Metallothionein: a game changer in histopathological diagnosis of Wilson disease

Wilson disease (WD) is a genetic disorder of copper metabolism caused by mutations in the ATP7B gene. Toxic copper accumulation leads to hepatic, neurologic, and psychiatric disorders with variable presentation. Metallothionein (MT) immunohistochemistry was proposed as a diagnostic marker.


Introduction
Wilson disease (WD) is an inherited disorder of copper metabolism caused by biallelic mutations in the ATP7B gene encoding a copper transporting adenosine triphosphatase. 1 Clinical features of WD are related to toxic accumulation of copper, primarily in the liver and the brain, leading to hepatic, neurologic, and/or psychiatric disorders.Corneal copper deposition results in the formation of Kayser-Fleischer rings (KFR). 2 The diagnosis of WD can be difficult, as both clinical criteria and genetic testing have significant limitations. 3Genetic diagnosis is complex and interpretation for rare variants is challenging as, for instance, the presence of biallelic pathogenic variants may not result in copper accumulation, 4 while some synonymous genetic variants may even be disease-causing in other individuals. 5However, early diagnosis is important, as effective treatment prevents the development of progressive disease and improves overall outcome.
Hepatic involvement may be asymptomatic until cirrhosis or may even initially present as acute liver failure.While the latter manifestation is histologically associated with confluent hepatocellular necrosis, more silent disease frequently shows steatotic liver disease (SLD) of variable activity leading to progressive liver fibrosis and cirrhosis. 6Histochemical staining of copper (e.g.rhodanine) and copper-associated protein (e.g.orcein) is currently used to support the histological evaluation of liver biopsies with suspected WD.As the biliary system is the major route for copper excretion, chronic biliary diseases (and others) result in hepatic copper accumulation and may give rise to positive staining. 7etallothioneins (MTs) are a group of cysteine-rich proteins with a high affinity for metals, effectively acting as an endogenous chelator.Under physiological conditions perivenular MT expression has been described in adult human liver. 7In mice with genetic deletion of the ATP7B gene, the cellular abundance of MT correlated with the endogenous copper detoxification capacity. 8Recently, Rowan et al. detected MT upregulation in liver tissue of WD patients (n = 20) and suggested that MT immunohistochemistry may be useful for screening. 9Here we defined its role within the repertoire of diagnostic tests used for evaluating potential WD.

P A T I E N T S
The databases of the Institute of Pathology, and the joint WD database of the Department of Gastroenterology and Hepatology and the Department of Paediatrics I, University Hospital Heidelberg, were reviewed for patients suspicious for WD (No. S-319/2010).Between 07/2001 and 12/2019, a total of 69 liver samples from 64 patients were identified fulfilling the Leipzig criteria (score ≥4; Table S1). 3Informed consent was available from each patient in the databases.Sequencing data of the ATP7B gene were available for 55 patients.A biallelic mutation of the ATP7B gene was known in 43 patients, of which 25 were compound heterozygotes.Only one ATP7B gene mutation was detected in 11 patients, and no mutation was found in one patient.The liver copper concentration was >250 lg/g dry weight in 40 out of 41 patients analysed, and a 24-h urinary copper excretion of >2-fold the upper normal limit was present in 31 of 52 patients analysed at the time of tissue sampling.The patients´characteristics are detailed in Table 1.The control cases included 78 patients with a hepatitic-type liver damage and 72 cases suffering from chronic cholestatic liver diseases, and 10 histologically normal liver samples (Table S2).Liver samples were stained for haematoxylin and eosin (H&E), periodic acit-Schiff (PAS) after diastase digestion, Prussian blue, modified Gomori's, orcein, and rhodanine.Pseudonymized formalin-fixed, paraffin-embedded samples were retrieved from the archive for immunohistochemical staining in cooperation with the National Centre for Tumour Diseases (NCT) tissue bank (project number: 3777).The study was approved by the local Ethics Committee (No. S-206/2005 and S-319/2010) and carried out in accordance with the Declaration of Helsinki.

