Serum biomarker for progranulin-associated frontotemporal lobar degeneration

Serum Biomarker for Progranulin-Associated Frontotemporal Lobar Degeneration Kristel Sleegers, MD, PhD,1–3 Nathalie Brouwers, MSc,1–3 Philip Van Damme, MD, PhD,4 – 6 Sebastiaan Engelborghs, MD, PhD,3,7,8 Ilse Gijselinck, MSc,1–3 Julie van der Zee, PhD,1–3 Karin Peeters, BSN,1–3 Maria Mattheijssens, BSN,1–3 Marc Cruts, PhD,1–3 Rik Vandenberghe, MD, PhD,5,6 Peter P. De Deyn, MD, PhD,3,7,8 Wim Robberecht, MD, PhD,4 – 6 and Christine Van Broeckhoven, PhD, DSc1–3

Objective: Mutations that lead to a loss of progranulin (PGRN) explain a considerable portion of the occurrence of frontotemporal lobar degeneration. We tested a biomarker allowing rapid detection of a loss of PGRN. Methods: We used an enzyme-linked immunosorbent assay to measure in serum the PGRN protein levels of six affected and eight unaffected carriers from within an extended Belgian founder family segregating the null mutation IVS1⫹5G⬎C. Further, we measured serum PGRN levels in 2 patients with another null mutation (a Met1 and a frameshift mutation), in 4 patients carrying a predicted pathogenic missense mutation and in 5 patients carrying a benign missense polymorphism, in 9 unaffected noncarrier relatives, and in 22 community controls. Results: Serum PGRN levels were reduced in both affected and unaffected null mutation carriers compared with noncarrier relatives ( pexact ⬍ 0.0001), and allowed perfect discrimination between carriers and noncarriers (sensitivity: 1.0; 1 ⫺ specificity: 0.0). Serum PGRN levels in Cys139Arg and Arg564Cys mutation carriers were significantly lower than in controls, but greater than in null mutation carriers, fitting the hypothesis of partial loss of function caused by these missense mutations. As expected, levels for carriers of benign missense polymorphisms were not significantly different from controls. Interpretation: Our results indicate that the serum PGRN level is a reliable biomarker for diagnosing and early detection of frontotemporal lobar degeneration caused by PGRN null mutations, and provided the first in vivo evidence that at least some missense mutations in PGRN may lead to a (partial) loss of PGRN. Ann Neurol 2009;65:603– 609

Frontotemporal lobar degeneration (FTLD) is a collective term for a number of focal neurodegenerative diseases that result from atrophy of the frontal or anterior temporal cortices, or both. When considered from a clinical, microscopic, or genetic perspective, marked heterogeneity exists. Clinically, FTLD presents with either changes in personality and social conduct (frontotemporal dementia) or language problems (primary progressive aphasia).1 Primary progressive aphasia can be further subdivided into a fluent variant (semantic dementia) and a nonfluent variant (progressive nonfluent aphasia). The microscopic neuropathological changes underlying these clinical phenotypes can, in turn, be subdi-

vided. In most FTLD patients, affected brain areas show specific neuronal inclusions. In the majority, these inclusions are ubiquitin immunoreactive and contain TAR-DNA binding protein (TDP43).2 In a smaller fraction of FTLD patients, deposits of hyperphosphorylated tau proteins are observed.3 Four genes are known to play a role in the cause of FTLD. Mutations in the microtubule-associated protein tau (MAPT) gene are a cause of FTLD with tauopathy,4 mutations in progranulin (PGRN5,6) and valosincontaining protein (VCP7) cause FTLD with TDP43positive inclusions, and mutations in chromatinmodifying protein 2B (CHMP2B8) are associated with FTLD with tau-negative, TDP43-negative, ubiquitin-

From the 1Neurodegenerative Brain Diseases Group, Department of Molecular Genetics, VIB (Flanders Institute for Biotechnology), University of Antwerp; 2Laboratory of Neurogenetics, Institute Born-Bunge; 3University of Antwerp, Antwerpen; 4Laboratory for Neurobiology and Vesalius Research Center, VIB, 5Department of Neurology , University Hospital Gasthuisberg; 6University of Leuven, Leuven; 7Laboratory of Neurochemistry and Behavior, Institute Born-Bunge; and 8Memory Clinic and Division of Neurology, ZNA Middelheim, Antwerpen, Belgium.

