Summations
-
1,242 differentially expressed genes including 70 microRNAs have been reported in people with LBD.
-
Several mitochondrial genes and genes involved in ubiquitin proteasome system and autophagy–lysosomal pathway were significantly downregulated in people with LBD.
-
Our functional analyses highlighted the importance of RNA-mediated gene silencing, neuregulin signalling, and neurotrophic factors in LBD molecular pathology.
Considerations
-
This systematic review has excluded studies that were not published in English. It did not include studies that investigated animal models or cell lines.
-
All included studies were small cross-sectional studies, and there was substantial heterogeneity among the included studies.
-
Majority of the included studies have employed relative quantification methods, so it was not possible to do combined analyses using their findings.
Introduction
Lewy body dementia (LBD) includes two overlapping clinical syndromes: dementia with Lewy bodies (DLB) and Parkinson’s disease dementia (PDD) (McKeith et al., Reference Mckeith, Boeve, Dickson, Halliday, Taylor, Weintraub, Aarsland, Galvin, Attems, Ballard, Bayston, Beach, Blanc, Bohnen, Bonanni, Bras, Brundin, Burn, Chen-Plotkin, Duda, EL-Agnaf, Feldman, Ferman, Ffytche, Fujishiro, Galasko, Goldman, Gomperts, Graff-Radford, Honig, Iranzo, Kantarci, Kaufer, Kukull, Lee, Leverenz, Lewis, Lippa, Lunde, Masellis, Masliah, Mclean, Mollenhauer, Montine, Moreno, Mori, Murray, O’brien, Orimo, Postuma, Ramaswamy, Ross, Salmon, Singleton, Taylor, Thomas, Tiraboschi, Toledo, Trojanowski, Tsuang, Walker, Yamada and Kosaka2017). DLB is the second most common neurodegenerative dementia and its incidence rate has been estimated as 112 per 100,000 for those over 65 years of age (Perez et al., Reference Perez, Helmer, Dartigues, Auriacombe and Tison2010; Walker et al., Reference Walker, Possin, Boeve and Aarsland2015). LBD leads to increased mortality (Oesterhus et al., Reference Oesterhus, Soennesyn, Rongve, Ballard, Aarsland and Vossius2014), earlier nursing home admissions, more frequent falls, worse quality of life, higher costs (Vossius et al., Reference Vossius, Rongve, Testad, Wimo and Aarsland2014), and more caregivers’ burden than Alzheimer’s disease (AD). DLB is underdiagnosed in many clinical settings and nearly 50% of people with DLB reportedly remain misdiagnosed in the UK (Freer, Reference Freer2017). Missing the diagnosis of LBD and treating associated neuropsychiatric symptoms with antipsychotic medications risk life-threatening adverse effects, such as neuroleptic malignant syndrome in people with LBD. Currently, we do not have a reliable biological fluid-based diagnostic biomarker or a disease-modifying drug for LBD. Improving our knowledge of molecular mechanisms underlying neurodegeneration in LBD is essential for discovering reliable diagnostic biomarkers and novel therapeutic targets for LBD (Walker et al., Reference Walker, Possin, Boeve and Aarsland2015; Velayudhan et al., Reference Velayudhan, Ffytche, Ballard and Aarsland2017).
Although the onset of most LBD appear sporadic, several studies have reported familial aggregation of LBD and its core clinical features (Tsuang et al., Reference Tsuang, Digiacomo and Bird2004; Nervi et al., Reference Nervi, Reitz, Tang, Santana, Piriz, Reyes, Lantigua, Medrano, Jimenez-Velazquez, Lee and Mayeux2011). The heritability of DLB has been estimated as 59.9%, and the genetic risk factors for DLB are likely to be independent from known risk variants of AD and Parkinson’s disease (PD) (Guerreiro et al., Reference Guerreiro, Escott-Price, Hernandez, Kun-Rodrigues, Ross, Orme, Neto, Carmona, Dehghani, Eicher, Shepherd, Parkkinen, Darwent, Heckman, Scholz, Troncoso, Pletnikova, Dawson, Rosenthal, Ansorge, Clarimon, Lleo, Morenas-Rodriguez, Clark, Honig, Marder, Lemstra, Rogaeva, ST George-Hyslop, Londos, Zetterberg, Barber, Braae, Brown, Morgan, Troakes, AL-Sarraj, Lashley, Holton, Compta, Van Deerlin, Serrano, Beach, Lesage, Galasko, Masliah, Santana, Pastor, Diez-Fairen, Aguilar, Tienari, Myllykangas, Oinas, Revesz, Lees, Boeve, Petersen, Ferman, Graff-Radford, Cairns, Morris, Pickering-Brown, Mann, Halliday, Hardy, Trojanowski, Dickson, Singleton, Stone and Bras2019). Two genome-wide association studies (GWAS) (Guerreiro et al., Reference Guerreiro, Ross, Kun-Rodrigues, Hernandez, Orme, Eicher, Shepherd, Parkkinen, Darwent, Heckman, Scholz, Troncoso, Pletnikova, Ansorge, Clarimon, Lleo, Morenas-Rodriguez, Clark, Honig, Marder, Lemstra, Rogaeva, ST George-Hyslop, Londos, Zetterberg, Barber, Braae, Brown, Morgan, Troakes, AL-Sarraj, Lashley, Holton, Compta, Van Deerlin, Serrano, Beach, Lesage, Galasko, Masliah, Santana, Pastor, Diez-Fairen, Aguilar, Tienari, Myllykangas, Oinas, Revesz, Lees, Boeve, Petersen, Ferman, Escott-Price, Graff-Radford, Cairns, Morris, Pickering-Brown, Mann, Halliday, Hardy, Trojanowski, Dickson, Singleton, Stone and Bras2018; Rongve et al., Reference Rongve, Witoelar, Ruiz, Athanasiu, Abdelnour, Clarimon, Heilmann-Heimbach, Hernandez, Moreno-Grau, DE Rojas, Morenas-Rodriguez, Fladby, Sando, Brathen, Blanc, Bousiges, Lemstra, Van Steenoven, Londos, Almdahl, Palhaugen, Eriksen, Djurovic, Stordal, Saltvedt, Ulstein, Bettella, Desikan, Idland, Toft, Pihlstrom, Snaedal, Tarraga, Boada, Lleo, Stefansson, Stefansson, Ramirez, Aarsland and Andreassen2019) and a genome-wide analysis of copy number variants (Kun-Rodrigues et al., Reference Kun-Rodrigues, Orme, Carmona, Hernandez, Ross, Eicher, Shepherd, Parkkinen, Darwent, Heckman, Scholz, Troncoso, Pletnikova, Dawson, Rosenthal, Ansorge, Clarimon, Lleo, Morenas-Rodriguez, Clark, Honig, Marder, Lemstra, Rogaeva, ST George-Hyslop, Londos, Zetterberg, Barber, Braae, Brown, Morgan, Troakes, AL-Sarraj, Lashley, Holton, Compta, Van Deerlin, Serrano, Beach, Lesage, Galasko, Masliah, Santana, Pastor, Diez-Fairen, Aguilar, Tienari, Myllykangas, Oinas, Revesz, Lees, Boeve, Petersen, Ferman, Escott-Price, Graff-Radford, Cairns, Morris, Pickering-Brown, Mann, Halliday, Hardy, Trojanowski, Dickson, Singleton, Stone, Guerreiro and Bras2019) have investigated the genetic associations of DLB. There has not been any specific GWAS investigating people with PDD. Genetic associations between LBD and the variants in APOE, GBA, SNCA, and MAPT have been replicated by two or more studies. Other reported genetic associations of LBD that need further replication include the variants in ADGRG7, BCHE-K, BCL7C, CHRFAM7A, CNTN1, GABRB3, LAPTM4B, mtDNA, MSR1, NME1, NME2, NOS2A, PDZD2, PSEN1, SCARB2, SPAG9, TFG, TREM2, UCHL1, and ZFPM1.
