The Endocannabinoid, Anandamide, Induces Cannabinoid Receptor-Independent Cell Death in Renal Proximal Tubule Cells

Background: The endocannabinoid (EC) system is well characterized in the central nervous system but scarcely studied in peripheral organs. In this paper, we newly identify the effect of the EC anandamide (AEA) upon renal proximal tubule cells. Methods: Measurement of lactate dehydrogenase (LDH) release after treatment of primary renal proximal tubule cells (RPTEC) and renal carcinoma cell line (Caki-1) with AEA, arachidonic acid (AA), ethanolamide (EtAm), EC receptor CB1 antagonist (AM251), CB2 receptor antagonist (SR144528), TRPV1 receptor antagonist (capsazepine), degradation enzyme fatty acid amide hydrolase (FAAH) antagonist (URB597), antioxidants GSH-EE; Trolox, GSH depletor BSO, membrane cholesterol depletor (MCD), apoptosis inhibitor zVAD, necroptosis inhibitor Nec-1 or ferroptosis inhibitor Fer-1. Western blot and qRT-PCR analysis plus determination of reactive oxygen species (ROS) via H2-DCFDA were performed. Histology for EC enzymes, N-acetylphosphatidylethanolamine-hydrolyzing phospholipase D (NAPE-PLD) and FAAH, as well as the determination of physiological levels of ECs in human and rat renal tissue via liquid chromatography were conducted. Results: AEA both dose- and time-dependently induces cell death in RPTEC and Caki-1 within hours, characterized by cell blebbing, not influenced by blocking the described EC receptors by AM251, SR144528, capsazepine or FAAH by URB597 or MCD. Cell death is mediated via ROS. There is no difference found in the histology of the enzymes FAAH and NAPE-PLD in human renal tissue with interstitial nephritis. Blocking of apoptotic, necroptotic or ferroptotic cell death does not lead to a reduction in LDH release in vitro. Conclusion: The endocannabinoid anandamide induces cell death in renal proximal tubule cell in a time- and dose-dependent manner. This pathway is mediated via ROS and is independent of cannabinoid receptors, membrane cholesterol or FAAH activity.


Introduction
Endocannabinoids (ECs) are endogenous lipids that have been gaining more interest during the last years as the exogenous equivalent is known as Δ 9 -THC (Δ 9 -tetrahydrocannabinol), the active psychotropic natural product of Cannabis [1].
The EC system is extensively studied in immune system function and the central nervous system, as it is involved in memory processing, regulation of pain, neuroprotection, addiction and appetite [2] [3] [4]. In peripheral organs like the liver, the endocannabinoid system contributes to cell death mechanisms that involve cyclooxygenase 2 (COX2) [5], as well as hepatic injury and fibrogenesis [6] [7]. The EC anandamide (AEA) can reach an intracellular level of up to 50 µM upon proinflammatory activation of cells [8] [9].
In renal tissue, the EC system has gained more interest during the last years and a role for the cannabinoid receptor 1 (CB1) in pathogenesis of interstitial fibrosis has been proposed, while the cannabinoid receptor 2 (CB2) has been regarded as protective in chronic renal disease [10] [11]. Ritter and colleagues proposed in 2016 the necessity of further studies to define and exploit the role of the EC system in the kidney [12]. In this study, we considered renal disease models resulting in cell loss, like nephritis or ischemic damage during kidney transplantation as well as a renal cell carcinoma (RCC). Understanding those mechanisms resulting in cell loss and the possible involvement of the EC system is an important step towards understanding disease entities.
For a long time, apoptosis or necrosis were the only forms of cell death known.
Currently, processes not classified as apoptosis or necrosis are referred to as regulated necrosis [13]. The identification of several other pathways leading to cell death, besides apoptosis and necrosis, has evolved during the last decade and is still growing. The recently identified cell death pathways in renal tissue are apoptosis, necroptosis and ferroptosis [14] [15] [16].
Apoptosis is cell death mainly regulated via caspases that determine cellular fate and can be subdivided into an intrinsic (regulated via cellular stress stimuli) and ex- tumor necrosis factor receptor (TNFR) or TNF related apoptosis inducing ligand-receptor (Trail-R)) [17]. zVAD (z-Val-Ala-Asp-(OMe)-Fluoromethyl Ketone) is a selective caspase inhibitor and, therefore, used to block apoptotic cell death.
Ferroptosis is associated with: ROS accumulation, activation of mitogen activated protein kinases (MAPK), release of arachidonic acid metabolites, glutathione (GSH) dependence as well as small mitochondria [20], and can be blocked by Ferrostatin-1 (Fer-1) through the reduction of lipid oxidation [15].
Hepatic cell death upon anandamide (AEA) treatment has been shown to be cannabinoid receptor-independent, as G protein-coupled receptors CB1, CB2 and transient receptor potential vanilloid receptor 1 (TRPV1) are not involved [21].
In this manuscript, we identify the effect of the endocannabinoid AEA upon renal proximal tubule cells.

