Background
Primary liver cancer is the fourth leading cause of cancer-related mortality worldwide, of which hepatocellular carcinoma (HCC) accounts for 80–90% [
1,
2]. About 800,000 people worldwide die each year from HCC [
3], which is often caused by viral hepatitis B or C, persistent alcohol abuse and nonalcoholic fatty liver disease. Given the global increase in obesity and type 2 diabetes, metabolic syndrome-related nonalcoholic fatty liver disease has become an increasingly risk factor for HCC [
4,
5]. As the early symptoms and features of HCC are not typical, more than 80% of patients with HCC cannot receive curative treatment [
6]. Therefore, the early diagnosis and treatment of HCC still require further exploration.
Previous studies have found that approximately 15% of cancers are attributable to microbial infections [
7]. The intestinal microbiome is the most common factor dominating tumor initiation, development, and therapeutic efficacy [
8‐
11]. Over the years, accumulating evidence has proved that polymorphic intratumoral microbiomes enable cancer hallmark capabilities and have been recognized as emerging mechanisms of tumorigenesis and progression [
12‐
14]. Various tumors, including HCC, were once considered sterile tissues. However, advances in sequencing technology have led to the identification and characterization of the microbial composition in tumor tissues. The classical technologies of 16S rRNA, 18S rRNA and whole genome assessments have been extensively applied to obtain the taxonomic profile of the microbiome. A new method, 2bRAD sequencing for microbiome (2bRAD-M), overcomes the challenge of low microbial biomass or severe DNA degradation in the detected samples [
15]. 2bRAD-M uses the type IIB restriction enzyme to perform qualitative and relative quantitative microbial analysis of the unique tags obtained after enzymatic digestion of the microbial genome. This method allows for the accurate generation of species-level taxonomic signatures for low microbial biomasses in tumor tissues and normal tissues [
16,
17].
Microbial metabolism is the essential characteristic and function of microbes, and most of their effects on the host are related to microbial metabolism. Previous researches have extensively explored the relationship between gut microbiota and metabolism through multi-omics integration analysis [
18‐
20], but clinical data on intratumoral bacteria and the relationship between the intratumoral microbiome and metabolome have not been widely studied.
Unlike gut microbiomes, tumor tissues are often of low biomass, and therefore we assessed bacterial 16S rRNA using the fluorescence in situ hybridization (FISH) method to verify the presence of bacteria in tumor and normal tissues of HCC patients. In addition, we constructed a mouse liver cancer model in situ and analyzed the intratumoral microbial community (2bRAD-M, n = 24) and metabolome (LC–MS, n = 24) from tumor tissues and normal liver tissues obtained from the HCC mouse model. The characterization of tumor microbiome and metabolome might provide new opportunities for developing novel biomarkers and therapeutic targets.
Materials and methods
Histology analysis
A total of 3 pairs of HCC tissues and adjacent normal tissues were collected from the First Affiliated Hospital, Zhejiang University School of Medicine. The study was conducted in accordance with the Declaration of Helsinki, and was approved by the Ethics Committee of the First Affiliated Hospital, Zhejiang University School of Medicine (No. IIT20210168B-R1).
HCC tissues and adjacent liver tissues from HCC patients were paraffin-embedded and sectioned at a 40 µm thickness. Hematoxylin and eosin (H&E) stained the cytoplasm and nucleus with contrasting colors to identify cellular components. For immunohistochemistry (IHC) staining, HCC sections were dewaxed with xylene and hydrated with absolute ethanol. The tissue sections were immersed in citric acid antigen retrieval buffer. Then, the samples were heated for 8 min to boiling, taken off the heat for 8 min, and then returned to a medium–low heat for 7 min. The slices were washed 3 times in phosphate buffer solution (PBS) (pH 7.4) with shaking on a decolorizing shaker for 5 min each time. Then, the tissue sections were immersed in 0.3% H2O2-methanol for 25 min, washed with PBS, and probed with the anti-Salmonella typhimurium lipopolysaccharide (LPS) antibody (Abcam, Cambridge, UK) and a rabbit polyclonal anti-Ki67 antibody (Abcam, Cambridge, UK) at 4 °C overnight. On the second day, the slices were washed with PBS, and horseradish peroxidase-conjugated goat anti-rabbit secondary antibody was added and incubated at room temperature for 1 h. Liver sections were scanned with panoramic MIDI (3DHISTECH, Budapest, Hungary).
16S rRNA fluorescence in situ hybridization (FISH)
FISH was performed to detect bacteria in HCC tissues. The probe used was the universal 16S rRNA probe EUB338 (5’-CY3-GCTGCCTCCCGTAGGAGT-3’), which is used to specifically bind to bacterial 16S rRNA. After precipitating the probes, sample slides were processed with prehybridization, hybridization, and posthybridization washes and DNA counterstaining, and the procedure was performed according to published studies [
13,
21].
