APOL1 risk genotype in European steroid-resistant nephrotic syndrome and/or focal segmental glomerulosclerosis patients of different African ancestries

Abstract

Background

Apolipoprotein L1 (APOL1) risk variants are strongly associated with sporadic focal segmental glomerulosclerosis (FSGS) in populations with African ancestry. We determined the frequency of G1/G2 variants in steroid-resistant nephrotic syndrome (SRNS)/FSGS patients with African or French West Indies ancestry in France and its relationships with other SRNS genes.

Methods

In a cohort of 152 patients (139 families), the APOL1 risk variants were genotyped by direct Sanger sequencing and pathogenic mutations were screened by next-generation sequencing with a panel including 35 SRNS genes.

Results

The two risk allele [high-risk (HR)] genotypes were found in 43.1% (66/152) of subjects compared with 18.9% (106/562) in a control population (P < 0.0001): 33 patients homozygous for APOL1 G1 alleles, 4 homozygous for G2 and 29 compound heterozygous for G1 and G2. Compared with patients in the low-risk (LR) group, patients in the HR group were more likely to originate from the French West Indies than from Africa [45/66 (68.2%) versus 30/86 (34.9%); P < 0.0001]. There were more familial cases in the HR group [27 (41.5%) versus 8 (11.4%); P < 0.0001]. However, causative mutations in monogenic SRNS genes were found in only 1 patient in the HR group compared with 16 patients (14 families) in the LR group (P = 0.0006). At diagnosis, patients in the HR group without other mutations were more often adults [35 (53.8%) versus 19 (27.1%); P = 0.003] and had a lower estimated glomerular filtration rate (78.9 versus 98.8 mL/min/1.73 m²; P = 0.02).

Conclusions

The HR genotype is frequent in FSGS patients with African ancestry in our cohort, especially in those originating from the West Indies, and confer a poor renal prognosis. It is usually not associated with other causative mutations in monogenic SRNS genes.

APOL1, duplication, focal segmental glomerulosclerosis, minimal change disease, steroid-resistant nephrotic syndrome

INTRODUCTION

Apolipoprotein L1 (APOL1)-associated nephropathies are a spectrum of related non-diabetic kidney diseases strongly associated with G1 and G2 coding risk variants in populations with African ancestry. It has been observed that African Americans have a 4-fold increased risk of end-stage renal disease (ESRD). Two studies identified a locus on chromosome 22 associated with non-diabetic ESRD, focal segmental glomerulosclerosis (FSGS) and human immunodeficiency virus (HIV)-associated nephropathy (which is manifested as collapsing glomerulopathy) and this was extended to hypertension-attributed ESRD [1, 2]. The major source of genetic risk for African American hypertension-attributed ESRD and FSGS was then localized to APOL1, encoding ApoL1 [3]. FSGS, HIV-associated nephropathy, hypertension-attributed ESRD, hypertension-attributed chronic renal failure and non-diabetic ESRD were associated with two APOL1 risk alleles: G1, comprising two missense variants (S342G and I384M), and G2, which bears a 6 base pair (bp) in-frame deletion leading to the deletion of two amino acids (N388_Y389del) [3–7]. G1 and G2 are mutually exclusive, never occurring on the same chromosome. In particular, these variants were strongly associated with sporadic FSGS in either the homozygous or compound heterozygous state with a recessive pattern of inheritance. One risk allele was associated with only a small increase in renal disease risk. Two risk alleles versus zero risk alleles yielded an odds ratio (OR) of 10–17 for FSGS [3, 4].

ApoL1 is the trypanolytic factor of human serum that confers resistance to Trypanosoma brucei brucei. In vitro studies revealed that only the kidney disease–associated ApoL1 variants lyse Trypanosoma brucei rhodesiense, one of the causes of human African sleeping sickness. The APOL1 gene may have undergone natural selective pressure in western Africa [3].

Previous studies have been performed in African Americans, but the prevalence of the APOL1 risk genotype in Europe is unknown. The aim of our study was to determine the frequency of APOL1 G1/G2 variants in a cohort of steroid-resistant nephrotic syndrome (SRNS) and/or FSGS patients with African or French West Indies ancestry in a French reference centre for inherited kidney diseases and to analyse associations with mutations in other known monogenic SRNS genes and with clinical outcome.

