Neuron-type specific aging-rate reveals age decelerating interventions preventing neurodegeneration

doi:10.21203/rs.3.rs-4360587/v1

Download PDF

Biological Sciences - Article

Neuron-type specific aging-rate reveals age decelerating interventions preventing neurodegeneration

https://doi.org/10.21203/rs.3.rs-4360587/v1

This work is licensed under a CC BY 4.0 License

Version 1

posted

You are reading this latest preprint version

Different types of neurons show distinct susceptibility to age-dependent functional decline and degeneration and are linked to different types of neurodegenerative disorders. The underlying reasons for different aging trajectories of distinct neuron types are poorly understood. Here, we employ aging clocks to assess whether distinct neurons differ in their biological age trajectory in C. elegans where the identity of each single neuron is known. We find that specifically ciliated sensory neurons with high neuropeptide expression show the most rapidly progressing biological age. The more rapidly aging neurons show a reduction of mitochondrial respiration and elevated protein translation gene expression. Reducing protein translation with cycloheximide effectively protected fast aging neurons. We show that the C. elegans neuronal aging pattern are highly correlated with human brain aging and contrasted by geroprotective interventions. We performed an in silico drug screen and identified known and novel neuroprotective small molecule compounds. We show that the natural occurring plant metabolite syringic acid and the piperazine derivative vanoxerine delay neuronal degeneration and propose that they could serve as neuroprotective interventions. We also identify neurotoxins that accelerate neurodegeneration indicating that this approach could reveal interventions as well as risk factors for neuronal aging.

Biological sciences/Developmental biology/Ageing

Biological sciences/Neuroscience/Neural ageing

Biological sciences/Computational biology and bioinformatics/Virtual drug screening

aging

transcriptomic clock

neurodegeneration

neuronal aging intervention

Age-related diseases are the leading causes of death in high-income countries with the highest prevalence of ischemic heart disease, Alzheimer’s Disease (AD) and other dementias, cancer of the respiratory system, intestinal cancer, kidney failure, and diabetes mellitus¹. The distinctive types of chronic diseases of aging indicate that different organs can become the weak-point of an organism that will cause its death and it was reported that nearly 20% of the population show accelerated aging in single organs compared to the rest of their body². In mice, molecular changes across multiple tissues occurring with aging have been described and unique molecular aging trajectories were identified^3–5. Indeed, different organs age at different and individual paces^2,6,7. How distinct cell types are differing in their susceptibility to age-related functional decline is largely unexplored.

Aging is also the highest risk factor for the development of neurodegenerative diseases like AD or Parkinson’s Disease (PD). The distinct neurodegenerative diseases are triggered by the functional decline of distinct neuron types. For instance, in AD, the first tissues degenerating are the parahippocampal gyrus and the olfactory bulb, followed by the degeneration of the hippocampus, leading to the characteristic clinical dementia symptoms^8–11. PD in contrast, affects mostly the dopaminergic innervation of the mid brain and, according to more recent studies, also the cerebellum, resulting in the perturbation of motility and the induction of tremor syndromes^12,13. Why and how distinct brain regions are differently affected in neurodegenerative diseases and whether different neuron classes would show intrinsically different aging trajectories is poorly understood. In order to identify neuron-specific aging trajectories, we employed the nematode Caenorhabditis elegans as experimental model with a well characterized neuronal system of 302 neurons in total, and ≥ 128 different neuron classes¹⁴. Importantly, the nematode system uniquely allows the assessment of the integrity of individual neurons during the normal aging process in vivo.

Aging trajectories are being increasingly well determined by aging clocks, which are predictive models of molecular signatures that estimate an individual’s chronological, and recently also biological age. Aging clocks based on DNA methylation or transcriptome data could predict biological age differences, resulting from genetic age acceleration or deceleration as well as lifespan-affecting treatments^2,15,16. Cellular heterogeneity and tissue composition might affect the prediction accuracy and interpretation of bulk aging clocks¹⁷, and recent single-cell transcriptomic clocks highlight cell-type specific aging pattern¹⁸. Still, aging clocks relying on bulk data provide a solid reference because they capture broader aging trajectories, summarizing molecular profiles from diverse cell types, and averaging-out outlier profiles, thus providing general insights into aging patterns and dynamics. While aging clocks have been used to identify biological aging differences across individuals, among distinct treatments, and recently also distinct organs, they have not been used to identify differences among distinct cell types.

Here, we asked whether distinct aging trajectories could be identified among the different and well characterized neuron types in C. elegans. We show that individual neuron types in young adult animals show a distinct biological age prediction that we find to correspond to the pace of their degeneration in vivo. Neuron types with a higher predicted age show the earliest neurodegeneration during adulthood while predicted young neuron types remain intact. The transcriptomic differences between fast and slow aging neuron types suggest protein translation as a crucial driver of the biological aging process. Slowing down translation with the established translation inhibitor cycloheximide (CHX) reduced the neuronal degeneration of fast aging neurons. We determined that the transcriptional pattern of biological age difference among these neuronal cell types are correlated with transcriptional pattern of human and mouse brain aging and anti-correlated with anti-aging interventions supporting the general validity of our approach. Using a transcriptomic resource incorporating thousands of different compounds in human cell lines (CMAP)¹⁹, we identified pharmacological compounds that we show to prolong the integrity of neurons in vivo. We demonstrate that the natural occurring plant metabolite syringic acid and the piperazine derivative vanoxerine delay neuronal degeneration and propose that they could serve as neuroprotective interventions. In reverse, our approach can also identify neurotoxins that accelerate neurodegeneration thus serving as risk assessment. Our data suggest that the mechanisms underlying neuron-type specific aging rates allow the identification of therapeutic interventions that could slow neuronal aging and prevent neurodegeneration.

To investigate whether different neuron classes age differently within an organism we applied our previously developed binarized transcriptomic biological age predictor (BitAge) ¹⁶ on pseudo-bulk data from the C. elegans Neuronal Gene Expression Map & Network (CeNGEN) dataset ²⁰, which comprises 128 distinct neuron classes from young adult worms. The BitAge clock is trained on the biological age of isogenic whole worm populations and highly accurately predicts not only the chronological age but especially the biological age. The 128 neuron age predictions range from ≈ 98 h in FLP neurons to ≈ 177 h in ADL neurons (Figure 1 A) suggesting that different neurons might indeed show an almost two-fold biological age difference in young adult worms. We independently confirmed the different age predictions of the neuron types in the neuronal pseudo-bulk data from young adult worms in a recent cell atlas of C. elegans aging (Calico) ²¹ which contains 67 of the 128 neuron classes from the CeNGEN dataset. Also, the Calico dataset of day 1 adult worms exhibits the same predicted age distribution (Supplement Figure 1A), and the BitAge predictions on both datasets are significantly correlated to each other (Supplement Figure 1B, Pearson correlation 0.64, p‑value = 5.92e-09). The youngest and oldest 10% of the 67 neuron classes (as predicted in the CeNGEN dataset) show a stronger correlation (Pearson correlation 0.75, p-value = 2e-03), while the middle group has a higher prediction variance. These results indicate that BitAge is able to predict age differences across different datasets robustly and that neurons indeed show biological age differences on the transcriptome level in young adult worms.

