Identification of Combinations of Plasma lncRNAs and mRNAs as Potential Biomarkers for Precursor Lesions and Early Gastric Cancer

Patients with gastric cancer (GC) are usually first diagnosed at an advanced stage due to the absence of obvious symptoms at an early GC (EGC) stage. Therefore, it is necessary to identify an effective screening method to detect precursor lesions of GC (PLGC) and EGC to increase the 5-year survival rate of patients. Cell-free RNA, as a biomarker, has shown potential in early diagnosis, personalised treatment, and prognosis of cancer. In this study, six RNAs (CEBPA-AS1, INHBA-AS1, AK001058, UCA1, PPBP, and RGS18) were analysed via real-time quantitative polymerase chain reaction (RT-qPCR) using the plasma of patients with EGC and PLGC to identify diagnostic biomarkers. The receiver operating characteristic (ROC) curve analysis was used to evaluate the diagnostic accuracy. Among the six RNAs, four lncRNAs (CEBPA-AS1, INHBA-AS1, AK001058, and UCA1) were upregulated and two mRNAs (PPBP and RGS18) were downregulated in the plasma of patients with PLGC and EGC. According to the findings of the ROC analysis, the four-RNA combination of INHBA-AS1, AK001058, UCA1, and RGS18 had the highest area under the curve (AUC) value for determining risk of GC in patients with PLGC and the six-RNA combination including CEBPA-AS1, INHBA-AS1, AK001058, UCA1, PPBP, and RGS18 had the highest AUC value for determining the risk of GC in patients with EGC. The results suggest the potential usefulness of noninvasive biomarkers for the molecular diagnosis of GC at earlier stages.


Introduction
Gastric cancer (GC) is the fifth most common malignant tumour and the fourth leading cause of cancer-related death globally. According to the World Health Organization's global cancer data, 1.09 million new cases and 770,000 deaths occurred due to GC in 2020 [1]. Due to the lack of awareness of early screening for tumours, lack of obvious symptoms in precancer and early cancer stages, and unavailability of more advanced means of early tumour screening, most patients get diagnosed after progressing to advanced stage GC. A survey shows that the 5-year survival rate of patients with early GC (EGC) is >90%, while that of patients with advanced GC is only approximately 30% [2]. erefore, attention must be paid to the timely detection and treatment of EGC.
Although gastroscopy combined with tissue biopsy is the "gold standard" method for tumour detection and diagnosis, endoscopy is limited by the type of equipment and physician techniques and the results vary [3]. Furthermore, because it is not possible to distinguish the heterogeneity of a tumour using tissue specimens, discrepancies often occur between gastroscopic results and pathological diagnosis [4]. Several noninvasive detection methods for GC are available, including circulatory biomarkers. However, currently known tumour antigens in serum, including carcinoembryonic antigen (CEA), carbohydrate antigen 19-9 (CA19-9), CA72-4, and pepsinogen [5,6], are of little use in GC diagnosis due to their low specificity and sensitivity [7]. us, identifying additional biomarkers is urgently needed to improve the diagnostic efficiency of EGC.
In the present study, we investigated the expression of four lncRNAs and two mRNAs in the plasma of patients with EGC and precursor lesions of GC (PLGC) using realtime quantitative polymerase chain reaction (RT-qPCR). Furthermore, the specificity and sensitivity of single and combination RNAs were evaluated for the diagnosis of EGC and PLGC. Our study indicated that the combination of plasma lncRNA and mRNA expression might be a useful biomarker for GC diagnosis.

Plasma Samples.
Plasma samples from healthy participants (n � 120) and patients diagnosed with PLGC (n � 119), EGC (n � 143), colorectal cancer (CRC)/esophageal cancer (EC) (n � 42), or GC pre-/postoperatively (n � 46) were provided by the First Medical Center of the Chinese PLA General Hospital (Beijing, China). None of the patients were administered radiotherapy or chemotherapy treatment prior to the operation. e basic information of the patients is listed in Table 1. All plasma samples were immediately stored at −70°C. Written informed consent was obtained from all patients. e Ethics Committee of the First Medical Center of the Chinese PLA General Hospital (Beijing, China) approved the use of samples for the present study (S2017-010-02).

Examination of Tumour Markers.
e expression levels of tumour markers CEA, alpha fetoprotein (AFP), and CA19-9 in plasma of patients were measured using the electrochemiluminescence immunoassay on the Roche Cobas 8000e602 analyser (Roche Diagnostics, Mannheim, Germany). e recommended normal reference range values for diagnostic purposes were 0-5.0 ng/ml for CEA, 0-20 ng/ml for AFP, and 0.1-37 U/ml for CA19-9. e examination was performed at the clinical laboratory of the First Medical Center of Chinese PLA General Hospital following standard procedures.