MATERIALS AND METHODS

Patients

We studied a French multicentre cohort of SRNS and/or FSGS patients with African or West Indies (Guadeloupe, Martinique, Haiti) and French Guiana ancestry. The eight patients originating from French Guiana were included in the West Indies group for further analysis. Patients’ blood samples were referred to our laboratory for a genetic analysis. We asked for recent follow-up to clinicians in charge of the patients, including final glucocorticoid and other immunosuppressive drugs sensitivity. Patients with HIV infection were excluded. Estimated glomerular filtration rate (eGFR) was calculated by the Schwartz formula in children and by the Modification of Diet in Renal Disease formula in adults. Glucocorticoid resistance was defined as a lack of complete response to 4 weeks of treatment with 60 mg/m² and methylprednisolone pulses in children and to 16 weeks of treatment with 1 mg/kg/day in adults.

APOL1 sequencing

Written informed consent was obtained from participants or their parents. DNA of patients was isolated from peripheral blood using standard procedures. Biological collection was approved by the Ethical Committee ‘Comité de Protection des Personnes Ile-De-France’.

Polymerase chain reaction (PCR) amplification at the end of exon 6 of APOL1 was performed in one PCR product. Sequencing was performed using a Big Dye terminator cycle sequencing kit and an ABI Prism 3500 XL DNA analyser (Thermo Fisher Scientific, Waltham, MA, USA). Direct Sanger sequencing allowed the detection of three single-nucleotide polymorphisms: the two missense variants p.Ser342Gly and p.Ile384Met (rs73885319 and rs60910145), called the G1 risk allele, and the 6 bp deletion p.del Asn388/Tyr389 (rs71785313), called G2.

Qualitative PCR assay for detection of duplication

Duplication at the APOL1 locus was searched in all patients [8]. APOL1 is localized between APOL2 and MYH9 and the APOL1 duplication (encompassing APOL1, APOL2 and the 11 first exons of MYH9) leads to an aberrant junction between APOL2 and MYH9. A PCR fragment was designed to detect the junction fragment for the APOL1 duplication. Forward and reverse primers were localized in the 3′-UTR of APOL2 (TCCTAGTGAAACCAATAAGCC) and in intron 11 of MYH9 (GTGAAAGTGCCTGACACG). A band of 288 bp indicated the presence of the APOL1 duplication in that sample.

Targeted exon sequencing

In 11 patients, a mutation in a gene involved in monogenic forms of SRNS was found by candidate gene Sanger sequencing. The other patients were tested using multiplex PCR and next-generation sequencing developed with Multiplicom. The panel included 35 genes for SRNS (Supplementary data, Table S1). For four patients, no more DNA was available. Patients who get complete remission with glucocorticoids or with other immunosuppressive treatment or who relapsed after renal transplantation were not tested with this panel. All pathogenic variants were verified by Sanger sequencing. All the variants identified were evaluated to determine their pathogenic character. We considered only those located in coding exons and the splice site region. We excluded the silent mutations, the splice variants that did not affect the splice site scores and all the variants present in the Exac database with a minor allele frequency >0.01. We screened for the missense variants with the three most commonly used bioinformatic predictors of the variants’ pathogenicity, namely PolyPhen-2 (http://genetics.bwh.harvard.edu/pph2/), sorting intolerant from tolerant (SIFT) (http://sift.jcvi.org/) and Mutation Taster (http://www.mutationtaster.org/) to detect highly deleterious mutations. We considered pathogenic nonsense, frameshift, essential splice, known mutations and missense variants with high prediction scores.

Control population

We used individual data available in the 1000 Genomes Project (http://www.internationalgenome.org/about/) to assess high-risk (HR) genotype frequency in the general population of different African origins.

Statistical analysis

Continuous values are reported as mean [standard deviation (SD)]. Dichotomous data are presented as percentages. Chi-squared or Fisher’s exact tests were applied for dichotomous and categorical data; the unpaired t-test was used to compare two continuous variables. Cumulative renal survival was analysed by Kaplan–Meier survival curves and log-rank tests. Two-tailed P-values  <0.05 were regarded as statistically significant. Statistical analyses were performed using InStat 3 and Prism 4 software (GraphPad Software, La Jolla, CA, USA).

RESULTS

Clinical data at diagnosis and during follow-up

We studied 152 SRNS and/or FSGS patients in 139 families with African ancestry. We identified 41 familial cases (26.8%) in 28 families.

Clinical data at diagnosis and during follow-up are shown in Table 1.

