Subjects

The research performed with human samples was approved by the Ethics Committee of the Hospital Universitari i Politècnic La Fe (Valencia, Spain; protocol code: 2019/0052; 22/05/2019) and was conducted in compliance with the principles of the Declaration of Helsinki, Good Clinical Practice guidelines, and local regulatory requirements. Informed consent was obtained from all participants prior to research initiation. All data were anonymised.

All of the experimental procedures involving animals were conducted in accordance with the European Union Guidelines for the Care (European Union Directive, 2010/63/EU) and the guidelines for the use of laboratory animals. The experimental design was approved by the Ethical Committee for Animal Testing of the University of Navarra (Pamplona, Spain; protocol code: CEEA/066-22; 02/09/2022).

Statistical power of the data was evaluated from the normalised counts of the experimental subjects, using Bioconductor’s RnaSeqSampleSize R package (https://www.bioconductor.org/packages/release/bioc/html/RnaSeqSampleSize.html). For a cohort of 20 individuals, a probability (power) of 0.70 was obtained, considering that 10% of miRNAs with a differential expression profile may have a relevant role.

Patients

A first cohort (discovery) including 20 in-depth phenotyped WD patients (Supplementary Table S1) was screened by high-throughput small RNA-seq (RNA-sequencing). For validation purposes, a second cohort (validation) and a third cohort (follow-up) comprising 22 and 25 WD patients, respectively (Supplementary Tables S1 and S2), were screened by quantitative PCR (qPCR). The discovery and follow-up cohorts comprised the same patients studied 3 years apart. Two patients included in the discovery cohort could not be further studied because they did not attend the medical check-up, while seven new patients diagnosed during these years and not evaluated in the discovery cohort were included in the follow up cohort. Individuals from the discovery and follow-up cohorts were supervised at Hospital Universitari i Politècnic La Fe (Valencia, Spain) and at Hospital General Universitari d’Elx (Alicante, Spain). Individuals from the validation cohort were recruited at Complejo Hospitalario Universitario Insular Materno Infantil de Gran Canaria (Las Palmas, Spain). These clinical series have been partially reported [6, 16]. Differential diagnosis is based on clinical features, biochemical criteria, presence of corneal Kaiser-Fleischer (KF) ring and other hepatic disturbances. All patients included met the following inclusion criteria: (1) Leipzig score ≥ 3 at diagnosis (without considering ATP7B mutations); and (2) conclusive genetic analysis. Other chronic liver diseases (hepatitis C or B, immune mediated liver diseases or alcohol related liver disease) were ruled out. We excluded patients who had undergone liver transplantation. Supplementary Tables S1 and S2 show the clinical and biochemical features of the patients.

Peripheral blood samples for plasma isolation and biochemical analysis were obtained during clinical follow-up visits. Biochemical determinations analysed using standard protocols (upon request): AST, aspartate transaminase; ALT, alanine aminotransferase; GGT, gamma-glutamyl transferase; bilirubin; AP, alkaline phosphatase; cholesterol; and triglycerides. For miRNA analyses, EDTA-blood samples were processed for plasma separation within 2–3 h following collection. After centrifugation at 1900×g for 10 min (4 °C), plasma fraction was aliquoted and stored at – 80 °C.

Healthy subjects

Plasma from controls were provided by the Biobanco para la Investigación Biomédica y en Salud Pública de la Comunitat Valenciana (PT13/0010/0064), integrated in the Spanish National Biobanks Network and in the Valencian Biobanking Network. Samples were processed following standard operating procedures. Demographic data from controls recorded in general health and lifestyle questionnaires were provided (Supplementary Table S3). Healthy subjects were matched by sex and age with patients, and those who registered neoplasms, cardiovascular, respiratory, bone, mental or endocrine disease risk factors were discarded.

Mice

Serum and liver samples of wild-type (WT) and Atp7b−/− (WD) mice were received from Dra. González-Aseguinolaza laboratory (Centro de Investigación Médica Aplicada, Pamplona, Spain). Atp7b−/− on a C57BL/6 J background were bred and maintained under pathogen-free conditions and genotyped at 3 weeks of age as described [14, 15]. These Atp7b−/− mice show no ATP7B expression in the liver and exhibit the typical biochemical and physiopathological alterations observed in WD patients, except for the neurological signs [12, 13]. Alterations comprise high copper excretion in urine, low holoceruloplasminemia, high serum transaminase levels and increased liver copper concentration with associated hepatocellular damage [12,13,14,15].

The cohort comprised 20 or 10 WD mice for the studies using serum or liver, respectively, with the same number of WT mice. Serum samples were collected at 6–7, 10, 12, 16, 20, 30 and 40 weeks old, while liver samples were obtained at 20, 30 and 40 weeks old. Samples were immediately frozen at -80ºC and no tissue preservatives were used.

