Hsrω and Other lncRNAs in Neuronal Functions and Disorders in Drosophila

Long noncoding RNAs (lncRNAs) have a crucial role in epigenetic, transcriptional and posttranscriptional regulation of gene expression. Many of these regulatory lncRNAs, such as MALAT1, NEAT1, HOTAIR, etc., are associated with different neurodegenerative diseases in humans. The lncRNAs produced by the hsrω gene are known to modulate neurotoxicity in polyQ and amyotrophic lateral sclerosis disease models of Drosophila. Elevated expression of hsrω lncRNAs exaggerates, while their genetic depletion through hsrω-RNAi or in an hsrω-null mutant background suppresses, the disease pathogenicity. This review discusses the possible mechanistic details and implications of the functions of hsrω lncRNAs in the modulation of neurodegenerative diseases.


Introduction
Long noncoding RNAs (lncRNAs) have well-programmed expression profiles in neurons, and many of them are now known to be associated with human neurodegenerative diseases [1][2][3][4]. LncRNAs impact neurodegeneration by regulating multiple cellular activities such as chromatin organization, epigenetic regulation, gene expression or signaling pathways. Some of the neurodegeneration associated with lncRNAs sequester RNA processing proteins to form liquid-liquid phase separation (LLPS)-based, membraneless structures which directly or indirectly affect widespread gene expression. For example, the ubiquitously expressing nuclear-enriched abundant transcript 1 (NEAT1) lncRNA in human recruits several RNA-binding proteins (RBPs) from the Drosophila behavior/human splicing (DBHS) protein family, such as PSF/SFPQ, NONO/P54NRB and PSPC1, to organize the paraspeckles, which act as molecular hubs for different cellular processes and whose misregulation can lead to neurodegeneration. [5][6][7]. Likewise, the abundant lncRNA metastasis-associated lung adenocarcinoma transcript 1 (MALAT1), also known as NEAT2, binds with a specific set of mRNA-processing proteins to form nuclear splicing speckles and interacts with miR-125b to inhibit neuronal apoptosis and suppress inflammatory cytokines in Alzheimer's disease [8,9].
The present review briefly discusses the roles of lncRNAs associated with neurological disorders in Drosophila melanogaster, with a focus on the hsrω lncRNAs. Similar to NEAT1 and MALAT1 lncRNAs, the hsrω lncRNAs in Drosophila and their functional analog, Sat III lncRNAs in humans, sequester a large set of RBPs to control transcription and posttranscriptional processing of mRNAs. The hsrω lncRNAs modulate the neurotoxicity of diseases associated with abnormal RNA processing such as amyotrophic lateral sclerosis (ALS) and polyQ. Likewise, Sat III lncRNA is associated with the frontal cortex in frontotemporal lobar degeneration (FTLD) disease in humans. Based on experimental evidence, this review considers the possible mechanistic understanding of involvement of hsrω lncRNAs in different fly models of human neurodegenerative diseases.

LncRNAs Associated with Neuronal Functions and Diseases in Drosophila
Drosophila has been widely used as a model system to examine the involvement of different genes including lncRNAs in brain development and neurodegenerative diseases. The neuron-specific lncRNA CASK regulatory gene (CRG) was found to control locomotory behavior in Drosophila by upregulating the expression of its neighboring gene, CASK [10], which is a protein-coding gene involved in walking behavior in Drosophila [11,12]. It is suggested that the CRG lncRNA upregulates CASK expression by facilitating the association of RNA Pol II transcription machinery at the promoter region of the CASK gene [10]. Further studies are required to understand the molecular mechanism and significance of CRG-dependent regulation of CASK expression in Drosophila. The lncRNA gene yellowachaete intergenic RNA (Yar) is located within a neural gene cluster of Drosophila, and loss of Yar lncRNA in Yar-null mutants causes shortened nighttime sleep bouts within a normal circadian sleep-wake cycle and down-regulates levels of sleep rebound following deprivation [13]. These sleeping abnormalities are rescued by the transgenic expression of Yar, which demonstrates that Yar is indeed required for sleep regulation in Drosophila [13,14]. Although the specific mechanism of regulation of the sleeping behavior in Drosophila by Yar is not clear, it is speculated to affect the stabilization or translation of a wide range of cytoplasmic mRNAs in the cell. The lncRNA produced by the iab-8 gene represses abd-A expression in neural cells by either producing a microRNA (miR-iab-8) or by transcriptional interference of RNA pol II at the 3 end of the iab-8 gene, which overlaps with the abd-A promoter [15]. Loss of iab-8 lncRNA causes sterility in both male and female flies, inducing behavioral defects. Knocking down iab-8 lncRNA expression affects the bending of the male abdomen and thereby prevents copulation with female flies, while female flies with knocked down iab-8 lncRNA cannot pass eggs through the oviduct [15].
