Clinical Review & Education
Review
Primary Familial Brain Calcification With Known Gene Mutations A Systematic Review and Challenges of Phenotypic Characterization Vera Tadic, MD; Ana Westenberger, PhD; Aloysius Domingo, MD; Daniel Alvarez-Fischer, MD; Christine Klein, MD; Meike Kasten, MD
IMPORTANCE In the past 2 years, 3 genes (SLC20A2, PDGFRB, and PDGFB) were identified as causative of primary familial brain calcification (PFBC), enabling genotype-specific phenotyping.
Supplemental content at jamaneurology.com
OBJECTIVES To provide a systematic literature review on the neuroimaging and clinical phenotype of genetically confirmed PFBC and summarize known pathophysiological mechanisms, to improve and harmonize future phenotype description and reporting by addressing data gaps, and to develop uniform definitions for clinical characterization. EVIDENCE REVIEW We systematically searched the MEDLINE database among articles published from January 1, 2012, through May 31, 2014, for the 3 genes and selected 25 articles from all records (n = 75) and from sources cited in the reference lists. Only genetically confirmed cases with individual clinical information were included, leaving 15 reports. Predefined categories for data extraction were different neurologic and psychiatric symptoms, imaging results, and age at onset (AAO). We also assessed availability of information to estimate possible bias. FINDINGS We included a total of 179 cases, 162 of which belong to 25 families. Availability of information ranged from 96.6% for ethnicity to 24.4% for AAO. All cases had calcifications on comprehensive cranial computed tomography, most frequently located in the basal ganglia (70.6%), subcortical white matter (40.8%), cerebellum (34.1%), or thalamus (28.5%). Mean (SD) AAO was 27.9 (22.3) years, and the AAO was comparable across genes (P = .77). The most frequently described signs were movement disorders, such as parkinsonism (12%) and dystonia (19%). Penetrance of the imaging phenotype was 100% compared with only 61% of the clinical phenotype. We propose a novel definition of disease status by specifying PFBC into genetic, clinical, and imaging phenotypes. Pathophysiological pathways converge on impaired phosphorus homeostasis and integrity of the blood-brain barrier. CONCLUSIONS AND RELEVANCE Especially in rare conditions, meta-analyses are the most suitable tool to extract reliable information on the natural course of a disease. For future analyses, we provide a minimal data set that can be used for systematic clinical and imaging data collection in PFBC and that will also improve informed counseling of patients.
Author Affiliations: Institute of Neurogenetics, University of Lübeck, Lübeck, Germany (Tadic, Westenberger, Domingo, AlvarezFischer, Klein, Kasten); Graduate School for Computing in Medicine and Life Science, University of Lübeck, Lübeck, Germany (Domingo); Department of Psychiatry, University of Lübeck, Lübeck, Germany (Alvarez-Fischer, Kasten). Corresponding Author: Christine Klein, MD, Institute of Neurogenetics, University of Lübeck, Ratzeburger Allee 160, 23538 Lübeck, Germany (christine.klein @neuro.uni-luebeck.de).
JAMA Neurol. 2015;72(4):460-467. doi:10.1001/jamaneurol.2014.3889 Published online February 16, 2015. 460
Section Editor: David E. Pleasure, MD. (Reprinted) jamaneurology.com
Copyright 2015 American Medical Association. All rights reserved.
Downloaded From: http://archneur.jamanetwork.com/ by a Penn State Milton S Hershey Med Ctr User on 05/22/2015
Primary Familial Brain Calcification
F
Review Clinical Review & Education
irst reports on intracerebral calcifications date back to the 19th century with an autopsy report from Delacour1 of an individual with bilateral cerebral calcifications most prominent in the striatum. However, the best-known case description is by Fahr2 in 1930, although the calcifications were localized in the white rather than the gray matter and, in retrospect, were likely secondary.3 Nonspecific calcifications in the brain are a common incidental and age-dependent finding on neuroimaging, with an estimated frequency of 1% to 20%.4-6 For brain calcification of unknown origin, the term idiopathic basal ganglia calcification (IBGC) was coined. However, with the discovery of the genetic causes of brain calcifications following an autosomal dominant trait, this disorder ceased to be idiopathic. Thus, the term IBGC is no longer appropriate for the disorder in patients with genetically confirmed disease; a recent report has suggested replacing IBGC with the term primary familial brain calcification (PFBC).7 The neuroimaging phenotype consists of bilateral calcifications in the brain, predominantly in the basal ganglia but also in the cerebellum, thalamus, and brainstem. The clinical phenotype encompasses neurologic and psychiatric features. The neurologic signs are variable in type and severity; those most commonly described are movement disorders, but seizures, headache, and cerebellar symptoms may occur. Psychiatric features are also variable and include cognitive impairment, mood disorders, psychotic symptoms, and obsessive-compulsive symptoms, although fewer data are available compared with neurologic signs.8 Although the exact pathophysiological features remain elusive, recent discoveries offer important clues toward the cause and disease mechanisms of PFBC.9 In particular, mutations in 3 different genes have been identified as causative of PFBC. First, SLC20A2 (OMIM 158378) was described in February 201210 as coding for type III sodium-dependent inorganic phosphate (P i ) transporter 2 (PiT-2). The second reported gene, PDGFRB (OMIM 173410), was described in January 201311 as encoding 1 of 2 receptors for plateletderived growth factor. The gene encoding its major ligand, PDGFB (OMIM 190040), was the third gene identified in September 2013.6 The different mutations in all 3 genes share loss