Urinary Protein-Biomarkers Reliably Indicate Very Early Kidney Damage in Children With Alport Syndrome Independently of Albuminuria and Inflammation

Abstract

Introduction: Alport syndrome (AS) is a hereditary type IV collagen disease. It starts shortly after birth, without clinical symptoms, and progresses to end-stage kidney disease early in life. The earlier therapy starts, the more effectively end-stage kidney disease can be delayed. Clearly then, to ensure preemptive therapy, early diagnosis is an essential prerequisite.

Methods: To provide early diagnosis, we searched for protein biomarkers (BMs) by mass spectrometry in dogs with AS stage 0. At this very early stage, we identified 74 candidate BMs. Of these, using commercial enzyme-linked immunosorbent assays (ELISAs), we evaluated 27 in dogs and 28 in children, 50 with AS and 104 healthy controls.

Results: Most BMs from blood appeared as fractions of multiple variants of the same protein, as shown by their chromatographic distribution before mass spectrometry. Blood samples showed only minor differences because ELISAs rarely detect disease-specific variants. However, in urine , several proteins, individually or in combination, were promising indicators of very early and preclinical kidney injury. The BMs with the highest sensitivity and specificity were collagen type XIII, hyaluronan binding protein 2 (HABP2), and complement C4 binding protein (C4BP).

Conclusion: We generated very strong candidate BMs by our approach of first examining preclinical AS in dogs and then validating these BMs in children at early stages of disease. These BMs might serve for screening purposes for AS before the onset of kidney damage and therefore allow preemptive therapy.

Keywords: C4 binding protein; collagen type XIII; dogs; early screening; hyaluronan binding protein 2; proteomics.

Graphical abstract

Figure 1

General time courses of disease biomarkers. There are several hypothetical BM types that are differently targeted by our approach. Type A: BMs applicable preclinically; increase indicates alterations starting in a preclinical stage and persisting under progression. Type B: truly preclinical BMs; indicating only preclinical alterations. Type C: early BMs; increase indicates alterations simultaneous with the development of early clinical signs, often overlooked or neglected (e.g., in AS, hematuria). Type D: late BMs; increase indicates subsequent alterations in manifested disease. This type (e.g., in AS, (micro)-albuminuria, decline of eGFR) is already in clinical use and was not part of our search. Type A and B were searched for in juvenile AS-mice (4 weeks¹³) and AS-dogs (7 weeks). This search might also detect type C. Human samples usually come from different points of individual disease development. Their evaluation might therefore miss type B BMs. However, all BMs detected (A, B, or C) are likely to be more suitable for early diagnosis of kidney injury than those currently applied. AS, Alport syndrome; BM, biomarker.

Figure 2

Biomarker candidates identified by mass spectrometry. Criteria defining a biomarker: Biomarker selection in mice, see reference. BM candidates in dogs were by weighted mass spectrometric data (Supplementary Methods, Formula 1). Peptides supporting the proteins in a 2D-sub-fraction from AM were compared by Sieve with those of the homologous 2D-sub-fraction from NM as well as those from CF with those from NF. The quantifier “ratio” is related to the concentration quotient (AM/NM, CF/NF) of each specific protein in the sub-fractions compared. n.a., not altered; n.f., not found; 0, no cluster in this direction; x, overall trend higher in AM than in NM and in CF than in NF; o, increased parts/protein variants in AM and CF; DIV/0: no protein detected in healthy controls (no ratio can be given); mean ratio, mean of all ratios with synergistic alteration of the protein (AM/NM or CF/NF). Red: candidate BMs; blue: some proteoforms might serve as candidate BMs; green: candidate BMs in mice, criteria not fulfilled in dogs. Grey filling: positive APR; black filling: negative APR. Altogether, 124 protein chains were more abundant in affected than in unaffected dogs, either in males or females or both, and 74 proteins fulfilled the definition for a BM candidate. AM, affected male; APR, acute phase protein; BM, biomarker; CF, carrier female; NF, unaffected female; NM, unaffected male.

Figure 2

Biomarker candidates identified by mass spectrometry. Criteria defining a biomarker: Biomarker selection in mice, see reference. BM candidates in dogs were by weighted mass spectrometric data (Supplementary Methods, Formula 1). Peptides supporting the proteins in a 2D-sub-fraction from AM were compared by Sieve with those of the homologous 2D-sub-fraction from NM as well as those from CF with those from NF. The quantifier “ratio” is related to the concentration quotient (AM/NM, CF/NF) of each specific protein in the sub-fractions compared. n.a., not altered; n.f., not found; 0, no cluster in this direction; x, overall trend higher in AM than in NM and in CF than in NF; o, increased parts/protein variants in AM and CF; DIV/0: no protein detected in healthy controls (no ratio can be given); mean ratio, mean of all ratios with synergistic alteration of the protein (AM/NM or CF/NF). Red: candidate BMs; blue: some proteoforms might serve as candidate BMs; green: candidate BMs in mice, criteria not fulfilled in dogs. Grey filling: positive APR; black filling: negative APR. Altogether, 124 protein chains were more abundant in affected than in unaffected dogs, either in males or females or both, and 74 proteins fulfilled the definition for a BM candidate. AM, affected male; APR, acute phase protein; BM, biomarker; CF, carrier female; NF, unaffected female; NM, unaffected male.

Figure 3

Comparison of BM concentrations in sample set 1 from patients with AS and healthy controls. Second samples were taken about 4 years after the first. Significances: P < 0.05 (light blue bars); P < 0.01 (dark blue bars); all other not significant. c, d: Values below the detection limit (= 0) are arranged close to the abscissa for visualization. (a) Lumican (LUM) in serum. Due to the small number of samples, dots indicate tendency. (b) Complement factor I (CFI) in serum. (c) Complement factor H (CFH) in urine; (d) Hyaluronan binding protein 2 (HABP2) in urine.

Figure 4

Receiver operating characteristic curves of urinary BMs, comparison of patients with Alport syndrome, thin basement membrane nephropathy, and benign familial hematuria (sample set 2) with controls. Blue: Collagen XIII (ColXIII); green: hyaluronan binding protein 2 (HABP2); magenta: complement C4 binding protein (C4BP); red: reference line. BM, biomarker.