Behavioural Effects and RNA-seq Analysis of Aβ42-Mediated Toxicity in a Drosophila Alzheimer’s Disease Model

Alzheimer’s disease (AD) is the most common neurological ailment worldwide. Its process comprises the unique aggregation of extracellular senile plaques composed of amyloid-beta (Aβ) in the brain. Aβ42 is the most neurotoxic and aggressive of the Aβ42 isomers released in the brain. Despite much research on AD, the complete pathophysiology of this disease remains unknown. Technical and ethical constraints place limits on experiments utilizing human subjects. Thus, animal models were used to replicate human diseases. The Drosophila melanogaster is an excellent model for studying both physiological and behavioural aspects of human neurodegenerative illnesses. Here, the negative effects of Aβ42-expression on a Drosophila AD model were investigated through three behavioural assays followed by RNA-seq. The RNA-seq data was verified using qPCR. AD Drosophila expressing human Aβ42 exhibited degenerated eye structures, shortened lifespan, and declined mobility function compared to the wild-type Control. RNA-seq revealed 1496 genes that were differentially expressed from the Aβ42-expressing samples against the control. Among the pathways that were identified from the differentially expressed genes include carbon metabolism, oxidative phosphorylation, antimicrobial peptides, and longevity-regulating pathways. While AD is a complicated neurological condition whose aetiology is influenced by a number of factors, it is hoped that the current data will be sufficient to give a general picture of how Aβ42 influences the disease pathology. The discovery of molecular connections from the current Drosophila AD model offers fresh perspectives on the usage of this Drosophila which could aid in the discovery of new anti-AD medications.


Introduction
Biolo gically, ageing is the progressive accumulation of various behavioural, anatomical, cellular, and molecular damage that results in the deterioration of physical and mental capabilities in addition to the increased risk of contracting age-related diseases and eventual death [1]. Today, societies worldwide are experiencing an increase in lifespan, with a global estimated life expectancy of 72.6 years and 67.2 years for females and males, respectively [2]. In 2020, the world population of older adults aged 60 years and above was reported to be 1 billion, and this is expected to reach 2.1 billion by 2050 [3]. Due to the continuously surging number of older adults, age-associated neurodegenerative disorders are becoming serious public health issues.
Alzheimer's disease (AD) is the most widespread neurodegenerative disease, covering at least 70% of dementia cases, and is associated with the irreparable and gradual decline of memory and cognitive functions as well as other symptoms such as motor deterioration and behavioural fluctuations [4]. As a multifactorial disease, there are many hypotheses behind the pathogenesis of AD. The most established principle involves the aggregation of amyloids, referred to as the amyloidogenic pathway. Briefly, the amyloid precursor protein (APP) is cleaved by β-secretase and subsequently by γ-secretase to excrete functional Aβ peptides. The two most common types of Aβ peptides are Aβ40 and Aβ42, with Aβ42 being more prone to aggregation and more toxic [5]. The aggregation of Aβ peptides results in oligomerization, fibrilization, and the subsequent formation of plaques. It is the build-up of these amyloid plaques that have become the hallmark occurrence of AD [6]. Other theories include the tau hypothesis, which states that the abnormal phosphorylation of tau proteins forms neurofibrillary tangles in the brain that interfere with neuronal function, and the oxidative stress hypothesis which suggests that oxidative injury to neurons and brain cells is responsible for the pathological changes in AD [7,8]. Regardless of available knowledge, the actual mechanism behind the disease remains unclear.
While it is irrefutable that human genetic exploration has enhanced the understanding of genes linked to AD, studies on human subjects are hindered by ethical and technical restraints. Therefore, animals are employed to mirror human diseases. The Drosophila melanogaster, otherwise known as the fruit fly, is commonly used to extensively analyse AD [9]. The model boasts characteristics such as a short lifespan, easy maintenance, a fully sequenced genome, and simple anatomy [6]. Out of 287 identified human disease genes, D. melanogaster carries homologs for at least 197 (69 %) of them [10]. Most importantly, the D. melanogaster is able to recapitulate many age-dependent behavioural aspects such as memory and motor capacities [11]. Interestingly, wild-type D. melanogaster lacks the Aβ gene region and does not produce the peptide; therefore, transgenic Drosophila inserted with the Aβ gene are perfect models to study AD amyloidosis [9]. Recently, D. melanogaster has been extremely useful in AD therapeutic discovery, such as probiotic studies [12,13] and polyphenol testing [14,15].
The transcriptome is the collection of the total RNA molecules in either a cell or a cell population. Distinct from the genome, where the content is generally fixed, the transcriptome is constantly shifting according to external environmental circumstances. By analyzing the transcriptome, differentially expressed genes (DEGs) that show statistically significant changes in expression levels between two or more conditions or treatments are identified. Although D. melanogaster has been proven to be an ideal model for AD, transcriptome research on the Drosophila AD model is limited. So far, recent AD studies using Drosophila revealed that cryptic slicing errors, in addition to dynamic, age-dependent activity between the brain transcriptome and proteome are triggered by tauopathy [16,17] while histone acetylation dysregulation directly contributes to amyloidosis [18]. Besides, available transcriptome studies on Drosophila utilized microarray techniques which are more prone to experimental inconsistency and background noise besides being incapable of fully sequencing the whole transcriptome compared to RNA-sequencing (RNA-seq) technology [19].
Thus, in this study, we propose a Drosophila AD model carrying the human Aβ42 gene as a suitable platform to investigate the behavioural changes in "young," "middle," and "old" age phases of AD Drosophila. Furthermore, we performed a comprehensive evaluation on "old" AD Drosophila compared to the Control Drosophila of the same age that is devoid of the Aβ42 gene to identify DEGs associated with the amyloidogenic pathway.

