Progranulin genetic variability contributes to amyotrophic lateral sclerosis

Progranulin genetic variability contributes to amyotrophic lateral sclerosis K. Sleegers, MD, PhD N. Brouwers, MSc S. Maurer-Stroh, PhD M.A. van Es, MD P. Van Damme, MD, PhD P.W.J. van Vught, MSc J. van der Zee, MSc S. Serneels T. De Pooter M. Van den Broeck M. Cruts, PhD J. Schymkowitz, PhD P. De Jonghe, MD, PhD F. Rousseau, PhD L.H. van den Berg, MD, PhD W. Robberecht, MD, PhD C. Van Broeckhoven, PhD, DSc Address correspondence and reprint requests to Prof. Dr. Christine Van Broeckhoven, Neurodegenerative Brain Diseases Group, VIB–Department of Molecular Genetics, University of Antwerp, Universiteitsplein 1, BE-2610 Antwerpen, Belgium christine.vanbroeckhoven@ ua.ac.be

Supplemental data at www.neurology.org Editorial, page XXX

ABSTRACT

Objectives: Null mutations in progranulin (PGRN) cause ubiquitin-positive frontotemporal dementia (FTD) linked to chromosome 17q21 (FTDU-17). Here we examined PGRN genetic variability in amyotrophic lateral sclerosis (ALS), a neurodegenerative motor neuron disease that overlaps with FTD at a clinical, pathologic, and epidemiologic level.

Methods: We sequenced all exons, exon-intron boundaries, and 5= and 3= regulatory regions of PGRN in a Belgian sample of 230 patients with ALS. The frequency of observed genetic variants was determined in 436 healthy control individuals. The contribution of eight frequent polymorphisms to ALS risk, onset age, and survival was assessed in an association study in the Belgian sample and a replication series of 308 Dutch patients with ALS and 345 Dutch controls.

Results: In patients with ALS we identified 11 mutations, 5 of which were predicted to affect PGRN protein sequence or levels (four missense mutations and one 5= regulatory variant). Moreover, common variants (rs9897526, rs34424835, and rs850713) and haplotypes were significantly associated with a reduction in age at onset and a shorter survival after onset of ALS in both the Belgian and the Dutch studies.

Conclusion: PGRN acts as a modifier of the course of disease in patients with amyotrophic lateral sclerosis, through earlier onset and shorter survival. Neurology® • • • GLOSSARY ALS ⫽ amyotrophic lateral sclerosis; FTD ⫽ frontotemporal dementia; FTDU ⫽ ubiquitin-positive subtype of frontotemporal dementia; HWE ⫽ Hardy-Weinberg equilibrium; SNPs ⫽ single nucleotide polymorphisms; Ub-ir ⫽ ubiquitin immunoreactive.

Amyotrophic lateral sclerosis (ALS) is a fatal neurodegenerative disease, which is neuropathologically characterized by ubiquitin immunoreactive (Ub-ir) inclusions in the perikaryon and proximal axon of surviving motor neurons in the anterior horns of the spinal cord and brainstem. Whereas in a small percentage of patients, ALS concurs with dementia, in up to 50% of patients disturbances of executive functions and behavioral changes can be demonstrated.1,2 Autopsy of patients with behavioral and executive abnormalities has shown ubiquitinated inclusions in hippocampus and neocortical neurons as well.3,4 This pathology is identical to the ubiquitin-positive subtype of frontotemporal dementia (FTDU), a neurodegenerative dementia characterized by behavioral changes. In FTDU, ubiquitin-immunoreactive inclusions can be observed in cytoplasm and nucleus