Next, we sought to replicate the neuronal age predictions with a method that is not based on the assumption of directed transcriptome changes but stochastic variation accumulation instead. Recently, we showed that all current aging clocks might be driven by accumulating stochastic variation, and that for biological age predictions almost no biological data is required ²². Instead, simulating the aging process by adding stochastic variation to a biological ground state is sufficient to build a predictor that is enabling chronological, as well as biological age predictions. We have shown that for transcriptomic data of C. elegans one biological sample as the starting point and stochastic variation drawn from a normal distribution is sufficient for accurate age predictions. This stochastic clock is thereby trained on a different concept than BitAge, which is trained to find biological age pattern in biological samples. Also, the predictions with this stochastic clock are significantly correlated with the BitAge predictions on the CeNGEN dataset (Supplement Figure 1C, Pearson correlation 0.65, p-value = 5.5e-17). Similar to Supplement Figure 1B, the 10% youngest and oldest 10% of the 128 CeNGEN neurons show a stronger correlation (Pearson correlation 0.87, p-value = 1.15e-08), while the middle group has a higher prediction variance. These results corroborate the biological age prediction differences, especially in the youngest and oldest neuron groups.

The aging transcriptome in species ranging from C. elegans to mice was recently shown to exhibit a gene length dependent transcriptional decline (GLTD), where long genes are downregulated with age, while short genes are upregulated ^23–25. The reduced expression of long genes likely results from the heightened susceptibility of long genes to accrue transcription-blocking DNA damage. In line with this feature of the aging transcriptome, the marker genes of the 10 % oldest predicted neurons are significantly shorter than expected by chance (adjusted p-value = 0.013), while the marker genes of the 10 % youngest neurons are significantly longer than expected (adjusted p-value = 0.037) (Figure 1B). These results indicate that even in chronologically young, but biologically old neurons a gene-length dependent transcriptome imbalance can be observed.

Taken together, biological and stochastic aging clock measurements and GLTD all suggest specific neuron types to age more rapidly than others.

Neuron specific age predictions are associated with differential degeneration

To assess whether the predicted neuron-specific age differences are associated with different degrees of neuron-specific cellular degeneration, we next chose three young (I2, OLL, PHC) and three old (ASI, ASJ, ASK) predicted neurons (Figure 2A) and scored their degeneration over the chronological age (Figure 2B). Green fluorescent protein (GFP) was expressed under promoters specific for those neurons (ASI, ASJ, ASK, and OLL) or, alternatively, under promoters specifically expressed in a subset of neurons to make identification and segmentation of the selected neurons easier (I2 and PHC) (Figure 2A). Macroscopic aberrations on the neurites were counted and subsequently, neurons were classified as healthy, damaged, or severely damaged. In accordance with our predictions, the three young predicted neurons show less degeneration than the old predicted neurons at all analyzed timepoints (Figure 2C). On the first day of adulthood, that is closest to the age of the nematodes used for transcriptomics and subsequently for our prediction, I2, OLL, and PHC exhibit a minimal degeneration offset between 10 – 20% of all nematodes analyzed. Upon aging, there is a slight, non-significant increases of the fraction of nematodes displaying degeneration (up to ≈ 35%) for the young predicted neurons. This fraction of degeneration is consistent with our previous observation of approximately 35% degeneration at day 7 of adulthood in URY neurons ²⁶, which we here predict to belong to the top 10 youngest neurons (Figure 1A).

The ASI, ASJ, and ASK neurons, which are predicted to be biologically older, exhibit >45% damaged neurons already on the first day of adulthood. Interestingly, there is a significant increase of neurodegeneration in ASI (p-value=0.027, Cohen's h=-1.02, i.e. a large effect) and ASK (p-value=0.049, Cohen’s h=-0.6, i.e. a medium effect) neurons upon aging which is in stark contrast to the young predicted neurons where no such elevated degeneration levels could be observed. These results suggest that the predicted biological age differences of the youngest, respective oldest neurons are biologically meaningful and can serve as a predictor of neuronal degeneration already at day 1 of adulthood.

Environmental exposition could be a discriminator for premature aging in neurons

Next, we aimed to understand potential commonalities among the youngest, as well as among the oldest neurons. For this, we adapted a hierarchical whole-animal connectome for C. elegans ²⁷ with a rough anatomical correspondence on the x-axis and directional flow of neuronal signaling on the y-axis and color-coded it with the predicted biological age (Figure 3A). The predicted oldest neurons cluster in the top middle part of the network and consist mostly of sensory neurons, while the youngest neurons cluster further to the right. 6 out of the 10 oldest neurons are amphid neurons (ADL, ASJ, ASK, ASG, ADF, ASI), the primary chemosensory organ of C. elegans, which is mostly ciliated²⁸. Indeed, comparing the 14 amphid neurons of the CeNGEN dataset with the remaining 114 neurons shows a significant increased biological age (Figure 3B, p-value=1.8e-07), which can be replicated in the Calico dataset (Supplement Figure 2A, p-value=4.2e-08), and with the stochastic data-based clock predictions (Supplement Figure 2B, p-value=7.2e-22). Amphid neurons, as part of the sensory system, express a variety of neuropeptides, neurotransmitters, receptors, and innexins to transmit the sensed cues. The number of expressed neuropeptides and receptors are significantly higher in amphid neurons (Supplement Figure 2C,D), while the number of neurotransmitter or innexins is not significantly changed (Supplement Figure 2E,F) ²⁹. Moreover, the number of neuropeptides and the number of receptors per neuron are significantly positively correlated with the predicted biological age in the CeNGEN dataset (Supplement Figure 2G,H), the Calico dataset (Supplement Figure 2I,J), and the predictions with the stochastic data-based clock (Supplement Figure 2K,L). Depletion of unc-31 leads to reduced neuropeptide release and exhibits a mild, but significant reduction of degeneration in ASI neurons (Supplement Figure 2M, p-value=0.03, Cohen’s h=0.36). The number of innexins and number of total synapses per neuron do not show a significantly positive, and potentially rather a negative correlation with the predicted ages (Supplement Figure 2N-S).

Amphid neurons are part of the ciliated neuron classes; comparing all 28 ciliated neurons with the remaining 100 neurons also shows a significant increased biological age (Figure 3C, p-value=0.0006), which can be replicated in the Calico dataset (Supplement Figure 2T, p-value=9e-05), and with the stochastic data-based clock predictions (Supplement Figure 2U, p-value=2.3e-14). The ciliated neurons can be further divided into five distinct classes dependent on where its cilia terminate²⁸. Neurons with cilia exposed to the environment show the highest predicted biological age (Figure 3D, 1-way ANOVA: 8.8e-06), while the other ciliated neuron classes are not significantly different from not-ciliated neurons. A similar effect can be observed in the Calico dataset (Supplement Figure 2V, 1-way ANOVA: 1.8e-08), and the predictions with the stochastic data-based clock (Supplement Figure 2W, 1-way ANOVA: 6.6e-18). These results indicate that the oldest neurons are functionally related and mostly consist of ciliated sensory neurons; that contact to the environment; and the production, potentially the translational load, of neuropeptides are associated with more rapid biological age progression.

Transcriptional Clustering identified reduced translation efficacy as potential driver of neuronal aging

We next sought to identify the age-related transcriptional patterns and signatures underlying the biological age distinctions across the 128 neuron classes. To mitigate data variance and extract overarching trends, we initially categorized the neurons into five distinct groups based on their predicted age, followed by a fuzzy clustering analysis. We identified 4 transcriptional clusters (Figure 4A, Supplement Figure 3A): Cluster 1 shows a general increase over the predicted age and is enriched for stress-induced pathways including DNA repair, response to DNA damage stimulus, transcription-related pathways, and synthesis of ribosomal components, while oxidative phosphorylation is under-represented (Figure 4B). Cluster 2 shows the highest expression in the youngest age group, generally declines over the predicted age time-course, and is enriched in mRNA processing, active translation, and oxidative phosphorylation. Cluster 3 is showing an increase until the last age group, in which it sharply declines, and is enriched in ribosomal genes and proteasome core complex genes. Cluster 4 is especially increased in the oldest age group and is enriched in cilia, immune response, and neuropeptide genes. Conversely, translation-related pathways and oxidative phosphorylation are under-represented. As shown above, the oldest age group is enriched in amphid neurons (ADL, ASJ, ASK, ASG, ADF, ASI, AWA, ASEL), which are mostly ciliated²⁸ and exposed to the environment (Figure 3). The strong enrichment of translation-related pathways (Cluster 1) in the highly expressed genes in the most rapidly aging neurons and the lower translation-related pathways (Cluster 4) in the most slowly aging neurons is consistent with recent studies on transcriptional changes with chronological age in the brain of different organisms^30–32.

