Clinical Significance of Various Drug-Sensitivity Markers in Patients with Surgically Resected Pulmonary Pleomorphic Carcinoma

Various drug-sensitivity markers are potentially responsible for tumor progression and chemotherapy resistance in cancer patients with both epithelial and sarcomatous components; however, the clinicopathological significance of drug-sensitivity markers in patients with pulmonary pleomorphic carcinoma (PPC) remains unknown. Here, we clarified the prognostic impact of these drug-sensitivity markers in PPC by performing immunohistochemical and clinicopathologic analyses of samples from 105 patients with surgically resected PPC in order to evaluate levels of vascular endothelial growth factor 2 (VEGFR2), stathmin 1 (STMN1), tubulin β3 class III (TUBB3), thymidylate synthetase (TS), topoisomerase II (Topo-II), glucose-regulated protein, and 78 kDa (GRP78)/binding immunoglobulin protein (BiP). We observed the rates of high expression for VEGFR2, STMN1, TUBB3, TS, Topo-II, and GRP78/BiP were 33% (39/105), 35% (37/105), 61% (64/105), 51% (53/105), 31% (33/105), and 51% (53/105) of the samples, respectively. Moreover, multivariate analysis identified VEGFR2 and GRP78/BiP as significant independent markers for predicting worse prognosis. These findings suggested elevated VEGFR2 and decreased GRP78/BiP levels as independent factors for predicting poor outcomes following surgical resection in patients with PPC.


Patient Characteristics and Immunohistochemistry
The percentages based on differential expression of the target markers are listed in Table 1, and patient demographics based on different markers expression are listed in Table 2. The median patient age was 69 years (range: 35-88 years), with 79 men (75%) and 26 women (25%). Eighty-four patients (80%) were smokers and 34 (32%) were diagnosed with stage I disease, 37 (35%) with stage II, 27 (26%) with stage III, and seven (7%) with stage IV. All patients were diagnosed using resected primary tumors. Histologic analysis revealed that 29 patients with PPC harbored a combination of carcinomatous and sarcomatous components. In the remaining 76 primary tumors, carcinomatous components were identified in 48 patients with adenocarcinoma, 13 with squamous cell carcinoma, 8 with adenosquamous cell carcinoma, 2 with small cell carcinoma, and 5 with large cell carcinoma. Of the sarcomatous components, 69 patients exhibited spindle-cell type, 10 giant-cell type, and 26 both spindle-and giant-cell types. The day of surgery was considered the start day for measuring postoperative survival, and the median follow-up period was 476 days (range: 30-4519 days).
We performed an immunohistochemical examination of the 105 primary samples from PPC patients, with representative findings for each of the markers presented in Figure 1. The percentages of high expression and average scores for VEGFR2, TUBB3, STMN1, Topo-II, TS, and GRP78/BiP are listed in Table 1.
Our results showed the following: elevated VEGFR2 level was significantly linked with STMN1 level; elevated TUBB3 level was significantly associated with gender, vascular invasion, and TS; elevated STMN1 was significantly associated with VEGFR2 and TS levels; elevated Topo-II level was significantly associated with age and the T factor; elevated TS level was significantly associated with TUBB3, STMN1, and GRP78/BiP level; and elevated GRP78/BiP level was significantly associated with the N factor and pathological stage ( Table 2).   We performed an immunohistochemical examination of the 105 primary samples from PPC patients, with representative findings for each of the markers presented in Figure 1. The percentages of high expression and average scores for VEGFR2, TUBB3, STMN1, Topo-II, TS, and GRP78/BiP are listed in Table  1.