University of Antwerp–CDE, Universiteitsplein 1, B-2610 Antwerpen, Belgium. E-mail: [email protected]

Address correspondence to Dr Van Broeckhoven, VIB–Department of Molecular Genetics, Neurodegenerative Brain Diseases Group,

Published in Wiley InterScience (www.interscience.wiley.com). DOI: 10.1002/ana.21621

Potential conflict of interest: Nothing to report. Additional Supporting Information may be found in the online version of this article. Received Jul 11, 2008, and in revised form Oct 29. Accepted for publication Nov 25, 2008.

© 2009 American Neurological Association

603

positive inclusions. For a long time, MAPT mutations were the only known genetic cause of FTLD; therefore, research on drug development and biomarkers has primarily been focused on tauopathy. The presence of a different (nontauopathy) disease entity in more than 50% of the patients may explain the limited success of these studies. With recently expanding knowledge on the distinct underlying pathophysiological mechanisms and neuropathological changes, it will become possible to develop drugs that target the underlying pathology specifically. Antemortem differentiation between the distinct molecular entities is pertinent to ensure success. Despite emerging evidence of differences in clinical manifestation between patients with tau-positive or tau-negative FTLD,9,10 overlap is considerable, making it impossible to predict the underlying pathomechanism for individual patients. Biomarkers are needed that reflect the central pathogenic process well, and that predict disease already in early stages, or even in the preclinical phase. Ideally, a good biomarker test is easy to perform, (relatively) noninvasive, and inexpensive. All confirmed pathogenic mutations in PGRN reported to date are loss-of-function (LOF) mutations that lead to a reduction in PGRN protein levels (http://www.molgen. ua.ac.be/FTDMutations/). We hypothesized that a pathogenic LOF PGRN mutation will be measurably manifested in serum with decreased protein levels, which might serve as a biomarker reflecting the central pathogenic process of PGRN-related FTLD. We aimed to test this hypothesis by measuring PGRN protein levels with an enzyme-linked immunosorbent assay (ELISA) that previously detected a significant decrease of PGRN in cerebrospinal fluid of PGRN null mutation carriers.11 We investigated three different null mutations, and nine missense mutations identified in Belgian FTLD and Alzheimer’s disease (AD) patients. The null mutations included PGRN IVS1⫹5G⬎C (for which haploinsufficiency is the result of intranuclear degradation of the unspliced transcript, causing an approximately 50% reduction in PGRN protein), Met1Ile (which prevents translation of the mutant transcript leading to an approximately 50% reduction in PGRN protein), and a frameshift mutation (Pro127ArgfsX2, with an approximately 50% reduction in PGRN protein because of nonsense-mediated messenger RNA decay).6 Of the missense mutations we identified in patients with different clinical diagnoses (Table), four were predicted to be pathogenic based on genetic, in silico, and in vitro data (Cys139Arg, Arg432Cys, Pro451Leu, and Arg564Cys),12–14 whereas the others are more likely to be benign missense polymorphisms (Asp33Glu, Leu261Ile, Ala324Thr, Arg433Trp, and Gly515Ala), because we had no evidence of pathogenicity (http:// www.molgen.ua.ac.be/FTDMutations/).12,13,15,16