I M M U N O H I S T O C H E M I S T R Y
Immunohistochemistry was performed on 4-micron thick sections using an automated immunostainer (Ventana BenchMark Ultra, Ventana Medical Systems, Tucson, AZ, USA).The settings included pretreatment with cell conditioner 1 for 32 min, incubation with an MT specific antibody (clone UC1MT, Abcam, Cambridge, MA, USA, dilution 1:800) at 36°C for 24 min, followed by an 8-min incubation with OptiView Universal Linker (Ventana) and visualization using DAB (8-min incubation, Opti-View HRP Multimer, Ventana).Based on the distribution of the immunosignal in the liver lobules, five staining patterns were discriminated: no staining, perivenular, periseptal, patchy (increased without evident zonal restriction, but not diffuse), and diffuse (staining of the entire lobule).In addition, staining intensity (no, weak, moderate, strong) and percentage of positive hepatocytes were evaluated.nominal variables (e.g.count data) was tested using Fisher's exact test (fisher.testfunction, stats package).Continuous variables were analysed with the nonparametric Mann-Whitney test (Wilcox.testfunction, stats package).In case of multiple testing, P-values were adjusted with the Bonferroni-Holm correction method (p.adjust function, stats package).ROC analysis including calculation of area under the curves (AUCs) and confidence intervals where performed in R with the pROC package. 10P < 0.05 was considered significant.Graphs were generated with the ggplot2 package.

E X P R E S S I O N O F M T I N L I V E R T I S S U E
A total of 229 formalin-fixed, paraffin-embedded liver specimens were analysed for MT expression.Any positive staining was recorded in 91% (n = 63/69) of WD samples and 64% (n = 103/160) of control cases (P < 0.01, Table 2).In particular, a diffuse MT staining (Figure 1A) was observed in 80% (n = 55/69) of WD patients but only 3% (5/160) of control cases (P < 0.01).This pattern had a sensitivity of 79.7% and a specificity of 96.6% for WD.A patchy pattern (Figure 1B) was seen in 34% (n = 52/160) of controls, but in only 7% (n = 5/69) of WD cases (P < 0.01).Periseptal MT expression (Figure 1C) was usually seen in advanced fibrosis/cirrhosis (F3/F4; n = 33/37).Periportal staining (Figure 1D) was mostly observed in patients with biliary disease (n = 19/27).Perivenular staining (Figure 1E) was seen in all entities, except alcohol-associated and metabolic dysfunction-associated steatotic liver disease.

M T I M M U N O S T A I N I N G I N L I V E R T I S S U E C A N B E U S E D A S A D I A G N O S T I C M
We applied ROC analysis to identify the optimal cutoff for MT immunostaining to detect WD patients.As shown in Figure 2 "at least moderate staining in >50% of hepatocytes" was the best histological criterion for WD.Cases showing this staining characteristic were considered MT-positive.Neglecting the staining intensity as a component of the potential biomarker would reduce its specificity by 5% and false-positives would occur in clinically relevant differential diagnoses like metabolic dysfunction-associated steatohepatitis (n = 1), autoimmune hepatitis (n = 1), or drug-induced acute liver failure (n = 3).Positive MT staining was observed in 81% of liver specimens of WD patients (n = 56/69), but only five control cases were sufficiently positive for MT immunostaining to resemble WD (Table 2).The sensitivity, specificity, and accuracy for diagnosis of WD in all patients were 81.3%, 96.9%, and 92.4%, respectively (Table 3), In contrast, rhodanine staining had a sensitivity of 34.4% and the sensitivity of orcein staining for the detection of WD was 26.2%.Of note, positive MT staining retained a high sensitivity (72.2%) in patients without relevant fibrosis (F0/F1), a subgroup of patients in whom histochemistry is usually negative (sensitivity: rhodanine 11.1%, orcein 0%, Figure 3).Thirty-five samples from treatment-na€ ıve patients were available.Sensitivity, specificity, and accuracy of positive MT staining for diagnosis of WD were 85.7%, 96.9%, and 94.9%, respectively.When focusing on the subgroup analysis of patients without available biallelic genotyping information (n = 9), the sensitivity of MT immunostaining dropped to 66.7% (still being much higher than histochemistry), while the specificity and accuracy remained excellent (96.9% and 95.3%, respectively).Small liver biopsies may be prone to sampling error, but MT immunostaining was also informative in biopsies <15 mm length and those including <10 portal tracts (Table 3).

