*Article* **First Evidence of the Protective Effects of 2-Pentadecyl-2-Oxazoline (PEA-OXA) in In Vitro Models of Acute Lung Injury**

**Aniello Schiano Moriello 1,2, Fiorentina Roviezzo <sup>3</sup> , Fabio Arturo Iannotti <sup>1</sup> , Giuseppina Rea <sup>4</sup> , Marco Allarà 1,2 , Rosa Camerlingo <sup>5</sup> , Roberta Verde <sup>1</sup> , Vincenzo Di Marzo 1,6,\* and Stefania Petrosino 1,2,\***


**Abstract:** Acute respiratory distress syndrome (ARDS) is a serious inflammatory lung disorder and a complication of SARS-CoV-2 infection. In patients with severe SARS-CoV-2 infection, the transition to ARDS is principally due to the occurrence of a cytokine storm and an exacerbated inflammatory response. The effectiveness of ultra-micronized palmitoylethanolamide (PEA-um) during the earliest stage of COVID-19 has already been suggested. In this study, we evaluated its protective effects as well as the effectiveness of its congener, 2-pentadecyl-2-oxazoline (PEA-OXA), using in vitro models of acute lung injury. In detail, human lung epithelial cells (A549) activated by polyinosinic–polycytidylic acid (poly-(I:C)) or Transforming Growth Factor-beta (TGF-β) were treated with PEA-OXA or PEA. The release of IL-6 and the appearance of Epithelial–Mesenchymal Transition (EMT) were measured by ELISA and immunofluorescence assays, respectively. A possible mechanism of action for PEA-OXA and PEA was also investigated. Our results showed that both PEA-OXA and PEA were able to counteract poly-(I:C)-induced IL-6 release, as well as to revert TGF-β-induced EMT. In addition, PEA was able to produce an "entourage" effect on the levels of the two endocannabinoids AEA and 2-AG, while PEA-OXA only increased PEA endogenous levels, in poly-(I:C)-stimulated A549 cells. These results evidence for the first time the superiority of PEA-OXA over PEA in exerting protective effects and point to PEA-OXA as a new promising candidate in the management of acute lung injury.

**Keywords:** acute respiratory distress syndrome; anti-inflammatory; endocannabinoids; fibrosis; lung epithelial cells; palmitoylethanolamide; 2-pentadecyl-2-oxazoline

### **1. Introduction**

Acute respiratory distress syndrome (ARDS) is the most devastating condition of acute lung injury, characterized by pulmonary edema, severe hypoxemia and impaired ability to eliminate CO2, and represents one among the most challenging clinical disorders of critical care medicine with high mortality [1–3]. A wide spectrum of risk factors associated with ARDS, such as pneumonia, bacterial or viral infection, transfusion of blood components, trauma, acute pancreatitis and drug reaction, can cause direct or indirect acute lung injury [2]. Among viral infections the current SARS-CoV-2 pandemic has arisen as a new

**Citation:** Schiano Moriello, A.; Roviezzo, F.; Iannotti, F.A.; Rea, G.; Allarà, M.; Camerlingo, R.; Verde, R.; Di Marzo, V.; Petrosino, S. First Evidence of the Protective Effects of 2-Pentadecyl-2-Oxazoline (PEA-OXA) in In Vitro Models of Acute Lung Injury. *Biomolecules* **2023**, *13*, 33. https://doi.org/10.3390/biom13010033

Academic Editor: Myron R. Szewczuk

Received: 1 December 2022 Revised: 20 December 2022 Accepted: 21 December 2022 Published: 24 December 2022

**Copyright:** © 2022 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https:// creativecommons.org/licenses/by/ 4.0/).

risk factor in as much as it can affect the lower respiratory tract by causing ARDS [4]. In particular, in the early stage of COVID-19, the first cells infected by SARS-CoV-2 are the nasal ciliated cells [5]. Successively, if innate or adaptive responses are not able to clear the virus, the latter can spread from the nasal cavity to the lung via inhalation and thus infect alveolar cells [6,7], causing a diffuse alveolar damage that might progress to ARDS. Patients with severe COVID-19 exhibit systemic hyper-inflammation characterized by a cytokine storm, including an excessive release of pro-inflammatory mediators, such as interleukin (IL)-1, IL-6, IL-8 and tumor necrosis factor-alpha (TNF-α) [8], which in turn is responsible for ARDS. In addition, in the lungs of patients with severe COVID-19 the expansion of fibroblasts, which determines a degree of fibrosis that increases over the course of the disease [8], was found, causing or worsening lung injury and failure.

Palmitoylethanolamide (PEA), originally identified in egg yolk and subsequently in a wide variety of food sources, is highly recognized as an endogenous bioactive lipid, given its presence in most cells and tissues, from both animals and humans [9]. PEA is synthesized "on demand" in conditions of current or potential damage and is endowed with antiinflammatory, analgesic and neuroprotective properties [10], which are mediated by several molecular and cellular mechanisms. One of these is the Autacoid Local Inflammation Antagonism (ALIA), through which PEA downregulates the degranulation of mast cells [11]. PEA is known to directly activate the peroxisome proliferator-activated receptor- (PPARα) [12] and the orphan G-protein-coupled receptor 55 (GPR55) [13] or to indirectly activate the cannabinoid receptors CB1 and CB2 [14,15] and the transient receptor potential vanilloid 1 (TRPV1) [16–18]. The indirect interaction of PEA with cannabinoid and vanilloid receptors is known as the "entourage effect", since it is due to PEA increasing the levels of the endocannabinoids and endovanilloids, i.e., anandamide (AEA) and 2-arachidonoyl-glycerol (2-AG). The effect depends upon the inhibition or down-regulation of the AEA-hydrolyzing enzyme fatty acid amide hydrolase (FAAH) [14] or the stimulation of the activity of the biosynthesizing enzyme diacylglycerol lipase (DAGL) [15]. PEA tissue concentrations are altered during different neuro-inflammatory disorders, suggesting that (i) increased levels might represent a compensatory mechanism to restore homeostasis, while (ii) decreased levels might contribute to the etiology of the disease [9,10,19]. For these reasons, the exogenous application of PEA could be required to potentiate the endogenous protective mechanisms, when the endogenous production of PEA is insufficient o lacking [20,21].

Recently, a natural congener of PEA, 2-pentadecyl-2-oxazoline (PEA-OXA), identified in both green and roasted coffee beans [22], has been reported to have anti-inflammatory and anti-nociceptive properties in an experimental model of acute inflammatory pain [23], to reduce neuroinflammation in an experimental model of Parkinson's disease [24] and to exert neuroprotective effects in different neuroinflammatory conditions associated with spinal and brain trauma in mice [25]. It has been hypothesized that PEA-OXA could exert its protective role by inhibiting the enzyme responsible for PEA catabolism, *N*-acyl-ethanolamine-hydrolyzing acid amidase (NAAA), resulting in an increase in the endogenous levels of PEA [23].

Therefore, given the ability of PEA and its congener PEA-OXA to exert important antiinflammatory and protective effects, the present work aimed to investigate their potential effectiveness in in vitro models of acute lung injury, reproducing the clinical conditions of SARS-CoV-2 infection. In particular, this study was based on two currently well-established facts, i.e., i) IL-6 is a key contributor of the cytokine storm observed in SARS-CoV-2 infectionassociated hyperinflammation and multiorgan failure [8,26], suggesting that this cytokine is a promising marker and an efficacious therapeutic target for the treatment of COVID-19 [26,27]; and ii) ultramicronized PEA (PEA-um) is a promising adjuvant to be assumed in the earliest stage of COVID-19, since a reduction in the inflammatory state has been demonstrated both in cultured murine alveolar macrophages activated by the SARS-CoV-2 Spike Protein [28] and in a randomized clinical trial [29] carried out with this formulation of PEA.

To achieve our objectives, here we used the human lung epithelial cell line A549 to reproduce: (i) a viral infection resulting in IL-6 cytokine release and (ii) Epithelial–Mesenchymal Transition (EMT) mechanisms resulting in lung fibrosis.

#### **2. Materials and Methods**

All reagents were purchased from Sigma-Aldrich (Milano, Italy) unless otherwise stated. Pentadecyl-2-oxazoline (PEA-OXA) and palmitoylethanolamide in an ultra-micronized formulation (referred to as PEA hereafter) were obtained from Epitech Group SpA (Saccolongo, Padova, Italy). Polyinosinic–polycytidylic acid (poly-(I:C)) was purchased from InvivoGen (Aurogene, Roma, Italy). 50 -Iodoresiniferatoxin (IRTX) and GW6471 were purchased from Tocris Bioscience (Space Import-Export, Milano, Italy). The deuterated standards—[2H]8-AEA, [2H]5-2-AG and [ <sup>2</sup>H]4-PEA—were purchased from Cayman Chemical (Cabru, Arcore, Italy). The human lung epithelial cell line (A549) was purchased from LGC Standards (Milano, Italy). The human IL-6 ELISA Kit and transforming growth factor beta (TGF-β) were purchased from Abcam (Prodotti Gianni, Milano, Italy). Total mRNA was isolated from A549 cells using Trizol Reagent (Thermo Fisher, Milano, Italy) following the manufacturer's instructions. cDNA preparation from RNA was performed using iScript Reverse Transcription enzyme (Biorad, Milano, Italy). Specific primer sequences were designed using Primer3 Software (https://primer3.ut.ee/, accessed on 28 November 2022) and synthetized by Eurofin (Milano, Italy).

#### *2.1. Cell Culture*

A549 cells were grown in Dulbecco's Modified Eagle Medium (DMEM) complemented with penicillin (400 U mL−<sup>1</sup> ), streptomycin (50 mg mL−<sup>1</sup> ) and 10% Fetal Bovine Serum (FBS), in the presence of a 5% CO<sup>2</sup> atmosphere at 37 ◦C, plated on 100 mm diameter Petri dishes.

#### *2.2. Poly-(I:C)-Induced Inflammatory Response in A549 Cells*

A549 cells were plated into 24-well culture dishes at a cell density of 2 <sup>×</sup> <sup>10</sup><sup>5</sup> cells per well for 1 day at 37 ◦C in a 5% CO<sup>2</sup> atmosphere. After 1 day, A549 cells were stimulated with poly-(I:C) (100 µg mL−<sup>1</sup> ) or vehicle (water) and incubated for 6 h at 37 ◦C in a 5% CO<sup>2</sup> atmosphere. Poly-(I:C)-stimulated A549 cells were treated with PEA-OXA (0.1, 1 and 10 µM), PEA (0.1, 1 and 10 µM) or vehicle (dimethyl sulfoxide or methanol, respectively) and incubated for the indicated time. Poly-(I:C)-stimulated A549 cells were also treated with a TRPV1 antagonist, IRTX (0.1 µM), or PPAR-α antagonist, GW6471 (1 µM), in the presence or absence of PEA-OXA (10 µM) or PEA (10 µM) and incubated for the indicated time. After 6 h, the supernatants were collected, and the amounts of produced IL-6 were measured by using a human IL-6 ELISA kit according to the manufacturer's instructions (Abcam) and by using a reader Glomax® Explorer (Promega, Milano, Italy). Data are expressed as picograms per milliliter of IL-6.

