Liver fluke granulin promotes extracellular vesicle-mediated crosstalk and cellular microenvironment conducive to cholangiocarcinoma

Abstract

Crosstalk between malignant and neighboring cells contributes to tumor growth. In East Asia, infection with the liver fluke is a major risk factor for cholangiocarcinoma (CCA). The liver fluke Opisthorchis viverrini secretes a growth factor termed liver fluke granulin, a homologue of the human progranulin, which contributes significantly to biliary tract fibrosis and morbidity. Here, extracellular vesicle (EV)-mediated transfer of mRNAs from human cholangiocytes to naïve recipient cells was investigated following exposure to liver fluke granulin. To minimize the influence of endogenous progranulin, its cognate gene was inactivated using CRISPR/Cas9-based gene knock-out. Several progranulin-depleted cell lines, termed ΔhuPGRN-H69, were established. These lines exhibited >80% reductions in levels of specific transcript and progranulin, both in gene-edited cells and within EVs released by these cells. Profiles of extracellular vesicle RNAs (evRNA) from ΔhuPGRN-H69 for CCA-associated characteristics revealed a paucity of transcripts for estrogen- and Wnt-signaling pathways, peptidase inhibitors and tyrosine phosphatase related to cellular processes including oncogenic transformation. Several CCA-specific evRNAs including MAPK/AKT pathway members were induced by exposure to liver fluke granulin. By comparison, estrogen, Wnt/PI3K and TGF signaling and other CCA pathway mRNAs were upregulated in wild type H69 cells exposed to liver fluke granulin. Of these, CCA-associated evRNAs modified the CCA microenvironment in naïve cells co-cultured with EVs from ΔhuPGRN-H69 cells exposed to liver fluke granulin, and induced translation of MAPK phosphorylation related-protein in naïve recipient cells in comparison with control recipient cells. Exosome-mediated crosstalk in response to liver fluke granulin promoted a CCA-specific program through MAPK pathway which, in turn, established a CCA-conducive disposition.

Keywords: Cellular crosstalk; Extracellular vesicle; Liver fluke granulin; Opisthorchis viverrini.

Fig. 1

Programmed CRISPR/Cas9 mutation of the human progranulin gene. Panel A, linear map of a pre-designed lentiviral CRISPR/Cas9 vector containing fused codon-optimized puromycin resistance marker (dark gray bar)-Cas9 (yellow bar) and green fluorescent protein (green bar) driven by the mammalian elongation factor alpha-1 promoter (dark blue arrow); guide RNA (gRNA) targeting human granulin exon 2 (red bar) is expressed from a single vector. The human U6 promoter (blue arrow) drove the gRNA. The vector backbone includes the 5′- and 3′-long terminal repeats (LTR) of the HIV-1 provirus (light gray blocks). Panel B, schematic representation of the partial human granulin gene precursor, huPGRN on chromosome 17: NC_000017.11 regions 44,345,086–44,353,106 (8021 bp) and protein structure. Nucleotide sequence in exon 2 encodes the N-terminus and part of the granulin/epithelin module (GEM) of progranulin; indicating locations of gRNA (4417–4438 nt; red colored-letter) predicted double-stranded break (DSB) (red arrow). Panels C–E, on target INDEL mutations in ΔPGRN-H69 analysis by NGS libraries and CRISPResso bioinformatic platform. Frequency distribution of position-dependent insertions (red bars) (C) and deletion (magenta) (D) the major INDELs; 6 and 2 bp deletions at the programmed CRISPR/Cas9 cleavage site. Other minor mutations including 1 bp insertion (red square) or base substitution (bold) were observed further (>10 bp) predicted cleavage site (E). (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

Fig. 2

Reduction of progranulin transcript and protein expression levels and cell proliferative effects of liver fluke granulin. Panels A and B, reduction of huPGRN transcription levels from ΔhuPGRN-H69 cell; red bar (∼70%) comparing with H69 reference (black bar). The huPGRN differential transcript after normalization with human GAPDH gene; mean ± SD, n = 3 (biological replicates); P < 0.0001 (****), unpaired t-test. Diminished level of huPGRN revealed by WB analysis using anti-PGRN antibody (64 kDa) compared with anti-GAPDH antibody (36 kDa). Progranulin levels in ΔhuPGRN-H69 cells (red bars) was reduced by ∼80%. The GAPDH levels were stable within three biological replicates. These decreased levels of progranulin were significantly different from H69 (black bars); n = 3 (biological replicates); P < 0.0001 (****), unpaired t-test. Panel C, real time monitoring of H69 vs ΔhuGRN-H69 cell proliferation before and after addition of liver fluke granulin. The lower normalized cell index (nCI) of ΔhuGRN-H69 cells (discontinuous red lines) compared with H69 cells (discontinuous gray lines) were monitored over 48 h. The nCI of ΔhuGRN-H69 cells was recovered as in H69 cells nCI value after addition of liver fluke granulin at 100 nM for 24 h, and higher than H69 from 24 to 48 h. The nCI signals are shown as mean ± SD, for ≥3 independent experiments, and assessed using two-way ANOVA with Dunnet’s multiple correction. Comparing 100 nM liver fluke granulin treatments on each cell type to untreated cells (either H69 vs 100 nM liver fluke granulin treated H69 or ΔhuPGRN-H69 vs 100 nM liver fluke granulin -treated ΔhuPGRN-H69): P < 0.05 (*); P < 0.01 (**); P < 0.001 (***). Comparing 100 nM liver fluke granulin-treated H69 to 100 nM liver fluke granulin-treated ΔhuPGRN-H69: P < 0.05 (#); P < 0.001 (###). (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

