Tau is a microtubule-associated protein involved in the maintenance of axonal stability. During AD progression, tau becomes post-translationally modified in several ways, with the most research interest directed toward changes in the phosphorylation state. Indeed, recent work has focused on excessive tau phosphorylation at specific sites as a biomarker of AD (Itoh et al. 2001, Aguillon et al. 2022, Johansson et al. 2023). Such an important role of tau in AD is noted in animal models that carry various tau mutations, such as P301L (Bukhari and Feany 2022) or R406W (Wittmann et al. 2001), amongst others. Abnormal accumulation and aggregation of tau leads to phenotypic changes in animals similar to those observed in humans. Therefore, the use of tau mutant Drosophila in the present study should be well-suited to identifying relevant molecular targets of cognitive reserve corresponding to those in AD (Bagyinszky et al. 2014, de Rojas et al. 2021).

AD is one of many causes of dementia, which is widely accepted as a progressive decline in two or more areas of cognition. Thus, individuals with AD are characterised by a range of phenotypes. Individuals suffering from very mild AD show variable deterioration of functional, cognitive and behavioural abilities (Atri 2019) that begin years before a clinical diagnosis. During the later stages, dysphoria, psychosis, aggression, wandering and improper sexual behaviour may also be observed (Mega et al. 1996). Notably, a typical phenotype is the sleep cycle disruption (Wang and Holtzman 2020), along with shifts in mood, increased anxiety, depression, withdrawal and apathy (Atri 2019), as well as locomotive dysfunction. Unlike late-onset AD, in which the medial lobe is most susceptible to atrophy, studies have suggested that the parietal and occipital cortex are most susceptible to atrophy in early-onset AD patients (Adriaanse et al. 2014). The observation of such phenotypes supplies an opportunity to use behavioural assessments (e.g. sleep patterns, motor function) to monitor Drosophila progression or degeneration.

The expression of specific genes has been suggested to be linked to cognitive reserve. Yegla and Foster (Yegla and Foster 2022) studied cognitive reserve-associated genes in the CA1 region of the hippocampus of young and old rats. Genes that have been linked to cognitive reserve include Fos, Hmgcr and Rgn. Usually, cognitive reserve genes are found to function as negative regulators of the aging process, affecting factors such as cognitive stress (Yegla and Foster 2022). Heuer and colleagues (Heuer et al. 2020) used a systems genetics approach to examine how gene expression relates to cognition in the aging process and AD. An examination of the molecular processes that affect cognitive reserve in mice found that molecular pathways related to microglia and astrocytes function as biomarkers for AD (Heuer et al. 2020). Specifically, a negative relationship between the expression of Fibroblast growth factor-2 (Fgf2) and short-term memory outcomes was observed. In contrast, Kiyota and colleagues (Kiyota et al. 2011) determined that Fgf2 may help prevent the development of AD pathology. Although studies disagree on the exact effects of the expression of Fgf2, it is clear that this gene plays a role in cognitive function during AD. In addition to Fgf2, many other genes have been associated with cognitive reserve. For example, Ramos-Miguel et al. examined components of the SNARE complex and found that STX1A, SYT12, SNAP25 and GABAergic STXBP1 were linked with stronger cognition despite AD pathology (Ramos-Miguel et al. 2021).

Cognitive reserve has also been studied with respect to other neurological conditions or pathologies in addition to AD (Afimova et al. 2013). They found that a higher frequency of the G allele of SNAP-25 was correlated with greater cognitive fitness and that a G allele might increase the likelihood of greater cognitive reserve (Afimova et al. 2013). It was also determined that the expression of CNTLN and PROK1 in the coronary arteries and the expression of PRSS50 in the atrial appendage interacted with deposition on episodic memory performance (Hohman et al. 2017). In addition, predicted gene expression levels interacted with amyloid deposition on executive functioning performance (Hohman et al. 2017). Overall, this suggests that subsets of genes impact resilience and demonstrate that cell-cycle regulation, angiogenesis and heme biosynthesis affect the progression of AD.

Many advantages exist for modelling neurodegenerative disease using Drosophila. Drosophila is simple to maintain, has a short life span and reproduces rapidly. Although Drosophila is a relatively simple organism, it engages in various complex types of behavior, such as courtship behaviour, learning, memory, social interaction, aggression, grooming and addiction (Cowan et al. 2011). Drosophila is an ideal model for studying AD because its genetics are tractable, a variety of gene disruption methods exist for Drosophila, rapid production of novel transgenic flies is possible and its genome has been sequenced and annotated. In addition, the cellular and molecular bases that drive Drosophila's biological processes are essentially the same as those of mammals, with over 75% of human disease-related genes having Drosophila orthologues (Cowan et al. 2011).

