Introduction
Hepatocellular carcinoma is the sixth most common malignancy and the third leading cause of cancer-related deaths worldwide, mainly due to liver cirrhosis caused by hepatitis virus, alcohol and fat accumulation [
1,
2]. As a cancer with a progressive disease course, without effective early intervention, the disease can cause an irreversible damage and extremely poor prognosis [
3,
4]. Despite the great progress achieved in various treatments, such as surgery, chemoradiotherapy, immunotherapy and targeted therapy, in the past few decades, the prognosis of HCC patients is still not optimistic due to the high metastasis rate and recurrence rate of this tumour type [
5‐
7]. Therefore, it is urgent to explore the molecular mechanisms that occur during the pathogenesis and progression of HCC.
The interior of tumours was originally thought to be sterile, especially in solid tumours. However, recent evidence suggests that intratumoral microbes form an important component of the tumour microenvironment (TME) and that these microbes are intricately involved in tumorigenesis, progression, and sensitivity to therapy in the local environment [
8]. Deborah Nejman comprehensively identified the presence of abundant intratumoral bacteria in breast, lung, ovarian, pancreatic, melanoma, bone, and brain tumours. The results revealed a tumour-specific microbial composition, and the metabolic pathways and clinical features of these microorganisms were closely related [
9]. At present, it is believed that intratumoral microorganisms affect tumour progression mainly by causing DNA damage, the activation of oncogenic pathways, and the regulation of the immune system in the microenvironment [
10,
11]. The above results show that investigating the importance of intratumoral bacteria is an emerging field, attracting the interest of researchers for its potential role as an intervention target in tumour diagnosis and therapy. It is reasonable to speculate that microorganisms within HCC can also participate in the progression and metastasis of HCC through the above pathways. However, few relevant studies have explored the tumour microbiome in HCC. Therefore, it is necessary and meaningful to investigate the presence, abundance and functions of intratumoral microorganisms in this cancer type.
In this study, we first performed fluorescence in situ hybridization (FISH) using HCC tissues to verify the presence of bacteria in HCC, followed by MiSeq using 99 HCC tissues and adjacent tissues, to comprehensively analyse microbial infiltration and changes in the metabolic pathways that occur in HCC. Some valuable conclusions were revealed. For instance, microbes in tumours may support tumour cell proliferation and invasion through increased fatty acid and lipid synthesis. In addition, certain microorganisms may be involved in the process of HCC in unique ways. Taken together, the results suggest that intratumoral microbes interact and complement tumour cells and the TME. We believe that our research supports further developments in this field.
Methods
Sample collection
We collected 99 HCC and paracancerous tissue samples from patients at the First Affiliated Hospital of Zhengzhou University (Zhengzhou, Henan, China). All studies using these samples were approved by the Ethics Committee of the First Affiliated Hospital of Zhengzhou University. The average age of patients included in the study was 54.19. 71.4% patients had HBV infection. 77.3% of the patients were single tumor, and 79.1% of the patients were classified as Child A. More detailed information was shown in Table
1. All tumor tissues and adjacent tissues are obtained by surgical resection.
Table 1
Clinical characteristics of patients included in this study
n | 44 | Grade = G3 (%) | 7 (22.6) |
Age (mean (SD)) | 54.19 (9.02) | BCLC (%) | |
HBV = YES (%) | 30 (71.4) | A | 2 (5.0) |
Child (%) | | A1 | 19 (47.5) |
ChildA | 34 (79.1) | A2 | 2 (5.0) |
ChildB | 6 (14.0) | A4 | 3 (7.5) |
ChildC | 3 (7.0) | B | 11 (27.5) |
Tumor number = single (%) | 34 (77.3) | D | 3 (7.5) |
T stage (%) | | Stage (%) | |
T1 | 33 (75.0) | I | 32 (72.7) |
T2 | 8 (18.2) | Ib | 2 (4.5) |
T3 | 2 (4.5) | II | 9 (20.5) |
TI | 1 (2.3) | IIIA | 1 (2.3) |
CA125-B (mean (SD)) | 53.58 (161.87) | CEA-B (mean (SD)) | 2.76 (2.61) |
Fluorescence in situ hybridization
Tissues were prepared into paraffin sections, added to prehybridization solution, and incubated at 37 °C for 1 h. Then, hybridization solution containing the EUB338 probe (Servicebio, G3016-3) was added, and samples were incubated at 42 °C overnight in an incubator and photographed under a microscope (NIKON DS-U3).
The probe sequence used was EUB338:5'-CY3-GCT GCC TCC CGT AGG AGT-3'.
