Age-dependent H3K9 trimethylation by dSetdb1 impairs mitochondrial UPR leading to degeneration of olfactory neurons and loss of olfactory function in Drosophila

Aging is characterized by a decline in essential sensory functions, including olfaction, which is crucial for environmental interaction and survival. This decline is often paralleled by the cellular accumulation of dysfunctional mitochondria, particularly detrimental in post-mitotic cells such as neurons. Mitochondrial stress triggers the mitochondrial unfolded protein response (UPRMT), a pathway that activates mitochondrial chaperones and antioxidant enzymes. Critical to the efficacy of the UPRMT is the cellular chromatin state, influenced by the methylation of lysine 9 on histone 3 (H3K9). While it has been observed that the UPRMT response can diminish with an increase in H3K9 methylation, its direct impact on age-related neurodegenerative processes, especially in the context of olfactory function, has not been clearly established. Using Drosophila, we demonstrate that an age-dependent increase in H3K9 trimethylation by the methyltransferase dSetdb1 reduces the activation capacity of the UPRMT in olfactory projection neurons leading to neurodegeneration and loss of olfactory function. Age-related neuronal degeneration was associated with morphological alterations in mitochondria and an increase in reactive oxygen species levels. Importantly, forced demethylation of H3K9 through knockdown of dSetdb1 in olfactory projection neurons restored the UPRMT activation capacity in aged flies, and suppressed age-related mitochondrial morphological abnormalities. This in turn prevented age-associated neuronal degeneration and rescued age-dependent loss of olfactory function. Our findings highlight the effect of age-related epigenetic changes on the response capacity of the UPRMT, impacting neuronal integrity and function. Moreover, they suggest a potential therapeutic role for UPRMT regulators in age-related neurodegeneration and loss of olfactory function.


Introduction
Aging is associated to a time-dependent organ dysfunction that increases the vulnerability of an organism to various forms of stress, ultimately leading to organism death [1][2][3][4][5][6][7] .Aging induces alterations across various physiological systems, with the olfactory system being particularly affected 1 .Olfaction, the sense of smell, is essential for detecting environmental odors crucial for feeding, reproductive and survival behaviors 2,3 .Importantly, dysfunction in olfaction is emerging as one of the early signs of neurodegenerative diseases, including Alzheimer's and Parkinson's disease 8,9,11 .Research on Drosophila melanogaster has shown that the age-related decline in odor response is influenced by the functional state of mitochondria in olfactory projection neurons (OPNs) 4 .This decline is accompanied by defects in neuronal integrity and a decrease in synaptic proteins 5, which correlates with a reduction in mitochondria and an increase in ROS 5 , but the underlying mechanisms has not been defined.
Under compromised mitochondrial integrity or function, cells engage in a transcriptional response known as the mitochondrial unfolded protein response (UPR MT ) 14 .This mitochondrial program can be activated by the impairment of the electron transport chain (ETC), alteration of mitochondrial dynamics, accumulation of unfolded proteins, reduction of mitochondrial DNA, or reduction of mitochondrial chaperones or protease 6-9, .Upon UPR MT pathway engagement, the transcription factor ATFS-1 (C.elegans ortholog of mammalian ATF5 and Drosophila crc) translocates from the mitochondria to the nucleus 10,11 .In the nucleus, ATFS-1, DVE-1, and UBL-5 interact to reorganize the chromatin structure, enabling activation of the nuclear transcription of mitochondrial chaperones, including hsp-60 and hsp-6, and the protease clpp-1.This coordinated transcriptional response restores mitochondrial function under stress conditions by metabolic adaptations, and enhancing mitochondrial biogenesis [12][13][14] .
Chromatin remodeling is crucial for UPR MT regulation, with the epigenetic state of lysine 9 on histone 3 (H3K9) serving as a critical regulator of the response [14][15][16][17] .Changes in H3K9 methylation by the methyltransferase MET-2, the C. elegans ortholog of human SETDB1, modify UPR MT -related loci exposure, modulating binding of UPR MT regulators DVE-1 and ATFS-1 21 .In addition, enzymes that remove methyl groups from H3K9 significantly influence UPR MT activation.For example, demethylases JMJD-3.1 and JMJD-1.2remove trimethylation from H3K9me3 and H3K27me3, enabling UPR MT activation 15,17 .Recent studies revealed the effects of H3K9me3 methylation on UPR MT activation and mitochondrial function across species.In C. elegans the epigenetic factors BAZ-2 and SET-6, which regulate H3K9me3 levels, have conserved roles in impacting aging processes through mitochondrial function 18 .In mice, deletion of Baz2b, a homologue of BAZ-2, shows beneficial effects on mitochondrial function and cognitive abilities, indicating a conserved mechanism across species that influences aging and healthspan through mitochondrial function and epigenetic regulation 18 .Accordingly, in the hippocampus of aged mice, H3K9me3 levels rise with age, a change associated with age-related cognitive decline 19 .Age-related increases in H3K9me3 have also been observed in the brains of aged Drosophila and muscle stem cells of aged mice 20,21 .While previous research has linked changes in methylation levels to age-related functional decline, the specific role of these epigenetic alterations in the context of neuronal degeneration through reduced UPR MT activation capacity remains to be fully elucidated.This study aims to bridge this gap by providing detailed insights into how epigenetic mechanisms, particularly methylation changes, directly contribute to the aging-associated loss of olfactory function and neuronal degeneration by impacting the UPR MT pathway and mitochondrial function.
