Community characteristics and relationship between gut microbiota and intratumoral microbiota in hepatocellular carcinoma

Abstract

Background: The combination of local therapy with lenvatinib and programmed cell death protein-1 (PD-1) inhibitors represents an emerging treatment paradigm for unresectable hepatocellular carcinoma (uHCC). Our study sought to investigate the interrelationship between gut microbiota and intratumoral microbiota in the context of triple therapy, with a view to identifying potential biological markers.

Methods: The gut microbial community profiles of patients with primary untreated hepatocellular carcinoma (HCC) and those treated with local therapy combined with lenvatinib and PD-1 inhibitors were analyzed by 16S rRNA gene amplicon sequencing. Additionally, microbial community profiles of tumor tissues of patients with HCC and normal liver tissues were analyzed.

Results: In our investigation, we observed that patients with HCC who received triple therapy exhibited a notable enhancement in the abundance of Actinobacteriota and a considerable decrease in Escherichia Shigella. Patients who received hepatic artery infusion chemotherapy (HAIC) in combination with levatinib and PD-1 inhibitors exhibited significantly elevated levels of Faecalibacterium prausnitzii and Bacteroides stercoris in comparison to those who received transarterial chemoembolization (TACE) in combination with levatinib and PD-1 inhibitors. Furthermore, a notable decline in microbial diversity was observed within HCC tumors in comparison to normal liver tissues. The gut and intratumoral microbiota in HCC patients exhibited a high degree of similarity to the microbes present at the phylum level.

Conclusions: Gut microbiota is connected to triple therapy with local therapy combined with lenvatinib and PD-1 inhibitors for HCC. These discoveries underscore the potential of utilizing gut microbiota and intratumoral microbiota as biomarkers, as well as the possibility of triple therapy in the management of HCC.

Keywords: gut microbiota; hepatic artery infusion chemotherapy; hepatocellular carcinoma; intratumoral microbiota; lenvatinib; programmed cell death protein-1 inhibitors; transarterial chemoembolization.

Figure 1

Gut microbial diversity and taxonomic composition in patients with primary untreated HCC patients with HCC treated with local therapy combined with lenvatinib and PD-1 inhibitors. (A-D) Between-group α-diversity indices from (A–D) are chao1, ace, shannon, simpson. (E, F) β-diversity differences between groups based on weighted and unweighted unifrac distance. (G) Between-group PCoA analysis based on Unweighted unifrac distance. (H) Between-group PLS-DA analysis i.e. partial least squares discriminant analysis. (I, J) Histograms of gut microbial species at the phylum and genus level, marked with "***" if P < 0.001, "**" if 0.001 <= P <= 0.01, "**" marker; if 0.01 < P <= 0.05 "*" marker, if P > 0.05 no marker. (K) LEfSe clustering plot of two groups of gut microbiota. (L) LEfSe LDA plot of two groups of gut microbiota (LDA score [log10]>2).

Figure 2

Gut microbial diversity and taxonomic composition in HCC patients with TACE combined with lenvatinib and PD-1 inhibitors and HCC patients with HAIC combined with lenvatinib and ICIs. (A-D) Between-group α-diversity indices from (A–D) are chao1, ace, shannon, simpson. (E, F) β-diversity differences between groups based on weighted and unweighted unifrac distance. (G) β-diversity differences between groups by PLS-DA analysis i.e., Partial Least Squares Discriminant Analysis. (H) Comparison of keystone species differences at the class level. (I) Comparative bar chart of key species differences at the species level. (J) LEfSe clustering plot of two groups of gut microbiota. (K) LEfSe LDA plot of two groups of gut microbiota (LDA score [log10]>2).

Figure 3

Microbial diversity and taxonomic composition of tumor tissues and normal liver tissues from HCC patients. (A-D) Intergroup α-diversity indices from (A–D) are chao1, ace, shannon, simpson. (E) Intergroup PCoA analysis based on weighted unifrac distance (F) Between-group PLS-DA analysis was partial least squares discriminant analysis (G) Intergroup NMDS analysis i.e. non-metric multidimensional scaling method, stress<0.2 its graphs have some interpretive significance (H, I) Comparative plots of microbial keystone species differences at the phylum and genus level, marked with "***" if P < 0.001 , "***" if 0.001 <= P <= 0.01, "**" marker; if 0.01 < P <= 0.05, then "*" marker; if P > 0.05 then no marker (J) LEfSe clustering plot of two groups of microbes (K) LEfSe LDA plot of two groups of microbes (LDA score [log10]>4.8).

Figure 4

The same and different of group U and I (A) Venn diagram of microorganisms of group U and group I. (B) Histogram of species composition of group T and U and I microorganisms at the phylum level , marked with “***” if P < 0.001, “**” if 0.001 <= P <= 0.01, “** “ marker; if 0.01 < P <= 0.05, “*” marker; if P > 0.05, no marker.

Figure 5

Functional prediction based on PICRUSt2. (A) Abundance prediction results of KEGG function of group T and group U bacterial communities at level2 obtained by PICRUST2, abundance threshold is 0.07. (B) Abundance prediction results of KEGG function in group T and group U bacterial communities at level3 obtained by PICRUST2, abundance threshold is 0.02. (C) After the prediction of the function of all samples was obtained by PICRUST2, the Wilcox test was used to analyze the difference function between group I and group L.