The amino acid mutations of the podocin in proteinuria: a meta-analysis

Abstract

While many previous studies have reported an association between the single-nucleotide polymorphisms (SNPs) of the podocin and proteinuria occurred, a conclusive relationship has not been defined in every oligoallelic state of amino acid (AA) mutations in podocin. In this study, we performed a meta-analysis of the published data to investigate the impact of the oligoallelic AA mutations of the podocin on proteinuria; a total 16 AA mutations were investigated for oligoallelic pathogenicity. Despite significant heterogeneity within some of the comparisons, the results revealed significantly higher risks of proteinuria in early-onset (onset age <16) individuals for five mutations (P118L, R138Q, R168H, V180M, and V260E), and in all onset ages individuals for five mutations (R138Q, G140X, R229Q, V260E, and V290M) compared to non-variant individuals. We also tested the steroid response in individuals with R229Q and E237Q. No statistically significant differences in the two mutations carrier rate were observed between steroid resistance patients and controls. No AA mutation was selected for meta-analysis on the recurrence of proteinuria after renal transplantation as lack of control data. In conclusion, our meta-analysis tested the pathogenicity of the oligoallelic AA mutations in podocin and suggested the potential causative mutations, and the alleles showing an association with protein susceptibility. The sensitivity and specificity of each causative mutation are pending further testing.

• Amino acid mutations
• meta-analysis
• proteinuria
• SNP

Introduction

Proteinuria is generally regarded as the principal risk factor for deterioration of renal function. The ongoing proteinuria results in progressive renal damage. More than half of the patients with macro albuminuria progress to end-stage renal disease (ESRD) within 10 years.1^,2 The glomerular filtration barrier injury is responsible for leakage of high-molecular-weight proteins. It can be secondary to a wide range of medical problems, including various cardiovascular, metabolic and autoimmune diseases, or occur as a heredity disorder as a result of gene mutation. Numerous advances in our understanding of gene variant and renal diseases have been witnessed in the past 10 years. More than 10 genes have been identified as pathogenic forms until now.3–5 Some genetic variants are highly correlated with proteinuria progression and relapse, therefore the identification of a mutation in a patient renders further immunosuppressive treatment unnecessary or disease recurrence in a kidney graft unlikely.6

Podocin, the protein consists of 383 amino acid (AA) and encoded by the NPHS2 gene, is essential for an intact kidney filtration barrier and usually mutated in patients with proteinuria. Its function is currently understood as a scaffold providing the necessary protein lipid micro-environment needed for proper function and signaling of the slit diaphragm protein complex.7–9 NPHS2 knock-out mice developed proteinuria during the antenatal period and died a few days after birth from renal failure caused by massive mesangial sclerosis.10 In previous reports, 20–40% of the familial, autosomal recessive, pediatric steroid-resistant nephrotic syndrome (SRNS) can be ascribed to mutations of the NPHS2 gene. Single-nucleotide polymorphism (SNP) is the main form among these mutations.11^,12. In summary, more than 3000 mutations of NPHS2 have been detected in previous study. Nearly 1% of them change codon that codes for different AA or AA deleted. Non-conservative mutation, include missense mutations and non-sense mutation, can render the resulting protein structural changed, but are not always pathogenic. As two or more mutations of these variants usually occur simultaneously, the pathogenic capacity of an oligoallelic form is often ignored in previous studies.13 The evidence of the oligoallelic mutations is weak due to a lack of adequate data and discrepancies between the published investigations. A meta-analysis may clarify the association between the mutations and susceptibility to proteinuria; therefore, we performed meta-analyses to comprehensively investigate the effects of the mutations of podocin on patients with proteinuria in a pooled study.

Methods

Search strategy

Systematic literature searches were conducted in the following electronic databases: PubMed, Embase, Web of Science, Chinese Biomedical Literature (CBM), and the Cochrane Library. The relevant articles were published before 15 January 2014. Non-English publications satisfying the inclusion criteria were translated into English and then assessed. The following search terms were used: (NPHS2 OR Podocin) AND proteinuria AND (polymorphism/polymorphisms OR variant/variants OR mutations/mutation). To identify additional potentially relevant publications, the related references from all the retrieved articles and reviews were manually searched. Only published studies with full-text articles were included in the meta-analysis. All data were entered into the Review Manager 5.0 software (Cochrane Collaboration, Germany) by one author (Lu Lu) and checked by another author (Xiao-ming Sun). Any disagreements were resolved by discussion between the two reviewers and by seeking the opinion of a third reviewer (Yi Yin) when necessary.