Inhibition of translation alleviates neurodegeneration in fast aging neurons

Based on the enrichment of active protein biosynthesis processes in the accelerated aging neurons, we aimed to test whether translational activity contributes to neurodegeneration. Hence, we treated nematodes for 24 h with the translation inhibitor cycloheximide (CHX) and scored neurite degeneration in ASK and ASJ neurons (from the group of old predicted neurons), as well as in I2 and OLL neurons (as representatives of the young predicted neurons). In the young predicted neurons, no effect of CHX or a DMSO-control treatment was observed (Figure 4C). In contrast, old predicted neurons exhibited significantly less neurite deterioration upon CHX-treatment, with a Cohen’s h of 1.4, i.e. large effect size, for the ASI, and a Cohen’s h of 0.79, i.e. medium to large effect size for the ASK neuron. These results indicate that translational activity in the old predicted neurons is responsible for the premature neurodegeneration that was observed.

Comparison with mammalian brain aging

In order to see whether the biological age-related transcriptional patterns of chronologically young adult C. elegans neurons (NeuronAge) might be conserved to higher organisms, we next compared the conserved KEGG pathway enrichments of NeuronAge with mouse and human datasets. We computed age-correlations of z-score normalized gene counts for all human brain regions in the GTEx dataset³³, the Tabula Muris Senis (TMS) dataset⁵, and an additional mouse Hypothalamus aging cohort (GSE157025). Similarly, we calculated the enriched pathways for several “anti-aging” treatments like young serum injections³⁴, the platelet factor PF4³⁵, sport in humans³⁶, and krilloil in C. elegans³⁷. An unbiased clustering analysis revealed that the aging-trajectories of C. elegans, mouse, and humans cluster together. We validated the clustering of NeuronAge by including the conserved pathway enrichments for NeuronAge on the Calico dataset. The trajectories of the anti-aging interventions formed a separate cluster that negatively correlates with the brain aging, irrespective of the organism (Figure 5). These results indicate that neuronal transcriptomic aging trajectories are conserved from nematodes to mice and humans and that known anti-aging treatments largely anti-correlate with the aging datasets supporting their geroprotective effectiveness.

Identification of drugs preserving neuronal function

As the NeuronAge trajectories cluster together with human neuronal aging trajectories, we sought to use transcriptome data to identify small molecule compounds that could delay neuronal aging. We used the transcriptome resource CMAP consisting of 470k transcriptomes of 19,841 different pharmacological compound treatments in human cell lines¹⁹. We focused on the 3,566 samples for the terminally differentiated neuronal cell line NEU, consisting of 2,467 different molecules (Figure 6A,B). Based on the NeuronAge prediction, this analysis identified both negatively correlated compounds (potentially neuro-protective / anti-aging active) and positively correlated compounds (potentially neuro-toxic / pro-aging active). Pathways enriched upon CHX treatment compared to control are significantly negatively correlated with pathways enriched in NeuronAge (Pearson Correlation -0.22), indicating that the beneficial effect of CHX that we saw is mirrored in the transcriptome. To identify neuro-protective small molecule compounds, we ranked the enriched pathways for all 170 molecules that remained after filtering and before using an absolute correlation threshold of 0.25 (see Source Data). The top anti-NeuronAge compound hits contain several (9 out of 16) for which a protective effect for neurons has been previously documented, thus validating our approach (Figure 6B). The glycogen synthase kinase 3 (GSK3) inhibitor AR-A014418 was shown to inhibit beta-amyloid induced neurodegeneration³⁸; the selective serotonin reuptake inhibitor fluoxetine protects against neurotoxicity and neurodegeneration^39–41; the PPAR-alpha activator gemfibrozil exhibits neuroprotective effects via upregulating pro-survival factors and suppressing inflammation⁴²; the kinase inhibitor sorafenib protects against neurodegeneration in C. elegans ⁴³; the selective aryl hydrocarbon receptor modulator 3,3'-diindolylmethane (DIM) is neuroprotective and promotes brain-derived neurotrophic factor (BDNF) ^44,45; the insulin-sensitizing agent rosiglitazone exhibits neuroprotective effects in the eye and the brain ^46–48; the p38 MAPK inhibitor SB202190 was shown to reduce hippocampal apoptosis and rescue spatial learning as well as memory deficits in rats⁴⁹; dibutyryl-cAMP-Na (dBcAMP) elevates cAMP levels and protects against neurodegeneration in stab wound or kainic acid injuries^50–52; and the catecholamine-O-methyltransferase inhibitor tolcapone was shown to improve cognitive function⁵³.

2 out of 15 “pro-NeuronAge” compounds were shown to be detrimental, while 2 are potentially protective. BAY-K8644 is known to be neurotoxic⁵⁴; and amiodarone induces neuronal apoptosis⁵⁵ and is known to induce adverse neurological effects⁵⁶. Tacedinaline/CI-994 is a class I histone deacetylase inhibitor correlates positively with NeuronAge, was shown to promote functional recovery following spinal cord injury⁵⁷, and to enhance synaptic and structural neuroplasticity⁵⁸. Of note, this effect might be due to a hormetic response^59–61. Likewise, resveratrol is potentially neuroprotective⁶² due to a hormetic response⁶³. More than half of the top hits have, however, not been tested in neurons. It is conceivable that a short-term “pro-NeuronAge” effect might be hormetic and anti-aging after more time has passed, potentially explaining the effect of resveratrol and tacedinaline.

In summary, 11 out of 31 compounds have neuroprotective evidence, out of which 9 are predicted to revert NeuronAge, i.e. “anti-NeuronAge”, with our in-silico approach, while 2 out of 31 compounds are known to be neurotoxic, both of which are predicted correctly to be “pro-aging”, giving weight to the potential that an in-silico screen has to identify novel compounds.

Identification of neuro-protective molecule compounds

Next, we aimed to determine whether compounds that we predicted to be anti-NeuronAge, i.e. neuroprotective, could indeed prevent the age-related functional decline of aging neurons. We chose two compounds, that were among the most strongly anti-correlated with NeuronAge patterns, BRD-K13195996 and vanoxerine (Figure 6B). The chemical identity of the phenolic compound BRD-K13195996 is 3-Hydroxy-4,5-dimethoxybenzoic acid, which is related to 4-Hydroxy-3,5-dimethoxybenzoic acid that is known as syringic acid. Syringic acid is a naturally occurring secondary compound derived from edible plants and fruits, among those olives, walnuts, and grapes – and furthermore red wine and honey ⁶⁴. A correlation between the anti-oxidative properties of syringic acid and reduced neurotoxicity following bisphenol A insult has recently been shown⁶⁵, yet no clear mechanism is reported so far⁶⁶. Given the dietary availability of syringic acid, we chose to test its effect on rapidly aging neurons in C. elegans. Vanoxerine is a potent dopamine uptake inhibitor and has been developed as cocaine-abuse medication ⁶⁷, and, moreover, vanoxerine was observed to impede colorectal cancer stem cell functions by repressing G9a expression ⁶⁸. Vanoxerine was so far not reported to exhibit neuroprotection or anti-aging effects.