As cognitive reserve has been defined as the maintenance of good cognition despite neurodegenerative pathologies, the success of this experiment can be determined by the extent to which flies demonstrate strong cognition while continuing to carry a tau mutation. Currently, tests such as the Mini-Mental State Exam are used to assess cognitive health in humans. This particular test determines cognitive health by examining how participants perform on assessments measuring factors such as orientation, working memory, memory, visuospatial skills and language (Farias et al. 2008). To accurately gauge the cognitive health of the flies in this experiment, it would be ideal to find and assess the equivalent of each of these six factors. In this study, a T-maze for the assessment of short-term memory (Dissel et al. 2017), courtship conditioning to assess long-term memory (Broughton et al. 2003) and a heat-maze (Foucaud et al. 2010) (analogue of the Morris water maze) to assess spatial memory would be used. A review by Gistelinck and colleagues (Gistelinck et al. 2012) describes the impact of tau overexpression on olfactory learning and memory as measured by aversive phototaxis suppression that can be used to monitor tau toxicity in Drosophila that overexpress human tau. It will be important to monitor the original tau mutant using the appropriate measure of cognition that most closely aligns with the model to confirm that experimental evolution is taking place with the particular human tau isoform overexpression system being used.

Experimental evolution is a method of studying the evolutionary process by examining the genetic changes that occur when natural selection is applied to a population of organisms within a lab environment (Izutsu et al. 2016). Numerous advantages exist when using Drosophila as a model organism for experimental evolution. They have a short generation time, are simple to maintain and abundant public genomic resources exist (Burke and Rose 2009). Selection can produce physiological changes quickly because it is simple to maintain Drosophila populations with many variations (Burke and Rose 2009). Several models exist for guiding Drosophila evolution that are available for the present study. A 2016 paper by Fuse and colleagues records how researchers in Japan kept a Drosophila population under constant darkness for 1,500 generations (Izutsu et al. 2016). Flies in this population adapted to their environment and displayed phototactic behaviour and increased fecundity (Izutsu et al. 2016). Although it is possible to allow mutations to occur naturally, the process can be speeded up by inducing mutagenesis. Currently, two methods exist for subjecting Drosophila to mutagenesis. The first method uses an alkylating agent, such as ethyl methanesulphonate, or EMS, to substitute a nucleotide base (Lin et al. 2014). It is cost-efficient and helpful in producing mutations of different natures in a single gene (Lin et al. 2014). The second method is irradiation, proposed here.

Mutant Drosophila flies that express human tau protein will be used (Wittmann et al. 2001). Several mutant flies have been previously characterised to have cognitive decline. This mutation is easily monitored by the continued presence of the rough eye phenotype. The tau mutant used is associated with early-onset familial dementia and significantly shortened the lifespan of afflicted flies as compared to wild-type tau (Wittmann et al. 2001). A mutation such as this would allow the wild-type flies to rapidly outcompete the mutant flies resulting in a shorter generation time. However, any of several well-characterised tau mutations may be used if impaired cognition and rough eye phenotype are displayed (Ambegaokar and Jackson 2010).

A stock of P1 mutants to provide a supply of females is to be maintained at a temperature of 25°C, at 60% humidity, at 12-hour light/dark cycles in order to be bred to males from each generation of experimental flies created. This is to maintain the presence of dominant mutations of tau between generations. These flies will be kept on a yeast, dark corn syrup and agar diet. Given that this study will be longitudinal over many generations, an awareness and strategy for managing potentially confounding variables is warranted with excellent work available from Piper and Partridge that informed the present design (Piper and Partridge 2016). In addition, in order to account for generational effects, early emergence and to maintain age-matched adult flies, we will use the methods employed by Linford and colleagues (Linford et al. 2013).

Initially, the tau mutant male Drosophila flies will be exposed to irradiation to initiate a non-selective mutation pattern. The goal is to expand the initial pool of mutations that may be further refined by the selective pressure of competing with wild-type flies for limited food/water. As a result of irradiation, it is possible that the tau mutation is lost. This is monitored by removing any flies that lose the rough eye phenotype from inclusion in the study.