Plasma RNA Preparation and RT-qPCR.
Plasma cfRNA was extracted from 200 μL of plasma samples using the miRNeasy Serum/Plasma kit (Qiagen, Valencia, CA) according to the manufacturer's instructions. Total RNA was reverse-transcribed using ImProm-II TM Reverse Transcriptase (Promega, Madison, WI) according to the manufacturer's instructions. All RNA and cDNA products were stored at −70°C. RT-qPCR was performed with the TB Green ® Premix Ex Taq TM (Takara, Shiga, Japan) using the Mx3000p instrument (Agilent Technologies, Santa Clara, CA) according to the manufacturer's instructions. e RNA information and primer sequences are listed in Tables S1 and S2. e expression level of each RNA was quantified using the △Ct method with the 18S rRNA gene as the endogenous control for data normalisation (Table S3), since its expression level did not significantly differ between the plasma samples of patients with GC and healthy controls [25].

Statistical Analyses.
e statistical analyses were performed using GraphPad Prism 5.0 and IBM SPSS Statistics for Windows, version 19.0 (IBM Corp., Armonk, NY). Student's t-test was used for the analysis of data with normal distribution and homogeneity of variance between the two groups, whereas Mann-Whitney U test was used for data with skewed distribution with homogeneity of variance. All statistical tests were two-tailed, and P < 0.05 indicated statistical significance. e specificity, sensitivity, and area under the curve (AUC) of each RNA were determined using receiver operating characteristic (ROC) curve analysis. In addition, the expression level of RNA was analysed using logistic regression analysis to determine the diagnostic value of multiple RNA combinations. e nomogram was formulated using the R 3.5.1 software ( e R Foundation for Statistical Computing, Vienna, Austria) with the rms statistical package.

Positive Rates of Traditional Markers for PLGC and EGC.
To test whether the traditional tumour markers could be used for the prediction of risk of GC in EGC and PLGC, we examined the levels of CEA, AFP, and CA19-9 in plasma of patients with EGC (including 129 patients) and PLGC (including 10 patients with gastricism and 101 patients with LGD or HGD) (Tables S4). We found lower positive rates for single and combined tumour markers in patients with PLGC than patients with EGC (Table 2). e positive rates of these tumour markers in both groups were lower than previously reported positive rates of tumour markers in GC patients with CEA (21.8%), AFP (5.0%), CA19-9 (24.4%), and their combinations (47.1%, including CA125) [26], which indicated that the traditional markers are unsuitable for the prediction of the risk of GC in patients with EGC and PLGC.

Analysis of the Expression Level of the Six RNAs in Plasma of Patients with PLGC.
Based on our previous studies using microarray analysis [22,24], we validated the expression levels of the six RNAs in plasma samples of 119 patients with PLGC and 120 healthy participants using RT-qPCR. Of these six RNAs, CEBPA-AS1, INHBA-AS1, AK001058, and UCA1 were significantly higher and PPBP and RGS18 were significantly lower in the plasma of patients with PLGC than in that of healthy participants (P < 0.001) (Figure 1).
To evaluate whether plasma levels of the six candidate RNAs could be used as diagnostic biomarkers for PLGC, ROC curve analyses were performed. e plasma levels of CEBPA-AS1, INHBA-AS1, AK001058, UCA1, PPBP, and RGS18 differed between patients with PLGC and healthy participants, with AUCs of the ROC curves at 0.651, 0.639, 0.741, 0.692, 0.721, and 0.773, respectively (Figure 2(a)). Furthermore, the AUC of the combined six RNAs was 0.805 (95% confidence interval (CI): 0.751-0.859), and a six-minus-RNA signature was constructed; the six-minus-PPBP RNA signature had an AUC (AUC � 0.819; 95% CI: 0.767-0.871) higher than that for the other five RNA combinations. e combination of INHBA-AS1, AK001058, UCA1, and RGS18 (I&A&U&R) RNA signature had an AUC (AUC � 0.820; 95% CI: 0.767-0.872) higher than that of the other four RNA combinations. We also constructed three-RNA and two-RNA signatures and randomly selected from the six RNAs. eir AUCs were also lower than that of the I&A&U&R RNA signature (Figure 2 erefore, four new RNA signatures, including INHBA-AS1, AK001058, UCA1, and RGS18, were constructed, and a nomogram was developed that incorporated the four significant risk factors for predicting GC. A total score was calculated with four RNAs, and the value of each of these variables was given a score on the point scale axis. A total score could be easily calculated by adding each single score, and by projecting the total score to the lower total point scale, we were able to estimate the probability of GC in patients with PLGC ( Figure 2(c)).

