Clinical Kidney Journal, 2021, vol. 14, no. 7, 1826–1834 doi: 10.1093/ckj/sfaa265 Advance Access Publication Date: 14 December 2020 Original Article Clinical Kidney Journal Downloaded from https://academic.oup.com/ckj/article/14/7/1826/6033662 by Fundacio Biblioteca Josep Laporte user on 22 February 2023 ORIGINAL ARTICLE Multidimensional inflammatory and immunological endotypes of idiopathic focal segmental glomerulosclerosis and their association with treatment outcomes Neus Roca1, Alvaro Madrid2, Mercedes Lopez3, Gloria Fraga3,4, Elias Jatem5,6, Jorge Gonzalez5,6, Cristina Martinez5 and Alfons Segarra5,6 1 Servicio Nefrologia Pediátrica, Hospital Universitari de Vic, Universitat de Vic, Barcelona, Spain, 2Servicio de Nefrologı́a Pediátrica, Hospital de Sant Joan de Déu de Barcelona, Barcelona, Spain, 3Servicio de Nefrologı́a Pediátrica, Hospital Vall d’Hebrón, Barcelona, Spain, 4Servicio de Nefrologı́a Pediátrica, Hospital de la Santa Creu i Sant Pau, Barcelona, Spain, 5Institut de Recerca Biomedica August Pi Sunyer, Lleida, Barcelona, Spain and 6Servicio de Nefrologia, Hospital Universitario Arnau de Vilanova, Lleida, Spain Correspondence to: Alfons Segarra; E-mail: alsegarr@gmail.com ABSTRACT Objectives. Idiopathic focal segmental glomerulosclerosis (FSGS) has been linked to immunological and inflammatory response dysregulations. The aim of this study was to find endotypes of FSGS patients using a cluster (CL) analysis based on inflammatory and immunological variables, and to analyse whether a certain endotype is associated with response to treatment with corticosteroids. Methods. This prospective observational study included patients with idiopathic FSGS diagnosed by kidney biopsy. Serum levels of soluble interleukin (IL)-1 receptor, tumoural necrosis factor alpha, Interferon gamma (IFNc), IL-6, IL-17, IL-12, IL-23, IL-13, IL-4, IL-5, IL-6, haemopexin (Hx), haptoglobin (Hgl), soluble urokinase-type plasminogen activator receptor (suPAR) and urinary CD80 (uCD80) were measured with enzyme-linked immunosorbent assay or nephelometry. T-helper lymphocyte populations and T-regulatory lymphocytes were analysed by flow cytometry. A factorial analysis followed by a k-means CL analysis was performed. Results. A total of 79 FSGS patients were included. Three CLs were identified. CL1 (27.8%) included IL-12, IL-17, IL-23 and a T helper 17 (Th17) pattern. CL2 (20.2%) included IL-4, IL-5, IL-13, immunoglobulin E and Th2 pattern. CL3 (51.8%) included IL-6, Hx, Hgl, suPAR and uCD80. There were no differences in age, gender, kidney function, albumin or proteinuria among CLs. About 42/79 patients (53.1%) showed cortico-resistance. The prevalence of cortico-resistance was significantly lower in CL2 (4/16, 25%) than in CL1 (16/26, 72.7%) and CL3 (22/41, 53.7%) (P ¼ 0.018), with no significant differences between CLs 1 and 3 (P ¼ 0.14). Received: 30.7.2020; Editorial decision: 07.12.2020 C The Author(s) 2020. Published by Oxford University Press on behalf of ERA-EDTA. V This is an Open Access article distributed under the terms of the Creative Commons Attribution Non-Commercial License (http://creativecommons.org/ licenses/by-nc/4.0/), which permits non-commercial re-use, distribution, and reproduction in any medium, provided the original work is properly cited. For commercial re-use, please contact journals.permissions@oup.com 1826 Multidimensional inflamatory and immunological nephrotic syndrome’s endotypes | 1827 Conclusions. Patients with FSGS and indistinguishable clinical presentation at diagnosis were classified in three distinct CLs according to predominant Th17, Th2 and acute inflammatory responses that display differences in clinical response to treatment with corticosteroids. Keywords: cluster analysis, endotypes, focal segmental glomerulosclerosis, idiopathic nephrotic syndrome, inflammatory response, interleukins, lymphocyte populations Downloaded from https://academic.oup.com/ckj/article/14/7/1826/6033662 by Fundacio Biblioteca Josep Laporte user on 22 February 2023 GRAPHICAL ABSTRACT INTRODUCTION still unknown and has been linked to the presence of a circulat- ing soluble factor capable of inducing irreversible podocyte in- The term idiopathic focal segmental glomerulosclerosis (FSGS) jury [4, 5]. Since some patients respond to treatment with refers to a clinicopathological entity that has diverse aetiologies corticosteroids and/or immunosuppressants [1, 2], it has been and pathogenic mechanisms, and a common histopathological suggested that, in certain cases, the pathogenesis of FSGS may expression characterized by the presence of sclerosis/hyalinosis be related to the activation of the inflammatory and/or the im- lesions in a variable percentage of glomeruli and affecting only mune response [6]. Studies analysing the inflammatory and im- one part of them [1]. This morphological definition allows a het- munological response in patients with nephrotic are scarce, erogeneous group of patients to be included under this same mainly focused on minimal change disease (MCD) and have led term. The response to corticosteroids has been identified as the main long-term outcome