604

Annals of Neurology

Vol 65

No 5

May 2009

Subjects and Methods Study Population We investigated serum PGRN levels of IVS1⫹5G⬎C PGRN mutation carriers in a large Belgian FTLD-U family (DR8) segregating this null mutation.6,17,18 The family has 10 branches extending over at least 5 generations and comprising at least 226 individuals, of whom 41 had a clinical or pathological diagnosis of FTLD-U, and 57 were obligate or genetically confirmed unaffected mutation carriers. All living patients and relatives older than 18 years were contacted and received information regarding the study design. Consequently, 6 affected mutation carriers (median age at onset, 67.5 [range, 45–76] years; 4 women), 8 unaffected mutation carriers (median age at onset, 43 [range, 34 – 69] years, 5 women), and 5 noncarrier first-degree relatives of distantly related patients within the DR8 pedigree could be included in this study. In addition, we included two Belgian FTLD-U patients with a different PGRN null mutation (DR118.1, PGRN Met1Ile; DR120.1, PGRN Pro127ArgfsX2) and four of their noncarrier relatives. Twelve missense mutation carriers were included, of whom 5 carried a predicted pathogenic missense mutation (Cys139Arg: n ⫽ 2; Arg432Cys: n ⫽ 1; Pro451Leu: n ⫽ 1; and Arg564Cys: n ⫽ 1). The likely benign missense polymorphisms included Arg433Trp (n ⫽ 4), Leu261Ile (n ⫽ 1, also carried Arg433Trp), Asp33Glu (n ⫽ 1), Ala324Thr (n ⫽ 1), and Gly515Ala (n ⫽ 1) (see the Table for clinical characteristics). Four additional noncarriers were included; these were relatives of other mutation carriers. Median age of all nine related noncarriers was 51 (range, 32– 84) years; four were women. In addition, we investigated PGRN serum levels in a larger group of random community control individuals (n ⫽ 22). All participants (or legal custodian in case of the patient) gave written, informed consent that their biomaterial could be used anonymously in scientific research without obtaining a personal result. For individual genetic counseling if desired, family members were referred to a governmental center for medical genetic counseling and DNA diagnostics. The medical ethical committees of the ZNA Middelheim and the University of Antwerp, Belgium, granted their approval for this study.

Data Collection Blood was drawn and processed within 4 hours after collection. Each tube was centrifuged 22.880g at 4°C for 10 minutes, and serum was separated and stored in 1ml aliquots in liquid nitrogen until further use in this study.

Progranulin Enzyme-Linked Immunosorbent Assay PGRN serum levels were measured by ELISA in duplicate, blinded for PGRN genotype and FTLD-U phenotype. Ninety-six-well plates were coated with monoclonal antiPGRN antibody (R&D Systems, Minneapolis, MN) and blocked with 1% bovine serum albumin, before samples and standard (recombinant human PGRN, R&D Systems) were loaded. Afterward, bound PGRN was detected using biotinylated polyclonal anti-PGRN antibody (R&D Systems), avidin-biotin complex, and orthophenylenediamine. Absorbance was measured at 490nm. The antibody recognizes fulllength PGRN, and not its cleavage products.11

Table. Serum Progranulin Levels and Characteristics of Patients Carrying a Progranulin Mutation Patient No.

Mutation

Serum PGRN (ng/ml)

Initial Clinical Diagnosis

Age at Onset (years)

Disease Durationa (years)

Sex

IVS1⫹5G⬎C

61.5

PNFA

45

2

F

Null mutations DR119.1 DR142.1

IVS1⫹5G⬎C

63.3

AD

66

6

F

DR25.1

IVS1⫹5G⬎C

81.0

FTD

69

6

F

DR26.1

IVS1⫹5G⬎C

52.8

PNFA

65

2

M

DR25.14

IVS1⫹5G⬎C

73.9

AD

76

1

F

DR25.5

IVS1⫹5G⬎C

48.7

FTD

70

2

M

DR120.1

Pro127ArgfsX2

87.0

PNFA

56

3

F

DR118.1

Met1Ile

61.1

FTD

62

3

F

b

Predicted pathogenic missense mutations DR197.1

Cys139Arg

102.1

AD

80

6

F

DR121.1

Arg432Cys

125.7

FTD

66

3

F

DR152.1

Pro451Leu

165.0

AD

74

3

F

DR144.1

Arg564Cys

90.6

AD

70

8

F

b

Predicted benign amino acid substitutions DR148.1

Asp33Glu

152.3

AD

81

2

F

DR196.1

Ala324Thr

139.1

AD

86

6

M

DR195.1

Leu261Ile/Arg433Trp

136.1

AD

78

4

F

DR198.1

Arg433Trp

145.6

AD

80

7

F

DR199.1

Arg433Trp

112.0

AD

74

6

F

DR92.1

Arg433Trp

175.1

AD

70

12

F

DR201.1

Gly515Ala

138.8

AD

89

3

F

a

At time of blood sampling. b Based on presence or absence in control individuals and on in silico data (evolutionary conservation and effect on protein structure and stability). PGRN ⫽ progranulin; PNFA ⫽ progressive nonfluent aphasia; AD ⫽ Alzheimer’s disease; FTD ⫽ frontotemporal dementia.