Table 1

Clinical data according to APOL1 risk allele status

Clinical data	Total, 152 patients (139 familiesa)	LR group, 86 patients (82 families)		HR group, 66 patients (59 families)		P-value
		Mutation, 16 patients (14 families)	No mutation, 70 patients (68 families)	Mutation, 1 patient (1 family)	No mutation, 65 patients (58 families)
French West Indies/African origin, n/N (%)	75/77 (49.3)	7/9 (43.7)	23/47 (32.8)	0/1	45/20 (69.2)	<0.0001b
Familial, n (%)	41 (26.8)	5 (31.2)	8 (11.4)	1	27 (41.5)	<0.0001b
Age at onset (years), mean (SD)	17.1 (14.0)	6.9 (9.2)	14.7 (14.1)	0	22.5 (12.8)	<0.0001b
						<0.0001c
						0.007d
Age at onset >18 years, n (%)	57 (37.5)	3 (18.7)	19 (27.1)	0	35 (53.8)	0.0026b
						0.01b
Proteinuria at onset (g/day), mean (SD)	7.4 (7.2)	7.4 (5.2)	7.6 (8.5)	2.5	7.6 (7.5)	NS
Albuminaemia at onset (g/L), mean (SD)	19.4 (10.4)	23.7 (11.9)	18.4 (13.7)	7	19.2 (7.8)	NS
eGFR at onset (mL/min/1.73 m2), mean (SD)	88.6 (37.9)	82.0 (31.6)	98.8 (32.7)	100	78.9 (43.2)	0.02b
HBP, n (%)	27 (17.8)	3 (18.7)	10 (14.3)	0	14 (21.5)	NS
Glucocorticoid treatment, n (%)	101 (66.4)	4 (25.0)	52 (74.3)	1	44 (67.7)	0.003c
						0.0004d
Steroid resistant, n (%)	95 (94.0)	4 (100)	46 (88.5)	1	44 (100)	0.03b
Follow-up (years), mean (SD)	6.2 (9.2)	5.3 (8.2)	7.5 (11.1)	0.4	5.2 (7.1)	NS
eGFR at last follow-upe (mL/min/1.73 m2), mean (SD)	97.5 (62.9)	103.2 (99.1)	107.9 (58.2)	100	85.1 (62.5)	0.02b
HBP at last follow-upf, n (%)	30 (19.7)	2 (12.5)	14 (20.0)	0	14 (21.5)	NS
ESRD, n (%)	45 (29.6)	7 (43.7)	17 (24.3)	0	21 (32.3)	NS
Age at ESRD (years), mean (SD)	22.2 (14.1)	13.4 (8.8)	17.9 (13.4)		27.9 (13.8)	0.03b
						0.01c
Renal transplantation, n (%)	26 (17.1)	4 (25.0)	11 (15.7)	0	11 (16.9)	NS
Age at renal transplantation (years), mean (SD)	24.7 (16.7)	11.5 (3.8)	21.8 (15.9)		32.2 (17.3)	0.03b
						0.01c
Relapse after transplantation, n (%)	5 (19.2)	0	3 (27.3)		2 (18.2)	NS

a

In one family, one patient was in the LR group and another in the HR group. In another family, one patient was in the LR group and three patients were in the HR group.

b

Patients in the LR group without mutation compared with patients in the HR group without mutation.

c

Patients in the LR group with mutation compared with patients in the HR group without mutation.

d

Patients in LR group with mutation compared with patients in the LR group without mutation.

e

If no ESRD.

f

Before ESRD.

HBP, high blood pressure.

Genetic data

APOL1 sequencing

Thirty-three patients were homozygous for APOL1 G1 alleles, 4 were homozygous for APOL1 G2 alleles and 29 were compound heterozygous for APOL1 G1 and G2 alleles. Fourteen patients were heterozygous for APOL1 G2 alleles and 15 heterozygous for APOL1 G1 alleles. The two risk allele (HR) genotype was considered as any combination of G1 and G2 and was found in 66 (43.1%) subjects. The HR genotype was found in 60% (45/75) of patients originating from the French West Indies and in only 27.3% (21/77) of patients originating from Africa. The other patients were considered at low risk (LR).