of function as the likely disease mechanism.12 Mutations in the SLC20A2 gene lead to an accumulation of Pi and subsequently to calcium phosphate deposition. Mutations in PDGFB and PDGFRB also result in calcification, however, through indirect processes (Figure 1). Although the receptor and its ligand are also expressed in neurons, data from selective knockout studies indicate impaired recruitment of pericytes to endothelial cells, resulting in a dysfunctional blood-brain barrier and thereby contributing to brain calcification.6 A recently published abstract13 suggested involvement of a fourth gene in the pathophysiological features of PFBC by reporting a mutation in XPR1 as the cause of disease in 1 family with PFBC and no mutations in any of the known PFBC genes. In keeping with previously proposed disease mechanisms, XPR1 appears to alter Pi homeostasis. However, this finding awaits peer-reviewed publication and independent replication. The gene discoveries were followed by a number of mutational screening studies and clinical reports of individual cases or families with confirmed genetic causes (ie, definitive PFBC), thus for the first time enabling a summary of available phenotypic information based on accurate genetic classification. This step calls for a sysjamaneurology.com
Figure 1. Proposed Disease Mechanism of Brain Calcification A
B
PDGFRB
PDGFB
PiT-2
Brain cell
Pi XPR1
Calcium phosphate deposits
A, Under physiological conditions, endothelial cells secrete platelet-derived growth factor B (PDGFB) that binds to PDGF receptor β (PDGFRB) on pericytes. This process is necessary for the recruitment of pericytes and correct proliferation and migration along the blood vessels. Inorganic phosphate (Pi) in the brain parenchyma is taken up by Pi transporter 2 (PiT-2). From the brain cell, Pi is secreted through XPR1. B, Under pathologic conditions, the PDGF ligand or receptor loses its normal function, leading to pericyte dysfunction and an altered blood-brain barrier. Pi accumulates in the brain parenchyma owing to its reduced uptake by PiT-2. The proposed disease mechanism of mutated XPR1 has not yet been published in detail. However, the mechanism was found to reduce efflux of phosphate from cells.13 Whether glial and/or neuronal cells participate in the diseased state of primary familial brain calcification remains unknown.
Figure 2. Flowchart of the Literature Review Literature search in PubMed Search terms: PFBC, IBGC, brain calcification, Fahr disease, OR calcification AND SLC20A2, PDGFRB, OR PDGFB
75 Records identified through database searching from January 1, 2012, to May 31, 2014
20 Records remained after removing double counts, articles without clinical information, and false-positive findings
5 Records included by reviewing reference lists and sources
25 Records included in qualitative synthesis
15 Records included in quantitative synthesis (meta-analysis) 179 Cases (90.5% familial and 4.5% sporadic cases)a
We identified 179 cases of primary familial brain calcification (PFBC), 162 of whom belonged to 25 families. IBGC indicates idiopathic basal ganglia calcification; PFBC, primary familial brain calcification. a
Information was unavailable for the remaining 5.0% of cases.
tematic literature review on the findings of comprehensive cranial computed tomography (CCT) and the neurologic and psychiatric phenotype of genetically confirmed PFBC. (Reprinted) JAMA Neurology April 2015 Volume 72, Number 4
Copyright 2015 American Medical Association. All rights reserved.
Downloaded From: http://archneur.jamanetwork.com/ by a Penn State Milton S Hershey Med Ctr User on 05/22/2015
461
Clinical Review & Education Review
Primary Familial Brain Calcification
Figure 3. Mutations in the 3 Genes Linked to Primary Familial Brain Calcification (PFBC) c.935-1G>A21 c.1086delC27 c.1101C>G15 c.1145G>A15 c.1399C>T5 c.1409delC10
c.82G>A17,21 c.124_126delGTG10
c.1470_1478delGCAGGTCCT21 c.1483G>A8
c.152C>T5 c.551C>T11,17
c.185T>C21 c.212G>A5
c.583_584delGT15
c.260_261delTC5
2
3
4
5
c.1802C>G/T10,15
c.1492G>A10 c.1506C>A15
c.581A>G17
6
c.1828_1831delTCCC15
7
8
9
10
11
c.1784C>T10
c.760C>T15
c.514A>T15 c.515delA20
c.1909A>C5
c.1523+1G>A15
c.1723G>A10,19
c.509delT15
c.1711G>A17
SLC20A2
c.431-1G>T11,17
c.1794+1G>A/C15
c.344C>T5
c.1652G>A15 c.1703C>T15
c.338C>G23 c.323T>C22
c.1527delT17 c.356T>C6 c.433C>T6 c.439C>T28
c.3G>A6
1
c.726G>C6
c.445C>T6
c.26T>G6
2
3
4
5
6
PDGFB
c.1973T>C11,17 c.2083C>T29
2
3
14
15
c.2959C>T11,17 c.3212A>T17
22
23
PDGFRB
A subset of mutation carriers is described as being clinically unaffected. This observation may reflect true reduced penetrance but also may be in keeping with incomplete reporting and/or unrecognized clinical symptoms and signs in carriers of mutations in PFBCrelated genes. Supporting the latter notion, the recent example of mutations in the parkin gene causing early-onset Parkinson disease has taught us that systematic clinical information is often a bottleneck despite the description of large numbers of mutation carriers for more than 15 years since its discovery.14 Because genetic causes of PFBC have been reported only recently, a systematic review shortly after the description of causative genes thus offers a unique opportunity to (1) summarize available phenotypic data; (2) assess the availability of the data, analyze whether the quality is sufficient, and suggest improvements where needed; and (3) open the discussion as to how to define clearly all important variables by 462
The diagrams depict the mutations reported in the articles taken for review and by Nicolas et al,26 which are analyzed in the Discussion section. We also included c.1101C>G (discussed by Hsu et al15) in the review, although it is a synonymous change, thus depicted in this figure. The deletion of the entire SLC20A2 gene is not depicted (discussed by Baker et al21) but is included in the review. Numbers in boxes represent exons.