Survival Analysis
The median disease-free survival (DFS) and overall survival (OS) for all patients were 443 days and 991 days, respectively. The median DFS and OS for patients with adenocarcinoma and non-adenocarcinoma components were 522 days and 1038 days, and 336 days and 507 days, respectively. Of the 105 patients, 60 died after the initial surgery. Univariate and multivariate analyses were performed in all patients (Table 3; Table 4), with univariate analysis identifying disease stage and VEGFR2, TUBB3, STMN1, Topo-II, TS, and GRP78/BiP levels as significant prognostic markers for OS in all patients and disease stage, lymphatic permeation, pleural involvement, and VEGFR2 and STMN1 levels as significant predictors for DFS in all patients ( Table 3). Application of a univariate log-rank test enabled screening of variables with a cut-off of p < 0.05, with subsequent multivariate analysis identifying disease stage and VEGFR2 and GRP78/BiP levels as independent prognostic factors for predicting a worse OS in all patients and disease stage, pleural involvement, and VEGFR2 and STMN1 levels as significant prognostic markers for DFS in all patients were (Table 3). Figure 2 shows the Kaplan-Meier survival curves for patients exhibiting high and low levels of VEGFR2, STMN1, TUBB3, TS, Topo-II, and GRP78/BiP, respectively.
We then performed univariate and multivariate analyses according to histological type. Univariate analysis identified disease stage and STMN1 and TS levels as significant prognostic markers for OS in patients with an adenocarcinoma component, and disease stage and VEGFR2 and TUBB3 levels as significant prognostic markers for OS in patients with a non-adenocarcinoma component. Additionally, significant predictors for DFS in patients with an adenocarcinoma component were STMN1 and GRP78/BiP levels, and significant predictors for DFS in patients with a non-adenocarcinoma component were disease stage and VEGFR2 level (Table 3). Furthermore, multivariate analysis confirmed disease stage as an independent prognostic factor for predicting worse OS in patients with an adenocarcinoma component and disease stage and VEGFR2 level as independent prognostic factors for predicting worse OS in patients with a non-adenocarcinoma component. Additionally, significant prognostic markers for DFS in patients with an adenocarcinoma component were STMN1 and GRP78/BiP levels and disease stage and VEGFR2 level in patients with a non-adenocarcinoma component (Table 4).

Discussion
To the best of our knowledge, this is the first study evaluating the clinicopathological significance of various drug-sensitivity markers, including VEGFR2, STMN1, TUBB3, TS, Topo-II, and GRP78/BiP, in patients with surgically resected PPC. We found that elevated VEGFR2 levels and decreased GRP78/BiP levels were independent prognostic variables for predicting a worse outcome after surgical resection. Additionally, the data suggested that VEGFR2 plays a crucial role in the survival of patients with non-adenocarcinoma-component PPC. PPC is recognized as a rare tumor with a dismal outcome, and there are few promising drug targets for its treatment; however, our findings suggest VEGFR2 and GRP78/BiP as potential target candidates.
In this study, VEGFR2 was expressed in 33% of all PPC patients, and the data suggested that elevated VEGFR2 level can be used as an independent prognostic variable for predicting a worse outcome after surgical resection. VEGFR2 is the main pro-angiogenic receptor for VEGF-A and plays a key role in tumor-induced angiogenesis [7]. These data suggest that VEGF-A and VEGFR2 play important roles in cancer-cell physiology. Interestingly, previous reports showed VEGFR2 expression in 30% to 91% of tumor cells [24,25] along with associated poor prognosis [24,26]. Another study showed that compared with other NSCLCs, 18F-fluorodeoxyglucose (18-FDG) uptake, glucose metabolism, angiogenesis, and cell proliferation were significantly elevated in patients with PPC [5], with 18F-FDG uptake significantly correlated with VEGF expression, which is an important prognostic indicator in various cancers [27]. Angiogenesis is essential for tumor growth, and enhanced vascular supplies of nutrients and oxygen are reflective of malignant potential. Our results suggested that VEGFR2 levels might offer prognostic value to aid in the identification of patients as potential responders to antiangiogenic therapies, such as bevacizumab and ramucirumab.
We found that GRP78/BiP was expressed in 51% of the PPC patients, and that a low level of GRP78/BiP represented an independent prognostic variable for predicting a worse outcome after surgical resection. GRP78/BiP is involved in protein folding and assembly in the ER and increased resistance to ER-stress-induced apoptosis [21][22][23], and its levels are reportedly elevated in various cancers and closely associated with malignancy, metastasis, and chemotherapy resistance [22,23]. A previous study suggested elevated GRP78/BiP level as a significant factor for predicting a favorable prognosis in patients with lung cancer [28], whereas a review indicated that elevated GRP78/BiP level correlated with drug resistance, tumor recurrence, and poor survival [23]. These findings suggest that GRP78/BiP induction could represent a therapeutic strategy targeting drug-resistant cells associated with lung, bladder, and breast cancers, whereas GRP78/BiP inhibition might be effective against resistant cells associated with gastric cancer, transformed fibroblasts, and epidermoid carcinoma. Although previous studies suggested GRP78/BiP induction or inhibition as a potential therapeutic strategy [22,23], further investigations focused on the discovery of small molecules related to modulation of ER-stress-related signaling and GRP78/BiP levels are required. Our findings indicated altered GRP78/BiP level as an ER-stress marker and involved in PPC pathogenesis and development, suggesting GRP78/BiP as a potentially promising molecular target for the treatment of PPC.
Furthermore, our data indicated that STMN1, TUBB3, TS, and Topo-II levels did not correlate significantly with OS in PPC patients; however, these markers might play an important role in PPC development. Our future work will examine relationships between currently used drugs and levels of these markers.
This study has several limitations. First, although samples were collected from multiple institutions, the number of patients was too low to confirm our results statistically. Moreover, due to the rarity of PPC, we were unable to validate our results by using another cohort; therefore, a similar study should be performed using a larger cohort. Second, we were unable to accurately detect the epithelial growth factor receptor (EGFR)-mutation status of the primary tumors. The multi-institutional retrospective nature of this study limited the detection of the EGFR mutation to the physicians' discretion; therefore, we were unable to collect complete data concerning the frequency of the mutation. Our previous study demonstrated that EGFR mutation was detected in~18% of patients with PPC, especially those with an adenocarcinomatous component; however, in those with a sarcomatoid component, KRAS mutations were not observed in all patients [5]. Additionally, Sartori et al. reported that EGFR mutations were not detected in a cohort of patients with sarcomatoid carcinomas [29]. Although it remains unclear whether EGFR-tyrosine kinase inhibitors are effective for treating PPC with EGFR mutations, further investigation is warranted to elucidate the potential of molecular-targeting therapy. In the present study, EGFR mutation was not evaluated in patients with PPC; however, future studies should focus on the relationship between drug-sensitive markers and EGFR mutation in patients with PPC. Third, although a previous report described GRP78/BiP in human non-small cell lung cancer [30], it remains unclear whether manipulation of VEGFR2 and/or GRP78/BiP levels within PPC cells represents an efficacious strategy to decrease tumor growth. Our data indicated a close association between VEGFR2 and GRP78/BiP and PPC metastasis, tumor invasiveness, and patient survival; however, whether their targeting would increase chemotherapeutic efficacy remains unclear. Further investigations should focus on the control of VEGFR2 and GRP78/BiP levels in PPC using in vitro and in vivo studies. Although few patients in our cohort had been treated with ramucirumab, a VEGFR2 inhibitor, this might represent a promising future treatment strategy.

