Article Figures & Data
Figures
- Figure 1.
Distribution of clustered fronto-cortical cells based on mRNA profiles. The enrichment of genetic markers for the indicated clusters can be found in Extended Data Figure 1-1. A, Clusters expressing Vglut1 or Gad1/2 were identified as pyramidal (Pyr) neurons or interneurons, respectively. Cells that did not express these genes were considered as “non-neuronal” cells. Phenotypic profiles of cells in each category indicated in the figure. B, Frequency of DEGs by THC and WIN 55,212-2 across the different clusters. CGE, caudal ganglionic eminence; Gad1/2, glutamate decarboxylase; MGE, medial ganglionic eminence; Vglut1, type I vesicular glutamate transporter. The lists of DEGs with the statistics for THC and WIN can be found in Extended Data Figures 1-2 and 1-3, respectively.
- Figure 2.
THC and WIN 55,212-2 modified expression of respiratory genes in mouse frontal cortex. A, Genes commonly and distinctively modified by THC and/or WIN 55,212-2 are indicated. The expression changes induced by THC and WIN on respiratory genes can be found in Extended Data Figures 2-1 and 2-2, respectively. Comparison of genes differentially expressed by THC and WIN across the different respiratory chain complexes in (B) Vglut1+ pyramidal neurons, (C) non-neuronal cells, and (D) Gad1/2+ neurons. CI, CIII, and CIV, proton-mobilizing complexes; CV, ATP synthase, Ass. F, assembly factor; cluster numbers indicated at bottom of histograms. The cell-specific effects of THC and WIN on the expression of NDUFB8 subunits in microglia and layer 2/3 neurons of the prelimbic cortex can be found in Extended Data Figure 2-3. E, Mean cellular expression of Cnr1 and of genes coding for indicated downstream effectors in pyramidal (C1, C6, C8) and microglial (C7) clusters were obtained using the Average Expression function in Seurat. Inset shows expression ratios for βarr:Gαi/o subunits calculated using all Gαi/o versus βarr1/βarr2 effectors (top) or using cell-prevalent Gαi/o versus βarr effectors (bottom). Note: Expression profiles for Cnr1 and downstream effectors across all other clusters are provided in Extended Data Figure 2-5 for comparison. Only microglia displayed trace amounts of Cnr2. F, Concentration response curves for WIN and THC were obtained for the indicated effectors using a heterologous expression system (HEK293 cells). All responses were normalized to the endogenous ligand 2-AG which was the standard. Pharmacodynamic parameters can be found in Extended Data Figure 2-4. Data correspond to 3–4 independent experiments with four replicates per experiment. Curves generated by WIN and THC at each effector were compared by two-way ANOVA which showed an effect of drug as indicated in the figures, an effect of concentration (p ≤ 0.001), and an interaction (p ≤ 0.001), except for CB1-Gαi2 (p = 0.8700) and CB2-βarr2 (p = 0.4520), where no interaction was observed. C1–18 refer to clusters shown in Figure 1. C1, pyramidal neurons of layers 2/3; C6, pyramidal neurons of layers 2/3; C8, pyramidal neurons of layer 5a; C3, astrocytes; C7, microglia; C11, polydendrocytes; C10, medial ganglionic eminence (MGE) interneurons expressing parvalbumin; C12, MGE interneurons expressing somatostatin; C13, caudal ganglionic eminence (CGE) interneurons expressing vasoactive intestinal polypeptide; C16, CGE interneurons expressing neuropeptide Y; C18, neurons expressing glutamic acid decarboxylase (Gad).
- Figure 3.
THC induces transcriptomic and structural changes to glutamatergic synapses in the frontal cortex of adolescent male mice. A, Gene ontology enrichment results for 203 non-respiratory genes differentially expressed by THC. Statistics can be found in Extended Data Figure 3-1. Categories associated with synapse-related terms are highlighted. B, Cellular distribution of THC DEGs associated with synapse function. In bold, genes (mentioned in text) that are functionally or structurally associated to glutamatergic synapses. C, Effect of THC on spine density of apical and basal dendrites in pyramidal neurons of layer 2/3 of prelimbic cortex. Histograms show mean ± SEM of five mice/condition. Five to six neurons were analyzed per mouse; spines on primary and secondary apical branches and on 4–7 basal dendrites/neuron were considered. Results corresponding to Mann–Whitney comparisons are indicated in the figure.