variable [2], but it has been found that to discordant results [7–11]. In some of them, an increase in the many patients do not respond well to the treatment with corti- levels of interleukin (IL)-6, IL-1 and tumour necrosis factor a costeroids or immunosuppressant drugs based on current clas- (TNFa) [4, 9, 11] has been described, but in others, the levels of sification, supporting the notion that FSGS is a multifactorial these molecules were not higher than those of healthy controls disorder with diverse and overlapping pathologies, sharing an [7, 8, 10]. Some recent studies found a relationship between identical phenotype at the time of diagnosis. In contrast to the FSGS and certain cytokines of inflammatory origin, like soluble phenotypic classifications, endotyping categorizes disease var- urokinase-type plasminogen activator receptor (suPAR) [12, 13] iants based on their underlying pathophysiologic mechanisms or cardiotrophine-like cytokine-1 [14], and others described a [3]. Therefore, identification of FSGS endotypes, determined by potential role for TNFa [15] and transforming growth factor-b distinct pathophysiological mechanisms, and further character- (TGF-b) [16]. The role of lymphocytes in the pathogenesis of idi- izing the FSGS uncontrolled by current treatment regimens may opathic nephrotic syndrome (INS) has been the focus of multi- provide a basis to understand disease causality and to develop ple investigations mainly focused on MCD [4, 5, 17], which have efficient management approaches. The pathogenesis of FSGS is found different patterns of T helper (Th) response. Some studies 1828 | N. Roca et al. have not shown differences between Th1/Th2 ratio [18], but suPAR were measured by ELISA (Quantikine R&D Systems, Inc., others have found a predominance of a Th2 pattern [19, 20] and MN, USA). The level of TNFa and interferon-c (IFNc), IL-17, IL-12 R had even provided evidence that IL-13, related to a Th2 re- and IL-23, were measured using the MILLIPLEXV MAP system sponse, is capable of inducing podocyte injury in experimental (catalogue number HCYTOMAG-60K Millipore Corp., MO, USA). models [19]. In all cases, these relationships have been incon- All ELISA analyses were performed in duplicate. In addition, to clusive. Instead of these ‘pathogenic-driven’ approaches, no analyse the reproducibility of the measures, in a sample of 15 studies have used unstructured approaches that incorporate patients, three or more determinations were made during the multi-dimensional data to identify clusters (CLs) of patients nephrotic outbreak phase, prior to the start any treatment. In all Downloaded from https://academic.oup.com/ckj/article/14/7/1826/6033662 by Fundacio Biblioteca Josep Laporte user on 22 February 2023 with presumably similar mechanisms of disease. cases, the variation coefficients of repeated measures were The aim of this study was to find endotypes of patients with <10%. FSGS by integrating multidimensional inflammatory and The Th subtypes and T-regulatory (Treg) cells were analysed immunological variables, using a CL analysis, and to analyse from peripheral blood samples using total heparinized blood the association between these endotypes and clinical response without performing any manipulation other than the incuba- to treatment with corticosteroids. tion with the different monoclonal antibodies. Total blood sam- ples were incubated at 37 C with the monoclonal antibodies conjugated with different fluorochromes, at the corresponding MATERIALS AND METHODS dilution and concentration and were analysed by flow Study design and population cytometry using the following monoclonal antibodies: anti-CD3, anti-CD4, anti-CXCR3, anti-CCR4, anti-CCR6, anti-CD25, anti- This prospective observational study, included patients with hCD127, anti-FoxP3 and anti-CD45RO (FACSCanto, Becton INS due to FSGS, diagnosed by renal biopsy between 2012 Dickinson, San Jose, CA, USA). Th subtypes and Treg cells were and 2019 in five tertiary referral hospitals. We included both defined according to the staining patterns summarized in paediatric and adult patients, who met all the following inclu- Supplementary Methods, Table S1. sion criteria: (i) diagnosis of FSGS performed by renal biopsy, containing more than eight glomeruli; (ii) absence of secondary aetiologies after conducting a systematic and protocolized Histopathological analysis of kidney biopsies study; (iii) absence of a family history of nephropathy; and All kidney biopsies were blindly analysed by two experienced (iv) no treatment with steroids, immunosuppressants, angio- nephropathologists who confirmed the morphological pattern tensin II blockers or statins at the time blood and urine samples of FSGS case by case. Kidney biopsies were stained with haema- were obtained. Patients with collapsing glomerulopathy were toxylin and eosin, periodic acid–Schiff-methenamine and excluded. Masson’s trichrome for morphological analysis and immunoflu- INS was defined by 24-h urinary protein excretion values orescence studies were carried out with antibodies against im- >3.5 g/day in adults and >40 mg/m2/h in children associated