Am. J. Hum. Genet. 65:59-67, 1999
`
`The Molecular Basis of Cystathionine 3-Synthase Deficiency in Dutch
`Patients with Homocystinuria: Effect of CBS Genotype on Biochemical and
`Clinical Phenotype and on Response to Treatment
`
`Leo A. J. Kluijtmans,~’* Godfried H. J. Boers/Jan P. Kraus,~
`Lambert P. W. J. van den Heuvel,~ Johan R. M. Cruysberg,:~ Frans J. M. Trijbels,~
`and Henk J. Blom~
`
`Departments of ~Pediatrics, 2Internal Medicine, and ~Ophthalmology, University Hospital Nijmegen, Nijmegen, the Netherlands; and
`4Department of Pediatrics, University of Colorado School of Medicine, Denver
`
`Summary
`
`Introduction
`
`Homocystinuria due to cystathionine 3-synthase (CBS)
`deficiency, inherited as an autosomal recessive trait, is
`the most prevalent inborn error of methionine metab-
`olism. Its diverse clinical expression may include ectopia
`lentis, skeletal abnormalities, mental retardation, and
`premature arteriosclerosis and thrombosis. This varia-
`bility is likely caused by considerable genetic heteroge-
`neity. We investigated the molecular basis of CBS defi-
`ciency in 29 Dutch patients from 21 unrelated pedigrees
`and studied the possibility of a genotype-phenotype re-
`lationship with regard to biochemical and clinical ex-
`pression and response to homocysteine-lowering treat-
`ment. Clinical symptoms and biochemical parameters
`were recorded at diagnosis and during long-term follow-
`up. Of 10 different mutations detected in the CBS gene,
`833T-~C (I278T) was predominant, present in 23 (55%)
`of 42 independent alleles. At diagnosis, homozygotes for
`this mutation (n = 12) tended to have higher homocy-
`steine levels than those seen in patients with other gen-
`otypes (n = 17), but similar clinical manifestations. Dur-
`ing follow-up, I278T homozygotes responded more
`efficiently to homocysteine-lowering treatment. After
`378 patient-years of treatment, only 2 vascular events
`were recorded; without treatment, at least 30 would
`have been expected (P < .01). This intervention in Dutch
`patients significantly reduces the risk of cardiovascular
`disease and other sequelae of classical homocystinuria
`syndrome.
`
`Received June 8, 1998; accepted for publication May 3, 1999; elec-
`tronically published June 4, 1999.
`Address for correspondence and reprints: Dr. Godfried H.J. Boers,
`Department of Internal Medicine, University Hospital Nijmegen,
`P.O. Box 9101, 6500 HB Nijmegen, the Netherlands. E-maih
`G.Boers@aig.azn.nl
`*Present affiliation: Department of Pharmacology, University of
`Pennsylvania School of Medicine, Philadelphia.
`© 1999 by The American Society of Human Genetics. All rights reserved.
`0002-9297/9916501-0010502.00
`
`Homocystinuria due to cystathionine 3-synthase (CBS;
`L-serine hydrolyase [adding homocysteine]) deficiency
`(MIM 236200) is the most common inborn error in
`methionine metabolism. CBS is a pyridoxal 5’-phosphate
`(PLP)-dependent enzyme and condenses homocysteine
`and serine to cystathionine, an irreversible step in trans-
`sulfuration (Mudd et al. 1995).
`Carson and Neill (1962) first described homocystin-
`uria in mentally retarded individuals in Northern Ire-
`land. Soon thereafter, it was shown that the primary
`defect in homocystinuria was an enzymatic defect of CBS
`(Mudd et al. 1964), with a recessive mode of inheritance
`(Finkelstein et al. 1964). The clinical manifestation of
`CBS deficiency is diverse, and four major organ systems
`are predominantly involved: the eye (high myopia and
`ectopia lentis), the skeleton (osteoporosis, scoliosis, and
`Marfanoid features), the vascular system (premature ar-
`teriosclerosis and thromboembolism), and the CNS
`(mental retardation, convulsions, and psychiatric dis-
`turbances).
