0021-972X/04/$15.00/0
`Printed in U.S.A.
`
`The Journal of Clinical Endocrinology & Metabolism 89(1):49–60
`Copyright © 2004 by The Endocrine Society
`doi: 10.1210/jc.2003-031021
`
`Two Prevalent CYP17 Mutations and Genotype-
`Phenotype Correlations in 24 Brazilian Patients
`with 17-Hydroxylase Deficiency
`
`MARIVAˆ NIA COSTA-SANTOS, CLAUDIO E. KATER, RICHARD J. AUCHUS, AND
`BRAZILIAN CONGENITAL ADRENAL HYPERPLASIA MULTICENTER STUDY GROUP
`Division of Endocrinology and Metabolism, Department of Medicine, Escola Paulista de Medicina, Federal
`University of Sao Paulo (M.C.-S., C.E.K.), Sao Paulo, Brazil 04039-034; and Division of Endocrinology and
`Metabolism, Department of Internal Medicine, University of Texas Southwestern Medical Center (R.J.A.), Dallas,
`Texas 75390-8857
`
`We performed molecular genetic analysis of 24 subjects from
`19 families with 17-hydroxylase deficiency in Brazil. Of 7 novel
`CYP17 mutations, 2 (W406R and R362C) account for 50% and
`32% of the mutant alleles, respectively. Both mutations were
`completely inactive when studied in COS-7 cells and yeast
`microsomes; however, phenotypic features varied among sub-
`jects. Some 46,XY individuals with these genotypes had am-
`biguous genitalia, and other subjects had normal blood pres-
`sure and/or serum potassium. We found mutations W406R and
`R362C principally in families with Spanish and Portuguese
`ancestry, respectively, suggesting that two independent
`founder effects contribute to the increased prevalence of 17-
`hydroxylase deficiency in Brazil. Mutations Y329D and P428L
`
`MUTATIONS IN THE CYP17 gene cause 17-hydroxylase
`
`deficiency (17OHD), a rare form of congenital adrenal
`hyperplasia (CAH) with an estimated incidence of about
`1:50,000 newborns (1). Individuals with 17OHD account for
`roughly 1% of all cases of CAH, and most reports involve
`isolated cases from consanguineous families (2). Since cloning
`of the CYP17 gene encoding cytochrome P450c17 (CYP17, 17␣-
`hydroxylase/17,20-lyase) (3), nearly 40 different mutations in
`CYP17 have been described (4–9), although a few more com-
`mon mutations reoccur in certain ethnic groups (10–12).
`The typical features of complete 17OHD were described
`almost 40 yr ago (13), as hypertension, hypokalemia, and
`sexual infantilism in phenotypic females. Subsequent reports
`identified 17OHD as a cause not only of incomplete male
`pseudohermaphroditism (14), but also sexual infantilism in
`46,XY subjects (15). The lack of adrenal 17␣-hydroxylase
`activity drives massive overproduction of the 17-deoxy-
`steroids 11-deoxycorticosterone (DOC) and corticosterone
`(B), which are the mineralocorticoids that cause hypertension
`and hypokalemia in 17OHD (4). Concomitant lack of gonadal
`17,20-lyase activity precludes sex steroid production and
`hence the development of the male phenotype in utero or of
`secondary sexual characteristics at puberty.
`
`Abbreviations: CAH, Congenital adrenal hyperplasia; CPR, cyto-
`chrome P450-oxidoreductase; DOC, 11-deoxycorticosterone; 17OHD,
`17-hydroxylase deficiency.
`JCEM is published monthly by The Endocrine Society (http://www.
`endo-society.org), the foremost professional society serving the en-
`docrine community.
`
`retained a trace of activity, yet the two individuals with these
`mutations had severe hypertension and hypokalemia. The
`46,XX female with mutation Y329D reached Tanner stage 5,
`whereas the 46,XY subject with mutation P428L remained sex-
`ually infantile. The severity of hypertension, hypokalemia,
`17-deoxysteroid excess, and sex steroid deficiency varied,
`even among patients with completely inactive CYP17 pro-
`tein(s). Spontaneous sexual development occurred only in
`46,XX females with partial deficiencies. We conclude that
`other factors, in addition to CYP17 genotype, contribute to the
`phenotype of
`individual patients with 17-hydroxylase
`deficiency. (J Clin Endocrinol Metab 89: 49–60, 2004)
`
`Nevertheless, there is considerable variation in the clinical
`and biochemical features of 17OHD (16), including the vari-
`ant of isolated 17,20-lyase deficiency (17, 18). The severity of
`clinical disease tends to be milder with mutations that retain
`partial catalytic activity in assays using heterologous expres-
`sion systems (4), but the age of onset of hypertension, the
`degree of hypokalemia, and the aldosterone production rate
`appear to vary, even among patients with mutations that
`completely inactivate the enzyme (2). However, because
`there have been no studies of multiple individuals bearing
`the same genotype who have been studied by the same
`investigators, it is not clear to what extent genotype alone
`determines phenotype in 17OHD.
