WO2016120632A1

WO2016120632A1 - Assay

Info

Publication number: WO2016120632A1
Application number: PCT/GB2016/050199
Authority: WO
Inventors: Simon AFFORD; Annika WILHELM
Original assignee: The University Of Birmingham
Priority date: 2015-01-30
Filing date: 2016-01-29
Publication date: 2016-08-04
Also published as: GB201501612D0

Abstract

The invention provides a method of diagnosis or prognosis of a chronic liver disease or inflammatory bowel disease (IBD) in a subject comprising providing a sample from the subject and measuring the amount of TWEAK (tumour necrosis factor-like weak inducer of apoptosis) in the sample, wherein if the amount of TWEAK is below a predetermined level, then there is an increased risk that the subject has a chronic liver disease or has a poor prognosis of the chronic liver disease, or if the amount is above a predetermined level then there is an increased risk that the subject has IBD. The invention also provides a method of determining the effect of treatment on a chronic liver disease comprising providing measuring the amount of TWEAK in a first sample from a subject, provided before the treatment and comparing it to an amount of TWEAK in a second sample from a subject provided after treatment, wherein an increase in TWEAK between the first and second samples indicates that the chronic liver disease is being treated. Assay kits for use in the invention are also provided.

Description

Assay

The invention provides a method of diagnosis and prognosis of chronic inflammatory liver diseases, such as primary sclerosing cholangitis (PSC) and inflammatory bowel disease (IBD), to methods of following the treatment of chronic inflammatory liver diseases and to assay kits for use in such assays.

TWEAK (Tumour necrosis factor-like weak inducer of apoptosis) is also known as TNFSF12, AP03L and CD255. It is a multifunctional cytokine implicated in physiological tissue regeneration and tissue repair (Winkles J. A. Nat. Rev. Drug. Discov (2008) 7 (5) 411-425, Cheng E. et al Frontiers in Immunology (2013) 4, Article 473, 1-13). TWEAK is initially synthesised as a membrane anchored protein but furin cleavage within the stalk region generates a secreted, soluble, isoform. TWEAK isoforms are known to bind a small surface receptor, Fnl4 which triggers a number of signalling pathways. Sustained Fnl4 signalling has been implicated in cerebral ischaemia, chronic inflammatory diseases and cancer. These include multiple sclerosis, rheumatoid arthritis, stroke and lupus nephritis. See also Burkley L.C. et al (Immunol. Rev. (2011), 244, 99-114).

TWEAK has also been implicated in atherosclerosis through Fnl4 and its effect on epithelial and endothelial cells. It has also been observed in kidney, muscle and brain (Burkly Supra).

In liver, the TWEAK/Fnl4 pathway is upregulated after partial hepatectomy suggesting that it is involved in liver regeneration.

The cDNA and proteins sequences of TWEAK are known in the art, as are antibodies against TWEAK and immunoassays for TWEAK (see for example Chicheportiche Y et al J. Biol. Chem (1997) 272, 32401-42410 and Marsters S.A. et al Curr. Biol (1998), 8, 525-528), incorporated herein by reference.

TWEAK assays are commercially available, for example from eBioscience Ltd, Hatfield, UK. There are a number of related chronic diseases, such as chronic inflammatory diseases of the liver.

Primary sclerosing cholangitis (PSC) is characterised by chronic inflammation and stricture formation of the biliary tree, progressing to cirrhosis and end stage liver failure. The majority (>80%) of patients with PSC develop inflammatory bowel disease (IBD), which is a factor associated with increased frequency of colorectal and hepatobiliary malignancy. The inflammatory, obstructive and pre-malignant components of the disease result in unpredictable individual outcomes and effective, non-invasive prognostic markers have been previously desperately lacking in the art. Chapman R. et al AASLD Practice Guidelines, Hepatel. (2010) 51 (2) 660 summarises the prior art diagnosis and management of PSC.

Primary Biliary Cirrhosis (PBC) is sometimes considered to be a model autoimmune disease because of its serological signature, antimitochondrial antibody (AMA) and specific bile duct pathology (Lindor K.D. et al Hepatology (2009) 50(1), 209-308).

Autoimmune hepatitis (AIH) is a disease that occurs when the body's immune system attacks cells of the liver. Anomolous presentation of HLA Class II on the surface of hepatocytes, causes a cell-mediated immune response against the body's own liver (Alvarez F. et al J. Hepatol., J. Hepatol. (1999) 31(5), 929-38).

Alcoholic Liver Disease (ALD) encompasses hepatic manifestations of alcohol overconsumption, including fatty liver, alcoholic hepatitis and chronic hepatitis with fibrosis or cirrhosis (O'Shea R.S. et al, Hepatology (2010), 51(1) 307-328).

Non-alcoholic steatohepatitis (NASH) resembles ALD but is not associated with excessive alcohol intake. It is characterised by inflammation of the liver with concurrent fat accumulation (Vuppalanchi R. et al, Hepatology (2009), 49(1) 306-17).