munoglobulin A (IgA), IgG, IgM, C3, fibrinogen and light chains, with hypalbuminaemia <3.5 g/dL and oedema. After diagnosis, and were processed for an electron microscope study. both the clinical follow-up and the treatment criteria were car- ried out following the recommendations indicated in the KDIGO Statistical analysis 2012 guidelines. The criteria for complete and partial remission were also defined according to those described in these Quantitative variables are expressed as mean 6 standard devia- guidelines, both for paediatric and adult patients [13]. tion. For dichotomous variables, a Chi-squared test was per- Idiopathic FSGS was diagnosed with the evidence of typical formed to determine the difference between groups. Data lesions in optical microscopy associated with diffuse podocyte distribution was tested for normality using a Kolmogorov– effacement in electron microscopy and after exclusion of Smirnov test or a Shapiro–Wilk test. Analysis of variance was secondary aetiologies including: reduction of renal mass, used to assess significant intergroup variability among more morbid obesity, HIV-associated nephropathy, heroin or cocaine than two groups and unpaired two-tailed Student’s t-test was use, parvovirus B19 infection, consumption of analgesics, used for between-group comparisons. Significance was ac- bisphosphonates or interferon, vesicoureteral reflux or obstruc- cepted at P < 0.05, which was adjusted using a Bonferroni cor- tive sleep apnoea. A genetic study to rule out pathogenic genetic rection for multiple comparisons. We analysed 23 different mutations was performed in all patients who showed cortico- variables related to inflammatory response, serum IL levels and resistance and were <35 years old, using a kidney panel Th cell polarization. Principal component analysis was per- previously described and validated [21]. Patients carrying formed to analyse relationships among variables to choose the known pathogenic mutations were excluded from the study. variables most relevant to the study. We retained significant All serum samples were obtained during the nephrotic components with eigenvalue 1. For CL analysis, continuous phase before starting any treatment. Blood samples were variables were standardized using z scores. Orthogonal rotation obtained in standard tubes with no additives, were centrifuged with Kaiser normalization was performed, and only variables at 1500g  10 min and stored at 80 C until used. Serum creati- with loadings of >0.4 were retained. Next, the subjects were nine was measured by the IDMS-traceable compensated sorted into groups by using a k-means method based on the cor- method (Hitachi Modular P-800 Roche Diagnostics, Germany). relation ratio and mixed principal component analysis (PCA). The estimated glomerular filtration rate (eGFR) was calculated The optimal number of CLs was determined by NbClust in R using the Schwartz equation in children and Chronic Kidney software package. Imputation of the missing values was per- Disease Epidemiology Collaboration (CKD-EPI) formula in formed with the k-Nearest Neighbour method and iterative ro- adults. Hemopexin (Hx) and haptoglobin (Hgl) were measured bust regression using the VIM package. Patients and variables by nephelometry (Coulter Biotek, Berlin, Germany). Urinary with >10% of missing data were removed. After classifying the CD80 (uCD80) was measured by ELISA [Human CD80 ELISA Kit patients according to the response to corticosteroid treatment, (B7-1) (ab256392), Cambridge, UK]. The serum levels of IL-6 and a univariate analysis was performed to analyse the differences Multidimensional inflamatory and immunological nephrotic syndrome’s endotypes | 1829 between the two groups. The cut-off values of quantitative vari- Table 1. Clinical and biochemical characteristics of FSGS patients ables associated with the response to treatment were calculated by receiver operating characteristic (ROC) curves, using Variables FSGS patients n ¼ 79 Youden’s index to select the optimal value. The relationship be- Age, median (IQR), years 35 (25–49) tween the response to corticosteroids and the type of CL was Gender: male, n (%) 37 (46.8) analysed by Chi-squared test. Steroid resistance, n (%) 42 (53.2) This study adhered to the parameters established by the Creatinine, mg/dL 1.1 (0.86–1.42) declaration of Helsinki. All patients gave their informed consent eGFR, mL/min/1.73 m2 88 (74.4–104.4) Downloaded from https://academic.oup.com/ckj/article/14/7/1826/6033662 by Fundacio Biblioteca Josep Laporte user on 22 February 2023 in writing and the bioethics committee of the corresponding Proteinuria, g/dL 10.8 (8.2–14.4) centre approved the study. Albumin, g/dL 2.48 (2–2.78) Th1, % 23 (22–25) TNFa, pg/mL 8.8 (5.4–14.8) RESULTS IFNc, pg/mL 14.4 (7.7–24.4) IL-12, pg/mL 664.5 (478–774) Patient selection Th2, % 34 (31–39) Ninety-eight patients were screened; however, 19 patients IL-4, pg/mL 3.6 (2.1–6.1) (19.4%) were excluded because of the following reasons: nine IL-5, pg/mL 4.5 (2.2–6) patients (9.2%) received steroids, immunosuppressants, angio- IL-13, pg/mL 4.1 (2.4–6.4) IgE, g/dL 49 (46–98) tensin II