The Web of Science (WOS) core database from Clarivate Analytics was deemed as the best for bibliometric analysis (Aggarwal et al., 2016; Ke et al., 2020), so it was used as the data source. The WOS core database was searched on May 6, 2023, for all articles related to the relationship between CRC and gut viruses. All relevant publications are collected primarily on the basis of titles (T1) and abstracts (AB), using the following search formula: #1: (AB = ((Rectal Neoplasm*) OR (Rectal Tumor*) OR (Rectal Cancer*) OR (Rectum Neoplasm*) OR (Rectum Cancer*) OR (Cancer of the Rectum) OR (Cancer of Rectum) OR (Colorectal Neoplasm*) OR (Colorectal Tumor*) OR (Colorectal Cancer*) OR (Colorectal Carcinoma*) OR (Colonic Neoplasm*) OR (Colon Neoplasm*) OR (Cancer of Colon) OR (Colon Cancer*) OR (Cancer of the Colon) OR (Colonic Cancer*))) OR (TI = ((Rectal Neoplasm*) OR (Rectal Tumor*) OR (Rectal Cancer*) OR (Rectum Neoplasm*) OR (Rectum Cancer*) OR (Cancer of the Rectum) OR (Cancer of Rectum) OR (Colorectal Neoplasm*) OR (Colorectal Tumor*) OR (Colorectal Cancer*) OR (Colorectal Carcinoma*) OR (Colonic Neoplasm*) OR (Colon Neoplasm*) OR (Cancer of Colon) OR (Colon Cancer*) OR (Cancer of the Colon) OR (Colonic Cancer*))); #2: (AB = (virus* OR virology* OR HPV OR EBV OR CMV OR JCV)) OR (TI = (virus* OR virology* OR HPV OR EBV OR CMV OR JCV)); Final dataset: #1 AND #2. The search articles were published in English from 2001 (1 January 2001) to 2023 (5 May 2023).

The selection criteria and literature selection process for this study are shown in Figure 1. Briefly, search formulas for a preliminary search were entered, followed by a review of the publications identified in the initial search, with the following inclusion criteria: (1) Written in English; (2) The types of literature included types are articles, but not letter, comments, reviews, or conference abstract; (3) The publications were from the WoSCC Citation Index Expanded (SCI-E) and Social Sciences Citation Index (SSCI) databases; (4) The search time span was from 2001 (1 January 2001) to 2023 (5 May 2023); (5) The manuscript was based on the theme of the relationship between CRC and gut viruses; and (6) To avoid bias caused by daily database updates, relevant literatures were searched and screened on the same day.

CiteSpace (6.2.R1) software, a literature visualization application tool created by Prof. Chaomei Chen, is a citation visualization analysis software that is gradually developed under the background of scientometrics and data visualization. Its main function is to convert a large amount of literature data into a visual map.

The parameters of CiteSpace are set as follows: the time slice is chosen from 2001 to 2023, the year of each slice is 1, and the criteria are chosen (g-index, g² ≤ k Σ_i ≤ gc_i,k∈ Z⁺,k = 25). The included literature was visualized and analyzed for country/region, institution, reference, and keywords, respectively.

In the generated plots, N denotes the number of network nodes, E denotes the number of connected lines, Density denotes the network density, and Modularity is the evaluation index of network modularity. The larger the Modularity Q value is, the better the clustering obtained by the network. Modularity Q value >0.3 means that the delineated clustering structure is significant. Silhouette value was used to measure the homogeneity of the network. Closer to 1 reflects the higher homogeneity of the network, and above 0.5 means that the clustering structure has reasonableness.

The main measurements are as follows. (1) The analysis of cooperation network among countries/regions, institutes of journals and (2) Cited literature analysis mainly included network diagram, timeline diagram, and references burst. Co-citation means that two (or more literatures) are cited by one or more subsequent literatures at the same time, and the two literatures are said to constitute a co-citation relationship, which is a research method used to measure the degree of relationship between literatures; (3) Keyword analysis mainly included keyword cluster analysis, keyword time zone map and keyword burst analysis. Keyword cluster plots focused on reflecting the structural characteristics between clusters and highlighting their key nodes and important connections. The keyword time zone diagram focused on showing the evolution of high-frequency keywords in terms of the temporal dimension (Zhang et al., 2021). The keyword burst analysis allowed for the exploration of rapidly growing topics in the field (Chen et al., 2014; Luo et al., 2021). A flowchart of the study design is shown in Figure 1.