PSC, PBC, AIH, ALD and NASH are chronic inflammatory diseases and are related by similar disease responses and for which TWEAK levels have now been found to be a marker. Inflammatory Bowel Disease (IBD) is a term used to describe Crohn's disease and ulcerative colitis, a group of inflammatory diseases of the colon and small intestine (Baumgart D.C. et al Lancet (2007) 369, 1627-40) for which TWEAK has also now been found to be associated.

Hepatocellular Carcinoma (HCC) is also called malignant hepatoma. It is often secondary to hepatitis B or C infection or cirrhosis.

The Inventors have unexpectedly found that TWEAK, especially soluble TWEAK can be used as a diagnostic and prognostic marker of chronic liver diseases. It has also been found to be a marker of irritable bowel disease.

A first aspect of the invention provides a method of diagnosis or prognosis of a chronic liver disease or IBD in a subject comprising providing a sample from the subject and measuring the amount of TWEAK in the sample, wherein if the amount of TWEAK is below a predetermined level, then there is an increased risk that the subject has a chronic liver disease or has a poor prognosis of the chronic liver disease, or if the amount is above a predetermined level then there is an increased risk that the subject has IBD. If the level is below the predetermined level then the subject may then be treated by liver transplant.

A further aspect provides a method of determining the effect of treatment on a chronic liver disease comprising measuring the amount of TWEAK in a first sample from a subject provided before the treatment, and comparing it to an amount of TWEAK in a second sample from a subject provided after treatment, wherein an increase in TWEAK between the first and second samples indicates that the chronic liver disease is being treated.

The chronic liver disease may be a chronic inflammatory liver disease.

The patient may be treated with anti-inflammatory drugs or cytoprotective bile acids (Ursodeoxycholic acids) which slow progression in a proportion of patients. Those that progress towards end stage disease require transplantation. A still further aspect provides a method of predicting whether a subject may need a liver transplant, comprising measuring an amount of TWEAK is a sample from the subject, wherein the amount of TWEAK is an indication of the risk of the subject needed a transplant. Typically the lower the level of TWEAK, the greater the risk of needing a transplant.

The disease may be IBD. The effect of treatment on IBD may similarly be monitored with a decrease in TWEAK to within the normal range showing efficacy of treatment of the IBD.

More typically the disease which is assayed for is a chronic inflammatory liver disease. Typically the disease is selected from autoimmune cholangiopathies, such as PSC or PBC, fatty liver diseases such as ALD or NASH; typically PSC, PBC, AIH, ALD, or NASH, more typically PSC or PBC, especially PSC.

Typically the TWEAK is soluble TWEAK. It may be monomeric or trimeric soluble TWEAK. Commercially available antibodies from eBioscience Ltd, bind to it. The antibody is typically polyclonal so it detects several epitopes on the antigen.

The amount of TWEAK in the sample may be assayed using a specific binding reagent. The specific binding reagent may be a TWEAK- specific antibody or a TWEAK- specific fragment thereof, such as Fab, F(ab')2 or svFv.

Antibodies against TWEAK are commercially available from a number of companies such as PeproTech EC Ltd London UK and eBioscience, Hadfield, UK.

Typically an immunoassay, such as ELISA or Western blotting, is used to detect the amount of TWEAK.

ELISA, for example uses antibodies to detect specific antigens. One or more of the antibodies used in the assay may be labelled with an enzyme capable of converting a substrate into a detectable analyte. Such enzymes include horseradish peroxidase, alkaline phosphatase and other enzymes known in the art. Alternatively, other detectable tags or labels may be used instead of, or together with, the enzymes. These include radioisotopes, a wide range of coloured and fluorescent labels known in the art, including fluorescein, Alexa fluor, Oregon Green, BODIPY, rhodamine red, Cascade Blue, Marina Blue, Pacific Blue, Cascade Yellow, gold; and conjugates such as biotin (available from, for example, Invitrogen Ltd, United Kingdom). Dye sols, metallic sols, chemiluminescent labels or coloured latex may also be used. One or more of these labels may be used in the ELISA assays according to the various inventions described herein, or alternatively in the other assays, labelled antibodies or kits described herein.

The construction of ELISA-type assays is itself well known in the art. For example, a "binding antibody" specific for the TWEAK is immobilised on a substrate. The "binding antibody" may be immobilised onto the substrate by methods which are well known in the art. TWEAK in the sample are bound by the "binding antibody" which binds the TWEAK to the substrate via the "binding antibody".

Unbound immunoglobulins may be washed away.

In ELISA assays the presence of bound immunoglobulins may be determined by using a labelled "detecting antibody" specific to a different part of the TWEAK of interest than the binding antibody.

Alternatively, the amount of mRNA encoding TWEAK may be measured, for example using quantitative PCR (qPCR). The TWEAK- specific binding agent may therefore be a TWEAK- specific primer. qPCR is generally known in the art. The sample is typically a liver biopsy for qPCR. Probes and primers for TWEAK

(TNFSF12) are commercially available from Roche, such as:

Probe#

Target Forward primer (5'→3') Reverse primer (5'→3') (Roche)

TNFSF12 ATCGCAGCCCATTATGAAGT CTCACTGTCCCGTCCACAC 85 The sample may be selected from whole blood, serum, plasma, urine, bile or a liver tissue biopsy sample, especially whole blood, serum or plasma.