A genetic screen for identifying the genes that interact with dFIG4, one of the causative genes for Charcot-Marie-Tooth disease (CMT), identified CR18854 lncRNA as a suppressor of CMT phenotypes in Drosophila [16,17]. The CR18854 gene, located at the 30D1 site, encodes a 2566 base-long hairpin RNA that generates endogenous short-interfering RNA [18]. The CR18854 lncRNA binds with Hrp59 and Staufen RBPs, and thus has widespread effects on gene expression at the post-transcriptional level. Loss of CR18854 lncRNA suppresses the enlarged lysosome phenotype induced by the fat body-specific knockdown of dFIG4 [18]. The CR18854 gene also shows a genetic interaction with Cabeza (dFUS), a Drosophila homolog of human FUS and one of the causing genes of amyotrophic lateral sclerosis (ALS) [18]. Another lncRNA, CR43467, was also identified as a suppressor of the rough-eye phenotype of dFIG4 and as a rescuer of the enlarged lysosome phenotype induced by fat body-specific knockdown of dFIG4 [19]. The lncRNAs of the hsrω gene also interact with the dFIG4 in neurons, which suggests that CR18854 lncRNAs and hsrω lncRNAs may interact with common factors such as Cabeza in CMT and ALS pathogenesis [16,17].

Role of Hsrω lncRNAs in Neuron Development and Neurodegenerative Diseases
The heat-shock RNA omega (hsrω) is a developmentally active, heat-shock inducible, noncoding gene conserved in all the known species of Drosophila [20][21][22]. Originally, the hsrω gene was believed to be composed of two exons, E1 (~475 bp) and E2 (~750 bp), separated by an intron (~700 bp), followed by a long stretch of 280 bp of tandem repeats extending for~5 kb to~15 kb of length ( Figure 1) [23,24], and it was believed to produce two primary transcripts, hsrω-n1 and hsrω-pre-c, after splicing form hsrω-n2 (nuclear) and hsrω-c (cytoplasmic) transcripts, respectively [23,[25][26][27]. The hsrω-c transcript localizes to cytoplasm while the hsrω-n1, hsrω-n2 and hsrω-pre-c transcripts remain confined to the nucleus. The hsrω-c, hsrω-n1, hsrω-n2 and hsrω-pre-c have been renamed at the FlyBase (www.flybase.org) as hsrω-RA, hsrω-RB, hsrω-RG and hsrω-RC, respectively ( Figure 1). Based on the high throughput sequencing data, FlyBase has added three additional RNAs (hsrω-RD, hsrω-RF and hsrω-RH) and three miRNAs to the list of noncoding RNAs produced by the hsrω gene ( Figure 1). Further investigations on the expression profiles and functions of these newly identified hsrω transcripts are required for a better understanding of the functions and significance of the hsrω gene and its conservation in Drosophila species.