`Biochemically, patients with CBS deficiency are char-
`acterized by severe hyperhomocysteinemia and homo-
`cystinuria, hypermethioninemia, and decreased plasma
`cysteine levels. Furthermore, the CBS activities measured
`in either liver biopsy specimens (Mudd et al. 1964), cul-
`tured fibroblasts (Uhlendorf and Mudd 1968), or phy-
`tohemagglutinin-stimulated lymphocytes (Goldstein et
`al. 1972) are mostly well below the range of CBS activ-
`ities observed in controls and heterozygotes for CBS de-
`ficiency. The first choice of therapy in CBS-deficient pa-
`tients consists of administration of supraphysiological
`doses of pyridoxine (vitamin B6), the precursor of PLP,
`the cofactor of CBS. A large international survey of >600
`patients with homocystinuria showed that ~50% of the
`patients responded to high doses of pyridoxine with a
`substantial reduction in blood homocysteine concentra-
`tions (Mudd et al. 1985). Pyridoxine-nonresponsive pa-
`tients usually are more severely affected than pyridoxine-
`
`59
`
`Sandoz Inc. IPR2016-00318
`Sandoz v. Eli Lilly, Exhibit 1116-0001
`
`

`
`60
`
`Table 1
`
`CBS Mutations in Dutch Homocystinurics
`
`DNA
`Mutation
`
`Amino Acid
`Substitution
`
`RFLP
`
`Frequencya
`
`R125W
`373C~T
`I152M
`456C~G
`C165Y
`494G~A
`V180A
`539T~C
`I278T
`833T~C
`1105C~Tb R369C
`1111G~A V371M
`1301C~A
`T434N
`1330G~A D444N
`1471C~Tb R491C
`
`AciI
`Sau3A
`BsoFI
`+HhaI
`÷BsrI
`HhaI
`+NlaIII
`None
`TaqI
`+BglII
`
`1/42
`3/42
`4/42
`1/42
`23/42
`2/42
`2/42
`1/42
`2/42
`2/42
`
`Am. J. Hum. Genet. 65:59-67, 1999
`
`tected in patients from Central and Eastern Europe
`(Kraus 1998).
`In the present study, we investigated the molecular
`basis of homocystinuria due to CBS deficiency in 29
`Dutch homocystinuria patients from 21 unrelated ped-
`igrees, and we studied a possible relationship between
`CBS genotype and biochemical and clinical phenotype.
`We therefore measured homocyst(e)ine concentrations
`at diagnosis, upon pyridoxine treatment, and upon max-
`imal treatment, with pyridoxine and folic acid, with or
`without betaine, and then recorded whether homocy-
`steine-lowering treatment had been able to prevent fur-
`ther clinical events or symptoms.
`
`a In independent alleles.
`Mutations were observed in cis in one patient.
`
`Patients, Material, and Methods
`
`responsive patients and are concomitantly treated with
`combinations of folic acid, hydroxycobalamin, and be-
`taine, to stimulate remethylation of homocysteine to
`methionine.
`Data on the clinical efficacy of homocysteine-lowering
`treatment are scarce. In pyridoxine-responsive patients
`and early-treated pyridoxine-nonresponsive patients,
`such treatment has reduced the number of initial throm-
`boembolic events (Mudd et al. 1985). A recent report
`by Wilcken and Wilcken (1997) showed that treatment
`that effectively lowered plasma homocysteine concen-
`trations markedly reduced the cardiovascular risk in a
`group of 32 patients with pyridoxine-responsive and
`pyridoxine-nonresponsive homocystinuria.
`The human CBS gene, which has been mapped to
`21q22.3 (Miincke et al. 1988), encodes a CBS subunit
`of 63 kD, which assembles into a homotetrameric pro-
`tein (Kraus et al. 1978). So far, >60 mutations have been
`detected in the CBS gene (Kraus 1998), and functional
`relevance has been tested for some of them in either a
`bacterial (de Franchis et al. 1994; Marble et al. 1994;
`Kluijtmans et al. 1996a) or a yeast (Kruger and Cox
`1995) expression system. Although most mutations seem
`to be private or restricted to only a few pedigrees, three
`mutations are relatively common among patients with
`homocystinuria. An 833T-~C transition (I278T) (Kozich
`and Kraus 1992) has been found in alleles from hom-
`ocystinuric patients of different ethnic backgrounds and
`has been reported to be associated with pyridoxine re-
`sponsiveness and a relatively mild clinical phenotype
`when present in homozygous state (Shih et al. 1995).
`On the other hand, a 919G-~A transition (G307S) is
`related to a more severe clinical phenotype and has been
`detected mainly in alleles from homocystinuric patients
`of Celtic origin (Gallagher et al. 1995). A third relatively
`common CBS mutation, an IVS11-2A-~C splice muta-
`tion, which results in skipping exon 12, has been de-
`
`Patients
`
`We studied 29 patients with homocystinuria due to
`CBS deficiency, from 21 unrelated pedigrees. Patients
`were initially diagnosed on the basis of clinical mani-
`festations of homozygous CBS deficiency, in combina-
`tion with a quantitative determination of severe hyper-
`homocysteinemia and hypermethioninemia. In two
`patients (19 and 29), the diagnosis was made only on
`the basis of severe homocystinuria, which was demon-
`strated by qualitative urine analyses. Homozygous CBS
`deficiency in these two patients has been confirmed by
`detection of a homozygous mutation (patient 19) and
`by the clinical manifestation of ectopia lentis (19 and
`29), never observed in heterozygotes for CBS deficiency.
