Cannabis compounds exhibit anti-inflammatory activity in vitro in COVID-19-related inflammation in lung epithelial cells and pro-inflammatory activity in macrophages

Cannabis sativa is widely used for medical purposes and has anti-inflammatory activity. This study intended to examine the anti-inflammatory activity of cannabis on immune response markers associated with coronavirus disease 2019 (COVID-19) inflammation. An extract fraction from C. sativa Arbel strain (FCBD) substantially reduced (dose dependently) interleukin (IL)-6 and -8 levels in an alveolar epithelial (A549) cell line. FCBD contained cannabidiol (CBD), cannabigerol (CBG) and tetrahydrocannabivarin (THCV), and multiple terpenes. Treatments with FCBD and a FCBD formulation using phytocannabinoid standards (FCBD:std) reduced IL-6, IL-8, C–C Motif Chemokine Ligands (CCLs) 2 and 7, and angiotensin I converting enzyme 2 (ACE2) expression in the A549 cell line. Treatment with FCBD induced macrophage (differentiated KG1 cell line) polarization and phagocytosis in vitro, and increased CD36 and type II receptor for the Fc region of IgG (FcγRII) expression. FCBD treatment also substantially increased IL-6 and IL-8 expression in macrophages. FCBD:std, while maintaining anti-inflammatory activity in alveolar epithelial cells, led to reduced phagocytosis and pro-inflammatory IL secretion in macrophages in comparison to FCBD. The phytocannabinoid formulation may show superior activity versus the cannabis-derived fraction for reduction of lung inflammation, yet there is a need of caution proposing cannabis as treatment for COVID-19.

The combination of phytocannabinoid standards at the ratios found in fraction F CBD (F CBD:std ) showed similar activity to F CBD in the lung epithelial cell model. Based on high-performance liquid chromatography (HPLC) and gas chromatography-mass spectrometry analysis, F CBD contains approximately 66% phytocannabinoids by total content. The phytocannabinoid assemblage included CBD (93.5%), CBG (6.1%) and a minute amount of THCV (0.4%) ( Table 1; Supplementary Fig. S1). Additionally, multiple terpenes were detected in F CBD (Table 1; Supplementary Fig. S3). A combination of phytocannabinoid standards at the ratios found in fraction F CBD (F CBD:std ) resulted in activity similar to that of the cannabis-derived fraction (IC 50 of 4.1 μg/mL for IL-6 and IL-8; Fig. 1e,f). CB2 inverse agonist attenuated F CBD and F CBD:std activity in lung epithelial cell model. Treatment with CB1 or CB2 inverse agonists (IA), TRPA1 blocker, TRPV1 or TRPV2 receptor antagonists did not affect F CBD or F CBD:std activity on IL-6 secretion (Fig. 3a). Using CB2 IA and TRPV1and TRPV2 antagonists to a lesser extent, with F CBD:std treatments increased IL-8 secretion in A549 cells in comparison to F CBD:std alone (Fig. 3b). CB1 and CB2 IA, TRPA1 blocker or TRPV1 or TRPV2 receptor antagonists did not affect F CBD activity on IL-8 secretion (Fig. 3b). Treatment with CB2 IA, TRPA1 blocker or TRPV1 or TRPV2 receptor antagonists did not affect IL-6 or IL-8 levels, except for CB1 IA that led to reduction in IL-8 and IL-6 levels ( Fig. 3; Supplementary Fig. S4). F CBD treatment lead to reduction in CCL2, CCL7, ACE2 and IL-7 gene expression in lung epithelial cell model. Quantitative PCR analysis demonstrated that F CBD or F CBD:std treatments reduced the mRNA steady state level of the pro-inflammatory cytokines CCL2 and CCL7 in TNFα treated A549 cells (Fig. 4a,b). However, the reduced expression of the two genes was less than those treated with dexamethasone ( Fig. 4a,b). F CBD:std treatment led to only a 1.3-fold reduction in the expression level of IL-7, whereas F CBD and dexamethasone reduced IL-7 expression substantially (2.6-and 2.7-fold, respectively; Fig. 4c). Moreover, F CBD , F CBD:std and dexamethasone treatments reduced the expression level of angiotensin I converting enzyme 2 (ACE2), F CBD to a greater extent than dexamethasone or F CBD:std (Fig. 4d). Expression levels of these genes were examined also at 4 h post-F CBD treatment. F CBD reduction of CCL2 and CCL7 gene expression was not apparent at 4 h, although at 4 h F CBD treatment substantially reduced ACE-2 and IL-7 expression levels ( Supplementary Fig. S5).
F CBD and F CBD:std attenuate expression of phagocytosis-associated receptors. F CBD treatment, but not F CBD:std , increased expression of FcγRII and CD36 in comparison to the vehicle control ( Fig. 7a,b). Treatment with ruxolitinib which inhibits monocyte activation (described by Ahmed et al. 14 ) reduced FcγRII expression (Fig. 7a), and palmitic acid (PA) reduced expression of CD36 (Fig. 7b), in agreement with 15 . Expression of SCARB1 was reduced by F CBD and roxulitinib, but not by F CBD:std (Fig. 7c).

