Time-resolved multi-omic analysis of paclitaxel exposure in human iPSC-derived sensory neurons unveils mechanisms of chemotherapy-induced peripheral neuropathy

The microtubule-stabilizing drug paclitaxel remains standard of care for various solid malignancies but frequently leads to chemotherapy-induced peripheral neuropathy (CIPN). CIPN is a leading cause for premature treatment termination and a significantly reduced quality of life in long-term cancer survivors. The molecular mechanisms of neuro-axonal degeneration, neuroinflammation and pain in patients treated with paclitaxel remain incompletely understood, and there are currently no predictive biomarkers or preventive treatments. We used human iPSC-derived sensory neurons exposed to paclitaxel to comprehensively model the pathophysiology of CIPN. Neurotoxicity was assessed over time using viability assays and sequential RNA sequencing as well as deep proteome and lipidomic analyses. We observed a time and dose-dependent decline of cell viability at clinically relevant paclitaxel doses. Sequential RNA sequencing defined JUN as an early immediate gene, followed by the overexpression of genes of the neuronal stress response (e.g., ARID5A, WEE1, DUSP16, GADD45A), neuronal injury and apoptotic pathways (e.g., ATF3, HRK, BBC3 [PUMA], BCL2L11 [BIM], CASP3), neuroinflammation and nociception (CALCB, MMP10, IL31RA, CYSLTR2, C3AR1, TNFRSF12A) and neuronal transduction (e.g., CAMK2A, STOML3, PIRT), while key enzymes of lipid biosynthesis were markedly downregulated (e.g., LSS, HMGCS1, HMGCR, DHCR24). Deep proteome analyses following 48 hours of exposure to 100nM paclitaxel revealed a strong correlation of differentially expressed RNA with proteins, and a marked degradation of essential axonal transport proteins such as kinesins, stathmins and scaffold proteins. Consistent with the downregulation of rate-limiting enzymes of lipid biosynthesis, lipidome analysis confirmed deregulation of neuronal lipid homeostasis. In summary, paclitaxel induces transcriptomic and proteomic signatures of the neuronal stress response, neuroinflammation, nociception and disturbed metabolism. These may explain, in part, the clinical phenotype of sensory loss, hypersensitivity and neuropathic pain frequently observed in patients suffering from CIPN, but constitute pharmacologically addressable targets. Overall design: 72 samples were included in RNA sequencing analyses from 12 different experimental units (plates): Seven time points from BIHi264-A (7 × 6-well plates) and 48-h samples from five cell lines (5 × 6-well plates). iPSC-DSN were maintained in geltrex coated 6-well plates at 10^6 cells/well, as described in the CDDis manuscript entitled above. After maturation for >60d, three wells were either treated with DMSO at 1/60.000 or paclitaxel at 100 nM, respectively. Investigated time points were 48h of paclitaxel incubation for all five cell lines (30 samples), and for BIHi264-A, the additional time points of 2h, 6h, 12h, 24h and 24h or 120h following removal of the drug after 48h incubation (s.c. 72h and 168h timepoints; in total, 42 samples). RNA was harvested with the Aurum™ Total RNA Mini Kit according to manufacturers' instructions. RNA sequencing was performed by Brooks Life Sciences Genewiz® with PolyA selection for RNA removal, 2x150 bp sequencing configuration and 20-30 million reads per sample. RNA Library Preparation and NovaSeq Sequencing by AzentaLifeSciences Genewiz: RNA samples were quantified using Qubit 4.0 Fluorometer (Life Technologies) and RNA integrity was checked with an RNA Kit on Agilent 5300 Fragment Analyzer (Agilent Technologies). RNA sequencing libraries were prepared using the NEBNext Ultra RNA Library Prep Kit for Illumina following the manufacturer's instructions (NEB, Ipswich). Briefly, mRNAs were first enriched with Oligo(dT) beads. Enriched mRNAs were fragmented for 15 minutes at 94 °C. First-strand and second strand cDNAs were subsequently synthesized. cDNA fragments were end repaired and adenylated at 3'ends, and universal adapters were ligated to cDNA fragments, followed by index addition and library enrichment by limited-cycle PCR. Sequencing libraries were validated using NGS Kit on the Agilent 5300 Fragment Analyzer (Agilent Technologies), and quantified by using Qubit 4.0 Fluorometer (Invitrogen). The sequencing libraries were multiplexed and loaded on the flowcell on the Illumina NovaSeq 6000 instrument according to manufacturer's instructions. The samples were sequenced using a 2x150 Pair-End (PE) configuration v1.5. Image analysis and basecalling were conducted by the NovaSeq Control Software v1.7 on the NovaSeq instrument. Raw sequence data (.bcl files) generated from Illumina NovaSeq was converted into fastq files and de-multiplexed using Illumina bcl2fastq program version 2.20. One mismatch was tolerated for index sequence identification. Computational Methods. RNA Seq reads were mapped to human genome (GRCh38.p7) with STAR-version 2.7.3a [100] using the default parameters. Reads were assigned to genes with featureCounts version 2.0.0 [101] with the following parameters: -t exon -g gene_id -s 0 -p, gene annotation - Gencode GRCh38/v25. The differential expression analysis was carried out with DESeq2-version 1.32.0 [102] using the default parameters. We kept genes that have at least 5 reads in at least 3 samples. Gene set enrichment analysis was carried out with R/tmod package version 0.50.07 using MSigDB gene sets [103]. For Reference Numbers: Please refer to CDDis manuscript CDDis-25-4526.