Introduction

Worldwide the prevalence for dementia is still increasing; age-standardized prevalence for those aged >60 years varies around 5–7 % in most world regions, and the prevalence doubles about every 5 years. It is estimated that 35.6 million people were living with dementia worldwide in 2010, with numbers expected to almost double every 20 years, to 65.7 million in 2030 and 115.4 million in 2050 (Prince et al. 2013). Although there seems to be some indication that the incidence of dementia may be decreasing in certain populations (Qiu et al. 2013), we are facing a large societal challenge.

It is known that dementia is caused by an interplay between genes, lifestyle and environmental factors (van der Flier and Scheltens 2005). Recent advances in genetic research have unveiled many potential genetic risk factors for all types of dementia. In this review we will focus on the heritability and genetic risk factors associated with the four most common types of dementia: Alzheimer’s disease (AD), Vascular Dementia (VaD), Frontotemporal Lobar Degeneration (FTLD), Lewy Body dementia (LBD). Among these four types, AD is the most prevalent (34 % before age 65, and 54 % after age 65), see Table 1 for more detailed prevalence number and heritability estimates per dementia subtype (van der Flier and Scheltens 2005; Dementia Report United Kingdom 2007; Prince et al. 2013).

Table 1 Prevalence (among patients with dementia) and heritability estimates per dementia type
Full size table

Gender differences in dementia prevalence clearly exist, while men suffer more from FTLD (before the age of 55), VaD (in between ages of 45 and 75), and LBD (across the age spectrum), women more often suffer from AD (Dementia Report United Kingdom 2007; Savica et al. 2013). Until recently, there was also a discrepancy in prevalence rates between developed and less developed countries, where particularly in South-East Asia VaD used to be the most prevalent dementia subtype. However due to increasing longevity and improved physical health, the prevalence of AD, with typically a later age of onset than VaD, increased in line with the aging population. Better physical health reduces cerebrovascular disease and hence the numbers with VaD. These changes also tend to shift the sex ratio towards a preponderance of female cases (Dementia Report United Kingdom 2007).

The heritability rates of the different dementia subtypes ranges from 40 to 80 % with early-onset (e.g. before the age of 65) dementia having a bigger genetic component than late-onset dementia (van der Flier and Scheltens 2005). Nevertheless, for all subtypes several genetic risk factors have already been identified and those will be covered in the next paragraphs.

Alzheimer’s Disease

AD is the most common neurodegenerative disorder and it is clinically defined by a slowly progressing loss of cognitive functions, primarily memory impairment. Neuropathologically the disease is characterized by the aggregation and deposition of beta amyloid (Aβ) peptide in the form of neuritic plaques and hypophosphorylated tau protein in the form of intracellular neurofibrillary tangles.

Particularly the early-onset form of AD (EOAD) is thought to be highly heritable and to run in families, whereas the late-onset variant (LOAD) takes on a more sporadic form. However, lately several studies have been able to find many genetic risk factors for LOAD as well.

For EOAD three major genetic loci have been identified: APP (amyloid precursor protein) on chromosome 21q, PSEN1 (presenilin 1) on chromosome 14q, and PSEN2 (presenilin 2) on chromosome 1q. Within these three genes over 200 mutations have been found (see the AD&FTD mutation database: http://molgen.ua.ac.be-admutations/).

These discoveries have led to the so called amyloid hypothesis, which postulates that the abnormal production of Aβ is the initial step in triggering the pathophysiological cascade that eventually leads to AD (Glenner and Wong 1984) and that other neuropathological hallmarks of AD, such as hyperphosphorylated tau and neurofibriliary tangles, vascular damage and inflammation are consequences rather than separate causes of the disease process. Clinically and neuropathologically EOAD and LOAD are much alike (Rosenthal and Kamboh 2014). Although LOAD is much more common than EOAD; it has been estimated that 99 % of all AD cases are LOAD cases (Karch et al. 2014).