receptor antagonists or statins, three patients (3%) did Th17, % 3.3 (2.4–24.5) not give their written consent, three (3%) patients were lost to Th1–Th17, % 6.5 (3.9–9.8) follow-up within the first 3 months after diagnosis and four Treg, % 2.1 (1.4–3) patients (4%) carried known pathogenic mutations in genes IL-17, pg/mL 7.2 (3–31) encoding proteins of glomerular filtration barrier (NPHS2: n ¼ 1; IL-23, pg/mL 35.6 (25.6–51.6) collagen IVa3: n ¼ 2 and LMB1X: n ¼ 1). The final study group in- sIL1R, pg/mL 2257 (1383.9–3085.4) cluded 79 patients. Of these patients, 17 (21.5%) were paediatric IL-6, pg/mL 3.9 (1.8–12.7) and 62 (71.5 %) were >18 years. Seventy patients showed the Hx, mg/mL 110.9 (49.1–324.9) classic variant (NOS), four patients showed a tip lesion and five Hgl, mg/dL 137.8 (90.8–338.8) patients showed a perihiliar variant (Figure 1). suPAR, ng/mL 3421.8 (2409.6–4830) uCD80, ng/mg creat 36 (16.4–56) Clinical and biochemical characteristics Values are expressed as median [interquartile range (IQR)] 25–75th quartile un- The main clinical and biochemical characteristics of the study less otherwise specified. group are summarized in Table 1. The correlation matrix among biochemical variables is shown in Supplementary those five components. Table 3 summarizes the variables in- Results, Table S2. cluded in each component in the rotated component matrix. Variables with scores <0.4 are not represented. Component 1 Principal component analysis comprised the percentage of Th2 lymphocytes and the levels of IL-4, IL-13 and IgE. Component 2 included the percentage of We first performed a PCA to reduce dimensionality of variables Th17 lymphocytes, the circulating levels of IL-17, IL-23 and IL-12 from a total of 23. Five components were retained, explaining and the percentage of regulatory T lymphocytes. Component 3 73.1% of all variance in the data. Table 2 summarizes the results included levels of TNFa, IL-6, Hx, soluble IL-1 receptor (sIL1R) of the PCA and the percentage of variability retained by each of and Hgl. Component 4 included IFNc and Component 5 included suPAR. Figure 2 shows the 3D scatter-plot graphic representa- Screened eligible patients: tion of the distribution of all these variables for the first three 98 patients with INS due to FSGS PCA that retained 59.1% of total variability. CL analysis 19 patients not enrolled: • 9 received treatment previously Figure 3 represents the graphic classification dendogram of • 3 refused consent patients in three differentiated CLs. Table 4 shows the distribu- • 4 detected mutations in tion of the biochemical variables analysed, in each and one of podocyte proteins these three CLs. The results of the analyses of variance to com- • 3 lost to follow-up pare the distribution of biochemical variables among CLs are summarized in Supplementary Results, Table S3. There were no 79 FSGS patients recruited: differences in age, gender, kidney function, albumin or protein- • 70 classic variant (NOS) uria among the three CLs. There were no significant differences • 4 tip lesion between adult and paediatric patients (Supplementary Results, • 5 perihiliar variant Table S1). CL1 included 22 patients (27.8%) and was character- ized by a significantly higher percentage of both Th17 lympho- cytes and circulating levels of IL-17, IL-12 and IL-23, as well as a percentage of regulatory T lymphocytes significantly lower than 17 pediatric 62 adults those of CL2 and CL3. CL2 included 16 patients (20.3%) and was characterized by a percentage of Th2 lymphocytes and levels of FIGURE 1: Flow chart of patient selection. IL-4, IL-5, IL-13 and IgE significantly higher than those of CL1 1830 | N. Roca et al. Table 2. Results of PCA and percentage of variability retained by each component for the five components with eigenvalues >1 Initial eigenvalues Extraction sums of squared loadings Rotation sums of squared loadings % of Cumulative Cumulative Component Total % of variance Cumulative % Total variance % Total % of variance % 1 6.152 30.762 30.762 6.152 30.762 30.762 4.746 23.729 23.7 2 3.792 18.962 49.724 3.792 18.962 49.724 3.616 18.080 41.8 3 2.185 10.923 60.647 2.185 10.923 60.647 3.473 17.364 59.1 Downloaded from https://academic.oup.com/ckj/article/14/7/1826/6033662 by Fundacio Biblioteca Josep Laporte user on 22 February 2023 4 1.357 6.785 67.432 1.357 6.785 67.432 1.438 7.190 66.3 5 1.139 5.694 73.126 1.139 5.694 73.126 1.353 6.763 73.1 Table 3. Rotated component matrix showing the variables retained in each component. Component 1.0 IL17 IL23 Variables 1 2 3 4 5 IL12 TH17 0.5 TH1 TH17 Th2 0.923 – – – – Component 2 IL-4 0.908 – – – – IL-13 0.852 – – – – sIL1R IFNγ IgE IgE 0.749 – – – – 0 TH2 TNFα suPAR Th17 0.639 0.785 – – – IL6 IL5 IL-23 – 0.774 – – – Hemopexin CD80 IL4 IL13 IL-12 – 0.753 – – – TH1 –0.5 Tregs – 0.670 – – – Haptoglobin Tregs IL-17 – 0.651 – – – Th1 – 0.532 – – – –1.0 IL-6 – – 0.818 – – –1.0 –1.0 –0.5 Hx – – 0.800 – – –0.5 0 0 0.5 0.5 nt 2 sIL1R – – 0.765 – – Compone 1.0 1.0 pone nt 2 Com TNFa – – 0.748 – – Hgl – – 0.720 – – FIGURE 2: 3D scatter-plot graphic representation of the distribution of variables IFNc – – – 0.809 according to the first three PCA that retained the 59.1% of total variability. suPAR – – – – 0.867 Variables with scores <0.4 are not represented. 