The neuronal receptor tyrosine kinase Alk is a target for longevity.

Abstract Inhibition of signalling through several receptor tyrosine kinases (RTKs), including the insulin‐like growth factor receptor and its orthologues, extends healthy lifespan in organisms from diverse evolutionary taxa. This raises the possibility that other RTKs, including those already well studied for their roles in cancer and developmental biology, could be promising targets for extending healthy lifespan. Here, we focus on anaplastic lymphoma kinase (Alk), an RTK with established roles in nervous system development and in multiple cancers, but whose effects on aging remain unclear. We find that several means of reducing Alk signalling, including mutation of its ligand jelly belly (jeb), RNAi knock‐down of Alk, or expression of dominant‐negative Alk in adult neurons, can extend healthy lifespan in female, but not male, Drosophila. Moreover, reduced Alk signalling preserves neuromuscular function with age, promotes resistance to starvation and xenobiotic stress, and improves night sleep consolidation. We find further that inhibition of Alk signalling in adult neurons modulates the expression of several insulin‐like peptides, providing a potential mechanistic link between neuronal Alk signalling and organism‐wide insulin‐like signalling. Finally, we show that TAE‐684, a small molecule inhibitor of Alk, can extend healthy lifespan in Drosophila, suggesting that the repurposing of Alk inhibitors may be a promising direction for strategies to promote healthy aging.

