RNA was extracted from Patient 1 and control fibroblast cell lines and cDNA was synthesized using the Superscript IV kit (ThermoFisher, Waltham, MA) according to the manufacturer’s protocol. Complementary DNA (cDNA) was amplified using polymerase chain reaction (PCR) with three sets of primers and sent for Sanger sequencing. Set 1 was proximal to the mutation (forward primer in exon 3: GAGAATGAGCAGACGGAGGA, reverse primer in exon 4: CAATGGACAGGGCGATGAAG). Set 2 surrounded the mutation (forward primer in exon 5: TCCAAGTCTGCAGTGGTGTT, reverse primer in exon 6: GGAAGCGACTCAGAGGCATA). Set 3 also surrounded the mutation (forward primer in exon 5: ATTATGTCTCCAAGTCTGCAGT, reverse primer in exon 7: AGAGTAGCGGACACCTTTCA). qRTPCR was performed using the Fast SYBR Green Master Mix (ThermoFisher, Waltham, MA) with the three sets of primers according to the manufacturer’s protocol. Delta cycle threshold (dCT) values were calculated by normalizing the cycle threshold (CT) value of each sample to the loading control (GAPDH). Double delta cycle threshold (ddCT) values were calculated by normalizing the dCT value of each patient sample to the corresponding control sample. Relative expression levels were then calculated for each primer set and compared using a one-way Analysis of Variance and post-hoc Tukey honest significant difference Test.

Fibroblasts from controls and Patient 1 were grown to confluence in 10 cm² tissue culture dishes in Dulbecco’s Modified Eagle Medium (ThermoFisher, Waltham, MA) with penicillin/streptomycin and 10% Fetal Bovine Serum (Sigma-Aldrich, St. Louis, MO) and analyzed for intracellular metals by inductively coupled plasma mass spectrometry (Trace Metals Laboratory, Harvard TH Chan School of Public Health, Boston, MA) as described previously [17]. The internal standard was 50 ppb indium. Briefly, cell samples were digested with 2 mL/g total wet weight nitric acid for 24 hours at room temperature. Ultrapure water was used for final sample dilution.

Patient 1 is a 9-year old female from the UAE with progressive generalized dystonia and brain mineralization. She was the product of an uncomplicated term pregnancy. Her initial motor development was delayed: she sat at 8 months, and never stood. She said her first words around 1 year of age. At 2 years of age, she developed progressive generalized dystonia. She also displayed failure to thrive and required a G-tube for supplemental nutrition. Her parents are first cousins, and she has 3 healthy brothers, ages 3, 13, and 14 years (Figure 1A). There are no additional affected family members. There were no significant environmental exposures to heavy metals.

We first evaluated her at the age of 5 years. At that time, she could speak about 40-50 words, but no phrases or sentences. Her speech was dysarthric. She could W-sit and crawl but could not walk. She had a generalized dystonia and a dystonic smile. Ophthalmology evaluation was normal.

Her initial brain magnetic resonance imaging (MRI) was significant for T1 hyperintensity of bilateral globi pallidi and substantia nigra (Figure 2A–F). There was no abnormal T2 hyperintensity. Susceptibility weighted imaging disclosed no evidence of abnormal susceptibility artifact. Computed tomography scan of the brain did not demonstrate any calcifications.

Initial laboratory parameters demonstrated an elevated serum manganese (Mn) level of 64.2 μg/L (normal range 4-16.5 ug/L). Liver function tests, serum copper, ceruloplasmin, calcium, parathyroid hormone, ferritin, and urine oligosaccharides were normal. The complete blood count (CBC) demonstrated neutropenia. Serum iron was elevated at 207 mg/dL. Hair analysis corroborated elevated manganese storage, with a hair Mn concentration of 516 ppb. By comparison, the mother’s hair Mn was 147 ppb and her brother’s hair Mn concentration was 214 ppb. Serum Mn levels in her younger brother were initially elevated at 44.1 μg/L, but subsequently normalized. Her mother’s serum Mn level was mildly elevated at 28.7 μg/L. Her father’s serum Mn level was normal. Manganese excretion in urine was undetectable. Blood and urine levels of additional divalent heavy metals (arsenic, mercury, lead, copper, and zinc) were normal. Liver ultrasound was normal, with no evidence of portosystemic shunting or arteriovenous malformation.