`Worldwide, the most common form of CAH is 21-hydrox-
`ylase deficiency (19), and the second most common form
`appears to be lipoid CAH in Japan and Korea (20) and 11-
`hydroxylase deficiency in the Middle East (21); founder ef-
`fects that yield a single prevalent mutation account for the
`high prevalence of these two disorders in their respective
`populations. In contrast, 17OHD appears to be the second
`most common form of CAH in Brazil (16, 22). Founder effects
`may also contribute to the high prevalence of 17OHD in
`Brazil, but the population of Brazil is among the most eth-
`nically heterogeneous in the world (23). The Portuguese set-
`tled Brazil beginning in the 1500s, and the indigenous Am-
`erindian people, Africans derived from the extensive slave
`trade, and waves of immigration from Italy, Spain, Germany,
`Asia, and The Netherlands contribute to the genetic diversity
`(23–25).
`
`49
`
`The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 14 September 2015. at 14:34 For personal use only. No other uses without permission. . All rights reserved.
`
`MYLAN PHARMS. INC. EXHIBIT 1027 PAGE 1
`
`

`
`50 J Clin Endocrinol Metab, January 2004, 89(1):49–60
`
`Costa-Santos et al. (cid:127) Two Common CYP17 Mutations in Brazil
`
`TABLE 1a. Basal and ACTH-stimulated adrenal steroid values in
`24 Brazilian patients with 17OHD
`
`The Brazilian Congenital Adrenal Hyperplasia Multi-
`center Study Group has had the opportunity to evaluate the
`clinical features of 30 subjects with 17OHD from 24 kindreds,
`the largest group of 17OHD cases studied by a single group.
`To provide insight into the phenotypic variations in 17OHD
`and to define the genetics of 17OHD in Brazil, we analyzed
`the CYP17 gene in these subjects. We now report the results
`of molecular genetic and functional analyses of
`the
`mutations.
`
`Subjects and Methods
`Subjects, clinical presentation, and hormonal evaluation
`
`Of 30 subjects in whom the diagnosis of 17OHD was established (at
`Escola Paulista de Medicina by C.E.K.), DNA was analyzed in 24, con-
`stituting the cohort for the genetic analysis. These 24 subjects derived
`from 19 kindreds, and consanguinity was known to occur in 6 of the 19
`families. The study protocol was approved by the committee on ethics
`in human research from Escola Paulista de Medicina (n.1703/98), and
`all patients provided written informed consent. Blood pressure was
`measured by aneroid sphygmomanometer in the seated position on at
`least three occasions. For diagnostic studies, blood samples were ob-
`tained before and 60 min after the iv injection of 250 ␮g cosyntropin
`[synthetic ACTH-(1–24)], and Table 1 lists the mean and ranges of basal
`and stimulated hormone values in these subjects. The diagnosis of
`17OHD was established by the reduced circulating concentrations of
`cortisol and gonadal steroids, elevated gonadotropins, and high [⬎3 sd
`above normal, with or without ACTH-(1–24) stimulation] concentra-
`tions of the diagnostic steroids DOC and/or B, as well as frequently
`elevated concentrations of 18-hydroxydeoxycorticosterone and 18-
`hydroxycorticosterone (16). In our subjects, basal hormone concentra-
`tions alone established the diagnosis. The clinical features are summa-
`rized in Table 2.
`
`DNA preparation, PCR, and sequencing
`
`DNA was extracted from peripheral leukocytes (Pure Gene DNA
`Isolation Kit D-5000, Gentra Systems). The 6.4-kb CYP17 gene was am-
`plified into 1–4 pieces from 0.5–1 ␮g genomic DNA using TaKaRa Ex
`Taq DNA polymerase (Takara Shuzo Co., Shiga, Japan) in 100-␮l reac-
`tions using buffer and deoxy-NTPs provided by the manufacturer and
`3% dimethylsulfoxide. The primers are listed in Table 3, and the loca-
`tions of the primers are illustrated in Fig. 1. To amplify 3- to 4-kb
`products, PCR parameters included 40 cycles of 3 min at 94 C, 1 min at
`65 C, and 3 min at 70 C. For amplification of the entire gene, the annealing
`time was increased to 1.5 min, and the extension parameters were 72 C
`for 5.5 min. The final PCR products were precipitated with ethanol and
`purified on 1% agarose gels using the QIAEX II kit (Qiagen, Chatsworth,
`CA). Amplicons were submitted for direct sequencing of the 8 exons and
`flanking intronic DNA by the dye termination method on a PE Applied
`Biosystems instrument (McDermott Center Sequencing Facility at Uni-
`versity of Texas Southwestern Medical Center, Dallas, TX). The muta-
`tions were identified by comparison with the GenBank sequence (ac-
`cession no. M19489) for CYP17 (3) using MacVector 6.5.3 (Accelrys
`Corp., San Diego, CA). Identified mutations were confirmed by se-
`quencing the product of a second PCR amplification in the opposite
`direction.
`
`Heterologous expression and enzyme assay
`
`The cDNAs for missense CYP17 mutations were generated by sequential
`PCR using overlapping mutagenic oligonucleotides (Table 3) with template
`plasmid pLW01-c17 and Ex Taq polymerase with 1% dimethylsulfoxide as
`previously described (26). The final PCR product was extracted with phe-
`nol-chloroform, precipitated with ethanol, digested with BamHI and EcoRI,
`gel-purified, ligated into the eukaryotic expression vector pcDNA3 (In-
`vitrogen, Carlsbad, CA), and later subcloned into yeast expression vector
`V10 (27). Each cDNA insert was sequenced in its entirety to ensure that only
`the desired mutations were introduced.
