1. Overview

Abstract:

The goal of this project is to determine the cell types and circuits that are active in the nucleus accumbens during opioid taking and seeking, and how these cells and circuits adapt during chronic opioid consumption. These findings will provide a more comprehensive understanding of opioid-induced changes in reward circuitry and inform the development of targeted interventions to combat opioid addiction and its long-term consequences.

Overdose deaths are on the rise, with 107,941 reported overdose deaths in the US in 2022 alone. Strikingly, 73,838 of those deaths were associated with synthetic opioids, most of which involved fentanyl. Drugs of abuse are thought to “hijack” synaptic plasticity mechanisms and drive molecular, synaptic, and circuit remodeling within the brain’s “reward circuit”- the ventral tegmental area (VTA)-nucleus accumbens (NAc) mesolimbic dopamine system. These neuroadaptive responses are hypothesized to underlie the development of addictive behaviors and to promote negative emotional and motivational states during withdrawal through strengthening anti-reward systems. Evidence suggests that there are dissociable circuits within the striatum and NAc that mediate distinct aspects of opioid reward, seeking, somatic withdrawal, and dependence and that each of these may be associated with their own forms of drug-induced plasticity mechanisms. Single-cell transcriptomic data has identified several novel classes of striatal medium spiny neurons (MSNs), however, limited studies on these cell types exist, especially in the context of opioid use. Additionally, single-cell sequencing is unable to capture key features of these newly identified subtypes, including topography and input-output organization. This proposal seeks to utilize the novel “Barcoded Anatomy Resolved by Sequencing” (BARseq) technique to address these technical challenges while elucidating the active cells and circuits during fentanyl taking and seeking. BARseq provides a means to trace thousands of single neurons, providing gene expression and topographical information of the cells in the injection site(s), and producing brain-wide projection patterns. In a diverse, heterogeneous, and anatomically indistinct region with multiple output projection pathways such as the NAc, this technique is essential to characterize the multiple distinct subclasses of MSNs and begin to understand their function in addiction. The studies proposed will provide novel mechanistic insights into the cells, circuits, and neuroadaptive mechanisms that mediate the progression of addiction, with greater precision than previously available, resulting in novel opportunities for developing cell-type specific treatments for addiction.

Confirmation of funding/facilities/ethics approvals:

The investigators have all of the appropriate approvals to perform the experiments listed in the proposal. Additionally, all investigators have access to all of the required equipment and resources necessary to complete the project.

Estimated timeline:

Months 1-4: Preliminary experiments to optimize experimental procedures including gene panel selection, optimization of viral titer, optimization of in situ scRNA sequencing methods to be performed in our home laboratory, optimization of tissue microdissection, library preparation, bulk RNA sequencing to be performed at the Stanford genomics core, and optimization of data processing and analytical pipelines.

Months 5-8: Preparation of samples: IV fentanyl self-administration, 4OHT injection, viral injection, brain collection, sectioning and microdissections.

Months 9-12: In situ scRNA sequencing of injection sites and library preparation and bulk sequencing of projection sites.

Months 13-24: Data processing and analysis.

2. Introduction

Medium spiny neurons (MSNs) are typically characterized by their expression of dopamine receptor (D)1 or D2, however, it has become increasingly clear that the standard D1/D2 phenotypes do not fully capture the role of MSNs in addiction, but instead distinct subclasses of MSNs mediate various aspects of addictive behavior.¹ Recently, single-cell transcriptomic data has identified several novel classes of MSNs that are not defined by the classical expression of DRs, including several non-canonical MOR+ MSNs^2–5. Heritability enrichment analysis identifies these cell types as being overrepresented in human psychiatric disease and importantly, these non-canonical MSN types show the most robust gene expression changes of all MSN types in humans with opioid use disorder^5,6. Little is known about the function of these non-canonical MSN subtypes under normal circumstances and importantly, they have yet to be explored in the context of opioid addiction. Although single-cell RNA sequencing (scRNAseq) suggests that these cell types are important for the physiological changes driving addictive processes, single-cell sequencing alone is unable to capture all the important features of these newly identified subtypes, including topography and input-output organization. Considerable transcriptional and spatial diversity of MSNs exists along the AP, ML, and DV axes of the striatum and NAc^4,7, making it difficult to identify where these cells might reside spatially. Furthermore, MSNs differ in their input-output organization, a feature which is not resolved through sequencing but might convey important differences between these subtypes as well as the circuits they are integrated in and the behavioral relevance of these cell types^8–10. Traditional methods of circuit tracing are laborious, time-consuming, notoriously low throughput, lack cellular specificity, and do not allow for accurate detection of complex projection patterns. Intersectional approaches utilizing logical gating exist, however these strategies can reduce labeling efficiency and can label off-target populations, resulting in confusing results^11–16. “Multiplexed Analysis of Projections by Sequencing” (MAPseq) and “Barcoded Anatomy Resolved by Sequencing” (BARseq) techniques address these technical challenges and provide a means to trace thousands of single neurons, providing single-cell gene expression data and topographical information of the cells in the injection site(s), and producing brain-wide projection patterns^17–20. My proposal is the first to examine the combined single-cell gene expression, spatial organization, and projectome of the striatum and NAc using BARseq, which is uniquely suited to perform this analysis. In addition, my proposed research will be the first to perform a holistic analysis of the cells and circuits active during fentanyl taking, seeking, and relapse and the subsequent gene expression and projection density changes that occur following chronic fentanyl consumption.