conserved hallmarks that both characterize aging and have the capacity to modulate healthy lifespan (López-Otín, Blasco, Partridge, Serrano, & Kroemer, 2013). Among these hallmarks, disrupted nutrient sensing pathways and intercellular communication have been highlighted by findings that several conserved signalling proteins in nutrient sensing pathways can modulate healthy lifespan. Many of these proteins lie in pathways downstream of receptor tyrosine kinases (RTKs) that coordinate not only growth and development but also adult metabolism and function.
For example, the insulin-like growth factor (IGF) receptor is an RTK whose signalling pathways have been extensively studied for their capacity to modulate healthy lifespan among diverse eukaryotic species (Alic & Partridge, 2011;Fontana, Partridge, & Longo, 2010).
Notably, multiple interventions that inhibit insulin/IGF signalling (IIS) can extend healthy lifespan. At the level of the receptors, extension of healthy lifespan has been observed for reduced production of insulin-like ligands in the vinegar fly Drosophila (Broughton et al., 2005;Grönke, Clarke, Broughton, Andrews, & Partridge, 2010), reduced signalling through the insulin/IGF receptor orthologue or its substrates in Drosophila (Clancy et al., 2001;Slack et al., 2010;Tatar et al., 2001), heterozygous deletion of the IGF-1 receptor in mice (Holzenberger et al., 2003), and homozygous deletion of the insulin receptor substrate Irs1 in mice (Selman et al., 2008). Downstream of RTKs, lifespan extension has been reported in Drosophila with inhibited function of the effector kinases PI3K or Ras (Slack et al., 2015;Slack, Giannakou, Foley, Goss, & Partridge, 2011), or over-expression of the transcription factor Foxo, whose activity is inhibited by IIS (Giannakou et al., 2004;Hwangbo et al., 2004). Excitingly, these pathways appear important for human longevity as well: candidate gene studies in centenarians have found enrichment for single-nucleotide polymorphisms in genes encoding the IGF-1 receptor (Suh et al., 2008) and Foxo3a (Flachsbart et al., 2009;Willcox et al., 2008).
These studies suggest that RTK-mediated signalling pathways are a promising direction for understanding aging across species and for uncovering therapeutic targets that can modulate the aging process itself.
While IIS has been a critical gateway for understanding the modulation of healthy aging, the possibility remains that other RTKs can exert similar effects. In humans, 58 RTKs have been identified with distinct ligands, tissue expression patterns and physiological functions (Lemmon & Schlessinger, 2010). In Drosophila, at least 20 genes encoding RTKs exist in the genome, most with clear mammalian orthologues; however, even in the well-studied field of Drosophila development, the function of many of these RTKs remains unclear (Sopko & Perrimon, 2013), and few have been studied for their roles in aging. In the field of cancer biology, however, a recurring role for mutations in many RTKs has made them a focus for a great deal of translational research. Among these, mutations in anaplastic lymphoma kinase (Alk) have been associated with lymphoma, neuroblastoma and non-small-cell lung cancers (Hallberg & Palmer, 2013), leading to the development of effective small molecule Alk inhibitors for clinical use (Kwak et al., 2010;Peters et al., 2017). This critical role of Alk in tumorigenesis has spurred a growing number of studies aiming to understand not only its pathological potentials but also its physiological functions.
Under basal conditions, Alk is expressed most highly in the nervous system, both in vertebrates, including zebrafish (Yao et al., 2013), and in invertebrates, including Drosophila (Cheng et al., 2011).
In vertebrates, recent studies have identified two activating ligands, ALKAL1 and ALKAL2 (Fadeev et al., 2018;Guan et al., 2015), whereas in Drosophila the single identified ligand is the secreted LDL repeat protein jelly belly (jeb) (Englund et al., 2003). Alk signalling is essential for a number of developmental processes: proper neuronal differentiation and survival in zebrafish (Yao et al., 2013), sparing of nervous system growth during nutrient deprivation in larval Drosophila (Cheng et al., 2011), regulation of body growth during nutrient deprivation in larval Drosophila (Okamoto & Nishimura, 2015), and neuronal circuit assembly in the developing Drosophila retina (Bazigou et al., 2007) and neuromuscular junction (Rohrbough & Broadie, 2010).
Alk signalling also plays important roles in adult nervous system function. Adult-onset Alk inhibition in neurons enhances associative memory in both wild-type and neurofibromatosis type 1 (NF1) disease model Drosophila (Gouzi, Bouraimi, Roussou, Moressis, & Skoulakis, 2018;Gouzi et al., 2011), and Alk knockout in mice increases adult hippocampal neurogenesis and enhances performance in novel object recognition tasks (Bilsland et al., 2008). These findings have led to the hypothesis that, in addition to its more canonical roles as an RTK in growth and nutrient sensing, Alk plays a specific role in constraining long-term memory formation (Gouzi et al., 2018).
These findings raise the possibility that other functions remain to be identified for Alk in the adult brain.
Here, we have asked whether Alk, like several other RTKs, modulates healthy lifespan in Drosophila. We build on several established tools to inhibit Alk signalling, including mutation of the jeb gene, RNAi knock-down of Alk, and expression of a dominant-negative Alk protein in adult neurons. In each case, we find that Alk inhibition can extend healthy lifespan in flies. Moreover, we find that inhibition of Alk signalling improves neuromuscular function of aging flies, extends survival under starvation or xenobiotic stressors, and improves night sleep consolidation. Finally, we report that TAE-684, a small molecule Alk inhibitor, can extend healthy lifespan in Drosophila, suggesting that Alk may be a promising pharmacological target in aging.