`The enzymatic activities of the four missense mutations were studied
`by transient transfection of COS-7 and HEK-293 cells with 1–2 ␮g of the
`
`Steroids
`
`Basal
`
`Post-ACTH
`
`Cortisol (␮g/dl)
`
`DOC (ng/dl)
`
`B (␮g/dl)
`
`2.3 ⫾ 2.9
`[0.1–11.9]
`
`6.2 ⫾ 7.9
`[0.1–19.7]
`
`291 ⫾ 124
`[120–504]
`
`415 ⫾ 159
`[150–737]
`
`Reference valuesa
`Basal
`Post-ACTH
`6–25
`18–42
`
`4–12
`
`12–61
`
`15.5 ⫾ 7.7
`[0.84–32.9]
`
`23.9 ⫾ 10.6
`[2.66–44.4]
`
`0.1–0.5 1.7–4.8
`
`18OHDOC (ng/dl)
`
`271 ⫾ 163
`[35–704]
`
`478 ⫾ 253
`[41–1045]
`
`0–10
`
`56–158
`
`18OHB (ng/dl)
`
`10–35
`
`84–174
`
`411 ⫾ 232
`261 ⫾ 139
`[43–855]
`[11–456]
`Values are the mean ⫾ SD [range]. 18OHB, 18-hydroxycorticoste-
`rone; 18OHDOC, 18-hydroxydeoxycorticosterone.
`a Basal and ACTH-stimulated reference values from Nichols In-
`stitute and from Kater et al. (25a); normal ranges for adolescents and
`adults are similar. Conversion factors for Systeme International units
`are: DOC, ⫻0.03026 nmol/liter; B, ⫻28.86 nmol/liter; 18OHDOC,
`⫻0.02886 nmol/liter; 18OHB, ⫻0.02759 nmol/liter; cortisol, ⫻27.59
`nmol/liter.
`
`TABLE 1b. Baseline gonadotropin and gonadal steroid values in
`24 Brazilian patients with 17OHD
`
`Gonadotropins/gonadal steroid
`
`LH (IU/liter)
`
`FSH (IU/liter)
`
`Testosterone (ng/dl)
`
`Patients
`
`46,XX
`51 ⫾ 31
`[12.5–79]
`
`87 ⫾ 40
`[38–170]
`
`46,XY
`50 ⫾ 28
`[12.7–87]
`
`77 ⫾ 42
`[20–164]
`
`30 ⫾ 16
`[12.5–65]
`
`18 ⫾ 16
`[2–48]
`
`Reference
`values
`
`2–15
`
`2–12
`
`⬍50a
`
`Estradiol (ng/dl)
`
`1.1 ⫾ 0.7
`[0.1–2.2]
`Values are the mean ⫾ SD [range].
`a Reference values for testosterone and estradiol are given for pre-
`pubertal children and are higher for Tanner stages 2–5. Conversion
`factors for Systeme International units are: testosterone, ⫻0.03467
`nmol/liter; estradiol, ⫻35.71 pmol/liter.
`
`1.5 ⫾ 1.0
`[0.3–3.1]
`
`⬍30a
`
`pcDNA3 expression vectors using FuGENE6 (3 ␮l) in 100 ␮l serum-free
`medium as previously described (26). Incubations with 0.1 ␮m [3H]pro-
`gesterone, -pregnenolone, or -17␣-hydroxypregnenolone (90,000 cpm;
`PerkinElmer Life Sciences, Norwalk, CT) for up to 16 h were repeated
`three times using COS-7 cells and were confirmed with an additional
`experiment using HEK-293 cells under similar assay conditions. In some
`cases, incubations were repeated with 0.01 ␮m steroids to increase assay
`sensitivity. Extraction, chromatography, and autoradiography were per-
`formed as previously described (28).
`The P450 content and enzymatic activities of the mutations were also
`studied in Saccharomyces cerevisiae strain W303B. Yeast were transformed
`with 1 ␮g expression vector V10 (empty, and with wild-type or mutant
`CYP17 cDNA) with or without pYcDE2-OR to provide cytochrome
`P450-oxidoreductase (CPR) (29), using the lithium acetate method as
`previously described (26). CO-reduced P450 difference spectra were
`performed by resuspending yeast harvested from 80 ml culture in 12 ml
`0.1 mm potassium phosphate (pH 7.4) with glucose, adding 3-ml aliquots
`to two cuvettes, and bubbling CO gas into the sample cuvette for 1 min
`
`The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 14 September 2015. at 14:34 For personal use only. No other uses without permission. . All rights reserved.
