The studies involving human participants were reviewed and approved by the office of the Chief Coroner at Ontario Forensic Pathology Service (Chief Forensic Pathologist, and Chief Coroner). Written informed consent for participation was not required for this study in accordance with the national legislation and the institutional requirements.

All statistical analyses were performed using GraphPad Prism^® (version 8.0 for Microsoft Windows, Graph Pad Software, San Diego, CA, United States) with a significant difference defined as a p-value of 0.05 or less. All results for mRNA expression and protein quantification were expressed as mean ± standard deviation (SD) and mean ± standard error of the mean (SEM), respectively. An unpaired t-test was used to determine significant differences in mRNA expression between the SARS-CoV-2-infected and mock Vero E6 cells. The non-parametric two-tailed Mann-Whitney test was used to assess the differences in protein expression between COVID-19 and control postmortem human lung tissues for absolute quantitative IHC analysis.

With the SARS-CoV-2-associated inflammatory response in Vero E6 cells, we observed that at the mRNA level, CYP3A4 was upregulated while UGT1A1 was downregulated (Figure 2B). However, none of the investigated transporters were significantly dysregulated (Figure 2C and Supplementary Figure S3B). Several disease-drug interactions associated with SARS-CoV-2 infection have been suggested (Kumar and Trivedi, 2021). For example, CYP3A4 involved in the metabolism of dexamethasone, a corticosteroid drug often administered to hospitalized COVID-19 patients, is known to be dysregulated by inflammation (Kumar and Trivedi, 2021; Dunvald et al., 2022). Also, nirmatrelvir, an antiviral drug used to treat SARS-CoV-2 infection is metabolized by CYP3A4 and is currently coadministered with ritonavir (an inhibitor of CYP3A4) to improve its bioavailability (Nwabufo and Bendayan, 2022). Therefore, it is important to further investigate how SARS-CoV-2-associated inflammatory response will affect the PK profile of nirmatrelvir-ritonavir combination therapy for the treatment of COVID-19. Furthermore, human clinical data suggests that remdesivir is extensively metabolized by CYP2C8, CYP2D6, and CYP3A4 (Yang, 2020) and since studies have shown that these enzymes are dysregulated by inflammation (Dunvald et al., 2022), the PK profile of remdesivir and dexamethasone could be altered by the inflammatory state observed in COVID-19. This may partly explain the variable clinical outcomes observed in several clinical studies that investigated the safety and efficacy of remdesivir (Cascella et al., 2022). In general, remdesivir was primarily approved by the US FDA for administration to COVID-19 patients requiring hospitalization, and proper dosing regimen and treatment course need to be further investigated to achieve the full benefits of remdesivir for patients at different stages of COVID-19. Additionally, caution needs to be taken when administering drugs to COVID-19 patients especially the patient population with comorbidities to mitigate clinically relevant disease–drug interactions. For example, UGT1A1 is highly polymorphic and can impact irinotecan (a prodrug used for small cell lung cancer chemotherapy) metabolite related-toxicity (Bandyopadhyay et al., 2021). Given that UGT1A1 mRNA expression was significantly downregulated in SARS-CoV-2-infected Vero E6 cells and patients with lung cancer have a greater than 7-fold higher risk of SARS-CoV-2 infection (Rolfo et al., 2022), further investigation is required to determine the effect of prescribing UGT1A1 candidate drugs to COVID-19 patients, especially the UGT1A1 poor metabolizers which accounts for about 10% of North Americans (Dean, 2018).