Many candidate gene studies on LOAD focused on genes that code for potential proteins that play a role in Aβ production. One of the best examples of that is the gene coding for apolipoprotein E (APOE), on chromosome 19q13, which was actually already discovered in linkage studies (Pericak-Vance et al. 1991). The APOE gene has three alleles; the ε2, ε3, and ε4 alleles. Of these the ε4 allele is associated with a 4-fold greater risk for LOAD than the ε3 allele (Genin et al. 2011). The less common ε2 allele appears to have some sort of protective effect and is associated with longevity (Olesen et al. 1997; Deelen et al. 2014). Some studies suggest that the APOE e4 allele does not explain all of the genetic risk for AD of this chromosomal region. Two other possible candidates that are in close proximity to APOE have therefore been suggested to also increase the risk for AD; TOMM40 (encoding translocase of outer mitochondrial membrane 40 homolog;(Roses et al. 2010; Takei et al. 2009)) and EXOC3L2 (exocyst complex component 3-like-2;(Seshadri et al. 2010)). These suggest that other biological mechanisms may be at play including mitochondrial dysfunction (Ferencz et al. 2012).

Of the hundreds of possible candidate genes that were identified based on the amyloid hypothesis only a few have consistently been shown to have a modest association with AD (odds ratios between 1.15 and 1.5 and p-value between 0.0001 and 0.01) (Karch and Goate 2014). These candidate genes include ACE (angiotensin-converting enzyme; (Kehoe et al. 1999)), ADAM10 (disintegrin and metalloproteinase domain-containing protein 10; (Kim et al. 2009a)), CHRNB2 (cholinergic receptor, nicotinic, beta2;(Cook et al. 2004)), DAPK1 (death-associated protein kinase 1;(Y. Li et al. 2006)), IL8 (interleukin 8;(Li et al. 2009)), MTHFR (methylenetetrahydropholate reductase;(Chapman et al. 1998)) and SORL1 (sortilin-related receptor;(Rogaeva et al. 2007)).

More recently, several genome-wide studies have been conducted to find new genetic risks for AD. From all amyloid-related candidate genes only APOE reached significance on a genome-wide level (with p-values down to 1 × 10−160; (Harold et al. 2009)). See Table 2 for the top genetic markers for AD that have all been identified in large genome-wide association studies (Lambert et al. 2009, 2013; Hollingworth et al. 2011; Naj et al. 2011; Seshadri et al. 2010; Harold et al. 2009). Pathway analysis has shown that many of these top genes play a role in the following possible pathophysiological mechanisms: cholesterol metabolism, immune response and endocytosis.

Table 2 Main genetic risk factors for Alzheimer’s disease
Full size table

Cholesterol Metabolism

There is a strong association between cholesterol and AD risk. For instance, high levels of cholesterol appear to accelerate the formation of beta-amyloid plaques and epidemiological studies showed increased risk of AD in older individuals with dyslipidaemia (Reitz 2013).

APOE plays a central role in cholesterol metabolism; APOE protein is involved in lipid transport and is highly expressed in the central nervous system and the liver. It binds to Aβ and influences the clearance of soluble Aβ and the Aβ aggregation (Leduc et al. 2010).

The ABCA7 (ATP-binding cassette sub-family A member 7) gene is part of the ATP binding transport family, which includes several hundreds of membrane transport proteins that utilize the energy of ATP to transport a specific group of ligands across the cell membrane. As such ABCA7 is responsible for lipid transport. ABCA7 also participates in macrophage uptake of Aβ, and removal of ABCA7 results in increased levels of insoluble Aβ and it has been shown to mediate APP processing (Rosenthal and Kamboh 2014).

The gene CLU encodes for clusterin, which is also mainly involved in clearance and aggregation of Aβ (DeMattos et al. 2002, 2004). Clusterin has also been suggested to play a role regulation of brain cholesterol and lipid metabolism, and the inhibition of neuronal apoptosis/potentiation of neuroprotection (Nuutinen et al. 2009) and oxidative stress (Trougakos and Gonos 2006). Clusterin also plays a key role in protecting the immune system from the detrimental effect of Aβ aggregation, which is an activator of the complement system (Hardy et al. 2011).