`
`MYLAN PHARMS. INC. EXHIBIT 1027 PAGE 2
`
`

`
`Costa-Santos et al. (cid:127) Two Common CYP17 Mutations in Brazil
`
`J Clin Endocrinol Metab, January 2004, 89(1):49–60 51
`
`W406R
`R362C
`R362C
`W406R/M1T
`Y329X
`W406R/R362C
`W406R
`W406R/R362C
`W406R
`Y329D/A–G
`W406R
`W406R
`W406R
`W406R
`R362C
`R362C
`P428L
`W406R
`W406R
`W406R
`W406R
`R362C
`R362C
`R362C
`
`Mutations
`
`3.1
`
`3.4
`3.2
`2.3
`2.6
`3.1
`2.9
`3.6
`2.8
`3.8
`2.0
`1.8
`3.4
`
`2.5
`2.6
`3.6
`3.2
`3.6
`2.4
`3.9
`3.6
`3.3
`
`90
`80
`90
`120
`110
`80
`100
`70
`100
`110
`102
`70
`120
`130
`130
`100
`97
`100
`90
`110
`105
`95
`100
`
`130
`120
`130
`160
`180
`125
`160
`110
`170
`145
`137
`115
`150
`200
`200
`145
`155
`140
`135
`145
`140
`130
`140
`
`F
`F
`F
`
`
`F
`F
`F
`F
`F
`F
`FF
`
`F
`F
`F
`F
`F
`FF
`FF
`F
`F
`F
`
`F
`
`(mEq/liter)
`
`K⫹
`
`(mmHg)
`
`DBP
`
`(mmHg)
`
`SBP
`
`Socialsex
`
`Fem
`Fem
`Fem
`Fem
`Fem
`Fem
`Fem
`Fem
`Fem
`Fem
`Fem
`Fem
`Fem
`Fem
`Fem
`Fem
`Fem
`Fem
`Fem
`Fem
`Fem
`Fem
`Fem
`Amb
`
`EGc
`
`B1P1
`B3P2d
`B3P2
`B2P1d
`B5P2d
`B2P1
`B1P1
`B1P1
`B3P2d
`B5P5
`B1P2d
`B1P1
`B1P1
`B1P1
`B3P2d
`B1P2
`B1P2
`B1P1
`B1P1
`B1P1
`B1P2
`B1P1
`B1P1
`B1P1
`
`XY
`XX
`XX
`XY/XXY
`XY
`XY
`XY/XO;CN
`XY
`XY
`XX
`XX
`XX
`XX
`XX
`XX
`CN
`XY
`XX
`XY
`XY
`XY
`XY
`XX
`XY
`
`Classical/MPH
`Classical
`GB
`2K⫹Myop/MPH
`Stroke/MPH
`Classical/MPH
`Classical/MPH
`Classical/MPH
`Classical/MPH
`Tanner5
`2K⫹Myop
`Classical
`NormalBP
`Classical
`Classical
`Stroke/MPH
`Arrh/MPH
`Classical
`Classical/MPH
`Classical/MPH
`HipFx/MPH
`NormalK⫹/MPH
`Classical
`Amb/MPH
`
`19
`27
`17
`40
`27
`14
`15
`19.7
`16
`27
`18
`17.4
`16.6
`15.7
`34
`34
`19
`11.8
`15.5
`14.4
`13.8
`13.7
`16.5
`10.3
`
`aClassical,Hypertension,hypokalemia,pubertaldelay;MPH,malepseudohermaphrodite;HipFx,hipfracture;Arrh,arrhythmia;Myop,myopathy;GB,gonadoblastomain
`
`dGonadalsteroidreplacementbegunbeforediagnosis.
`cEG,Externalgenitalia;Amb,ambiguousgenitalia;Fem,female/infantile.
`bCN,Chromatin(buccalsmearinactiveX)negative;case17,1cellin39wasXXY.
`
`streakovaries.
`
`1.8–4.1
`
`3.2
`
`70–120
`
`100
`
`110–200
`
`146
`
`10.3–40
`
`18
`
`Range
`Median
`
`19
`18-2
`18-1
`17
`16
`15
`14
`13
`12
`11
`10-2
`10-1
`
`8
`9
`
`7-2
`7-1
`
`6
`
`5-2
`5-1
`
`1-2
`1-1
`
`2
`
`
`34
`
`Tannerstage
`
`Karyotypeb
`
`Remarkablefeaturesa
`
`Age(yr)
`
`Kindredandsubjects
`
`TABLE2.Clinicalcharacteristicsof24Braziliansubjectswith17OHDatdiagnosis
`
`The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 14 September 2015. at 14:34 For personal use only. No other uses without permission. . All rights reserved.
`
`MYLAN PHARMS. INC. EXHIBIT 1027 PAGE 3
`
`

`
`52 J Clin Endocrinol Metab, January 2004, 89(1):49–60
`
`Costa-Santos et al. (cid:127) Two Common CYP17 Mutations in Brazil
`
`TABLE 3. Oligonucleotide primers for DNA amplification and sequencing
`
`Primer no.
