Interferon γ treatment increases endocannabinoid and related N‐acylethanolamine levels in T84 human colon carcinoma cells

Abstract

Background and Purpose

Endocannabinoids and related N‐acylethanolamines (NAEs) are involved in regulation of gut function, but relatively little is known as to whether inflammatory cytokines such as IFNγ affect their levels. We have investigated this in vitro using cultures of T84 colon cancer cells.

Experimental Approach

T84 cells, when cultured in monolayers, differentiate to form adult colonic crypt‐like cells with excellent permeability barrier properties. The integrity of the permeability barrier in these monolayers was measured using transepithelial electrical resistance (TEER). NAE levels were determined by ultra‐performance liquid chromatography‐tandem mass spectrometric analysis. Expression of the enzymes involved in NAE and 2‐arachidonoylglycerol (2‐AG) turnover were assessed with qPCR.

Key Results

IFNγ treatment for 8 or 24 h increased levels of both endocannabinoids (anandamide and 2‐AG) and the related NAEs. The treatment did not affect the rate of hydrolysis of either anandamide or palmitoylethanolamide by intact cells, and in both cases, fatty acid amide hydrolase (FAAH) rather than NAE‐hydrolysing acid amidase (NAAA) was mainly responsible for the hydrolysis of these NAEs. IFNγ treatment reduced the TEER of the cells in a manner that was not prevented by inhibition of either FAAH or NAAA but was partially reversed by apical administration of the NAE palmitoylethanolamide.

Conclusion and Implications

IFNγ treatment mobilized endocannabinoid and related NAE levels in T84 cells. However, blockade of anandamide or NAE hydrolysis was insufficient to negate the deleterious effects of this cytokine upon the permeability barrier of the cell monolayers.

Linked Articles

This article is part of a themed section on 8^th European Workshop on Cannabinoid Research. To view the other articles in this section visit http://onlinelibrary.wiley.com/doi/10.1111/bph.v176.10/issuetoc

Abbreviations

2‐AG

2‐arachidonoylglycerol

ABHD

αβ hydrolase domain

AEA

anandamide

BHT

butylhydroxytoluene

CUDA

12‐[[(cyclohexylamino)carbonyl]amino]‐dodecanoic acid

EPEA

eicosapentaenoyl ethanolamide

FAAH

fatty acid amide hydrolase

LEA

linoleoyl ethanolamide

MAGL

monoacylglycerol lipase

NAAA

N‐acylethanolamine‐hydrolysing acid amidase

NAEs

N‐acylethanolamines

NAPE‐PLD

N‐acyl phosphatidylethanolamine PLD

OEA

oleoylethanolamide

PEA

palmitoylethanolamide

SEA

stearoylethanolamide

TEER

transepithelial electrical resistance

URB597

cyclohexylcarbamic acid 3′‐carbamoylbiphenyl‐3‐yl ester

Introduction

The endocannabinoid system, comprising the cannabinoid (CB) receptors, their endogenous ligands (anandamide, AEA, and 2‐arachidonoylglycerol, 2‐AG) and the biosynthetic and degradative machinery for these ligands, plays a variety of regulatory roles both in the brain and the periphery (Ligresti et al., 2016). In the gut, for example, the endocannabinoid system acts to limit the severity of symptoms produced by experimental colitis (Massa et al., 2004; Storr et al., 2008; Sałaga et al., 2014). AEA belongs to a class of lipids termed N‐acylethanolamines (NAEs), which include biologically active compounds such as palmitoylethanolamide (PEA) and oleoylethanolamide (OEA). PEA acts as an endogenous anti‐inflammatory agent and has clinical efficacy in pain conditions (Gabrielsson et al., 2016).

The canonical pathway for NAE synthesis is from membrane phosphatidylethanolamine‐ and phosphatidylcholine‐containing phospholipids by a two‐step process: transacylation to form N‐acyl phosphatidylethanolamine (NAPEs) followed by hydrolysis to the corresponding NAEs by the enzyme NAPE‐PLD. However, other pathways have been described (Ueda et al., 2013). NAEs are hydrolysed to their corresponding fatty acids by the enzymes fatty acid amide hydrolase (FAAH) and NAE‐hydrolysing acid amidase (NAAA) (Ueda et al., 2013). 2‐AG is synthesized by a pathway involving diacylglycerol lipases, and metabolism is again by hydrolysis, of which monoacylglycerol lipase (MAGL) is a key enzyme, although FAAH can also hydrolyse this endocannabinoid (Ueda et al., 2013). Selective inhibitors of these AEA/NAE and 2‐AG hydrolytic enzymes have been shown to be efficacious in experimental models of colitis and colon cancer (Izzo et al., 2008; Storr et al., 2008; Alhouayek et al., 2011, 2015; Sałaga et al., 2014). With respect to the colitis models, the end points measured were the macroscopic changes, inflammatory cell infiltration and disruption of intestinal cell motility. However, no direct measurements of the gut permeability barrier function, known to be disturbed in inflammatory bowel disorders, were reported in these studies.

