Background
Although the surgical techniques and chemotherapy/radiotherapy regimens are greatly improved, cancer is still a serious worldwide public health issue and is the second leading cause of death in the United States [
1]. In 2015, 4,292,000 new cancer cases and 2,814,000 cancer deaths are projected to occur in China [
2]. Therefore, identifying novel biomarkers for early diagnosis and prognosis is necessary for a better control of cancer.
High-throughput genomic platforms revealed that numerous sites within the human genome are transcribed into noncoding transcripts (ncRNAs), which have recently emerged as critical regulators of gene expression. Long noncoding RNAs (lncRNAs) are a class of ncRNAs longer than 200 nucleotides without protein-coding capacity [
3]. Several studies have shown that dysregulated lncRNA expression in various types of cancers is involved in tumorigenesis and cancer metastasis through silencing tumor suppressors or activation of oncogenes at epigenetic, transcriptional and post-transcriptional levels [
4‐
8]. LncRNAs have been implicated as promising biomarkers for diagnosis and prognosis in several cancers [
9‐
11].
HOXA11-AS, the homeobox A11 antisense lncRNA, is located near the homeobox A11 (HOXA11) gene [
12]. Human homeobox (HOX) gene clusters are essential for morphological development, which determine the identity of body segments. Recently studies suggested abnormal expression of HOX in various cancers. Protein coding genes are located on the sense strand of the HOXA gene clusters, while noncoding genes are located on the antisense strand [
13]. The lncRNA HOXA11-AS was first identified in cervical cancer [
14], and upregulated HOXA11-AS expression was closely associated with tumor progression and poor prognosis. Then the role of HOXA11-AS in cancer progression was identified in many other types of cancers, including gastric cancer [
15], epithelial ovarian cancers [
12], bladder cancer, cervical cancer [
14] and glioma [
16]. In vitro and in vivo assays evaluating the effects of HOXA11-AS alterations revealed a complex integrated phenotype affecting cell growth, apoptosis, migration, invasion and stemness maintenance through multiple biologic processes, such as epithelial–mesenchymal transition (EMT) [
13,
17].
However, most studies reported so far are limited in discrete outcome and sample size. Therefore, we conducted a quantitative meta-analysis to clarify the prognostic and clinicopathological significance of lncRNA HOXA11-AS expression in patients with cancer.
Discussion
Recently, substantial evidence has demonstrated that lncRNAs are important regulatory molecules in tumorigenesis. It has been reported that lncRNA FAL1 induces epithelial-to-mesenchymal transition (EMT) in non-small cell lung cancer (NSCLC) [
33]. In gastric cancer, lncRNA H19 promotes cell proliferation and inhibits cell apoptosis [
34]. Li et al. [
35] showed that LncRNA HOTAIR enhanced invasion and metastasis of breast cancer. Moreover, lncRNA HOXA11-AS has been found to induce tumor progression and stemness maintenance in cervical cancer [
14]. Many lncRNAs are identified in various types of cancers, and their dysregulated expression is correlated with tumor progression and patients’ survival, thus making them promising markers for the diagnosis and prognosis. For example, the lncRNAs TUG1 [
36], UCA1 [
37], SPRY4-IT1 [
38], HULC [
9], HOTTIP [
39], H19 [
40] and HOTAIR [
41] were found to be novel promising biomarkers for poor prognosis in human cancers.
A previous meta-analysis combining seven studies has reported that high lncRNA HOXA11-AS expression was a significant indicator for poor OS and PFS from a total of 608 individuals [
42]. Here we conducted independently a meta-analysis to investigate whether lncRNA HOXA11-AS can be served as a prognostic biomarker in human cancers. The present comprehensive meta-analysis pooled a total of nine independent cohorts with 1320 cancer patients, and the results indicated that high HOXA11-AS expression was a prognostic risk factor for OS in cancer patients (HR = 1.86, 95% CI 1.39–2.48).
Subgroup analysis, meta regression analysis and sensitivity analysis were performed due to significant heterogeneity across these studies. The results of subgroup analysis suggested that sample size (more than 100 or fewer than 100) and preoperative treatment (No or unclear) altered the significance of HOXA11-AS prognostic role in OS (HR = 1.69, 95% CI 0.89–3.21 vs. HR = 2.37, 95% CI 1.40–4.02; and HR = 2.28, 95% CI 1.68–3.10 vs. HR = 1.32, 95% CI 0.99–1.76, respectively). However, meta regression analysis failed to identify the source of the significant heterogeneity in these above four covariates (
p > 0.05). Besides, the influence analysis indicated that there were two cohorts from Wang et al. [
26] and Zhang et al. [
30] impacting the pooled HR and its 95% CI apparently. After excluding these two outlier cohorts above, the pooled results were in line with the full meta-analytic results (pooled HR = 2.45, 95% CI 1.85–3.25) without any significant heterogeneity across the remaining studies (
p = 0.445; I
2 = 0). The reason why the study from Wang et al. [
26] caused obvious heterogeneity might owe to the much younger age of patients suffering from osteosarcoma or the different origin of the tumor cells. As for the cohorts from Zhang et al. [
30], data were from the Cancer Genome Atlas (TCGA), lacking sufficient clinicopathological information for patients’ normalization. However, the results of the cohort from Zhang et al. [
30] was in accordance with the full meta-analytic results. We inferred that the biological types of carcinoma and different patients’ basic status might have notable influence. Moreover, the prognostic significance of HOXA11-AS in PFS was evaluated in two studies with 215 patients, and the results indicated that patients with high HOXA11-AS expression were possible to have significantly shorter PFS (pooled HR = 2.47, 95% CI 1.29–4.75) with no significant heterogeneity. Furthermore, by combining HRs from Cox multivariate analyses, we found that HOXA11-AS was an independent prognostic factor of OS and PFS in cancer patients.