Immune System

Inflammatory involvement in AD has already long been hypothesized to be one of the underlying pathophysiological mechanisms of the disease (Ridolfi et al. 2013). Inflammatory processes may be neurotoxic (Bates et al. 2009) and markers of inflammation have been shown to be associated with amyloid plaques (McGeer and McGeer 2001) Moreover, it was found that patients on NSAIDS have a lower risk to develop AD (Ridolfi et al. 2013).

The CR1 gene codes for the complement receptor 1 (CR1) is the main receptor of the complement C3b protein and is a regulator of complement activation (RCA). As such it has numerous functions in the immune system (Khera and Das 2009). Activation of this receptor can protect against Aβ-induced neurotoxicity and may reduce the accumulation or promote the clearance of amyloid and degenerating neurons (Rogers et al. 2006).

CD33 (Siglec-3) belongs to a class of immune cell surface receptors called sialic acid-binding immunoglobulin-like lectins. CD33 triggers immune cell–cell interactions through endocytosis. It has been shown that CD33 expression is increased in AD brains as is the number of CD33-positive microglia (Karch et al. 2012).

Endocytosis

Endocytosis is the process by which cells absorb molecules from outside the cell that are otherwise too large for them. Endocytosis regulates many processes, such as nutrient uptake, cell adhesion and migration, growth, neurotransmission and drug delivery. The following genes are thought to play a role in endocytosis and AD.

PICALM (Phosphatidylinositol binding clathrin assembly protein) plays a role in clathrin-mediated endocytosis, synaptic transmission, and the removal of apoptotic cells (Harel et al. 2008) . In the brain the PICALM protein is predominantly expressed in endothelial cells where it could play a role in Aβ transport into the bloodstream (Baig et al. 2010). PICALM may thus be involved in Aβ clearance, which is supported by the finding that individuals carrying the risk allele of the PICALM top SNP show lower levels of Aβ in cerebrospinal fluid (Schjeide et al. 2011).

BIN1 (Bridging integrator 1) encodes several isoforms of an adaptor protein involved in receptor mediated endocytosis (Pant et al. 2009). Additionally, it could indirectly also have an effect on Aβ production and/or the clearance of Aβ from the brain.

The neuronal sorting receptor SORL1 (sortilin receptor 1) is involved in the trafficking of APP (Rogaeva et al. 2007) and has also been suggested to bind with lipoproteins and mediate endocytotic processes.

Taken together these confirmed AD risk genes explain about 7–8 % of the genetic variance in AD, and it has been estimated that 5.8 % can be attributed solely to the APOE gene. Whereas, in total it is estimated that about 33 % of the phenotypic variance in AD can be explained by genetics (Ridge et al. 2013). This suggests that besides the APOE gene, all other genetic risk factors identified so far have a very small effect and more importantly that there are still many risk genes to be discovered.

Vascular Dementia

Subcortical vascular dementia or cerebral small vessel disease is a common cause of disability in the elderly. A clinical diagnosis of vascular dementia (VaD) is based on a clinical representation of dementia, a history of ischemic or hemorrhagic vascular disease or hypoperfusive ischemic cerebral infarcts and a close temporal relation between the cerebrovascular disease and onset of dementia symptoms (Roman et al. 1993). Compared to AD, VaD is more often characterized by attention and concentration deficits, bradyphrenia and word finding problems. However the symptoms largely rely on the location of cerebrovascular disease and hence it is a very heterogeneous disease (Pohjasvaara et al. 2003).

A genetic basis for VaD is difficult to investigate, because of the variety in clinical presentations of the disorder. Most patients with VaD have the sporadic form, which has a very low heritability (<1 %) (Bergem et al. 1997; Gatz et al. 2010). However there are two highly heritable forms of VaD; cerebral autosomal dominant arteriopathy with subcortical infarct and leukoencephalopathy (CADASIL) and familial cerebral amyloid angiopathy (FCAA).