### *2.3. RNA Extraction and Quantitative PCR (qPCR)*

Total RNA was isolated from A549 cells using TRIzol Reagent (cat# 15596018, Life Technologies, Milano, Italy) and reacted with DNase-I (cat# AMPD1, Merk, Milano, Italy) for 15 min at room temperature, followed by spectrophotometric quantification. Subsequently, the RNA integrity number (RIN) for each RNA sample was analyzed on an Agilent 2100 bioanalyzer (Roma, Italy). Purified RNA was reverse-transcribed by the use of the iScript cDNA Synthesis Kit (cat# 1708841, Bio-Rad, Milano, Italy). Quantitative PCR (qPCR) was carried out in a real-time PCR system CFX384 (Bio-Rad, Milano, Italy) using the SYBR Green PCR Kit (Cat# 1725274, Bio-Rad for mRNAs) detection technique and specific primer sequences (Table 1).

Quantitative PCR was performed on independent biological samples (*n* = 3). Each sample was amplified simultaneously in quadruplicate in a one-assay run with a nontemplate control blank for each primer pair to control for contamination or primer–dimer formation, and the cycle threshold (Ct) value for each experimental group was determined. The housekeeping gene ribosomal protein S16 was used to normalize the Ct values, using the 2−∆Ct formula.


**Table 1.** List of primer sequences used in the qPCR analysis.

#### *2.4. Quantification by Liquid Chromatography–Atmospheric Pressure Chemical Ionization–Mass Spectrometry (LC-APCI-MS) of the Endogenous AEA, 2-AG and PEA Levels in A549 Cells*

A549 cells, plated in 6-well culture dishes at a cell density of 9 <sup>×</sup> <sup>10</sup><sup>5</sup> cells per well, were stimulated with poly-(I:C) (100 µg mL−<sup>1</sup> ) or vehicle (water) and treated in the presence or absence of PEA-OXA (10 µM), PEA (10 µM) or vehicle (dimethyl sulfoxide and methanol, respectively) and incubated for 6 h at 37 ◦C in a 5% CO<sup>2</sup> atmosphere. After 6 h, cells and supernatants were collected and homogenized in a solution of chloroform/methanol/Tris-HCl 50 mM, pH 7.4 (2:1:1 by vol.) containing 10 pmol of [2H]8-AEA and 50 pmol of [ <sup>2</sup>H]5-2-AG and [2H]4-PEA as internal standards. The lipid-containing organic phase was dried, weighed and pre-purified by open-bed chromatography on silica gel. Fractions derived by eluting the column with a solution of chloroform/methanol (90:10 by vol) were analyzed by LC-APCI-MS by using a Shimadzu (Shimadzu, Kyoto, Japan) High Performance Liquid Chromatography (HPLC) apparatus (LC-10ADVP) coupled with a Shimadzu (LCMS-2020) quadrupole MS via a Shimadzu APCI interface. LC-APCI-MS analyses of AEA, 2-AG and PEA were performed in the selected ion-monitoring (SIM) mode [30,31], using m/z values of 356 and 348 (molecular ion + 1 for deuterated and undeuterated AEA), 384.35 and 379.35 (molecular ion + 1 for deuterated and undeuterated 2-AG), and 304 and 300 (molecular ion + 1 for deuterated and undeuterated PEA). The AEA, 2-AG and PEA levels were determined on the basis of their area ratio with the internal standard signal areas to provide the amounts in pmol mg–1 of the lipid extract.

#### *2.5. TGF-β-Induced Epithelial–Mesenchymal Transition in A549 Cells*

A549 cells, harvested at 80% confluence and plated into 24-well culture dishes, were stimulated with TGF-β1 (2 ng mL−<sup>1</sup> ) or vehicle (PBS) and incubated for 72 h at 37 ◦C in a 5% CO<sup>2</sup> atmosphere. TGF-β-stimulated A549 cells were treated with PEA-OXA (10 µM), PEA (10 µM) or vehicle (dimethyl sulfoxide and methanol, respectively) and incubated for the indicated time. After 72 h, A549 cells were fixed with 70% ethanol/0.1% Triton for 30 min at 4 ◦C, treated with 5% BSA for 60 min at room temperature and then stained with primary antibodies, rabbit anti-human cytokeratin (clone ab9377, Abcam) and mouse anti-human vimentin V9 (clone ab8069, Abcam), overnight at 4 ◦C. The secondary antibodies, goat antirabbit Alexa fluor488 (Cell Signaling Technology Danvers, MA, USA) and goat anti-mouse Alexa fluor594 (Cell Signaling Technology Danvers, MA, USA), were incubated for 60 min at 4 ◦C, and DAPI (Sigma, Milano, Italy), used to stain the nucleus, was incubated for 7 min at room temperature. Appropriate isotype controls were used. The images were acquired with a fluorescence microscope (Zeiss, Milano, Italy) and AxionCam MRc5 (Zeiss, Milano, Italy).

#### *2.6. Data Analysis*

Each experiment was performed in at least 3–4 independent biological samples for each group. Data were expressed as means ± standard error of the mean (SEM). Statistical analyses were performed using GraphPad Prism software version 9.0 (GraphPad Software Inc., San Diego, CA, USA). One-way analysis of variance (ANOVA) followed by Tukey's multiple comparison test was used for the analysis. *p* values < 0.05 were considered statistically significant. Figures were generated in GraphPad Prism software version 9.0.

#### **3. Results** *3.1. PEA-OXA and PEA Reduce Poly-(I:C)-Induced Release of Il-6 in Lung Epithelial Cells*

**3. Results**

*2.6. Data Analysis*

#### *3.1. PEA-OXA and PEA Reduce Poly-(I:C)-Induced Release of Il-6 in Lung Epithelial Cells* A549 cells stimulated with poly-(I:C) (100 μg mL-1 for 6 h) and treated with the vehi-

images were acquired with a fluorescence microscope (Zeiss, Milano, Italy) and Axion-

Each experiment was performed in at least 3–4 independent biological samples for each group. Data were expressed as means ± standard error of the mean (SEM). Statistical analyses were performed using GraphPad Prism software version 9.0 (GraphPad Software Inc., San Diego, CA). One-way analysis of variance (ANOVA) followed by Tukey's multiple comparison test was used for the analysis. *p* values < 0.05 were considered statistically significant. Figures were generated in GraphPad Prism software version 9.0.

*Biomolecules* **2022**, *12*, x FOR PEER REVIEW 5 of 13

Cam MRc5 (Zeiss, Milano, Italy).

A549 cells stimulated with poly-(I:C) (100 µg mL−<sup>1</sup> for 6 h) and treated with the vehicle of PEA-OXA or PEA significantly released IL-6, as compared to vehicle-stimulated A549 cells (Figure 1). PEA-OXA and PEA, at the highest concentration tested (10 µM), reduced the release of IL-6 from poly-(I:C)-stimulated A549 cells, as compared to poly-(I:C)-stimulated A549 cells treated with the vehicle of PEA-OXA or PEA (Figure 1). PEA-OXA (10 µM) was more effective (1.6-fold) than PEA (10 µM) in reducing the release of IL-6 from poly-(I:C) stimulated A549 cells (53% and 33% of inhibition, respectively) (Figure 1). No effect on IL-6 release was observed when A549 cells were treated with PEA-OXA or PEA alone (0.1, 1 and 10 µM), i.e., in the absence of poly-(I:C), as compared to vehicle-treated A549 cells (Figure 1). cle of PEA-OXA or PEA significantly released IL-6, as compared to vehicle-stimulated A549 cells (Figure 1). PEA-OXA and PEA, at the highest concentration tested (10 μM), reduced the release of IL-6 from poly-(I:C)-stimulated A549 cells, as compared to poly- (I:C)-stimulated A549 cells treated with the vehicle of PEA-OXA or PEA (Figure 1). PEA-OXA (10 μM) was more effective (1.6-fold) than PEA (10 μM) in reducing the release of IL-6 from poly-(I:C)-stimulated A549 cells (53% and 33% of inhibition, respectively) (Figure 1). No effect on IL-6 release was observed when A549 cells were treated with PEA-OXA or PEA alone (0.1, 1 and 10 μM), i.e., in the absence of poly-(I:C), as compared to vehicle-treated A549 cells (Figure 1).

**Figure 1.** PEA-OXA and PEA reduce IL-6 release from poly-(I:C)-stimulated A549 cells. (**a**) IL-6 release was measured after stimulation of A549 cells with poly-(I:C) (100 μg mL-1 ) in the presence or absence of PEA-OXA (0.1, 1 and 10 μM) for 6 h at 37 °C in a 5% CO<sup>2</sup> atmosphere; (**b**) IL-6 release was measured after stimulation of A549 cells with poly-(I:C) (100 μg mL-1 ) in the presence or absence of PEA (0.1, 1 and 10 μM) for 6 h at 37 °C in a 5% CO<sup>2</sup> atmosphere. Each bar shows the mean ± SEM of independent experiments (*n* = 4). *p*-values were determined by ANOVA followed by Tukey's multiple comparisons test. \*\*\*\* *p* < 0.0001 and \* *p* < 0.05. **Figure 1.** PEA-OXA and PEA reduce IL-6 release from poly-(I:C)-stimulated A549 cells. (**a**) IL-6 release was measured after stimulation of A549 cells with poly-(I:C) (100 µg mL−<sup>1</sup> ) in the presence or absence of PEA-OXA (0.1, 1 and 10 µM) for 6 h at 37 ◦C in a 5% CO<sup>2</sup> atmosphere; (**b**) IL-6 release was measured after stimulation of A549 cells with poly-(I:C) (100 µg mL−<sup>1</sup> ) in the presence or absence of PEA (0.1, 1 and 10 µM) for 6 h at 37 ◦C in a 5% CO<sup>2</sup> atmosphere. Each bar shows the mean ± SEM of independent experiments (*n* = 4). *p*-values were determined by ANOVA followed by Tukey's multiple comparisons test. \*\*\*\* *p* < 0.0001 and \* *p* < 0.05.

#### *3.2. The Anti-Inflammatory Effect of PEA-OXA and PEA Is not Reverted by Antagonism at the TRPV1 or PPAR- Receptors in Lung Epithelial Cells 3.2. The Anti-Inflammatory Effect of PEA-OXA and PEA Is Not Reverted by Antagonism at the TRPV1 or PPAR-α Receptors in Lung Epithelial Cells*

In untreated A549 cells, we a found robust mRNA expression of PPAR and a low relative mRNA expression of TRPV1 (Figure 2). Instead, no mRNA expression of CB2 was In untreated A549 cells, we a found robust mRNA expression of PPARα and a low relative mRNA expression of TRPV1 (Figure 2). Instead, no mRNA expression of CB2 was found in untreated A549 cells (Figure 2). Importantly, the highest mRNA expression levels in this cell line were found for NAAA (Figure 2).

When A549 cells were stimulated with poly-(I:C) (100 µg mL−<sup>1</sup> for 6 h) and treated with a selective TRPV1 (IRTX, 0.1 µM) or PPAR-α (GW6471, 1 µM) receptor antagonist, IL-6 release was comparable to that observed in poly-(I:C)-stimulated A549 cells treated with the vehicle (Figure 3a). When poly-(I:C)-stimulated A549 cells were co-treated with PEA-OXA (10 µM) and IRTX (0.1 µM) or GW6271 (1 µM), IL-6 release was comparable to that observed in poly-(I:C)-stimulated A549 cells treated with PEA-OXA (10 µM) (Figure 3a). Likewise, when poly-(I:C)-stimulated A549 cells were co-treated with PEA (10 µM) and IRTX (0.1 µM) or GW6271 (1 µM), IL-6 release was comparable to that observed in poly-(I:C)-stimulated A549 cells treated PEA (10 µM) (Figure 3b). No effect was observed on IL-6 release when A549 cells were treated with the antagonists alone, i.e., in the absence of poly-(I:C), as compared to vehicle-treated A549 cells (Figure 3a).

in this cell line were found for NAAA (Figure 2).

