Newborn screening for cerebrotendinous xanthomatosis is the solution for early identification and treatment[S]

Cerebrotendinous xanthomatosis (CTX) is a progressive metabolic leukodystrophy. Early identification and treatment from birth onward effectively provides a functional cure, but diagnosis is often delayed. We conducted a pilot study using a two-tier test for CTX to screen archived newborn dried bloodspots (DBSs) or samples collected prospectively from a high-risk Israeli newborn population. All DBS samples were analyzed with flow injection analysis (FIA)-MS/MS, and 5% of samples were analyzed with LC-MS/MS. Consecutively collected samples were analyzed to identify CTX-causing founder genetic variants common among Druze and Moroccan Jewish populations. First-tier analysis with FIA-MS/MS provided 100% sensitivity to detect CTX-positive newborn DBSs, with a low false-positive rate (0.1–0.5%). LC-MS/MS, as a second-tier test, provided 100% sensitivity to detect CTX-positive newborn DBSs with a false-positive rate of 0% (100% specificity). In addition, 5β-cholestane-3α,7α,12α,25-tetrol-3-O-β-D-glucuronide was identified as the predominant bile-alcohol disease marker present in CTX-positive newborn DBSs. In newborns identifying as Druze, a 1:30 carriership frequency was determined for the c.355delC CYP27A1 gene variant, providing an estimated disease prevalence of 1:3,600 in this population. These data support the feasibility of two-tier DBS screening for CTX in newborns and set the stage for large-scale prospective pilot studies.

community in Galilee (15). In this community a 1:11 carrier frequency was determined for the pathogenic founder c.355delC CYP27A1 gene variant (15).
A low-risk effective oral therapy for CTX is available in the form of CDCA and CA, the main bile acids deficient in CTX. Treatment with CDCA is considered the standard of care for CTX and has been shown to normalize the biochemical abnormalities and halt progression of the disease (16,17). Treatment of infants and children affected with CTX has been described, including treatment of presymptomatic cases (6,(18)(19)(20). In a limited number of children, treatment with CA was reported as a less hepatotoxic alternative to CDCA (6). Treatment of patients with advanced neurological disease does not appear to reverse the neurological impairment (17). A recent study demonstrated that patients who began CDCA treatment after age 25 were significantly more limited in ambulation and more cognitively impaired compared with those who started treatment earlier (21). Therefore, it is essential to diagnose and treat CTX as early as possible, something that will likely only be achieved through implementation of universal newborn screening for this disorder.
Biochemical approaches to screen newborn dried bloodspots (DBSs) for CTX have recently been described (22)(23)(24). The defect in CTX results in the accumulation of bile acid precursors, including the ketosterols, 7-hydroxy-4-cholesten-3-one and 7,12-dihydroxy-4-cholesten-3-one (712C4). DeBarber and colleagues have demonstrated that elevated 712C4 can be measured in CTX newborn DBSs using LC-MS/MS (22,23). Analysis of this marker can be used to discriminate CTX-positive from CTX-negative newborn DBSs (22,23), although the necessity for LC and the cost of ketosterol derivatization reagent limits the applicability of this methodology as a first-tier test. The ketosterol, 712C4, is metabolized to form bile alcohol species, such as 5-cholestane-3,7,12-triol, or is further hydroxylated to form 5-cholestane-tetrol isomers. The 5cholestane-3,7,12,25-tetrol was reported to be the highest concentration tetrol isomer present in CTX blood (25,26), primarily present in the form of a glucuronide conjugate (25,26). Recently, in a validation study using 200 premature and term newborn DBSs, Vaz et al. (24) demonstrated that negative mode flow injection analysis (FIA)-MS/ MS analysis of elevated 5-cholestane-tetrol glucuronide isomer species (CTGS) could be used to screen newborn DBSs for CTX. As CTGS can also be elevated in cholestatic liver disease, the metabolite ratio of CTGS over tauro (t)dihydroxycholanoic acid species (predominantly tCDCA) was used to enhance the sensitivity of testing, a ratio strongly elevated in CTX, but not in cholestasis (24).
Here, we present biochemical and genetic data from a prospective pilot study screening DBSs from a high-risk newborn population for CTX.