Genetic association studies cannot clarify the functional implications of the identified genetic associations. Gene expression studies investigating ribonucleic acid (RNA) levels (Segundo-Val & Sanz-Lozano, Reference Segundo-Val and Sanz-Lozano2016) are necessary for quantifying transcriptional changes and for understanding the effects of gene expression regulation and alternative splicing. Unlike genetic associations, gene expression changes are dynamic and tissue-specific. As gene expression differs with disease progression, gene expression studies help staging diseases and identifying RNA-based therapeutic targets (Harries, Reference Harries2019). Gene expression studies in people with LBD have highlighted the importance of alternative splicing of α-synuclein in the molecular pathogenesis of LBD. Increased expression of α-synuclein-112 and decreased expression of α-synuclein-126 in the post-mortem frontal cortices of people with LBD have been reported (Beyer et al., Reference Beyer, Humbert, Ferrer, Lao, Carrato, Lopez, Ferrer and Ariza2006). Gene expression profiling of post-mortem LBD brains and biological fluids of people living with LBD can advance our molecular-level mechanistic understanding of neurodegeneration in LBD. This will facilitate identifying reliable diagnostic biomarkers and novel therapeutic targets for LBD. Considering the need for a comprehensive summary of all available evidence from the gene expression studies in people with LBD, we aimed to conduct the first systematic review on this topic.
Materials and methods
Study design
The protocol for this systematic review has been registered in the international prospective register of systematic reviews (PROSPERO protocol CRD42017080647; available at http://www.crd.york.ac.uk/PROSPERO/display_record.php?ID=CRD42017080647).
Search strategy
We systematically searched the following five online databases: MEDLINE/PubMed (since 1946), EMBASE (since 1974), PsycINFO (since 1806), Web of Science (since 1900), and OpenGrey (since 2004). The search strategy included combinations of population search terms and exposure search terms. The population search terms were (‘Lewy’ OR ‘Parkinson*’) AND ‘Dementia’. The exposure search terms included (Gene* AND express*) OR (RNA) OR (qPCR) OR (RNA AND Seq*). Reference lists of the studies included in the review were explored for identifying other potentially eligible studies. All studies that were published on or before 1 January 2018 were considered. Studies that were not published in English were not included.
Eligibility criteria
We included all gene expression studies that satisfied the following eligibility criteria: (i) they were human studies. Studies on animal models and in vitro studies investigating human tissue derived cell lines were excluded; (ii) they presented original research data; (iii) participants in at least one study group were clinically diagnosed to have DLB or PDD or LBD. Studies that solely included people with other types of dementia or PD without dementia were excluded; (iv) there was a control group in which LBD was clinically ruled out; and (v) they investigated expression levels of at least one gene.
Study selection
We merged our search results across the databases and removed duplicates using the RefWorks software (ProQuest LLC, USA). We excluded the abstracts that did not mention investigating gene expression changes in people with LBD. We attempted retrieving full texts of all potentially eligible abstracts and assessed the eligibility of the full-text papers. The studies that failed to meet one or more of the eligibility criteria were excluded. When a conference abstract was not accompanied by its full text, we requested further details from the corresponding author, if the contact information was provided. If the corresponding author did not respond to our request within 14 days, we excluded that abstract.
Quality assessment
We assessed the quality of eligible studies using a tool (Supplementary Table 1), adapted from the quality of genetic association studies tool (Q-Genie) (Sohani et al., Reference Sohani, Meyre, DE Souza, Joseph, Gandhi, Dennis, Norman and Anand2015; Sohani et al., Reference Sohani, Sarma, Alyass, DE Souza, Robiou-DU-Pont, LI, Mayhew, Yazdi, Reddon, Lamri, Stryjecki, Ishola, Lee, Vashi, Anand and Meyre2016). The tool assessed the following 11 dimensions: (i) the rationale for study, (ii) selection and definition of people with LBD, (iii) selection and comparability of comparison groups, (iv) technical assessment of gene expression, (v) non-technical aspects of assessment of gene expression, (vi) other sources of bias, (vii) sample size and power, (viii) a priori planning of statistical analyses, (ix) statistical methods and control for confounding, (x) testing of assumptions and inferences for gene expression analyses, and (xi) appropriate interpretation of the study results. Each dimension was scored on a scale from one (poor) to seven (excellent), so the total scores could range from 11 to 77.
Data extraction
We extracted the following data: (i) population characteristics including their mean age, ethnicity, and severity of illness, (ii) sample size in each study group, (iii) case definition, (iv) investigated genes and their transcripts, (v) investigated tissue, (vi) methods for analysing gene expression changes, (vii) differential fold changes between the study groups with their p-values, (viii) statistical correction for multiple testing, and (ix) statistical analyses addressing the effects of potential confounders.
Data synthesis
A descriptive synthesis was carried out using the extracted data and major findings of each included study. We have synthesised the data by listing the reported gene expression changes in post-mortem brain tissue and in biological fluids of people with DLB or PDD. We have summarised the reported findings on alternative splicing and on differentially expressed microRNA (miRNA) in people with DLB or PDD separately.
Data analyses
Functional implications of identified differentially expressed genes (DEGs) (p < 0.05) were analysed by the Ingenuity Pathway Analyses (IPA) using the Ingenuity knowledge base (Ingenuity, Redwood City, USA). IPA is a powerful functional analysis tool that helps identifying potential biomarkers within the context of biological systems. Our IPA analysis settings included stringent filters with only experimentally observed relationships, and we identified disrupted functional pathways and dysfunctional molecular networks after Benjamani–Hochberg false discovery rate (FDR) at 5% correction in people with LBD.