Cell culture
The murine HCC cell line Hepa1-6 was obtained from American Type Culture Collection (ATCC, USA). Hepa1-6 cells were cultured in high glucose Dulbecco’s modified Eagle’s medium (Sigma-Aldrich, USA) supplemented with 10% fetal bovine serum (Sigma-Aldrich, USA) and 1% penicillin- streptomycin (Thermo Fisher Scientific, USA). Prior to experiments, cells were maintained in an incubator at 37 °C in a 5% CO2 atmosphere.
Orthotopic HCC mouse model
The animal experiment was approved by the Animal Experimental Ethics Committee of The First Affiliated Hospital, Zhejiang University School of Medicine (project number 20221072). Six-week-old WT C57BL/6 J male mice (n = 12) were obtained from the Experimental Animal Center of Zhejiang Academy of Medical Sciences. All animals were housed under specific pathogen-free conditions at a constant temperature (22 ± 2 °C) with a 12-h daylight/darkness cycle, and they had free access to standard rodent feed and tap water until the end of the experiment.
After a week of adaptive feeding, the mice were anesthetized and placed on a laboratory table in the supine position. After disinfecting the abdomen of the mice, sterilized small forceps and ophthalmic scissors were used to open the mouse abdomen along the middle line of the abdomen to avoid unnecessary bleeding. The left lobe of the liver was removed using sterile swabs. A total of 3 × 105 Hepa1-6 cells in 10 µl Corning Matrigel (Matrigel:PBS = 1:4) were injected into the left lobe of the liver, which was then placed back into the abdominal cavity. The surgical site was sutured and disinfected with iodophor. Continuous monitoring and care were given to mice after surgery. After 2 weeks, the HCC tissues and paired adjacent liver tissues were harvested from mice for further analysis. During the processes of model construction, liver samples handling, transportation and sequencing, strict aseptic procedures were followed to eliminate potential bacterial contamination.
2bRAD sequencing for microbiome (2bRAD-M)
2bRAD-M is a microbial diversity analysis technique based on 2b-RAD technology [
22], which performs qualitative and relative quantitative analysis of microorganisms by unique tags obtained after enzymatic cleavage of microbial genomes by type IIB restriction enzymes. A database containing unique tags of each microorganism (2b-Tag-DB) was used for qualitative analysis [
15,
22]; that is, all microbial species that had unique tags were screened. The 2b-Tag-DB was established again for the quantitative microorganisms, and the relative quantitative analysis was carried out; that is, the microbial species obtained in the previous step were subsequently screened, and the abundance was estimated according to the distribution of unique tags.
Microbial genomes were electronically cleaved using the BcgI restriction enzyme to extract raw reads. The clean reads for each sample were searched separately in the 2bRAD-M database (
http://github.com/shihuang047/2bRAD-M) to obtain microbial annotation information for that sample. The clean reads of each sample were retrieved in the new database using a secondary library built using the genomes of the microorganisms that might be present, and the relative abundance of each microorganism in the sample was calculated using the special formula [
15,
17]. Finally, the samples were annotated and summarized at different classification levels, such as phylum, genus and species. The R package Vegan v2.5.7 was used to perform PCoA ordination.
Liquid chromatography-mass spectrometry (LC–MS) technology [
23] was used in this nontargeted metabolomics research [
24,
25]. The experimental process mainly included metabolite extraction, LC–MS detection and data analysis. Samples were stored at −80 °C after collection until analysis. In brief, HCC tissues and liver tissues were mixed with 300 μL of precooled acetonitrile and 200 mg of ceramic beads. Next, the mix was homogenized and centrifuged (4 °C, 12,000 rpm) for 10 min. The supernatant was centrifuged (4 °C, 12 000 rpm, 10 min) again and then filtered using 0.22 μm syringe filters before analysis. The QExactiveTM HF mass spectrometer was operated in positive and negative polarity mode with a spray voltage of 3.5 kV, sheath gas flow rate of 35 psi, capillary temp of 320 °C, and aux gas flow rate of 10 L/min [
26,
27].
The raw data were first analyzed using Compound Discoverer 3.1 (CD3.1) software. The peaks were simply screened, and were aligned according to retention time deviation and mass deviation for different samples to make the identification more accurate. Subsequently, all the peaks were extracted based on the set ppm, signal-to-noise ratio (S/N), additive ions and other information, and the peak area was quantified. Then, spectroscopic processing and database retrieval were conducted to obtain qualitative and quantitative results of metabolites by comparing high-resolution secondary spectral databases mz Cloud and mz Vault and Mass List primary database retrieval, and then quality control was carried out on the data to ensure the accuracy and reliability of data results. Next, the metabolites were subjected to principal component analysis (PCA) and partial least squares discriminant analysis (PLS-DA). Hierarchical clustering (HCA) and metabolite correlation analysis were used to reveal the relationships between samples and metabolites. The KEGG database was used to identify potential biological pathways [
28].