Population cages will be used to each house 5,000 total flies. Population cages may be constructed on-site or purchased from Carolina Biological Supply Company. Two primary cages will be used in this experiment. Each cage will consist of a combination of 2,500 AD mutant male flies carrying a tau mutation and 2,500 sterile male wild-type Drosophila. Flies will be introduced into the cage environment as adults. The cage environment will be made as natural as possible, with limited food and water. Each cage is monitored twice per day to remove and count dead flies.

The control cage will contain Population 1 (the original, non-irradiated tau mutant fly). Mutant flies that survive to the 90-99% threshold are removed from the cage and bred with females from the initial stock of tau mutants. The resulting generation will repopulate the population cage and compete with a new generation of wild-type male flies. Population 1 will represent natural evolutionary pressure and provide a control for changes in lifespan that may be caused by variables like breeding effects or other confounds (Jafari 2010, Promislow et al. 2022).

The experimental cage will contain Population 2 (the irradiated mutant flies). Population 2 will also consist of 2500 mutant flies that compete against 2500 wild-type flies for limited food and water. When between 90 and 99% of the mutants have died, the surviving mutants will be removed and bred with females from the original stock of tau mutants. Half of the resulting generation of mutants will be irradiated to introduce new mutations. No mutagenesis will be induced in the remaining half of the generation. Both the irradiated and non-irradiated flies will be placed together in the population cage to compete against a new generation of wild-type males. As a result, the experimental population cage will consist of a mix (1:1) of the offspring of surviving male mutant flies that are not irradiated and the offspring of surviving male mutant flies that are irradiated (introducing additional mutations). This process of repeated irradiation and mixing 1:1 of the irradiated and non-irradiated flies will occur with every generation. Thus, new mutations will regularly be introduced to compete with the prior group of successful flies. Observation of the rough eye phenotype will be used to assess mutation maintenance in AD mutant flies. See Fig. 1 for a graphical representation of the method to be employed to drive experimental evolution over generations of directed evolution.

Figure 1.  

Flow chart representation of the evolutionary process proposed. The initial population of tau mutant flies is divided into two groups, Population 1 and Population 2. Population 2 functions as a control, representing evolution in the absence of additional mutations, but under selective pressure. Population 2 is placed into a population cage that includes 2500 tau mutant flies that will compete with an equal number of wild-type flies for limited food and water. When between 90 and 99% of the mutants have died, the surviving mutants are bred with female flies from the initial stock to produce a new generation. The males of this new generation are placed back into the population cage to compete against a new group of wild-type male flies.

Population 1 functions as the experimental population. Two thousand five hundred tau mutants from Population 1 will be irradiated to induce random mutagenesis on the tau mutant background. These 2500 flies will then be placed into a population cage to compete against an equal number of wild-type flies. When between 90 and 99% of these mutants have died, the surviving mutants are bred with the female flies from the initial stock. Two thousand five hundred flies will be removed from the resulting generation and divided into two equal groups. One group will receive treatment with irradiation before being placed into the population cage and the second group will receive no additional mutagenic treatment before being placed into the population cage. Together, these two groups will compete against 2500 new wild-type male flies for limited resources. Such a strategy ensures that new mutations are introduced while increasing the likelihood that beneficial mutations that have already occurred are maintained in the population. This process is then repeated until increased longevity is seen based on statistical comparison.

Multiple techniques are available for assessing cognition. Aversive phototoxic suppression (T-maze) has been used by others (Seugnet et al. 2008,Dissel et al. 2017) to determine short-term memory. This method is based on the flies' ability to remember which chamber in a T-maze holds an aversive stimulus (e.g. an illuminated vs. a darkened chamber). Cognition may be assessed by examining long-term memory using courtship conditioning in flies (Broughton et al. 2003). Spatial organisation may be used to assess its level of reserve. An example of this using a mouse model would be the Morris water maze. A 2015 study (Granger et al. 2016) examined sex-based differences in cognitive reserve in a mouse model using the Morris water maze. The researchers associated reserve with using effective navigational strategies to complete the maze successfully. Male mice more efficiently switched to spatial learning strategies instead of systematic learning strategies, allowing them to better compensate for amyloid pathology (Granger et al. 2016). Analogues for the Morris water maze exist that use a heat maze to assess spatial learning in Drosophila. For example, it was shown that Drosophila use visual cues to increase their navigational efficiency (Foucaud et al. 2010). See Fig. 2 for a graphical representation of the generalised scheme to assess successful generational improvements in cognition using the assessments described.

Figure 2.  