F CBD increases silica particl internalization in macrophages.
Imaging flow cytometry analysis showed that F CBD increased the percentage of macrophage cells that internalized FNP (Table 3; Supplementary

Discussion
We have identified a CBD rich fraction (F CBD ) from the inflorescence extract of a high CBD cannabis strain with immune-modulation activity in alveolar epithelial and macrophage cell models. F CBD reduced IL-8 and IL-6 secretion in alveolar epithelial cells. IL-8 is one of the cytokines that characterizes the cytokine storm in severe COVID-19 patients; IL-6 is a prominent cytokine also involved in the cytokine storm and is secreted during the disease from alveolar epithelial cells 3 . In addition to CBD, F CBD contained CBG and minute amount of THCV. The IC 50 of a combinations of active phytocannabinoid standards (F CBD:std ) at the relative concentrations found in F CBD were similar to that of the original fraction in the alveolar epithelial cell model. Treatment with CBD by itself led to a reduction in IL-6 and IL-8 levels in an inverse bell-shaped dose-response in alveolar epithelial cells; i.e., only 3 µg/mL was active whereas other CBD concentrations exhibited lower or no cell activity. These results are in line with an earlier publication suggesting that CBD has a bell-shaped dose-response for anti-inflammatory activity by Gallily et al. 16 . Notably, F CBD (i.e., combination of CBD with CBG and THCV) led to a dose-dependent response rather than a bell-shaped dose-response. These results are in accordance with 16 , suggesting that the addition of other phytomolecules to CBD (crude cannabis extract in the case of 16 ) prevented its bell-shaped dose-response. The CBD bell-shaped dose-response is associated with a narrow therapeutic window, which is difficult to use effectively in clinical therapy. Therefore, the fact that F CBD has a dose-dependent response makes it better suited than CBD for patient care.
CBD is a negative allosteric modulator of CB1 signaling 17 . TRPA1 is a receptor in alveolar epithelial cells involved in the pathogenesis of several airway diseases including chronic obstructive pulmonary disease and asthma 18 . Both TRPV1 and TRPV2 interact with phytocannabinoids, including CBD, CBG and THCV 19 . Also, TRPV1, TRPV2 and TRPA1 were found to be associated with pulmonary inflammation 20 . Nevertheless, cotreatment with CB1 IA, TRPA1 blocker or TRPV1 or TRPV2 antagonist had no substantial effect on F CBD and www.nature.com/scientificreports/ F CBD:std activity. Only co-treatment with CB2 IA affected F CBD:std activity on IL-8 secretion. The involvement of receptors in F CBD and F CBD:std activity remains to be demonstrated. In addition to reducing IL-6 and IL-8 levels, F CBD and F CBD:std reduced the expression levels of CCL2 and CCL7 in alveolar epithelial cells by 6 h treatment. The systemic cytokine profiles detected in severe COVID-19 patients includes increased production of inflammatory chemokines such as CCL2 21 . Moreover, CCL2 and CCL7 were shown to be abundant in bronchoalveolar fluid from severe COVID-19 patients and were associated with recruitment of monocytes into the lungs 21 . Our results suggest that treatment with F CBD or F CBD:std may lead to reduced secretion of inflammatory cytokines associated with the disease, and possibly to a reduction of macrophage recruitment during the cytokine storm. However, dexamethasone was more effective than F CBD in reducing both CCL2 and CCL7 expression. IL-7 was shown to raise lymphocyte counts in septic patients with low absolute lymphocyte counts 22 and to restore protective immunity in patients that suffer from CD4+ T cell deficiency (e.g., as in the case of HIV infection 23 ). It was suggested that treatment against SARS-CoV-2 infections should also attempt to increase IL-7 levels 22 . The fact that F CBD:std reduced IL-7 expression only to a minor extent in comparison to dexamethasone or F CBD suggests that using purified phytocannabinoids may have an advantage over cannabis-derived fractions for COVID-19-like inflammation.
The ACE2 receptor is a part of the dual renin-angiotensin system (RAS) 24 . ACE2 was shown to be involved with SARS-CoV-2 human infection; the ectodomain of the S protein of SARS-CoV-1 binds to the peptidase domain of ACE2 with relatively high affinity 25 . In cells of patients with severe symptoms of COVID-19, ACE2 was substantially upregulated 199-fold; this upregulation was suggested to be one of the factors leading to disruption of the RAS, as ACE2 is a part of the counteracting hypotensive axis of RAS. The increase in ACE2 and other key RAS components is predicted to elevate bradykinin levels in multiple tissues, leading to increases in vascular permeability and hypotension; the latter is highly associated with severe COVID-19 patients 26 . Indeed, a negative correlation was identified between ACE2 gene expression and COVID-19 mortality 27 . F CBD reduced the expression level of ACE2 at 4 and 6 h post treatment. F CBD:std and dexamethasone also reduced ACE2 expression but to a lesser extent. However, the ability of F CBD to reduce ACE2 expression should be examined at both the protein and functional levels (e.g., binding of the viral protein) to fully determine the effect F CBD may have on ACE2-related treatment of COVID-19 patients. In any case, such reduction of ACE2 expression should be considered with care as the advantages and disadvantages of this reduction are disputed 24 .