Results
Included studies
Fig. 1 presents the Preferred Reporting Items for Systematic Review and Meta-Analyses (PRISMA) (Moher et al., Reference Moher, Liberati, Tetzlaff, Altman and Group2010) flow chart describing the process of identifying all eligible studies. We identified and screened 2,379 papers after removing the duplicates and found 31 papers eligible to be included in this systematic review. Among the 31 included studies, 23 measured gene expression changes using quantitative polymerase chain reaction (qPCR). There were three studies using gene expression microarrays (Stamper et al., Reference Stamper, Siegel, Liang, Pearson, Stephan, Shill, Connor, Caviness, Sabbagh, Beach, Adler and Dunckley2008; Nelson et al., Reference Nelson, Wang, Janse and Thompson2018; Santpere et al., Reference Santpere, Garcia-Esparcia, Andres-Benito, Lorente-Galdos, Navarro and Ferrer2018) and three studies using next-generation RNA sequencing (RNA-Seq) (Henderson-Smith et al., Reference Henderson-Smith, Corneveaux, DE Both, Cuyugan, Liang, Huentelman, Adler, Driver-Dunckley, Beach and Dunckley2016; Hoss et al., Reference Hoss, Labadorf, Beach, Latourelle and Myers2016; Pietrzak et al., Reference Pietrzak, Papp, Curtis, Handelman, Kataki, Scharre, Rempala and Sadee2016). One study employed northern blotting (Shyu et al., Reference Shyu, Kao, Chou, Hsu and Soong2000) and another used RNase protection assay (Bychkov et al., Reference Bychkov, Gurevich, Joyce, Benovic and Gurevich2008). Supplementary Table 2 presents the quality assessment scores of all included studies. Their quality assessment total scores ranged from 40 to 56 (median = 49), and there were 23 studies with quality assessment total scores above 45.
The Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) flow chart.
Overall, 369 individual people with LBD and 368 controls were included in these 31 studies. Among them, there were 294 people with DLB and 75 people with PDD. Most of the studies have investigated post-mortem brain tissue that obtained from various brain banks. Investigated regions of brain tissue include frontal cortex, temporal cortex, motor cortex, medial temporal gyrus, superior parietal gyrus, superior frontal gyrus, substantia nigra, hippocampus, anterior cingulate cortex, locus coeruleus, caudate nucleus, and pons. All but one study have reported the details of brain bank and the investigated region of brain tissue. There were only three studies (Shyu et al., Reference Shyu, Kao, Chou, Hsu and Soong2000; Funahashi et al., Reference Funahashi, Yoshino, Yamazaki, Mori, Mori, Ozaki, Sao, Ochi, Iga and Ueno2017; Salemi et al., Reference Salemi, Cantone, Salluzzo, Giambirtone, Spada and Ferri2017) that have investigated peripheral blood samples, and there was only one study (Muller et al., Reference Muller, Kuiperij, Versleijen, Chiasserini, Farotti, Baschieri, Parnetti, Struyfs, DE Roeck, Luyckx, Engelborghs, Claassen and Verbeek2016) that has investigated cerebrospinal fluid (CSF). Supplementary Table 3 presents the list of all 1,242 reported DEGs in people with DLB. There were 1,236 reported DEGs in post-mortem brain tissue and six DEGs in biological fluids have been reported so far.
DEGs in post-mortem DLB brains
Table 1 presents a summary of studies that have investigated gene expression changes in post-mortem DLB brains. α-synuclein encoding SNCA total gene expression levels did not differ significantly in people with DLB, but its shorter isoforms have been found to be significantly upregulated. Reported DEGs in post-mortem DLB brains included genes involved in protein signalling, folding, and degradation. UCHL-1 encoding ubiquitin C-terminal hydrolase L1, SNCAIP encoding synphilin-1, and PRKN encoding Parkin contribute to the ubiquitin proteasome system (UPS) that is essential for the regulation and removal of misfolded proteins. Statistically significant downregulation of UCHL-1 and PRKN and significant upregulation of SNCAIP have been reported in DLB brains (Ciechanover et al., Reference Ciechanover, Orian and Schwartz2000; Shimura et al., Reference Shimura, Hattori, Kubo, Mizuno, Asakawa, Minoshima, Shimizu, Iwai, Chiba, Tanaka and Suzuki2000; Barrachina et al., Reference Barrachina, Castano, Dalfo, Maes, Buesa and Ferrer2006). The autophagy–lysosomal pathway (ALP) is crucial for protein degradation (Ginns et al., Reference Ginns, Mak, KO, Karlgren, Akbarian, Chou, Guo, Lim, Samuelsson, Lamarca, Vazquez-Derose and Manning-Bog2014). GBA gene encoding lysosomal enzyme beta-glucocerebrosidase was significantly downregulated in DLB brains (Chiasserini et al., Reference Chiasserini, Paciotti, Eusebi, Persichetti, Tasegian, Kurzawa-Akanbi, Chinnery, Morris, Calabresi, Parnetti and Beccari2015). β-site amyloid precursor protein (APP) cleaving enzyme encoding BACE1 was found to be significantly upregulated in DLB (Coulson et al., Reference Coulson, Beyer, Quinn, Brockbank, Hellemans, Irvine, Ravid and Johnston2010). Moreover, genes involved in synaptic regulation and neurotransmission, such as TH, ADRA2C and ADRA1D, were significantly differentially expressed in post-mortem DLB brains, but these findings have not been replicated so far (Szot et al., Reference Szot, White, Greenup, Leverenz, Peskind and Raskind2006). Furthermore, HSP70 and HSP27 were found to be upregulated up to threefold in DLB brains (Cantuti-Castelvetri et al., Reference Cantuti-Castelvetri, Klucken, Ingelsson, Ramasamy, Mclean, Frosch, Hyman and Standaert2005; Outeiro et al., Reference Outeiro, Klucken, Strathearn, Liu, Nguyen, Rochet, Hyman and Mclean2006).