`Furthermore, the parents of patient 29 showed an ab-
`normal response to a methionine-loading test, compa-
`rable to that observed in obligate heterozygotes. Hom-
`ocysteine-lowering treatment was initiated in all patients
`immediately after the diagnosis had been made. Each
`patient was seen on a regular basis (once or twice each
`year) by two of us (Boers and Cruysberg), and the bi-
`ochemical efficacy of homocysteine-lowering therapy
`was determined by measurement of homocyst(e)ine and
`methionine in blood. At diagnosis and during long-term
`follow-up, clinical manifestations were recorded by
`means of routine clinical, radiographic, or scintigraphic
`examination procedures.
`
`Pyridoxine Responsiveness
`
`Pyridoxine responsiveness was examined after 6
`weeks of treatment with vitamin P~, 750 rag/day in
`adults or 200-500 rag/day in children. Patients in whom
`non-protein-bound serum homocysteine had decreased
`to <20 ~mol/L, or total plasma homocysteine (protein-
`and non-protein-bound) to <50 ~mol/L, were classified
`as pyridoxine responsive. All other patients were cate-
`gorized as pyridoxine-nonresponsive homocystinurics.
`
`Sandoz Inc. IPR2016-00318
`Sandoz v. Eli Lilly, Exhibit 1116-0002
`
`

`
`Kluijtmans et al.: Molecular Basis of CBS Deficiency
`
`61
`
`Biochemical Analysis
`
`Homocystine, homocysteine-cysteine-mixed disulfide,
`and methionine levels in serum were determined as de-
`scribed by Boers et al. (1983). The amount of
`non-protein-bound homocysteine was calculated as the
`sum of twice the concentration of homocystine plus the
`concentration of the homocysteine-cysteine-mixed di-
`sulfide moiety. Since 1990, total plasma homocysteine
`concentrations (i.e., the total amount of protein- and
`non-protein-bound homocysteine moieties) were deter-
`mined as described by Te Poele-Pothoff et al. (1995).
`CBS activities in extracts of cultured fibroblasts were
`measured as initially described by Fowler et al. (1978),
`with some modifications (Boers et al. 1985). These CBS
`activities were measured with and without addition of
`1 mM PLP to the incubation mixture.
`
`Mutation Analysis
`
`The procedure for mutation analysis of the CBS gene
`in patients with homocystinuria has been described else-
`where (Kluijtmans et al. 1996a). In brief, genomic DNA
`was isolated from peripheral blood leukocytes (Miller
`et al. 1988) and stored at 4°C. Total RNA was extracted
`from cultured fibroblasts by the method of Chomczynski
`and Sacchi (1987) and was stored as an ethanol precip-
`itate at -80°C. We used 1-5 /~g total RNA for first-
`strand cDNA synthesis with Superscript II Reverse Tran-
`scriptase (Life Technologies). First-strand cDNA was
`used as a template in PCR amplification reactions to
`amplify the CBS-encoding region in multiple overlapping
`cDNA fragments. These fragments were subsequently
`sequenced on an ABI 377 automated DNA sequencer
`(Applied Biosystems) with the Taq Dye Deoxy Termi-
`nator Cycle Sequencing Kit. All cDNA fragments were
`sequenced on both strands, and mutations were con-
`firmed at the genomic DNA level by restriction enzyme
`analysis or DNA sequencing. Restriction enzymes were
`purchased from Life Technologies or from New England
`Biolabs and were used according to the manufacturers’
`recommendations. Screening for 833T-~C was per-
`formed as described elsewhere (Kluijtmans et al. 1996b).
`Using this procedure, we were able to discriminate be-
`tween the real 833T-~C carriers and those with an
`844ins68 duplication variant (Sebastio et al. 1995).
`
`Cloning and Expression of Mutations
`
`cDNA fragments containing the presumed functional
`mutations were amplified by PCR and subcloned into
`an expression cartridge as described elsewhere (de Fran-
`chis et al. 1994). Recombinant clones were selected and
`sequenced to verify the integrity of the cloned fragment.
