Progressing Preemptive Genotyping of CYP2C19 Allelic Variants for Sickle Cell Disease Patients

Abstract

Aims: Interindividual variability in drug response and adverse effects have been described for proton pump inhibitors, anticonvulsants, selective serotonin reuptake inhibitors, tricyclic antidepressants, and anti-infectives, but little is known about the safety and efficacy of these medications in patients with sickle cell disease (SCD). We genotyped the CYP2C19 gene which has been implicated in the metabolism of these drugs in an SCD patient cohort to determine the frequencies of reduced function, increased function, or complete loss-of-function variants. Materials and Methods: DNAs from 165 unrelated SCD patients were genotyped for nine CYP2C19 (*2, *3, *4, *5, *6, *7,*8, *12, and *17) alleles using the iPLEX^® ADME PGx multiplex panel. Results: Three CYP2C19 alleles (*2, *12, and *17) were detected with the following frequencies: 0.209, 0.006, and 0.236, respectively. The predicted phenotype frequencies were distributed as extensive (31.5%), intermediate (24.8%), poor (5.5%), ultrarapid (30.3%), and unknown metabolizers (7.9%). Discussion: Prognostic genotyping is potentially useful for identifying SCD patients with allelic variants linked to proven clinical pharmacokinetic consequences for several drugs metabolized by the CYP2C19 gene. However, the main challenge to implementing a genetics-guided prescribing practice is ensuring concordance between CYP2C19 genotypes and metabolic phenotypes in SCD patients.

Introduction

Sickle cell disease (SCD) is a genetic disorder affecting largely populations of African ancestry worldwide. Painful, acute vaso-occlusive crisis is the hallmark symptom of SCD and is the precursor of diverse clinical comorbidities affecting various organ systems, including cerebrovascular events, seizure, inflammation, infection, chronic pain, and chronic depression (Lee, 2013; Elmariah et al., 2014). Proton pump inhibitors, anticonvulsants, anti-infectives, selective serotonin reuptake inhibitors, and tricyclic antidepressants are drugs commonly prescribed for SCD patients for these SCD-related morbidities. Interindividual variability in pharmacologic response and adverse drug response has been described for these drugs, but little is known about the safety and efficacy of these medications in patients with SCD (Gardiner and Begg, 2006; Lee, 2013). The highly polymorphic CYP2C19 enzyme is involved in the metabolism and variability in response for these drug classes. Of the 36 allelic variants reported for the CYP2C19 enzyme, at least 12 variants have no enzymatic activity (www.cypalleles.ki.se/cyp2c19.htm, accessed: May 10, 2016). Based on the activity levels of these allelic variants, four distinct metabolic phenotypes are identified: ultrarapid metabolizers (UMs), extensive metabolizers (EMs), intermediate metabolizers (IMs), and poor metabolizers (PMs). PMs are compound heterozygous for different inactivating alleles or homozygous for an inactivating variant, and may display variation in the severity of functional enzyme deficiencies. IMs carry one functional allele and one nonfunctional allele, but may demonstrate a wide range of levels of enzyme activity. EMs have two functional alleles. Ultrarapid metabolizers carry multiple copies of functional alleles.

Preemptive genotyping of allelic variant functional activity level could be used to determine SCD patients' metabolic profiles for CYP2C19 drugs. Preemptive genotyping anticipates current and future medication prescription needs of patients as opposed to this practice, whereby genotyping is performed only when clinically indicated (He and McLeod, 2012; Scott et al., 2012). Indeed, preemptive pharmacogenetic genotyping programs linked to patients' electronic health records are available for oncologic, cardiovascular, and psychiatric disorders in major academic medical centers (Verschuren et al., 2012; Rasmussen-Torvik et al., 2014; Birdwell et al., 2015; Dunnenberger et al., 2015; Hicks et al., 2015).The frequency of CYP2C19 polymorphic alleles and predicted metabolic profiles displays distinct interracial and ethnic variation (Dandara et al., 2001; Hamdy et al., 2002; Allabi et al., 2003; Luo et al., 2006; Matimba et al., 2009; Babalola et al., 2010; Kearns et al., 2010; Man et al., 2010; Martis et al., 2013). African Americans with SCD constitute a patient group, for whom the preemptive determination of CYP2C19 metabolic genotypes may potentially provide lifelong applicable information for selection of appropriate dosages of CYP2C19 drugs and identification of individuals pharmacogenetically prone to unsatisfactory drug response or side effects (He and McLeod, 2012; Scott et al., 2012). CYP2C19 data could facilitate quantification and clinical assessment of pharmacogenetic risk in SCD patients. However, to our knowledge, CYP2C19 allelic frequency and genotype data for African American patients with SCD are unavailable (Babalola et al., 2010). Therefore, the primary purpose of this study was to describe the frequency of CYP2C19 allelic variants, genotypes, and predicted metabolic phenotypes in an African American SCD patient cohort and check for correspondence with previous studies in populations of African ancestry.