**Figure 2***.* mRNA expression levels of PEA targets (PPAR, TRPV1, CB2) and PEA-catabolizing enzyme (NAAA) in A549 cells. Bar chart with individual points showing the mRNA expression levels of the indicated proteins (PPAR, TRPV1, CB2 and NAAA) measured in A549 cells. Each bar shows the mean ± SEM of 3 independent biological samples. Data are expressed using the 2-ct formula. **Figure 2.** mRNA expression levels of PEA targets (PPARα, TRPV1, CB2) and PEA-catabolizing enzyme (NAAA) in A549 cells. Bar chart with individual points showing the mRNA expression levels of the indicated proteins (PPARα, TRPV1, CB2 and NAAA) measured in A549 cells. Each bar shows the mean <sup>±</sup> SEM of 3 independent biological samples. Data are expressed using the 2−∆ct formula. *Biomolecules* **2022**, *12*, x FOR PEER REVIEW 7 of 13

found in untreated A549 cells (Figure 2). Importantly, the highest mRNA expression levels

**Figure 3.** TRPV1 and PPAR- antagonists do not revert the anti-inflammatory effect of PEA-OXA and PEA in poly-(I:C)-stimulated A549 cells. (**a**) IL-6 release was measured after that A549 cells were stimulated with poly-(I:C) (100 μg mL-1 ) and treated with IRTX (0.1 μM) or GW6471 (1 µM) in the presence or absence of PEA-OXA (10 μM), for 6 h at 37 °C in a 5% CO<sup>2</sup> atmosphere; (**b**) IL-6 release was measured after A549 cells were stimulated with poly-(I:C) (100 μg mL-1 ) and treated with IRTX (0.1 μM) or GW6471 (1 µM) in the presence or absence of PEA (10 μM), for 6 h at 37 °C in a 5% CO<sup>2</sup> atmosphere. Each bar shows the mean ± SEM of independent experiments (*n* = 4). The *p*-values were determined by ANOVA followed by Tukey's multiple comparisons test. \*\*\*\* *p* < 0.0001 and \* *p* < 0.05. **Figure 3.** TRPV1 and PPAR-α antagonists do not revert the anti-inflammatory effect of PEA-OXA and PEA in poly-(I:C)-stimulated A549 cells. (**a**) IL-6 release was measured after that A549 cells were stimulated with poly-(I:C) (100 µg mL−<sup>1</sup> ) and treated with IRTX (0.1 µM) or GW6471 (1 µM) in the presence or absence of PEA-OXA (10 µM), for 6 h at 37 ◦C in a 5% CO<sup>2</sup> atmosphere; (**b**) IL-6 release was measured after A549 cells were stimulated with poly-(I:C) (100 µg mL−<sup>1</sup> ) and treated with IRTX (0.1 µM) or GW6471 (1 µM) in the presence or absence of PEA (10 µM), for 6 h at 37 ◦C in a 5% CO<sup>2</sup> atmosphere. Each bar shows the mean ± SEM of independent experiments (*n* = 4). The *p*-values were determined by ANOVA followed by Tukey's multiple comparisons test. \*\*\*\* *p* < 0.0001 and \* *p* < 0.05.

#### *3.3. Effect of PEA and PEA-OXA Treatment on AEA, 2-AG and PEA Endogenous Levels, in Poly-(I:C)-Stimulated A549 Cells 3.3. Effect of PEA and PEA-OXA Treatment on AEA, 2-AG and PEA Endogenous Levels, in Poly-(I:C)-Stimulated A549 Cells*

When A549 cells were stimulated with poly-(I:C) under the same conditions shown above (100 μg mL-1 for 6 h), the endogenous levels of AEA, 2-AG and PEA did not change, as compared to those in A549 cells stimulated with vehicle (Figure 4). By contrast, when poly-(I:C)-stimulated A549 cells were treated with PEA (10 μM), the endogenous levels of AEA and 2-AG were significantly increased by 4-fold and 1.5-fold, respectively, compared When A549 cells were stimulated with poly-(I:C) under the same conditions shown above (100 µg mL−<sup>1</sup> for 6 h), the endogenous levels of AEA, 2-AG and PEA did not change, as compared to those in A549 cells stimulated with vehicle (Figure 4). By contrast, when poly-(I:C)-stimulated A549 cells were treated with PEA (10 µM), the endogenous levels of AEA and 2-AG were significantly increased by 4-fold and 1.5-fold, respectively, compared

to those in poly-(I:C)-stimulated A549 cells treated with the PEA vehicle (Figure 4b,c). In addition, the levels of PEA were significantly increased by 96-fold when poly-(I:C)-stimulated A549 cells were treated with PEA (10 µM), as compared to those in poly-(I:C)-stim-

PEA were significantly increased by 18-fold when poly-(I:C)-stimulated A549 cells were treated with PEA-OXA (10 µM), as compared to those in poly-(I:C)-stimulated A549 cells treated with the vehicle of PEA-OXA (Figure 4a). No statistically significant variation in the endogenous levels of AEA and 2-AG was observed when poly-(I:C)-stimulated A549 cells were treated with PEA-OXA (10 µM) (Figure 4a), compared to poly-(I:C)-stimulated A549 cells treated with the PEA-OXA vehicle, although a trend towards the enhancement

of the AEA levels was observed (Figure 4a).

to those in poly-(I:C)-stimulated A549 cells treated with the PEA vehicle (Figure 4b,c). In addition, the levels of PEA were significantly increased by 96-fold when poly-(I:C)-stimulated A549 cells were treated with PEA (10 µM), as compared to those in poly-(I:C)-stimulated A549 cells treated with the vehicle of PEA (Figure 4a). The endogenous levels of PEA were significantly increased by 18-fold when poly-(I:C)-stimulated A549 cells were treated with PEA-OXA (10 µM), as compared to those in poly-(I:C)-stimulated A549 cells treated with the vehicle of PEA-OXA (Figure 4a). No statistically significant variation in the endogenous levels of AEA and 2-AG was observed when poly-(I:C)-stimulated A549 cells were treated with PEA-OXA (10 µM) (Figure 4a), compared to poly-(I:C)-stimulated A549 cells treated with the PEA-OXA vehicle, although a trend towards the enhancement of the AEA levels was observed (Figure 4a). *Biomolecules* **2022**, *12*, x FOR PEER REVIEW 8 of 13

**Figure 4.** Variation of the levels of PEA, AEA and 2-AG in poly-(I:C)-stimulated A549 cells treated with PEA and PEA-OXA. PEA (**a**), AEA (**b**) and 2-AG (**c**) levels were quantified by LC–MS; after that, A549 cells were stimulated with poly-(I:C) (100 μg mL-1 ) in the presence or absence of PEA (10 μM) and PEA-OXA (10 μM) for 6 h at 37 °C in a 5% CO<sup>2</sup> atmosphere. Each bar shows the mean ± SEM of independent experiments (*n* = 3). The *p*-values were determined by ANOVA followed by Tukey's multiple comparisons test. \*\*\*\* *p* < 0.0001, \*\*\* *p* < 0.001 and \* *p* < 0.05. **Figure 4.** Variation of the levels of PEA, AEA and 2-AG in poly-(I:C)-stimulated A549 cells treated with PEA and PEA-OXA. PEA (**a**), AEA (**b**) and 2-AG (**c**) levels were quantified by LC–MS; after that, A549 cells were stimulated with poly-(I:C) (100 µg mL−<sup>1</sup> ) in the presence or absence of PEA (10 µM) and PEA-OXA (10 µM) for 6 h at 37 ◦C in a 5% CO<sup>2</sup> atmosphere. Each bar shows the mean ± SEM of independent experiments (*n* = 3). The *p*-values were determined by ANOVA followed by Tukey's multiple comparisons test. \*\*\*\* *p* < 0.0001, \*\*\* *p* < 0.001 and \* *p* < 0.05.

#### *3.4. PEA-OXA and PEA Block TGF--Induced Epithelial–Mesenchymal Transition in Lung Epithelial Cells 3.4. PEA-OXA and PEA Block TGF-β-Induced Epithelial–Mesenchymal Transition in Lung Epithelial Cells*

We previously reported that A549 cells cultured in the absence of TGF-1 maintain a classic epithelial morphology appearing short, spindle-shaped and triangle-shaped, whereas following incubation with TGF-1 (2 ng mL-1) for 72 h, these cells assume an elongated shape, with many cells losing contact with their neighboring cells and displaying a long spindle shape typical of a fibroblast-like morphology, an effect known as Epithelial–Mesenchymal Transition (EMT) [32]. Consistent with these morphological observations, alterations in the expression and distribution of cytokeratin and vimentin were evidenced here by immunofluorescence assays. Indeed, these observations showed that, under the same experimental conditions, TGF-1 induced the down-regulation of cytokeratin (Figure 5a, panels B and T; Figure 5b) and the up-regulation of vimentin (Figure 5a, panels H and T; Figure 5c), when compared to vehicle-incubated A549 cells (Figure 5a, panels A, G and S; Figure 5b,c). Treatment of TGF-1-incubated A549 cells with PEA-OXA (10 µM) or PEA (10 µM) reverted the down-regulation of cytokeratin (Figure 5a, panels C, U, D and V; Figure 5b) and the up-regulation of vimentin (Figure 5a, panels I, U, J and V; Figure 5c). No cell alteration in the expression and distribution of cytokeratin and vimentin was evidenced by immunofluorescence assay when A549 cells were treated with PEA-OXA (10 µM) alone (Figure 5a, panels E, K and W; Figure 5b,c), i.e., in the absence of TGF-1, as compared to vehicle-treated A549 cells (Figure 5a, panels A, G and S; Figure 5b,c). Likewise, no cell alteration in the expression and distribution of vimentin was evidenced by immunofluorescence assay when A549 cells were treated with PEA (10 µM) We previously reported that A549 cells cultured in the absence of TGF-β1 maintain a classic epithelial morphology appearing short, spindle-shaped and triangle-shaped, whereas following incubation with TGF-β1 (2 ng mL−<sup>1</sup> ) for 72 h, these cells assume an elongated shape, with many cells losing contact with their neighboring cells and displaying a long spindle shape typical of a fibroblast-like morphology, an effect known as Epithelial– Mesenchymal Transition (EMT) [32]. Consistent with these morphological observations, alterations in the expression and distribution of cytokeratin and vimentin were evidenced here by immunofluorescence assays. Indeed, these observations showed that, under the same experimental conditions, TGF-β1 induced the down-regulation of cytokeratin (Figure 5a, panels B and T; Figure 5b) and the up-regulation of vimentin (Figure 5a, panels H and T; Figure 5c), when compared to vehicle-incubated A549 cells (Figure 5a, panels A, G and S; Figure 5b,c). Treatment of TGF-β1-incubated A549 cells with PEA-OXA (10 µM) or PEA (10 µM) reverted the down-regulation of cytokeratin (Figure 5a, panels C, U, D and V; Figure 5b) and the up-regulation of vimentin (Figure 5a, panels I, U, J and V; Figure 5c). No cell alteration in the expression and distribution of cytokeratin and vimentin was evidenced by immunofluorescence assay when A549 cells were treated with PEA-OXA (10 µM) alone (Figure 5a, panels E, K and W; Figure 5b,c), i.e., in the absence of TGF-β1, as compared to vehicle-treated A549 cells (Figure 5a, panels A, G and S; Figure 5b,c). Likewise, no cell alteration in the expression and distribution of vimentin was evidenced by immunofluorescence assay when A549 cells were treated with PEA (10 µM) alone (Figure 5a,

alone (Figure 5a, panels L and X; Figure 5c), as compared to vehicle-treated A549 cells (Figure 5a, panels G and S; Figure 5c). Instead, PEA (10 µM) alone showed a per se effect on the expression and distribution of cytokeratin (Figure 5a, panels F and X; Figure 5b),

panels L and X; Figure 5c), as compared to vehicle-treated A549 cells (Figure 5a, panels G and S; Figure 5c). Instead, PEA (10 µM) alone showed a per se effect on the expression and distribution of cytokeratin (Figure 5a, panels F and X; Figure 5b), as compared to vehicle-treated A549 cells (Figure 5a, panels A and S; Figure 5b). *Biomolecules* **2022**, *12*, x FOR PEER REVIEW 9 of 13

