CircRNA circ-PDCD11 is highly expressed in lung large-cell carcinoma and predicts poor survival

Abstract Background Circ-PDCD11 (hsa_circ_0019853, 461 bp) has been characterized as an oncogenic circRNA in breast cancer, while its function in other cancers is unclear. In this study, we explored the role of circ-PDCD11 in lung cancer. Methods Plasma samples were obtained from patients with lung large-cell carcinoma (LLCC, n = 40), lung squamous cell carcinoma (LSCC, n = 40), lung adenocarcinoma (LA, n = 40) and small-cell lung cancer (SCLC, n = 40) as well as healthy controls (Control, n = 40). Paired tumor and nontumor tissue samples were obtained from all patients. Expression of circ-PDCD11 in these samples was determined by RT-qPCR. The role of plasma circ-PDCD11 in the diagnosis of LLCC was analyzed with ROC curve. A five-year follow-up was performed to analyze the role of plasma circ-PDCD11 in the prognosis of LLCC. Results Plasma circ-PDCD11 was specifically upregulated in LLCC but not in other lung cancer types, compared to the controls. Increased circ-PDCD11 expression in tumor tissues compared to nontumor tissues was only observed in LLCC patients but not in other lung cancer types. Increased plasma circ-PDCD11 levels effectively separated LLCC patients from patients with other types of cancers. High plasma circ-PDCD11 levels were closely correlated with poor survival of LLCC patients. Plasma circ-PDCD11 levels were closely correlated with tumor metastasis, but not tumor size of LLCC. Conclusion CircRNA circ-PDCD11 is highly expressed specifically in LLCC and predicts poor survival.


Introduction
The lung is one of the most important organs of the human body. Malignancies in lung has been recognized as the most common cause of cancer deaths in both females and males [1,2]. In most countries, lung cancer affects about 60 out of 100,000 people each year with a decreased onset age in recent years [3,4]. Without proper treatment, most lung cancer patients will die within 7 months after initial diagnosis [5]. Lung cancer is a group of heterogeneous malignancies and cannot be treated by unified treatment protocol [6,7]. Therefore, each subtype of lung cancer should be investigated separately to develop specific therapies.
Lung cancer can be divided into small (SCLC) and nonsmall (NSCLC) cell lung cancer [8,9]. NSCLC can be further divided into lung large-cell carcinoma (LLCC), lung squamous cell carcinoma (LSCC) and lung adenocarcinoma (LA) [8]. LLCC usually starts near the periphery of the lung and originates from large cells, which are bigger than typical cancer cells [10]. Although LLCC only accounts for about 10% of all lung cancers and is relatively less studied compared to other subtypes [10]. Compared to other lung cancer types, LLCC grows and spreads more rapidly, leading to worse overall survival [11]. Other than routine chemotherapy and radiation therapy, LLCC can also be treated with targeted therapy, while specific targets remain lack for this malignancy [12]. Without the capacity of protein-coding, circRNAs may indirectly affect protein synthesis to participate in human diseases, including lung cancers [13,14]. However, studies of the role of circRNAs in lung cancers mainly focus on other types of lung cancers other than LLCC [13,14]. Circ-PDCD11 (hsa_-circ_0019853, 461 bp) is a circRNA that is generated by backsplicing of exon 27 to exon 24 of PDCD11 pre-mRNA [15]. Circ-PDCD11 is known to promote breast cancer progression by sponging miR-432-5p to promote LDHA expression [15]. However, the role of circ-PDCD11 in other malignancies is unclear. Our preliminary sequencing analysis revealed the altered expression of circ-PDCD11 in LLCC. Therefore, we further explored the role of circ-PDCD11 in LLCC.

Patients and healthy controls
This study was conducted in compliance with the Declaration of Helsinki. The subjects of this study were lung large-cell carcinoma (LLCC) patients (15 females and 25 males, n ¼ 40, 15 cases of stage I/II and 25 cases of III/IV, 55.5 ± 6.9 years, negative for synaptophysin, chromogranin, CD56, and INSM1 neuroendocrine markers), lung squamous cell carcinoma (LSCC) patients (15 females and 25 males, n ¼ 40, 13 cases of stage I/II and 27 cases of III/IV, 55.9 ± 6.4 years), lung adenocarcinoma (LA) patients (15 females and 25 males, n ¼ 40, 14 cases of stage I/II and 26 cases of III/IV, 55.8 ± 6.3 years), small cell lung cancer (SCLC) patients (n ¼ 40, 12 cases of stage I/II and 28 cases of III/IV, 55.7 ± 7.1 years), as well as healthy controls (Control, 15 females and 25 males, n ¼ 40, 55.8þ/6.5 years) who were admitted to the Second Affiliated Hospital of Shandong University of Traditional Chinese Medicine from January 2014 to March 2016 (Ethics Committee of this hospital approved this study; Ethics Approval No. YIC72RE). All patient and control groups had similar age and gender distributions. All cancer patients were diagnosed by imaging analysis and biopsies. Healthy controls were enrolled after they completed systemic physiological examinations and showed all parameters within the normal range. All cancer patients and controls singed informed consent.