Materials and Methods

Human subjects

The study participants were randomly selected patients with SCD receiving care at the Georgia Regents University Comprehensive Sickle Cell Center outreach clinics in southeastern Georgia. The study was approved by the Georgia Regents University Institutional Review Board. Written informed consent or assent was obtained from each patient before inclusion into the study. Study participants were recruited between January 2011 and January 2013. Medical records of the study participants were reviewed to assess SCD genotype, clinical, and medical data.

CYP2C19 genotyping

Whole blood samples (10 mL in tubes containing EDTA) were collected from the study participants in steady state. Genomic DNA was extracted using the Puregene^® DNA Purification Kit (Qiagen) according to the manufacturer's instructions. We used the iPLEX^® ADME PGx multiplex panel (Sequenom, Inc.) to genotyped CYP2C19 *2, *3, *4, *5, *6, *7,*8, *12, and *17 alleles across all study participants. The genotype profiles were reported as heterozygous, homozygous, and homozygous variants, or no “call.” The iPLEX ADME PGx multiplexed panel uses Sequenom Bioscience's iPLEX biochemistry with specific ADME oligo multiplex mixes on the MassARRAY^® system to simultaneously interrogate 192 biologically relevant polymorphisms in 36 pharmacogenes. After running the reactions, mutations were detected, quantified, and genotype reports automatically created using TYPER software. TYPER software assigns the wild-type (*1) CYP2C19 alleles in the absence of other detectable variant alleles (http://bioscience.sequenom.com/iplex-adme-pgx-panel). The CYP allele designations refer to those defined by the Cytochrome P450 Allele Nomenclature Committee (Sim and Ingelman-Sundberg, 2006).

Statistical analysis

The primary outcome measure was genotype frequencies. CYP2C19 allele frequencies were presented with 95% confidence interval. Genotype frequencies were presented as percentage of the study cohort with 95% confidence interval. The observed genotype frequencies were compared with those expected for concordance with Hardy–Weinberg equilibrium using the χ² test.

Results

Demographic, clinical, and medical characteristics

This study elucidates the allelic variants of the CYP2D19 in an SCD cohort. A total of 165 SCD patients (82 males) were recruited. The study participants were all African Americans. Race was self-reported by the subjects. The study participants' demographic features, clinical characteristics, and disease comorbidities are summarized in Table 1. The subjects ranged in age from 16 to 61 years and their body mass index ranged from 15.3 to 38.4. SCD genotype frequencies were distributed as SS (97.5%), SB Thal° (1.8%), and S-Los Angeles (0.6%), respectively. Ten subjects died due to disease complications during the course of the study.

Table 1.

Demographic, Medical, and Clinical Characteristics of SCD Patient Cohort

	No. (%) of subjects, n (%)
Sex
Male	82 (49.6)
Female	83 (50.4)
Age, years
Median
Male	30.5 (16–57)
Female	31.7 (18–61)
Ethnic origin
African American	165 (100)
SCD genotype
SS	161 (97.5)
SB Thal°	3 (1.8)
S-Los Angeles	1 (0.6)
BMI
Median
Male	23.1 (15.3–35.2)
Female	23.7 (16.5–38.4)
Disease comorbidities
Avascular necrosis	31 (18.1)
Pulmonary hypertension	15 (9.09)
Cardiomyopathy	15 (9.09)
Acute chest syndrome	13 (7.98)
Stroke	11 (6.66)
COPD	10 (6.06)
Chronic pain	10 (6.06)
Renal disease	10 (6.06)
Leg ulcer	8 (4.84)
Psychiatric disorders	6 (3.63)
Deep vein thrombosis	6 (3.63)
Liver disease	5 (3.03)
Dysrhythmia	3 (1.81)
Neurologic complications	3 (1.81)
Asthma	2 (1.21)
Congestive heart failure	2 (1.21)
Diabetes	1 (0.6)
Other complications	8 (4.84)
Deceased from disease complications	10 (6.06)

BMI, body mass index; COPD, chronic obstructive pulmonary disease; SCD, sickle cell disease.