The WOS database literature was analyzed, and the CiteSpace software was run with countries as nodes, and the top 3 countries in 57 countries were the United States (107 articles), China (78 articles), and Germany (33 articles). Through data analysis, it was found that China and the United States can be said to be the most core countries in the network, but it can also be seen that the international community’s exchange of relevant research on the relationship between gut virus and CRC is relatively distant. Germany primarily collaborates closely with the United Kingdom, Scotland, Turkey, and Switzerland; China primarily collaborates closely with Australia and Singapore (Figure 2A); The United States collaborates closely with Argentina, Netherlands, South Korea, and Canada.

The Research Institute Collaboration and Co-Emergence Network displays the capacity for interdisciplinary research among research institutions and the collaboration between research institutions. A total of 398 institutions published relevant literature, including the German Helmholtz Association (10 articles), the University of California (9 articles), the Federal Higher Education System of Pennsylvania (8 articles), the University of California, San Diego (7 articles), and Baylor University Medical Center (7 articles) published in the United States (Figure 2B). The most prolific institutions are mainly from the United States and Germany.

Among the 356 studies on microbiota and CRC, the visual map generated with “reference” as the node can obtain 728 nodes and 1,721 connected literatures in the network (Figure 3A). Log-likelihood ratio (LLR) can be used to cluster the cited literature and generate 12 groups of cluster labels (#0–11), respectively: “jc virus” (#0), “human papillomavirus” (#1), “cancer phenotype” (#2), “epigenetics” (#3), “hepatitis b virus” (#4), “inflammatory bowel disease” (#5), “bacteriophages” (#6), “gut microbiome” (#7), “crc biomarkers” (#8), “nested-pcr” (#9), “colon neoplasm” (#10) and “APC” (#11). The modularity Q is 0.9319, which indicates that the clustering of the network is reasonable; the silhouette is 0.9773, which indicates that the clustering has good homogeneity.

Based on the co-cited literature, the noun terms in the keywords were extracted, the clustering was annotated, and the timeline diagram of the co-cited literature was obtained (Figure 3B). The distribution from left to right in the center of the node in the Figure represents the year, in which the cited literature was first published and reflected the temporal characteristics and evolution trend of the cited literature (Wang et al., 2021). It can be found that there are some clusters that are currently active, such as “human papillomavirus” (#1), “hepatitis b virus” (#4), “inflammatory bowel disease” (#5), “bacteriophages” (#6), and “gut microbiome” (#7), which represent research hotspots in this field.

In addition, as shown in Figure 3C, the top 25 literatures with the strongest citation bursts based on citation burst values were identified (See details in Table 1). The 25 studies mainly examined the association of specific viruses, including HPV, CMV, and JCV, with CRC and its oncogene mutations, as well as two statistical reports on CRC morbidity and mortality. The first citation burst occurred in 2002 and was published by Ricciardiello et al. (2001). It evaluated the difference in TCR of 8 pairs of JCV isolated from colon cancer tissue and non-tumor epithelium. It was found that the only JCV strain present in the human colon was Mad-1, and only a variant with a single 98 bp sequence was found in cancerous tissue, which may be involved in the development of chromosomal instability (Ricciardiello et al., 2001). The strongest burst in 2010 came from a paper published in 2007 by Damin et al. (2007) which indicated that HPV DNA was detected in colorectal specimens from 60 rectal cancer patients (83.3%), while HPV DNA (p < 0.001) was not detected in tissues of non-malignant control groups, and HPV16 was found to be the most common type of virus. The second most frequently cited article is a large international study published in Cancer Cause Control by Gornick et al. that examined 37 human papillomavirus (HPV) in the DNA of paraffin-embedded or frozen tissues in three different populations of CRC patients (279 cases) and normal paracancerous tissues (30 cases) in the United States (73), Israel (106) and Spain (100 cases). All samples were negative for all types of HPV using both the GP5+/GP6+ PCR reverse line blot method and the SPF10 INNO-LiPA method (Gornick et al., 2010). The 3 papers published in 2018 continued to explode, which means that they have received a lot of attention in the field recently. The most recent burst occurred in 2021 and has been continued for 2 years now. There are mainly three literatures as follows. Nakatsu et al. (2018) found a significant increase in the diversity of intestinal bacteriophage communities in CRC patients and altered interactions between bacteriophages and oral bacterial symbionts in fecal samples from CRC patients. In addition to this, Hannigan et al. (2018) proposed that bacteriophage communities are associated with CRC and may influence cancer progression by altering bacterial host communities. It is not difficult to notice that in recent years, the field has paid more attention to the role of bacteriophages in the occurrence and development of CRC.