We first compared the frequency of the APOL1 HR genotype in our cohort with a control population originating from Africa using individual data available in the 1000 Genomes Project (http://www.internationalgenome.org/about/). The HR frequency was significantly higher in FSGS patients [43.1% (66/152)] than in controls [18.9% (106/562)] {P < 0.0001; odds ratio [OR] 3.3 [95% confidence interval (CI) 2.2–4.8]}. Then we compared the HR genotype frequency in patients of French West Indies ancestry with a control population from the West Indies (Barbados). The HR genotype frequency was significantly higher in FSGS patients [60% (45/75)] compared with controls [13.5% (13/96) [P < 0.0001; OR 9.6 (95% CI 4.5–20.1)]. Finally, we compared the HR genotype frequency between FSGS patients originating from Africa with controls originating from West Africa (Sierra Leone and Gambia). The APOL1 HR genotype frequency was significantly higher in FSGS patients [27.3% (21/77)] than in controls [15.1% (30/198)] [P = 0.02; OR 2.1 (95% CI 1.1–3.9)].

Duplication of the APOL1 gene

An APOL1 duplication including APOL1, APOL2 and a part of MYH9 has been observed at the APOL1 locus and may increase susceptibility to kidney disease [8]. The duplication is revealed by the presence of a junction fragment and the genotype for APOL1 is identified by Sanger sequencing. We searched for this APOL1 duplication in the whole cohort and identified it in three families. In one family with four affected individuals (Family 1) we detected a heterozygous duplication of the APOL1 gene in two patients who thus had three copies of APOL1 (G1/G2/G0), but not in the two other patients tested. A duplication was also found in one patient who was G0 homozygous (G0/G0/G0) and in one patient who was G1 homozygous (G1/G1/G1), without a more severe phenotype than other patients.

Mutations in known monogenic SRNS genes

To analyse the link between the APOL1 genotype and mutations in known monogenic SRNS genes, we screened the patients for causative mutations (Supplementary data, Table S1). Pathogenic mutations were found in 17 patients (11.2%) from 15 families (10.8%), almost exclusively belonging to the APOL1 LR group (Table 2). Only one individual in the APOL1 HR group presenting with congenital nephrotic syndrome also had a monogenic disorder, being compound heterozygous for two NPHS1 mutations, in addition to the APOL1 G1/G1 two risk alleles. In contrast, in the APOL1 LR group, 16 individuals in 14 families had disease-causing mutations. A total of 10 patients presented with autosomal recessive diseases, including 5 patients from 3 families already known as familial cases: 5 patients from 3 families bore mutations in NPHS1, 2 patients in NPHS2, 1 patient in LAMB2 and another one in PLCE1. Two COQ2 variants were present in one patient with one clearly pathogenic frameshift variant and one missense variant of unknown significance. The patient, born from consanguineous parents, presented with SRNS at 27 years of age and reached ESRD 2 years later, without neurological symptoms, which is not the usual clinical presentation associated with COQ2 mutations [32]. Further clinical investigations and familial DNA sequencing are required to assess the segregation and the role of these COQ2 variants in the development of the disease. Six patients presented with mutations leading to autosomal dominant diseases: an INF2 mutation in one patient, a WT1 mutation in three patients, a LMX1B mutation in one patient and a TRPC6 mutation in one patient. These six patients presented as sporadic cases and require further familial investigations to determine if they carry de novo mutations.