CYP2C19 alleles and genotype frequencies

Table 2 showed our cohort CYP2C19 allelic frequencies, genotypes, and predicted metabolic phenotype frequencies. We genotyped nine CYP2C19 alleles (*2, *3, *4, *5, *6, *7, *8, *12, and *17) across all study subjects. The CYP2C19* 1 is considered the wild type with a normal enzyme activity. The abnormal CYP2C19*2 (splicing defect) and *3 (premature stop codon) alleles are the most prevalent alleles associated with the PM profile across racial populations (Scott et al., 2012). Other defective (loss-of-function) alleles include the CYP2C19 *4, *5, *6, *7, *8, *9, and *10. The function of the CYP2C19*12-*16, *18-*26 alleles is unknown or indeterminate, but the *17 allele is deemed an increased function allele (Scott et al., 2012). Regarding the CYP2C19-predicted metabolic phenotypes, individuals are classified as EM (homozygous for the CYP2C19*1); IM (heterozygous with either *1 or *17 plus a reduced or absent enzyme function allele); PM (two loss-of-function alleles); and ultrarapid metabolizer (or homozygous genotype for the *17 or heterozygous for the *1 and *17) (Scott et al., 2012). We identified three variant alleles and seven genotypes of the CYP2C19 in our cohort. The CYP2C19*1 wild-type allele frequency was 0.545. Of the three variant alleles identified, the *17 occurred in the highest frequency (0.236). This was followed by *2 (0.209) and *12 (0.006), respectively. The CYP2C19*1, *2, *12, and *17 frequencies were in concordance with Hardy–Weinberg equilibrium (p-value 0.261). Fifteen subjects (9.0%) were homozygous for the *17/*17 genotype and 35 subjects (21.2%) were heterozygous for the *1/*17. Thirty-nine subjects (23.6%) had the *1/*2 genotype, two subjects (1.2%) were heterozygous for the *1/*12 allele, and nine subjects (5.4%) were homozygous for the *2/*2 allele. The CYP2C19*3,*4,*5, *6, *7, and *8 alleles were not detected. Based on genotype frequencies, the assigned predicted metabolic phenotypes in the cohort were distributed as ultrarapid (UM/30.3%/50 subjects), extensive (EM/31.5%/52 subjects), intermediate (IM/24.8%/41 subjects), and poor (PM/5.5%/9 subjects) metabolizers, respectively. Because the phenotypic consequences of the *2/*17 genotype may fall between the EM and IM phenotype dependent on the CYP2C19 substrate, we assigned this combination to the unknown phenotype category (UNK/7.9%/12 subjects), together with one subject with the indeterminate *17/UNK genotype.

Table 2.

CYP2C19 Allelic and Genotype Frequencies

Allele	Genetic alteration	Enzyme activity	Number of alleles	Allelic frequency	95% CI range
*1	Wild type	Normal	180	0.545	0.483 to 0.590
*2	Splicing defect	None	69	0.209	0.165 to 0.253
*3	Stop Codon	None	0	0.000	0.000 to 0.000
*4	Missense	None	0	0.000	0.000 to 0.000
*5	Missense	None	0	0.000	0.000 to 0.000
*6	Missense	None	0	0.000	0.000 to 0.000
*7	Splicing defect	None	0	0.000	0.000 to 0.000
*8	Missense	None	0	0.000	0.000 to 0.000
*9	Missense	Decreased	0	0.000	0.000 to 0.000
*12	Missense	Unknown	2	0.006	−0.002 to 0.014
*17	Increased function	Increased	78	0.236	0.191 to 0.282
UNK	—	—	1	0.003	−0.003 to 0.009
Total			330	1.0