Gut disorders such as Crohn's disease and ulcerative colitis have a strong inflammatory component, with cytokines such as IFNγ being involved in their pathogenesis (Groschwitz and Hogan, 2009). There is ample published evidence that immune responses can be modulated by ligands interacting with CB receptors, be they synthetic, plant‐derived or endogenous (see Tanasescu and Constantinescu, 2010; Kaplan, 2013; Turcotte et al., 2016). However, there is less information available with respect to the converse, that is, whether inflammatory cytokines affect endocannabinoid signalling, and most information concerns the effects of cytokines upon CB receptor expression (Maccarrone et al., 2001; Börner et al., 2008; Jean‐Gilles et al., 2015). Given the role of the endocannabinoid system in mitigating the symptoms of colitis, it is conceivable that a cytokine‐induced deficit in endocannabinoid levels could be involved in the pathogenesis of gut disorders and that restoration of this deficit with FAAH or NAAA inhibitors could play a part in their therapeutic effects in the colitis models. As an initial step in investigating this possibility, the present study has investigated the effect of IFNγ upon the levels of the endocannabinoids, related NAEs and the catabolic enzymes FAAH and NAAA in a human colon carcinoma cell line; determined whether blockade of these enzymes increases NAE levels in these cells cultured under inflammatory conditions and determined whether such a blockade reverses the permeability deficits of the cells produced by the inflammatory conditions.

Methods

Cell culture

Human T84 colon carcinoma cells (American Type Culture Collection, Rockville, MD, USA), selected for their capacity to form polarized, tight monolayers on semipermeable support (Ou et al., 2009) and frozen directly after expansion, were cultured at 37°C in a 5% CO₂ environment in DMEM/F12 medium supplemented with 8% FBS, 1% PEST and 1% L‐glutamine. Throughout, cells were kept in culture for 10 passages and used from passage 2 to 10 after defrosting the cryotubes. T84 cells were subcultured after partial digestion with 0.25% trypsin and counted with Trypan Blue before plating.

Extraction and quantification of NAEs and 2‐AG by ultra‐performance liquid chromatography‐tandem mass spectrometric analysis

In the first set of experiments, T84 cells were seeded overnight in six‐well plates at a density of 1 × 10⁶ cells per well. Medium was then replaced with either fresh medium or medium containing IFNγ (10 ng·mL⁻¹) for 8 or 24 h and then incubated for a further 20 min with 0 or 2 μM ionomycin. After the second incubation period, the cell plates were immediately placed on ice and scraped with 2 × 1 mL of methanol using a rubber policeman. Cell debris of the methanol extracts in Falcon tubes was sedimented by centrifugation at 2000 g for 15 min (4°C), and the methanol extracts were stored at −80°C until assayed (never longer than 2 days to avoid chemical decomposition of the metabolites).

In the second set of experiments, essentially the same protocol was followed, although in this case, the cells were incubated with medium containing IFNγ (10 ng·mL⁻¹) for 22 h, after which either vehicle, URB597 (1 μM final concentration) or pentadecylamine (30 μM final concentration) was added and the samples were incubated for a further 2 h. Extracts were then prepared as described above.

The methanolic extracts were diluted with milliQ water to a final concentration of 5% methanol (v/v). To samples of these extracts were added 10 μL of internal standard solutions [deuterated lipids, of which those relevant for the present study are AEA‐d₈, PEA‐d₄ (used as internal standard for PEA and linoleoyl ethanolamide; LEA), OEA‐d₄ (used as internal standard for OEA, eicosapentaenoyl ethanolamide; EPEA and palmitoleoyl ethanolamide; POEA), stearoylethanolamide (SEA)‐d₃ and 2‐AG‐d₈] and 10 μL antioxidant solution [0.2 mg·mL⁻¹ butylhydroxytoluene (BHT)/EDTA in methanol/water (1:1)] and then applied to Waters Oasis HLB SPE cartridges (200 mg of sorbent, 30 μm particle size) (for details of the complete internal standards mixture, see the original methodology paper of Gouveia‐Figueira and Nording, 2015). The cartridges were washed with ethyl acetate (2 mL), followed by methanol (2 × 2 mL) before being conditioned with wash solution (WS, 95:5 v/v water/MeOH with 0.1% acetic acid, 2 × 4 mL). After the sample containing internal standard and antioxidant solutions was loaded, cartridges were washed with 2 × 4 mL of WS and eluted with 4 mL acetonitrile, followed by 2 mL of methanol and 1 mL of ethyl acetate. Eluates were concentrated with a MiniVac system (Farmingdale, NY, USA), reconstituted in 100 μL of methanol and vortexed. Solutions were transferred to LC vials with low‐volume inserts, 10 μL of 12‐[[(cyclohexylamino)carbonyl]amino]‐dodecanoic acid (CUDA; 50 ng·mL⁻¹) was added, and analysis was performed immediately using an Agilent ultra‐performance liquid chromatography system (Infinity 1290) coupled with the electrospray ionization source in positive mode to an Agilent 6490 Triple Quadrupole system equipped with the iFunnel Technology (Agilent Technologies, Santa Clara, CA, USA). Analyte separation was achieved with a Waters BEH C₁₈ column (2.1 × 150 mm, 2.5 μm particle size). The mobile phase consisted of (A) 0.1% acetic acid in MilliQ water and (B) acetonitrile:isopropanol (90:10). The following gradients were employed: 0–2.0 min 30–45% B, 2.0–2.5 min 45–79% B, 2.5–11.5 min 79% B, 11.5–12 min 79–90% B, 12–14 min 90% B, 14–14.5 min 90–79% B, 14.5–15.5 min 79% B, and 15.6–19 min 30% B. Precursor ions, [M + H]⁺ and [M‐H]⁻, product ions, multiple reaction monitoring (MRM) transitions and optimal collision energies were established for each analyte. ESI conditions were as follows: capillary and nozzle voltage at 4000 and 1500 V, drying gas temperature 230°C with a gas flow of 15 L·min⁻¹, sheet gas temperature 400°C with a gas flow of 11 L·min⁻¹, the nebulizer gas flow was 35 ψ, and iFunnel high and low pressure RF at 150 and 60 V (positive mode). The dynamic MRM option was performed for all compounds with optimized transitions and collision energies. The MassHunter Workstation software was used to manually integrate all peaks. For further details, see Gouveia‐Figueira and Nording (2015). It should be noted that a second run of the analytes was performed with the electrospray ionization source in negative mode to obtain data on oxylipin concentrations. These data have been published elsewhere (Alhouayek et al., 2018).