Statistical Analysis Duplicate serum PGRN levels were averaged for further analysis. Nonparametric Mann–Whitney U test was used to compare quantitative traits (serum PGRN levels, age) between two groups (carrier status), and Kruskal–Wallis test for comparison of three groups. Exact p values were computed. Correlations between age and serum PGRN levels were assessed using Spearman’s rho. Diagnostic performance of the ELISA was assessed by a receiver operating characteristics curve. All analyses were performed using SPSS 14.0 software (SPSS, Chicago, IL).

Results Serum Progranulin Levels in Affected Carriers of the IVS1⫹5G⬎C Founder Mutation The median serum PGRN level in FTLD-U patients carrying the founder mutation was 62.4ng/ml, and ranged from 48.7 to 81.0ng/ml (Fig 1). This was significantly different ( pexact ⬍ 0.0001) from the serum

PGRN levels measured in healthy noncarrier relatives from within the same family, for which the median was 135.3ng/ml, ranging from 110.9 to 144.4ng/ml, and from community control individuals (median, 226.5ng/ ml; range, 101–387ng/ml; pexact ⬍ 0.0001). We observed no pattern of association between serum PGRN levels and initial clinical diagnosis of the FTLD-U patients (frontotemporal dementia, progressive nonfluent aphasia, AD; see the Table). Age at onset varied widely in this founder FTLD-U family (see Supplementary Fig), but there was no evidence of correlation between onset age and serum PGRN levels (Spearman’s rho, 0.26; p ⫽ 0.6). Serum Progranulin Levels in Unaffected Carriers of the IVS1⫹5G⬎C Founder Mutation Unaffected carriers had average serum PGRN levels comparable with their affected relatives (see Fig 1). The

Sleegers et al: Serum Progranulin in FTLD

605

Serum Progranulin Levels in Missense Mutation Carriers Next, we measured serum PGRN levels in 12 individuals carrying either a predicted pathogenic missense mutation or a benign missense polymorphism; 10 were clinically diagnosed with AD, 1 had FTLD, and 1 was a healthy first-degree relative carrying the Cys139Arg mutation (see the Table). Serum PGRN levels in carriers of benign missense polymorphisms (Asp33Glu, Leu261Ile, Ala324Thr, Arg433Trp, or Gly515Ala) were not significantly different from control individuals ( p ⫽ 0.2; median, 142.7ng/ml, ranging from 112.0 –175.1ng/ml; Fig 2) and were significantly greater than serum PGRN levels of null mutation carriers ( pexact ⬍ 0.0001). Patients carrying a pathogenic missense mutation (Cys139Arg, Arg432Cys, Pro451Leu, or Arg564Cys) had intermediate levels between null mutation carriers and control in-

Fig 1. Box plot for serum enzyme-linked immunosorbent assay (ELISA) of progranulin (PGRN) for affected and unaffected IVS1⫹5G⬎C carriers and noncarriers. Serum levels of PGRN protein are depicted in nanograms per milliliter. Solid lines in the boxes are median levels, box lengths display the interquartile range, whiskers indicate minimum and maximum levels per group, and open circles depict outliers. *Exact p ⬍ 0.0001 compared with noncarriers.

