Hum Genet (2009) 125:401411

DOI 10.1007/s00439-009-0631-z

ORIGINAL INVESTIGATION

PHEX analysis in 118 pedigrees reveals new genetic clues

in hypophosphatemic rickets
Cline Gaucher Odile Walrant-Debray
Thy-Minh Nguyen Laure Esterle
Michle Garabdian Frdric Jehan

Received: 23 December 2008 / Accepted: 23 January 2009 / Published online: 15 February 2009
Springer-Verlag 2009

Abstract Familial hypophosphatemic rickets is a rare disease, which is mostly transmitted as an X-linked dominant
trait, and mutations on the phosphate regulating gene with
homologies to endopeptidases on the X-chromosome
(PHEX) gene are responsible for the disease in most familial
cases. In this study we analyzed PHEX in a large cohort
of 118 pedigrees representing 56 familial cases and 62 sporadic cases. The high-resolution melting curves technique
was tested as a screening method, along with classical
sequencing. PHEX mutations have been found in 87% of
familial cases but also in 72% of sporadic cases. Missense
mutations were found in 16 probands, two of which being
associated with other PHEX mutations resulting into truncated proteins. By plotting missense mutations described so
far on a 3D model of PHEX we observed that these mutations focus on two regions located in the inner part of the
PHEX protein. Family members of 13 sporadic cases were
analyzed and a PHEX mutation was detected in one of the
apparently healthy mother. These results highlight the
major role of PHEX in X-linked dominant hypophosphatemic rickets, and give new clues regarding the genetic

Electronic supplementary material The online version of this

article (doi:10.1007/s00439-009-0631-z) contains supplementary
material, which is available to authorized users.
C. Gaucher (&) O. Walrant-Debray T.-M. Nguyen
L. Esterle M. Garabdian F. Jehan
Inserm U561, Hpital Saint-Vincent-de-Paul,
82 avenue Denfert-Rochereau, 75014 Paris, France
e-mail: celine.gaucher@parisdescartes.fr
C. Gaucher
EA2496, Dental School,
Universit Paris-Descartes, Paris, France

analysis of the disease. A screening of the diVerent family

members should be mandatory when a PHEX mutation is
assessed in a sporadic case and the search for another
PHEX mutation should be systematically proceed when
facing a missense mutation.

Introduction
Familial hypophosphatemic rickets was Wrst described by
Albright et al. (1937) as a vitamin D resistant rickets. Later
on, several forms of hypophosphatemic rickets have been
identiWed depending on their inheritance (autosomal or
X-linked, dominant or recessive), of which the most frequent form is the X-linked dominant disease (Winters et al.
1958). In 1995 the PEX gene, latter renamed phosphate
regulating gene with homologies to endopeptidases on the
X-chromosome (PHEX), was identiWed (OMIM-300550) as
the main cause of the X-linked dominant hypophosphatemic rickets (OMIM-307800) (HYP Consortium 1995).
Afterwards, FGF23 was identiWed as the cause of the
Autosomal Dominant Hypophosphatemic Rickets (ADHR
Consortium 2000) and DMP1 (Lorenz-Depiereux et al.
2006a, b) and SLC34A3 (Lorenz-Depiereux et al. 2006a, b)
as causes of recessive autosomal hypophosphatemic
rickets. Klotho gene may also be involved, as hypophosphatemic rickets has been observed in a child with a translocation causing an increase in
-Klotho levels (Brownstein
et al. 2008).
The phenotype of X-linked dominant hypophosphatemic
rickets, although variable in its expressivity, is characterized
by: rickets with bone deformities, short stature, dental
anomalies, and at the biological level, hypophosphatemia
with low renal phosphate reabsorption, normal serum calcium level with hypocalciuria, normal or low serum level

123

402

of 1,25-(OH)2D3, normal serum level of PTH and increased

activity of serum alkaline phosphatases (Thakker and
ORiordan 1988).
The PHEX gene is located on chromosome Xp22.1
22.2 in humans. The coding region spans 2,250 bp (22
exons) and encodes for a 749 amino acid transmembrane
endopeptidase that shares signiWcant homologies with
the members of the Neprilysin (M13) zinc-metallopeptidases family (namely Neprilysin, Kell antigen and
Endothelin Converting Enzyme 1 and 2). The Neprilysin
(M13) zinc-metallopeptidases proteins are characterized
by a short intra-cytoplasmic domain, a single transmembrane hydrophobic domain and a large extracellular
region containing ten conserved cysteine residues. SpeciWc domains are also highly conserved: a zinc-binding
motif (HxxTH), a co-ordinating zinc binding site (GENIAD), and a VNAxY motif, in contact with the subtrate
peptides (Francis et al. 1997; Turner and Tanzawa 1997;
Bland et al. 2008). PHEX is mainly expressed in osteoblasts, osteocytes and odontoblasts (Ruchon et al. 2000;
Thompson et al. 2002). It has been shown to interact,
directly or not, with the phosphatonins matrix extracellular phosphoglycoprotein (MEPE) and FGF23 (Rowe
2004). Yet, PHEX substrates have not been clearly
identiWed.
In the seven studies made since 1995 on large cohorts of
hypophosphatemic patients, PHEX mutations were found in
41 to 95% of the cases with a similar rate of mutation
observed overall in familial cases, 58% of 224 patients, and
in sporadic cases, 50% of 78 patients (Francis et al. 1997;
Holm et al. 1997; Rowe et al. 1997; Dixon et al. 1998;
Tyynismaa et al. 2000; Cho et al. 2005; Ichikawa et al.
2008). So far, more than 200 diVerent mutations have been
described, most of them being reported in the PHEX
database (http://www.phexdb.mcgill.ca). Small deletions,
small insertions, nonsense mutations, missense mutations
as well as splice site mutations are spread all along the 22
exons, including the 3-UTR region in the 22nd exon
(Ichikawa et al. 2008). Large deletions (Francis et al. 1997;
Holm et al. 1997; Dixon et al. 1998), pseudo-exons
(Christie et al. 2001) or mosaic (Goji et al. 2006) have also
been identiWed.
In this article, the PHEX gene was analyzed using classical sequencing in 209 patients with hypophosphatemic
rickets, representing 118 families, and high-resolution
melting curves analysis was evaluated as a possible screening method. Our results point out the frequent involvement
of PHEX in familial cases (87%), but also in sporadic cases
(72%) and in late onset cases, and the necessity of a systematic screening of the probands parents, even when
asymptomatic. They show the utility of a complete
sequencing of the 22 PHEX exons when a missense mutation has been found.