Inclusion criteria

Studies were selected based on the following inclusion criteria: (1) at least one AA mutation of podocin was detected in the study. (2) The study was a case–control study or a cohort study of renal diseases with proteinuria. (3) The age of the patients at disease onset was described as either early-onset (onset age <16 years) or all onset ages. (4) The diagnostic criteria of diseases were described in the study and were assessed by professional physicians or biopsies.

Exclusion criteria

Studies were excluded according to the following criteria: (1) a non-human population was mentioned. (2) The study was a case report or a familial investigation. (3) The proteinuria was secondary to anon-renal cause (overflow proteinuria, drug side effect, excessive fluid intake, etc.). (4) The presence of mutations was not assessed in all individuals. (5) All individuals with variants were compound status.

Data extraction and synthesis

Two independent investigators (Lu Lu and Xiao-ming Sun) performed the data extraction using the inclusion and exclusion criteria described above. All discrepancies were resolved by discussion and, if required, participation by the third author (Yi Yin). The following information was extracted from each study: the identical AA mutations, first author’s surname, year of publication, phenotype of diseases, geographical location of the study population, and the number of cases and controls with each AA mutation. Allele frequencies were calculated for the case and control groups from the corresponding genotype distribution. Patients with the compound mutations were removed from the analysis and listed on the tables. The two investigators’ results were compared, and disagreements were resolved by discussion.

Statistical analysis

The statistical analyses were performed using Stata Version10.0 (Stata Corp LP, College Station, TX) software. The included studies were categorized by disease phenotype and the AA mutations. Then, the comparisons of mutations carriers were made between cases and controls by pooling data. Only an AA mutation included by more than one studies was selected to the analyses. The association between the AA mutation and the disease phenotype was determined by calculating the respective odds ratio (OR) with a 95% confidence interval (CI). To validate the meta-analysis results, a sensitivity analysis was performed by the sequential omission of individual studies. Heterogeneity of the effect sizes was evaluated using the Q statistic and the I² statistic.14 A fixed-effects model was used when the p value was greater than 0.05 and I² was less than 50%; otherwise, a random effects model was used. The significance of the pooled OR was determined using a Z test, and a p value of less than 0.05 was considered significant. Publication bias was simulated using the Harbord linear regression method. To evaluate our results, two indirect in silico predictors, Mutation Taster (http://www.mutationtaster.org/) and PolyPhen2 (http://genetics.bwh.harvard.edu/pph2/index.shtml) were used for reference. The above work was completed by two authors and checked by the third author.

Results

Characteristics of the included studies

A total of 236 possibly relevant published articles were identified following the retrieval strategy outlined above. After strict review, 185 of these publications were excluded and 53 were selected for our meta-analysis. Figure 1 shows the selected articles included 64 AA mutations from eight exons of NPHS2 (from 143 comparative trials), of which 59 were studies on proteinuria susceptibility (Table 1), eight were studies on steroid response (Table 2) and five were studies on the recurrence of proteinuria after renal transplantation (Table 3). Of the included trials, 60 were conducted in Europe from 13 countries (Belgium, Cyprus, Finland, France, German, Greece, Hungary, Italy, Netherland, Spain, Poland, the UK, and the Czech Republic), 41 were Western Asian and Arabian trials (Egypt, Saudi Arab, and Turkey), 26 were North American trials (the US and Canada), ten were South Asian trials (India and Pakistan), two were from Mexican, and two were form Japan. A total of 45,839 samples (including 22,725 patients and 23,114 healthy controls) were used for the analysis. As only a AA mutation included by more than one studies can be pooled and analyzed, we selected 16 mutations (P20L, Q39X, P118L, P118L, R138X, G140X, V165X, R168H, L169P, V180M, R229Q, A242V, V260E, V268G, A284V, and V290M) on proteinuria susceptibility and two mutations (R229Q and E237Q) on steroid response. As a reference, only synonymous mutations of podocin were found in East Asian patients with proteinuria (163 Chinese, 136 Japanese, and 70 Koreans) and the five relevant studies were not included in the analysis described above.15–21

Figure 1. The included oligoallelic AA mutations on podocin. Human podocin consists of 383 amino acids from 8 exons. A total of 64 AA mutations were analyzed in our study, including 46 AA substitutions and 18 AA deletions. The PHB domain was indicated.