Assay of AEA and PEA uptake and hydrolysis

Cells were seeded overnight at a density of 1.5 or 2.5 × 10⁵ cells per well in 24‐well plates. Uptake of [Ara‐³H]AEA and [Pa‐³H]PEA was determined using the method of Rakhshan et al. (2000) as modified by Sandberg and Fowler (2005). Briefly, cells were washed twice with 400 μL of pre‐warmed KRH buffer (composition: 120 mM NaCl, 4.7 mM KCl, 2.2 mM CaCl₂, 10 mM HEPES, 0.12 mM KH₂PO₄, 0.12 mM MgSO₄, pH 7.4) with and without 1% BSA, after which 300 μL of pre‐warmed KRH buffer with 0.1% fatty acid‐free BSA and 50 μL of test compound or vehicle was added and the cells were incubated for 10 min at 37°C. [Ara‐³H]AEA or [Pa‐³H]PEA (50 μL, diluted with the corresponding non‐radioactive NAE to give a final concentration of 104 nM, in KRH buffer with 0.1% fatty acid‐free BSA) was added, and the cells were incubated for 4 or 15 min, as appropriate, at 37°C. After the incubation, the buffer was removed, and the wells washed three times with ice‐cold KRH with 1% BSA after which 500 μL of 0.2 M NaOH was added, and the cells solubilized by heating in a warm oven. Aliquots (300 μL) were transferred to scintillation vials together with 4 mL scintillation fluid and analysed for tritium content by liquid scintillation with quench correction. Retention of the ligand by the plastic wells was assessed concomitantly using the same assay but in the absence of cells.

[Et‐³H]AEA and [Et‐³H]PEA hydrolysis by the cells was measured using the same protocol as above using the same incubation concentrations of radioligand, but reactions were stopped by addition of 600 μL of activated charcoal solution (120 μL activated charcoal +480 μL 0.5 M HCl) (Boldrup et al., 2004) followed by mixing and centrifugation (1200 x g for 10 min at room temperature) of 600 μL aliquots. Aliquots (200 μL) of the aqueous phase were then transferred to scintillation vials together with 4 mL scintillation fluid and analysed for tritium content by liquid scintillation with quench correction. Blanks were wells without cells.

Transepithelial electrical resistance (TEER) measurements

Cells were seeded in transwells at a density of 2.5 × 10⁵ cells per well. The medium (500 and 1500 μL in the apical and basolateral compartments, respectively) was changed every day. The cells were used when TEER, measured using an Evohm voltameter coupled to the endohm‐12 chamber (World Precision Instruments, Hitchin, UK), was consistently >900 Ω.cm² (usually days 9–10). Initially (t0), TEER was measured and then the medium in the basolateral compartment was replaced with 1.5 mL fresh medium or medium containing 10 ng·mL⁻¹ IFNγ, as appropriate. The medium in the apical compartment was replaced with 500 μL of medium containing either vehicle (DMSO, 0.1% v/v) or the test compounds (PEA, final concentration 10 μM, URB597 1 μM and pentadecylamine 30 μM), and the wells were incubated for 24 h, after which TEER was re‐measured. In some cases, PEA (final concentration 10 μM) was added to the basolateral, rather than the apical compartment. In some experiments, cell lysates were then collected for mRNA measurements by RT‐qPCR.

