Profiling of the tumor-associated microbiome in patients with hepatocellular carcinoma

Abstract

Background: Tumor tissues have been shown to host a diverse array of bacteria, suggesting a link between the intratumoral microbiota and the development and progression of cancer. The aim of this explorative study was to perform microbiome analysis in liver tumor and to evaluate its relationship with cancer stage and survival outcome.

Results: We conducted an exploratory study on a cohort of 20 hepatocellular cancer patients from the SORAMIC trial. Patients were divided into curative and palliative groups according to treatment type (local ablation, alone or combined with systemic therapy). The V1-V2 regions of 16 S rRNA were sequenced starting from archival tissues. Amplicon Sequence Variants (ASVs) were taxonomically assigned to the upper (UGI) or lower (LGI) gastrointestinal tract. Bacteria were identified in both tumoral and non-tumoral tissues, showing higher diversity and correlation between diversity and shorter survival in the palliative group (S. aureus p < 0.05; B. parvula p < 0.01; A. chinensis p < 0.01). Both therapy groups were enriched with the genus Bacilli, including Streptococcus spp., Gemella haemolysans and Helicobacter pylori, commonly found in UGI. The results suggested that among palliative patients and those with shorter survival, G. haemolysans was more prevalent, while H. pylori was more often found in curative patients with longer survival. However none of the results were significantly different (p > 0.05). A higher microbiome biodiversity was associated with an increased number of lesions (Hoylesella, Agathobacter, Sphingobium, Cardiobacterium, Photobacterium and Serratia, all with p < 0.01).

Conclusions: The presence of bacteria, predominantly from communities of the UGI, suggests their translocation into liver tissue due to impaired barrier function of the upper gut or the ascending pathway along the biliary duct system. The intratumoral prevalence of bacteria with proinflammatory and oncogenic potential suggests their potential role in HCC pathomechanisms.

Keywords: Helicobacter pylori; Hepatocellular carcinoma; Interventional radiology; Liver; Microbiota.

Fig. 1

Representative images of the intratumor localization of bacteria in FFPE liver tissues. Images show the results of immunochemical staining for lipopolysaccharide (LPS) and lipoteichoic acid (LTA) in hepatocellular cancer tumor, stromal and adjacent tissues to detect gram-negative and gram-positive bacteria, respectively. Bacteria identified in tumor cells are denoted with red arrows, stromal cells with green arrows and hepatocytes with orange arrows. Scale bars, 0.060 mm and 0.020 mm. Magnification: 1x in the upper panels and 20x in the tumor, stroma and adjacent panels

Fig. 2

Co-occurrence bacterial network in FFPE liver tissues. Size fonts, color fonts and color nodes denote the level of degree (number of neighbors) of each node, and edge colors denote rho values. Level of significance: p value < 0.05 (correlations were calculated with Spearman’s rank correlation coefficient (rho ≥ 0.4) after false discovery rate correction via the Benjamin‒Hochberg procedure)

Fig. 3

Richness as a function of the degree (number of neighbors) of the different members of the bacterial community network. Genera detected with a degree of ≤ 5 in each FFPE liver tissue sample are indicated in brackets

Fig. 4

Overall bacterial diversity (top) and unsupervised hierarchical clustering of prevalent taxa detected in tumor (HCC) and adjacent tissues via the Bray‒Curtis dissimilarity index (bottom). Both the top and bottom panels display the results from FFPE liver tissues at the taxonomic ranks of class (A, B) and genus (C, D). Bootstrapping analysis is reported in SF7

Fig. 5

Principal component analysis (PCA). The analysis was performed starting from tumor (red) and adjacent (black) FFPE liver tissues, with each group name positioned at its respective data centroid

Fig. 6

Principal component analysis (PCA) and indicator species analysis (ISA). The analysis was performed on FFPE liver tissues considering comparisons of the following clinical characteristics: curative vs. palliative (top left), survival > 16 months vs. < 16 months in the curative group (top right) and in the palliative group (bottom left), respectively, and tumor vs. adjacent tissues (bottom right). Level of significance: significant differences in the abundances of bacterial species within groups are indicated by gray boxes and marked with * for p < 0.05 and ** for p < 0.001 (differences between predefined groups were assessed using an unsupervised hierarchical clustering using Bray-Curtis algorithm and indicator species analysis)