Plasma and tissue samples
All LLCC, LSCC, LA and SCLC patients and healthy controls were fasted overnight prior to the extraction of fasting blood. About 3 ml fasting blood was extracted from each participant in an EDTA tube. All samples were in centrifuged for 20 min at 1200 g to collect supernatant (plasma) sample. Paired tumor and non-tumor tissue samples were collected from all LLCC, LSCC, LA and SCLC patients either through biopsies or by dissecting the resected tumors from patients who received surgical resection of the primary tumors. Both plasma and tissue samples were immediately subjected to RNA preparation.

Follow-up
The 40 LLCC patients were monitored every month either through outpatient visit or by phone call to record their survival. Survival conditions were recorded and summarized to prepare survival curves.

RNA isolation
Fresh samples after collection were immediately subjected to the preparation of RNA samples using Trizol reagent (Invitrogen). After precipitation with methanol, RNA samples were washed twice with 70% ethanol, dissolved in nucleasefree water and digested by DNase I to remove gDNA. RNA samples were then subjected to 5% urea-PAGE gel electrophoresis to analyze RNA integrity. RNA concentration was determined by nanodrop 2000 and diluted to about 1000 ng/ul using nuclease-free water.

RT-qPCR
With about 1 ul RNA sample as template, SSRT III kit (Invitrogen) was applied to prepare cDNA samples. The quality of cDNA samples was first analyzed by amplifying GAPDH through PCR. Only cDNA samples with high quality were used in the subsequent assays. To determine the expression of circ-PDCD11, cDNA samples were subjected to qPCR with 18S rRNA as the internal control. Ct values were normalized using the 2 À᭝᭝Ct method to quantify the expression of circ-PDCD11. Primer sequences were circPDCD11 forward 5 0 -GGG GCTATGTAGGGTCCAT-3 0 and reverse 5 0 -AGGGTCTTGGAGGAA TCTG-3 0 and 18S rRNA forward 5 0 -GTAACCCGTTGAACCCCAT-3 0 and reverse 5 0 -CCATCCAATCGGTAGTAGC-3 0 [15].

Statistical methods
GraphPad Prism 6 (GraphPad, USA) was applied to plot images and compare datasets. Differences between two groups were compared by Student's t-test. Correlations were analyzed by Pearson's correlation coefficient. The 40 LLCC patients were divided into high and low circ-PDCD11 level groups (n ¼ 20, cutoff value ¼ median plasma level of circ-PDCD11). Associations between patients' clinical data and plasma circ-PDCD11 levels were analyzed by Chi-squared test. Survival curves were plotted for both high and low circ-PDCD11 level groups and compared using log-rank test. The diagnostic value of plasma circ-PDCD11 for LLCC was analyzed by performing ROC curve analysis with LLCC patients as the true positive cases and LSCC, LA and SCLC patients or controls as the true negative cases. p < .05 was statistically significant.

Plasma circ-PDCD11 in different types of lung cancer patients
Circ-PDCD11 levels in plasma samples from LLCC, LSCC, LA and SCLC patients as well as healthy controls were determined using RT-qPCR. Our data illustrated that plasma circ-PDCD11 was specifically upregulated in LLCC but not in other lung cancer types compared to the controls (Figure 1, p < .01). Therefore, circ-PDCD11 may specifically participate in LLCC.
Circ-PDCD11 levels in paired tissue samples from lung cancer patients RNA samples were prepared from paired tumor and nontumor tissue samples from LLCC, LSCC, LA and SCLC patients and subjected to RT-qPCR to determine the expression of circ-PDCD11. Increased circ-PDCD11 expression in tumor tissues compared to non-tumor tissues was only observed in LLCC patients (Figure 2(A), p < .01), not LSCC (Figure 2(B)), LA (Figure 2(C)) and SCLC (Figure 2(D)) patients.

Diagnostic value of plasma circ-PDCD11 for LLCC
Because altered expression of circ-PDCD11 was only observe in LLCC patients, we then explored the diagnostic value of circ-PDCD11 for LLCC by performing ROC curve analysis, in which LLCC patients were the true-positive cases and LSCC, LA and SCLC patients or controls were the true negative cases. It was observed that, using increased plasma levels of circ-PDCD11 as a biomarker, LLCC patients were separated from LSCC (Figure 3(A)), LA (Figure 3(B)) and SCLC (Figure 3(C)) patients and healthy controls (Figure 3(D)). It is worth noting that, using increased expression levels of circ-PDCD11 in LLCC tissue as a biomarker, LLCC patients were also separated from LSCC (Supplemental Figure 1(A)), LA (Supplemental Figure  1(B)) and SCLC (Supplemental Figure 1(C)).