| Reduced jeb expression extends lifespan
The jelly belly (jeb) gene encodes the single activating ligand for Alk in Drosophila (Englund et al., 2003). Because deletion of other RTK ligands in Drosophila has previously been shown to extend healthy lifespan (Grönke et al., 2010), we asked whether deletion of jeb would have a similar effect on lifespan. The jeb k05644 allele has been well characterized as a jeb loss-of-function allele due to insertion of the P{lacW} element in the first intron of jeb (Weiss, Suyama, Lee, & Scott, 2001). Flies carrying this allele are healthy and viable as heterozygotes, but homozygotes show early larval lethality due to failed fusion of the visceral mesoderm (Englund et al., 2003;Weiss et al., 2001). We obtained the corresponding Bloomington Stock Center strain (BL10576), and upon back-crossing the strain to standardize genetic background, we observed that the stock contained two insertions of P{lacW} on the second chromosome, one whose mini-white transgene produced an orange eye colour and another whose mini-white transgene produced a red eye colour. We isolated each insertion as an independent strain, and both were homozygous lethal. We determined by PCR from genomic DNA that only the orange-eye-producing insertion was within the expected location in the jeb gene ( Figure 1a). We confirmed by quantitative real-time PCR (qPCR) that the isolated jeb k05644 allele produced a significant ~25% reduction in jeb mRNA levels in heterozygotes ( Figure 1b). Notably, the independent insertion of P{lacW} produced a significant ~35% increase in jeb mRNA levels, potentially due to selection for genetic modifiers that could compensate for partial loss of jeb in the original stock.
We next tested whether heterozygous loss of jeb would extend healthy lifespan. We found that heterozygous jeb k05644 female flies were significantly long-lived compared to their wild-type siblings ( Figure 1c, median lifespan + 8.3% and p = .0019 vs. wild-type). This lifespan extension was reproducible in independent experiments in female flies but not in males ( Figure S1). This sexual dimorphism in lifespan extension is reminiscent of other studies reporting greater lifespan extensions in females than in males for rapamycin treatment in flies (Bjedov et al., 2010), dietary restriction in flies (Regan et al., 2016), S6K1 mutation in mice (Selman et al., 2009), and lowered insulin-like signalling in both flies (Clancy et al., 2001;Tatar et al., 2001) and mice (Selman et al., 2008). Finally, as a control, we found no significant change in lifespan in female flies heterozygous for the extra insertion of P{lacW} compared to their wild-type siblings (Figure 1d, median lifespan −8.6% and p = .25 vs. wild-type). While not significant, these flies were, if anything short-lived, consistent with the increased expression of jeb. These results suggest that reduced levels of jeb, and therefore reduced signalling through Alk, can extend healthy lifespan in flies.

| Neuronal knock-down of Alk extends lifespan
To directly test whether reduced Alk signalling can extend lifespan, we next turned to RNAi knock-down of Alk expression. Because Alk is necessary for proper development of both the visceral mesoderm (Englund et al., 2003) and the nervous system (Bazigou et al., 2007;Cheng et al., 2011), we restricted RNAi knock-down of Alk to adulthood by using the GeneSwitch inducible expression system, which uses a modified GAL4 protein to drive expression of UAS transgenes only in the presence of the activating drug RU-486 (Osterwalder, F I G U R E 1 Heterozygous loss of jeb extends lifespan. (a) The original stock (BL10576) containing the jeb k05644 mutation contained two insertions of P{lacW} that became apparent upon back-crossing. Primers designed as shown amplified either P{lacW} independent of its insertion site (Set 1) or P{lacW} in the annotated insertion site in the jeb gene (Set 2). PCR from genomic DNA using these primers confirmed the presence of the jeb k05644 insertion or the secondary extra insertion in each back-crossed stock. (b) qPCR from whole-fly RNA showed reduced jeb mRNA levels in female w Dah ;jeb k05644 /+; flies compared to w Dah ;+/+; flies, with increased jeb mRNA levels in w Dah ;extra-insertion/+; flies compared to w Dah ;+/+; flies. n = 6 biological replicates of three whole flies per replicate for each genotype; p values are from Dunnett's multiple comparison tests between the indicated groups.  (Chintapalli, Wang, & Dow, 2007;Leader, Krause, Pandit, Davies, & Dow, 2018); within the nervous system, single-cell RNA-Seq data sets indicate that Alk gene is more highly expressed in neurons than in glial cells (Davie et al., 2018). We therefore used a ubiquitous driver, Actin5c-GeneSwitch (Act-GS), and an enhanced neuronal driver, elav-Gene-Switch (elav-GS Tricoire , (Latouche et al., 2007)) to reduce Alk gene expression in all tissues or specifically in neurons. We first confirmed by qPCR that Alk RNAi driven by each GeneSwitch system significantly reduced Alk mRNA levels ( Figure 2b,c, 48% decrease for Act-GS and 58% decrease for elav-GS).
We then assessed the lifespan of flies with ubiquitous or neuron-specific Alk knock-down starting in adulthood. We found that RU-486-treated Act-GS > Alk RNAi female flies were significantly long-lived compared to their uninduced siblings treated with vehicle control food (Figure 2d, median lifespan + 5.8% and p = .00014 vs. uninduced control). This lifespan extension was reproducible in independent experiments and was not due to activation of the Act-GS driver itself, as driver-alone control flies showed no significant difference in lifespan when treated with RU-486 ( Figure S2a,b).
We found similar effects when restricting knock-down of Alk to neurons: RU-486-treated elav-GS > Alk RNAi female flies were significantly long-lived compared to their uninduced siblings (Figure 2e, median lifespan + 6.4% and p = .0020 vs. uninduced control), a finding that was again reproducible in independent experiments and was not due to activation of the elav-GS driver itself ( Figure S2c,d). These results suggest that reduced Alk expression can extend healthy lifespan when restricted to adult neurons.