`CBS expression was induced upon addition of isopropyl-
`~-thiogalactopyranoside (Kozich and Kraus 1992). The
`
`CBS assay was performed without addition of bovine
`serum albumin to the incubation mixture, to which cys-
`tathionine was added at a final concentration of 2 raM.
`
`Statistics
`
`Differences in clinical manifestation of CBS deficiency
`and efficacy of homocysteine-lowering treatment be-
`tween separate groups were assessed by Yates’ corrected
`x~ test. Differences in age at diagnosis, homocysteine,
`and methionine concentrations were assessed by non-
`parametric Wilcoxon-Mann-Whitney U tests. All P val-
`ues reported are two-tailed, and P < .05 was considered
`statistically significant.
`
`Results
`
`Study Group
`
`Twenty-nine homocystinuria patients from 21 pedi-
`grees were included in this study. Six families had two
`affected siblings; one family had three affected siblings.
`The male-female ratio was 16:13. The diagnosis of CBS
`deficiency was established at a mean age of 26 years
`(median, 23 years [range 4-60]).
`
`Genetic Basis of Homocystinuria
`
`The CBS gene of the patients with homocystinuria was
`analyzed for mutations, either by direct sequencing of
`reverse transcription-PCR-amplified fragments (in 14
`patients) or by RFLP analysis of genomic DNA frag-
`ments, to screen for previously recognized mutations (in
`15 patients). The molecular basis of homocystinuria was
`resolved in 25 (86%) of the 29 patients. Overall, 10
`different mutations were found (table 1), of which the
`833T-~C mutation was the most prevalent. This muta-
`tion was observed in 23 (55%) of 42 independent alleles.
`In four patients, including one in whom the entire cDNA
`was sequenced, only one mutation in heterozygous state
`was found; the mutation in the second allele has yet to
`be found (table 2).
`
`CBS Activities in Cultured Fibroblasts
`
`CBS activities were measured in extracts of cultured
`fibroblasts in 12 healthy controls, in 9 CBS-deficient pa-
`tients homozygous for I278T, and in 12 patients with
`another CBS genotype (table 2). Without PLP addition
`to the incubation mixture, the mean ( _+ SD) CBS activity
`in homozygotes was .17 ( _+ .37) nmol cystathionine/mg
`protein/h, <2.5% of the control mean (7.4 [ _+ 5.1] nmol
`cystathionine/mg protein/h). One homocystinuric pa-
`tient (18) clearly exhibited CBS activities, in the range
`of obligate heterozygotes, and has been described else-
`where (Kluijtmans et al. 1996a).
`
`Sandoz Inc. IPR2016-00318
`Sandoz v. Eli Lilly, Exhibit 1116-0003
`
`

`
`62
`
`Table 2
`
`Am. J. Hum. Genet. 65:59-67, 1999
`
`The Molecular Basis of Homocystinuria Due to CBS Deficiency
`
`CBS ACTIVITYa
`
`(NMOL CYSTATHIONINE/
`PYRIDOXINE MG PROTEIN/H)
`
`SIBSHIP
`
`PATIENT CODE RESPONSIVENESS
`
`PEP~
`
`+ PEP
`
`MUTATION
`
`Allele 1
`
`Allele 2
`
`1
`
`8
`9
`10
`11
`12
`13
`14
`
`15
`
`16
`17
`
`18
`19
`
`20
`21
`
`1
`2
`3
`4
`5
`6
`7
`8
`9
`10
`11
`12
`13
`14
`15
`16
`17
`18
`19
`20
`21
`22
`
`23
`24
`25
`26
`27
`28
`29
`
`HvE51
`HvE53
`vU58
`AB58
`RB61
`Gr32
`BGr28
`JJ68
`WSJ63
`BH50
`HH49
`MC62
`JC70
`RvD66
`HvD67
`YvH75
`AP67
`JL75
`MK66
`LK61
`LH80
`GMV29
`
`CE38
`RB58
`ADZ46
`JM61
`JU78
`JR42
`RP72
`
`+
`+
`+
`
`+
`+
`
`+
`+
`ND
`ND
`
`ND
`
`+
`
`ND
`+
`
`+
`+
`+
`+
`+
`ND
`ND
`
`ND
`.06
`ND
`.32
`ND
`0
`0
`ND
`ND
`.05
`ND
`.08
`.06
`.25
`0
`0
`0
`1.74
`.62
`.12
`.18
`.09
`
`0
`0
`.06
`0
`.07
`.03
`ND
`
`ND
`0
`ND
`.31
`ND
`.12
`.24
`ND
`ND
`.38
`ND
`1.03
`1.08
`.31
`0
`0
`.24
`2.60
`.77
`.20
`.26
`.67
`
`C165Y
`C165Y
`C165Y
`C165Y
`C165Y
`I278T
`I278T
`I278T
`I278T
`I278T
`I278T
`I152M
`I152M
`R125W
`R125W
`I278T
`C165Y
`D444N
`I152M
`I278T
`I278T
`R369C +
`R491C
`0 I278T
`0
`I278T
`.42
`V180A
`.12
`I278T
`0
`I278T
`.08
`I278T
`ND
`I278T
`
`I278T
`I278T
`I278T
`Unidentified
`Unidentified
`I278T
`I278T
`I278T
`I278T
`I278T
`I278T
`I278T
`I278T
`T434N
`T434N
`I278T
`C165Y
`D444N
`I152M
`I278T
`I278T
`R369C +
`R491C
`I278T
`I278T
`Unidentified
`V371M
`I278T
`V371M
`Unidentified
`
`NOTE.--A plus sign (+) indicates pyridoxine responsive, and a minus sign ( ) indicates pyridoxine non-
`responsive; ND not determined.