123

Hum Genet (2009) 125:401411

Patients and methods

Patients
209 patients (135 females and 74 males) have been
included in this study, representing 118 probands divided in
56 familial cases and 62 sporadic cases. Most probands
(n = 106) have European origin, 10 have North African origin, 2 have Caribbean origin and one was from Asia. Most
patients had an early onset of the disease (before 5 years of
age) but three patients had a late onset (from 10 to 12 years
of age). Inclusion criteria were hypophosphatemia associated with tubular phosphate wasting (low serum phosphate
and low TmP/GFR), bone deformities and radiological
signs of rickets. Exclusion criteria were the presence of
tubulopathy, hypercalciuria, or hyperparathyroidism before
the onset of treatment. The dominant inheritance was established in all familial cases except in two in which brothers
and sisters had hypophosphatemia but their parents were
asymptomatic. The relatives of sporadic probands for
whom we had DNA samples were included in the study to
assess the de novo status of the mutation. All patients had
signed an informed consent to the genetic study in accordance with the local ethic committee.
DNA analysis
Genomic DNA was prepared from blood using commercial
extraction kits. The 22 exons with their adjacent intronic
sequences and the 3UTR region Xanking the polyadenylation site were ampliWed by polymerase chain reaction
(PCR). Primers (Invitrogen or Operon) were designed
(Electronic supplementary Table online) from the PHEX
gene sequence (RefSeq NM_000444) to perform PCR
ampliWcation of at least 50 bp of intron sequence downstream and upstream of exons. PCRs were performed on a
Crocodile III thermocycler (Appligene) or a Light Cycler
LC480 (Roche Diagnostics corporation) with 0.5 U of Taq
DNA Polymerase (QBiogene) and its Incubation Mix TPol
buVer. PCR conditions were as followed: 3 min at 95c then,
35 cycles at 95C, 30 s denaturation; 52C, 20 s annealing;
72C, 35 s extension; and 5 min at 72C. Electrophoresis of
the PCR products were performed on 1% agarose gels and
puriWcation on a Chargeswitch PCR Clean-up kit (Invitrogen). All sequencings were performed with internal primers (Electronic supplementary Table online) and carried out
by the Institut Fdratif de Recherche 116 Sequencing
Facility (IFR116, Paris, France).
High resolution melting (HRM) curves analysis
Polymerase chain reaction ampliWcations followed by
HRM were performed on 11 exons using the LightCycler

Hum Genet (2009) 125:401411

403

Table 1 PHEX mutations found in this study

Exon

cDNA (NM_000444)a

Protein

Mutation type

Inheritanceb

Referencesd

c.20_21insAG

p.Ser7ArgfsX24

Frameshift

c.117delA

p.Leu39LeufsX1

Frameshift

c.142C>T

p.Gln48X

Nonsense

c.230C>T

p.Cys77Ser

Missense

Fc

c.254G>T

p.Cys85Phe

Missense

c.293_294 ins GAAGA

p.Met98MetfsX10

Frameshift

c.349+2T>C

NA

Splice site

c.350-1G>T

NA

Splice site

c.413T>C

Leu138Pro

Missense

1, 3

c.415delT

p.Tyr139IlefsX5

Frameshift

10

c.436+4A>C

NA

Splice site

11

c.437-3C>G

NA

Splice site

12

c.538delT

p.Trp180GlyfsX52

Frameshift

13

c.621T>A

p.Tyr207X

Nonsense

14

c.[=; 665T>C]

p.Leu222Pro

Missense (mosaicism)