I hypothesize that adaptations in the activity, connectivity, and gene expression of distinct subclasses of striatal MSNs will underlie the transition to compulsive fentanyl use to promote drug-seeking and negative affective states during withdrawal. The funds received from this proposal will help collect preliminary data towards the first aim of my project, which is to determine the identity, gene expression, and projectome of medium spiny projection neurons active during fentanyl self-administration (SA), extinction, and reinstatement.

3. Methods

Experimental Methods:

Fentanyl SA: TRAP2 x LSL-Flp mice in which 4-hydroxytamoxifen (4OHT) inducible creERT promotes the permanent expression of Flp (Fig. 2) will be implanted with an indwelling jugular vein catheter (JVC) (Instech) in the right jugular vein connected to a mid-scapular vascular access button (Instech). Following the surgery, mice will be allowed to recover for 7-14 days and tested for patency using ketamine flush solution before beginning fentanyl SA. Mice will be trained to SA fentanyl (2.5μg/kg in 18μl over 4s) via lever responding paired with a cue light, lever retraction, and a 40s timeout period in operant chambers (MedAssociates). The experimental timeline and advancement criteria was adopted from Slosky et al. 2022²¹, and is shown in Figure 1. Figure 2 shows the timeline for induction of the TRAP system and injection of barcoded sindbis virus. Four experimental timepoints will be assayed as follows: (1) early fentanyl SA, (2) late fentanyl SA, (3) beginning of extinction training, and (4) during the cue-induced reinstatement session. Yoked controls self-administering saline will also be TRAPd at each time point to control for any neural activity that occurs due to context.

BARseq/MAPseq Sample Preparation: Animals will be injected with barcoded sindbis virus 24hrs after TRAPing to allow adequate time for Flp expression. Barcoded sindbis virus will be injected into the NAc and dorsal striatum (DS) (180nl/site) according to procedures described in Chen et al. 2019¹⁸. Brains will be harvested 24 hours later. The injection site will be cryosectioned at 20um while the projection target areas will be sectioned at 200um and micro-dissected on slides. Targets will include the VTA, GPe, EPN, SNr, SNc, VP, extended amygdala complex, LHb and LH. MAPSeq: MAPseq will be used to analyze the expression of barcodes in the projection sites according to procedures outlined in Kebschull et al. 2016¹⁷. Briefly, RNA from dissected target regions will be isolated, reverse transcribed, PCR amplified, and sent out for Illumina sequencing (Stanford genomics core). BARseq: Gene expression and somatic barcode expression at the injection sites will be sequenced by Illumina sequencing chemistry as described in Chen et al. 2019, and Sun et al. 2021^18–20,22 using imaging-based sequencing with images acquired on a Nikon confocal (in our lab). MAPseq/BARseq Data Processing: MAPseq data will be processed to generate a barcode matrix using the Zador lab MAPseq analysis pipeline available on github with procedures first described in Kebschull et al. 2016^17,23. For BARseq data, images of gene targets and barcodes will be analyzed using n2v (available on github²⁴), BarDensr²⁵, cellpose²⁶, cellranger (10x genomics), and seurat²⁷. Barcodes will be matched to the barcodes identified in projection patterns through MAPseq. BARseq/MAPseq Data Analysis: Matlab, python, and R will be used to perform dimensionality reduction (generation of UMAP plots, Louvain Community Detection²⁸), cell annotation and reference mapping (scArches²⁹), hierarchical clustering analysis (SigClust³⁰), spatial gene expression, and differential gene expression analysis (DEseq2³¹) with methods similar to those described in Huemos et al. 2023 and Luecken and Theis 2019^32,33. Matlab scripts used in Chen et al. 2019 and Sun et al. 2021^18,20 will be modified and used to perform downstream analysis. To (1) identify cell types, topographical organization, and projection patterns of activated MSNs during each phase of SA behavior, I will first determine whether specific populations of cells are activated or whether there is broad activation regardless of cell type identity and compare the activated cell types between experimental groups. Next, I will map the TRAPd cells along with their cell type identity onto slices to identify the spatial distribution of activated cells during each context and determine if there are interactions between the topographical organization and cell identity of the activated cells during each context. To (2) characterize changes in single-cell gene expression that occur in ensembles activated by each context, Above, I will have already identified the cell types activated during each context. I will analyze how gene expression in these ensembles changes over the course of the SA task. The results of the single-cell gene expression analysis will help to identify neuroadaptations that occur due to chronic remifentanil consumption and will inform mechanistic studies addressing how these neuroadaptations contribute to persistent remifentanil taking and seeking. To (3) measure differences in projection density to efferent targets due to opioid consumption, I will use standard MAPseq analysis techniques to perform unbiased clustering of barcodes in the projection sites to determine if any differences in the projection strength to each area exists. I will also perform these analyses in the populations identified in (1) to determine if projection strength or targeting in ensembles activated by any of the SA contexts changes over time.

