Refining Frontotemporal Dementia With Parkinsonism Linked to Chromosome 17:  Introducing FTDP-17 (MAPT) and FTDP-17 (PGRN) FREE

Frontotemporal dementia with parkinsonism is one of the major degenerative dementia syndromes (Table 1), particularly for those who begin experiencing cognitive, behavioral, or motor changes before age 65 years. Advances in immunocytochemistry and molecular genetics have greatly expanded our knowledge of the disorders (and their associated dysfunctional proteins) that can manifest as dementia with or without parkinsonism (Table 2). No disease-altering treatment has been identified as yet for any of the neurodegenerative disorders that can manifest clinically as FTDP (Table 2). The development of potential therapies requires knowledge about the pathophysiology of the varying disorders. The identification of causative genes offers opportunities to quickly learn about the pathophysiologic processes involved in neurodegeneration, and drug testing can proceed relatively quickly using transgenic mouse models that are designed to mimic the human disease. Several groups of investigators have focused on families carrying mutations that cause FTDP.

The hunt for causative genes in FTDP was largely spearheaded by the first Frontotemporal Dementia and Parkinsonism Linked to Chromosome 17 Consensus Conference in Ann Arbor in 1996, for which FTDP linked to chromosome 17 (FTDP-17) was the major focus.¹Soon thereafter in 1998, mutations in MAPTwere identified.²During the 8 years since this discovery, 41 mutations in MAPThave been found,³and many other issues relating to FTDP-17 due to mutations in MAPThave been characterized (Table 3).⁴No sex predilection has been identified. The typical age of onset varies between 25 and 65 years. Penetrance appears to be close to 100%, though individuals living into old age without symptoms have been observed in families with at least 1 mutation (exon 10 + 16).²The duration of symptoms from onset to death is typically 3 to 10 years. Symptomatology usually involves executive dysfunction and altered personality and behavior, with aphasia and parkinsonism evolving in many individuals. Memory impairment occurs less frequently as the primary presenting feature, and visuospatial impairment and limb apraxia are quite rare. Motor neuron disease is also infrequent, though several cases have been reported. Most patients carry 1 or more of the syndromic diagnoses listed in Table 1, particularly frontotemporal dementia (FTD) with or without parkinsonism, progressive nonfluent aphasia, or primary progressive aphasia. Rarely, the syndromes of mild cognitive impairment, probable Alzheimer disease, semantic dementia, or corticobasal syndrome are manifested. Few cases have been diagnosed with amyotrophic lateral sclerosis (ALS), and there are no reports of patients with mutations in MAPTwho were diagnosed with posterior cortical atrophy or dementia with Lewy bodies. Over time, most patients develop other clinical features such that 2 or more syndromes can be applied, reflecting the progressively expanding involvement of other brain regions.⁵

Structural neuroimaging studies show frontal and/or temporal atrophy, either symmetric or asymmetric⁶; parenchymal signal changes on magnetic resonance imaging are either absent or very mild.⁷A similar topography of abnormalities is typically seen on single-photon emission computed tomography and positron emission tomography scans, often with basal ganglia and/or thalamic hypoperfusion or hypometabolism. Pathologically, cortical atrophy is as indicated on imaging studies, with the maximally affected cortical gyri sometimes described as having a “knife edge” appearance. Tau-positive inclusions in neurons (eg, neurofibrillary tangles, neuronal threads, and Pick bodies) and/or glia (eg, astrocytic plaques and oligodendroglial coiled bodies) are always present on histologic examination, sometimes accompanied by argyrophilic grains. These tau-positive inclusions are often in a distribution such that patients would be pathologically identified as having corticobasal degeneration, progressive supranuclear palsy, argyrophilic grain disease, or Pick disease if the presence of an MAPTmutation was not known.

The mutations presumably cause disease either through disrupting the alternative splicing of MAPTexon 10 and thereby altering the relative levels of tau isoforms with 4 or 3 microtubule binding repeats, or they directly decrease the ability of tau to bind to and promote microtubule assembly and/or increase tau-filament formation. No disease-altering treatments exist yet for the tauopathies, though kinase inhibitors and microtubule tau stabilizers have shown promise in in vitro and animal model studies.^8- 9

A significant minority of patients with FTDP-17—many of whom were the focus of discussion at the meeting in Ann Arbor in 1996—had no identifiable mutations in MAPT, nor did they have any tau-positive inclusions at autopsy.^10- 13The recent identification of mutations in PGRNin all of these remaining chromosome 17–linked families and in many other kindreds (Figure)^14- 27has now solved this decade-long conundrum, a few atypical FTDP phenotypes have been redefined, and an amazing freak of nature has been realized. The PGRNgene is only 1.7 Mb centromeric to MAPTon chromosome 17, demonstrating an intriguing example of how 2 apparently different genes can cause a very similar phenotype and reside so near to each other on the same chromosome.²⁷

