Whole-exome sequencing reveals genomic landscape of intrahepatic cholangiocarcinoma and identifies SAV1 as a potential driver

Zhou, Zheng-Jun; Ye, Yu-Hang; Hu, Zhi-Qiang; Hou, Yue-Ru; Liu, Kai-Xuan; Sun, Rong-Qi; Wang, Peng-Cheng; Luo, Chu-Bin; Li, Jia; Zou, Ji-Xue; Zhou, Jian; Fan, Jia; Song, Cheng-Li; Zhou, Shao-Lai

doi:10.1038/s41467-024-54387-8

Download PDF

Article
Open access
Published: 17 November 2024

Whole-exome sequencing reveals genomic landscape of intrahepatic cholangiocarcinoma and identifies SAV1 as a potential driver

Zheng-Jun Zhou^1,2^na1,
Yu-Hang Ye^1,2^na1,
Zhi-Qiang Hu^1,2^na1,
Yue-Ru Hou³^na1,
Kai-Xuan Liu^1,2,
Rong-Qi Sun^1,2,
Peng-Cheng Wang^1,2,
Chu-Bin Luo^1,2,
Jia Li^1,2,
Ji-Xue Zou²,
Jian Zhou ORCID: orcid.org/0000-0002-2118-1117^1,2,
Jia Fan^1,2,
Cheng-Li Song³ &
…
Shao-Lai Zhou ORCID: orcid.org/0000-0002-8526-5221^1,2

Nature Communications volume 15, Article number: 9960 (2024) Cite this article

8533 Accesses
5 Citations
Metrics details

Subjects

Abstract

Intrahepatic cholangiocarcinoma (ICC) is the second most common primary hepatic malignancy after hepatocellular carcinoma, with poor prognosis and limited treatment options. The genomic features of ICC in Chinese patients remain largely unknown. In this study, we perform deep whole-exome sequencing of 204 Chinese primary ICCs and characterize genomic alterations and clonal evolution, and reveal their associations with patient outcomes. We identify six mutational signatures, including Signatures A and F, which are highly similar to previously described signatures linked to aristolochic acid and aflatoxin exposures, respectively. We also identify 13 significantly mutated genes in the ICC samples, including SAV1. We find that SAV1 was mutated in 2.9% (20/672) of 672 ICC samples. SAV1 mutation is associated with lower SAV1 protein levels, higher rates of tumor recurrence, and shorter overall patient survival. Biofunctional investigations reveal a tumor-suppressor role of SAV1: its inactivation suppresses Hippo signaling, leading to YAP activation, thereby promoting tumor growth and metastasis. Collectively, our results delineate the genomic landscape of Chinese ICCs and identify SAV1 as a potential driver of ICC.

Convergent genomic diversity and novel BCAA metabolism in intrahepatic cholangiocarcinoma

Article Open access 19 April 2023

Integrative analysis of multiple genomic data from intrahepatic cholangiocarcinoma organoids enables tumor subtyping

Article Open access 16 January 2023

Prognostic factors in patients with intrahepatic cholangiocarcinoma

Article Open access 17 August 2024

Introduction

Intrahepatic cholangiocarcinoma (ICC) is the second most common primary hepatic malignancy after hepatocellular carcinoma (HCC) and has an increasing incidence worldwide^1,2. Advances in diagnostic modalities and clinical screening have made early detection and curative resection of ICC possible; however, long-term patient survival is still poor because of high relapse rates^3,4. Recently, several pioneering studies expanded our understanding of ICC pathogenesis at the molecular level by delineating the landscape of genetic alterations that underlie ICC carcinogenesis^{5,6,7,8,9,10,11}. However, most of these studies contained limited numbers of ICC samples, and the patients were from multiple countries with different epidemiological and ancestral backgrounds. Furthermore, most of the studies used targeted sequencing, and whole-exome sequencing (WES) or whole-genome sequencing was only performed with low coverage. More importantly, exhaustive follow-up was performed on very few of the patients. These shortcomings may impede the discovery of potential drivers of ICC, as well as hinder the exploration of potential etiological factors contribute to genetic alterations and their association with patient outcomes.

In this work, we aim to systematically define genomic alterations in Chinese patients with ICC. We perform deep WES on 204 Chinese primary ICCs and analyze the associations of genomic alterations with tumor progression and patient outcomes. Focusing on a potential driver of ICC, SAV1, we perform Sanger sequencing to evaluate all the SAV1 coding exons in an additional 468 ICC specimens. We also investigate the functional effects and associated mechanisms of SAV1 on tumor growth and metastasis. Collectively, our results delineate the genomic landscape of Chinese ICCs and identify SAV1 as a potential driver of ICC.

Results

Overview of genomic alterations in patients with primary ICC

To discover the genomic alterations in Chinese patients with ICC, we performed WES of tumor and matched non-cancerous liver-tissue samples from 204 patients with ICC. The average sequencing depth was 445.9-fold for tumors and 138.7-fold for normal tissues (Supplementary Data 1). We identified a total of 21,772 somatic coding SNVs and 1302 indels, with 2–1628 mutations per tumor genome (Supplementary Data 2). We performed Sanger sequencing to validate 1756 randomly selected somatic non-synonymous mutations, which revealed a high true-discovery rate (95.9%). There was an average of 3.77 (range = 0.07–54.27) somatic coding mutations per Mb (Fig. 1A).

**Fig. 1: Global genomic alterations in 204 Chinese primary ICCs.**

The WES results indicated that C > A transversions and C > T and T > C transitions were ubiquitous in all of the ICCs (Fig. 1A), a feature shared by other ICC cohorts^8,9,11. In addition, we identified 17 amplified segments, which harbored several known oncogenes, including MYC, ERBB2, and CCND1. We also identified 24 lost segments, which harbored tumor suppressors such as TP53 (17p13) and CDKN2A/B (9q21) (Fig. 1A; Supplementary Fig. 1A, B; Supplementary Data 3). In addition, we identified whole-genome doubling in 44 (21.6%) of the 204 tumors (Supplementary Data 4).

We correlated the tumor mutation burden (TMB) and copy number variation (CNV) burden with patient prognosis. The results showed that high TMB was predictive of low recurrence-free survival (RFS) but had no ability to predict OS. In addition, CNV burden showed no associations with patient prognosis, either for OS or RFS (Supplementary Fig. 1C).

Mutational signatures

We applied NMF and identified six mutation signatures in the 204 ICC samples (Signatures A–F; Fig. 1B, C; Supplementary Fig. 2A; Supplementary Data 5). Signature A was characterized by dominant T > A mutations and was highly similar to the previously described Signature 22 (cosine correlation similarity = 0.99), which is known to result from exposure to aristolochic acid and to be associated with a high mutational burden¹². Consistent with that, Signature A was a predominant signature in the hypermutated ICC samples. Signature B, characterized by C > G mutations, was similar to the previously described Signature 13, which has been proposed as an APOBEC-driven hypermutated phenotype. Signatures C and E were highly similar to the previously described signatures 1 and 5, respectively (correlation coefficient = 0.96 and 0.93, respectively), which have been associated with aging. Signature D showed similarity to the previously described Signature 20 (correlation similarity = 0.85), which was previously confirmed to be driven by defective DNA mismatch repair and associated with high numbers of indels at mono/polynucleotide repeats. Consistent with that, Signature D was mainly identified in four ICCs with high numbers of indels. Signature F, characterized by C > A mutations, showed strong similarity to the previously described Signature 24 (correlation similarity = 0.94), which is associated with exposure to aflatoxin in cancers.