Here, we employed behavioral, molecular, and morphological methodologies to investigate whether epigenetic regulation of UPR MT is linked to neurodegeneration in the aging brain and its involvement in age-associated olfactory decline.To this end, we utilized the OPNs in the adult Drosophila antennal lobe (AL), which exhibit age-related neurodegeneration correlating with functional neuronal decline 5 .Our results demonstrate that with aging, there is a decline in the response capacity of the UPR MT in OPNs, functionally associated with a dSetdb1-dependent increase in H3K9me3 levels.Genetic inhibition of dSetdb1 reduces H3K9me3 levels, enabling the activation of UPR MT , restoring mitochondrial oxidation to youthful levels, and preventing age-associated degeneration of OPNs.This effect is particularly evident in the somas located in the antennal lobe (AL) and the presynaptic connections of the lateral horn (LH) in the Drosophila brain.Importantly, maintaining UPR MT activation during aging preserved olfactory function.These findings underscore the critical role of epigenetic regulation, specifically through dSetdb1 and H3K9me3, in modulating neuronal integrity and sensory function during aging.

1.
UPR MT response capacity decreases with aging in the Drosophila antennal lobe.
To study the modulation of UPR MT along aging and its association to olfactory function, we first generated reporters based on the expression of the fluorescent protein dsRed under the promoters of chaperones hsp60, which specifically responds to UPR MT stimuli in across species and has been effectively used to indicate UPR MT activation [22][23][24][25][26] .We focused on the antennal lobe (AL), the functional homolog of the vertebrate olfactory bulb, where olfactory projection neurons process olfactory sensory inputs.In young flies (0 dpe), a low-intensity signal from the Hsp60::dsRed reporter was detected under control conditions, which significantly increased upon exposure to the UPR MT activator paraquat (PQ, Fig. 1A and B).Importantly, this increase in Hsp60::dsRed signal was not induced by non-specific mitochondrial stressor tunicamycin, which activates UPR ER and the ER-stress specific reporter Xbp1::GFP 27 (Fig. 1A-D).To further evaluate the specificity of the UPR MT reporter, we downregulated the UPR MT nuclear activators dve, ubl, and crc.Pan-neuronal downregulation of these UPR MT activators significatively reduced the Hsp60::dsRed response to PQ compared to control flies (Fig. 1E and F), indicating that this novel reporter can be used to specifically monitor UPR MT activity.We next used the Hsp60::dsRed reporter to evaluate UPR MT activity during aging in the Drosophila AL.Compared to the robust signal triggered by PQ in young flies, aged flies (45 dpe) did not exhibit Hsp60::dsRed reporter activation in AL neurons after PQ (Fig. 1G-H).Importantly, no significant changes in GFP-labelled neuronal volume were observed in aged versus young flies or after PQ treatment (Fig. 1I).This data suggests that the ability to trigger UPR MT activity declines with advanced age.
We then explore the age-dependent endogenous expression of UPR MT -associated chaperones Hsp60 and Hsc70-5 using Scope, a single-cell gene expression repository of brain cells from Drosophila at different ages (http://scope.aertslab.org) 28.The Drosophila brain consists of three major groups of neurons: glutamatergic, GABAergic and cholinergic neurons, with the latter being the most abundant in the AL (Fig. 1J).Single-cell data for cholinergic neurons show that both Hsp60A and Hsc70-5 expression levels decrease in aged flies compared to young animals (Fig. 1K-L).Interestingly, this was also observed in glutamatergic neurons but not in GABAergic neurons, suggesting differences in UPR MT activation in different neuronal populations.These findings suggest that in the Drosophila AL, aging is associated with a decline in UPR MT activity and chaperone expression, particularly in cholinergic neurons.

2.
Epigenetic regulation of the UPR MT by dSetdb1 in the AL of Drosophila brain.

It has been previously demonstrated that trimethylation of H3K9 increases during Drosophila
aging 20 , a phenomenon that mirrors observations in other species.To assess whether methylation levels of H3K9 can modulate UPR MT activation in flies, we studied flies with pan-neuronal knockdown of dSetdb1, a specific H3K9 methyltransferase.Our data demonstrates that panneuronally downregulating dSetdb1 prevents the age-associated increase in H3K9 trimethylation in homogenates of Drosophila heads (Fig. 2A).We then investigated the role of H3K9 trimethylation in UPR MT activation by examining the Hsp60::dsRed reporter in flies with a ubiquitous loss of function of dSetdb1.Control flies exhibited similar basal levels of Hsp60::dsRed signal in the AL of young and aged flies.Similarly, we observed no differences in the reporter signal for young flies in which dSetdb1 was downregulated (Fig. 2B), consistent with the low levels of H3K9 trimethylation observed in young animals (Fig. 2A).However, the Hsp60::dsRed signal in aged dSetdb1 mutants was significantly higher compared to age-matched control flies (Fig. 2B-C).These results were further confirmed using a second UPR MT reporter based on Hsc70-5 expression (Fig. 2D-E).These data suggest that dSetdb1 contributes to age-dependent H3K9 trimethylation, and its reduced function in the AL of aged flies correlates with a basal increase in UPR MT activity.
To understand the relevance of H3K9me3-related genes in a neuron-specific context, we then analyzed single-cell data from Scope to assess the expression levels of dSetdb1, as well as the H3K9 demethylases Utx and Kdm2 15,29 .In vAChT neurons, dSetdb1 expression remains constant throughout aging (Fig. 2G).However, both Utx and Kdm2 exhibit an age-dependent decrease in expression (Fig. 2H and I).Together, this data suggests that the age-dependent reduction in H3K9 demethylation enzymes could be associated with higher levels of H3K9me3 in the aged Drosophila brain, which in turn might contribute to the age-related decrease in UPR MT activity.

Age-dependent decline in olfactory function depends on the epigenetic modulation of the UPR MT .