Prognostic value of plasma circ-PDCD11 for LLCC
The 40 LLCC patients were divided into high and low circ-PDCD11 level groups (n ¼ 20). Survival curves were plotted for both high and low circ-PDCD11 level groups and compared using log-rank test. It was observed that high plasma circ-PDCD11 levels were closely correlated with poor survival of LLCC patients (Figure 4). It is worth noting that circ-PDCD11 expression levels were also closely correlated with patients' poor survival (p < 0.05, data not shown).

Associations between patients' clinical data and plasma circ-PDCD11 levels
The 40 LLCC patients were divided into high and low circ-PDCD11 level groups (n ¼ 20). Associations between patients' clinical data and plasma circ-PDCD11 level were analyzed by  Chi-squared test. Plasma circ-PDCD11 were closely correlated with tumor metastasis, but not tumor size of LLCC (Table 1). Therefore, circ-PDCD11 upregulation in LLCC may be involved in tumor metastasis.

Discussion
This study analyzed the involvement of circ-PDCD11 in different types of lung cancer. We found that circ-PDCD11 was only highly expressed in LLCC, but not in other types of lung cancer. In addition, circ-PDCD11 may serve as a potential diagnostic biomarker and prognostic biomarker for LLCC.
Circ-PDCD11 in a recent study was found to be highly expressed in triple-negative breast cancer, and to promote tumor growth by increasing lactate production, glucose uptake, extracellular acidification and ATP generation via sponging miR-432-5p, suggesting its oncogenic role in this  Survival curves were plotted for both high and low circ-PDCD11 level groups and compared using log-rank test. malignancy [15]. The involvement of circPDCD11 in other types of cancers is unclear. This study showed that in both plasma and tumor tissues, circPDCD11 is only highly expressed in LLCC, but not other types of lung cancer. Our study suggested that circPDCD11 is specifically involved in LLCC and different types of lung cancer may require the involvement of different molecular factors. Interestingly, Chisquared test showed that circPDCD11 in plasma is only closely correlated with patients' tumor metastasis, but not tumor growth. We speculate that circPDCD11 may mainly regulate tumor metastasis in LLCC. These results are consistent with the observation that high levels of circPDCD11 were closely associated with patients' poor survival since tumor metastasis directly affects patients' survival. Our observation is in opposition to the role of circPDCD11 in breast cancer, in which circPDCD11 plays an oncogenic role by promoting tumor growth. Therefore, the role of circPDCD11 in tumor growth and metastasis should be further analyzed. At present, there are no specific biomarkers for LLCC. Histopathological biopsy is still the most widely used method to confirm LLCC. However, this method cannot be used in large-scale screening. We showed that plasma circPDCD11 can be used as a biomarker to effectively separate LLCC patients from other types of lung cancer and healthy controls. Different subtypes of lung cancer are usually distinguished by microscopic analysis [16], which requires the collection of tumor tissues. After synthesis, circRNAs may be released into blood stream [13,14]. Therefore, plasma levels of circRNAs may reflect levels of circRNA in tumors and monitoring plasma levels of cricRNAs may assist cancer diagnosis. This study provided a novel way to distinguish lung cancer subtypes by measuring the plasma circPDCD11 levels. This novel biomarker may be applied in the large-scale screening of LLCC. Moreover, high plasma circPDCD11 levels were closely correlated with the poor survival of LLCC patients. Therefore, circPDCD11 may serve as a diagnostic and prognostic biomarker for LLCC. Although circPDCD11 in LLCC tissues showed higher accuracy in the diagnosis of LLCC compared to plasma circPDCD11, tumor tissues are hard to obtain. Therefore, plasma circPDCD11 should be used for large-scale cancer screening. However, this study is limited by the relatively small sample size. The conclusion of this study should be verified by future studies with bigger sample size.
In conclusion, circPDCD11 is overexpressed in LLCC, not other lung cancer subtypes. Increased plasma circPDCD11 levels could serve as a potential diagnostic and prognostic biomarker for LLCC.

Ethical approval
This study passed the review board of Ethics Committee of The Second Affiliated Hospital of Shandong University of Traditional Chinese Medicine. Informed consent was obtained from all individual participants included in the study.

Disclosure statement
No potential conflict of interest was reported by the author(s).

Funding
The author(s) reported there is no funding associated with the work featured in this article.

Data availability statement
The data that support the findings of this study are available on request from the corresponding author.