15

c.732+1_732+2delGT

NA

Splice site

16

c.733-1G>T

NA

Splice site

19

c.849+1G>T

NA

Splice site

17

c.849+6 insT

NA

Splice site

18

c.871C>T

p.Arg291X

Nonsense

F, F,S

4, 5, 6, 7

20

c.931C>T

p.Gln311X

Nonsense

Fc

21

c.979_980insCT

p.Tyr327LeufsX5

Frameshift

22

c.1002insG

p.Pro335AlafsX13

Frameshift

23

c.1036T>G

p.Tyr346Asp

Missense

24

c.1042A>T

p.Lys348X

Nonsense

25

c.1078_c.1079+2delAAGT

NA

Splice site

26

10

c.1092C>G

p.Asn364Lys

Missense

27

10

c.1092C>A

p.Asn364Lys

Missense

28

10

c.1109_1110ins GAA

p.Met730_Val371insLys

In frame insertion

29

10

c.1152T>G

p.Tyr384X

Nonsense

S, F, S

30

11

c.1185delG

p.Gly395GlyfrX12

Frameshift

31

11

c.1208G>A

p.Trp403X

Nonsense

1, 8

32

11

c.1263del G

p.Met421IlefsX2

Frameshift

33

12

c.1319_1321 delAGG

p.Glu440del

In frame deletion

34

12

c.1403delA

p.Lys468ArgfsX15

Frameshift

35

13

c.1434T>A

p.Tyr478X

Nonsense

Fc

36

13

c.1471del G

p.Asp491ThrfsX23

Frameshift

37

13

c.1482+2delAAinsT

NA

Splice site

38

14

c.1522C>T

p.Gln508X

Nonsense

39
40

Mutation #

14

c.1586+3_1586+6del GAGT

NA

Splice site

S, F

3, 4, 8

14

c.1586_c.1586+1delAG

NA

Splice site

41

15

c.1601C>T

p.Pro534Leu

Missense

1, 5, 6, 8, 10, 11

42

15

c.1645C>T

p.Arg549X

Nonsense

F, Fc

1, 8

43

15

c.1645+1G>A

NA

Splice site

S, F, S

4, 9

44

15

c.1645+5G>A

NA

Splice site

45

16

c.1683G>A

p.Trp561X

Nonsense

46

16

c.1699C>T

p.Arg567X

Nonsense

9, 11

47

16

c.1700+1G>A

NA

Splice site

48

123

404

Hum Genet (2009) 125:401411

Table 1 continued
Exon

cDNA (NM_000444)a

Protein

Mutation type

Inheritanceb

Referencesd

Mutation #

17

c.1735G>A

p.Gly579Arg

Missense

F, F, S, S

1, 5, 6, 8, 9

49

17

c.1765_1768delAATG

p.Asn589ValfsX29

Frameshift

50

17

c.1768+2 insT

NA

Splice site

51

18

c.1779T>A

p.Tyr593X

Nonsense

52

18

c.1782_1783insTGAT

p.Lys595X

Nonsense

S, F

53

18

c.1806G>A

p.Trp602X

Nonsense

54

18

c.1843insA

p.Thr615AsnfsX5

Frameshift

55

18

c.1857_1858insGATT

p.Asn620AspfsX1

Frameshift

56

18

c.1895_1899+18

NA

Splice site

57

F, S

58

delTAAATGTGAGTACAACTGTGGCT
18

c.1899+1G>A

NA

Splice site

18

c.1899+5G>A

NA

Splice site

59

19

c.1919T>C

p.Leu640Pro

Missense

60

19

c.1929_1930insAAAT

p.Ile644LysfsX4

Frameshift

61

19

c.1953_1954insGAAGCTTGG

p.Arg651_
Glu652insGluAlaTrp

In frame insertion

62

19

c.1965+1G>A

NA

Splice site

3, 11

63

20

c.1970A>G

p.Tyr657Cys

Missense

64

20

c.1989_1990delCA

p.Asp663GlufsX5

Frameshift

65

20

c.2018_2036
delTACCAGGCATCACATTCAC

p.Leu673ProfsX7

Frameshift

66

20

c.2040_2041delCA

p.Asn680LysfsX41

Frameshift

67

20

c.2063_2064insGTTA

p.Tyr688X

Nonsense

68

21

c.2071-2A>C

NA

Splice site

69

21

c.2092-2093insAGAC

p.Pro698GlnfsX18

Frameshift

70

21

c.2104C>T

p.Arg702X

Nonsense

S, F

1, 6

71

21

c.2138_2139ins C

p.Pro713ProfsX4

Frameshift

72

22

c.2147-1G>T

NA

Splice site

73

22

c.2150T>A

p.Val717Asp

Missense

74

22

c.[=; 2155G>T]

p.Gly719Cys

Missense
(mosaicism)

75

22

c.2160_2175

p.Ala720AlafsX14

Frameshift

76

delAATTAGTAACTTTGAA
22

c.2171_2172insAACT

pPhe724X

Nonsense

77

22

c.2239C>T

p.Arg747X

Nonsense

S, S

3, 5, 6, 8, 10

78

NA for not applicable

a
Nucleotide +1 corresponds to the A of the ATG initiation codon reported in Genbank (accession number NM_000444)
b
F is for familial inheritance S for sporadic case. Each letter represents a distinct proband
c
Mutation of this patient has been reported previously in Rowe et al. 1997
d
References are the following: 1, Rowe et al. 1997; 2, Tyynismaa et al. 2000; 3, Filisetti et al. 1999; 4, Holm et al. 1997; 5, Francis et al. 1997; 6,
Dixon et al. 1998; 7, Sato et al. 2000; 8, Popowska et al. 2000; 9, Ichikawa et al. 2008; 10, Cho et al. 2005; 11, Song et al. 2007

480 (Roche Diagnostics Corporation). For male patients,

who possess only one copy of PHEX (1 X-chromosome),
we added equivalent amount of genomic DNA from a male
control (1 X-chromosome) indeed not mutated in the 11
exons (conWrmed by sequencing). DNA samples were distributed on 96-well plate and each 10
l reaction contained
0.5
l DNA (100 ng/
l), 1
l of each forward and reverse

123

primers (5
M), 1.7
l of sterile water, and 5
l of HRM

Master Mix (Roche Diagnostics Corporation) containing
Fast Start DNA Polymerase, reaction buVer, dNTPs and
HRM Dye. The PCR program was: initial denaturation
95c 10 s, followed by 40 cycles of: 10 s denaturation at
95C, 15 s step-down annealing 58 to 51C, (0.8C stepsize) and 15 s extension at 72C. HRM consisted of an

Hum Genet (2009) 125:401411

405

Table 2 Mutations frequencies

in reports with large cohorts

Proband:
origin

Mutations
(total) (%)

Rowe et al. (1997)

Eur/N-Afr
N = 106

55

99

56

28

Francis et al. (1997)

European
N = 43

74

29

82

14

57

Dixon et al. (1998)

European
N = 68

45.5

46

52

22

32

This study

Eur/N-Afr
N = 118

79

54a

91

60a

73

Four pedigrees have not been

included in this table because
dominant transmission could not
be proven in 2 familial cases and
no family history was collected
in 2 sporadic cases

initial denaturation at 95C followed by 1 min of cooling

down to 40C. Then, continuous Xuorescent acquisition (25
acquisitions per degree Celsius) was perform during the
melting phase starting from 40 to 95C. Melting curves were
analyzed with the LC480 Gene Scanning software 3.5
(Roche Diagnostics Corporation). All PCR products obtained
through HRM experiments were puriWed, and sequenced
with internal primers (Supplementary Table online).
Genotyping
The c.2239C>T (p.Arg747X) variation frequency was
assessed using Xuorescent hybridization probe melting
curves in a control cohort of 350 healthy children with no
known alterations of phosphate metabolism. PCR ampliWcation was performed on the LC480 apparatus (Roche
Diagnostics Corporation) using the Roche Genotyping mix.
Probes and speciWc primers were designed by Tib
MolBiol: forward PCR primer 5-TCAGGGTCAATGG
GCAATTAGTA-3; reverse PCR primer 5-AAATGAAA
GTCTCCAGGCCTA; hybridization probes: 5-CAGCTA
CCAGAGTCAGCAGGA-FL and 5-LCRed640-TCCATG
CCTCTGTTCATCGTGGAA-PH. The PCR conditions
were: initial denaturation 95C, 10 min; then 40 cycles of
denaturation 95C, 10 s, annealing 48C, 15 s, extension
72C, 20 s. The melting curves were performed after denaturation at 95C (1 min) followed by a rapid cooling down
to 37C and a continuous increase in temperature up to
70C (5 Xuorescence acquisition per degree Celsius. All
curves were analyzed with the LC480 Genotyping software
3.5.

