High Rate of Mutations of Adhesion Molecules and Extracellular Matrix Glycoproteins in Patients with Adult-Onset Focal and Segmental Glomerulosclerosis

(1) Background: Focal and segmental glomerulosclerosis (FSGS) is a pattern of injury that results from podocyte loss in the setting of a wide variety of injurious mechanisms. These include both acquired and genetic as well as primary and secondary causes, or a combination thereof, without optimal therapy, and a high rate of patients develop end-stage renal disease (ESRD). Genetic studies have helped improve the global understanding of FSGS syndrome; thus, we hypothesize that patients with primary FSGS may have underlying alterations in adhesion molecules or extracellular matrix glycoproteins related to previously unreported mutations that may be studied through next-generation sequencing (NGS). (2) Methods: We developed an NGS panel with 29 genes related to adhesion and extracellular matrix glycoproteins. DNA was extracted from twenty-three FSGS patients diagnosed by renal biopsy; (3) Results: The average number of accumulated variants in FSGS patients was high. We describe the missense variant ITGB3c.1199G>A, which is considered pathogenic; in addition, we discovered the nonsense variant CDH1c.499G>T, which lacks a Reference SNP (rs) Report and is considered likely pathogenic. (4) Conclusions: To the best of our knowledge, this is the first account of a high rate of change in extracellular matrix glycoproteins and adhesion molecules in individuals with adult-onset FSGS. The combined effect of all these variations may result in a genotype that is vulnerable to the pathogenesis of glomerulopathy.


Introduction
Focal and segmental glomerulosclerosis (FSGS) is one of the most common causes of adult-onset nephrotic syndrome alongside membranous nephropathy, and it is the primary glomerular disease that leads to end-stage renal disease (ESRD) in the United States [1,2]. Despite the numerous research studies conducted in recent years, an ideal therapy is still difficult to find, and a high percentage of patients eventually progress to ESRD because of years of unsatisfactory care and severe accompanying morbidity. In fact, 10 years following the diagnosis of primary FSGS, 45% of individuals develop ESRD [3]. A total of 16 samples of normal renal tissue received at the Pathology Department of the Marqués de Valdecilla University Hospital in Santander were selected as the control group; these corresponded to renal wedges (from organ donors) and radical nephrectomies (mainly due to renal cancer) of patients with preserved renal function. Subsequently, after a thorough review of the criteria, the control group comprised 15 samples, as one sample was not selected due to a diagnosis of IgA GN at another hospital.
In addition, CNVs were compared with publicly available control databases, such as the Genome Aggregation Database (gnomAD), the Database of Genomic Variants (DGV), and databases for pathogenic CNVs, such as ClinVar.
Based on Online Mendelian Inheritance in Man (OMIM; https://www.omim.org, accessed on 4 March 2022), we distinguished between nephropathic and phenocopy genes. As a result, "nephropathic genes" were defined as genes that were recognized in OMIM as causing "nephropathy". In contrast, "phenocopy genes" were those genes in OMIM found to produce a syndromic or unrelated illness. (Supplementary Table S1). Where a case was not yet documented in OMIM, recent research pointing to the gene's causal pathogenic role in the nephrotic syndrome was considered to stratify cases as nephropathic vs. phenocopy [18,19].

Sample Preparation and Mutational Analysis
The renal biopsy was performed by percutaneous strategy under ultrasound guidance. Renal tissues were processed, formalin fixed, and paraffin embedded according to standard procedures for histological diagnosis.
Genomic DNA was extracted from formalin-fixed paraffin-embedded (FFPE) material. From the FFPE material, 5 sections of 3 µm were taken and collected in a 1.5 mL Eppendorf tube. First, tissue deparaffinization was carried out by adding 160 µL of Deparaffinization Solution (QIAGEN, Hilden, Germany) and incubating at 56 • C for 5 min at 700 rpm. Next, DNA extraction was carried out using Cobas ® DNA Sample Preparation Kit (Roche Diagnostics, Basel, Switzerland) according to the manufacturer's protocol (Supplementary dataset).
Once DNA was extracted from the samples, nucleic acid quantification was performed using the Qubit 1X dsDNA HS Assay Kit (ThermoFisher Scientific, Waltham, MA, USA); according to the manufacturer, samples at a concentration of at least 2 µg/uL were considered valid for sequencing. Finally, all the samples were diluted up to 1 ng/µL to begin the preparation of libraries.