Importantly, our analysis revealed that elevated expression level of HOXA11-AS was significantly related to advanced TNM stage and poor tumor differentiation. However, there was no significant association between HOXA11-AS expression and age or gender of patients. In addition, two studies from Sun et al. and Cui et al. suggested that elevated expression level of HOXA11-AS was significantly related to more distant metastasis (OR = 5.782, 95% CI 1.579–21.173). However, the study from Kim et al. showed the insignificant association between HOXA11-AS expression and lymphovascular invasion (OR = 1.556, 95% CI 0.680–3.558), or recurrence (OR = 1.688, 95% CI 0.661–4.309).
Mechanisms underlying the regulatory role of HOXA11-AS in tumorigenesis and tumor progression have been extensively investigated in various types of cancer. Kim et al. found that HOXA11-AS promoted cancer stem cells (CSCs) self-renewal and EMT in cervical cancer cells through regulating the expression of SOX2, Oct-4, Nanog, E-cadherin, β-catenin, and Vimentin [
14]. Sun et al. [
31] revealed that HOXA11-AS served as a critical effector in gastric cancer tumorigenesis and progression via HOXA11-AS/miR-1297/EZH2 crosstalk. Study from Chen et al. indicated that HOXA11-AS promoted EMT in NSCLC through interacting with EZH2 and DNMT1 and repressing miR-200b expression [
27]. In osteosarcoma, HOXA11-AS functions as a competing endogenous RNA and regulates ROCK1 expression by sponging miR-124-3p, thus promoting cell proliferation and metastasis [
28]. A better understanding of the functions of HOXA11-AS in human cancer may help the development of new prognostic and therapeutic strategies.
Recent studies have reported dysregulated HOXA11-AS expression in multiple types of cancers, and demonstrated the underlying mechanisms on lncRNA HOXA11-AS regulating the malignant phenotypes of cancer cells. The present meta-analysis indicated that HOXA11-AS, an oncogenic lncRNA, was a promising biomarkers for prognosis estimation. However, there are limited studies on the regulatory role of HOXA11-AS in the conversion from precancerous lesions to malignancies. For now, it’s quite difficult to figure out the functional importance of HOXA11-AS expression during tumorigenesis. Therefore, more relevant studies are warranted to fill the gap and promote the clinical application of HOXA11-AS as early detectable indicator.
Compared to other biomarkers for cancer, lncRNA HOXA11-AS has been growing to be a promising prognostic biomarkers for reasons as followed: (1) our study demonstrated that overexpressed HOXA11-AS was an independent unfavorable prognostic factor of OS and PFS in cancer patients; (2) many studies revealed that lncRNAs reflected more tumor biological characteristics through interacting with protein coding genes and miRNAs, thus affecting cell growth, apoptosis, and metastasis, etc.
However, the current meta-analysis had some limitations calling for cautious interpretation of the results. First, only eight studies published in full-text were pooled for our meta-analysis, and the data of three cohorts were from databases online (such as TCGA and CCGA). Second, the cut-off value of high and low HOXA11-AS expression level were different among studies, although most of them were set to median. Third, HRs of five cohorts could not be directly obtained from the publications. Moreover, calculating HRs and corresponding 95% CIs through survival curves might not be precise enough. Forth, most of the included studies reported positive results so that our results might overestimate the prognostic significance of HOXA11-AS in cancer to some degree. In case of the significant publication bias, we used trim and fill analysis, and only found another four studies unpublished. However, the filled meta-analytic results supported our original results. Finally, cancer is a complicate and heterogeneous disease. It’s quiet difficult to extrapolate conclusions on the biologic role of HOXA11-AS during tumorigenesis, especially when we take various solid tumors in account. And grading, staging and the investigation of precancerous lesions are important aspects to define the regulatory role of HOXA11-AS in the early or late stages of cancer. Therefore, larger-scale, multi-center, and high-quality studies are warranted to validate our findings.
Authors’ contributions
SM and LA collected and analyzed the data, wrote the paper; FF analyzed the data; CS and YH conceived and designed this study, analyzed the data, wrote the paper; and all authors reviewed the paper. All authors read and approved the final manuscript.