median of the average serum PGRN levels in unaffected carriers was 57.0ng/ml (range, 32.7–72.5ng/ml). These serum PGRN levels were significantly lower ( pexact ⬍ 0.0001) than in noncarrier relatives, as well as in community control individuals. There were no significant differences in age between unaffected carriers and noncarriers ( pexact ⫽ 0.5), and within the group of unaffected carriers, we observed no correlation between age at blood collection and serum PGRN levels (Spearman’s rho, ⫺0.405; p ⫽ 0.3). Serum Progranulin Levels in Frameshift and Met1 Mutation Carriers Of the FTLD-U patients carrying a frameshift or Met1 mutation, serum PGRN levels were 61.1 and 87.0ng/ml (DR118.1 and DR120.1, see the Table), both in line with the observed values for the founder null mutation carriers. A receiving operator characteristics analysis including all null mutation carriers and all control individuals suggested that a cutoff chosen at 94ng/ml (the average of the greatest level in carriers and the lowest level in noncarriers), would perfectly predict a null mutation with a sensitivity of 1.0 and a 1 ⫺ specificity of 0.0; that is, none of the noncarriers had levels less than this cutoff value (area under the curve, 1.0; p ⬍ 0.001).

606

Annals of Neurology

Vol 65

No 5

May 2009

Fig 2. Box plot for serum enzyme-linked immunosorbent assay (ELISA) of progranulin (PGRN) for patients and control individuals. White box depicts serum levels of community control individuals (n ⫽ 22) and relative control individuals (n ⫽ 9); dark gray box depicts serum levels in all patients tested (Alzheimer’s disease, frontotemporal dementia, or progressive nonfluent aphasia); and light gray boxes show the patient serum levels split up by type of mutation (benign missense polymorphism, pathogenic missense mutation, and null mutation). Asterisk indicates a significantly lower serum PGRN level when compared with control individuals, at p ⫽ 0.023; double asterisks indicate a significantly lower serum PGRN level when compared with control individuals, at p ⫽ 2e⫺8). Serum PGRN levels are measured in nanograms per milliliter. Solid lines in the boxes are median levels, box lengths display the interquartile range, and whiskers indicate minimum and maximum serum levels per group.

Fig 3. Serum progranulin (PGRN) levels for individual carriers relative to median control levels. Orange squares indicate benign missense polymorphisms; blue triangles indicate predicted pathogenic missense mutations based on in silico and/or in vitro data; purple triangles indicate affected null mutation carriers; and green squares indicate unaffected null mutation carriers. Clinical diagnoses are indicated per mutation carrier. Black striped line indicates the median serum PGRN level of relative control individuals sampled following the same strict protocol as the patients, with the interquartile range demarcated by the dotted black lines. Red striped line indicates the median serum PGRN level of community control individuals, with the interquartile range demarcated by the dotted red lines.

dividuals or benign missense polymorphism carriers (median, 113.9ng/ml; range, 90.6 –175.1ng/ml; see Fig 2), being significantly lower than control individuals ( pexact ⫽ 0.023) and significantly greater than null mutation carriers ( pexact ⫽ 0.005), suggestive of a partial loss of PGRN protein. When considering the individual PGRN levels (Fig 3), a high variation is apparent, with a marked reduction in serum PGRN levels for two of the predicted pathogenic missense mutations (Cys139Arg and Arg564Cys), whereas the Pro451Leu and Arg432Cys carriers had levels comparable with the benign missense polymorphism carriers. The serum PGRN level of the healthy first-degree relative carrying Cys139Arg (94.4ng/ml) was in the same intermediate range as that of the AD patient carrying this mutation, further underscoring a loss of PGRN associated with this mutation (see Fig 3). Receiver operating characteristics analysis indicated that a randomly chosen carrier of one of the four predicted pathogenic missense mutations has an 87.7% chance of having a lower PGRN serum level than a randomly chosen control individual ( p ⫽ 0.007).