Design of the Targeted Sequencing Panel
We designed a custom panel for the targeted sequencing of DNA extracted from FFPE material using AmpliSeq technology (ThermoFisher Scientific), which is the standard for FFPE sequencing.
Primer design was performed for the FFPE tissue to generate 175 bp length amplicons, covering 92.65 kb over 29 genes with a total 910 amplicons.

Library Preparation
Library preparation was performed automatically using the Ion AmpliSeq kit for Ion Chef DL8 (ThermoFisher Scientific), with a 31-amplification-cycle program of 4 min. Once amplified, libraries were quantified and diluted to equimolar concentrations.

Sequencing
Sequencing was performed using Ion S5 (ThermoFisher Scientific) with a 520 chip, which generated a minimum of 5 million useful reads to determine germinal coverage to at least ×100.

Sequencing Data Collection
Only cases with a coverage greater than or equal to 160 reads were selected for analysis. In those patients where the library had a depth of less than 160 reads, the entire sequencing process was repeated.

Clinical and Pathological Characteristics
A total of 23 patients were included in the study (clinical and histopathological characteristics are summarized in Table 1). Of these patients, 74% were male, and 26% were female. The median age was 54 years (range 18-82 years); mean age: 54 years), and all 23 patients were adults ≥18 years old.
In almost all cases (74%; n = 17), classical or NOS (not otherwise specified) FSGS was the predominant morphological variant observed on biopsy. A perihilar variant was identified in four patients (17%; n = 4), while cellular and tip variants were observed in one patient each.

Mutational Analysis
Of the 29 genes studied, variants were found in all of them, except in the RHOA and LPAR5 genes, although many of them were silent mutations (Supplementary Table S2). UTR regions as well as non-coding regions, which do not affect splicing, were not considered. All the raw sequencing data can be accessed from the European Nucleotide Archive (ENA) using this link: https://www.ebi.ac.uk/ena/data/search?query=PRJEB53199 (accessed on 31 May 2022).
Initially, we examined all non-silent mutations, specifically noting the presence of alterations in the FN1 and TNC genes in all FSGS patients. The frequencies of variations in the different genes analyzed are shown in Figure 1. Variations were observed in both the control group and in all patients, although the average number of cumulative variants (Single Nucleotide Polymorphisms, SNPs) was higher in FSGS patients (16.5 SNPs) compared with the control group (mean of 11.8 SNPs).
We then shifted our focus to non-benign variants; specifically, those classified as pathogenic, probably pathogenic, and uncertain significant variants ( Table 2). Two patients were found to have the pathogenic missense variant ITGB3c.1199G>A, and five FSGS patients were discovered to have the likely pathogenic nonsense variant Variations were observed in both the control group and in all patients, although the average number of cumulative variants (Single Nucleotide Polymorphisms, SNPs) was higher in FSGS patients (16.5 SNPs) compared with the control group (mean of 11.8 SNPs).
We then shifted our focus to non-benign variants; specifically, those classified as pathogenic, probably pathogenic, and uncertain significant variants ( Table 2). Two patients were found to have the pathogenic missense variant ITGB3c.1199G>A, and five FSGS patients were discovered to have the likely pathogenic nonsense variant CDH1c.499G>T, which has no Reference SNP (rs) Report. The missense variants COL1A2c.1015A>C, ITGB1c.1807A>T, ITGB3c.2351C>T, LAMC1c.148T>C, LPAR3c.524C>G, LPAR4c.259C>G, LPAR4c.260T>C, LPAR6c.227A>T, LPAR6c.998T>C, and WASF3c.934G>C were classified as uncertain significant variants. None of the gene variants listed above were detected in the control group, but the difference was not statistically significant. Five patients had the CDH1c.499G>T mutation (patients 1, 2, 5, 7, and 9): two women and three men, with a mean age of 52.6 years. Two of them also had the ITGB3c.1199G>A mutation (patients 2 and 7). The first patient was diagnosed with autoimmune hemolytic anemia and is currently being studied for interstitial lung disease (ILD) and receiving treatment with Rituximab. The other patient developed an advanced chronic disease with subsequent transplantation and recurrence of FSGS in the control biopsy. The rest of the patients with the CDH1c.499G>T mutation are currently under control without immunosuppressive treatment and have mild CKD.
The remaining variants were deemed benign, with many having similar frequencies in the overall European population (Supplementary Table S4). Additionally, we identified new variants that have not yet been annotated and therefore have no Reference SNP (rs) Report (Table 3). On the other hand, when we searched for a correlation between non-benign mutations and disease severity, as measured by the level of proteinuria, we observed a tendency toward a greater accumulation of mutations in the subnephrotic range of proteinuria ( Figure 2). Furthermore, to directly correlate these mutations with glomerulopathies, we considered the stages of CKD. According to the statistical analysis, the number of mutations is not dependent on the stage of chronic kidney disease (p = 0.1164). However, on average, stage 3a has 2.42 more mutations compared with stage 1. We created a graph showing the distribution of mutations by genes and CKD stages (Figure 3).
On the other hand, when we searched for a correlation between non-benign mutations and disease severity, as measured by the level of proteinuria, we observed a tendency toward a greater accumulation of mutations in the subnephrotic range of proteinuria ( Figure 2). Furthermore, to directly correlate these mutations with glomerulopathies, we considered the stages of CKD. According to the statistical analysis, the number of mutations is not dependent on the stage of chronic kidney disease (p = 0.1164). However, on average, stage 3a has 2.42 more mutations compared with stage 1. We created a graph showing the distribution of mutations by genes and CKD stages (Figure 3).