Typically the normal levels of TWEAK, for example for a healthy patient without the disease is 942 pg/ml (IQR 885-1044) for serum. These may be used as the predetermined levels.

A level of typically above a predefined serum level elevated above the normal range ( see earlier)/ml suggests the presence of IBD.

A level of below 275 pg/ml (IQR 111-410) OR <330 pg/ml for serum is the diagnostic of the chronic inflammatory liver disease, or is prognostic of a poor outcome.

A poor prognostic outcome suggests that the subject should be treated for the disease or alternatively triaged.

The method may be used in combination with one or more other assays previously used for liver disease. For example, the UKELD (U.K. model for end-stage liver disease) tests creatinine, the prothrombin time normalised ratio (INR), and bilirubin. Aspartate amniotransferase (AST), Alanine aminotransferase (ALT), alkaline phosphatase (ALP) and gamma glutamyl transpeptidase (GGT) may also be used. Sodium, IgG and albumin may also be tested for.

Liver- specific antoantibodies may be detected. For example, antinuclear antibodies (ANA), anti smooth muscle antibodies (ASMA) and perinuclear anti neutrophil cytoplasmic antibodies (pANCA) are generally known in the art as are assay for such autoantibodies also generally known in the art. Such autoantibodies may be detected using, for example antibodies, or fragments thereof, specific for such autoantibodies.

The subject is typically a human subject.

Assay kits for use in the invention are also provided. The invention provides an assay kit for use in the method of the invention, wherein the disease is a chronic inflammatory liver disease, the kit comprising a TWEAK- specific binding agent and one or more reagents for the detection of liver function or liver disease.

It may comprise an assay kit according to the invention, comprising a reagent for the detection of an autoantibody selection from ANA, ASMA and pANCA.

The kit may comprise an assay kit according to the invention, comprising a reagent for specifically measuring one or more of AST, ALP, ALT, bilirubin, creatinine, prothrombin or gamma glutamyl transpeptidase.

The reagents may be specific for measuring the analyte and may include one or more analyte specific binding reagents, such as antibodies or fragments thereof or may include substrates specific for the analyte.

The invention will now be described by way of example only with reference to the following figures:

Figure 1 Soluble TWEAK levels in serum samples from healthy controls, IBD and PSC patients

Soluble TWEAK levels were measured by ELISA of serum samples from healthy controls (HC, n=22) and patients with IBD (n=24) or with PSC (n=64) (A). sTWEAK levels stratified by PSC disease stages of pre-cirrhotic (n=22) and cirrhotic (n=42) (B). (* * p<0.01, * *** p<0.0001; Mann-Whitney U test). The distribution of values was significantly different across the whole population (p<0.0001 (A only) Kruskal- Wallis test).

Figure 2 Accuracy of sTWEAK at predicting transplant-free survival of PSC patients Area under the receiver operator characteristic (AUROC) curve analysis was performed to determine the optimal cut-off level of sTWEAK (330 ng/mL) as a predictor of liver transplant/death in PSC patients (n=64). Figure 3 Lower sTWEAK levels are associated with poorer transplant- free survival in PSC patients

Kaplan-Meier survivorship estimates for patients with PSC (n=64) having high (>330 pg/mL) and low (<330 pg/mL) sTWEAK.

Figure 4 Soluble TWEAK levels in serum samples from healthy controls (HC) and patients with NASH, ALD, PBC, PSC, IBD or HCC (hepatocellular carcinoma).

ELISA to detect soluble TWEAK

Serum samples were collected from 64 newly diagnosed PSC patients attending the Liver Clinic (Queen Elizabeth hospital, Birmingham, UK) between January 2010 and November 2013. Clinical and laboratory data were collected at time of diagnosis and each subsequent clinic visit (3-12 monthly). Severity of liver disease was classified as pre-cirrhotic and cirrhotic. Patients with active ascending cholangitis, evident cholangiocarcinoma or colonic cancer were excluded from sampling.

Serum samples from patients with IBD alone (n=24) and healthy individuals (HC) without evidence of liver or gastrointestinal disease (n=22) served as controls. The distribution of intestinal inflammatory activity in IBD patients was classified as predominantly large bowel (colitis), small bowel or peri-anal disease.

Soluble TWEAK levels were quantified using human TWEAK instant ELISA (eBioscience, Hatfield, UK) according to manufacturer's instructions. Briefly, pre- coated standard and sample wells of the provided 96-well plate were hydrated with dfhO. Then, 50 μΐ_^ of human serum diluted 1 in 2 with sample diluent or undiluted cell supernatant were added. After a 3 hour incubation on a microplate shaker at 400 RPM the plate was washed six times with wash buffer. Any residual wash buffer was removed by inverting the plate and gently tapping it on a paper towel after the last wash. Upon addition of TMB substrate solution, a blue colour developed which was monitored with a Synergy HT microplate reader (BioTek Instruments, Potton, UK) set at 620nm until the highest standard had reached an optical density of about 0.9. Stop solution was added once colour development had occurred (typically 8 min) and absorbance was measured with the microplate reader set at 450 nm with a reference wavelength set to 620 nm. Each sample was run in duplicate and the sTWEAK concentration was determined by comparison to the standard concentration curve. Results are presented as medians unless otherwise stated. The statistical significance between medians was assessed using Kruskall-Wallis test or non-parametric Mann- Whitney U-test. Correlations between continuous variables were assessed with the Spearman's rho. The Chi-squared test was used to study significance between categorical variables. This analysis was carried out using Prism v6.0a software (GraphPad, CA, USA)