Metabolizer phenotype	Genotype	Number of subjects (%)	Observed frequency (%)	Predicted frequency (%)
Ultrarapid metabolizer
Heterozygous	*1/*17	35 (21.2)	21.2	25.4
Homozygous	*17/*17	15 (9.1)	9.1	5.6
	Sub total	50 (30.3)
Extensive metabolizer
	*1/*1	52 (31.5)	31.5	29.8
Intermediate metabolizer
	*1/*2	39 (23.6)	23.6	23.8
	*1/*12	2 (1.2)	1.2	0.7
	Subtotal	41 (24.8)
Poor metabolizer
	*2/*2	9 (5.4)	5.5	4.4
Unknown	*2/*17	12 (7.2)	7.3	9.9
	*17/UNK	1 (0.6)	0.6	0.1
	Subtotal	13 (7.8)
	Total	165

CI, confidence interval; UNK, unknown.

Since the frequency of CYP2C19 polymorphic alleles displays distinct interracial and ethnic variation, naturally, we expected our cohort CYP2C19 allelic frequencies to mirror that of healthy African Americans (Dandara et al., 2001; Hamdy et al., 2002; Allabi et al., 2003; Luo et al., 2006; Matimba et al., 2009; Kearns et al., 2010; Man et al., 2010; Martis et al., 2013). However, as depicted in Table 3, the frequency data for variant CYP2C19 alleles beyond those normally tested (*2,*3, and recently *17) for African Americans are very limited. To our knowledge, only a single study conducted in an African population has published data on the prevalence of the *2 and *3 alleles in SCD patients (Babalola et al., 2010). Recently, Martis et al. (2013) reported the CYP2C19*15 variant in a large African American cohort. Man et al. (2010) reported the CYP2C19*9, *10, and *13 alleles in African subjects. However, none of these alleles were observed in our cohort. The frequency (23%) for the *17 allele is the highest, to date, reported in African Americans (Martis et al., 2013). Our cohort exhibited a high frequency of the *17/*17 ultra-rapid genotype (9.1%) compared to (2.8%) reported in Martis et al. (2013). However, no differences were detected for the *1/*17 (UM), *2/*2 (PM), and *2/*17* (UNK) genotypes in the cohort (21.2%, 5.5%, and 7.3%) and healthy African Americans (20.4%, 4.8%, and 9.2%), respectively (Martis et al., 2013). Our study reports the frequency of the CYP2C19*12 allele for the first time in an African American cohort. As depicted in Table 3, the limited studies suggest that to date, the CYP2C19 *2,*3, *9,*12, *13,*15, and *17 are the variants of potential interest for preemptive genotyping in populations of African descent (Alessandrini et al., 2013).

Table 3.

CYP2C19 Frequencies in Previously Studied Populations

	Allele frequencies
Racial and ethnic group	N	*1	*2	*3	*4	*5	*6	*7	*8	*9	*10	*12	*13	*15	*17	UKN	Ref.
African American
African American	165	0.545	0.209	0.000	0.000	0.000	0.000	0.000	0.000	0.000	0.000	0.006	0.000	0.000	0.236	0.003	This study
African American	250	0.594	0.194	0.004	0.000	0.000	0.000	0.000	0.000	0.000	0.000	0.000	0.012	0.014	0.182	—	Martis et al. (2013)
African American	114	—	—	—	—	—	—	—	—	—	—	—	—	—	0.210	—	Kearns et al. (2010)
African American	236		0.182	0.008	—	—	—	—	—	—	—	—	—	—	—	—	Luo et al. (2006)
African	250	0.769	0.192	0.002	0.000	0.000	0.000	0.000	0.000	0.014	0.060	0.000	0.016	0.000	0	0.012	Man et al. (2010)
African ethnic group
Luo	30	—	0.070	0.000	—	—	—	—	—	—	—	0.040	0.030	0.050	—	—	Matimba et al. (2009)
Ibo	20	—	0.330	0.000	—	—	—	—	—	0.000	—	0.000	0.040	0.000	—	—	Matimba et al. (2009)
Shona	15	—	0.230	0.000	—	—	—	—	—	0.000	—	—	0.030	0.000	—	—	Matimba et al. (2009)
Venda	9		0.170	0.000	—	—	—	—	—	0.060	—	—	0.000	0.000	—	—	Matimba et al. (2009)
African national group
Nigerian	158	0.845	0.155	0.000	—	—	—	—	—	—	—	—	—	—	—	—	Babalola et al. (2010)
Nigerian SCD	115	0.843	0.157	0.000	—	—	—	—	—	—	—	—	—	—	—	—	Babalola et al. (2010)
Tanzanian	106	0.821	0.179	0.000	—	—	—	—	—	—	—	—	—	—	—	—	Dandara et al. (2001)
Beninese	119	0.870	0.130	0.000	—	—	—	—	—	—	—	—	—	—	—	—	Allabi et al. (2003)
Egyptian	247	0.888	0.110	0.002	—	—	—	—	—	—	—	—	—	—	—	—	Hamdy et al. (2002)

(—), not screened; UKN, undetermined.

