Think classical homocystinuria if the genetic test did not confirm Marfan syndrome: Late diagnosis and phenotypic variability in adult siblings with classical homocystinuria

Abstract

Classical homocystinuria is an inherited metabolic disease of homocysteine metabolism due to biallelic pathogenic variants in CBS. The biochemical hallmark is elevated homocysteine and methionine levels. The treatment consists of betaine supplementation and protein restricted diet. We report two adult siblings with late diagnosis of classical homocystinuria, a variable phenotype and good response to the treatment.

Patient 1 is a 29-year-old female with a history of myopia, Marfanoid habitus with significant kyphoscoliosis, anxiety and a psychotic episode. Clinical exome sequencing identified compound heterozygous pathogenic variants in CBS (c.209+1G>A; c.992C>T (p.Ala331Val)). She had markedly elevated homocysteine (298 μmol/L) and methionine (1040 μmol/L) levels. Her brain magnetic resonance spectroscopy revealed a low n-acetyl-aspartic acid peak. She was started on betaine supplementation, and a protein-restricted diet (0.8 g/kg/day) leading to significant decrease in her homocysteine (37 μmol/L) and methionine (49 μmol/L) levels. Patient 2 is a 27-year-old female (younger sibling) with a history of anxiety, one generalized tonic-clonic seizure and a dural sinus thrombosis in neuroimaging. She had both familial CBS variants and markedly elevated homocysteine (152 μmol/L) and methionine (560 μmol/L) levels, which were improved significantly on betaine supplementation and the protein-restricted diet. Both siblings had average range intellectual abilities. Higher homocysteine levels may result in severe skeletal, and central nervous system phenotypes.

1. Introduction

Classical homocystinuria (OMIM #236200) is one of the inherited metabolic diseases of sulfur amino acid metabolism. It is caused by biallelic pathogenic variants in CBS, which encodes cystathionine β-synthase (CBS). The CBS is an essential enzyme for the conversion of homocysteine to cystathionine (trans‑sulfuration) and pyridoxal-5′-phosphate is the cofactor for this enzymatic reaction. Methionine is converted to homocysteine via de-methylation and homocysteine is converted to methionine via re-methylation. Folic acid and vitamin B12 are cofactors for the re-methylation of homocysteine to methionine [1]. The prevalence of classical homocystinuria is estimated between 1:200,000 and 1:335,000 [2]. The highest incidence of classical homocystinuria is 1:1800 in the Qatar population due to a founder pathogenic variant (c.1006C>T; p.Arg336Cys) [3,4].

Clinical features of classical homocystinuria predominantly involve four organ systems: ophthalmologic (ectopia lentis and/or severe myopia), skeletal (tall and slender status with a Marfanoid habitus, kyphoscoliosis, osteoporosis), vascular (thromboembolism) and central nervous system (developmental delay, intellectual disability, psychiatric problems) [2]. Two different phenotypes have been reported: 1. pyridoxine responsive; 2. pyridoxine non-responsive. The pyridoxine responsive phenotype presents with isolated thromboembolic events in adulthood, and patients are responsive to high dose pyridoxine therapy with significant improvements in their homocysteine levels. The pyridoxine non-responsive phenotype presents with early onset multisystem disease [5].

Patients with classical homocystinuria have markedly elevated homocysteine and methionine levels. Identification of biallelic pathogenic variants in CBS confirms the molecular diagnosis of classical homocystinuria [2]. Methionine and/or homocysteine in blood dot spots are used as biomarkers for identification of classical homocystinuria in the neonatal period as part of newborn screening programs [4,6].

Once the molecular genetic diagnosis of classical homocystinuria is established, an oral pyridoxine challenge test is performed by giving 100–500 mg pyridoxine for 3–5 days. If there is >30 % decrease in homocysteine level compared to baseline, this is a pyridoxine responsive phenotype which is treated using high dose pyridoxine. Failure to response to pyridoxine is characteristic of the pyridoxine non-responsive phenotype which is treated using betaine supplementation, a methionine- or protein-restricted diet. Additionally, supplementations of folic acid, vitamin B12 and pyridoxine are used in the treatment of classical homocystinuria as they are cofactors in the re-methylation and trans‑sulfuration pathways [1].