Drosophila Husbandry
All Drosophila stocks employed in this study can be found in Flybase (http:// fybase. bio. india na. edu). The wild type Oregon-R which is devoid of the Aβ42 gene (#5), the pan-retinal Glass Multiple Reporter-GAL4 (GMR-GAL4) (BDSC#1104), the ubiquitous driver Actin5C-GAL4 (BDSC#4414) and the UAS-Aβ42 ( Control lines were obtained by crossing the wild-type Oregon-R with the specific GAL4 line of the specific experiment. On the other hand, the AD Drosophila lines were obtained by crossing UAS-Aβ42 with the same GAL4 line to produce transgenic GAL4-Aβ42 Drosophila that expressed Aβ42 in the resultant tissue. For the neurotoxicity assay involving the adult eye, GMR-OreR was used as the Control while GMR-Aβ42 is the AD Drosophila line. On the other hand, the survivability analysis, negative geotaxis assay, and RNA-seq utilized Actin5C-OreR as the Control as well as Actin5C-Aβ42 as the AD Drosophila line.

Adult Eye Imaging for the Neurotoxicity Assay
For scanning electron microscopy (SEM), whole Drosophila from the experimented group were fixed overnight at 4 °C in McDowell-Trump fixative (Sigma-Aldrich, MO, USA). The samples were then rinsed with phosphate buffer and post-fixed in 1 % osmium tetroxide (Sigma-Aldrich, MO, USA) for an hour. The samples were rinsed again with distilled water and dehydrated in a series of increasing ethanol concentrations for 15 min each and in hexamethyldisilazane (Sigma-Aldrich, MO, USA, Cat no.: 440191) for 10 min before being left to air-dry overnight in a desiccator. Samples were mounted and gold-coated before being observed under SEM (Hitachi Ltd., Tokyo, Japan). The degree of ommatidia deformation was computed from each SEM image through Flynotyper (https:// flyno typer. sourc eforge. net) which quantifies the phenotypic score (P-score) ratio, which is a measure of eye deformation severity. The experiment was done with three males and three females per experimental Drosophila line.