**Figure 5.** TGF-1-induced epithelial–mesenchymal transition in A549 cells as evidenced by immunofluorescence approaches is avoided by the treatment with PEA-OXA or PEA. (**a**) Immunofluorescence staining was examined for the following markers: cytokeratin, an epithelial marker (green fluorescence, panels A–F), and vimentin, a mesenchymal marker (red fluorescence, panels G–L). DAPI staining was included to visualize the cell nucleus (blue fluorescence, panels M–R). In the merged images the co-expression and co-distribution of the markers are visualized (panels S–X). Quantifications of cytokeratin (**b**) and vimentin (**c**). The cells were captured with a 40 × microscope **Figure 5.** TGF-β1-induced epithelial–mesenchymal transition in A549 cells as evidenced by immunofluorescence approaches is avoided by the treatment with PEA-OXA or PEA. (**a**) Immunofluorescence staining was examined for the following markers: cytokeratin, an epithelial marker (green fluorescence, panels A–F), and vimentin, a mesenchymal marker (red fluorescence, panels G–L). DAPI staining was included to visualize the cell nucleus (blue fluorescence, panels M–R). In the merged images the co-expression and co-distribution of the markers are visualized (panels S–X). Quantifications of cytokeratin (**b**) and vimentin (**c**). The cells were captured with a 40 × microscope objective (Bar = 20 µm). Each bar shows the mean ± SEM of independent experiments (*n* = 3). The *p*-values were determined by ANOVA followed by Tukey's multiple comparisons test. \*\*\*\* *p* < 0.0001, \*\*\* *p* < 0.001, \*\* *p* < 0.01 and \* *p* < 0.05.

#### **4. Discussion**

Since the effectiveness of ultra-micronized PEA i) in attenuating acute lung inflammation, reducing immune cell infiltration and cytokine release in an acute lung injury model induced by lipopolysaccharide (LPS) [33] and ii) in inhibiting the pro-inflammatory response activated by SARS-CoV-2 spike protein in cultured murine alveolar macrophages has been already demonstrated [28], in this study we investigated for the first time if PEA-OXA, a congener of PEA endowed with anti-inflammatory activity [23], could be as effective as PEA in counteracting the IL-6 production and lung cell failure typical of COVID-19. For this purpose, we used two different in vitro models of acute lung injury. In the first in vitro model, we used a human lung epithelial cell line (A549) activated by a stable synthetic double-stranded RNA (poly-(I:C)) that can bind to toll-like receptor 3 (TLR3) with high affinity [34] to stimulate the pathophysiological viral disease state and reproduce the cell signaling pathways typical of the cytokine storm in terms of IL-6 overproduction. Our results demonstrated that both PEA-OXA and PEA were able to counteract IL-6 release induced by poly-(I:C) in A549 cells. Interestingly, PEA-OXA resulted to be more efficacious than PEA (at the same concentration tested and in terms of percent of the effect of poly-(I:C) alone) at counteracting poly-(I:C)-induced IL-6 cytokine release. It is possible that PEA-OXA exerts a stronger protective effect against IL-6 than PEA because it might act through a dual mechanism of action, both PEA-mediated and PEA-independent. In fact, we also investigated the mechanism of action through which PEA-OXA and PEA could exert their anti-inflammatory effects in A549 cells. Our results indicated that the inhibitory effect of PEA-OXA and PEA on poly-(I:C)-induced IL-6 cytokine release was not reverted by an antagonism at TRPV1 and PPAR-α receptors, although these receptors, and particularly the latter one, were strongly expressed in A549 cells. This suggests a non-TRPV1- and non-PPAR-α-mediated mechanism for the actions of the two molecules. Therefore, we investigated the ability of PEA-OXA and PEA to modulate the endogenous levels of endocannabinoids (AEA and 2-AG), as well as, in the case of PEA-OXA, the endogenous levels of PEA, in poly-(I:C)-stimulated A549 cells. Our results confirmed the existence of an "entourage" effect of PEA on the endogenous levels of 2-AG [15] and demonstrated for the first time the ability of PEA to also increase the endogenous levels of AEA in an inflammatory condition. In addition, we confirmed the ability of PEA-OXA to increase the endogenous levels of PEA in inflammatory conditions, an effect that might be exerted by inhibiting the enzyme responsible for PEA degradation (i.e. NAAA) [23]. In line with this hypothesis, we found a strong mRNA expression of NAAA in untreated A549 cells. This finding was not surprising, since NAAA is known to be abundantly expressed in lung cells and tissues [35]. Intriguingly, despite the fact that PEA-OXA significantly (18-fold) elevated the endogenous levels of PEA, this was not sufficient for this compound to also trigger the elevation of the endocannabinoid levels, suggesting that the elevation of PEA cellular levels beyond a certain threshold is necessary to induce an "entourage" effect. Indeed, and not surprisingly, exogenous PEA administration to the cells caused a much higher elevation of the cellular PEA levels (96-fold) than PEA-OXA at the same concentration. These results, taken together, suggest that the anti-inflammatory effects of PEA-OXA and PEA may be partially mediated by bioactive endogenous lipids (i.e., endocannabinoids in the case of PEA, and PEA in the case of PEA-OXA). Further studies will be needed to further clarify the mechanism(s) of the anti-inflammatory actions described for PEA and PEA-OXA in lung cells.

In the second in vitro model, we used A549 cells activated by TGF-β in order to reproduce the lung fibrosis that is a critical feature of chronic lung diseases and a serious complication of SARS-CoV-2 infection. In particular, TGF-β modulates lung tissue morphogenesis and differentiation by inducing the development of EMT, an important cellular process in chronic respiratory diseases. Our results suggest that following exposure to TGF-β, A549 cells acquire a fibroblast-like morphology characterized by a decrease of the epithelial marker expression and an increase of the mesenchymal marker expression. The immunofluorescence analysis showed the presence of EMT, characterized by a reduction

in cytokeratin-positive staining and an increase in vimentin-positive staining, which was inhibited by the treatment with both PEA-OXA and PEA. These results are in agreement with previous data, in which ultra-micronized PEA inhibited the inflammation response and lung fibrosis in mice subjected to idiopathic pulmonary fibrosis [36].

#### **5. Conclusions**

In summary, in this study we reported for the first time the protective effects of PEA-OXA and PEA in counteracting the inflammatory response induced by poly-(I:C), as well as in reverting the fibrosis induced by TGF-β. Our results also evidence a greater effectiveness of PEA-OXA over PEA and point to PEA-OXA as a new and promising candidate in the management of acute lung injury caused by conditions induced by a cytokine storm.

**Author Contributions:** Conceptualization, A.S.M., V.D. and S.P.; methodology, G.R., M.A., R.C. and R.V.; formal analysis, F.A.I., A.S.M. and S.P.; investigation, A.S.M., F.R., F.A.I. and S.P.; data curation, S.P.; writing—original draft preparation, A.S.M., F.R. and S.P.; writing—review and editing, V.D. and S.P.; supervision, V.D. and S.P.; project administration, S.P. All authors have read and agreed to the published version of the manuscript.

**Funding:** This research received no external funding.

**Institutional Review Board Statement:** Not applicable for studies not involving humans or animals.

**Data Availability Statement:** Not applicable here.

**Acknowledgments:** The authors thank Francesco Della Valle for his foresight, excellent and invaluable contribution to scientific research, opening new perspectives from pharmacological research to the pharmaceutical industry. S.P. also dedicates this work to the memory of her beloved mother who died of COVID-19.

**Conflicts of Interest:** A.S.M., M.A. and S.P. are employees of the Epitech Group SpA. S.P. and V.D. are co-inventors on patents on Adelmidrol and/or PEA, respectively, which are unrelated to the present study. The other authors declare no other conflict of interest.

#### **References**


**Disclaimer/Publisher's Note:** The statements, opinions and data contained in all publications are solely those of the individual author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to people or property resulting from any ideas, methods, instructions or products referred to in the content.

**Amira Elfarnawany 1,2 and Faramarz Dehghani 1,\***


**Abstract:** Chemotherapy-induced peripheral neuropathy (CIPN) is a common side effect of several chemotherapeutic agents, such as Paclitaxel. The main symptoms of CIPN are pain and numbness in the hands and feet. Paclitaxel is believed to accumulate in the dorsal root ganglia and free nerve endings. Novel therapeutic agents might help to mitigate or prevent Paclitaxel toxicity on dorsal root ganglion (DRG) neurons. Thus, we used primary DRG neurons as a model to investigate the potential neuroprotective effects of the endocannabinoid-like substance, palmitoylethanolamide (PEA). DRG neurons were isolated from cervical to sacral segments of spinal nerves of Wister rats (6–8 weeks old). After isolation and purification of neuronal cell populations, different concentrations of Paclitaxel (0.01–10 µM) or PEA (0.1–10 µM) or their combination were tested on cell viability by MTT assay at 24 h, 48, and 72 h post-treatment. Furthermore, morphometric analyses of neurite length and soma size for DRG neurons were performed. Adverse Paclitaxel effects on cell viability were apparent at 72 h post-treatment whereas Paclitaxel significantly reduced the neurite length in a concentrationdependent manner nearly at all investigated time points. However, Paclitaxel significantly increased the size of neuronal cell bodies at all time windows. These phenotypic effects were significantly reduced in neurons additionally treated with PEA, indicating the neuroprotective effect of PEA. PEA alone led to a significant increase in neuron viability regardless of PEA concentrations, apparent improvements in neurite outgrowth as well as a significant decrease in soma size of neurons at different investigated time points. Taken together, PEA showed promising protective effects against Paclitaxel-related toxicity on DRG neurons.

**Keywords:** peripheral neuropathic pain; neurotoxicity; dorsal root ganglion neurons; palmitoylethanolamide; paclitaxel; neurite length; soma size; MTT assay

#### **1. Introduction**

Chemotherapy-induced neuropathic pain (CINP) is a dose-limiting side effect of some anticancer drugs, such as bortezomib, cisplatin, oxaliplatin, paclitaxel, thalidomide, and vincristine [1]. The incidence of CINP in patients ranges from 12.1% to 96.2%, depending on the chemotherapeutic agent used and the type of cancer treated [2]. Taxanes are a class of chemotherapy drugs that promote tubulin polymerization into highly stable intracellular microtubules and cause cell death by intermixing with microtubules via normal cell division [3,4]. Paclitaxel is a Taxane derivative that has been used successfully as a first-line treatment for a variety of solid tumors, including ovarian cancer, breast cancer, cervical cancer, lung carcinomas, and other solid tumors [5–7].