CADASIL is the most prominent inherited form of vascular dementia, which is an autosomal dominant disease that is rare with a prevalence of only 2 in 100,000. The main clinical features include migraine with aura, stroke, mood disturbances, and cognitive decline, with a mid-life (30s–60s) adult onset (Federico et al. 2005). Genetic testing is the gold standard for the diagnosis. CADASIL is caused mostly by missense mutations in the NOTCH3 (Neurogenic locus notch homolog protein 3) gene (Federico et al. 2005). NOTCH3 is a receptor located on the wall of vascular smooth vessel cells, as such Disruption of Notch3 functioning can lead to the self-destruction (apoptosis) of these cells. Damage to vascular smooth muscle cells is thought to cause recurrent strokes and VaD (Wang et al. 2008; Schmidt et al. 2012).

FCAA is also an autosomal dominant disease that is characterized by recurrent stroke, movement impairments, and VaD with memory loss and personality changes. As for CADASIL it is a rare disorder and so far exact prevalence rates are unknown. Like other cerebrovascular disorders, FCAA usually occurs in mid-life (between ages 55 and 65) (Maia et al. 2007). Mutations in the APP (amyloid precursor protein), CST3 (Cystatin 3), or ITM2B (Integral membrane protein 2B) genes lead to the aggregation of proteins, which form amyloid plaques in the vessel walls. The amyloid plaques damage brain cells, eventually causing cell death and impairing various parts of the brain (Federico et al. 2012).

Several candidate genes have already been identified for sporadic VaD and most of the genes play a role in lipid metabolism (MMP1;MMP3;MMP9;APOE;PON1;PON2;VLDLR;SREBF2), angiotensin (AGT; ACE) and inflammation (IL1A; IL1B; IL6; TNF;TGFB1) (for extensive review see (Lee and Kim 2013)). However most of these genetic findings have not been replicated and/or show only weak associations (0.01 < p < 0.05).

So far the strongest evidence comes from studies combining multiple genetic loci, suggesting that there is some epistasis going on. For instance, a combined genotype of the two genes TNF β1 (tumor necrose factor receptor b1) and AGT (angiotensin) showed a stronger association with VaD than their individual genotypes (Kim et al. 2009b). Furthermore a strong synergistic epistasis was also suggested between the genes FAM134B (family with sequence similiarity 134, variant B) and TNFRSF19 (tumor necrosis factor receptor, super family, member 19) (Kong et al. 2011).

Recently two genome-wide association studies (GWAS) on sporadic VaD have been undertaken (Schrijvers et al. 2012; Kim et al. 2013). The study with 84 patients and 200 controls from Korea did not result in any genome-wide significant hit (Kim et al. 2013), but there was a suggestive hit, that could be replicated using gene-wide analysis in the SYK (spleen tyrosine kinase) gene. Whereas the other study among 67 patients and 5700 controls did identify a novel locus on chromosome X near the AR (androgen receptor) gene. Both these findings do not seem to be related to any of the candidate pathways, such as lipid metabolism or inflammation.

Frontotemporal Lobar Degeneration

Frontotamporal lobar degeneration (FTLD) is an anatomopathological descriptive term referring to a cluster of disorders that are characterized by the atrophy of the frontal and anterior temporal lobes. Apart from the overlap in atrophy, FTLD disorders are a clinically, genetically and pathologically heterogeneous group of disorders.

FTLD is the second most common form of early-onset dementia, after AD. In approximately 75 % of the patients the onset of FTLD occurs before the age of 65 years. In the age group between 45 and 65 years, the prevalence of FTLD is estimated to be between 10 and 30 per 100,000 (Dementia Report United Kingdom 2007).

FTLD can manifest as two clinically distinct subtypes: a behavioral subtype and a language disturbances subtype (Karageorgiou and Miller 2014). The behavioral subtype of FTLD (bvFTLD) is characterized by severe changes in behavior and personality such as disinhibition, apathy, loss of empathy, stereotypic behavior and loss of social competence. Cognitively, executive functions are impaired, while during the early stages of the disease memory and visuospatial skills are well preserved. bvFTLD accounts for more than 50 % of the FTLD patients. The onset is typically before the age of 65, with an average onset of 58 years (Kirshner 2014).

