Analysis of complete and incomplete response to neoadjuvant therapy in rectum cancer patients

Abstract

Aim:

Bioinformatic analysis of the differentially expressed genes (DEGs) of the two groups of patients is the main aim of this project.

Background:

Neoadjuvant therapy is a method to treatment disease before local surgery. This method has been administrated to treatment rectum cancer; complete and incomplete responses are observed in patients.

Methods:

The DEGs of gene expression profiles of patients with complete and incomplete response to treatment with neoadjuvant therapy were explored from recorded data in Gene Expression Omnibus (GEO) database and evaluated via protein-protein interaction (PPI) analysis and gene ontology assessment.

Results:

Among 606 significant DEGs, EGFR, JUN, CDK4, IGF1, and APP were pointed out as the key genes that are involved in response model to neoadjuvant therapy while “GnRH signaling pathway” and “regulation of multicellular organism growth” were introduced as the major involved groups of biological terms.

Conclusion:

It can be concluded that, there is genetic variation between rectum cancer patients and efficacy enhancement of treatment need considering precision medicine perspective.

Introduction

Neoadjuvant therapy initially demonstrated as a systemic therapy before local treatment. This method is developed and extended to radiotherapy before surgery. Evidences indicate that, neoadjuvant therapy plays significant role in treatment of various types pf cancers (1). Several advantages for neoadjuvant therapy versus breast cancer are documented; reduction of size of tumor before surgery in response patients, decrease of axillary surgery, and possible observing efficacy of treatment (2). Investigation indicates that, higher rate of R0 resections is observed in patients with esophageal cancer which were characterized with pathologic complete response to neoadjuvant therapy. Recurrences are decreased and 5-year overall survival and disease-free survival are improved in these treated patients. R0 resection refers to absence of tumor after surgery (3, 4). Rectal cancer treatment via neoadjuvant therapy demonstrated complete response of considerable numbers of patients to treatment. However, it is reported that up-regulation of L1CAM and IGF2 is associated with lessened response to neoadjuvant therapy (5).

Since treatment is accompanied with gene expression change in several genes, study of the differentially expressed genes can improve our understanding of molecular mechanism of the applied therapeutic methods (6, 7). Study of radiotherapy resistance in rectum adenocarcinoma showed, many genes including three hubs (PLAGL2, ZNF337 and ALG10) have been dysregulated. Sensitivity of chemotherapeutic drugs, expression of various disease-related genes, and tumor immune infiltration are attributed to these three hub genes (8). A meta-analysis study demonstrated that, there is significant association between down-regulation of miR-451 and development of colorectal cancer (9).

PPI network analysis as a bioinformatic tool attracted attention of researcher to explore molecular mechanism of many diseases and therapeutic methods. In this method, among many individuals, a few numbers of critical genes or proteins highlight and introduce as the main player which effect treatment or progress of diseases (10, 11). In the present study, gene expression profiles of patients with complete and incomplete response to treatment are extracted from GEO database and evaluated via PPI network analysis to find the major genes which are involved in response to neoadjuvant therapy.

Methods

Data extraction and pre-evaluation analysis

Neoadjuvant therapy and rectal cancer were searched in GEO database and the RNA-seq series analysis (GSE242786) was explored. The related data are published by Chong Z et al (12). The gene expression profiles of patients with complete response to treatment were compared with the individuals of patients with incomplete response via GEO2R program. Uniform Manifold Approximation and Projection (UMAP) plot was applied to confirm comparability of the samples. If the samples be separated from controls in UMAP plot, the identified DEGs can be analyzed via more investigations. The significant DEGs were determined and visualized via Venn diagram.

PPI network analysis

The significant DEGs were assessed via PPI network analysis to find the critical genes. The queried DEGs were included in a network via “Protein query” of STRING database by Cytoscape software v 3.10.1. To maximize interactions between the nodes of PPI network, a confidence score of 0.2 was considered. The main connected component of PPI network was analyzed by “Network Analyzer” plugin of Cytoscape software. Since the important DEGs play central role in the analyzed network, a few numbers of DEGs as hub genes were selected based on a degree value cutoff of mean + 2 (standard deviation). Relationship between degree value and betweenness centrality was visualized via scattered plot. To validate hub nodes, a degree distribution diagram was provided. The top 5% of nodes based on betweenness centrality were identified as bottlenecks. The common hubs and bottlenecks were pointed out as hub-bottleneck nodes.