25 Dendrogram using average linkage (between groups) Rescaled distance cluster combine 20 15 10 5 0 71 72 69 53 48 49 66 68 58 45 70 32 78 25 29 35 46 62 74 51 59 79 60 63 33 43 44 64 41 52 54 39 67 50 61 47 56 73 76 77 40 2 3 1 19 20 18 16 21 17 22 8 13 15 5 9 12 4 6 10 14 7 11 57 75 65 42 23 24 36 37 55 26 27 30 31 28 34 38 80 79 78 77 76 75 74 73 72 71 70 69 68 67 66 65 64 63 62 61 60 59 58 57 56 55 54 53 52 51 50 49 48 47 46 45 44 43 42 41 40 39 38 37 36 35 34 33 32 31 30 29 28 27 26 25 24 23 22 21 20 19 18 17 16 15 14 13 12 11 10 9 8 7 6 5 4 3 2 1 0 Y FIGURE 3: Classification dendogram of patients showing the existence of three differentiated CLs. Multidimensional inflamatory and immunological nephrotic syndrome’s endotypes | 1831 Table 4. Characteristics of FSGS patients according to CL differentiation Variables CL1, n ¼ 22 CL2, n ¼ 16 CL3, n ¼ 41 P1 P2 P3 Clinical characteristics Age, years 34.5 (28–47.5) 39 (32.5–58) 34.5 (16–46.5) 0.755 1 0.404 Cortico-resistance, n (%) 16 (72.7) 4 (25) 22 (53.7) 0.018 0.140 0.018 Biochemical characteristics Creatinine, mg/dL 1.1 (0.75–1.3) 1 (0.8–1.4) 1.1 (0.9–1.4) 1 1 1 eGFR, mL/min/1.73 m2 Downloaded from https://academic.oup.com/ckj/article/14/7/1826/6033662 by Fundacio Biblioteca Josep Laporte user on 22 February 2023 84.9 (74.1–97.3) 90 (81.8–113.8) 87.9 (73.6–104.4) 1 0.818 1 Proteinuria, g/dL 10.1 (7.8–13.2) 13.3 (9.1–18) 9.3 (8.2–14) 0.207 1 0.155 Albumin, g/dL 2.44 (2.18–2.6) 2.15 (1.5–3.3) 2.49 (2.2–2.8) 1 1 1 Th1, % 23 (20–24) 22 (22–26) 23.5 (22–25.7) 1 0.015 0.063 TNFa, pg/mL 5.85 (4.9–7.5) 5.4 (2.9–7.3) 14.1 (10.8–18.7) 1 0.000 0.000 IFNc, pg/mL 12 (4.9–26) 10.5 (4–16.9) 16.7 (10.3–24.4) 1 1 1 IL-12, pg/mL 760 (709–881.8) 580 (504.5–665) 552 (348–771) 0.000 0.000 0.403 Th2, % 32.5 (30–37.3) 61 (37–67.5) 34 (32–37) 0.000 1 0.000 IL-4, pg/mL 4.3 (2.7–4.9) 13 (4.6–14.8) 2.65 (2–3.9) 0.000 1 0.000 IL-5, pg/mL 3.5 (2–4.7) 6.9 (2.3–8.6) 4.6 (2.4–5.7) 0.000 1 0.000 IL-13, pg/mL 4.2 (3.7–5.6) 11 (4.4–15.3) 3.7 (1.3–5.2) 0.000 0.710 0.000 IgE, mg/mL 47 (45–48) 143 (48.5–160) 50.5 (45.2–98) 0.000 0.292 0.000 Th17, % 30.9 (25.1–34.7) 3 (2.1–3.8) 3 (2.1–3.7) 0.000 0.000 1 Th1–Th17, % 13.3 (7.2–17.6) 5.6 (2–8.6) 5 (3.8–8.7) 0.000 0.000 0.506 Treg, % 0.8 (0.6–1) 2.9 (2.1–3.2) 2.5 (1.9–3.7) 0.000 0.000 1 IL-17, pg/mL 47.1 (38.1–54.9) 4.7 (2.3–8) 4.7 (2.8–7.9) 0.000 0.000 0.737 IL23, pg/mL 57.3 (35.1–63.6) 37 (25.3–48.8) 32 (23–42) 0.000 0.000 1 sIL1R, pg/mL 2294.9 (1496.2–2916.6) 1857.8 (1162.1–2345) 2296.2 (1328.5–3483) 1 1 1 IL-6, pg/mL 3 (1.8–3.9) 2 (1.2–5.6) 12.4 (3.9–19.6) 1 0.002 0.011 Hx, mg/mL 48.4 (28.4–72.7) 49 (40.1–116.5) 320.6 (136.9–440.4) 0.374 0.000 0.001 Hgl, mg/dL 69 (33.6–127.3) 116.8 (61.8–213.1) 321.4 (147.05–425.3) 0.543 0.000 0.000 suPAR, ng/mL 3413.5 (2497.8–4765.9) 3421 (2703.3–4830) 3444 (2373–4672.5) 1 1 0.647 uCD80, ng/mg creat 32.3 (16.1–45.2) 32.7 (6–56.5) 37.2 (23–78.5) 1 0.029 0.008 Values are expressed as median (interquartile range) 25–75th quartile unless otherwise specified. The bold values are the p values with stadistical significance (p<0.05). Table 5. Patient characteristics according to the response to corticosteroid treatment Variables Cortico-sensitive, n ¼ 37 Cortico-resistant, n ¼ 42 P-value Age, years 34 (26–38) 41 (23.3–54) 0.34 Gender: male, n (%) 21 (56.8) 16 (43.2) 0.09 Creatinine, mg/dL 1 (0.9–1.2) 1.2 (0.8–1.5) 0.12 eGFR, mL/min/1.73 m2 90 (81.8–107.6) 85.9 (73.4–99.4) 0.13 Proteinuria, g/dL 12.1 (8.2–14.5) 9.9 (8.4–12.8) 0.34 Albumin, g/dL 2.56 (2–2.8) 2.41 (2–2.6) 0,85 Th1, % 24 (22–26.5) 22.5 (21–24.3) 0.13 TNFa, pg/mL 8.7 (4.5–14.7) 9 (5.7–15.1) 0.62 IFNc, pg/mL 14.4 (9.5–26.2) 12.8 (5–24) 0.28 IL-12, pg/mL 618 (483–716.5) 726 (466–834) 0.046 Th2, % 36 (30.5–59) 34 (31.8–37.3) 0.043 IL-4, pg/mL 3.6 (2.4–11.6) 3.6 (1.9–4.6) 0.007 IL-13, pg/mL 4.8 (2.4–9.5) 3.9 (2–5.8) 0.12 IL-5, pg/mL 4.2 (2.2–7.3) 4.5 (2–5.9) 0.13 IgE, mg/mL 51 (46.5–143) 47 (45–77.3) 0.16 Th17, % 3.1 (2.1–6.3) 4.3 (2.4–25.9) 0.18 Th1–Th17, % 6.5 (3.5–9.2) 6.5 (4.3–14.7) 0.07 Treg, % 2.7 (1.8–3.1) 1.8 (0.8–2.6) 0.06 IL-17, pg/mL 4.8 (2.3–9) 9.8 (4.4–38.5) 0.023 IL-23, pg/mL 33 (23–46) 43 (29.8–56.9) 0.043 sIL1R, pg/mL 2133.1 (1141.1–2916.6) 2257.1 (1615.1–3185.7) 0.63 IL-6, pg/mL 2.7 (1.2–8.5) 10.7 (3–17) 0.83 Hx, mg/mL 110.9 (46.3–152.9) 114.5 (49–364.7) 0.18 Hgl, mg/dL 116.8 (70.8–321.4) 139.3 (99.5–371) 0.37 suPAR, ng/mL 3421.8 (2407.5–4915.9) 3206.7 (2457.1–4124.6) 0.34 uCD80, ng/mg creat 34 (15.5–55.5) 36.5 (17.6–81.8) 0.17 Values are expressed as median (interquartile range) 25–75th quartile unless otherwise specified. The bold values are the p values with stadistical significance (p<0.05). 