e-Pub ahead of print at www.neurology.org From the Neurodegenerative Brain Diseases Group (K.S., N.B., J.v.d.Z., S.S., T.D.P., M.V.d.B., M.C., C.V.B.) and Neurogenetics Group (P.D.J.), Department of Molecular Genetics, and SWITCH Laboratory (S.M.-S., J.S., F.R.) VIB, University of Antwerp (K.S., N.B., J.v.d.Z., S.S., T.D.P., M.V.d.B., M.C., P.D.J., C.V.B.), Antwerpen; University Brussels (VUB) (S.M.-S., J.S., F.R.); Department of Neurology (M.A.v.E., P.W.J.v.V., L.H.v.d.B.), Rudolf Magnus Institute of Neuroscience, University Medical Center Utrecht, the Netherlands; Division of Neurology (P.V.D., W.R.), University Hospital Gasthuisberg, University of Leuven (KUL); and Division of Neurology (P.D.J.), University Hospital Antwerpen, Belgium. Supported by the Zenith Award of the Alzheimer Association USA, Stichting Alzheimer Onderzoek-Belgium, the EU contract LSHMCT2003-503330 (APOPIS), the InterUniversity Attraction Poles (IAP) program P5/19 and P6/43 of the Belgian Federal Science Policy office (BELSPO), the Fund for Scientific Research-Flanders (FWO-F), and the Special Research Fund (BOF) of the University of Antwerp. K.S., P.V.D., and M.C. hold a postdoctoral fellowship and N.B. a doctoral fellowship of FWO-F, and J.v.d.Z. a doctoral fellowship of the Institute for Science and Technology-Flanders (IWT-F). S.M.-S. is recipient of a Marie Curie Intra-European Fellowship. L.H.v.d.B. is supported by a grant from the Netherlands Organization for Health Research and Development. W.R. is supported by the Research Council of the University of Leuven, and through the E. von Behring Chair for Neuromuscular and Neurodegenerative Disorders. Disclosure: The authors report no conflicts of interest. Copyright © 2008 by AAN Enterprises, Inc.

1

of neurons in layer II of the frontal and temporal neocortex and fascia dentata of the hippocampus.5 The observation that TDP-43 is a major component of these inclusions in both ALS and FTDU6,7 further strengthens the hypothesis that both disorders are part of a clinicopathologic spectrum. But ALS and FTD do not only show pathologic similarities. Epidemiologic studies have shown an increased risk of dementia in first- and second-degree relatives of patients with ALS,8 and families are known segregating both FTD and ALS.9 Whereas in the latter the presence of ubiquitin pathology was not investigated, together these data suggest that ALS and FTDU are part of a continuous spectrum of ubiquitin-associated neurodegenerative diseases.10 Recently, null mutations in the gene encoding the growth factor progranulin (PGRN) were shown to cause FTDU linked to 17q21 (FTDU-17; a phenotype that may include parkinsonism or signs of motor neuron disease) through a partial loss of transcript or protein.11,12 PGRN encodes a secreted precursor protein composed of a signal peptide and 7.5 tandem repeats of highly conserved motifs containing 12 cysteines that can be proteolytically cleaved to form a family of granulin peptides.13,14 Progranulin is a widely expressed growth factor with a range of biologic functions.15,16 How a 50% loss of PGRN causes FTDU-17 remains to be elucidated, Table

Description of Belgian and Dutch study samples Belgian ALS patients (n ⫽ 230)

Dutch ALS patients (n ⫽ 308)

Age at onset, y, mean ⫾ SD

57.6 ⫾ 12.3

57.9 ⫾ 11.6

Gender (% men)

63

60.3

Survival after first symptoms*

28 (6–111)

27.6 (1.5–273.6)

Site of onset (% spinal)†

67.3

80.8

Belgian controls (n ⫽ 436)

Dutch controls (n ⫽ 345)

Age at inclusion, y, mean ⫾ SD

58.7 ⫾ 15.8

60 ⫾ 11.7

Gender (% men)

44

54.8

*Median (range) survival time until death in months, based on complete survival time available for 183 Belgian and 130 Dutch patients. †Data available for 219 Belgian and 300 Dutch patients. 2

Neurology •••

but haploinsufficiency suggests a role of PGRN in neuronal survival. On the other hand, PGRN has anti-inflammatory properties while granulin peptides have proinflammatory properties, and the upregulation of PGRN in activated microglia suggests PGRN might be involved in neurodegeneration through neuroinflammation.17 Several lines of evidence have indicated that PGRN plays a role in the pathomechanism of ALS as well. In lumbar spinal cord of patients with ALS, e.g., a marked upregulation of PGRN transcripts has been observed,18 and PGRN has been shown to stimulate expression of VEGF, a growth factor for motor neurons,19 suggesting that genetic variation in PGRN might be underlying the spectrum of ALS and FTDU. Here we report a systematic mutation screening and association analysis of PGRN in patients with ALS and control individuals, aiming to examine the contribution of genetic variability in PGRN to ALS pathogenesis. MATERIALS AND METHODS Study groups. Two sets of patients and control individuals were included in this study; descriptives are shown in the table. A total of 230 sporadic Belgian patients with ALS were recruited at the university hospitals of Leuven and Antwerpen, Belgium, with a diagnosis of definite, probable, or laboratory supported probable ALS (according to El Escorial criteria, available at http://wfnals.org). None of the patients carried a SOD1 mutation. The control group consisted of 436 community control persons who were neurologically healthy and of Belgian descent. Informed consent was obtained from all individuals. The presence of population substructure was excluded based on 27 randomly selected microsatellite markers in Structure 2.1.20 The medical ethical committees of the Universities of Antwerp and Leuven granted approval for the study. Dutch patients with ALS (n ⫽ 308) were recruited through the University Medical Center Utrecht, The Netherlands, and diagnosed according to El Escorial criteria. The control sample consisted of 345 healthy individuals of Dutch descent. The institutional ethical committee of the University Medical Center Utrecht approved the study.