Survivability Analysis and Negative Geotaxis Assay
The CApillary FEeder (CAFE) [20] assay was used throughout the survivability analysis and negative geotaxis assay. Liquid feed was prepared similarly to solid feed with the exception of cornmeal and agar. Tested Drosophila were gathered within 24 h of eclosure and transferred to the CAFE assay with a maximum of 10 flies per vial. Males and females were separated during the transfer. Dead Drosophila were counted while remaining live Drosophila were moved to clean vials with fresh liquid feed daily. The maximum lifespan indicates the day when the last Drosophila in the cohort dies. The experiment was carried out in triplicates with n≃50 Drosophila per replicate for male and female samples each while combined samples are pooled samples of 50 males and 50 females per replicate. Lifespans of experimental cohorts were compared to Actin5C-OreR.
For the negative geotaxis assay, males and females were tested separately. Approximately 10 experimental Drosophila were transferred into the climbing tubes and allowed to acclimatize to the surroundings for 10 min. Next, the climbing tubes were knocked against a flat surface three times to knock down the Drosophila. The Drosophila were given 10 s to scale the wall of the tube. The experiment was video recorded and analysed using the online software Toxtrac [21]. The experiment was executed in triplicates.

Total RNA Extraction and RNA-Sequencing
Total RNA samples of the specific Drosophila lines were extracted using a combination of TRIzol (Invitrogen, MA, USA) and the RNeasy Mini kit (Qiagen, Hilden, Germany). Briefly, ten Drosophila samples were homogenized in 300 μL of TRIzol. Subsequently, 80 μL of chloroform was added and the mixture was vortexed for 20 s. The mixture was centrifuged for 10 min at 10,000 g at 4°C to separate the layers. The aqueous layer was transferred to a sterile tube and added with 200 μL of absolute ice-cold isopropanol with careful resuspension. Sample clean-up was carried out using the MinElute Cleanup Kit (Thermo Fisher Scientific, MA, USA) followed by TurboTM DNase Kit according to the manufacturer's protocol. The total RNA quality was evaluated via agarose gel electrophoresis, NanoDrop 2000/2000c (Thermo Fisher Scientific, MA, USA), and 2100 Bioanalyzer (Agilent, CA, USA). High-quality total RNA (≥ 5 μg; ≥ 200 ng/μL and OD 260/280 = 1.8 to 2.2) were employed for library construction. Each replicate included five males and five females. Library construction and sequencing were performed by Apical Group (Singapore). The libraries were generated using TruSeq® RNASample Preparation (Illumina, CA, USA). The transcriptomic samples from all conditions were sequenced via the Illumina HiSeq PE150 platform. Raw data (raw reads) were processed through In-house Perl scripts. Library construction criteria include Q20 > 95 %, Q30 > 90 %, an error rate of 0.01 %, GC content between 30 and 70 %, effective rate > 90 %, and a sequencing depth of >30 million reads per sample.

Differential Expression Analysis
The reads were uploaded onto the Galaxy server (https:// usega laxy. org/). The BioProject ID for the submitted reads in the database is PRJNA921963. Ascension numbers for Control (Actin5C-OreR) samples are SRR23047823, SRR23047822, and SRR23047821. Ascension numbers for Aβ42 (Actin5C-Aβ42) samples are SRR23047820, 1 3 SRR23047819 and SRR23047818. The sample reads that passed the quality check were mapped to the reference genome for Drosophila melanogaster (https:// zenodo. org/ record/ 11851 22/ files/ Droso phila_ melan ogast er. BDGP6. 87. gtf) via RNA STAR (Galaxy Version 2.7.8a+galaxy0) using the "Paired-end" and "unstranded strand" options [22]. The programme featureCounts (Galaxy Version 2.0.1+galaxy2) was used to count the number of reads mapped to each gene [23]. The count outputs from featureCounts were compiled into a single matrix file in csv format. The DESeq2 R package version 3.6.3 was used to perform the differential expression analysis [24]. A filtering criterion with a threshold of more than 5 counts for each gene for at least 2 samples was applied. Genes with adjusted P values (padj) of less than 0.05 and log2FoldChange (log2FC) of more than 1 and less than 1 were determined as DEGs.

GO and KEGG Enrichment Analysis of DEGs
The online DAVID Knowledgebase v2022q3 (https:// david. ncifc rf. gov/) [25] was used to obtain the Gene Ontology (GO) terms, GO clusters, and Kyoto Encyclopedia of Genes and Genomes (KEGG). The DEGs were characterized into three GO categories: biological processes (BP), cellular components (CC), and molecular functions (MF). REVIGO (http:// revigo. irb. hr/) was used to visualize the GO terms [26]. The KEGG (Kyoto Encyclopedia of Genes and Genomes) webserver is a database that aids in the comprehension of high-level functions and roles of the biological system using molecular-level data generated by high throughput systems [27].