Results
Mutation screening
Among the 118 probands, 93 patients (79%) displayed a
PHEX mutation (Table 1) representing 78 diVerent mutations. To the best of our knowledge, 60 of them have not
been yet reported. When dividing the 116 probands with a

Familial
cases N

Mutations
(%)

Sporadic
cases N

Mutations
(%)

known familial history between familial (n = 56) and sporadic (n = 60) cases, PHEX mutations were found respectively in 49 (87.5%) and 44 (73%) of the cases. These
percentages are higher than those found in the three largest
reported cohorts, especially for sporadic cases (Table 2).
The repartition of the diVerent types of mutation in our
cohort is: 28% of point mutations leading to a nonsense
mutation, 30% of insertion/deletion (ins/del) leading either
to a frameshift or to an inframe ins/del of amino acid, 23%
of mutations (point mutations or ins/del) leading to disrupted splice sites, and 19% of point mutations leading to a
missense mutation. These percentages are similar from
those reported in the PHEX database: 18, 42, 19 and 21%
respectively.
High resolution melting (HRM) curve analysis
As an alternative to classical PCR followed by sequencing,
which we have performed on all genomic DNA samples,
we have also tested another approach of mutation detection,
the high resolution melting curve analysis (HRM) available
on the LC480 apparatus (Roche). This has been done in a
subset of 66 patients for 11 exons (exons 6, 10, 13, 14, 15,
16, 17, 19, 20, 21, 22). In parallel and in a blind manner,
all PCR products were analyzed by classical sequencing
in order to evaluate the relevance of this new method.
For male patients, possessing only one copy of PHEX
(1 X-chromosome), we added an equivalent amount of
genomic DNA from a male control with no mutation in
analyzed exons.
A total of 726 PCR reactions were analyzed by HRM.
Figure 1 shows two representative examples of HRM
analysis, performed in exon 15 (Fig. 1a) and 20 (Fig. 1b). The
HRM protocol allowed the Wnding of 23 mutations (conWrmed by classical sequencing), 4 false negatives (mutations not detected by HRM), 138 false-positives (suspected
mutations) and 559 true-negatives. Among the 4 false negatives, there were 3 point-mutations and 1 del/ins mutations.
For the HRM method, we determined a sensibility and a
speciWcity of respectively 0.85 and 0.79; and positive and
negative predictive values of respectively 0.14 and 0.99.

123

406

Hum Genet (2009) 125:401411

Fig. 1 Mutations detection by HRM. Mutations detection by HRM

was performed with the LC480 Gene Scanning software 3.5 (Roche
Diagnostics Corporation). HRM was performed following PCR on the
LC480 apparatus with primers described in the Electronic supplementary Table online. A HRM detection in exon 15. Two DNA samples

with true single nucleotide mutations are indicated as well as false positive samples. B Labeled curves correspond to mutated DNA in HRM
analysis of exon 20. Mutation c.1970A>G was not detected by HRM
analysis and was put in the false negative group

Mutation analysis

2000; Cho et al. 2005) (Table 2). Because of its high frequency among hypophosphatemic patients, we tested the
possibility of a polymorphism. The screening of 350
healthy children (200 males and 150 females, representing
500 chromosomes) in a control cohort showed no variation.
We discovered 16 missense mutations, 8 being new
(Table 1). When plotting all reported missense mutations
including ours (n = 42) on PHEX amino acid sequence
alignment, missense mutations occur predominantly on
residues strictly conserved in mammals with only one
exception, and in chicken with two exceptions (Electronic
supplementary Figure online).

Among the 93 diVerent mutations, we found 35 small insertions or deletions ranging from 1 to 23 bp. Among these
mutation, 24 led to a frameshift mutation, 3 to an inframe insertion or deletion, and 8 to a splice site disruption, (Table 1). The in-frame mutations generated a 1amino acid insertion (p.Met730_Val371insLys), a 1-amino
acid deletion (p.Glu440del) and a 3-amino acid insertion
(p.Arg651_Glu652insGluAlaTrp) respectively. Fourteen
other splice site disruptions were generated by single nucleotide variations. Most of the 22 disrupted splice sites occur
in the 3-donor site (only 3 were in the 5-acceptor site) and
were located within the 3/+5 bp regions of the splice
sites, and therefore are likely to be the cause of the disease,
according to the Information for the Coordinates of Exons
(ICE, Chong et al. 2004). However one patient displayed a
more distal disruption further in the nucleotide sequence:
c.849 + 6insT (Table 1). The ADN of this patient was
sequenced in the 22 exons and the 3UTR region conWrming the absence of additional PHEX mutation. Interestingly,
this patient had a late onset of the disease.
We found 20 diVerent nonsense mutations leading to a
truncation of the PHEX protein and generating PHEX fragments ranging from a 48 (p.Gln48X) to a 747 amino acid
molecule (p.Arg747X). The latter mutation has been
already described in several studies (Francis et al. 1997;
Dixon et al. 1998; Filisetti et al. 1999; Popowska et al.