PGRN sequencing. A mutation analysis was performed of PGRN non-coding exon 0, coding exons 1 to 12, and a conserved 5= regulatory region in intron 0 as previously described.12 Twenty ng of gDNA were PCR amplified and amplification products were sequenced on an Applied Biosystems 3730xl DNA Analyzer using the Big Dye Terminator v3.1 Cycle Sequencing kit (Applied Biosystems, Foster City, CA). PGRN SNP genotyping. All common single nucleotide polymorphisms (SNPs; minor allele frequency ⬎5% [table

Figure 1

PGRN rare genetic variants observed in Belgian patients with amyotrophic lateral sclerosis

Schematic representation of PGRN variants observed in patients only (upper half) and variants observed in patients and control individuals (lower half). Coding mutations are presented with their cDNA position numbered according to the largest PGRN transcript (GenBank Accession Number NM_002087.2) and starting at translation initiation site ⫹ 1. Mutations in the 5= regulatory region and intronic variants are given by their gDNA position relative to the reverse complement of GenBank Accession Number AC003043, starting at nt 1. Missense mutations are indicated in bold, 5= regulatory variants in italics.

E-5 on the Neurology® Web site at www.neurology.org]) except IVS4 ⫹ 24G⬎A (r2 ⫽ 0.989 with IVS3-47-46insGTCA) were genotyped in the Dutch replication set by Sequenom MassARRAY genotyping according to the manufacturer’s protocol (Sequenom). Spectrum analysis and genotype scoring was performed using Typer 3.3 software (Sequenom).

Allele sharing analysis. In two patients with ALS and one patient with Alzheimer dementia21 carrying the p.Ala324Thr mutation we genotyped 14 microsatellite markers spanning an 8 cM region around PGRN as previously described.22

In silico prediction of pathogenicity. The effect of missense mutations on protein function was estimated in silico using the Sorting Intolerant From Tolerant (SIFT v.2) program,23 using different input methods and outcome thresholds as described.24 For structural modeling of mutations located in granulin domains, the full structure of a granulin domain was reconstructed.24 The effect of mutations on stability of granulin domains was evaluated using FoldX, introducing a penalty for forming or breaking disulfide bonds.25 MatInspector analysis26 was performed to estimate the effect of 5= regulatory variants on putative transcription factor binding sites. A core similarity cut off of 1 was used.

Statistical analysis. Deviations from Hardy-Weinberg equilibrium (HWE) were excluded using the HWE program for both Belgian and Dutch samples.27 For all SNPs with a minor allele frequency ⬎5%, ␹2 statistics and logistic regression analysis adjusted for age and gender were performed in SPPS 12.0, to examine the contribution of each detected common variant to the susceptibility to ALS. To study genotype-phenotype correlations, additional analyses were performed for patients with a spinal or bulbar onset of ALS. The effect of the polymorphisms on age at onset in patients was assessed in a univariate analysis of variance, adjusted for gender and site of onset. The effect of the polymorphisms on survival after onset of the first symptoms was tested with a Cox proportional hazards model. Hazard ratios (HR) were calculated with their 95% CIs, adjusted for gender, site of,

Figure 2

PGRN missense mutations relative to protein sequence

Schematic localization of missense mutations in the PGRN protein relative to the granulin domains.

and age at first symptoms. Survival times of patients who were still alive at last follow-up (47 Belgian patients and 178 Dutch patients) were right-censored. Pairwise LD, computed in Haploview,28 revealed two blocks of increased LD, one spanning the 5= regulatory region, the second covering intron 2 through to the 3= untranslated region. Haplotype frequencies were estimated in these LD blocks using a progressive EM insertion algorithm computing both maximum likelihood estimates of haplotype probabilities and posterior probabilities of pairs of haplotypes for each subject. Haplotype associations, based on these haplotype probabilities, were investigated with score statistics. A sliding window analysis was performed with 2-SNP windows. Both haplotype estimation and analysis were performed in Haplo Stats version 1.2.2.29 Empirical p values were computed based on 1,000 random permutations of patient and control labels, to avoid false positive findings due to multiple testing.

