New risk factors and molecular landscapes of hepatic angiosarcoma in the Taiwanese population†

Hepatic angiosarcoma is a rare, highly aggressive malignancy of the liver. The tumorigenesis of hepatic angiosarcoma has been relatively understudied in terms of aetiology and molecular properties. A recent study published in The Journal of Pathology revealed a strong association between hepatic angiosarcoma incidence and chronic kidney disease, particularly in end‐stage renal disease using population‐based data from the National Health Insurance Research Database in Taiwan and an institutional cohort. The study also revealed enrichment in the mutational signature of aristolochic acid exposure and is the first reported observation of this mutational signature in human sarcomas. © 2023 The Pathological Society of Great Britain and Ireland.

No conflicts of interest were declared.
In a recent issue of The Journal of Pathology, Huang et al [1] reported that chronic kidney disease (CKD) and end-stage renal disease (ESRD) are risk factors for hepatic angiosarcoma (HAS) in the Taiwanese population.A significant proportion of ESRD-associated HAS analysed in this study exhibited high tumour mutation burden with a predominance of A:T>T:A transversionenriched COSMIC mutational signature SBS22 [1], which is indicative of exposure to aristolochic acid (AA) [2,3].
In this study, the authors observed that HAS was the most common subtype of visceral angiosarcoma (13/28 cases, 46%) within a cohort of 77 angiosarcomas of various locations and that there was a significant enrichment of patients with ESRD among HAS patients [1].This observation was consistent with the authors' findings from the Taiwanese National Health Insurance Research Database (NHIRD), which revealed that HAS made up a significant proportion (59%) of visceral angiosarcomas and that CKD/ESRD was a risk factor for HAS.While chronic liver diseases and diabetes mellitus (DM) were also independently overrepresented in patients with HAS, further analysis indicated that the prevalence rates of DM and chronic liver disease were not significantly higher in CKD/ESRD-associated HAS compared to HAS patients with normal renal function, demonstrating that these factors are not required for the development of CKD/ESRD-associated HAS [1].As such, the authors hypothesised that an uncharacterised aetiological agent of HAS in Taiwanese patients was also inducing a concomitant deterioration of renal function.
Using whole-exome sequencing, the authors analysed 27 HAS tumour-normal specimen pairs and observed a high tumour mutation burden (TMB) in these cancers relative to other cancers profiled by The Cancer Genome Atlas (TCGA).While the authors had not explored immune cell signatures or neoantigen expression in this dataset, high-TMB liver and urothelial cancers have been shown to exhibit elevated PD-L1, higher numbers of neoantigens, and increased immune cell infiltration [4].The high TMB observed in HAS therefore suggests that ESRD-associated HAS patients may be candidates for immune checkpoint blockade therapy.
Mutational signature analysis showed a dominant contribution of the SBS22-like signature in 59% (16/27) of patients, while a dominant contribution of the SBS1-like signature (due to spontaneous deamination of 5methylcytosine) occurred in 16% (5/27) of patients.Additionally, SBS22-like A:T>T:A transversion mutations were most enriched in cases with concomitant ESRD, followed by cases with CKD and cases with normal renal function.Frequent SBS22-like mutations were observed in the tumour suppressor genes TP53 and ATRX, suggesting that these mutations play important roles in HAS tumourigenesis.Notably, mutational signature analysis of non-HAS primary angiosarcoma tumour samples in an earlier study did not reveal dominance of the AA mutational signature SBS22, instead displaying predominantly the signatures of CpG demethylation (SBS1, ageing), of UV light exposure (SBS7), and of defective mismatch repair [5].
This paper follows up on several similar studies conducted using the Taiwanese NHIRD, where systematic prescription records of AA-containing herbal products have allowed detailed investigations, revealing that the consumption or handling of AA-containing herbal products were associated with increased risks of urinary tract cancer [6][7][8], kidney failure [9,10], and renal cell carcinoma [11].This study provides important evidence that AA exposure is a possible aetiological agent of HAS and reports for the first time the SBS22 signature in human sarcomas in addition to a number of previously described human carcinomas with SBS22 (Figure 1).The authors' findings that SBS22-like mutations were frequently observed in key tumour suppressor genes suggest that AA exposure and associated resulting mutations play a causal role in the carcinogenesis of HAS.
Intriguingly, while the mechanisms of AA carcinogenicity and genotoxicity have been well described using in vitro and in vivo systems (reviewed in [12]), it remains unclear whether the carcinogenic effects of AA are cellor organ-specific.In human cancers, the enrichment of SBS22-like mutations in cancer driver genes is frequently observed in urothelial carcinoma of the upper urinary tract (UTUC) [3] and some liver malignancies [13,14], suggesting a causal role in these cancer types.In contrast, they made up the minority of driver gene mutations in renal cell carcinoma (RCC), suggesting that AA-associated mutations play a bystander role in these tumours [15].Correspondingly, the authors reported concurrent urothelial cancers in 6% (five out of 83) of the studied cohort of HAS patients (five out of 83, or 6%) and postulated that AA might be exerting a tumorigenic field effect [1].While this concurrence rate was deemed 'unusually high', it is intriguing that the majority of HAS patients evaded other tumour types that are more frequently associated with AA exposure (Figure 1) given the dominance of the SBS22-like mutational signature in 59% of the sequenced cases.It remains to be seen whether this observation may be due to rare genetic alterations in this group of HAS patients that result in their predisposition to HAS.However, further Figure 1.The growing family of AA-associated cancer sites.The figure summarises observations of the SBS22-like signature in wholegenome or whole-exome sequencing reports of renal cell carcinoma (RCC), oesophageal squamous cell carcinoma (ESCC), hepatocellular carcinoma (HCC), intrahepatic cholangiocarcinoma (ICC), upper tract urothelial carcinoma (UTUC), and bladder cancers (reviewed in [12]), as well as the present study [1].The circle sizes show the relative proportions of tumours reported to harbour the AA-associated mutational signature (SBS22) among all the anatomical site-specific samples analysed.Parts of the figure use images obtained from Servier Medical Art by Servier, licensed under a Creative Commons Attribution 3.0 Unported License (https://creativecommons.org/licenses/by/3.0/).

2
F Chung and J Zavadil studies are warranted to determine whether the carcinogenic effects of AA in HAS may be modulated by other risk factors for hepatic cancers such as viral infections, genetic makeup, or alcohol consumption [16] and to establish by the use of independent, highly-specific exposure markers such as aristolactam-DNA adducts that the reported SBS22-like signature does not originate from other chemical exposures inducing similar mutational patterns (discussed in [12] and references therein).
It is especially concerning that the mutational signature indicative of AA exposure continues to be observed in an increasing number of tumour types and with increasing frequency, despite AA-containing herbal products being prohibited (in Taiwan since 2003).While this could be due to more widespread practice of genomic profiling, particularly in previously underrepresented populations [12], it may also be due to ongoing and hitherto unknown routes of AA-exposure or similar genotoxic chemicals.Environmental contamination of soil and groundwater with AA has been shown to be stable, with subsequent studies indicating that AA may accumulate in food crops cultivated in AA-contaminated soil [12].
In sum, this study by Huang et al [1] provided evidence of a strong association between HAS and chronic kidney disease, particularly ESRD.Additionally, genome-scale analysis of ESRD-associated and non-ESRD-associated HAS tumours revealed a relatively high mutation burden, indicating that these patients may be good candidates for immunotherapy.This study also provided the first documented report of the AA mutational signature SBS22 in human sarcomas.