- Figure 4.
THC decreases the complexity of dendritic arborization of L2/3 pyramidal neurons in prefrontal cortex of adolescent male mice. A, Characteristic morphology of GFP-labeled L2/3 PYR neurons in prelimbic cortex of adolescent Thy-1 male mice that received vehicle (CTL) or THC. Histograms display (B) total dendrite length, (C) number of branch points, and (D) branch depth. They correspond to mean ± SEM of five mice/condition. Four neurons displaying apical and basal dendrites were reconstituted per mouse. Results corresponding to Mann–Whitney comparisons are indicated in the figure. Intersections of (E) apical dendrites within a 360 μm radius from soma and (F) basal dendrites within 200 μm radius from soma from the soma of prelimbic L2/3 pyramidal neurons in CTL and THC-treated mice were compared by two-way ANOVA. Apical: showed and effect of distance from soma F(17,676) = 25.73, p < 0.001; an effect of treatment F(1,676) = 58.03, p < 0.001 and an interaction F(17,676) = 2.144, p = 0.005. Basal showed and effect of distance from soma F(9,676) = 74.46, p < 0.001; an effect of treatment F(1,380) = 187.4, p < 0.001 and an interaction F(9,380) = 12.65, p < 0.001. Post hoc comparisons with Sidak's test: *p < 0.01; **p < 0.001; ***p < 0.0001.
- Figure 5.
WIN 55,212-2 induces transcriptomic changes to glutamatergic synapses in the frontal cortex of adolescent male mice. A, Gene ontology enrichment results for 149 non-respiratory genes differentially expressed by WIN. Statistics can be found in Extended Data Figure 5-1. Categories associated with synapse-related terms are highlighted. B, Cellular distribution of WIN DEGs associated with synapse function. In bold, genes (mentioned in text) that are functionally or structurally associated to glutamatergic synapses.
- Figure 6.
Spatial coexpression between nine cell-specific gene panels and (A) 13 nominally significant THC-related genes that passed the two-stage consistency filtering (FDR-corrected for 13 × 9 observations) and (B) 18 nominally significant WIN-related DEG that passed our two-stage consistency filtering (FDR-corrected for 18 × 9 observations). A visual schematic of the analytic pipeline can be found in Extended Data Figure 6-1. The spatial correlation coefficients between group differences in cortical thickness and gene expression for THC-related consistent and inconsistent DEGs can be found in Extended Data Figures 6-2 and 6-3, respectively. Statistics for virtual histology for THC-related consistent and inconsistent DEGs can be found in Extended Data Figures 6-4 and 6-5, respectively. The spatial correlation coefficients between group differences in cortical thickness and gene expression for WIN-related consistent and inconsistent DEGs can be found in Extended Data Figures 6-6 and 6-7, respectively. Statistics for virtual histology for WIN-related consistent and inconsistent DEGs can be found in Extended Data Figures 6-8 and 6-9, respectively. Inter-regional coexpression of THC and WIN nominally significant genes can be found in Extended Data Figure 6-10.
- Figure 7.
Gene ontology enrichment results for the genes that were differentially expressed in mice and passed our two-stage filtering process in humans when exposed to a cannabinoid receptor agonist. Enrichment for biological process, cellular compartment, and molecular function for 37 THC-related consistent genes for Δ-9-tetrahydrocannabinol (A) and 33 consistent genes for WIN 55,212-2 (B). All statistics for the consistent THC-related genes, both pooled and separated into neuronal and non-neuronal categories, are available in Extended Data Figures 7-1 and 7-2, respectively. Statistics for GO analysis for inconsistent THC-related DEGs, both pooled and separated into neuronal and non-neuronal categories, are available in Extended Data Figures 7-3 and 7-4, respectively. Statistics for the consistent WIN-related genes, both pooled and separated into neuronal and non-neuronal categories, are available in Extended Data Figures 7-5 and 7-6, respectively. Statistics for GO analysis for inconsistent WIN-related DEGs, both pooled and separated into neuronal and non-neuronal categories, are available in Extended Data Figures 7-7 and 7-8, respectively.
Tables
- Table 1.