The optimal sTWEAK cut-off levels to predict short-term transplant free survival were calculated by a receiver operating characteristic (ROC) curve analysis. Transplant-free survival was assessed using the Kaplan-Meier method and the log- rank test was performed to compare the survival curves of individual groups. Univariate and multivariate Cox proportional hazard models were used to determine impact of clinical parameters on short-term transplant free survival. The reported results included hazard ratios (HR) and 95% confidence intervals.

sTWEAK levels are low in patients with PSC

To investigate sTWEAK levels in a chronic liver disease setting, serum samples were acquired from 64 patients with PSC, 24 patients with IBD with no evidence of concurrent PSC and from 22 healthy individuals. A summary of the general characteristics is shown in Table 1. Patients with PSC had a median age of 36 years, IBD patients had a median age of 35 years and healthy controls of 34 years at the time of the blood sample. Most PSC patients were diagnosed with IBD (n=51, 79.6%). Immunosuppressive medication was taken by 25 (39.1%) PSC patients and 9 (37.5%) IBD patients.

Soluble TWEAK serum levels as measured by a commercially available ELISA were significantly lower in PSC patients compared to healthy controls (p<0.01) and IBD patients (p<0.0001; Figure 1A). The median levels of sTWEAK in patients with PSC were 275 pg/mL (Interquartile range; IQR 111-410), in the healthy control group 942 pg/mL (IQR 885-1044) and in patients with IBD 1130 pg/mL (IQR 959-1429). Stratifying by disease severity, PSC patients with cirrhosis demonstrated significantly lower sTWEAK levels than those with pre-cirrhosis [189 pg/mL (IQR 67-299) versus 632 pg/mL (IQR 391-862), respectively; p< 0.001; Figure IB].

Soluble TWEAK levels in patients with PSC were then compared to demographic and clinical data (Table 2and 3). A significant positive correlation between sTWEAK and albumin was detected {Spearman >=0.252 /?<0.05; Table 2). The median of albumin concentration in the investigated PSC patients was 42 g/L (IQR 37-45), which is still in the normal range. No other tested variables such as age, colitis or immunosuppression demonstrated an association with sTWEAK levels (Table 2 and 3).

Controls IBD PSC

Number of patients 22 24 64

Gender, male (%) 12 (54.5) 12 (50.0) 40 (62.5)

Age, years 34 (28-37) 35 (25-55) 36 (26-57)

Immunosuppression :

Steroids (%) N/A 5 (20.8) 3 (4.7)

Azathioprine (%) N/A 4 (16.7) 5 (7.8)

Steroids+Azathioprine (%) N/A 0 (0.0) 7 (10.9)

Steroids+MMF (%) N/A 0 (0.0) 10 (15.6)

None (%) N/A 14 (58.3) 39 (60.9)

Table.1 Characteristics of healthy individuals and patients with IBD and PSC Abbreviation: MMF=Mycophenolate mofetil

Covariate RHO p value

Age, years -0.124 0.329

Laboratory findings*:

AST (iU/L) -0.107 0.403

ALP (ratio to ULN) -0.243 0.053

ALT (iU/L) -0.021 0.885

Albumin (g/L) 0.252 0.045

Bilirubin (μηιοΙ/L) -0.241 0.055

Platelet count (xlO^A9/L) 0.058 0.653

Na+(mmol/L) -0.106 0.406

Creatinine (μηιοΙ/L) 0.193 0.126

IgG (g/L) 0.018 0.895

UKELD -0.119 0.359

Table 2 Continuous variables correlating with sTWEAK levels

Abbreviations: UKELD=United Kingdom model for end-stage liver disease. Numbers in bold represent significant p values at the /?<0.05 level {Spearman correlation). *At time of taking sample

Median sTWEAK concentration,

Characteristics ng/mL (IQR)

Yes No p value

Gender, male 278 (83-362) 267 (162-540) 0.549

Colitis 304 (110-564) 253 (133-280) 0.239

Active colitis 248 (83-409) 278 (127-410) 0.715

Laboratory findings*:

INR above normal, >1.2 248 (67-314) 290 (129-493) 0.121

UKELD > 49 268 (138-319) 277 (98-625) 0.372

Medication:

UDCA 268 (110-464) 280 (133-358) 0.987

Immunosuppression 358 (234-522) 248 (98-331) 0.067

Antibiotics** 260 (129-309) 278 (114-518) 0.356

Table 3 Factors associated with sTWEAK levels

Abbreviations: IQR= interquartile range, INR=international normalised ratio, UKELD=United Kingdom model for end- stage liver disease, UDCA= Ursodeoxycholic acid. *At time of taking sample ** if taken in the past 3 month {Mann-Whitney U test) Low sTWEAK levels are a predictor of adverse prognosis in patients with PSC

