Twenty-one novel mutations identified in the COL4A5 gene in Chinese patients with X-linked Alport’s syndrome confirmed by skin biopsy

Abstract

Background. The clinical and pathological features of Alport syndrome are characterized by abnormalities in the basement membrane collagen network which are composed of the α3, α4 and α5 chains of type IV collagen and usually associated with hearing loss and ocular lesions. The predominant form (85% of AS) is inherited as X-linked mode (XLAS) caused by mutations encoding the α5 chain of type IV collagen gene, COL4A5. Different mutations in the COL4A5 gene have been reported widely, but only a few mutations were identified in Chinese patients.

Methods. We studied 71 Chinese patients from 35 unrelated families with XLAS confirmed by skin biopsy. Genomic DNA was extracted from peripheral blood of all patients. All 51 exons of the COL4A5 gene were screened by direct sequencing for the probands.

Results. A total of twenty-five identified gene mutations were considered to be pathogenic, including 1 nonsense, 1 splice-site, 1 complex rearrangement, 5 small deletions, 2 small insertions and 15 missense mutations. Twenty-one mutations have not been reported previously.

Conclusions. We have identified 25 pathogenic mutations in 35 Chinese families with XLAS. Skin biopsy is effective for the diagnosis of XLAS.

Alport syndrome, COL4A5 gene, direct sequencing, skin biopsy

Introduction

Alport syndrome (AS), a common hereditary renal disease, is responsible for ∼1% of end-stage renal disease (ESRD) in Western countries [ 1 ]. It is characterized by haematuria and progressive renal failure with or without sensorineural hearing loss and ocular abnormalities. About 85% of AS is dominant and X chromosome linked (XLAS, OMIM 301050) [ 2 ]. XLAS is caused by mutations in the COL4A5 gene encoding the α5 chain of type IV collagen [ 3 ]. In XLAS patients, α5 chain is always simultaneously absent from the glomerular basement membrane (GBM) and the basement membrane underlying the epidermis (EBM), indicating that a skin biopsy may help to diagnose XLAS [ 4–6 ].

More than 400 different mutations in the COL4A5 gene have been reported in the literature, but very few of these mutations were identified in China [ 7–12 ]. Based on our previous work, we carried out this study to investigate the features of COL4A5 gene mutations in 71 Chinese XLAS patients from 35 unrelated families confirmed by skin biopsy.

Materials and methods

Patients and controls

Inclusion criteria.

Briefly, the inclusion of subject families was based on the classical diagnostical criteria of AS: positive family history of haematuria with or without progression to ESRD, progressive sensorineural hearing loss, characteristic ocular changes (lenticonus and/or maculopathy), characteristic ultrastructural changes of the GBM, absence or mosaic loss of α3, α4, α5 (IV) chains in the GBM and absence or mosaic loss of α5 chain in the EBM.

Based on the criteria, 71 patients (35 men and 36 women) with XLAS confirmed by skin biopsy from 35 unrelated Chinese families were enrolled from January 2004 to April 2009. Normal members in these families were also requested to take part in this study. One hundred healthy persons (50 women and 50 men) were used as control subjects.

The protocol of this study was approved by the ethics committee of Shanghai

Ruijin Hospital. All study participants gave written informed consent before enrolment in this study.

Mutation analysis

Genomic DNA was extracted from peripheral blood with a GenElute blood genomic DNA kit (Sigma-Aldrich, St Louis, MO) according to the manufacturer’s instructions. Forty-eight pairs of oligonucleotide primers were designed for amplifying all 51 exons and flanking intronic regions of the COL4A5 gene (NC_000023.10) [ 13 ]. Polymerase chain reaction (PCR) was performed in 25 μL of solution containing 30 ng genomic DNA, 0.15 μL of 5 U/mL Ex Taq polymerase (Takara Bio, Ohtsu, Japan), 2.0 μL of 2.5 × 10 ⁻³ mol/L deoxyribonucleotide triphosphate, 2.5 μL of 10 × Ex Taq buffer, and 1 μl of 5 pmol/L of each primer. PCR was performed with an initial denaturation step at 95°C for 5 min, subsequently followed by 35 cycles with denaturation at 95°C for 45 s, annealing at 48–57°C for 45 s and elongation at 72°C for 45 s. These Polymerase chain reaction (PCR) products were purified and directly sequenced on an ABI 3700-automated DNA sequencer (Perkin-Elmer Applied Biosystems, Foster City, CA).