Quantitative Reverse Transcription PCR
To validate the transcriptomic analysis, quantitative realtime PCR (qRT-PCR) was carried out using five upregulated DEGs and five downregulated DEGs along with a housekeeping gene. Table 1 shows the list of qRT-PCR primers used for the 10 DEGs and one housekeeping gene in this study. First-strand complementary DNA (cDNA) was synthesised from total RNA using the iScript™ Reverse Transcription Supermix (Bio-rad, CA, USA) according to the manufacturer's protocol. Next, qRT-PCR was performed using iTaq Universal SYBR Green Supermix (Bio-rad, CA, USA) in a CFX96 real-time thermal cycler (Bio-rad, CA, USA) following protocol from the manufacturer. The experiment was done in triplicate.

Aβ42 Expression in the Eye Resulted in Severe Rough Eye Phenotype
The Drosophila eye morphology was analysed as a means to understand the neurodegenerative consequences of Aβ42 expression in the eye neurons. A common Drosophila eye is made up of 800 hexagonal elements called ommatidia which are positioned in an accurate crystalline arrangement to construct a concave form as depicted in Fig. 1 Ai for the Control GMR-OreR eye. Since each ommatidium has eight photoneurons, any malformations in the eye morphology could be seen as deformities in the neurons [33]. These distortions in the eye morphology are termed the rough eye phenotype (REP). Contrariwise, GMR-Aβ42 eyes ( Fig. 1B and 1Bii) demonstrated merged ommatidia with holes, which bring about a "glazed" morphology compared to GMR-OreR ( Fig. 1 Ai and 1Aii). P-score ratios of the eyes were assessed, with higher P-score ratios (over 1.5) indicating a more severe REP (Fig. 1C). GMR-Aβ42 eyes had a significantly higher (P=0.0067) P-score than GMR-OreR; thus, displaying Aβ42's neurotoxicity on the AD Drosophila eyes.

Drosophila-Expressing Aβ42 Demonstrated Shortened Lifespan
Next, the long-term expression of Aβ42 on AD Drosophila was investigated. This was achieved using the Actin5C-GAL4 driver, which promotes ubiquitous expression in muscle tissues. For females ( Fig. 2A) and males (Fig. 2B), Actin5C-Aβ42 lines had shorter lifespans compared to the control Actin5C-OreR. From Table 2, male Actin5C-Aβ42 had a maximum lifespan of 19 days after eclosure which was 51.66 % shorter than the control that lived to a maximum lifespan of 41 days after eclosure. Similarly, female Actin5C-Aβ42 experienced a 43.90 % reduction in maximum lifespan of 23 days after eclosure compared to the Control which had a maximum lifespan of 41 days after eclosure. This confirmed that Aβ42 negatively affects the lifespan of AD Drosophila.
To visualize the different phases of the Drosophila lifespan reared using the CAFE assay, the lifespan data of females and males were pooled to give a combined lifespan graph (Fig. 2C). This average total lifespan was divided into three stages: "young," "middle," and "old" age phases with boundaries for Actin5C-OreR at day 12.5 (33.3% mortality) and day 23 (66.86% mortality) while boundaries for Actin5C-Aβ42 were at day 7.5 (35.0% mortality) and day 11.5 (68.5% mortality). By differentiating the three phases, three time points based on the Actin5C-Aβ42 graph were pinpointed for the negative geotaxis analysis: days 5 (15.92% mortality), 10 (57.32% mortality), and 15 (84.08% mortality) after eclosure were chosen to represent "young", "middle", and "old" age phases, respectively. The three time points were used for the negative geotaxis assay.