As sTWEAK levels were closely related to a more advanced disease stage in our patient cohort, the predictive value of sTWEAK levels for transplant-free survival was investigated. First, AUROC curve analysis was performed to identify optimal serum sTWEAK levels for potential prediction of outcome Figure 2. The AUROC curve analysis revealed an optimal cut-off level of 330 pg/mL with a sensitivity of 86% and specificity of 56%. The area under the curve was 0.754 (p=0.001). By using the optimal cut-off threshold PSC patients were separated into two groups (low sTWEAK, <330 pg/mL; high sTWEAK, >330 pg/mL;Table 4). No significant differences were detected between the low and high sTWEAK groups and most clinical factors. However, as expected significantly more patients with cirrhosis were in the low sTWEAK group compared to the high sTWEAK group (90 vs. 26%, respectively, /?<0.001). In addition, patients in the low sTWEAK group had significantly higher ALP ratios compared to patients in the high soluble TWEAK group [2.1 (IQR 1.1-3.6) vs. 1.0 (IQR 0.7-2.3) ALP ratio to upper limit of normal (ULN), respectively, /?<0.05]. Bilirubin levels were also significantly higher in the low sTWEAK group [28 (IQR 14-53) vs. 12 (IQR 9-34) μιηοΙ/L, respectively, /?<0.05]. The group with high sTWEAK levels had significantly more patients on immunosuppression compared to the low sTWEAK group (57% vs. 29%, respectively, /?<0.05). <330 pg/mL >330 pg/mL p value

Number of patients (%) 41 (64) 23 (36) N/A

Gender, male (%) 27 (66) 13 (57) 0.592

Age, years 36 (27-57) 40 (24-57) 0.370

Cirrhosis (%) 37 (90) 6 (26) < 0.001

No. with colitis (%) 31 (76) 20 (87) 0.346

Colitis, active (%) 8 (20) 4 (17) 1.000

Laboratory findings*:

AST (iU/L) 74 (56-110) 80 (34-112) 0.580

ALP (ratio to ULN) 2.1 (1.1-3.6) 1.0 (0.7-2.3) 0.030

ALT (iU/L) 70 (44-113) 103 (35-156) 0.780

Albumin (g/L) 41 (37-44) 43 (40-45) 0.220

Bilirubin (μmol/L) 28 (14-53) 12 (9-34) 0.030

Platelet count (xlO^A9/L) 187 (110-298) 245 (168-310) 0.380

INR above normal, >1.2 (%) 19 (46) 4 (17) 0.055

Na+(mmol/L) 141 (139-144) 141 (140-142) 0.522

Creatinine ( mol/L) 68 (59-77) 71 (66-84) 0.145

IgG (g/L) 17 (13-20) 14 (13-20) 0.710

UKELD (continuous) 48 (46-52) 46 (44-51) 0.160

UKELD > 49 (%) 19 (46) 6 (26) 0.275

Medication:

UDCA (%) 33 (81) 18 (78) 1.000

Immunosuppression (%) 12 (29) 13 (57) 0.038

Antibiotics** (%) 9 (22) 1 (4) 0.081

Table 4 Characteristics of PSC patients in relation to low and high sTWEAK levels

PSC patients were divided into two groups on the basis of high sTWEAK (>330 pg/mL) and low sTWEAK (<330 pg/mL) levels in serum samples. Data is expressed as median (interquartile range). Numbers in bold represent significant p values at the /?<0.05 level. Abbreviations: INR=international normalised ratio, UKELD=United Kingdom model for end-stage liver disease, UDCA= Ursodeoxycholic acid. *At time of taking sample ** if taken in the past 3 month (for continuous variables Mann- Whitney U test otherwise Fisher's exact test)

During the follow up period of 29 (IQR 8-41) months, 29 (45.3%) patients reached the endpoint. Of those patients 18 (28.1%) were transplanted, 7 (10.9%) died and 4 (6.3%) patients died despite transplantation. Kaplan-Meier survival analysis demonstrated a significant difference in transplant-free survival within a 3-year follow-up interval between the low sTWEAK and high sTWEAK groups (log rank test /?<0.01, Figure 3). Within the low sTWEAK group, 25 out of 41 patients (61%) reached an endpoint whereas only 4 out of 23 patients (17%) of the high sTWEAK group reached an endpoint within the follow up period. On univariate Cox analysis, low sTWEAK levels were a significant predictor of death or transplantation (HR: 4.013, 95% CI 1.395-11.545; p<0M) (Table 5). Besides low sTWEAK levels, additional predictors of poor outcome were elevated baseline ALP (HR: 1.136, 95% CI 1.029-1.254; /?=0.011), hypoalbuminaemia (HR: 1.164, 95% CI 1.092-1.241; /?<0.001), elevated bilirubin (HR: 1.004, 95% CI 1.002-1.006; /?<0.01), INR>1.2 (HR: 3.497, 95% CI 1.572-7.819; /?<0.01), low sodium levels (HR: 1.176, 95% CI 1.054- 1.314; /?<0.01) and immunosuppression (HR: 3.501, 95% CI 1.415-8.662; /?<0.01). In addition, the UKELD score as continuous variable (HR: 1.172, 95% CI 1.105-1.243; /?<0.001) and as dichotomous variable (HR: 7.576, 95% CI 3.300-17.241; /?<0.001) with the clinical cut-off point of 49 was predictive of transplantation/death.