x FOR PEER REVIEW 3 of 11 produced by the hsrω gene ( Figure 1). Further investigations on the expression profiles and functions of these newly identified hsrω transcripts are required for a better understanding of the functions and significance of the hsrω gene and its conservation in Drosophila species. Most studies on the hsrω gene have so far focused primarily on functions of the 280 bp repeats containing nucleus-restricted hsrω-n lncRNAs (hsrω-RB and hsrω-RG); since there has been no specific information on the longest hsrω-RF transcript, it is not known if its functions overlap with those of the hsrω-RB and hsrω-RG transcripts. Since the 280 bp repeats have been found to be localized by in-situ hybridization, either in the omega speckles or at the site of transcription [27][28][29][30][31], it may be presumed that the hsrω-RF transcripts are also present in the omega speckles. Accordingly, the hsrω-RB, hsrω-RG and hsrω-RF transcripts are referred together as hsrω-n lncRNAs. The hsrω-n lncRNAs sequester multiple hnRNPs, chromatin remodeling factors, components of the nuclear membrane, and some disease-associated RBPs (Table 1) to form the nucleoplasmic omega speckles [27][28][29][30][31][32]. The ~1.9 kb long nuclear hsrω-pre-c lncRNA is spliced to remove the ~700 b long omega intron to produce the ~1.2 kb cytoplasmic hsrω-c (hsrω-RA) lncRNA. The hsrω lncRNAs carry a potentially translatable 23-27 amino-acid-long open reading frame (ORFω) in Exon 1 [21,23,33]. Although the sequence of ORFω is divergent, the organization of exons, ORFω and the omega intron of the hsrω gene is highly conserved in different species of Drosophila, which suggests that either translational activity of hsrω-c or the small omega peptide product may have an important role in cellular functions [21]. A systemic study on cellular functions of hsrω-c lncRNA and its omega peptide will help to explore the absolute functional potential of the hsrω gene in Drosophila. Most studies on the hsrω gene have so far focused primarily on functions of the 280 bp repeats containing nucleus-restricted hsrω-n lncRNAs (hsrω-RB and hsrω-RG); since there has been no specific information on the longest hsrω-RF transcript, it is not known if its functions overlap with those of the hsrω-RB and hsrω-RG transcripts. Since the 280 bp repeats have been found to be localized by in-situ hybridization, either in the omega speckles or at the site of transcription [27][28][29][30][31], it may be presumed that the hsrω-RF transcripts are also present in the omega speckles. Accordingly, the hsrω-RB, hsrω-RG and hsrω-RF transcripts are referred together as hsrω-n lncRNAs. The hsrω-n lncRNAs sequester multiple hnRNPs, chromatin remodeling factors, components of the nuclear membrane, and some disease-associated RBPs (Table 1) to form the nucleoplasmic omega speckles [27][28][29][30][31][32]. The~1.9 kb long nuclear hsrω-pre-c lncRNA is spliced to remove thẽ 700 b long omega intron to produce the~1.2 kb cytoplasmic hsrω-c (hsrω-RA) lncRNA. The hsrω lncRNAs carry a potentially translatable 23-27 amino-acid-long open reading frame (ORFω) in Exon 1 [21,23,33]. Although the sequence of ORFω is divergent, the organization of exons, ORFω and the omega intron of the hsrω gene is highly conserved in different species of Drosophila, which suggests that either translational activity of hsrω-c or the small omega peptide product may have an important role in cellular functions [21]. A systemic study on cellular functions of hsrω-c lncRNA and its omega peptide will help to explore the absolute functional potential of the hsrω gene in Drosophila. Preferentially present on ecdysone-induced sites, also present in omega speckles [30,35] Sxl HuR Involved in sex determination and dosage compensation, present in omega speckles [30,41] ISWI SMARCA1 Catalytic sub-unit of chromatin remodeling complex transitionally localized in omega speckles, essential for omega speckle maturation and localization [40,42] Hsp83 Hsp90 Heat-shock protein, present in omega speckles and accumulates at hsrω gene site (93D) during heat shock [43] Fatima and Lakhotia, Unpublished Nuclear matrix protein, present in omega speckles and accumulates at hsrω gene site (93D) during heat shock [31,44] Megator TPR Nuclear matrix protein, present in omega speckles and accumulates at hsrω gene site (93D) during heat shock [31,45] Since hsrω is a heat-shock-inducible gene, initial studies were based on the characterization of hsrω gene expression regulation under various cell stress conditions. RNA:RNA in situ hybridization studies suggest that hsrω lncRNAs are ubiquitously expressed with a tissue-specific variation in their levels [30,46,47]. The hsrω lncRNAs play an important role in maintaining the cell physiology as mutants with disrupted hsrω lncRNA functions, as in hsrω-null, hsrω-RNAi, ISWI-null or Hrb87F-null show delayed development, developmental lethality, short life span, reduced fecundity and reduced stress tolerance [26,31,40,48,49]. Following the discovery of a central role of these transcripts in sequestering RBPs in omega speckles, functions of hsrω lncRNAs were investigated in different biological processes that are influenced by abnormal functions of RBPs. Since abnormal functions of RBPs severely affect neuronal development and function, the roles of hsrω lncRNAs have been studied in different neurodegenerative diseases such as polyQ and ALS.