Ethics and Data Management

Data sharing: The investigator acknowledges their willingness to share data and materials with other eligible investigators through academically established means. In order to further the advancement of research, data collected from experiments specified in this proposal will be published and freely available through PubMed Central. In addition, I will present my unpublished data at annual conferences such as Biological Psychiatry and the American College of Neuropsychopharmacology, facilitating dissemination of knowledge to the scientific community and to the public. Genomic Data Sharing Plan: Raw target area sequencing data will be deposited to Sequence Read Archive (SRA) and raw in situ sequencing images will be deposited at the Brain Image Library (BLI). Bioinformatic resources generated in this study will be deposited to publicly available databases. Processing scripts used for analysis will be deposited on Mendeley Data. All data generated by this study will be presented annually at the meetings and published in peer-reviewed journals and will be shared upon request with qualified academic investigators for non-commercial research.

4. Pilot Data

Adult C57BL/6J mice were implanted with jugular vein catheters and trained to self-administer remifentanil (0.1mg/kg) through operant lever responding (Fig. 3). Lever responding increased as the fixed ratio schedule increased, indicating that mice would exert more effort to earn the remifentanil reward as the SA task progressed. During extinction training, removal of drug-associated cues and the drug reward rapidly reduced lever responding (n=8 mice) (Fig. 3A). Presentation of drug-associated cues in the absence of a drug reward during a cue-induced reinstatement test significantly increased lever responding compared to the average of the last three days of extinction training (n= 6 mice) (Fig. 3B,C). The data demonstrate the investigator’s ability to catheterize mice and perform drug-SA studies.

5. Budget

Barcoded sindbis virus	$2,000
In Situ scRNA sequencing reagents	$15,000
Bulk RNA sequencing library preparation reagents and sequencing costs	$5,000
Computer upgrades/data storage devices	$1,500
Misc. Consumables	$1000
Animal costs	$500

6. References

1. Reiner, B. C. et al. A single-nucleus transcriptomic atlas of medium spiny neurons in the rat nucleus accumbens. Sci. Rep. 14, 18258 (2024).

2. Saunders, A. et al. Molecular Diversity and Specializations among the Cells of the Adult Mouse Brain. Cell 174, 1015-1030.e16 (2018).

3. Tran, M. N. et al. Single-nucleus transcriptome analysis reveals cell-type-specific molecular signatures across reward circuitry in the human brain. Neuron 109, 3088-3103.e5 (2021).

4. Chen, R. et al. Decoding molecular and cellular heterogeneity of mouse nucleus accumbens. Nat. Neurosci. 24, 1757–1771 (2021).

5. Gayden, J. et al. Integrative multi-dimensional characterization of striatal projection neuron heterogeneity in adult brain. BioRxiv (2023) doi:10.1101/2023.05.04.539488.

6. Phan, B. N. et al. Single nuclei transcriptomics in human and non-human primate striatum in opioid use disorder. Nat. Commun. 15, 878 (2024).

7. Roman, K. M. et al. Spiny projection neurons exhibit transcriptional signatures within subregions of the dorsal striatum. Cell Rep. 42, 113435 (2023).

8. Smith, J. B. et al. Genetic-Based Dissection Unveils the Inputs and Outputs of Striatal Patch and Matrix Compartments. Neuron 91, 1069–1084 (2016).

