The dominant route of delivery of cannabis is via inhalation of combustion smoke, resulting in exposure to phytocannabinoids and activation of the endocannabinoid system [2–4]. The airway epithelium represents the first line of defence in the human lung against inhaled insults, including cannabis smoke. To better understand how the epithelium is able to respond to cannabis smoke exposure in the context of the endocannabinoid system, we performed a characterisation study using bioinformatic and complementary protein analysis approaches. We provide data supporting expression of CB₁, CB₂, and TRPV1 at the protein level in the human airway epithelium in situ and in vitro. We support our observations of protein expression by demonstrating that CNR1 and CNR2 transcripts are present in human airway epithelial cells using in situ hybridisation methods. The demonstration that these receptors were present warranted an exploration into the endocannabinoid system downstream of the receptors. Using 1090 unique patient samples of airway epithelial cells curated from publicly available datasets, we demonstrated in healthy subjects that the gene expression levels of the endocannabinoid system show minor differences between females relative to males. In contrast, we demonstrated with samples from people with asthma or individuals with COPD that disease status appears to be a strong driver of endocannabinoid system gene expression, with no interaction with sex. Lastly, we validated the bioinformatic approach by demonstrating that TRPV1, a top candidate upregulated at the gene level in our studies, is also upregulated at the protein level in people with asthma. Collectively, our results confirm the expression of the endocannabinoid system in human airway epithelial cells and that disease status impacts expression to a greater extent than sex, which may have functional consequences in distinct populations of cannabis consumers.

The legalisation of cannabis in multiple jurisdictions on a global scale reduces barriers for individuals to consume cannabis for either medicinal or recreational purposes. The dominant route of cannabis delivery is through inhalation of smoke from plant combustion [2–4]. Inhaled cannabis smoke travels through the upper and lower airways, with airway epithelial cells being a major site of first contact. We have demonstrated with an in vitro model of airway epithelial cell culture that cannabis smoke induces a concentration-dependent reduction in airway epithelial cell viability, barrier function, while promoting proinflammatory cytokine secretion [31, 32]. Complementary profiling of human epithelial cells isolated via bronchial brushings has demonstrated cannabis consumption-dependent elevations in TLR5, TLR6, and TLR9 gene expression [105]. A limitation of these studies is the lack of mechanistic interrogation into the role that the endocannabinoid system contributed to the observed functional consequences of cannabis smoke. To begin to implicate the endocannabinoid system in epithelial cell functions, an in vitro approach has used direct cannabinoid administration to airway epithelial cells independent of combustion, showing functional consequences with altered barrier function mediated by mechanisms dependent and independent of cannabinoid receptors [35, 36]. Our demonstration that multiple endocannabinoid system components are expressed at the gene level in airway epithelial cells with CB₁, CB₂, and TRPV1 confirmed at the protein level, is consistent with this system have the potential to play a role in mediating cannabis smoke-induced effects. Collectively, our data and existing literature suggest that the dominant form of cannabis consumption, cannabis smoke inhalation, is able to induce functional responses in airway epithelial cells that may be influenced by the endocannabinoid system.

Cannabis use patterns have been reported to differ among males and females, with females consuming less in overall quantity and frequency [2–4]. The endocannabinoid system is in turn regulated by sex hormones with diverse interactions at levels of receptors, enzymes, and signalling molecules [106, 107]. In the context of lung health and disease, sex-dependent lung physiology is observed as well as asthma and COPD disease incidence and progression [102, 103]. In light of the potential for interactions between sex, user practices, and the endocannabinoid system, we performed a bioinformatic analysis that stratified for male and female sex status. In healthy individuals, we observe that only 7 of 32 of our endocannabinoid gene signature candidates were differentially regulated between males versus females (2 upregulated, and 5 downregulated). In our analysis, the most significant differentially expressed gene between sexes was ABHD6 (increased in female samples). Increased female expression of ABHD6 has been previously reported in immune cells, with only female cells showing oestrogen- or progesterone-dependent induction of ABHD6 gene levels [108]. Our observations of minor sex-dependent effects on the expression of the 32-gene endocannabinoid signature contrast with sex differences in the expression of CB₁ and CB₂ protein expression in heart tissue, where CB₁ receptors are more highly expressed in females and CB₂ receptors more highly expressed in males [98]. Our observations of limited impact of sex on the 32-gene endocannabinoid signature expression in healthy subjects were conserved in people with asthma and individuals with COPD. In contrast to sex status, disease status drove a large effect for differential expression of the 32-gene endocannabinoid signature, with the majority of genes dysregulated (21 of 32 genes in people with asthma and 26 of 32 genes in those with COPD). Collectively, these findings suggest that asthma or COPD status impacts expression of the endocannabinoid system, whereas sex status plays a comparatively smaller role.