Cell Death Determination
Detection of cell death was performed using the cytotoxicity detection kit (Roche Applied Sciences, Mannheim, Germany) to measure lactate dehydrogenase (LDH) release. 10 µL of cell culture free medium were added to 90 µL of LDH reaction mixture. After an incubation time of 20 minutes in the dark, the assay was measured at 490 nm on a 96 well microtiter plate with the help of an ELISA reader (SPECTROstar Omega, BMG Labtech, Ortenberg, Germany). As a negative control culture medium was used-The negative control was acquired by adding 1 µl TritonX to the cells, thereby, causing cell death and subsequent rupture of cellular membranes, therefore representing the 100% LDH release control.

Experimental Kidney Transplantation
Allogeneic renal transplantation was performed in male Dark Agouti (DA (RT1 av1 )) to Lewis (LEW (RT1 l )) rat strain combination at an age of about 2 month, while transplants between Lewis rats (isotransplantation) and non-transplanted Lewis rats served as control [23]. Recipients of isografts were sacrificed 10 days after transplantation; recipients of allografts were sacrificed 6 to 8 days after transplantation. Kidneys were removed and snap-frozen followed by HPLC-MS/MS analysis. Range of morphological alterations in allo-and isotransplanted kidneys were equivalent to those published before [23].
Animal experiments were approved, according to the German national laws

Human Biopsies
Formalin-fixed and paraffinized human renal biopsies were selected from the archives of the Department of Pathology, University Hospital of Cologne, Germany. Histologic evaluation was based on independent analyses by two staff pa-

Liquid Chromatography
Liquid chromatography was performed as previously published [25].

Reactive Oxygen Species (ROS)
Detection of ROS was performed as previously described [26].

Statistical Analysis
All data represents the mean of three independent experiments. Densitometric analysis was performed with the ImageJ program, according to the manufacturer's protocols, to quantify Western blots. GraphPrism 5 (GraphPad Software, La Jolla, Calif., USA) was used to calculate statistical significances with the Student's unpaired t-test (*p < 0.05, **p < 0.01, ***p < 0.001).

AEA Mediates Cell Death in RPTEC and Caki-1 in a Dose Dependent Manner
It is known that AEA can induce cell death in hepatic stellate cells [21]. Caki-1 PRTEC (EtAm) and arachidonic acid (AA) were used for treatment as well to prove that the degradation products of AEA are not toxic to the cells, as it is known from literature that free fatty acids can present a source of cellular toxicity [30] [31] [32]. RPTEC have a higher sensitivity against AEA-induced cell death as the dose-dependent LDH release is already visible after 2 h of treatment. EtAm as degradation product of AEA, as well as ethanol (untreated vehicle control), did not cause any increase in LDH release. We observed an increase in LDH release after treating RPTEC with 25 µM AA up to 30% compared to positive control TritonX. RPTEC, as primary cells, seem to have a higher susceptibility for LDH loss after AA treatment. Compared to AEA treatment, LDH release does not increase over time but stays constant at approximately 30% (data not shown).
To further confirm the findings obtained with LDH release and to represent the findings with another method, time-lapse videos were produced during 4 h of treatment with 70 µM AEA (Figures 1(e)-(h)) in both Caki-1 and RPTEC. The characteristically morphological blebbing of the cells indicates a non-apoptotic type of cell death (Figure 1(f), Figure 1(h)).
LDH release and morphology of Caki-1 and RPTEC led to the conclusion that renal cells undergo cell death upon AEA treatment. As an increase in the concentration of AEA leads to an exponential increase in LDH release, it is not likely that the cell death results from free acid. Additionally, treatment with the degradation products of AEA, namely arachidonic acid and ethanolamide, or ethanol as a vehicle control did not lead to an increase in LDH release above vehicle control levels. An influence upon cell death due to higher free acids or alcohol is not likely to cause increased cell membrane damage.
Since the effect of AEA in the following experiments should result in clear readout of LDH release, the concentration of 70 µM AEA was used for all other experiments as it obtained the best signal-to-noise ratio. The LDH release amounts to approximately 70% and therefore represents an adequate level of cellular leakage. Additionally, higher doses of AEA can be found in proinflammatory processes within organs and are pathologically possible.