Patients
Patients with histologically confirmed PPC and who underwent surgical resection at multiple institutions between August 2001 and October 2015 (n = 105) were enrolled in this study. Pleomorphic carcinoma was diagnosed according to the 2015 World Health Organization Classification of Tumours [3]. Diagnoses were confirmed using light microscopy and immunohistochemistry. PPC was defined as NSCLC containing at least 10% sarcomatoid components. Surgically resected primary tumors (n = 105) were included in this study in accordance with institutional guidelines and the Helsinki Declaration. The institutional review boards of all participating institutions approved this study. This research has been approved by Gunma University ethic committee on 24 November 2017 (ethic code: 1385). The currently collected tumor samples used in this study were performed using those of our previous study [31].

Statistical Analysis
Statistical analyses were performed using Student's t-test and the χ 2 test for continuous and categorical variables, respectively. The Kaplan-Meier method was used to estimate survival as a function of time, and survival differences were analyzed by the log-rank test. OS represented the time from tumor resection to death from any cause, and DFS represented the time between tumor resection and the first episode of disease progression or death. Univariate and multivariate survival analyses were performed using a Cox proportional hazards model and a logistic regression model for radical surgery. A p < 0.05 was considered statistically significant. All statistical analyses were performed using GraphPad Prism software (v.4.0; GraphPad Software, San Diego, CA, USA) and JMP Pro software (v.12.0; SAS Institute Inc., Cary, NC, USA).

Conclusions
Analysis of resected samples from PPC patients indicated elevated levels of VEGFR2 and attenuated levels of GRP78/BiP as independent predictive factors for poor outcomes after surgical resection. These findings suggested that the management of VEGFR2 and GRP78/BiP levels might enhance the treatment of PPC.