Demographics for Saguenay Youth Study adolescents as a function of cannabis experimentation before age 16
Males (n = 475) Cannabis exposure Not exposed (29.5%) Exposed (70.5%) Age of exposure NA 13.84 (1.62) Pregnancy duration ± 1 SD 39.22 (1.52) 38.89 (1.87) % born at term (>37 WG) 86.5% 80.0% Birth weight ± 1 SD (g) 3,481 (469.7) 3,407 (508.5) Tanner stage 3.18 (0.86) 3.71 (0.73) Household CAD income (% households) <$15,000 4.9% 10.8% $20,000–$30,000 5.8% 5.0% $30,000–$40,000 9.8% 6.5% $40,000–$50,000 10.4% 13.7% $50,000–$60,000 13.8% 12.2% $60,000–$70,000 11.7% 12.2% $70,000–$80,000 9.5% 10.8% $80,000–$85,000 9.2% 6.5% >$85,000 24.8% 22.3% Parental education (% individuals) Father Mother Father Mother Some primary school 0.7% 0.6% 0.8% 0.8% Primary school 0.0% 0.0% 0.0% 0.0% Some high school 13.2% 15.6% 19.2% 22.1% High school 37.6% 32.4% 43.1% 41.2% Some college 9.2% 9.3% 5.4% 9.2% College 20.8% 23.7% 16.2% 19.1% Some university 5.6% 5.0% 4.6% 2.3% Bachelor's degree 8.6% 9.7% 7.7% 4.6% Graduate degree 4.3% 3.7% 3.1% 0.8% The analytic pipeline performed using the human homologs can be found in Extended Data Table 1-1. The list of human homologs of DEGs can be found in Extended Data Table 1-2. The magnitude and the distribution of group differences in cortical thickness in male adolescents can be found in Extended Data Tables 1-3 and 1-4. SD, standard deviation; WG, weeks of gestation; CAD, Canadian Dollar.
Table 1-1
Analytic pipeline between inter-regional variations in cortical thickness as a function of cannabis exposure before the age of 16 in Saguenay Youth Study adolescent males and gene expression for THC-related and WIN-related differentially expressed genes in peri-adolescent male mice. Download Table 1-1, TIF file.
Table 1-2
Lists of human homolog ligand-related differentially expressed genes in mice. Download Table 1-2, XLSX file.
Table 1-3
Magnitude of the age-adjusted group difference between non-exposed adolescent males and cannabis-exposed adolescent males for 34 left cortical regions of the Desikan-Killiany atlas. Download Table 1-3, XLSX file.
Table 1-4
Distribution of group differences in thickness. The blue dashed lines represent the average cannabis-related thickness across 34 regions while the black full lines mark 0. Negative values indicate reduced cortical thickness in not-exposed adolescent males against cannabis-exposed adolescent males before adjusting (A) after adjusting (B) for age. Download Table 1-4, TIF file.
Figure 1-1
Enrichment of genetic markers for indicated cluster. Download Figure 1-1, XLSX file.
Figure 1-2
Differentially expressed genes by THC in mice for each Cluster. Download Figure 1-2, XLSX file.
Figure 1-3
Differentially expressed genes by WIN in mice for each Cluster. Download Figure 1-3, XLSX file.
Figure 2-1
Expression changes induced by THC on respiratory genes. The log of fold change is indicated by the grey tone (0: white; 1: dark grey). Download Figure 2-1, XLSX file.
Figure 2-2
Expression changes induced by WIN on respiratory genes. The log of fold change is indicated by the grey tone (0: white; 1: dark grey). Download Figure 2-2, XLSX file.
Figure 2-3
THC and WIN display cell specific effects on the expression of NDUFB8 subunits by microglia and Layer 2/3 neurons of the prelimbic cortex. An experimenter blind to treatment conditions counted in 3D the puncta corresponding to immunoreactivity (IR) of the NDUFB8 subunit I in Complex I (green, white arrows). (A) NDUFB8-IR within microglial silhouettes. Microglia were labelled with an antibody that recognizes transmembrane protein TMEM119 (red). (B) NDUFB8-IR within somatic volume of neurons revealed by NeuN (red). (C) Histogram shows the average number of puncta ± SEM counted in neurons and microglia of CTL and treated mice as indicated. The number of mice per condition is indicated at the bottom of each histogram. The number of microglia assessed per condition were CTL: 385 microglia, WIN: 209 microglia and THC: 260 microglia. The number of neurons assessed per condition were CTL: 1135 neurons, WIN: 300 neurons and THC: 425 neurons. Data were analyzed by Two-way ANOVA, which indicated and overall effect of cell type F(1, 34) = 18.24; p < 0.0001), no overall effect of treatment F(2, 34) = 2.044; p = 0.1451) and an interaction F(2, 34) = 5.037; p = 0.0121). Post-hoc comparisons using Tukey’s test showed a significant downregulation of microglial NDUFB8 by THC (p = 0.0279) but not WIN (p = 0.9996), and a trend for downregulation of neuronal NDUFB8 by WIN (p = 0.0857) but not THC (p = 0.9718). Download Figure 2-3, TIF file.