proposing a minimal data set for the comprehensive and systematic characterization of PFBC.
Methods We performed a systematic literature review according to the PRISMA guidelines (Preferred Reporting Items for Systematic Reviews and Meta-analyses). The search in MEDLINE included all publications in English from January 1, 2012, through May 31, 2014, and applied the following search terms: PFBC, IBGC, brain calcification, Fahr disease, OR calcification AND SLC20A2, PDGFRB, OR PDGFB. All cases with a positive genetic test result and individual clinical information were summarized. Because information was often incomplete, we first assessed the availabil-
JAMA Neurology April 2015 Volume 72, Number 4 (Reprinted)
Copyright 2015 American Medical Association. All rights reserved.
Downloaded From: http://archneur.jamanetwork.com/ by a Penn State Milton S Hershey Med Ctr User on 05/22/2015
jamaneurology.com
Primary Familial Brain Calcification
Review Clinical Review & Education
Table 1. Demographic and Clinical Characteristics of PFBC Mutation Groups PFBC Mutationa Characteristic
SLC20A2b
PDGFRB
PDGFB
No. (%) of cases
119 (66.5)
AAE, mean (SD) [% with information available]
45.6 (21.7) [50.4]
44.8 (18.8) [100]
33.6 (18.0) [100]
AAO, mean (SD) [% with information available]
31.7 (26.0) [24.4]
16.3 (6.7) [26.6]
22.9 (15.7) [53.1]
AAD, mean (SD) [% with information available]
40.5 (37.7) [3.4]
NA
NA
Sex, No. (% with information available)
16 (8.9)
32 (17.9)
73.1
93.3
100
Male
46 (38.7)
8 (53.3)
17 (53.1)
Female
41 (34.5)
6 (40.0)
15 (47.0)
115 (96.6)
15 (100)
28 (87.5)
White
49 (41.2)
15 (100)
25 (78.1)
Asian
65 (54.6)
Ethnicity, No. (% with information available)
Hispanic Other Clinical status, No. (% with information available)
0 1 (0.8)
0
0
0
3 (9.4)
0
0
104 (87.4)
15 (100)
30 (93.8)
Symptomatic
71 (59.7)
7 (46.7)
23 (71.9)
Asymptomatic
33 (27.7)
8 (53.3)
7 (21.9)
22 (18.5)
0
7 (221.9)
6 (5.0)
0
0
16 (13.4)
2 (13.3)
3 (9.4)
Cerebellar symptoms
7 (5.9)
2 (13.3)
4 (12.5)
Epilepsy
3 (2.5)
0
0
Clinical symptoms Dystonia PKD (dystonia subcategory) Parkinsonism
Headache or migraine Fatigue
16 (13.4) 6 (5.0)
3 (20.0)
14 (43.8)
0
5 (15.6)
Other neurologic symptoms
10 (8.4)
1 (6.7)
5 (15.6)
Dementia
18 (15.1)
0
5 (15.6)
6 (5.0)
1 (6.7)
3 (9.4)
13 (10.9)
3 (20.0)
9 (28.1)
Depression Other psychiatric disorders
ity of information to estimate possible bias (Figure 2 and eAppendix in the Supplement). In the literature on IBGC or PFBC, the term affected is usually defined as positive findings on a CT scan even if no clinical symptoms or signs were detected.7,8,15 To date, no clear nomenclature exists for individuals with positive findings on CT scans and neurologic or psychiatric symptoms. A candidate term is symptomatic; however, this term is in conflict with established definitions of affected and symptomatic in clinical ratings, whereby individuals with signs found during the examination are classified as affected and those who are aware of their clinical signs are classified as symptomatic.16 For clarity, we restricted the terms to affected, preceded by the aspect of interest, that is, genetic, clinical, or imaging. Information on CT findings was adopted from the articles without our own assessment. Moreover, some publications highlight the importance of discriminating primary and secondary calcifications, for example, in the context of hyperparathyroidism and via specific laboratory tests, but this information is available only in a subset of studies. Our statistical analysis is restricted to descriptive and basic univariate comparisons performed with χ2 tests for categorical variables and MannWhitney tests for continuous variables. jamaneurology.com
Abbreviations: AAD, age at diagnosis; AAE, age at examination; AAO, age at onset; NA, not available; PFBC, primary familial brain calcification; PKD, paroxysmal kinesigenic dyskinesia. a
Unless otherwise indicated, data are expressed as number (percentage) of cases.
b
The large family with PFBC described by Baker et al21 carried an SLC20A2 mutation and a THAP1 mutation. Whether the phenotypic expression in the family is a result of the SLC20A2 or of the THAP1 mutation or a combination thereof cannot be determined; therefore, we did not include this family in our statistical analysis.