We report two adult siblings who were diagnosed with classical homocystinuria in their twenties. The older sibling presented with Marfanoid features, while the younger sibling presented with a vascular phenotype. They had good response to the current standard therapy with excellent improvements in their homocysteine and methionine levels. These patients illustrate the phenotypic variability of classical homocystinuria within the same family.

2. Material, methods and results

We received a signed case report consent form from both patients to report their de-identified information in this manuscript. We reviewed Electronic Patient Charts. All biochemical and molecular genetic investigations were performed in a clinical biochemical and molecular genetic laboratories. The 2D T1-weigthed and 2D T2-weighted images, chemical shift image (CSI) multi-voxel brain magnetic resonance spectroscopy (MRS) were acquired on 3-Tesla Siemens Skyra (Erlangen, Germany) at University of Alberta Hospital, Edmonton in Patient 1. Standard of care T1-weighted images with fluid inversion recovery (FLAIR) (acquired in sagittal plane) were used for tissue segmentation (gray matter (GM)/white matter (WM)/cerebral spinal fluid (CSF)). T2-weighted images (acquired in axial oblique plane) were used for the assessment of voxel placement during analysis. CSI MRS (semi-adiabatic localization by adiabatic selective refocusing (sLASER) sequence, axial-oblique plane, resolution 6.25 × 6.25 × 10 mm, 8 × 8 voxel matrix, TR = 1520 ms, TE = 135 ms) was used to assess n-acetyl aspartic acid (NAA), creatine and choline compounds in white matter regions of basal ganglia. CSI MRS data were analyzed using FSL-MRS [re; first, T1-weighted images were segmented to GM/WM/CSF using fsl_anat (FSL v 6.0.7.11) [7], and this segmentation was provided as one of the inputs for FSL-MRS fitting. Basis set for the CSI sequence was generated using MRI Cloud (https://braingps.mricloud.org/mrs-cloud) [8,9]. Two regions of interest (ROIs) were defined as a mask (in CSI space) using fsl_eyes, including only voxels with good spectra quality; WM ROIs, where WM content >80 % and GM ROIs, where GM content >50 %. Spectra fitting was done for each of the masked ROIs. Based on the report generated by FSL-MRS fitting, we report a ratio of NAA compound to total choline (NAA/tCho) and NAA to total creatine (NAA/tCr). A 27-year-old female with PGM1 related congenital disorders of glycosylation (PGM1-CDG) was used as a control for MRS evaluation [10].

2.1. Patient 1

This is a 29-year-old female who has a long-standing history of scoliosis and pectus excavatum that led to a referral to our medical genetic clinic at the age of 12 years. Her physical examination features were suggestive of Marfan syndrome. No genetic investigations were performed at that time. She was re-referred to our medical genetic clinic at the age of 26 years with the suspected diagnosis of Marfan syndrome and underwent targeted next generation sequencing panel for aortopathy (24 genes) due to mitral valve prolapse and Ghent criteria of 8 (a score of ≥7 is considered a positive systemic score for Marfan syndrome), which did not reveal any variants in those genes. At the age of 27 years, she had a psychotic episode leading to an admission to a psychiatric inpatient unit. During that admission, a medical genetics consultation was requested, and she underwent a clinical exome sequencing. Her exome sequencing result revealed two pathogenic variants (c.209+1G>A and c.992C>T; p.Ala331Val) in CBS (segregated by parents), confirming the molecular genetic diagnosis of classical homocystinuria. She was referred to metabolic genetic clinic for the management of her classical homocystinuria. In her physical examination, she had long face, malar hypoplasia, high-arched palate, soft stretchable skin, pectus deformity, scoliosis, positive thumb and wrist signs, and flat foot at the age of 27 years. Her height was at 55th percentile (164 cm) and her expected mid-parental height was 165.5 cm. We summarized her clinical features as per organ and/or system involvement below:

2.1.1. Skeletal phenotype

She had marked scoliosis and pectus excavatum in childhood. She underwent scoliosis corrective surgery with posterior instrumentation of T2 to S1, posterior lumbar interbody fusion at L5-S1 and posterolateral osteotomy thoracolumbar spine at the age of 12 years. Her X-ray revealed 75° Cobbs angle at thoracolumbar junction and 40° spine curve at midthoracic spine (Fig. 1). Her bone mineral density revealed a low bone density for age with Z score of −1.2 in left hip; Z score of −2.2 in left femoral neck and Z score of −1.5 in right distal forearm at the age of 27 years.