`a Cystathionine f!-synthase activity expressed in nmol cystathionine formed/mg protein/h.
`b Mean ( 2 SD) CB$ activity in fibroblasts of healthy controls: 7.4 ( 2 5.1).
`
`The mean CBS activity in fibroblasts of homozygotes
`for I278T (n = 9) versus patients with other genotypes
`(n = 12) was 0.05 _+ 0.07versus 0.13 _+ 0.18 nmol cys-
`tathionine/mg protein/h (P = .28), and 0.13 _+ 0.14 ver-
`sus 0.42 m 0.38 nmol cystathionine/mg protein& (P =
`.07) in the assay without and with, respectively, 1 mM
`PLP. Patient 18 was excluded in these calculations.
`
`In Vitro Expression of Mutations
`
`The functional relevance of seven mutations was in-
`vestigated in an Escbericbia coli expression system: six
`constructs contained a single mutation, and one con-
`struct contained the ll05C~T (R369C) and 1471C~T
`(R491C) mutations in cis. All mutated constructs, except
`the one containing 1330G~A (D444N) (Kluijtmans et
`al. 1996a), showed a reduction in CBS activity of>90%,
`demonstrating the detrimental effects of each construct
`on CBS activity (fig. 1). Two mutations (373C~T
`
`[R125W] and 1301C-~A [T434N]) have not yet been
`functionally tested.
`
`Pyddoxine Responsiveness
`
`Fourteen (48%) of 29 patients were classified as pyr-
`idoxine responsive and 9 (31%) patients as pyridoxine
`nonresponsive, on the basis of the criteria described in
`the Patients, Material, and Methods section. In six pa-
`tients (12, 13, 16, 21, 28, and 29), responsiveness to
`pyridoxine alone could not be assessed. In view of their
`extremely high homocysteine levels at diagnosis, these
`patients were treated directly with a combination of ther-
`apeutic regimens (pyridoxine and folic acid, with or
`without betaine). Pyridoxine responders were diagnosed
`at a mean age of 29 years (median, 26 years [range,
`7-54 years]), and nonresponders at a mean age of 19
`years (median, 16 years [range, 4-30 years]; P = .08).
`Seven (58%) of 12 homozygotes for the I278T mu-
`
`Sandoz Inc. IPR2016-00318
`Sandoz v. Eli Lilly, Exhibit 1116-0004
`
`

`
`Kluijtmans et al.: Molecular Basis of CBS Deficiency
`
`63
`
`rosis, scoliosis, or Marfanoid features, and 13 (45%)
`patients had complications in the vascular system. The
`CNS was involved in 16 (55%) patients, and psychiatric
`illness was noticed in only 4 (14%) patients. No signif-
`icant differences in clinical presentation of CBS defi-
`ciency were observed between I278T homozygotes and
`patients with other genotypes (data not shown).
`
`Biochemical Characteristics at Diagnosis
`
`At diagnosis, non-protein-bound homocysteine con-
`centrations had been measured in 21 patients and total
`homocysteine concentrations in 6 patients. In two pa-
`tients, no baseline homocysteine blood levels were avail-
`able; in these two cases the initial diagnosis had been
`made in 1975, by means of qualitative examination of
`urine only. The mean non-protein-bound homocysteine
`concentration was 135/~mol/L (range, 42-266/~mol/L;
`n = 21) and mean total plasma homocysteine concen-
`tration was 240/~mol/L (range, 134-299 /~mol/L; n =
`6). Mean serum methionine concentration was 130
`/~mol/L (range, 52-549/~mol/L; n = 23). Homozygotes
`for the I278T mutation tended to have a higher mean
`homocysteine concentration than homocystinuria pa-
`tients with other genotypes (160 _+ 73 /~mol/L versus
`116 _+ 57 /~mol/L non-protein bound; P = .16),
`whereas methionine concentrations were not signifi-
`cantly different between both genotype groups (89 _+
`35/,mol/L versus 152 _+ 166/,mol/L; P = .49).