Aβ42-Expressing Drosophila Had Impaired Climbing Abilities
While cognitive degradation and AD have always been linked, the condition is also marked by a decrease in motor coordination that worsens with age [34]. When knocked to the bottom of the vial, Drosophila have an innate negative geotaxis response in which they strive to scale the wall of their surroundings. This escape response is one of the numerous Drosophila behaviours that deteriorate with age, making this the ideal analysis to effectively research ageing and locomotion loss in relation to Aβ42 expression [35].
By culturing Drosophila using the CAFE assay, the phenotypic differences between Control Actin5C-OreR Drosophila and AD Drosophila were observed. Overall, males of Control (Fig. 3A) were faster with a climbing average speed of 15.2 mm/s (5 days after eclosure), 18.1 mm/s (10 days after eclosure), and 19.9 mm/s (15 days after eclosure) than their female counterparts (Fig. 3B)

RNA-seq Analysis
Following the results from the survivability analysis and negative geotaxis assay, the time point of 15 days after eclosure was chosen for RNA sample collection. Two different sample groups composed of Actin5C-OreR (denoted as control) and Actin5C-Aβ42 (denoted as Aβ42) with three biological replicates each were collected during this time point. The quality and quantity of the RNA samples are shown in Table 3.
To obtain an overview of the transcriptome changes between the Control and Aβ42 samples, a principal component analysis (PCA), heat map, and volcano plot were performed (Fig. 4). The PCA (Fig. 4A) and heat map (Fig. 4B) graphs depicted a clear clustering of the Control samples as well as another clustering for the Aβ42 samples. The volcano plot (Fig. 4C) showed that many genes from the Aβ42 samples were significantly differentially expressed compared to the Control samples with the most statistically significant genes scattered at the top, and the most upregulated genes are labelled red to the right while the most downregulated genes are labelled red to the left. From a total of 13585 genes, 1496 (11.01 %) genes were significantly differentially expressed (padj > 0.05) between the Aβ42 samples against the Control samples. Out of these DEGs (Supplementary Data 1), 1398 (10.29 %) were upregulated and 98 (0.72%) were downregulated.

GO and KEGG Analyses
The genes that showed a substantial difference in expression levels were then categorised according to the GO terms. A total of 298 GO terms (Supplementary Data 2) were obtained from the upregulated DEGs with 163 GO terms under BP (Fig. 5A), 66 CC terms (Fig. 5B), and 69 MF terms (Fig. 5C). On the other hand, downregulated DEGs had a total of 38 GO terms, of which 9 were BP terms (Fig. 5D), 4 were CC terms (Fig. 5E), and 25 were MF terms (Fig. 5F).
These upregulated and downregulated GO terms were clustered to get a more precise outlook on the GO terms. A total of 23 clusters were obtained from upregulated DEGs.

Validation Via RT-qPCR
To corroborate the RNA-seq data, we performed RT-qPCR to assess the expressions of 10 genes: five upregulated and five downregulated. All evaluated genes showed profiles that were similar to the RNA-seq (Fig. 6A). Pearson correlation (Fig. 6B) was used to test for the correlation between the RNA-seq data and qRT-PCR. The coefficient of determination (R 2 ) which identifies the strength of correlation between the two data sets was above 90 % (R 2 =93 %), indicating a high level of correlation; therefore, validating the accuracy of the RNA-seq results.