We performed a systematic mutation analysis of PGRN through direct sequencing of all coding exons including exon-intron boundaries and of 5= and 3= regulatory regions in a series of 230 Belgian patients diagnosed with ALS and in 436 Belgian control individuals. We identified 17 rare variants in 29 patients (9 exonic and 4 intronic variants, and 3 in the 5= and 1 in the 3= regulatory regions; tables E-1, E-2, and E-3). Of these rare variants, 11 were absent from 872 control chromosomes (figure 1). RESULTS PGRN sequencing.

Missense mutations. The nine exonic sequence variants included four missense mutations that were absent from Belgian control individuals (table E-1). Of those, c.329G⬎A (Arg110Gln) and c.371T⬎C (Ile124Thr) were novel missense mutations, whereas c.970G⬎A (Ala324Thr) and c.1253G⬎A (Arg418Gln) have previously been reported although their pathogenic nature remained unclear (http://www.molgen.ua.ac.be/ FTDmutations). All four missense mutations were located in or at the border of a granulin domain (figure 2), and none of the wild-type residues were highly conserved, although Ile124Thr is located at a position that is moderately conserved between granulin domains, containing either an Ile or a Val residue (figure E-1). SIFT analysis predicted that Arg110Gln, AL324Thr, and Arg418Gln were unlikely to affect protein function (table E-1), but Ile124Thr was predicted to affect protein function in four out of six tests. Limiting the SIFT analysis of Ile124Thr to 13 sequences specifically validated for human PGRN revealed a SIFT score of 0.01, predicting an intolerant change. Structural modeling revealed an average effect of ⫺0.35 ⫾ 0.03 kcal/mol of Ile124Thr and 0.36 ⫾ 0.01 kcal/mol of Ala324Thr on the stability of the granulin domain. We previously detected Ala324Thr also in Neurology •••

3

one Belgian patient with Alzheimer dementia.21 Allele sharing revealed that both ALS patients share alleles of microsatellite markers flanking PGRN, spanning a shared region of maximum 2.8 Mb. One patient with ALS shared alleles at seven consecutive markers with the Alzheimer patient, spanning ⬍6 Mb (table E-4).

Figure 3

Age at onset and survival after disease onset by IVS2 ⫹ 21G⬎A genotype

Clinical phenotype. Of the four missense muta-

tions, three were identified in four women, and one (Ile124Thr) in a man. All women had a spinal onset of the disease, whereas the man had a bulbar onset. The onset ages varied between patients, ranging from 53 years (Ala324Thr) to 74 years (Arg110Gln). Likewise, symptom duration varied from 16 months (Ile124Thr) to ⬎88 months (Ala324Thr). Even between the two patients carrying Ala324Thr, onset age and disease duration varied widely. One patient showed first symptoms at age 62 years, and died after 28 months; the other patient had first symptoms at age 53 years, and was still alive after 88 months. Association analysis. Genotype data of eight fre-

quent SNPs (minor allele frequency ⬎5%) were extracted from the sequencing data obtained in both Belgian patients and Belgian control individuals for genetic association analyses (table E-5). Genotype frequencies did not differ between patients and control individuals for any of the individual SNPs, neither in crude nor in age or gender adjusted logistic regression analyses (table E-5). Haplotype based association analysis was performed in the two distinct blocks of PGRN with high LD. We observed no significant global or haplotype-specific differences between patients with ALS and control individuals in either LD block (global p values 0.34 [block 1] and 0.63 [block 2]; data not shown). Genotype–phenotype correlation. Given the clini-

cal heterogeneity of ALS, we evaluated the effect of PGRN SNPs on site of onset (bulbar or spinal), age at onset, and survival. We did not observe an association between a single SNP or haplotype and site of onset. However, age at onset was reduced in carriers of the rare allele of IVS2 ⫹ 21G⬎A (mean difference: 7.7 years [95% CI: 3.2 to 12.1 years]; p ⫽ 0.001) (figure 3A), and to a lesser extent in carriers of the rare allele of IVS347-46insGTCA and IVS4 ⫹ 24G⬎A (pairwise r2 ⫽ 0.989; mean difference: 4.1 years [95% CI: 0.9 to 7.4 years]; p ⫽ 0.013). In a haplotype based setting, a haplotype in the second LD block tagged by the frequent alleles at IVS2 ⫹ 21G⬎A, IVS3-47-46insGTCA, and IVS4 ⫹ 24G⬎A showed an association with a later age at onset 4