Hierarchical clustering, based on the prevalence of the mutation signatures, identified nine groups of tumors (Msig-clusters 1–9; Fig. 1C) characterized by enrichment of specific mutation signatures. For example, Msig-clusters 1 and 6 were characterized by Signature A, as they contained more mutations that fit Signature A than other clusters. Furthermore, Msig-cluster 3 was characterized by Signature D, Msig-cluster 4 was characterized by signature F, and Msig-cluster 5 was characterized by Signature B. Associations between the signature-based clusters and clinical background and genetic alterations were also revealed (Supplementary Fig. 2B). Msig-clusters 1 and 6 exhibited significant correlations with higher TMB (P = 0.029 and P = 0.003, respectively), while Msig-clusters 8 and 9 were correlated with lower TMB (P = 9.59 × 10⁻⁵ and P = 0.021, respectively). Msig-clusters 4 and 5 demonstrated higher TMB (P = 8.97 × 10⁻⁵ and P = 0.003, respectively) and higher CNV burden (P = 0.038 and P = 0.035, respectively). Especially, we confirmed correlations between Msig-cluster 5 and the presence of hepatolithiasis (P = 0.017), implicating AID/APOBEC enzymes in cholangitis-induced tumorigenesis. In addition, Msig-clusters 3 and 9 respectively exhibited correlations with the presence of ARID1A mutations or IDH1 mutations (P = 0.005 and P = 0.027, respectively).

Significantly mutated protein-coding genes and associations with patient outcomes

We used MutSigCV to identify cancer driver genes¹³. Thirteen genes were significantly enriched with damaging mutations (q < 0.1; Fig. 2A; Supplementary Data 6): TP53, KRAS, IDH1, ELF3, SAV1, BAP1, ARID1A, PBRM1, PTEN, ARID2, BRD7, CDKN2A, and ACVR2A. We also identified oncogenes (MYC, ERBB2, and CCND1) with copy-number amplification and tumor-suppressor genes (CDKN2A and CDKN2B) with copy-number deletion. We then compared the mutation frequencies in the 13 genes to those in a combined TCGA (n = 30) and MSK (n = 158) cohort. We found that the mutation frequencies in several of the 13 genes were higher in our ICC cohort than in the TCGA and MSK cohort, including TP53 (30% versus 16%), KRAS (14% versus 6%), ELF3 (5% versus 1%), and SAV1 (3% versus 0%), whereas the mutation frequencies of IDH1 (10% versus 26%) and PBRM1 (7% versus 13%) were slightly lower in our ICC cohort than in the TCGA and MSK cohort (Fig. 2B).

**Fig. 2: Significantly mutated protein-coding genes and associations with patient outcome.**

We next correlated somatic mutations in specific genes with patient outcomes. We found that somatic mutations in TP53, KRAS, ELF3, and SAV1 were correlated with patient OS and/or RFS (Supplementary Fig. 3). When we looked at the combination of these four genes (TP53, KRAS, ELF3, and SAV1; Fig. 2C), the cumulative recurrence rate was highest (92.9%) in patients with mutations in two or more of the four genes, which was significantly higher than the cumulative recurrence rate in patients with mutation in only one of the four genes (76.3%) or in none of the four genes (60.9%; Fig. 2D). Similarly, the 5-year OS among patients with mutations in two or more of the four genes was 7.7%, which was significantly lower than that among patients with mutation in only one of the four genes (27.2%) or in none of the four genes (45.3%; Fig. 2E).

Clonal diversity and evolution of recurrent somatic mutations

To estimate genetic diversity based on WES of all the ICC samples, we used an analysis pipeline described in a previous study¹⁴ to compute a Shannon index for each patient^15,16, which estimates not only the number of clones, but also their relative abundance. A patient with a single clone would have a Shannon index of 0, with increasing Shannon index corresponding to a greater number of clones and a smaller dispersal of CCFs (Fig. 3A). We divided the patients into a “clonal equilibrium” group, defined by a Shannon index higher than the median value among all patients, and a “clonal dominance” group, defined by a Shannon index below the median value (Fig. 3A). Kaplan–Meier survival analysis showed that clonal dominance was associated with significantly shorter OS and RFS compared with clonal equilibrium. Univariate and multivariate analyses revealed that the Shannon index was an independent prognostic factor for both OS and RFS (Fig. 3B; Supplementary Data 7).

We next deduced the approximate sequential order in which the somatic alterations occurred during ICC development by calculating how often a recurrent non-synonymous mutation resided on a clonal event versus a subclonal event. This analysis confirmed that BAP1, ARID1A, and TP53 mutations were the earliest mutations to evolve, followed by PBRM1, ELF3, and KRAS mutations. By contrast, ARID2 and IDH1/2 mutations were mostly restricted to subclonal events, which occurred at later time points (Fig. 3C).

Somatic mutation of SAV1 was associated with reduced SAV1 expression and predictive of patient outcomes after curative resection

We evaluated all coding exons of SAV1 in an additional 468 pairs of ICCs and normal tissue samples by Sanger sequencing. We identified SAV1 somatic mutations in 14 of the 468 ICCs. Thus, in the overall total of 672 ICCs included in our study, we identified 20 SAV1 somatic mutations in 20 different patients (Fig. 4A; Supplementary Fig. 4; Supplementary Data 8). Specifically, 14 (70%) of the 20 SAV1 mutations were truncating mutations, which suggested a possible tumor-suppressor role of SAV1 in ICC.

**Fig. 4: Clinical significance of *SAV1* mutation in 672 ICCs.**

We further evaluated SAV1 expression by immunohistochemistry in all 672 ICCs. The results showed that SAV1 expression was down-regulated among all the tumor samples compared with that in the matched non-tumor liver samples (Fig. 4B). Moreover, the patients with SAV1 somatic mutation showed a further decrease in tumor SAV1 expression (Fig. 4B).

We then correlated SAV1 somatic mutation and expression with the patients’ clinical characteristics and outcomes. SAV1 somatic mutation was correlated with increased tumor size and poor tumor differentiation. The SAV1 expression level was correlated with higher CA19-9 and GGT levels, increased tumor size, and lymphatic metastasis (Supplementary Data 9). Among the 672 patients with ICC, those with SAV1 somatic mutation or reduced SAV1 expression exhibited shorter OS and RFS compared with patients without those features (Fig. 4C). Univariate and multivariate analyses confirmed that SAV1 somatic mutation and low expression were independent prognostic factors for OS and RFS (Supplementary Data 10).

Structural analysis of the effects of SAV1 mutations on protein structure

We obtained the full-length structure of human SAV1 based on the prediction provided in the AlphaFold predicted-structure database (https://alphafold.ebi.ac.uk/), and showed the structure of 198–383 residues region in this study. We also built a structural model of the SAV1–MST2 heterotetramer to more intuitively understand the significance of dimerization of the WW12 domain to the function of SAV1 (Supplementary Methods).

The structural and biochemical results of a recent study confirmed that SAV1 can interact with MST2 and form a heterodimer through the SARAH domain¹⁷, which is essential for the interaction between SAV1 and MST kinase¹⁸. Two SAV1 molecules interact through their WW12 domains, composed of WW1 and WW2, to form a homodimer; WW2 is mostly responsible for the formation of the dimer interaction interface¹⁹. Furthermore, two SAV1–MST2 heterodimers further dimerize through the SAV1 WW12 domain to form a heterotetramer, which phosphorylates the downstream LATS/MOB complex and promotes Hippo activation¹⁹.

The R47X, N58fs, E154fs, and P203fs mutations in our ICC samples caused deletion of the WW12 and SARAH domains. The deletion of WW12 is predicted to hinder the formation of the SAV1 homodimer, and the absence of the SARAH domain is likely to impede the interaction between SAV1 and MST kinase to form a heterodimer (Supplementary Fig. 5B). In addition, the L231fs, R233X, R242X, and E247X mutations in our ICC samples is predicted to destroy the integrity of the WW2 domain and caused deletion of the SARAH domain. Deletion of WW2 is expected to cause SAV1 to fail to form a homodimer (Supplementary Fig. 5B). The absence of the SARAH domain is predicted to impede the interaction between SAV1 and MST kinase. Similarly, the P271fs, Y273fs, R313X, R318X, and T334fs mutations caused the deletion of the SARAH domain, which is essential for the interaction between SAV1 and MST kinase (Supplementary Fig. 5B). Therefore, all of the above mutations could affect SAV1-mediated downstream signal transduction.

SAV1 functions as a tumor-suppressor gene in ICC

Our results suggested a possible tumor-suppressor role for SAV1 in ICC. To test that hypothesis, we performed WES on five ICC cell lines (HuCCT1, CCLP1, RBE, HCCC-9810, and SG231). All of the ICC cell lines were identified as having WT SAV1 using the same filter criteria used for the ICC samples. The WES results for the cell lines were validated by Sanger sequencing. Western blots and qRT-PCR confirmed that SAV1 mRNA and protein levels in the five ICC cell lines, especially HCCC-9810 and SG231, were lower than those in H-69 cells (Fig. 5A).