123

Mutation distribution in the PHEX gene

The 78 diVerent mutations found in our cohort are
distributed along the 22 exons or their adjacent intronic
sequences, and no mutation was found in the 3-UTR
region, or in the part of 5-UTR region we studied. We have
analyzed mutation distribution along the PHEX coding
sequence (2,250 bp including stop codon) where intronic
mutations were attributed to the closest nucleotide of the
nearest exon. For both the PHEX database and our cohort,
we have calculated a rolling average of mutation frequency
in a 100 bp-window, with a 20 bp step increment (Fig. 2).
As noticed previously (Filisetti et al. 1999), mutations are
unevenly distributed along the PHEX gene. Three obvious
regions with high density of mutations were observed in

Hum Genet (2009) 125:401411

20

Mutation frequency per 100bp

Fig. 2 Mutation repartition

along the PHEX gene. All PHEX
mutations, from our study or
from the PHEX database, were
plotted as a rolling average of
mutation frequency in a 100 bpwindow with a 20 bp step increment along the 2,247 bp of the
coding region. Intronic mutations aVecting splice sites have
been connected to the closest
base of the closest exon. The 22
coding exons (tube) are shown
with proportional size. Cysteine
residues and motives implicated
in zinc binding and catalytic
activities are shown under the
3D bar representing the PHEX
amino acid sequence

407

15
PHEX database
This report

10

0
0

400

800

1200

1600

2250

2000

5
1 2

10

11

12

13

14 1516 17

18

19 20

21 22

Cysteine residue

NH2

VNAFY HEFTH

GENIAD

Mutation positions:
In PHEX database
In this report

400

800

1200

1600

2000

2250 bp

Mutation location in PHEX cDNA (bp)

our cohort and were similarly found when analyzing mutations from the PHEX database (Fig. 2).
On a three-dimension (3D) model of the PHEX protein,
we analyzed all PHEX mutations (presently observed and
previously reported) leading to one amino acid variation in
the protein sequence, including missense mutations and the
few in-frame insertions or deletions of one amino acid
(Fig. 3). This 3D model of PHEX (Modbase, Pieper et al.
2004) is based on the crystal structure of Neprilysin (ref,
PDB reference 1dmtaA in the RCSB protein databank) the
closest homolog of PHEX in the M13-peptidase family. In
this model, like in the Neprilysin crystal structure, amino
acids 154 comprising the intracellular and transmembrane
regions are absent. Among the ten cysteine residues highly
conserved in mammals seven have been subjected to missense mutations in the PHEX protein (Electronic supplementary Figure online). Because all cysteine residues of
PHEX are likely to be involved in stabilizing the protein
structure and because disruption of disulWde bonds would
alter the 3D structure of the protein, the missense mutations
involving cysteine residues and the three missense mutations adjacent to cysteine residues were not analyzed on the
3D model. The remaining missense mutations are located
in the inner part of the cocoon-shaped molecule, with two
dense regions at each extremity of PHEX (Fig. 3a). The
Wrst one contains 25 missense mutations and surrounds
the proteolytic domain well characterized in NEP, with the
GENIAD, VNAFY and HExxE motives highly conserved
in the M13 peptidases family (Oefner et al. 2000; Bianchetti et al. 2002) (Fig. 3c). The second one regroups 16
other missense mutations on the other side of the protein

(Fig. 3b). In contrast, nonsense mutations plotted on the

same 3D model of PHEX are randomly located (data not
shown).
PHEX variants not causing the disease
Splice site disruption, nonsense or in/del are likely to be the
cause of the disease but question remains as to the relationship between single nucleotide replacements and the occurrence of the disease. We were faced with 3 diVerent
situations in our cohort
1. We found a mutation, c.1206A>G, which is silent for
glutamine 402 (Glu402Glu). But the relationship
between this mutation and the disease is not clear as the
patient also displays a missense mutation c.1735G>A
(Glu579Arg). The c.1206A>G silent mutation has not
been found in any other patient, and has never been
described in the literature.
2. We found a c.505G>A variant, which transforms a
valine into a methionine in position 169 (V169M) in 2
diVerent pedigrees. But, in each pedigree, another
PHEX mutation was found. One patient of the Wrst
family also carries a c.20922093ins mutation leading
to a frameshift (p.Pro698GlnfsX18); a patient of the
second family carries both mutations c.505G>A and a
c.538delT on the same exon, the latter mutation leading
also to a frameshift (p.Trp180GlyfsX52). In this family, the two aVected persons carry the c.538delT frameshift mutation along with the V169M missense
mutation while three asymptomatic members of the
family are carrying the V169M mutation alone (Fig. 4).

123

408

Hum Genet (2009) 125:401411

Fig. 3 Location of missense

mutations in a 3D PHEX model.
Coordinates of the 3D structure
of the human PHEX protein
were obtained from Modbase
(modbase_model_98a03d65
16af4f7e7433f926aa7be1a1;
http://modbase.compbio.
ucsf.edu). All missense mutations described so far, at the best
of our knowledge, were visualized on the 3D structure using
MacPyMol 0.99rc6

c.538delT mutation
V169M mutation
I-1

II-1

I-2

II-2

II-3

II-4

3. We analyzed PHEX in both parents in 13 of the apparent sporadic cases to conWrm the de novo appearance
of the mutation. But we found evidence for a maternal
transmission in one case (Y688X). This mother displayed no evident clinical or biological signs of hypophosphatemic rickets: no history of bone deformities in
childhood, height and serum phosphorus in the lower
range, and good general health.
Polymorphisms

III-1

III-2

Fig. 4 Genealogical tree of the family sharing the c.538delT and

V169M variants Only patients II-3 and III-2 (arrows) have hypophosphatemic rickets

Of interest, analysis of the family tree shows that both

mutations are carried by the same allele.
As a consequence of these results, all probands displaying a
missense mutation have been sequenced for the 22 exons,
including the 3UTR region of exon 22, to verify they were
not carrying any other PHEX mutation.

