Correlation of Clinicopathological Features and IL6 Expression in Tumor Budding of Colon Adenocarcinoma

Background: Interleukin-6 (IL6) is one of the main cytokines produced by cancer-associated broblasts (CAFs). IL6 is linked with cancer progression and poor prognosis by activating cancer cells and modifying the cancer microenvironment. However, little is known about the expression of IL6 in tumor budding (TB) and its association with TB in colon adenocarcinoma (CA). Methods: The clinicopathological and prognostic signicance of IL6 in TB was examined using a tissue microarray consisting of 36 patient samples of TB in CA. IL6 mRNA was detected by RNAscope kit. Patients were stratied into negative and positive IL6 expression groups. Results: IL6 expression was overwhelmingly observed in CAFs but was negligible in cancer cells. In the IL6-positive group in CAFs, TB grade was higher than in the IL6-negative group (P=0.0161). There was a signicant difference in overall survival (OS) between CA cases in the IL6-positive group and the IL6-negative group (log rank test, P=0.0367). Cox proportional hazard regression model revealed that the IL6-negative group (OR = 0.25; 95% CI: 0.05–0.96; P=0.0440) had better OS for CA than the IL6-positive group. Conclusions: TB may be affected by IL6 expression, and IL6 expression in CAFs at TB may make IL6 an important prognostic marker.


Background
Colorectal cancer (CRC) has increasing morbidity and mortality worldwide and is a global health problem [1]. Despite the high prevalence of colorectal cancer, the pathological mechanisms remain largely unknown [2]. However, many prognostic factors for colorectal cancer have been studied. In particular, the tumor budding (TB) region is a unique site and is known to be deeply involved in metastasis and invasion [3]. It has been demonstrated that TB is involved in EMT, which is known to be affected by the surrounding microenvironment of cancer [4,5]. Cancer-associated broblasts (CAFs) have an important role in the cancer microenvironment, and IL6 produced by CAFs is involved in various processes [6]. We focused on the microenvironment in TB. IL6 is an important cytokine but has not been studied in TB. We investigated the clinicopathological characteristics of IL6 expression using RNAscope, a recently developed ISH technique with high sensitivity.

Patients and materials
This study was conducted in accordance with the Declaration of Helsinki and was approved by the ethics committee of Shinshu University School of Medicine (approval no. 4088). Among 115 colon adenocarcinoma (CA) cases surgically resected at Shinshu University Hospital, Matsumoto, Japan between 2010 and 2012, stage I-III cases with TB were selected. Clinicopathological data were obtained from medical records. Materials used for evaluation were archived formalin-xed para n-embedded tissues. According to the report of Lugli et al., TB was classi ed into Bd1 (0-4 buds), Bd2 (5-9 buds), and Bd3 (≥10 buds) [3]. Furthermore, Bd1 and Bd2 were de ned as low-grade TB and Bd3 was de ned as high-grade TB. In the budding area, the score of in ammatory cell in ltration (tumor-in ltrating lymphocytes, TILs) was measured. According to the report of Ropponen et al., the TIL scores were: none, 0; mild, 1; moderate, 2; and marked, 3 [7]. TIL scores were classi ed into low-grade scores 0 and 1 and high-grade scores 2 and 3.

TMA construction
A tissue microarray (TMA) was prepared from para n blocks containing su cient tumor. The TMA was 3 mm in diameter and contained a fully analyzable TB region. The TB region was de ned as an area with a single cell or a detached group of tumor cells consisting of ve cells or fewer, and was selected based on the morphology of the hematoxylin and eosin (H&E)-stained slide [8]. The generation of the TMA was in accordance with our previous report [9].

Statistical analysis
Pearson's chi-squared test, log-rank test, and Cox proportional hazard regression analysis were analyzed by JMP Statistics software version 13 (JMP, Tokyo, Japan). A P-value less than 0.05 was considered signi cant.