Summary of gene expression studies in people with dementia with Lewy bodies
DLB, dementia with Lewy bodies; PDD, Parkinson’s disease dementia; Ctrl, controls; qPCR, quantitative polymerase chain reaction; RNA-Seq, next-generation RNA sequencing; TV, transcript variant; DEG, differentially expressed genes; FC, frontal cortex; PL, peripheral leucocytes; BT, unspecified brain tissue; TC, temporal cortex; MTG, medial temporal gyri; SPG, superior parietal gyri; SFG, superior frontal gyri; SN, substantia nigra; Hc, hippocampus; ACC, anterior cingulate cortex; LC, locus coeruleus; CN, caudate nucleus
DEGs in post-mortem PDD brains
Six studies have investigated gene expression changes in post-mortem PDD brains (Bychkov et al., Reference Bychkov, Gurevich, Joyce, Benovic and Gurevich2008; Stamper et al., Reference Stamper, Siegel, Liang, Pearson, Stephan, Shill, Connor, Caviness, Sabbagh, Beach, Adler and Dunckley2008; Beyer et al., Reference Beyer, Domingo-Sabat, Santos, Tolosa, Ferrer and Ariza2010; Beyer et al., Reference Beyer, Ispierto, Latorre, Tolosa and Ariza2011; Henderson-Smith et al., Reference Henderson-Smith, Corneveaux, DE Both, Cuyugan, Liang, Huentelman, Adler, Driver-Dunckley, Beach and Dunckley2016; Hoss et al., Reference Hoss, Labadorf, Beach, Latourelle and Myers2016) (Table 2). A RNA-Seq study that investigated the posterior cingulate cortical transcriptomics of people with PDD has reported statistically significant upregulation of genes associated with protein folding pathways, such as HSP40 and DNAJB1, and downregulation of genes associated with hormonal activity, ion transport, nerve growth, and cytoskeleton structure (Henderson-Smith et al., Reference Henderson-Smith, Corneveaux, DE Both, Cuyugan, Liang, Huentelman, Adler, Driver-Dunckley, Beach and Dunckley2016). Inflammation-associated CSF3 and SELE were significantly upregulated, and PENK, CRH, and SST were significantly downregulated in people with PDD (Henderson-Smith et al., Reference Henderson-Smith, Corneveaux, DE Both, Cuyugan, Liang, Huentelman, Adler, Driver-Dunckley, Beach and Dunckley2016). Another study that investigated the gene expression changes in posterior cingulate cortices of people with PDD using gene expression microarrays identified 556 DEGs (p < 0.01) in PDD. There was downregulation of genes involved in neurite growth and cell adhesion, such as KIF21A, DYNC2LI1, and TBCA (Stamper et al., Reference Stamper, Siegel, Liang, Pearson, Stephan, Shill, Connor, Caviness, Sabbagh, Beach, Adler and Dunckley2008). Another study that employed RNase protection assay has reported significant upregulation of G protein-coupled receptor (GPCR) pathways-related ARRB2 (p < 0.001), ARR3 (p < 0.001), GRK3 (p < 0.01), and GRK5 (p < 0.05) in post-mortem PDD brains (Bychkov et al., Reference Bychkov, Gurevich, Joyce, Benovic and Gurevich2008).
Summary of gene expression studies in people with Parkinson’s disease dementia
PDD, Parkinson’s disease with dementia; Ctrl, controls; qPCR, quantitative polymerase chain reactions; RNA-Seq, next-generation RNA sequencing; TC, temporal cortex; CN, caudate nucleus; Po, pons; PB, peripheral blood; BG, basal ganglia; PCC, posterior cingulate cortex; TV, transcript variant; DEG, differentially expressed genes.
DEGs in biological fluids of people with LBD
There were only three studies which have investigated gene expression changes in biological fluids of people with DLB. Two of them have investigated peripheral leucocytes (Funahashi et al., Reference Funahashi, Yoshino, Yamazaki, Mori, Mori, Ozaki, Sao, Ochi, Iga and Ueno2017; Salemi et al., Reference Salemi, Cantone, Salluzzo, Giambirtone, Spada and Ferri2017), and the third has investigated CSF (Muller et al., Reference Muller, Kuiperij, Versleijen, Chiasserini, Farotti, Baschieri, Parnetti, Struyfs, DE Roeck, Luyckx, Engelborghs, Claassen and Verbeek2016). There was only one study that specifically investigated gene expression changes in biological fluids of people with PDD (Shyu et al., Reference Shyu, Kao, Chou, Hsu and Soong2000). SNCA total gene expression levels did not differ significantly, but its isoform SNCA-126 level was significantly upregulated in peripheral leucocytes of people with DLB (Funahashi et al., Reference Funahashi, Yoshino, Yamazaki, Mori, Mori, Ozaki, Sao, Ochi, Iga and Ueno2017). Another study has assessed expression levels of 11 mitochondrial genes in peripheral leucocytes and found significant (p < 0.05) downregulation of MT-ATP8, MT-CO2, MT-CO3, and MT-ND2 in people with DLB (Salemi et al., Reference Salemi, Cantone, Salluzzo, Giambirtone, Spada and Ferri2017). Only MIR-125B was significantly downregulated in CSF of people with DLB (p = 0.03) (Muller et al., Reference Muller, Kuiperij, Versleijen, Chiasserini, Farotti, Baschieri, Parnetti, Struyfs, DE Roeck, Luyckx, Engelborghs, Claassen and Verbeek2016). Moreover, expression levels of HSP70 did not differ significantly in peripheral mononuclear blood cells of people with PDD (Shyu et al., Reference Shyu, Kao, Chou, Hsu and Soong2000).
Potential miRNA biomarkers for LBD
Four studies have reported differential expression of miRNAs in people with DLB (Hebert et al., Reference Hebert, Wang, Zhu and Nelson2013; Muller et al., Reference Muller, Kuiperij, Versleijen, Chiasserini, Farotti, Baschieri, Parnetti, Struyfs, DE Roeck, Luyckx, Engelborghs, Claassen and Verbeek2016; Pietrzak et al., Reference Pietrzak, Papp, Curtis, Handelman, Kataki, Scharre, Rempala and Sadee2016; Nelson et al., Reference Nelson, Wang, Janse and Thompson2018) and two more have investigated differentially expressed miRNAs in people with PDD (Henderson-Smith et al., Reference Henderson-Smith, Corneveaux, DE Both, Cuyugan, Liang, Huentelman, Adler, Driver-Dunckley, Beach and Dunckley2016; Hoss et al., Reference Hoss, Labadorf, Beach, Latourelle and Myers2016). Seventy differentially expressed miRNAs in people with LBD have been identified so far (Table 3) (Supplementary Table 3). MIR-125B was significantly differentially expressed in both post-mortem DLB temporal cortices and CSF of people living with DLB (Hebert et al., Reference Hebert, Wang, Zhu and Nelson2013; Muller et al., Reference Muller, Kuiperij, Versleijen, Chiasserini, Farotti, Baschieri, Parnetti, Struyfs, DE Roeck, Luyckx, Engelborghs, Claassen and Verbeek2016). Differential expression levels of 36 miRNAs in prefrontal cortices could distinguish PDD from PD with 81.2% sensitivity and 88.9% specificity (Hoss et al., Reference Hoss, Labadorf, Beach, Latourelle and Myers2016). Moreover, a study that investigated anterior cingulate cortices of people with DLB using RNA-Seq has identified 14 potential upstream regulatory miRNAs after appropriate multiple testing correction (Pietrzak et al., Reference Pietrzak, Papp, Curtis, Handelman, Kataki, Scharre, Rempala and Sadee2016) (Table 3).
Summary of studies that have investigated miRNA expression changes in people with Lewy body dementia
DLB, Dementia with Lewy bodies; Ctrl, controls; PD, Parkinson’s disease; PDD, Parkinson’s disease with dementia; AD, Alzheimer’s disease; qPCR, quantitative polymerase chain reactions; RNA-Seq, next-generation RNA sequencing; CSF, cerebrospinal fluid; TG, temporal gyrus; PFC, prefrontal cortex; ACC, anterior cingulate cortex; MC, motor cortex.