Sequence analysis and mutation identification were performed by Sequencing Analysis software (Perkin-Elmer Applied Biosystems). For missense mutations, SIFT (Sorting Intolerant From Tolerant) and PolyPhen were adopted to assess their pathogenic possibilities [ 14 , 15 ]. The novel substitutions in or near the splicing junctions were evaluated for their potential functional role using NNSplice, SpliceView and NetGene2 software programmes [ 16–18 ].

All samples of the probands with potential pathogenic mutations were amplified by separate investigators and tested again for confirmation. Then the mutational sites were amplified and sequenced in other patients and normal family members.

Statistical analysis

Data are expressed as mean ± SD. The independent samples t -test was used to compare the differences between male and female patients’ onset age. A P-value <0.05 was considered significantly different. Statistical analysis was completed with SPSS ver. 13.0 software (SPSS Software, Chicago, IL).

Results

Mutations and polymorphisms identified in COL4A5 gene

Twenty-five mutations in 25 different families were considered to be pathogenic, with a mutation detection rate of 71.4% (25/35), showing 80% (24/30) and 20% (1/5) for male and female probands, respectively. These mutations included 1 nonsense mutation, 1 splicing mutation, 1 complex rearrangement mutation, 5 small deletions, 2 small insertions and 15 different missense mutations ( Table 1 ). None of the novel mutations were found in the control group.

Table 1.