| Neuronal expression of dominant-negative Alk extends lifespan
As an additional means of reducing Alk signalling, we turned to a UAS-driven dominant-negative Alk transgene (UAS-Alk DN ), in which only the extracellular and transmembrane domains of Alk are expressed without the intracellular signalling domains (Bazigou et al., 2007). Because of the neuron-specific effects of Alk knock-down described above, we restricted expression of UAS-Alk DN to adult neurons using the canonical neuronal GeneSwitch driver elav-GS 301 (Osterwalder et al., 2001). We first confirmed by qPCR that Alk DN driven by elav-GS significantly increased Alk mRNA levels when measured using primers directed against the extracellular region of Alk, without producing any change in Alk mRNA levels when for RNAi knock-down of Alk, these results suggest that reduced Alk signalling in adult neurons is sufficient to extend lifespan in females.

| Dominant-negative Alk expression improves locomotor behaviour, stress resistance and night sleep consolidation
In addition to extending lifespan, reduced IIS and TOR signalling have been associated with preserved neuromuscular function with age (Alic et al., 2014), starvation resistance (Bjedov et al., 2010;Slack et al., 2010), resistance to xenobiotic toxins (Slack et al., , 2011, and improved sleep consolidation (Metaxakis et al., 2014). To determine whether reduced Alk signalling in adult neurons would produce a similar preservation of function with age, we assessed the negative geotaxis (climbing) ability of flies with neuron-specific Alk DN expression starting in adulthood. As a control, we first tested activation of the elav-GS driver itself and found that driver-alone control female flies showed no significant difference in the decline of climbing To assess whether improvements in climbing ability were due to overall increased hyperactivity, and to measure whether Alk DN -expressing flies had additional behavioural improvements, we next quantified activity and sleep patterns over the course of 24 hr (12 hr of light during the day cycle followed by 12 hr of dark in the night cycle). We found that Alk DN -expressing female flies showed no evidence of hyperactivity, but instead showed significantly reduced activity during the night cycle (Figure 5a,b).
We next quantified sleep, defined as continuous bouts of 5 or more minutes of inactivity, and found that Alk DN   Moreover, reduced Erk signalling is one of the essential downstream mediators of enhanced longevity from reduced insulin-like signalling in Drosophila (Slack et al., 2015). We therefore assessed Erk phosphorylation by Western blot from adult fly heads and found that female flies expressing Alk DN starting in adulthood showed reduced levels of Erk phosphorylation with no significant change in total Erk levels (Figure 6a,b). This is consistent with previous studies observing reduced Erk phosphorylation in flies with Alk DN or Alk RNAi expression driven by constitutive neuronal drivers (Gouzi et al., 2011).