Table 2

Mutations in SRNS genes identified

Family	G1	G2	Gene	Exon (zygosity, segregation)	Mutations	Exac	PPH2	SIFT	MT	Age at onset (years)	ESRD	Age at ESRD (years)	Extrarenal signs	References
2 (3153)	G1/G1	0/0	NPHS1	2 (het, p)	c.106 delG-p.Ala36Profs*6	NA	NA	NA	NA	0 (10 days)	No		Microcephaly, pulmonary valve stenosis	[9]
				20 (het, m)	c.2728 T>C-p.Ser910Pro	NA	0.959	0.01	DC 1					[9, 10]
3 (2934)	0/0	0/0	NPHS1	4 (het, m)	c.479 G>C-p.Cys160Ser	NA	1	0	DC1	0.03	No			[11, 12]
				27 (het, p)	c.3478 C>T-p.Arg1160*	8/120 000	NA	NA	NA					[13–15]
4 (1075)	0/0	0/0	NPHS1	11 (het)	c.1379 G>A-p.Arg460Gln	1/120 000	0.476	0.33	PM 0.960	0.4	Yes	12.6		[16, 17]
				22 (het)	c.2928 G>T-p.Arg976Ser	8/120 000	0.873	0	DC 0.953					[15, 17, 18]
5 (2464, 2467, 2468)	0/0	G2/0	NPHS1	13 (hom, m, p)	c.1756 A>G-p.Arg586Gly	NA	0.005	0.01	DC 0.758	2.4	No			[10, 19]
	0/0	G2/0	NPHS1	13 (hom, m, p)	c.1756 A>G-p.Arg586Gly	NA	0.005	0.01	DC 0.758	7.2	No			[10, 19]
	0/0	0/0	NPHS1	13 (hom, m, p)	c.1756 A>G-p.Arg586Gly	NA	0.005	0.01	DC 0.758	0.09	No			[10, 19]
6 (1547)	0/0	0/0	NPHS2	6 (hom, m, p)	c.781 G>T-p.Glu261*	NA	NA	NA	NA	0.9	Yes	1.1	Pulmonary artery stenosis	No
7 (3901)	0/0	0/0	NPHS2	6 (hom, m, p)	c.779 T>A-p.Val260Glu	2/120 000	0.991	0	DC 1	0	No			[20–22]a
8 (580)	0/0	0/0	LAMB2	7 (hom, m, p)	c.736 C>T-p.Arg246Trp	1/120 000	1	0	DC 1	0.25	No		Pierson syndrome	[23, 24]a
9 (520)	0/0	0/0	PLCE1	18 (het)	c.4519_4520 delinsT -p.Glu1507*	NA	NA	NA	NA	5.5	Yes	6.1		No
				24 (het)	c.5475 A>C-p.Leu1825Phe	NA	0.998	0.03	DC 0.993					No
10 (3585)	0/0	0/0	COQ2	1 (het)	c.223_249delins AAGGA- p.Arg75Lysfs*90	NA	NA	NA	NA	27	Yes	29		No
				3 (het)	c.644 G>A-p.Gly215Glu	5/120 000 (5/10 000 Africans)	0.637	0.01	DC 1					No
11 (571)	0/0	0/0	INF2	2 (het)	c.311 G>T-p.Cys104Phe	NA	0.999	0	DC 1	10.3	Yes	14.9	Pes cavus, sensitive deficiency; hypoactive deep tendon reflexes; muscle weakness	[25]
12 (1389)	0/0	0/0	TRPC6	8 (het)	c.2150 T>C-p.Ile717Thr	NA	0.914	0	DC 1	4.0	No			No
13 (3786)	0/0	0/0	WT1	9 (het)	c.1432 + 5 G>A-Splice	NA	NA	NA	NA	NA	Yes	9		[26–28]
14 (3252)	G1/0	0/0	WT1	9 (het)	c.1432 + 4 C>T-Splice	NA	NA	NA	NA	23.1	Yes	23.8		[29–31]
15 (1749)	0/0	G2/0	WT1	3 (het)	c.799 del-p.Val267Serfs*19	NA	NA	NA	NA	19.3	Yes	21.8		No
16 (2110)	0/0	0/0	LMX1B	4 (het)	c.725 C>G-p.Ser242Trp	NA	0.999	0	DC 1	2.2	No			No

a

Patient already reported in this publication.

Hom, homozygous; het, heterozygous; comp het, compound heterozygous; p, paternal; m, maternal; NA, not available; MT, Mutation Taster; PPH2, PolyPhen-2.

Thus a mutation was less frequently found in patients in the APOL1 HR group than in the LR group (P = 0.0006). For further analysis, patients with and without identified mutations were analysed separately.

Comparison of patients characteristics according to their APOL 1 status

Clinical data are shown in Table 1.

Patients in the HR group were more likely to originate from the French West Indies than from Africa than those in the LR group [45/66 (68.2%) versus 30/86 (34.9%); P < 0.0001]. The exact geographical origin for patients from Africa was available in all but seven patients (six families). In the LR group, only three patients from the same family originated from East Africa, one from Central Africa and the others from West Africa. In the HR group, one patient originated from Central Africa and the other patients originated from West Africa. Two patients in the HR group and one patient in the LR group presented with sickle cell disease.

At diagnosis, patients in the HR group without mutation were older than patients in the LR group without mutation [22.5 years (SD 12.8) versus 14.7 (14.1); P < 0.0001] and were more often >18 years of age than in the LR group without mutation [35 (53.8%) versus 19 (27.1%); P = 0.003]. As expected, patients in the LR group with mutation were younger at diagnosis than patients without mutation in the LR and HR groups.

