borderline of the OSCC invasive margin. Assessment of bMF focuses on a very narrow restricted area of the invasive margin and does not require additional IHC analysis. Morphologically, bMF is similar to the presence of CAFs; however, CAFs are detected as myofibroblasts by the positive-expression of α-SMA, requiring IHC for detection. In addition, evaluation methods for CAFs vary among different reports (11), making it difficult to use CAF evaluation as a standardized biomarker. We did not confirm the expression of α-SMA in bMF in the present study; however, the biological functions of bMF might be similar to those of CAFs. Another unique point about bMF is that it is limited to 100 μm from the borderline lesion. In some cases, bMF can be observed in only borderline lesions, with no fibrosis observed in distant areas (Figure 1A), indicating that proximity to the lesion is important for fibrosis. CAFs function in a pro-tumor stage, through secretion of tumor growth factors and angiogenesis factors including EGF, HGF, VEGF, CXCL12, and MMPs (6). CAFs enhance tumor growth by suppression of CTLs through expression of Fas-L and secretion of iNOS, suggesting that CAFs induce an immune-suppressive TME (12). We observed a lower tumor infiltrating lymphocyte (TIL) infiltration in bMF-positive cases than in bMF-negative cases (data not shown). Together, this suggests that bMF, potentially composed of CAFs and collagen fibers, enhances tumor growth directly through secretion of growth factors and indirectly through suppression of anti-tumor immunity.
In this study, we identified bMF as a novel independent prognostic factor alongside standard prognostic factors such as tumor size, lymph node metastasis, tumor grade, and mode of invasion. Tumor size, mode of invasion, and bMF were all indicated as independent poor prognostic factors by multivariate analysis in our cohort; however, lymph node metastasis and tumor grade did not reach statistical significance, as described previously (4, 13). Considering that bMF is focused on the TME borderline, bMF might represent a tumor-stroma borderline TME, which is not representative of the tumor center or distant interstitial area. Our data present for the first time the importance of borderline TME, especially in small size (T1+T2) and histologically high grade (grade 2+3) cases. This suggests that bMF evaluation might be a powerful biomarker for small OSCCs. In our study, mode of invasion was also statistically significantly correlated with PFS by multivariate analysis. However, mode of invasion focuses on the invasion pattern of cancer cells and not on the TME. Thus, evaluating both the mode of invasion (as a measure of cancer cell invasive ability) and bMF as a measure of the borderline TME might provide a more accurate predictive marker for OSCC.
In summary, we defined borderline microenvironment fibrosis and identified it as a novel independent poor prognostic factor of OSCC.
Acknowledgements
This work was supported by the Project for Cancer Research and Therapeutic Evolution (P-CREATE) from the Japan Agency for Medical Research and Development (AMED) for T. Torigoe.
Footnotes
Authors' Contributions
Kei Tsuchihashi, Munehide Nakatsugawa, Yoshihiko Hirohashi and Toshihiko Torigoe wrote the manuscript. Kei Tsuchihashi, Munehide Nakatsugawa, Hiroko Asanuma and Yoshihiko Hirohashi performed experiments. Kei Tsuchihashi, Munehide Nakatsugawa, Jun-ichi Kobayashi, Takashi Sasaya, Rena Morita, Terufumi Kubo, Takayuki Kanaseki, Tomohide Tsukahara, Yoshihiko Hirohashi and Toshihiko Torigoe analysed data. Tadashi Hasegawa, Hiroshi Hirano, Akihiro Miyazaki and Toshihiko Torigoe supervised the study. All authors reviewed the manuscript.
Conflicts of Interest
The Authors declare that they have no competing interests in relation to this study.
- Received June 4, 2020.
- Revision received June 23, 2020.
- Accepted June 24, 2020.
- Copyright© 2020, International Institute of Anticancer Research (Dr. George J. Delinasios), All rights reserved