Bacterial DNA extraction and sequencing
All of the samples were subjected to the same procedures for DNA extraction and PCR amplification by the same laboratory staff. Each sample was suspended in 790 μL of sterile lysis buffer (4 M guanidine thiocyanate; 10% n-lauroyl sarcosine; 5% n-lauroyl sarcosine-0.1 M phosphate buffer [pH 8.0]) in a 2-mL screw-cap tube containing 1 g glass beads (0.1 mm BioSpec Products, Inc., USA). This mixture was vortexed vigorously and incubated at 70 °C for 1 h. After incubation by bead beating for 10 min at maximum speed, DNA was extracted using the manufacturer’s instructions for bacterial DNA extraction using the E.Z.N.A.®Stool DNA Kit (Omega Biotek, Inc., GA), with the exception of lysis steps, and the product was stored at − 20 °C until further analysis. The extracted DNA obtained from each sample was used as the template to amplify the V3 ~ V4 region of 16S rRNA genes.
The primers F1 and R2 (5′-CCTACGGGNGGCWGCAG-3′ and 5′-GACTACHVGGGTATCTAATCC-3′) corresponding to positions 341 to 805 in the Escherichia coli 16S rRNA gene were used to amplify the V3 ~ V4 region of each sample by PCR. The PCR experiments were run in an EasyCycler 96 PCR system (Analytik Jena Corp., AG, Germany) using the following program: 3 min of denaturation at 95 ℃ followed by 21 cycles of 0.5 min at 94 ℃ (denaturation), 0.5 min of annealing at 58 ℃, and 0.5 min at 72 ℃ (elongation), with a final extension at 72 ℃ for 5 min. The products from different samples were indexed and mixed at equal ratios for sequencing using the MiSeq platform (Illumina Inc., USA) according to the manufacturer’s instructions.
Sequencing data processing and OTU (Operational Taxonomic Unit) cluster annotation
Paired-end sequence data were obtained based on MiSeq sequencing. According to the complementary region (overlap) between the PE (paired end) reads, the paired reads were merged into a single sequence. Quality control filtering was performed on data regarding the quality of reads and the effect of merging. Samples were distinguished according to the index sequences and primer sequences at both ends of the sequence to obtain high-quality effective sequences and to correct the sequence orientation.
OTU (operational taxonomic unit) is a label artificially set for a taxonomic unit (strain, genus or species) to facilitate analysis in population genetics research [
12]. To identify the number of species, genera and other information in the sequencing results obtained from a sample, after removing the single sequences without repeats, we classified the sequences into taxonomic units, namely, OTUs, based on a similarity value of 97%. Chimeric sequences were removed during the clustering process to obtain the representative sequence of an OTU. Subsequently, by aligning the 16S bacterial and archaeal ribosome databases Silva3 (Release 138
http://www.arb-silva.de) [
13], we performed OTU species annotation based on the QIIME platform (
http://qiime.org/scripts/assign_taxonomy.html). The RDP classifier Bayesian algorithm was used to perform taxonomic analysis on the representative sequences of OTUs with a similarity level of 97%, and the community composition of each sample was counted at each classification level (phylum, class, order, family, genus, and species) [
14].
Alpha diversity was defined by the Chao, Ace, Shannon, and Simpson indices, which were calculated using mothur (version v.1.42.1,
http://www.mothur.org) [
15]. Beta diversity was assessed by unweighted and weighted UniFrac distance matrices and visualized by principal coordinate analysis (PCoA). QIIME software was used to calculate the beta diversity distance matrix, and the R software ‘vegan’ package was employed for PCoA analysis and visualization.
LEfSe is a data analysis method based on linear discriminant analysis (LDA) effect size [
16]. Specifically, we first used the nonparametric factorial Kruskal‒Wallis (KW) sum-rank test to establish differential microbiota. Next, linear discriminant analysis (LDA) was applied to estimate the magnitude of the role of species abundance in the differential effect. The algorithm emphasizes statistical significance and biological relevance. LEfSe analysis was performed with a web-based tool (
http://huttenhower.sph.harvard.edu). In this study, LDA > 2.5 was considered statistically and biologically significant. Subsequently, using Qiime software (
http://qiime.org), a random forest algorithm was used to identify OTUs with significant differences between groups, with 1000 trees for modelling and fivefold cross-validation to estimate the size of generalization error. The default settings were used for the rest of the parameters. All biochemical index data were obtained using the last blood sample collected from the patient before surgery. Any correlations with the abundance of microorganisms were established by Spearman’s method. Due to the small number of patients with Child‒Pugh scores B and C, we pooled the two groups.
Discussion
Interactions between microbes and the human body have been well documented; microbes can affect a variety of physiological functions, including metabolism and the immune system [
17]. These interactions are also observed in cancer. Martel et al. suggested that at least 20% of cancers are influenced by the microbiota [
18]. Therapeutic responsiveness (e.g., to immunotherapy and chemotherapy) has been observed to depend on the gut microecosystem in animal experiments and cancer patients [
19‐
23]. Previous studies have focused on the interaction between gut microbes and specific diseases, and although multiple noninvasive biomarkers of particular diseases have been developed, due to the limitation that the microbiota play indirect roles in disease, the precise identification of the causal relationship between gut microbes and diseases, including cancer, remains a challenge [
24,
25]. Intratumoral microbes play a more direct role on a more local scale than gut microbes to influence tumour progression [
8]. Such microbes have been shown to interact with tumour cells in diseases such as lung, colorectal and skin cancers [
11,
26‐
28]. Mechanistically, intratumoral microbes act through three principal mechanisms: (1) their metabolites regulate oncogenes or oncogenic pathways; (2) they promote DNA damage and gene mutation; and (3) they regulate immune responses in the microenvironment [
8,
10,
11]. Overall, these effects inhibit the antitumour response. Nejman et al. demonstrated that intratumoral microbiota are tumour specific, which may imply that the metabolic pathways in which microorganisms are involved in different microenvironments can differentially affect tumour cells [
9]. Therefore, to further clarify the mechanism underlying HCC progression, it is necessary to explore the intratumoral microbes associated with HCC.