mRNA extraction and RT‐qPCR

The Dynabeads® mRNA direct purification kit was used to extract mRNA according to the manufacturer's instructions. Briefly, inserts were washed with PBS, and lysis/binding buffer was added before the plates were stored at −80°C. The extracted mRNA was quantified in Nanodrop Lite® (Thermo Fisher Scientific), diluted in TE buffer to a concentration of 5 ng·μL⁻¹ (10 μL total) and mixed with the 2XRT master mix (High capacity cDNA reverse transcription kit). Reverse transcription was run using a Life Touch thermal cycler (BIOER, Hangzhou, China) with 1/10 diluted cDNA loaded into the assay plates. qPCR was performed with an Eco Real‐Time PCR instrument (Illumina Inc., San Diego, CA, USA) using KAPA SYBR FAST qPCR kit Master Mix for the PCR reactions and with each sample measured in duplicate. An initial holding stage of 3 min at 95°C followed by 45 cycles consisting of denaturation at 95°C for 3 s and annealing/extension at 60°C for 30 s was used for amplification, and products were analysed by performing a melting curve at the end of the PCR reaction. Data were normalized to the 60S ribosomal protein L19 (RPL19). The ΔCt [i.e. threshold cycle (Ct) of the gene of interest minus the corresponding Ct of the reference gene (RPL19)] is reported here, where a ΔCt of −1 corresponds to a doubling of the mRNA. The primers used are shown in Table 1.

Table 1.

Primer sequences used for RT‐PCR

Gene	Product	Forward primer (5′ to 3′)	Reverse primer (5′ to 3′)
RPL19	RPL19	CAC ATC CAC AAG CTG AAG GCA	CTT GCG TGC TTC CTT GGT CT
NAPEPLD	NAPE‐PLD	ACT GGT TAT TGC CCT GCT TT	AAT CCT TAC AGC TTC TTC TGG G
FAAH	FAAH	CAC ACG CTG GTT CCC TTC TT	GGG TCC ACG AAA TCA CCT TTG A
NAAA	NAAA	ATG GAG CGT GGT TCC GAG TT	AGG CTG AGG TTT GCT TGT CCT
DAGLA	DAGLα	CCC AAA TGG CGG ATC ATC G	GGC TGA GAG GGC TAT AGT TAG G
DAGLB	DAGLβ	TCA GGT GCT ACG CCT TCT C	TCA CAC TGA GCC TGG GAA TC
MGLL	MAGL	GGA AAC AGG ACC TGA AGA CC	ACT GTC CGT CTG CAT TGA C
ABHD6	ABHD6	GAT GTC CGC ATC CCT CAT AAC	CCA GCA CCT GGT CTT GTT TG
ABHD12	ABHD12	GGC AGA AAG CTC TAT AGC ATC G	CCT GTA GCC AAG GTC TGA ATG

Data and statistical analyses

The data and statistical analysis comply with the recommendations on experimental design and analysis in pharmacology (Curtis et al., 2015). For technical and other reasons, the group sizes for some of the experiments are unequal. Unbalanced samples are more sensitive to violation of the assumptions of equal variances and normality of residues, and so we have used permutation methods rather than standard t‐tests / ANOVAs when possible, since they are robust to such violations. Mixed model analyses were conducted using the function lme in the package nlme for the R statistical programme versions 3.3.2–3.4.1 (R Core Team, 2016). Fisher's randomization (permutation) tests using 10⁶ iterations (except when a complete enumeration was feasible) were conducted using the functions independence_test, aovp and permTS in the coin, lmperm and perm packages, respectively, for R.

Statistical significance was set at P<0.05. Note, however, that in the lipidomic and qPCR experiments summarised in Figs. 1 and 4, the use of a simple P<0.05 does not allow for identification of potential false positives due to multiple testing. In these cases, we have used a 5% false discovery rate (Benjamini & Hochberg, 1995) to determine the critical values of P. In theory, adjusted P values can be calculated (for example with the p.adjust function in R), but because readers may have their own preferences as to the best way to deal with multiple testing in exploratory data, giving the unadjusted P values in supplementary Tables S1 and S2 allows the reader to make her/his own calculations.

Figure 1.

Levels of 2‐AG and NAEs in T84 extracts following treatment with IFNγ and ionomycin. Cells were treated with IFNγ (10 ng·mL⁻¹) for either 8 or 24 h after which ionomycin (Iono; 0 or 2 μM) was added and the cells incubated for a further 20 min. Shown are scatterplots with the bars representing means, n = 8–9. The experiment was designed so that all groups were n = 9, but some values were lost for technical reasons. Data points shown as triangles were identified as possible outliers but included in the statistical analyses. Data points shown as red squares were identified as probable outliers, and the statistical analyses were undertaken both with and without their inclusion. The full statistical analysis is given in Table S1, Supporting information.