Having established that the decline in UPR MT activity in aged Drosophila is linked to elevated levels of H3K9me3, we then explored the potential link between UPR MT activation and olfactory function in Drosophila.The ability to discern between odors diminishes with age in flies, a phenomenon quantifiable through the olfactory T-maze (Fig. 3A-B).Therefore, we genetically downregulated the UPR MT transcriptional activators dve, ubl or crc and studied the ability of flies to discriminate odors throughout their lifespan.Remarkably, young flies with the knockdown of the nuclear activators of the UPR MT exhibited a reduced olfactory capacity to discriminate an abrasive odor compared to control animals (Fig. 3C).Aged flies with the knockdown of dve, ubl or crc did not show significantly different olfactory function compared to age-matched controls or to young flies from the same genotype (Fig. 3C).These findings underscore the crucial role of UPR MT transcriptional activators in preserving olfactory discrimination.We next explored the impact of the epigenetic regulation of UPR MT in neuronal functionality.To this end, we generated flies with pan-neuronal knockdowns of the H3K9 methyltransferase dSetdb1 and demethylases Kdm2 or Utx.Consistent with our previous observations, young flies with downregulated dSetdb1 did not show a difference in H3K9me3 levels compared to controls.In contrast, downregulation of demethylases Utx or Kdm2 led to increased H3K9me3 levels, highlighting their distinct regulatory roles (Fig. 3D).To determine if H3K9me3 levels alter neuronal functionality in the olfactory system, we assessed olfactory function in flies with downregulation of dSetdb1, Utx, or Kdm2.In young flies, pan-neuronal knockdown of dSetdb1 showed no significant difference from control flies.However, aged dSetbd1 mutant flies exhibited improved olfactory function, with no significant difference when compared with young flies (Fig. 3E).
On the other hand, pan-neuronal knockdown of Utx or Kdm2 impaired olfactory function in young flies, with an odor discrimination capacity similar to aged control flies (Fig. 3E), indicating that agedependent increases in H3K9me3 progressively affects olfactory function.

Epigenetic modulation of the UPR MT influences olfactory function in an OPN-cell autonomous manner.
As olfactory function is a complex behavior dependent on multiple central and peripheral neuronal populations, we investigated whether age-related changes in H3K9 methylation, UPR MT and olfactory function were specifically associated with cholinergic OPNs.We first assessed agedependent changes in H3K9me3 trimethylation in olfactory projection neurons (OPNs).To this end, we selectively label the plasma membrane of OPNs by expressing the CD8::GFP fusion protein using the OPN-specific driver GH146-Gal4.We then performed immunofluorescence to evaluate H3K9me3 levels specifically in ToPro3-positive nuclei located in GFP-positive neurons.Using this method, we observed an increase in trimethylation in aged OPNs compared to young ones (Fig. 4A-C).We then evaluated the regulation of dSetdb1 in aging-associated OPNs trimethylation by generating flies carrying the knockdown of dSetdb1 specifically in cholinergic OPNs.Remarkably, H3K9 trimethylation levels in aged flies with OPN-specific dSetdb1 knockdown were reversed and with no significant difference from young control flies (Fig. 4A-C).
To further explore the neuronal-specificity of the UPR MT effect, we assessed Hsp60::dsRed reporter activity specifically in aged OPNs.The response of the UPR MT sensor in CD8::GFP tagged neurons increased in young flies treated with PQ compared to animals treated with vehicle.However, this response to the mitochondrial stressor was diminished in aged animals (Fig. 4D-E).Importantly dSetdb1 knockdown in OPNs, increased reporter activity in response to PQ in aged flies (Fig. 4D-E).
Single cell expression analysis specifically in cholinergic OPNs using Scope revealed a decrease in the UPR MT -associated chaperones Hsp60 and Hsc70-5 in aged OPNs (Fig. 3F-H), with constant levels of dSetdb1 and lower levels of Kdm2 (Fig. 4I-J).Importantly, this data suggests that the observed increase in trimethylation levels within OPNs may be associated with a decline in demethylase activity as flies age.Underlaying significant implications for the regulation of UPR MT and the overall epigenetic landscape in aged neurons.
Having demonstrated that dSetdb1 is essential for the increase in H3K9me3 in aged flies, preventing the activation of UPR MT specifically in OPNs, we next evaluated olfactory function.
Knockdown of dSetdb1 only in cholinergic OPNs improved olfactory function in aged flies compared to controls.We then assessed if this improvement in olfactory function was dependent of UPR MT .To this end, dve or crc were knockdown specifically in cholinergic OPNs in dSetdb1-deficient flies.
Importantly, knockdown of dve or crc reversed the maintenance of olfactory function induced by dSetdb1 knockdown, mirroring the olfactory capacity of control aged flies (Fig. 4L).Additionally, downregulation of dve and crc only in OPNs impaired olfactory function in young flies, to levels comparable to that of aged control flies.

Downregulation of dSetdb1 in OPNs restores age-associated mitochondrial morphological abnormalities and reduces mROS levels.
As changes in UPR MT activation can influence mitochondrial morphology and function, potentially triggering degenerative mechanisms, we investigated mitochondrial morphology in OPNs by expressing a mitochondrially targeted GFP.We examined mitochondrial morphology in three distinct compartments of OPNs, including cell bodies in the AL, the axonal tract, and the presynaptic terminal-enriched lateral horn (LH, Fig. 5A).In the AL of aged flies, a marked decrease in total mitochondrial volume was observed, along with increased mitochondrial fragmentation and sphericity.Notably, the targeted downregulation of dSetdb1 within OPNs mitigated these agerelated changes, resembling the values observed in young control flies.Surprisingly, dSetdb1 knockdown also resulted in reduced mitochondrial fragmentation in young flies, suggesting a potential disruption of the mitochondrial network when compared with age-matched controls (Fig. 5B-F).Within the axonal tract, aged control axons exhibited a significant reduction in mitochondrial volume when compared to younger flies (Fig. 5G-H).Targeted knockdown of dSetdb1 effectively maintained mitochondrial volume.Lastly, no significant changes in mitochondrial parameters were observed in the LH of aged flies for both genotypes (Fig. 5I-J, and Supplementary Fig. 1).