Dunvald and coworkers recently summarized studies reporting altered mRNA expression of clinically relevant DMETs including CYP1A2, CYP2B6, CYP2C8, CYP2C9, CYP2C19, CYP2D6, CYP3A4, PGP, BCRP, and MRP2 by proinflammatory cytokines such as IL-6, IL-1β, TNF-α, and IFN-γ in a dose-dependent manner in primary cultures of human hepatocytes (Dunvald et al., 2022). For example, the mRNA expression of the hepatic efflux transporters P-gp, MRP2, and BCRP is typically low, with a maximum downregulation of 2-fold, in response to IL-6, IL-1β, TNF-α, or IFN-γ at doses of 1–100 ng/mL (Ramsden et al., 2015; Moreau et al., 2017; Ning et al., 2017; Dunvald et al., 2022). However, the concentrations of the investigated inflammatory markers in SARS-CoV-2 infected Vero E6 cells are probably more physiologically relevant compared to the artificial stimulation of inflammatory response by the administration of toxins such as lipopolysaccharides, or direct administration of proinflammatory cytokines such as IL-6 and IL-1β—both of which may result in supraphysiological responses. Furthermore, species (human and monkey) and organ (kidney and liver) differences in DMETs expression, may be responsible for the lack of dysregulation of the above-mentioned DMETs. For example, a previous study found that CYP2C18 mRNA expression is unaffected by cytokine administration due to limited hepatic expression (Aitken and Morgan, 2007).

We localized for the first time, the cellular expression of uptake transporters—ENT1 and ENT2 in human lung tissues by IHC. We found that both ENT1 and ENT2 were primarily localized in inflammatory cells of lung tissues from COVID-19 patients (Figure 4; Table 2), making them potentially liable to SARS-CoV-2-mediated increases in expression. Interestingly, ENT1 and ENT2 may be prone to downregulation with SARS-CoV-2 infection due to the presence of acute lung injury and hypoxia in some COVID-19 patients (Johnson, 2022). Indeed, previous studies reported a significant downregulation of ENT1 and ENT2 expression in lung epithelial and endothelial cells in both acute lung injury (Morote-Garcia et al., 2013) and hypoxia (Eltzschig et al., 2005). More so, increased extracellular adenosine levels due to acute lung injuries (Eckle et al., 2009) suggest a potential competitive inhibition of ENTs-mediated transport processes. This is clinically important because ENT1 and ENT2 are involved in the uptake of the COVID-19 drugs - remdesivir and molnupiravir, making this uptake pathway potentially liable to dysregulation with SARS-CoV-2 infection and may partly explain the limited and variable clinical efficacy observed for remdesivir (Johnson, 2022).

To the best of our knowledge, our study is the first to reveal the cellular localization of CYP1A2, CYP2C8, CYP2C9, and CYP2D6 by IHC in human lung tissues. Previous studies investigating the expression of DMEs in the human respiratory system, have mostly assessed expression at the mRNA level, used non-intact lung samples such as microsomes and bronchial specimens, or failed to uncover cellular localization (Hukkanen et al., 2002). The human lung tissue is highly heterogeneous, making microsomes, bronchial, and other non-intact lung specimens an inaccurate description of DMETs expression in different pulmonary cell types. For example, the mRNA expression of CYP2C8 was previously found in both bronchial and peripheral lung tissue samples (Macé et al., 1998), but its protein expression and cellular localization were not reported. An IHC study identified CYP2B6 in human Clara cells (Mori et al., 1996) whereas CYP2C19 and CYP3A4 proteins were detected in serous cells of bronchial glands (Yokose et al., 1999) but expression in other pulmonary cell types was not reported. Our study uncovered the pulmonary cellular localization of CYP2B6, CYP2C8, CYP2C19, and CYP3A4 (Figure 4; Table 2).