Neurology •••

(A) Age at onset by IVS2 ⫹ 21G⬎A genotype. Mean age at onset is given with 95% CI for Belgian patients homozygous for the G allele (black bar) and carriers of the A allele (gray bar) of IVS2 ⫹ 21G⬎A. *p ⫽ 0.001. (B) Survival after disease onset by IVS2 ⫹ 21G⬎A genotype. A Kaplan-Meier plot of survival is given for Belgian patients homozygous for the wild type GG (dashed line) and carriers of the A allele (black line). Y-axis denotes cumulative survival, X-axis denotes survival time in months after disease onset (t ⫽ 0).

(haplotype frequency 62%; haplotype-specific p ⫽ 0.026), whereas two haplotypes tagged by the rare allele at IVS2 ⫹ 21G⬎A were associated with an earlier age at onset (haplotype-specific p ⫽ 0.023 and 0.038). A sliding window analysis revealed a 2-SNP window of IVS2 ⫹ 21G⬎A and IVS3-47-46insGTCA to be associated with age at onset (p ⫽ 0.005) (figure E-2). Moreover, carriers of the rare allele of IVS2 ⫹ 21G⬎A had a shorter survival after onset of ALS (HR 1.70 [95% CI 1.10 to 2.64]; p ⫽ 0.017) (figure 3B). Although nonsignificant, an HR of similar magnitude was observed for carriers of the rare allele of IVS3-4746insGTCA and IVS4 ⫹ 24G⬎A (HR 1.82 [95% CI 0.84 to 3.95]; p ⫽ 0.1). In order to confirm these findings, we replicated the association analysis in a Dutch sample of patients with ALS and controls. As in the Belgian group, we did not observe significant association between single SNPs or haplotypes and disease status or site of onset, and no association

was found with age at onset, but we could replicate an association between PGRN and survival after onset. Patients homozygous for the rare allele at IVS3-47-46insGTCA had a shorter survival after onset of ALS (HR 2.29 [95% CI 1.15 to 4.55]; p ⫽ 0.018). DISCUSSION We report a systematic screening and association study of PGRN in a series of 230 Belgian patients with ALS and 436 control individuals. We detected 11 mutations in patients, of which five were predicted to affect protein sequence or levels. Associations between three consecutive common SNPs (IVS2 ⫹ 21G⬎A [rs9897526], IVS3-47_-46insGTCA [rs34424835], IVS4 ⫹ 24G⬎A [rs850713]) and disease onset and survival further indicated the presence of a PGRN variant (in linkage disequilibrium with these SNPs) that has a significant effect on disease progression, which could be replicated in an independent series of Dutch patients with ALS and controls. The observation of a putative modifying effect on disease progression is of clear interest in the pathogenesis of ALS, since at a clinical level, ALS is known to have a wide variability in onset age and disease duration, even within autosomal dominant families with a known mutation.30 One intronic PGRN SNP (IVS2 ⫹ 21G⬎A) was consistently associated with both age at onset and survival in the Belgian sample, whereas two other SNPs (IVS3-47_-46insGTCA and IVS4 ⫹ 24G⬎A) in LD with IVS2 ⫹ 21G⬎A and a two-SNP haplotype including IVS2 ⫹ 21G⬎A and IVS3-47_-46insGTCA also showed association with age at onset. Moreover, IVS3-47_-46insGTCA showed association with a reduction in survival after onset in the Dutch replication sample. Our data suggest that PGRN might act as a modifier of motor neuron loss in ALS. Conceivably, the function of PGRN in neuronal survival is altered or reduced, leading to a more rapid loss of motor neurons. The threshold of motor units required to be lost before ALS becomes overt might be reached sooner, and after the disease has become clinically evident, the progression may be faster. None of the associated SNPs affected consensus splice sites or alternative splice sites, and different SNPs were associated with survival in the Belgian and Dutch studies, arguing against their direct role in pathogenicity, but rather indicating that a true modifying variant must exist in LD with these SNPs. In the mutation analysis we did not detect variants that might explain this association, but of interest is

