Screening prognosis-associated key eRNAs and target genes in UCEC

One thousand five hundred eighty eRNA expression data was extracted from the samples of UCEC in TCGA database, as well as the survival and clinical phenotypes dataset, and 120 eRNA gene expressions were found associated with the UCEC patients’ overall survival according to the results from the Kaplan–Meier log-rank test (p < 0.05). Spearman rank correlation coefficients were then used here to identify the significant correlations between the prognostic eRNAs and their predicted target genes thus helping with the further screening (Spearman’s rank correlation coefficient R > 0.5, p < 0.05). Finally, the relevant information of the 48 key eRNAs and their target genes was summarized in Table 2. Especially, the UCEC patients with high expression of eRNA IGFBP7-AS1 showed better overall survival performance compared to those low IGFBP7-AS1 expression ones (Fig. 1A, p < 0.05). In addition, as indicated in Fig. 1B, IGFBP7-AS1 had a closely positive correlation with IGFBP7 (R = 0.51, p < 0.001).

Table 2 List of overall survival associated genes derived from enhancersFig. 1

Impact of IGFBP7-AS1 and IGFBP7 on Uterine Corpus Endometrial Carcinoma (UCEC) A Kaplan–Meier curve analysis of high-level and low-level IGFBP7-AS1 expression groups B Scatter plot showing the significant correlation between IGFBP7-AS1 and IGFBP7 levels

Data validation of IGFBP7-AS1 and IGFBP7 levels by reverse transcription quantitative PCR (RT-qPCR)

qRT-PCR was used to measure the expression levels of IGFBP7-AS1 and IGFBP7 in UCEC samples and paired adjacent samples. Down-regulation of IGFBP7-AS1 and IGFBP7 was seen in tumors when comparing it with adjacent samples, in accordance with its expression trend in the TCGA dataset (Fig. 2A, B). A significantly positive correlation was seen between these two genes in tumor tissues (R = 0.776, p < 0.001, Fig. 2C).

Fig. 2

Quantification of IGFBP7-AS1 and IGFBP7 levels by RT-qPCR. A, B Expression levels of IGFBP7-AS1 and IGFBP7 in UCEC samples and paired adjacent samples. C Significant correlation between IGFBP7-AS1 and IGFBP7 levels in tumor samples

Correlation of IGFBP7-AS1 expression with clinical features and prognosis survival of UCEC

Using data from GEPIA database, we compared the IGFBP7-AS1 and IGFBP7 expression levels across UCEC normal and cancer tissues, and it was clear that they all had decreased expression in tumor tissue (Fig. 3A, B). Subsequently, we explored the relationship of IGFBP7-AS1 and IGFBP7 expression with clinicopathologic variables of UCEC and found that the expression level of IGFBP7-AS1 was significantly different in different age (Fig. 3C), tumor grade (Fig. 3D) and histological type (Fig. 3E) group; however, clinical stage showed no significant correlation with IGFBP7-AS1 levels (Fig. 3F). Besides, IGFBP7 was found related to patient age (Fig. 3G), tumor grade (Fig. 3H), but had no evidence for its correlation with histological type (Fig. 3I) and clinical stage (Fig. 3J). Using Cox regression, we further investigated the relationship between IGFBP7-AS1 and OS, along with the clinicopathologic characters before. The results were listed in Table 3, and some parameters, including clinical stage, tumor grade and IGFBP7-AS1 levels, turned out to be linked with OS both in univariate and multivariate regression. Figure 3K, showing a forest boxplot from Multivariate analysis, suggested that mRNA IGFBP7-AS1 was an independent favorable prognostic factor to UCEC. Since the variable age and histological type were insignificant in the multivariate analysis, another multivariate Cox model with only the significant ones of clinical stage, tumor grade and IGFBP7-AS1 levels was conducted to verify the results, in which all the three variables remained to be significant, supporting the finding of IGFBP7-AS1’s independent prognosis for UCEC (Supplementary Table 1).