Efflux transporters such as P-gp and BCRP have previously been identified in human lung tissues (Cordon-Cardo et al., 1990; Van der Valk et al., 1990; Scheffer et al., 2002; Fetsch et al., 2006). In contrast to an earlier study that found BCRP expression only in bronchial epithelial cells and capillaries (Scheffer et al., 2002), later work demonstrated staining of alveolar pneumocytes and negligible staining of the bronchial epithelial cells (Fetsch et al., 2006). These findings are not in agreement with the results of this study. In our study, BCRP was primarily localized in the cytoplasm of endothelial cells in both COVID-19 and control tissues (Figure 4; Table 2). In previous reports, P-gp localization was shown on the luminal surface of bronchial and bronchiolar epithelial cells (Cordon-Cardo et al., 1990; Van der Valk et al., 1990). P-gp was stained in alveolar macrophages whereas staining of alveolar epithelial cells was dependent on the antibody employed (Van der Valk et al., 1990). On the contrary, our findings showed that P-gp is robustly localized in the alveolar epithelial cells (Figure 4; Table 2) of both COVID-19 and control tissues. MRP1 was initially found in the apical membrane of the cytoplasm of bronchial epithelial cells (Flens et al., 1996); however, two later studies confirmed MRP1 localization in the basolateral membrane (Bréchot et al., 1998; Scheffer et al., 2002) which is consistent with its localization in other tissues (Gustavsson et al., 2016). In our study, MRP1 expression in the bronchial epithelium is cytoplasmic (Figure 4; Table 2). Moreover, MRP1 was predominantly localized in the cytoplasm and nucleus of alveolar epithelial cells of both COVID-19 and control postmortem human lung tissues (Figure 4; Table 2). MRP2 positivity was previously found in the apical membrane of the bronchial and bronchiolar epithelial layers (Sandusky et al., 2002; Scheffer et al., 2002). We observed a robust localization in the alveolar epithelial cells of both COVID-19 and control tissues (Table 2). We also found MRP2 expression in bronchial epithelial cells in a higher number for controls (3/5) compared to the COVID-19 (1/10) human lung tissues; however, it was more circumferential not apical membranous compartmentalization (Figure 4; Table 2). Our study demonstrated that circumferential membrane compartmentalization in alveolar epithelial and endothelial cells was the major expression site for MRP2 (Figure 4; Table 2). This circumferential compartmentalization is indicative of potential bidirectional efflux transport processes mediated by MRP2 in human lung tissues compared to the anticipated unidirectional efflux transport.

Typically DMETs are localized in the cytoplasm and cell membrane, however, our study found that some DMETs were localized in atypical cellular compartments (Figure 4; Table 2). For example, our study detected MRP1 in the cytoplasm and nucleus. Although transporters are localized at the cell membrane, they are also present in cytoplasmic organelles such as the Golgi apparatus, rough endoplasmic reticulum, nuclear envelope, and mitochondria. For instance, a previous IHC study found MRP2 expression in both cytoplasm and cell membrane of tumor cells (Yamasaki et al., 2011), and it is anticipated that in the cytoplasm, MRP2 may not function as an efflux pump (Evers et al., 1998). Therefore, further investigation is required to determine whether differences in pulmonary cellular compartmentalization affect the structure and function of the implicated DMETs.

From the absolute quantitation, we observed significant differences in the expression of P-gp and MRP1 between the COVID-19 and control human lung tissues (Figure 6B). P-gp is involved in the transport of nirmatrelvir, remdesivir, and dexamethasone while MRP1 transports lopinavir and ritonavir (Nwabufo and Bendayan, 2022)—indicating a potential alteration in pulmonary drug PK/PD profile in COVID-19 patients. The lack of observable significant dysregulation in the protein expression of other DMETs in COVID-19 human lung tissues (Figures 5, 6) could be attributed to the potential variation in the stage of COVID-19 between cases, as well as comorbidities (Table 1), ongoing medications, and genetic polymorphisms in the expression of DMETs for both COVID-19 and control cases. These variables could not be further addressed due to the paucity of premortem clinical information. The small sample size, and other diseases present in the control samples may also have an impact on the expression of inflammatory markers, regulatory proteins, and DMETs, possibly resulting in a negligible dysregulation in their expression between COVID-19 and control tissues. Drug toxicity, cardiovascular diseases, and head trauma are the reported causes of death for the 5 control cases (Table 1), and these disorders can induce inflammatory response and dysregulate the expression of DMETs through regulatory proteins (Wu and Lin, 2019; Stanke-Labesque et al., 2020; Dunvald et al., 2022).