We next assessed mitochondrial oxidation levels as a surrogate marker of mitochondrial function 30,31 .To this end, we employed the UAS-MitoTimer construct, a mitochondrial oxidation reporter 31,32 .This tool relies on the expression of a green fluorescent protein that transitions to red fluorescence upon oxidation 31 .Compared to young control flies, we observed an increase in mitochondrial oxidation in older control flies within the AL (Fig. 5K-L), axonal tract (Fig. 5M-N), and the LH (Fig. 5O-P).Remarkably, downregulation of dSetdb1 in OPNs reversed the age-related mitochondrial oxidation in the three analyzed neuronal regions.Our results indicate that the downregulation of dSetdb1 in OPNs, activating UPR MT during aging, not only reverses ageassociated changes in mitochondrial morphology but also effectively prevents the accumulation of mitochondrial oxidation.

Epigenetic regulation of UPR MT by dSetdb1 modulates age-dependent neurodegeneration of OPNs.
Loss of neuronal function is often linked to degenerative structural changes in the neuronal circuit 33,34 , a phenotype extensively associated with mitochondrial dysfunction.Therefore, we investigated whether UPR MT activation regulates neuronal integrity throughout aging in OPNs.To assess the impact of UPR MT modulation in neuronal integrity, we first assessed changes in neuronal number throughout aging by counting nuclei from GFP-positive OPNs.In control flies, aging resulted in a significant decrease in the number of OPNs.Remarkably, this age-related neuronal loss was prevented in flies where dSetdb1 was downregulated only in OPNs (Fig. 6A-B).We next evaluated axonal integrity GFP-labelled OPNs.In aged control flies, a reduction in axonal integrated density was observed when compared to young control flies, consistent with the previously noted decrease in the total number of neurons.Notably, dSetdb1 knockdown protected against the decline in axonal integrity of OPNs associated with aging (Fig. 6C-D).

It has been previously demonstrated that the age-associated decline in olfactory function in
Drosophila is associated to the loss of synapses in OPNs 5 .Thus, we focused on the LH, a region enriched in presynaptic connections of OPN neurons 5,35 .Control flies exhibited a significant decrease in LH integrated density throughout aging, which was prevented by downregulation of Setdb1 in OPNs (Fig. 6E-F).This data suggests that H3K9-dependent UPR MT activation plays a crucial role in maintaining neuronal integrity within this presynaptic enriched region.To gain a more detailed insight into synaptic zones, we employed the Brp::GFP reporter, a fusion protein that specifically accumulates in presynaptic buttons, facilitating visualization and quantification of presynaptic puncta 36 .Our analysis revealed that both total volume and number of GFP puncta in the LH of aged control flies were reduced compared to their younger counterparts (Fig. 6G-I).When dSetdb1 was downregulated in OPNs, we observed a decrease in this age-dependent integrated density decline.Notably, there was no significant difference between the number of Brp::GFPpositive puncta in young and aged dSetdb1 knockdown flies (Fig. 6I).
These findings collectively demonstrate that the targeted downregulation of dSetdb1 plays a causal role in preserving neuronal numbers and axonal integrity in OPNs.This intervention not only maintains presynaptic densities but also actively contributes to the demethylation of H3K9 and the activation of UPR MT , which are integral to the preservation of olfactory function during aging.By modulating these key processes, our results establish a direct link between the epigenetic regulation by dSetdb1 and the mitigation of age-related neurodegeneration in OPNs, underscoring the potential of targeted epigenetic interventions in maintaining neural health in olfactory system.

Discussion
The UPR MT plays a critical role in preserving mitochondrial homeostasis 37 .Therefore, any change in its activation potential, whether caused by physiological or pathological factors, could impact mitochondrial function [38][39][40][41] .Although it is widely accepted that olfactory function declines with age, diverse factors have been associated with this age-related impairment [42][43][44][45][46] .Our findings indicate that the age-dependent decline of olfactory function in Drosophila is associated to a decrease in the activation capacity of the UPR MT in olfactory neurons.Crucially, the reduction in UPR MT activation is linked to an increase in H3K9me3, which is dependent on the methylation activity of dSetdb1.
Importantly, targeting this methylation process can effectively prevent age-related neuronal degeneration and restore the loss of olfactory function associated with aging.
Our study uncovers a novel aspect of epigenetic regulation in the aging process, emphasizing the specific role for dSetdb1 in modulating H3K9me3 levels and suppressing of the UPR MT .While previous research has established the impact of H3K9 methylation on the UPR MT , primarily through the actions of the demethylases JMJD 3.1 and JMJD 1.3, as well as methyltransferases such as MET-2 (dSetdb1 orthologue), SET6, BAZ2 in worms and mice respectively 14,15,41 , the specific function of Setdb1 in this context remained unclear.Our analysis indicates that in the AL of Drosophila brain, the absence of dSetdb1, also known as eggless, leads to a reduction in H3K9me3 in aged flies, suggesting its role as a tri-methyltransferase that acts as an epigenetic modulator of UPR MT during aging.This finding aligns with evolutionary conservation in epigenetic regulation across species.Supporting this, research shows that in C. elegans, knocking out met-2 also reduces H3K9me3, suggesting a conserved mechanism 47 .However, previous research on the role of met-2 in UPR MT regulation only focused on H3K9Me2 and did not extended their examination to H3K9me3 14 Additionally, the role of SET6 in tri-methylating H3K9 and reducing UPR MT in mice further highlights a possible conserved epigenetic pathway influencing aging across distinct species 41 .