Discussion
In our cohort of patients with adult-onset FSGS, we observed for the first time a high rate of mutations in adhesion molecules and extracellular matrix glycoproteins, with the high prevalence of non-silent mutations the key result of our investigation, although the limited sample size does not allow for statistical significance. Both in the general population (European population, dbSNP, 1000 Genomes) and in our control group, this rate of alteration is lower than in the FSGS patient group (Supplementary Table S5). Similar to   teinuria ( Figure 2). Furthermore, to directly correlate these mutations with glomerulopathies, we considered the stages of CKD. According to the statistical analysis, the number of mutations is not dependent on the stage of chronic kidney disease (p = 0.1164). However, on average, stage 3a has 2.42 more mutations compared with stage 1. We created a graph showing the distribution of mutations by genes and CKD stages (Figure 3).

Discussion
In our cohort of patients with adult-onset FSGS, we observed for the first time a high rate of mutations in adhesion molecules and extracellular matrix glycoproteins, with the high prevalence of non-silent mutations the key result of our investigation, although the limited sample size does not allow for statistical significance. Both in the general population (European population, dbSNP, 1000 Genomes) and in our control group, this rate of alteration is lower than in the FSGS patient group (Supplementary Table S5). Similar to

Discussion
In our cohort of patients with adult-onset FSGS, we observed for the first time a high rate of mutations in adhesion molecules and extracellular matrix glycoproteins, with the high prevalence of non-silent mutations the key result of our investigation, although the limited sample size does not allow for statistical significance. Both in the general population (European population, dbSNP, 1000 Genomes) and in our control group, this rate of alteration is lower than in the FSGS patient group (Supplementary Table S5). Similar to this study, other genetic studies have previously reported mutations in adhesion molecules and extracellular matrix glycoproteins in patients with FSGS [23][24][25].
Most genetic research has established a direct correlation between mutations in podocyte and collagen COL4A (A3/A4/A5) genes and the development of familial FSGS. This correlation persists even in cases of adult-onset FSGS, where a monogenic cause of FSGS could be identified in up to 29% of cases [7][8][9]26,27]. A similar rate of mutation was reported in adult-onset steroid-resistant nephrotic syndrome, with a higher age of onset of FSGS related to a lower rate of single-gene identification [28].
Mutations in the INF2 gene have been reported in familial FSGS of both Caucasian and Asian ancestry [23,24], while Marx et al. identified a novel nonsense variant in the PODXL gene in a three-generation family with an atypical glomerular nephropathy resembling FSGS [25]. Although these studies found mutations in specific molecule groups, they did not investigate the entire set of genes that constitute these groups.
To perform a comprehensive analysis of these genes, we utilized NGS technology, which allowed us to rapidly study multiple genes in a single experiment at high resolution [16]. This approach facilitates the identification of new molecules or pathways involved in or contributing to the development of FSGS. However, interpreting new variants without a known clinical significance can be challenging, and additional functional tests are necessary to link these variants to FSGS [7,16,23].
It is noteworthy that in our study, all the FSGS patients showed alterations in tenascin and fibronectin 1, which have previously been linked to podocyte injury [29][30][31]. Additionally, a large percentage of our patients exhibited mutations in integrins (78%) and laminins (87%), which have been described in other studies [32,33] that we will discuss below.
