Novel serological neo-epitope markers of extracellular matrix proteins for the detection of portal hypertension

Abstract

Background: The hepatic venous pressure gradient (HVPG) is an invasive, but important diagnostic and prognostic marker in cirrhosis with portal hypertension (PHT). During cirrhosis, remodelling of fibrotic tissue by matrix metalloproteinases (MMPs) is a permanent process generating small fragments of degraded extracellular matrix (ECM) proteins known as neoepitopes, which are then released into the circulation.

Aim: To investigate their potential as plasma markers for detection of PHT.

Methods: Ninety-four patients with alcoholic cirrhosis and 20 liver-healthy controls were included. Clinical and laboratory data of the patients were collected. All patients received HVPG measurement with blood sampling. In these samples, the following degradation or formation markers were measured: C1M (type I-collagen), C3M and PRO-C3 (type III collagen), C4M and P4NP 7S (type IV collagen), C5M (type V collagen), C6M (type VI collagen), BGM (biglycan), ELM (elastin), CRPM (CRP).

Results: All ECM markers except for CRPM correlated significantly with HVPG. Interestingly, C4M, C5M and ELM levels were significantly higher in patients with HVPG >10 mmHg. Multiple regression analysis identified PRO-C3, C6M and ELM as significant determinants, while the models A and B including PRO-C3, ELM, C6M and model for end-stage liver disease (MELD) provided better description of PHT (r = 0.75, P < 0.0001). The models provided odds ratios of >100 for having clinical significant PHT.

Conclusions: These novel non-invasive extracellular matrix markers reflect the degree of liver dysfunction. The different degrees of portal hypertension correlated with these circulating neoepitopes. Using a single blood sample, these neoepitopes in combination with MELD detect the level of portal hypertension.

Figure 1

Protein fingerprint markers stratified according to the degree of portal hypertension (HVPG) range: HVPG < 10 mmHg (n = 23), HVPG equal 10–16 mmHg (n = 28), HVPG ≥16 mmHg (n = 37) compared with controls (n = 14). (a) Data for C3M, C6M, CPRM, ELM and P3NP are shown. (b) Data for BGM, C1M, C4M, C5M and P4NP 7S are shown. All data are shown as geometric mean values ± SEM. Asterisks indicate significant difference between the specific groups compared with controls (*P < 0.05; ** P < 0.01; ***P < 0.001).

Figure 2

Combination of plasma biomarkers in a linear regression algorithm correlated with HVPG or stratified according to clinical relevant HVPG range in cirrhotic patients and controls. (a) Model A combining ECM markers only correlated with HVPG. (b) Model B combining ECM markers and the MELD score to HVPG. (c) Model A and (d) Model B stratified according to portal hypertension ranges.

Figure 3

Models A and B correlated with HVPG in patients with a HVPG<10 mmHg (n = 23) (a and b, respectively) in cirrhotic patients.

Figure 4

Area under the receiver operating curve (AUROC) plots for models A and B for the separation of HVPG≤10 mmHg vs. HVPG>10 mmHg.