Circulating miRNA-enriched total RNA isolation

Total RNA enriched for circulating miRNA fraction was extracted from 400 µL human plasma, 10 µL mouse serum and 20 mg mouse liver per subject and column using miRNeasy Mini kit (Qiagen, Hilden, Germany), following manufacturer’s protocol (miRNeasy Mini Handbook 11/2020). Aliquots from plasma/serum samples were thawed on ice and centrifuged (16,000×g, 5 min, 4 °C) to remove cell debris prior to RNA isolation. During isolation from mouse serum, 5.1 × 108 copies of synthetic cel-miR-39 (IDT, Leuven, Belgium) were added into each sample to be used as exogenous miRNA reference.

Small RNA library preparation and high-throughput sequencing

Individual small RNA libraries were prepared from 6 µL of total RNA enriched for circulating miRNA fraction isolated from human plasma samples with NEBNext Multiplex Small RNA Library Prep kit (New England Biolabs, Ipswich, MA, USA). To improve final miRNA sequencing yield, library prep protocol was optimised according to dilution 1:5 of 3′ and 5′ SR RT primer and 16× PCR cycles for indexing. Pre- and post-gel size selection PCR-indexed libraries were purified with Nucleospin Gel and PCR Clean-Up (Macherey–Nagel, Düren, Germany) following manufacturer’s protocol (Macherey–Nagel—03/2023, Rev. 08) for PCR clean-up and DNA purification from polyacrylamide gel, respectively.

Pre- and post-gel size selection library QC (Quality Control) was performed using High Sensitivity D1000 ScreenTape in a TapeStation 4200 (Agilent Technologies, Santa Clara, CA, USA). Post-gel purified libraries were quantified by qPCR with KAPA Library Quantification Kit (Kapa Biosystems, Wilmington, MA, USA) and pooled for sequencing in Illumina HiSeq 2500 (1 × 50 bp, v4).

Bioinformatics analysesPrimary analysis

FASTQ raw data files obtained from small RNA-seq were processed for computational analysis. Before and after removing adapter sequences and selecting reads by size with Cutadapt (v2.6), QC with FastQC (v0.11.8) was performed to check if reads size distribution and quantity was compatible with mature miRNA (16–28 bp) and sufficient for differential representation analyses. Selected reads were aligned against pre-miRNA sequences (hairpin) from the latest miRBase v22 using Bowtie (v1.1.2). Counts of reads matching hairpin arms were obtained with Subread (v1.6.0) and a custom GFF (General Feature Format) file with 5p and 3p coordinates corresponding to annotated human mature miRNAs in miRBase v22.

Differential expression analysis and functional enrichment

To identify differential expressed mature miRNAs, edgeR package was used [17]. Starting from extracted mature miRNA counts, those low-represented were filtered before setting per-sample library size for normalisation by the trimmed mean of M values (TMM) method. QLF (Quasi-Likelihood F) and LRT (Likelihood Ratio Test), both considered generalised linear models (GLM), were applied to determine differentially expressed mature miRNAs in patients compared to controls (design 1), and in a second approach, including the covariables sex and age, to adjust comparisons with controls (design 2). miRNAs with a false discovery rate (FDR) < 0.05 were considered as significantly deregulated.

miRNA detection by quantitative PCR

To obtain cDNA from mature miRNAs, a 2–2.5 µL total RNA enriched for circulating miRNA fraction isolated from human plasma or murine serum or 10 ng if isolated from tissue sample was used with the TaqMan Advanced miRNA cDNA Synthesis Kit (Applied Biosystems, Foster City, CA, USA).

Levels of mature miRNAs of interest were detected by qPCR with TaqMan Advanced (Applied Biosystems, Foster City, CA, USA) probes (Supplementary Table S4) and master mix in a LightCycler 480 II (Roche, Mannheim, Germany) programmed. Pre-amplified cDNA was diluted 1/10 and 2.5 µL per reaction were used as template for qPCR, scaling reaction components to a 10 µL final volume.

Relative miRNA expression levels in case sample (human or murine) compared to control sample was calculated as \(^-\text\Delta \text\right)}\), being \(\Delta Ct=}_}-}_}\). As reference miRNAs for qPCR analysis in human plasma samples, endogenous levels of miR-16-5p and miR-484 were used, while for mouse serum, endogenous miR-484 and exogenous cel-miR-39 spike-in levels were considered to normalise data. Finally, for mouse liver samples, endogenous miR-16-5p was used as reference miRNA.

Statistical analyses

Shapiro–Wilk test was applied to biochemical parameters to check for normality distribution, and as all resulted negative, non-parametric Wilcoxon signed-rank test was used. Paired t test was performed to assess differences of relative miRNAs expression levels, represented as log2FC, between cases and controls. For in-group comparisons, two-sample t test was used instead.

For murine samples, Shapiro–Wilk and Levene tests were applied to check for normality distribution and homoscedasticity of miRNA expression levels represented as log2FC. Two-way mixed ANOVA (analysis of variance between-sex factor and within-age factor) and pairwise post hoc t test were used to determine differences between groups.

The Spearman’s Rho (Rs) coefficient of correlation was used to establish the association of relative miRNAs expression levels with biochemical parameters in human plasma and in murine serum, and additionally, with Leipzig score (including ATP7B mutations) and age in human plasma.