R E L A T I O N O F T H E M T I M M U N O S T A I N I N G T O E S T A B L I S H E D D I A G N O S T I C T E S T S F O R W D
To identify factors that may be associated with the failure of MT immunostaining to detect WD patients, we compared MT-positive and MT-negative WD cases.While we did not observe a difference in terms of age at diagnosis, sex, therapy, histological appearance of liver disease, or fibrosis stage between either groups, MTnegative WD patients had significantly more neurologic disorders compared to MT-positive cases (50% versus 13%, P < 0.05; Table 4).Furthermore, KFRs were significantly more prevalent in MT-negative cases (50% versus 15%, P < 0.05).We did not detect a difference

Metallothionein immunostaining in Wilson disease 941
Five control cases were MT-positive.Three cases occurred in secondary sclerosing cholangitis related to chronic biliary obstruction (n = 1) or intensive care unit cholangiopathy (n = 2; Table 2) and one patient had PSC.In the last case, progressive familial intrahepatic cholestasis3 (PFIC3) was defined by genetic testing.Initially, this latter patient was suspected to suffer from WD and was treated accordingly.As no biochemical normalization was achieved, treatment was modified from D-penicillamine to trientine, which was then supplemented by zinc.A steady decrease in liver copper content was detected in three serial biopsies (Figure 5A-C) taken over a period of 101 months (1022, 232, and 58 lg/g, respectively).The last biopsy taken after 8 months of dual trientine and zinc treatment became MT-positive (Figure 5C).
Anecdotally, another patient (not included in the above analysis due to a Leipzig score <4) had a positive family history for WD and decreased ceruloplasmin in serum (<0.10 g/L), while the most prevalent ATP7B gene mutation (p.H1069Q) was not present and liver copper content was even decreased (21 lg/ g dry weight).Interestingly, no MT staining was recorded in this patient (Figure 5D).

P E R F O R M A N C E O F M T I M M U N O H I S T O C H E M I S T R Y I N A M O D I F I E D L E I P Z I G S C O R E
Considering the high diagnostic performance of the MT immunohistochemistry for the detection of WD outlined above, we evaluated its potential role within a multiparameter score.According to the European >18 lmol/L 6% (n = 3) -