Table 1. The included oligoallelic AA mutations on proteuria susceptibility.

Exon 1	AA Changing	P20L	E30Q	Q39L	Q39X	G42R	P45X	E87K	P89T
	No. of studies [Ref.]	9 [A]	142	135	242^,43	243	145	146	142
	No. of participants	3458	345	345	582	1284	16	120	345
Exon 2	AA Changing	G92C	G97S	I108V	M115T	P118L	W122L	K126N
	No. of studies [Ref.]	147	148	149	142	4 [B]	142	150
	No. of participants	484	350	872	345	1154	345	245
Exon 3	AA Changing	V127X	Q129X	I136L	R138Q	R138X	L139R	G140X	L142P	R146X
	No. of studies [Ref.]	142	143	145	20 [C]	324^,42^,51	152	4 [D]	152	153
	No. of participants	344	237	16	6900	688	15	1034	15	97
Exon 4	AA Changing	D160G	Y162X	V165X	R168H	L169P
	No. of studies [Ref.]	147	154	342^,55^,56	243^,57	4 [E]
	No. of participants	484	111	668	359	1859
Exon 5	AA Changing	V180M	F185X	I192V	R196X	L203P	A208T	A213T	F216Y	R229Q	E237Q	R238X	A242V
	No. of studies [Ref.]	6 [F]	158	142	119	142	159	142	145	22 [G]	156	145	4 [H]
	No. of participants	1927	279	344	113	345	169	345	16	6815	270	16	1803
Exon 6	AA Changing	G257E	V260E	E264Q
	No. of studies [Ref.]	156	34^,43^,50	160
	No. of participants	270	711	370
Exon 7	AA Changing	V268G	V268X	H276L	A284V	R286X	K289X	V290M	R291Q
	No. of studies [Ref.]	255^,56	142	142	5 [I]	161	142	5 [J]	158
	No. of participants	323	345	345	1144	103	343	1692	279
Exon 8	AA Changing	A301X	S302X	A317X	R322Q	Q328R	P341S	L347X
	No. of studies [Ref.]	142	156	162	142	156	150	156
	No. of participants	345	270	90	345	270	245	269

Notes: The No. of participants was adjusted by excluding compound station samples. Capital letters were used instead of Ref. with number more than three: A [11,35,44,48,51,56,63]; B [42,47,55,56]; C [4,12,42–44,47,49,51,53,56,57,59,61–68]; D [24,51,61,63]; E [42,49,58,63]; F [42,43,49,56,61,66]; G [11,41,42,44,46,48–50,53,55–57,60,63,67,69–74]; H [44,56,58]; I [11,31,43,56,61,65]; J [12,42,48,49,62].

Table 2. Characteristics of the studies that have evaluated the effect of the oligoallelic AA mutations on the steroid response.

								Case			Control
Mutation	Exon	Year	References	Phenotype	Onset age	Region	Excluded	WT	Het	Hom	WT	Het	Hom
P20L	Exon 1	2004	Ruf et al.56	NS	Early-onset	US		190	0	0	126	1	0
A41G	Exon 1	2007	He et al.75	FSGS	All onset ages	Canada		62	1	0	15	0	0
G42R	Exon 1	2007	He et al.75	FSGS	All onset ages	Canada		62	1	0	15	0	0
A61V	Exon 1	2007	Gbadegesin et al.25	NS	Early-onset	US		41	1	0	10	0	0
R138Q	Exon 3	2004	Ruf et al.56	NS	Early-onset	US	12	158	2	18	126	1	0
R229Q	Exon 5	2012	Reiterova et al.48	NS	All onset ages	Czech Republic	1	32	2	1	12	2	0
R229Q	Exon 5	2007	He et al.75	FSGS	All onset ages	Canada	1	57	6	0	12	2	0
R229Q	Exon 5	2007	Gbadegesin et al.25	NS	Early-onset	US		40	2	0	10	0	0
E237Q	Exon 5	2007	He et al.75	FSGS	All onset ages	Canada		63	0	0	14	1	0
E237Q	Exon 5	2004	Ruf et al.56	NS	Early-onset	US		189	1	0	126	1	0
E310K	Exon 8	2007	He et al.75	FSGS	All onset ages	Canada		62	1	0	15	0	0