With closer inspection, how similar are the clinical phenotypes associated with mutations in MAPTand PGRN? While our knowledge of the full spectrum of clinical, radiologic, and pathologic issues in FTDP associated with mutations in PGRNis still evolving, interesting findings have already emerged that allow comparisonsbetween MAPT's and PGRN's mutation-associated characteristics (based on published findings to date and unpublished data from our group) (Table 3). The frequency of mutations in PGRNin FTD series is similar to that in MAPT.¹⁸With at least 35 mutations identified to date,³almost as many mutations in PGRNhave been discovered in less than 1 year than in the 8 years since the initial identification of mutations in MAPT. The mode of inheritance follows an autosomal dominant pattern but with reduced penetrance (only 90% of carriers develop symptoms by age 70 years).¹⁸There are multiple known PGRNmutation carriers who are asymptomatic in their 70s, and at least 1 known affected individual developed symptoms after age 80 years. The clinical features and particularly the syndromic diagnoses have been more variable than in MAPTmutation carriers, with not only behavioral and cognitive features commonly present, but also memory impairment, limb apraxia, parkinsonism, and visuospatial dysfunction, leading to cases being diagnosed with mild cognitive impairment, Alzheimer disease, Parkinson disease, Parkinson disease with dementia, and dementia with Lewy bodies in addition to FTD with or without parkinsonism and 1 of the progressive aphasia syndromes.²⁵The diagnosis of corticobasal syndrome has also been particularly frequent in the cases reported thus far, while no patient with a definite pathogenic PGRNmutation has been reported to date with an ALS phenotype.

As one would expect, based on the clinical features of apraxia and visuospatial dysfunction, greater parietal involvement is clearly present in many PGRNmutation cases, which is also reflected on imaging and pathologic studies. In some cases, rather striking signal changes on magnetic resonance imaging are present,²⁵which is rarely seen in MAPTmutation carriers. Another curious observation is the tendency in some kindreds for the same cerebral hemisphere to be maximally involved in most or all affected members of a family, such as a progressive aphasia syndrome with maximal left hemisphere involvement^22,26,28and the corticobasal syndrome with or without FTD features with maximal right hemisphere involvement^25,29; to our knowledge, this tendency has not been noted among any kindreds with MAPTmutations.

On histologic examination, the consistent finding is FTLD with ubiquitin-positive inclusions with neuronal intranuclear inclusions.^{14,16- 18,20- 26}Immunostaining directed against progranulin stain normal structures within neurons and activated microglia. However, the ubiquitinated inclusions are not progranulin immunoreactive; rather, transactive response DNA–binding protein 43 was very recently discovered to be a major ubiquitinated protein in both neuronal cytoplasmic and intranuclear inclusions in PGRNmutation cases.³⁰Moreover, transactive response DNA–binding protein 43 is also present in neuronal ubiquitin-positive inclusions in FTLD with ubiquitin-positive inclusions, FTLD with motor neuron disease, and idiopathic ALS.³¹

Also contrasting with MAPTis the mechanism of disease with PGRNmutations—all PGRNmutations identified thus far create functional null alleles that cause a partial reduction in progranulin production or haploinsufficiency.^14,16,18This disease mechanism may allow a more straightforward approach for treatment by either replacing progranulin or using drugs to increase production or secretion of progranulin from the remaining normal PGRNallele.

The net effect of 2 genes linked not only by proximity but also by most overlapping and expanding features requires refinements in our conceptual framework and nomenclature in FTDP. An obvious solution to this problem is to simply refine the term by including reference to the genetic cause of the disease in each case, and thus FTDP-17 could be subdivided into FTDP-17 (MAPT) and FTDP-17 (PGRN). This approach has the advantage of employing a now widely used, if not always completely appropriate, clinical terminology, refining it to reflect that ultimately these conditions are defined by their genetics rather than their clinical or pathological phenotypes. The scientific community has clearly just begun to expand the characterization and refine the nomenclature of familial disorders linked to chromosome 17.

Correspondence:Bradley F. Boeve, MD, Department of Neurology, Mayo Clinic, 200 First St SW, Rochester, MN 55905 (bboeve@mayo.edu).

Accepted for Publication:January 14, 2007.

Author Contributions:Study concept and design: Boeve and Hutton. Acquisition of data: Boeve and Hutton. Analysis and interpretation of data: Boeve and Hutton. Drafting of the manuscript: Boeve and Hutton. Administrative, technical, and material support: Boeve and Hutton.

Financial Disclosure:None reported.

Funding/Support:This study was supported by grants AG06786, AG16574, AG11378, and AG07216 from the National Institute on Aging; by the Robert H. and Clarice Smith and Abigail Van Buren Alzheimer's Disease Research Program of the Mayo Foundation; and by the Fund for Scientific Research–Flanders.

Additional Contributions:We thank our many collaborators within and outside the Mayo Foundation, particularly Rosa Rademakers, PhD, for her critical review of this paper, and her and Matt Baker's assistance in providing the sequence chromatograms for Figure 1. We thank the staff of the Mayo Clinic Alzheimer's Disease Research Center for their assistance in characterizing participants, and we particularly thank the members of the many kindreds with MAPTand PGRNmutations for participating in neurodegenerative disease research.