On the basis of keyword co-occurrence, the keywords were clustered, and a total of 14 clusters were collected, namely “human cytomegalovirus” (#0), “simian virus” (#1), “antibody” (#2), “gut microbiome” (#3), “constitutive activation” (#4), “case–control study” (#5), “agnoprotein” (#6), “kidney transplantation” (#7), “human papillomavirus” (#8), “liver metastasis” (#9), “inflammatory bowel disease” (#10), “drug resistance” (#11), “hsv” (#12), and “jc virus” (#13). From Figure 4A, there are many connections between nodes, which proves that keywords in this field are highly co-presentable. The label numbers in the keyword cluster are inversely proportional to the size of the cluster, and the largest cluster label is human cytomegalovirus (#0), which indicates that the relationship between human cytomegalovirus and CRC has been studied in depth.

As shown in Figure 4B, the research on the relationship between gut virus and CRC in the past decade gradually focused on the study of virus interaction and molecular mechanism. In recent years, herpes simplex virus, hepatitis b virus, t-cells, and the gut microbiome began to appear.

Based on the software’s analysis, a number of keywords with burst have been identified and the top 25 keywords for these burst were selected (Figure 4C), including cytosine deaminase, EBV, microsatellite instability, JCV, progressive multifocal leukoencephalopathy, DNA, DNA sequences, identification, receptor, risk, herpes simplex virus, T cells, Virus, replication, immunotherapy, oncolytic vaccinia virus, oncolytic virus, apoptosis, inflammatory bowel disease, gene therapy, HBV, pancreatic cancer, gastric cancer, gut microbiome, and protein.

This paper summarized the prevalence of HPV, CMV, JCV and other specific virus populations in CRC and how they affect the cellular and microbial microenvironment, which presumably contributes to the promotion of CRC. HPV is a group of circular double-stranded DNA viruses with epitheliophilic properties (Dunne and Park, 2013). Bodaghi et al. (2005) examined cancerous and para-cancerous tissues from 107 CRC samples and found that 36% (38/107) of the samples were infected with HPV16, the most common type of infection. Malekpour Afshar et al. (2018) detected positive HPV DNA in 22.6% (19/84) of CRC samples, with the most prevalent types being HPV types 51 and 56. The study by Giuliani et al. (2008) suggested that HPV and three polyomaviruses, namely SV40, BKV and JCV, were rarely detected simultaneously in the same sample. In addition, there was a correlation between HPV infection and certain clinical features. HPV16 infection was more common in stage I patients than in advanced patients (stages II, III, and IV) (Chen et al., 2012). HPV infection in CRC was different in left site (40%) and right site (20%) (Ambrosio et al., 2023). However, there is heterogeneity in the outcome of HPV infection in CRC. A case–control study by Burnett-Hartman et al. (2011) included a total of 609 colorectal tissue samples from individuals with colorectal adenomas (n = 167), hyperplastic polyps (n = 87), and polypo-free controls (n = 250), none of which had HPV DNA via PCR detection. A meta-analysis revealed that HPV prevalence was 11.2% (294/2630) in adenocarcinomas and 5.1% (21/415) in adenomas in 37 studies (Baandrup et al., 2014). HPV prevalence varies by geographic region, with the highest rates in South America, Asia, and the Middle East (Baandrup et al., 2014).

The possible role of HPV in the development of CRC has been extensively explored. A retrospective study by Lorenzon et al. (2011) found that HPV infection was associated with c-myc amplification, k-ras mutations, and p53 polymorphisms or mutations as potential mechanisms of CRC development. Chen T. H. et al. (2012) discovered that E6 protein was more expressed in HPV16 DNA-positive tumors, and E6 protein may reduce the expression levels of p21 and Mdm2 by inhibiting the transcriptional activity of p53. HPV-related molecular targets may be important mediators of HPV promoting carcinogenesis. The immune landscape and tumor flora of HPV-associated CRC deserve to be deeply explored. According to Ambrosio et al. (2023), HPV-positive tumors exhibited higher abundances and species types and lower abundances in specific genera, the most relevant of which was Bacteroides. HPV-negative CRC samples were found in 3 families (Caulobacteraceae, Campylobacteraceae, and Xanthomonadaceae) and 7 genera (such as Achromobacter, Bacteroides, and Brevundimonas), which illustrated that bacteria may influence the interaction between HPV and CRC. Meanwhile, in HPV-associated cancers, tumor environmental immune surveillance is reduced, but lymphocyte cytotoxic responses are enhanced (Ambrosio et al., 2023).