Unfortunately, peripheral neuropathic pain (PNP) is a common side effect of Paclitaxel treatment affecting around 70.8% (95% CI 43.5–98.1) of patients [8]. The incidence ranges from 30 to 50% after a single dose and rises to more than 50% after a second dose [9]. Hyperalgesia, allodynia, and sporadic burning, shooting, numbness, spasm, and prickling sensations are some of CINP signs, and these can drastically lower the patient's quality

**Citation:** Elfarnawany, A.; Dehghani, F. Palmitoylethanolamide Mitigates Paclitaxel Toxicity in Primary Dorsal Root Ganglion Neurons. *Biomolecules* **2022**, *12*, 1873. https://doi.org/ 10.3390/biom12121873

Academic Editors: Rosalia Crupi and Salvatore Cuzzocrea

Received: 22 November 2022 Accepted: 10 December 2022 Published: 14 December 2022

**Publisher's Note:** MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

**Copyright:** © 2022 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https:// creativecommons.org/licenses/by/ 4.0/).

of life [10,11]. Chemotherapy-induced peripheral neuropathy (CIPN) is predominately a sensory axonopathy and neuronopathy, and the sensory neurons residing in dorsal root ganglions (DRGs) are the primary targets. Therefore, DRG explants have been shown to represent a good, simple, and well-accepted model for studying peripheral neuropathy induced by antineoplastic agents [12]. The ability of DRG explants to outgrow neurites in vitro when exposed to nerve growth factor (NGF), as well as the interference with neurite elongation by toxic substances, is the basis for their use in drug neurotoxicity assessment [12–14].

The neurotoxic effect of Paclitaxel on neurite length of DRG was shown to be doseand time-dependent [15,16], and DRG dissociated post-mitotic neurons were observed to die by necrosis [15]. Paclitaxel also caused the enlargement of neuronal cell bodies, and suppression of DRGs neuritis [17]. Paclitaxel has shown to demonstrate concentrationand time-dependent effects on vesicular trafficking and membrane localization of Nav1.7 in sensory axons of DRGs, providing a possible mechanistic explanation for increased excitability of primary afferents and pain [18]. Paclitaxel was reported to alter intracellular trafficking in both *Drosophila* and mouse models of CIPN by inducing recycling defects in mouse DRG neurons in vitro [19]. Currently, tricyclic antidepressants and analgesic drugs such as amitriptyline, morphine, gabapentin, and duloxetine display limited efficacy for preventing and alleviating paclitaxel-induced peripheral neuropathic pain and/or suffering of patients from serious side effects [20–23]. As a result, finding novel therapeutic agents that can mitigate or prevent Paclitaxel neurotoxicity on DRG neurons is very crucial.

The endocannabinoid system (ECS) is an important biological system that regulates and balances a wide range of physiological functions in the body, making it a target for many drugs and therapies [24]. Modulating the ECS activity showed promising therapeutic effects in a wide range of diseases and pathological conditions, including neurodegenerative, cardiovascular, and inflammatory disorders, obesity/metabolic syndrome, cachexia, chemotherapy-induced nausea and vomiting, tissue injury, and pain [25]. Palmitoylethanolamide (PEA), an endogenous fatty acid amide analogue of the endocannabinoid anandamide, has an important role in tissue protective mechanisms [26,27]. PEA was discovered nearly 5 decades ago in lipid extracts of various natural products, and its anti-inflammatory and antinociceptive properties were later described [28].

There is evidence for PEA synthesis during inflammation and tissue damage. PEA has a variety of beneficial effects, including the relief of inflammation and pruritus, and is effective in the control of neurogenic and neuropathic pain [29]. The hypothesized theories for PEA's mode of action include modulating endocannabinoid signaling and indirectly activating cannabinoid receptors via "entourage" effects [30–33].

PEA acts primarily through the direct activation of the nuclear receptor PPAR-α [34]. After the activation of PPAR-α receptor, a chain of events leads to suppression of pain and inflammatory signals, including the inhibition of the release of pro-inflammatory cytokines such as IL-1β and IL-6 [35]. Previous studies showed a PEA-mediated protection of dentate gyrus granule cells during secondary neuronal damage, which was mediated by PPAR-α activation and influenced by reduction in inflammatory processes [36]. In a chronic constriction injury model of neuropathic pain, repeated PEA treatment (30 mg/kg) not only decreased edema and macrophage infiltrates, but also declined the decrease in axon diameter and myelin thickness [37]. However, research on studying the protective role of PEA against the toxicity of Paclitaxel on DRG neurons is still lacking.

In the present study, the effects of different Paclitaxel and PEA concentrations were investigated, either individually or in combination, on the viability, morphology, and neurite length of primary DRG neurons at various time points. We hypothesized that PEA might reduce the neurotoxicity induced by Paclitaxel on DRG neurons in a concentrationor/and time-dependent manner.

#### **2. Materials and Methods**

#### *2.1. Ethics Statement*

All animal experiments were carried out in accordance with the policy on ethics and the policy on the use of animals in neuroscience research, as specified in directive 2010/63/EU of the European Parliament and of the Council of the European Union on the protection of animals used for scientific purposes and were approved by local authorities for laboratory animal care and use (State of Saxony-Anhalt, Germany, permission number: I11M27).

#### *2.2. Materials*

Experiments were conducted with Palmitoylethanolamide (PEA, Tocris Bioscience, cat No. 0879-10 mg, Bristol, UK), Paclitaxel (Taxol equivalent, Invitrogen, cat No. P3456- 5 mg, Schwerte, Germany), Nerve Growth Factor-2.5S from the murine submaxillary gland (NGF, Sigma Aldrich, Merck, cat No. N6009-4X 25 µg, St. Louis, MO, USA) and glial cell-derived neurotrophic factor (GDNF, Sigma-Aldrich, cat No. SRP3309-10 µg, St. Louis, MO, USA), Uridine (Uridin, Sigma-Aldrich, U3003-5 g, Darmstadt, Germany), and 5- Fluoro-2-deoxyuridine (FudR, Sigma-Aldrich, cat No. F0503-100 mg, Darmstadt, Germany). PEA and Paclitaxelwere dissolved in DMSO to obtain stock solutions of 10 mM PEA and 1 mM Paclitaxel and stored at −20 ◦C, while NGF and GDNF dissolved in 0.1 % Bovine Serum Albumin (BSA, Sigma-Aldrich, cat No. A7906-10 g, St. Louis, MO, USA). A total of 20 mM uridine/5-fluorodeoxyuridine (UFdU) stock solution was prepared by mixing 48.8 mg uridine and 49.2 mg 5-fluorodeoxyuridine in 10 mL distilled water, and 100 µL aliquots were prepared and frozen at −20 ◦C. Notably, controls contained the similar highest concentration of DMSO (0.1%) to exclude any effects on investigated parameters.

#### *2.3. Isolation and Preparation of DRG Neurons*

DRG tissues isolated from 6–8 weeks of age Wister rats. In brief, rats were deeply anesthetized with isoflurane (Florene, 100% (*V*/*V*), 250 mL, Abcam, cat No. B506, Carros, France) by inhalation and sacrificed by decapitation with a commercial guillotine. Under aseptic conditions, the vertebral column was isolated and carefully cleared from all surrounding muscle, fat, and other soft tissue. The spinal cord was then exposed and scooped out. Following the dorsal roots. DRGs were localized, removed, collected from intervertebral foramina at both sides, and placed in a 3 mL sterile dish containing Hanks balanced salt solutions without Mg2+/Ca2+ (HBSS, Invitrogen, REF. 24020-091, Schwerte, Germany). Dorsal root neuronal culture was prepared according to a previously published protocol [38] with some modifications. Briefly, isolated DRGs were enzymatically digested in the first enzymatic solution, which contained 60 U/mL papain solution (Sigma-Aldrich, cat No. P4762-100 mg, St. Louis, MO, USA), 3 µL of 80 mg/mL saturated sodium hydrogen carbonate solution (NaHCO3, Merck, cat No. k22399729, Darmstadt, Germany), and 0.6 mg/mL L-Cysteine (L-Cys, Sigma-Aldrich, Cat No. C7352-25 g, St. Louis, MO, USA) dissolved in 1.5 mL of HBSS without Mg2+/ Ca2+. Afterwards, DRGs were incubated in a papain solution for 15 min in a 37 ◦C water bath, then incubated in a second solution which consisted of 4 mg/mL collagenase type II solution (CLS2, Sigma-Aldrich, Cat No.C6885-1 gm, St. Louis, MO, USA) and 4.6 mg/mL dispase type II (Dispase II, Sigma-Aldrich, Cat No. D4693-1 gm, St. Louis, MO, USA) solution in 3 mL HBSS without Mg2+/ Ca2+. The DRGs were mixed gently in collagenase solution and incubated again for 15 min in a water bath at 37 ◦C.

The resulting cell suspension was centrifuged at 200× *g* for 1 min. The collagenase solution was carefully aspirated, and the DRGs were washed with 2ml of titration media consisting of high glucose Dulbecco's Modified Eagle Medium (DMEM, Invitrogen, Ref. 41965-039, Schwerte, Germany) containing 10% heat-inactivated Fetal Bovine Serum (FBS, Invitrogen, REF. 10270-106, Schwerte, Germany). The DRGs were triturated 10–15 times by using p1000 pipette tips until the cell suspension became cloudy. Bovine serum albumin (BSA) was used for purification (15% (*W*/*V*) BSA solution) to obtain nearly pure neurons without myelin debris. After trituration, single-cell suspensions from DRGs were

centrifuged through 15% (*W*/*V*) BSA solution in DMEM, 3 mL of 15% BSA solution: 1 mL of cell suspension in a 15 mL conical tube at 300 g for 8 min at room temperature (RT) to separate sensory neurons in the pellet from non-neuronal cells and debris [39]. The BSA solution was removed and the pellet containing neurons was re-suspended in 1 mL of culture medium consisting of 445 mL of F12 medium (1X, Invitrogen, REF.11765-054, Schwerte, Germany), 50 mL of FBS and 5 mL of 0.1 mg/mL streptomycin/penicillin (Sigma Aldrich, cat No. P4333/100 mL, Darmstadt, Germany). The cell suspension was filtered by a 40 µm cell strainer (SARSTEDT, cat No. D-51588, Schwerte, Germany) to obtain single-cell suspensions and remove undigested tissue debris. The BSA solution was removed and the pellet containing neurons was re-suspended in 1 mL of culture medium consisting of 445 mL of F12 medium (1X, Invitrogen, REF.11765-054, Schwerte, Germany), 50 mL of FBS and 5 mL of 0.1 mg/mL streptomycin/penicillin (Sigma Aldrich, cat No. P4333/100 mL, Darmstadt, Germany). The cell suspension was filtered by a 40 µm cell strainer (SARSTEDT, cat No. D-51588, Schwerte, Germany) to obtain single-cell suspensions and remove undigested tissue debris. *2.4. Seeding and Growth of DRGs Neurons* 

Coverslips 12 mm round were pre-coated with 2 mg/mL Poly-D-lysine (PDL, Sigma

times by using p1000 pipette tips until the cell suspension became cloudy. Bovine serum albumin (BSA) was used for purification (15% (*W*/*V*) BSA solution) to obtain nearly pure neurons without myelin debris. After trituration, single-cell suspensions from DRGs were centrifuged through 15% (*W*/*V*) BSA solution in DMEM, 3 mL of 15% BSA solution: 1 mL of cell suspension in a 15 mL conical tube at 300 g for 8 min at room temperature (RT) to separate sensory neurons in the pellet from non-neuronal cells and debris [39].