We applied either compounds to nematodes for a 24 h short-term treatment. We assessed neurite degeneration in ASJ and ASK neurons (exemplarily for the old predicted neurons) and observed a significantly reduced deterioration for both compounds, with Cohen’s h ranging from 0.8 to 0.97, i.e. large effect sizes (Figure 6C). Applying either of the compounds to nematodes showed no significant adverse effects an OLL neurons (Supplement Figure 4A). This indicates that both compounds interfere with the physiological degeneration process of the old predicted neurons and are able to restore a healthy neuron state.

Next, we assessed whether our NeuronAge compound predictions could also identify neurotoxic compounds and hence serve for compound risk assessment. We tested the 5-HT1A serotonin receptor antagonist WAY-100635, for which so far no adverse effects on neuron health have been reported, in nematode I2 and OLL neurons (as representatives of healthy young neurons). We observed that WAY-100635 induced significant neurite deterioration in I2 (p-value=0.02, Cohen’s h=-0.76) but not in OLL (p-value=0.08, Cohen’s h=-0.34) neurons (Figure 6C). This indicates that WAY-100635 does not have an indiscriminate effect on all neurons but is more selective, potentially depending on surface receptor expression, presentation, or specific neuronal metabolism patterns.

Taken together, we could validate the anti-NeuronAge compound prediction method by identifying known neuroprotectors as well as discovering previously unknown neuroprotective molecules. In reverse, a positively correlated NeuronAge prediction could identify neurotoxic compound properties.

Why distinct neuron types exhibit different susceptibility to age-dependent degeneration and the associated neurodegenerative diseases has remained largely unclear. While differences in inter-individual aging are commonly known, differential aging of organs within the same organism, and aging variance between cells of the same tissue have recently been observed ^2–5. Here, we aimed to understand the aging pace of distinct neuron types to elucidate whether and why specific neurons age faster than others. We employed an aging clock approach to predict the age of distinct neuron types. Aging clocks are increasingly useful for measuring biological age and hold the promise to assessing an individual’s biological age. Employing the ‘BiT age’ and the ‘stochastic aging clock’ on the single neuron transcriptomics dataset (CeNGEN) we predicted the biological age of the 128 distinct neuron types in C. elegans. We observed that the youngest predicted neurons' biological ages were roughly the chronologic age of the nematodes, while the oldest predicted neurons’ biological ages were about 1.5-fold as old. The age dependent decline of long gene expression confirmed the distinct biological age of certain cell types at the same chronological age of the animal. The nematode model provides the distinct advantage that the integrity of single neurons could be followed during aging in live animals. We indeed observed that the age-dependent degeneration of specific neurons corresponds to the prediction of the aging clocks evidencing their reliability in identifying neuron-type specific aging trajectories.

In the nematode the identity and connectivity of each individual neuron is known, thus allowing to address how their distinct biological age is linked to their biological function. We found that particularly ciliated sensory neurons show an accelerated age trajectory. This might be linked to their exposure to the environment but could also indicate their functional requirement during larval development where the sensing of environmental condition, for instance for deciding to enter dauer stage amid food scarcity or overcrowding, is pivotal for survival. Similar to the exposed ciliated neurons in the nematode, olfactory neurons in the nose of higher organisms are constantly exposed to the environment. Indeed, in humans olfactory capacity is known to decline with age and olfactory dysfunction is an early sign of neurodegenerative diseases⁶⁹. Our results shed light on the possibility that this exposure to the environment might lead to a faster pace of aging and subsequent neurodegeneration.

We show that the neuron-type specific biological age differences can be used to determine neuroprotective transcriptome compositions. We identify distinct transcriptional patterns over the neuronal biological aging course and predicted that reduction of the translational load might reduce the pace of neuronal aging. Age-related changes in the translational machinery and the translation rate can be observed across various species and downregulation of translation has been shown to extend lifespan and healthspan parameters as evidenced by dietary restriction studies, downregulation of mTOR, or CHX treatment ⁷⁰. Recently, it has been shown that stochiometric changes of the ribosome are prominent in the aging brain of Nothobranchius furzeri, leading to enrichment in protein aggregates in old brains ³⁰, which is in line with the over-enrichment of ribosomal proteins in the insolublome of old C. elegans ⁷¹. Consistent with prediction and literature, we observed that transient inhibition of translation by CHX treatment is sufficient to reduce degeneration of the fast-aging neuronal cell types.

As neuron function is highly conserved from nematode to human, we tested whether age-dependent transcriptome changes might be similar. Indeed, we found that the transcriptional patterns of nematode neuronal biological age differences are significantly correlated with mouse and human brain aging trajectories, and anti-correlated with known anti-aging interventions such as young plasma treatment or sport. This conservation shows the potential of identifying conserved mechanisms that underlie the aging trajectories and might determine the susceptibility of specific neuron types to undergo degeneration and potentially contribute to specific neurodegenerative diseases.

We used the conservation of transcriptome age-trajectories to in silico screen for novel compounds using the human CMAP dataset ¹⁹. Such approaches are highly valuable as recent studies employed a transcriptome-based approach for drug screening using the CMAP resource to successfully identify geroprotective compounds that either induce a ‘youthful’ state as predicted through an age-classification approach leveraging the GTEx ³³ transcriptomic dataset ⁷², by mimicking longevity FOXO3 overexpression ⁷³, or a “youthful” matreotype ⁷⁴. Here, we extended this approach further and used transcriptomic data from C. elegans to compare to the CMAP resource to identify neuro-protective small molecule compounds. This approach successfully picked up nine known geroprotective interventions and also identified molecule compounds whose effect on neuroprotection was previously unexplored. By testing the top scoring compounds, we indeed found that they extend the integrity of fast aging neurons indicating a biological age-deceleration. Conversely, the pro-aging prediction revealed neurotoxic effects of compounds and could thus be highly valuable in risk assessment.

As proof of concept, we determined that syringic acid and vanoxerine effectively preserved the integrity of fast aging neurons. Syringic acid indeed has been suggested to exert neuroprotective effects through its antioxidant properties⁶⁶. Both vanoxerine and WAY-100635 have been developed to treat cocaine addiction but our data indicate starkly contrasting effects on neuronal aging. The high-affinity dopamine reuptake inhibitor vanoxerine was initially developed as antidepressant and has entered clinical trials for treatment of cocaine addiction⁷⁵ and, based on its property as ion channel blocker, for atrial fibrillation or atrial flutter⁷⁶. The 5-HT1A serotonin receptor antagonist WAY-100635 has been tested preclinically for cocaine addiction⁷⁷ and treatment of depressive disorders⁷⁸. We propose that NeuronAging clocks and the aging-associated gene expression responses we determined here could be useful in risk assessment given the strong similarities we show between the C. elegans neuronal aging trajectories and human brain aging

Taken together, we here define the biological basis for the distinct susceptibility of neurons to undergo age-dependent degeneration. We establish the utility of employing aging clocks to identify neuron-type specific aging rates and based on their transcriptome profiles reveal conserved aging pattern also present in human brain aging. We show that this approach is suitable for identifying neuroprotective molecules and propose that they could be useful in delaying neuronal aging and protect from age-associated degeneration.

C. elegans culture.

Nematodes were cultured on nematode growth medium (NGM) agar plates at 20 °C under standard conditions unless stated otherwise. All age statements given in in this publication consider the first day of adulthood as day 1. A complete strain list can be found in the supplement.

