Simultaneous targeting PI3K and PERK pathways promotes cell death and improves the clinical prognosis in esophageal squamous carcinoma

Shao-Qi Wang 1, Xiao Wang 2, Kai Zheng 3, Kai-Sheng Liu 2, Shao-Xiang Wang 4, Cong-Hua Xie 5

Highlights
•PI3Ks are upregulated in esophageal cancer tissues and affect survival of patients.
•PI3K inhibitor influences the mTOR and PERK/eIF2α/ATF4 pathways.
•PERK/eIF2α/ATF4 pathway affects the ESCC growth.
•Targeting mTOR and PERK pathways has therapeutic and prognostic values.

Abstract
PI3K pathway is an important anti-tumor target, but its effect and mechanism is not clear in esophageal squamous cell carcinoma (ESCC). By analysis of the Cancer Genome Atlas (TCGA) datasets, we found that PI3Ks level were significantly upregulated in human esophageal cancer tissues compared with that in non-cancer tissues. The alteration of PI3K can significantly affect the overall patient survival in ESCC but not in esophageal adenocarcinoma (EAC). We found that the classic PI3K inhibitor LY294002 obviously inhibited the canonical mammalian target of rapamycin (mTOR) pathway and restrained the growth of ESCC with less toxicity to normal cells. Besides, LY294002 inhibited noncanonical PKR-like ER kinase (PERK)/elF2α/ATF4 pathway as well. Both siRNA and the small molecule inhibitor GSK2656157 against PERK/elF2α/ATF4 pathway can significantly inhibit the growth of ESCC. More importantly, GSK2656157 aggravated the inhibitory effect of LY294002 on cell growth, colony formation, and apoptosis induction of ESCC. In addition of dual high expression of PI3K and PERK pathways in the ESCC patients, the difference of overall survival (OS) was more significant than using PI3K alone. These results indicated that dual targeting of PI3K and PERK pathways might improve clinical prognosis and enhance the treatment of ESCC patients.

Introduction
The molecular target therapy is a great progress in tumor treatment field [1]. The frequent activation of the PI3K/AKT/mTOR pathway in cancer, and its crucial role in survival, proliferation, invasion, angiogenesis, and metastasis of tumor cells, has made it a much desired target for pharmacologic intervention [2], [3]. Several agents have been developed to target Akt/PI3K pathways, including PI3K, PI3K/mTOR, or Akt inhibitors [4]. LY294002 is the first synthetic small molecule inhibitor known to inhibit PI3K pathway. Afterward, other PI3K inhibitors are emerging in an endless stream [5]. Investigation in inhibition of PI3K pathway in lung, head and neck cancers is very common [6], [7], however, relevant research of PI3K pathway in esophageal cancer is scarce [8]. Likewise, there is no remarkable clinical contribution and regulation mechanism in the literature related to inhibition of PI3K pathway in treatment of esophageal cancer.

TCGA (the Cancer Genome Atlas) is currently the largest cancer genome information database [9] and includes the whole genome information (gene mutation, copy number, mRNA expression, etc.), cancer types, morphometric grading, survival period and other key information related to cancer samples with strong clinical guiding significance [10]. The TCGA associated data to the esophageal cancer (updated in 2016) which is a collection of clinical data from 185 cancer patients, including 52% cases of ESCC and 48% cases of esophageal adenocarcinoma (EAC). The database is clinically significance to attract people’s attention to use it as a complete reference for cancer genome.

In this study, we categorized the esophageal cancer patients to squamous carcinoma and adenocarcinoma according to the classified TCGA data. The expression of PI3K gene in esophageal cancer tissue and normal tissue adjacent to carcinoma was investigated. The PKR-like ER kinase (PERK), a transmembrane protein resided in the endoplasmic reticulum (ER), plays a crucial role in the ER stress response [12]. By using the cell biology methods, we found that inhibition of PI3K pathway influenced canonical mTOR pathway and affected noncanonical PERK pathway as well. More importantly, dual high expression of the key genes of PI3K and PERK pathways had better clinical predictive value. Moreover, the single and combination inhibitory effects of PI3K and PERK inhibitors on cell growth, colony formation and apoptosis induction of ESCC were investigated.

Section snippets
TCGA analysis
The clinical features and survival data of esophageal cancer patients were obtained from the TCGA database and mRNA expression pattern of PI3Ks was taken from cBioPortal for Cancer Genomics [13]. Mutations, putative copy number alterations and mRNA were assessed for PI3Ks in the esophageal carcinoma using the cBioPortal Analysis platform. The primary search parameters included mutations, CNA and mRNA expression with the default setting.

Reagents and anti-bodies
Effect of high expression of PI3Ks on patients’ survival
We used the routine analysis of different genes for TCGA in the esophageal cancer database and found that the mRNA levels of PI3Ks (PIK3CA, B, D, G) were expressed higher in esophageal cancer tissue compared to normal adjacent tissue (Fig. 1A). Among these PI3Ks, PIK3CB had the highest mRNA expression in esophageal cancer tissues (P = 0.004). PI3Ks genome alterations in ESCC patients mainly included DNA amplification and mRNA upregulation, while the gene mutation was not common (Fig. 1B).

Discussion
In this study, we preliminary studied the effect and mechanism of PI3K in ESCC. Afterward simultaneous targeting of PI3K and PERK pathways was investigated (Fig. 4H). Interestingly, the mRNA of PI3K was highly expressed in both types of esophageal cancer, whereas only affected the patients’ survival with ESCC. As consistently reported previous [15], although esophageal cancer GSK2656157 includes squamous carcinoma and adenocarcinoma, it should be treated differently.

Acknowledgment
This work was supported by grants in part from the National Natural Science Foundation of China (No. 81602625), Guangdong Natural Science Foundation (No. 2017A010105013 and No. 2016A030310096), the Pearl River S&T Nova Program of Guangzhou (No. 201710010011), China Postdoctoral Science Foundation (No. 2015M572414) and Shenzhen Science and Technology Project (No. JCYJ20150324141711568; No. JCYJ20170302145059926).