Mutations identified in COL4A5 gene and patients’ major clinic data a

											Renal biopsy
	Mutations	Nucleotide change	FH	Exon	Gender	Age at study/onset	Haemat-uria	Protein-uria	Scr, μmol/L	Age at ESRD	LM	EM	α3,5 chains in GBM	Hearing loss	Ocular lesions
	Nonsense
AS-20/II-1	p.R1677X	c.5029C>T	Y	51	M	26/6	Mi+Ma	3.0 g	475	21				Y	NE
	Complex rearrangement
AS-5/III-1		c.2587_2606delins 75bp	Y	31	M	34/17	Mi	1.5 g	86		FSGS	VT		Y	NE
AS-5/II-1					F	54/36	Mi	2+	120					NE	NE
	Frame shift												(−)
AS-14/II-1	p.P263 fsX345	c.788delC	Y	14	M	20/11	Mi	NS	942	17	FSGS			NE	NE
AS-16/II-1	p.W1594fsX1638	c.4780_4781delTG	Y	49	M	15/6	Mi	3+	80		MPG		(−)	NE	NE
AS-21/I-1	p.G1412fsX1547	c.4235delG	N	46	M	5/4	Mi+Ma	2+	23		MCD			N	N
AS-30/II-1	p.W1648fsX1660	c.4942_4943insT	Y	50	M	7/2	Mi	1+	Nor		MPG			N	N
AS-30/I-1					F	29/16	Mi	1+	Nor					N	N
AS-31/IV-1	p.Q1016rfsX1020	c.3046delC	Y	35	M	9/1	Mi+Ma	1+	45		MCD	VT,L		Y	N
AS-31/III-3					F	34/20	Mi	1+	54					Y	N
AS-31/III-4					F	31/8	Mi	3+	91					NE	Y
AS-31/III-5					M	29/7	Mi	NS	168					Y	Y
AS-31/II-2					F	57/NK	Mi	1+	145					Y	Y
AS-31/II-3					F	55/18	Mi	1+	75					Y	Y
AS-33/II-1	p.G159fsX161	c.476delG	Y	9	M	18/6	Mi+Ma	4+	1469	14	MPG			Y	NE
AS-35/II-1	p.G192fsX203	c.573_574insGA	Y	10	M	5/3	Mi	2+	Nor		MCD	VT		N	N
AS-35/I-1					F	28/22	Mi	4+	76					NE	NE
	Splicing
AS-3/II-1		c.438+2T>C	Y	In7	M	28/24	Mi	2.9 g	204		FSGS	VT	(−) b	Y	NE
	Missense
AS-1/II-1	p.G875R	c.2623G>C	Y	31	M	17/6	Mi+Ma	NS	Nor		MPG		(−)	Y	NE
AS-2/III-1	p.G722R	c.2164G>C	Y	28	M	9/2	Mi	(−)	50					NE	NE
AS-2/II-1					F	41/37	Mi	(−)	Nor					N	NE
AS-2/II-2					F	37/33	Mi	(−)	72					N	NE
AS-4/II-1	p.G1427A	c.4280G>C	Y	46	M	17/12	Mi	(−)	Nor		FSGS		(−) b	NE	NE
AS-6/II-1	p.G1205S	c.3613G>A	Y	41	M	26/22	Mi	NS	789	25	FSGS		(−)	Y	N
AS-6/I-1					F	48/44	Mi	2+	73					NE	NE
AS-7/II-2	p.G869R	c.2605G>A	Y	31	M	21/18	Mi	2.8 g	77		FSGS	VT	(−)	N	NE
AS-7/II-1					F	26/20	Mi	(−)	Nor					NE	N
AS-9/II-3	p.G672S	c.2014G>A	Y	26	M	20/12	Mi	NS	126					N	Y
AS-9/I-1					F	50/26	Mi	1+	Nor					NE	NE
AS-9/II-1					F	30/21	Mi	(−)	Nor					NE	NE
AS-9/II-2					F	22/22	Mi	(−)	Nor					NE	NE
AS-9/III-1					M	6/6	Mi	1+	Nor					NE	NE
AS-11/III-1	p.G612D	c.1835G>A	Y	25	M	5/3	Mi	1+	50		MCD			NE	NE
AS-11/II-2					F	33/24	Mi	1+	Nor					NE	NE
AS-12/II-1	p.G1433V	c.4298G>T	Y	47	M	20/8	Mi	3+	Nor					N	N
AS-17/II-1	p.G426R	c.1276G>A	Y	20	M	45/26	Mi	1.6 g	129					N	NE
AS-18/III-5	p.R1677L	c.5030G>T	Y	51	M	24/13	Mi	NS	83		FSGS	VT,L		Y	N
AS-18/II-5					F	50/24	Mi+Ma	3+	72					Y	NE
AS-18/II-7					F	43/NK	Mi	2+	64					NE	NE
AS-18/III-1					F	41/10	Mi	3+	64					Y	NE
AS-18/III-2					F	31/NK	Mi	(−)	59					NE	NE
AS-18/III-4					F	27/NK	Mi	(−)	60					NE	NE
AS-18/IV-1					M	21/7	Mi	NS	1137	20				Y	NE
AS-19/III-1	p.G144R	c.430G>C	Y	7	M	34/29	Mi	NS	1391	31			(−)	Y	N
AS-19/II-2					F	58/NK	Mi	1+	85					NE	NE
AS-19/III-2					M	31/29	Mi	NS	85		FSGS	VT,L		Y	N
AS-19/IV-1					F	5/5	Mi	1+	42					NE	NE
AS-22/I-1	p.G828E	c.2483G>A	N	30	F	8/2	Mi+Ma	2+	47					NE	NE
AS-25/II-1	p.G582R	c.1744G>A	Y	24	M	22/18	Mi	NS	106		FSGS	VT		N	N
AS-28/I-1	p.G899V	c.2696G>T	N	32	M	35/26	Mi	3.2 g	674	30	FSGS		(−)	Y	Y
AS-29/II-1	p.T1480R	c.4439C>G	Y	47	M	23/21	Mi	NS	444	22	FSGS		(−)	N	NE
	Suspected
AS-34/II-2		Exons 2–10 not amplified	Y		M	37/7	Mi	NS	188		FSGS			Y	Y
AS-26/III-1		c.3017−4 A>G	Y		M	7/5	Mi	2+	Nor					NE	NE