| Dominant-negative Alk expression reduces the expression of insulin-like peptides
We next went on to assess Erk phosphorylation in both female and male head samples in parallel. Here, we again observed a significant decrease in Erk phosphorylation in Alk DN -expressing females, but not in males ( Figure S6a- Figure S6e-h), suggesting that other upstream signalling pathways may contribute more to Akt phosphorylation than Alk does.
These results prompted us to examine other downstream signalling pathways that may contribute to the lifespan extension observed with Alk inhibition. In larval Drosophila, Alk signalling has been found to inhibit the forkhead-box transcription factor Foxo, thereby relieving inhibition on the expression of the Drosophila insulin-like peptide 5 (dilp5) in insulin-producing neurons (Okamoto & Nishimura, 2015). and dilp5 transcription in the brain (Bai, Kang, & Tatar, 2012).
To confirm the connection between reduced neuronal Alk signalling and reduced dilp2/dilp5 expression, we assessed head RNA from flies with ubiquitous expression of Alk RNAi , where we observed a similar reduction in mRNA levels for dilp2 and dilp5 ( Figure S7a,b).
To determine whether these effects could be mediated cell-autonomously within insulin-producing neurons, we next drove expression of Alk DN specifically in these neurons using a dilp2-GS driver. Here, we found that Alk DN expression within insulin-producing neurons was sufficient both to reduce dilp2 and dilp5 expression and to produce a small but significant lifespan extension ( Figure S7c-e). These results suggest that cell-autonomous Alk modulation of dilp2 and dilp5 transcription within insulin-producing neurons could provide a potential mechanism by which reduced Alk signalling can extend healthy lifespan.

| Effects of neuronal dominant-negative Alk expression on the abdominal fat body
To extend our findings on communication between neuronal Alk signalling and fat-body-derived insulin-like peptides (Figure 6d

| The Alk inhibitor TAE-684 extends lifespan
Mutations in Alk that result in its mis-expression and/or constitutive activation have been described in a number of human cancer types, leading to the development of clinically approved small molecule Alk inhibitors (Hallberg & Palmer, 2013). In aging research, one promising direction to translation of basic research findings is the repurposing of existing drugs for use as interventions against aspects of the aging process itself (Dönertaş, Valenzuela, Partridge, & Thornton, 2018;Fuentealba et al., 2019;Ziehm et al., 2017). We therefore chose to test three small molecule Alk inhibitors for their effect on longevity: two drugs (alectinib and crizotinib) that have been clinically approved for treatment of Alk-positive non-small-cell lung cancer (Kwak et al., 2010;Peters et al., 2017), and one molecule (TAE-684) that is not yet clinically approved but has been previously used in studies of Alk function in mammalian and Drosophila disease models (Galkin et al., 2007;Gouzi et al., 2011Gouzi et al., , 2018. We treated flies starting at day 2 of adulthood with 100 nM, 1 µM or 10 µM of each inhibitor. For alectinib and crizotinib, we found no significant change in lifespan for wild-type w Dah female flies treated with the highest dose of 10 µM (Figure 7a To attempt to explain why TAE-684, but neither alectinib nor crizotinib, could extend lifespan in Drosophila, we hypothesized that amino acid differences between the human and Drosophila Alk sequences could underlie the differential effectiveness of the small molecules, particularly if these differences occur in amino acids near the drug binding site. Crystal structures are available for alectinib, crizotinib and TAE-684 in complex with human Alk (Bossi et al., 2010;Friboulet et al., 2014;Sakamoto et al., 2011). We used these structures to identify the amino acids within 3.5 A of each compound's binding site in the human protein and then used aligned protein sequences to identify the orthologous amino acids in the Drosophila protein. Drosophila Alk shares 34% amino acid identity overall with human Alk, with higher amino acid identity in the cytoplasmic domains (52%-58% identity) (Lorén et al., 2001). Consistent with high levels of amino acid identity, we found only a single amino acid difference among the six nearest amino acids to the binding site for TAE-684 (83% identity). However, we found lower levels of identity for the amino acids in the vicinity of the binding sites for alectinib (25%) and crizotinib (80%) (Figure 7d-f). While these findings do not exclude the possibility that these drugs can inhibit Drosophila Alk, they may explain some of the differential effectiveness we saw for these three compounds in their ability to extend lifespan in Drosophila.
To further investigate the effects of TAE-684, we examined  (Chintapalli et al., 2007). To test whether Alk inhibition in other tissues would also be sufficient to extend lifespan, we used the S 1 106-GS driver to express Alk RNAi in the fat body and gut of adult flies. We found that Alk inhibition in these tissues was also sufficient to extend lifespan ( Figure S10e), suggesting that Alk inhibition by small molecules may be able to extend lifespan by acting in non-neuronal tissues.