Figure 2-4
Pharmacodynamic parameters describing THC and WIN 55,212-2 responses at CB1 and CB2 receptors. Download Figure 2-4, XLSX file.
Figure 2-5
Mean cellular expression of Cnr1 and genes coding for indicated signaling effectors in all clusters identified. Note: Only microglia (C7) displayed trace amounts of Cnr2. Download Figure 2-5, TIF file.
Figure 3-1
Enrichment results for non-respiratory genes for differentially expressed genes by THC. Download Figure 3-1, XLSX file.
Figure 5-1
Enrichment results for non-respiratory genes for differentially expressed genes by WIN. Download Figure 5-1, XLSX file.
Figure 6-1
Virtual histology procedure. Reprinted from Navarri et al., 2023 (PMID: 36963244). Download Figure 6-1, TIF file.
Figure 6-2
Correlation between inter-regional profiles of group differences in cortical thickness (age-adjusted) and gene expression for 37 THC-related consistent differentially expressed. Download Figure 6-2, XLSX file.
Figure 6-3
Correlation between inter-regional profiles of group differences in cortical thickness (age-adjusted) and gene expression for 135 (out of the 149 genes) THC-related DEG that did not pass our two-stage consistency filtering and were used for gene ontology analysis. The following 14 genes were not in Human Allen Brain Database: BMERB1, COLGALT1, HSP90AA1, TRANK1, PHYHIPL, POLR2 M, LOC102724788, NTM, COX14, ZFHX3, ARPP21, GPCPD1, SCHIP1, CELF4. Download Figure 6-3, XLSX file.
Figure 6-4
Inter-regional co-expression of 13 consistent nominally significant THC-related DEG with nine cell-specific genes (corrected for 117 comparisons using a False Discovery Rate procedure). The group difference sign refers to the directionality of the spatial coexpression relationship between absolute age-adjusted cortical thickness (between cannabis-exposed adolescents males and non-exposed adolescent males from the Saguenay Youth Study cohort) and gene expression from the Allen Brain Atlas. Download Figure 6-4, XLSX file.
Figure 6-5
Inter-regional co-expression of 30 inconsistent nominally significant THC-related DEG with nine cell-specific genes (corrected for 270 comparisons using a False Discovery Rate procedure). The group difference sign refers to the directionality of the spatial coexpression relationship between absolute age-adjusted cortical thickness (between cannabis-exposed adolescents males and non-exposed adolescent males from the Saguenay Youth Study cohort) and gene expression from the Allen Brain Atlas. Download Figure 6-5, XLSX file.
Figure 6-6
Pearson correlation coefficients between inter-regional profiles of group differences in cortical thickness (age-adjusted) and gene expression for 33 WIN-related DEG that passed our two-stage consistency filtering and were used for gene ontology analysis. Download Figure 6-6, XLSX file.
Figure 6-7
Correlation between inter-regional profiles of group differences in cortical thickness (age-adjusted) and gene expression for 96 (out of the 108 genes) WIN-related DEG that passed our two-stage consistency filtering and were used for gene ontology analysis. The following 12 genes were not in Human Allen Brain Database: ATP5PF, UQCR10, TRAPPC13, PKM, PCDHA2, BMERB1, NAA15, ATP5PO, UQCC2, ATP5MF, SCHIP1, and ETNPPL. Download Figure 6-7, XLSX file.
Figure 6-8
Inter-regional co-expression of 18 consistent nominally significant WIN-related DEG with nine cell-specific genes (corrected for 162 comparisons using a False Discovery Rate procedure). The group difference sign refers to the directionality of the spatial coexpression relationship between absolute age-adjusted cortical thickness (between cannabis-exposed adolescents males and non-exposed adolescent males from the Saguenay Youth Study cohort) and gene expression from the Allen Brain Atlas. Download Figure 6-8, XLSX file.