The language subtype of FTLD is called primary progressive aphasia (PPA) and consists of a nonfluent/agrammatic variant and a semantic variant. Nonfluent/agrammatic PPA is characterized by effortful speech and grammatical errors, with relatively preserved language comprehension. While semantic PPA is characterized by impaired comprehension and conceptual knowledge with concomitant development of anomia, while speech production is spared. Each PPA subtype accounts for approximately 20–25 % of FTLD patients (Kirshner 2014).

Based on family studies it has been estimated that FTLD has a heritability of 30–50 %. Heritability varies across the different subtypes of FTLD (Rohrer et al. 2009).

In the following section we will discuss the most consistent genetic findings for FTLD. The number of mutations found in these genes and it’s frequencies are described in Table 3. The majority of FTLD cases seems to be related to mutations in two genes; MAPT (the microtubule-associated protein tau;(van Swieten and Spillantini 2007)) and PGRN (progranuline;(Baker et al. 2006)). Other genes that frequently have been shown to play a role in FTLD are CHMP2B (charged multivesicular body protein 2B,(Skibinski et al. 2005)), VCP (valosin containing protein;(Weihl et al. 2008)) and C9orf72 (Gijselinck et al. 2012) (See Table 4.).

Table 3 Main genetic risk factors FTLD
Full size table
Table 4 Main genetic risk factors for Lewy body dementias
Full size table

The neuropathologic subtype that is characterized by tau-positive pathologic inclusion is caused by mutations in the MAPT gene. Over 70 variants of MAPT have been identified that lead to deletions, missense, silent and splice cite mutations (http://www.molgen.ua.ac.be-FTDmutations). These mutations lead to tau pathology due to loss of function of the microtubules or from toxic gain of function (e.g. increased aggregation properties). Tau inclusions can usually be found in neurons and glial cells of the frontal and temporal cortex and the hippocampus. The mean onset of FTLD in patients with MAPT mutations is around 55 years of age. Clinically most MAPT mutations lead to bvFTLD (Ferrari et al. 2011). Besides mutations, there are also some functional haplotypes of the MAPT gene, but these are mostly related to other diseases than FTLD, including Parkinson’s Disease (PD), AD and amyotrophic lateral sclerosis (Ferrari et al. 2011).

The other common neuropathologic subtype of FTLD is characterized by TAR DNA-binding positive of 43 kDa (TDP-43) inclusions and has been associated with mutations in the PGRN gene as well as rarer mutations in other genes such as TARDBP (a gene coding for TDP-43) (Sieben et al. 2012). PGRN is involved in several biological processes such as cell cycle progression, cell growth regulation and inflammation. Currently over 140 variants of the PGRN gene have been reported in FTLD (http://www.molgen.ua.ac.be-FTDmutations). Clinically, individuals carrying PGRN mutations show predominantly apathy, language impairment and psychotic symptoms. The mean onset of the disease is about 60 years of age. Neuronal cell loss is usually apparent in the frontal, temporal and inferior parietal cortices of the dominant hemisphere. PGRN mutations are also frequently found in other neurodegenerative disorders, such as AD, PD and corticobasal syndrome (Le Ber 2013).

In FTLD with TDP-43-related pathology a genetic association has also been found with the gene C9orf72 (chromosome 9 open reading frame 72). Clinically, most cases are bvFTLD. The mean onset of the disease is around 55 years of age and a higher number of repeat expansions is associated with earlier age of onset.

The C9orf72 gene mainly plays a role in FTLD with motor neuron disease and FTLD with ALS (Devenney et al. 2014). The gene encodes for a protein with unknown function. Pathologically C9orf72-related FTLD has besides TDP-43pathology also neuronal cytoplasmic inclusions (NCI). These NCI are seen in the granular cells of the hippocampus and in patients with motor neuron disease (MND) also in the hypoglossal nuclei and motor neurons of the ventral horn of the spinal cord (van der Zee et al. 2013).