Discussion
The development of Aβ plaques in the brain is a defining feature of AD. The major target in this investigation was Aβ42, which is recognised as the most aggressive and toxic class of Aβ peptides. Using the Drosophila melanogaster as the model, three main characteristics of AD consequences were chosen: neuron degradation [36], premature death [37], and loss of mobility [38]. These characteristics correlate to three experiments: (1) the neurotoxicity analysis that observed the compound eye structures of the Drosophila, (2) the survival assessment that assessed the life expectancy of the Drosophila, and (3) the negative geotaxis assay that evaluated the locomotive capabilities of the Drosophila. Expression of Aβ42 in the adult Drosophila retina through the GMR-GAL4 driver distorted the eye structures, leading to the REP [33]. This was apparent in the current findings whereby GMR-Aβ42 eyes had irregularly shaped ommatidia that were fused and perforated, giving the eye a "glazed" exterior compared to the Control GMR-OreR that had uniformly-shaped ommatidia arranged in even columns across the eye, resulting in a concave and "egg-like" appearance. Likewise, the Actin5C-Aβ42 demonstrated decreased lifespan compared to the control Actin5C-OreR, which recapitulated the disease impact in humans. Based on the lifespan graph of total Actin5C-Aβ42, days 5, 10, and 15 after eclosure correspond to the "young," "middle," and "old" age phases, respectively of the AD Drosophila and were selected as time points for the negative geotaxis assay. Notably, the control Actin5C-OreR was at a different age phase during these three time points. In fact, Actin5C-OreR was still in its "young" age phase on days 5 and 10 after eclosure whereas Actin5C-OreR was in its "middle" age phase on day 15 after eclosure. It might be due to its younger age phase that Actin5C-OreR exhibited increasing average speeds throughout the 15 days while the "older" Actin5C-Aβ42 experienced consistent decrements in average speeds during the course of the three time points. This established that Aβ42 has an effect on locomotive capabilities when ubiquitously expressed in Drosophila muscles. Additionally, it is possible that Aβ42 hastened the ageing process of the AD Drosophila.
The transcriptome of Actin5C-Aβ42 was then analysed in comparison to the Control Actin5C-OreR. The climbing abilities of the two Drosophila lines were the most significantly different on day 15 after eclosure; as such, this time point was selected for transcriptomic analysis. Among the upregulated KEGG pathways were carbon metabolism as well as the glycolysis pathway. It has been shown that Aβ is dispersed variably in different areas of the brain, with greater deposition detected in regions that rely heavily on glycolysis [39]. Key enzymes associated with the pathway were found to be elevated in AD. For instance, post-mortem brains of AD patients showed increased phosphofructokinase and pyruvate kinase activities [40]. In CHME-5 human microglial cells incubated with AD plasma, upregulated glycolysis enzymes included phosphoglycerate kinase, pyruvate dehydrogenase, and pyruvate kinase [41]. This correlates with our current data where these enzymes were among the upregulated DEGs which indicate a dysregulation in gene expression and catalytic activity of multiple glycolytic enzymes in AD pathology. Moreover, it is believed that as AD worsens, increased glycolytic enzyme proteins reflect deficits in mitochondrial function [42].
As seen in the upregulated GO annotation clusters 2, 3, and 4 as well as the KEGG oxidative phosphorylation pathway, the AD Drosophila also presented upregulation of genes involved in microtubule-based processes and ATPase activities. This upregulation contradicts the tau hypothesis whereby hyperphosphorylated tau loses its function in reinforcing the microtubule structure [43]. However, subsequent examination of the DEGs revealed that the majority of these genes affect the sperm process. Axonemes, which are motile cytoskeletal structures in the tail, guide sperm motility. They consist of (9 + 2) microtubules, molecular motors (dyneins), and their regulatory structures [44]. High APP-like protein 2 (APLP2) expression levels were observed in the human sperm. APLP2 and APP were detected throughout the sperm but were most abundant at the tail [45]. As Aβ42 was expressed in all actin cells for the current transcriptomic analysis, it is likely that Aβ42 expression was also increased in the sperm. Only a few papers have so far addressed the connection between APP and sperm. Regardless, the findings of this study suggested that Aβ42 may affect sperm motility which could impact male fertility.
Besides that, the KEGG longevity regulating pathway was found to be upregulated. This pathway is a group of genes and biochemical reactions that control ageing and lifespan, for example, cellular and metabolic processes, signalling pathways, DNA repair, and stress responses [46]. In the current study, heat shock protein (HSP) genes such as Hsc70-2 and Hsp70Bb that were upregulated in this pathway. Many heat shock proteins were found to be highly expressed in older adult neurodegenerative patients and aged animals [47,48]. HSPs chaperone the conversion processes of nonnative proteins back to their native state [49]. It was posited that Hsp70 overexpression is the immune system's response to counteract Aβ aggregation [50]. Consequently, augmented HSP concentrations are used as biomarkers for conditions such as central nervous system ischemia and AD which are considered protein-misfolding diseases or proteopathy [51]. Despite that, the increase in HSP expression may be insufficient to nullify the extensive proteotoxic stress caused by AD [52]. Nevertheless, the increased HSP expression and misfolded protein GO terms in the Actin5C-Aβ42 samples against the Control implied that the occurrence of Aβ42 is a misfolded protein.