a

NS, nephrotic syndrome; FH, family history; M, male; F, female; Y, yes; N, no; Mi, microscopic haematuria; Ma, macroscopic haematuria; NK, not known; NE, not examined; In, intron; VT, variable thickening; L, lamellation; Nor, normal.

b

Fluorescence intensity decreased obviously.

One nonsense mutation, changed from C to T of the first base of codon 1677 in exon 51, leads to premature stop codon in a male patient from family AS-20. This mutation may cause a truncated protein without the last two conserved cysteine residues in the non-collagenous (NC1) domain. A splice-site variation (c.438+2 T → C in Intron 7) was demonstrated in a male patient from family AS-3. The silico analyses showed that this variation may affect the splicing procedure and be pathogenic ( Table 3 ). Five small deletions (c.476delG, c.787delC, c.3046delC, c.4235delG and c.4778_4779delTG) and two small insertions (c.573_574insGA and c.4942_4943insT) were identified in 14 patients from seven different families. All theses mutations resulted in frame shift and anticipated stop codon. The products of the c.476delG, c.787delC, c.3046delC (sequence shown in Figure 1 ) and c.573_574insGA mutations were predicted to be truncated proteins without typical collagenous Gly-X-Y repeat and NC1 domain. The mutations of c.4235delG, c.4778_4779delTG and c.4942_4943insT may mainly affect the NC1 domain ( Table 1 ). One complex rearrangement mutation was found in a male proband and his mother from family AS-5. Twenty base pairs (c.2587_2606) in exon 31 were deleted and substituted by another 75 bp (sequence shown in Figure 2 ). Fifteen different missense mutations were found in the subjects. Thirteen of them showed glycine substitutions in sequences coding for Gly-X-Y repeats of the collagenous domain. Two missense mutations (c.4439C>G in exon 47 and c.5030G>T in exon 51) changed the highly conserved sequence in the NC1 domain. All of these missense mutations seem to be pathogenic by the analysis using prediction software SIFT and/or PolyPhen ( Table 2 ). In addition, exons 2–10 of one patient (AS-34/II-2) was unable to be amplified. Since he had a classical family history and symptoms, we hypothesized a large deletion existed ( Table 1 ). However, the lacking of a female carrier in this family made it difficult to confirm this large deletion by performing quantitative PCR on the female patients.

Fig. 1.

Pedigree diagram and sequence results of AS-31.

Fig. 2.

Pedigree diagram and sequence results of AS-5.

Table 2.

Effects of 15 missense mutations predicted by the software programmes a

Base change	Amino acid change	SIFT	PolyPhen
c.2623G>C	p.G875R	Affect	Unknown
c.2164G>C	p.G722R	Affect	Unknown
c.4280G>C	p.G1427A	Affect	PossDam
c.3613G>A	p.G1205S	Affect	Unknown
c.2605G>A	p.G869R	Affect	Unknown
c.2014G>A	p.G672S	Affect	Unknown
c.1835G>A	p.G612D	Affect	Unknown
c.4298G>T	p.G1433V	Affect	ProbDam
c.1276G>A	p.G426R	Affect	Unknown
c.5030G>T	p.R1677L	Affect	ProbDam
c.430G>C	p.G144R	Affect	Unknown
c.2483G>A	p.G828E	Affect	Unknown
c.1744G>A	p.G582R	Affect	Unknown
c.2696G>T	p.G899V	Affect	Unknown
c.4439C>G	p.T1480R	Tolerated	PossDam

a

PossDam, possibly damaging; ProbDam, probably damaging.