VCP (Valosin-containing protein) is involved in several cellular processes like cell cycle regulation, suppression of apoptosis, DNA damage response and protein degradation (Ferrari et al. 2011). Mutations in the VCP gene cause inclusion body myopathy (IBM) associated with Paget’s disease of the bone (PGD) and FTLD (Weihl et al. 2008). This combination, IBMPFD, is an autosomal dominant disease. IBMFPD usually presents it-self clinically first by development of disabling weakness with a mean onset of 45 years of age, second patients have osteolytic lesions and third language and behavioral deficits become apparent, which usually happens around age 55. Neuropathologically, IBMPFD is characterized by tau-negative and ubiquitin positive inclusions.

CHMP2B (Charged multivesicular body protein 2b) plays a role in endosomal trafficking, cellular signaling and autophagy. Genetic variability in CHMP2B is found in a small proportion of FTLD cases. Clinically, patients usually present with bvFTLD, starting with early personality changes, aphasia characterized by a reduction in spontaneous speech, leading eventually to mutism. The average age of onset is around 58 years of age (Momeni et al. 2006). Neuropathologically, enlarged vacuoles in cortical neurons in the frontal temporal parietal and occipital cortices have been found (Urwin et al. 2010).

So far two GWA on FTLD have been published (Van Deerlin et al. 2010; Ferrari et al. 2014). In the first study by van Deerlin and colleagues, patients with TDP-43 neuropathology were investigated and a genome-wide significant hit was found for a locus near the gene TMEM106B (Transmembrane protein 106B) on chromosome 7p21.

The second GWA study reported a genome-wide significant hit on chromosome 6 near HLA (human leucocyte antigen) locus for three SNP’s (rs9268877, rs9268856, rs1980493). The HLA genes play a role in the immune system. In total there are about 200 HLA genes that are all situated near each other on chromosome 6. The HLA complex helps the immune system distinguish the body’s own proteins from proteins made by foreign invaders such as viruses and bacteria. For the behavioral variant of FTLD they also found a new locus on chromosome 11 near the RAB38 (Ras-related protein ) and CTSC (Cathepsin) genes. This latter locus plays a role in lysosomal biology and is as such involved in inflammatory processes. Interestingly, these GWAS study did not find evidence for any of the previously identified candidate genes, suggesting that FTLD is a heterogenic disease that has different genetic underpinnings among families.

Lewy Body Dementia

Lewy body dementia (LBD) is one of the most common forms of dementia with a prevalence rate of 4 % in the general population (6 % for men and 3 % for women) (Dementia Report United Kingdom 2007). LBD encompasses Parkinson disease dementia (PDD) and dementia with Lewy bodies (DLB) (McKeith et al. 1996; Hanson and Lippa 2009). It has been estimated that 25 to 50 % of patients with LBD have symptoms of PD. The symptoms of DLB include sleep problems, hallucinations (mostly visual), deficits in executive function, attention, memory domains and visuo-spatial ability (Kehagia et al. 2010), and are accompanied within the first year by Parkinsonian motor symptoms. In PDD, cognitive deficits and/or dementia are secondary to parkinsonism with motor symptoms present for a minimum of 1 year (“1 year rule” for clinical studies) (Mollenhauer et al. 2010). If the disease is in its full-blown stage then usually no differentiation is made between the two disease entities exist, neither clinically nor neuropathologically. LBD is associated with higher rate of mortality and is considered more costly to society than AD (Oesterhus et al. 2014; Bostrom et al. 2007; Hely et al. 2008), making it imperative to find the etiology of the disease. A recent twin study did however not show a strong support for a genetic contribution to LBD (C. S. Wang et al. 2009). Yet, others have demonstrated that LBD aggregates in families and may have an autosomal inheritance pattern (Nervi et al. 2011; Bogaerts et al. 2007). A few consistent genetic markers have been identified for LBD (Table 3), and these will be discussed in the next section.