Flow chart representing the assessment of improved longevity. Increased longevity may be a result of mutations that have induced cognitive reserve or it may be a result of mutations linked to genes impacting longevity itself. The extent to which increased lifespan may be attributed to improved cognition may be assessed using a T-maze or courtship rituals. If improved cognition is the driving factor behind increased longevity, then next-generation sequencing is used to determine which mutations have impacted cognition. If longevity is increased, but cognition is not improved, then further testing and sequencing are required to determine which genes have been affected by the process of mutagenesis.

Short-term memory will be assessed through Aversive Phototaxic Suppression (Dissel et al. 2017). First, flies will be placed in a T-maze and choose between light and dark rooms. Then, an adverse stimulus (quinine/humidity) is paired with the light room. Over a series of trials, the phototaxis of the fly is observed. Ideally, experimental evolution will drive mutant flies to have corrected wild-type like memory.

The long-term memory of the flies will also be assessed through analysis of courtship conditioning. This method was employed by Dissel and colleagues (Dissel et al. 2017) and it is based on exposing a male fly to a non-receptive female. This same fly is later exposed to a receptive female to examine his associative memories.

To determine the extent to which the experimental system drives evolution towards improved cognitive reserve, survival curves will be used as an end-point measure. Fig. 3 shows a theoretical set of expected survival curves as experimental evolution drives revertants to become more like wild-type. Survival curves in aging studies across animal models, including Drosophila, are well-established as appropriate indicators of change in longevity. In addition, the method of data collection proposed includes the possibility of other measures of evolutionary success, including median lifespan, mean lifespan and maximum lifespan, amongst others (Blumenstiel et al. 2009, Haelterman et al. 2014, Ziehm et al. 2015). Progress towards increased survival will be assessed using the Kaplan-Meier estimation method. A Kaplan-Meier curve will be constructed for each generation of mutant flies and compared to the control group using a log-rank test with significance set to p = 0.05. The control comparison is the generational matching group of the original isogenic tau mutant fly that is maintained with the same generational frequency. By selecting and propagating only those mutant flies that survive in a population of wild-type, it is expected that the genetic variables that allow cognitive reserve will be enriched with each generation.

Figure 3.  

Simulated data for the change in survival rates over time. As the generations increase from very early to late, we expect that the survival of the irradiated flies will increase and look more like the curve for wild-type flies. Observation of a trend towards statistical insignificance compared to the initial mutant fly would strongly indicate that the population was evolving (the curve is shifting to the right). The determination that the population has evolved to be like wild-type will be at the point in which there is no statistical difference (p > 0.05) between the wild-type and mutant flies as illustrated as Ready to Validate.

Using next-generation sequencing methods, the genomes of these surviving Drosophilas can be compared to ascertain the mutations responsible for cognitive reserve development. In this study, the flies resulting from experimental evolution would be pooled and their genome sequenced and compared to the original tau mutant to identify the genes altered by experimental evolution. Sequencing technologies, such as those by Oxford Nanopore, Pacific Biosciences and others, when combined with various algorithms to assemble such data, have the ability to provide complete genome information (Ludwig et al. 2022). Examples in which Drosophila genomes have been compared include work by Blumenstiel and colleagues (Blumenstiel et al. 2009), who used Illumina next-generation sequencing to identify EMS-induced mutations affecting eggshell morphology in Drosophila. Similarly, Lee and colleagues (Lee et al. 2016) used whole genome sequencing to identify EMS-induced mutations affecting competition between wild-type and RpS3 heterozygous cells. Feng and colleagues have also successfully explored the genomic differences amongst various Drosophila species using NGS methods (Feng et al. 2023). While the approach is focused on whole genome comparisons, with the prior knowledge of putative cognitive reserve genes, more specific comparisons could also be completed. The successful completion of this approach would support more traditional experimental approaches to determine the mechanistic basis and potential pharmacological strategies to regulate the predicted molecular pathways.

We have suggested a novel approach to understanding the genetic pathways that underlie cognitive reserve. Although we have focused on AD, this approach could be applied to other neurodegenerative disorders with an accurate animal model. In particular, with evidence for the existence of cognitive reserve in Dementia with Lewy Bodies and Frontotemporal Dementia (Perneczky et al. 2007, Borroni et al. 2009, Carli et al. 2021), researchers could apply the same method. Results from such studies would provide insights into the specific molecular pathways that regulate the different phenotypic expressions of these diseases as they relate to cognitive decline or reserve. The power of experimental evolution could similarly be used by selection of animals that demonstrate increased resistance to the cognitive decline accompanying the aging process. Such a finding would profoundly impact our understanding and ultimately reveal cellular processes that underlie age-related cognition decline.