Importance of alternative splicing in LBD
Expression levels of multiple isoforms of SNCA, SNCB, PRKN, and APP have been evaluated in people with DLB. Four alternatively spliced transcripts of SNCA, SNCA-98, SNCA-112, SNCA-126, and SNCA-140 have been studied. SNCA-98 was expressed 2.7 times more in frontal cortices of people with DLB (p < 0.05) (Beyer et al., Reference Beyer, Domingo-Sabat, Lao, Carrato, Ferrer and Ariza2008). Similarly, SNCA-112 expression levels were upregulated in people with DLB by twofold, when compared with controls without cognitive impairment (p = 0.002), and by threefold, when compared with people with AD (p < 0.001). Fourfold downregulation (p < 0.001) in frontal cortices and twofold upregulation in peripheral leucocytes of SNCA-126 in people with DLB have been reported (Beyer et al., Reference Beyer, Humbert, Ferrer, Lao, Carrato, Lopez, Ferrer and Ariza2006; Funahashi et al., Reference Funahashi, Yoshino, Yamazaki, Mori, Mori, Ozaki, Sao, Ochi, Iga and Ueno2017). Significant (p = 0.008) downregulation of SNCA-140 in DLB brains (Beyer et al., Reference Beyer, Lao, Carrato, Mate, Lopez, Ferrer and Ariza2004a) has been reported, but another study has failed to replicate this finding in frontal cortices and caudate nuclei of people with DLB (Beyer et al., Reference Beyer, Domingo-Sabat, Santos, Tolosa, Ferrer and Ariza2010). Upregulation of SNCA-140 levels in temporal cortices and its downregulation in caudate nuclei of people with PDD has been reported (Beyer et al., Reference Beyer, Ispierto, Latorre, Tolosa and Ariza2011). Two transcript variants (TVs) of SNCB have been investigated in people with LBD (Snyder et al., Reference Snyder, Mensah, Hsu, Hashimoto, Surgucheva, Festoff, Surguchov, Masliah, Matouschek and Wolozin2005; Beyer et al., Reference Beyer, Domingo-Sabat, Santos, Tolosa, Ferrer and Ariza2010). Expression levels the alternatively spliced isoforms, SNCB-TV1 and TV2, varied across different brain regions. SNCB-TV1 and TV2 were significantly downregulated in temporal and frontal cortices, and SNCB-TV1 was significantly upregulated in caudate nuclei of people with DLB. SNCB-TV1 and TV2 were significantly upregulated in temporal cortices and caudate nuclei of people with PDD (Beyer et al., Reference Beyer, Domingo-Sabat, Santos, Tolosa, Ferrer and Ariza2010; Beyer et al., Reference Beyer, Ispierto, Latorre, Tolosa and Ariza2011). Levels of both SNCB transcripts were significantly reduced in pons of people with PDD (Beyer et al., Reference Beyer, Ispierto, Latorre, Tolosa and Ariza2011).
Three major isoforms of APP, APP-695, APP-751, and APP-770 have been investigated in people with DLB. People with DLB had significantly higher levels of APP-770 (p < 0.05) and significantly lower levels of APP-695. Results on APP-751 expression levels were inconsistent (Beyer et al., Reference Beyer, Lao, Carrato, Mate, Lopez, Ferrer and Ariza2004b; Barrachina et al., Reference Barrachina, Dalfó, Puig, Vidal, Freixes, Castaño and Ferrer2005). Moreover, significantly (p = 0.02) lower levels of PRKN-TV7 have been reported in people with DLB (Humbert et al., Reference Humbert, Beyer, Carrato, Mate, Ferrer and Ariza2007). Furthermore, another study has investigated the importance of alternative splicing in posterior cingulate cortices of people with PDD using RNA-Seq and SpliceSeq (Ryan et al., Reference Ryan, Cleland, Kim, Wong and Weinstein2012) software (Henderson-Smith et al., Reference Henderson-Smith, Corneveaux, DE Both, Cuyugan, Liang, Huentelman, Adler, Driver-Dunckley, Beach and Dunckley2016). There was 2.6-fold overexpression of alternatively spliced isoforms of RELA, compared to its main transcript, in people with PDD. More than threefold downregulation of alternatively spliced isoforms of ATXN2 in PDD was found (Henderson-Smith et al., Reference Henderson-Smith, Corneveaux, DE Both, Cuyugan, Liang, Huentelman, Adler, Driver-Dunckley, Beach and Dunckley2016).
Functional analyses of reported DEGs in LBD
We investigated the functional implications of the 1,242 reported DEGs (p < 0.05) using IPA. Movement disorders (p = 2.42 × 10−9), disorders of basal ganglia (p = 2.71 × 10−9), schizophrenia (p = 3.29 × 10−8), immune response of brain (p = 3.48 × 10−7), neuronal death and survival (p = 7.34 × 10−6), tauopathy of hippocampus (p = 2.96 × 10−5), neuronal morphology (p = 4.28 × 10−5), synaptic transmission (p = 3.93 × 10−4), and tauopathy of amygdala (p = 4.02 × 10−4) associated genes were significantly enriched among the reported DEGs in people with LBD. Fig. 2(A) presents the molecular pathways that were significantly enriched among the LBD DEGs after Benjamini–Hochberg FDR correction at 5%. Initiation of protein translation-related eukaryotic initiation factor-2 (EIF2) signalling, neuronal maintenance-related neuregulin signalling, oxidative damage, apoptosis, and neuronal survival-related PEDF signalling, signal transduction via phosphorylation and cell survival-related mechanistic target of rapamycin (mTOR) signalling, and gene transcription regulating High Mobility Group-B1 (HMGB1) signalling pathways-associated genes were significantly enriched among the LBD DEGs after FDR correction (Supplementary Table 4). The genes, associated with neuroinflammation signalling pathway, were significantly enriched among LBD DEGs after FDR correction and they included 21 upregulated and 14 downregulated DEGs. Moreover, movement of myeloid cells-related genes were significantly enriched among the downregulated LBD DEGs (z = -2.24; p = 3.23 × 10−4).
Functional analyses of reported differentially expressed genes (DEGs) in people with Lewy body dementia (LBD). (A) Canonical pathways were enriched among the reported DEG in people with LBD. Green represents downregulated genes and red represents upregulated genes. Yellow line presents the p-values after Benjamini–Hochberg false discovery rate (5%) correction. (B) A dysfunctional network of reported DEG may lead to amyloidopathy and tauopathy in LBD. (C) A dysfunctional molecular network involving α-synuclein, presenilin 1, and tyrosine hydroxylase may explain α-synuclein aggregation and neurodegeneration in LBD. (D) A dysfunctional molecular network involving brain-derived neurotrophic factor, dicer 1, and argonaute highlights the importance of neurotropic factors and RNA silencing complexes in the pathophysiology of LBD. (B–D). Green represents downregulated genes and red represents upregulated genes. Solid lines represent direct interactions and dotted lines represent indirect interactions.