Fig. 1.

A) Chest Xray in Patient 1 shows severe scoliosis with a Cobbs angle of 75° at thoracolumbar junction. B) Chest Xray in Patient 1 after scoliosis corrective surgery.

2.1.2. Central nervous system phenotype

She completed a college education specializing in child and youth care. She has an anxiety disorder that was diagnosed at the age of 25 years and has been on antidepressants since then. At the age of 27 years, she developed an acute episode characterized by an acute distress, dysregulation and disorganization of her thoughts after losing a family member. She had a history of hallucinations, delusions and suicidal ideas. She presented to emergency room and was admitted to a psychiatric inpatient unit for 17 days. She was diagnosed with a brief psychotic episode and was started on a low-dose olanzapine therapy for 4 weeks. Her Generalized Anxiety Disorder 7-item (GAD-7) score was 11, indicating moderate anxiety disorder. She underwent neuropsychological assessment using Wechsler Adult Intelligence Scale-Fourth Edition (WAIS-IV) at the age of 28 years after her molecular genetic diagnosis of classical homocystinuria (Table 1). Her overall intellectual reasoning abilities were average (WAIS-IV General Ability Index = 108, 70th percentile; WAIS-IV Perceptual Reasoning Index = 112, 79th percentile and WAIS-IV Verbal Comprehension Index = 103, 58th percentile). Her scores were generally average with respect to academic skills, attention, working memory, and visual-spatial abilities (between 32nd to 68th percentile). However, she had low scores for her verbal and language skills, processing speed, memory, and executive functioning (as low as (<0.1 percentile). On self-report questionnaires she endorsed significant ongoing mental health symptoms including depression and anxiety. Her neuropsychological assessment results did not indicate an intellectual developmental disorder or a specific learning disorder.

Table 1.

Neuropsychological assessment results are summarized in Table 1 for Patient 1 and Patient 2.

Test	Ability Measured	Patient 1 (percentiles)	Patient 2 (percentiles)
WAIS-IV General Ability Index	Overall intellectual abilities	70	73
WRAT-IV Word Reading	Single word reading	27	34
WRAT-IV Math Computation	Written math calculation	37	42
WAIS-IV Working Memory Index	Auditory working memory	5	21
WAIS-IV Digit Span	Digit repetition and rearrangement	16	25
WAIS-IV Arithmetic	Mental math	5	25
WMS-III Spatial Span	Visual working memory	25	75
Finger Tapping Right Hand	Simple motor speed	6	21
Finger Tapping Left Hand	Simple motor speed	0.4	19
Grooved Pegboard Right Hand	Fine motor coordination	8	16
Grooved Pegboard Left Hand	Fine motor coordination	6	2
WAIS-IV Processing Speed Index	Graphomotor processing speed	34	84
WAIS-IV Symbol Search	Visual scanning	75	84
WAIS-IV Digit-Symbol Coding	Complex processing speed	9	75
Trail Making Test Part A	Visual sequencing	4	16
WAIS-IV Verbal Comprehension Index	Overall verbal abilities	58	58
WAIS-IV Vocabulary	Expressive vocabulary	75	50
WAIS-IV Similarities	Verbal reasoning	37	63
Controlled Oral Word Association	Phonemic verbal fluency	5	32
Animal Fluency	Semantic verbal fluency	53	32
Boston Naming Test	Picture naming	8	9
WAIS-IV Perceptual Reasoning Index	Overall visual-spatial abilities	79	84
WAIS-IV Block Design	Hands-on visual reasoning	84	63
WAIS-IV Matrix Reasoning	Abstract visual reasoning	63	95
RAVLT Learning	List learning efficiency	7	53
RAVLT Short Delay Recall	Immediate list recall	30	45
RAVLT Long Delay Recall	Delayed list recall	32	32
WMS-IV Logical Memory I	Immediate story recall	5	63
WMS-IV Logical Memory II	Delayed story recall	5	37
RCFT Delayed Recall	Delayed visual recall	1	2
RCFT Delayed Recognition	Recognition of complex figure	19	1
CVMT Learning	Visual learning efficiency	32	32
CVMT Delayed Recognition	Recognition of simple figures	<0.1	37
Trail Making Test Part B	Speeded set shifting	16	70
WCST Errors	Reasoning in response to feedback	37	73
DKEFS Color-Word Interference	Response inhibition	2	75
DKEFS Color-Word Switching	Response inhibition/set shifting	16	50