123

By screening exonic (including the 3-UTR and part of the

5-UTR sequences) and adjacent intronic sequences, we have
found 6 polymorphisms, all located in introns. One of them is
not reported in the SNP database (Table 3) and has an allele
frequency of 3.5% in our cohort. There is reported frequency
data for four of the other observed SNPs (http://www.ncbi.
nlm.nih.gov/SNP; Table 3), and their allele frequencies in
out cohort are similar to those described. We have found no
report of frequency data for the SNP rs60807057, which is a
poly (A) variation number (with 12 or 10 adenine residues).
An allele frequency of 28% is found for the (A)10 repeat and
72% for the (A)12 repeat in our cohort (Table 3).

Hum Genet (2009) 125:401411

Table 3 PHEX polymorphisms and their frequencies in
this study versus in the European
population

409

Location

5UTR

Intron 2

Intron 7

Intron 8

Intron 17

Intron 18

c.133

c.188-47

c.849+138

c.934+46

c.1769-10c

c.1900-20

Base change

c>t

c>t

g>a

a>g

>t

del tt

SNP number

rs5951494

rs178720

rs6629449

rs3752433

rs60807057

In our cohort

10%

20%

2.5%

3.5%

30%

28%

In European population

8%

12.520.8%

6.7%

25.037.5%

Allele frequency

Patients negative for PHEX mutation

Among the 16 sporadic cases negative for PHEX mutation
there were 6 males and 10 females. In one male patient we
failed to obtain any PCR products for two consecutive
exons, and another one was found to bear a FGF23 mutation (R179W) (data not shown). Two of the female had a
late onset of the disease.
Five familial cases, all of them being mother to daughter
transmissions, are negative for PHEX mutations.

Discussion
The present genetic study of a large cohort of hypophosphatemic patients (209 patients representing 118 pedigrees)
highlights the major role played by mutations in the PHEX
gene as the cause of familial hypophosphatemic rickets.
PHEX mutations were found in 79% of all probands, a frequency similar to that reported by Francis et al., 74% of 43
pedigrees (Table 2), but higher than those reported in the
two other large cohorts reported so far, 45 and 55% of 68
and 106 pedigrees, respectively (Rowe et al. 1997; Dixon
et al. 1998). In the subset of familial cases showing a
dominant inheritance pattern compatible with X-linked
transmission (n = 54), we found PHEX mutations in 91% of
the probands. PHEX mutations were also found in the two
families with no evidence of dominant transmission (several patients among brothers and sisters but asymptomatic
parents). Interestingly, the Wve familial cases with no mutation were all females with mother to daughter transmission,
and thus, neither the hypothesis of large deletions nor
mosaic mutation in the PHEX gene can be excluded.
Pseudo-exons as well could explain a part of these negative
familial cases. In the subset of sporadic cases (n = 60),
73% had PHEX mutations. This conWrms that PHEX mutations are not restricted to familial cases compatible with
X-linked dominant inheritance, although the percentage of
patients with no mutation found in the PHEX gene still
remains higher for sporadic cases than familial cases (27
vs. 12.5%). Among the 16 negative sporadic cases, one
male is suspected to carry a large deletion covering more
than 2 exons, one boy was found to bear a FGF23 mutation,

already described (ADHR Consortium 2000), and one girl

was later diagnosed as presenting a tumor-induced osteomalacia. Thus, 23% of the sporadic cases remain unexplained, leaving the possibility of other causal genes in
which a mutation with recessive inheritance could be
responsible for hypophosphatemic rickets. This frequency
of sporadic cases with no PHEX mutations is much lower
than those reported earlier, from 43 to 72%, and points out
PHEX as the major genetic cause of hypophosphatemic
rickets, even in sporadic cases. Moreover the evidence of a
mother-to-son transmission in one sporadic case is an
important issue for genetic counseling, and parents from
sporadic cases should be systematically screened for the
PHEX gene even if they display no classical sign of hypophosphatemic rickets. Our Wndings also point out the possible involvement of PHEX mutations in late onset cases,
since one out of these 3 patients had a mutation in an
intronic region of PHEX 2 bp away from the splice site
consensus sequence. Along this line, a L555P missense
mutation was found in a large family with adult-onset
hypophosphatemia (Econs et al. 1998). Therefore the possibility of a PHEX mutation should be explored in all patients
with hypophosphatemic rickets, including those with a late
onset of the disease, in as much as such Wndings would alleviate the long-term search for tumors in these latter
patients.
The high percentage of PHEX mutation in the whole
cohort, and particularly in the sporadic population, may
be due in part to the selection of patients based on a thorough screening with the following exclusion criteria: associated signs of tubulopathy, such as aminoaciduria,
proteinuria, -microglobulinuria, that is found in patients
with CLCN5 mutations (Lloyd et al. 1996), and/or hypercalciuria, that is found in patients with mutations of kidney
ionic transporters or channels like CLCN5 and NaPiIIc
(Bergwitz et al. 2006; Lorrenz-Depiereux et al. 2006b); and
hyperparathyroidism before initiation of treatment, which is
found in patients with DMP1 mutations or translocation in
the vicinity of the Klotho gene.
If mutations disrupting the overall integrity of the PHEX
protein (nonsense, ins/del leading to frameshift, and splice
site disruption) are likely to be the cause of the disease, the
role played by the other types of mutations is more