Existing research underscores the beneficial role of UPR MT in maintaining cellular integrity, primarily through maintaining mitochondrial function, a critical factor for healthy aging [48][49][50] .Indeed, the overexpression of the histone demethylases JMJD-1.2 and JMJD-3.1 extend lifespan in C. elegans 40 .Conversely, reduced expression of UPR MT nuclear effectors ATFS-1, UBL-5, and DVE-1, as well as demethylases JMJD-1.2 and JMJD-3.1,compromises lifespan and cellular viability 15,48,51- 53 .Recent work reveals the fine tunned regulation of UPR MT along aging, particularly through the H3K9 methyltransferase SET-6 and the epigenetic reader BAZ-2, which modulate gene expression related to mitochondrial health and stress responses, essential for neuronal viability 54 .Importantly, beyond the impact on lifespan, UPR MT regulation has profound implications for neuronal function.In C. elegans, mitochondrial function was found to influence pharyngeal pumping (eating) and defecation rates, crucial for lifespan 50 .Additionally, deficits in dopamine-dependent behaviors were observed in pdr-1 and pink-1 mutants, indicative of neuronal dysfunction without neuronal loss.This dysfunction is exacerbated by the downregulation of atfs-1, which is critical for UPR MT 52 .In a mammalian context, Baz2b ablation enhanced mitochondrial function in the hippocampus and cerebellum in older mice, suggesting its role in modulating age-related cognitive decline 41 .This was accompanied by improved performance in locomotion, reflecting preserved motor functions in aged mice.Beyond these, UPR MT regulates hippocampal neural stem cell aging, with implications for cognitive functions 55 and affects skeletal muscle aging, as exercise improves coordination of UPR MT and mitophagy in aging skeletal 56 .Moreover, it plays a critical role in fertility and reproductive aging 57 suggesting that UPR MT disruption can pivot healthy aging towards pathological states.Our data indicate that reducing H3K9me3 levels during aging to enhance UPR MT activation is beneficial for olfactory function and support the prevailing hypothesis that modulation of epigenetic regulators, which suppress UPR MT transcriptional activation, constitutes a viable therapeutic approach for ameliorating mitochondrial dysfunction associated with aging.
Brain aging exhibit distinct regional variations across multiple levels, including gene expression, organelle, and neuronal function [58][59][60] .Our analysis of gene expression from single cell studies reveals neuronal-specific transcriptional expression for H3K9-regulating enzymes such as dSetdb1, utx, and kdm2 in aged vAChT, vGlut, and Gad1 neurons.It has been previously demonstrated that H3K9me3 levels are not uniform across the brain, varying based on brain regions or neuronal types [61][62][63] .Neurodegenerative conditions, characterized by the selective degeneration of specific neurons and their projections, exhibit differential neuronal vulnerability that is intricately linked to variations in neuronal morphology, activity patterns, and gene expression profiles within these affected structures [64][65][66] .Our data suggests that the age-dependent reduction in H3K9 demethylation enzymes within specific neuronal populations may contribute to differential neuronal vulnerability along aging, which deserves further exploration.
The increase we had shown in the levels of H3K9me3 in the brains of aged fruit flies aligns with prior studies reporting a rise in H3K9me3 in aged Drosophila heads 33 .Studies performed in C.
elegans have revealed that as age progresses, there is an increase in the expression of the H3K9me3 methyltransferase SET-6 and the epigenetic reader BAZ-2.Remarkably, inhibiting their expression has been linked to preservation of pharyngeal pumping in these organisms 41 .In mice, administering an inhibitor for the histone methyltransferase SUV39H1, which is responsible for the trimethylation of H3K9, was found to mitigate age-associated cognitive decline and augment dendritic spines in the hippocampus 67 .Such findings support the notion that diminishing H3K9me3 levels might enhance functionality of different brain modules during aging.While our findings suggest that reducing hypermethylation in aging could potentially enhance UPR MT response, it is critical to acknowledge that methylation processes also govern a vast of other cellular and neuronal functions 68 .For instance, histone methylation plays a crucial role in gene expression regulation, cellular differentiation, and even neuronal activity 69,70 , all of which could be inadvertently impacted by broad-spectrum epigenetic interventions.This pleiotropic nature of methylation underscores the importance of a targeted approach.
Mitochondria play a central role as primary sensors for degenerative stimuli 71 .With aging, neurodegeneration is often preceded by mitochondrial dysfunction, which manifests as morphological changes, including swelling, fragmentation, reduced volume, and increased oxidative stress [72][73][74][75][76][77][78][79][80] .Consistent with prior studies in aged flies 5 , we did not observe an increase in fragmentation within axons of the OPNs and lateral horn (LH), which could be attributed to the specific neuronal types under investigation 81 .Interestingly, our observations align with those from other studies 5 , as we identified an age-related deterioration in Drosophila's olfactory circuits, which coincides with a rise in oxidative mitochondria.Current research underscores the central role of UPR MT activation in orchestrating mitochondrial morphology and function [82][83][84][85][86] .Importantly, our study demonstrated that genetic inhibition of dSetdb1 restored youthful levels of H3K9me3, enabling UPR MT activation to restore mitochondrial morphology and oxidative status.This maintenance of cellular viability through UPR MT activation parallels findings from a recent study, which revealed that mild mitochondrial dysfunction-dependent UPR MT activation, protects cardiomyocytes against cardiac ischemia-reperfusion injury in a mouse model 9895 .Also, NAD+ activation of the UPR MT rejuvenated muscle stem cells in aged mice 82 , highlighting the role of UPR MT in altering aging markers in stem cells and extending lifespan.
Our research highlights the crucial role of UPR MT regulation in the age-related decline of olfactory function.Focusing in the olfactory system in aged Drosophila, we have demonstrated detrimental effects of epigenetic changes on mitochondrial function, impacting neuronal survival.The importance of olfaction extends beyond sensory perception, with human olfactory processing is intricately linked to emotions and memories, mediated by the limbic system and cerebral cortex 87,88 .
A compromised sense of smell is not only associated with depression in a significant number of cases 89 but also frequently precedes the onset of age-related neurodegenerative diseases such as Alzheimer's 90,91 and Parkinson's disease 92,93 .Our findings underscore the need for further exploration of the UPR MT pathway and its epigenetic regulation as a potential target for developing interventions to mitigate the decline in neuronal function associated with the aging process.Protein Quantification and western blotting.For immunostaining of H3K9me3, samples were prepared as described previously3.Briefly, samples were frozen in liquid nitrogen and ground to a fine powder using a pestle fitted to a 1.5ml Eppendorf Centrifuge tube filled with RIPA buffer, which included 50nM Tris, 150 mM NaCl, 1 mM EDTA, 0.1% of Nonodet P-40 (NP-40), 0.25% of Sodium Deoxycholate, and 0.02% of sodium azide in ddH20.RIPA buffer was supplemented with phenylmethylsulphonyl fluoride (PMSF -Sigma) and a protease inhibitor cocktail (Sigma, P8340).