Analysis of cultured skin fibroblasts for heavy metal content in Patient 1 compared to 2 healthy controls demonstrated similar levels of manganese and iron, but significantly increased levels of zinc (p < 0.05, Student’s t-test) (Table 2). Sequencing and deletion/duplication analysis of the SLC30A10 gene was negative [7, 8, 10, 12]. Whole exome sequencing demonstrated a homozygous intronic variant in the SLC39A14 gene, c.751-9C>G (NM_001128431, [Hg19] chr8: 22273273 C>G) (Figure 1B), which was not found in the gnomAD database. Both parents and two of the three unaffected siblings were heterozygous carriers, and the third unaffected sibling was homozygous for the wild type allele. Gel electrophoresis suggested both normal and abnormal splicing in Patient 1 fibroblast cells, and sequencing of cDNA demonstrated inclusion of the entire intronic sequence between exons 5 and 6. qRT-PCR analysis on fibroblast cells confirmed aberrant splicing, demonstrating significantly decreased SLC39A14 transcript levels around the mutation compared to proximal to the mutation (Figure 1C–E).

Patient 1 has been followed intermittently for 4 years at our institution, although she and her family returned to the UAE for long periods between hospitalizations. During this time, her blood manganese levels have ranged between 49.3-79.6 μg/L (Figure 3). Beginning at the age of 5 years, we treated her with repeated 5-day cycles of chelation with parenteral edetate calcium disodium (CaNa₂EDTA) (40 mg/kg/day) over a 4-year period. The CaNa₂EDTA was well tolerated, except for a couple of occasions when it had to be held because of worsening neutropenia. The CaNa₂EDTA effected a urine Mn diuresis from negligible concentrations pretreatment to >25 μg/L; however, serum Mn was not significantly altered. We attempted to coordinate monthly courses of chelation, but because of practical limitations with an international patient, she would often go several months or more between courses. Each course of chelation was associated with a small window of subjective improvement in dystonia and motor functioning, typically lasting weeks.

Two oral chelants, dimercaptosuccinic acid (DMSA, 200 mgs PO BID for 28 days) and d-penicillamine (125 mg PO BID for 30 days), were trialed and proved ineffectual in increasing urinary Mn excretion or lowering blood Mn levels. The two oral chelants were given in separate trials, with a substantial wash-out period in-between their use. We attempted a treatment trial with para-amino salicylic acid (PAS), a drug historically used for tuberculosis with some evidence of benefit in occupational studies of Mn toxicity [18, 19] and animal studies [20, 21]. The patient was not able to swallow the medication, and it could not be administered through her G-tube or effectively dissolved because of its enteric coating.

In addition, since the age of 7 years, the patient has been maintained on a reduced manganese diet using modular formulas. At the age of 8 years, we introduced days with a “Mn free diet”, or days of the week when she receives negligible Mn from oral food. To achieve this, she is entirely fed via manganese-depleted formula for these days, and is supplemented with a manganese free multivitamin. Her macronutrient and micronutrient profile on this diet is summarized in Supplementary Tables 1–3. We began with 2 days a week of “Mn free diet”, and then ultimately increased to 3 days a week. We considered supplementation with iron to decrease gastrointestinal manganese absorption, but opted not to because of baseline high-normal serum iron levels.

There were 2 longer periods of time when the patient went off diet and parenteral chelation therapy while temporarily returning home to the UAE due to logistical hurdles in securing the customized Mn depleted formula. There was a 6-month period between ages 6-7 years, and another 6-month period between ages 7-8 years. Both of these treatment hiatuses were associated with a marked clinical deterioration from which she did not entirely recover following subsequent chelation treatments.