Discussion
About 40 years ago, Janssens et al. identified MT as a copper-associated protein in livers with primary biliary cholangitis 12 and higher concentrations of MT were subsequently reported in patients with WD. 13 Potentially, due to the lack of reliable antibodies during this early time of evolving immunohistochemistry in diagnostic pathology, MT immunostaining was forgotten until it was recently rediscovered. 9T is a large intracellular molecule, capable of binding specific metals like cadmium, copper, and zinc independently of the ATP7B gene status.Elevated concentrations of copper in hepatocytes may induce MT expression to prevent oxidative cell damage.ROC analysis confirmed the cutoff criterion (at least moderate staining in >50% of hepatocytes) described by Rowan et al. 9 MT immunostaining revealed an excellent diagnostic performance, both in treated and untreated patients.The high MT expression level in hepatocytes may explain the low sensitivity of routine histochemical stains (e.g.rhodanine, orcein) for the diagnosis of WD, as MT-bound copper is in the reduced cuprous (Cu + ) form, whereas detection by histochemical stains requires the oxidized cupric (Cu 2+ ) form. 7lthough WD frequently presents as SLD, it should be considered as a differential diagnosis in all patients with chronic hepatitis and cirrhosis of unknown aetiology.WD is also a likely differential diagnosis in all cases of acute liver failure associated with nonimmune hemolytic anaemia. 2,11In liver biopsies showing these kinds of damage pattern, we did not detect any MT-positive cases.Another rare differential diagnosis is PFIC3.Five treatment-na€ ıve PFIC3 patients were analysed by MT immunohistochemistry; all were negative, implying that an ABCB4 gene mutation does not induce MT expression.Thus, MT immunostaining is able to differentiate WD from clinically relevant differential diagnosis, already in patients with early, prefibrotic disease.Due to its superior performance, MT immunohistochemistry may substitute for rhodanine histochemistry as an optional parameter of the Leipzig score.Moreover, 97% of WD patients in our series would be correctly classified by MT immunostaining, if hepatic copper concentration would not be available (or testing was not performed yet).MT immunostaining seems also helpful in cases with already available liver biopsy, in which WD was not primarily considered as a differential diagnosis and the noninvasive parameters ceruloplasmin in serum and 24-h urinary copper excretion were not performed yet.Finally, MT immunostaining may be of great help, if ATP7B genotyping takes too long, is too expensive, or interpretation of the genotype is hard.
Although MT immunostaining may be considered the new standard for histological evaluation of WD, it has limitations.False-positives occurred in cases of secondary sclerosing cholangitis (n = 3), PSC (n = 1), and PFIC3 (n = 1).While integration of a patient's history, clinical, and histological findings will aid in correct interpretation of cases with sclerosing cholangitis, the latter case was iatrogenic due to 8 months of zinc treatment of this PFIC3 patient before the MTpositive liver biopsy was obtained.][16][17] Thus, false-positive findings in MT immunohistochemistry can be expected in patients treated with zinc preparations.Fittingly, all three WD patients receiving zinc were positive.Moreover, 3 out of 14 D- penicillamine-treated and 4 out of 12 trientine-treated patients showed negative MT immunostaining, suggesting that these drugs do not induce hepatocellular MT expression, but MT negativity could eventually be related to treatment.Thus, the diagnostic interpretation of MT immunostaining should be restricted to treatment-na€ ıve patients, in which it has the best diagnostic performance (Table 3).
Perivenular MT expression was described in normal liver. 7Independent of the underlying disease, periseptal staining was mostly seen in livers with advanced fibrosis/cirrhosis, possibly indicating that this pattern may be a consequence of parenchymal extinction, as previously described for glutamine synthetase expression. 18Interestingly, we did not detect any perivenular MT staining in patients with alcoholassociated or metabolic dysfunction-associated steatotic liver disease, which frequently show a predominantly perivenular fatty change.Thus, fat accumulation may affect (visualization of) hepatocellular MT expression and may pose difficulties in evaluating MT immunostaining (Figure 1F).
MT-negative cases were significantly more prevalent in WD patients with neurologic disorders and the presence of KFR.Hepatic copper quantification was only available for four patients with neurologic presentation and six patients with KFR.In these small subgroups, we could not detect a significant difference between patients with and without neurological presentation or KFR, respectively.The 24-h copper excretion was significantly higher in MT-positive than in MT-negative WD patients, while the hepatic parenchymal copper concentration was similar between both groups, raising the possibility that other hepatocellular proteins may act as copper-binding proteins in MT-negative patients. 19,20Furthermore, metallothionein abundance is regulated not only by the tissue metal content (zinc and copper), but also by the cellular redox state, serum growth factors, or by ligands of the Farnesoid X Receptor (FXR) nuclear receptor. 21Importantly, a negative MT immunostaining does not rule out WD.

