Overview: Last year we established research methodology and protocols, built an infrastructure of hardware and software, formed collaborative arrangements, trained a team of scientists and support personnel, and performed over 150 sequencing runs on the Illumina HiSeq 2000. Over 15 billion bases of transcriptome sequence information have been generated, consequently, the main effort has been devoted to intensive analysis of the resulting data sets. We have sequenced the transcriptomes of physiologically or genetically labeled pain-sensing neurons sorted by FACS, neurons in dorsal spinal cord during peripheral inflammation and models of rheumatoid arthritis, inflamed peripheral tissue, and axotomized DRG and dorsal and ventral spinal cords. We have sampled multiple time points to follow the evolution and resolution of the intervention with enough samples at each point to permit statistical comparison. Because we sorted for certain neuronal populations we know which genes are in the pain-sensing neurons and which are in mainly non-pain-sensing neurons such as proprioceptive primary afferents. The ability to form incisive hypotheses regarding pain physiology is greatly advanced by this type of neuron-specific information and we now have quantitative information on all the genes that mediate DRG and sensory and motor spinal cord functions. TRPV1 Transcriptome: One important focus of the neuronal sorting experiments is a particular subpopulation of DRG neurons that express a multifunctional thermo- chemo- pH- and lipid-responsive ion channel called TRPV1. This ion channel is also gated by capsaicin, the active ingredient in hot pepper. Previous experiments demonstrated that the potent capsaicin analog resiniferatoxin (RTX) can control cancer pain in dogs and humans. Because of this crucial role, we want to know everything possible about TRPV1-expressing DRG neurons. We isolated TRPV1 neurons by genetic labeling or physiological activation and then performed deep sequencing of the mRNA content using next-gen RNA-Seq. The genetic method expressed a fluorescent marker allowing the TRPV1 DRG neurons to be isolated by FACS. A second strategy was to isolate by pharmacological activation. We loaded primary DRG neurons with a calcium sensitive dye, stimulated them with RTX and sorted the neurons that displayed RTX-induced increases in fluorescence. We also killed the cells either genetically or by microinjection of RTX. Our first paper (in preparation) outlines the transcriptome results from the genetically labeled TRPV1 neurons and ganglia in which the TRPV1 neurons had been deleted by expression of diphtheria toxin or microinjection of RTX. This has provided comprehensive new information on genes expressed by a clinically important population of nociceptive neurons. Analgesia transcriptome: One of the most interesting aspects of the transcriptome analyses is quantitative insight provided by next-gen RNA-Seq. We now know the quantitative relationships between the exact genes that mediate the actions of known analgesic drugs such as morphine, clonidine, lidocaine, ibuprofen, and gabapentin. It has not been clear which paralogs or subunits of drug binding receptors are expressed by different tissues in the pain pathway, yet this becomes clear when expression values for all the relevant genes are obtained quantitatively, at the same time, and with excellent reproducibility between animals and treatments. Additional analgesic targets: The transcriptome experiments also point to new targets for potential analgesic drug development. New targets that are highly differentially expressed in the TRPV1 population include an orphan GPCR, a lipid-binding GPCR, and a leukotriene receptor. In some cases prototype agonists or antagonists are available although their analgesic potential has not been explored. In another example, we observe that the Mu opioid receptor is expressed exclusively in the DRG and not in the dorsal spinal cord. This allows us to conclude that epidural or intrathecal opioid analgesia is solely mediated by a presynaptic action on DRG neurons. Amplification of ongoing studies: The RNA-Seq results also inform and amplify hypothesis-driven studies from our and other groups. In a collaborative work with NIAAA, we observe that certain lipids are TRPV1 agonists. Using the transcriptome databases, we extracted the quantitative expression data for all the genes involved in lipid transport, generation, degradation, and the cognate receptors for the relevant lipids from sequencing of skin, DRG and dorsal spinal cord. Differential expression levels therein provided insight into new enzymes that generate a particular, yet previously unrecognized, family of lipids that may be very important for TRPV1 activation. Canine ganglionic transcriptome: We have nearly completed the canine tissue collection for the cancer pain transcriptome study. The ganglion and spinal cord tissue have been obtained from controls and animals with osteosarcoma that were euthanized because of inadequate pain control or treated with resiniferatoxin and tissues obtained at autopsy. This study was undertaken to test for genes activated by nociceptive input from naturally occurring bone cancer and modulated by treatment. We can also make comparison to parallel studies in mouse, rat and human, although the exact models or cancer problems will be different. This is a unique set of data that will provide new insight into the transcriptomics of cancer pain in a species with a cancer pain problem that is very similar to the human. Cross-species comparison: Another project we are in the process of completing is a species comparison of the trigeminal ganglion transcriptome. The trigeminal ganglia and the medulla (medullary dorsal horn) are equivalent to DRG and spinal cord for the face and head. We have obtained and analyzed these two tissues in collaboration with Dr. Joel Kleinman formerly of NIMH. We have trigeminal ganglion sequence information from mouse, rat, and monkey and soon the dog. Comparisons show several remarkable species differences in degree of expression. This study is providing a new level of cross-species validation of potential therapeutic targets and mechanisms that aid in ascertaining the predictive capability of translational animal models. Summary: The data sets acquired over the past year provide unprecedented and extremely fine grained detail on gene expression in pain sensing circuits. This may seem complicated but the basic goal is to understand how we sense pain and how we may control it when necessary. There are a wide variety of painful stimuli that can be encountered in our environment and different neurons exist to sense these different types of pain signals. We are trying to figure out exactly what molecules the different types of pain sensing neurons make and how they work together to do their job. We will use this information to understand pain signaling and how to control it. Taken together these data will provide a transformative new resource for the pain research community and will allow a new much more precise assessment of experimental manipulations and verification of experimental results.
| Sapio, Matthew R; Neubert, John K; LaPaglia, Danielle M et al. (2018) Pain control through selective chemo-axotomy of centrally projecting TRPV1+ sensory neurons. J Clin Invest 128:1657-1670 |