Our IPA upstream analyses and causal network analyses revealed that inhibition of TCF7L2 encoding a transcription factor (p = 4.44 × 10−21) and of neurotrophic BDNF (p = 7.39 × 10−6) were likely upstream biological causes leading to the reported gene expression changes in people with LBD (Kramer et al., Reference Kramer, Green, Pollard and Tugendreich2014). Furthermore, they highlighted the importance of RNA-mediated gene silencing in the molecular pathology of LBD. AGO2 (p = 2.22 × 10−8) encoding protein argonaute-2 that is essential for the formation RNA-induced silencing complex (RISC), and DICER1 (p = 7.57 × 10−6) encoding dicer that cleaves miRNA and small interfering RNA (siRNA) and activates RISC were identified as potential upstream biological causes for the reported gene expression changes in LBD. Our IPA network analyses (Supplementary Table 5) showed that several reported DEGs in LBD, including APBA2, HTRA2, ENO2, and MIR17 directly or indirectly interact with APP and Tau encoding APP and MAPT [Fig. 2(B)]. This dysfunctional molecular network can explain neurodegeneration in LBD and varying degrees of comorbid AD pathology, reported in many post-mortem LBD brains [Fig. 2(B)]. PD- and DLB-associated genes, GBA, PRKN, and SCARB2, directly interact with SNCA that interacts with L-3,4-dihydroxyphenylalanine (L-DOPA) synthesising tyrosine hydroxylase gene (TH) and with PSEN1 encoding an essential protein for γ-secretase complex that cleaves β-amyloid from APP. Reported downregulation of PSEN1 can increase α-synuclein aggregation independent of its γ-secretase activity, and associated downregulation of SNCB further increases α-synuclein aggregation that leads to formation of Lewy bodies (Winslow et al., Reference Winslow, Moussaud, Zhu, Post, Dickson, Berezovska and Mclean2014). Fig. 2(C) presents a dysfunctional molecular network including differential expression of these genes and their interactions. Fig. 2(D) presents the complex interactions between neurotrophic BDNF and RNA-mediated gene silencing, regulated by DICER1, AGO2, and associated miRNAs. This dysfunctional molecular network can influence gene expression of many downstream genes and can impact neuronal survival and maintenance in people with LBD.
Discussion
This is the first systematic review of all gene expression studies that have investigated people with LBD. We have listed all reported DEGs in people with LBD and have investigated their functional implications. Our functional analyses advance our understanding of molecular mechanisms underlying neurodegeneration in LBD. The strengths of this systematic review include its broad eligibility criteria, following PRISMA guidelines, and searching multiple databases including grey literature. Its limitations are excluding studies that were not published in English, not including studies that investigated animal models or cell lines, and substantial heterogeneity among the included studies. All included studies were small, and the smallest study has included only four people with LBD (Hebert et al., Reference Hebert, Wang, Zhu and Nelson2013). The studies have not reported sample size estimation or power analysis, so type-II error is likely. They differed widely on their population characteristics, case definitions, selection of controls, methods for measuring gene expression changes, and statistical analyses. Many studies did not employ statistical corrections for multiple testing. Majority of the studies have employed relative quantification qPCR, so it was not possible to do combined analyses using their findings. Moreover, another RNA-Seq study that investigated anterior cingulate and dorsolateral prefrontal cortical transcriptomics of people with DLB and PDD was published after the completion of this systematic review in June 2019 (Rajkumar et al., Reference Rajkumar, Bidkhori, Shoaie, Clarke, Morrin, Hye, Williams, Ballard, Francis and Aarsland2020). It identified 12 genome-wide significant DEGs (MPO, SELE, CTSG, ALPI, ABCA13, GALNT6, SST, RBM3, CSF3, SLC4A1, OXTR, and RAB44) in people with LBD.
Although α-synuclein aggregation is the key initial step in the formation of Lewy bodies (Beyer et al., Reference Beyer, Domingo-Sabat and Ariza2009), α-synuclein encoding SNCA total gene expression levels often do not differ significantly in people with LBD. Two of its alternatively spliced shorter isoforms, SNCA-98 and SNCA-112, have increased propensity for aggregation (Beyer et al., Reference Beyer, Humbert, Ferrer, Lao, Carrato, Lopez, Ferrer and Ariza2006), and they were found to be significantly upregulated in DLB brains (Beyer et al., Reference Beyer, Domingo-Sabat, Lao, Carrato, Ferrer and Ariza2008). This highlights the need for further in-depth investigation of RNA biology, alternative splicing, and expression levels of individual transcripts in people with LBD and other α-synucleinopathies (Lee & Trojanowski, Reference Lee and Trojanowski2006). Upregulation of SNCA-126 could be detected in peripheral leucocytes of people living with DLB (Funahashi et al., Reference Funahashi, Yoshino, Yamazaki, Mori, Mori, Ozaki, Sao, Ochi, Iga and Ueno2017), and diagnostic biomarker potential of peripheral levels of SNCA transcripts warrant further research. Moreover, alternative splicing of β-synuclein encoding SNCB also plays an important role in the pathology of α-synucleinopathies (Gamez-Valero & Beyer, Reference Gamez-Valero and Beyer2018). β-synuclein prevents α-synuclein inhibiting proteasomes, and it inhibits further α-synuclein aggregation (Snyder et al., Reference Snyder, Mensah, Hsu, Hashimoto, Surgucheva, Festoff, Surguchov, Masliah, Matouschek and Wolozin2005). Expression levels of the two SNCB transcripts differ between DLB brains with and without co-existent AD pathology (Beyer et al., Reference Beyer, Domingo-Sabat, Santos, Tolosa, Ferrer and Ariza2010). SNCB-TV1 and TV2 were significantly downregulated in temporal cortices of people with DLB and they were significantly upregulated in temporal cortices of people with PDD (Beyer et al., Reference Beyer, Domingo-Sabat, Santos, Tolosa, Ferrer and Ariza2010; Beyer et al., Reference Beyer, Ispierto, Latorre, Tolosa and Ariza2011). Further investigation of expression levels of individual transcripts of SNCB may lead to molecular subtyping of LBD (Beyer et al., Reference Beyer, Domingo-Sabat, Santos, Tolosa, Ferrer and Ariza2010).