IL6 expression in cancer stroma
In the TB region, IL6-expressing cells were mainly identi ed in cancer stroma. These IL6-expressing cells were spindle-shaped and were considered as CAFs (Fig. 1A, 1C). In four cases, IL6 expression could not be detected in the cancer stroma. Thirteen cases could be recognized as the IL6-high expression group. There was no tendency in the distribution of expressing cells in the stroma. However, there was almost no IL6 expression in the cancer cells in the TB region. Cancer cells throughout the TMA core also had little IL6 expression. Thirty cases were completely negative for IL6 expression in cancer cells. IL6 expression in the cancer cells was faint and had no characteristic distribution. No cases could be recognized as IL6positive.
Association between IL6 expression and clinicopathological characteristics As presented in Table 1, the clinicopathological characteristics of patients with CA are described in Table 1. In the IL6-positive group, TB grade was higher than in the IL6-negative group (P = 0.0161). There was no signi cant difference between the IL6-positive group and the IL6-negative group in terms of age, sex, vascular invasion, histological grade, TILs, or TNM stage. Asterisk (*) indicates a signi cant difference between groups (P < 005).
IL6negativity predicts better prognosis of CA To clarify the impact of IL6 expression, Kaplan-Meier analysis with log-rank test was used to evaluate the association between IL6 expression and OS in CA (Fig. 2). The IL6-negative group (median OS, 1980 (range, 1771-2531) days) had signi cantly better OS than the IL6-positive group (median OS, 1556 (range; 1212-2377.5) days) (log-rank test, P = 0.0367).
A Cox proportional hazard regression model revealed the relationship between clinicopathological factors and OS (Table 2). These results revealed that the IL6-negative group (OR = 0.25; 95% CI: 0.05-0.96; P = 0.0440) had better OS for CA than the IL6-positive group.  [14]. In our study, no association was found between IL6 and in ammation expressed as TILs, possibly because of the method of evaluation and the number of cases. However, there are no reports of IL6 expression in CAFs in CRC. Nonetheless, there are reports of IL6 expression from CAFs in several other carcinomas [15] [16]. Qiao et al. reported that IL6 expression from CAFs is associated with poor prognosis in esophageal squamous cell carcinoma [16]. This is the rst report on IL6 expression from CAFs in the TB region, and indicates that IL6 expression is a poor prognostic factor.
TB grade was previously reported to be associated with prognosis [17]. In our study, TB grade was not related to prognosis, possibly because of the small number of samples. The TB region strongly affects metastasis and invasion. Although the mechanism of TB involvement in prognosis is unclear, the involvement of EMT has been reported in recent years [5]. TB in CRC has been shown to upregulate mesenchymal markers and known inducers of EMT, such as the transcription factors ZEB1 and ZEB2 [18]. However, another report revealed that TB shows downregulation of E-cadherin but does not share other regulatory changes common to EMT, suggesting that TB formation may occur by other mechanisms [19] [20]. Yamada et al. reported that ZEB1, an EMT protein, is highly expressed in stroma near TB [20]. Our study demonstrates that IL6 expression is correlated with TB grade. As mentioned above, its involvement of TB and EMT is speculated [20]. EMT and IL6 expression in the cancer stroma are known to be involved in miR-34A suppression [21]. This fact proves an indirect link between TB and IL6. However, IL6-affected TB may be directly involved in EMT IL6-affected TB may be directly involved in EMT.
There are several studies of IL6 in CRC, but these mostly focused on IL6 expression in cancer cells [22] [23]. Although many reports indicate that IL6 expression in cancer cells is associated with poor prognosis [24] [25], one report demonstrated that IL6 expression at other sites confers a favorable prognosis [26]. Meanwhile, Nagasaki et al. reported that IL6 expression is higher in CAFs than in cancer cells when comparing cancer cells and CAFs isolated from human CRC [12]. In our study, IL6 expression has been largely identi ed in the stroma corresponding to CAFs, and IL6 expression in cancer cells is negligible. Therefore, although IL6 produced by CAFs seems to have a strong effect on prognosis, further investigation is necessary. Many reports have examined IL6 expression by immunostaining [24] [25] [26], but there may be many nonspeci c reactions. Thus, RNA in situ measurement may provide more accurate information.
There are several limitations of our study. An increased number of cases would enable more accurate information to be obtained. In addition, expression analysis of IL6 receptor in cancer cells in the TB area should be performed.
Taken together, inhibition of IL6 expression may be a potential therapeutic strategy for the treatment of cancers in which IL6 from CAFs may have important effects.

Conclusions
Our results reveal the relationship between IL6 expression of CAFs and TB in CA. A further study is warranted to con rm these ndings.
Abbreviations IL6, interleukin-6; CAFs, cancer-associated broblasts; TB, tumor budding; CA, colon adenocarcinoma Declarations Ethics approval and consent to participate This study was approved by the ethics committee of Shinshu University School of Medicine (Approval Code: 4088). The requirement of informed consent was waived, and an opt-out method was used due to the retrospective design of the study. The investigation was conducted in compliance with the Helsinki Declaration.

Consent for publication
Not applicable.

Availability of data and materials
All data generated and analyzed during the current study are available from the corresponding author on reasonable request.

Competing interests
The authors declare that they have no competing interests.

Funding
This study was partially supported by the Hokuto Foundation for Bioscience (grant awarded to T.U.). The funding body had no role in the study design, collection, analysis, or interpretation of data, or manuscript writing. The authors declare no con icts of interest.
Authors' contributions KS participated in the design of the study, performed the pathological analysis, and drafted the manuscript. TU and MI helped with the pathological analysis. TU performed statistical analysis. TN and YT conducted immunohistochemistry. KS and YM examined the clinical data of cases. HO and TU critically revised the draft for important intellectual content. All authors have read and approved the manuscript.