Human subject research considerations and sample collection
Newborn DBSs were obtained from the archives of the Israeli Ministry of Health Central Newborn Screening Laboratory or were prospectively collected from newborns born at two hospitals in the Northern Israel region, the Galilee and Ziv Medical Centers. Informed consent was obtained from all families to test the samples for CTX. In the prospective group, additional DBS samples were collected at the time of routine newborn screening. Families were informed that, in the event of a CTX-positive determination, they would be contacted for diagnostic confirmation, genetic counseling, and medical care to be provided by the Galilee Medical Center Institute for Human Genetics. The study was approved by the Institutional Review Boards (IRBs) of the hospitals and by the central IRB committee of the Israeli Ministry of Health. Non-newborn CTX-positive DBSs were also obtained from participants enrolled in an IRB-approved study at Oregon Health and Science University (OHSU) with informed consent obtained from all study participants. Residual blood samples submitted to the OHSU Sterol Analysis Diagnostic Laboratory for confirmation of CTX were also used with IRB approval.

DNA extraction from DBSs and genetic analysis
Genomic DNA was extracted using 3 × 3 mm punches with a QIAamp DNA MicroKit (QIAGEN, Hilden, Germany) following the protocol for isolation of genomic DNA from DBSs.
The genomic DNA samples were genetically screened for the common pathogenic genetic variants found in the Druze or Moroccan Jewish populations as previously described (13,14,27). In brief, this consisted of DNA amplification followed by restriction enzyme analyses. Specifically, for the Druze population, CTX allele detection was accomplished by amplification of a 262 bp fragment from genomic DNA containing the c.355delC variant in exon 2 of CYP27A1, with sense primer (5′-GCTTGGCCCAGT-TATTCAGTTTTG-3′) and antisense primer (5′-GCCCTGTTC-CAGTCCCTTCAG-3′). The PCR products were subsequently digested with Fnu4HI restriction enzyme (New England Biolabs, Ipswich, MA) and analyzed on 3.5% agarose gels. For Moroccan Jews, all three known pathogenic genetic variants were tested; the c.1016 C>T (pThr339Met) variant by amplification of a 382 bp fragment from exon 5 in CYP27A1 using sense primer (5′-TGCTTTCACAGGGAAGAAGC-3′) and antisense primer (5′-ACAAAAGGACCATCCCCTGT-3′). The PCR products were digested with HpyCH4III restriction enzyme (New England Biolabs) and analyzed on 3.5% agarose gels. To detect the other two Moroccan Jewish genetic variants (14), direct sequencing of the targeted variants in exon 4 was performed using sense primer (5′-GAGGACACCGTGACCTTCGT-3′) and antisense primer (5′-TAGGGAACTGGTTCAGGTTG-3′). Sequences were determined using the BigDye Terminator cycle sequencing kit (Applied Biosystems, Foster City, CA) and analyzed using the ABI PRISM 3500 XL genetic analyzer (Applied Biosystems).

Preparation of samples for MS/MS analysis
Quality control (QC) and calibrator DBS samples were generated using working dilutions of 5-cholestane-25-tetrol-glucuronide (from stock solution stored in 85:15 acetonitrile:water) and 712C4 in methanol. The working dilutions were spiked into whole blood that was spotted onto filter paper, dried, and stored desiccated at 80°C. Glycine and taurine conjugates of CDCA and CA were measured using calibrators generated by spiking charcoal-stripped serum with authentic standards spotted onto filter paper and dried. An internal standard solution containing 50 ng of CDCA-d5-24-glucuronide, 1 ng of 712C4-d9, and 0.5 ng each of t-CDCA-d5 and glyco-CA in 300 l of methanol was added to each DBS. Punches (3.2 mm) from calibrators, QCs, blanks, and unknown samples were extracted in 96-well plates by shaking for 75 min at 45°C. After centrifugation at 3,000 rpm for 5 min, 200 l of the supernatant were transferred to a new plate and dried using a SpeedVac. The extract was resuspended in 200 l of a 75:25 mixture of water:acetonitrile before analysis. QC samples were analyzed with each plate for FIA-MS/MS and LC-MS/MS analysis; calibrators were analyzed for LC-MS/MS analysis.