Statistical analysis
Statistical analyses were performed using SPSS (version 26) and the R software (version 4.0). Comparisons between two groups were calculated by Student's t test. In addition, spearman correlation analysis was used to calculate the relationship between two groups based on relative abundance. P < 0.05 was considered statistically significant.
Discussion
With the development of intratumor microbiome assessment techniques, more and more intratumor microbiome characteristics have been identified in different types of tumors [
8,
30,
31]. In 2020, Nejman et al. detected intratumoral bacteria in seven cancer types and indicated the presence of intratumor bacteria in tumors and immune cells. In addition, the intratumor bacteria were significantly different among the cancer types [
13]. Similar to these results, a negative relationship between the intratumoral bacterial load and lymphocyte infiltration was observed in nasopharyngeal carcinoma. Furthermore, the intratumoral bacterial load was associated with the survival rate, including disease-free survival, overall survival and distant metastasis-free survival [
32]. A recent study showed that intratumor bacteria could markedly inhibit lung metastasis in a breast tumor mouse model [
33]. Taken together, intratumor bacteria could play essential roles in tumor initiation and development, and the potential function and targeted therapy for intratumor bacteria is worth exploring.
Increasing evidence showed that the intratumoral microbiota could play essential roles in HCC progression. Xue et al. assessed the characteristics of bacteria in 47 paired HCC and liver tissues using 16S rRNA sequencing and explored the potential association among differentially expressed genes and metabolites [
14]. Based on the analysis of normal liver, peritumoral, and HCC tissue samples, the features of diversity, structure, and abundance were described, and a prognostic prediction model was built in the HCC cohort [
12]. In this study, we adopted 2bRAD-M and LC–MS tools to determine the potential association between the intratumoral microbiota and metabolites in mice. Our findings showed that the intratumoral microbiota and metabolites were evidently different in HCC and paired nontumor tissues. To sum up, the features of the intratumoral microbiota might represent potential strategies for HCC diagnosis and treatment, and the intratumoral microbiota could provide clues for therapeutic decision-making in HCC.
Studies have shown significant differences between cancer and normal tissues, including between genes, metabolites, and RNA methyladenosine [
34‐
37]. Our research classified the Top20 differential metabolites based on their potential sources into two categories: probably derived from the host, and probably derived from both the host and microorganism. Furthermore, we evaluated the relationship between the tumor microbiota and metabolites in a liver cancer mouse model. The results indicated a significant positive correlation between the metabolites D-( −)-Glutamine and DL-Glutamine, probably derived from the host, and Pseudomonas koreensis and Pseudomonas psychrotolerans. It suggests that the microbiota may indirectly influence the abundance of metabolites by affecting the host's metabolic status. In recent years, the role of glutamine metabolism imbalance in the pathogenesis of HCC has been increasingly emphasized [
37,
38]. Functionally, decreased GOT2 expression has been found to promote glutaminolysis through glutamine metabolism and contribute to HCC progression [
38]. In addition, Wei et al. found that HMGB1 regulated glutamine metabolism in HCC cell through dual mechanisms. On the one hand, HMGB1 could promote glutamine synthetase expression via the mTORC2-AKT-C-MYC pathway. On the other hand, HMGB1 could inhibit glutamate dehydrogenase by inducing the mTORC1 pathway to down-regulate SIRT4 [
39]. These findings highlight the importance of studying tumor metabolism as biomarkers for HCC diagnosis and prognosis prediction.
Accumulated evidence suggests that there is a close relationship between intratumoral bacteria and metabolite [
40,
41]. Microbiota derived metabolites play a role in regulating the tumor microenvironment [
42]. Citrulline, for example, is a metabolite that derived from both host and microbial sources [
43,
44]. Citrulline is known to have two synthetic pathways that are primarily completed in the small intestine [
45]. In our study, citrulline has strong correlation with
muribaculaceae bacterium isolate 102 HZI,
Ralstonia sp UNC404CL21Col and
Allobacillus sp SKP48, suggesting that these bacteria in tumors may be involved in the metabolic pathway of citrulline in the liver. Although the evidence for the association between the microbiota and metabolites is relatively weak, it provides some clues for us to further explore the specific function and mechanism of metabolites derived from intratumoral bacteria.
In conclusion, we utilized 2bRAD-M and LC–MS to describe the features of bacteria and metabolites in an HCC mouse model. Our findings might provide clues for further study of the potential function of bacteria and metabolites in tumor tissues.
Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit
http://creativecommons.org/licenses/by/4.0/. The Creative Commons Public Domain Dedication waiver (
http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated in a credit line to the data.
Publisher's Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.