We also identified two novel variations with unknown pathogenesis in the splice junctions: c.3107−4A>G in Intron 35 (AS-26/III-1) and c.1948+13T>G in Intron 25 (AS-16/II-1). These two variations were not identified in 100 control subjects, but silicon analysis showed that they had little effect for the splicing procedure ( Table 3 ). Sixteen single nucleotide polymorphisms were also found ( Table 4 ).

Table 3.

Effects of three variations in the splicing junctions predicted by the software programmes

	NNSplice 0.9		NetGene2		SpliceView
Variations	Natural splice site	Mutant splice site	Natural splice site	Mutant splice site	Natural splice site	Mutant splice site
c.438+2T>C	0.97	Not detected	0.64	Not detected	81	Not detected
c.3017−4A>G	0.90	0.77	0.77	0.77	81	81
c.1948+13T>G	1.00	1.00	0.93	0.93	92	92

Table 4.

SNPs identified in COL4A5 gene

Nucleotide change	Position	SNP databases	Frequency (/35)	Protein change
c.1095G>C	E19	rs2272945	4	c.1095G>C (p.=)
c.1331T>G	E20	rs2272946	3	p.I444S
c.2055T>C	E27	rs7884085	4	c.2055T>C (p.=)
c.2349G>A	E29	rs3747408	4	c.2349G>A (p.=)
c.3513A>G	E39	rs2273051	2	c.3513A>G (p.=)
c.4275C>T	E46	rs61746140	4	c.4275C>T (p.=)
c.438+36G>T	In7	rs73526282	2
c.609+21T>C	In10	rs6622333	17
c.891+37A>G	In15	rs2294543	3
c.891+83_891+84isnACTT	In15	rs3215506	4
c.1339+181A>G	In20	NEW	3
c.1587+136A>G	In23	rs4308887	19
c.2509+50T>G	In30	rs2179674	4
c.2509+145T>A	In30	rs2064379	4
c.2768−11A>G	In32	NEW	11
c.3107−100G>A	In35	rs3761615	1
c.3246+67A>G	In36	rs3747409	4
c.3553+42T>C	In39	rs28465565	1

E, exon; In, intron.

Phenotypes in patients carrying COL4A5 mutations

Clinic data of 51 patients in 25 families carrying COL4A5 mutations are shown in Table 1 . In detail, the average onset age in 28 male patients were 12.3 ± 9.0 years old, which were younger than 18 female patients (21.6 ± 11.2 years, P = 0.004). Five female patients had not been identified until our family investigations. Twenty-two probands had a definitive family history, while the other three seemed to be sporadic. All the patients had microscopic haematuria, while only seven patients had gross haematuria (5 males and 2 females). Proteinuria was present in 44 patients, among them 11 (all males) patients had nephrotic syndrome. Eight male patients had ESRD. Eighteen patients showed hearing loss among 30 examined cases, and only six patients had ocular lesions among 19 examined cases ( Table 1 ).

Pathological change

Renal biopsy was performed in 19 cases, showed 4 minimal-change nephropathy (MCD), 11 focal segmental glomerulosclerosis (FSGS) and 4 mesangial proliferative glomerulonephritis by light microscope. Electron microscopy examination was done in eight patients, variable thickening of the GBM was observed in their renal specimens and GBM lamellation was found in three patients ( Table 1 ). Immunofluorescent staining of type IV collagen chains in GBM was tested in 10 male patients, which indicated that absence of α3,5(IV) chains was observed in eight of them, the dramatically decreased immune fluorescence intensity of α3,5(IV) chains was observed in 2 patients (AS-3/II-1 and AS-4/II-1). The absence of the α5 (IV) chain in the EBM was observed in all 28 male patients, and heterozygous loss of α5 (IV) chains was observed in all 23 female patients.