| D ISCUSS I ON
We have shown that Alk inhibition, by jeb mutation, RNAi knock-  (Bjedov et al., 2010;Clancy et al., 2001;Regan et al., 2016;Tatar et al., 2001), and in mice from both genetically heterogeneous and homogeneous backgrounds (Harrison et al., 2009;Holzenberger et al., 2003;Selman et al., 2008Selman et al., , 2009). This suggests that mortality in female and male animals may be due to distinct life-limiting pathologies in each sex (Regan & Partridge, 2013). Crucially, the effects of IIS on aging can be sexually dimorphic in humans as well: polymorphisms in IIS pathway genes have been associated with increased old-age survival in women but not men in a Dutch population study of elderly people (van Heemst et al., 2005). Taken together, these studies consistently underscore the importance of considering whether interventions that extend healthy lifespan have sexually dimorphic effects, as appears to be the case for inhibition of Alk signalling.
One notable feature of Alk among other RTKs is its relative specificity of expression in the nervous system. While research on longevity to date has underscored an important role for the primary metabolic tissues (including the gut and adipose tissue) in the modulation of lifespan, an increasing number of studies have defined an additional role for the nervous system in modulating organism-wide metabolism and longevity. In Drosophila, neuronal over-expression of a dominant-negative form of the insulin receptor is sufficient to extend lifespan (Augustin et al., 2018) and to maintain electrical synapse function in aging neuronal circuits (Augustin et al., 2017). Intracellular signalling pathways also play a role in neuronal modulation of lifespan: neuronal over-expression of the nutrient sensor AMP-activated protein kinase (AMPK), or of the autophagy-promoting kinase Atg1, is sufficient to extend lifespan and preserve climbing ability with age in Drosophila (Ulgherait, Rana, Rera, Graniel, & Walker, 2014). The mechanisms underlying these neuron-specific protective effects may be shared with the effects of inhibition of Alk signalling, as both neuronal AMPK and Atg1 over-expression lead to a reduction in dilp2 and dilp5 mRNA expression (Ulgherait et al., 2014), similar to our observed effects for Alk inhibition (Figure 6). Our observed reduction in dilp2 and dilp5 expression is also consistent with previous studies of Alk function in the insulin-producing neurons of the brain, where inhibition of Alk signalling can activate Foxo, which in turn negatively regulates dilp5 transcription (Okamoto & Nishimura, 2015). Notably, however, there may be both shared and divergent mechanisms among neuronal interventions that extend healthy lifespan and stress resistance: we observe a protection against starvation stress for flies that express Alk DN in neurons (Figure 4), in contrast to the sensitivity to starvation seen for flies that over-express AMPK in neurons (Ulgherait et al., 2014). Our results also suggest that non-cell-autonomous effects of neuronal Alk inhibition are likely to play a role in modulating lifespan, as we observe an increase in fat-body-derived dilp6 mRNA when Alk DN is expressed in neurons ( Figure 6). Notably, previous studies have found that over-expression of dilp6 in the abdominal fat body can extend healthy lifespan and reduce brain mRNA levels of dilp2 and dilp5 (Bai et al., 2012), suggesting that a feedback loop of nutrient sensing pathways between the brain and fat body could be an important mechanism underlying healthy lifespan extension. Our results from knock-down of Alk in the fat body ( Figure S10) provide additional support for an important role of Alk signalling not only in the brain but also in the fat body.
When examining the effects of small Alk molecule inhibitors, we found no effect of either alectinib or crizotinib on Drosophila lifespan, and a small but significant effect of TAE-684 (Figure 7). This is in agreement with recent studies showing no effect of crizotinib on wild-type Drosophila lifespan (Kim et al., 2019) but consistent effects of TAE-684 on developmental growth patterns and adult associative memory performance (Gouzi et al., 2011(Gouzi et al., , 2018. As suggested by our structural analysis of the binding sites for each small molecule inhibitor, one potential reason for these results may be the number of conserved amino acid residues between human and Drosophila Alk drug binding sites for each small molecule. Other factors could also explain some of our results: for example, CNS penetrance differs among Alk inhibitors (Peters et al., 2017). However, our results also suggest that Alk inhibitors can extend lifespan without producing an observable decrease in phosphorylated Erk in head extracts, so other tissues including the fat body may play an important role ( Figure S10). Moreover, drug efficacy is influenced by other factors including concentration, delivery method and food composition, so our results do not preclude the possibility that alectinib and crizotinib can effectively inhibit Drosophila Alk in other conditions. Taken together, our findings support careful consideration of the many factors influencing drug efficacy, including the likelihood of substrate binding in different animal species (Ziehm et al., 2017), for future pharmacological studies in model organisms.
While we observed pro-longevity and stress-resistance phenotypes of Alk inhibition in this study, strategies aiming to inhibit Alk are not without potential concerns. Previous studies have shown that Alk inhibition in adult animals can enhance associative memory in flies (Gouzi et al., 2011) and novel object recognition memory in mice (Bilsland et al., 2008), which can be seen as beneficial but may also suggest that Alk plays a role in suppressing inappropriate memory formation under basal conditions (Gouzi et al., 2018). Clinically, Alk inhibitors are well tolerated (Hallberg & Palmer, 2013), but unexplained side effects including visual disturbances (Kwak et al., 2010;Peters et al., 2017) suggest the need for caution before pursuing strategies requiring chronic treatment with Alk inhibitors.
Taken together, our results show that inhibition of Alk by either genetic or pharmacological means is sufficient to extend healthy lifespan and promote stress resistance in Drosophila. The highly conserved nature of Alk signalling among invertebrates and vertebrates lends support to further studies of Alk in mammalian aging and in models of age-related neurodegenerative diseases. Finally, our results are encouraging for future drug repurposing studies in aging research, as existing clinically approved pharmacological inhibitors of Alk and other RTKs could have the capacity to extend healthy lifespan while potentially delaying functional decline and the onset of age-related diseases.
The wild-type stock Dahomey was collected in 1970 in Dahomey (now Benin) and has since been maintained in large population cages with overlapping generations on a 12h:12h light:dark cycle at 25ºC.

