CK‐18 cell death markers improve the prediction of histological remission in autoimmune hepatitis during biochemical remission

Incomplete histological remission of autoimmune hepatitis (AIH) is associated with a reduced long‐term survival and an increased relapse rate even during biochemical remission (BR). The aim of this international multicentre study was to explore the diagnostic fidelity of cytokeratin‐18 cell death markers to noninvasively detect incomplete histological remission. Thereby, cytokeratin‐18 cell death marker M65 but not ALT and immunoglobulins was significantly higher in patients with incomplete histological remission (mHAI ≥ 4) compared to those with mHAI ≤ 3. M65 levels > 305 U/L, identified in the training cohort, facilitated the noninvasive detection of incomplete histological remission with a sensitivity of 75% and negative predictive value of 86% in the validation cohort. While BR with M65 < 305 U/L suggested complete histological remission (86%), BR with M65 > 305 U/L reduced the rate of histological remission to 60%. In conclusion, M65 may help to better select patients for or to reduce surveillance liver biopsies in the future.


| INTRODUC TI ON
Autoimmune hepatitis (AIH) is a chronic progressive autoimmune disorder. Although the natural disease course is detrimental, immunosuppressive therapy can dramatically improve the life expectancy.
In order to prevent histological disease progression, complete normalization of liver enzymes and immunoglobulin G (IgG) is recommended by current guidelines. [1][2][3] This biochemical remission (BR) can be achieved in 40%-80% of patients. 2,4,5 However, persistent necro-inflammatory activity in the liver (modified hepatitis activity index, mHAI ≥ 4) during BR was associated with a reduced long-term prognosis in a recent retrospective study. 6 The current guidelines suggest a liver biopsy for the assessment of histological remission for the evaluation of treatment withdrawal after at least 3 years of therapy with at least 2 years of BR. 2,3 The same cut-off of mHAI ≥ 4 is associated with relapses after stopping immunosuppressive therapy. 7 So far, noninvasive clinical tools to assess persistent histological inflammation during BR are completely lacking.
Beyond the classical liver enzymes like aminotransferases, the hepatocyte cell death can be measured by the release of cytokeratin-18 (CK-18) fragments. 8 While caspase-cleaved CK-18 indicates apoptosis, measured by the M30 ELISA, total CK-18 is also released during necrosis, measured by the M65 ELISA. The clinical utility of CK-18 cell death biomarkers was studied in multiple acute and chronic liver diseases. Thereby, cell death markers were helpful to noninvasively identify steatosis, fibrosis and the presence of metabolic (dysfunction)-associated fatty liver diseases (MAFLD). 9,10 They further improved the prediction of the prognosis in acute and chronic liver failure. [11][12][13] Additionally, higher cell death biomarkers were associated with more advanced liver fibrosis and a worse prognosis in primary biliary cholangitis (PBC). 14 In AIH, cell death biomarkers were associated with elevations of liver enzymes in a recent study that did not assess the histological disease activity. 15 The current multicentre study assessed the predictive capacity of CK-18 cell death biomarkers to noninvasively detect a relevant persistent necro-inflammatory activity in the liver (mHAI ≥ 4) in comparison to routine laboratory parameters in adult patients with AIH that achieved BR under maintenance immunosuppressive therapy.

| Patients
We included 60 adult patients with AIH without evidence for variant syndromes, who had a surveillance liver biopsy under ongoing immunosuppressive therapy while they were in BR as defined by alanine aminotransferase (ALT) and IgG levels within the normal range (Hamburg/Germany: n = 2; Hannover/Germany: n = 32; Larissa/ Greece: n = 26). In order to assess the diagnostic fidelity of M65 for the non-invasive detection of mHAI ≥ 4, cohort was divided into a training (n = 31) and a validation cohort (n = 29). The patient data at the time point of the surveillance biopsy are summarized in Table 1.
This study was approved by the local research Ethics Committee of Hannover Medical School. Written informed consent was obtained from all patients from the prospective biomaterial repository of Hannover Medical School (approval number 5582). The use of material and data from external patients in the multicentre cohort was approved by the respective local ethical committees.

| Histology
Biopsies were processed and histological scoring for the mHAI and fibrosis was performed by an experienced liver pathologist in a blinded fashion. 16

| Serological detection of caspase-cleaved and total cytokeratin-18
Blood samples were preferentially taken at the day of admission before the performance of the liver puncture or within the first 24 hours thereafter. Blood plasma was cryo-conserved at −80°C.
Caspase-generated neoepitopes of CK-18 were measured in cryo-conserved plasma, taken within 24 hours before or after the biopsy, as recently published. 9 In short, the M30-Apoptosense and M65 ELISAs were used according to the manufacturer's instructions (Peviva, Bromma, Sweden).
The other laboratory parameters (ALT and IgG) were measured in the course of the routine clinical surveillance of the patients.

| Statistical analysis
Statistical analysis was performed with SPSS 15.0 and GraphPad Prism 5. Mann-Whitney U tests were used for comparison of two groups. The area under the receiver operating curve (AUC) was calculated and the Youden's index was used for the identification of cut-off values. P-values below .05 (two-tailed) were considered statistically significant in all analyses.
In summary, histological remissions is found in 86%-92% of patients with BR and M65 < 305 U/L, while histological remission is found in only 56%-60% of patients with BR and M65 > 305 U/L in the training and validation cohorts ( Figure S2).

| D ISCUSS I ON
Liver biopsy remains the 'gold standard' for the assessment of staging (fibrosis) and grading (inflammation) of chronic liver diseases in order to guide and individualize therapy. However, liver biopsies have always a bleeding risk and a probability of sampling error, while they are expensive and unpopular among patients. Therefore, noninvasive methods are being extensively investigated for the assessment of staging and grading in chronic liver diseases.
While transient elastography (TE) can be used to identify MAFLD/nonalcoholic steatohepatitis patients with advanced fibrosis, the method cannot distinguish between inflammatory activity in nonfibrotic patients. 17 Hartl et al 18 showed that liver fibrosis and inflammation can also be measured noninvasively with TE in AIH patients. In this trial, at diagnosis and especially in clinical settings of severe hepatitis, TE rather correlated with inflammation than with fibrosis and vice versa. After prolonged immunosuppressive therapy, TE correlated with fibrosis failing to detect persistent inflammation. 18 Dhaliwal et al from Sheffield were the first to report a negative impact of incomplete histological remission (mHAI ≥ 4) in the liver during BR on the survival of AIH patients. 6 Although confirmation from other centres is pending, the Sheffield cohort has a unique size and a unique long-term follow-up over several decades. 6,19 They reported higher ALT levels in patients with persistent histological activity during BR, but the difference in the aminotransferases in both cohorts was just 5-6 U/L within the normal range. 6  In summary, CK-18 cell death marker M65 could be a promising sensitive, noninvasive surveillance tool for the individual guidance of a long-term therapy in AIH in the majority of patients who reach BR.
M65 may help to reduce surveillance liver biopsies in AIH patients with a good treatment response.