Her brain magnetic resonance imaging (MRI) was normal. Her brain MRS showed 27.5 % decrease in NAA/tCho in white matter (WM ROI) and 12.8 % decrease in NAA/tCho in gray matter (GM ROI) compared to age and sex matched control (Fig. 2).

Fig. 2.

A) Brain magnetic resonance spectroscopy (chemical shift imaging, sLASER with TE = 135 ms) in Patient 1 shows decreased NAA/tCho. Regions of interest in white matter (green outline) and gray matter (red outline) are shown overlaid on T2 weighted image for the Patient 1 (P1) and age-matched control (control). B) Fitted spectra in white matter region are shown for the Patient 1 (P1) and control (control).

2.1.3. Ophthalmologic phenotype

She had myopia, diagnosed at the age of 9 years. There was no lens subluxation.

2.1.4. Vascular phenotype

There was no history of thromboembolic events.

2.1.5. Biomarkers and treatment outcome

Her baseline homocysteine level was 298 μmol/L (reference range 4.9–13.7 μmol/L) and plasma methionine was 1040 μmol/L (reference range 14–40 μmol/L). Her pyridoxine challenge test (500 mg for 3 days) showed a 3 % decrease in her homocysteine level and was classified as a pyridoxine non-responsive phenotype. She was started on betaine (6 g/day), vitamin B12 (1 mg/day), folic acid (5 mg/day) and pyridoxine (200 mg/day). As her homocysteine levels were > 50 μmol/L, her betaine dose was gradually increased to 10 g/day. Due to ongoing high homocysteine levels, she was started on a protein-restricted diet (0.8 g/kg/day). Her homocysteine level decreased to <50 μmol/L. Her biomarkers on the treatment are summarized in Fig. 3. She tolerated the treatment well. She did not have any thromboembolic events. She did not report any improvements in her anxiety disorder.

Fig. 3.

Plasma biomarkers in response to treatment for Patient 1 and 2 are depicted in Fig. 3.

A) Patient 1: Homocysteine and methionine levels were decreased 3 % in response to pyridoxine challenge test. She was started on betaine with gradual increases in the dose. At 5 months of treatment, plasma homocysteine level was >50 μmol/L. On protein restricted diet homocysteine level was decreased to <50 μmol/L.

B) Patient 2: Homocysteine and methionine levels were decreased gradually with gradual increase of betaine dose, and on protein restricted diet. At 10^th month of treatment her homocysteine and methionine levels were gradually increased reaching to >100 μmol/L. At that time her protein intake was increased to 0.9 mg/kg/day due to weight loss. When protein intake was decreased to 0.8mg/kg/day, her homocysteine and methionine levels were improved.

2.2. Patient 2

This is a 27-year-old female, younger sister of Patient 1. She underwent familial variant test confirming the molecular genetic diagnosis of classical homocystinuria at the age of 25 years after her older sister's diagnosis. Her past medical history was remarkable for one generalized tonic-clonic seizure and dural sinus thrombosis in neuroimaging at the age of 17 years. Her physical examination showed a prominent forehead, low set posteriorly rotated small ears, long fingers and toes, and hypertelorism at the age of 25 years. We summarized her clinical features as per organ/system involvement below:

2.2.1. Central nervous system/vascular phenotype

She had a generalized tonic-clonic seizure at the age of 17 years. Her brain computer tomography (CT) revealed dural sinus thrombosis (Fig. 4) that was attributed to oral hormonal contraceptive. She was started on oral anti-coagulant (rivaroxaban) therapy and anti-seizure medication (keppra), and both were discontinued after 6 months. There was no history of recurrent seizures since then.

Fig. 4.