Discussion In this study, we showed that serum PGRN levels of FTLD-U patients carrying a PGRN null mutation are significantly reduced compared with noncarriers. This corresponds well with the loss of activity of one of two alleles of PGRN leading to PGRN haploinsufficiency in carriers of a null mutation.5,6 These results support the use of serum PGRN levels as a biomarker for diagnosis of PGRNrelated FTLD-U, because they correctly reflect the underlying molecular disease entity. Moreover, as a diagnostic test, it is relatively noninvasive (eg, when compared with cerebrospinal fluid) and easy to perform. Furthermore, in this study, the range of serum PGRN levels observed in null mutation carriers did not overlap with that of noncarriers, predicting an excellent diagnostic performance. Serum PGRN levels were already reduced in unaffected carriers, with levels in the same range as serum levels of the patients manifesting FTLD-U, suggesting that the measurement of serum PGRN levels can straightforwardly identify presymptomatic at-risk individuals even in the absence of knowledge of the underlying PGRN mutation.

Sleegers et al: Serum Progranulin in FTLD

607

This has important ethical implications comparable with presymptomatic genetic screening for known PGRN mutations, especially in the current absence of a cure or prevention. But in the advent of targeted treatment or prevention, preclinical detection will be crucial in the proper management of this disease at an early stage. Another inference of the equally low serum PGRN levels in affected and unaffected carriers is that the wide variability of onset age observed in FTLD-U families segregating a PGRN mutation17,19 appears not to result from differences in serum PGRN levels, though modest correlations could have been missed in this relatively small sample size. However, while in the process of revising our article, a PGRN plasma study was published,20 confirming our initial observation, and arguing against a role of PGRN levels as modifier of onset age. We previously predicted that at least some of the PGRN missense mutations observed in clinically diagnosed FTLD or AD patients were potentially pathogenic because of partial loss of mutant PGRN after protein misfolding and degradation in the endoplasmic reticulum.12,13 In vitro experiments confirmed our in silico prediction data showing that some PGRN missense mutations lead to a partial loss of mutant protein (45–70% vs 100% for null mutations14). In contrast with the complete loss of mutant protein in case of null mutations leading to autosomal dominant FTLD-U, this less pronounced reduction of PGRN may act to increase susceptibility to various forms of neurodegeneration such as AD. Here, we showed in vivo reduced serum PGRN levels for two such predicted pathogenic missense mutations, Cys139Arg and Arg564Cys, that we observed in clinically diagnosed AD patients. The Cys139Arg mutation was also recently reported in a patient clinically diagnosed with early-onset familial FTLD.21 This mutation disrupts one of the six disulfide bridges in granulin domain F (grnF), responsible for the typical folding of each of granulin domain of PGRN, known as the granulin Cys-fold. The Arg564Cys mutation introduces an extra Cys residue that may compete with existing disulfide bridges. Our observation of partially reduced PGRN levels in serum of Cys139Arg and Arg564Cys carriers corroborates with our predictions regarding pathogenicity of these mutations. On the other hand, two other predicted pathogenic missense mutations had serum PGRN levels indistinguishable from those of benign missense polymorphisms, even though the Arg432Cys mutation in vitro resulted in reduced secretion of PGRN.14 Although these two mutations gave insignificant in vivo data in our ELISA, we cannot yet exclude a pathogenic role for these missense mutations. Perhaps they act by disrupting the function of the respective granulin rather than the holoprotein, or perhaps patient carriers have a genetic background that offsets the reduced secretion of PGRN in the presence of the missense mutation. As expected, the serum PGRN levels for the benign missense