Role of Hsrω lncRNA in polyQ Expansion Disorders
The first indication of a potential role of hsrω lncRNAs in neurodegeneration was obtained in a genetic screen of the factors that regulate spinocerebellar ataxia type 1 (SCA-1) neurotoxicity in Drosophila [50]. SCA-1 is a neurodegenerative disease caused by the expansion of the polyglutamine (polyQ) tract in the ataxin-1 protein. In a genetic screen, two mutant alleles of the hsrω gene, hsrω 05241 and P292, were identified as enhancers of SCA-1-induced neurotoxicity [50]. Afterwards, UAS/GAL4-based overexpression alleles of the hsrω gene, viz. EP3037 and EP93D, were also reported to enhance the neurodegeneration caused by the expression of expanded polyQ (127Q) or of expanded huntingtin protein (Htt-ex1p-93Q) in the developing eyes of Drosophila [51]. Interestingly, a null allele of Hrb87F, an hnRNP associated with hsrω-n lncRNA in omega speckles (Table 1), also dominantly enhanced the neurodegeneration. Despite the strong genetic interaction between the hsrω gene and the polyQ-expanded transgene, neither hsrω-n lncRNA nor the associated hnRNPs were found to display any distinct colocalization with the polyQ inclusion bodies in Drosophila eye disc cells [51]. Further studies revealed that the RNAi-based selective depletion of hsrω-n lncRNAs using eye-specific GMR-GAL4 drivers dramatically suppressed the polyQ pathogenesis and restored pigmentation and ommatidial arrays in adult eyes [52]. Loss of hsrω-n lncRNA suppressed eye-specific degeneration in a variety of expanded polyQ backgrounds such as the 127Q, ataxin-1 Q82 (SCA1), MJDTR-Q78 (SCA3) or Httex1p Q93 (Huntington's disease) fly models [52]. The RNAi-mediated depletion of hsrω lncRNA reduced the polyQ aggregates without reducing the mutant mRNA levels, which suggests that hsrω lncRNAs affect either the polyQ mRNA translation or polyQ protein stability. Interestingly, down-regulation of hsrω-n transcripts had only a marginal effect on neuropathy caused by the over-expression of wild-type or mutant tau protein in flies [52], suggesting a selective role of these lncRNAs in modulating neurodegeneration [53]. Modulation of activities of the CREB-binding protein (CBP) and the Drosophila inhibitor of apoptosis protein 1 (DIAP1) by the hsrω lncRNAs have been suggested to be some of the causal factors that ameliorate polyQ toxicity [54].

Role of Hsrω lncRNAs in ALS Disease
Amyotrophic lateral sclerosis (ALS) is one of the most common adult-onset motor neuron diseases in which both upper and lower motor neurons have a progressive loss that results in muscle weakness, paralysis and premature death. Although 90-95% of ALS cases are sporadic and only 5-10% are familial, most of them have alterations in RNA metabolism. ALS is a clinical outcome of the malfunctioning of different RNAbinding proteins (RBPs) including FUS, C9ORF72, TDP-43, hnRNPA1, ATXN2, ANG and TAF15 [55][56][57]. The dysfunctional forms of these proteins get mislocalized in the cytoplasm and form pathogenic aggregates. Loss of hsrω lncRNAs in motor neurons also causes ALS-like phenotypes in Drosophila [36]. The pan-neuronal ELAV-GAL4-driven RNAi-based depletion of hsrω-n transcripts induces anatomical defects in presynaptic terminals of motor neurons that impair locomotion in larval as well as adult flies and shorten their life span [36]. Similarly, the motor neuron-specific D42-GAL4-driven loss of hsrω lncRNA reduces the terminal synapse branch length, branch number and boutons number at neuromuscular junctions (NMJ), which suggests that hsrω lncRNAs have an important role in the development of NMJ in Drosophila.