Lewy bodies are complex structures and they are made of more than 80 distinct proteins (Wakabayashi et al., Reference Wakabayashi, Tanji, Odagiri, Miki, Mori and Takahashi2013). Available gene expression findings in people with LBD highlight the defects in molecular networks clearing abnormal proteins than overexpression of a few pathogenic genes. Optimal functioning of ALP and UPS is essential for the degradation of misfolded proteins (Higashi et al., Reference Higashi, Moore, Minegishi, Kasanuki, Fujishiro, Kabuta, Togo, Katsuse, Uchikado, Furukawa, Hino, Kosaka, Sato, Arai, Wada and Iseki2011). Prior studies, which mainly focused on expression levels of genes associated with AD or PD, have indicated dysfunctional ALP and UPS contributing to neurodegeneration in LBD. They have reported statistically significant downregulation of UCHL-1, PRKN, and GBA in people with LBD (Ciechanover et al., Reference Ciechanover, Orian and Schwartz2000; Shimura et al., Reference Shimura, Hattori, Kubo, Mizuno, Asakawa, Minoshima, Shimizu, Iwai, Chiba, Tanaka and Suzuki2000; Barrachina et al., Reference Barrachina, Castano, Dalfo, Maes, Buesa and Ferrer2006; Chiasserini et al., Reference Chiasserini, Paciotti, Eusebi, Persichetti, Tasegian, Kurzawa-Akanbi, Chinnery, Morris, Calabresi, Parnetti and Beccari2015). Recent transcriptomic studies have advanced our understanding of dysfunctional molecular networks involving ALP and UPS in people with LBD (Stamper et al., Reference Stamper, Siegel, Liang, Pearson, Stephan, Shill, Connor, Caviness, Sabbagh, Beach, Adler and Dunckley2008; Henderson-Smith et al., Reference Henderson-Smith, Corneveaux, DE Both, Cuyugan, Liang, Huentelman, Adler, Driver-Dunckley, Beach and Dunckley2016; Hoss et al., Reference Hoss, Labadorf, Beach, Latourelle and Myers2016; Pietrzak et al., Reference Pietrzak, Papp, Curtis, Handelman, Kataki, Scharre, Rempala and Sadee2016; Nelson et al., Reference Nelson, Wang, Janse and Thompson2018; Santpere et al., Reference Santpere, Garcia-Esparcia, Andres-Benito, Lorente-Galdos, Navarro and Ferrer2018; Rajkumar et al., Reference Rajkumar, Bidkhori, Shoaie, Clarke, Morrin, Hye, Williams, Ballard, Francis and Aarsland2020). Decreased expression of GBA impairs lysosomal protein degradation and leads to α-synuclein aggregation and neurotoxicity in stem cell-derived neurons (Mazzulli et al., Reference Mazzulli, XU, Sun, Knight, Mclean, Caldwell, Sidransky, Grabowski and Krainc2011). Aggregated α-synuclein can set off a vicious cycle by inhibiting neuronal lysosomal activity further (Mazzulli et al., Reference Mazzulli, XU, Sun, Knight, Mclean, Caldwell, Sidransky, Grabowski and Krainc2011). SCARB2 encodes a lysosomal membrane protein that transports GBA to lysosomes (Gan-Or et al., Reference Gan-OR, Dion and Rouleau2015) and its downregulation may impair the ALP further in people with LBD. GBA, PRKN, and SCARB2 directly interact with SNCA and they interact indirectly with TH, SNCB, and PSEN1. This dysfunctional network and the decreased expression of UCHL1 that is essential for the hydrolysis of misfolded proteins by neuronal UPS (Saigoh et al., Reference Saigoh, Wang, Suh, Yamanishi, Sakai, Kiyosawa, Harada, Ichihara, Wakana, Kikuchi and Wada1999; Shibata et al., Reference Shibata, Motoi, Tomiyama, Ohnuma, Kuerban, Tomson, Komatsu, Hattori and Arai2012) exacerbate α-synuclein aggregation and cytoplasmic accumulation of other misfolded proteins.
Significantly decreased expression levels of mitochondrial genes involved in energy metabolism have been reported in people with LBD. Significant downregulation of MT-ATP8, MT-CO2, MT-CO3, and MT-ND2 could be measured in peripheral leucocytes of people with DLB (Salemi et al., Reference Salemi, Cantone, Salluzzo, Giambirtone, Spada and Ferri2017). Reduced levels of mitochondrial complex I activity and oxygen uptake in DLB brains have been reported (Navarro et al., Reference Navarro, Boveris, Bandez, Sanchez-Pino, Gomez, Muntane and Ferrer2009; Swerdlow, Reference Swerdlow2011). Moreover, prior genetic association studies have found significant associations of LBD with mtDNA haplogroup H (Chinnery et al., Reference Chinnery, Taylor, Howell, Andrews, Morris, Taylor, Mckeith, Perry, Edwardson and Turnbull2000) and TFAM encoding mitochondrial transcription factor A (Gatt et al., Reference Gatt, Jones, Francis, Ballard and Bateman2013). A recent RNA-Seq study and subsequent analysis of metabolic reprogramming in LBD brains by genome-scale metabolic modelling (Sertbas et al., Reference Sertbas, Ulgen and Cakir2014) have highlighted the importance of mitochondrial dysfunction in LBD pathology (Rajkumar et al., Reference Rajkumar, Bidkhori, Shoaie, Clarke, Morrin, Hye, Williams, Ballard, Francis and Aarsland2020). Mitochondrial dysfunction may set off a vicious cycle by generating more reactive oxygen species, which can lead to more mitochondrial oxidative damage (Spano et al., Reference Spano, Signorelli, Vitaliani, Aguglia and Giometto2015). Reactive oxygen species and consequent oxidative stress lead to α-synuclein aggregation that in turn impair more mitochondria (Spano et al., Reference Spano, Signorelli, Vitaliani, Aguglia and Giometto2015). Further studies are warranted for investigating associated molecular mechanisms as well as the biomarker and therapeutic potential of mitochondrial transcripts in LBD.
Our functional analyses of reported DEGs highlighted the importance of RNA-mediated gene silencing, neuregulin signalling, and downregulation of neurotrophic factors in the molecular pathology of LBD. Downstream regulatory effects of decreased expression of neurotrophic BDNF may explain many reported gene expression changes in people with LBD (Kramer et al., Reference Kramer, Green, Pollard and Tugendreich2014). BDNF interacts with AGO2 and DICER1, which were found to be upregulated in people with LBD. AGO2 is essential for the formation of RISC and DICER1 is important for the activation of RISC. Consequent RNA-mediated gene silencing may lead to downregulation of several downstream genes that related to neuronal survival and maintenance in people with LBD. Moreover, several reported DEGs in LBD interact with APP and MAPT. BACE1 expression levels were significantly upregulated in people with DLB (Coulson et al., Reference Coulson, Beyer, Quinn, Brockbank, Hellemans, Irvine, Ravid and Johnston2010) and there is a two-way relationship between α-synuclein aggregation and β-amyloid secretion (Roberts et al., Reference Roberts, Schneider and Brown2017). These findings may explain varying degrees of co-existent amyloid and Tau pathology in people with LBD.