Note: NS, Nephrotic syndrome; FSGS, focal segmental glomerulosclerosis; Excluded, the samples were excluded as compound station; WT, wild type; Het, heterozygous; Hom, homozygous.

Table 3. Characteristics of the studies that have evaluated the effect of the oligoallelic AA mutations on the recurrence of proteinuria after renal transplantation.

								Case			Control
Mutation	Exon	Year	References	Phenotype	Onset age	Region	Excluded	WT	Het	Hom	WT	Het	Hom
P20L	Exon 1	2004	Weber et al.76	SRNS	Early-onset	France		1	1	0		N/A
R138Q	Exon 3	2014	Holmberg et al.6	NS	Early-onset	Finland	1	3	0	2		N/A
R138Q	Exon 3	2003	Bertelli et al.77	FSGS	Early-onset	Italy		7	0	2		N/A
R138X	Exon 3	2014	Holmberg et al.6	NS	Early-onset	Finland		4	0	2		N/A
R138X	Exon 3	2004	Weber et al.76	SRNS	Early-onset	France	1	30	0	1		N/A
S211T	Exon 5	2003	Bertelli et al.77	FSGS	Early-onset	Italy		1	1	0		N/A
T326X	Exon 8	2004	Weber et al.76	SRNS	Early-onset	France		2	1	0		N/A
L347X	Exon 8	2014	Holmberg et al.6	NS	Early-onset	Finland		5	0	1		N/A
L347X	Exon 8	2004	Ruf et al.56	NS	Early-onset	US	1	23	0	1	80	0	0

Note: NS, Nephrotic syndrome; FSGS, focal segmental glomerulosclerosis; SRNS, steroid-resistant nephrotic syndrome; Excluded, the samples were excluded as compound station; WT, wild type; Het, heterozygous; Hom, homozygous; N/A, not applicable as non-human population was mentioned.

Sensitivity analysis

A sensitivity analysis was performed to explore the influence of the trial quality on the magnitude of the effect. We used the METANINF command in Stata to investigate the influence of each individual study on the overall meta-analysis summary estimate.22 Three subgroups were analyzed using a random-effects model, and 24 subgroups were analyzed using a fixed-effects model. For all of the pooled analyses, the significance of the OR was not greatly influenced by omitting any single study. This result demonstrates the stability of the data in our analysis.

Association of the AA mutations with proteinuria susceptibility

A total 99 of the included studies reported 16 AA mutations for the proteinuria patients and controls. In our meta-analysis, we did not find any significant associations between eight AA mutations and proteinuria across the study populations (P20L, Q39X, R138X, V165X, L169P, A242V, V268G, and A284V). The ORs of these AA mutations were not statistically significant (p > 0.05). In the all onset age populations, the ORs that were estimated for proteinuria susceptibility were statistically significant in five mutations: R138Q (OR 10.640, 95% CI 5.937–19.067, p < 0.001), G140X (OR 4.945, 95% CI 1.120–21.829, p = 0.035), R229Q (OR 1.450, 95% CI 1.003–2.097, p = 0.048), V260E (OR 8.466, 95 % CI 1.524–47.038, p = 0.015), and V290M (OR 4.417, 95% CI 1.042–18.718, p = 0.044). In the early-onset (onset age < 16) populations, the ORs that were estimated for proteinuria susceptibility were also statistically significant in five mutations: P118L (OR 5.539, 95% CI 1.336–22.965, p = 0.018), R138Q (OR 11.555, 95% CI 6.112–21.845, p < 0.001), R168H (OR 10.760, 95% CI 1.091–106.16, p = 0.042), V180M (OR 7.997, 95% CI 2.478–25.805, p = 0.001), and V260E (OR 13.359, 95% CI 1.587–112.417, p = 0.017). Three mutations (R140X, R229Q, and V290M) were determined the pathogenic effect in all onset age populations but not in the early-onset populations. The results of Polyphen-2 and Mutation taster were consistent: All selected mutations have been identified to disease causing except for P20L (benign). The details of the results above are listed in Table 4.