`
`Sequence
`
`Oligo pairs and amplicons
`
`Oligonucleotide primer
`CYP17 genea
`c17geneS1ab
`c17geneAS1
`I3S1
`I4AS1
`I1AS1
`I1S1
`I3AS1
`I4S1
`I6AS1
`I6S1
`I5S1
`I7AS1
`Promotor
`c17PS1
`c17PAS1
`c17PS2
`c17PAS2
`c17PS3
`Mutagenesisc
`5⬘-AATGAGAAGGAGCGGCACCAGCCGGATCAG-3⬘
`W406R-S1
`18
`5⬘-ATCCGGCTGGTGCCGCTCCTTCTCATTGTG-3⬘
`W406R-AS1
`19
`5⬘-CGAGAGGTGCTTTGCCTCAGGCCCGTGGCC-3⬘
`R362C-S1
`20
`5⬘-ACGGGCCTGAGGCAAAGCACCTCTCGGATG-3⬘
`R362C-AS1
`21
`5⬘-CAGCTCATCTCACTGTCAGTAAGCTATTTG-3⬘
`P428L-S1
`22
`5⬘-TAGCTTACTGACAGTGAGATGAGCTGGGTC-3⬘
`P428L-AS1
`23
`5⬘-AAGAAGAAGCTCGACGAGGAGATTGACCAG-3⬘
`Y329D-S1
`24
`5⬘-GTCAATCTCCTCGTCGAGCTTCTTCTTCAC-3⬘
`Y329D-AS1
`25
`5⬘-TCAGCAAAAAACCCCTCAAGACCCG-3⬘
`pLW01-AS1
`26
`a Ix, Intron x; S and AS, sense and antisense, respectively (see Fig. 1 for primer locations).
`b Primer c17geneS1 was modified (underlined C) to c17geneS1a (see Results).
`c T7 (sense) and pLW01-AS1 (antisense) oligonucleotides were used as terminal primers. The mutated base pair is underlined.
`
`1
`2
`3
`4
`5
`6
`7
`8
`9
`10
`11
`12
`
`13
`14
`15
`16
`17
`
`5⬘-CTCCACCGCTGTCTATCTTGCCTGCC-3⬘
`5⬘-CTCTAAATCTGTGTTGTGGGGCCAC-3⬘
`5⬘-GCTGGAGAAGCAAAATGGAAGAAGGGTGG-3⬘
`5⬘-CCTACTATGTGCCAGGTTCTCTGCTTG-3⬘
`5⬘-TGGTCTGAAGACCTGAACCAATCCC-3⬘
`5⬘-CAAGAGTGGGGTGGATGGGTGTGAG-3⬘
`5⬘-GATTGGGGACAATGTCAGGGTCTAC-3⬘
`5⬘-GAGTGTCACAGATGGGGCTCCTTCC-3⬘
`5⬘-TGGGCTGGCAAGCAGTGAATGCATC-3⬘
`5⬘-GCATGAGGCTGAGCAAGGAAGGGAG-3⬘
`5⬘-CCTCTCCTGGGCTTACACACACTAG-3⬘
`5⬘-AGCAGAGTCCAGGCTCGCTGTGTG-3⬘
`
`5⬘-GAAGGGACTGCTGGAGCCATGGCAG-3⬘
`5⬘-GGAGGGGGTGTAAGAACAGGGAGAG-3⬘
`5⬘-GCCCTTTGTCCTTTCCCTCAGAAGC-3⬘
`5⬘-CAGCAAGAGAGCCACGAGCTGCCAC-3⬘
`5⬘-ACCTATCTCTCCCTTCCCTTCCACC-3⬘
`
`6.4 kb
`
`3.0 kb (1⫹4, 2⫹3)
`0.25 kb (3⫹4)
`0.4 kb (1⫹5)
`
`0.6 kb (6⫹7)
`0.75 kb (8⫹9)
`
`1.0 kb (2⫹10)
`1.5 kb (11⫹12)
`
`0.9 kb (13⫹14)
`
`0.9 kb (15⫹16)
`
`0.9 kb (5⫹17)
`
`0.5 kb (18⫹26)
`1.2 kb (T7⫹19)
`0.6 kb (20⫹26)
`1.1 kb (T7⫹21)
`0.4 kb (22⫹26)
`1.3 kb (T7⫹23)
`0.6 kb (24⫹26)
`1.1 kb (T7⫹25)
`1.7 kb (T7⫹26)
`
`representation of
`FIG. 1. Schematic
`the human CYP17 gene, indicating the
`approximate locations and sizes of ex-
`ons (numbered boxes), oligonucleotide
`primers (arrows, numbered as ex-
`plained in Table 3), and identified mu-
`tations (labeled asterisks).
`
`(26). Using the same suspension of whole yeast used for CO-reduced
`spectra, substrate-induced difference spectra were recorded with up to
`40 ␮m progesterone as previously described (30).
`Microsomes were prepared from 1 liter yeast culture grown to an A600
`of 1.0–1.8 in defined medium by sonication of spheroplasts as previously
`described (26), and protein content was determined by colorometric
`assay. Microsomes containing CPR and wild-type CYP17 (25 ␮g protein)
`or the mutations (250 ␮g) were incubated at 37 C with 0.1 ␮m [3H]pro-
`gesterone, -pregnenolone, or -17␣-hydroxypregnenolone for 60 min in
`200 ␮l 50 mm potassium phosphate, pH 7.4, with 1 mm NADPH. Ex-
`traction, chromatography, and autoradiography (26, 29) and immuno-
`blotting were performed as previously described (28).