Unlike AD (Perry, Reference Perry2004), available gene expression studies have provided inconsistent evidence for the presence of chronic neuroinflammation in people with LBD. Inflammation-associated genes were significantly upregulated in people with PDD (Stamper et al., Reference Stamper, Siegel, Liang, Pearson, Stephan, Shill, Connor, Caviness, Sabbagh, Beach, Adler and Dunckley2008). However, a gene expression microarray study and a RNA-Seq study have documented statistically significant downregulation of several inflammation-associated genes, including IL1B, IL2, IL6, CXCL2, CXCL3, CXCL8, CXCL10, and CXCL11 in post-mortem DLB brains (Santpere et al., Reference Santpere, Garcia-Esparcia, Andres-Benito, Lorente-Galdos, Navarro and Ferrer2018; Rajkumar et al., Reference Rajkumar, Bidkhori, Shoaie, Clarke, Morrin, Hye, Williams, Ballard, Francis and Aarsland2020). TNF was upregulated only in rapidly progressive DLB (Garcia-Esparcia et al., Reference Garcia-Esparcia, Lopez-Gonzalez, Grau-Rivera, Garcia-Garrido, Konetti, Llorens, Zafar, Carmona, Del Rio, Zerr, Gelpi and Ferrer2017) and another study has found upregulation of IL6 in hippocampi of people with DLB (n = 5) (Imamura et al., Reference Imamura, Hishikawa, Ono, Suzuki, Sawada, Nagatsu, Yoshida and Hashizume2005). However, these findings have not been replicated. Moreover, a recent transcriptomic and proteomic study has reported lack of evidence for microglia-mediated neuroinflammation in post-mortem pulvinar of people with DLB (Erskine et al., Reference Erskine, Ding, Thomas, Kaganovich, Khundakar, Hanson, Taylor, Mckeith, Attems, Cookson and Morris2018). Optimal microglial activation is essential for neuronal survival and synaptic plasticity (Chen et al., Reference Chen, Jalabi, Hu, Park, Gale, Kidd, Bernatowicz, Gossman, Chen, Dutta and Trapp2014). Decreased expression of inflammation-associated genes leading to impaired neuronal survival rather than chronic neuroinflammation may explain neurodegeneration in DLB (Rajkumar et al., Reference Rajkumar, Bidkhori, Shoaie, Clarke, Morrin, Hye, Williams, Ballard, Francis and Aarsland2020). The differential expression levels of inflammation-associated genes may help distinguishing DLB from PDD and AD, and their diagnostic and prognostic biomarker potential warrant further research.
Each neurodegenerative disorder is hypothesised to have its own unique peripheral miRNA signature (Sheinerman et al., Reference Sheinerman, Toledo, Tsivinsky, Irwin, Grossman, Weintraub, Hurtig, Chen-Plotkin, Wolk, Mccluskey, Elman, Trojanowski and Umansky2017). We have listed 70 miRNAs that were found to be differentially expressed in people with LBD, and their biomarker and therapeutic potential need further investigation. Identifying differentially expressed miRNAs in post-mortem LBD brains advances our molecular-level mechanistic understanding. However, discovery of clinically adoptable diagnostic biomarkers requires identifying differentially expressed miRNAs in biological fluids of people living with LBD (Muller et al., Reference Muller, Kuiperij, Versleijen, Chiasserini, Farotti, Baschieri, Parnetti, Struyfs, DE Roeck, Luyckx, Engelborghs, Claassen and Verbeek2016). Despite their vulnerability for degradation, RNA, especially miRNA, remain stable in biological fluids by being either bound to protein complexes or encapsulated within blood cells or extracellular vesicles (EVs) (Taylor & Gercel-Taylor, Reference Taylor and Gercel-Taylor2013). Only one study has evaluated miRNA expression changes in biological fluids of people with LBD (Muller et al., Reference Muller, Kuiperij, Versleijen, Chiasserini, Farotti, Baschieri, Parnetti, Struyfs, DE Roeck, Luyckx, Engelborghs, Claassen and Verbeek2016) and there has not been any systematic research investigating the EV RNA expression levels in people with LBD. CSF small EVs can transmit α-synuclein aggregation in vitro (Stuendl et al., Reference Stuendl, Kunadt, Kruse, Bartels, Moebius, Danzer, Mollenhauer and Schneider2016). Small EVs can cross blood–brain barrier (Schiera et al., Reference Schiera, Di Liegro and Di Liegro2015) and they transport RNA between brain and peripheral circulation. Diagnostic biomarker potential of small EV RNA, enriched for neuronal origin, is increasingly recognised (Mustapic et al., Reference Mustapic, Eitan, Werner, Berkowitz, Lazaropoulos, Tran, Goetzl and Kapogiannis2017; Van Giau & An, Reference Van Giau and An2016), and the need for more studies investigating small EV RNA in people with LBD cannot be overemphasised.
Notwithstanding the extent of research on gene expression changes in people with PD without dementia, there is a scarcity of studies investigating these changes in people with PDD. The nosological validity of DLB and the diagnostic boundaries between DLB and PDD continue to be debatable because the clinical presentations of advanced stages of DLB and PDD are often identical (Postuma et al., Reference Postuma, Berg, Stern, Poewe, Olanow, Oertel, Marek, Litvan, Lang, Halliday, Goetz, Gasser, Dubois, Chan, Bloem, Adler and Deuschl2016). Prevailing sparse evidence that comes almost exclusively from post-mortem brain tissue of people with clinically advanced PDD have indicated limited overlap between DEGs in DLB and PDD. Transcriptomes of DLB and PDD may display more pronounced differences during earlier clinical stages. We suggest more transcriptomic studies investigating biological fluids of people living with DLB and PDD for advancing our understanding of molecular differences between these two clinically overlapping disorders. All studies, included in this systematic review, have expressed gene expression changes at only one point of time. Gene expression changes are dynamic and they differ with disease progression, so we suggest future gene expression studies investigating the longitudinal gene expression changes in biological fluids of people living with LBD.
Supplementary material
To view supplementary material for this article, please visit https://doi.org/10.1017/neu.2020.13
Acknowledgements
We thank the Student Selected Components (SSC) programme of King’s College London, London, UK, for funding this systematic review. We thank the authors and the participants of the included studies.
Authors’ contributions
APR conceived this study, and both APR and AC designed the systematic review protocol. AC reviewed the literature, identified eligible studies, and completed necessary quality assessment and data extraction. AC and APR interpreted the findings of the included studies. APR performed the Ingenuity Pathway Analyses. AC wrote the initial draft. Both authors were involved in further critical revisions of the manuscript and they have approved the final version of the manuscript.
Financial support
This research was supported by the Student Selected Components (SSC) programme of King’s College London, London, UK
Statement of interest
Both authors declare that they do not have any competing interests.
