HO‐1 in lymph node metastasis predicted overall survival in patients with esophageal squamous cell carcinoma receiving neoadjuvant chemoradiation therapy

Abstract Background Lymph node metastasis is one of the pivotal factors of the clinical outcomes of patients with esophageal cancer receiving neoadjuvant chemoradiation therapy (NACRT). Both the nuclear factor‐erythroid 2‐related factor 2 (Nrf2) signaling pathway and heme oxygenase‐1 (HO‐1) are frequently upregulated in various human malignancies and associated with resistance to chemoradiation therapy, subsequently resulting in adverse clinical outcomes. However, the Nrf2 and HO‐1 status in lymph node metastasis and their differences between primary and metastatic lesions are unknown. Aims To examine the levels of Nrf2 signaling proteins and HO‐1 in primary and metastatic lesions of patients with esophageal squamous cell carcinoma using immunohistochemistry. Methods and Results We immunolocalized Nrf2 signaling proteins in 69 patients with lymph node metastases, who received NACRT with 5‐fluorouracil and cisplatin before esophagectomy. We also compared the findings between primary and metastatic lesions. Residual lymph node metastases were detected in 30 patients and among them, both primary and metastatic lesions were available for evaluation in 25 patients. Subsequently, we correlated the results with patients' survival. Nrf2, HO‐1, and the Ki‐67 labeling index were all significantly lower in the patients with lymph node metastases than in those with primary tumors. Carcinoma cells with high HO‐1 levels were significantly associated with pathological resistance to NACRT. These results suggested that overall and disease‐free survival of esophageal squamous cell carcinoma were significantly associated with both pN2 and high HO‐1 levels, respectively. Conclusions Protein expression in the Nrf2 pathway was significantly lower in patients with lymph node metastases than in those with primary lesions. HO‐1 levels in lymph node metastases could be used to predict the eventual clinical outcome of patients with esophageal cancer receiving NACRT.


| INTRODUCTION
Esophageal cancer is one of the most lethal gastrointestinal malignancies. 1 Esophageal squamous cell carcinoma (ESCC) is the most frequent type in Japan and other East Asian countries. The standard therapy for locally advanced ESCC is radical resection with extensive lymph node (LN) dissection following neoadjuvant chemotherapy (NAC). 2 Radiation in conjunction with neoadjuvant chemoradiation (NACRT) is an alternative treatment for locally advanced ESCC. [3][4][5] LN metastasis is a well-known adverse prognostic factor in ESCC, especially for patients receiving NAC or NACRT. 6,7 The histopathological tumor regression grade (TRG) in LN metastases predicts clinical outcomes of patients receiving NAC. 8,9 However, TRG has been evaluated mostly in primary tumors. 10 In addition, there are no standard methods for evaluating the histological efficacy of preoperative therapy in LN metastases.
The nuclear factor-erythroid 2-related factor 2 (Nrf2) signaling pathway plays key roles in regulating antioxidative protein expression. 11 Among those proteins, heme oxygenase 1 (HO-1) exerts antioxidant and apoptotic effects, facilitating carcinoma cell proliferation and resistance to anti-tumor therapy. 12 In addition, both Nrf2 and HO-1 are frequently up-regulated and correlated with tumor progression, aggravation, resistance to treatment, and adverse clinical outcomes in various human malignancies. [13][14][15][16][17] In ESCC, Nrf2 overexpression in primary tumors is correlated with NACRT resistance and adverse clinical outcomes. 18,19 However, the status of Nrf2 in LN metastases in patients receiving NACRT and the differences in Nrf2 between primary and LN metastatic lesions are unknown. Therefore, we intended to clarify the following: (1) the status of antioxidant proteins in LN metastases of patients with ESCC receiving NACRT and its correlation with histological TRG, (2) the differences of antioxidant proteins between primary tumors and metastases, and

| ESCC cases
We examined 69 patients who were diagnosed with stage II-IV ESCC between 2011 and 2015 at the Tohoku University Hospital, Sendai, Japan. All the patients in this study had received NACRT before esophagectomy with regional LN dissection. Residual primary tumors were identified by histological examination in surgical specimens of 50/69 patients. Of the 50 patients, residual LN metastases were detected by histological examination in 30 patients. Five cases with no residual LN metastatic lesions seen during an extensive pathological evaluation were excluded from this study.
Response Evaluation Criteria in Solid Tumors (RECIST), version 1.1 was used to evaluate the clinical therapeutic effects of NACRT. 20 The eighth edition of the American Joint Committee on Cancer/Union for International Cancer Control Tumor-Node-Metastasis (TNM) staging system for esophageal carcinoma was used for tumor staging. 21 The overall survival (OS) and disease-free survival (DFS) rates were determined as the duration from the initial surgery to demise or cancer recurrence, respectively, or last censoring.

| Ethics approval and consent to participate
The study protocol was approved by the Ethics Committee of the Tohoku University School of Medicine (Accession No. 2020-1-87), and informed consent was obtained from all patients prior to surgery through signed consent forms.

