Hereditary transthyretin-related amyloidosis is frequent in polyneuropathy and cardiomyopathy of no obvious aetiology

Abstract Background Hereditary Transthyretin-Related Amyloidosis, a clinically heterogeneous autosomal dominant disease caused by pathogenic variants in the TTR gene, is characterized by the deposition of insoluble misfolded protein fibrils. The diagnosis, especially in non-endemic areas, is typically delayed by 4–5 years; a misdiagnosis due to clinical heterogeneity is common. The study objective was to define the prevalence of Hereditary Transthyretin-Related Amyloidosis in patients with polyneuropathy and/or cardiomyopathy of no obvious aetiology. Method A multicenter observational “Epidemiological analysis for the hereditary Transthyretin-Related AMyloidosis”—TRAM study was performed in Germany, Austria, and Switzerland. Results A total of 5141 participants were recruited by 50 neurologic and 27 cardiologic specialized centres. Genetic analysis demonstrated a 1.1% Hereditary Transthyretin-Related Amyloidosis positivity rate among patients with polyneuropathy and/or cardiomyopathy of not obvious aetiology. Twenty-one various TTR variants (TTR-positive) were identified. Body Mass Index was lower in the TTR-positive patients as an indicator for the involvement of the autonomic nervous system; the age of onset of clinical manifestations was higher in TTR-positive patients. There were no other genotype-phenotype correlations or the prevalence of specific clinical manifestations in TTR-positive patients. Conclusions Our data support the fact that Hereditary Transthyretin-Related Amyloidosis is underdiagnosed in polyneuropathy and cardiomyopathy patients. Routine implementation of genetic testing is recommended in patients with unexplained polyneuropathy and/or cardiomyopathy to accelerate the earlier diagnosis and the time-sensitive treatment initiation. KEY MESSAGES More than 5.000 participants with CM and/or PNP of no obvious aetiology were recruited in the observational “Epidemiological analysis for the hereditary Transthyretin-Related AMyloidosis” TRAM study and screened for pathogenic TTR variants. The study demonstrated >1% of patients with CM and/or PNP of unclear aetiology are positive for a pathogenic TTR variant. Routine genetic testing is recommended in patients with unexplained CM and/or PNP to accelerate the initial diagnosis and timely treatment initiation.


Introduction
Hereditary transthyretin-related amyloidosis (hATTR) is an autosomal dominant condition caused by a pathogenic variant in the TTR gene, that is located on chromosome 18q12.1 and comprises 4 exons and 5 introns [1]. Pathogenic variants in the TTR gene, coding for the protein transthyretin, result in destabilization and misfolding of this protein. To date, more than 150 TTR pathogenic variants are described which cause amyloid formation [2]. Still, the penetrance is not clear for most of the variants. hATTR has been considered an endemic disease with early or late-onset in Japan, Sweden, and Portugal.
hATTR manifests with a variety of clinical presentations of which cardiomyopathy (CM) and polyneuropathy (PNP) are the most common [3,4]. Due to the heterogeneity of the phenotype, the diagnosis is delayed by 4-5 years, especially in non-endemic areas, and the misdiagnosis is a frequent burden for such patients [5][6][7].
hATTR should be considered in older persons who have been hospitalized for heart failure, elevated troponin or N-terminal pro-brain natriuretic peptide (NT-proBNP) and intolerance of angiotensin-converting enzyme (ACE) inhibitors, angiotensin receptor blockers, or b blockers [8]. Lumbar spinal stenosis [9], previous orthopaedic procedures [10], spontaneous biceps tendon rupture [11], carpal tunnel syndrome [12] may be early indicators of hATTR in patients with progressive disabilities. The most typical presentation of hATTR PNP is a progressive, length-dependent, mixed sensory and motor peripheral polyneuropathy, which usually begins with the loss of thermal and pain sensation in the feet, slowly ascends the limbs, and is typically associated with variable autonomic disturbances [1].
Usually, hATTR diagnosis requires the histologic demonstration of amyloid deposits in biopsy specimens. However, initial genetic testing is a fast, cheap (due to the small gene size), and reliable analysis to prove for hATTR. The most prevalent variant associated with amyloid PNP is NM_000371.3:c.148G > A (p.Val50Met) [13,14]. Affected patients with this genotype display a variable disease penetrance which is influenced by origin, gender, parental gene transmission, and age [1,15]. The most common variant involved in the cardiac form of hATTR is NM_000371.3:c.424G > A (p.Val142Ile) [16,17]. The NM_000371.3:c.238A > G (p.Thr80Ala) variant is another frequent cause of amyloid PNP and CM [18].
Within Europe, the incidence of hATTR is highly variable. There are large areas in Portugal and Sweden where the disease is endemic. The described prevalence for Portugal is 192 per 1,000,000 and for Sweden it is 26 per 1,000,000 inhabitants [13]. Smaller endemic areas have been identified in Cyprus (56 per 1,000,000 inhabitants) and Majorca (50 per 1,000,000 inhabitants) [19,20]. In the rest of Europe, cases of hATTR are mainly seen as sporadic [21]. However, up to now, only a few systematic prospective studies describing the hATTR prevalence among larger groups of patients have been published [22,23].
The aim of the study was therefore based on the NGS-sequencing of the TTR gene to investigate the hATTR prevalence in a large risk group: patients with PNP and/or CM of no obvious aetiology. In the case of the detection of a pathogenic variant in TTR, patients obtain the diagnosis for hATTR for which causative treatments are available.