`
`Results
`
`Mutation analysis
`Seven CYP17 gene mutations were found, none of which
`has been described previously. We found 5 missense muta-
`tions in exons 1, 6 (two), 7, and 8; a nonsense mutation in exon
`
`6; and an AG to CG mutation at g.2306 in the splice acceptor
`site of intron 2 (Table 4 and Fig. 1). Mutation W406R in exon
`7 was the most common, accounting for half of the mutant
`alleles,
`including 11 homozygotes. Mutation R362C ac-
`counted for almost one third of the mutant alleles with 7
`homozygotes, and 2 subjects were compound heterozygotes
`for W406R plus R362C. Together, mutations W406R and
`R362C accounted for 23 of 28 (82%) of the alleles identified
`in 25 (52%) and 16 (33%) of the 48 sequenced CYP17 genes,
`respectively (Table 4 and Fig. 2). One subject was homozy-
`gous for P428L, and another was homozygous for Y329X.
`One 46,XX female who reached Tanner stage 5 (Table 2) was
`a compound heterozygote for Y329D and the AG to CG
`substitution in the splice acceptor site of intron 2, and 1
`subject was heterozygous for M1T and W406R.
`Only the M1T mutation altered the restriction map of the
`
`The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 14 September 2015. at 14:34 For personal use only. No other uses without permission. . All rights reserved.
`
`MYLAN PHARMS. INC. EXHIBIT 1027 PAGE 4
`
`

`
`Costa-Santos et al. (cid:127) Two Common CYP17 Mutations in Brazil
`
`J Clin Endocrinol Metab, January 2004, 89(1):49–60 53
`
`Exon/Intron
`
`Homozygotes
`
`Heterozygotesb
`
`Affected allelesc
`
`7
`6
`8
`6
`6
`1
`Intron 2
`
`11
`7
`1
`1
`0
`0
`0
`
`3 (2⫻ R362C; 1⫻ M1T)
`2 (2⫻ W406R)
`0
`0
`1 (AG/CG splice)
`1 (W406R)
`1 (Y329D)
`
`14 (50)
`9 (32)
`1 (3.5)
`1 (3.5)
`1 (3.5)
`1 (3.5)
`1 (3.5)
`
`28 (100)
`
`TABLE 4. CYP17 mutations in 24 Brazilian subjects
`
`Mutationsa
`Amino acid
`W406R
`R362C
`P428L
`Y329X
`Y329D
`M1T
`AG/CG splice (position 2306)
`
`Nucleotide
`TGG/CGG
`CGC/TGC
`CCG/CTG
`TAC/TAG
`TAC/GAC
`ATG/ACG
`AG/CG
`
`1
`2
`3
`4
`5
`6
`7
`
`Total
`a Base substitutions are underlined.
`b The second mutation affecting the compound heterozygotes is in parentheses, such that each subject is listed twice.
`c Allele frequency is calculated based on one allele per homozygote (see text). Percentages are given in parentheses.
`
`20 subjects
`
`4 subjects
`
`FIG. 2. Brazilian CYP17 mutations. A, Electro-
`pherograms corresponding to homozygous and het-
`erozygous patients for the common mutations
`W406R and R362C. PCR-amplified DNA was puri-
`fied and submitted for direct sequencing using oli-
`gonucleotide I6S1 (for W406R) or I4S1 (for R362C) as
`described in the text. B, PmlI digest from patient and
`family members bearing the M1T mutation. Half of
`the DNA, PCR-amplified using oligonucleotides
`c17PS3 and I1AS1 (1 kb amplicon), is digested by
`PmlI, indicating heterozygosity for M1T in all three
`family members. The sequences of this region for the
`wild-type and mutant alleles are shown with the
`PmlI site underlined.
`
`surrounding region, introducing a PmlI site (CACGTG). Half
`of the DNA that was amplified from the affected proband
`with this mutation, from her mother, and from her half-sister
`was digested by PmlI at the expected site, indicating that all
`3 are heterozygous for this mutation (Fig. 2). For the family
`members of all other kindreds, zygosity was determined by
`sequencing amplified DNA from the region surrounding the
`mutation(s). DNA from all 29 available parents and from 36
`of 46 available siblings (78%) contained 1 copy of the same
`mutation found in the affected family member.
`After sequencing the exons and flanking intronic DNA
`from 24 patients, we consistently observed 5 differences from
`the CYP17 sequence M19489 deposited in GenBank (3): 1) a
`polymorphism at D283 (GAT to GAC) in exon 5, 2) a third
`C at the CC in positions ⫺26 to ⫺28 at the 3⬘ end of intron
`2, 3) a third C at the CC in positions ⫺3 to⫺5 at the 3⬘ end
`of intron 3, 4) an A to T substitution in position ⫺5 at the 3⬘
`end of intron 7, and 5) a C in place of the T 34 bp upstream
`from the ATG start codon. This last difference was incorpo-
`rated into primer c17geneS1a, which gave better PCR am-
`plifications than primer c17geneS1 (Table 3). These differ-
`
`ences, except for 2 and 5 above, have been noted previously
`(31), suggesting that these minor changes correspond to the
`correct sequence in our population.