Throughout the text, N refers to the number of separate experiments conducted with two to six replicates. For the NAE and 2‐AG quantification experiments, n refers to the number of separate samples extracted and analysed from different wells. In these experiments, box and whisker plots were undertaken using GraphPad Prism v7.0b for the Macintosh (GraphPad Software Inc., San Diego, CA, USA) to identify anomalous values. Confidence intervals of the means were also obtained using this programme. In the Figures, possible outliers (≥1.5 × interquartile range above or below the 25% and 75% values) are shown as triangles but were included in the analyses. Probable outliers (≥3 × interquartile range above or below the 25% and 75% values) are shown as red squares in the Figures, and were excluded from the analyses.

Materials

[Ethanolamine‐1‐³H]AEA ([Et‐³H]AEA, 2.22 GBq·mmol⁻¹), [arachidonyl‐5,6,8,9,11,12,14,15‐³H]AEA ([Ara‐³H]AEA, 7.4 GBq·mmol⁻¹), [ethanolamine‐1‐³H]PEA ([Et‐³H]PEA, 0.74 GBq·mmol⁻¹) and palmitoyl [9,10‐³H] ethanolamide ([Pa‐³H]PEA, 2.22 GBq·mmol⁻¹) were obtained from American Radiolabeled Chemicals Inc. (St. Louis, MO.). AEA, PEA, OEA, SEA, eicosapentaenoyl ethanolamide (EPEA), palmitoleoyl ethanolamide (POEA), linoleoyl ethanolamide (LEA), 2‐AG, AEA‐d₈, PEA‐d₄, OEA‐d₄, SEA‐d₃, 2‐AG‐d₈, 12‐[[(cyclohexylamino)carbonyl]amino]‐dodecanoic acid (CUDA), butylhydroxytoluene (BHT) and cyclohexylcarbamic acid 3′‐carbamoylbiphenyl‐3‐yl ester (URB597) were purchased from Cayman Chemical Co. (Ann Arbor, MI, USA). Ionomycin, pentadecylamine and diacerein were obtained from Sigma‐Aldrich (St. Louis, MO, USA). Recombinant human IFNγ, cell culture reagents [medium (DMEM/F12, GIBCO cat. no. 11330–057), trypsin, FBS, PEST and L‐glutamine] and Dynabeads® mRNA direct kit were purchased from Life technologies (Thermo Fisher Scientific, Waltham, MA USA). Transwell permeable supports cat. no. 3401 were obtained from Corning Inc. (Corning, NY, USA). High capacity cDNA reverse transcription kit was purchased from Applied Biosystems, Thermo Fisher Scientific. KAPA SYBR® FAST qPCR master mix was purchased from KAPA Biosystems, (Wilmington, MA, USA).

Nomenclature of targets and ligands

Key protein targets and ligands in this article are hyperlinked to corresponding entries in http://www.guidetopharmacology.org, the common portal for data from the IUPHAR/BPS Guide to PHARMACOLOGY (Harding et al., 2018), and are permanently archived in the Concise Guide to PHARMACOLOGY 2017/18 (Alexander et al., 2017).

Results

Effect of IFNγ and ionomycin treatment upon NAE levels in T84 cells

T84 cells in six‐well plates were treated with IFNγ (10 ng·mL⁻¹) for 8 or 24 h and then incubated for a further 20 min with 0 or 2 μM of the calcium ionophore ionomycin prior to collection of cells, extraction of lipids and analysis of NAEs and 2‐AG by ultra‐performance liquid chromatography‐tandem mass spectrometric analysis. Seven NAEs and 2‐AG were detected in the T84 cells (Figure 1). There were some probable outliers found (one for OEA, one for SEA and three for LEA). For OEA and SEA, inclusion of the outliers gave essentially the same statistical significance levels as seen when they were excluded (both are shown in Table S1). For LEA, however, this was not the case, and so we have elected not to interpret the data for this lipid. For the other lipids (with the possible exception of EPEA), there was a significant main effect of IFNγ, the cytokine treatment causing increased levels. A significant main effect of ionomycin treatment to increase levels was seen for 2‐AG, SEA and POEA, but not for AEA or PEA. A significant decrease in levels was seen for OEA and a significant interaction IFNγ × ionomycin was seen for EPEA (Figure 1, Table S1).

Uptake and hydrolysis of AEA and PEA in intact T84 cells

A time‐dependent accumulation of both [Ara‐³H]AEA and [Pa‐³H]PEA was seen (Figure 2A, B). In both cases, the accumulation was significantly reduced by the selective FAAH inhibitor URB597 (1 μM, Kathuria et al., 2003), but not by the NAAA inhibitor diacerein (50 μM, Petrosino et al., 2015). The reduction produced by URB597 is consistent with published data from other cell lines, where FAAH controls the rate of uptake by preserving the relative concentrations of extra‐ and intracellular NAE (Deutsch et al., 2001). Consistent with the uptake experiments, the cells were able to hydrolyse both [Et‐³H]AEA and [Et‐³H]PEA (Figure 2C) in a manner sensitive to inhibition by URB597 but not diacerein. PEA is a very good substrate for NAAA (Ueda et al., 2001), and so the lack of effect of diacerein is perhaps surprising. In consequence, we tested a second NAAA inhibitor, pentadecylamine (30 μM, Yamano et al., 2012). The compound did not inhibit the hydrolysis of [Et‐³H]AEA but produced a small (~25%) inhibition of the hydrolysis of [Et‐³H]PEA (Figure 2D). The hydrolysis of [Et‐³H]AEA and [Et‐³H]PEA in T84 cells was not significantly affected by the 24 h treatment with IFNγ (10 ng·mL⁻¹) (Figure 2E).