Results We identified 15 reports5,6,8,10,11,15,17-25 meeting our inclusion criteria. These reports provided information on a total of 179 cases for analysis, 162 of which belonged to 25 families (Figure 2). The mutations described in each report are depicted in Figure 3. Most cases carried a mutation in SLC20A2 (131 of approximately 437 cases undergoing screening), 12 of which had an additional mutation in the THAP1 gene (all from the same family). Mutations in the more recently identified PDGFRB gene (15 of approximately 47 cases undergoing screening) and PDGFB gene (33 of approximately 90 cases undergoing screening) were not only less frequently screened for their existence but also seemed to be less common (Table 1). Available information was incomplete for all items apart from genetic status (inclusion criterion, genetic phenotype) and varied widely, ranging from 96.6% for ethnicity to 24.4% for age at onset (AAO) (Figure 4A). Most of the studies reexamined patients for the newly identified mutations from cases and study samples reported previously.27-29 Study designs included family studies, clinical samples, and cohort studies. (Reprinted) JAMA Neurology April 2015 Volume 72, Number 4
Copyright 2015 American Medical Association. All rights reserved.
Downloaded From: http://archneur.jamanetwork.com/ by a Penn State Milton S Hershey Med Ctr User on 05/22/2015
463
Clinical Review & Education Review
Primary Familial Brain Calcification
Figure 4. Several Issues of Known Gene Mutation in Primary Familial Brain Calcification (PFBC) A Availability of information
100 90 80
B
Affected regions in CCT
Age at
Basal Ganglia
Yes, but no severity code Mild Moderate Severe
Cerebellum
Yes, but no severity code Mild Moderate Severe
Examination Onset Diagnosis
Cases, %
70 60 50
PDGFRB
Region
40
PDGFB
Thalamus
Yes, but no severity code Mild Moderate Severe
Subcortical White Matter
Yes, but no severity code Mild Moderate Severe
30 20 10 0 PDGFRB
SLC20A2
PDGFB
Gene C
SLC20A2
Clinical status according to mutation
Affected cases
0
10
20
30
40
50
60
70
80
90 100
Cases, %
Unaffected cases
80 D Frequency of symptoms
70
SLC20A2
60
45
50
40
PDGFRB PDGFB
35 40
Cases, %
Cases, %
50
30 20
30 25 20 15 10
10
5
0
0 SLC20A2
PDGFRB
Gene
PDGFB
Dystonia Parkinsonism Cerebellar Headache symptoms and migraine
Fatigue
Other Dementia Depression Other neurologic psychiatric symptoms disorders
Clinical Symptom
A, Availability of information for ages at examination, onset, and diagnosis. B, Affected regions in cranial computed tomography (CCT). C, Clinical status according to gene mutation in PFBC. D, Frequency of clinical symptoms in PFBC.
All cases included in this review had calcifications on CCT examination, although the CCT results were not always explicitly stated in the report. The calcifications were most frequently located in the basal ganglia (70.6%), subcortical white matter (40.8%), cerebellum (34.1%), and thalamus (28.5%) (Table 2 and Figure 4B). In most cases, the basal ganglia were affected by calcifications, and other anatomic regions were affected in addition to the basal ganglia. The only cases with calcifications exclusively outside the basal ganglia are from 1 family, 1 case carrying an SLC20A2 mutation, and 1 case with a PDGFB mutation.6,15,29 Of all cases with positive CCT findings (imaging phenotype), 64.1% were clinically affected (68.9% of the women and 60.9% of the men). Furthermore, no obvious association between the site of calcification on CCT and the type of clinical sign was observed; for example, cerebellar symptoms co-occurred with basal ganglia calcifications more frequently than with cerebellar calcifications. Mean (SD) AAO was 27.9 (22.3) years (33.6 [26.0] years for women and 22.2 [16.0] years for men; Mann-Whitney P = .22). Mean age at examination was 42.6 (21.1) years; AAO was compa464
rable across genes (P = .77). The phenotypes were variable, ranging from asymptomatic to severely affected cases (Figure 4C), for example, severe dementia or highly disabling movement disorders. The most frequently described signs were movement disorders that were predominantly extrapyramidal, such as parkinsonism (11.8% overall) and dystonia (19.1%). Other described symptoms included paroxysmal kinesigenic dyskinesia with known genes excluded (6 cases [3.4%]) and cerebellar symptoms (7.3%). In cases with information on AAO, we observed a tendency toward more frequent dystonic symptoms in younger cases (10 of 34 [29%] in those with AAO
Review
Primary Familial Brain Calcification With Known Gene Mutations A Systematic Review and Challenges of Phenotypic Characterization Vera Tadic, MD; Ana Westenberger, PhD; Aloysius Domingo, MD; Daniel Alvarez-Fischer, MD; Christine Klein, MD; Meike Kasten, MD
IMPORTANCE In the past 2 years, 3 genes (SLC20A2, PDGFRB, and PDGFB) were identified as causative of primary familial brain calcification (PFBC), enabling genotype-specific phenotyping.