While podocyte degradation is a critical step in the pathophysiology of FSGS, the environment may also contribute to the development of segmental lesions. Research has shown that alterations in adjacent epithelial parietal and mesangial cells also play a role in FSGS [34]. Extracellular matrix glycoproteins and adhesion molecules also contribute to podocyte health [11,29,35,36], and animal and cell culture studies have linked alterations in these molecules to nephrotic syndrome and podocyte damage [30][31][32][37][38][39][40]. Moreover, mutations in adhesion proteins such as laminin β2, integrin α3, and integrin β4 in humans have been associated with steroid-resistant nephrotic syndrome due to a lack of appropriate adhesion of podocytes to the glomerular basement membrane [13,[41][42][43]. Experimental and human FSGS exhibit altered expression of extracellular matrix proteins, including laminin-1, perlecan, collagen type IV-2, laminin-2, agrin, and collagen type IV-4, that are produced by parietal epithelial cells and podocytes [32]. Therefore, these variations in extracellular matrix proteins may possibly have an impact on the histopathologic type of FSGS [44].
In addition to other signals, adhesion molecules and extracellular matrix glycoproteins are components of a network that involves redundant interactions between molecules [29]. The heterogeneity of research and the difficulty in identifying a single common etiology of a disease that is frequently syndromic is explained by this complicated network [26,45].
It is important to note that, in our study, no mutation in a single gene was detected that could explain FSGS, but each patient had a variable number of non-silent mutations (9 to 22 non-silent mutations per patient) (Supplementary Table S3), which could partially explain the predisposition to the disease.
Multiple gene variants may interact to promote podocyte damage caused by different non-Mendelian forms of FSGS [9]. Therefore, each individual mutation affects only a few patients, and the addition of mutations that may promote these minor injuries contributes to the development of segmental lesions. In this sense, a greater number of mutations in adhesion molecules and extracellular matrix glycoproteins could favor the occurrence of FSGS by increasing susceptibility to the disease [9]. It is interesting that several authors have noted a synergistic effect of various FSGS mutations. Frese et al. demonstrated that carriers of type IV collagen (COL4A5) gene mutations with related polymorphisms in the slit diaphragm genes experience severe forms of FSGS [46], while Bullich et al. report a similar result with COL4A3 mutations [17].
Notably, we cannot be sure of the exact role of the variants we found, and we classified most of them as benign or "of uncertain significance" [21], understanding that further studies will be needed to understand the pathogenic role of adhesion molecule and extracellular matrix glycoprotein mutations in the development of FSGS.
Several limitations should be mentioned. Although our study sheds light on the prevalence of mutations in adhesion molecules and extracellular matrix glycoproteins in adult-onset FSGS patients, it is limited by a small sample size. Furthermore, the lack of a well-defined classification system for adult-onset FSGS and the absence of validated methods for distinguishing non-familial FSGS [22] make it difficult to draw definitive conclusions. While we excluded patients with secondary forms of FSGS and suspected familial forms, it is still possible that some patients had these forms. Additionally, the absence of electron microscopy analysis prevented us from differentiating between primary and adaptive forms of FSGS. Further research is needed to fully understand the complex pathogenesis of this heterogeneous disease.

Conclusions
We were able to identify a high rate of mutation in adhesion molecules and extracellular matrix glycoproteins in all of the selected adult patients with FSGS.
We found mutations without a Reference SNP (rs) Report. Two patients were found to have the pathogenic missense variant ITGB3c.1199G>A, and five FSGS patients were discovered to have the likely pathogenic nonsense variant CDH1c.499G>T. None of the gene variants were detected in the control group; however, this difference was not statistically significant.