There was also an increase in expression levels of antimicrobial peptide (AMP) genes comprising attacin, diptericin, and drosocin as observed in upregulated GO annotated cluster 5. According to a previous study, the upregulation of innate immune molecules, including AMPs, rises proportionally with age and this is ascribable to the individual's increasing vulnerability to infections [53]. Congruently, healthy ageing Drosophila without the Aβ gene experience progressively increasing AMP expressions as they age [54]. The overexpression of AMP genes in Drosophila causes more damage by inducing cytotoxic effects. An over-activated central nervous system was found to be accountable for the initiation of neurodegeneration [55]. In another Drosophila AD model expressing human Aβ42 specifically in the brain, the AMP genes exhibited significantly amplified expression levels in contrast to the Control [54]. In the current study, as the Actin5C-Aβ42 and Control samples were of the same age (15 days after eclosure), the upregulation of  Fig. 6 Verification of RNAsequencing results with quantitative RT-qPCR. A Trend comparison of quantitative RT-PCR against RNA-seq of five upregulated differentially expressed genes and five downregulated differentially expressed genes. P values show significance between Aβ42 and Control at *P<0.05, **P<0.005, and ***P<0.0005. B Pearson correlation between the log2FC expressions of RT-qPCR against RNA-sequencing data AMP genes and the pathways controlling longevity in the current transcriptomic analysis suggested that Aβ42 expression accelerated Drosophila senescence which is consistent with the results of the survival assessment.
Conversely, while AMP genes were elevated, the expressions of microsome and heme via cytochrome P450 genes such as Cyp316a1 and Cyp4d8 were inversely expressed. Cytochrome P450 (CYP) enzymes aid in drug metabolism in the liver and are also active in the neurovascular unit and brain cells [56]. When a foreign body is detected, the innate immune response releases AMPs and initiates the neuroinflammation process [57]. This inflammatory mechanism works in tandem with the downregulation of the main xenobiotic metabolising CYP enzyme system [58]. In an APP23 mouse AD model, the inhibition of Cyp46a1 expression was followed by an abundance of Aβ peptides, which caused a build-up of brain cholesterol and eventually neuronal death [59]. Therefore, this study's results of upregulated AMP genes provide pieces of evidence that the innate immune response contributes to AD pathogenesis.
Similarly, in downregulated GO annotated cluster 1, protease and proteolysis activities were reduced with many genes such as CG11912 (orthologous to human chymotrypsin C) and LysX found to be downregulated. The process of conjugating ubiquitin protein with the Lys residue of a misfolded protein or damaged organelle is known as ubiquitination. The target substrate is tagged by the lysosomal pathway or the ubiquitin-proteasome system (UPS) [60]. A significant risk factor for AD is the decline in proteasome activity. This was observed in AD patient brain tissues where chymotrypsin and post-glutamyl peptidase activities were significantly lower than age-matched controls [61]. Furthermore, 20S proteasomal activity in erythrocytes was significantly reduced compared with healthy subjects [62]. Correspondingly, it was described that soluble Aβ oligomers are able to suppress proteasome activity [63]. Thus, it could be postulated that hyperphosphorylated tau and soluble Aβ forms are responsible for the decreased UPS system activity in AD.

Conclusions
By employing three behavioural assays, this work has shown that the expression of Aβ42 in Drosophila melanogaster has significant negative effects on its ocular structures, lifespan, and motor capabilities. Results from RNA-sequencing analysis suggest that Aβ42 expression dysregulates pathways related to carbon metabolism, oxidative phosphorylation, antimicrobial peptides, and longevity-regulating pathways. Additionally, the potential association between Aβ42 and male fertility was evinced. As certain AD medications show adverse reactions towards male reproductive function [64] as well as possible gender differences in human drug metabolism [65], the current data might be beneficial to AD treatment studies in formulating drugs that address these issues. Accordingly, due to its relatively straightforward genome and numerous conserved genetic pathways with humans, the transcriptome data from AD Drosophila can aid in understanding the molecular mechanisms underlying the disease and, by extension, discovering new treatment targets. In essence, AD is a complex neurological disease with various factors influencing its pathogenesis. Although there are many more theories and factors that may contribute to the aetiology, it is hoped that the current findings will be sufficient to provide an overview of how Aβ42 impacts the disease pathology.