Discussion

To the best of our knowledge, this study is the first to identify the frequency of genetic polymorphisms in the CYP2C19 gene in an African American SCD patient cohort. This information is useful for the proper design of pharmacogenetic study of CYP2C19 substrates for rational prescribing practice in this patient population. The CYP2C19 metabolizes several psychotropics, proton pump inhibitors, anti-infectives, and anticonvulsants often prescribed for SCD comorbidities (Bertilsson, 1995; Furuta et al., 2004; Sim et al., 2006; Simon et al., 2009; Sibbing et al., 2010; Hodgson et al., 2014; Luk et al., 2014). For the antiplatelet drug, clopidogrel, genotype-directed guidelines have been provided by the Clinical Pharmacogenetics Implementation Consortium (Robert et al., 2012; Scott et al., 2013).

Preemptive genotyping and clinical application of functional information of CYP2C19 allelic variants could empower clinicians to communicate pharmacologic risk and drug response prediction to SCD patients using biological evidence, as opposed to explaining statistical risk without biological significance, and possibly inform drug–dose adjustments according to the genotype (Schildcrout et al., 2012; Sparkenbaugh and Pawlinski, 2013). For instance, with regard to treatment of peptic ulcer disease with standard doses of proton pump inhibitors, we could extrapolate that 30% of our cohort who are UMs (*1/*17 and *17/*17) would experience reduced acid-inhibitory effect, which may result in therapeutic failure. Also, with our cohort being 31.5% EM and 30.3% UM, greater than half of our subjects will theoretically have significantly decreased drug expose for CYP2C19 substrates such as, diazepam, citalopram, amitriptyline, and clomipramine as the concentrations of these drugs may decrease by rapidly converting into inactive metabolites with a possible reduction of therapeutic effects. Interestingly, for the 8% of our cohort with unknown predicted metabolic phenotypes, the pharmacologic risks for CYP2C19 substrates could not be determined, making these individuals candidates for clinical vigilance and alternative drug choices. Thus, for SCD patients prone to multiorgan disease-related comorbidities and polypharmacy, preemptive genotyping provides readily available genetic information to guide therapy (Leeder, 2015). Even normal CYP2C19 metabolic genotype result has clinical utility for SCD patients. Clinicians could use normal metabolic profile results to reassure patients of the safety of the drugs prescribed and likelihood to work for them, thereby potentially promoting adherence. In addition, SCD patients could share their pharmacogenetic results with their other health provider to ensure optimal prescribing for future medications (Haga and Mills, 2015).

To be sure, the utility of genotype-guided prescribing practice for CYP2C19 substrates in SCD patients will depend on CYP2C19 enzyme expression, and genotype-metabolic phenotype concordance, which may be influenced by SCD. Generally, the amount of active CYP2C19 depends on the level of gene expression. Acute pain crisis in SCD is the quintessential inflammatory state (Sparkenbaugh and Pawlinski, 2013). Interestingly, recent reports have associated inflammation response with a decreased CYP2C19 activity in patients with severe inflammatory disease. The decrease in drug metabolism can be up to 70% and is mediated by proinflammatory cytokines (Vet et al., 2011). Thus, independent of metabolic genotype, systematic inflammation associated with SCD pain may downregulate CYP2C19 enzyme expression in SCD patients and lead to “acquired” PM status in patients either with homozygous (*1/*1) or heterozygous (*1/*2) genotype. Downregulation of the CYP2C19 enzyme has been reported in solid tumor and multiple myeloma cancer patients due to tumor-associated inflammatory factors (Williams et al., 2000; Helsby et al., 2008; Vet et al., 2011; Burns et al., 2014). Thus, it is likely that during vaso-occlusive pain crisis, SCD patients may have increased exposure to drugs because of decreased clearance and an increased risk of adverse drug effects (Shah and Smith, 2015). Furthermore, inflammation could also affect the expression and activity of drug transporters, influence protein binding, capillary permeability, cardiac output, and liver blood flow. Collectively, these processes may have important pharmacokinetic and pharmacodynamic consequences and corresponding phenotypic disposition (Shah and Smith, 2015). Thus, given the numerous variables that could cause discordance in drug metabolic genotype–phenotype relationship, extrapolating allelic frequencies and genotypes of the CYP2C19 enzyme from healthy African Americans to SCD patients is not recommended.