123

410

questionable, namely missense mutations, ins/del of one or

few amino acid residues, or mutations not strictly located in
donor-acceptor consensus sequence splice sites. The following observations support the hypothesis that at least
several of the observed missense mutations induce a loss of
function of the PHEX protein and, consequently, the hypophosphatemic rickets: (1) all missense mutations described
in this study and in the PHEX database involve residues
highly conserved among mammals; (2) some missense
mutations occur relatively frequently in unrelated pedigrees, for example p.Pro534Leu and p.Gly579Arg have
been found up to four times in the present cohort and up to
Wve times in the literature, although one cannot exclude that
same pedigrees have been involved in diVerent reports; (3)
seven of the missense mutations aVect cysteine residues
involved in disulWde bonds essential for the folding of the
protein (out of the 10 present in the PHEX protein); (4)
missense mutations are not randomly distributed in the 3D
PHEX molecule but are mainly found in two regions
located in the inner part of the PHEX protein; (5) one of the
missense mutations (L555P) was found to segregate with
the disease in a large family with an adult onset type of
hypophosphatemia (Econs et al. 1998); (6) Several missense mutations (as the p.Cys85Phe and p.Gly579Arg in
our cohort) have been shown to alter the traYcking of
PHEX through the endoplasmic reticulum, the glycosylation, or the folding conformation of the protein (Sabbagh
et al. 2001, 2003). But our results also suggest that some
missense mutations are not associated with the disease, for
example Val169Met in the present study. So the Wndings of
missense mutation should always be followed by a complete search of mutation in the whole PHEX gene.
Finally, and as an alternative to classical PCR followed
by sequencing, we have tested the HRM approach of mutation detection. By using this method in a subset of patients,
we detected 82% of the 24 mutations presented by the
patients in exon by sequencing analysis, involving point
mutations and small deletion/insertions. Among the 4 undetected mutations were point mutations and insertions/deletions as well. The good negative predictive value suggests
that the HRM approach may be a relevant screening
method for the analysis of large cohorts of patients when
one gene with numerous exons is frequently involved, as is
the case for patients with hypophosphatemic rickets due to
PHEX alteration.
In conclusion, this analysis of the PHEX gene in a large
cohort of hypophosphatemic patients highlights new
genetic issues. Accurate inclusion and exclusion criteria
must be deWned before the genetic analysis. This prescreening will increase the percentage of PHEX mutations
found in sporadic cases. Missense mutations, even if
located on highly conserved residue of the protein, may not
be the cause of the disease, and their Wndings do not

123

Hum Genet (2009) 125:401411

preclude a complete analysis of the 22 exons and their

intronic junctions. In addition, patients with a late onset
form of the disease and even apparently healthy family
members of sporadic cases may bear a PHEX mutation. So
PHEX sequencing should be proposed to all sporadic cases,
whatever their age at the onset of the disease, and to their
parents.
Acknowledgments We thank Roche Diagnostics Corporation for
providing the HRM software, and all patients, as well as their families
and medical practitioners, who eagerly contributed to the work.

References
Albright F, Butler AM, Bloomberg E (1937) Rickets resistant to vitamin D therapy. Am J Dis Child 54:529547
Bergwitz C, Roslin NM, Tieder M, Loredo-Osti JC, Bastepe M,
Abu-Zahra H, Frappier D, Burkett K, Carpenter TO, Anderson D,
Garabedian M, Sermet I, Fujiwara TM, Morgan K, Tenenhouse
HS, Juppner H (2006) SLC34A3 mutations in patients with hereditary hypophosphatemic rickets with hypercalciuria predict a key
role for the sodiumphosphate cotransporter NaPi-IIc in maintaining phosphate homeostasis. Am J Hum Genet 78:179192
Bianchetti L, Oudet C, Poch O (2002) M13 endopeptidases: new
conserved motifs correlated with structure, and simultaneous phylogenetic occurrence of PHEX and the bony Wsh. Proteins
47:481488
Bland ND, Pinney JW, Thomas JE, Turner AJ, Isaac RE (2008) Bioinformatic analysis of the neprilysin (M13) family of peptidases
reveals complex evolutionary and functional relationships. BMC
Evol Biol 8:16
Brownstein CA, Adler F, Nelson-Williams C, Iijima J, Li P, Imura A,
Nabeshima Y, Reyes-Mugica M, Carpenter TO, Lifton RP (2008)
A translocation causing increased alpha-klotho level results in
hypophosphatemic rickets and hyperparathyroidism. Proc Natl
Acad Sci USA 105:34553460
Cho HY, Lee BH, Kang JH, Ha IS, Cheong HI, Choi Y (2005) A clinical and molecular study of hypophosphatmic rickets in children.
Pediatr Res 58:329333
Chong A, Zhang G, Bajic V (2004) Information for the coordinates of
exons (ICE): a human splice sites database. Genomics 84:762
766
Christie PT, Harding B, Nesbit MA, Whyte MP, Thakker RV (2001)
X-linked hypophosphatemia attributable to pseudoexons of the
PHEX gene. J Clin Endocrinol Metab 86:38403844
Consortium ADHR (2000) Autosomal dominant hypophosphataemic
rickets is associated with mutations in FGF23. Nat Genet 26:345
348
Dixon PH, Christie PT, Wooding C, Trump D, GrieV M, Holm I, Gertner JM, Schmidtke J, Shah B, Shaw N, Smith C, Tau C, Schlessinger D, Whyte MP, Thakker RV (1998) Mutational analysis of
PHEX gene in X-linked hypophosphatemia. J Clin Endocrinol
Metab 83:36153623
Econs MJ, Friedman NE, Rowe PSN, Speer MC, Francis F, Strom M,
Oudet C, Smith JA, Minomiya JT, Lee BE, Bergen H (1998) A
PHEX Gene mutation is responsible for adult-onset vitamin
D-resistant hypophosphatemic osteomalacia: Evidence that the
disorder is not a distinct entity from X-linked hypophosphatemic
rickets. J Clin Endocrinol Metab 83:34593462
Filisetti D, Ostermann G, von Bredow M, Strom T, Filler G, Ehrich J,
Pannetier S, Garnier JM, Rowe PSN, Francis F, Julienne A, Hanauer A, Econs MJ, Oudet C (1999) Non-random distribution of