that all three associated SNPs are in LD with rs5848 located in the 3= UTR. Rs5848 has been implicated in differential PGRN expression in a database-mining experiment investigating allele frequency imbalances between SNPs in redundant expressed sequence tags and regulatory cis-acting variants in LD with these SNPs, which signifies cis-acting variation in gene expression and potential consequences on phenotypic variation.31 In the PGRN sequencing we detected four missense mutations in five patients (5/230, 2.2%), which were all located in or at the border of granulin domains. Although it is unclear whether these mutations affect protein function, and whether disruption of one granulin domain is sufficient to cause a clinical phenotype, SIFT analysis predicted that at least one mutation, Ile124Thr, was close to the threshold of affecting protein function. Ile124Thr was observed in a patient with a bulbar onset of ALS, and a rapid disease progression. The other three missense mutations were predicted to be tolerated in SIFT analyses, but this evolutionary conservation analysis does not exclude pathogenicity. The three mutations (Arg110Gln, Ile124Thr, and Arg418Gln) located at the borders of granulin domains could conceivably interfere with proteolytic cleavage (e.g., by altering affinity for elastase or secretory leukocyte protease inhibitor17), thus disturbing the balance between PGRN and granulin peptides in the inflammatory response that is pertinent in neurodegeneration. For example, Arg418Gln is located at the C-terminal border of granulin C, and albeit not highly conserved, there seems to be a preference of positively charged amino acids at the C-terminal borders of the granulin domains, suggesting that a mutation at this position might alter cleavage. Structural modeling revealed that Ile124Thr had a weakly stabilizing and Ala324Thr a destabilizing effect on the granulin domain structure. Altered structural stability could directly affect protein function, either by interfering with protein–protein interactions or by misfolding and early degradation. Whereas the latter seems implausible because of the weak effect of both mutations on stability, changed interactions might be of interest, for example in light of the reported stimulating effect of PGRN on VEGF.19 Additional evidence for a pathogenic role for Ala324Thr is provided by our observation that the two ALS carriers shared a 2.8 Mb haplotype surrounding PGRN with a patient with Alzheimer dementia, suggesting a distant common ancestor. This founder effect is in favor of a pathogenic nature of this mutation. Also the Neurology •••

5

other three missense mutations had minor allele frequencies of ⬍0.1% in Belgian control individuals supporting that they are pathogenic in nature rather than a rare innocent polymorphism. Further, a mutation in the 5= regulatory region (EX 0, g.96061A⬎G) that we observed in one patient is likely to abolish EVI1, PAX2, and ISRE transcription factor binding sites, and create a NFAT binding site. Whether these mutations are truly pathogenic remains to be confirmed in functional analyses assessing protein function or transcriptional activity. However, we recently obtained evidence for a role of PGRN missense mutations in the pathogenesis of FTD, at least one of which (p.Arg432Cys) was shown to segregate with disease.24 It is important to mention that in contrast to FTDU-17, we did not find null mutations in ALS. In a previous study of patients with ALS, no PGRN null mutations were observed,32 and in a French FTD series, null mutations were never observed in patients with FTD with concurrent motor neuron disease.33 Together these data might suggest that PGRN haploinsufficiency does not contribute significantly to ALS pathogenesis. Rather, null mutations and missense mutations or common polymorphisms in PGRN might give rise to a different, albeit related, clinical phenotype of either FTDU or ALS through different mutational mechanisms. This has already been documented for other genes, such as the peripheral myelin protein 22 (PMP22) gene in demyelinating diseases. Haploinsufficiency of PMP22, through deletion of the gene or through rapid degradation of mutant truncated protein, causes the relatively mild phenotype of hereditary neuropathy with liability to pressure palsy, whereas duplication of PMP22 causes Charcot-Marie-Tooth disease type 1 (CMT1), and missense mutations are associated with a more severe CMT1 phenotype and Dejerine-Sottas syndrome.34 Comparably, heterozygous missense mutations in senataxin (ALS4) cause juvenile ALS35 whereas nonsense and frameshift mutations are predominantly found in an autosomal recessive form of spinocerebellar ataxia (ataxia-ocular apraxia 2).36,37 Further, common genetic polymorphisms in genes in which null mutations cause disease may affect related traits, disease susceptibility, or severity. In ␤-thalassemia, e.g., the clinical phenotype may be altered when a polymorphism (e.g., ⫺158G␥ C¡ T, which is associated with increased levels of fetal hemoglobin in healthy individuals) in the promoter of the ␥-globin gene is co-inherited with a null mutation.38 6