Supplemental content at jamaneurology.com
OBJECTIVES To provide a systematic literature review on the neuroimaging and clinical phenotype of genetically confirmed PFBC and summarize known pathophysiological mechanisms, to improve and harmonize future phenotype description and reporting by addressing data gaps, and to develop uniform definitions for clinical characterization. EVIDENCE REVIEW We systematically searched the MEDLINE database among articles published from January 1, 2012, through May 31, 2014, for the 3 genes and selected 25 articles from all records (n = 75) and from sources cited in the reference lists. Only genetically confirmed cases with individual clinical information were included, leaving 15 reports. Predefined categories for data extraction were different neurologic and psychiatric symptoms, imaging results, and age at onset (AAO). We also assessed availability of information to estimate possible bias. FINDINGS We included a total of 179 cases, 162 of which belong to 25 families. Availability of information ranged from 96.6% for ethnicity to 24.4% for AAO. All cases had calcifications on comprehensive cranial computed tomography, most frequently located in the basal ganglia (70.6%), subcortical white matter (40.8%), cerebellum (34.1%), or thalamus (28.5%). Mean (SD) AAO was 27.9 (22.3) years, and the AAO was comparable across genes (P = .77). The most frequently described signs were movement disorders, such as parkinsonism (12%) and dystonia (19%). Penetrance of the imaging phenotype was 100% compared with only 61% of the clinical phenotype. We propose a novel definition of disease status by specifying PFBC into genetic, clinical, and imaging phenotypes. Pathophysiological pathways converge on impaired phosphorus homeostasis and integrity of the blood-brain barrier. CONCLUSIONS AND RELEVANCE Especially in rare conditions, meta-analyses are the most suitable tool to extract reliable information on the natural course of a disease. For future analyses, we provide a minimal data set that can be used for systematic clinical and imaging data collection in PFBC and that will also improve informed counseling of patients.
Author Affiliations: Institute of Neurogenetics, University of Lübeck, Lübeck, Germany (Tadic, Westenberger, Domingo, AlvarezFischer, Klein, Kasten); Graduate School for Computing in Medicine and Life Science, University of Lübeck, Lübeck, Germany (Domingo); Department of Psychiatry, University of Lübeck, Lübeck, Germany (Alvarez-Fischer, Kasten). Corresponding Author: Christine Klein, MD, Institute of Neurogenetics, University of Lübeck, Ratzeburger Allee 160, 23538 Lübeck, Germany (christine.klein @neuro.uni-luebeck.de).
JAMA Neurol. 2015;72(4):460-467. doi:10.1001/jamaneurol.2014.3889 Published online February 16, 2015. 460
Section Editor: David E. Pleasure, MD. (Reprinted) jamaneurology.com
Copyright 2015 American Medical Association. All rights reserved.
Downloaded From: http://archneur.jamanetwork.com/ by a Penn State Milton S Hershey Med Ctr User on 05/22/2015
Primary Familial Brain Calcification
F
Review Clinical Review & Education
irst reports on intracerebral calcifications date back to the 19th century with an autopsy report from Delacour1 of an individual with bilateral cerebral calcifications most prominent in the striatum. However, the best-known case description is by Fahr2 in 1930, although the calcifications were localized in the white rather than the gray matter and, in retrospect, were likely secondary.3 Nonspecific calcifications in the brain are a common incidental and age-dependent finding on neuroimaging, with an estimated frequency of 1% to 20%.4-6 For brain calcification of unknown origin, the term idiopathic basal ganglia calcification (IBGC) was coined. However, with the discovery of the genetic causes of brain calcifications following an autosomal dominant trait, this disorder ceased to be idiopathic. Thus, the term IBGC is no longer appropriate for the disorder in patients with genetically confirmed disease; a recent report has suggested replacing IBGC with the term primary familial brain calcification (PFBC).7 The neuroimaging phenotype consists of bilateral calcifications in the brain, predominantly in the basal ganglia but also in the cerebellum, thalamus, and brainstem. The clinical phenotype encompasses neurologic and psychiatric features. The neurologic signs are variable in type and severity; those most commonly described are movement disorders, but seizures, headache, and cerebellar symptoms may occur. Psychiatric features are also variable and include cognitive impairment, mood disorders, psychotic symptoms, and obsessive-compulsive symptoms, although fewer data are available compared with neurologic signs.8 Although the exact pathophysiological features remain elusive, recent discoveries offer important clues toward the cause and disease mechanisms of PFBC.9 In particular, mutations in 3 different genes have been identified as causative of PFBC. First, SLC20A2 (OMIM 158378) was described in February 201210 as coding for type III sodium-dependent inorganic phosphate (P i ) transporter 2 (PiT-2). The second reported gene, PDGFRB (OMIM 173410), was described in January 201311 as encoding 1 of 2 receptors for plateletderived growth factor. The gene encoding its major ligand, PDGFB (OMIM 190040), was the third gene identified in September 2013.6 The different mutations in all 3 genes share loss of function as the likely disease mechanism.12 Mutations in the SLC20A2 gene lead to an accumulation of Pi and subsequently to calcium phosphate deposition. Mutations in PDGFB and PDGFRB also result in calcification, however, through indirect