Hum Genet (2009) 125:401411

mutations in the PHEX gene, and under-detected missense mutations at non-conserved residues. Eur J Hum Genet. 7(5):615619
Francis F, Strom TM, Hennig S, Boddrich A, Lorenz B, Brandau O,
Mohnike KL, Cagnoli M, SteVens C, Klages S, Borzym K, Pohl
T, Oudet C, Econs MJ, Rowe PS, Reinhardt R, Meitinger T,
Lehrach H (1997) Genomic organization of the human PEX gene
mutated in X-linked dominant hypophosphatemic rickets.
Genome Res 7:573585
Goji K, Ozaki K, Sadewa AH, Nishio H, Matsuo M (2006) Somatic
and germline mosacism for a mutation of the PHEX gne can lead
to transmission of X-linked hypophosphatemic rickets that mimics an autosomal dominant trait. J Clin Endocrinol Metab 91:365
370
Holm IA, Huang X, Kunkel LM (1997) Mutational analysis of the PEX
gene in patients with X-linked hypophosphatemic rickets. Am
J Hum Genet 60:790797
Hyp consortium (1995) A gene (PEX) with homogies to endopeptidases is mutated in patients with X-linked hypophosphatemic rickets.
Nat Genet 11:130136
Ichikawa S, Traxler EA, Estwick SA, Curry LR, Johnson ML,
Sorenson AH, Imel EA, Econs MJ (2008) Mutational survey of
the PHEX gene in patients with X-linked hypophosphatemic rickets. Bone 43:663666
Lloyd SE, Pearce SH, Fisher SE, Steinmeyer K, Schwappach B,
Scheinman SJ,Harding B, Bolino A, Devoto M, Goodyer P,
Rigden SP, Wrong O, Jentsch TJ, Craig IW, Thakker RV (1996)
A common molecular basis for three inherited kidney stone diseases. Nature 1;379(6564):445449.
Lorenz-Depiereux B, Bastepe M, Benet-Pags A, Amyere M,
Wagenstaller J, Mller-Barth U, Badenhoop K, Kaiser SM,
Rittmaster RS, Shlossberg AH, Olivares JL, Loris C, Ramos FJ,
Glorieux F, Vikkula M, Jppner H, Strom TM (2006a) DMP1
mutations in autosomal recessive hypophosphatemia implicate a
bone matrix protein in the regulation of phosphate homeostasis.
Nat Genet 38:12481250
Lorenz-Depiereux B, Benet-Pages A, Eckstein G, TenenbaumRakover Y, Wagenstaller J, Tiosano D, Gershoni-Baruch R, Albers N, Lichtner P, Schnabel D, Hochberg Z, Strom TM (2006b)
Hereditary hypophosphatemic rickets with hypercalciuria is
caused by mutations in the sodium-phosphate cotransporter gene
SLC34A3. Am J Hum Genet. 78(2):193201
Oefner C, DArcy A, Hennig M, Winkler FK, Dale GE (2000) Structure of human neutral endopeptidase (Neprilysin) complexed with
phosphoramidon. J Mol Biol 296:341349
Pieper U, Eswar N, Braberg H, Madhusudhan MS, Davis F, Stuart AC,
Mirkovic N, Rossi A, Marti-Renom MA, Fiser A, Webb B,
Greenblatt D, Huang C, Ferrin T, Sali A (2004) MODBASE, a
database of annotated comparative protein structure models, and
associated resources. Nucleic Acids Res 32:D217D222

411
Popowska E, Pronicka E, Suiek A, Jurkiewicz D, Rowe P, Rowinska
E, Krajewska-Walasek M (2000) X-linked hypophosphatemia in
Polish patients. 1. Mutations in the PHEX gene. J Appl Genet
41:293302
Rowe PSN (2004) The wrickkened pathways of FGF23, MEPE and
PHEX. Crit Rev Oral Biol Med 15:264281
Rowe PS, Oudet CL, Francis F, Sinding C, Pannetier S, Econs MJ,
Strom TM, Meitinger T, Garabedian M, David A, Macher MA,
Questiaux E, Popowska E, Pronicka E, Read AP, Mokrzycki A,
Glorieux FH, Drezner MK, Hanauer A, Lehrach H, Goulding JN,
ORiordan JL (1997) Distribution of mutations in the PEX gene
in families with X-linked hypophosphataemic rickets (HYP).
Hum Mol Genet 6:539549
Ruchon AF, Tenenhouse HS, Marcinkiewicz M, Siegfried G, Aubin
JE, DesGroseillers L, Crine P, Boileau G (2000) Developmental
expression and tissue distribution of Phex protein: eVect of the
Hyp mutation and relationship to bone markers. J Bone Miner Res
15:14401450
Sabbagh Y, Boileau G, DesGroseillers L, Tenenhouse HS (2001) Disease-causing missense mutations in the PHEX gene interfere with
membrane targeting of the recombinant protein. Hum Mol Genet
10:15391546
Sabbagh Y, Boileau G, Campos M, Carmona AK, Tenenhouse HS
(2003) Structure and function of disease-causing missense mutations in the PHEX gene. J Clin Endocrinol Metab 88:22132222
Sato K, Tajima T, Nakae J, Adachi M, Asakura Y, Tachibana K,
Suwa S, Katsumata N, Tanaka T, Hayashi Y, Abe S, Murashita
M, Okuhara K, Shinohara N, Fujieda K (2000) Three novel
PHEX gene mutations in Japanese patients with X-linked hypophosphatemic rickets. Pediatr Res 48:536540
Song HR, Park JW, Cho DY, Yang JH, Yoon HR, Jung SC (2007)
PHEX gene mutations and genotype-phenotype analysis of Korean patients with hypophosphatemic rickets. J Korean Med Sci
22:981986
Thakker RV, ORiordan JL (1988) Inherited forms of rickets and
osteomalacia. Baillieres Clin Endocrinol Metab 2:91157
Thompson DL, Sabbagh Y, Tenenhouse HS, Roche PC, Drezner MK,
Salisbury JL, Grande JP, Poeschla EM, Kumar R (2002) Ontogeny
of Phex/PHEX protein expression in mouse embryo and subcellular localization in osteoblasts. J Bone Miner Res 17(2):311320
Turner AJ, Tanzawa K (1997) Mammalian embrane metallopeptidases :
NEP, ECE, Kell and PEX. FASEB J 11:355364
Tyynismaa H, Kaitila I, Nnt-Salonen K, Ala-Houhala M, Alitalo T
(2000) IdentiWcation of Wfteen novel Phex gene mutations in Wnnish
patients with hypophosphatemic rickets. Hum Mutat 15:383384
Winters RW, Graham JB, Williams TF (1958) A genetic study of
familial hypophosphatemia and vitamin D-resistant rickets with a
review of the literature. Medicine 37:97142

123