Methods
The "Epidemiological analysis for the hereditary TransthyRetin-related AMyloidosis" (TRAM) study was approved by Ethics Committee at the Medical Faculty of the University of Rostock: approval A 2018-0069, 19.04.2018. Ethics approval was obtained for each participating centre in Germany, Austria, and Switzerland. A list of all Ethics committees that have approved the study in all three countries is provided (Supplementary Information).
The study was registered in ClinicalTrials.gov with the number NCT03237494.

Study design
The TRAM study is an observational epidemiological study that has been conducted in Germany, Austria, and Switzerland. Participants with PNP and/or CM of no obvious aetiology from 77 centres with either a neurologic (n ¼ 50) or cardiologic (n ¼ 27) specialty were enrolled (Supplementary Table 1S).

Eligibility criteria
The inclusion criteria are as follows: (1) informed consent is obtained from the participant; (2) the participant is diagnosed with CM and/or PNP of no obvious aetiology; (3) the participant is aged between 18 and 85 years; (4) the participant has not undergone chemotherapy for any solid or hematological malignancies. There are no specific exclusion criteria.
All enrolled participants have visited the study site one time. At this visit, a Case Report Form (CRF) was completed, and a blood sample was drawn and applied on a Dried Blood Spot (DBS) filter CentoCard V R . The blood sample was used for genetic analysis to screen for hATTR.

Clinical information
All clinical information is documented by the physician in an electronic CRF that asks for physical examinations, demographic details, current symptoms, medication, and family history. The following current symptoms and concomitant medication were questioned by the physician with symptoms of PNP grouped at the top and symptoms of CM grouped at the bottom of the table (Table 1). Height and weight were used to calculate the Body Mass Index (BMI) of each subject and echocardiography was used to assess left ventricular thickness. All CRFs are stored with appropriate measures to maintain data privacy regulations (all data are encrypted in transit with TLS1.2). CENTOGENE GmbH operates core services in an external Data Centre. Based on multiple certifications, the Data Centre provider ensures compliance to applicable international legislation (e.g. GDPR and HIPPA) and a high level of IT Security and Business Continuity. The Data Centre is certified for different quality standards: ISO 9001-Quality Management System; ISO 27001-Information Security Management System.

Genetic testing
All enrolled participants have undergone genetic testing for hATTR. DNA from the CentoCard V R was extracted using the Extracta DNA Prep Kit according to the manufacturer's instructions (QuantaBio). Afterward, the concentration of the DNA was measured with the Quantifluor system in a Fluoroskan Ascent TM FL Microplate Fluorometer (ThermoFisher). The extracted DNA was used in a multiplex PCR generating 12 TTR-specific amplicons (Primer sequences and PCR program in Annexe 1). PCR products from both pools are pooled together and purified using AMPure XP magnetic beads (Beckman Coulter) in a 1:1 PCR product: beads ratio to remove unbound primers. A second PCR amplification is performed adding Illumina-compatible and patient-specific barcoded index primers to each amplicon. After another purification round, the samples are pooled together and sequenced via Next-Generation Sequencing with a 300 cycle v2 MiSeq kit and a MiSeq Instrument (Illumina Inc.). The resulting 150 bp long paired-end reads are aligned with bioinformatic methods and compared to the reference sequence TTR: NM_000371.3 to identify single nucleotide variants (SNVs). All variants are classified according to the American College of Medical Genetics (ACMG) into Class 1 (pathogenic), Class 2 (likely pathogenic), Class 3 (variance of uncertain significance, VUS), Class 4 (likely benign), and Class 5 (benign), only class 1, 2, and 3 are reported via a medical report. These classifications were assessed using existing information from HGMD V R [24], ClinVar [25], and the in-house platform CentoMD [26].
In cases where no relevant SNVs were detected by the above described NGS, an in-house developed deletion/duplication analysis was conducted, using endpoint quantification by Real-Time PCR. These RT-PCRs are conducted on a LightCycler 480 II (Roche) with VeriQuest Fast SYBR Green qPCR Master Mix (ThermoFisher) and specifically designed primers targeting each TTR exon. The expression of each TTR exon was evaluated by a relative quantification analysis (DDCT-Method) with GUSB serving as reference gene, and thereby showing whether a deletion or duplication is present.

Statistical methods
Descriptive statistical analysis methods (absolute and relative frequencies, mean, standard deviation, median, and range) were used to describe the total study population and to compare TTR-positive and TTRnegative patients. As exploratory statistical tests, v 2 -Test was used to test differences in nominal data, and t-test for independent samples was used to test continuous variables between TTR-positive and TTR-negative patients. We also report effect sizes and 95% confidence intervals (CI), namely Cohen's d for continuous data and Cramer's V for nominal data. As statistical Software SPSS (IBM Corp. Released 2019) [27] and R [28], package effect size [29] was used. A twosided significance level of a ¼ 0.05 was considered. However, since no adjustment for multiple testing was applied in this exploratory study, interpretation of pvalues should be done cautiously. Interpretation of results is therefore primarily based on effect sizes and 95% CI.

Study participants
The

Genetic findings
Sequencing of the TTR gene in the 5141 study participants revealed the presence of a reportable heterozygous variant in 55 subjects (1.1% of all; subsequently referred to as TTR-positives). Forty-two of these (76.4%) were diagnosed with Class 1, five (9.1%) with Class 2, and eight (14.5%) with Class 3 variant ( Figure  1(A')). There was a total of 21 distinct variants; twelve pathogenic, two likely pathogenic, and seven of uncertain significance (Figure 1(A")). Nineteen variants were predicted to result in missense alteration at protein level; two affect intronic splice motifs. The two variants, c.148G > A (p.Val50Met) and c.424G > A (p.Val142Ile), were outstandingly frequent, while all other variants were observed between one and three times each (Figure 1(B)). The clinical and demographic characteristics of each TTR variant previously described in the literature [2,25,[30][31][32][33][34][35][36][37][38][39][40] and found in our study are shown in Table 2.

Characteristics of TTR-positive vs. TTR-negative
In the first step, the study participants were genetically stratified into 55 TTR-positive vs. 5086 TTR-negative, next test was performed whether these subgroups differed in any of the available clinical and demographic parameters. The distribution of sex was similar in TTR-positive and TTR-negative with a slightly higher proportion of males in the TTR-positive [69 vs. In contrast, none of the 26 individually documented symptoms was substantially more frequent in either of the two groups ( Figure 3). However, as odds ratios (TTR-positive over TTR-negative) were still >1 for 20 of the 26 symptoms [e.g. anaemia, history of carpal tunnel syndrome, diarrhoea, dyshidrosis, abnormal (fluttering) heartbeat, and unintentional weight loss], we also compared the average number of symptoms per subject between the groups. Although not substantially but the symptoms load was higher in the TTR-positive group [mean difference: 1.3, 95% CI: À0.1-2.7, p ¼ .074, Cohen's d: 0.34, 95%CI: 0.05-0.63, Figure 2(F)]. Decision-trees analyses that considered the absence/presence of the individual symptoms did not identify a model in which both sensitivity and positive prediction value were >50% (data not shown).
The TTR variant c.424G > A (p.Val142Ile) was only in fourth place in prevalence in Europe (1.4%) (with most patients from France), however the most frequent in the USA among African Americans [41]. The c.238A > G (p.Thr80Ala) variant, most frequent in Northwestern Ireland [34], in our study, occurred in 2 (4%) subjects. European TTR variants, c. [31] were not found in our study. The variants c.424G > A (p.Val142Ile), c.88T > C (p.Cys30Arg), c.206C > T (p.Thr69Ile), c.212A > G (p.Glu71Gly), and c.250T > C (p.Phe84Leu) identified in our study were not previously described in Germany. Gender distribution in our study is in line with the majority of previous studies where proportions of men are reported from 61 to 86% [42]. However, only 52% of the male are among hATTR patients in the THAOS registry [31].
The differences between our data and the data of the THAOS registry, including its German cohort, can be explained by different mosaic geographic (within Germany) distributions of hATTR participants both by prevalence and by variants.
Clinical phenotype of hATTR in the THAOS subjects was CM in 14.9%, CM þ PNP in 21.1%, and PNP in 64.0%. The proportion of the PNP phenotype in the cohort was driven by Portugal, Sweden, 79.6 and 64.1%, respectively. In Germany and Denmark most subjects had CM, 51 and 78%, respectively [31], which is in contrast to our study revealing the PNP phenotype in 51% of participants, CM phenotype in 20%, and mixed presentation in 29% (Figure 2(D)). This difference can be explained as the THAOS study included a large portion of wild-type ATTR which is predominantly associated with CM, whereas wild-type ATTR subjects were excluded from our study.
In our study c.148G > A (p.Val50Met) historically known for its PNP clinical pattern [2,[30][31][32] caused pure CM in 2 (12%) cases and CM þ PNP clinical presentation in 5 (29%) cases. Known as CM variant, c.424G > A (p.Val142Ile) [2,31,32,34,41], presented in our study as PNP in 1 (13%) case, mixed CM/PNP pattern in 2 (25%) cases. All three participants with the variant c.233T > A (p.Leu78His) presented with PNP, instead of CM as described in the literature [2,32]. In the majority of c.148G > A (p.Val50Met) subjects in Portugal the onset is typically at 30-40 years, "early" onset, while the majority of patients with c.148G > A (p.Val50Met) in Sweden has the onset after 50 years "late" onset [21]. A "late" and "early" distribution of the onset among patients with the c.148G > A (p.Val50Met) variant was reported in Japan [43]. Of 24 subjects from Germany with c.148G > A (p.Val50Met) variant in THAOS registry 13 (54%) had "early" onset, 11 (46%) "late" onset [31]. Two out of 17 (17%) of our participants with c.148G > A (p.Val50Met) variant had "early" Portuguese-type, the other (83%) had "late" Swedish-type of disease onset. Although it has been reported previously that the early onset is not typical for non-c.148G > A (p.Val50Met) variants [44,45], in our study two subjects with c.424G > A (p.Val142Ile) and c.338T > C (p.Val113Ala) variants had the onset at 30 and 34 years, accordingly. Such is highly likely due to various screening strategies: Testing the general population vs. risk cohort (CM and/or PNP individuals). Thus, the TRAM study complements the previous studies in terms of clinical manifestations and age of onset of various variants of hATTR.
A substantially lower BMI in the TTR-positive group was detected. Autonomic manifestations are present in 50-80% of hATTR patients, most frequently patients demonstrate diarrhoea, constipation, postural hypotension, urinary incontinence, or erectile dysfunction [46]. Different variants might reflect different likelihoods for autonomic dysfunction. In the THAOS registry, the prevalence of autonomic dysfunction was higher in patients with c.148G > A (p.Val50Met) [31]. Modified BMI (mBMI, BMI multiplied by serum albumin level to compensate for edema) value has demonstrated a strong correlation with survival in patients with c.148G > A (p.Val50Met) variant [47]. The difference in BMI between TTR positive and negative patients is because the spectrum of "unknown" diseases manifested by PNP and CM (TTR negative) is not characterized by the involvement of the autonomic nervous system.

Strengths and limitations
Our study is the largest systematic cohort in an hATTR risk group with more than 5000 participants demonstrating CM, PNP, and both CM/PNP of no obvious aetiology.
A potential limitation of our study is the lack of a longitudinal component to investigate how clinical symptoms progress over the time or with the treatment.

Conclusions and health care implications
Our study showed >1% of patients with CM and/or PNP of unclear aetiology are positive for a pathogenic variant in TTR gene. This confirms the need for mandatory genetic testing and counselling for all patients in this group. ticipant to take part in this study. We thank all Principal Investigators and their study team members (Supplementary Table).
Funding The study was funded by Alnylam Pharmaceuticals Inc., Cambridge, Massachusetts, US. The funders of this study had no role in the study design, data analysis, data interpretation, or writing of the manuscript.

Funding
The study was funded by Alnylam Pharmaceuticals Inc., Cambridge, MA, USA. The funders of this study had no role in the study design, data analysis, data interpretation, or writing of the manuscript.

Data availability statement
The data that support the findings of this study are available upon reasonable request from the corresponding author, J.H. (email: joerg.hartkamp@centogene.com). The data are not publicly available due to containing information that could compromise the privacy of research participants.