Fig. 3

IGFBP7-AS1 and IGFBP7 expression in UCEC and their correlations with various clinicopathology features and prognosis survival. A IGFBP7-AS1 expression in UCEC tumor and normal tissues (the red for tumor tissues and the black for normal tissues). B IGFBP7 expression in UCEC tumor and normal tissues (the red for tumor tissues and the black for normal tissues). C The expression of IGFBP7-AS1 in different age, D tumor grade, E histological type and F clinical stage. G The expression of IGFBP7 in different age, H tumor grade, I histological type and J clinical stage. K forest plot of Multivariate COX regression analysis about age, stage, histological type, tumor grade and IGFBP7-AS1 expression

Table 3 Univariate analysis and multivariate analysis for the association of IGFBP7-AS1 expression with OS of UCEC patientsEnrichment analysis

GO enrichment analysis and KEGG pathway analysis were performed here so as to provide more information about the biological function of IGFBP7-AS1. Figure 4A, B were the bubble plots about the GO terms and KEGG signaling pathways of IGFBP7-AS1. Top 10 terms for biological process (BP), cellular component (CC), molecular function (MF) was extracted as the most important ones associated with IGFBP7-AS1——leukocyte migration in BP and extracellular matrix in CC were demonstrated to have strong correlation with IGFBP7-AS1. The highest-ranking signaling pathways in KEGG were PI3K − Akt signaling pathway, Neuroactive ligand − receptor interaction, Cytokine − cytokine receptor interaction and MAPK signaling pathway, which means IGFBP7-AS1 may regulate leukocyte migration through these pathways to influence on the occurrence and development of tumors. Moreover, Neuroactive ligand − receptor interaction and Cytokine − cytokine receptor interaction were all immune related pathways, which suggested the connection between IGFBP7-AS1 and immune activities.

Fig. 4

GO term and KEGG pathway analysis in UCEC. A bubble plot of GO term analysis B of KEGG pathway analysis

IGFBP7-AS1 expression correlated with immune status

The tumor immune microenvironment (TME) plays an essential role in activating heterogeneity among cancer cells, thus stimulating multidrug resistance and resulting in the occurrence, development, and metastasis of tumors. The tumor immune microenvironment has been a critical part in tumor research so we calculated the stromal score, immune score, ESTIMATE score and tumor purity of each tumor sample by ESTIMATE and compared them in high and low IGFBP7-AS1 and IGFBP7 expression groups. A high stromal and immune score or low tumor purity usually contributed to better prognosis and response to immunotherapy. The results indicated that increased IGFBP7-AS1 and IGFBP7 expression was closely related with higher stromal score, immune score, estimate score and lower tumor purity (Fig. 5A-H). So high levels of IGFBP7-AS1 and IGFBP7 may be beneficial for immunotherapy effect of UCEC.

Fig. 5

TME analysis for IGFBP7-AS1 and IGFBP7. A The IGFBP7-AS1 expression differences in Stromal score, B Immune score, C ESTIMATE score and D Tumor purity. E The IGFBP7 expression differences in Stromal score, F Immune score, G ESTIMATE score and H Tumor purity

To have a better knowledge of their association with immune related cells, we employed CIBERSORT to further process the biological role of IGFBP7-AS1 in TME. Results in boxplot of Fig. 6A indicated that T cells CD4 memory activated, T cells follicular helper, T cells regulatory (Tregs), T cells gamma delta, Macrophages M1, Macrophages M2 and Mast cell resting were the immune cells affected by IGFBP7-AS1 expression. Among them, T cells CD4 memory activated, T cells follicular helper, T cells gamma delta, Macrophages M1 and Macrophages M2 showed low proportion in the high IGFBP7-AS1 expression groups compared to the low. In contrast, T cells regulatory (Tregs) proportion and Mast cell resting were clearly up-regulated in high-level IGFBP7-AS1 group.