Discussion

To date, >400 mutations from different ethnic origins have been reported in AS, while the data from Chinese patients were insufficient considering its large population. We previously identified five novel mutations from 16 Chinese families by PCR–DGGE analysis and direct sequencing [ 13 ]. In this study, the mutation detective rate was 71.4% (25/35, two suspected mutations were not included), which is consistent with other studies based on genomic DNA [ 7–11 , 19–22 ]. Four of our 25 mutations (c.2605G>A, c.788delC, c.1276G>A and c.5029C>T) were reported previously, while the other 21 mutations are novel findings to the best of our knowledge [ 6 , 7 , 10 , 11 ].

The COL4A5 gene comprises 250 kb of genomic DNA and contains 51 exons, encoding a 6.5-kb transcript. Mutations were found to be spread along the entire gene and no hot spot was found. The 25 mutations identified in this study were located in 20 different exons, similar to the findings from other researches. Most of the reported mutations in the COL4A5 gene are small DNA lesions (nonsense, missense, small deletions or insertions), while gross rearrangements may account for 10–15% of cases (HGMD database). Based on our method, small nucleotide changes account for the major mutations and only one complex rearrangement mutation was found in the family of AS-5. In exon 31 of the male proband and his mother, 20 bp (c.2587−2606) were deleted and substituted by another 75 bp. We reviewed the sequence and found that it was been shifted in the COL4A5 gene complementary strand (c.2677+1203–2677+1277). Similar complex rearrangement mutation was rarely reported [ 22 ].

We also detected some nucleotide variations in COL4A5 gene, including six variations in the exons, such as c.2349G>A (p.=), c.1095G>C (p.=), c.2055T>C (p.=) c.3513A>G (p.=), c.4275C>T (p.=) and c.1331T>G (p.I444S). We believe that these variants are not pathogenic because their appearance rates are very high in our patients and other reports [ 7 , 8 , 21 ]. The variations of c.1948+13T>G and c.3107−4 A>G are both in the splicing junctions, which were not found in the 100 control subjects in our study. But they are not in the highly conserved splice signals ‘AG’ or ‘GT’, and the prediction software programmes showed they had little effect on the splicing junctions. So we deduced that their effect on splicing procedure should be confirmed on cDNA level.

The α5 chain contains 1685 amino acids, which were divided into a NC1 domain at the carboxy-terminal end, a central collagenous domain consisting of Gly-X-Y repeats and a short amino-terminal domain [ 23 ]. Mutations of gross rearrangement, nonsense, frameshift may result in the failure to synthesize complete α5 chains which were similar to one nonsense, one complex rearrangement and seven frameshift mutations in our study. Missense mutations, particularly replacement of a glycine residue in the Gly-X-Y repeat, were the most frequent mutant type, accounting for 52% (13/25) in our study. Such mutations may induce a kink in folding and forming of the collagen triple helix, which makes them exhibit increased susceptibility to photolytic attacking [ 24 ]. The amino acid of threonine in the site of 1480 and arginine in the site of 1677 are both highly conserved in the NC1 domain. Their changes may involve the NC1 function and interfere in the heterotrimer formation [ 25 ].

Type IV collagen consists of six chains (α1–α6) which form at least three types of heterotrimers in mammalian basement membranes. For example, (α1) ₂ α2 is found in all the basement membranes, α3–α4–α5 heterotrimer is the predominate component of GBM and (α5) ₂ α6 is present in Bowan’s capsule BM and EBM, but not in the GBM. The α5 chains are expressed in both GBM and EBM, which makes them suitable to detect XLAS patients by skin biopsy. In our experience of ∼1000 cases, skin biopsy is specific for the diagnosis of XLAS. Its sensitivity was reported in ∼75% [ 26 ]. In this study, we also examined 16 family members whose skin biopsies were normal but could be AS patients from pedigree analysis. However, results showed that no mutation was found. It did not mean the sensitivity of skin biopsy was nearly 100%. We considered this discrepancy to be the limited number of cases in our study.