| Survival analysis
Lifespan assays were carried out as described in detail in (Piper & Partridge, 2016). From the eggs collected for each set of parental crosses, the progeny that emerged as adults within a 24 hr window were collected and allowed to mate for 48 hr, after which they were separated into single-sex vials containing either standard food or drug-or vehicle-containing food (GeneSwitch and Alk inhibitor ex-

| Climbing (negative geotaxis) analysis
For each genotype analysed in climbing assays, five vials of control food and five vials of food containing RU-486 (200 µM), each containing 15 flies, were housed side-by-side in a single DrosoFlipper.
Flies were maintained as in lifespan studies and then were filmed for climbing assays once to twice per week. For climbing assays, the flies were kept in DrosoFlippers and transferred to empty vials on each side of the flipper, creating a standard vertical column 20cm in height for each set of flies. Flies were tapped to the bottom of the vials and allowed to climb upwards for 15 s before a still camera image was captured. The heights of individual flies were then assessed by manual multi-point selection in Fiji software (Schindelin et al., 2012), with each height in pixels calibrated to a height in cm from a ruler placed next to the vials during filming.

| Starvation and DDT stress assays
For all stress assays, flies were maintained exactly as in lifespan studies until 14 days of age. For starvation assays, flies were then transferred to vials containing 1.5% agar containing either RU-486 (200 µM) or 2 ml/L ethanol as a control condition. Flies were then transferred to fresh vials three times per week, with deaths and censors scored once per day until day 7, then two to three times per day starting at day 7. For DDT xenobiotic stress assays, flies were transferred to vials containing 1.5% agar for 5 hr before being transferred to vials containing 0.03% DDT (Sigma) mixed into standard fly food.
Deaths were scored two to three times per day without any further transfers to fresh vials.