LC-MS/MS method
Polarity switching was used for LC-MS/MS analysis. The ESI interface negative mode settings and MRM transitions monitored for quantification were the same as for FIA-MS/MS. The ESI interface was also operated in the positive mode with the following settings: CUR 35, TEM 600, GS1 40, GS2 20, IS 3,000, and CAD medium. Positive mode MRM transitions monitored for quantification of 712C4 were as follows: 712C4, m/z 417.3→253.1; and 712C4-d 9 , m/z 426.3→253.1. LC-MS/MS analysis was performed using a 50 × 2.1 (i.d.) mm, 1.8 m HSST3 column (Waters, Milford, MA). The 3 min gradient method used mobile phase [solvent A (water) and solvent B (acetonitrile), both at 0.1% formic acid] delivered at 0.5 ml/min. Solvent B was increased from 25% to 90% over 1 min. The column was washed at 90% solvent B for 1.25 min and then equilibrated for 0.75 min. The column temperature was 50°C.

Newborn DBS sample collection
Three CTX-positive DBSs from Druze newborns known to be homozygous for the pathogenic founder c.355delC CYP27A1 gene variant (27), as well as known carrier newborn DBSs, were retrieved from the Israeli Newborn Screening Laboratory archives. The newborn DBSs had been stored for 3-8 years refrigerated (4°C) and were from children previously identified as being affected with CTX or as carriers through prenatal genetic testing (or testing shortly after birth), performed as part of a community-wide screening program for genetic diseases established and overseen in the region by Falik-Zaccai et al. (15). The archived newborn DBSs were analyzed along with DBSs prospectively collected from a high-risk newborn population born at the Galilee and Ziv Medical Centers.

Genetic analysis
Genetic analysis of CTX-causing founder genetic variants common among the Druze and Moroccan Jewish populations was performed for around 55% of newborn DBS samples consecutively collected in a prospective manner. Among 366 Druze newborn DBS samples tested for the c.355delC CYP27A1 gene variant, 13 carrier samples were identified, providing a frequency of 1:30 carriership in this sample of Druze-origin population. Based on the Hardy-Weinberg equation, these data predict a disease prevalence of 1:3,600 in the Druze population screened for pilot study.

FIA-MS/MS method performance
QC samples were generated by spiking low tCDCA whole blood with 5-cholestane-25-tetrol-glucuronide authentic standard at nominal concentrations of 500 and 5,000 ng/ml (low and high QCs). The mean observed GTCS/tCDCA ratio for low QC was close to the lowest metabolite ratio observed for a CTX newborn sample and for high QC ratio was comparable to a mid-range metabolite ratio for a CTX adult sample. The between-run precision data for the GTCS/tCDCA ratio determined for QC samples injected across different FIA-MS/MS batches was <24% relative standard deviation (RSD). Based on the precision data, a reasonable ratio cut-off for first-tier screening using our instrumentation was designated as 0.361 (70% of lowest CTX newborn ratio observed).

