Natural history and genotype‐phenotype correlation of pantothenate kinase‐associated neurodegeneration

Abstract Aims To investigate the natural history and genotype‐phenotype correlation of pantothenate kinase‐associated neurodegeneration. Methods We collected data of patients with PKAN by searching from available publications in English and Chinese. Patients diagnosed in our center (Peking University First Hospital) were also included. The difference in natural history and genotype between early‐onset (<10 year of age at onset) and late‐onset patients (≥10 year of age at onset) with PKAN was compared. Results A total of 248 patients were included. The median age at onset was 3.0 years in the early‐onset group and 18.0 years in the late‐onset group. Dystonia in lower limbs was the most common initial symptom in both groups. In the early‐onset group, the median interval between the disease onset and occurrence of oromandibular dystonia, generalized dystonia, loss of independent ambulance was 6.0 years, 5.0 years, and 5.0 years. The corresponding values in late‐onset group were 1.0 year, 4.0 years, and 6.0 years. About 20.0% died at median age of 12.5 years and 9.5 years after the onset in early‐onset group. About 2.0% of the late‐onset patients died during the follow‐up. A total of 176 mutations were identified. Patients carrying two null alleles in PANK2 showed significantly earlier age of disease onset and progressed more rapidly to loss of independent ambulance. Conclusions This study provided a comprehensive review on the natural history and genotype of 248 patients with PKAN. The results will serve as a historical control data for future clinical trial on PKAN.

independently until 15-40 years after onset. 2 However, there are patients with early onset but slow progression or late onset with rapid progression. The typical manifestation of neuroimaging is the "eye-of-the-tiger" sign ( Figure 1). In 2001, biallelic recessive mutations in PANK2 were revealed in PKAN. 3 The mutations result in a decreased activity of pantothenic acid kinase 2, which is a key regulatory enzyme in coenzyme A production by pantothenic acid, possibly leading to the reduction of coenzyme A 3 and accumulation of its substrates, cysteine, which may chelate iron accumulation. 4 The reported prevalence was 1-3/1 000 000, 4 but the accurate prevalence of PKAN is unclear yet.
Understanding of the natural history of PKAN is essential for decision-making on disease management by physicians and setups of future clinical trials. However, there are few large sample studies. We summarized the natural history and genotypic spectrum of 248 patients with PKAN by searching and analyzing the data from available publications. Patients diagnosed in our center (Pediatric Department, Peking University First Hospital) were also included and pooled for the analysis. We compared the differences in natural history and genotype between early-onset and late-onset patients with PKAN.

| Literature searching strategy
Case series studies or case reports on PKAN containing information on the clinical and genotypic spectra of patients were searched from the following databases: PubMed, Embase, Cochrane Library, Web of Science, China National Knowledge Infrastructure (CNKI), and WAN FANG DATA (the latter two are Chinese public searching databases).
The publication was in English or Chinese. The searching deadline was July 5, 2019.

| Inclusion criteria of patients from our center
Patients who met all the following criteria were enrolled: (a) dystonia; (b) "eye-of-the-tiger" sign on brain MRI; and (c) biallelic pathogenic or likely pathogenic variants in PANK2. The study was approved by the Medical Ethics Committee of Peking University First Hospital.
Informed consent was obtained from the parents of the children.

| Data collection
The following information of the patients was collected: basic information (gender, age, and country), symptoms at onset and during the progression, examination (fundus, cranial MRI, and cranial CT), and mutations in PANK2 in each individual. Patients were divided into early-onset (<10 year of age at onset) and late-onset groups (≥10 years of age at onset).