In this study, we assessed the microenvironment of HCC and confirmed the presence of bacteria in immune cells using FISH technology. Specifically, we performed MiSeq sequencing using 99 HCC and paracancerous tissues based on 16S rRNA sequence technology, which yielded a comprehensive map of the microbes within HCC and paracancerous tissues. The results showed that the microbial community diversity, including alpha and beta diversity, was significantly higher in HCC tissues than in paracancerous tissues. However, the dominant species of the microflora (Proteobacteria, Firmicutes, Actinobacteria, etc.) did not differ between the two groups, but their levels were slightly different. We conjectured that certain microorganisms with low relative abundance may contribute more to the difference between the two groups. The significant elevations in
Enterobacteriaceae,
Neisseria and
Fusobacterium abundance in tumour tissues caught our attention. A high abundance of
Enterobacteriaceae is often associated with higher levels of inflammation, which may be related to the ability of this type of microorganism to utilize inflammatory byproducts (such as nitrate) in the microenvironment as energy sources, which is an ability that is not found in competing bacteria [
29‐
31]. Moreover, virulence factors secreted by
E. coli, such as cytolethal distending toxin (CDT), further aggravate the inflammatory response and directly induce DNA damage [
32]. Considering that inflammation is a recognized risk factor for cancer, this may be a potential mechanism through which Enterobacteriaceae is involved in the progression of HCC [
33]. Similarly,
Fusobacterium plays a role in proinflammatory processes [
34]. Moreover, a unique ability of
Fusobacterium may be more involved in the progression of HCC: the bacterium can shuttle noninvasive bacteria into the cytoplasm of host cells [
35]. Considering the direct contact that occurs between intratumoral microbes and tumour cells in the TME, this ability could theoretically have a greater influence. This indicates that even bacteria with levels that are not significantly different between HCC and paracancerous tissue can be involved in tumour progression by a mechanism that remains to be elucidated; uncovering this potential mechanism is obviously a great challenge to the pursuit of a full understanding of the interaction between intratumoral microbes and tumours. In addition, the increased abundance of certain tumour-promoting bacteria, such as
Neisseria, and the low infiltration of antitumour bacteria, such as
Pseudomonas, may further influence HCC progression [
36,
37]. The research conducted by Rolandas Gedgaudas et al. proved that intestinal permeability in patients with portal hypertension was significantly higher than that in the healthy control group, which was accompanied by a high abundance of some bacteria in peripheral blood [
38]. These bacteria mainly included Enterobacteriaceae, Shigella and many other kinds of bacteria, which is consistent with the high abundance of bacteria in HCC tissues observed in our study. This finding indicates that the progression of portal hypertension is accompanied by bacterial translocation, which can partly explain the source of intratumoral bacteria in HCC tissue.
Another interesting phenomenon is that functional enrichment analysis showed that the metabolic pathway in which microorganisms associated with HCC were involved in featured significant enhancement of fatty acid and lipid biosynthesis. It has been confirmed that changes in lipid metabolism occur in rapidly proliferating cancer cells [
39]. Cancer cells transfer more carbon to fatty acids for membrane and signalling molecule biosynthesis than normal cells to maintain rapid cell growth [
40]. Previous studies have shown that tumour cells tend to synthesize fatty acids de novo [
41]. However, the entry of pyruvate into the TCA cycle is inhibited due to the hypoxic TME, and the resulting reduction in fatty acid synthesis may be compensated by the increased uptake of exogenous lipids [
42,
43]. Based on our research, we speculate that microbial metabolites in the TME may provide another plausible source of fatty acids and lipids for cancer cells, which in turn promotes their proliferation and invasion.
Even within the HCC group, different clinical features led to differential microbial communities. For example, the abundance of Veillonella and Alloprevotella was significantly increased in the TME of HBV-related HCC tissues. Higher Child‒Pugh scores tended to favour the high abundance of Sphingosineum and Erysipelas. Based on the current data, we are not able to precisely interpret the relationship between changes in the abundance of specific bacteria and clinical features. Certainly, these microbes affect changes in the clinical characteristics of patients in their own unique ways. Collectively, the microbial community associated with HCC causally interacts with the unique TME, influencing the progression of HCC through multiple mechanisms. While we cannot accurately assess the relationship between the two due to the lack of related research, we believe that our findings provide a basis and guidance for further exploration in this field.
Publisher's Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.