In the first phase of the disease, a specific adaptive immune response is needed to eliminate the virus and to prevent disease progression to more severe stages 2 . Indeed, the dysfunction of alveolar macrophages are among the abnormal characteristics in some severe COVID-19 patients 28 , and an abundance of increased inflammatory monocyte-derived macrophages replaces tissue-resident alveolar macrophages in patients with severe disease 21 . Additionally, during SARS-CoV-1 infections that provoke a disease course similar to those seen during infection with SARS-CoV-2 21 , a marked reduction in macrophages phagocytosis activity was detected 29 . Also, phagocytosis was important in the antibody-mediated elimination of SARS-CoV-1 in a mouse model 30 .  www.nature.com/scientificreports/ Notably, F CBD and F CBD:std , and CBD to a lesser extent, led to a marked increase in macrophage polarization and to cell actin remodeling that corresponds to the growth of filopodia-like membrane structures 31 . F CBD reduced expression of SCARB1; SCARB1 encodes SR-B1 that is a scavenger receptor (class B) and is also responsible for phagocytosis of silica particles in macrophages 32 . However, F CBD treatment also led to an increase in FcγRII and CD36 gene expression. Phagocytosis is initiated by the ligation of Fcγ receptors to IgG-opsonins on the target  www.nature.com/scientificreports/ cell 33 , whereas CD36 expression in macrophages was shown to be involved with lung fibrosisin in mice 34 . Alveolar macrophages play an important role in Fc receptor-mediated responses during acute virus infections and in phagocytosis-mediated clearance of respiratory virus infections 35,36 . CD36 is an important scavenger receptor for phagocytosis of Streptococcus pneumoniae, a primary bacterial agent associated with pneumonia, which is down regulated by influenza 37 . Indeed, F CBD led to a marked increase in the internalization of silica particles by macrophages, and in so doing, increased levels of phagocytosis. Possibly, the increase in macrophage polarization and phagocytosis, and the upregulation of FcγRII and CD36 expression in these cells following F CBD treatment may facilitate phagocytosis-mediated clearance of respiratory viruses, and benefit the first phase of the immune response to SARS-CoV-2. However, it should be noted that macrophages themselves can be infected by the virus, as SARS-CoV-1 infects macrophages as a result of IgGmediated phagocytosis that requires FcγRII receptor signaling pathways 38 . Advantages and disadvantages of increasing macrophage phagocytosis activity should be carefully considered 3,21 .
Notably, although F CBD:std treatment increased macrophage polarization, it did not increase the phagocytosisassociated gene expressions, nor phagocytosis. Hence, additional active compounds in the cannabis-derived F CBD and not in the phytocannabinoid standard mix that composed F CBD:std are responsible for this increased gene expression and phagocytosis activity. Indeed, F CBD contained multiple terpenes, some including γ-Curcumene and Guaiol at considerable percentages. The presence of terpenes in F CBD may account for the differences in activity between F CBD and F CBD:std .
During the second phase of COVID-19, pneumonia patients exhibit features of macrophage activation syndrome (MAS) in which macrophages play a major pro-inflammatory role by releasing pro-inflammatory cytokines such as IL-6, IL-8 and CCL2 3 . Moreover, subsets of macrophages in patients with COVID-19 were found to express genes associated with IL-6, whereas expression of IL-6 was again associated with severe depletion of lymphocytes from the spleen and lymph nodes 21 . Notably, F CBD led to a marked increase of IL-8 expression and IL-8 protein levels in macrophages. It also led to an increase in IL-6 expression levels, above that induced by PMA 39 . These results suggest a substantial, in vitro, pro-inflammatory role for F CBD in macrophages. However, F CBD;std was less active in ILs induction, again demonstrating a notable difference between F CBD and F CBD:std , which may originate from the presence or absence, respectively, of terpenes.
To conclude, treatment with cannabis compounds CBD, CBG and THCV may have clinical value in reducing cytokine secretion in lung epithelial cells. However, treatment with F CBD containing terpenes in addition to these phytocannabinoids substantially induced macrophage phagocytosis and increased their IL levels. Yet, to confirm more specifically the pro-inflammatory effect of F CBD in macrophages it is necessary to perform the same experiments on primary alveolar macrophages (e.g., from mice). Nevertheless, these results suggest a pro-inflammatory role for cannabis extract that is higher than that of the phytocannabinoid standard mix. The latter maintained anti-inflammatory activity in the alveolar epithelial cells with relatively reduced proinflammatory activity in macrophages. Hence, the mix of phytocannabinoids shows superior activity versus the Figure 6. Representative examples of confocal images of macrophages following treatment with the control and F CBD (7 µg/mL). Control (0.7% v/v methanol) treatment served as the solvent (vehicle) control. Cell were stained for F-actin (EasyProbes™ ActinRed 555 Stain, red stain), and nuclei (Hoechst, blue stain); n = 5, in each biological replicate multiple cells were examined (see Table 2). Membrane filopodia-like structures are marked with white arrows. FNP, fluorescent-labeled silica 50-100 nm particles.