Drosophila
Homogenized samples were incubated at 4°C for 1h and sonicated by sonicator (QSonica) at 40% of equipment maximal amplitude with 3 pulses in 1 min.Sonicated Samples were centrifuged at 500g to pellet the debris, and all supernatant was transferred to a new tube and then centrifuged for

Dissection of adult Drosophila brains and confocal microscopy. Flies of the desired genotype
were collected in groups of 15 and placed in vials containing Drosophila medium.Flies were anesthetized using a CO 2 pad.Using Dumont forceps #5, flies were held from the thorax and dipped in the sylgard petri dish filled with cold PBS.Brains were isolated by removing the exoskeleton from the fly´s head and carefully removing the esophagus and air sacs of the flies' brains as previously described 4 .For live imaging of endogenous fluorescence signal (Fig. 1A, C, E, 2A, 3C, E, 4H, 6A, C, E and 8C) brains were fixed in 2% paraformaldehyde with 0.1% Triton X-100 (Sigma, T9284).for 20 minutes and then changed to a 4% paraformaldehyde, 0.1% Triton X-100 solution for 20 more minutes, then washed for 10 min 3 times in PBS Triton X-100 at 0.1% followed by 3 quick washes with PBS only.Brains were mounted in VectaShield antifade mounting medium (Vector, H1000) for later visualization in an SP8 confocal microscope.For all experiments, all brains were imaged on the same day.For immunostaining (Fig. 4J, 5A,B,G,I, 7A,C, and 8A), brains were dissected as described previously5.Brains were fixed for 1h at room temperature in 4% PFA with 0.5% Triton X-100, washed 3 times for 10 minutes with PBS with 0.5% Triton X-100, and blocked for 1h with Normal goat Serum (NGS) (Cell Signaling, 5425S) at 5% in PBS, 0.5% Triton X-100 and stained overnight at 4ºC with primary and after 3 washes in PBS, 0.5% Triton X-100 with secondary antibodies using the same conditions.The secondary antibody was washed 6 times for 10 min each with PBS, 0.5% Triton X-100 a 3 quick washes in PBS before mounting.Washed brains were placed in a stripe of 15 to 20 µl of Vectashield antifade mounting medium (Vector, H1000) on cover glass (Deltalab, D10004), and imaging was performed in confocal microscope SP8 using the 63x objective with digital zoom necessary for desired resolution.Fluorescence intensity for each channel was adjusted using control flies to the point that no saturation was observed, then the same parameters were used for all images.Images of antennal lobe sections of Drosophila brains and OPNs were taken at a depth of 10µm using a Z stack separation of 0.6µm.Primary antibodies used were anti-GFP (Invitrogen, 1:1000), rabbit anti-H3K9me3 (Abcam, ab8898; 1:500), and secondary antibody donkey anti-Rabbit 555 (Thermofisher, 1:1000).Oxidation: Mitochondrial oxidation was assessed using MitoTimer, a fluorescent protein that shifts from green to red upon oxidation.The analysis involved processing both the green and red channels and determining their ratio.In this context, oxidized mitochondria appeared red, while healthy mitochondria were green, following the methodology outlined in a previous study 31,32 .
Nuclei number and H3K9me3 quantification: For quantification of the specific signal in Olfactory Projection Neurons, a surface of GFP-labeled OPNs was rendered, then To-Pro3 labeled nuclei signal was masked and the surface rendered, and finally, the signal for H3K9me3 in the nuclei of OPNs was quantified as described previously 5 .Neuronal Integrity: Neuronal degeneration of OPNs was analyzed by rendering the surface of GFP-labeled OPNs in AL, the distal part of the axon, and axonal terminals in the LH, Integrated density was calculated to determine the amount of signal intensity.Presynaptic puncta Quantification: For quantifying presynaptic puncta, we employed the Brp::GFP reporter, a fusion of Brunchpilot and GFP.This reporter accumulates in presynaptic buttons, allowing visualization.After rendering the GFP surface, we applied a mask to the GFP channel and counted the puncta, using a threshold ratio of 300µm in the masked channel to ensure accuracy as described previously 36 .Data was plotted and analyzed using GraphPad Prisms 9 Software.To compare the interaction between age and genotype/treatment, two-way ANOVA and for more than two groups, analysis of variance with Dunnet multiple comparisons was performed using GraphPad Prism RNA extraction, RT-PCR and qPCR.Total RNA was extracted either from 2 whole flies or a minimum of 20 heads using a typical trizol extraction following the next steps: tissue was placed on a 1.5 ml Eppendorf tube and 300 µl of trizol reagent were added before homogenization with a plastic pestle 60 times.At this point sample can be stored at -80ºC.700 µl of trizol were added and samples were centrifuged at 13000 rpm for 10 min at 4ºC to pellet the tissue debris.Supernatant was transferred to a fresh 1.5 tube.300ul of chloroform was added, and samples were shaked vigorously by hand 15 times.Then incubation of the samples at room temperature for 5 min.