Table 4. Meta-analysis of the single mutations on NPHS2 and association with proteinuria/steroid response.

				Heterogeneity test				Odds ratio				H test
Mutation	Phenotype	Onset age	No. of studies	Q	I^2	p	Model	OR	(95%CI)	p	p	Mutation taster	Polyphen2 (Score)
P20L	Proteinuria	All onset ages	7	19.87	69.80%	0.003	R	1.467	(0.274–7.846)	0.654	0.513	DC	B (0.098)
	Proteinuria	Early-onset	2	2.86	65.10%	0.091	R	0.518	(0.015–17.841)	0.716	–
Q39X	Proteinuria	Early-onset	2	1.48	32.50%	0.223	F	2.386	(0.251–22.678)	0.449	–	DC	N/A
P118L	Proteinuria	Early-onset	4	2.32	<0.01%	0.509	F	5.539	(1.336–22.965)	0.018*	0.031*	DC	PD (0.974)
R138Q	Proteinuria	All onset ages	20	12.3	<0.01%	0.872	F	10.640	(5.937–19.067)	<0.001*	0.029*	DC	PD(1.000)
	Proteinuria	Early-onset	17	11.79	<0.01%	0.758	F	11.555	(6.112–21.845)	<0.001*	0.029*
R138X	Proteinuria	Early-onset	3	0.52	<0.01%	0.77	F	1.458	(0.238–8.916)	0.683	0.499	DC	N/A
G140X	Proteinuria	All onset ages	4	0.34	<0.01%	0.842	F	4.945	(1.120–21.829)	0.035*	0.621	DC	N/A
	Proteinuria	Early-onset	3	1.02	<0.01%	0.797	F	3.136	(0.532–18.484)	0.207	0.588
V165X	Proteinuria	Early-onset	3	0.38	<0.01%	0.826	F	1.651	(0.276–9.880)	0.583	0.426	DC	N/A
R168H	Proteinuria	Early-onset	2	0.05	<0.01%	0.832	F	10.760	(1.091–106.16)	0.042*	–	DC	PD (1.000)
L169P	Proteinuria	All onset ages	4	1.23	<0.01%	0.746	F	2.339	(0.493–11.103)	0.285	0.239	DC	PD (0.999)
	Proteinuria	Early-onset	3	0.72	<0.01%	0.698	F	1.613	(0.255–10.204)	0.612	0.239
V180M	Proteinuria	Early-onset	6	8.3	39.80%	0.14	F	7.997	(2.478–25.805)	0.001*	0.821	DC	PD (0.879)
R229Q	Proteinuria	All onset ages	21	33.16	39.70%	0.032	R	1.450	(1.003–2.097)	0.048*	0.365	DC	PD (0.903)
	Proteinuria	Early-onset	9	11.31	29.30%	0.185	F	0.975	(0.621–1.533)	0.914	0.276
A242V	Proteinuria	All onset ages	4	0.14	<0.01%	0.987	F	1.164	(0.758–1.788)	0.738	0.754	DC	PD (0.992)
	Proteinuria	Early-onset	2	0.01	<0.01%	0.903	F	1.472	(0.152–14.248)	0.487	–
V260E	Proteinuria	All onset ages	3	1.09	<0.01%	0.581	F	8.466	(1.524–47.038)	0.015*	0.866	DC	PD (0.998)
	Proteinuria	Early-onset	2	0.12	<0.01%	0.724	F	13.359	(1.587–112.417)	0.017*	0.735
V268G	Proteinuria	Early-onset	2	0.03	<0.01%	0.865	F	1.558	(0.159–15.289)	0.704	–	DC	PD (0.993)
A284V	Proteinuria	All onset ages	4	0.55	<0.01%	0.908	F	3.528	0.725–17.179	0.119	0.821	DC	PD (1.000)
	Proteinuria	Early-onset	3	0.39	<0.01%	0.822	F	3.014	0.496–18.327	0.231	0.856
V290M	Proteinuria	All onset ages	5	2.9	<0.01%	0.575	F	4.417	1.042–18.718	0.044	0.314	DC	PD (0.998)
	Proteinuria	Early-onset	4	2.02	<0.01%	0.568	F	3.443	0.707–16.767	0.126	0.465
R229Q	Steroid response	All onset ages	3	0.21	<0.01%	0.900	F	0.682	0.212–2.192	0.520	0.128	DC	PD (0.903)
E237Q	Steroid response	All onset ages	2	0.99	<0.01%	0.320	F	0.273	0.038–1.979	0.199	–	DC	PD(0.774)