*Biomolecules* **2022**, *12*, x FOR PEER REVIEW 4 of 19

#### *2.4. Seeding and Growth of DRG Neurons* Aldrich, cat No. P6407, St. Louis, MO, USA) and 0.2 mg/mL laminin (Sigma Aldrich, cat

Coverslips 12 mm round were pre-coated with 2 mg/mL Poly-D-lysine (PDL, Sigma Aldrich, cat No. P6407, St. Louis, MO, USA) and 0.2 mg/mL laminin (Sigma Aldrich, cat No. L2020-1 mg, St. Louis, MO, USA) for at least 1 h or overnight in 4 ◦C, then washed one time with distilled H2O directly before seeding the cells in culture medium. DRG neuronal cells (5000 cells in 50 µL culture medium) were then pre-seeded onto the center of the coated coverslips for 2 h in an incubator with 37 ◦C and 5% CO2. Then, 1 mL of warm culture medium adjusted at pH 7.4 containing 50 ng/mL NGF and 20 ng/mL GDNF (which is essential for growing neuritis of neurons) and 20 µM UFdU (for inhibiting the growth of any remains of supporting cells in culture) was gently added to the wells, and the cells were maintained again at 37 ◦C with 5% CO2. The growth and morphology of neurons were monitored after 2, 24, 48, and 72 h to detect the suitable time of treatment (Figure 1a). No. L2020-1 mg, St. Louis, MO, USA) for at least 1 h or overnight in 4 °C, then washed one time with distilled H2O directly before seeding the cells in culture medium. DRGs neuronal cells (5000 cells in 50 µL culture medium) were then pre-seeded onto the center of the coated coverslips for 2 h in an incubator with 37 °C and 5% CO2. Then, 1 mL of warm culture medium adjusted at pH 7.4 containing 50 ng/mL NGF and 20 ng/mL GDNF (which is essential for growing neuritis of neurons) and 20 µM UFdU (for inhibiting the growth of any remains of supporting cells in culture) was gently added to the wells, and the cells were maintained again at 37 °C with 5% CO2. The growth and morphology of neurons were monitored after 2, 24, 48, and 72 h to detect the suitable time of treatment (Figure 1a).

**Figure 1.** Morphological features and treatment protocols of DRG neurons. (**a**) Representative images show the morphology and growth of DRG neurons at different time points after BSA purification. Scale bars = 50 µm. (**b**,**c**) treatment protocols for studying the effects of Paclitaxel or /and palmitoylethanolamide (PEA) on DRG neurons viability and morphology (Neurite length measurement) respectively, at 24, 48, and 72-h post-treatment. **Figure 1.** Morphological features and treatment protocols of DRG neurons. (**a**) Representative images show the morphology and growth of DRG neurons at different time points after BSA purification. Scale bars = 50 µm. (**b**,**c**) treatment protocols for studying the effects of Paclitaxel or /and palmitoylethanolamide (PEA) on DRG neurons viability and morphology (Neurite length measurement) respectively, at 24, 48, and 72-h post-treatment.

#### *2.5. Cell Viability (MTT Assay)*

DRG neurons were treated 24 h after seeding. Cells were treated with different concentrations of Paclitaxel (0.01, 0.1, 1, 10 µM) and PEA (0.1, 1, 10 µM), either individually

or simultaneously combined to study the effects on cell viability. Paclitaxel concentrations were selected based on the literature [15–19,40,41] as well as PEA [42–44]. DRG neurons (4–5 <sup>×</sup> <sup>10</sup><sup>4</sup> cells/well) in 96 well plates were treated with different concentrations of Paclitaxel and PEA alone or in combination for 24, 48, and 72 h. (Figure 1b). Then, cell viability (%) was measured at the different time points using MTT assay. Four hours before termination of experiments at different time points, 3-(4,5-dimethylthiazol-2-yl)- 2,5-diphenyltetrazolium bromide solution (MTT, Invitrogen, cat. No M6494, 5 mg/mL, Eugene, OR, USA) was added. Cells were further incubated for 4 h at 37 ◦C and 5% CO2. After removing MTT solution, formazan crystals dissolved in 100 µL of dimethyl sulfoxide (DMSO, Sigma-Aldrich, cat No. D4540-500 mL, Lyon, France) were added and, after another 20 min absorbance values, were measured at wavelengths (540 nm and 720 nm) by a microplate reader (SynergyTMMx, BioTek Instruments, Winooski, VT, USA). DRG neurons cultured in normal media free of Paclitaxel or/and PEA were used as control groups. Controls contained the similar highest concentration of DMSO (0.1%) to exclude any solvent effects on cell viability. All experiments were performed three times independently with 2–3 technical replica for each treatment.

#### *2.6. Immunofluorescence Staining and Microscopy*

To investigate the effects of various treatments on the morphology of DRG neurons, cells (2–4 <sup>×</sup> <sup>10</sup><sup>3</sup> cells/well) were seeded on 12 mm sterile coverslips in a 24-well plate (Greiner Bio-One, Cat No. 662160, Frickenhausen, Germany), cultured for 24 h until most neurites outgrew, and then treated with different concentrations of Paclitaxel or PEA, either alone or in combination (Figure 1c). At the end of each time point, the cells were fixed with 4% paraformaldehyde (PFA, AppliChem, cat No.141451.1211, Darmstadt, Germany) for 15 min at RT and immediately subjected to immunofluorescence or stored in 1 × PBS at 4 ◦C until further use. For immunofluorescence staining, fixed cells were washed 3 times with 0.02 M PBS for 10 min before unspecific bindings were blocked by incubating cells in normal goat serum (NGS, Sigma Aldrich, cat No. G9023-10 mL, Taufkirchen, Germany, 1:20) in 0.02 M PBS/0.3% (*v*/*v*) Triton) for 30 min. Afterward, cells were incubated with neuronal marker mouse anti-β-III tubulin antibody (TUBB3, Biolegend, San Diego, cat No: 801201, CA, USA, 1:1000) overnight for labelling the cytoskeleton of neurons. Coverslips were thereafter washed thrice for 10 min in PBS, incubated with the secondary antibody goat anti-mouse Alexa Fluor® 488 conjugated (Life Technologies, cat No. 2066710, Darmstadt, Germany, 1:200) for 1 h, washed again 3 times with PBS, and stained with DAPI (40 ,6- Diamin-2-phenylindol, Sigma-Aldrich, Munich, Germany, cat No. D9542) for visualization of nuclei. The stained cells were washed in distilled water and covered with DAKO fluorescence mounting medium (DAKO, Agilent Technologies, Inc., Santa Clara, CA 95051, USA). The DRG neurons photomicrographs were captured by using a Leica DMi8 (Wetzlar, Germany) microscope, and five images were randomly taken from each coverslip. The experiment was performed 3 times independently.

#### *2.7. Image Analysis and Determination of Neurite Lengths and Soma Sizes*

Measurement of neurite length as a marker for investigating the neurotoxicity of DRG neurons was assessed by using Neurite Tracer, a plugin for ImageJ software version v1.52 used for automated neurite tracing as previously described [45] with some adjustments (Figure S1). Briefly, a sample image pair from cultures of DRG neurons fluorescently labelled with TUBB3 as neuronal marker and DAPI as nuclear marker were opened in Image J (v1.46r (National Institutes of Health, Laboratory for Optical and Computational Instrumentation, University of Wisconsin, Madison, WI, USA) and converted to 8-bit grayscale, and then individually opened in neurite tracer plugin. Large bright objects (somats of neurons) were removed from all images by application of Fiji software version 2.9.0 (accessed 15 January 2022) (https://imagej.net/Fiji/Downloads). Thereafter, the resulting images were inserted to neurite tracer. Afterwards, the threshold was adjusted manually before starting the automated tracing of neuritis. Images with the traced neuritis were merged with RGB original images to ensure the reliability and accuracy of the tracing process. Afterwards, the number of neurons was determined by using a multi-point tool of ImageJ. Finally, traced neuritis lengths were normalized with the numbers of neurons to calculate the neurite length/cell. To determine the size of neuronal somata, soma areas of neurons were selected, and soma areas were measured. The results were normalized with those from the control group. using a multi-point tool of ImageJ. Finally, traced neuritis lengths were normalized with the numbers of neurons to calculate the neurite length/cell. To determine the size of neuronal somata, soma areas of neurons were selected, and soma areas were measured. The results were normalized with those from the control group. *2.8. Statistical Analysis* 

*Biomolecules* **2022**, *12*, x FOR PEER REVIEW 6 of 19

Computational Instrumentation, University of Wisconsin, Madison, WI, USA) and converted to 8-bit grayscale, and then individually opened in neurite tracer plugin. Large bright objects (somats of neurons) were removed from all images by application of Fiji software version 2.9.0 (accessed 15 January 2022) (https://imagej.net/Fiji/Downloads). Thereafter, the resulting images were inserted to neurite tracer. Afterwards, the threshold was adjusted manually before starting the automated tracing of neuritis. Images with the traced neuritis were merged with RGB original images to ensure the reliability and accuracy of the tracing process. Afterwards, the number of neurons was determined by

#### *2.8. Statistical Analysis* Data analysis and visualization were carried out by using GraphPad Prism

Data analysis and visualization were carried out by using GraphPad Prism (GraphPad Software version 8.0.1 for Windows, La Jolla, CA, USA). The normal distribution of data was assessed by use of the Kolmogorov–Smirnov test. The effect of treatments on viability and neurite length of DRG neurons was assessed using one-way ANOVA (analysis of variance) followed by the Bonferroni multiple comparisons test (*p* < 0.05). An alpha level of 0.05 was used for all tests. (GraphPad Software version 8.0.1 for Windows, La Jolla, CA, USA). The normal distribution of data was assessed by use of the Kolmogorov–Smirnov test. The effect of treatments on viability and neurite length of DRG neurons was assessed using one-way ANOVA (analysis of variance) followed by the Bonferroni multiple comparisons test (*p* < 0.05). An alpha level of 0.05 was used for all tests. **3. Results** 

#### **3. Results**

#### *3.1. Characterization of DRG Neuronal Cells* DRG neurons cultures were examined under a light microscope at various time

DRG neurons cultures were examined under a light microscope at various time points (2, 24, 48, and 72 h) to track their growth and morphology. After 2 hours, neuron somas appeared round, bright, and refractile, with a large nucleus (Figure 1a). Three distinct subpopulations (small, medium, and large neurons) based on soma diameter were observed, (Figure 1a). Most of the DRG neurons extended long thin neuritis after 24 h of cells seeding, while, after 48 and 72 h of culturing, all sensory neurons had long neuritis which connected and formed networks together (Figure 1a). points (2, 24, 48, and 72 h) to track their growth and morphology. After 2 hours, neuron somas appeared round, bright, and refractile, with a large nucleus (Figure 1a). Three distinct subpopulations (small, medium, and large neurons) based on soma diameter were observed, (Figure 1a). Most of the DRG neurons extended long thin neuritis after 24 h of cells seeding, while, after 48 and 72 h of culturing, all sensory neurons had long neuritis which connected and formed networks together (Figure 1a).