Samples were centrifuged at 13,000 rpm for 15 min at 4ºC.The upper phase was carefully transferred to a new RNAse-free tube, without touching the interphase.700uL of isopropanol were added to precipitate the RNA and samples were incubated for 5 minutes at room temperature or 1 hour at -20ºC.Centrifuge at 12,000 rpm for 15 min at 4ºC.Discard the supernatant and RNA pellet was washed with 1 ml of ethanol 70% prepared with miliQ quality H2O.Centrifuge at 13000 rpm for 10 min at 4ºC.Air dry the pellet briefly, resuspend the pellet in an appropriate volume of MiliQ water (20 to 50 µl).The RNA concentration was measured for each sample in duplicate using a NanoQuantMultiskan spectrophotometer and the purity of the sample was evaluated using the 260/280nm absorbance ratio.Reverse transcriptase-PCRs (RT-PCR) were carried out using the iScript RT Kit for cDNA synthesis (BioRad, 1708841).Where the final mix include 1000ng of RNA sample and 5µL of iScript RT-PCR mix with reverse transcriptase, final mix was carried to a final volume of 20ul.The cDNAs obtained were then used as a template for real-time PCRs carried out using the Eva Green qPCR Master Mix (SolisBioDyne, 08242510.6) in a StepOne Plus Machine (Applied Biosystem).The final PCR mixture (20ul) contained 1ul of cDNA, 4 ul of 5xFirePol MasterMix, 0.2umoles of each primer and was carried out to the final volume of 20ul.The thermal profile used for the reaction included a 2-minute heat activation of the enzyme at 95ºC, followed by 35 cycles of denaturation at 95ºC for 15 seconds and annealing/extension at 58ºC for 60 seconds, followed by melt analysis ramping at 58-90ºC.Negative control was conformed of water instead of cDNA and were included in each plate.Relative transcript levels were assessed using the Comparative CT method and expression values were normalized to 28S ribosomal expression, used as an internal control.One biological replicate corresponded to a homogenized solution of 20 fly heads minimum.For statistical analysis when comparing genotype and treatment, two-way ANOVA tests were performed with analysis of variance and Dunnet multiple comparisons for more than two groups using GraphPad Prism 6. P-value: **** p < 0.0001; *** p < 0.001; ** p < 0.01, * p<0.05 and ns > 0.05.
Strains and Culture.Strains carrying the following transgenes were obtained from the Bloomington Drosophila Stock Center (BDSC) from Indiana University: UAS-Mito-GFP (BL#8443, encodes the 31 amino acid mitochondrial import sequence from human cytochrome C oxidase subunit VIII fused to the N-terminus of the Green fluorescent protein), UAS-MitoTimer (BL#57323, encodes a mitochondrial targeting sequence with a roGFP which turns its fluorescence from green to red when oxidized ), Elav-Gal4 -Gal4 (BL#485), Actin-Gal4 (BL#9431), GH146-Gal4 (BL#1104), GH146,GFP (BL#36500); RNAi lines for UPR MT genes including crc (BL# 25985), dve (BL#26225), ubl (BL#65893), and RNAi lines for UPR MT associated H3K9 methylation enzymes dSetdb1 (BL# 31352), dSetdb1 loss of function (BL#30566), Utx (BL#34076), and Kdm2 (BL#33699).Wild-type flies used as controls are Canton S (BL#64349), and RNAi Control flies from the Transgenic RNA Interference Project (TRIP) (BL#35787).Only female flies were used for all experiments to avoid genetic variation and aggressive behavior from males.All fly stocks were maintained on a standard Drosophila medium which consisted of 112.5g of Molasses, 35g of dry yeast, 90g of corn flour, 9g of agar, 2.5g of Tegosept diluted in 10ml of ethanol 95%, and 6ml of propionic acid per 1L of water; at 25ºC and under a circadian cycle of 12h of light and 12h of darkness.Treatment with Mitochondrial Stressor Paraquat.Experiments requiring mitochondrial stressor paraquat (Sigma, 36541) were performed by supplementing standard Drosophila medium with 100µl of paraquat diluted in dH2O at 10µM.After PQ addition, vials must be airdried before use.Groups of flies were exposed to paraquat-supplemented medium for 48hrs before experiments were conducted.Olfactory functional assay.Olfactory T-maze was used to perform the olfactory behavioral test based on Hussain et al. 2018.Briefly, 15 flies are presented with an abrasive odor 0.1M of Hydroxychloroquine, 3-octanol (Sigma, W358118), or a pleasant odor of 2.3-butanedione (Sigma, B85307) at the end of one arm of the T-maze and at the end of the opposite arm flies are exposed to control solution (vehicle only).Flies have 60 seconds to discriminate between odors and go to an arm of the T-maze.At the end of the 60 seconds, an image is acquired, and flies in both arms are counted.The olfactory preference index consists of ((Flies in Experimental Odor -Flies in Vehicle Odor) / (Total flies in the experiment)).The olfactory preference index is calculated for every trial, and every n in the graph corresponds to the mean of 5 trials of 15 flies each.For statistical analysis when comparing genotype and treatment, two-way ANOVA tests were performed with analysis of variance and Dunnet multiple comparisons for more than two groups using GraphPad Prism 6. P-value: **** p < 0.0001; *** p < 0.001; ** p < 0.01, * p<0.05 and ns > 0.05.
14 min at 13000g at 4C.The upper soluble phase was transferred to a new 1.5ml Eppendorf for membrane and plasma proteins and pellet, and 200ul of liquid phase was kept for nuclear proteins.Pellet was dissolved in the liquid phase by pipetting.Quantification of samples was performed using the Pierce BCA Protein Assay Kit (Thermo Scientific, 23225) under the manufacturer's instructions.Samples were boiled in SDS sample buffer for 15 min, separated on an SDS-PAGE gel, transferred, and revealed using BioRad TransBlot and ChemiDoc, respectively.Primary antibodies used were rabbit anti-H3K9me3 1:2000 (Abcam, ab8898), and loading control anti-Tubulin 1:1000 (Thermofisher, MA1-744).Secondary antibodies were anti-rabbit conjugated with Horse Radish Peroxidase (HRP) 1:1000 (Thermofisher).The stained membranes were briefly incubated in luminol and scanned using ChemiDoc (BioRad).One biological replicate corresponded to a homogenized solution of 20 fly heads minimum.The normalized H3K9me3 levels were calculated by normalizing the ratio of H3K9me3 and loading control to that of young control samples.The significance of the interaction between genotypes and time was calculated by a two-way ANOVA test with Dunnet multiple comparisons using GraphPad Prism 9. P-value: **** p < 0.0001; *** p < 0.001; ** p < 0.01, * p<0.05 and ns > 0.05.