Note: Model of meta: R = random-effects model, F = fixed-effects model; DC, disease causing; PD, Probably damaging; N/A, not applicable; *The p value was less than 0.05, and the OR was considered statistically significant.

Association of the mutations with steroid response and recurrence of proteinuria

Five of the included studies reported two mutations (R229Q and E237Q) allele frequencies for the SRNS patients and steroid-sensitive nephrotic syndrome (SSNS) controls. In our meta-analysis, we did not find any significant associations between the two mutations and steroid resistance across all of the study populations. Furthermore, even the R138Q, the R138X, and the L347X were both mentioned twice in our included studies, only one trial focusing on the L347X set up a control group contained 80 patients without recurrence. In the analysis for the recurrence of proteinuria after renal transplantation, no AA mutation was selected due to lack of control data. The results of Polyphen-2 and Mutation taster show that, except P20L in exon 1, all other selected mutations have been identified to disease causing. The details of the results above are listed in Table 4.

Publication bias

Although only three of the subgroups (R138Q, All onset ages group; R138Q, Early-onset group; R229Q, All onset ages group) contained more than 10 studies, Harbord tests were performed to assess the publication bias in this meta-analysis (Table 4). The results did not reveal obvious evidence of asymmetry in 24 of the analyses because the p values of the Harbord tests were greater than 0.1. However, the p value of the Harbord test was less than 0.1 for three analyses: the frequency of the P118L on proteinuria susceptibility in early-onset populations (p = 0.013), the frequency of the R138Q on proteinuria susceptibility in all onset ages populations (p = 0.029) and in early-onset populations (p = 0.029). This result suggests that there was a bias in these studies, but the in sufficient number of studies (only four studies included) may have caused the deviation in the bias. The number of studies was too small to conduct the Harbord test in seven of our analysis groups, for which p values could not be calculated.

Discussion

Podocin plays an important role in the slit diaphragm of podocytes. Mutations in this protein gene NPHS2 can cause nephrotic syndrome, such as focal segmental glomerulosclerosis (FSGS) or minimal change disease (MCD), and always related to steroids resistance.7^,23^,24 The mainstream opinions consider homozygous or compound heterozygous mutations in NPHS2 account for 50% of the cause of proteinuria.13^,25 However, the AA mutations were always compound in a patient and trans-associated with each other. The oligoallelic state of each AA mutation was hard to evaluate due to small samples in most experiments. In our study, the pooled results show eight mutations (P118L, R138Q, G140X, R168H, V180M, R229Q, V260E, and V290M) possess the causative effect for proteinuria with oligoallelic state. Of these mutations, three (G140X, R229Q, and V290M) were confirmed their effect in all onset ages subgroup but not statistically significant in early-onset age subgroup. It suggests that the three mutations may be associated with late-onset disease.