608

Annals of Neurology

Vol 65

No 5

May 2009

polymorphism carriers were in line with those of noncarriers. It is important to mention that the Cys139Arg mutation carriers would not have been recognized using the diagnostic cutoff value we calculated for null mutation carriers. Larger scale studies will have to define a threshold serum PGRN level below which the presence a pathogenic missense mutation should be considered, and follow-up functional studies are warranted. It should be noted that the serum PGRN levels observed in the relative control individuals sampled within families of null mutation carriers were lower and less variable than the serum PGRN levels measured in community control individuals. The limited variability in the former control group might be because of the fact that some of these control individuals were related (albeit distantly in the large founder family). More likely, their values were more similar because of a more rigidly followed sampling protocol we used in this study. Two other studies recently addressed the lack of a noninvasive biomarker for FTLD-U.22,23 One study reported that TDP43 levels measured by ELISA in plasma were increased in 46% of FTLD patients.23 Levels were also increased in 22% of AD patients and in 8% of cognitively healthy individuals. TDP43 measurements will be important in the initial diagnostic workup by discriminating between the two major proteinopathies underlying FTLD, but measurement of PGRN levels will allow refining the diagnostic process by direct detection of those patients who have PGRNrelated TDP43-opathy. The other study reported reduced PGRN messenger RNA expression in blood of two carriers of a PGRN frameshift mutation confirming LOF by nonsense-mediated messenger RNA decay.22 In contrast, the PGRN ELISA test measures reduction of serum PGRN levels independent of the nature of the LOF mutation. Another advantage of the serum ELISA test for PGRN is its easier applicability in a hospital setting and potential to monitor treatment efficacy in trials specifically aiming at restoring PGRN levels (eg, by administering exogenous PGRN). This is particularly pertinent given the tight physiological regulation of PGRN, with overexpression of this growth factor causing malignancies.15 Currently, in the absence of a treatment for PGRNrelated FTLD-U and of sufficient knowledge on why some PGRN null mutation carriers do not develop symptoms of disease until late in life, the primary application of this assay is distinguishing mutation carriers from noncarriers, similar to genetic testing. A benefit of this assay is, however, that in a nonreferral hospital where genetic testing is not readily available, it may expedite diagnosis and referral to a tertiary hospital. Moreover, genetic testing for LOF mutations is complex because they are characterized by a broad range of mutation types (eg, nonsense, frameshift, exon deletions, whole-gene deletions, and missense mutations) that can

lead to LOF at both the transcript and protein levels. Therefore, the serum PGRN ELISA test may help to identify null PGRN mutation carriers that might otherwise have been missed in standard molecular DNA diagnostic testing. Furthermore, we foresee several additional applications of this technique. It may expedite the diagnosis of seemingly sporadic FTLD patients with PGRN LOF mutations,24 and it may facilitate the development of causative-based treatment by defining more homogenous patients groups for drug trials. Furthermore, because PGRN LOF mutations have also been identified in patients with a clinical diagnosis of FTLD, AD, or Parkinson’s disease caused by clinical heterogeneous expression of these mutations,12,13,15,16 the test may also be applicable in differential diagnosis of these neurodegenerative disorders. Also, molecular DNA diagnostic screening of PGRN may demonstrate new PGRN mutations with an uncertain pathogenic effect on PGRN protein hampering a straightforward risk interpretation for the carrier. Here, serum PGRN levels may help to provide a rapid in vivo indication of a potential effect on PGRN production. Further studies, including larger samples and different PGRN mutations, are needed to appraise its performance as a diagnostic or predictive test, and to accurately define a diagnostic threshold value. Nonetheless, our data obtained in the Belgian FTLDU founder family have shown that serum PGRN ELISA is a promising biomarker for diagnosing or predicting PGRN-linked FTLD-U, with an excellent sensitivity and specificity. This work was supported by the Medical Foundation Queen Elisabeth (GSKE) (C.V.B., M.C.), the Foundation for Alzheimer Research (SAO/FRMA)–Belgium #06606 to M.C., #06651 to K.S., the InterUniversity Attraction Poles (IAP) (CVB, WR, PPDD) program P6/43 of the Belgian Federal Science Policy Office (BELSPO), the Fund for Scientific Research–Flanders (FWO-F; K.S., P.V.D, J.v.d.Z., S.E., N.B., I.G.), the Special Research Fund (BOF) (K.S.) of the University of Antwerp, the Research Council of the University of Leuven (W.R.), and the E. von Behring Chair for Neuromuscular and Neurodegenerative Disorders (W.R.).

We are grateful to the participants of this study for their cooperation and acknowledge the contribution of the BioBank of the Institute Born-Bunge. References 1. Neary D, Snowden JS, Gustafson L, et al. Frontotemporal lobar degeneration: a consensus on clinical diagnostic criteria. Neurology 1998;51:1546 –1554. 2. Neumann M, Sampathu DM, Kwong LK, et al. Ubiquitinated TDP-43 in frontotemporal lobar degeneration and amyotrophic lateral sclerosis. Science 2006;314:130 –133. 3. Cairns NJ, Bigio EH, Mackenzie IR, et al. Neuropathologic diagnostic and nosologic criteria for frontotemporal lobar degeneration: consensus of the Consortium for Frontotemporal Lobar Degeneration. Acta Neuropathol 2007;114:5–22.