The potential disease-specific response to cannabis exposure as a result of dysregulated endocannabinoid signalling is relevant as cannabinoids have and are being pursued for bronchodilatory and immunomodulatory properties [7, 8, 104, 109]. Furthermore, in populations of people with asthma and those with COPD, cannabis consumption is not avoided and is shown to interact with disease progression. The reported benefits of cannabis exposure on lung function in people with asthma may be selective for this population based on endocannabinoid signalling. Indeed, if TRPV1 is a dominant receptor for responses downstream of inhaled cannabis, the signalling mediated in airway epithelial cells from people with asthma may be augmented relative to healthy controls. Our observation of elevated TRPV1 in airway epithelial cells from people with asthma is consistent with a previous report demonstrating correlation with asthma severity [110]. A limitation of our study is that we did not explore other cell types for TRPV1, which is also expressed in airway smooth muscle cells and can modulate smooth muscle contraction [111]. In contrast, the lack of observed benefit on lung function and association with advanced pathology in COPD subjects may be a result of a distinct expression profile of the endocannabinoid system [10, 11, 104]. Our results and those in the literature suggest that a universal response to inhaled cannabis by healthy subjects and those with asthma or COPD should not be assumed and cautions translation of safety and efficacy studies performed in healthy individuals to those with underlying asthma or COPD.

Our study is heavily focused on using deposited datasets generated by microarray gene expression technology. Microarrays are low cost, cover large numbers of transcripts, and benefit from a standardised format for analysis and public deposition of data. As a result of these benefits, gene expression data for large, independent cohorts are available for curation and data-mining purposes if appropriate measures are taken to normalise and integrate datasets from diverse user groups. Despite benefits of curated microarray datasets, there exist limitations of our approach. Specifically, microarray technology is susceptible to a poor signal-to-noise ratio, making transcripts expressed at low levels difficult to detect, suggesting that we may be under-estimating the signal or introducing high variance of these transcripts. To address this limitation, the SCAN method deployed functions to reduce technical and across-sample variation and to increase the signal-to-noise ratio while maintaining the ability to detect differentially expressed genes [91]. As in all analyses, the sample size will dictate the statistical power, which can only be as large as the available datasets. In our case, the sample sizes for females and males in our asthma and COPD datasets were smaller than our healthy cohorts, which may have limited our ability to detect any sex-specific effects on endocannabinoid gene expression. Finally, in order to compare data across multiple experiments, the microarray chip technology used needs to be considered with additional normalisation methods required for data generated on different platforms. Different microarray technologies will have unique probe compositions and quality control protocols which can lead to systematic biases between experiments, though these batch-specific effects can be addressed using batch correction techniques and normalisation [112]. Despite these highlighted limitations, the curation of multiple microarray gene expression datasets from diverse cohorts of study subjects can be aligned with normalisation methods to minimise batch and cross-platform effects and maximising sample sizes to detect differential gene expression patterns.

The advances of our study are accompanied by several limitations that need to be discussed. All analyses performed used human lung samples or datasets that did not contain information on cannabis consumption practices of the study subjects or body mass index, and we did not stratify by age. Furthermore, we broadly categorised people with asthma and COPD without stratifying by disease severity. It is possible that cannabis consumption impacts the expression of endocannabinoid system components examined in our study, a possibility we are not able to interrogate. Related to this possibility, a key finding of our study is the elevation of TRPV1 in people with asthma, which may be regulated differently with and without cannabis exposure [59]. Future studies examining healthy subjects and those with asthma or COPD in the context of cannabis exposure will address this limitation. We also only focused on airway epithelial cells, a dominant aspect of the respiratory mucosal defence system in the lungs. Observations in airway epithelial cells may be cell-specific and not translate to other structural or immune cells. Lastly, our archived human lung tissue samples used for immunohistochemistry and in situ hybridisation were collected from noninvolved regions of the lung taken during resection in individuals undergoing surgery as part of lung cancer management. It is possible that the underlying status of lung cancer impacts the expression profiling of the endocannabinoid system. Our analysis of nonlung cancer microarray datasets and immunoblot confirmation using healthy and diseased (asthma) samples attempted to corroborate observations made from the samples taken during lung resections.

In summary, we demonstrate that endocannabinoid system components are expressed in human airway epithelial cells at the gene level and CB₁, CB₂, and TRPV1 are expressed at the protein level. We demonstrate that gene expression patterns for a 32-gene endocannabinoid signature are differentially expressed between healthy individuals and those with asthma or COPD, whereas only minor differences are observed between sexes. We confirm that our bioinformatic approach for analysis of gene expression has the potential to reflect corresponding protein expression level changes as demonstrated by elevated TRPV1 protein expression in human airway epithelial cells from people with asthma relative to healthy controls. Our study lays a foundation with primary human lung samples from well-defined patient populations to justify exploring the functional consequences of endocannabinoid system signalling in human airway epithelial cells in both health and disease. The complete functions of the endocannabinoid system in airway epithelial cells remain to be defined.