Interestingly, our study indicted a significantly higher percentage of African American EM (31.3%) compared to 23.2% EM in the Martis study, raising the possibility that within African American patient subgroups, some CYP2C19 alleles may be underrepresented or overrepresented because of founder effect (Martis et al., 2013). For instance, SCD patients are reported to have a slightly higher frequency of the CYP2D6 gene deletions compared to healthy African Americans (Yee et al., 2013). CYP2D6 gene deletions are associated with the PM phenotype and impaired ability to convert opioids such as codeine and hydrocodone into their active analgesic forms and as such are implicated in SCD pain phenotype and analgesic drug response. Undoubtedly, to understand variability in drug response linked to CYP2C19 enzyme in SCD patients, further work should focus on measurement of CYP2C19 activity, especially for psychotropic medications frequently used as adjuvant analgesics for SCD pain (Darbari et al., 2011). These medications have been shown to have infrequent, but potentially important hematologic side effects in SCD patients (Jerrell et al., 2011).

There are limitations to our study that we acknowledged. The method used to genotype the CYP2C19 alleles does not differentiate between the CYP2C19*1B, *1C, and *9. Thus, it is likely that some individual classified at EM may harbor the reduced-function CYP2C19*9 allele. Also, genotyping only allows identification of the CYP2C19 allelic status and is independent of environmental influences. This feature could likewise be considered a drawback as this approach provides no information on the actual level of enzyme activity that may depend on numerous modulating influences as noted above. Phenotyping on the other hand, that is, the direct analysis of metabolite ratios using appropriate probes would take into account such factors. Thus, pharmacokinetic study remains the gold standard for discerning individual SCD patients' drug metabolic phenotypes. Nonetheless, preemptive CYP2C19 genotyping aided by appropriate pharmacokinetics and pharmacodynamics studies brings into sharp relief the potential for genomic-based dosing strategies in SCD patients in the same way as pharmacogenetics guides oncological use of biological agents (Anwar et al., 2015).

Conclusion

In conclusion, preemptive genotyping is potentially useful for identifying SCD patients with allelic variants linked to proven clinical pharmacokinetic consequences for several drugs metabolized by the CYP2C19 gene. Indeed, guidelines and treatment algorithms have been developed to assist clinicians in genotype-directed anti-platelet therapy (Roberts et al., 2012) and other prescribing algorithms are currently underway (Hicks et al., 2015). Such guidelines, primarily driven by an increasing evidence base, categorize patients based on their CYP2C19 genotypes to guide drug dosing. However, the main challenge to implementing genetics-guided prescribing practice for SCD patients is ensuring concordance between CYP2C19 genotypes-metabolic phenotypes.

Acknowledgments

The authors wish to thank the nursing staff at the Georgia Regents University Comprehensive Sickle Cell Clinic. This work was accepted for presentation in abstract form at the 9^th Annual Sickle Cell and Thalassemia Advanced Conference, October 7–9, 2015, at the Guy's and St. Thomas' Hospital, London, England.

Funding Sources

This study was supported, in part, by a grant from the National Institute of Nursing Research, NIH, grant no. K01NR012465 and the National Institute on Minority Health and Health Disparities, NIH, grant no. P20MD003383. The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Institute for Nursing Research or the National Institute on Minority Health and Health Disparities.

Authors' Contributions

Conceived and designed the experiments: C.J., M.L., and A.K. Participated in research design: C.J., N.B., and A.K.; performed the experiments: N.P. Analyzed the data: C.J. and H.X. Wrote or contributed to the writing of the article: C.J., N.B., A.K., and M.L.

Author Disclosure Statement

No competing financial interests exist.