As a herpesvirus group DNA virus, HCMV usually infects humans and sets up a lifelong infection. Most infected people have no clinical symptoms, but under certain conditions, multiple organs and systems can cause serious disease (Yu et al., 2023). Nagel et al. (2023) assessed a CRC incidence of 1.159% (165 cases) in the HCMV group and 2.845% (405 cases) in the control group. HCMV infection was found to be associated with a statistically significant reduction in CRC incidence. Through quantitative and histological analysis, Chen H. P. et al. (2012) found that the detection rate of HCMV DNA in tumor tissues (69/163) was significantly higher than that in neighboring non-tumor tissues (14/163), and the virus copy number of HCMV in tumor tissues was also significantly higher than that in non-tumor tissues. Moreover, HCMV was more likely to infect CRC epithelial cells (Chen and Chan, 2014). HCMV that has cell transformation and possible carcinogenic effects affects a variety of cellular functions, including regulating angiogenesis, promoting cell proliferation, and inhibiting signaling pathways for apoptosis and anticancer immunity (Yu et al., 2023). For example, HCMV is one of the inducers of COX-2, and the overexpression of COX-2 in CRC tissues is associated with poor prognosis (Zhang and Sun, 2002). HCMV-infected cells directly induce angiogenesis by secreting VEGF and other angiogenic factors (Caposio et al., 2011). HCMV infection induces high levels of cyclin, phosphorylated Rb and p53, thus leading to cell cycle arrest (Jault et al., 1995). Therefore, the virus may play a regulatory role in the development of CRC.

JCV, a member of the polyomavirus family, is a double-stranded cyclic DNA virus. JCV infection is very common in both the upper and lower digestive tracts of humans in immunocompetent individuals (Ricciardiello et al., 2000). Jung et al. (2008) explored the role of JCV in precancerous lesions (adenomas) and its relationship with microsatellite instability (MSI). In 61 adenomas carrying JCV sequence, 8% (5/61) were MSI-H, and in 12 adenomatous polyps expressing T-Ag, 8% (1/12) were MSI-H. Furthermore, T-Ag is specifically expressed in the nuclei of these precancerous lesions. Goel et al. (2006) found the JCV T-Ag (JCVT) DNA sequence in 77% (77/100) of CRCs, and the expression of T-Ag in CRC was significantly correlated with genomic instability and methylation of multiple gene promoters, which suggested that JCV may be involved in the occurrence of CRC through multiple mechanisms of genetic and epigenetic instability. Nosho et al. (2009) detected 35% (271/766) JCVT expression in CRCs and found that JCVT was independently associated with p53 expression and chromosomal instability (CIN) in colon cancer through multivariate analysis, but not significantly associated with CPG island methylation phenotype (CIMP). The Wnt/β-Catenin signaling pathway is closely related to the development of CRC (Wong and Pignatelli, 2002). Based on Enam et al. (2002), the T-Ag and the late helper protein Agnoprotein were expressed in CRC. β-catenin appeared in the nucleus and coexisted with T-Ag. β-catenin and T-Ag had a synergistic role in inducing c-myc promoter transcription, which demonstrated that T-Ag may regulate the Wnt signaling pathway through its interaction with β-catenin. Therefore, DNA viruses such as CMV, HPV, and JCV can alter the Wnt/β-catenin pathway in a variety of ways and further affect pathways associated with CRC (Marongiu and Allgayer, 2022).

Besides, phages can regulate gut flora and the immune system (Shuwen and Kefeng, 2022). Single-stranded DNA containing M13 phages is highly immunogenic and can specifically target tumor cell surfaces. CEA specific M13 phage induced tumor regression in a mouse model by specifically binding to CRC cells and activating immune cells and anti-tumor immune responses (Murgas et al., 2018). Combination therapy with M13 phage targeting Fusobacterium nucleatum (Fn) and silver nanoparticles can effectively inhibit the development of CRC by remodeling Fn, reducing immunosuppressive cells (MDSCs), and activating antigen-presenting cells (APCs) to remodel the tumor immune microenvironment (Dong et al., 2020). However, it was difficult to establish a single causal link between specific viral infections and CRC, because the gut microbiome is a highly complex ecosystem involving multiple interactions of bacteria, viruses, bacteriophages, and other microorganisms.