Fig. 6

Relationship of IGFBP7-AS1 expression with immune related cells and immune infiltration scores of TILs and immune-related substances. A The boxplots of immune cells proportion in different IGFBP7-AS1 expression groups (the red for high expression group and the blue for low expression group of IGFBP7-AS1). B Scatterplot of IGFBP7-AS1 correlated with B Cells memory, C B cells native, D Macrophages M1, E Macrophages M2, F Mast cells resting, G T cells follicular helper, H T cells gamma delta and I T cells regulatory (Tregs). J Scatterplot of IGFBP7 correlated with memory B cells, K Dendritic cells activated, L Macrophages M0, M Mast cells resting, N Plasma cells, O T cells CD4 memory resting, P T cells follicular helper and Q T cells gamma delta. R The boxplots of immune infiltration cell score in different IGFBP7-AS1 and S IGFBP7 expression groups

We further investigated the detailed correlation between immune cells and expression of IGFBP7-AS1 and IGFBP7. Levels of IGFBP7-AS1 expression were inversely correlated with memory B cells, Macrophages M1, Macrophages M2, T cells follicular helper and T cells gamma delta, whereas native B cells, Mast cells resting and T cells regulatory (Tregs) were positively associated with IGFBP7-AS1 expression (Fig. 6B-I). At the same time, IGFBP7 expression levels were positively correlated with levels of Mast cells resting, Plasma cells, T cells CD4 memory resting and T cells gamma delta, and diversely correlated with memory B cells, Dendritic cells activated, Macrophages M0 and T cells follicular helper (Fig. 6J-Q).

Tumor-infiltrating lymphocytes (TILs) play a key role in sentinel lymph node status and overall survival rate prediction [19]. Immune infiltration scores of TIL types and immune-related substances were calculated in order to determine if the IGFBP7-AS1 and IGFBP7 connected with the immune infiltration. We assessed differences between high- and low- levels in immune infiltration cell and marked differences were shown in aDCs, CCR, CD8 + T cells, Check point, Cytolytic activity, HLA, iDCs, Mast cells, Neutrophils, T helper cells, TIL, Treg and Type II IFN Reponses between IGFBP7-AS1 subgroups. Except DCs, MHC class I, NK cells, Th2 cells and Type I IFN Response, IGFBP7 expressed significantly different in other immune cells (Fig. 6R, S).

IGFBP7-AS1 expression correlated with immune checkpoint genes expression

Then, we analyzed the immune checkpoint genes expression with IGFBP7-AS1 and IGFBP7. We could find in the pictures plotted in Fig. 7A, B, that these immune checkpoints related genes (CD200, NRP1, LAIR1, CD244, CD40LG, CD48, CD28, HAVCR2, TNFSF14, HHLA2, CD70, CD27, TNFRSF4, TNFSF15, TNFRSF9) expressed differently between high- and low-expression groups of both IGFBP7-AS1 and IGFBP7. PD-1 is a protein on the surface of T and B cells that acts as an essential role in regulating the immune system's response to the cells in the human body by down-regulating the immune system and developing self-tolerance by suppression of T cell inflammatory activity [20]. CTLA4 is a protein receptor that works as a member of immune checkpoints and has effect on immune responses downregulation [32]. They were essential immune checkpoints in carcinoma proliferation research and immune checkpoint blockade therapy by targeting PD-1and CTLA4 revealed promising clinical effects. These key modulators were detected to express more accompanied with higher IGFBP7-AS1 levels (Fig. 7C, D).

Fig. 7

Correlations of IGFBP7-AS1 and IGFBP7 with immune checkpoint genes, TMB and m6A genes. A The immune checkpoints related genes expression in different IGFBP7-AS1 and B IGFBP7 expression groups. C The scatter plot showing the correlation between IGFBP7-AS1 and CTLA4 levels. D The scatter plot showing the correlation between IGFBP7-AS1 and PD1 levels. E The correlation scatterplot of TMB with IGFBP7-AS1 expression and F IGFBP7 expression. G The m6A genes expression in different IGFBP7-AS1 and H IGFBP7 expression groups