Mutations in the synuclein family have been known to play a role in PD, and recently they have received attention in LBD as well (Meeus et al. 2012; Fuchs et al. 2007). A duplication in the a-synyclein gene (SNCA) was recently reported in a DLB patient, but these type of mutations are rare (Meeus et al. 2012). Although duplications may be rare, association studies have found SNPs within α-, β-, and γ-synuclein genes to be associated with increased risk of DLB (Nishioka et al. 2010; Bras et al. 2014). It is not surprising that these genes may play a significant role in LBD as synucleins are one of the major components of Lewy bodies, a hallmark of the disease (George 2002; Miklya et al. 2014). Yet, their role remains to be determined, as some SNCA SNPs are not associated with Lewy body pathology (Wider et al. 2012), and others are associated with Lewy body pathology, but in AD (Linnertz et al. 2014).

Mutations in glucocerebrosidase (GBA), such as N370S and L444P, on the other hand are more common in LBD. Pathogenic mutations have been observed in up to 7.6 % of DLB cases (Tsuang et al. 2012; Mata et al. 2008), and found to increase the risk of DLB (Tsuang et al. 2012). Associated with the presence of cortical Lewy bodies (Clark et al. 2009) the specificity to LBD is more pronounced with GBA, as mutations in this gene are not related to AD type pathology (Clark et al. 2009), nor are they observed frequently in AD (Tsuang et al. 2012). Lysosomal dysfunction has been highlighted as a potential hallmark of PD. As such, GBA may play a prominent role as it encodes the lysosomal enzyme glucocerebrosidase (Yang et al. 2014). In conjunction with recent findings of an association between scavenger receptor class B member 2 (SCARB2) and DLB lysosomal dysfunction may play a vital role in the pathogenesis of LBD. Further studies will need to confirm this association (Bras et al. 2014).

Another prominent player in LBD is the MAPT H1 haplotype. Suggested to regulate tau-transcription the MAPT gene is one of the strongest predictors of dementia in PD (Williams-Gray et al. 2013; Seto-Salvia et al. 2011). The H1 haplotype is also associated with postural instability (Williams-Gray et al. 2013) a known risk factor for PDD (Moore and Barker 2014). None of the MAPT haplotypes have any influence in AD or DLB (Seto-Salvia et al. 2011), suggesting that the influence is specific to PDD. The MAPT gene has also been found to regulate α-synuclein in the brainstem (Colom-Cadena et al. 2013), middle frontal and inferior parital cortex (Wider et al. 2012), supporting a link between MAPT haplotypes and synucleinopathies.

The leucine rich repeat kinase 2 (LRRK2) gene plays an important role in the formation of Lewy bodies (Zhu et al. 2006), yet its role in LBD remains to be elucidated. Meeus and colleagues (Meeus et al. 2012) reported an LRKK2 (R1441C) mutation in a patient with PDD while a large case–control study was not able to show a strong association between LRKK2 and DLB (Bras et al. 2014). Further studies should consider potential interactions with SNCA, as LRKK2 was found to interact with SNCA in relation to Lewy body pathology (Linnertz et al. 2014).

To date only one genome-wide linkage study has been performed among patients with familial LBD. A locus on chromosome 2 q35-q36 was identified, however none of the genes in this region could explain the relation with LDB (Bogaerts et al. 2007).

Some of the genes commonly associated with AD have been implicated in LBD. Novel presenelin mutations in PSEN1 A79 and PSEN2 R71W were recently observed in DLB patients (Meeus et al. 2012). These mutations could potentially lead to α-synuclein accumulation as seen in mouse models for PSEN1 (Winslow et al. 2014). Nevertheless, these mutations can result in different clinical presentations, suggesting that differential diagnosis between AD, PDD and DLB is not feasible based on these genes (Meeus et al. 2012). APOE, one of the primary genetic risk factors for AD has also been associated with LBD. In a recent large case–control study with up to 788 DLB cases, APOE was the strongest genetic risk factor for DLB followed by SNCA (Bras et al. 2014). However, a recent large meta-analysis has failed to find convincing support for an APOE influence in PDD (Wider et al. 2012).

Overall there seem to be several prominent players when it comes to the genetic influence in LBD. Yet, it becomes evident from reviewing the literature that larger studies are warranted, as the field remains lagging behind the AD/PD field.