Neurology •••

While in our study PGRN null mutations apparently do not contribute substantially to the occurrence of ALS compared to FTDU, potential pathogenic missense and 5= regulatory mutations were detected in 6/230 patients with ALS (2.6%). Perhaps more importantly, we observed that genetic variability in PGRN potentially modified disease progression in ALS. Replication of the association findings will remain necessary in additional independent ALS patient samples, but our data indicate that PGRN genetic variability acts as a modifier of the course of disease rather than as an important cause of ALS. If it can be confirmed that PGRN levels contribute to the wide variability in disease progression in ALS, it might also serve as a target for therapeutic interventions. Received May 21, 2007. Accepted in final form August 13, 2007.

REFERENCES 1. Lomen-Hoerth C, Murphy J, Langmore S, Kramer JH, Olney RK, Miller B. Are amyotrophic lateral sclerosis patients cognitively normal? Neurology 2003;60:1094–1097. 2. Ringholz GM, Appel SH, Bradshaw M, Cooke NA, Mosnik DM, Schulz PE. Prevalence and patterns of cognitive impairment in sporadic ALS. Neurology 2005;65:586–590. 3. Wightman G, Anderson VE, Martin J, et al. Hippocampal and neocortical ubiquitin-immunoreactive inclusions in amyotrophic lateral sclerosis with dementia. Neurosci Lett 1992;139:269–274. 4. Mackenzie IR, Feldman H. Neuronal intranuclear inclusions distinguish familial FTD-MND type from sporadic cases. Dement Geriatr Cogn Disord 2004;17:333–336. 5. Mackenzie IR, Feldman HH. Ubiquitin immunohistochemistry suggests classic motor neuron disease, motor neuron disease with dementia, and frontotemporal dementia of the motor neuron disease type represent a clinicopathologic spectrum. J Neuropathol Exp Neurol 2005;64:730–739. 6. Arai T, Hasegawa M, Akiyama H, et al. TDP-43 is a component of ubiquitin-positive tau-negative inclusions in frontotemporal lobar degeneration and amyotrophic lateral sclerosis. Biochem Biophys Res Commun 2006;351:602–611. 7. Neumann M, Sampathu DM, Kwong LK, et al. Ubiquitinated TDP-43 in frontotemporal lobar degeneration and amyotrophic lateral sclerosis. Science 2006;314: 130–133. 8. Majoor-Krakauer D, Ottman R, Johnson WG, Rowland LP. Familial aggregation of amyotrophic lateral sclerosis, dementia, and Parkinson’s disease: evidence of shared genetic susceptibility. Neurology 1994;44: 1872–1877. 9. Hosler BA, Siddique T, Sapp PC, et al. Linkage of familial amyotrophic lateral sclerosis with frontotemporal dementia to chromosome 9q21-q22. JAMA 2000; 284:1664–1669. 10. Talbot K, Ansorge O. Recent advances in the genetics of amyotrophic lateral sclerosis and frontotemporal

11.

12.

13.

14.

15.

16.

17.

18.

19.

20.

21.

22.

23.