Material and methods
Extract preparation. High CBD Cannabis sativa strain Arbel (IMC, Israel) inflorescence was extracted using ethanol as described previously 41 and decarboxylated by heating the dried crude extract to 220 °C for 10 min. The dried decarboxylated extract was weighed, and then resuspended in absolute methanol (volume of solvent added according to the desired concentration) and filtered through a 0.45 μm syringe filter.

Extract fractionation.
A flash chromatography apparatus equipped with a diode array detector was used to fractionize the decarboxylated crude extract. An Ecoflex C-18 80 g (Flash Pure, Buchi, C-18, 50 µm spherical, max. pressure 180 psi) column was used for separation, with methanol and water as the mobile phase, as suggested by the manufacturer.
Chemical analyses. HPLC and gas chromatograph with mass selective detector (GCMS 8860 and GC/ MSD 5977B, Agilent) analysis was carried out as previously described 41 . Qualitative and quantitative analysis of the phytocannabinoids in fractions was done in comparison to the standard calibration curves obtained from dissolving standards in methanol at different concentrations from 0 to 25 µg/mL.

Standard/material preparation and use.
The cannabinoid standards at a concentration of 1 mg/mL in methanol used in this study included cannabidiol (CBD, Restek catalog no. 34011) cannabigerol (CBG, Restek catalog no. 34091) and tetrahydrocannabivarin (THCV, Restek catalog no. 34100). Inverse agonists (IA) to cannabinoid receptors type 1 (CB1) and 2 (CB2) used were Abcam products: CB1 (AM251, ab120088), CB2 (SR144528, ab146185), as was the TRPA1 blocker (HC-030031, ab120554). All IAs, as well as the TRPV1 antagonist (Abcam ab141772) and TRPV2 antagonist (Tranilast 1098/10) were dissolved in dimethyl sulfoxide (DMSO) at a concentration of 10 mM. Phorbol 12-myristate 13-acetate (PMA) (P1585; Sigma Aldrich, USA) was dissolved in DMSO at the stock concentration of 5 µg/mL. Dexamethozone (D4902; Sigma Aldrich, USA) was dissolved in methanol at the stock concentration of 1000 µg/mL. Ruxolitinib JAKAVI was dissolved in DMSO at the concentration of 5000 µg/mL, confirmed with GCMS and HPLC and was used at a final concentration of 2% (v/v). TNFα (300-01A; PeproTech, Rocky Hill, NJ, USA) was dissolved in water at the stock concentration of 100 µg/mL. Determination of IL levels and cell viability. IL-6 and IL-8 levels were determined as described previously 43 with the following modifications: A549 cells were plated at 5 × 10 4 cells per well in DMEM complete media (400 µL) in 24-well cell culture plate. They were allowed to attach and grow at 37 °C in air and 5% CO 2 in a humidified incubator overnight with complete DMEM, and then the media was replaced with serum free DMEM. Cell excitation was performed with 300 ng/mL TNFα. Treatments were performed with cannabis crude extract, fraction or combination of compounds together with 100 µL serum free DMEM. IL-6 and IL-8 www.nature.com/scientificreports/ secretion levels were analyzed after 4 or 6 h of incubation for A549 or KG1 cell lines, respectively. Supernatant samples were collected and tested using IL-6 and IL-8 ELISA kits (DY206 and DY208 respectively, R&D Systems, Minneapolis, MN, USA). Dexamethasone was used as a positive control. For cell viability, an Alamar Blue (resazurin) assay was performed on each well as described previously 43 . For dose response assays, data points were connected by non-linear regression lines of the sigmoidal dose-response relation. GraphPad Prism version 6.1 (https ://www.graph pad.com/scien tific -softw are/prism /, GraphPad Software Inc., San Diego, USA) was employed to produce dose-response curves and IC50 doses were calculated using nonlinear regression analysis.