Summary and Conclusions

Most dementia subtypes are complex heterogenic disorders that are not caused by one sole dominant genetic mutation, but are rather the result of the interplay between environment and several genes. Moreover, the most important risk factor for all types of dementia is increasing age; with the increase of 10 years, from 65 to 75 years of age, resulting in an increased risk of dementia that’s beyond of what is conveyed by any of the described genetic risk alleles.

It should be noted that many patients do not have a clear clinical presentation, because frequently patients show some overlap between several subtypes of dementia and sometimes the differential diagnosis may be rather difficult and somewhat arbitrary as seen in LBD. The most common example of this is the so-called mixed type dementia, which is a combination of Alzheimer’s disease and vascular dementia. Patients usually show typical AD neuropathology accompanied by a vascular component (Lathe et al. 2014; Magaki et al. 2014). Given that vascular risk factors, such as hypertension, dyslipidemia and diabetes are risk factors for both types of dementia, also suggests that there may be some common pathophysiology (Altman and Rutledge 2010).

Figure 1 shows the overlap in the major underlying pathology of the four dementia subtypes. Overlap in genetic risks is seen in FTLD and LBD, where genetic mutations in the MAPT gene both play a role. Indeed tau-pathology plays a role in both disorders and there may even be some overlap in α-synuclein involvement (Kara et al. 2014). Tau-pathology is a candidate in all of the described dementia subtypes, but it should however be noted that many different forms of tau exist and that it depends on the dementia subtype which form of tau is (for extensive review please see (Tang et al. 2014; Spillantini and Goedert 2013)).

Fig. 1
figure 1

Overlap in major underlying pathologies of dementia based on their genetic associations

Full size image

Similarly, overlap between LBD and AD has also been suggested, as mutations in APP, PSEN1 and PSEN2 are found in both disorders. Moreover, comorbid LBD and AD is quite common (Magaki et al. 2014). It is suggested that there is some sort of spectrum between AD and PDD, with DLB in the middle of these two disorders, as both Aβ and α-synuclein play a role in DLB, while Aβ involvement in PDD is rare, as is α-synuclein in AD (Meeus et al. 2012).

As can be seen neuroinflammation is considered to play a role in all types of dementia. It is most likely not the main cause of neurodegeneration, but it could serve as a catalyst of other ongoing neurodegenerative processes (Jones et al. 2010; Ridolfi et al. 2013). Whereas cholesterol metabolism is thought to be mostly related to the onset of AD and VaD (Reitz et al. 2004), and much less to the other subtypes. Two recent studies even showed that cholesterol does not seem to play a role in onset of PDD and LBD (Gudala et al. 2013; Borroni et al. 2006).

Indeed a recent GWAS study suggests that there is some overlap in the neuropathological features of AD and related dementias (Beecham et al. 2014) The authors related AD loci with both core and comorbid neuropathological features of AD and found that core features such as neurofibrillary tangles and neuritic plaques were associated with ABCA7, BIN1, cASS4, MEF2C and PICALM, while comorbid neuropathological features such as lewy body disease and hippocampal sclerosis were associated with SORL1 and PTK2B respectively. APOE was associated with both core and comorbid neuropathological features, including neurofibrillary tangles and neuritic plaques, lewy body disease, cerebral amyloid angiopathy (CAA), but not hippocampal sclerosis and vascular brain injury. As such, it seems that AD loci are associated not only with core neuropathological features of AD, but also comorbid neuropathological features.

In this review we focused on genetic mutations that can increase the risk for dementia, whereas it would also be of interest to examine genetic mutations that can protect an individual for the onset of dementia or that may interact beneficially with lifestyle factors. Recently, a GWA study among Icelandic people showed that a rare mutation in the APP gene decreases the risk for dementia and AD in particular (Jonsson et al. 2012). Also, a GWA study on longevity identified some new loci that may play a role in healthy aging (Deelen et al. 2014). These new protective genetic variations may be likely candidates to study the prevention of dementia.

Finally, because dementia is not caused by only genetics, more studies are needed that investigate epistasis and gene by environment interaction. So far some attempts have been made (Gusareva et al. 2014; Alam et al. 2014; Kong et al. 2011), but these studies require replication.