Figure 2.

Uptake (panels A and B) and hydrolysis (panels C–E) of AEA and PEA in T84 cells cultured in 24‐well plates. In panels A and B, the data are scatterplots, N = 6, with the bars representing the means, of the cellular retention of 104 nM [Ara‐³H]AEA (panel A) and [Pa‐³H]PEA (panel B) at the incubation times shown. Veh, vehicle; URB1, 1 μM URB597; D50, 50 μM diacerin; U + D, combination of the two compounds. For the values excepting the data for the wells alone, three‐way ANOVA using permutation probabilities (10⁶ iterations) and the function aovp in the package lm Perm version 2.1.0 for R gave P <0.05 for AEA uptake: incubation time, URB597, incubation time × URB597; PEA uptake: incubation time, URB597, incubation time × URB597; all other P values were >0.05. Panels C (N = 5) and D (N = 7; P30, 30 μM pentadecylamine) show hydrolysis as % of control following 15 min incubation with 104 nM substrate. In panel D, the 95% confidence interval for P30 was 86–118 for AEA and 58–95 for PEA. Panel E shows the hydrolysis of AEA (N = 8) and PEA (N = 6) in cells treated for 24 h with IFNγ (10 ng·mL⁻¹) prior to assay. The values are expressed as % of the corresponding hydrolysis seen in control samples assayed concomitantly. In both cases, the 95% confidence intervals straddled 100% (88–122% for AEA; 87–110% for PEA).

Effects of URB597 and pentadecylamine upon 2‐AG and NAE levels in IFNγ‐treated T84 cells

The data above indicate that T84 cells express functionally active FAAH and NAAA, but that FAAH rather than NAAA is predominantly responsible for the hydrolysis of exogenous AEA and PEA in the intact cells. In order to determine whether the same is true for endogenous AEA and PEA, IFNγ‐treated T84 cells in six‐well plates were treated with either vehicle, 1 μM URB597 or 30 μM pentadecylamine for 2 h prior to collection of cells, extraction of lipids and measurement of 2‐AG and NAE levels. The data are shown in Figure 3. URB597 treatment increased levels of all NAEs, but not of 2‐AG. Small, but significant, increases in NAE levels were also seen after pentadecylamine treatment of the cells, with the exception of LEA.

Figure 3.

Levels of 2‐AG and NAEs in T84 extracts following treatment with IFNγ (10 ng·mL⁻¹, 24 h) and either vehicle (V), 1 μM URB597 (U) or 30 μM pentadecylamine (P) for the last 2 h. Shown are scatterplots with the bars representing means. The experiment was designed so that all groups were n = 12, but some values were lost for technical reasons, primarily incomplete evaporation of the methanol in the samples prior to analysis. These samples were discarded, to leave n = 10, 9 and 8 for vehicle, URB597 and pentadecylamine respectively. Data points shown as triangles were identified as possible outliers but included in the statistical analyses. There were no probable outliers. For all NAEs, a one‐way permutation ANOVA (function independence test in the coin package for R) with 10⁶ iterations gave P values <0.05, allowing post hoc bivariate examinations of the individual groups. *P < 0.05; NS, not significant: two‐tailed permutation tests (complete enumeration; function permTS in the perm package for R; Bonferroni corrections). In the case of 2‐AG, the permutation ANOVA was not significant (P > 0.05).

Effects of IFNγ treatment upon barrier properties and mRNA levels of NAE and 2‐AG metabolic enzymes for T84 cell monolayers

T84 cells, when cultured to confluence as monolayers, differentiate spontaneously, develop tight junctions on the apical side and have excellent barrier functions that can be quantitated by measurement of the TEER. Treatment of these monolayers with inflammatory cytokines such as IFNγ affects both the structure and the function of these cells, resulting in an increased permeability of the monolayers visible by a reduction in the TEER (Zolotarevsky et al., 2002; Bruewer et al., 2005). Here, T84 cells in inserts (TEER >900 Ω.cm⁻²) were treated with vehicle or IFNγ (10 ng·mL⁻¹, basolaterally) for 24 h, after which the TEER was again measured. At t0, the TEER values were 1159 ± 67 and 1232 ± 79 Ω.cm² for controls and IFNγ‐treated cells respectively (means ±95% confidence intervals, n = 12). At t24, the corresponding values were 1177 ± 73 and 1031 ± 47 Ω.cm² respectively. A linear mixed model matching for time gave a significant interaction time × treatment, due to the difference between the t24 values for control and IFNγ‐treated cells (P<0.05, permutation test). The corresponding P value at t0 was not significant.