Salinization of silicon dioxide surfaces with APTES.
To prepare the silica dispersion, 1 g of silica was added to 40 mL of methanol and stirred. Then, APTES (0.0045 mol) was slowly added to the solution. The reaction was carried out at ambient temperature for 45 min. After silanization, 50-100 nm or 30-70 nm particles were collected by centrifugation (9000 rpm, 10 min) washed 4 times with water, and dried at 35 °C under vacuum for 3 h 44 .
Labeling of amine functionalized silica nanoparticles with 5(6)-carboxyfluorescein and IgG. Quantitative real-time PCR. Quantitative real time PCR (qPCR) was carried out as described previously 41 .
Briefly, cells were treated with cannabis compounds or methanol (0.7% v/v) as vehicle control for 4 or 6 h. Cells were then harvested and total RNA was extracted. RNA was reverse-transcribed, primers were designed and PCR was performed. The expression of each target gene was normalized to the expression of Hypoxanthine Phosphoribosyltransferase 1 (HPRT1) mRNA in the 2 −ΔΔCt and is presented as the ratio of the target gene to HPRT mRNA, expressed as 2 −ΔCt , where Ct is the threshold cycle and ΔCt = Ct Target − Ct HPRT1. Experiments were repeated three times. The primers were: www.nature.com/scientificreports/ Imaging flow cytometry. Differentiated macrophages from KG1 cells (10 × 10 5 cells/plate; seeded on 6-well plate culture dish) were replaced with 2 mL of 5% FBS-IMDM media with FNP, ENP, or ENPG (40 µg/ mL) and incubated at 37 °C for 4 h for phagocytosis. The cells were detached from the surface of the plate using a trypsin 0.25%:EDTA 0.05% solution (03-052-1A, Biological Industries, Israel) for 3 min, washed with DMEM complete media, centrifuged and transferred to 50 µL cold PBS kept on ice. Cells were analyzed by multispectral imaging flow cytometry (ImageStream markII flow cytometer; Amnis Corp, part of EMD Millipore, Seattle, WA, USA). Fluorescence intensity of the Fluorescein labeled silica beads was measured in channel 2 of the cytometer (480 nm ex, 560 nm em). An 60× magnification with Olympus UplanFLN 60× dry objective 0.9NA was used for all samples. At least 4000 cells were collected for each sample and data were analyzed using a dedicated image analysis software (IDEAS 6.2; AmnisCorp). Cells were gated for single cells using the area and aspect ratio features, and for focused cells using the Gradient RMS feature. Cropped cells were further eliminated by plotting the cell area of the bright field image against the Centroid X feature (the number of pixels in the horizontal axis from the left corner of the image to the center of the cell mask). Cells were further gated for cells that were positive (for ENP, ENPG or FNP). Because of their larger size, only FNP beads could be further analyzed for bead internalization vs. those attached to the cell surface. This was done using the intensity feature (the sum of the background-subtracted pixel values within the masked area of the image) and max pixel (the largest value of the subtracted background pixel). FNP internalization was calculated by the internalization feature, i.e. the ratio of the intensity inside the cell to the intensity of the entire cell, mapped to a log scale. To define the internal mask for the cell, the object mask of the brightfield image was eroded by 8 pixels. Cells with an internalization score higher than 0.33 were gated as cells with internalized FNP. Statistical analysis. Data were processed using the JMP statistical package (https ://www.jmp.com/en_us/ home.html, SAS Inc, NC, USA). Comparisons between two groups were made using the Student's T-Test. Comparisons between more than 2 groups were made with analysis of variance (ANOVA) followed by Tukey-Kramer's honest significant difference (HSD) test as post hoc. Values are shown as mean ± standard error (sem). P values ≤ 0.05 were considered significant.

Data availability
The datasets generated during and/or analysed during the current study are available from the corresponding author on reasonable request.