LC-MS/MS method performance
To calculate analyte concentrations, calibrators were run with each batch and curves generated using a weighted (1/x) least-squares linear regression for peak area ratio (analyte/internal standard) plotted against nominal concentration. Calibration curves for measurement of 5cholestane-25-tetrol-glucuronide in DBSs demonstrated linearity across the range of 250-10,000 ng/ml with correlation coefficients r 2 > 0.95. The between-run accuracy was within ±22% and precision <21% for mean 5-cholestane-25-tetrol-glucuronide concentration measured across LC-MS/MS batches (see supplemental Table S1). This data was generated using CDCA-d5-24-glucuronide as internal standard. Stable isotope-labeled 5-cholestane-25-tetrolglucuronide internal standard was not commercially available, but would likely improve the method performance.
Autosampler carryover was acceptable with <20% peak area for 250 ng/ml calibrator present after injection of 10,000 ng/ml calibrator (the highest concentration in untreated CTX-positive DBSs was 10,154 ng/ml).
A stock solution of 5-cholestane-25-tetrol-glucuronide in acetonitrile:water (85:15) was stable frozen at 80°C for at least 6 months (compared with solid compound stored at 80°C). The 5-cholestane-25-tetrol-glucuronide spiked into whole blood and spotted onto filter paper was stable frozen at 20°C for up to 6 months. For DBSs stored at 20°C or 40°C, the concentrations for weekly time points were within ±20% of the initial concentration for up to 3 weeks.
Matrix effects for LC-MS/MS quantitation of 5cholestane-25-tetrol-glucuronide were determined by comparing the peak area for standard spiked into filter-paper extract to peak area for standard spiked into DBS extracts from unaffected individuals (29). Matrix effects were observed within ±34% with a peak area RSD <15% (n = 4 different individuals). When these samples were analyzed using FIA-MS/MS, increased ion suppression was observed for 5-cholestane-25-tetrol-glucuronide relative to tCDCA, indicating that a stable isotope-labeled internal standard analog should optimally be used for each compound.  Potential method interference for LC-MS/MS quantitation of 5-cholestane-25-tetrol-glucuronide was evaluated by examination of the peak shape, shoulder, and area ratio for two MRM transitions acquired [quantifier and qualifier ions (30); m/z 75 and 85, respectively]. Data from reversedphase analysis of CTX-positive DBSs demonstrated that multiple putative CTGS peaks were extracted (with the same precursor ion mass and characteristic m/z 75 and 85 glucuronide product ions; see Fig. 1). The predominant CTGS isomer present co-chromatographed with authentic 5-cholestane-25-tetrol-glucuronide standard using a fast 3 min gradient, or shallower gradient performed over 30 min, and possessed similar MS/MS product ion data (see supplemental Fig. S1); nevertheless there is the possibility of other isomeric species with alternate stereochemistries co-eluting.
The percent distribution of CTGS calculated using m/z 75 product ion peak area was similar in treated or untreated CTX-positive DBSs, including newborn CTX-positive DBSs (see Table 1). As authentic standards were not available for each CTGS, the percent distribution value assumes similar ionization efficiency and m/z 75 product ion fragment intensity for each species (in support of this, the ion ratio for m/z 75 and 85 product ions was comparable for each CTGS).

FIA-MS/MS analysis of CTGS/tCDCA ratio in DBSs as a test for CTX
FIA-MS/MS analysis was used in a blinded manner to determine the CTGS/tCDCA ratio in all DBS samples (see Table 2 and Fig. 2). CTX-positive newborn DBSs from the Newborn Screening Laboratory archives could be identified with 100% sensitivity using the CTGS/tCDCA ratio, but with a very low false positive rate (0.1-0.5% depending where the cut-off value was placed). The lowest ratio of CTGS/tCDCA determined in CTX-positive newborn DBSs was 0.516 and the highest ratio in unaffected newborn DBSs was 0.634, with overlap between the ratios in CTXpositive and unaffected newborn DBSs. Around 5% of samples, including samples with a ratio >0.361, were subjected to LC-MS/MS analysis.