Expression of Plasma RNAs Is Associated with Clinicopathological Features and Grade of Dysplasia in Patients with PLGC.
e relationship between the six RNA levels in the  Table 3. e expression levels of INHBA-AS1, AK001058, UCA1, and PPBP were associated with age, and the expression level of RGS18 was associated with sex. Furthermore, we divided PLGC into gastricism, low-grade dysplasia (LGD), and high-grade dysplasia (HGD) and found that the expression of these plasma RNAs correlated with the degree of the lesion but not with gastricism (Figure 3, * P < 0.05).

Combination of the Six Plasma RNAs Might Predict EGC.
We validated the six RNAs using the same method in 143 patients with EGC, and the results were similar to those of patients with PLGC ( Figure 4). ROC curve analyses indicated that the plasma levels of CEBPA-AS1, INHBA-AS1, AK001058, UCA1, PPBP, and RGS18 differed between patients with EGC and healthy participants, with AUCs of the ROC curves at 0.587, 0.606, 0.681, 0.683, 0.739, and 0.801, respectively ( Figure 5(a)). Furthermore, the AUC of the combined six-RNA panel was 0.845 (95% CI: 0.798-0.892). We also constructed five-RNA signatures, four-RNA signatures, three-RNA signatures, and two-RNA signatures and randomly selected from the six RNAs. e resultant AUCs were lower than that for the six-RNA signature ( Figure 5(b)). erefore, the six-RNA panel, including CEBPA-AS1, INHBA-AS1, AK001058, UCA1, PPBP, and RGS18 and a nomogram that incorporated the six significant     risk factors for predicting GC, were constructed. A total score was calculated with six RNAs by adding each single score. By projecting the total score to the lower total point scale, we were able to estimate the probability of GC in patients with EGC ( Figure 5(c)).

Expression of Plasma RNAs Is Not Associated with the Clinicopathological Features and Differentiation in Patients with EGC.
e relationship between RNA levels in the plasma and the clinicopathological features of patients with EGC are shown in Table 4. ere was no statistical correlation between the expression levels of the six RNAs and age, sex, and tumour size in patients with EGC. We then divided EGC into low differentiated adenocarcinoma, moderately differentiated adenocarcinoma, and highly differentiated adenocarcinoma. Although the expression of some plasma RNAs such as INHBA-AS1 and AK001058 appeared to be correlated with the LGD, low-grade dysplasia; HGD, high-grade dysplasia. * P < 0.05, * * P < 0.01, * * * P < 0.001, * * * * P < 0.0001.  Journal of Oncology differentiation degree of GC, there was no regularity in the plasma RNA expression variation in the differential stages between patients with GC and healthy controls ( Figure 6). Furthermore, we analysed the expression of plasma RNAs in patients with GC pre-and postoperatively. Four RNA levels were not corrected following the operation for GC; however, the expressions of INHBA-AS1 and UCA1 were significantly lower postoperatively than preoperatively (P < 0.001) (Figure 7).

Plasma RNAs Might Be Expressed in Other Types of Gastrointestinal Cancers.
To verify whether the abovementioned six RNAs could distinguish other types of gastrointestinal cancers, plasma samples from 42 patients with other types of gastrointestinal cancers were assessed, including 11 patients with CRC and 31 patients with EC. As shown in Figure 8, the expression levels of the six RNAs showed similar pattern for patients with EC/CRC and those with EGC. However, the expression levels of CEBPA-AS1, INHBA-AS1, AK001058, and UCA1 were significantly different in the plasma of patients with EC/CRC and patients with EGC. e expression levels of the four lncRNAs were lower in patients with EGC than in healthy controls, and they were much lower in patients with EC/CRC than in healthy controls. Furthermore, although RGS18 was different between patients with EGC and EC, PPBP was not different between patients with EGC and EC/CRC, the expression levels of the two mRNAs were lower in patients with EC/CRC than in patients with EGC but were higher than that in healthy controls (Figure 8). ese results are only suggestive due to the small sample size of the study, and a comparative analysis of a large sample sized study is required.