**Fig. 5: Identification of *SAV1* as a tumor-suppressor gene in ICC.**

Next, we knocked down SAV1 in HuCCT1 cells (Fig. 5A). Biofunctional investigations revealed that the knockdown of SAV1 resulted in an increase in ICC cell proliferation, colony formation, and invasive ability (Fig. 5B–D). In vivo ICC mouse models showed that SAV1 knockdown accelerated tumor growth and metastasis (Fig. 5E).

We generated lentiviral constructs to re-express WT and selected mutant SAV1 variants in SG231 cells (Fig. 5A). The results showed that overexpression of WT SAV1 substantially suppressed ICC cell proliferation, colony formation, and invasion ability. By contrast, expression of the SAV1 mutants failed to cause these effects in whole or in part, suggesting that these SAV1 mutations were inactivating mutations (Fig. 5B–D).

Next, we used the SAV1 mutants together with WT SAV1 for in vivo study. In agreement with the in vitro results, WT SAV1 significantly suppressed tumor growth and metastasis, whereas the SAV1 mutants yielded a larger tumor volume and increased metastasis compared with the WT SAV1 (Fig. 5E). These results further suggested that the SAV1 mutations were inactivating types, supporting the notion that SAV1 is a tumor-suppressor gene in ICC and that certain somatic mutations abolish its function and its tumor-inhibitory effect.

SAV1 inactivation suppresses Hippo signaling, leading to YAP activation

In mammals, the scaffold protein SAV1 promotes activation of the Hippo kinase cascade²⁰. Therefore, we tested the effect of SAV1 on Hippo signaling in ICC cells. Western blot analysis showed that knockdown of SAV1 in HuCCT1 cells reduced YAP Ser127 phosphorylation. Conversely, ectopic expression of SAV1 increased YAP Ser127 phosphorylation in SG231 cells (Fig. 6A). Consistent with the YAP phosphorylation, the phosphorylation of the cytosolic kinases upstream of YAP (MST1, MST2, and LATS1) was also decreased in HuCCT1 cells after knockdown of SAV1, whereas it was increased in SG231 cells after restoration of SAV1 expression (Fig. 6A). Immunofluorescence staining showed that knockdown of SAV1 resulted in nuclear accumulation of YAP in HuCCT1 cells, whereas overexpression of SAV1 in SG231 cells led to increased levels of cytoplasmic YAP and reduced levels of nuclear YAP (Fig. 6B, C). These results suggested that SAV1 triggers a Hippo kinase cascade that leads to phosphorylation, cytoplasmic retention, and inactivation of YAP. We also confirmed that compared with WT SAV1, mutated SAV1 failed to trigger this kinase cascade, which suggests that SAV1 positively modulates Hippo signaling and that its inactivation due to mutation suppresses Hippo signaling, leading to YAP activation in ICC cells (Fig. 6A–C). Clinical ICC samples with SAV1 somatic mutation tended to display nuclear accumulation of YAP, as revealed by immunohistochemistry staining (Fig. 6D), which further indicated that SAV1 inactivation suppresses Hippo signaling and leads to YAP activation in ICC cells. Furthermore, YAP knockdown or verteporfin (a YAP inhibitor) treatment alleviated the SAV1 knockdown-induced increases in in vitro cell proliferation, colony formation, and invasion (Fig. 7A–D) and in vivo tumor growth and metastasis in HuCCT1 cells (Fig. 7E, F). Moreover, in SG231 cells, overexpression of S112A mutant YAP (a nonphosphorylatable YAP mutant), but not that of WT YAP, reversed the inhibitory effect of SAV1 on in vitro cell proliferation, colony formation, and invasion and in vivo tumor growth and metastasis (Fig. 7B–F). These results suggest that SAV1 inactivation promotes ICC growth and metastasis by activating YAP.

**Fig. 7: SAV1 inactivation promotes ICC growth and metastasis by activating YAP.**

Discussion

In this study, we delineated the genomic landscape, including somatic SNVs/indels and CNV, in 204 patients with ICC. Through NMF and hierarchical clustering, we identified the Msig-cluster 1 and 6 groups, which were characterized by Signature A, and the Msig-cluster 4 group, which was characterized by Signature F. Signatures A and F were highly similar to the previously described Signature 22 (aristolochic acid) and 24 (aflatoxin), respectively, which is different than the results of previous ICC cohort studies conducted in other countries and regions^9,10,11. Exposures to aristolochic acid or aflatoxin have been revealed as important environmental risk factors for HCC development, especially in China, and are known to contribute to hypermutation in cancer^11,12,21. Our results confirm that exposures to aristolochic acid or aflatoxin are also important environmental risk factors for ICC development in China, and that the SNV burden is higher in ICCs potentially linked to aristolochic acid or aflatoxin exposure than in other ICCs. In addition, we identified Signature D mainly in four ICC samples with higher numbers of indels, indicating similarity to the previously described Signature 20, which was confirmed to be driven by defective DNA mismatch repair. Hence, we propose that the four patients with Signature D would likely benefit from immunotherapy.

We identified 13 genes that were significantly mutated in ICC. About half of these genes, including TP53, KRAS, and IDH1, presented a different mutation frequency in our ICC cohort than in a combined TCGA and MSK cohort. We propose that this might be due to differences in racial or etiological factors. We also identified mutations in TP53, KRAS, ELF3, and SAV1 that were correlated with patient OS and/or RFS. The combined mutational profile of these four genes showed strong predictive value and correlated with OS and RFS. Although some other genes such as IDH1, BAP1 and PBRM1 have been confirmed as drivers of ICC, they have no ability to predict ICC prognosis. Our results suggest that TP53, KRAS, ELF3, and SAV1 may play roles in ICC growth and metastasis.

We also demonstrated the role of clonal diversity in predicting ICC outcomes. We showed that clonal dominance could independently confer a higher risk of recurrence and shorter OS, which is consistent with the results of previous studies^14,22. Although some other studies found that higher numbers of clones were associated with worse patient prognosis^23,24, they did not investigate the contribution of relative clone size to prognosis, which might explain the contradiction between previous findings and our results. We propose that the presence of a dominant clone at the time of diagnosis could reflect the capacity of ICC to undergo rapid clonal turnover upon acquisition of a newly acquired adaptive mutation, leading to tumor recurrence¹⁴.

Among the genes whose mutational frequency correlated with patient OS and/or RFS, SAV1, which was mutated in about 3% of the 672 ICC samples in our cohort, was not previously known as a driver of ICC. We did not find SAV1 to be significantly mutated in the TCGA cohort, possibly because of limited sample size, low sequencing coverage, or differences in racial or etiological factors. A search of cancer genomics datasets available at www.cbioportal.org showed that apart from two nonsense mutations (C697T and C724T), the other SAV1 somatic mutations were identified in our ICC cohort. Furthermore, we found that the majority of SAV1 mutations in our cohort were truncating or splice-site mutations. Although lack of biochemical experiments to support our structure model, through structural analysis and inference, we speculatively predicted that these mutations can block the normal downstream signal transduction process.

In addition to these truncating or splice-site mutations, we also identified 3 missense mutations: S94F, H229, and T363. H229 is located near the junction between the WW1 and WW2 domains. In wild-type (WT) SAV1, the side chain of H229 interacts with L199 and L231 through two stable hydrogen bonds (Supplementary Fig. 5A). L199 is located at the beginning of the WW1 domain, and the interaction between L199 and H229 contributes to the stability of WW1 conformation. L231, along with H229, is located in a loop region that also contains the R233 residue, which is important for the dimerization function of WW2 (Supplementary Fig. 5A, B)¹⁹. An R233Q mutation was previously shown to weaken the dimerization of WW12¹⁹. Similarly, in the H229R mutant that we identified in ICC, the flexibility of the side chain of the substituted R229 residue is predicted to weaken the ability of the residue to form stable hydrogen bonds with L199 and L231. The resulting lack of hydrogen bond interaction between R229 and L199 probably weakens the conformational stability of the WW1 domain. Moreover, the lack of interaction between R229 and L231 is predicted to affect the stability of the loop where R233 is located (Supplementary Fig. 5A, B), thus affecting the ability of R233 to stabilize the dimerization of WW2 and preventing SAV1 from forming a stable heterotetramer with MST2. The reduced phosphorylation function of the unstable heterotetramer results in disruption of the normal downstream signal transduction process.

T363 is located in the middle of the SARAH domain. The side chain of T363 interacts with the Q359 residue through a stable hydrogen bond (Supplementary Fig. 5A). Although T363 is not a part of the interface where SAV1 interacts with MST2, the hydrogen bond between T363 and Q359 is important for the secondary structure stability of the α-helix of the SARAH domain, which supports the hydrophobic interaction between L361 and L365 of SAV1 and L448 and L452 of MST2¹⁷. Moreover, previous studies showed that the stability of the SARAH structure is essential for the interaction between SAV1 and MST kinase¹⁸. In the T363K mutant, K363 is predicted not to form a stable hydrogen bond with Q359, which is expected to weaken the secondary structure stability of the SARAH domain, further weaken the stability of the interaction between SAV1 and MST2 to form a heterotetramer, and thus influence the activation of the Hippo pathway.

SAV1 is a human homolog of Salvador that acts as a scaffolding protein and contains two protein–protein interaction modules known as WW domains¹⁹. Down-regulation of SAV1 in several types of cancer including HCC²⁵, clear cell renal cell carcinoma²⁶, and colorectal cancer²⁷ can promote cell proliferation, anti-apoptosis, migration, and invasion. Furthermore, previous studies in fish and flies have shown a cholangio-suppressive role of SAV1 orthologs^28,29. The functional impact of genetic mechanisms affecting SAV1 is still unknown, however, especially in cancers such as ICC. Although a previous TCGA publication identified SAV1 deletions in 5% of ICC patients, the sample size was small, with only two SAV1 deletions identified in two patients³⁰. Another previous study revealed focal deletions at 14q22.1, which contain coding regions of SAV1 and occur in 12% of ICCs³¹, our results did not identify recurrent genomic deletions involving SAV1 (Supplementary Fig. 1A, B; Supplementary Data 3), which suggests that CNV might not the main factor that leads to SAV1 inactivation. Nevertheless, we showed that SAV1 was significantly mutated in our ICC cohort, resulting reduced SAV1 expression. Moreover, we demonstrated that SAV1 somatic mutation and down-regulation were both predictive of tumor recurrence and shorter OS. Importantly, we revealed through gain-of-function and loss-of-function studies that SAV1 plays a tumor-suppressor role in ICC. We also confirmed that SAV1 inactivation by somatic mutation promotes ICC growth and metastasis.

SAV1 is one of the core kinase components of the Hippo signaling pathway in mammals²⁰. Our results further confirm its role in modulating the Hippo pathway in ICC: SAV1 inactivation suppresses Hippo signaling, leading to YAP activation. Through in vitro and in vivo functional analyses, we confirmed that SAV1 inactivation promotes ICC growth and metastasis by activating YAP. Inactivation of the Hippo pathway has been shown to be associated with poor prognosis in HCC²⁵, yet its role in ICC was largely unknown. Our results revealed that SAV1 mutation leads to inactivation of the Hippo pathway, which in turn promotes ICC growth and metastasis. In addition, our results revealed a tumor-suppressor role for verteporfin, which exhibited an effect similar to that of YAP knockdown in SAV1-inactivated ICC cells. Verteporfin is already approved by the FDA for eye disease indications such as macular degeneration, so it would be relatively easy to adopt for ICC treatment. Hence, our finding of ICC inhibition by verteporfin offers exciting possibilities for the treatment of SAV1-mutated ICC. Notably, the FDA-approved application of verteporfin is based on photodynamic therapy to eliminate neovascularization of blood vessels. Light activation is required for verteporfin to inhibit neovascular angiogenesis, but it is not required for verteporfin to disrupt the interaction between YAP and TEAD³². Therefore, verteporfin might have a dual function inhibiting both angiogenesis and YAP activity, both of which can contribute to inhibiting the progression of SAV1-mutated ICC.

Taken together, our results delineate the genomic landscape that characterizes Chinese ICCs and identify SAV1 as a potential driver.

Methods

Patients and follow-up

The Research Ethics Committee of Zhongshan Hospital granted ethical approval for the use of human subjects (Y2018-060). Each participant gave informed consent in signed form for inclusion in the study. We enrolled and continuously collected tumor and matched non-cancerous liver-tissue samples from 672 Chinese patients with primary ICC who received curative resection from 2010 to 2016 in the Department of Liver Surgical Oncology of Zhongshan Hospital, Fudan University, Shanghai, China. Patients that received palliative surgeries or prior interventions (such as trans-hepatic artery embolization, chemotherapy, or radiotherapy) or that developed other primary malignancies or inflammatory diseases during follow-up were excluded from the study. A total of 204 ICCs were subjected to WES, and the remaining 468 ICCs were subjected to Sanger sequencing. Curative resection was defined as complete resection of tumor nodules, with cancer-free tumor margins shown by histologic examination, and resection of regional lymph nodes, including the hilar, hepatoduodenal-ligament, and caval lymph nodes, with no cancerous thrombus in the portal vein (main trunk or major branches), hepatic veins, or bile duct³³. Patients with further lymph node involvement were considered to have distant metastasis and were excluded from the study³⁴. Tumor differentiation was graded histologically according to the Edmondson–Steiner criteria³⁵. Liver function was graded according to the Child–Pugh system. Tumor stage was determined according to the 2017 International Union against Cancer TNM system. The clinicopathologic characteristics of the patients are listed in Table 1.

Table 1 Clinicopathologic characteristics of patients with intrahepatic cholangiocarcinoma

Full size table

The present study includes follow-up data collected through December 2018. The follow-up procedures are described in detail elsewhere^36,37. We diagnosed tumor recurrence on the basis of computed tomography scans, magnetic resonance imaging, digital subtraction angiography, and elevated serum carbohydrate antigen 19-9 (CA19-9) level, with or without histological confirmation³⁸. We defined disease-free survival as the interval between curative surgery and any diagnosis of recurrence (intrahepatic or extrahepatic)³⁹. We defined overall survival (OS) as the time from the date of curative surgery until death or the end of follow-up. The surviving patients were censored at the end of follow-up.

DNA preparation, DNA capture, and sequencing

Snap frozen tissue samples from tumor and matched non-cancerous liver were obtained and embedded in OCT compound, sectioned by a cryostat, and stained by hematoxylin and eosin. We performed macrodissection to enrich the tumor fraction relative to the dominant stromal component and other normal cells. DNA was extracted using a general protocol for genome sequencing. Preparation of sequencing libraries and DNA capture methods were carried out according to the manufacturer’s protocols.

Whole-exome sequencing

Snap-frozen tissue samples from 204 primary tumors and matched non-cancerous liver from 204 ICC patients were subjected to whole exome sequencing. Sequencing libraries were constructed using a modification of the KAPA Library Preparation Kit. Briefly, 1 μg genomic DNA was sheared to an average fragment size of 200 bp using a Biorupter (Diagenode, Belgium). Fragments were purified using AMPureXP beads (Beckman Coulter Inc, Brae, CA, USA) to remove small products. Then, the DNA was subjected to three enzymatic steps: end repair, A-tailing, and ligation to Illumina paired-end indexed adapters, as outlined in the KAPA Library Preparation Kit protocol (Kapa, Inc.). The libraries were subjected to minimal PCR cycling and were quantified using the Qubit 2.0 Fluorometer (Thermo Fisher, Germany). Libraries were combined into pools for solution phase hybridization using the Roche NimbleGen SeqCap EZ exome enrichment kits V3 (Roche NimbleGen Inc). The captured libraries were analyzed on Agilent’s 2100 Bioanalyzer, and the DNA concentrations were measured using the Qubit 2.0 Fluorometer (Thermo Fisher, Germany) and then sent for sequencing to generate 2 × 150 bp paired-end reads using a HiSeq or Novaseq platform (Illumina, Inc.).

Data quality control

Sequence artifacts; including reads containing adapter contamination, low-quality nucleotides, and unrecognizable nucleotides (N); undoubtedly set a barrier to subsequent reliable bioinformatics analysis. Hence, quality control is an essential step and is applied to guarantee a meaningful downstream analysis. The steps of data processing were as follows:

(1)
Discard a pair of reads if either read contains adapter contamination;
(2)
Discard a pair of reads if it contains poly-N;
(3)
Discard a pair of reads if the proportion of low-quality (Phred quality <5) bases is over 50% in either read.

Reads mapping and detection of somatic genetic alterations

Valid sequencing data was mapped to the reference human genome (UCSC hg19) using the Burrows-Wheeler Aligner (BWA)⁴⁰ software to get the original mapping results stored in BAM format. We performed local realignment of the original BAM alignment using the GATK2⁴¹ and then marked duplicate reads using Sambamba⁴².

Somatic SNVs were detected by muTect1 and muTect2⁴³, and somatic indels were detected by Strelka⁴⁴. High-confidence somatic mutations were called if the following criteria were met: (1) the tumor samples were covered sufficiently (≥10×) and the normal samples were covered ≥8× at the genomic level; (2) the variants were supported by at least 4% of the total reads in the tumor and less than 1% of the total reads in the normal tissue; (3) the variants were supported by at least three reads in the tumor. (4) Only SNVs agreed on by both tools (muTect1 and muTect2) were retained. ANNOVAR⁴ was performed to do annotation for the Variant Call Format obtained in the previous effort.

Tumor copy number variation detection, ploidy, and purity prediction

We used FACETS⁴⁵, an allele-specific copy number analysis tool to detect genome-wide total, allele-specific, and integer DNA copy number, and predict tumor ploidy. The GISTIC (V2.0) algorithm was used to infer recurrently amplified or deleted genomic regions⁴⁶. Genes with a total copy number greater than the gene-level median ploidy were considered gains. Genes with more than twice the median ploidy were considered amplifications. Genes with less than the median ploidy were considered losses. Genes with a total copy number of 0 were considered deletions.

Spectrum and signatures of somatic mutations

We grouped all identified somatic single-nucleotide variants (SNVs) according to the 16 possible flanking nucleotide contexts into 96 trinucleotides (mutated base plus its sequence context)^47,48. We then used a non-negative matrix factorization (NMF) module to decipher the underlying mutation signatures⁴⁹. The overall number of mutation signatures was selected based on the reproducibility of the signatures and low overall reconstruction error. We then compared the inferred signatures to the COSMIC signatures by cosine similarity. We performed hierarchical clustering of the tumors based on the contributions of the mutational signatures in each patient⁵⁰.

Significantly mutated genes

For significantly mutated genes analysis, all identified somatic SNVs and indels were subjected MutSigCV¹³ to identify significantly mutated genes (false discovery rate [FDR] <0.1).

Cancer cell fraction estimation and mutation cluster analysis

For each somatic mutation, the VAF was calculated using the number of reads supporting the variant allele (Rmut) and the number of reads supporting the reference allele (Rnorm; namely, VAF = Rmut/(Rmut + Rnorm)). Then, we calculated the value of cancer cell fractions (CCF) as follows: VAF = p*CCF/(CPNnorm (1−p) + p*CPNmut), where CPNmut indicates the local copy number in the tumor, CPNnorm indicates the local copy number in the normal controls (usually assumed to be 2 except for sex chromosomes), and p indicates the tumor purity in each sequenced sample. The VAF was defined as the VAF of each somatic mutation. CCF is represented as a distribution between 0 and 1. Then all somatic mutations (SNVs and indels) identified were applied to infer the mutation cluster according to PyClone (a Bayesian clustering method)⁵¹.

Estimation of Shannon index

To assess the clonal diversity in each patient, we used the Shannon index of ecological diversity⁵². After we inferred mutation clusters using PyClone, we filtered out clusters with a single mutation and those involving only silent mutations. We then used the remaining mutation clusters to quantitate the Shannon diversity index in each patient¹⁴.

In this analysis, mutations were assumed to accumulate in a linear way, with each mutation giving rise to a new cell population¹⁴. We ordered the mutation clusters by increasing cancer cell fraction (CCF) in each patient. The raw size s_i of each of the N mutations in each patient was estimated as follows:

$$\,{s}_{1}=\, {{CCF}}_{1}$$

(1)

$${s}_{i}={{CCF}}_{i}-{{CCF}}_{i-1}$$

(2)

The relative abundance of each population p_i was then computed as the size of the mutation s_i divided by the sum of the sizes of all N mutations in the patient.

$${p}_{i}=\frac{{s}_{i}}{{\sum}_{i}{s}_{i}}$$

(3)

The resulting relative abundances p_i were then used to derive the Shannon diversity index for the patient.

$$Shannon\,Index=\, {\sum }_{i}{p}_{i}\,\log ({p}_{i})$$

(4)

Sanger sequencing

We randomly selected 1600 somatic non-synonymous SNVs and 156 indels identified by WES. We used Sanger sequencing to validate the selected mutations. All mutations presented in Fig. 2A were also validated by Sanger sequencing. In addition, all coding exons of SAV1 were further screened in an additional 468 ICC samples. Sanger sequencing primers were designed using the Primer3 software (http://frodo.wi.mit.edu/). All mutations identified in tumors were confirmed by independent PCR and Sanger sequencing in the specific tumors and paired normal tissues to determine their somatic nature.

Modeling of the SAV1 and SAV1–MST2 complex structures

Because there was no previously determined full-length SAV1 structure, we obtained the full-length structure of SAV1-HUMAN based on the prediction provided in the AlphaFold predicted-structure database (https://alphafold.ebi.ac.uk/). There was no stable secondary structure for residues 1–197 in the predicted structure, so only the structure of residues 198–383 was shown in the present study. In addition, because of the flexibility of the connection-region loop between the WW1 and WW2 domains, we adjusted the deflection angle of the connection region between the WW1 and WW2 domains with reference to the highly homologous crystal structure of the WW12 domain of mouse SAV1 (PDB code: 7BQG). Based on the crystal structure of the dimer in the WW12 domain (PDB Code: 7BQG) and the crystal structure of the SAV1–MST2 heterodimer complex (PDB Code: 6AO5), we built a structural model of the SAV1–MST2 heterotetramer to more intuitively understand the significance of WW12 dimerization to the function of SAV1. The structural model of the SAV1-MST2 heterotetramer was generated in PyMOL (https://pymol.org/2/). Specifically, to obtain the WW12 domain dimer as described in previous work¹⁹, a WW12 domain dimer was generated using the generate symmetry mates module in PyMOL. Then, the SAV1 structure modeled by AlphaFold was aligned with the WW12 domain dimer in PyMOL to obtain the SAV1 dimer. Subsequently, after aligning the SAV1-MST2 heterodimer complex (PDB Code: 6AO5) with the SAV1 dimer, the angle of the flexible loop between the WW12 domain and the SARAH domain of SAV1 was manually adjusted in PyMOL to ensure that the distance between the two MST2 molecules was reasonable. Finally, a schematic diagram of the structure of the SAV1-MST2 heterotetramer was presented.

Cell lines, animals, and lentiviral vector

One human immortalized, nonmalignant cholangiocyte cell line: H-69, and five ICC cell lines were used in this study: H-69, HuCCT1, SG231, and CCLP1 were kindly provided from Dr. Robert Anders at Johns Hopkins, while HCCC-9810 and RBE cell lines were purchased from the Institute of Biochemistry and Cell Biology (Chinese Academy of Sciences). Four-to-six-week-old male NOD-Prkdc^scid IL2rg^tm1/Bcgen mice were obtained from the Beijing Biocytogen Co., Ltd and maintained under specific pathogen-free conditions. Humane care was provided for all animals in concordance with the criteria described in the “Guide for the Care and Use of Laboratory Animals” (National Institutes of Health publication 86-23, revised 1985). The maximal tumor size permitted by The Research Ethics Committee of Zhongshan Hospital is 2 cm in length for subcutaneous tumor and the tumors of all animals in this study didn’t exceed the criteria.

The following lentiviral vectors were purchased from Shanghai GeneChem Co: wild-type and mutant SAV1 expression vectors and corresponding control lentiviral vector (Ubi-MCS-SV40-firefly_Luciferase-IRES-Puromycin), and the shRNA-SAV1 and its negative control (hU6-MCS-Ubiquitin-firefly_Luciferase-IRES-puromycin). The Ubi-MCS-SV40-firefly_Luciferase-IRES-Puromycin-SAV1 wild-type and mutant vectors were transfected into SG231 cells with low intrinsic SAV1 levels. The hU6-MCS-Ubiquitin-firefly_Luciferase-IRES-puromycin-shRNA-SAV1 was transfected into HuCCT1 cells with high intrinsic SAV1 expression. The Ubi-MCS-SV40-firefly_Luciferase-IRES-Puromycin and hU6-MCS-Ubiquitin-firefly_Luciferase-IRES-puromycin lentiviral vectors were used as controls. Stably transfected clones were validated by immunoblotting.

Cell proliferation, colony formation, and matrigel invasion assays

Cells (2000 cells/well) were seeded in 100 μL of media in a 96-well plate. Then, 10 μL CCK-8 solution (Dojindo) was added to the cells at the indicated time points. The cells were then incubated for an additional 2 h. The numbers of viable cells were determined by measurement of absorbance at 450 nm.

To assess the colony formation abilities of the cells, 500–1000 cells were seeded into each well of six-well plates and incubated at 37 °C for 12–16 days. Then, the cells were fixed with 100% methanol and stained with 0.1% crystal violet. Image-Pro Plus v6.2 (Media Cybernetics) was used to count the megascopic cell colonies.

To assay cell invasion, 24-well Transwell plates with an 8 μm pore size (Minipore) were precoated with Matrigel (BD Biosciences). Then, 100 μL Dulbecco’s modified Eagle medium (DMEM) with 1% fetal bovine serum (FBS) containing 1 × 10⁵ cells was added to the upper chamber. The lower chamber contained 600 μL DMEM with 10% FBS. After 48 h, both the Matrigel and the remaining cells in the upper chamber were removed. Cells that had invaded the lower surface of the membrane were fixed using 4% paraformaldehyde and then stained with Giemsa. Cells from five microscopic (200×) fields were counted.

In vivo assays for tumor growth and metastasis

For tumor growth assay, 100 μL serum-free DMEM and Matrigel (BD Biosciences; 1:1) containing 1 × 10⁷ ICC cells were injected subcutaneously into the upper left flank region of mice. For tumor metastasis assay, mice were placed in a restrainer, and tumor cells (1 × 10⁶ cells in 200 μL serum-free DMEM) were injected through the tail vein using an insulin needle. In some experiment groups for verteporfin treatment, intraperitoneal injection of verteporfin (100 mg/kg) thrice weekly was performed starting 7 days after inoculation. The mice were monitored every 5 days and were sacrificed after 5 weeks. Bioluminescence imaging was performed using an IVIS Lumina K Series III, and image radiance values were normalized using Living Image (Perkinelmer).

RNA isolation and qRT-PCR

Total RNA was extracted from cells using Trizol reagent (Invitrogen) according to the manufacturer’s instructions. The mRNA expression in the cells was assessed with qRT-PCR using an ABI7900HT instrument (Applied Biosystems). Quantitative RT-PCR (qRT-PCR) was performed using the SYBR PrimeScript RT-PCR Kit (Takara Bio). GAPDH was used as an internal control for qRT-PCR. Relative mRNA levels were estimated on the basis of Ct values and normalized according to GAPDH expression using the following equation: 2^−ΔCt(ΔCt = Ct [target gene]−Ct [GAPDH]). All experiments were performed in triplicate.

Western blot and immunofluorescence analysis

Western blotting was performed as described previously⁵³. Briefly, proteins from total cell lysates were separated by 10% SDS-PAGE. The proteins were then transferred to polyvinylidene difluoride (PVDF) membranes. After washing and blocking steps, the membranes were incubated with primary antibodies. After washing, the membranes were incubated with horseradish peroxidase-conjugated secondary antibodies. Antibody binding was detected using enhanced chemiluminescence assays.

For immunofluorescence assays, cells cultured on glass slides were fixed in 4% paraformaldehyde for 15 min. Subsequently, the cells were permeabilized with 0.1% Triton X-100 for 15 min at room temperature, washed with phosphate buffered saline (PBS), and blocked with PBS containing 1% (w/v) bovine serum albumin (BSA) and 0.15% (w/v) glycine (BSA buffer) for 1 h at room temperature. The cells were then treated with primary antibody for 2 h at room temperature. A negative control (primary antibody omitted) was included on every slide. The cells were then washed with BSA buffer and incubated with 2 μg/mL Alexa Fluor 488-conjugated goat anti-mouse antibody (Molecular Probes, Eugene, OR) for 1 h at room temperature. After rinsing in PBS, the slices were counter-stained with diamidino phenylindole and examined by fluorescence microscopy (Leica Microsystems Imaging Solutions, Cambridge, UK).

Immunohistochemistry (IHC) and evaluation of immunohistochemical variables

IHC staining was performed using an avidin-biotin-peroxidase complex as described previously⁵⁴. Briefly, rehydration and microwave antigen retrieval were performed. Then, the slides were incubated with monoclonal antibodies at 4 °C overnight. Next, the slides were incubated with secondary antibody (GK500705, Gene Tech) at 37 °C for 30 min. The slides were stained with 3,3′-diaminobenzidine and then counter-stained with Mayer’s hematoxylin. Slides that were treated identically but without primary antibody served as negative controls.

IHC staining was analyzed by three independent investigators who were blinded to patient characteristics. Any discrepancies were resolved by consensus. Five representative microscope fields were photographed under high-power magnification (200×) using the Leica QWin Plus v3 software. Each image was captured using identical settings. Image-Pro Plus v6.2 software (Media Cybernetics, Inc.) was used to determine density. In each photograph, the integrated optical density of positive SAV1 and YAP staining was determined. The ratio of that density to the total area of each photograph was calculated as the density.

Statistical analysis

Statistical analyses were performed in the R 3.6.2 environment or using SPSS 16.0 for Windows. The data were expressed as the mean ± SD of three independent experiments unless otherwise specified. Student’s t-test was used to compare quantitative data between groups. The chi-square or Fisher’s exact test was used to compare categorical data. The Kaplan–Meier method was used to calculate both the OS and the cumulative recurrence rates. Differences were analyzed by the log-rank test. All statistical tests were two-sided, and P < 0.05 were considered statistically significant.

Reporting summary

Further information on research design is available in the Nature Portfolio Reporting Summary linked to this article.

Data availability

The raw sequence data reported in this paper have been deposited in the Genome Sequence Archive (Genomics, Proteomics & Bioinformatics 2021) in National Genomics Data Center (Nucleic Acids Res 2022), China National Center for Bioinformation/Beijing Institute of Genomics, Chinese Academy of Sciences (GSA-Human: HRA008513). The accession number HRA008513 contains all raw sequencing data (204 WES data) with restrictedly controlled access, which is a property of Zhongshan Hospital, Fudan University, provided to the scientists or researchers through a restricted-access agreement that prevents sharing the dataset with a third party or publicly. Researchers could send data requests for scientific purpose via the website [https://bigd.big.ac.cn/gsa-human/browse/HRA008513] after registration. The correspondence author will check the requests with caution and make the decision whether to permit or not, in 1 month upon receiving the requests. The data will be available for 1 month once access has been granted by the correspondence author. Source data are provided with this paper.