`
`Heterologous expression, enzyme assay, and
`difference spectroscopy
`To determine whether the mutant enzymes retained any
`residual 17␣-hydroxylase activity, cDNAs bearing the four
`missense mutations (W406R, R362C, P428L, and Y329D)
`were constructed and subcloned into mammalian and yeast
`expression vectors pcDNA3 and V10, respectively. COS-7
`cells transiently transfected with pcDNA3 containing the
`wild-type CYP17 cDNA metabolized progesterone to the
`expected 4:1 mixture of 17␣-hydroxyprogesterone and 16␣-
`hydroxyprogesterone (Fig. 3) (32, 33). In contrast, COS-7 cells
`expressing the CYP17 mutations W406R and R362C pro-
`duced only the same background metabolites as mock-trans-
`fected cells (Fig. 3). Unlike mutations W406R and R362C,
`mutation Y329D always exhibited a small amount (⬃5%) of
`residual activity when expressed in COS-7 cells, and muta-
`
`The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 14 September 2015. at 14:34 For personal use only. No other uses without permission. . All rights reserved.
`
`MYLAN PHARMS. INC. EXHIBIT 1027 PAGE 5
`
`

`
`54 J Clin Endocrinol Metab, January 2004, 89(1):49–60
`
`Costa-Santos et al. (cid:127) Two Common CYP17 Mutations in Brazil
`
`CO-reduced difference spectrum. However, CO-reduced dif-
`ference spectra using yeast expressing the four missense
`mutations did not demonstrate detectable P450 absorbances
`(Fig. 4C). Similarly, yeast expressing mutations Y329D and
`P428L did not form a type I difference spectrum in the pres-
`ence of up to 40 ␮m progesterone (not shown). The lack of
`discernable spectral changes upon addition of CO or pro-
`gesterone indicates that only a small fraction of the mutant
`CYP17 proteins is expressed in a functional form in yeast.
`Immunoblots using fresh yeast microsomes containing the
`four missense mutations all contained immunoreactive pro-
`tein that comigrated with wild-type human CYP17 at ap-
`proximately 57 kDa as well as degradation products (Fig.
`4D). The amount of full-length protein remained relatively
`constant in microsomes containing wild-type CYP17 after
`multiple freeze-thaw cycles. In contrast, the quantity of full-
`length CYP17 protein declined rapidly with freeze-thawing
`or warming in sodium dodecyl sulfate sample buffer for the
`four missense mutations, and products of lower mass in-
`creased in parallel, presumably through proteolysis (not
`shown). We conclude from these data that all four missense
`mutations impair activity primarily by destabilizing the en-
`zyme structures, thus impairing the capacity to incorporate
`and/or retain heme. For mutations Y329D and P428L, a
`sufficient portion of the protein molecules remain properly
`folded to exhibit catalytic activity at least transiently, but this
`activity is barely detectable using sensitive radiochemical
`assays at low substrate concentrations. Mutations Y329D and
`P428L are examples of partial, combined deficiencies in both
`17␣-hydroxylase and 17,20-lyase activities.
`
`Correlation of genotype and phenotype: mutations W406R
`and R362C
`We compared the phenotypic characteristics in the 11 ho-
`mozygotes for W406R and the 7 homozygotes for R362C
`whose CYP17 enzymes are completely inactive in heterolo-
`gous assay systems. Although DOC concentrations pre- and
`post-ACTH-(1–24) administration were higher in W406R ho-
`mozygotes than in subjects homozygous for R362C (P ⬍
`0.01), blood pressure and circulating concentrations of po-
`tassium or other hormones were similar in the two groups
`(Table 5). Aldosterone values were low in both groups, and
`in fact, plasma aldosterone values in untreated subjects were
`uniformly suppressed regardless of genotype (data not
`shown). Thus, despite equally inactive CYP17 enzymes, ho-
`mozygotes for mutations W406R and R362C showed some
`trends to phenotypic differences, and clinical features varied
`even among subjects with these two common mutations
`(Tables 2 and 5). Although variations in blood pressure
`and potassium values may be influenced by dietary and
`environmental differences, the range of genital differentia-
`tion among these male pseudohermaphrodites remains
`unexplained.
`The finding of the W406R or R362C allele in an affected or
`obligate heterozygous subject who recently emigrated from
`the Iberian peninsula would argue against these mutations
`arising de novo in Brazil. We studied the family of a com-
`pound heterozygote for W406R and R362C (Fig. 5) whose
`deceased father was born in Spain (DNA not available). The
`
`FIG. 3. Progesterone metabolism by transiently transfected COS-7
`cells expressing the cDNA for CYP17, wild type or mutations W406R,
`R362C, P428L, and Y329D. A representative autoradiogram of chro-
`matographed steroid products is shown, with the locations of steroid
`standards indicated. The film was intentionally overexposed to reveal
`trace 17␣-hydroxylation by mutations P428L and Y329D, so that
`by-products of progesterone metabolism by endogenous enzymes ap-
`pear more abundant than is typical. Similar results were obtained in
`three separate experiments with COS-7 cells and in one identical
`experiment using HEK-293 cells. Prog, Progesterone; 17OHP and
`16OHP, 17␣-hydroxy- and 16␣-hydroxyprogesterone, respectively.