Neurite imaging

Nematodes were synchronized by L4 picking and grown on standard NGM for one day, four days, or seven days. For imaging nematodes were placed in a drop of 250 mM NaN₃ on a 2% agarose pad. Imaging was performed on a Zeiss Imager.M2 at 400fold magnification. Z-Stacks of nematode heads / tails were acquired employing 2 µm step width. Acquisition time was set between 300 ms to 3 s per plane to achieve a good signal to noise ratio and was strongly depending on the imaged expression strain.

Scoring of neurite degeneration

Recorded Z-stack images of neurons were analyzed by hand, counting blebs, large spherical outgrowths, branching, breaks, and necrosis on the dendrites of the analyzed neurons. Images were classified according to the degree of aberration: necrotic neurons, broken or truncated neurons, neurons with ≥ 10 blebs, or ≥ 3 outgrowths were scored as ‘severely damaged’; neurons with 5 – 9 blebs, or 2 outgrowths were classified as ‘mildly damaged’; and neurons with less < 5 blebs or < 2 outgrowths were classified as ‘healthy’. See Figure 2B for exemplary images.

Compound treatment

Standard NGM plates seeded with OP50 were coated with compounds by dropwise adding compounds, dissolved in 300 – 1000 µl medium, directly to the plate’s surface. Plates were dried for at least 1 h before transferring L4 stage nematodes onto them. Nematodes were incubated with the compounds for 24 h and then used for neurite imaging. Final concentration of compounds was: Cycloheximide – 2 mM; Syringic acid – 2.5 mM; Vanoxerine – 10 nM; WAY-100635 – 25 nM. Control nematodes were incubated on appropriate solvent control coated plates (either water or ≤ 5 ‰ DMSO).

BitAge prediction

The BitAge clock¹⁶ was used as described previously. Briefly, each sample was binarized, i.e. genes higher than the median expression value within each sample after removing genes with zero counts, were set to 1, and the remaining genes to 0. The BitAge coefficients for the 576 clock genes are added up for all genes in a given sample that is 1 after binarization. After adding the BitAge intercept the results show the predicted biological age.

Stochastic-data based clock

The stochastic-data based clock was used as described previously²². Briefly, each sample was log10-transformed after the addition of one pseudo-count. Subsequently, the samples were min-max normalized to bring each sample within the range [0,1], and then binarized as described above. The normalized counts were then added up for all 1010 stochastic-data based clock genes (see Source Data). Note that the stochastic clock might result in slightly different genes every-time a clock is trained.

Gene length analysis

First, we downloaded the differential expressed genes for each neuron and all other cells in the CeNGEN²⁰ dataset (https://cengen.shinyapps.io/CengenApp/) with the statistical test “Wilcoxon on single cells”. This gives a list of genes with log fold changes, and percentage expression in the specific neuron and all other cells. We further filtered this list of significant genes by requiring that the gene is expressed in at least 90% of cells of the specific neuron and at most 10% in all other cell types. 39 neurons had no genes with these requirements, i.e. 89 neurons were used for further analysis. Next, we used the marker genes of the 10% oldest (31 genes), respective 10% youngest neurons (33 genes) and calculated the density distribution of the log10-normalized gene lengths. To calculate a two-sided permutation test, we calculated the median log10 gene length of the genes and compared it to the 100.000 permutation median. The permutation for the old marker genes used 31 genes, the permutation for the young marker genes 33 genes out of all marker genes of the 89 neurons (303 genes).

Neuronal connectome mapping

We downloaded and adapted the Cytoscape file from Cook et al.²⁷ by adding head neurons, deleting non-neuronal cell-types, and color-coding neurons by their predicted Age with BitAge on the CeNGEN dataset.

Median total number of synapses calculation

The NeuroType.xlsx file was downloaded from https://www.wormatlas.org/neuronalwiring.html. For each of the 128 neuron classes we summed up the median total number of synapses in the head, tail, and mid-body.

Fuzzy clustering

To cluster general trajectories over the predicted age, we first summarized the gene expression into 5 bins: 1) [97-110], 2) (110,120], 3) (120,130], 4) (130,140], 5) (140, 180]. Within each bin, we computed the median expression level for each gene. To make the genes comparable and bring them onto the same scale we calculated the z-score across the 5 bins for each of the 9950 genes with non-zero standard deviation. Next, we used fuzzy clustering with Mfuzz v2.58.0 ⁷⁹ to identify trajectories across the predicted age bins. The elbow method was computed with the Dmin function of Mfuzz and indicated an optimal number of 4 clusters. The genes belonging to each cluster were subsequently used for a pathway enrichment analysis with clusterprofiler v4.9.2.002 ⁸⁰, with maxGSSize=500 and the list of all 9950 genes as the background gene list.

Heatmap

We processed several public datasets for the heatmap:

TPM normalized gene expression values for human brain tissues from the GTEx v8 data⁸¹ release (i.e. Amygdala, Anterior Cingulate Cortex, Caudate Basal Ganglia, Cerebellar Hemisphere, Cerebellum, Cortex, Frontal Cortex, Hippocampus, Hypothalamus, Nuclear Accumbens Basal, Putamen Basal Ganglia, Substantia Nigra) were correlated with the chronological age (the midpoints of the publicly available age bins).
The mouse aging time-course for Hypothalamus data from GSE157025 was downloaded and a gene-wise correlation with the chronological age calculated.
Whole brain data from the Tabula Muris Senis (GSE132040) was downloaded and edgeR v3.40.2 ⁸² to calculate normalized expression values. The normalized expression values were correlated to the chronological age.
Differentially expressed genes for mouse Hippocampus data treated with the platelet factor PF4 or saline control were downloaded from GSE173254. We multiplied the logFCs by -1 to always compare treatment vs. control, instead of control vs. treatment.
Differentially expressed genes for mouse Hippocampus data treated with young serum or sham were downloaded from GSE234667.
C. elegans whole worm data treated with Krill oil from GSE207152. We used edgeR v3.40.2 ⁸² to calculate normalized expression values and calculated z-scores for each gene over all samples. The z-score normalized expression values were used for a regression model: , where is the intercept term, is the coefficient for the Age variable, is the coefficient for the Krilloil-treatment variable, and is the coefficient for the interaction between Age and Krilloil, i.e. the difference in the slope over age.
Differential expressed genes upon physical activity were downloaded from the supplementary data from PMID: 30927700.

The Pearson correlation values of all genes of 1.) - 3.), the logFC of all genes for 4.)-5.), the coefficients for 6.) were used to calculate enriched pathways analysis with fgseaMultilevel from the fgsea R package⁸³ with nPermSimple=1000 for all conserved KEGG pathways between C. elegans, mouse, and humans. For 7.) the “anti-aging/AD” genes were split into genes that are up-, respective down-regulated upon exercise. Both gene sets were used for an enrichment analysis with enricher from enrichplot v1.18.0 with maxGSSize=500, minGSSize=5, and the all genes quantified in the supplementary data from PMID: 30927700 as the background gene list. For both enrichment analyses the enrichment fold change, i.e. number of observed genes per pathway divided by the number of expected genes per pathway, was calculated. Finally, the fold changes were combined, i.e. pathways with a bigger fold change enrichment in the downregulated genes were multiplied by -1.

The clustering was done on the normalized enrichment scores for 1.)-6.) and the fold change enrichment score for 7.) with the Ward method and a correlation distance matrix.