Image quantification.For image quantification 3D reconstruction of labeled structures was performed in Imaris Software.The surface of the desired signal was rendered.And the following parameters where quantified.Integrated density: This parameter represents a cumulative metric of the fluorescence signal within a specified region, denoting the aggregate of signal intensity and its spatial distribution.It is computed by multiplying the average fluorescence intensity by the volume of the signal-bearing domain, thereby yielding a singular value that encapsulates both the concentration and extent of the fluorescent activity.This approach ensures normalization for variations in the volume of the assessed Region of Interest, thereby facilitating a more accurate comparative analysis across samples.. Mitochondrial Morphology: Mitochondrial changes during aging in OPNs were analyzed by rendering the surface of mitochondrial reporter mitoGFP expressed specifically in the OPNs.Total mitoGFP volume (µm 3 ), mitoGFP puncta number, mitoGFP fragmentation index which corresponds to volume (µm 3 ) per area, (µm 2 ) mitoGFP sphericity index, mitoGFP average size (µm 3 ), and integrated density were analyzed.Mitochondrial 9. P-value: **** p < 0.0001; *** p < 0.001; ** p < 0.01, * p<0.05 and ns > 0.05.Single Cell RNA-seq Data analysis.To analyze gene expression in different cells population between the Drosophila aging brain, we used SCope (http://scope.aertslab.org)or the "ScopeLoomR" package in R with the scRNA-seq data "Aerts_Fly_AdultBrain_Filtered_57k.loom"under accession code GEO:GSE107451.We compared the AUC (Area Under the Curve) values derived from SCope, which indicate the activity levels of genes under regulons across diverse cellular populations.These values, reflects the combined activity of gene sets regulated by specific transcription factors, allowing to infer changes in gene expression.By assessing the AUC values for specific genes of interest-namely Hsp60A, Hsc70-5, dSetdb1, Utx, and Kdm2-we could quantitatively evaluate their activity in their respective regulon within distinct neuronal populations, including cholinergic (vAChT), glutamatergic (vGlut), GABAergic (Gad1), and olfactory projection neurons (OPN).This approach allowed us to quantify a proxy for gene abundance and enabling a nuanced understanding of the regulatory mechanisms at play.For a comprehensive understanding of the technical underpinnings and applications of SCope in single-cell transcriptomics, we refer to the work by Davie et al. (2018) 28 , which established a single-cell transcriptome atlas of the aging Drosophila brain.All data was tabulated in R and then plotted using GraphPad Prism 9 for statistical analysis.

Figure 6 .
Figure 6.dSetdb1 Knockdown Preserves Neuronal Integrity and Synaptic Density in the Aging Drosophila OPNs.(A) GFP-labeled olfactory projection neurons (OPNs) in the AL of 0 and 45 dpe control flies and flies with knockdown of dSetdb1 and dve.GFP-positive OPNs are gray, and the right panel shows merged channels with nuclei labeled with ToPro3 in cyan.Scale bar, 20µm.(B) Quantification of nucleus count in GH146-positive OPNs.The graph shows Two-way ANOVA with Bonferroni's multiple comparisons between 0 and 45 dpe shows that control flies had a significant decrease in nucleus count (p=0.0233,n=10).dSetdb1 RNAi (p>0.9999,n=10) Control vs. dSetdb1 RNAi of 45 dpe, (p=0.0398,n=10).(C) Orthogonal view of the 3D reconstruction of the distal axonal tract of OPNs tagged with GFP.Panel shows show the combination of axis Y and X, Z and X, and Z and Y. Panel shows axons from control flies, knockdown flies for dSetdb1.Scale bar, 5µm.(D) Quantification of axonal integrated density of axons shown in C. Two-way anova multiple comparison between Control flies show a significant decrease in axonal integrated density from 0 to 45 dpe (p<0.0001,n=10).dSetdb1 RNAi (p>0.9999,n=10).(E) Representative images of orthogonal view of the 3D reconstruction of GFP-tagged OPNs in the LH.Images show the combination of axis Y and X, Z and X, and Z and Y. Panel shows LH from 0 and 45 dpe control flies, knockdown flies for dSetdb1 and dve.Scale bar, 10µm.(F) Quantification of GFP integrated density in the LH of images shown in E. Two-way ANOVA with Bonferroni's multiple comparisons test shows a significant decrease in GFP volume in the LH of control flies from 0 to 45 dpe (p<0.0001,n=10).And no change for dSetdb1 RNAi (p>0.9999,n=10).(G) Orthogonal view of representative 3D reconstruction images of Brp::GFP-labeled presynaptic densities in LH of 0 and 45 dpe flies bearing the dSetdb1 GH146 knockdown.Scale bar, 20µm (H Quantification of BrpGFP integrated density of images shown in G. LH Brp::GFP integrated density, there was a significant decrease in Brp::GFP integrated density in control flies from 0 to 45 dpe (p=0.0031,n=6), while dSetdb1 RNAi flies did not show a significant change (p=0.8823,n=6).(I) Quantification of number of presynaptic densities labeled with BrpGFP in the LH of flies bearing the dSetdb1 knockdown shown in G.Control flies showed a reduction in the number of presynaptic densities labeled with Brp::GFP from 0 to 45 dpe (p<0.0001,n=6), and dSetdb1 RNAi no significant change (p>0.9999,n=6).White and red bars represent 0 and 45 dpe flies, respectively.n = independent fly brain.Pvalue: **** p < 0.0001; *** p < 0.001; ** p < 0.01, * p<0.05 and ns > 0.05.