dementia: common pathways in neurodegenerative disease. Hum Mol Genet 2006;15 suppl 2:R182–R187. Baker M, Mackenzie IR, Pickering-Brown SM, et al. Mutations in progranulin cause tau-negative frontotemporal dementia linked to chromosome 17. Nature 2006;442:916–919. Cruts M, Gijselinck I, van der ZJ, et al. Null mutations in progranulin cause ubiquitin-positive frontotemporal dementia linked to chromosome 17q21 Nature 2006; 442:920–924. Bhandari V, Palfree RG, Bateman A. Isolation and sequence of the granulin precursor cDNA from human bone marrow reveals tandem cysteine-rich granulin domains. Proc Natl Acad Sci USA 1992;89:1715–1719. Zhu J, Nathan C, Jin W, et al. Conversion of proepithelin to epithelins: roles of SLPI and elastase in host defense and wound repair. Cell 2002;111:867–878. Daniel R, He Z, Carmichael KP, Halper J, Bateman A. Cellular localization of gene expression for progranulin. J Histochem Cytochem 2000;48:999–1009. He Z, Bateman A. Progranulin (granulin-epithelin precursor, PC-cell-derived growth factor, acrogranin) mediates tissue repair and tumorigenesis. J Mol Med 2003;81:600–612. Ahmed Z, Mackenzie IR, Hutton ML, Dickson DW. Progranulin in frontotemporal lobar degeneration and neuroinflammation. J Neuroinflammation 2007;4:7. Malaspina A, Kaushik N, de Belleroche J. Differential expression of 14 genes in amyotrophic lateral sclerosis spinal cord detected using gridded cDNA arrays. J Neurochem 2001;77:132–145. Tangkeangsirisin W, Serrero G. PC cell-derived growth factor (PCDGF/GP88, progranulin) stimulates migration, invasiveness and VEGF expression in breast cancer cells. Carcinogenesis 2004;25:1587–1592. Pritchard JK, Stephens M, Donnelly P. Inference of population structure using multilocus genotype data. Genetics 2000;155:945–959. Brouwers N, Nuytemans K, van der Zee J, et al. Clinical expression of progranulin in an extended founder family includes Alzheimer and Parkinson diagnoses. Arch Neurol 2007;64:1436 –1446. van der Zee J, Rademakers R, Engelborghs S, et al. A Belgian ancestral haplotype harbours a highly prevalent mutation for 17q21-linked tau-negative FTLD. Brain 2006;129:841–852. Ng PC, Henikoff S. Accounting for human polymorphisms predicted to affect protein function. Genome Res 2002;12:436–446.

24.

25.

26.

27. 28.

29.

30.

31.

32.

33.

34.

35.

36.

37.

38.

van der Zee J, Le BI, Maurer-Stroh S, et al. Mutations other than null mutations producing a pathogenic loss of progranulin in frontotemporal dementia. Hum Mutat 2007;28:416. Czaplewski C, Oldziej S, Liwo A, Scheraga HA. Prediction of the structures of proteins with the UNRES force field, including dynamic formation and breaking of disulfide bonds. Protein Eng Des Sel 2004;17:29–36. Cartharius K, Frech K, Grote K, et al. MatInspector and beyond: promoter analysis based on transcription factor binding sites. Bioinformatics 2005;21:2933–2942. Terwilliger J, Ott J. Handbook of Human Genetic Linkage. Baltimore: Johns Hopkins UP; 1994. Barrett JC, Fry B, Maller J, Daly MJ. Haploview: analysis and visualization of LD and haplotype maps. Bioinformatics 2005;21:263–265. Schaid DJ, Rowland CM, Tines DE, Jacobson RM, Poland GA. Score tests for association between traits and haplotypes when linkage phase is ambiguous. Am J Hum Genet 2002;70:425–434. Majoor-Krakauer D, Willems PJ, Hofman A. Genetic epidemiology of amyotrophic lateral sclerosis. Clin Genet 2003;63:83–101. Ge B, Gurd S, Gaudin T, et al. Survey of allelic expression using EST mining. Genome Res 2005;15:1584– 1591. Schymick J, Yang Y, Andersen P, et al. Progranulin mutations and ALS or ALS-FTD phenotypes. J Neurol Neurosurg Psychiatry 2007; 78:754–756. Le Ber I, van der Zee J, Hannequin D, et al. Progranulin null mutations in both sporadic and familial frontotemporal dementia. Hum Mutat 2007;28:846–855. Nelis E, Haites N, Van Broeckhoven C. Mutations in the peripheral myelin genes and associated genes in inherited peripheral neuropathies. Hum Mutat 1999;13: 11–28. Chen YZ, Bennett CL, Huynh HM, et al. DNA/RNA helicase gene mutations in a form of juvenile amyotrophic lateral sclerosis (ALS4). Am J Hum Genet 2004;74: 1128–1135. Duquette A, Roddier K, McNabb-Baltar J, et al. Mutations in senataxin responsible for Quebec cluster of ataxia with neuropathy. Ann Neurol 2005;57:408–414. Moreira MC, Klur S, Watanabe M, et al. Senataxin, the ortholog of a yeast RNA helicase, is mutant in ataxia-ocular apraxia 2. Nat Genet 2004;36:225–227. Thein SL. Genetic modifiers of beta-thalassemia. Haematologica 2005;90:649–660.

Neurology •••

7