CMAP

The CMAP resource uses the L1000 array, which measures 978 landmark transcripts, which can be used to infer most of the remaining transcriptome with high accuracy¹⁹. Here we used all available L1000 datasets for a human differentiated neuronal cell line. Despite only measuring a subset of landmark genes and inferring the rest, the CMAP resource has shown to be highly valuable for drug screens. We downloaded the aggregated level 5 L1000 Connectivity Map¹⁹ data from GSE92742 and did a pathway enrichment analysis with fgseaMultilevel from the fgsea R package⁸³ with nPermSimple=1000 and the conserved KEGG pathways between human and C. elegans for each of the samples in the level 5 dataset. To compare it to the NeuronAge trajectory, we first calculated the z-score of each gene that has at least some gene counts across the 128 neurons of the CeNGEN dataset. We correlated these z-score normalized genes with the biological age prediction from BitAge. The resulting Pearson correlation values for all genes were used for a pathway enrichment analysis with fgseaMultilevel and the conserved KEGG pathways between human and C. elegans. To identify whether any compound might revert the NeuronAge gene expression trajectory on the pathway level, we correlated the normalized enrichment scores (NES) of the NeuronAge KEGG pathway enrichment analysis with all NES that we calculated from the CMAP dataset. Next, we filtered for only compounds that were tested in the neuronal cell line (“NEU”). Compounds had to be tested at least twice, with all measurements resulting in correlations in the same direction. Additionally, we filtered for compounds that were measured at 24h and 6h and took only those compounds that showed a stronger correlation into the same direction at the 24h timepoint compared to the 6h timepoint. Lastly, we filtered out those compounds that had no information available at PubCHEM and used a correlation threshold of 0.25, respective -0.25.

Statistics

All data are presented as mean ± SD. Number of cohorts (ℕ), individuals (n), and technical replicates (N) is stated in the figures and their respective figure legends. The applied statistical tests are mentioned in the figure legends and the respective p-values are directly reported in the diagrams. All statistics were done two-sided if not stated otherwise. Independent t-tests were calculated with Python’s Scipy ⁸⁴ v1.5.1 stats.ttest_ind function. Kruskall-Wallis tests were calculated with Python’s Scipy v1.5.1 stats.kruskall() function. One-way ANOVA’s were calculated with Python’s pingouin ⁸⁵v0.3.6 anova function and the parameter ss_type=2. Cohen’s h ⁸⁶ as a measure of effect size was calculated by hand with Python’s numpy ⁸⁷ v1.18.5. Plots were generated with Python’s seaborn ⁸⁸ v.0.11.0, matplotlib ⁸⁹ v.3.3.0, or GraphPad Prism 9. Boxplots are shown with the center line depicting the median, the box limits the bottom, respective top quartiles, and the whiskers the 1.5x interquartile range. Scatterplots showing a linear regression model fit are shown with a 95% confidence interval.

Data availability

The unfiltered TPM counts and the Cell Marker list was downloaded from the CENGEN dataset were assessed at https://cengen.shinyapps.io/CengenApp/ .

The Calico dataset was downloaded from https://c.elegans.aging.atlas.research.calicolabs.com/data .

The neuron-specific information was assessed at https://www.wormatlas.org/ .

The gene length information was downloaded from https://wormbase.org/ .

The CMAP data were downloaded from GSE92742.

Data for the heatmap were downloaded either from the GEO database: GSE157025, GSE132040, GSE173254, GSE234667, GSE207152. From the Supplementary data from PMID: 30927700. Or the GTEx v8 database: https://gtexportal.org/home/downloads/adult-gtex/bulk_tissue_expression

Acknowledgement

We thank the Regional Computing Center of the University of Cologne (RRZK) for providing computing time on the DFG-funded High Performance Computing (HPC) system CHEOPS as well as support. Worm strains were provided by the National Bioresource Project (supported by The Ministry of Education, Culture, Sports, Science and Technology, Japan), the Caenorhabditis Genetics Center (funded by the NIH National Center for Research Resources, US), and the C. elegans Gene Knockout Project at the Oklahoma Medical Research Foundation (part of the International C. elegans Gene Knockout Consortium). We express our gratitude to Piali Sengupta for sharing PY6457 strain.

C.G. was supported by the Peter and Traudl Engelhorn Foundation. D.H.M. is member of the Cologne Graduate School of Ageing Research. B.S. acknowledges funding from the Deutsche Forschungsgemeinschaft (Reinhart Koselleck-Project 524088035, FOR 5504 project 496650118, SCHU 2494/3-1, SCHU 2494/7-1, SCHU 2494/10-1, SCHU 2494/11-1, SCHU 2494/15-1, CECAD EXC 2030 – 390661388, and GRK2407), the Deutsche Krebshilfe (70114555), Deutsche José Carreras Leukämie-Stiftung (DJCLS 04 R/2023), and the John Templeton Foundation Grant (61734).

Author contributions

Conceptualization: C.G., D.H.M., B.S.; Experimentation: C.G.; Bioinformatics: D.H.M; Data analysis: C.G., D.H.M.; Manuscript writing: C.G., D.H.M, B.S.; Visualization: C.G., D.H.M; Supervision: B.S.; Funding acquisition: B.S.

Declaration of Interests

All authors declare no competing interests.