At t24, cell lysates were collected for qPCR measurement of enzymes related to endocannabinoid and NAE synthesis and degradation with RPL19 as reference gene. The data are shown in Figure 4. Note that in the figure, the data are expressed as the untransformed ∆Ct, and a change in −1 unit reflects a doubling in the mRNA level. There were no significant differences between control and IFNγ‐treated cell monolayers with respect to the expression levels of the NAE synthetic enzyme NAPE‐PLD and the catabolic enzymes FAAH and NAAA. The data with FAAH and NAAA are consistent with the unchanged rates of hydrolysis seen following IFNγ treatment of the T84 cells in 24‐well plates (Figure 1E). 2‐AG is synthesized primarily by DAG lipases A and B and hydrolysed by MAGL and to a lesser extent by αβ hydrolase domain (ABHD) proteins 6 and 12 (Ueda et al., 2013). The mRNA levels of DAGLA and ABHD12 were not significantly affected by IFNγ‐treatment. However, a robust decrease in mRNA for MAGL was seen (Figure 3). The difference in the mean ∆Ct values was 1.51, which corresponds to a ~65% reduction in mRNA levels as a result of the IFNγ‐treatment.

Figure 4.

Levels of mRNA for NAE and 2‐AG synthetic and catabolic enzymes in T84 cell monolayers cultured on inserts after treatment for 24 h with IFNγ (10 ng·mL^‐1). The individual ΔCt values (N=9) are shown, with the means indicated by the bars. *Indicates a significant effect of IFNγ, calculated using a two‐tailed permutation test (complete enumeration; function permTS in the perm package for R) upon implementation of a 5% false discovery rate. The unadjusted P values are given in Table S2, Supporting information to allow readers who prefer other ways of correcting for multiple comparisons to be able so to do.

Effect of URB597, pentadecylamine and PEA upon barrier deficits produced by IFNγ in T84 cells

Given that IFNγ treatment mobilizes NAEs (Figure 1) and that the NAEs AEA and PEA can influence TEER in Caco‐2 human colon carcinoma cells (Alhamoruni et al., 2010; Karwad et al., 2017), we hypothesized that inhibition of FAAH and/or NAAA could be sufficient to affect the barrier deficit induced by IFNγ. In addition to the control and IFNγ‐treated T84 cells described above, TEER was measured initially and after 24 h of incubation with T84 cells in inserts treated either with vehicle or with IFNγ (10 ng·mL⁻¹, basolaterally) with and without 1 μM URB597 (apically), 30 μM pentadecylamine (apically) or PEA (10 μM apically or basolaterally).

The data obtained at t24, expressed for simplicity as a percentage of the corresponding value at t0, are shown in Figure 5. The effect of IFNγ was a clear increase in permeability, which was not significantly changed by any of the additional treatments. It can be argued that modest changes in TEER produced by these treatments might be hidden within the variability of the response to IFNγ. This issue can be avoided by calculating the data as % reversal of the change in TEER (i.e. 100‐100×[∆IFN + treatment‐∆control]/[∆IFN‐∆control], where ∆ refers to the difference in Ω.cm² between t0 and t24 for each condition) which removes the variability of the response to IFNγ. In this case, the mean % reversals (with 95% confidence intervals and N in brackets) were as follows: URB597–10 (−35 to +15, n = 12); pentadecylamine 4 (−24 to +32, n = 12). In contrast, addition of PEA apically produced a partial reversal [+36 (+6 to +66, n = 13)], whereas the compound given basolaterally did not [+14 (−33 to +61, n = 8)]. Karwad et al. (2017) have also reported beneficial effects of PEA upon the TEER of Caco‐2 human carcinoma cells following treatment with IFNγ and tumour necrosis factor α, although in this case, the effect was for basolateral rather than apical treatment of the cell.

Figure 5.

Effects of URB597 (1 μM), pentadecylamine (30 μM) and PEA (10 μM) upon the permeability of IFNγ‐treated T84 cells. IFNγ (10 ng·mL^‐1) was added to the basolateral (“bl”) side and the test compounds were added to the apical (“ap”) side unless otherwise shown. TEER data at t=24 h are expressed as % of the corresponding t=0 value. The dotted lines indicate the mean values for C (control, i.e. no IFNγ) and V (vehicle, i.e. treatment with IFNγ alone) samples. Comparing V with C, a two‐sided permutation test with 10⁶ iterations (function permTS in the package perm for R) gave a P value <0.05. For the (IFNγ + test compound)‐treated samples, a permutation ANOVA (function independence_test in the coin package for R) with 10⁶ iterations gave a non‐significant P value. However, a significant effect of PEA, when given apically, was seen when the data were expressed as % reversal of the change in TEER (P<0.05, two‐tailed one sample t‐test vs a value of 0%, i.e. no reversal; for details, see Results section).

Discussion

The present study has investigated the effects of IFNγ upon endocannabinoid levels and barrier properties of T84 cells. The main findings are that the cytokine increases the levels of the endocannabinoids and related NAEs; that both AEA and PEA, when given exogenously, are predominantly metabolized by FAAH rather than NAAA; but that inhibition of these enzymes does not block the reduction of TEER produced by IFNγ treatment. These data are considered in turn.