Gene ontology analysis

The biological terms related to the identified hub-bottlenecks were explored from GO_BiologicalProcess-EBI-UniProt-GOA-ACAP-ARAP_25.05.2022_00h00 and KEGG_25.05.2022 sources via ClueGO plugin of Cytoscape software.

Statistical analysis

An adjusted p-value of <0.05 was applied to find the significant DEGs. A corrected p-value via Bonferroni step down correction method was applied to determine the biological terms.

Results

Pre-evaluation analysis

UMAP plot (Figure 1) indicates that, gene expression profile of one sample of Incomplete Response group is similar to profiles of individuals in Response group. Analyses revealed this profile belongs to GSM770486 gene set (Figure 1). As depicted in Venn diagram (Figure 2), there are 606 significant DEGs which separate the Complete and Incomplete Response groups from each other. After deletion of the uncharacterized individuals, a total number of 576 DEGs were remained for more analysis.

Figure 1.

UMAP plot of gene expression profile analysis of patients with rectal cancer with complete and incomplete response to treatment with neoadjuvant therapy. Up; comparison between the GSM7770474-8 as Complete Response group members and GSM7770479-83 and GSM777085-8, as Incomplete Response group individuals. Down; comparison between the GSM7770474-8 as Complete Response group samples and GSM7770479-83, GSM777085, and GSM7770487-8 as Incomplete Response group patient samples. “nbrs” refers to number of neighbors the gene expression profile is compared with it.

Figure 2.

Venn diagram of comparison between the GSM7770474-8 as Complete Response group samples and GSM7770479-83, GSM777085, and GSM7770487-8 as Incomplete Response group patient samples.

PPI network analysis

576 significant DEGs were included in PPI network. 532 individuals were recognized by STRING database. The PPI network including 4 isolated and a main connected component of 528 was formed. The central part of PPI network is presented in Figure 3. A degree value of 61 was determined as cutoff to introduce 26 hub genes. Relationship between degree value and distribution of degree value are demonstrated in Figures 4 and 5. A total number of 26 genes were identified as bottlenecks. There were 18 common nodes between hubs and bottlenecks which introduced as hub-bottlenecks (Table 1).

Figure 3.

Central part of the main connected component

Figure 4.

Scatter plot for the main connected component. The top point in right and up sides of diagram refers to EGFR that is the top hub and bottleneck node.

Figure 5.

Distribution of degree value related to the elements of the studied PPI network.

Table 1.

List of hub-bottlenecks; BC refers to betweenness centrality

No.	Display Name	Degree	BC
1	EGFR	174	0.117
2	JUN	141	0.053
3	APOE	110	0.040
4	IGF1	103	0.027
5	APP	86	0.027
6	PRKACB	85	0.035
7	CDK4	84	0.027
8	RPLP0	82	0.015
9	HSPA9	79	0.020
10	EGR1	76	0.015
11	RPS14	75	0.010
12	UBA52	73	0.029
13	PHGDH	67	0.015
14	PRDX1	67	0.011
15	TBP	64	0.013
16	EIF4A3	64	0.011
17	CS	63	0.017
18	PHB	62	0.011

Gene ontology assessment

Biological terms related to the 18 hub bottlenecks are shown in Figure 6. As depicted in Figure 6, 11 groups of biological terms are connected to the 11 hub-bottleneck genes including (EGFR, JUN, APOE, IGF1, APP, PRKACB, CDK4, RPLP0, EGR1, EIF4A3, and PHB). Percentage of terms per group of biological terms is shown in Figure 7. “GnRH signaling pathway” and “regulation of multicellular organism growth” are appeared as the main groups of biological terms which were associated to the 7 hub-bottleneck genes (EGFR, JUN, IGF1, APP, PRKACB, CDK4, and EGR1) (see Figures 7 and 8).

Figure 6.

The biological terms related to the 18 hub-bottlenecks.

Figure 7.

%terms per group for gene ontology analysis.

Figure 8.

The major groups of biological terms and the related genes.