Conclusion
Positive MT immunostaining represents an excellent biomarker for the diagnosis of WD and should be incorporated in the diagnostic work-up of liver biopsies from all patients with a clinical or histological differential diagnosis of WD.It should replace rhodanine staining as part of a modified Leipzig score and can be used alternatively to hepatic copper quantification.
A N A L Y S I S Statistical analysis was performed in R (version 4.2.2,Vienna, Austria).The statistical independence of Ó 2023 The Authors.Histopathology published by John Wiley & Sons Ltd., Histopathology, 83, 936-948.
Metallothionein immunostaining in Wilson disease 939 between the levels of ceruloplasmin (0.10 AE 0.01 versus 0.13 AE 0.03 g/l, P > 0.05), copper (8.5 AE 1.1 versus 9.6 AE 1.6 lmol/l, P > 0.05), and zinc (12.4 AE 1.1 versus 11.2 AE 0.7 lmol/L, P > 0.05) in serum between MT-positive and MT-negative WD patients, respectively.However, the 24-h copper excretion was significantly higher in MT-positive than in MT-negative WD patients (8.18 AE 2.23 versus 5.71 AE 3.36 lmol/ d, P < 0.05; Figure4), while the hepatic parenchymal copper concentration was similar between both groups (1091.0AE 57.7 versus 1094.9AE 314.9 lg/g dry weight, P > 0.05).M T I M M U N O S T A I N I N G I N C L I N I C A L A P P L I C A T I O NIn the diagnostic setting, there are three histological scenarios in which WD is a relevant differential diagnosis:(1) acute liver failure, particularly in young patients, (2) SLD, and (3) unexplained cirrhosis.None of the control cases of acute liver failure and SLD analysed here showed positivity of MT immunostaining.Furthermore, only one primary sclerosing cholangitis (PSC) patient among the cirrhotic controls (n = 42) was MT-positive, while 79% (n = 19/24) of the cirrhotic WD cases were positive in MT immunohistochemistry.

Figure 2 .
Figure 2. ROC analysis of MT immunostaining for the detection of WD patients.At least moderate staining in >50% of hepatocytes (purple line) has the highest specificity for the detection of WD patients and only a slightly reduced sensitivity compared to any staining in >50% of hepatocytes (blue line).Although strong MT expression (yellow line) has a high specificity, the associated strong reduction of sensitivity excludes its use as a diagnostic test for WD.

Figure 3 .
Figure 3. MT immunostaining outperforms rhodanine and orcein histochemistry.(A) WD: Strong diffuse MT immunoreactivity of hepatocytes is evident at low magnification.(B) Sensitivity of the MT immunohistochemistry, rhodanine, and orcein histochemistry for the detection of WD at different fibrosis stages.(C) WD, same patient: The localization of the insert is indicated by the black box.Insert: At high magnification rhodanine-positive deposits are detected in periseptal hepatocytes.(D) WD, same patient: The black box represents the insert: At high magnification orcein-positive granules (indicated areas) are focally detected in hepatocytes.

Figure 4 .
Figure 4. Clinical chemistry in MT-positive and MT-negative WD patients.(A) Ceruloplasmin serum levels and (B) copper serum levels are not statistically significant between both groups.(C) 24 h urinary copper excretion is significantly lower in MT-negative compared to MTpositive patients.(D) Hepatic copper content is not significantly different between MT-positive and MT-negative WD patients.

Figure 5 .
Figure 5. MT staining in patients without WD.(A) PFIC3: Perivenular MT expression in initial liver biopsy.(B) PFIC3, same patient after 6 years of D-penicillamine treatment for assumed WD: Perivenular MT expression.(C) PFIC3, same patient after 8 months of dual trientine and zinc treatment: Diffuse and strong MT immunostaining.(D) Unaffected sibling with positive family history for WD: Negative MT immunostaining.Liver copper content was normal and the most prevalent ATP7B gene mutation (p.H1069Q) was not detected.Scale bar = 200 lm.

Table 1 .
Patients' characteristics of the WD cohort

Table 4 .
Comparison of MT-positive and MT-negative WD patients