Fig. 3

EV-derived H69 cell characterization and PGRN expression. Panel A, the visualization of EVs from H69 cells by in situ hybridization using fluorophore 488-labeled anti-CD81 antibody (green dot, panel A) around DAPI-stained cell nuclei (blue color). The H69-derived EV-like particle sizes ranged from 55 to 80 nm. The protein composition of EVs were determined using the Exosome Antibody Array (SBI System Biosciences). Panel B, dark spots indicate presence of the target protein (upper panel). Absence of a spot for GM130 confirmed the absence of cellular contaminants. The intensity in gray scale was also plotted (lower panel). Panel C, WB analysis for the exosome-specific markers; strongly positive signal for CD9 and CD81 were found. Panels D and E, both the evRNA and evProtein from ΔhuPGRN-H69 showed reduction of ∼90% differential transcript levels [red bar, panel D] after normalization with GAPDH in comparison to H69 [black bar] (red bar, panel D). The levels of progranulin were reduced by ∼90%; unpaired t-test, P < 0.0001 (****), n = 3. Panels F–H, show an induction of CCA-related mRNAs carrying cholangiocyte derived-EVs after exposure to liver fluke granulin. The heat map was plotted based on the differential transcript fold change as determined using the CCA-related gene Array using H69 as the reference from H69 with liver fluke granulin treatment (panel F) and ΔhuPGRN-H69, ΔhuPGRN-with liver fluke granulin treatment (panel H). Similar quantities of evRNAs from each group were probed individually with 88 CCA targets in triplicate; there were slight differences (P < 0.05 by 2way ANOVA gene profiles (55 from 88 transcripts were detected) in ΔhuPGRN-H69 vs H69 (panel A) by absent of DPYD, ESR, KDM3A, LEF1, PRKC9, PTPN13, SERPINA3, SERRPINE2 and SOX11 (panel H). However, most of these genes were recovered after exogenous liver fluke granulin treatment of ΔhuPGRN-H69 (panels G and H). The CCA mRNA profiles after liver fluke granulin treatment of EVs derived from ΔhuPGRN-H69 and H69 were markedly dissimilar following exposure to liver fluke granulin with P < 0.0001, 2way ANOVA (panels F and G). Panel G, shows the CCA-related evRNAs after exposure to liver fluke granulin in ΔhuPGRN-H69 compared with its mRNA profile (P < 0.0001 with two-tailed, paired t-test). The statistical significance was calculated from differential fold changes in transcription. The liver fluke granulin (rOv-GRN-1) treatment group of cells showed 1–15 folds (gradient red bar) level of induction of transcription. The ‘1’ indicates baseline (white bar) level expression, and ‘0’ indicated the absence of expression (gradient blue bar). The transcripts from non- or treated- rOv-GRN-1treatment were not detected (no baseline signal) indicate in black bar and gray-cross bar, respectively. The heat map was plotted from the average transcript level ran in triplicates using Prism v8. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

Fig. 4

Uptake of extracellular vesicles and paracrine transfer of mRNAs from donor to naïve recipient cells. Panels A and B, representative fluorescence micrographs showing uptake of PKH26-labeled EVs into recipient cholangiocytes; blue—nuclei staining with the NucBlu Live Cell Stain ReadyProbe, red—PKH26-labeled EVs. Pitstop 2 blocked clathrin-dependent endocytosis of EVs (C). Merged images shown in lower panel with bright field. Magnification, 40×; scale bar, 10 µm. Arrows indicate internalized EVs in naïve cells. Panels D to F, cholangiocytes were grown to 80% confluence after which they were co-cultured with EVs from liver fluke granulin-activated ΔhuPGRN-H69 cells for 24 h. The cell lysate (500 ng) was used for 17 targets (panel D indicates the key of antibody spot in duplicate) on the human MAPK phosphorylation antibody array; ab211061 (Abcam). The relative density of the spots was quantified by densitometry analysis (panels E and F) (RFU). Red boxes and red bars represent the predominant proteins that significantly higher than control groups (grey bars) namely, GSK3a, MKK3, mTOR, p38, p53, p70S6K and RSK1. The MAPK phosphorylation protein pattern of these protein revealed statistically significant differences in comparison with its control; P < 0.0001, 2way ANOVA. Data represent the average of two individual sets per sample. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)