1832 | N. Roca et al. 100 80 Downloaded from https://academic.oup.com/ckj/article/14/7/1826/6033662 by Fundacio Biblioteca Josep Laporte user on 22 February 2023 Sensitivity 60 40 IL12 IL17 IL23 20 IL4 TH2 TH17 0 0 20 40 60 80 100 100-specificity Variable AUC SE 95% CI Cut-off value Sensitivity (95% CI) Specificity (95% CI) IL12 pg/mL 0.638 0.0641 0.522 to 0.744 > 698 58.5 (42.1–73.7) 75.6 (58.8–88.2) IL23 pg/mL 0.658 0.0619 0.542 to 0.762 > 49 40.5 (25.6–56.7) 89.1 (74.6–97) IL4 pg/mL 0.619 0.0654 0.502 to 0.726 ≤ 4.9 83.3 (68.6–93) 45.9 (29.5–63.1) IL17 pg/mL 0.681 0.0617 0.566 to 0.782 > 9.2 57.1 (41.0–72.3) 81 (64.8–92) TH2 (%) 0.562 0.0690 0.445 to 0.674 ≤ 38 83.3 (68.6–93) 43.2 (27.1–60.5) TH17 (%) 0.612 0.0644 0.495 to 0.720 > 9.35 40.5 (25.6–56.7) 83.8 (68.0–93.8) FIGURE 4: ROC curves to analyse the capacity of variables that showed significant differences in the univariate analysis to predict cortico-resistance. and CL3. CL3 included 41 patients (51.9%) and was characterized high scores for Component 1, corresponding to a Th2 pattern, by the absence of polarization of Th lymphocyte subpopulations showed a significantly lower frequency of cortico-resistance. and by levels of TNFa, IL-6, Hx, Hgl, suPAR and CD80 signifi- cantly higher than those of CL1 and CL2. DISCUSSION Variables associated with the response to treatment The results of our study indicate that, at the time of diagnosis and in the absence of pharmacological interferences, the in- with corticosteroids flammatory response and the polarization of Th cell subtypes Overall, 42/79 patients (53.1%) showed corticosteroid resistance. do not follow a single pattern common to all patients with idio- Table 5 summarizes the baseline characteristics of patients, pathic FSGS. From the 23 variables initially analysed, our classified according to the response to corticosteroid treatment. patients could be classified into three CLs with well- There were no differences in age, gender, eGFR or albumin be- differentiated signatures. CL1 can be considered as a predomi- tween cortico-sensitive and cortico-resistant patients. Patients nant Th17 cell response, with lower percentage of Treg cells and with cortico-resistance showed higher levels of IL-17, Th17 and higher levels of IL-17 and IL-23. CL2 is characterized by a pre- Th1–Th17 cells and lower levels of Th2 cells and IL-4, IL-5 and dominant Th2 response pattern, with increased levels of IL-4, IL-13 than patients with cortico-sensitivity. Figure 4 shows the IL-5, IL-13 and IgE, and CL3 comprised patients with no results of ROC curves performed to analyse the capacity of each polarization of T-cell response and with increased levels of of these variables to predict cortico-resistance, including the proteins and mediators of the acute inflammatory response. area under the curve and their associated Youden’s values. The When comparing these three CLs, there were significant differ- prevalence of cortico-resistance was significantly lower in CL2 ences in both inflammatory and immunological profiles among (4/16, 25%) than in CL1 (16/22, 72.7%) and CL3 (22/41, 53.7%) (P ¼ them, but no differences in the clinical profile of the disease at 0.018). Figure S1 in the Supplementary Results shows a 3D the time of diagnosis in terms of age, gender distribution, serum graphic representation of the distribution of patients according albumin, eGFR or proteinuria. to their respective factor scores for the first three PCA, differen- The role of lymphocytes in the pathogenesis of INS has been tiating cortico-sensitive and cortico-resistant cases. In this analysed in several studies [4, 5, 17, 22, 23]. Most of them have graphic representation, it can also be seen that patients with focused on MCD and have found different types of polarization Multidimensional inflamatory and immunological nephrotic syndrome’s endotypes | 1833 of the Th response. Some studies have not shown differences in The main strength of our study is the inclusion of a large ho- Th1/Th2 ratio [18], whereas others have found a predominance mogeneous sample of patients, studied at the time of diagnosis of a Th2 pattern [19, 20] and had even provided evidence that without therapeutic interferences, using a multidimensional IL-13, related to a Th2 response, is capable of inducing podocyte marker analysis of the inflammatory and immunological injury in experimental models [19]. In our cohort, 20% of response. When interpreting our results, it must be taken into patients (CL2) showed such a predominant Th2 response. This account that we measured the levels of cytokines and Th cell type of response was not the most prevalent, but showed a fre- subtypes in the bloodstream, and this is not necessarily repre- quency of cortico-resistance significantly lower than that ob- sentative of what occurs at the level of the lymphoid organs. Downloaded from https://academic.oup.com/ckj/article/14/7/1826/6033662 by Fundacio Biblioteca Josep Laporte user on 22 February 2023 served in CL1 and CL3. These results are in agreement with The differentiation of Th0 ‘naı̈ve’ lymphocytes into mature Th previously published data in which an association between a lymphocytes, takes place at a local level in the lymphoid organs, predominant Th2 pattern and the response to corticosteroids is driven by signals given by the antigen-presenting cells [17, 22, also described [18], and might be linked to the inhibitory effect 32]. IL-12 leads to differentiation in Th1 cells, which produce of corticosteroids in IL-4, IL-5 