Many researchers analysed the correlations between the phenotype and genotype of XLAS [ 8 , 21 , 27 , 28 ]. The majority of reports considered large rearrangements, premature stop, frameshift, donor splice site and mutations involving the NC1 domain to cause severe type with ESRD at ∼20 years of age; non-glycine XY-missense, glycine-XY mutations involving exons 21–47 may cause the moderate-to-severe type, while glycine-XY mutations involving exons 1–20 cause the moderate type with ESRD at ∼30 years. But for female patients, the situation was more complicated presumably because of the influence of random X-chromosomal inactivation [ 28 ].

Our results are generally based on the classification, which showed that five patients with nonsense (AS-21/II-1), frameshift (AS-14/II-1, AS-33/II-1), and NC1-domain (AS-18/IV-1, AS-29/II-1) mutations entered ESRD at ∼20 years old. This seemed to be earlier than the three patients with glycine-XY mutations involving exon 32 (AS-28/I-1), exon 41 (AS-6/II-1) and exon 7(AS-19/III-1) who entered ESRD at ∼30 years old. Hearing loss was observed in 72.2% (13/18) of the patients with the severe type, while only 41.7% (5/12) in the moderate-severe and moderate group. Ocular lesions were found in 36.4% (4/11) in the severe type, while 25% (2/8) in the moderate-severe and moderate type, respectively.

Many patients in our study were examined through family investigations. It was difficult to perform electric listening tests, fundus examination and renal biopsing for these patients. So our clinical data was relatively limited to building a conclusive correlation. More importantly, many patients in this study were children and had not yet developed renal failure, hearing loss or ocular lesions. Therefore, follow-up studies are critical for learning the phenotype–genotype correlations.

Due to the lack of messenger RNA (mRNA) samples in most patients, we performed direct sequencing and found mutations in 71.4% (25/35) of our patients, which is higher than our previous findings and some other studies by PCR–single-strand conformation polymorphism. This may possibly be due to the fact that our patients were all confirmed with α5 chain absence on EBM by skin biopsy, and other inherited nephritis including familial FSGS and IgA nephropathy can be excluded. Autosomal dominant and recessive AS caused by mutations in COLA3 and COLA4 gene can be excluded as well. But mutations in deep introns and large deletions, especially in female patients, could not be identified partially due to the technical limitations. Recently, several groups tried to detect mutations by reverse transcriptase-PCR using mRNA extracted from leukocytes, skin fibroblasts or hair roots [ 5 , 29–31 ]. They reported that the detection rate was slightly increased from 77% up to 84%. Some other new techniques such as semi-quantitative PCR and multiplex ligation-dependent probe amplification were also used for gene detection and the results were promising [ 32–34 ]. But direct sequencing is a reliable method, especially when only genomic DNA can be obtained.

In conclusion, we have identified 25 mutations, including 21 novel mutations in the COL4A5 gene in 35 Chinese families with XLAS confirmed by skin biopsy in this study.

We are grateful to the patients and their family members for their participation in the study. We thank Professor Bin Han in the National Gene Research Center for technical support. This study was supported by a grant from the National Natural Science Foundation of China (no.81070568), Leading Academic Discipline Project of Shanghai Health Bureau (05β001) and Shanghai Leading Academic Discipline Project (No.T0201).

Conflict of interest statement. We declare that we have no financial and 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, the manuscript entitled ‘Twenty-one novel mutations identified in COL4A5 gene in Chinese patients with X-linked Alport’s syndrome confirmed by skin biopsy’.

Author notes

*

These authors contributed equally to this work.