Brain CT imaging in Patient 2 showed sinus venous thrombosis. The clot burden affects much of the anterior and superior aspects of the superior sagittal sinus, with extension into its posterior aspect.

She graduated from nursing school and has been working as a nurse in emergency room. She has history of anxiety disorder and panic attack since the age of 17 years old. She was started on antidepressant treatment at the age of 20 years, which led to improvements in her symptoms. Her GAD-7 score was 4 (minimal anxiety) prior to treatment of classical homocystinuria. She underwent neuropsychological assessment at the age of 25 years after the molecular genetic diagnosis of classical homocystinuria (Table 1). Her overall intellectual reasoning abilities were average (WAIS-IV General Ability Index = 112, 79th percentile; WAIS-IV Perceptual Reasoning Index = 115, 84th percentile and WAIS-IV Verbal Comprehension Index = 103, 58th percentile). She acknowledged mild health-related anxiety on interview. Her neuropsychological assessment results did not indicate an intellectual developmental disorder or a specific learning disorder.

Her brain MRI revealed a mild ill-defined nonspecific T2/FLAIR hyperintense changes within the deep white matter of the parietal lobes and there was no evidence of dural venous sinus thrombosis at the age 25 years. A request for brain MRS was declined by our neuroradiology department.

2.2.2. Skeletal phenotype

There was no history of scoliosis. Her bone mineral density was normal for age (Z-score: 0.9 in left total hip and Z-score: −1.6 at left femoral neck) at the age of 25 years.

2.2.3. Ophthalmologic phenotype

She wore glasses for myopia since the age of 8 years. She underwent corrective laser eye surgery for myopia at the age of 22 years. There was no lens subluxation.

2.2.4. Biomarkers and treatment outcome

Her homocysteine level was 152 μmol/L and plasma methionine level was 560 μmol/L at the time of the molecular genetic diagnosis of classical homocystinuria. She was started on betaine (6 g/day), vitamin B12 (1 mg/day), folic acid (5 mg/day) and pyridoxine (200 mg/day). As her homocysteine level was >100 μmol/L, her betaine dose was increased to 10 g/day. Due to persistent high homocysteine levels (>100), a protein-restricted diet (0.8 g/kg/day) was started. Her homocysteine level decreased to <50 μmol/L. Her biomarkers on the treatment are summarized in Fig. 3. She tolerated the treatment well. She reported subjective improvements in her anxiety disorder. Her GAD-7 score was 3 (improved from 4) on treatment of classical homocystinuria. She did not have any new thromboembolic events.

3. Discussion

We report two adult siblings with classical homocystinuria with a pyridoxine non-responsive phenotype. They were diagnosed by clinical exome sequencing and by positive family history in their 20s. Patient 1 presented with a childhood-onset skeletal phenotype, whereas younger sibling (Patient 2) presented with adolescent onset central nervous system/vascular phenotype. They both had history of anxiety disorder, moderate in Patient 1 and mild in Patient 2. Additionally, Patient 1 presented with a psychotic episode in her 20s. They both had normal cognitive functions in their neuropsychological assessments. Both patients tolerated a protein restricted diet leading to improvements in their homocysteine levels (<50 μmol/L) and normalization of their methionine levels. It was previously reported that patients with pyridoxine non-responsive phenotype have homocysteine levels that are >100 μmol/L on the treatment [11]. To the best of our knowledge, we report excellent response to the current standard treatment of classical homocystinuria in both siblings despite late diagnosis and pyridoxine non-responsive phenotype.