Cannabinoid Receptors in Caki-1 and RPTEC
Cannabinoid receptors are expressed in several peripheral organs at the mRNA and protein levels. Here, cannabinoid receptor mRNA and protein expression were investigated in Caki-1 and RPTEC. PCR results lead to the conclusion that Caki-1 and RPTEC express CB1, CB2 and TRPV1 (Figure 2(b)). qRT-PCR results in Caki-1 and RPTEC underline the results of the PCR analysis. There is transcription of the most common cannabinoid receptors CB1, CB2 and TRPV1 (Figure 2(a)).
Western blot analysis revealed that Caki-1 express CB2 irrespective of the treatment with 70 µM AEA for 4 h. An increase in TRPV1 protein was observed after AEA treatment. There is no protein for CB1 detectable (Figure 2

AEA-Induced Cell Death in Caki-1 and RPTEC Is Not Mediated via Cannabinoid Receptors or FAAH
As cannabinoid receptors are expressed by Caki-1 and RPTEC, the dependence of LDH release and cell death was analyzed by selectively blocking: CB1 with AM251, CB2 with SR144528, TRPV1 by capsazepine and the AEA degradation enzyme fatty acid amide hydrolase (FAAH) by URB597 (Figure 3(a), Figure   3(b)).
LDH release is significantly increased in cells pretreated with the CB1 blocker (AM251) or the CB2 blocker (SR144528) and followed by subsequent AEA treatment in Caki-1 cells. Blocking of TRPV1 (capsazepine) and FAAH (URB597) did not change the LDH release as compared to AEA treatment alone ( Figure   3(a)). LDH release in RPTEC is not altered when CB1 is blocked by AEA treat- ment. The application of inhibitors of CB2, TRPV1, and FAAH after AEA treatment leads to a significant increase in LDH release (Figure 3(b)).
AEA-induced cell death seems to be independent of the known cannabinoid receptors and FAAH. It is possible that the increase in LDH release and therefore the cell death induced by the inhibitors alone can be explained by the protective functions of cannabinoid receptors and by the role of FAAH in the metabolism of arachidonic acid.

AEA-Induced Cell Death Depends on ROS but Not on Membrane Cholesterol
As cannabinoid receptors and FAAH do not participate in cell death processes in  (Figure 4(a)).
LDH release by RPTEC after GSH-EE treatment is slightly increased to approximately 37%. Pretreatment of RPTEC with GSH-EE or BSO and AEA over 4 h lead to a significant increase of LDH release. Treating RPTEC with Trolox and 70 µM AEA for 4 h significantly reduced the LDH release towards the control level of LDH release found in untreated RPTEC (Figure 4(b)).
To verify these results, we used H 2 DCFDA, a well-known method to visualize ROS formation within the cytoplasm by fluorescence microscopy. We performed H 2 DCFDA in RPTEC (Figure 4(c), Figure 5). Results show that there is a significant increase of H 2 DCFDA signal after treatment of RPTEC with Trolox and 70 µM AEA over 4 h (Figure 4(c)). Additionally, the H 2 DCFDA signal significantly decreases after treatment with BSO and 70 µM AEA over 4 h, which reflects the results of LDH release. Pretreatment with GSH-EE in RPTEC did not lead to any significant changes of the H 2 DCFDA signal.

AEA-Induced Cell Death in RPTEC Cannot Be Blocked by Common Cell Death Inhibitors
To characterize the dominant type of cell death in RPTEC, apoptosis was blocked by zVAD, whereas necroptotic cell death was prevented by Necrostatin-1 (Nec-1) and ferroptosis by Ferrostatin-1 (Fer-1) (Figure 3(c)). Blocking of three known cell death pathways by zVAD, Nec-1, and Fer-1 did not lead to a significant reduction of the LDH release in RPTECs. Using a combination of all

Acute Renal Injury in Humans
The endocannabinoid system is a complex metabolic system and the degradation of lipids, like AEA, is strictly regulated. Therefore, it is important to determine the enzyme production of both synthesis and degradation by NAPE-PLD and FAAH in the processing of AEA.
It is difficult to extrapolate systemic physiological effects of ECs by cell culture studies. To analyze the potential pathological importance of this system in diseases with acute renal injury, these two enzymes were studied in two different renal disease entities. Using human renal biopsies, control kidneys (Suppl. Figure 1(a)-(e)) were compared to those from patients with interstitial nephritis associated with tubulitis, interstitial fibrosis, and tubular atrophy (Suppl. Figure 1(f)-(j)). Both enzymes were produced in renal proximal tubuli but not in distal tubuli. However, their level remained unchanged in interstitial nephritis vs. the control kidneys.

