Chronic pain, a significant global health crisis affecting over 1.2 billion people, presents a major medical challenge due to its complex mechanisms and the limitations of current treatments. Existing therapies often offer only partial relief, are accompanied by adverse effects, and carry the risk of addiction. These treatments primarily focus on pain signal modulation rather than addressing the underlying cellular and molecular causes of chronic pain. A deeper understanding of the molecular changes driving chronic pain progression is therefore crucial for developing innovative, non-addictive treatment strategies. Pain sensation is initiated peripherally by nociceptors, whose cell bodies reside in dorsal root ganglia (DRG) and trigeminal ganglia (TG). These nociceptors are categorized into subtypes based on their neurochemical profiles (peptidergic, non-peptidergic, C-Low Threshold mechanoreceptor, and somatostatin-positive). Single-cell RNA sequencing (scRNA-seq) offers a powerful tool to study the intricate functional diversity of nociceptors in pain development and chronification. Building upon previous research and integrating diverse scRNA-seq and single-nuclei RNA-seq datasets, the researchers created the iPain atlas (comprising iPainDRG and iPainTG) to comprehensively map the molecular signatures of various chronic pain models, aiming to identify shared mechanisms across these models. The iPain atlases, accessible via the CELLXGENE browser, serve as a valuable resource for studying chronic pain.

The study builds upon existing research on single-cell transcriptomics of sensory neurons and chronic pain. Previous work by Usoskin et al. (2015) established a classification of sensory neuron types using large-scale scRNA-seq. Renthal et al. (2020) explored transcriptional reprogramming of sensory neurons after axonal injury. Other studies (Wang et al., 2021; Avraham et al., 2021; Zhang et al., 2022; Sharma et al., 2020; Parpaite et al., 2021; Yang et al., 2022; Jia et al., 2022; Nguyen et al., 2017; Liu et al., 2022) provided additional scRNA-seq data on DRG and TG neurons in various pain models. The iPain atlas integrates these datasets to provide a more comprehensive understanding of chronic pain development.

The iPain atlas was constructed by integrating in-house and publicly available single-cell and single-nucleus RNA sequencing data from various sources (detailed in Table 1). Data integration was performed using scVI and scANVI models, with imputation of missing features using MultiVI. The analysis focused on the nociceptive lineage in DRG, which includes several pain models with consistent time-points up to 28 days post-injury. UMAP visualization was used to identify pain states ('woPain' and 'wPain') and refine them into microstates (Reference, Moving, Pain, Recovery, Lasting Pain). Differential gene expression analysis identified driver genes responsible for the state transitions. Cell-cell communication analysis, employing tensor decomposition, was conducted to elucidate interactions between nociceptors, immune cells, and satellite glial cells. Cellular senescence was assessed using the SenMayo gene set, identifying SASP-positive cells. In vivo validation was performed using SA-β-galactosidase staining in DRG after nerve injury. Human DRG data (from patients with chronic pain and diabetic neuropathy) were analyzed using the SenMayo score and bulk-to-single-cell deconvolution. The therapeutic potential of senolytics (Navitoclax, Venetoclax, PROTAC Bcl-XL degrader) was evaluated in mice with CCI-induced chronic pain, assessing mechanical and thermal sensitivity and general health parameters. Behavioral tests (von Frey, thermal paw withdrawal, pinprick, open field, elevated plus maze, beam walk, rotarod) were used to assess pain and motor function.

The iPain atlas revealed a dynamic progression of nociceptor states following injury. Sox11 was identified as a key transcription factor driving the transition to and persistence of the 'Lasting Pain' state. Cell-cell communication analysis implicated the p53 pathway in this process. Importantly, a senescence phenotype, characterized by high expression of SASP genes, was consistently observed in a subset of nociceptors across various rodent pain models and in human patients with chronic pain or diabetic neuropathy. The SenMayo z-score was significantly higher in all pain models and in human samples, irrespective of sex or age. Treatment with senolytic compounds effectively reduced mechanical hypersensitivity in mice with CCI-induced chronic pain without significant side effects or impact on other sensory modalities or general health, and this was associated with the removal of senescent cells.

This study demonstrates the development of a senescence phenotype in nociceptors as a key driver of chronic pain. The consistent observation of this phenomenon across multiple pain models and in human samples indicates its significance as a potential therapeutic target. The success of senolytic treatment in mice suggests that selectively removing senescent nociceptors could provide a novel approach to chronic pain management. While previous research has linked senescence to chronic pain, this study identifies it as an early and sex-independent mechanism, distinct from previously reported findings primarily focused on the spinal cord. The iPain atlas provides a valuable resource for future research into the mechanisms of chronic pain and for identifying additional therapeutic targets.

The iPain single-cell transcriptomic atlas provides a comprehensive resource for understanding chronic pain mechanisms. This study reveals a previously unknown role for nociceptor senescence in driving chronic pain, offering a promising new therapeutic target. Senolytic treatment effectively reduced pain in mice, demonstrating translational potential for human patients. Further research should focus on validating these findings in larger human cohorts and optimizing senolytic therapies for clinical use. The iPain atlas itself can be instrumental in this process, facilitating further exploration of chronic pain mechanisms.

The study primarily focused on rodent models and used deconvolution methods for human data analysis, potentially limiting the precision of cellular-level interpretation. Further, the human data analysis focused on DRG samples and there is a current lack of available TG samples or datasets from human pain sufferers. Although the study demonstrated positive results with senolytics, further investigation is needed to optimize the treatment strategy, including dosage and duration, for clinical applications. The relatively small number of animals in the senolytic treatment experiments might impact the overall statistical power.