Subsequent multivariate analysis by Cox proportional hazards regression demonstrated that sTWEAK levels below 330 pg/mL remained to be an independent predictor of death or transplantation after adjustment for other significant variables (adjusted HR: 3.628, 95% CI 1.033-12.746; /?=0.044) (Table 6A). The strong predictive value of sTWEAK was also retained when the UKELD score was included in the multivariate analysis as continuous variable (adjusted HR: 4.124, 95% CI 1.224-13.890, /?=0.022, Table 6B) or dichotomous variable with the cut-off point of 49 (adjusted HR: 3.944, 95% CI 1.160-13.407, /?=0.028,Table 6C). Other factors such as elevated baseline ALP levels were also predictive of transplantation/death when the UKELD score was included in the multivariate analysis as continuous or dichotomous variable but not when the UKELD score was excluded (Table 6). In addition, hypoalbuminaemia was a significant predictor of transplantation/death when the UKELD score was excluded or applied as dichotomous variable but not as continuous variable. Furthermore, low sodium levels were predictive of transplantation/death when UKELD was excluded (Table 6A). As low sodium levels are part of the UKELD formula they were not included in analysis containing UKELD.

HR (95% CI) p value

Gender, male 1.298 (0.601-2.802) 0.506

Age, years 1.021 (0.998-1.044) 0.168

Colitis 1.926 (0.375-2.288) 0.868

Active colitis 0.922 (0.350-2.430) 0.870

Laboratory findings*:

AST (iU/L) 1.000 (0.999-1.002) 0.771

ALP (ratio to ULN) 1.136 (1.029-1.254) 0.011

ALT (iU/L) 1.000 (0.997-1.003) 0.833

Low albumin (g/L) 1.164 (1.092-1.241) <0.001

Bilirubin (μπιοΙ/L) 1.004 (1.002-1.006) <0.001

Low platelets (xlO^A9/L) 1.001 (0.998-1.004) 0.557

INR above normal/1.2 3.497 (1.572-7.813) 0.002

Low Na+ (mmol/L) 1.176 (1.054-1.314) 0.004

Creatinine (μπιοΙ/L) 0.989 (0.963-1.015) 0.408

IgG (g/L) 1.011 (0.964-1.060) 0.661

UKELD (continuous) 1.172 (1.105-1.243) <0.001

UKELD > 49** 7.576 (3.300-17.241) <0.001

Medication:

UDCA 0.895 (0.339-2.367) 0.824

Immunosuppression 3.501 (1.415-8.662) 0.007 sTWEAK*** 4.013 (1.395-11.545) 0.010

Table 5 Univariate analysis of clinical parameters for the prediction of transplant-free survival in patients with PSC

Numbers in bold represent significant p values at the /?<0.05 level or below. Abbreviations: CI= confidence interval, INR=international normalised ratio, UKELD=United Kingdom model for end- stage liver disease, UDCA= Ursodeoxycholic acid. *At time of taking sample ** clinically used cut-off point *** cut-off at <330 pg/mL

HR (95% CI) p value

A) UKELD excluded

ALP (ratio to ULN) n.s. n.s

Low albumin (g/L) 1.131 (1.056-1.211) <0.001

Bilirubin (μπιοΙ/L) n.s. n.s.

INR above normal/1.2 3.067 (1.323-7.143) 0.009

Low Na+ (mmol/L) n.s. n.s.

Immunosuppression n.s. n.s. sTWEAK* 3.628 (1.033-12.746) 0.044

B) UKELD included as continuous variable

ALP (ratio to ULN) 1.130 (1.007-1.268) 0.037

Low albumin (g/L) n.s. n.s.

UKELD 1.102 (1.003-1.210) 0.043

Immunosuppression n.s. n.s. sTWEAK* 4.124 (1.224-13.890) 0.022

C) UKELD included as a dichotomous variable

ALP (ratio to ULN) 1.134 (1.006-1.279) 0.039

Low albumin (g/L) 1.106 (1.020-1.199) 0.014

UKELD > 49** 3.658 (1.342-9.968) 0.011

Immunosuppression n.s. n.s. sTWEAK* 3.944 (1.160-13.407) 0.028

Table 6 Multivariate analysis of clinical parameters for the prediction of transplant-free survival in patients with PSC

Multivariate analysis using backward stepwise elimination. Numbers in bold represent significant p values at the /?<0.05 level or below. Abbreviations: n.s= not significant CI= confidence interval, INR=international normalised ratio, UKELD=United Kingdom model for end-stage liver disease. * cut-off at <330 pg/mL **clinically used cut-off point

This demonstrates that sTWEAK serum levels are altered in patients with PSC and that sTWEAK levels below 330 pg/mL are an independent predictor of transplantation or death. This is the first time that sTWEAK levels are studied in a cohort of PSC patients.