Discussion
This study newly describes the effect of increased levels of the endocannabinoid AEA upon renal proximal tubule cells during proinflammatory processes. Caki-1 and RPTEC were chosen for two reasons: i) as cell culture models to represent renal proximal tubule cells, and ii) to use primary cells that represent a more physiological representation of renal morphology. The endocannabinoid AEA can be measured under physiological conditions in human and rat renal tissue (Suppl. Figure 2). As the EC system is present in renal proximal tubule cells, pathological increase of AEA during proinflammatory processes is likely. In comparison to other organs, like the liver, kidney cells represent a vast variety of different cell types therefore representing one of the most complex organs in mammals. A homogenous culture model has the advantage of representing one cell type without an effect of dilution due to the different cell regulatory mechanisms of numerous cell types. Additionally, it is possible to vary cell culture treatments in different concentrations (like the one used in vitro concentration of AEA) to ensure an optimal signal-to-noise ratio without fearing an increased systemic effect on the respective treatments.
Compared to hepatic stellate cells, in which AEA induces necrosis [21], renal cells undergo cell death via another, more complex mechanism. This cell death is dose-dependent and morphologically defined by cell blebbing, which is highly characteristic for non-apoptotic cell death. In contrast, cell budding is a characteristic feature of apoptotic cell death [17]. Cell blebbing is a feature of non-apoptotic cell death. The influence of Trolox within the cell and the opposite effects of GSH-EE and BSO, respectively, on LDH release and rapid ROS formation is an additional aspect pointing towards ferroptotic cell death [33].
As seen in Figure 4, Trolox and AEA treated RPTEC show a decrease in LDH release but an increase in H 2 DCFDA as compared to AEA treatment alone. According to literature in previous studies in hepatic stellate cells and AEA [21], we should expect a decrease in both LDH release and H 2 DCFDA after treatment of Trolox and AEA. Nevertheless, there seems to be some compensatory effect.
Even though ROS increases after Trolox and AEA treatment, LDH leakage significantly decreases. Caki-1 cells, as a cell line with RCC background, show a different behavior towards ROS modulation. There is no significant decrease in LDH release measureable after treatment of BSO; GSH-EE or Trolox together with AEA. Caki-1 cells seem to be more adept to deal with cellular stress presented by ROS.
To further extend our understanding of the potential the role of this system in acute renal injury, we analyzed the expression level of NAPE-PLD and FAAH in two renal disease entities (Suppl. Figure 1). Surprisingly, we did not find any significant changes between the control tissues and the diseased kidneys despite inflammatory alterations in interstitial nephritis or immunologic destruction of proximal tubuli in rat renal transplants, which is in contrast to the aforementioned findings after toxic [15] or metabolic [34] tubular injury.
Martin-Sanchez and colleagues verify ferroptosis as the dominant cell death mechanism in acute kidney injury [15]. Acute injury occurring in diabetic nephropathy with subsequent tubular inflammation can lead to a local upregulation of the endocannabinoid system [34]. According to the acquired results, our data show a non-apoptotic type of cell death which cannot completely be characterized with the help of the methods utilized. The absence of LDH reduction after blocking LDH release with common cell death blockers against apoptosis, necroptosis, and ferroptosis does not necessarily lead to rejection of the evidence towards the analyzed cell death mechanisms. This could be due to the fact that the process of, for example, ferroptotic cell death might be regulated downstream of the effect of Fer-1 [33].
AEA induces cell death in a dose-and time-dependent manner in renal proximal tubule cells. An effect in vitro can be measured at concentrations above 25 µM. Select 70 µM AEA with an LDH release of 70% is especially important in experimental setups modifying ROS or blocking cell death receptors. As the baseline LDH release is on average at 12% (which is expected for adherent growing cells) selecting AEA levels resulting in an LDH release in between positive and negative control and biological levels of AEA are increased in proinflammatory processes. This effect is not mediated via cannabinoid receptors, FAAH or membrane cholesterol but is mediated via ROS. AEA-induced cell death leads to a blebbing of cells and additionally to ROS activation, so that ferroptotic cell death is most likely. Furthermore, this mechanism is not a general response to acute tubular injury but seems to be activated only in very specific tubular damage such as inflicted by specific toxins or metabolic alterations.
In summary, the endocannabinoid AEA induces cannabinoid receptor-and membrane cholesterol-independent cell death in renal proximal tubule cells via ROS. While AEA can induce cell death, the detailed mechanism behind this process has to be further evaluated in the future using newly developed blockers and methods to differentiate non-apoptotic cell death mechanisms.