LC-MS/MS analysis of 5-cholestane-25-tetrolglucuronide, tCDCA, and 712C4 in DBSs as a test for CTX
Use of LC-MS/MS to determine the 5-cholestane-25tetrol-glucuronide/tCDCA ratio, and to quantify 712C4 in DBSs, provided 100% sensitivity to detect the CTX-positive newborn samples with no overlap between the 5-cholestane-25-tetrol-glucuronide/tCDCA ratio and the 712C4 concentration determined for CTX-positive compared with other newborn DBSs (see Table 3 and Fig. 2). LC-MS/MS allowed identification as positive 3/3 DBSs from CTX-affected newborns, 0/9 DBSs from identified carrier newborns, 0/15 cholestasis newborn DBSs, and 0/61 unaffected newborn DBSs. Additional test validation data were generated by analyzing DBSs from children/adults affected with CTX (see Tables 2 and 3) and peroxisome biogenesis disorders in the Zellweger spectrum (PBD-ZSDs). DBSs from individuals with PBD-ZSDs may have an elevated CTGS/ tCDCA ratio (23). These samples can be discriminated from CTX-positive DBSs using the tauro-trihydroxycholestanoic acid to CTGS ratio (a metabolite ratio decreased in CTX, but not in PBD-ZSDs) (23). Our data indicate that 712C4 was not elevated in these samples and might also be a useful marker to discriminate PBD-ZSDs from CTXpositive DBSs (see Fig. 2C and the numeric data in Table 3).