With respect to the effects upon AEA and 2‐AG, our data are to our knowledge the first report using colorectal cancer cells, although there is a very recent report showing that treatment of human Caco‐2 colorectal cells with the combination of IFNγ (24 h) and TNF‐α (16 h starting at t = 8 h) produced an approximate doubling of PEA and OEA levels (Karwad et al., 2017). Incubation with the calcium ionophore ionomycin further increased the levels of 2‐AG, SEA, EPEA and POEA but not of AEA and PEA. At first sight, the data with AEA and PEA are perhaps surprising, given that the initial enzymes in the NAE synthetic pathway are activated by calcium. However, there are several NAE synthetic pathways (Ueda et al., 2013). In this respect, Leung et al. (2006) reported that the genetic deletion of NAPE‐PLD reduced SEA, OEA and PEA levels in the brain, whereas AEA levels were unchanged. Conversely, AEA and other NAE levels in mouse peritoneal exudate cells (>80% macrophages) are not affected by incubation with the calcium ionophore A23187 (Kuwae et al., 1999). Thus, the present data suggest that in T84 cells, AEA and PEA levels are increased by IFNγ primarily by effects upon calcium‐independent pathways.

A related question is how IFNγ treatment increases NAE and 2‐AG levels. In the case of 2‐AG, we found a robust reduction in the mRNA of MAGL, which, if translated into a reduction in activity, would certainly explain the increase in 2‐AG levels following cytokine treatment. In contrast, NAAA and FAAH levels and combined hydrolytic activity in intact cells were unaffected by the treatment, the latter finding in contrast to the situation in lymphocytes, where IFNγ treatment reduces to about half the FAAH activity (Maccarrone et al., 2001). Our findings suggest that IFNγ is most likely to influence the synthesis rate, rather than the hydrolysis rate of AEA. This may be due to an up‐regulation in a synthesis protein other than NAPE‐PLD, but an alternative without alteration in enzyme expression would be to increase the precursor levels. Indeed, treatment with IFNγ + TNFα produces marked changes in the fatty acid composition of phosphatidylcholines and phosphatidylethanolamines in membrane fractions from tight junction regions of T84 cells (Li et al., 2008). Whether or not such changes are sufficient to account for the effects of IFNγ upon NAE levels awaits elucidation.

The second novel finding is that although T84 cells express both FAAH and NAAA, FAAH is primarily responsible for the hydrolysis of exogenous AEA and PEA. The result with AEA is expected, but PEA has a high affinity for NAAA (Ueda et al., 2001), and treatment of mice with an NAAA‐selective inhibitor increases PEA levels in the colon (Alhouayek et al., 2015). NAAA is primarily located in lysosomes, and it is possible that extracellular PEA, after accumulation by the cells, is not partitioned into this organelle. However, although endogenous PEA (and other NAE, except LEA) levels in IFNγ‐treated cells are significantly increased by the NAAA inhibitor pentadecylamine, the increases are smaller than those found with the FAAH inhibitor URB597, which would suggest that the latter enzyme is a more important determinant of NAE levels in these cells.

The third novel finding concerns the effects of FAAH and NAAA inhibitors, or rather lack thereof, upon the reduction in TEER produced by IFNγ. No significant effect was seen, in contrast to the partial reduction produced by apically administered PEA. As mentioned in the Introduction, FAAH and NAAA inhibitors produce beneficial effects in vivo in models of colitis (Storr et al., 2008; Sałaga et al., 2014; Alhouayek et al., 2015), but the end points in these studies, such as macroscopic changes, inflammatory cell infiltration and disruption of intestinal cell motility, give no information as to permeability barrier changes. T84 cells resemble adult colonic crypt cells (Dharmsathaphorn et al., 1984; Madara et al., 1987), and so our data, if extrapolated to the situation in vivo, would suggest that effects upon the permeability barrier provided by these cells do not contribute to the beneficial effects of FAAH and NAAA inhibitors in animal models of colitis.

Author contributions

M.A. and C.J.F conceived and designed the experiments and analysed the data. M.A., L.R. and S.G.‐F. performed the experiments. C.J.F. wrote the paper.

Conflict of interest

The authors declare no conflicts of interest.

Declaration of transparency and scientific rigour

This Declaration acknowledges that this paper adheres to the principles for transparent reporting and scientific rigour of preclinical research recommended by funding agencies, publishers and other organisations engaged with supporting research.

Supporting information

Table S1 P values for the data shown in Fig. 1.

Table S2 Mean, SD, % of control and P values for the data shown in Fig. 4.

Alhouayek, M. , Rankin, L. , Gouveia‐Figueira, S. , and Fowler, C. J. (2019) Interferon γ treatment increases endocannabinoid and related N‐acylethanolamine levels in T84 human colon carcinoma cells. British Journal of Pharmacology, 176: 1470–1480. 10.1111/bph.14135.