Figure 6-9
Inter-regional co-expression of 21 inconsistent nominally significant WIN-related DEG with nine cell-specific genes (corrected for 162 comparisons using a False Discovery Rate procedure). The group difference sign refers to the directionality of the spatial coexpression relationship between absolute age-adjusted cortical thickness (between cannabis-exposed adolescents males and non-exposed adolescent males from the Saguenay Youth Study cohort) and gene expression from the Allen Brain Atlas. Download Figure 6-9, XLSX file.
Figure 6-10
(A) Inter-regional co-expression of 30 nominally significant THC-related DEGs that did not pass the two-stage consistency filtering with cell-specific genes. FDR-corrected for 270 comparisons (30 genes × 9 cell panels). (B) Inter-regional co-expression of 21 nominally significant WIN-related DEGs that did not pass the two-stage consistency filtering with cell-specific genes. FDR-corrected for 189 comparisons (21 genes × 9 cell panels). Download Figure 6-10, TIF file.
Figure 7-1
Gene ontology results for lists of consistent THC-related differentially expressed genes for biological process (BP), cellular compartments (CC) and molecular function (MF). Gene ratio indicates the enrichment for a given term and p-values underwent a Benjamini-Hochberg procedure. Download Figure 7-1, XLSX file.
Figure 7-2
Gene ontology results for lists of THC-related differentially expressed consistent genes for biological process (BP), cellular compartments (CC) and molecular function (MF). Gene ratio indicates the enrichment for a given term and p-values are corrected following a Benjamini-Hochberg procedure. Top row refers to consistent neuronal genes (n = 20), middle row to consistent non-neuronal genes (n = 21), and the bottom row to all consistent genes (n = 37). Download Figure 7-2, TIF file.
Figure 7-3
Gene ontology results for lists of inconsistent THC-related differentially expressed genes for biological process (BP), cellular compartments (CC) and molecular function (MF). Gene ratio indicates the enrichment for a given term and p-values underwent a Benjamini-Hochberg procedure. Download Figure 7-3, XLSX file.
Figure 7-4
Gene ontology results for lists of inconsistent THC-related differentially expressed genes for biological process (BP), cellular compartments (CC) and molecular function (MF). Gene ratio indicates the enrichment for a given term and p-values are corrected following a Benjamini-Hochberg procedure. Top row refers to consistent neuronal genes (n = 79), middle row to consistent non-neuronal genes (n = 79), and the bottom row to all consistent genes (n = 149). Download Figure 7-4, TIF file.
Figure 7-5
Gene ontology results for lists of consistent WIN-related differentially expressed genes for biological process (BP), cellular compartments (CC) and molecular function (MF). Gene ratio indicates the enrichment for a given term and p-values underwent a Benjamini-Hochberg procedure. Download Figure 7-5, XLSX file.
Figure 7-6
Gene ontology results for lists of WIN-related differentially expressed consistent genes for biological process (BP), cellular compartments (CC) and molecular function (MF). Gene ratio indicates the enrichment for a given term and p-values corrected following a Benjamini-Hochberg procedure. Top row refers to consistent neuronal genes (n = 31), middle row to consistent non-neuronal genes (n = 2), and the bottom row to all consistent genes (n = 33). Download Figure 7-6, TIF file.
Figure 7-7
Gene ontology results for lists of inconsistent WIN-related differentially expressed genes for biological process (BP), cellular compartments (CC) and molecular function (MF). Gene ratio indicates the enrichment for a given term and p-values underwent a Benjamini-Hochberg procedure. Download Figure 7-7, XLSX file.
Figure 7-8
Gene ontology results for lists of WIN-related differentially expressed inconsistent genes for biological process (BP), cellular compartments (CC) and molecular function (MF). Gene ratio indicates the enrichment for a given term and p-values are corrected following a Benjamini-Hochberg procedure. Top row refers to consistent neuronal genes (n = 95), middle row to consistent non-neuronal genes (n = 16), and the bottom row to all consistent genes (n = 108). Biological process and molecular function analysis could not be computed for the inconsistent non-neuronal WIN-related DEG given the gene list. Download Figure 7-8, TIF file.