A natural history study of 629 patients with classical homocystinuria reported variable phenotypes between pyridoxine non-responsive and pyridoxine responsive phenotypes. Patients with classical homocystinuria and pyridoxine non-responsive phenotype presented with childhood onset cognitive dysfunction, lens dislocation, Marfanoid features, osteoporosis and thromboembolic events. Patients with the pyridoxine responsive phenotype presented with thromboembolic events, and/or lens dislocation in adolescence or adulthood [12]. In a multicenter study, 328 patients were reported for their phenotypes and their response to pyridoxine challenge test. There was a statistically significant difference for their average age of diagnosis between pyridoxine non-responsive (in the first decade of life) and extreme pyridoxine responsive phenotypes (in the 3rd-4th decades of life). Developmental delay and learning difficulties were reported in 53 % of patients with pyridoxine non-responsive phenotype, whereas none of the patients with extreme pyridoxine responsive phenotype had these symptoms. Interestingly vascular (27 %) and psychiatric phenotypes (11 %) were reported in less than one-third of the patients with pyridoxine non-responsive phenotype. Whereas vascular phenotype was reported in >50 % of the patients with pyridoxine responsive phenotype leading to their diagnosis after their first event [5]. Central nervous system/vascular phenotype was present in our Patient 2 with pyridoxine non-responsive phenotype. Unfortunately, homocysteine level was not measured during the first central nervous system/vascular event. Her diagnosis was only confirmed due to positive family history of classical homocystinuria in her older sibling. Both patients in our study had normal cognitive functions by their neuropsychological assessments. It seems that phenotypes are a continuum between pyridoxine non-responsive and pyridoxine responsive phenotypes in classical homocystinuria.

Thirty families with classical homocystinuria were reported with same phenotypes within the same family [13]. Three siblings with classical homocystinuria were reported with Marfanoid features and markedly elevated homocysteine levels (>200 μmol/L). Two of those siblings (both females) had lens subluxation. Younger sister had a progressive movement disorder characterized by bradykinesia, and chorea from the age of 16 years, whereas older sister had rigidity at the age of 25 years. Their younger brother did not have a lens subluxation or a movement disorder [14]. In another study, 13 families with classical homocystinuria were reported. There was up to 70 μmol/L difference between their homocysteine levels [15]. Response to pyridoxine challenge test was similar between siblings within the same family [12,15]. As there were reports of the similar response to pyridoxine challenge test within the same family, we did not perform this test in Patient 2 in our study. Patient 1 had childhood onset skeletal phenotype, who had markedly elevated homocysteine level (22 times of the upper reference range). Whereas younger sibling presented with thromboembolic event in adolescence, who had moderate to markedly elevated homocysteine level (11 times of the upper reference range). We report two siblings with variable phenotypes and age of onset within the same family. This may be due to the differences in their homocysteine levels and longer exposure time to higher homocysteine levels leading to skeletal phenotype in Patient 1. As we have no repeated measurements of homocysteine levels in both siblings since childhood, we do not know if this is the reason for variable phenotypes.

NAA is synthesized in neurons and plays roles for myelinogenesis, synaptic plasticity, and neuroprotection. It also serves as an osmolyte for controlling water distribution, and nitrogen removal from the central nervous system. NAA is used as a marker of the functional integrity of neuronal mitochondrial metabolism. Its reduction indicates impaired brain energy metabolism [16,17]. Elevated homocysteine activates the N-methyl-d-aspartate (NMDA) receptor and causes decrease in NAA in the central nervous system [18]. Elevated homocysteine also inhibits mitochondrial oxidative phosphorylation and results in production of reactive oxygen species in neuron cells leading to neurotoxicity [19]. In a study, 113 patients with coronary heart disease (without a history of stroke or transient ischemic attack) underwent brain MRS and 32 of them had marginally elevated homocysteine level (13.5 ± 5.6 μmol/L). NAA in brain MRS was significantly lower in patients with marginally elevated homocysteine levels [20]. An infant with 5-methyl-tetrahydrofolate reductase (MTHFR) deficiency had a low NAA in brain MRS at the age of 10 months. Her initial homocysteine level was 154 μmol/L [21]. Interestingly, our Patient 1 had low NAA in brain MRS who had markedly elevated homocysteine level. The NAA/tCho ratio was 27.5 % lower than the control in white matter (Supplemental Table 1). To the best of our knowledge, we report low NAA in brain MRS in a patient with a classical homocystinuria for the first time.