Discussion
Traditional serum markers such as CEA, CA19-9, APF, and CA125 are used widely to diagnose GC, although the positive rates of combination of these markers were 50% at most in GC [26] and much lower in EGC [5]. In recent years, cfRNA as a biomarker for early diagnosis has become a research hotspot in noninvasive diagnosis of gastric cancer, including miRNAs [27], lncRNAs [22] and mRNAs [24]. In this study, we investigated the combination of lncRNAs and mRNAs for the prediction of EGC, and we found that the combination of six RNAs (CEBPA-AS1, INHBA-AS1, AK001058, UCA1, PPBP, and RGS18) for predicting EGC and the      Journal of Oncology combination of four RNAs (INHBA-AS1, AK001058, UCA1, and RGS18) for PLGC showed a better prediction than with lncRNA or mRNA alone. As previously reported, compared with healthy participants, the expressions of lncRNA CEBPA-AS1, INHBA-AS1, AK001058, and UCA1 were significantly higher [15,22,28] and the expressions of PPBP and RGS18 were significantly lower [24] in the plasma of patients with EGC, indicating these as potential biomarkers for GC [15,22,24]. However, our results showing that the combination of these plasma lncRNAs and mRNAs rather than a single lncRNA or mRNA alone can more efficiently distinguish patients with EGC from healthy participants. us, the combination of different RNA molecules may predict cancer occurrence and improve the diagnostic efficiency when compared with RNA molecules used alone, as diagnostic markers. Another study  reported that the combination of lncRNA H19, MEG3, and miRNA miR-675-5p improved the diagnostic efficiency compared with using each alone [27]. Although previous studies showed that most biomarkers of GC are associated with cancer differentiation [14,17,29,30], and some RNAs in this study including CEBPA-AS1, IHNBA-AS1, and AK001058 exhibited increasing trends with tumour progression, the expressions of these RNAs were not associated with GC differentiation in heathy participants which should be closed to the high differentiation in patients with GC. Our study results suggest the need to consider healthy controls when assessing the relationship between expressions of biomarkers and cancer differentiation.
In the stomach, chronic atrophic gastritis and dysplasia are two main precursor lesions preceding the development of GC [31]. However, although several studies have reported that miRNAs let-7 and miR-421 can be used for predicting PLGC [32,33], little is known about the biomarkers of PLGC. To investigate the early diagnosis of EGC, using plasma samples from patients with PLGC, the combination of four RNAs including INHBA-AS1, AK001058, UCA1, and RGS18 was analysed and found to be a candidate biomarker for the diagnosis of PLGC and indicated an increased risk of GC. Interestingly, the expressions of some of these RNAs were related to the age and sex of patients with PLGC but not those with GC, suggesting the effects of clinicopathological features on precancerous development of GC. Meanwhile, their expressions were associated with the degree of dysplasia but not with gastricism, which is in accordance with previous reports of different lncRNA expression profiles between EGC and gastritis tissues [34] and indicate an important tumorigenic transformation between gastricism and dysplasia-this requires further investigation.
We also investigated whether these plasma RNAs were differentially changed in other gastroenteric tumours including CRC and EC. We found most of them with similar patterns in CRC and EC but significantly differentially expressed compared to EGC. None of these RNAs was reported in patients with EC, while some were not reported in patients with CRC. Besides lncRNA UCA1 is upregulated and promotes CRC proliferation via the miR-143/MYO6 axis or via RNA-RNA interactions [35,36]; INHBA-AS1 is highly expressed in CRC and promotes proliferation by sponging miR-422a to increase the AKT1 axis [37]; AK001058 is significantly increased in CRC and promotes proliferation, invasion, migration, and prolongs the S stage of CRC by regulating the methylation of ADAMTS12 [38].
ese results indicate that lncRNAs function in CRC and may be specific biomarkers for CRC. Together with previous reports of UCA1 upregulation in liver and pancreatic cancers [39,40] as well as CEBPA-AS1 in liver cancer [41], our results suggest that these plasma RNAs can be universal tumour markers for the diagnosis of gastrointestinal cancers. However, additional experimental studies are required to verify their expression in other gastrointestinal cancers besides GC, with a sufficient sample size to ensure consistent and repeatable results.

Conclusions
Our findings suggest that the plasma cfRNAs were differentially expressed in PLGC and EGC and that they are useful noninvasive tumour markers for the molecular diagnosis of GC. e combination of lncRNAs and mRNAs provided a higher diagnostic efficiency than single RNAs. is study has the potential to provide new insights into the application of cfRNAs as a noninvasive biomarker in large-scale screening of EGC at a precursor stage and as a general biomarker for diagnoses of gastrointestinal cancers.

Data Availability
e data used to support the findings of this study are available from the corresponding author upon request.

Conflicts of Interest
e authors declare that they have no conflicts of interest regarding the publication of this paper.

Supplementary Materials
Table S1: information on lncRNA CEBPA-AS1, INHBA-AS1, AK001058, UCA1, and mRNA PPBP and RGS18. Table S2: sequences of primers used in the present study. Table S3: expression of plasma RNAs in patients with PLGC and EGC in the present study. Table S4: