Article Text
PDF
Abstract
Objectives Familial Mediterranean fever (FMF) can be complicated by AA amyloidosis (AAA), though it remains unclear why only some patients develop amyloidosis. We examined the gut microbiota composition and inflammatory markers in patients with FMF complicated or not by AAA.
Methods We analysed the gut microbiota of 34 patients with FMF without AAA, 7 patients with FMF with AAA, 19 patients with AAA of another origin, and 26 controls using 16S ribosomal RNA gene sequencing with the Illumina MiSeq platform. Associations between bacterial taxa and clinical phenotypes were evaluated using multivariate association with linear models statistical method. Blood levels of interleukin (IL)−1β, IL-6, tumour necrosis factor-α and adipokines were assessed by ELISA; indoleamine 2,3-dioxygenase (IDO) activity was determined by high-performance liquid chromatography.
Results Compared with healthy subjects, specific changes in faecal microbiota were observed in FMF and AAA groups. Several operational taxonomic units (OTUs) were associated with FMF. Moreover, two OTUs were over-represented in FMF-related AAA compared with FMF without AAA. Additionally, higher adiponectin levels and IDO activity were observed in FMF-related AAA compared with FMF without AAA (p<0.05).
Conclusion The presence of specific changes in faecal microbiota in FMF and in FMF-related AAA suggests that intestinal microorganisms may play a role in the pathogenesis of these diseases. These findings may offer an opportunity to use techniques for gut microbiota manipulation.
- familial Mediterranean fever
- AA amyloidosis
- microbiota
- adipokines
- indoleamine 2,3-dioxygenase
Statistics from Altmetric.com
Key messages
What is already known about this subject?
Several risk factors for AA amyloidosis (AAA) in familial Mediterranean fever (FMF) have previously been found, including environmental factors.
Some data point toward the influence of the environment on AAA occurrence, particularly the gut microbiota.
What does this study add?
FMF was associated with specific changes in faecal microbiota, suggesting that the gut microbiota might be involved in the clinical expression of FMF.
In patients with FMF, amyloidosis was independently associated with a specific alteration in the microbiota composition, suggesting that intestinal microorganisms may play a role in AAA pathogenesis.
How might this impact on clinical practice or future developments?
The presence of specific changes in faecal microbiota in patients with FMF and AAA may allow for the use of techniques for gut microbiota manipulation to prevent inflammation and amyloidosis occurrence.
Introduction
Familial Mediterranean fever (FMF) is the most common monogenic autoinflammatory disease, secondary to mutations in the MEFV gene, which encodes the protein pyrin.1 AA amyloidosis (AAA) is the most serious complication of FMF and is characterised by the presence of extracellular deposits of an amorphous substance, including an amyloidogenic derivative of the serum amyloid-associated (SAA) protein.2 Other diseases, mainly chronic idiopathic inflammatory diseases (such as chronic inflammatory rheumatic diseases or inflammatory bowel diseases), autoinflammatory diseases other than FMF, chronic infections and obesity, can be complicated by AAA.3 The alarmins S100A8 and A9 have been studied as inflammatory biomarkers for many of these diseases,4 whereas changes in adipokine levels are observed in obesity, including proinflammatory leptin and adiponectin, which has an anti-inflammatory effect on atherosclerosis.5 6
The gut microbiota has a key role in several physiological functions, particularly in host energetic and vitamin metabolism and protection against pathogens, due to components of these bacteria and/or the production or processing of metabolites.7–9 For example, tryptophan metabolism represents a significant pathway by which the gut microbiota influences the host. These metabolites can be of microbial or endogenous origin and are mainly metabolised by indoleamine 2,3-dioxygenase (IDO, EC 1.13.11.52), the activity of which is stimulated by several proinflammatory components, such as lipopolysaccharide, tumour necrosis factor (TNF)-α and interleukin (IL)-1 and IL-6. Thus, a proinflammatory environment will induce overactivation of IDO, which is responsible for many immunomodulatory activities and a shortage of tryptophan.10 Nevertheless, overactivation of IDO can also have proinflammatory effects under certain conditions.11–14
To date, it remains unclear why only some patients with FMF develop AAA. Several risk factors have been found, such as male sex and genetic factors such as specific mutations in the MEFV gene (M694V homozygosity, associated with a more severe phenotype, including more frequent and long attacks and higher dose of colchicine needed15 16) or in the gene encoding SAA.17–19 More interestingly, some risk factors are environmental, such as the country of residence.17 20 Association between the single-nucleotide polymorphism Arg753Gln in Toll-like receptor 2, which belongs to the microbe recognition pathway, has also been reported.21 Other data highlight the influence of the environment on AAA occurrence, particularly the gut microbiota. Indeed, SAA acts as an opsonin, and intestinal epithelial cells can synthesise it in response to the gut microbiota.22 23 Moreover, a high-fat diet can induce AAA in a mouse model overexpressing hepatic SAA,24 and several bacteria can produce amyloid-enhancing factors that can cross species barriers and may be transmitted by ingestion.2 25–27 Digestive tract involvement was found in nearly all cases of amyloidosis in autopsy series,28 and evidence of gut microbiota involvement in the pathogenesis of another disease involving amyloid plaques, Alzheimer's disease, is accumulating, both in vitro and in vivo, in animal models as well as in humans.29–32
Based on these findings, the gut microbiota appears to be a credible candidate for both the overproduction of the SAA protein and synthesis of amyloid-enhancing factors in FMF. Here, we investigate the gut microbiota and several proinflammatory markers, cytokines, adipokines and tryptophan metabolites in patients with FMF, with or without AAA, in patients with non-FMF-related AAA and in healthy controls. We tested the hypothesis of specific dysbiosis in patients with FMF and in those with FMF complicated by AAA.
Methods
Study design and control groups
This cross-sectional study included patients with FMF followed up in the national reference centre for FMF and inflammatory amyloidosis in Tenon Hospital, Paris, and fulfilling the Tel Hashomer criteria with at least one MEFV mutation.33 Patients were diagnosed as having AAA-complicating FMF in the presence of proteinuria with biopsy-proven AAA (positive immunohistochemistry with an anti-SAA antibody or proteomics). The inclusion period ranged from April 2016 to March 2018. Demographic and clinical data were retrieved from the patients' clinical files.
Two other control groups were assembled: patients with AAA secondary to a cause other than FMF and healthy controls; the latter did not have a chronic disease that required systemic treatment. None of the patients took antibiotics in the 6 weeks preceding stool sampling. Stool and blood samples were collected at the same time.
All patients provided their informed consent.
Patient and public involvement
This research was conducted without patient involvement.
MEFV genotyping
Genomic DNA was isolated from peripheral leucocytes using standard procedures. Direct sequencing of the MEFV gene was performed using the Sanger method. Exons and flanking intronic sequences (NM_000243) were PCR amplified from genomic DNA. The primer sequences used are available on request. The PCR products were sequenced using the Big Dye Terminator reaction kit (Applied Biosystems, Foster City, California, USA) and a 96-capillary ABI Prism sequencer and were then analysed using SeqScape V.2.6 software (Applied Biosystems).
Gut microbiota analysis
Stool samples were immediately cooled at 4°C and then stored at −80°C within 6 hours until DNA extraction. Following a previously described method,11 the stool samples were resuspended in 250 µL of 4 M guanidine thiocyanate and 40 µL of 10% N-lauroyl sarcosine; 500 µL of 5% N-lauroyl sarcosine was then added. DNA was extracted by mechanical disruption of the microbial cells in a FastPrep instrument (MP Biomedicals) after the addition of 500 mg of 0.1 mm diameter glass beads. The nucleic acids were recovered from the clear lysate by alcohol precipitation. The DNA was stored at −20°C until 16S ribosomal RNA (rRNA) gene sequencing.
Microbial diversity was determined for each sample by amplifying the V3 and V4 hypervariable regions of the 16S rRNA gene (5′-TACGGRAGGCAGCAG-3′ (sense) and 5′-CTACCNGGGTATCTAAT-3′ (antisense)) using a 16S-amplicon library preparation protocol (Metabiote, GenoScreen, Lille, France). According to the manufacturer’s protocol, 16S rRNA gene PCR was performed using 5 ng of genomic DNA with 192 bar-coded primers (Metabiote MiSeq Primers) at a final concentration of 0.2 µM and an annealing temperature of 50°C for 30 cycles. The PCR products were purified using an Agencourt AMPure XP-PCR Purification system (Beckman Coulter, Brea, California, USA), quantified according to the manufacturer's protocol and multiplexed at equal concentrations. A 300 bp paired-end sequencing protocol with the Illumina MiSeq sequencing platform (Illumina, San Diego, California, USA) was applied for 16S rRNA gene sequencing at GenoScreen, Lille, France. Raw paired-end reads were quality filtered using the PRINSEQ-lite PERL script and were assembled using fast length adjustment of short reads to improve genome assemblies with a minimum overlap of 30 bases and 97% overlap identity; primer sequences were removed using CutAdapt.
The sequences were then demultiplexed and quality filtered using the ‘quantitative insights into microbial ecology’ (QIIME V.1.9.1) software package, and Illumina reads were joined using the fastq-join method (https://expressionanalysis.github.io/ea-utils/). We used the UCLUST algorithm to assign operational taxonomic units (OTUs) to sequences with a 97% threshold of pairwise identity, and the sequences were classified taxonomically using the Greengenes reference database. Rarefaction was performed (13 000 reads per sample) and used to compare the abundances of OTUs across samples.
Serum levels of inflammatory proteins, cytokines and adipokines, and IDO activity
C reactive protein (CRP) and SAA protein levels were determined by immunonephelometry using an IMMAGE analyser (Beckman Coulter, Villepinte, France). The IL-6 level was assessed by chemiluminescence enzyme immunoassay using a LUMIPULSE analyser (Fujirebio, Tokyo, Japan). The blood levels of other inflammatory cytokines (IL-1β and TNF-α), adipokines (leptin and adiponectin) (Quantikine, R&D Systems, Oxford, UK) and S100A8/A9 proteins (Bühlmann, Amherst, New Hampshire, USA) were measured by ELISA according to the manufacturers' instructions. For these circulating biomarkers, the control group was formed in part by 17 healthy male volunteers from the French study METASPERME (N°IDRCB: 2011-A01052-39) and by 7 healthy women volunteers presented in a previous publication.34 None of the healthy controls were diabetic or obese.
Tryptophan and kynurenine levels were evaluated by high-performance liquid chromatography with a coulometric detection system (ESA Coultronics; ESA Laboratories, Chelsford, Massachusetts, USA). The kynurenine:tryptophan ratio was determined from the kynurenine and tryptophan absolute concentrations and was used as a marker for IDO activity. The values for our cohort were compared with 48 age-matched and sex-matched healthy controls.
Statistical analyses
The Bray-Curtis index was used for β-diversity analysis, and the results are presented in the form of principal coordinate analysis, where each sample is coloured according to the disease. Statistical significance between the studied groups was evaluated using the non-parametric analysis of similarities test with 9999 permutations. Shannon and Chao1 indexes were applied to assess α-diversity, and statistical significance was evaluated using the nonparametric Mann-Whitney test. Relative abundances and associations between bacterial taxa and clinical and biological data were assessed using the multivariate association with linear models (MaAsLin) statistical method,35 taking into consideration the effects of potential confounding factors such as age, sex, body mass index (defined as weight in kilogram divided by height squared), smoking status and treatment. Only OTUs present in more than 50% of the samples were considered in the analysis.
GraphPad Prism Version 7 software (GraphPad Software, San Diego, California, USA) was used for statistical analyses of biological parameters. Categorical variables are reported as percentages and were compared using the χ² or Fisher’s test, according to expected frequencies. Continuous variables were expressed as medians and quartile 1–quartile3 values and were analysed using the bilateral Student t-test for data normally distributed or the nonparametric Mann–Whitney test when the two groups were compared. When more than two groups were compared, one-way analysis of variance with Tukey's post hoc test or the non-parametric Kruskal-Wallis test followed by Dunn's post hoc test was employed. The Spearman correlation coefficient was calculated to determine correlations between two continuous variables. Associations were considered significant if the p value was <0.05 and the q value (ie, the false discovery rate using the Benjamini-Hochberg correction method) was <0.25.
Results
Study populations
We performed gut microbiota analysis on patients with FMF without (n=34, including 29 in FMF remission) or with (n=7) AAA, 19 patients with AAA secondary to another hereditary disease (n=4; ie, pyrin-associated autoinflammation with neutrophilic dermatosis, Fabry disease, tumour necrosis receptor-associated periodic syndrome and other genetic aetiologies), to inflammatory rheumatic diseases (n=4; ie, synovitis, acne, pustulosis, hyperostosis and osteitis syndrome, rhupus syndrome, psoriatic rheumatism and rheumatoid arthritis), to obesity (n=3), to an unknown reason (n=4) or to miscellaneous causes (n=4; ie, Crohn's disease, Waldenstrom's macroglobulinemia with MyD88 mutation, HIV infection and Schnitzler syndrome) and 26 healthy controls (table 1).
Clinical, demographic and renal function data for the study population
Regarding the patients with FMF without AAA, 31 (91%) carried homozygous or compound heterozygous mutations in the MEFV gene. Conversely, all patients with FMF with AAA were homozygous (M694V/M694V, n=6, or M680I/M680I, n=1). The treatments that each group of patients underwent are described in online supplementary tables S1 and S2.
Biological data, including acute phase proteins (CRP, SAA and S100A8/A9), proinflammatory cytokines (IL-1β, IL-6 and TNF-α), adipokines (leptin and adiponectin) and IDO activity, are depicted in figure 1. All patient groups had higher blood levels of IDO, S100A8/A9, IL-1β, IL-6 and TNF-α than did the controls. The patients with FMF complicated by AAA had higher adiponectin levels and IDO activity than did the patients with FMF without AAA (p<0.05). No correlation was found between the glomerular filtration rate estimated by the modification of diet in renal disease (MDRD) formula and IDO activity or adiponectin in patients with FMF with AAA (Spearman correlation test, p=0.33 and 0.32, respectively). When pooling data, IDO activity positively correlated with creatininemia, IL-6, TNF-α, adiponectin and age. In addition, adiponectin was correlated with SAA and CRP (online supplementary figure S1).
Supplemental material
Biological data among the different studied groups. Blood dosages of CRP (A), SAA (B), S100A8/A9 proteins (C), IL-1β (D), IL-6 (E), TNF-α (F), leptin (G), adiponectin (H) and IDO activity (I) are shown as medians and quartiles 1–3 among patients with FMF without (FMF–AAA) or with (FMF+AAA) AAA, patients with AAA of another cause (AAA–FMF) and healthy controls. Differences between each group were evaluated by Student's t-test for data normally distributed or the nonparametric Mann-Whitney test. The association was considered significant if the p value was less than 0.05 and if the Q value (ie, the false discovery rate using the Benjamini-Hochberg correction) was less than 0.25. *p <0.05, **p <0.01, ***p <0.001. AAA, AA amyloidosis; CRP, C reactive protein; FMF, familial Mediterranean fever; IL, interleukin; SAA, serum amyloid A; TNF, tumour necrosis factor.
Bacterial dysbiosis in FMF and AAA
Compared with healthy subjects, significant decreases in α-diversity were observed in patients with FMF globally (online supplementary figure S2) and specifically in those without amyloidosis and in patients with non-FMF-related AAA, as assessed by the Shannon and Chao1 indexes (p<0.05, figure 2A). In addition, β-diversity analysis showed significant differences between the healthy controls and these two groups, suggesting a significantly different microbiota composition (p<0.05 and<0.001, respectively; figure 2B). Although no significant difference was observed regarding patients with FMF with AAA, the statistical power was limited by the small number of subjects in this group.
Supplemental material
Altered bacterial microbiota composition and diversity in patients with FMF and AAA. (A) Microbial richness and evenness were calculated based on the Shannon index, microbial richness was calculated based on the Chao1 index, and statistical P significance was evaluated using the nonparametric Mann-Whitney test. (B) β-diversity. Principal coordinate analysis of the Bray-Curtis distance with each sample coloured according to the study population. PC1, PC2 and PC3 represent the three principal coordinates that captured most of the diversity, with the fraction of the diversity captured given as a percentage. differences between the study groups were evaluated by the nonparametric analysis of similarity test (analysis of similarities, 9999 permutations). The global composition of the bacterial microbiota at the phylum (C) and family (D) levels was expressed as the relative abundance for each group. Differences in bacterial taxon abundance between patients with FMF without AAA and healthy controls (E) and between patients with FMF with or without AAA (F) were calculated using the multivariate association with linear models statistical method. The fold change was calculated by dividing the mean abundance in each studied group. Only taxa present in more than 50% of the samples were used. The association was considered significant if the p value was less than 0.05 and if the Q value (ie, the false discovery rate using the Benjamini-Hochberg correction) was less than 0.25. *p<0.05, **p<0.01, ***p<0.001. AAA, AA amyloidosis; FMF, familial Mediterranean fever.
As expected, most bacteria identified belonged to the phyla Firmicutes, Bacteroidetes, Proteobacteria and Actinobacteria and to the families Lachnospiraceae and Ruminococcaceae (figure 2C,D).36 ,37
Differential bacterial composition in patients with FMF with and without AAA
After multivariate association testing with the MaAsLin statistical method to control for the effects of potential confounding factors, several OTUs belonging to the order Clostridiales were specifically associated with FMF (figure 2E). Nonetheless, no bacterial taxon was independently associated with AAA, likely because the different types of underlying disease that led to AAA have specific impacts on the gut microbiota, thus increasing heterogeneity among patients with amyloidosis. When restricting the analysis to patients with FMF only, two OTUs belonging to Clostridiales were specifically associated with AAA (figure 2F).
Discussion
In this study, we compared the microbiota and several biological markers of patients with FMF with or without AAA and patients with non-FMF-related AAA and healthy controls. FMF was associated with gut microbiota dysbiosis characterised by a decrease in α-diversity and a significantly altered composition. Moreover, among patients with FMF, AAA was associated with some alteration in the gut microbiota with two increased OTUs, as well as higher adiponectin levels and IDO activity.
Environmental factors have long been recognised as influencing the disease phenotype of FMF.17 20 38 Moreover, small intestinal bacterial overgrowth (SIBO) has been associated with unresponsiveness to colchicine, and decontamination therapy by rifaximin leads to a decrease in FMF attacks.39 Apart from abdominal pain during FMF attacks, none of our patients with FMF had clinical features suggestive of SIBO. To our knowledge, the only previous study of the gut microbiota in patients with FMF reported increases in Enterobacteriaceae, Acidaminococcaceae, Ruminococcus and Megasphaera and a decrease in Roseburia in 12 patients with FMF in remission compared with healthy controls.40 We did not confirm these results, which may be due to the use of different statistical methods, such as the lack of multivariate analysis or adjustment for multiple comparisons in the study by Khachatryan et al, to fewer patients recruited (with 13/19 being homozygous and only from Armenia) or to other environmental or genetic factors. Among the OTUs we found, the genera Blautia and Coprococcus are short-chain fatty acid-producing bacteria that are decreased in Crohn's disease.41 Interestingly, an increase in systemic concentrations of short-chain and long-chain fatty acids, produced by the gut microbiota, has also been observed in FMF, suggesting a leaky gut.42 43 As already described in other diseases,44 we may assume that a leaky gut in FMF, with the entry of environmental components into the systemic circulation, may exacerbate inflammation or may even trigger amyloid nucleation.
The gut microbiota has never been studied in AAA. In our study, two OTUs belonging to Clostridiales, including one from the genus Blautia, were over-represented in patients with AAA secondary to FMF compared with patients with FMF without amyloidosis. Interestingly, although its exact role has not been explored, Blautia has also been associated with Alzheimer's disease.45 In contrast, no bacterial taxon was associated with non-FMF-related AAA, a result that might be explained by several reasons.
The great heterogeneity in patients with AAA in terms of underlying inflammatory disease and of ongoing treatments, and these treatments might have modified their microbiota. However, colchicine used in patients with FMF does not appear to influence the gut microbiota, at least in in vitro experiments.46 Moreover, dysbiosis has been shown to be disease specific, such as in the different phenotypes of rheumatic inflammatory diseases or inflammatory bowel diseases.47 48 Therefore, it is possible that the occurrence of AAA is promoted by the disappearance or the emergence of certain bacterial taxa, differing for each underlying inflammatory disease but with common features. It would be very interesting to compare the gut microbiota between patients suffering from a specific disease (other than FMF) with or without AAA.
Transient involvement of the gut microbiota in AAA pathogenesis. It is conceivable that the transient presence of bacteria producing amyloid-enhancing factors is sufficient, within the context of an inflammatory disease, to generate nucleation and to induce a vicious circle. Indeed, SAA protein synthesis is dependent on proinflammatory cytokines; in turn, SAA can induce production of IL-1β, IL-6, TNF-α and IL-17A in several cell types.49
Analysis of the gut microbiota from stool samples.
A difference in bacterial functions and not in bacterial concentrations. Meta-omics studies would be a very interesting approach to address this possibility.
The preponderance of other environmental and/or genetic factors.
Previous studies on cytokines in FMF found increases in IL-6, IL-12, IL-17, IL-18, soluble IL-2 receptor, interferon-γ and TNF-α during an attack, as well as ongoing subclinical inflammation during remission.50–52 We confirmed that IL-6 and TNF-α, but also IL-1β levels, were significantly higher in patients with FMF than in controls. To our knowledge, no cytokine studies have been performed on patients with AAA. Although CRP and SAA protein levels were similar between patients with FMF without AAA and patients with non-FMF-related AAA, patients with AAA exhibited higher IL-6 and TNF-α levels. This is consistent with the fact that AAA is secondary to chronic inflammation and explains the efficacy of anticytokine therapies for this disease.2
A significant increase in IDO activity was observed in patients with AAA. This might be secondary to the kidney failure associated with AAA, as IDO activity is increased in kidney failure.53 This may also be explained by the fact that IDO activity is dependent on proinflammatory cytokines. Moreover, the tryptophan shortage induced by increased IDO activity enhances the synthesis of proinflammatory cytokines by macrophages, possibly inducing a vicious circle in AAA.10 12
Similarly, adiponectin levels were higher in patients with AAA than in patients with FMF without AAA, despite the fact that levels of TNF-α and IL-6, which inhibit adiponectin synthesis, are increased in these patients.6 These differences might be explained in part by kidney involvement.6 There are reports of adiponectin exerting proinflammatory activities in various tissues and contexts, and it thus may also be involved in AAA pathogenesis.54
The high levels of adiponectin and IDO activity in patients with AAA may be evaluated as diagnostic biomarkers to identify those at risk of AAA among patients with an inflammatory disease. Furthermore, in addition to proteinuria and creatininemia, they might be useful as intermediate markers in therapeutic follow-up for AAA. Finally, if the involvement of IDO in AAA is confirmed, the use of IDO inhibitors, currently in clinical trials in oncology, may be implemented.55
Conclusion
This study emphasises the association between FMF and gut microbiota dysbiosis characterised by a decrease in α-diversity and a significant alteration in composition. Although these results are descriptive, they suggest that the gut microbiota might be involved in the clinical expression of FMF. In patients with FMF, amyloidosis was independently associated with a specific alteration in the microbiota composition, suggesting that the gut microbiota may play a role in AAA pathogenesis. These data need to be further consolidated in mechanistic and interventional studies.
Supplemental material
Supplemental material
References
Footnotes
HS and SG-L are joint senior authors.
Handling editor Josef S Smolen
Presented at This work was previously presented at the 77th Congress of the French National Society of Internal Medicine 2018.
Collaborators The AA Amyloidosis Study Group includes Serge Amselem (Service de Génétique Médicale, Hôpital Trousseau, Assistance Publique des Hôpitaux de Paris, Paris, France), Camille Louvrier (Service de Génétique Médicale, Hôpital Trousseau, Assistance Publique des Hôpitaux de Paris, Paris, France), Léa Savey (Service de Médecine Interne, Hôpital Tenon, Assistance Publique des Hôpitaux de Paris, Paris, France), Joris Galland (Service de Néphrologie, Hôpital Tenon, Assistance Publique des Hôpitaux de Paris, Paris, France), Nicolas Martin Silva (Service de Médecine Interne, CHU de Caen, Caen, France), Antoine Hankard (Service de Médecine Interne, CHU de Caen, Caen, France), Alexandre Cez (Service de Néphrologie et Dialyse, Hôpital Tenon, Assistance Publique des Hôpitaux de Paris, Paris, France), Pierre-Antoine Michel (Service de Néphrologie et Dialyse, Hôpital Tenon, Assistance Publique des Hôpitaux de Paris, Paris, France), David Saadoun (Service de Médecine Interne et Immunologie Clinique, Groupe Hospitalier Pitié Salpétrière, Assistance Publique des Hôpitaux de Paris, Paris, France), Bertrand Knebelmann (Service de Néphrologie et de Transplantation, Hôpital Necker, Assistance Publique des Hôpitaux de Paris, Paris, France), Alexandre Hertig (Service de Néphrologie, Hôpital Tenon, Assistance Publique des Hôpitaux de Paris, Paris, France), Corinne Isnard Bagnis (Service de Néphrologie, Groupe Hospitalier Pitié Salpétrière, Assistance Publique des Hôpitaux de Paris, Paris, France), Tristan Legris (Service de Néphrologie, Hôpital de la Conception, Assistance Publique des Hôpitaux de Marseille, Marseille, France) and Xavier Belenfant (Service de Néphrologie, Centre Hospitalier Intercommunal André Grégoire, Montreuil, France).
Contributors SD, SG-L, HS, AA and GG designed the study, initiated this work and wrote the report. SD, SG-L and HS performed all statistical analyses. SD, SF, J-PB, J-ML, JC, SJ, MS, TF, SG-L and HS contributed to the sample preparation and carried out the experiments. SG-L, HS, AA, J-JB, AK and GG provided and cared for the study patients. All authors made substantial contributions to the acquisition of data, revised the article critically and gave the final approval of the manuscript to be submitted.
Funding This work was financed by a grant from Groupe Pasteur Mutualité and from the French Amyloidosis Association.
Competing interests None.
Patient consent for publication Not required.
Ethics approval The study was approved by the ethics committee (Comité de Protection des Personnes – Ile de France VI, n°DC-2015–2586) and was conducted in compliance with good clinical practices and the Declaration of Helsinki principles.
Provenance and peer review Not commissioned; externally peer reviewed.
Data availability statement All data relevant to the study are included in the article or uploaded as supplementary information.