Discussion

Our analysis led to introduce seven critical genes and two groups of biological terms that differentiate complete and incomplete response to neoadjuvant therapy in rectal cancer. As depicted in Figure 8, EGR1 and PRKACB are not connected to “regulation of multicellular organism growth” group of biological terms. So, EGFR, JUN, CDK4, IGF1, and APP were pointed out as the key involved genes in model of response to neoadjuvant therapy in rectum cancer. Epidermal growth factor receptor (EGFR) is the top up-regulated hub-bottleneck gene (see table 1). EGFR is a transmembrane glycoprotein that is involved in regulating cellular proliferation, cell survival, and cell differentiation. Up-regulation of EGFR in various types of tumor cells and its role in resistance to chemotherapy and radiation treatment is confirm by researchers (13). Surgery and chemotherapy are pointed out as the first-line interventions in treatment of colorectal cancer. Investigations demonstrated that, EGFR inhibitors such as Cetuximab or specific tyrosine kinase inhibitors can be considered as drugs against colorectal cancer (14). Targeting EGFR in cancers by drugs and natural products as a therapeutic method is highlighted by experts (15, 16).

The second up-regulated key gene is Jun proto-oncogene, AP-1 transcription factor subunit (JUN). It is reported that, c-JUN, JUNB and JUND are members of JUN family genes which are involved in activation of protein-1 transcription factor complexes. Investigations discovered that, activation of protein-1 transcription factor complexes effects essential cellular processes such as cell growth, cell differentiation, and neoplastic transformation. Significant increased level of c-JUN and JUNB in human colorectal adenocarcinomas is documented by researchers (17). As described in result section, apolipoprotein E (APOE) was a hub-bottleneck gene that were connected to the five groups of biological terms (see Table 1 and Figure 6). It is pointed out that, JUN-APOE-LRP1 axis Jun-APOE-(low-density lipoprotein receptor-related protein 1 (LRP1)) axis is involved in tumor metastasis in colorectal cancer (18).

Cyclin dependent kinase 4 (CDK4) is another down-regulated key gene. Since, proliferative abnormalities always are seen as imperfections in cell cycle control, CDK4 in association with CDK6, cyclin D, a cyclin-dependent kinase inhibitor (INK4), and retinoblastoma protein (RB) (CDK4/6–cyclin D–INK4–RB pathway) that is involved in regulation of cell cycle, plays role in cancer promotion. Hyperactivation of CDK4/6 in a number of human cancers is documented (19). Based on literature, inhibition of CDK4 activation leads to cell-cycle arrest and preventing cell growth in colorectal cancer (20).

The fourth key gene is insulin like growth factor 1 (IGF1) which is up-regulated in complete response group versus incomplete response group of patients. based on evidences, IGF1 beside insulin, play role in regulating of metabolism and its dysregulation in colorectal cancer is confirmed (21). There is a document that demonstrated inhibition of colon cancer resistance to regorafenib via linsitinib and aspirin as the IGF1-R antagonists (22).

The last key up-regulated hub-bottleneck gene is amyloid beta precursor protein (APP). Role of APP in tumor cell function such as colon carcinoma cells is investigated and reported by researchers (23). Binding of APP to ligand of EGFR (EGF) is documented by da Rocha JF et al (24).

“GnRH signaling pathway” was appeared as the largest pathway in the present study (see Figure 7 and 8). Human gonadotropin-releasing hormone (GnRH) is the natural ligand of human gonadotropin-releasing hormone receptor (GnRHR). Expression pf GnRH and GnRHR in larynx, liver, pancreas, lymphoma, colon, skin, kidney, brain, and blood and tumor derived from prostate, endometrium, breast, and ovary is reported by researchers (25). As depicted in Figure 8, all associated genes are connected to this group. The second group of biological term is “regulation of multicellular organism growth”. Multicellularity has progressed via cell-to-cell adhesion, communication, cooperation, and specialization between several eukaryotic families to form fungi, plants, and animals; a transition from simple to complex system (26). It seems, response to neoadjuvant therapy is associated with essential cellular functions. Finally, the findings can be considered for experimental examination and validation to apply in medical fields.

Conclusion

In conclusion, five key genes were introduced as the responsible individuals for pattern of response to neoadjuvant therapy in rectal cancer. This finding may refer to genetic variation among human population and need to precision medicine perspective about treatment of diseases. Targeting the introduced genes accompany by neoadjuvant therapy may enhance efficacy of treatment against rectum cancer. This finding can be considered for other diseases via more investigation in the field. Experimental evidences and evaluation of other disorders and diseases are recommended to improve the findings of this project.

Conflict of interests

There is no conflict of interest for authors of this article.