Organization, W. H. World health statistics 2019: Monitoring health for the SDGs, sustainable development goals. (World Health Organization, 2019, 2019).
Oh, H. S.-H. et al. Organ aging signatures in the plasma proteome track health and disease. Nature 624, 164–172 (2023).
Schaum, N. et al. Ageing hallmarks exhibit organ-specific temporal signatures. Nature 583, 596–602 (2020).
Pálovics, R. et al. Molecular hallmarks of heterochronic parabiosis at single-cell resolution. Nature 603, 309–314 (2022).
Tabula Muris Consortium. A single-cell transcriptomic atlas characterizes ageing tissues in the mouse. Nature 583, 590–595 (2020).
Tian, Y. E. et al. Heterogeneous aging across multiple organ systems and prediction of chronic disease and mortality. Nat. Med. 29, 1221–1231 (2023).
Ahadi, S. et al. Personal aging markers and ageotypes revealed by deep longitudinal profiling. Nat. Med. 26, 83–90 (2020).
Van Hoesen, G. W., Augustinack, J. C., Dierking, J., Redman, S. J. & Thangavel, R. The parahippocampal gyrus in Alzheimer’s disease. Clinical and preclinical neuroanatomical correlates. Ann. N. Y. Acad. Sci. 911, 254–74 (2000).
Echávarri, C. et al. Atrophy in the parahippocampal gyrus as an early biomarker of Alzheimer’s disease. Brain Struct. Funct. 215, 265–71 (2011).
Braak, H. & Braak, E. Neurofibrillary changes confined to the entorhinal region and an abundance of cortical amyloid in cases of presenile and senile dementia. Acta Neuropathol. 80, 479–86 (1990).
Murphy, C. Olfactory and other sensory impairments in Alzheimer disease. Nat. Rev. Neurol. 15, 11–24 (2019).
Li, T., Le, W. & Jankovic, J. Linking the cerebellum to Parkinson disease: an update. Nat. Rev. Neurol. 19, 645–654 (2023).
Surmeier, D. J. Determinants of dopaminergic neuron loss in Parkinson’s disease. FEBS J. 285, 3657–3668 (2018).
White, J. G., Southgate, E., Thomson, J. N. & Brenner, S. The structure of the nervous system of the nematode Caenorhabditis elegans. Philos. Trans. R. Soc. Lond. B. Biol. Sci. 314, 1–340 (1986).
Lu, A. T. et al. Universal DNA methylation age across mammalian tissues. Nat. aging 3, 1144–1166 (2023).
Meyer, D. H. & Schumacher, B. BiT age: A transcriptome-based aging clock near the theoretical limit of accuracy. Aging Cell 20, e13320 (2021).
Tomusiak, A. et al. Development of a novel epigenetic clock resistant to changes in immune cell composition. bioRxiv [Preprint] (2023) doi:https://doi.org/10.1101/2023.03.01.530561.
Buckley, M. T. et al. Cell-type-specific aging clocks to quantify aging and rejuvenation in neurogenic regions of the brain. Nat. aging 3, 121–137 (2023).
Subramanian, A. et al. A Next Generation Connectivity Map: L1000 Platform and the First 1,000,000 Profiles. Cell 171, 1437-1452.e17 (2017).
Taylor, S. R. et al. Molecular topography of an entire nervous system. Cell 184, 4329-4347.e23 (2021).
Roux, A. E. et al. Individual cell types in C. elegans age differently and activate distinct cell-protective responses. Cell Rep. 42, 112902 (2023).
Schumacher, B., Meyer, D. & Meyer, D. H. Accurate aging clocks based on accumulating stochastic variation. Research Square (2023).
Stoeger, T. et al. Aging is associated with a systemic length-associated transcriptome imbalance. Nat. aging 2, 1191–1206 (2022).
Ibañez-Solé, O., Barrio, I. & Izeta, A. Age or lifestyle-induced accumulation of genotoxicity is associated with a length-dependent decrease in gene expression. iScience 26, 106368 (2023).
Gyenis, A. et al. Genome-wide RNA polymerase stalling shapes the transcriptome during aging. Nat. Genet. 55, 268–279 (2023).
Gallrein, C. et al. Novel amyloid-beta pathology C. elegans model reveals distinct neurons as seeds of pathogenicity. Prog. Neurobiol. 198, 101907 (2021).
Cook, S. J. et al. Whole-animal connectomes of both Caenorhabditis elegans sexes. Nature 571, 63–71 (2019).
Inglis, P. N., Ou, G., Leroux, M. R. & Scholey, J. M. The sensory cilia of Caenorhabditis elegans. WormBook 8, 1–22 (2007).
Altun, Z. F. Neurotransmitter Receptors in C. elegans. WormAtlas (2011) doi:10.3908/wormatlas.5.202.
Kelmer Sacramento, E. et al. Reduced proteasome activity in the aging brain results in ribosome stoichiometry loss and aggregation. Mol. Syst. Biol. 16, e9596 (2020).
Ximerakis, M. et al. Single-cell transcriptomic profiling of the aging mouse brain. Nat. Neurosci. 22, 1696–1708 (2019).
Buchwalter, A. & Hetzer, M. W. Nucleolar expansion and elevated protein translation in premature aging. Nat. Commun. 8, 328 (2017).
GTEx Consortium. Human genomics. The Genotype-Tissue Expression (GTEx) pilot analysis: multitissue gene regulation in humans. Science 348, 648–60 (2015).
Fitz, N. F. et al. Extracellular Vesicles in Young Serum Contribute to the Restoration of Age-Related Brain Transcriptomes and Cognition in Old Mice. Int. J. Mol. Sci. 24, 12550 (2023).
Schroer, A. B. et al. Platelet factors attenuate inflammation and rescue cognition in ageing. Nature 620, 1071–1079 (2023).
Berchtold, N. C. et al. Hippocampal gene expression patterns linked to late-life physical activity oppose age and AD-related transcriptional decline. Neurobiol. Aging 78, 142–154 (2019).
SenGupta, T. et al. Krill oil protects dopaminergic neurons from age-related degeneration through temporal transcriptome rewiring and suppression of several hallmarks of aging. Aging (Albany. NY). 14, 8661–8687 (2022).
Bhat, R. et al. Structural insights and biological effects of glycogen synthase kinase 3-specific inhibitor AR-A014418. J. Biol. Chem. 278, 45937–45 (2003).
Zhang, F. et al. Fluoxetine protects neurons against microglial activation-mediated neurotoxicity. Parkinsonism Relat. Disord. 18 Suppl 1, S213-7 (2012).
Chung, E. S. et al. Fluoxetine prevents LPS-induced degeneration of nigral dopaminergic neurons by inhibiting microglia-mediated oxidative stress. Brain Res. 1363, 143–50 (2010).
Li, I.-H. et al. Study on the neuroprotective effect of fluoxetine against MDMA-induced neurotoxicity on the serotonin transporter in rat brain using micro-PET. Neuroimage 49, 1259–70 (2010).
Ivraghi, M. S. et al. Neuroprotective effects of gemfibrozil in neurological disorders: Focus on inflammation and molecular mechanisms. CNS Neurosci. Ther. 30, e14473 (2024).
Liu, Z. et al. Inhibitors of LRRK2 kinase attenuate neurodegeneration and Parkinson-like phenotypes in Caenorhabditis elegans and Drosophila Parkinson’s disease models. Hum. Mol. Genet. 20, 3933–42 (2011).
Rzemieniec, J., Wnuk, A., Lasoń, W., Bilecki, W. & Kajta, M. The neuroprotective action of 3,3’-diindolylmethane against ischemia involves an inhibition of apoptosis and autophagy that depends on HDAC and AhR/CYP1A1 but not ERα/CYP19A1 signaling. Apoptosis 24, 435–452 (2019).
Lee, B. D. et al. 3,3’-Diindolylmethane Promotes BDNF and Antioxidant Enzyme Formation via TrkB/Akt Pathway Activation for Neuroprotection against Oxidative Stress-Induced Apoptosis in Hippocampal Neuronal Cells. Antioxidants (Basel, Switzerland) 9, (2019).
Normando, E. M. et al. The retina as an early biomarker of neurodegeneration in a rotenone-induced model of Parkinson’s disease: evidence for a neuroprotective effect of rosiglitazone in the eye and brain. Acta Neuropathol. Commun. 4, 86 (2016).
Zhao, Z. et al. Rosiglitazone Exerts an Anti-depressive Effect in Unpredictable Chronic Mild-Stress-Induced Depressive Mice by Maintaining Essential Neuron Autophagy and Inhibiting Excessive Astrocytic Apoptosis. Front. Mol. Neurosci. 10, 293 (2017).
Mishra, J., Chaudhary, T. & Kumar, A. Rosiglitazone synergizes the neuroprotective effects of valproic acid against quinolinic acid-induced neurotoxicity in rats: targeting PPARγ and HDAC pathways. Neurotox. Res. 26, 130–51 (2014).
Yang, S. et al. Protective effects of p38 MAPK inhibitor SB202190 against hippocampal apoptosis and spatial learning and memory deficits in a rat model of vascular dementia. Biomed Res. Int. 2013, 215798 (2013).