`
`Response to Homocysteine-Lowering Treatment
`
`Long-term homocysteine-lowering therapy (mean
`term, 13 years [range, 1-29 years]) consisted of maxi-
`mally 750 mg pyridoxine. Fifteen (56%) patients were
`concomitantly treated with 5 mg folic acid per day, and
`eight (30%) patients also with 6 g betaine per day. Only
`one patient (17) had a methionine-restricted diet, with
`a methionine content of 600 mg/d. Intramuscular injec-
`tions with 1 mg hydroxycobalamin every 1-2 mo were
`given to four patients, because of the development of a
`vitamin B,2 deficiency. Two patients could not be fol-
`lowed after diagnosis had been made: patient 6 refused
`treatment and patient 13 moved to another country.
`Hence, follow-up was recorded in 27 (93%) of the pa-
`tients with homocystinuria. The mean length of follow-
`up was 11 years (range, 1-20 years; n = 11) in homo-
`zygotes for the I278T mutation and 16 years (range,
`3-27 years; n = 16) in patients with other genotypes
`(P = .09).
`Biochemically, pyridoxine treatment resulted in a
`marked decrease in homocysteine concentrations of
`90% in homozygotes for the I278T mutation and of
`67% in patients with other genotypes (P<.02). Ex-
`tended intervention with folic acid, with or without be-
`taine, further decreased homocysteine concentrations in
`
`Sandoz Inc. IPR2016-00318
`Sandoz v. Eli Lilly, Exhibit 1116-0005
`
`Figure I CBS activities measured in an E. coli expression system.
`Mutations were introduced as described in the Patients, Material, and
`Methods section. Bacterial lysates were assayed for CBS activity, and
`the mean CBS activity of a control construct (40.3 nmol cystathionine
`formed/mg protein/h) has been set to 100%,. Each bar represents an-
`other mutated construct and is the mean of two independently per-
`formed CBS assays.
`
`tation showed in vivo pyridoxine responsiveness, 3
`(25%) were nonresponders, and, in 2 (17%) patients,
`this specific responsiveness could not be assessed. In 17
`individuals with other genotype combinations, including
`compound heterozygotes for I278T, these numbers were
`7 (41%), 6 (35%), and 4 (24%), respectively (x2=
`0.13; P = .7). Conversely, in 14 pyridoxine-responsive
`patients, 18 (64%) of 28 alleles carried the I278T mu-
`tation (7 homozygotes, 4 heterozygotes), versus 6 (33%)
`I278T alleles in 9 nonresponsive patients (3 homozy-
`gotes) (X2 = 1.39; P = .5). There was an absolute con-
`cordance of pyridoxine responsiveness between siblings.
`
`Clinical Characteristics at Diagnosis
`
`The clinical manifestation of homocystinuria due to
`CBS deficiency in this study group is very diverse and is
`depicted in detail in table 3. At diagnosis, some patients
`(2, 13, and 15) showed virtually no clinical symptoms
`and were investigated because of a homocystinuric sib-
`ling. In other patients (6, 19, 22, 23, and 25), all four
`major organ systems were involved. Ocular abnormal-
`ities and skeletal abnormalities were the most consistent
`findings among these 29 patients with homocystinuria:
`25 (86%) suffered from either high myopia or ectopia
`lentis. Twenty-six (90%) patients exhibited osteopo-
`
`

`
`64
`
`Table 3
`
`Am. J. Hum. Genet. 65:59-67, 1999
`
`Clinical Presentation of CBS Deficiency at Diagnosis
`
`CLINICAL PRESENTATION
`
`SIBSHIP PATIENT Myopia
`
`Ectopia Osteo- Marfanoid Arterio- Thrombo- Mental
`Lentis porosis Scoliosis Featuresa sclerosisb embolism~ Retardationa Convulsions Psychosis Depression
`
`÷
`
`+
`+
`
`+
`
`÷
`
`÷
`
`1
`
`2
`
`3
`
`4
`
`5
`
`6
`
`7
`
`8
`9
`10
`11
`12
`13
`14
`15
`16
`17
`18
`19
`20
`21
`
`1
`2
`3
`4
`5
`6
`7
`8
`9
`10
`11
`12
`13
`14
`15
`16
`17
`18
`19
`20
`21
`22
`23
`24
`25
`26
`27
`28
`29
`
`+
`
`+
`+
`
`+
`+
`+
`+
`+
`+
`+
`
`+
`
`+
`+
`
`+
`+
`+
`+
`+
`+
`+
`+
`+
`+
`+
`
`a Manifested by at least extreme height and arachnodactyly.
`b Peripheral arterial disease in patients 7, 10-12, and 28; cerebrovascular disease in patients 12 and 25; and coronary artery disease in patients
`6 and 28.
`~ Deep venous thrombosis in patients 1, 7, 19, 23, and 25; pulmonary embolia in patients 9 and 22; and sinus sagittalis thrombosis in patient
`21.
`d Severe mental retardation in patients 17 and 19; mild mental retardation (not less than 80) in 10 other patients, as indicated.
`
`both genotype groups, by 84% and 47%, respectively
`(P < .05). Homocysteine concentrations normalized (i.e.,
`non-protein-bound homocysteine <20/xmol/L, or total
`plasma homocysteine <50 /xmol/L) in 21 (78%) of 27
`patients. Methionine concentrations decreased in both
`genotype groups upon pyridoxine therapy, by 56% and
`33%, respectively (P = .15). Additional treatment with
`folic acid and betaine, to stimulate homocysteine re-
`methylation, left methionine concentrations in homo-
`zygotes for the I278T mutation virtually unchanged,
`whereas methionine levels substantially increased in pa-
`tients with other genotypes.
`During long-term treatment, 9 (82%) of 11 homo-
`zygotes for I278T continuously presented non-protein-
`bound homocysteine concentrations <20/xmol/L or total
`homocysteine concentrations <50/xmol/L, versus only 6
`(38%) of 16 patients with another genotype (P = .06).
`Mean ( _+ SD) non-protein-bound homocysteine during
`
`long-term follow-up was 13 _+ 13 /~mol/L in I278T
`homozygotes versus 23 _+ 18 /xmol/L in patients with
`other genotypes (P = .31). For total homocysteine levels,
`these concentrations were 30 _+ 16 /xmol/L and 67 _+
`38/xmol/L, respectively (P < .01 ). Despite extended ther-
`apy, four patients (14, 15, 17, and 18), all without the
`I278T mutation, persistently had total plasma homo-
`cysteine >100/xmol/L.
`On the basis of genotype classification, 121 patient-
`years of treatment were recorded in 11 homozygotes for
`the I278T mutation, during which, in one patient (7),
`peripheral arteriosclerosis, already present at diagnosis,
`proceeded to development of an abdominal aortic an-
`eurysm that required vascular surgery. During 257 pa-
`tient-years of treatment in patients with another geno-
`type, one patient (25) who suffered from a
`cerebrovascular event before diagnosis died, at age 42
`years, of a myocardial infarction. In summary, in 378
`
`Sandoz Inc. IPR2016-00318
`Sandoz v. Eli Lilly, Exhibit 1116-0006
`
`

`
`Kluijtmans et al.: Molecular Basis of CBS Deficiency
`
`65
`
`patient-years of homocysteine-lowering treatment, only
`two vascular events were recorded. From the data of
`Mudd et al. (1985), an expected number of primary
`vascular events of 2 per 25 years can be calculated for
`untreated patients. Without treatment in our group of
`27 homocystinurics, at least 30 vascular events would
`have been expected (P < .01). During long-term follow-
`up, no first involvement of any involved organ system
`was noticed. Furthermore, no deterioration of clinical
`symptoms was recorded, except for recurrent depressive
`episodes in two patients (2 and 15).
`
`Discussion
`
`In the present study, 42 independent homocystinuric
`alleles were investigated, which led to the identification
`of 10 different mutations (table 1). The 833T~C (I278T)
`mutation, first recognized by Kozich and Kraus (1992)
`in heterozygous state in a pyridoxine-responsive patient,
`was found to be the most prevalent mutation in Dutch
`homocystinurics: 23 (55%) of the 42 independent Dutch
`homocystinuric alleles carried this mutation.