IGFBP7-AS1 Had a Significant Negative Correlation with TMB

TMB (tumor mutational burden) is a genetic characteristic of tumorous tissue that can provide valuable information for our cancer study and relevant treatment. It is defined as the number of non-inherited mutations per Mb of investigated genomic sequence and plays a key role in cancer prognosis and response to tumor immunotherapy treatment [11]. We calculated the TMB and discovered its connection with IGFBP7-AS1 and IGFBP7. The results suggested that both of IGFBP7-AS1 and IGFBP7 had a significant negative correlation with TMB. The more expression of IGFBP7-AS1 or IGFBP7, the less expression of TMB (Fig. 7E, F).

IGFBP7-AS1 expression correlated with m6A genes expression

m6A(6-methyladenine), a common type of mRNA methylation, impacts on the cancer development for its versatile functions in various physiological processes [38]. The connections between m6A and numerous cancer types have been indicated in many reports involving breast cancer, prostate cancer, stomach cancer and so on. There are some genes (HNRNPC, RBM15, METTL14, YTHDC2, WTAP, YTHDF1, YTHDC1, FTO, YTHDF2) proved to regulate the modification levels of m6A and then affect the cancer cell proliferation. These genes all expressed variously in different IGFBP7-AS1 and IGFBP7 levels groups (Fig. 7G, H). Except for FTO with IGFBP7-AS1, most of the m6A genes were in down-expressed condition with high levels of IGFBP7-AS1 and IGFBP7.

Pan-cancer verification of IGFBP7-AS1 and IGFBP7 levels

Comprehensive omics analysis of IGFBP7-AS1 across 33 cancers was conducted to verify IGFBP7-AS1’s role in diverse cancers and except to provide robust evidence for potential tumor research. We obtained the expression of IGFBP7-AS1 and IGFBP7 across 33 cancer samples and respective normal tissues from TCGA project. IGFBP7-AS1 and IGFBP7 expression differences could be found in BLCA, BRCA, CESC, CHOL, COAD, GBM, HNSC, KICH, KIRC, KIRP, LUAD, LUSC, PRAD, THCA and UCEC. Besides, IGFBP7 levels differed in SARC, STAD and IGFBP7-AS1 differed in LIHC as well (Supplementary Fig. 1A, B).

Next, we identified the prognostic value of IGFBP7 − AS1 for pan-cancer. Patients in the high-expression group survived longer than those in the low-expression group in LAML, LUAD and UCEC and the outcomes of LGG, MESO and STAD were opposite in Kaplan–Meier survival curves (Supplementary Fig. 1C-H). Furthermore, univariate Cox proportional hazard regressions were modeled to understand the altered expression of IGFBP7 − AS1 and IGFBP7 with patient overall survival, and the directions of prognostic effect were manifested varied depending on cancer types (Supplementary Fig. 1I).

We first investigated the association of IGFBP7 − AS1 and IGFBP7 expression with six subtypes of immune infiltration: C1 (wound healing), C2 (INF-r dominant), C3 (inflammation), C4 (lymphopenia dominant), C5 (immunologically quiet), and C6 (TGFβ dominant) [40]. Both IGFBP7 − AS1 and IGFBP7 were strongly correlated with immune subtypes in pan-cancer (Fig. 8A). Next, we evaluated whether a link between IGFBP7 − AS1 and IGFBP7 and expression of genes recognized as checkpoint components existed. Co-expression with immune checkpoint related genes in pan-cancers was exhibited in heatmaps of Fig. 8B, C. High positive correlations could be seen in tumors like CHOL, ESCA, LGG, LIHC, LUAD, LUSC and so on while negative ones were mainly appeared in MESO, LAML and THYM.