The Cabeza or dFUS protein of Drosophila is a homolog of the human FUS protein. FUS is an hnRNP with diverse roles in transcriptional and post-transcriptional regula-tion of gene expression [58]. Similar to many RBPs, the Cabeza also binds with hsrω lncRNAs in omega speckles, and depletion of hsrω lncRNA enhances its cytoplasmic localization leading to loss of Cabeza functions in the nucleus [36]. This suggests that hsrω-n lncRNAs in omega speckles act as nuclear anchors for Cabeza and thus facilitate Cabeza's nuclear functions. The role of hsrω lncRNAs in the ALS model of Drosophila was further investigated by a transgenic expression of FUS. FUS is largely a soluble protein but forms cytoplasmic inclusion bodies in a range of FUS proteinopathies such as ALS-FUS in motor neuron disease or FTLD-FUS in frontotemporal lobar degeneration, etc. The FUS-induced neurotoxicity in ALS is rescued by RNAi-mediated depletion of hsrω lncRNAs in Drosophila [59]. The GMR-GAL4-driven FUS expression in Drosophila optical neurons induces loss of pigmentation with aberrant eye morphology. These phenotypes were reversed by loss of hsrω lncRNAs, which resulted in the elimination of soluble FUS aggregates through the formation of cytoplasmic, nontoxic, insoluble inclusions of FUS-LAMP1 (lysosome-associated membrane protein 1) [59]. LAMP1 is the most abundant protein on the lysosome membrane and is used as a marker of autophagy. The insoluble inclusions of FUS-LAMP1 in hsrω-lncRNA-depleted cells are suggested to be degraded via autophagy [59]. This suggests a novel function of hsrω-n lncRNA in sequestering the toxic FUS protein from a soluble state to a harmless insoluble aggregate. To understand how hsrω regulates the localization and stability of FUS, arginine methylation of FUS, which is known to control its cellular localization and/or solubility, was investigated [60,61]. The hsrω lncRNAs were found to regulate the abundance of both type I and type II Drosophila arginine methyltransferases (DARTs) that control the methylation of FUS [62]. The arginine demethylation of FUS by DART1 and DART5 is the fundamental modification underlying the hsrω-knockdown-dependent suppression of FUS toxicity. However, how hsrω lncRNA regulates the expression of DARTs awaits further investigation.
The neurotoxicity caused by ALS-inducing factor TAR DNA-binding protein 43 (TDP-43) is also modulated by hsrω lncRNA levels. TDP-43 is a major protein associated with inclusion bodies in ALS and frontotemporal lobar degeneration (FTLD-TDP) [63]. The transgenically expressed TDP-43 binds with hsrω lncRNA in omega speckles and accumulates at the hsrω gene site, the 93D locus, during heat shock [37,38]. The level of hsrω lncRNA is upregulated~two-fold in TDP-43 over-expressing neurons, and loss of hsrω lncRNAs in hsrω-RNAi partially mitigates TDP-43 over-expression-associated neurodegeneration in eyes [37]. Studies based on the Drosophila polytene chromosomes and high-throughput ChIP-seq suggest that TDP-43 targets RNA Pol II super elongation complex (SEC) components ELL and Lilli at the hsrω gene site to elevate its expression [37,38]. This suggests that hsrω lncRNAs are functional targets of TDP-43, and therefore, altered expression of hsrω modulates TDP-43-mediated neurodegeneration. It would be interesting to examine the roles of SEC factors in the induction of hsrω lncRNA expression in polyQ or FUS-ALS disorders.

Hsrω lncRNAs as a Hub for RBPs to Modulate Neurodegeneration
Based on the experimental evidence, hsrω lncRNAs emerged as modulators of neurodegenerative diseases that are caused by abnormal RNA processing such as polyQ, FUS-ALS or TDP43-ALS. The RNAi-mediated depletion of hsrω-n lncRNAs upregulated the association of hnRNPs with CBP, enhanced the level of DIAP1 through its association with Hrb57A hnRNP and improved the proteasomal activity in polyQ diseases [28,[52][53][54]64]. Likewise, the loss of hsrω lncRNAs also enhanced the FUS-LAMP1 interaction and DARTs expression in FUS-ALS disease [36,65]. The suggestion that hsrω-n lncRNAs act as a hub for regulating the cellular dynamics of different proteins associated with RNA metabolism [28,29,31] may explain how hsrω-n lncRNAs affect localization, interactions and functions of client proteins. The mechanistic details of these events are still unknown.