Rao, M. S. & Abd-El-Basset, E. M. dBcAMP Rescues the Neurons From Degeneration in Kainic Acid-Injured Hippocampus, Enhances Neurogenesis, Learning, and Memory. Front. Behav. Neurosci. 14, 18 (2020).
Abd-El-Basset, E. M. & Rao, M. S. Dibutyryl Cyclic Adenosine Monophosphate Rescues the Neurons From Degeneration in Stab Wound and Excitotoxic Injury Models. Front. Neurosci. 12, 546 (2018).
Shohami, E. et al. The Ras inhibitor S-trans, trans-farnesylthiosalicylic acid exerts long-lasting neuroprotection in a mouse closed head injury model. J. Cereb. Blood Flow Metab. 23, 728–38 (2003).
Apud, J. A. et al. Tolcapone improves cognition and cortical information processing in normal human subjects. Neuropsychopharmacology 32, 1011–20 (2007).
Tran, N. K. C. et al. Ginsenoside Re blocks Bay k-8644-induced neurotoxicity via attenuating mitochondrial dysfunction and PKCδ activation in the hippocampus of mice: Involvement of antioxidant potential. Food Chem. Toxicol. 178, 113869 (2023).
Liao, R. et al. Amiodarone-Induced Retinal Neuronal Cell Apoptosis Attenuated by IGF-1 via Counter Regulation of the PI3k/Akt/FoxO3a Pathway. Mol. Neurobiol. 54, 6931–6943 (2017).
Algharably, E. A. el-H., Di Consiglio, E., Testai, E., Kreutz, R. & Gundert-Remy, U. Prediction of the dose range for adverse neurological effects of amiodarone in patients from an in vitro toxicity test by in vitro-in vivo extrapolation. Arch. Toxicol. 95, 1433–1442 (2021).
Zhang, S., Fujita, Y., Matsuzaki, R. & Yamashita, T. Class I histone deacetylase (HDAC) inhibitor CI-994 promotes functional recovery following spinal cord injury. Cell Death Dis. 9, 460 (2018).
Gräff, J. et al. Epigenetic priming of memory updating during reconsolidation to attenuate remote fear memories. Cell 156, 261–76 (2014).
Shao, L.-W. et al. Histone deacetylase HDA-1 modulates mitochondrial stress response and longevity. Nat. Commun. 11, 4639 (2020).
McIntyre, R. L., Daniels, E. G., Molenaars, M., Houtkooper, R. H. & Janssens, G. E. From molecular promise to preclinical results: HDAC inhibitors in the race for healthy aging drugs. EMBO Mol. Med. 11, e9854 (2019).
Zeyn, Y. et al. Histone deacetylase inhibitors modulate hormesis in leukemic cells with mutant FMS-like tyrosine kinase-3. Leukemia 37, 2319–2323 (2023).
Griñán-Ferré, C. et al. The pleiotropic neuroprotective effects of resveratrol in cognitive decline and Alzheimer’s disease pathology: From antioxidant to epigenetic therapy. Ageing Res. Rev. 67, 101271 (2021).
Calabrese, E. J., Mattson, M. P. & Calabrese, V. Resveratrol commonly displays hormesis: occurrence and biomedical significance. Hum. Exp. Toxicol. 29, 980–1015 (2010).
Srinivasulu, C., Ramgopal, M., Ramanjaneyulu, G., Anuradha, C. M. & Suresh Kumar, C. Syringic acid (SA) ‒ A Review of Its Occurrence, Biosynthesis, Pharmacological and Industrial Importance. Biomed. Pharmacother. 108, 547–557 (2018).
Helli, B. et al. The Protective Effects of Syringic Acid on Bisphenol A-Induced Neurotoxicity Possibly Through AMPK/PGC-1α/Fndc5 and CREB/BDNF Signaling Pathways. Mol. Neurobiol. (2024) doi:10.1007/s12035-024-04048-0.
Ogut, E., Armagan, K. & Gül, Z. The role of syringic acid as a neuroprotective agent for neurodegenerative disorders and future expectations. Metab. Brain Dis. 37, 859–880 (2022).
Rothman, R. B., Baumann, M. H., Prisinzano, T. E. & Newman, A. H. Dopamine transport inhibitors based on GBR12909 and benztropine as potential medications to treat cocaine addiction. Biochem. Pharmacol. 75, 2–16 (2008).
Bergin, C. J. et al. The dopamine transporter antagonist vanoxerine inhibits G9a and suppresses cancer stem cell functions in colon tumors. Nat. cancer 5, 463–480 (2024).
Fatuzzo, I. et al. Neurons, Nose, and Neurodegenerative Diseases: Olfactory Function and Cognitive Impairment. Int. J. Mol. Sci. 24, (2023).
Takauji, Y. et al. Restriction of protein synthesis abolishes senescence features at cellular and organismal levels. Sci. Rep. 6, 18722 (2016).
Xie, X. et al. Quantification of Insoluble Protein Aggregation in Caenorhabditis elegans during Aging with a Novel Data-Independent Acquisition Workflow. J. Vis. Exp. 46, 248–256 (2020).
Janssens, G. E. et al. Transcriptomics-Based Screening Identifies Pharmacological Inhibition of Hsp90 as a Means to Defer Aging. Cell Rep. 27, 467-480.e6 (2019).
McIntyre, R. L. et al. Inhibition of the neuromuscular acetylcholine receptor with atracurium activates FOXO/DAF-16-induced longevity. Aging Cell 20, e13381 (2021).
Statzer, C. et al. Youthful and age-related matreotypes predict drugs promoting longevity. Aging Cell 20, e13441 (2021).
Gorelick, D. A., Gardner, E. L. & Xi, Z. X. Agents in development for the management of cocaine abuse. Drugs 64, 1547–1573 (2004).
Piccini, J. P. et al. Randomized, double-blind, placebo-controlled study to evaluate the safety and efficacy of a single oral dose of vanoxerine for the conversion of subjects with recent onset atrial fibrillation or flutter to normal sinus rhythm: RESTORE SR. Hear. Rhythm 13, 1777–1783 (2016).
Kohtz, A. S., Zhao, J. & Aston-Jones, G. Serotonin signaling in hippocampus during initial cocaine abstinence drives persistent drug seeking. J. Neurosci. 44, e1505212024 (2024).
Kosari-Nasab, M. et al. Serotonin 5-HT1A receptors modulate depression-related symptoms following mild traumatic brain injury in male adult mice. Metab. Brain Dis. 34, 575–582 (2019).
Kumar, L. & E Futschik, M. Mfuzz: a software package for soft clustering of microarray data. Bioinformation 2, 5–7 (2007).
Wu, T. et al. clusterProfiler 4.0: A universal enrichment tool for interpreting omics data. Innov. (Cambridge 2, 100141 (2021).
GTEx Consortium. The Genotype-Tissue Expression (GTEx) project. Nat. Genet. 45, 580–5 (2013).
Robinson, M. D., McCarthy, D. J. & Smyth, G. K. edgeR: a Bioconductor package for differential expression analysis of digital gene expression data. Bioinformatics 26, 139–40 (2010).
Korotkevich, G. et al. Fast gene set enrichment analysis. bioRxiv (2021) doi:10.1101/060012.
Virtanen, P. et al. SciPy 1.0: fundamental algorithms for scientific computing in Python. Nat. Methods 17, 261–272 (2020).
Vallat, R. Pingouin: statistics in Python. J. Open Source Softw. 3, 1026 (2018).
Cohen, J. Statistical Power Analysis for the Behavioral Sciences. (Elsevier, 1977). doi:10.1016/C2013-0-10517-X.
Harris, C. R. et al. Array programming with {NumPy}. Nature 585, 357–362 (2020).
Waskom, M. L. seaborn: statistical data visualization. J. Open Source Softw. 6, 3021 (2021).
Hunter, J. D. Matplotlib: A 2D graphics environment. Comput. Sci. \& Eng. 9, 90–95 (2007).

There is NO Competing Interest.

SuppFigure1.pdf
Supplementary Figure 1
SuppFigure2.pdf
Supplementary Figure 2
SuppFigure3.pdf
Supplementary Figure 3
SuppFigure4.pdf
Supplementary Figure 4
SupplementaryFigureLegendsUpdateDavid30042024.docx
Supplementary Figure Legends

Download PDF

Version 1

posted

You are reading this latest preprint version

Neuron-type specific aging-rate reveals age decelerating interventions preventing neurodegeneration

Status:

Version 1

Abstract

Figures

Introduction

Results

Neuron specific age predictions are associated with differential degeneration

Environmental exposition could be a discriminator for premature aging in neurons

Transcriptional Clustering identified reduced translation efficacy as potential driver of neuronal aging

Inhibition of translation alleviates neurodegeneration in fast aging neurons

Comparison with mammalian brain aging

Identification of drugs preserving neuronal function

Identification of neuro-protective molecule compounds

Discussion

Methods

Declarations

References

Additional Declarations

Supplementary Files

Status:

Version 1