| Statistical analysis
The enumeration data were expressed as frequency and percentage. The measurement data were expressed as median (min-max).
Chi-square test was performed to compare the difference of genotype between the early-and late-onset patients. Mann-Whitney U test was conducted to evaluate the relationship between the genotype and age of onset, and to compare the differences in time of occurrence of the major symptoms between the early-and late-onset patients. P < .05 was considered statistically significant.
Survival analysis was used to compare the progression of motor disability between the early-and late-onset patients and the progression among the patients with 0, 1, and 2 null alleles as well.
Variables with a significance level < .1 in the survival analysis were F I G U R E 1 Neuroimaging of patients with PKAN. A, T2-weighted MRI of the brain showed a specific pattern of hyperintensity (indicated by the arrow) within the hypointense medial globus pallidus ("eye-of-the-tiger" sign). B, Brain CT showed calcification in medial globus pallidus, which was indicated by the arrow then tested in multivariate logistic regression analysis. Because detailed clinical data were not described in some of the studies, we used "n" to represent the number of available patients for the analysis. We summarized the available data when we analyzed the proportion, age of symptoms, and the interval between disease onset and symptoms. Missing data were removed when regression analysis was performed. When we analyzed the distribution of null allele in early-onset patients and late-onset patients, and the correlation between the age of onset and genotype, and multivariate logistic regression was performed, we excluded cases with only one mutation detected.

| General information of the patients
A total of 248 patients (132 males and 116 females) were included.
Of these patients, 242 were from 66 reported studies (Supporting Information), and 6 were from our center. Of the 248 patients, 128 were early-onset patients, and 120 were late-onset patients.
Among the early-onset patients, the median age at onset was

| Onset of the disease
As for the initial symptom, dystonia in lower limbs was the most common in both early-and late-onset patients, followed by dystonia in upper limbs. Some patients initially presented with vision loss, generalized dystonia, involuntary movement, or cognitive impairment.
Developmental delay (DD) was more common in early-onset patients before disease onset. DD was seen in 36.4% (24/66) of the early-onset patients. Of the 25 patients with onset before 2 years of age, DD was the initial symptom in 5 patients, whereas majority of late-onset patients (95.5%, 42/44) showed normal developmental milestones before disease onset.

| Progression of dystonia
The occurrence of dystonia of limbs, oromandibular dystonia, generalized dystonia, and loss of function of independent walking was compared between the early-and late-onset patients,

| Psychological and behavior disturbance
Compulsive behavior, emotional liability, anxiety, depression, or attention deficiency was reported in both early-onset and late-onset groups, with slightly more common in early-onset patients. About 44.9% (48/107) of early-onset and 38.8% (38/98) of late-onset patients were reported to present with psychological and behavior problems (P = .412).

| Cognitive decline
The assessment of cognitive function may be interfered by the limited mobility in PKAN. After the disease onset, cognitive decline was reported in 54.0% (34/63) of the early-onset patients and 33.8% (24/71) of the late-onset patients (P = .025).
Few studies reported calcification in globus pallidus. Seven out of eleven patients (63.6%) showed calcification in brain CT. All patients with calcification harbored biallelic mutations.  Figure 3A). Most of the mutations were located in the core catalytic region of PANK2. c.1583C>T (p.T528M) was the most frequent mutation, which was identified in 23 patients.