| NACRT and surgery
The patients were administered a continuous intravenous infusion of 5-fluorouracil (400 mg/m 2 /day) for 24 h on days 1-5 and 8-12, and cisplatin (40 mg/m 2 ) for 2 h on days 1 and 8. They also received radiotherapy (total of 30 Gy in 15 fractions over 3 weeks). The radiation field was long and T-shaped and contained supraclavicular and mediastinal LNs in the cases of cervical or upper and middle thoracic tumors, and perigastric LNs in those with lower thoracic tumors. In patients who underwent thoracoscopic esophageal subtotal excision, gastric tube reconstruction was performed by hand-assisted laparoscopy or open laparotomy, and cervical esophagogastric anastomosis with regional LN dissection were performed subsequently.

| Evaluation of therapeutic response
The histopathological findings of the resected specimens, including the invasive length of the primary lesions, status of LN metastases, and histological differentiation, were examined independently by two of the authors (F.F. and H.S.). Dissected regional LNs were classified were still discernible; Grade 3: entire lesion displaying coagulative necrosis and/or replaced by interstitial fibrosis, and no histologically identified viable carcinoma cells. 10 Grades 0 and 1 were considered "ineffective" and Grades 2 and 3 "effective." 18,19 The ratio of "effective" LN cases to all pp-MLNs was calculated and defined as the chemoradiation therapy effective rate (CRER). The CRER allowed the classification of individual patients into either thelow-CRER (LCRER, ≥0% and <50%) group or high CRER (HCRER, ≥50%) group. 8 LCRER was considered "ineffective" and HCRER "effective." 8

| Immunohistochemistry
Metastatic LNs with the largest volume of residual tumors were selected after a careful histological review. Tissues had been fixed in neutral 10% formalin and embedded in paraffin wax. Details of immunostaining, including primary antibodies, methods of antigen retrieval, and buffers used, are summarized in Table S1. A product of oxidative DNA damage caused by hydroxyl radicals, 8-hydroxydeoxyguanosine , was used to evaluate the levels of reactive oxygen species (ROS) in carcinoma cells. 22,23 Briefly, 4 μm-thick tissue sections were mounted on clean glass slides and deparaffinized with xylene and ethanol. Sections were then heated for antigen retrieval, as described in Table S1. To inhibit nonspecific antibody binding, the treated sections were incubated with 10% normal rabbit serum (Histofine Kit; Nichirei Bioscience, Tokyo, Japan) for Nrf2 and 8-OHdG, or 10% normal goat serum (Histofine Kit; Nichirei Bioscience) for HO-1, at 24 C for 30 min. The sections were then incubated with primary antibodies overnight at 4 C. which is an oxidative stress marker was also used to study ROS levels in carcinoma cells. 22,23 Nrf2, HO-1, and 8-OHdG were all semi-quantitatively assessed using the H-score, which represents the ratio of immuno-positive cells to the overall carcinoma cells in the examined lymph nodes, and multiplied by the relative immunointensity (0, negative; 1, weak; 2, moderate; 3, marked), ranging from 0 to 300. 24,25 Nuclear Ki-67 immunoreactivity was scored by counting the proportion of immuno-positive cells. 13 We tentatively determined the optimal cut-off values of the histological response of LN metastases using a receiver operating characteristic (ROC) curve. 24  below the cut-off were considered as having "low expression," while those with scores greater than the cut-off value as having "high expression." Stromal cells and lymphocytes were used as positive controls during immunohistochemistry. 26

| Post-NACRT histological TRG and antioxidant proteins in LN metastases
Representative micrographs of the histological TRG of LN metastases are shown in Figure 1. Multiple pp-MLNs were detected in 64.0% (16/25) of the patients (Table S2). In addition, the TRG of pp-MLNs varied in 75.0% (12/16) of those with multiple pp-MLNs (Table S2). Among the ineffective group, 86.7% (13/15) were also classified as NACRT-ineffective, based on the features of their primary tumors. However, 40% (4/10) of the effective group were reclassified as NACRT-ineffective (Table S3). Representative micrographs of Nrf2, HO-1, 8-OHdG, and Ki-67 are shown in Figure 2. In the non-cancerous esophageal squamous epithelium within the radiation field, the relative immuno-intensity of Nrf2 and HO-1 was weak, which made assessment of the findings difficult ( Figure 2). Clinical resistance to NACRT was significantly associated with high HO-1 expression (P = .028; Table 1).

| Overall survival and disease-free survival analysis
The results of the Kaplan-Meier analysis of patients with LN metastases, classified according to HO-1 levels in LN metastases and primary tumors, are summarized in Figure 4. The 5-year OS rate was significantly lower in patients with high HO-1 levels (P = .018). In the primary tumor group, the 5-year OS and DFS rate were significantly lower in patients with high HO-1 levels than in those with low levels.
However, there were no significant differences in the 5-year DFS of the patients (Figure 4). In the univariable Cox analysis, OS was significantly correlated with pN2 (P = .031) and high HO-1 levels (P = .029; Table 2). In addition, DFS was also significantly correlated with pN2 (P = .004; Table 2). However, in the multivariable analysis, none of the variables were independent prognostic factors (