processes (Figure 1). Although the receptor and its ligand are also expressed in neurons, data from selective knockout studies indicate impaired recruitment of pericytes to endothelial cells, resulting in a dysfunctional blood-brain barrier and thereby contributing to brain calcification.6 A recently published abstract13 suggested involvement of a fourth gene in the pathophysiological features of PFBC by reporting a mutation in XPR1 as the cause of disease in 1 family with PFBC and no mutations in any of the known PFBC genes. In keeping with previously proposed disease mechanisms, XPR1 appears to alter Pi homeostasis. However, this finding awaits peer-reviewed publication and independent replication. The gene discoveries were followed by a number of mutational screening studies and clinical reports of individual cases or families with confirmed genetic causes (ie, definitive PFBC), thus for the first time enabling a summary of available phenotypic information based on accurate genetic classification. This step calls for a sysjamaneurology.com
Figure 1. Proposed Disease Mechanism of Brain Calcification A
B
PDGFRB
PDGFB
PiT-2
Brain cell
Pi XPR1
Calcium phosphate deposits
A, Under physiological conditions, endothelial cells secrete platelet-derived growth factor B (PDGFB) that binds to PDGF receptor β (PDGFRB) on pericytes. This process is necessary for the recruitment of pericytes and correct proliferation and migration along the blood vessels. Inorganic phosphate (Pi) in the brain parenchyma is taken up by Pi transporter 2 (PiT-2). From the brain cell, Pi is secreted through XPR1. B, Under pathologic conditions, the PDGF ligand or receptor loses its normal function, leading to pericyte dysfunction and an altered blood-brain barrier. Pi accumulates in the brain parenchyma owing to its reduced uptake by PiT-2. The proposed disease mechanism of mutated XPR1 has not yet been published in detail. However, the mechanism was found to reduce efflux of phosphate from cells.13 Whether glial and/or neuronal cells participate in the diseased state of primary familial brain calcification remains unknown.
Figure 2. Flowchart of the Literature Review Literature search in PubMed Search terms: PFBC, IBGC, brain calcification, Fahr disease, OR calcification AND SLC20A2, PDGFRB, OR PDGFB
75 Records identified through database searching from January 1, 2012, to May 31, 2014
20 Records remained after removing double counts, articles without clinical information, and false-positive findings
5 Records included by reviewing reference lists and sources
25 Records included in qualitative synthesis
15 Records included in quantitative synthesis (meta-analysis) 179 Cases (90.5% familial and 4.5% sporadic cases)a
We identified 179 cases of primary familial brain calcification (PFBC), 162 of whom belonged to 25 families. IBGC indicates idiopathic basal ganglia calcification; PFBC, primary familial brain calcification. a
Information was unavailable for the remaining 5.0% of cases.
tematic literature review on the findings of comprehensive cranial computed tomography (CCT) and the neurologic and psychiatric phenotype of genetically confirmed PFBC. (Reprinted) JAMA Neurology April 2015 Volume 72, Number 4
Copyright 2015 American Medical Association. All rights reserved.
Downloaded From: http://archneur.jamanetwork.com/ by a Penn State Milton S Hershey Med Ctr User on 05/22/2015
461
Clinical Review & Education Review
Primary Familial Brain Calcification
Figure 3. Mutations in the 3 Genes Linked to Primary Familial Brain Calcification (PFBC) c.935-1G>A21 c.1086delC27 c.1101C>G15 c.1145G>A15 c.1399C>T5 c.1409delC10
c.82G>A17,21 c.124_126delGTG10
c.1470_1478delGCAGGTCCT21 c.1483G>A8
c.152C>T5 c.551C>T11,17
c.185T>C21 c.212G>A5
c.583_584delGT15
c.260_261delTC5
2
3
4
5
c.1802C>G/T10,15
c.1492G>A10 c.1506C>A15
c.581A>G17
6
c.1828_1831delTCCC15
7
8
9
10
11
c.1784C>T10
c.760C>T15
c.514A>T15 c.515delA20
c.1909A>C5
c.1523+1G>A15
c.1723G>A10,19
c.509delT15
c.1711G>A17
SLC20A2
c.431-1G>T11,17
c.1794+1G>A/C15
c.344C>T5
c.1652G>A15 c.1703C>T15
c.338C>G23 c.323T>C22
c.1527delT17 c.356T>C6 c.433C>T6 c.439C>T28
c.3G>A6
1
c.726G>C6
c.445C>T6
c.26T>G6
2
3
4
5
6
PDGFB
c.1973T>C11,17 c.2083C>T29
2
3
14
15
c.2959C>T11,17 c.3212A>T17
22
23
PDGFRB
A subset of mutation carriers is described as being clinically unaffected. This observation may reflect true reduced penetrance but also may be in keeping with incomplete reporting and/or unrecognized clinical symptoms and signs in carriers of mutations in PFBCrelated genes. Supporting the latter notion, the recent example of mutations in the parkin gene causing early-onset Parkinson disease has taught us that systematic clinical information is often a bottleneck despite the description of large numbers of mutation carriers for more than 15 years since its discovery.14 Because genetic causes of PFBC have been reported only recently, a systematic review shortly after the description of causative genes thus offers a unique opportunity to (1) summarize available phenotypic data; (2) assess the availability of the data, analyze whether the quality is sufficient, and suggest improvements where needed; and (3) open the discussion as to how to define clearly all important variables by 462
The diagrams depict the mutations reported in the articles taken for review and by Nicolas et al,26 which are analyzed in the Discussion section. We also included c.1101C>G (discussed by Hsu et al15) in the review, although it is a synonymous change, thus depicted in this figure. The deletion of the entire SLC20A2 gene is not depicted (discussed by Baker et al21) but is included in the review. Numbers in boxes represent exons.