and IL-13 gene transcription [24]. IL-2 and IFNc, while IL-4, IL-13 and IL-2 lead to Th2 cells, which Recent independent studies highlight the possible patho- produce IL-4, IL-13 and IL-5, and the signal induced by IL-6, TGF-b genic role of an imbalance between Th17 cells and Tregs in the and IL-21 induces differentiation to Th17 cells, which produce pathogenesis of FSGS [17, 25, 26]. Th17 cells induce autoimmune IL-17, IL-22, IL-6 and TNFa and exert a powerful pro-inflammatory and inflammatory injuries with the production of cytokines effect [25, 26]. In our study, ILs like IL-17, IL-13 and IL-4, aggregated such as IL-17, IL-22 and TNFa. Tregs have an anti-inflammatory into the same CL together with their corresponding synthetizing role and suppress the activation of the immune system by re- cells. In other cases, however, the IL and cytokine levels measured leasing IL-10 and TGF-b1 [4]. Th17 cells and their related cyto- in the bloodstream did not concur with that is known to occur kines have been found increased and Treg cells decreased in locally in the lymphoid organs. This explains the absence of asso- serum of INS patients with relapse in comparison with healthy ciation between IL-12 and IFNc with Th1 cells and the absence of controls [25, 26]. IL-17 has been found to be increased in kidney association between Th17 cells and IL-6. biopsies of FSGS patients in comparison with other histological In conclusion, our data indicate that the patterns of the in- patterns and it has also been positively related with glomerulo- flammatory response and Th cell polarization found in patients sclerosis index. It has been hypothesized that this cytokine may with FSGS at the time of diagnosis can be classified into three promote renal injury-causing alterations in nephrin phosphory- different CLs. The most frequent pattern is consistent with an lation and apoptosis of podocytes [25]. In our cohort, 30% of activated inflammatory response with no polarization of Th cell patients (CL1) showed this Th17-type response pattern. The third CL (CL3) observed in our is the one that included subtypes. The second pattern in order of frequency is consistent the higher number of patients and was characterized by with a predominant Th17-driven response. Finally, the least increased levels of IL-6, TNFa, Hx, Hgl and suPAR that, taken to- prevalent pattern has all the characteristics of a Th2-driven re- gether, indicate an activation of the acute-phase inflammatory sponse. Patients with a predominant Th2 response showed the response that was not associated to any defined type of polari- best response rate to corticosteroids, while Th17 or IL-1/IL-6 zation of Th cell subpopulations. IL-1, IL-6 and TNFa are the responses were associated with a higher prevalence of cortico- main cytokines released by monocytes in response to infections resistance. Together, these data provide useful information for or tissue injuries. These cytokines cause a systemic effect by future studies in order to analyse the pathogenic significance of the stimulation of the acute-phase response in the liver [27]. each of the identified response patterns. Altogether, the two CLs that were associated with signs of increased activation of the inflammatory response, whether SUPPLEMENTARY DATA mediated by Th17 cells or by the IL-1/IL-6 pathway (CL1 and CL3), were associated with a similar prevalence of cortico- Supplementary data are available at ckj online. resistance that, in both cases, was significantly higher than that observed in CL2. These data are in agreement with those - CONFLICT OF INTEREST STATEMENT described in previous studies in which increased serum of IL-6, Hx and Hgl, have been associated with steroid resistance [28, None declared. 29], and with the data from a recent study by our group, showing that increased circulating levels of IL-6, Hx and REFERENCES Hgl were associated with steroid-resistance in patients with MCD [30]. 1. Eddy AA, Symons JM. Nephrotic syndrome in childhood. Our data provide a comprehensive view on the pleomorphic Lancet 2003; 362: 629–639 immune and/or inflammatory response that can be found at 2. Rosenberg AZ, Kopp JB. Focal segmental glomeruloesclero- the time of diagnosis in patients with FSGS but they do not sis. Clin J Am Soc Nephrol 2017; 12: 502–517 allow a direct pathogenic role to be attributed to any of the 3. Shalhoub RJ. Pathogenesis of lipoid nephrosis: a disorder of observed response patterns. Even though there is experimental T-cell function. Lancet 1974; 304: 556–560 evidence indicating that both IL-17 and IL-13 can induce injury 4. Araya C, Diaz L, Wasserfall C et al. T regulatory cell function to podocytes [19, 31], there is no evidence enough to support in idiopathic minimal lesion nephrotic syndrome. Pediatr that these cytokines are directly involved in podocyte injury in Nephrol 2009; 24: 1691–1698 human disease. Furthermore, in clinical conditions associated 5. McCarthy ET, Sharma M, Savin VJ. Circulating permeability fac- with increased IL-17 levels, such as seronegative spondyloar- tors in idiopathic nephrotic sı́ndrome and focal segmental glo- thropathies, a higher