The congruency in the pulmonary cellular localization of SARS-CoV-2 infection, inflammatory markers, regulatory proteins, and DMETs, as well as the several COVID-19-related pulmonary pathologies of the COVID-19 cases suggest a potential dysregulation of pulmonary DMETs which may manifest in regulated clinical studies. We observed that SARS-CoV-2 infection and inflammatory response are distinctly localized to the cytoplasm of lymphocytes in COVID-19 compared to control human lung tissues (Figure 4; Table 2). The detection of SARS-CoV-2 viral protein in the control human lung tissues is indicative of the limited specificity of the antibody; however, clinical testing confirmed the presence of SARS-CoV-2 infection in the COVID-19 cases. Again, underlying diseases in the control cases are probably responsible for the observed inflammatory response. Interestingly, we observed a similar trend in the localization of xenosensing regulatory proteins; for example, the two master xenosensing regulatory proteins - PXR and CAR are both distinctly expressed in the cytoplasm of lymphocytes in COVID-19 human lung tissues but not in the control cases (Figure 4; Table 2). Also, we observed a similar trend in the cellular localization of some DMETs including CYP2C9, CYP2C19, CYP2D6, and MRP1 (Figure 4; Table 2). These observations suggest that SARS-CoV-2 infection could result in an inflammatory response that could activate xenosensing regulatory proteins, which could then dysregulate the expression of DMETs in lymphocytes. However, our study quantified global pulmonary protein expression for the investigated biomarkers to get a better representation of SARS-CoV-2-DMETs interactions. Moreover, it is practically challenging to quantify the investigated biomarkers in the lymphocytes of human lung tissues except when pulmonary lymphocyte isolates are used as specimens for the study.

Furthermore, we observed that the presence of inflammatory cells is a major driving force in the localization of DMETs between COVID-19 and control tissues (Figure 4; Table 2). In general, COVID-19 lung tissues had more inflammatory cells which also expressed DMETs compared to the control lung tissues. Notably, CYP3A4 and ENT2 were strongly localized in lymphocytes in COVID-19 compared to control human lung tissues (Figure 4; Table 2). This suggests a potential SARS-CoV-2-mediated increase in the expression of DMETs through the recruitment of inflammatory cells and may have implications for pulmonary drug PK/PD profiles. For example, CYP3A4 and ENT2 are involved in the disposition of remdesivir (Nwabufo and Bendayan, 2022) and may be susceptible to an altered pulmonary PK/PD profile in the context of SARS-CoV-2 infection. Additionally, the observed COVID-19-related pulmonary pathologies could also reduce the expression of DMETs. For example, DAD could alter the integrity of alveolar epithelial cells—which also house more than 90% of the investigated DMETs including CYP3A4 and P-gp. Our recent paper provides strategies for achieving optimal clinical efficacy and safety amidst SARS-CoV-2-associated inflammatory response (Nwabufo and Bendayan, 2022) and should be considered in the clinical decision-making process for COVID-19 drugs.

In conclusion, our study has shown for the first time that SARS-CoV-2 infection dysregulates clinically relevant DMEs—CYP3A4 and UGT1A1 at the mRNA level, and efflux transporters—P-gp and MRP1 at the protein level in Vero E6 cells and postmortem human lung tissues, respectively. We uncovered the human pulmonary localization of DMETs that are also involved in the disposition of COVID-19 drugs, and showed that inflammatory response is the driving force for the discrepancy in the localization of DMETs between COVID-19 and control human lung tissues. We observed that at the cellular level, DMETs could potentially be dysregulated by SARS-CoV-2-associated inflammatory responses and lung injuries. Further investigation of SARS-CoV-2-mediated dysregulation of human pulmonary DMETs and its potential implication in controlling the safety and efficacy of promising COVID-19 drugs as well as possible unexpected adverse drug reactions in COVID-19 patients on polypharmacy is needed. Further research is required to determine the spatial distribution and disposition of promising COVID-19 drugs at the cellular level in human lung tissues, and mass spectrometry imaging may offer an appealing analytical platform to further investigate this aspects (Nwabufo and Aigbogun, 2022).