9. Crittenden, J. R. & Graybiel, A. M. Basal Ganglia disorders associated with imbalances in the striatal striosome and matrix compartments. Front. Neuroanat. 5, 59 (2011).

10. Graybiel, A. M. & Matsushima, A. Striosomes and matrisomes: scaffolds for dynamic coupling of volition and action. Annu. Rev. Neurosci. 46, 359–380 (2023).

11. Hughes, A. C. et al. A single-vector intersectional AAV strategy for interrogating cellular diversity and brain function. Nat. Neurosci. 27, 1400–1410 (2024).

12. Fenno, L. E. et al. Targeting cells with single vectors using multiple-feature Boolean logic. Nat. Methods 11, 763–772 (2014).

13. Jeong, M. et al. Viral vector-mediated transgene delivery with novel recombinase systems for targeting neuronal populations defined by multiple features. Neuron 112, 56-72.e4 (2024).

14. Fenno, L. E. et al. Comprehensive Dual- and Triple-Feature Intersectional Single-Vector Delivery of Diverse Functional Payloads to Cells of Behaving Mammals. Neuron 107, 836-853.e11 (2020).

15. Saleeba, C., Dempsey, B., Le, S., Goodchild, A. & McMullan, S. A student’s guide to neural circuit tracing. Front. Neurosci. 13, 897 (2019).

16. Harris, J. A. et al. Anatomical characterization of Cre driver mice for neural circuit mapping and manipulation. Front. Neural Circuits 8, 76 (2014).

17. Kebschull, J. M. et al. High-Throughput Mapping of Single-Neuron Projections by Sequencing of Barcoded RNA. Neuron 91, 975–987 (2016).

18. Chen, X. et al. High-Throughput Mapping of Long-Range Neuronal Projection Using In Situ Sequencing. Cell 179, 772-786.e19 (2019).

19. Chen, X., Sun, Y.-C., Church, G. M., Lee, J. H. & Zador, A. M. Efficient in situ barcode sequencing using padlock probe-based BaristaSeq. Nucleic Acids Res. 46, e22 (2018).

20. Sun, Y.-C. et al. Integrating barcoded neuroanatomy with spatial transcriptional profiling enables identification of gene correlates of projections. Nat. Neurosci. 24, 873–885 (2021).

21. Slosky, L. M. et al. Establishment of multi-stage intravenous self-administration paradigms in mice. Sci. Rep. 12, 21422 (2022).

22. Chen, X. BARseq - high-throughput cell typing with in situ sequencing v2. (2022) doi:10.17504/protocols.io .81wgbp4j3vpk/v2.

23. GitHub - ZadorLaboratory/mapseq-processing: Code/scripts for processing raw MAPseq data. https://github.com/ZadorLaboratory/mapseq-processing .

24. GitHub - juglab/n2v: This is the implementation of Noise2Void training. https://github.com/juglab/n2v .

25. Chen, S. et al. BARcode DEmixing through Non-negative Spatial Regression (BarDensr). PLoS Comput. Biol. 17, e1008256 (2021).

26. Stringer, C., Wang, T., Michaelos, M. & Pachitariu, M. Cellpose: a generalist algorithm for cellular segmentation. Nat. Methods 18, 100–106 (2021).

27. Hao, Y. et al. Dictionary learning for integrative, multimodal and scalable single-cell analysis. Nat. Biotechnol. 42, 293–304 (2024).

28. Blondel, V. D., Guillaume, J.-L., Lambiotte, R. & Lefebvre, E. Fast unfolding of communities in large networks. J. Stat. Mech. 2008, P10008 (2008).

29. Lotfollahi, M. et al. Mapping single-cell data to reference atlases by transfer learning. Nat. Biotechnol. 40, 121–130 (2022).

30. Liu, Y., Hayes, D. N., Nobel, A. & Marron, J. S. Statistical Significance of Clustering for High-Dimension, Low–Sample Size Data. J. Am. Stat. Assoc. 103, 1281–1293 (2008).

31. Love, M. I., Huber, W. & Anders, S. Moderated estimation of fold change and dispersion for RNA-seq data with DESeq2. Genome Biol. 15, 550 (2014).

32. Luecken, M. D. & Theis, F. J. Current best practices in single-cell RNA-seq analysis: a tutorial. Mol. Syst. Biol. 15, e8746 (2019).

33.Heumos, L. et al. Best practices for single-cell analysis across modalities. Nat. Rev. Genet.24, 550–572 (2023).

Funding

Changes in circuits of motivation following fentanyl addiction

Funding Progress

Nonprofit Recipient

Funders

Author Updates

Topics

DOI