incidence of FSGS has not been described, meruloesclerosis. Clin J Am Soc Nephrol 2010; 5: 2115–2121 nor has it been described in patients with chronic rhinosinusi- 6. Shao XS, Yang XQ, and Zhao XD et al. The prevalence of Th17 tis, asthma or nasal polyposis, in which IL-13 levels are cells and FOXP3 regulate T cells (Treg) in children with primary increased. nephrotic syndrome. Pediatr Nephrol 2009; 24: 1683–1690 1834 | N. Roca et al. 7. Russell C, Baillie JK. Treatable traits and therapeutic targets: human focal segmental glomerulosclerosis serum on glo- goals for systems biology in infectious disease. Curr Opin Syst merular filtration barrier. Transl Res 2015; 166: 384–398 Biol 2017; 2: 140–145 21. Goumenos DS, Tsakas S, Meguid S et al. Transforming 8. Camici M. The nephrotic syndrome is an immunoinflamma- growth factor-beta(1) in the kidney and urine of patients tory disorder. Med Hypotheses 2007; 68: 900–905 with glomerular disease and proteinuria. Nephrol Dial 9. Shimoyama H, Nakajima M, Naka H et al. Up-regulation of Transplant 2002; 17: 2145–2152 interleukin-2 mRNA in children with idiopathic nephrotic 22. Kaneko K, Tuchiya K, Fujinaga S et al. Th1/Th2 balance in syndrome. Pediatr Nephrol 2004; 19: 1115–1121 childhood idiopathic nephrotic syndrome. Clin Nephrol 2002; 10. Bustos C, Gonzalez E, Muley R et al. Increase of tumour 58: 393–397 Downloaded from https://academic.oup.com/ckj/article/14/7/1826/6033662 by Fundacio Biblioteca Josep Laporte user on 22 February 2023 necrosis factor a synthesis and gene expression in periph- 23. Colucci M, Corpetti G, Emma F et al. Immunology of idio- eral blood mononuclear cells of children with idiopathic pathic nephrotic sı́ndrome. Pediatr Nephrol 2018; 33: 573–584 nephrotic syndrome. Eur J Clin Invest 1994; 24: 799–805 24. Yildiz B, Cetin N, Kural N et al. CD19þCD23þ cells, 11. Kanai T, Shiraishi H, Yamagata T et al. Th2 cells predominate CD4þCD25þ T cells, E-selectin and interleukin-12 levels in in idiopathic steroid-sensitive nephrotic syndrome. Clin Exp children with steroid sensitive nephrotic syndrome. Ital J Nephrol 2010; 14: 578–583 Pediatr 2013; 39: 42 12. Suranyi MG, Guasch A, Hall BM et al. Elevated levels of tumor 25. Kapojos JJ, Poelstra K, Borghuis T et al. Regulation of plasma necrosis factor-a in the nephrotic syndrome in humans. Am hemopexin activity by stimulated endothelial or mesangial J Kidney Dis 1993; 21: 251–259 cells. Nephron Physiol 2004; 96: 1–10 13. Cho MH, Lee HS, Choe BH et al. Interleukin-8 and tumor ne- 26. Cheung PK, Klok PA, Baller JFW et al. Induction of experimen- crosis factor-alpha are increased in minimal change disease tal proteinuria in vivo following infusion of human plasma but do not alter albumin permeability. Am J Nephrol 2003; 23: hemopexin. Kidney Int 2000; 57: 1512–1520 260–266 27. Yang J, Zhang BL. Value of determination of haptoglobin and 14. Rizk MK, El-Nawawy A, Abdel-Kareem E et al. Serum inter- alpha1-antitripsina in predicting response to glucocorticoid leukins and urinary microglobulin in children with idio- therapy in children with primary nephrotic syndrome. pathic nephrotic syndrome. East Mediterr Health J 2005; 11: Zhong Dang Dai Zhi 2015; 17: 227–231 993–1002 28. Cara-Fuentes G, Johnson RJ, Reiser J et al. CD80 and suPAR in 15. Wang L, Li Q, Wang L et al. The role of Th17/IL-17 in the path- patients with minimal change disease and focal segmental ogenesis of primary nephrotic syndrome in children. Kidney glomerulosclerosis: diagnostic and pathogenic significance: Blood Press Res 2013; 37: 332–345 response. Pediatr Nephrol 2014; 29: 1467–1468 16. Meijers B, Poesen R, Claes K et al. Soluble urokinase receptor 29. Lyngbæk S, Sehestedt T, Marott JL et al. CRP and suPAR are is a biomarker of cardiovascular disease in chronic kidney differently related to anthropometry and subclinical organ disease. Kidney Int 2015; 87: 210–216 damage. Int J Cardiol 2013; 167: 781–785 17. Peng Z, Mao J, Chen X et al. Serum suPAR levels help differen- 30. Roca N, Martinez C, Jatem E et al. Activation of the acute in- tiate steroid resistance from steroid-sensitive nephrotic flammatory phase response in idiopathic nephrotic syndrome: syndrome in children. Pediatr Nephrol 2015; 30: 301–307 association with clinicopathological phenotypes ad with re- 18. Lai KW, Wei CL, Tan LK et al. Overexpression of interleukin- sponse to corticosteroids. Clin Kidney J 2002; 39: 1143–1152 13 induces minimal-change-like nephropathy in rats. J Am 31. El Hussiny M, Mohamed M, Barakat, AL et al. A. Effect of IL-6 Soc Nephrol 2007; 18: 1476–1485 C-174G polymorphism on response to steroid therapy in 19. Eknoyan G, Lameire N. KDIGO clinical practice guideline for Egyptian children with nephrotic syndrome. Eur J Pharm Med glomerulonephritis. Kidney Int Suppl 2012; 2: 141 Res 2018; 5: 146–152 20. Sharma M, Zhou J, Gauchat JF et al. Janus kinase 2/signal 32. Braun CM, Huang S. Corticosteroid modulation of human, transducer and activator of transcription 3 inhibitors attenu- antigen-specific Thl and Th2 responses. J Allergy Clin ate the effect of cardiotrophin-like cytokine factor 1 and Immunol 1997; 100: 400–407