`
`tion P428L yielded a trace of 17␣-hydroxyprogesterone in
`most experiments (Fig. 3).
`The activity and spectral properties of the mutations were
`further characterized in yeast. Yeast microsomes containing
`wild-type human CYP17 and human P450-oxidoreductase
`are an abundant source of enzyme activity (29), as demon-
`strated by the pattern of [3H]pregnenolone metabolism in
`Fig. 4. As was observed when the mutations were expressed
`in COS-7 and HEK-293 cells, the 17␣-hydroxylase and 17,20-
`lyase activities of microsomes containing mutations W406R
`and R362C did not exceed background levels (Fig. 4A). These
`data confirm that mutations W406R and R362C are not active
`under these in vitro conditions. In contrast, microsomes con-
`taining mutations Y329D and P428L metabolized some pro-
`gesterone and pregnenolone by 17␣-hydroxylation, and both
`mutant enzymes demonstrated little 17,20-lyase activity even
`in the presence of cytochrome b5 (Fig. 4B). The 17␣-hydroxy-
`lation rates for Y329D and P428L were too low to determine
`meaningful kinetic constants, but we estimate that mutations
`Y329D and P428L retain less than 5% of the wild-type
`activity.
`Intact yeast that express wild-type human CYP17 afford
`CO-reduced difference spectra with a peak at 450 nm, de-
`rived from the CO adduct with the heme thiolate in CYP17
`(34). This difference spectrum, characteristic of all cyto-
`chromes P450, indicates the presence of properly folded,
`functional CYP17 protein, and we obtain 50–100 nmol of
`spectroscopically active wild-type human CYP17 per liter of
`culture using vector V10 in strain W303B (26). Likewise,
`addition of progesterone to a suspension of yeast expressing
`wild-type human CYP17 yields a type I substrate binding
`spectrum (30), and these difference spectra can be used to
`monitor the expression of functional protein in yeast. Be-
`cause mutations Y329D and P428L retain some activity, some
`of the protein must fold and incorporate heme properly. If
`these molecules form a stable CO adduct typical of active
`cytochromes P450, we would observe a peak at 450 nm in the
`
`The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 14 September 2015. at 14:34 For personal use only. No other uses without permission. . All rights reserved.
`
`MYLAN PHARMS. INC. EXHIBIT 1027 PAGE 6
`
`

`
`Costa-Santos et al. (cid:127) Two Common CYP17 Mutations in Brazil
`
`J Clin Endocrinol Metab, January 2004, 89(1):49–60 55
`
`FIG. 4. Activities and spectral proper-
`ties of mutant CYP17 enzymes in S. cer-
`evisciae. A and B, Autoradiograms of
`thin layer chromatograms (overexposed
`as in Fig. 3) showing pregnenolone (A)
`or 17␣-hydroxypregnenolone (B) me-
`tabolism by yeast microsomes contain-
`ing human CPR and wild-type (WT)
`or mutant CYP17 as indicated. Preg,
`Pregnenolone; DHEA, dehydroepian-
`drosterone; 17Preg, 17␣-hydroxypreg-
`nenolone. B, Purified cytochrome b5 (5
`pmol) was added where indicated. C,
`CO-reduced difference spectra from
`yeast expressing wild-type CYP17 (top),
`P428L (middle), or Y329D mutation
`(bottom). The cursor is set at 450 nm,
`and the bar represents 0.05 absorbance
`units (AU). D, Immunoblot of yeast mi-
`crosomes containing wild-type (WT)
`CYP17 and for the four missense mu-
`tations (30 ␮g protein/lane). The arrow
`denotes bands corresponding to full-
`length CYP17 proteins, which migrate
`near the chemically modified 62-kDa
`protein standard (26).
`
`father migrated to Brazil and married a Brazilian of Portu-
`guese ancestry, and the mother was heterozygous for allele
`R362C. Of their children, one with 17OHD bore one copy of
`each mutation, one was a heterozygote for W406R, one is
`wild type at both alleles, and one died of an unknown cause.
`The genetics within this kindred suggest that the W406R
`mutation arose in Spain from an ancestor common to many
`Brazilians bearing mutation W406R.
`
`Discussion
`In this study we report the largest series of 17OHD subjects
`studied in a single country by 1 group of investigators. In
`1994, Kater and Biglieri (16) first observed an unusual large
`number of Brazilian cases of 17OHD, and 17OHD remains
`the second most common cause of CAH in Brazil (22). The
`finding in the present study that 82% of the mutant alleles can
`be explained by 2 mutations suggests founder effects for the
`CYP17 mutations, an unusual result for an autosomal reces-
`sive disease in a country with such an extensive racial ad-
`mixture (24). Only 3 recurring CYP17 mutations have been
`described previously: a CATC duplication following Ile479
`found in Canadian Mennonites and Dutch Frieslanders (10),
`the deletion of Phe at codon 53 or 54 in Japan (11); and
`deletion of residues 487–489 in East Asia (35). The fact that
`founder effects can be manifested in a country with great
`ethnic heterogeneity also suggests a high coefficient of in-
`breeding in local areas, even though known c