proposing a minimal data set for the comprehensive and systematic characterization of PFBC.
Methods We performed a systematic literature review according to the PRISMA guidelines (Preferred Reporting Items for Systematic Reviews and Meta-analyses). The search in MEDLINE included all publications in English from January 1, 2012, through May 31, 2014, and applied the following search terms: PFBC, IBGC, brain calcification, Fahr disease, OR calcification AND SLC20A2, PDGFRB, OR PDGFB. All cases with a positive genetic test result and individual clinical information were summarized. Because information was often incomplete, we first assessed the availabil-
JAMA Neurology April 2015 Volume 72, Number 4 (Reprinted)
Copyright 2015 American Medical Association. All rights reserved.
Downloaded From: http://archneur.jamanetwork.com/ by a Penn State Milton S Hershey Med Ctr User on 05/22/2015
jamaneurology.com
Primary Familial Brain Calcification
Review Clinical Review & Education
Table 1. Demographic and Clinical Characteristics of PFBC Mutation Groups PFBC Mutationa Characteristic
SLC20A2b
PDGFRB
PDGFB
No. (%) of cases
119 (66.5)
AAE, mean (SD) [% with information available]
45.6 (21.7) [50.4]
44.8 (18.8) [100]
33.6 (18.0) [100]
AAO, mean (SD) [% with information available]
31.7 (26.0) [24.4]
16.3 (6.7) [26.6]
22.9 (15.7) [53.1]
AAD, mean (SD) [% with information available]
40.5 (37.7) [3.4]
NA
NA
Sex, No. (% with information available)
16 (8.9)
32 (17.9)
73.1
93.3
100
Male
46 (38.7)
8 (53.3)
17 (53.1)
Female
41 (34.5)
6 (40.0)
15 (47.0)
115 (96.6)
15 (100)
28 (87.5)
White
49 (41.2)
15 (100)
25 (78.1)
Asian
65 (54.6)
Ethnicity, No. (% with information available)
Hispanic Other Clinical status, No. (% with information available)
0 1 (0.8)
0
0
0
3 (9.4)
0
0
104 (87.4)
15 (100)
30 (93.8)
Symptomatic
71 (59.7)
7 (46.7)
23 (71.9)
Asymptomatic
33 (27.7)
8 (53.3)
7 (21.9)
22 (18.5)
0
7 (221.9)
6 (5.0)
0
0
16 (13.4)
2 (13.3)
3 (9.4)
Cerebellar symptoms
7 (5.9)
2 (13.3)
4 (12.5)
Epilepsy
3 (2.5)
0
0
Clinical symptoms Dystonia PKD (dystonia subcategory) Parkinsonism
Headache or migraine Fatigue
16 (13.4) 6 (5.0)
3 (20.0)
14 (43.8)
0
5 (15.6)
Other neurologic symptoms
10 (8.4)
1 (6.7)
5 (15.6)
Dementia
18 (15.1)
0
5 (15.6)
6 (5.0)
1 (6.7)
3 (9.4)
13 (10.9)
3 (20.0)
9 (28.1)
Depression Other psychiatric disorders
ity of information to estimate possible bias (Figure 2 and eAppendix in the Supplement). In the literature on IBGC or PFBC, the term affected is usually defined as positive findings on a CT scan even if no clinical symptoms or signs were detected.7,8,15 To date, no clear nomenclature exists for individuals with positive findings on CT scans and neurologic or psychiatric symptoms. A candidate term is symptomatic; however, this term is in conflict with established definitions of affected and symptomatic in clinical ratings, whereby individuals with signs found during the examination are classified as affected and those who are aware of their clinical signs are classified as symptomatic.16 For clarity, we restricted the terms to affected, preceded by the aspect of interest, that is, genetic, clinical, or imaging. Information on CT findings was adopted from the articles without our own assessment. Moreover, some publications highlight the importance of discriminating primary and secondary calcifications, for example, in the context of hyperparathyroidism and via specific laboratory tests, but this information is available only in a subset of studies. Our statistical analysis is restricted to descriptive and basic univariate comparisons performed with χ2 tests for categorical variables and MannWhitney tests for continuous variables. jamaneurology.com
Abbreviations: AAD, age at diagnosis; AAE, age at examination; AAO, age at onset; NA, not available; PFBC, primary familial brain calcification; PKD, paroxysmal kinesigenic dyskinesia. a
Unless otherwise indicated, data are expressed as number (percentage) of cases.
b
The large family with PFBC described by Baker et al21 carried an SLC20A2 mutation and a THAP1 mutation. Whether the phenotypic expression in the family is a result of the SLC20A2 or of the THAP1 mutation or a combination thereof cannot be determined; therefore, we did not include this family in our statistical analysis.