References

Siegel, R. L., Miller, K. D. & Jemal, A. Cancer statistics, 2019. CA Cancer J. Clin. 69, 7–34 (2019).
Article PubMed Google Scholar
Bridgewater, J. et al. Guidelines for the diagnosis and management of intrahepatic cholangiocarcinoma. J. Hepatol. 60, 1268–1289 (2014).
Article PubMed Google Scholar
Rizvi, S. & Gores, G. J. Pathogenesis, diagnosis, and management of cholangiocarcinoma. Gastroenterology 145, 1215–1229 (2013).
Article CAS Google Scholar
Mavros, M. N., Economopoulos, K. P., Alexiou, V. G. & Pawlik, T. M. Treatment and prognosis for patients with intrahepatic cholangiocarcinoma: systematic review and meta-analysis. JAMA Surg. 149, 565–574 (2014).
Article PubMed Google Scholar
Ong, C. K. et al. Exome sequencing of liver fluke-associated cholangiocarcinoma. Nat. Genet. 44, 690–693 (2012).
Article PubMed CAS Google Scholar
Jiao, Y. et al. Exome sequencing identifies frequent inactivating mutations in BAP1, ARID1A and PBRM1 in intrahepatic cholangiocarcinomas. Nat. Genet. 45, 1470–1473 (2013).
Article PubMed PubMed Central CAS Google Scholar
Chan-On, W. et al. Exome sequencing identifies distinct mutational patterns in liver fluke-related and non-infection-related bile duct cancers. Nat. Genet. 45, 1474–1478 (2013).
Article PubMed CAS Google Scholar
Zou, S. et al. Mutational landscape of intrahepatic cholangiocarcinoma. Nat. Commun. 5, 5696 (2014).
Article PubMed CAS Google Scholar
Nakamura, H. et al. Genomic spectra of biliary tract cancer. Nat. Genet. 47, 1003–1010 (2015).
Article PubMed CAS Google Scholar
Jusakul, A. et al. Whole-genome and epigenomic landscapes of etiologically distinct subtypes of cholangiocarcinoma. Cancer Discov. 7, 1116–1135 (2017).
Article PubMed PubMed Central CAS Google Scholar
Wardell, C. P. et al. Genomic characterization of biliary tract cancers identifies driver genes and predisposing mutations. J. Hepatol. 68, 959–969 (2018).
Article PubMed CAS Google Scholar
Ng, A. W. T. et al. Aristolochic acids and their derivatives are widely implicated in liver cancers in Taiwan and throughout Asia. Sci. Transl. Med. 9, eaan6446 (2017).
Article PubMed Google Scholar
Lawrence, M. S. et al. Mutational heterogeneity in cancer and the search for new cancer-associated genes. Nature 499, 214–218 (2013).
Article ADS PubMed PubMed Central CAS Google Scholar
Cerrano, M. et al. Clonal dominance is an adverse prognostic factor in acute myeloid leukemia treated with intensive chemotherapy. Leukemia 35, 712–723 (2021).
Article PubMed CAS Google Scholar
Maley, C. C. et al. Classifying the evolutionary and ecological features of neoplasms. Nat. Rev. Cancer 17, 605–619 (2017).
Article PubMed PubMed Central CAS Google Scholar
Bertucci, F. et al. Genomic characterization of metastatic breast cancers. Nature 569, 560–564 (2019).
Article ADS PubMed CAS Google Scholar
Bae, S. J. et al. SAV1 promotes Hippo kinase activation through antagonizing the PP2A phosphatase STRIPAK. eLife 6, e30278 (2017).
Article PubMed PubMed Central Google Scholar
Callus, B. A., Verhagen, A. M. & Vaux, D. L. Association of mammalian sterile twenty kinases, Mst1 and Mst2, with hSalvador via C-terminal coiled-coil domains, leads to its stabilization and phosphorylation. FEBS J. 273, 4264–4276 (2006).
Article PubMed CAS Google Scholar
Lin, Z., Xie, R., Guan, K. & Zhang, M. A WW tandem-mediated dimerization mode of SAV1 essential for Hippo signaling. Cell Rep. 32, 108118 (2020).
Article PubMed PubMed Central CAS Google Scholar
Yu, F. X., Zhao, B. & Guan, K. L. Hippo pathway in organ size control, tissue homeostasis, and cancer. Cell 163, 811–828 (2015).
Article PubMed PubMed Central CAS Google Scholar
Zhang, W. et al. Genetic features of aflatoxin-associated hepatocellular carcinoma. Gastroenterology 153, 249–262.e242 (2017).
Article PubMed CAS Google Scholar
Turajlic, S. et al. Tracking cancer evolution reveals constrained routes to metastases: TRACERx renal. Cell 173, 581–594.e512 (2018).
Article PubMed PubMed Central CAS Google Scholar
Andor, N. et al. Pan-cancer analysis of the extent and consequences of intratumor heterogeneity. Nat. Med. 22, 105–113 (2016).
Article PubMed CAS Google Scholar
Bochtler, T. et al. Clonal heterogeneity as detected by metaphase karyotyping is an indicator of poor prognosis in acute myeloid leukemia. J. Clin. Oncol. 31, 3898–3905 (2013).
Article PubMed Google Scholar
Sohn, B. H. et al. Inactivation of Hippo pathway is significantly associated with poor prognosis in hepatocellular carcinoma. Clin. Cancer Res. 22, 1256–1264 (2016).
Article PubMed CAS Google Scholar
Matsuura, K. et al. Downregulation of SAV1 plays a role in pathogenesis of high-grade clear cell renal cell carcinoma. BMC Cancer 11, 523 (2011).
Article PubMed PubMed Central CAS Google Scholar
Jiang, J. et al. SAV1 represses the development of human colorectal cancer by regulating the Akt-mTOR pathway in a YAP-dependent manner. Cell Prolif. 50, e12351 (2017).
Article PubMed PubMed Central Google Scholar
Brandt, Z. J., Echert, A. E., Bostrom, J. R., North, P. N. & Link, B. A. Core Hippo pathway components act as a brake on Yap and Taz in the development and maintenance of the biliary network. Development 147, dev184242 (2020).
Article PubMed PubMed Central CAS Google Scholar
Lee, K. P. et al. The Hippo-Salvador pathway restrains hepatic oval cell proliferation, liver size, and liver tumorigenesis. Proc. Natl Acad. Sci. USA 107, 8248–8253 (2010).
Article ADS PubMed PubMed Central CAS Google Scholar
Farshidfar, F. et al. Integrative genomic analysis of cholangiocarcinoma identifies distinct IDH-mutant molecular profiles. Cell Rep. 18, 2780–2794 (2017).
Article PubMed PubMed Central CAS Google Scholar
Sia, D. et al. Integrative molecular analysis of intrahepatic cholangiocarcinoma reveals 2 classes that have different outcomes. Gastroenterology 144, 829–840 (2013).
Article PubMed CAS Google Scholar
Liu-Chittenden, Y. et al. Genetic and pharmacological disruption of the TEAD-YAP complex suppresses the oncogenic activity of YAP. Genes Dev. 26, 1300–1305 (2012).
Article PubMed PubMed Central CAS Google Scholar
Zhou, S. L. et al. CXCL5 contributes to tumor metastasis and recurrence of intrahepatic cholangiocarcinoma by recruiting infiltrative intratumoral neutrophils. Carcinogenesis 35, 597–605 (2014).
Article PubMed CAS Google Scholar
Zhou, X. D. et al. Intrahepatic cholangiocarcinoma: report of 272 patients compared with 5,829 patients with hepatocellular carcinoma. J. Cancer Res. Clin. Oncol. 135, 1073–1080 (2009).
Article PubMed Google Scholar
Wittekind, C. [Pitfalls in the classification of liver tumors]. Pathologe 27, 289–293 (2006).
Article PubMed CAS Google Scholar
Zhou, S. L. et al. Genomic sequencing identifies WNK2 as a driver in hepatocellular carcinoma and a risk factor for early recurrence. J. Hepatol. 71, 1152–1163 (2019).
Article PubMed CAS Google Scholar
Zhou, S. L. et al. Tumor-associated neutrophils recruit macrophages and T-regulatory cells to promote progression of hepatocellular carcinoma and resistance to Sorafenib. Gastroenterology 150, 1646–1658.e17 (2016).
Article PubMed CAS Google Scholar
Zhou, Z. et al. Tumor-associated neutrophils and macrophages interaction contributes to intrahepatic cholangiocarcinoma progression by activating STAT3. J. Immunother. Cancer 9, e001946 (2021).
Article PubMed PubMed Central Google Scholar
Zhou, S. L. et al. A positive feedback loop between cancer stem-like cells and tumor-associated neutrophils controls hepatocellular carcinoma progression. Hepatology 70, 1214–1230 (2019).
Article PubMed CAS Google Scholar
Li, H. & Durbin, R. Fast and accurate short read alignment with Burrows-Wheeler transform. Bioinformatics 25, 1754–1760 (2009).
Article PubMed PubMed Central CAS Google Scholar
McKenna, A. et al. The Genome Analysis Toolkit: a MapReduce framework for analyzing next-generation DNA sequencing data. Genome Res. 20, 1297–1303 (2010).
Article PubMed PubMed Central CAS Google Scholar
Tarasov, A., Vilella, A. J., Cuppen, E., Nijman, I. J. & Prins, P. Sambamba: fast processing of NGS alignment formats. Bioinformatics 31, 2032–2034 (2015).
Article PubMed PubMed Central CAS Google Scholar
Cibulskis, K. et al. Sensitive detection of somatic point mutations in impure and heterogeneous cancer samples. Nat. Biotechnol. 31, 213–219 (2013).
Article PubMed PubMed Central CAS Google Scholar
Saunders, C. T. et al. Strelka: accurate somatic small-variant calling from sequenced tumor-normal sample pairs. Bioinformatics 28, 1811–1817 (2012).
Article PubMed CAS Google Scholar
Shen, R. & Seshan, V. E. FACETS: allele-specific copy number and clonal heterogeneity analysis tool for high-throughput DNA sequencing. Nucleic Acids Res. 44, e131 (2016).
Article PubMed PubMed Central Google Scholar
Mermel, C. H. et al. GISTIC2.0 facilitates sensitive and confident localization of the targets of focal somatic copy-number alteration in human cancers. Genome Biol. 12, R41 (2011).
Article PubMed PubMed Central Google Scholar
Alexandrov, L. B. et al. Signatures of mutational processes in human cancer. Nature 500, 415–421 (2013).
Article PubMed PubMed Central CAS Google Scholar
Helleday, T., Eshtad, S. & Nik-Zainal, S. Mechanisms underlying mutational signatures in human cancers. Nat. Rev. Genet. 15, 585–598 (2014).
Article PubMed PubMed Central CAS Google Scholar
Alexandrov, L. B., Nik-Zainal, S., Wedge, D. C., Campbell, P. J. & Stratton, M. R. Deciphering signatures of mutational processes operative in human cancer. Cell Rep. 3, 246–259 (2013).
Article PubMed PubMed Central CAS Google Scholar
Fujimoto, A. et al. Whole-genome mutational landscape and characterization of noncoding and structural mutations in liver cancer. Nat. Genet. 48, 500–509 (2016).
Article PubMed CAS Google Scholar
Roth, A. et al. PyClone: statistical inference of clonal population structure in cancer. Nat. Methods 11, 396–398 (2014).
Article PubMed PubMed Central CAS Google Scholar
Lee, J. Y. et al. Lobular carcinomas in situ display intralesion genetic heterogeneity and clonal evolution in the progression to invasive lobular carcinoma. Clin. Cancer Res. 25, 674–686 (2019).
Article PubMed CAS Google Scholar
Zhou, S. et al. Tacrolimus enhances the invasion potential of hepatocellular carcinoma cells and promotes lymphatic metastasis in a rat model of hepatocellular carcinoma: involvement of vascular endothelial growth factor-C. Transpl. Proc. 43, 2747–2754 (2011).
Article CAS Google Scholar
Zhou, Z. J. et al. HNRNPAB induces epithelial-mesenchymal transition and promotes metastasis of hepatocellular carcinoma by transcriptionally activating SNAIL. Cancer Res. 74, 2750–2762 (2014).
Article PubMed CAS Google Scholar

