Cancer risk, autoimmune comorbidity and autoantibodies in primary sclerosing cholangitis

Abstract

Primary sclerosing cholangitis (PSC) is a rare cholestatic liver disease affecting the bile ducts primarily, and the liver parenchyma secondarily. The disease is progressive but highly variable between diseased. Some progress quickly to liver cirrhosis and end-stage liver disease, with need of liver transplantation (LTX), while in others the disease may stay quiescent for years. There is a close association to inflammatory bowel disease (IBD), and especially ulcerative colitis (UC). Comorbidity with autoimmune disease (AID) together with elevated levels of autoantibodies (AAB) is also commonly described. Pathogenesis in PSC is ambiguous, where genetic studies link PSC closer to other AIDs than IBD. The importance of the comorbidities and AABs seen in PSC is, despite numerous studies, not well understood.

The most devastating threat in PSC is the high and unpredictable risk of developing cancer, especially hepatobiliary cancer (HBC), but also colorectal cancer (CRC) for those with a concomitant IBD. Cholangiocarcinoma (CCA) has a poor prognosis with a low estimated survival. All individuals with PSC are recommended yearly surveillance. Better understanding of risks and pathogenesis in PSC may help in, clinical evaluation, and prognostication. This thesis aimed to contribute to the knowledge on both cancer risk and comorbidity in PSC, along with an assessment on the role of the colon for disease development. Evaluation on family autoimmune inheritance and a complete proteomic analysis on autoantibodies linked to different phenotypes of the disease may contribute to a better understanding the pathogenesis.

A large cohort of individuals with a well-defined diagnosis of PSC (n=1 768) in Sweden were matched on sex, age, and municipality to a random reference cohort 10:1 from the general population in Sweden. Data on first-degree relatives, diagnoses of cancer, AID, LTX and death were retrieved from the Swedish national registers. This cohort may have comprised the prevalent population of individuals with PSC in Sweden which together with high coverage in Swedish registers are a stable fundament for these studies. This cohort was used for Studies I and III and serum from a smaller part of the cohort (n=500) was used for Study IV.

In Study I we estimated risk of cancer in PSC, with the hypothesis that PSC increases the risk for other than gastrointestinal cancers. Start of follow up commenced at date of diagnosis and the cohort was followed until December 31, 2016, or any of the following events: cancer diagnosis, LTX, first emigration date or death. We reproduced previous findings with high rates of HBC and CRC. Higher risks for pancreatic cancer and lymphoma were also detected. Analyzing the potential confounder of IBD in Study I led us to the aim of Study II; to evaluate the role of the bowel for development of future PSC. We evaluated all UC patients in Sweden (n=61 993) between 1990-2018 where detection of UC was performed according to a validated algorithm. In the cohort, 5 577 colectomies were performed, at which time point matching was performed and follow-up began. Matched comparators were achieved from the remaining noncolectomized UC cohort. The total cohort was followed over a median time of 13 years covering a period where great medical evolvement in IBD emerged. Overall, the incidence of PSC remained the same over the years, and we detected no difference between those with and without remaining colon. In a subgroup analysis discrepancies on incidence of PSC, related to age at UC diagnosis and colectomy as well as period of colectomy and time to colectomy, were detected. Another factor that we speculated may have influenced risk of cancer in Study I was the autoimmune linkage. In Studies III and IV we intended to tackle the longstanding idea of PSC as an AID by evaluating autoimmune comorbidity in both individuals with PSC and their first-degree relatives. For a part of the PSC cohort, a full-genome screening of AABs was performed. High odds for several AIDs were seen in PSC and in their first-degree relatives, which remained in first-degree relatives also when removing cases with the disease of interest. We quantified 1 153 selected autoantibodies, in 500 PSC and 220 controls as an explorative initiative. AABs were present in 5-10% of the cohort with a large variability. No single AAB marker was detected but rather clusters of AABs associated to different phenotypes and severity of disease.

In conclusion we confirmed previous known high risks of HBC and CRC along with an increased risk for lymphoma which has not been presented previously. Removal of the colon was not seen to prevent development of PSC on a group level, but the importance of colectomy for subgroups shall not be ruled out. Autoimmune comorbidity in PSC is common but does not seem to influence severe outcome. Both autoimmune heredity in PSC and occurrence of AABs support pathogenetic findings where some clusters of AAB may be of prognostic value.