| Activity and Sleep analysis
Activity and sleep were assessed using the Drosophila Activity Monitor (DAM2) system and DAMSystem3 data acquisition software (Trikinetics). At 13 days of age, individual mated female flies were placed into tubes with food containing either  or 2 ml/L ethanol as a control condition. These tubes were loaded into DAMs and placed into an incubator (Percival) set at 25°C, 65% humidity and a 12h:12h light:dark cycle. After two days of acclimatization, data were obtained from a 24-hr period on the third day

| PCR from genomic DNA
To identify the presence of P{lacW} in the jeb gene, PCR was carried out following manufacturer's instructions (All-In Taq

| Quantitative real-time PCR (qPCR)
Total RNA was isolated from whole flies, heads or abdomen carcasses (fat bodies) using standard TRIzol (Invitrogen) protocols. RNA samples were treated with Turbo DNAse (Invitrogen) and converted to cDNA using oligod(T) primers and Superscript II reverse transcriptase (Invitrogen). Quantitative RT-PCR was performed using Power SYBR Green PCR Master Mix (ABI) in the Quant Studio 6 Flex system. Relative quantities of transcripts were determined using the relative standard curve method normalized to Tub84B or RpL9.
Alk primers were designed to amplify either the extracellular region

| Western blots
For protein extracts, fly heads or bodies were homogenized in RIPA buffer (New England Biolabs), with the addition of Complete Mini protease inhibitors (Roche) and PhosSTOP phosphatase inhibitors (Roche). The BCA Protein Assay Kit (Pierce) was used to calculate relative protein concentrations among the samples.
Equal quantities of protein for each sample were then separated on 4%-12% NuPAGE Bis-Tris gels (Invitrogen) and transferred to a PVDF membrane.
Membranes were then probed with secondary anti-mouse or anti-rabbit antibodies (Abcam ab6789 or ab6721, 1:10,000) for 1 hr at room temperature. Blots were developed using Luminata Crescendo or Forte (Millipore) and the ImageQuant LAS 4000 system. Densitometric analysis of blot images was carried out using Fiji software (Schindelin et al., 2012).

| Triglyceride (TAG) Assays
Triglyceride Infinity Reagent (Thermo Scientific) was used to measure total triglyceride content from homogenates of whole single flies. Values were normalized to the protein content per fly as measured by the Pierce BCA Protein Assay Kit (Thermo Scientific).

| Conservation of binding sites for Alk inhibitors
Crystallographic structures of Alk in complex with alectinib, crizotinib and TAE-684 were downloaded from the Protein Data Bank (http://www.rcsb.org). Amino acid residues within 3.5 A of the Alk inhibitors were visualized using the software VMD 1.9.4 (Humphrey, Dalke, & Schulten, 1996). The sequence of the human Alk (UniProt ID Q9UM73) and the Drosophila Alk (UniProt ID Q7KJ08) were aligned using the UniProt Align function, and the binding residues were manually located in the sequence alignment. The sequence identity of the binding site was calculated as the percentage of residues in the human Alk binding site conserved in the Drosophila orthologue.

| Statistical analysis
Statistical analysis was carried out in Microsoft Excel, GraphPad Prism 6.0a, or R using the "survival" package (Terry Therneau, https://CRAN.R-proje ct.org/packa ge=survival). The statistical test used for each experiment is indicated in the figure legend. Log-rank tests on lifespan data were performed in Microsoft Excel (template available at http://piper lab.org/resou rces/) or in R. Linear regression analysis was performed on climbing data, and ANOVA, Mann-Whitney and t test analyses were performed on activity, sleep, qPCR and Western blot data in GraphPad Prism. For all statistical tests, p < .05 was considered significant.

ACK N OWLED G M ENTS
We thank M. Ahmad and G. Vinti for maintenance of the Drosophila laboratory facilities in the Institute of Healthy Ageing at UCL. We also thank the wider fly community for generous sharing of reagents and stocks, particularly the Bloomington Drosophila Stock Centre (NIH

DATA AVA I L A B I L I T Y S TAT E M E N T
The raw data supporting this study are available in Mendeley Data at http://dx.doi.org/10.17632 /k392r cs98d.1