Download references

Acknowledgements

This study was jointly supported by the National Natural Science Foundation of China (No. 82372985 to S.L.Z., No. 82373418 to Z.J.Z., No. 82273247 to C.L.S., No. 82173260 to S.L.Z., No. 82072681 to Z.J.Z., No. 82003082 to Z.Q.H.), Shanghai Technical Standard Program (21DZ2201100 to S.L.Z.) and Shanghai Medical Innovation Research Project (22Y11907300 to S.L.Z.).

Author information

These authors contributed equally: Zheng-Jun Zhou, Yu-Hang Ye, Zhi-Qiang Hu, Yue-Ru Hou.

Authors and Affiliations

Department of Liver Surgery and Transplantation, Zhongshan Hospital, Fudan University, Shanghai, China
Zheng-Jun Zhou, Yu-Hang Ye, Zhi-Qiang Hu, Kai-Xuan Liu, Rong-Qi Sun, Peng-Cheng Wang, Chu-Bin Luo, Jia Li, Jian Zhou, Jia Fan & Shao-Lai Zhou
Liver Cancer Institute, Zhongshan Hospital, Fudan University, Shanghai, China
Zheng-Jun Zhou, Yu-Hang Ye, Zhi-Qiang Hu, Kai-Xuan Liu, Rong-Qi Sun, Peng-Cheng Wang, Chu-Bin Luo, Jia Li, Ji-Xue Zou, Jian Zhou, Jia Fan & Shao-Lai Zhou
Institute of Cancer Stem Cell, Dalian Medical University, Dalian, Liaoning, China
Yue-Ru Hou & Cheng-Li Song

Authors

Zheng-Jun Zhou
View author publications
Search author on:PubMed Google Scholar
Yu-Hang Ye
View author publications
Search author on:PubMed Google Scholar
Zhi-Qiang Hu
View author publications
Search author on:PubMed Google Scholar
Yue-Ru Hou
View author publications
Search author on:PubMed Google Scholar
Kai-Xuan Liu
View author publications
Search author on:PubMed Google Scholar
Rong-Qi Sun
View author publications
Search author on:PubMed Google Scholar
Peng-Cheng Wang
View author publications
Search author on:PubMed Google Scholar
Chu-Bin Luo
View author publications
Search author on:PubMed Google Scholar
Jia Li
View author publications
Search author on:PubMed Google Scholar
Ji-Xue Zou
View author publications
Search author on:PubMed Google Scholar
Jian Zhou
View author publications
Search author on:PubMed Google Scholar
Jia Fan
View author publications
Search author on:PubMed Google Scholar
Cheng-Li Song
View author publications
Search author on:PubMed Google Scholar
Shao-Lai Zhou
View author publications
Search author on:PubMed Google Scholar

Contributions

Z.J.Z., Y.H.Y., Z.Q.H. and S.L.Z. performed the experiments; Y.R.H., S.L.Z., and C.L.S. analyzed the data; K.X.L., R.Q.S., P.C.W., C.B.L., J.L., and J.X.Z. provided the samples; Z.J.Z., C.L.S. and S.L.Z. wrote the paper; J.Z. and J.F. commented on the study and revised the paper; C.L.S. and S.L.Z. obtained funding and designed the research.

Corresponding authors

Correspondence to Cheng-Li Song or Shao-Lai Zhou.

Ethics declarations

Competing interests

The authors declare no competing interests.

Peer review

Peer review information

Nature Communications thanks Rory Smoot, and the other, anonymous, reviewer(s) for their contribution to the peer review of this work.

Additional information

Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Supplementary information

Supplementary Information

Description of Additional Supplementary Files

Supplementary Data 1-11

Reporting Summary

Source data

Source Data

Rights and permissions

Open Access This article is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License, which permits any non-commercial use, sharing, 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 you modified the licensed material. You do not have permission under this licence to share adapted material derived from this article or parts of it. 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-nc-nd/4.0/.

Reprints and permissions

About this article

Cite this article

Zhou, ZJ., Ye, YH., Hu, ZQ. et al. Whole-exome sequencing reveals genomic landscape of intrahepatic cholangiocarcinoma and identifies SAV1 as a potential driver. Nat Commun 15, 9960 (2024). https://doi.org/10.1038/s41467-024-54387-8

Download citation

Received: 27 January 2022
Accepted: 07 November 2024
Published: 17 November 2024
DOI: https://doi.org/10.1038/s41467-024-54387-8

This article is cited by

Review: Progress of the NLRP3 inflammasome in tumours and perspectives for cholangiocarcinoma
- Zhaoqin Zhuo
- Yuwei Xie
- Chengzhan Zhu
Cell Communication and Signaling (2025)