Previous studies reported that higher homocysteine levels were associated with lower intellectual quotient (IQ) [[22], [23], [24]]. The IQ was significantly (p < 0.001) lower in affected siblings with classical homocystinuria (IQ of <70) compared to their healthy siblings (average or high average IQ) [25]. The IQ was significantly higher in patients with classical homocystinuria diagnosed by positive newborn screening (IQ of 97.00; range 89–110) or by positive family history (IQ of 98.50; range 84–116) compared to patients diagnosed symptomatically (IQ of 79.00; range 39–113) [26]. Its seems that early diagnosis improves neurodevelopmental outcomes. Marginal to mild elevations of homocysteine levels (<20 μmol/L) resulted in cognitive dysfunction and dementia compared to individuals with normal homocysteine levels. The vascular brain lesions (e.g., increased signal intensities in cerebral white matter) were attributed to cognitive dysfunction [[27], [28], [29]]. Both of our patients showed below the average working memory index score (Patient 1 at 5th percentile and Patient 2 at 21st percentile). Both of our patients exhibited deficits in their simple motor speed and fine motor coordination, consistent with previous studies. Processing speed, memory, and executive functioning were below average in Patient 1 with higher homocysteine levels, whereas her younger sister had average scores with lower homocysteine levels. It seems that there are specific deficits in patients with classical homocystinuria that needs to be studied.

Neurodevelopmental disorders and abnormalities in neurotransmitter pathways were reported in inherited metabolic disorders with psychiatric presentations. Acute psychiatric episodes were likely due to an imbalance between glutaminergic and GABAergic neurotransmitters [30]. It was reported that elevated homocysteine levels affected the GABAergic signaling and altered NMDA-mediated glutamatergic transmission by activating the NMDA receptors in the central nervous system [18]. These changes led to an imbalance between glutaminergic and GABAergic neurotransmitters and psychiatric presentations in classical homocystinuria [31]. This might be a possible explanation for an acute psychotic episode in Patient 1 and anxiety disorder in both patients without significant neurodevelopmental disorders.

It has been reported that elevated homocysteine levels affect the bone metabolism by increasing intracellular calcium; by agonizing NMDA receptor leading to disruption of the mitochondrial membrane potential; and by increasing reactive oxygen species (ROS). The ROS-mediated oxidative pathways led to apoptosis affecting osteoblast activity leading to decrease in osteoblastogenesis, decreased mineralization and osteoporosis. Increased ROS and increased oxidative stress produced superoxide anions and reduced nitric oxide synthesis leading to decreased bone blood flow and activation of osteoclastogenesis, further contributing to osteoporosis [32]. It is likely that markedly elevated homocysteine level for a long period of time might have contributed to severe early onset skeletal phenotype in Patient 1 compared to her younger sister who does not have skeletal phenotype.

It has been reported that elevated homocysteine levels cause thrombosis by endothelial cell desquamation; by smooth muscle cell proliferation, and intimal thickening; by activation of thrombotic systems (activation of factor V, protein C, thrombomodulin expression and inhibition of tissue plasminogen activator); and by decreased synthesis of nitric oxide [33]. Our Patient 2 had a vascular phenotype and was on oral hormonal contraceptive, which may have contributed to her early onset vascular phenotype despite she had lower homocysteine levels than her older sister.

In conclusion, we report two siblings with classical homocystinuria with variable phenotypes, late diagnosis and good response to treatment. There were significant improvements in biomarkers despite late treatment onset and pyridoxine non-responsive phenotype. It seems that the phenotypes are a continuum between pyridoxine responsive and pyridoxine non-responsive phenotypes. Higher homocysteine levels are likely to result in worse outcomes in skeletal, and central nervous system phenotypes. Further studies are required to assess the homocysteine levels and their effects on the phenotypic severity in classical homocystinuria.

CRediT authorship contribution statement

Randa Sultan: Writing – review & editing, Writing – original draft, Methodology, Data curation. Jordan Urlacher: Writing – review & editing, Writing – original draft, Data curation. Taryn Athey: Writing – review & editing, Writing – original draft. Peter Kannu: Writing – review & editing, Writing – original draft. Peter Seres: Writing – review & editing, Writing – original draft, Data curation. Saadet Mercimek-Andrews: Writing – review & editing, Writing – original draft, Supervision, Methodology, Investigation, Data curation, Conceptualization.

Consent

The case report consent was signed by the patients. Our institution does not require approval for case reports.

Declaration of competing interest

All Authors declare that they have no conflicts of interest.

Appendix A. Supplementary data

Supplementary Table 1 shows choline, n-acetyl aspartic acid and creatine level for age appropirate control and for Patient 1.

Data availability

Data will be made available on request.