*3.2. Effects of Paclitaxel or/and PEA on Cell Viability of DRG Neurons* 

#### *3.2. Effects of Paclitaxel or/and PEA on Cell Viability of DRG Neurons* DRGs neurons were treated with different concentrations of Paclitaxel for 24, 48, and

*3.1. Characterization of DRG Neuronal Cells* 

DRG neurons were treated with different concentrations of Paclitaxel for 24, 48, and 72 h, and we found a significant reduction in the viability of cells at only 72 h post-treatment, regardless of Paclitaxel concentrations, compared to the untreated control group (*p* < 0.001) (Figure 2). Paclitaxel's effects on neuron viability were obviously time-dependent but not concentration-dependent. PEA, as expected, showed no statistically significant effect on the viability of cells in comparison to the untreated control group (*p* > 0.05) at 72 h post-treatment (Figure S2). 72-h, and we found a significant reduction in the viability of cells at only 72 h post-treatment, regardless of Paclitaxel concentrations, compared to the untreated control group (*p* < 0.001) (Figure 2). Paclitaxel's effects on neuron viability were obviously time-dependent but not concentration-dependent. PEA, as expected, showed no statistically significant effect on the viability of cells in comparison to the untreated control group (*p* ˃ 0.05) at 72 h post-treatment (Figure S2).

**Figure 2.** Effects of different concentrations of Paclitaxel on viability (%) of DRG neurons at different time points. Application of different concentrations of Paclitaxel showed no influence on the viability of neurons at (**a**) 24 h and (**b**) 48 h, whereas, at (**c**) 72 h post-treatment, Paclitaxel significantly reduced the viability of cells compared to control (\*\*\* *p* < 0.001). The asterisk denotes significant results regarding the respective measurement indicated with the bar. Values are served as mean ± SEM of three independent experiments performed in triplicate. SEM: Standard error mean.

The effects of combined treatments (Paclitaxel plus PEA) were compared to the effect of Paclitaxel alone on viability (%) at 72 h post-treatment. A significant increase was observed for almost all combinations of Paclitaxel (0.01–10 µM) plus PEA (0.1–10 µM) compared to cells treated with Paclitaxel alone. A significant effect was missed only for the combination (10 µM Paclitaxel + 1 µM PEA vs. 10 µM Paclitaxel) (*p* < 0.05) (Figure 3). Notably, the effect of PEA against Paclitaxel was clearly concentration independent. observed for almost all combinations of Paclitaxel (0.01–10 µM) plus PEA (0.1–10 µM) compared to cells treated with Paclitaxel alone. A significant effect was missed only for the combination (10 µM Paclitaxel + 1 µM PEA vs. 10 µM Paclitaxel) (*p* < 0.05) (Figure 3). Notably, the effect of PEA against Paclitaxel was clearly concentrationindependent.

*Biomolecules* **2022**, *12*, x FOR PEER REVIEW 7 of 19

**Figure 2.** Effects of different concentrations of Paclitaxel on viability (%) of DRG neurons at different time points. Application of different concentrations of Paclitaxel showed no influence on the viability of neurons at (**a**) 24 h and (**b**) 48 h, whereas, at (**c**) 72 h post-treatment, Paclitaxel significantly reduced the viability of cells compared to control (\*\*\* *p* < 0.001). The asterisk denotes significant results regarding the respective measurement indicated with the bar. Values are served as mean ± SEM of three independent experiments performed in triplicate. SEM: Standard error mean.

The effects of combined treatments (Paclitaxel plus PEA) were compared to the effect of Paclitaxel alone on viability (%) at 72 h post-treatment. A significant increase was

**Figure 3.** Effects of different concentrations of Paclitaxel (**a**) 0.01 µM, (**b**) 0.1 µM, (**c**) 1 µM, and (**d**) 10 µM either alone or in combination with different concentrations of PEA on viability (%) (mean ± SEM) of DRG neurons at 72 h post-treatment by using MTT assay. The asterisk indicates a significant increase in viability of DRG neurons treated with different Paclitaxel concentrations in combination with different concentrations of PEA at 72 h post-treatment compared to cells treated with **Figure 3.** Effects of different concentrations of Paclitaxel (**a**) 0.01 µM, (**b**) 0.1 µM, (**c**) 1 µM, and (**d**) 10 µM either alone or in combination with different concentrations of PEA on viability (%) (mean ± SEM) of DRG neurons at 72 h post-treatment by using MTT assay. The asterisk indicates a significant increase in viability of DRG neurons treated with different Paclitaxel concentrations in combination with different concentrations of PEA at 72 h post-treatment compared to cells treated with Paclitaxel only (\* *p* < 0.05, \*\* *p* < 0.01, and \*\*\* *p* < 0.001). Data are (mean ± SEM) of three independent experiments performed in duplicate.

#### *3.3. Effects of Paclitaxel or/and PEA on Morphology of DRG Neurons*

Toxic hallmarks of Paclitaxel were observed on the morphology of neurons such as suppression in neurite lengths of neurons, swelling of neuronal cell bodies, as well as retraction and blebbing formation at the distal endings of neurites (Figure 4a). To verify and quantify the Paclitaxel and PEA effects, two different endpoints were assessed, namely neurite length and soma size.

experiments performed in duplicate.

namely neurite length and soma size.

Paclitaxel only (\* *p* < 0.05, \*\* *p* < 0.01, and \*\*\* *p* < 0.001). Data are (mean ± SEM) of three independent

Toxic hallmarks of Paclitaxel were observed on the morphology of neurons such as suppression in neurite lengths of neurons, swelling of neuronal cell bodies, as well as retraction and blebbing formation at the distal endings of neurites (Figure 4a). To verify and quantify the Paclitaxel and PEA effects, two different endpoints were assessed,

*3.3. Effects of Paclitaxel or/and PEA on Morphology of DRG Neurons* 

**Figure 4.** Effects of different Paclitaxel concentrations on morphology, neurite length, and soma size of DRG neurons. (**a**) Immunofluorescence staining of DRG neurons treated with different Paclitaxel concentrations (0.01, 0.1, 1, and 10 µM) labeled with anti-mouse beta III Tubulin antibody after 24, 48, and 72 h. Different concentrations of Paclitaxel had toxic effects leading to a reduction in neurite length and an increase in soma area (yellow arrows) at all time points. Additionally, other characteristics of Paclitaxel toxicity on neuronal morphology are visible, including swellings and blebbing at distal ends of neuritis (red arrows). Nuclei were counterstained with DAPI. Five to eight regions were recorded randomly per each coverslip. Scale bars = 75 µm. (**b**–**d**) significant suppression in neurite lengths of DRG neurons treated with different Paclitaxel concentrations in comparison with the control group (\*\* *p* < 0.01, \*\*\* *p* < 0.001, \*\*\*\* *p* < 0.0001) at 24, 48, and 72-h post-treatment, respectively. Data are (mean ± SEM) of three independent experiments performed with (10–15) replicates. **Figure 4.** Effects of different Paclitaxel concentrations on morphology, neurite length, and soma size of DRG neurons. (**a**) Immunofluorescence staining of DRG neurons treated with different Paclitaxel concentrations (0.01, 0.1, 1, and 10 µM) labeled with anti-mouse beta III Tubulin antibody after 24, 48, and 72 h. Different concentrations of Paclitaxel had toxic effects leading to a reduction in neurite length and an increase in soma area (yellow arrows) at all time points. Additionally, other characteristics of Paclitaxel toxicity on neuronal morphology are visible, including swellings and blebbing at distal ends of neuritis (red arrows). Nuclei were counterstained with DAPI. Five to eight regions were recorded randomly per each coverslip. Scale bars = 75 µm. (**b**–**d**) significant suppression in neurite lengths of DRG neurons treated with different Paclitaxel concentrations in comparison with the control group (\*\* *p* < 0.01, \*\*\* *p* < 0.001, \*\*\*\* *p* < 0.0001) at 24, 48, and 72-h post-treatment, respectively. Data are (mean ± SEM) of three independent experiments performed with (10–15) replicates.

#### 3.3.1. Neurite Length

The treatment with the four different concentrations of Paclitaxel resulted in a significant reduction in neurite length 24 h after treatment when compared to the non-treated control group (*p* < 0.05) (Figure 4b). At 48 and 72 h post-treatment, all studied Paclitaxel groups had an apparent reduction in neurite length except for 0.01 µM Paclitaxel relative to control group (Figure 4c,d). Interestingly, Paclitaxel effects on neurite length were clearly time- and concentration-dependent. No alterations in morphology and neurite length were found in DRG neurons treated with PEA in comparison to the vehicle control group (*p* > 0.05) (Figure S3a–c).

The three different concentrations of PEA co-applied with 0.01 µM Paclitaxel had no significant protective effects on the neurite lengths of neurons at all investigated timelines compared to the 0.01 µM Paclitaxel group (*p* > 0.05; Figure S4). All combined groups of PEA with 0.1 µM Paclitaxel did not cause any significant increase in neurite length at any time point when compared to the 0.1 µM Paclitaxel group alone (*p* > 0.05); however, at 72 h after application, 1 µM PEA only plus 0.1 µM Paclitaxel resulted in an apparent increase in neurite length of DRG neurons compared to 0.1 µM Paclitaxel group (*p* < 0.05; Figure S5).

However, PEA concentrations (0.1, 1, and 10 µM) showed a significant protective effect on neurite outgrowth of DRG neurons when combined with 1 µM Paclitaxel and compared with 1 µM Paclitaxel alone at 24 and 72-h post-treatment. A total of 0.1 µM PEA combined with 1 µM Paclitaxel had a significant protective effect on neurite lengths of DRG neurons 48 h after treatment when compared to Paclitaxel alone (*p* < 0.05) (Figure 5). *Biomolecules* **2022**, *12*, x FOR PEER REVIEW 10 of 19

**Figure 5.** Showing protective effects of different PEA concentrations (0.1, 1, and 10 µM) co-applied with 1 µM Paclitaxel on neurite lengths of DRG neurons compared to 1 µM Paclitaxel alone at 24, 48, and 72 h post-treatment. (**a**) Representative microphotographs of DRG neurons stained with beta III Tubulin antibody for soma and neuritis (green) and DAPI for nuclei (blue). Scale bars = 75 µm. (**b**) Bar graphs indicated a significant increase in neurite lengths of neurons treated with different concentrations of PEA at 24 h and 72 h post-treatment compared to cells treated with Paclitaxel only, while at 48 h post-treatment only 0.1 µM PEA demonstrated a significant increase in neurite length against 1 µM Paclitaxel (\* *p* < 0.05, \*\* *p* < 0.01, \*\*\* *p* < 0.001, \*\*\*\* *p* < 0.0001). Data are (mean ± SEM) of three independent experiments performed in 10–15 replicates. The asterisk denotes significant results regarding the respective measurement indicated with bar charts. **Figure 5.** Showing protective effects of different PEA concentrations (0.1, 1, and 10 µM) co-applied with 1 µM Paclitaxel on neurite lengths of DRG neurons compared to 1 µM Paclitaxel alone at 24, 48, and 72 h post-treatment. (**a**) Representative microphotographs of DRG neurons stained with beta III Tubulin antibody for soma and neuritis (green) and DAPI for nuclei (blue). Scale bars = 75 µm. (**b**) Bar graphs indicated a significant increase in neurite lengths of neurons treated with different concentrations of PEA at 24 h and 72 h post-treatment compared to cells treated with Paclitaxel only, while at 48 h post-treatment only 0.1 µM PEA demonstrated a significant increase in neurite length against 1 µM Paclitaxel (\* *p* < 0.05, \*\* *p* < 0.01, \*\*\* *p* < 0.001, \*\*\*\* *p* < 0.0001). Data are (mean ± SEM) of three independent experiments performed in 10–15 replicates. The asterisk denotes significant results regarding the respective measurement indicated with bar charts.