Fig. 8

Correlations with immune subtypes, immune checkpoints and TILs. A Correlation between gene expression and immune infiltration subtypes in patients of pan-cancer. B, C Heatmaps about the correlation between IGFBP7-AS1 and IGFBP7 expression levels and acknowledged immune checkpoints’ mRNA expression. D, E Heatmaps about the relationship between IGFBP7-AS1 and IGFBP7 expression and immune infiltration cells

Also, the relationship with immune infiltration cells were analyzed here, and results were presented in Fig. 8D, E, which suggested that IGFBP7 − AS1 and IGFBP7 were strongly associated with immune cells and most of the correlations were positive.

We have done TME analysis of IGFBP7 − AS1 and IGFBP7 in UCEC, so in this section we intended to confirm IGFBP7 − AS1 and IGFBP7’s role in tumor progression and immune response in pan-cancer. Stromal, immune, ESTIMATE score and tumor purity of 33 pan-cancer were shown in Fig. 9A-D. IGFBP7 − AS1 and IGFBP7 expression were positively relevant to stromal score in majority pan-cancers with exceptions of SARC and STAD. In immune score analysis, IGFBP7 − AS1 and IGFBP7 still showed positive correlation with immune score across most of cancers, with the obvious exception of THYM, LAML, and IGFBP7 − AS1 in MESO, OV, SARC, SKCM. Up-level of ESTIMATE score and down-level tumor purity usually came with high levels of IGFBP7 − AS1 and IGFBP7(exceptions were TGCT, LAML, SARC in ESTIMATE score or tumor purity).

Fig. 9

Correlations of IGFBP7-AS1 and IGFBP7 expression with TME, TMB, MSI and tumor stem cells in multiple cancer. A The relationship between gene expression and stromal score, B immune score, C ESTIMATE score and D tumor purity. E, F Correlation between TMB and IGFBP7-AS1 and IGFBP7 expression. G, H Correlation between MSI and IGFBP7-AS1 and IGFBP7 expression. I The relationship between gene expression and RNAss and (J) DNAss

After that, we focused on the association with TMB and MSI across 33 pan-cancer and the correlation was exhibited in radar plot Fig. 9E-H. Most of the cancer including BLCA, BRCA, CESC, COAD, DLBC, HNSC, KIRC, LIHC, LUAD, PAAD, PRAD, SARC, STAD, UCEC, UVM had a significantly negative correlation with IGFBP7 − AS1 in terms of TMB, apart from THYM, which showed reverse result with significance. As for IGFBP7, its expression was significant associated with TMB in 21 out of 33 cancers, and most of the trends were similar with IGFBP7 − AS1. The results were consistent with the TMB analysis in UCEC.

MSI (microsatellite instability) is a good marker for determining a prognosis for cancer treatments, and the elevated MSI may be an indicator of higher tumor risk for the reason that the tumor’s disrupted function increases the gene instability [16]. The MSI radar plots reflected that in COAD, LIHC, STAD, as well as UCEC, IGFBP7 − AS1 was acted as negative correlation with MSI, contrary to the condition in HNSC and MESO. While for IGFBP7, the significant associations with MSI turned out to be all negative in DLBC, GBM, LIHC, LUAD, LUAC, PAAD, STAD and UCEC.

Expression-based RNA stemness score (RNAss) and methylation-based DNA stemness score (DNAss) were used to reflect the features of tumor stem cells [28]. Higher stemness index values are associated with stimulate biological processes in tumor stem cells and promoted tumor dedifferentiation [5]. At the same time, we found IGFBP7 − AS1 and IGFBP7 were negatively correlated with RNAss and DNAss across most of the cancers (Fig. 9I, J), and a decrease in tumor stemness was usually linked with better survival. All these evidences can help to explain the IGFBP7 − AS1 and IGFBP7’s mechanisms in modulating the overall survival.

Using CellMiner database, we assessed the influence of IGFBP7 − AS1 and IGFBP7 on drug sensitivity, and such research could help us to provide better precision treatment for patients. We chose the significantly relevant ones according to P values here. With the increase in the expression of IGFBP7 − AS1, the less sensitive the cells were to chemotherapeutic drugs involving Veliparib, Mocetinostat, CYC-116 and Alisertib (Supplementary Fig. 2).