PSC is characterised by chronic inflammation of the biliary epithelium (Eaton et al., 2013, Gastroenterol. 145(3), 521-536). Significantly higher levels of TNF-a for example were detected in the serum of PSC patients (NEVMAN et al., 2002, J. Gastroent. Hepatol. 17(2), 196-202). In addition, cirrhotic patients with a variety of chronic liver diseases had increased serum levels of inflammatory cytokines such as IL-6 and IL-Ιβ compared to pre-cirrhotic patients (Tilg et al., 1992, Gastroenterol. 103(1), 364-274). Therefore it was surprising that sTWEAK serum levels were significantly reduced in PSC patients compared to healthy controls and that the sTWEAK levels were even lower in PSC patients with cirrhosis than in patients without cirrhosis. Similarly to my findings in PSC patients, reduced sTWEAK levels were also detected in other diseases associated with increased inflammation including chronic heart failure, chronic kidney disease and type 1 diabetes (Chorianopoulos et al., 2009, Eur. J. Heart Failure 11(11) 1050-1056; Yilmaz et al., 2009, Am. Soc. Nephrol. CJASN 4(11) 1716- 1723), (Llaurado et al., 2012, PLoS one 7(8)). However, sTWEAK levels were not consistently decreased in all diseases as research by Park et al. has demonstrated significantly higher levels in patients with rheumatoid arthritis (Park et al., 2008, Scan. J. Rheumatol. 37(3) 173-178), which is in accordance with the data from the IBD patients that also had higher levels of sTWEAK compared to healthy controls. However, comparison of PSC patients with or without IBD demonstrated no difference in sTWEAK levels.

Why sTWEAK levels are reduced rather than increased in patients with PSC is not known. A pathological role of TWEAK in liver disease has been demonstrated in previous chapters. The Applicants have demonstrated that livers from patients with PSC express high TWEAK and Fnl4 protein levels compared to normal livers. HSCs increase Fnl4 expression after injury sensitising them to TWEAK stimulation and enhancing proliferation and release of inflammatory cytokines such as CCL5 (data not shown). Additionally they have shown a significant reduction of liver fibrosis in TWEAK KO mice. In addition, LPCs express Fnl4 and bone-marrow derived macrophages induce ductular reaction via TWEAK (Bird et al., 2013, PNAS 110(6) 6542-6547). Therefore, through some unknown mechanism, sTWEAK levels might be reduced to unsuccessfully prevent further tissue damage. A potential way through which sTWEAK is reduced is through CD 163, a TWEAK scavenger receptor. CD163-positive inflammatory macrophages and monocytes can bind and internalise TWEAK leading to low TWEAK concentrations (Bover et al., 2007, J. Immunol. 178(12) 8183-94; Moreno et al., 2009, Atherosclerosis 207(11) 103-110). Research has shown that patients with liver cirrhosis had significantly higher serum soluble CD 163 levels than healthy controls (Holland-Fischer et al., 2011, Gut, 60(10) 1389-1393). In addition, high soluble CD 163 and low sTWEAK concentrations have been correlated to type 1 diabetes mellitus diagnosis and cardiovascular mortality in peripheral arterial disease (Llaurado et al., 2012 Supra; Urbonaviciene et al., 2011 Atherosclerosis 219(2) 892-899). Another possible explanation for lower sTWEAK levels is that increased Fnl4 expression in the liver contribute excess binding of sTWEAK leading to lower serum sTWEAK concentration but potentially more TWEAK signalling, inducing more ductular reaction and fibrosis and therefore a more severe liver disease. Correlations between TWEAK or Fnl4 expression with the severity of disease have been investigated previously (Affo et al., 2012 Gut 62(3) 452-460; Li et al., 2013APJCP 14(6) 3509-3514). A study by Affo et al demonstrated that acute hepatitis patients with higher hepatic Fnl4 mRNA expression had a worse 90-day survival rate than patients with lower Fnl4 mRNA levels (Affo et al., 2012 Supra). They also detected that patients with acute hepatitis had lower sTWEAK levels but did not show whether it was predictive of disease outcome. In addition, high hepatic expression levels of Fnl4 were also associated as a predictor of poorer surgical outcome in patients with HCC (Li et al., 2013 Supra). However, investigating hepatic Fnl4 expression would not be feasible, as it would require an invasive biopsy in order to serve as a predictor of disease outcome.

Instead, the data has shown that sTWEAK levels in patient serum samples are predictive of transplant-free survival in a multivariate regression model despite including the UKELD score as a variable. PSC patients with low sTWEAK levels were more likely to die or require a transplant. In accordance with that, research by Richter et al also demonstrated that sTWEAK levels in serum samples could be used as a predictive tool of mortality in patients with advanced non-ischemic heart failure (Richter et al., 2010 Atherosclerosis 213(2) 545- 548). The data also demonstrated that serum albumin and INR, when the UKELD score was excluded as a variable, were also independently associated with transplant-free survival. Other research has also shown that serum albumin improved prediction of waiting list mortality in patients with liver cirrhosis when added to the Model for End-Stage Liver Disease (MELD) used in America (Myers et al., 2013 PLoS 8 (1) e51926). The prothrombin time expressed as INR is used to assess clotting and is included worldwide in almost all the prognostic models regarding transplantation for liver disease. Considering that patients with liver disease have a more complex coagulopathy and that the method for calibrating the INR calculation is based on serum levels from people with stable anticoagulation, the usefulness of INR seems to be suboptimal (Tripodi et al., 2007 Hepatology 46(2) 520-527). However despite these limitations, INR is still considered a valid measure (Kamath and Kim, 2009 Clinics in Liver Disease 13(1) 63-66).