The hsrω-n lncRNAs associated RBPs in omega speckles are involved in different steps of mRNA processing including splicing, export, localization, stability and translation ( Figure 2A) [27,49,[66][67][68]. The omega speckles are not formed following the loss of hsrω-n lncRNAs through hsrω-RNAi or in hsrω-null mutants or due to the absence of Hrb87F (Hrp36) hnRNP in Hrb87F-null flies [27,29]. Under these conditions, the omega speckleassociated RBPs get released in the nucleoplasm leading to an increase in the pool of free RBPs in functional compartments [31,53]. This increase in the titer of free RBPs is suggested to change the RBPs' interactome in the nucleoplasm [28,29,31], which would ultimately globally modulate the RNA metabolism in the cell ( Figure 2B). The primary cause of ALS is a reduction in the amount of functionally active forms of FUS or TDP-43 in the nucleus. Loss of hsrω-n lncRNA releases these RBPs from omega speckles so that improvement in their availability for nuclear RNA processing rescues the neurotoxicity. Regulation of the demethylation of FUS by hsrω lncRNAs [62] is another path through which these lncRNAs affect neurotoxicity ( Figure 2B). These studies suggest that hsrω-n lncRNAs act at multiple levels to modulate the functions of RBPs in neurodegenerative diseases. cause of ALS is a reduction in the amount of functionally active forms of FUS or TDP-43 in the nucleus. Loss of hsrω-n lncRNA releases these RBPs from omega speckles so that improvement in their availability for nuclear RNA processing rescues the neurotoxicity. Regulation of the demethylation of FUS by hsrω lncRNAs [62] is another path through which these lncRNAs affect neurotoxicity ( Figure 2B). These studies suggest that hsrω-n lncRNAs act at multiple levels to modulate the functions of RBPs in neurodegenerative diseases.

Future Perspectives
The current understanding is that hsrω-n lncRNAs modulate neurodegenerative disorders by regulating the dynamic availability of diverse RBPs associated with omegaspeckles (Table 1). RBPs are essential for maintaining the transcriptome through their

Future Perspectives
The current understanding is that hsrω-n lncRNAs modulate neurodegenerative disorders by regulating the dynamic availability of diverse RBPs associated with omega-speckles (Table 1). RBPs are essential for maintaining the transcriptome through their wide-ranging effects on pre-mRNA processing, mRNA transport, localization, translation and the decay of RNAs in the cell. Abnormal functions of RBPs and consequent widespread defects in RNA processing are common pathogenic underpinnings for several neurological disorders such as spinal muscular atrophy (SMA), Alzheimer's disease (AD), amyotrophic lateral sclerosis (ALS), frontotemporal dementia (FTD), multiple sclerosis (MS), congenital myasthenic syndrome (CMS) and fragile X-associated tremor/ataxia syndrome (FXTAS) [69][70][71][72][73].
A comprehensive list of proteins associated with omega speckle remains unknown. A systematic molecular approach to characterize the omega-speckle components may unravel some novel hsrω lncRNA-associated factors or pathways with roles in neuronal development and differentiation. Likewise, the roles of hsrω lncRNAs, other than the nuclear ones, in neurodegeneration are still unknown. A systematic study to examine the expression and functions of all seven known lncRNAs of the hsrω gene in neurodegenerative diseases will provide a comprehensive information about the diverse hsrω lncRNAs in RBP-associated neurodegenerative diseases.
Studies on hsrω lncRNA would also be useful to explore the therapeutic potential of Sat III lncRNA, a human functional analogue of the Drosophila hsrω lncRNA [35], which also shows elevated expression in TDP-43-mediated FTLD disease [37]. Sat III lncRNA also accumulates RBPs such as HSF1, HAP, CBP, Sam68, C2PA, 9G8, ASF/SF2 and SRp30 at the site of their transcription (9q12 at chromosome 9) during heat-shock treatment to form the nuclear stress bodies (nSBs) in human cells [74]. This indicates a similarity in functions of omega speckles and nSBs in sequestering different RBPs during stress [35,75]. The underlying mechanistic details about factors and pathways that regulate the expression and functions of hsrω lncRNAs in ALS will help in understanding the role of Sat III lncRNA in human ALS pathogenesis. Thus hsrω lncRNAs can be used as potential models to understand the mechanism through which lncRNAs are involved in either the establishment and progression or the rescue of various diseases and developmental defects.