| D ISCUSS I ON
PKAN is classified into classic and atypical type on the basis of both age of onset and disease progression. However, there are patients with early onset but slow progression or late onset with rapid progression, which were classified as "intermediate form" in some studies. 5 In our study, patients were classified into early-onset (before The "eye-of-the-tiger" sign is the typical manifestation in neuroimaging, which is caused by the accumulation of iron in globus pallidus (abnormally low T2WI signal), tissue necrosis, and edema (high signal in T2WI). Seven out of eleven patients (63.6%) showed calcification in brain CT, which was probably associated with extracellular calcium deposits in globus pallidum. 8 In contrast to idiopathic basal ganglia calcification, the calcification of patients with PKAN was limited to globus pallidus. However, CT was not F I G U R E 2 Comparison of disease progression between the early-onset and late-onset patients. (A,B) represented the scatter plots of the symptoms in the disease course of the early-onset (A) and late-onset patients (B), respectively. The abscissa represented the symptoms, and the ordinate denoted the age at which each symptom was shown. The dot corresponded to each individual, and the red triangle indicated the median. OMD, oromandibular dystonia; LD, limb dystonia; GD, general dystonia; and LOA, loss of independent ambulation (C). The comparison of the time of occurrence for each symptom after disease onset between early-onset and late-onset patients. **denotes P < .05, and "ns" indicates no statistical difference. D, Survival curve of ability of independent ambulation from disease onset in early-onset and lateonset patients. E, Survival curve of ability of independent ambulation with age in early-onset and late-onset patients. F, Survival curve of survival patients with age in early-onset and late-onset patients considered a routine examination for patients suspected of PKAN; hence, the incidence of basal ganglion calcification might be underestimated.
As for the genotype-phenotype correlation, we demonstrated that patients with two null alleles in PANK2 showed significantly earlier age of disease onset and progressed more rapidly to loss of independent ambulance. The proportion of patients carrying two null alleles was significantly higher in the early-onset patients than that in late-onset patients. PANK2 mutations likely decrease the function or abundance of PANK2 protein, thereby decreasing the production of coenzyme A, which plays an important role in ATP and fatty acid synthesis and in neurotransmitter metabolism.
PANK2 dysfunction also results in the accumulation of cysteine and other substrates, possibly leading to oxidative stress injury of neurons. PANK2 mutations are postulated to cause mitochondrial dysfunction because PANK2 enzyme localizes to the mitochondria in the human brain. 9 The lower the activity of the enzyme residue, the earlier the age of onset in the previous study 10  represented the regulatory region, 212-570aa represented the catalytic core region. Black, red, blue, and green represent missense mutation, frameshift mutation, nonsense mutation, and small fragment deletion, respectively. B, The proportion of patients carrying 0, 1, and 2 null alleles in early-onset and late-onset patients. In the early-onset patients, 27.5% (33/120) carried two null alleles, and 15.8% (19/120) had one null allele, whereas, in the late-onset patients, only 0.8% (1/115) carried two null alleles, and 20.9% (24/115) carried one null allele. C, The comparison of numbers of null alleles with the age of onset. The media age of disease onset in patients with 0 null allele, one null allele, and two null alleles was 11 y (0.33-52.0), 13.0 y (1.0-48.0) and 2.5 y (1.0-11.0). Patients with two null alleles showed earlier age of disease onset. "**"denotes P < .05, "***" denotes P < .001, "ns" indicates no statistical difference III clinical trial, there was no significant difference in primary endpoint between treatment groups and placebo groups according to Retrophin's announcement.
There are some limitations in this study. First, the original data were incomplete in some of the publications. Therefore, it was not possible to fully and comprehensively analyze the phenotype of the 248 patients. Second, most of studies were retrospective, which probably affect the accuracy of the data.

| CON CLUS IONS
This study provided a comprehensive review on the natural history and genotype of 248 patients with PKAN. The time curve of progression in dystonia varied between the early-onset and lateonset patients. Behavior problems were not uncommon in both early-and late-onset patients. Cognitive decline was reported, but interference of motor handicap on the assessments of cognitive function needs to be considered. A total of 176 mutations were summarized. We found that patients with two null alleles in PANK2 showed significantly earlier age of disease onset. This suggested a correlation between the genotype and phenotype. The results of this study may serve as a historical control data for future clinical trial on PKAN.

This study was supported by National Science and Technology
Major Project of the Ministry of Science and Technology of China (Grant nos. 2017ZX09304029-006).

CO N FLI C T O F I NTE R E S T
The authors declare no conflict of interest.

E TH I C A L S TA N DA R D
This study was approved by the institutional review board of the ethics committee of Peking University First Hospital and has therefore been performed in accordance with the ethical standards laid down in the 1964 Declaration of Helsinki and its later amendments. The parents of all participants had been provided written informed consent for the use of the children's information for scientific purposes.