We cannot confirm an oligoallelic causal link between the other seven mutations (P20L, Q39X, R138X, V165X, L169P, A242V, and A284V) and the proteinuria. However, only the P20L was assessed as the benign mutation by the mutation taster and the Polyphen2, while the others were identified as disease causing. Perhaps it is because these mutations play as modifiers other than oligoallelic causative factors. For example, A242V and A284V were considered causal in trans-association to R229Q,26–28 and NPHS1 polymorphism may enhance pathogenicity of P118L in podocin.29 Furthermore, the sample size is usually small in control groups. In fact, using data from small sample size in control groups (compared with case groups) have the disadvantage that allele frequencies could be underestimated when merging data.30 With the improvement of population genetics databases in different ethnic groups, using data from whole-genome databases as controls may solve this problem. By the way, only the synonymous mutations, such as A318A and L346L, are found in East Asian patients. It suggests that the mutations in podocin may not be responsible for the proteinuria in East Asian populations.

Among the pathogenic mutations, the R229Q was a hot spot and a controversial issue. This AA exchange with arginine to glutamine substitution is attributed to G to A nucleotide substitution at position 686 in exon 5. It was first reported that R229Q decrease nephrin binding to podocin and cause SRNS by Tsukaguchi et al.31 However, subsequent studies denied its independent pathogenicity in oligoallelic state. Mainstream views generally considered R229Q leads to proteinuria only in association with a pathogenic mutation on the second NPHS2 allele.32^,33 In our study, we corrected the selective data by excluding patients with compound R229Q variant, and the pooled results also denied the associations between the R229Q variant and proteinuria in the early-onset study populations, but not in late-onset patients. Although the differences cannot be explained now, we must take into account the sample size and the onset age in these previous studies. The number of participants are usually too little to estimate the causative effect as the low allele frequencies (4–7% in Europeans and 0–1.5% in African and Asian populations), especially for the more rare oligoallelic or homozygous events.34^,35. Of the included 21 studies, only three derived positive results. The sample size in case groups is often several times bigger than in healthy control groups, it may lead to an underestimated result. When pooled all data of oligoallelic events, the meta-analysis result do show a statistical significance. On the other hand, some experiments in vitro show that the R229Q exert a deleterious effect on podocin only trans-associated with another pathogenic mutation, but the subjects were adolescent or children (age < 16 years).36 These experiments cannot exclude the hypothesis that the level of podocin is sufficient to maintain glomerular function despite its low expression or properties altered when patients were young. People with R229Q may be pre-disposing to mechanical instability and deterioration of the filtration barrier as a long-term effect. Therefore, we suppose oligoallelic R229Q may possess causative effect, and the age and the compound mutations are both amplification factors. Further studies are warranted to comprehensively investigate the precise relationship.

In general, mutations in the NPHS2 gene are implicated in an autosomal-recessive form of SRNS and recurrence of proteinuria after renal transplantation.6^,37 Identification of podocin mutations was considered not only save the patients from unnecessary treatment with cortisol agents but may also assess the prognosis of renal transplant.24^,38–40 In our study, however, we found no significant link between steroid resistance and the presence of two including mutations (R229Q and E237Q). This result is consistent with some previous studies.41 Our results show the oligoallelic mutations of podocin cannot currently be considered as potential predictors to the effect of steroid treatment. We also summarized the mutations in regard to the recurrence of proteinuria after renal transplantation. Unfortunately, only one study set up a control group including 80 patients without recurrence of proteinuria. Despite the absence of the meta-analysis, three mutations (R138Q, R138X, and L347X) are mentioned in more than one study and occur with significantly high frequency in case groups. R138X and R138Q are generally considered as the most common mutations in patients with recurrence, which leads to a truncated podocin protein.6 Further studies are warranted to investigate the link between recurrence of proteinuria after renal transplantation and mutations in podocin in large sample sizes.

It should be noted there might be possible limitations in the meta-analysis design. In our raw data, the sample merged familial and sporadic patients together, and the diagnosis was usually based on biopsy other than the presence of proteinuria. It may cause selection bias when pooled. Nonetheless, our meta-analysis tested the pathogenicity of the oligoallelic AA mutations in podocin and suggested the potential causative mutations, and the alleles showing an association with protein susceptibility. The sensitivity and specificity of each causative mutation are pending further testing.

Declaration of interest

We declare that we have no financial or personal relationships with other people or organizations that can inappropriately influence our work; there is no professional or other personal interest of any nature or kind in any product, service and/or company that could be construed as influencing the position presented in, or the review of, this manuscript. This research is supported by the National Science Foundation of China (No. 81173457 and No. 81001674).