Age at onset according to APOL1 HR alleles or mutations in SRNS genes is presented in Figure 1. In the APOL1 HR patients, renal disease was diagnosed at a broad age range of 0–64 years. However, in very young children with disease onset before 1 year of age, a mutation was more likely to be identified. APOL1 HR-associated FSGS tended to present in children and young adults, with the majority of cases presenting between 6 and 40 years of age.

FIGURE 1

Percentage of patients with mutation SRNS genes or APOL1 variant detected in relation to age of onset of proteinuria. Histograms indicate fraction (in percentage) of patients with mutation detected in n patients per N patients examined per age group (on top of histograms).

Patients without mutation in the HR group had a lower eGFR at diagnosis than patients without mutation in the LR group [78.9 (SD 43.2) versus 98.8 (32.7) mL/min/1.73 m²; P = 0.02]. Mean proteinuria and albuminaemia were similar in both groups.

Familial cases

Among the families with several affected individuals, mutations in monogenic SRNS genes were identified in only six patients from four families, as shown above. In the patients without mutation, there were more familial cases in the APOL1 HR group than in the LR group [27 (41.5%) versus 8 (11.4%); P < 0.0001].

Among the 10 families (of 18) in which several DNA samples were available from affected patients, we checked whether the APOL1 alleles were segregating with the disease. It was the case in eight families. Surprisingly, in two families, only 3/4 and 1/2 patients had the two risk allele genotype. In the first family (Family 1; Table 3), one patient reached ESRD due to FSGS but did not carry any APOL1 risk alleles, whereas three other affected patients carried two risk alleles (Figure 2; Family 1 pedigree). Discrepancies within this family could not be explained by the APOL1 duplication, as shown above. In the other family (Family 17; Table 4), one patient had two risk alleles, whereas his affected brother had only one risk allele. No mutation in monogenic SRNS genes was identified in these patients.

FIGURE 2

Family 1 pedigree. PU, proteinuria; HU, haematuria; MCD, minimal change disease at renal biopsy.

Table 3

Family 1: clinical presentation and APOL1 risk allele status

Cases	Sex	APOL1 risk alleles	At diagnosis			At last follow-up
			Age (years)	eGFR	Nephrotic syndrome	Age (years)	eGFR	Dialysis	Transplantation (age)
III 1 (3285)	M	G1/G2	63	Normal	No, PU	NA	NA	No	No
III 4 (3122)	F	G1/G2/G0	52	82	No, PU	56	81	No	No
III 9 (3250)	F	G1/G2/G0		Normal	No, PU	45	80	No	No
III 12 (3172)	M	G0/G0	38	NA	NA	43		Yes	Yes (43 years)

M, male; F, female; PU, non-nephrotic proteinuria; NA, not available.

Table 4

Family 17: clinical presentation and APOL1 risk allele status

Cases	Sex	APOL1 risk alleles	At diagnosis			At last follow-up
			Age (years)	eGFR	Nephrotic syndrome	Age (years)	eGFR	Dialysis (age)	Transplantation (age)
1 (1019)	M	G2/G2	15	4	Yes	17		Yes (15 years)	Yes (17 years)
2 (1037)	M	G2/0	8.5	157	No, PU	9	126	No	No

M, male; F, female; PU, non-nephrotic proteinuria.

Histological data

The histological pattern of patients in both groups is shown in Table 5. A total of 89 patients presented with FSGS on renal biopsy, including 46 patients of French West Indies origin and 43 from Africa. The APOL1 HR genotype frequency among these FSGS patients was 49.4% (63% and 32% among patients of French West Indies and Guyana and Africa origin, respectively). Numerically more patients presented with collapsing FSGS in the HR group and more patients with minimal change disease in the LR group, but this was not statistically significant.

Table 5

Histological pattern according to APOL1 risk allele

Histological pattern	All	LR group		HR group
		Mutation	No mutation	Mutation	No mutation
No. of patients	152	16	70	1	65
Collapsing FSGS, n (%)	10 (6.7)	0	3 (4.2)		7 (10.8)
Tip lesion, n (%)	1 (0.6)	0	1 (1.4)		0
FSGS not otherwise specified, n (%)	78 (51.3)	7 (43.7)	35 (50.0)		36 (55.4)
Mesangial proliferation, n (%)	5 (3.3)	0	4 (5.7)		1 (1.5)
Minimal change disease, n (%)	33 (21.7)	3 (18.7)	21 (30.0)		9 (13.8)
Diffuse mesangial sclerosis, n (%)	1 (0.6)	0	1 (1.4)		0
Terminal kidney, n (%)	3 (1.9)	0	1 (1.4)		2 (3.1)
NA, n (%)	21 (13.8)	6 (37.5)	4 (5.7)	1	10 (15.4)

NA, complete biopsy report not available.