4. Hutton M, Lendon CL, Rizzu P, et al. Association of missense and 5⬘-splice-site mutations in tau with the inherited dementia FTDP-17. Nature 1998;393:702–705. 5. 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. 6. Cruts M, Gijselinck I, van der Zee J, et al. Null mutations in progranulin cause ubiquitin-positive frontotemporal dementia linked to chromosome 17q21. Nature 2006;442:920 –924. 7. Watts GD, Wymer J, Kovach MJ, et al. Inclusion body myopathy associated with Paget disease of bone and frontotemporal dementia is caused by mutant valosin-containing protein. Nat Genet 2004;36:377–381. 8. Skibinski G, Parkinson NJ, Brown JM, et al. Mutations in the endosomal ESCRTIII-complex subunit CHMP2B in frontotemporal dementia. Nat Genet 2005;37:806 – 808. 9. Bian H, Grossman M. Frontotemporal lobar degeneration: recent progress in antemortem diagnosis. Acta Neuropathol 2007; 114:23–29. 10. Boeve BF, Hutton M. Refining frontotemporal dementia with parkinsonism linked to chromosome 17: introducing FTDP-17 (MAPT) and FTDP-17 (PGRN). Arch Neurol 2008;65:460 – 464. 11. Van Damme P, Van Hoecke A, Lambrechts D, et al. Progranulin functions as a neurotrophic factor to regulate neurite outgrowth and enhance neuronal survival. J Cell Biol 2008;181:37– 41. 12. Brouwers N, Sleegers K, Engelborghs S, et al. Genetic variability in progranulin contributes to risk for clinically diagnosed Alzheimer disease. Neurology 2008;71:656 – 664. 13. van der Zee J, Le Ber I, Maurer-Stroh S, et al. Mutations other than null mutations producing a pathogenic loss of progranulin in frontotemporal dementia. Hum Mutat 2007;28:416. 14. Shankaran S, Capell A, Hruscha A, et al. Missense mutations in the progranulin gene linked to frontotemporal lobar degeneration with ubiquitin-immunoreactive inclusions reduce progranulin production and secretion. J Biol Chem 2008;283:1744 –1753. 15. Cruts M, Van Broeckhoven C. Loss of progranulin function in frontotemporal lobar degeneration. Trends Genet 2008;24: 186 –194. 16. Nuytemans K, Pals P, Sleegers K, et al. Progranulin variability has no major role in Parkinson’s disease genetic etiology. Neurology 2008;71:1147–1151. 17. Brouwers N, Nuytemans K, van der Zee J, et al. Alzheimer and Parkinson diagnoses in progranulin null mutation carriers in an extended founder family. Arch Neurol 2007;64:1436 –1446. 18. 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. 19. Rademakers R, Baker M, Gass J, et al. Phenotypic variability associated with progranulin haploinsufficiency in patients with the common 1477C3 T (Arg493X) mutation: an international initiative. Lancet Neurol 2007;6:857– 868. 20. Ghidoni R, Benussi L, Glionna M, et al. Low plasma progranulin levels predict progranulin mutations in frontotemporal lobar degeneration. Neurology 2008;71:1235–1239. 21. Bernardi L, Tomaino C, Anfossi M, et al. Novel PSEN1 and PGRN mutations in early-onset familial frontotemporal dementia. Neurobiol Aging (in press). 22. Coppola G, Karydas A, Rademakers R, et al. Gene expression study on peripheral blood identifies progranulin mutations. Ann Neurol 2008;64:92–96. 23. Foulds P, McAuley E, Gibbons L, et al. TDP-43 protein in plasma may index TDP-43 brain pathology in Alzheimer’s disease and frontotemporal lobar degeneration. Acta Neuropathol 2008;116:141–146. 24. 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.

Sleegers et al: Serum Progranulin in FTLD

609