She has also been managed symptomatically for her dystonia with oral trihexyphenidyl and botulinum toxin injections. In addition, she underwent tendon lengthening surgery. She was previously trialed on baclofen, clonazepam, and L-dopa/carbidopa with little clinical benefit. She is being considered for deep brain stimulation.

Despite these interventions, her serum manganese levels have remained essentially unchanged (Figure 3) and she has demonstrated a gradual disease progression. She currently has more difficulty W-sitting and crawling. Her dysarthria has progressed. She reaches with a significant action tremor. Her intelligence has scored within the average range on formal neuropsychological evaluations, within the limits imposed by her motor impairment. She receives some of her nutrition though a G-tube because of feeding difficulties.

Serial brain MRIs, the most recent of which was at the age of 7 years, have demonstrated stable mineralization of brain parenchyma evidenced on her most recent imaging as T1 hyperintensity of basal ganglia, hypothalamus, anterior commissure, amygdala, cerebral peduncles (including the substantia nigra), dorsal pons, superior cerebellar peduncles, dentate nuclei, cerebellar folia, and pyramids. Abdominal MRI did not demonstrate hepatic manganese accumulation, and liver function tests in blood have remained normal.

Patient 2 presented at 2 years old for evaluation of a progressive movement disorder and loss of ambulation skills. She was born full term via non-spontaneous vaginal delivery, weighing 3.1 kg. There were no post-birth complications. Her mother received prenatal care. There were no maternal or fetal infections noted. The pregnancy was complicated by gestational diabetes, which was well controlled throughout the pregnancy.

Her early development was normal. At the age of 18 months, she first started having trouble walking. The parents first noticed that she started to limp and then started falling frequently. This progressed to her walking on the tips of her toes and then over a few months, she became unable to walk at all. A few weeks later, she was no longer able to sit independently.

Along with this regression in motor milestones, her parents also noticed increased stiffness in all of her extremities and odd positioning of her upper extremities and fingers. Symptoms have continued to progress in a caudal to rostral direction—she subsequently began to develop problems with chewing food and sticking out her tongue. She also lost the ability to say a few words, such as “dada” and “mama”, which she was able to do before the regression started. She is still able to swallow but the difficulty in chewing has resulted in poor weight gain. There is no history of seizures.

There is an older sister (10 years old) and younger brother (6 months old) who are unaffected. The parents are healthy and deny consanguinity. Two maternal uncles died at 18 and 19 years of age of an unspecified liver condition. There are no other early deaths in the family. There is a paternal niece with dystonia diagnosed by another provider at the same hospital, but no genetic testing was pursued.

Patient 2 was seen by specialists in her home country (UAE) and baseline investigation included a gene panel for pantothenate kinase-associated neurodegeneration (PKAN), mitochondrial DNA sequencing, full gene sequencing of SLC30A10, and whole exome sequencing, which were all negative. Her MRI showed abnormal T1 shortening within the globus pallidus, brainstem, cerebellum, and, to a lesser extent, within the surrounding cerebral white matter (Figure 2G–I).

Additional testing included normal electroencephalography and normal biochemical labs, including plasma amino acids, urine organic acids, serum lactate, plasma carnitine, and plasma acylcarnitines. Initial testing demonstrated an elevated serum manganese level of 78 μg/L. Liver function tests, serum copper, ceruloplasmin, calcium and ferritin were normal. Zinc level was elevated at 123 mcg/dL (normal range: 29-115 mcg/dL). CBC and serum iron were normal. The NBIA sequencing panel completed by the Knight Diagnostic Laboratory at OHSU identified the same homozygous variant in intron 5 of SLC39A14, c.751-9C>G.

Feeding and speech evaluations performed in the United States demonstrated decreased oral motor skills, using suckling, munching, and suction to eat. She did not exhibit tongue lateralization and mastication to manipulate food. A soft pureed diet was recommended as well as a swallow study. Patient 2 returned to the UAE and received chelation therapy for a year but parents decided not to continue with it.