Evolution according to APOL1 status

All patients were steroid resistant in the HR group compared with 46 (88.5%) in the LR group without mutation (P = 0.03) (Table 1). Furthermore, two patients were sensitive to other immunosuppressive drugs in the HR group without mutation compared with seven in the LR group without mutation.

Nineteen percent of transplanted patients (five patients) relapsed after transplantation, with early relapse (before 7 days) in three patients (two in the LR group and one in the HR group) and late relapse (after 1 year) in one patient in the LR group. In the remaining patients of the HR group, FSGS recurrence was attributed to an APOL1 HR variant in the donor. Altogether, 19 patients can be classified retrospectively as immune forms, 16 in the LR group and 3 in the HR group (P = 0.01).

At the last follow-up, eGFR was lower in the HR group without mutation compared with the LR group without mutation [85.1 (SD 62.5) versus 107.9 (58.2) mL/min/1.73 m²; P = 0.02] and renal survival tended to be worse in the HR group than in the LR group by Kaplan–Meier analysis (Figure 3; P = 0.06). In the HR group, renal survival was not statistically different between patients of African origin and patients from the West Indies (Supplementary data, Figure S1). Patients were older at ESRD in the HR group without mutation [27.9 (SD 13.8) versus 17.9 (13.4) years; P = 0.03].

FIGURE 3

Renal survival in the HR and LR groups by Kaplan–Meier analysis (P = 0.06).

DISCUSSION

We analysed APOL1 risk allele genotype frequency in a large cohort of 152 FSGS patients of African ancestry living in Europe. The two risk allele genotype was frequent, found in 43% of patients in the whole cohort and in 60% of patients originating from the French West Indies, which was significantly higher than in controls. Until now, the only available data in FSGS patients were obtained in African American patients. The frequency of the two risk allele genotype in patients originating from the French West Indies in our study is similar to that observed in FSGS African American patients: the APOL1 risk genotype was present in 60–72% of self-identified African American FSGS patients in previous studies compared with 12.5–13% in controls [3, 4, 33–35].

To analyse the relative roles of APOL1 and of the genes known to be involved in FSGS/SRNS in these patients, we screened for pathogenic mutations with a panel including 35 SRNS genes. Of note, an important percentage of patients had a familial disease in the APOL1 HR group, whereas a mutation in SRNS genes was found in only one. Most of the mutations in monogenic SRNS genes were found in young patients in the APOL1 LR group. Because of the high frequency of G1 and G2 and the large increased risk of FSGS associated with these variants, families in which multiple individuals have FSGS may have APOL1-associated disease rather than one of the rarer Mendelian forms of disease. Accordingly, it has been shown that the relatives of African Americans with non-diabetic ESRD are enriched for APOL1 risk variants [36].

Interestingly, in two families, not all affected individuals carried the two risk allele genotype and no other mutation was identified. In two of these patients a duplication of APOL1 was detected, but not in two other affected members of the family [8]. The extra copy of APOL1 may be of either genotype G0 or G1. It is unknown whether G0G1, G1G1 and G0G0 haplotypes behave like G0 or G1, or neither. In the article by Ruchi et al. [8], the APOL1 duplication was significantly more frequent in kidney disease cases with an apparent G0G1 genotype than among controls with an apparent G0G1 genotype. Additional studies are needed to determine the functional effects of bearing more than two copies of APOL1, especially three risk alleles, as in one of our patients, since in a recent experimental model the severity of the disease correlated with the level of expression of the risk allele [37].

Patients in the HR group were more likely to originate from the French West Indies than from Africa. Among patients originating from Africa, most patients originated from West Africa. This is concordant with the prevailing hypothesis that G1 and to a lesser degree G2 renal risk alleles rose to high frequencies in West Africa because of positive selection by T. brucei rhodesiense, the causal agent of acute human African trypanosomiasis [38]. The variants are found outside Africa at high frequencies in persons of African ancestry. The trans-Atlantic slave trade resulted in the forced diaspora of millions of Africans from the Atlantic coastal regions of Africa to the Caribbean and the Americas [38]. The high allele frequencies of APOL1 G1 and G2 variants in African Americans and in the Caribbean reflect the APOL1 G1 and G2 frequencies along the Atlantic coast of Africa. The higher frequency observed in patients originating from the French West Indies compared with patients originating from Africa might be due to the fact that a high proportion of them originated from West Central Africa, in particular Nigeria, where a very high prevalence of the HR genotype is observed in the general population compared with other countries of West Africa [39].