`Seven mutated constructs have thus far been func-
`tionally assayed in an E. coli expression system, which
`shows that all mutations tested, except D444N (Kluijt-
`mans et al. 1996a), had a detrimental effect on CBS
`activity. The latter indicates that these mutations are
`likely disease-causing ones and not benign polymor-
`phisms. The R125W and T434N mutations, detected in
`the two siblings of pedigree 7, still need to be analyzed
`in an expression system. Another mutation (R125Q) af-
`fecting the same arginine residue has been found inde-
`pendently by two other groups and was shown to in-
`activate CBS completely (Marble et al. 1994; Sebastio
`et al. 1995). The R369C and R491C mutations in cis
`showed a relatively high residual CBS activity of ~10%
`of the control mean and are compatible with a relatively
`late onset of disease. Recently, R369C was postulated
`to be a rare polymorphism not affecting CBS function
`in a yeast expression system (Kim et al. 1997). The sep-
`aration of both mutations by in vitro mutagenesis and
`expression in an E. coli expression system will clarify
`their individual effect on CBS activity.
`We did not detect the 919G~A mutation, which is
`very frequent among patients of Celtic origin with pyr-
`idoxine-nonresponsive homocystinuria (Gallagher et al.
`1995). A 797G~A mutation, a frequent cause of pyri-
`doxine-responsive homocystinuria in Norwegian pa-
`tients (Kim et al. 1997), was also not observed in these
`Dutch patients with homocystinuria. In a survey of 14
`Italian families with predominantly pyridoxine-respon-
`sive homocystinuria patients, Sebastio et al. (1995) also
`did not detect the Celtic 919G-~A mutation and fre-
`quently observed the 833T-~C and 341C-~T transitions
`(in 9 and 4 of 36 alleles, respectively). Unlike 833T-~C,
`
`the other common CBS mutations may be of more recent
`genetic origin and have not spread through different
`populations yet. On the other hand, our observation of
`different haplotypes in homozygotes for 833T~C mu-
`tation indicates that there is not one common ancestor
`haplotype in these Dutch patients (data not shown).
`At diagnosis, no significant differences were observed
`in either biochemical or clinical expression of homo-
`cystinuria between homozygotes for I278T and patients
`with another genotype. Obviously, the homocystinuric
`phenotype is not only dependent on CBS genotype but
`is also influenced by other genetic and/or environmental
`factors. Recently, both factor V Leiden (Mandel et al.
`1996) and thermolabile methylenetetrahydrofolate re-
`ductase (Kluijtmans et al. 1998) have been postulated
`as genetic factors modifying thromboembolic risk in
`homocystinurics, and, obviously, other factors may exist
`as well. Furthermore, although we did not observe mul-
`tiple mutations in cis by sequencing analysis (except in
`patient 22), additional mutations in certain alleles might
`have influenced our genotype-phenotype correlation and
`might explain why our I278T homozygotes are more
`severely affected than those described by Shih et al.
`(1995).
`In the present study, we subdivided our patients into
`only two groups with regard to pyridoxine responsive-
`ness, to maintain statistical power in our analysis. On
`the basis of their biochemical in vivo response to large
`doses of pyridoxine, 48% (n = 14) of the Dutch hom-
`ocystinurics were categorized as responsive and 31%
`(n = 9) as nonresponsive. However, according to the
`classification of Brenton and Cusworth (1971), only two
`(7%) patients (17 and 18) of these nine would be un-
`equivocal nonresponders; the others (24%) would be
`considered partially pyridoxine responsive. Not only do
`our observations strongly deviate from those of Mudd
`et al. (1985), who observed virtually equal proportions
`of pyridoxine-responsive and -nonresponsive patients,
`but they also clearly indicate the relatively mild nature
`of CBS deficiency in the Netherlands. The absence, in
`the Netherlands, of a newborn screening program for
`homocystinuria, by which patients with pyridoxine-non-
`responsive homocystinuria, in particular, are detected
`(Mudd et al. 1985, 1995), might have contributed to
`this aberrant distribution. Nonresponders may have re-
`mained undiagnosed and lost to follow-up, because of
`early mortality or admittance to an institution for men-
`tally retarded patients. The high prevalence of I278T
`among Dutch patients with homocystinuria might also
`explain the high pyridoxine responsiveness among these
`individuals. Seven homozygotes for this mutation
`showed pyridoxine responsiveness, and three individuals
`were classified as nonresponders. The latter finding
`seems to be in contradiction with the finding of pyri-
`doxine responsiveness in three homocystinuria patients
`
`Sandoz Inc. IPR2016-00318
`Sandoz v. Eli Lilly, Exhibit 1116-0007
`
`

`
`66
`
`Am. J. Hum. Genet. 65:59-67, 1999
`
`homozygous for the I278T mutation reported elsewhere
`(Sebastio et al. 1995; Shih et al. 1995), although in those
`reports no clear definitions of responsiveness were pro-
`vided. The homocysteine concentrations upon pyrido×-
`ine treatment in our I278T homozygotes classified as
`nonresponders were only marginally above the cut-off
`values of