Biallelic exonic mutations in the manganese transporter SLC39A14 have recently been implicated in a pediatric-onset neurodegenerative disorder associated with brain manganese accumulation and clinical signs of manganese neurotoxicity, including parkinsonism-dystonia. Previously, 11 patients from 7 families had been reported with exonic missense, nonsense, and frameshift mutations[13–15]. In this article, we report 2 additional unrelated patients with a novel homozygous intronic variant that leads to aberrant splicing. Inclusion of the intron between exons 5 and 6, confirmed by cDNA sequencing, results in a stop codon at the beginning of the intron and suggests there is thus less functional protein. All 13 patients presented in infancy or early childhood with rapidly progressive dystonia, loss of developmental milestones, and/or bulbar dysfunction, and all had elevated levels of Mn in blood or serum.

The function of the SLC39A14 transporter has not been fully elucidated; however, it is presumed to function in hepatic Mn uptake for biliary excretion through SLC30A10 [8]. Aberrant SLC39A14 function likely results in insufficient clearance of manganese from blood, and subsequent accumulation in brain. It is not known what additional transport roles it may have in other organs or cellular compartments. The SLC39A14 protein localizes to the plasma membrane, specifically to the basolateral domain in polarized hepaocytes where it likely functions in hepatic Mn uptake from blood, and also localizes intracellularly where it may play a role in intracellular trafficking [8, 22, 23]. In addition to manganese toxicity in brain, it is not yet clear if intracellular manganese deficiency in other cellular compartments may contribute to the pathogenesis of this disorder. Arguing against this, our patient did not have any abnormalities on transferrin mass spectrometry to suggest dysfunction of the manganese dependent enzyme β-1,4-galactosyltransferase that is seen in SLC39A8 gene related disorder.

There have been case reports that SLC30A10 can be effectively treated with IV chelation with calcium disodium edetate and iron supplementation, the latter to reduce intestinal manganese absorption through competitive inhibition across shared transporters. The previous report on this disorder suggested biochemical, clinical, and neuroimaging improvement with monthly chelation with intravenous calcium disodium edetate [10, 18, 24]. In terms of SLC39A14-associated maganism, the previous reports suggest variable clinical response with chelation with CaNaEDTA. In three patients, chelation led to a reduction in Mn levels; one patient started on chelation at 17 years old continued to deteriorate, a second patient had reported worsening of cervical dystonia and discontinued treatment, and a third patient had some clinical improvement in terms of dystonia and motor abilities [13–15]. A fourth patient started on IV chelation at five years old had major clinical improvement and was able to walk again [13]. Patient 1 in this article was also treated with IV chelation over a four year period, and although urine diuresis of Mn was observed, there was no change in serum Mn levels and her disease continued to progress. Thus, further investigation is needed to understand the variable clinical response, which may be influenced by the age at which treatment is started, disease severity, the specific mutation, or other factors. One limitation of parenteral CaNa₂ EDTA is its inability to cross the blood brain barrier, and so its ability to remove brain manganese may be quite limited. Another practical limitation was that parenteral chelation in our patient required hospitalization, which made consistent monthly courses difficult to arrange. It is possible that this treatment may have been more effective if the frequency of administration was increased.

PAS, an anti-tuberculous drug that has some chelating properties for manganese, has potential for treatment of this disorder since it can cross the blood brain barrier [18–21]. Further, it is more practical to administer since it is an oral medication and can be taken continuously. Unfortunately, our patient was not able to tolerate PAS because she could not swallow capsules and they could not be administered through her G-tube or dissolved given their enteric coating.

For Patient 1, despite significant dietary restriction of manganese and multiple courses of parenteral chelation, we were not able to lower serum manganese levels to an appreciable degree or effect any significant improvement in disease course. Further, brain imaging did not demonstrate any improvement in established mineralization. However, it is likely that this treatment has slowed the progression of her disease as evidenced by her dramatic clinical deterioration when stopping her therapies for 6 month intervals. We have lost contact with Patient 2 as she returned to the UAE and we have not been able to sustain communication with her family.