The mean age at nephrotic syndrome onset was 22.3 years in the HR group in our cohort, higher than in the Nephrotic Syndrome Study Network (NEPTUNE) study (17.0 years) but lower than in another FSGS cohort (31.7 years) [4, 35]. We showed that APOL1-associated FSGS in our cohort presented at a broad age range of 0–64 years, but in the majority of patients the disease was diagnosed at 6–40 years. Together these data support the idea that FSGS among APOL1 HR individuals is a disease presenting during childhood and early adulthood.

Most patients had normal blood pressure in both groups. This demonstrates that FSGS associated with the APOL1 HR genotype is distinct from hypertensive nephropathies associated with the APOL1 HR genotype. Baseline and last follow-up eGFRs were significantly lower among patients with the APOL1 HR genotype, which is consistent with the faster progression rate that has been observed in these individuals in two previous studies [4, 5]. In our study, renal survival also tended to be worse in patients with the APOL1 HR genotype. Due to its prognostic value, we suggest adding APOL1 in gene panels used in SRNS. As APOL1-associated FSGS has a propensity to progress rapidly to ESRD, the question of treatment is crucial. In a previous study, the frequency of glucocorticoids sensitivity was similar in subjects with two APOL1 risk alleles [12/42 (29%)] and in subjects with zero or one APOL1 risk allele [5/15 (33%)] [4]. Furthermore, the APOL1 risk genotype did not seem to affect response to other immunosuppressive treatment regimens [5]. In the NEPTUNE study, the APOL1 HR genotype was associated with a 69% reduction in the probability of complete remission at any time, independent of histologic diagnosis [35]. In our study, 94% of patients were steroid resistant and were sent for genetic testing. In the HR group, all patients were steroid resistant and only a few could be retrospectively considered as immune forms. This is concordant with the clinical and experimental data showing the role of the intracellular expression of APOL1 in the podocytes, including decreased graft survival after renal transplantation in patients receiving at-risk allografts [40–42].

In conclusion, the two risk allele genotype is found in 43% of patients with African ancestry in our cohort in Europe, especially in patients originating from the West Indies. It is more frequent in young adult patients, but is also found in children, and is associated with a poor renal prognosis. The APOL1 HR genotype is usually not associated with other causative mutations in known monogenic SRNS genes and families in which multiple individuals have FSGS may have APOL1-associated disease.

ACKNOWLEDGEMENTS

The authors thank all the clinicians who referred their patients for genetic analysis: Dr Belenfant, Montreuil; Dr Galichon, Tenon Hospital, Paris; Dr Glotz, Saint Louis Hospital, Paris; Dr Guigons, Limoges; Dr Hue, CHU Guadeloupe; Dr Peraldi, Saint Louis Hospital, Paris; Dr Pillebout, Saint Louis Hospital, Paris; Dr Pouteil Noble, Lyon; Dr Ranchin, Lyon; Dr Tabbianeni, Montreuil; Dr Ulinski, Trousseau Hospital, Paris, France.

FUNDING

The research leading to these results has received funding from the Investments for the Future Programme (grant ANR-10-IAHY-01 to C.A.) and from the European Union’s Seventh Framework Programme (FP7/2007-2013) grant 305608 (EURenOmics).

AUTHORS’ CONTRIBUTIONS

O.G., O.B., M.J.T., C.A. and A.S. performed genetic experiments; contributed to study conception and design; performed analysis and interpretation of data; drafted and revised the article; provided intellectual content of critical importance to the work described and approved the final version of the article to be published. B.K., J.D., A.K., J.H. and C.D. took care of patients, contributed to analysis and interpretation of data, drafted and revised the article, provided intellectual content of critical importance to the work described and approved the final version of the article to be published. C.F. and O.A. performed genetic experiments, contributed to analysis and interpretation of data, drafted and revised the article, provided intellectual content of critical importance to the work described and approved the final version of the article to be published.

CONFLICT OF INTEREST STATEMENT

None declared.

Author notes

Corinne Antignac and Aude Servais contributed equally to this work.