DISCUSSION
We present here feasibility data from a prospective pilot study screening DBSs from a high-risk newborn population for CTX that supports use of a two-tier screening approach (the proposed screening algorithm is outlined in Fig. 3). The approach consists of first-tier negative mode FIA-MS/MS analysis to determine the metabolite ratio CTGS/tCDCA as described by Vaz et al. (24). This method provided 100% sensitivity to detect CTX-positive newborn samples with a false positive rate for unaffected newborn DBSs of 0.1-0.5% depending on the cut-off used. For this validation study, a subset of around 5% of samples was subjected to analysis with LC-MS/MS (including samples with a ratio >0.361). LC-MS/MS analysis provided 100% sensitivity to detect CTX-positive newborn samples with a 0% false positive rate (100% specificity). Any positive samples identified using second-tier LC-MS/MS testing would undergo confirmatory CYP27A1 gene sequencing (Fig. 3). Adoption of a twotier approach has proven to be a successful means to reduce false positive results for other newborn screening applications (31) and is considered an acceptable solution to lower false positive rates that could otherwise pose a burden on newborn screening laboratories, clinicians, and families.
Newborn screening programs routinely ascertain, as a quality measure, whether patients diagnosed with disorders later in life escaped detection by newborn screening  Data are expressed as mean (range of results) for m/z 75 product ion. Bold type indicates the 5-cholestane-25-tetrol-glucuronide/tCDCA ratio values that are provided in Fig. 2A.
a Although nine carrier samples overall were identified as part of the MS/MS data analysis (seven prospectively, two previously identified and retrieved from archives), based on the carriership frequency determined for c.355delC CYP27A1 gene variant, we would estimate at least 16 additional unidentified carrier samples to be present in the Druze DBSs biochemically screened.
(false negatives). The Galilee Medical Center is a regional center of expertise for treatment of CTX. To the best of our knowledge, and with the caveat that CTX is not routinely diagnosed in infancy or childhood, no new pediatric CTX cases were identified in the Northern Galilee region since the pilot study started.
A limitation of this pilot study is that, although eighthinch (3.2 mm) DBS punches were analyzed, a more sensitive LC-MS/MS instrument than found in many newborn screening laboratories was used for the study. In the Vaz et al. (24) pilot study, a lower sensitivity LC-MS/MS instrument was used, but quarter-inch DBS punches were required.
LC-ESI-MS/MS methodology was previously described utilizing keto derivatization to enable sensitive isotope dilution measurement of 712C4 in DBSs from CTX-positive newborns (22,23). One advantage of this approach was the ability to synthesize stable isotope-labeled internal standard using labeled derivatization reagent (22). For the pilot study, we used custom-synthesized 712C4-d9 internal standard for isotope dilution measurement of 712C4 without keto derivatization. Including the data we report here, elevated 712C4 has been measured in six CTXpositive newborn DBSs (with a mean value of 222 ng/ml, range 42-431 ng/ml; and a mean control value of 10.2, Data are expressed as mean (range of results) for m/z 75. Bold type indicates the 5-cholestane-25-tetrol-glucuronide/tCDCA ratio and 712C4 that are provided in Fig. 2B and C, respectively. a Outside validated linear analytical measurable range. b For one CTX-positive newborn sample, the mean calculated tCDCA concentration was 262 ng/ml (from n = 3 replicate analyses), with a resultant mean 5-cholestane-25-tetrol-glucuronide/tCDCA ratio of 0.4546 (RSD 29.9%, n = 3). The mean 5-cholestane-25-tetrol-glucuronide concentration for this sample was 492 ng/ml (RSD 9.7%, n = 3) and mean 712C4 concentration was 46.5 ng/ml (14.5%, n = 3).
c All non-CTX samples were <250 ng/ml 5-cholestane-25-tetrol-glucuronide except one carrier cholestasis sample with a 5-cholestane-25-tetrolglucuronide value of 728 ng/ml. Fig. 3. Two-tier screening approach to screen newborn DBSs for CTX. If the second-tier biochemical positive result is confirmed by genetic testing, or the genetic testing result is ambiguous, families should be contacted for follow-up, which should provide diagnostic and medical evaluation, genetic counseling, and referral to all needed specialists. range 0-16.5). The laboratory performance metrics for this disease marker thus far are: true positive cases identified, six; false positive cases, zero.
Measurement of 5-cholestane-25-tetrol-glucuronide and 712C4 in blood may be a useful tool for diagnosis and therapeutic monitoring in CTX (see data for untreated and treated CTX adults/children in Table 3). Measurement of 5-cholestanol in blood has routinely been used for this purpose and is offered by a number of clinical laboratories worldwide. More recently, measurement of 712C4 has been made available for clinical use as a more specific test for diagnosis of CTX (health care providers seeking information and assistance should contact the authors).
From 2003 onward, Falik-Zaccai et al. (15) at the Galilee Medical Center participated in extensive outreach efforts in nearby Druze-populated regions, using genetic testing to systematically screen communities for severe genetic disorders, including CTX, and implementing the first genetic screening program of its kind in an isolated Druze community in the Galilee. In this community, a 1:11 carrier frequency was determined for the c.355delC CYP27A1 gene variant (15). For this pilot study, we determined a 1:30 carrier frequency for the c.355delC CYP27A1 gene variant in Druze newborns born at the Galilee and Ziv Medical Centers. Based on the carrier frequency of 1:30, we have estimated a disease prevalence of 1:3,600 in this Druze newborn population; therefore, we did not necessarily expect to find a true positive during this proof-of-principle pilot newborn screening study. Our goal is to continue screening DBSs collected from newborns born in this region. In moving toward universal newborn screening for CTX, data from stepwise implementation of newborn screening, such as the screening for CTX we describe here, performed in a high-risk newborn population, will be extremely valuable. In addition, to test validation data, data can be generated regarding diagnostic follow-up and treatment of screen positive cases, as well as potential benefits of early detection and treatment.
In conclusion, the prospective pilot study data we present supports the feasibility of newborn screening for CTX; the perfect solution for early identification and timely treatment of a disorder that can otherwise present in the neonatal period as cholestatic jaundice that can progress to severe life-threatening liver disease. Without treatment there is neurodegenerative progression over time in CTX that may not be reversed by treatment. Compared with the current delay in diagnosis and treatment, early identification through newborn screening and treatment from birth onward would greatly benefit those affected with this metabolic disorder, effectively providing a functional cure for CTX.