Results We identified 15 reports5,6,8,10,11,15,17-25 meeting our inclusion criteria. These reports provided information on a total of 179 cases for analysis, 162 of which belonged to 25 families (Figure 2). The mutations described in each report are depicted in Figure 3. Most cases carried a mutation in SLC20A2 (131 of approximately 437 cases undergoing screening), 12 of which had an additional mutation in the THAP1 gene (all from the same family). Mutations in the more recently identified PDGFRB gene (15 of approximately 47 cases undergoing screening) and PDGFB gene (33 of approximately 90 cases undergoing screening) were not only less frequently screened for their existence but also seemed to be less common (Table 1). Available information was incomplete for all items apart from genetic status (inclusion criterion, genetic phenotype) and varied widely, ranging from 96.6% for ethnicity to 24.4% for age at onset (AAO) (Figure 4A). Most of the studies reexamined patients for the newly identified mutations from cases and study samples reported previously.27-29 Study designs included family studies, clinical samples, and cohort studies. (Reprinted) JAMA Neurology April 2015 Volume 72, Number 4
Copyright 2015 American Medical Association. All rights reserved.
Downloaded From: http://archneur.jamanetwork.com/ by a Penn State Milton S Hershey Med Ctr User on 05/22/2015
463
Clinical Review & Education Review
Primary Familial Brain Calcification
Figure 4. Several Issues of Known Gene Mutation in Primary Familial Brain Calcification (PFBC) A Availability of information
100 90 80
B
Affected regions in CCT
Age at
Basal Ganglia
Yes, but no severity code Mild Moderate Severe
Cerebellum
Yes, but no severity code Mild Moderate Severe
Examination Onset Diagnosis
Cases, %
70 60 50
PDGFRB
Region
40
PDGFB
Thalamus
Yes, but no severity code Mild Moderate Severe
Subcortical White Matter
Yes, but no severity code Mild Moderate Severe
30 20 10 0 PDGFRB
SLC20A2
PDGFB
Gene C
SLC20A2
Clinical status according to mutation
Affected cases
0
10
20
30
40
50
60
70
80
90 100
Cases, %
Unaffected cases
80 D Frequency of symptoms
70
SLC20A2
60
45
50
40
PDGFRB PDGFB
35 40
Cases, %
Cases, %
50
30 20
30 25 20 15 10
10
5
0
0 SLC20A2
PDGFRB
Gene
PDGFB
Dystonia Parkinsonism Cerebellar Headache symptoms and migraine
Fatigue
Other Dementia Depression Other neurologic psychiatric symptoms disorders
Clinical Symptom
A, Availability of information for ages at examination, onset, and diagnosis. B, Affected regions in cranial computed tomography (CCT). C, Clinical status according to gene mutation in PFBC. D, Frequency of clinical symptoms in PFBC.
All cases included in this review had calcifications on CCT examination, although the CCT results were not always explicitly stated in the report. The calcifications were most frequently located in the basal ganglia (70.6%), subcortical white matter (40.8%), cerebellum (34.1%), and thalamus (28.5%) (Table 2 and Figure 4B). In most cases, the basal ganglia were affected by calcifications, and other anatomic regions were affected in addition to the basal ganglia. The only cases with calcifications exclusively outside the basal ganglia are from 1 family, 1 case carrying an SLC20A2 mutation, and 1 case with a PDGFB mutation.6,15,29 Of all cases with positive CCT findings (imaging phenotype), 64.1% were clinically affected (68.9% of the women and 60.9% of the men). Furthermore, no obvious association between the site of calcification on CCT and the type of clinical sign was observed; for example, cerebellar symptoms co-occurred with basal ganglia calcifications more frequently than with cerebellar calcifications. Mean (SD) AAO was 27.9 (22.3) years (33.6 [26.0] years for women and 22.2 [16.0] years for men; Mann-Whitney P = .22). Mean age at examination was 42.6 (21.1) years; AAO was compa464
rable across genes (P = .77). The phenotypes were variable, ranging from asymptomatic to severely affected cases (Figure 4C), for example, severe dementia or highly disabling movement disorders. The most frequently described signs were movement disorders that were predominantly extrapyramidal, such as parkinsonism (11.8% overall) and dystonia (19.1%). Other described symptoms included paroxysmal kinesigenic dyskinesia with known genes excluded (6 cases [3.4%]) and cerebellar symptoms (7.3%). In cases with information on AAO, we observed a tendency toward more frequent dystonic symptoms in younger cases (10 of 34 [29%] in those with AAO