Including the UKELD score in the multivariate analysis as continuous or dichotomous variable showed that it is also predictive of patients at risk requiring transplantation. This was expected as the scoring system is currently used to put patients on the transplant list. However using the UKELD score to decide on the urgency of transplantation may not be useful for patients with PSC and other forms of cholestatic liver disease as such patients face unique clinical features that would better reflect the severity of illness for example ascending cholangitis due to biliary strictures (Goldberg et al., 2011 Liver Transplantation 17(11) 1355- 1363). ALP was only significant when UKELD score was applied as continuous or dichotomous variable but not when it was excluded from the calculation. ALP fluctuations are frequent during the clinical course, influenced by a variety of confounding variables (Stanich et al., 2011 Digestive and Liver Disease 43(4) 309-313). Nevertheless, ALP remains to be an important indicator of disease severity and higher ALP correlated with lower sTWEAK levels, which is in accordance with the predictive ability of sTWEAK. Furthermore, a larger number of patients in the high sTWEAK group were on immunosuppressant but a benefit of immunosuppressive medication in inhibiting disease progression remains yet to be determined (Hirschfield et al., 2013 The Lancet 382 (9904) 1587-1599).

Although sTWEAK levels seem to be an independent predictor of transplant-free survival in patients with PSC it remains to be determined whether the prognostic values of sTWEAK are disease specific or whether it also applies to other cirrhotic liver diseases such as PBC and ALD. Nevertheless, sTWEAK could serve as a predictive tool in patients with PSC and improve the way patients are allocated on the waiting list. A pragmatic approach could be to adopt sTWEAK measurements as part of the routine diagnostic measurements and possibly include it as a variable of the UKELD scoring system.

In conclusion, lower sTWEAK levels distinguish patients with PSC compared to those with IBD alone and healthy subjects. Moreover, sTWEAK inversely correlates with disease severity and is an excellent predictor of poorer clinical outcome. However, TWEAK has been shown to be involved in liver disease processes and therefore is most probably a pathogenic cytokine than only a biomarker in patients with PSC.

Figure 4 shows further supporting data for serum soluble TWEAK levels with NASH, ALD, PBC, PSC, IBD and HCC, compared to healthy controls showing s TWEAK levels for patients with the diseases.

Claims

1. A method of diagnosis or prognosis of a chronic liver disease or inflammatory bowel disease (IBD) in a subject comprising providing a sample from the subject and measuring the amount of TWEAK (tumour necrosis factor-like weak inducer of apoptosis) in the sample, wherein if the amount of TWEAK is below a predetermined level, then there is an increased risk that the subject has a chronic liver disease or has a poor prognosis of the chronic liver disease, or if the amount is above a predetermined level then there is an increased risk that the subject has IBD.

2. A method of determining the effect of treatment on a chronic liver disease comprising providing measuring the amount of TWEAK in a first sample from a subject, provided before the treatment and comparing it to an amount of TWEAK in a second sample from a subject provided after treatment, wherein an increase in TWEAK between the first and second samples indicates that the chronic liver disease is being treated.

3. A method of determining the risk that a subject will need a liver transplant, comprising measuring the amount of TWEAK in a sample from a subject.

4. A method according to claims 1 to 3, wherein the TWEAK is soluble TWEAK.

5. A method according to claim 4, wherein the TWEAK is trimeric or monomeric.

6. A method according to claims 1 to 5, wherein the disease is chronic inflammatory liver disease and the disease is selected from Primary Sclerosing Cholangitis (PSC), Primary Biliary Cirrhosis (PBC), Autoimmune hepatitis (AIH), Alcoholic Liver Disease (ALD) and non-alcoholic steatohepatitis (NASH), most typically PSC.

7. A method according to claims 1 to 5, wherein the chronic liver disease is hepatocellular carcinoma (HCC).

8. A method according to any preceding claim, comprising determining the amount of TWEAK with a specific binding agent for TWEAK.

9. A method according to claim 8, wherein the specific binding agent is an anti-TWEAK antibody or a TWEAK- specific fragment thereof.

10. A method according to any preceding claim, wherein the sample is selected from whole blood, plasma, serum, urine, bile or a liver tissue biopsy sample.

11. An assay kit for use in the method according to claims 1 or 4 to 10, wherein the disease is a chronic inflammatory liver disease, the kit comprising a TWEAK- specific binding agent and one or more reagents for the detection of liver function or liver disease.

12. An assay kit according to claim 11, comprising a reagent for the detection of an autoantibody selected from ANA, ASMA and PANCA.

13. An assay kit according to claims 10 or 11, comprising a reagent for measuring one or more of AST, ALP, ALT, bilirubin, creatinine, prothrombin or gamma glutamyl transpeptidase.