Regarding the protective effects of different PEA concentrations against 10 µM Paclitaxel, we found 0.1 or 1 µM PEA combined with 10 µM Paclitaxel showed a significant increase in neurite lengths of DRG neurons at 24, 48, and 72-h post-treatment compared to cells treated with 10 µM Paclitaxel only (*p* < 0.05) (Figure 6). Meanwhile, the 10 µM PEA plus 10 µM Paclitaxel group revealed a significant increase in neurite lengths only at 24 h post-treatment in comparison to the 10 µM Paclitaxel group (*p* < 0.05) (Figure 6). Regarding the protective effects of different PEA concentrations against 10 µM Paclitaxel, we found 0.1 or 1 µM PEA combined with 10 µM Paclitaxel showed a significant increase in neurite lengths of DRG neurons at 24, 48, and 72-h post-treatment compared to cells treated with 10 µM Paclitaxel only (*p* < 0.05) (Figure 6). Meanwhile, the 10 µM PEA plus 10 µM Paclitaxel group revealed a significant increase in neurite lengths only at 24 h post-treatment in comparison to the 10 µM Paclitaxel group (*p* < 0.05) (Figure 6).

**Figure 6.** Effects of different PEA concentrations (0.1, 1, and 10 µM) combined with 10 µM Paclitaxel at 24, 48, and 72 h post-treatment. (**a**) Representative immunofluorescence images show DRG neurons labeled with beta III Tubulin antibody (green) and DAPI for nuclei (blue). Scale bars = 75 µm. (**b**) A significant increase in neurite length of neurons was found in groups treated with different concentrations of PEA at 24 h only compared to cells treated with Paclitaxel only. At 48 and 72-h post-treatment, 0.1 µM PEA or 1 µM PEA combined with 10 µM Paclitaxel demonstrated **Figure 6.** Effects of different PEA concentrations (0.1, 1, and 10 µM) combined with 10 µM Paclitaxel at 24, 48, and 72 h post-treatment. (**a**) Representative immunofluorescence images show DRG neurons labeled with beta III Tubulin antibody (green) and DAPI for nuclei (blue). Scale bars = 75 µm. (**b**) A significant increase in neurite length of neurons was found in groups treated with different concentrations of PEA at 24 h only compared to cells treated with Paclitaxel only. At 48 and 72-h post-treatment, 0.1 µM PEA or 1 µM PEA combined with 10 µM Paclitaxel demonstrated a significant increasing effect on the neurite lengths in comparison with 10 µM Paclitaxel alone (\* *p* < 0.05, \*\* *p* < 0.01, \*\*\* *p* < 0.001, \*\*\*\* *p* < 0.0001). Data represented as (mean ± SEM), and the experiments were performed at least 3 independent times and 10–15 replicas. The asterisk denotes significant results regarding the respective measurement indicated with the bar graphs.

#### 3.3.2. Soma Size neurons 24 h post-treatment when compared to the control group (*p* < 0.05) (Figure 7a).

3.3.2. Soma Size

The four different concentrations of Paclitaxel led to an increase in the soma size of neurons 24 h post-treatment when compared to the control group (*p* < 0.05) (Figure 7a). At 48 and 72-h, all investigated groups treated with Paclitaxel showed apparent enlargements in areas of neuronal somata except for 0.01 µM of Paclitaxel when compared to the control group (*p* < 0.05) (Figure 7b,c). Effects of Paclitaxel on soma size of neuronal cell bodies were obviously time- and concentration-dependent. Treatment with PEA alone demonstrated no significant effects on the size of neuronal bodies at any time point when compared to control group (*p* > 0.05) (Figure S3d–f). At 48 and 72-h, all investigated groups treated with Paclitaxel showed apparent enlargements in areas of neuronal somata except for 0.01 µM of Paclitaxel when compared to the control group (*p* < 0.05) (Figure 7b,c). Effects of Paclitaxel on soma size of neuronal cell bodies were obviously time- and concentration-dependent. Treatment with PEA alone demonstrated no significant effects on the size of neuronal bodies at any time point when compared to control group (*p* > 0.05) (Figure S3d–f).

a significant increasing effect on the neurite lengths in comparison with 10 µM Paclitaxel alone (\* *p* < 0.05, \*\* *p* < 0.01, \*\*\* *p* < 0.001, \*\*\*\* *p* < 0.0001). Data represented as (mean ± SEM), and the experiments were performed at least 3 independent times and 10–15 replicas. The asterisk denotes sig-

The four different concentrations of Paclitaxel led to an increase in the soma size of

nificant results regarding the respective measurement indicated with the bar graphs.

*Biomolecules* **2022**, *12*, x FOR PEER REVIEW 12 of 19

**Figure 7.** Effects of different Paclitaxel concentrations on soma size of DRG neurons at different time points. Bar charts show a significant increase in the soma size of DRG neurons after the application of different Paclitaxel concentrations at (**a**) 24 h, (**b**) 48 h, and (**c**) 72 h post-treatment compared to the control (\*\* *p* < 0.01, \*\*\* *p* < 0.001, \*\*\*\* *p* < 0.0001). Asterisks denote significant results regarding the respective measurement indicated with the bar. Values are served as mean ± SEM of **Figure 7.** Effects of different Paclitaxel concentrations on soma size of DRG neurons at different time points. Bar charts show a significant increase in the soma size of DRG neurons after the application of different Paclitaxel concentrations at (**a**) 24 h, (**b**) 48 h, and (**c**) 72 h post-treatment compared to the control (\*\* *p* < 0.01, \*\*\* *p* < 0.001, \*\*\*\* *p* < 0.0001). Asterisks denote significant results regarding the respective measurement indicated with the bar. Values are served as mean ± SEM of three independent experiments, n = 30–45 replicates. SEM: Standard error mean.

three independent experiments, n = 30–45 replicates. SEM: Standard error mean. The effects of different combined groups of Paclitaxel plus PEA on soma size of DRG neurons were investigated in comparison to the cells treated with Paclitaxel alone. An increase in soma size was found for 0.01 µM Paclitaxel only at 24 h after treatment compared to control group. A total of 10 µM PEA combined with 0.01 µM Paclitaxel was the only group that demonstrated a significant decrease in neurons soma size in comparison The effects of different combined groups of Paclitaxel plus PEA on soma size of DRG neurons were investigated in comparison to the cells treated with Paclitaxel alone. An increase in soma size was found for 0.01 µM Paclitaxel only at 24 h after treatment compared to control group. A total of 10 µM PEA combined with 0.01 µM Paclitaxel was the only group that demonstrated a significant decrease in neurons soma size in comparison to Paclitaxel group (*p* < 0.05) at 24 h after application, while 0.1 and 1 µM PEA did not show any protective effects against 0.01 µM Paclitaxel group (*p* > 0.05) at the same time point (Figure S6).

to Paclitaxel group (*p* < 0.05) at 24 h after application, while 0.1 and 1 µM PEA did not show any protective effects against 0.01 µM Paclitaxel group (*p* > 0.05) at the same time point (Figure S6). The 0.1 and 10 µM PEA co-applied with 0.1 µM Paclitaxel revealed an apparent decrease in neuronal somata sizes, whereas 1 µM PEA had no effect compared to neurons treated with Paclitaxel only at 24 h after treatment (*p* < 0.05) (Figure 8a). At 48 h post-treatment, 1 and 10 µM of PEA combined with 0.1 µM of Paclitaxel were the only The 0.1 and 10 µM PEA co-applied with 0.1 µM Paclitaxel revealed an apparent decrease in neuronal somata sizes, whereas 1 µM PEA had no effect compared to neurons treated with Paclitaxel only at 24 h after treatment (*p* < 0.05) (Figure 8a). At 48 h posttreatment, 1 and 10 µM of PEA combined with 0.1 µM of Paclitaxel were the only groups with a significant neuroprotective effect on somata sizes when compared to 0.1 µM Paclitaxel (*p* < 0.05) (Figure 8a). At 72 h after application, all PEA concentrations combined with 0.1 µM Paclitaxel showed considerable protectant action except for the 10 µM PEA + 0.1 µM Paclitaxel group when compared to Paclitaxel only (Figure 8a).

groups with a significant neuroprotective effect on somata sizes when compared to 0.1 µM Paclitaxel (*p* < 0.05) (Figure 8a). At 72 h after application, all PEA concentrations combined with 0.1 µM Paclitaxel showed considerable protectant action except for the 10 µM PEA + 0.1 µM Paclitaxel group when compared to Paclitaxel only (Figure 8a). Treatment of neurons with the three different concentrations of PEA co-applied with 1 µM Paclitaxel demonstrated a significant protective effect on somata sizes when compared to cells exposed to 1 µM Paclitaxel alone at 24- and 72-h time points (*p* < 0.05) (Figure 8b). At 48 h post-treatment, 1 µM of Paclitaxel in combination with 10 µM of PEA was the only group without any protective effects, whereas the other two combined groups led to a strong decrease when compared to individual Paclitaxel treated cells (*p* < 0.05) (Figure 8b).

**Figure 8.** Effects of different PEA concentrations (0.1, 1, and 10 µM) in combination with different concentrations of Paclitaxel (**a**) 0. 1 µM, (**b**) 1 µM, and (**c**) 10 µM on soma sizes of DRG neurons at 24, 48, and 72-h post-treatment. The combined groups of Paclitaxel plus PEA demonstrated a varied significant decrease in DRG neuronal cell bodies in comparison with neurons treated with Paclitaxel only (\* *p* < 0.05, \*\* *p* < 0.01, \*\*\* *p* < 0.001, \*\*\*\* *p* < 0.0001). Data represented as (mean ± SEM), and the experiments were performed at least 3 independent times with n = 30–45 replicas. The asterisk denotes significant results regarding the respective measurement indicated with the bar graphs.

The three concentrations of PEA co-applied with 10 µM Paclitaxel showed statistically significant protective effects on DRG neurons' cell bodies against the 10 µM Paclitaxel group at 24 h after treatment (*p* < 0.05) (Figure 8c). Similarly, both concentrations, 0.1 and 1 µM of PEA, combined with 10 µM of Paclitaxel showed a significant neuroprotective effect on soma sizes, while the combination of 10 µM Paclitaxel and 10 µM PEA did not have any protective effect at 72 h post-treatment compared to Paclitaxel alone (*p* < 0.05) (Figure 8c). The 10 µM Paclitaxel plus 1 µM PEA group demonstrated a significant protective effect; however, 0.1 and 10 µM of PEA combined with 10 µM Paclitaxel groups did not reveal any protective effect on the size of cell bodies of neurons when compared to neurons treated only with 10 µM Paclitaxel at 48 h after treatment (*p* < 0.05) (Figure 8c).

Overall, the neuroprotective actions of PEA against the induced toxicity of Paclitaxel on soma size of DRG neurons were time and concentration independent (*p* < 0.05).

#### **4. Discussion**

Peripheral neuropathy (PN) is one of the most common side effects of Paclitaxel, affecting up to 97% of all gynecological and urological cancer patients [46,47]. Paclitaxel causes cell death in cancer cells by interfering with mitosis via microtubule stabilization; however, Paclitaxel also affects the peripheral nervous system, causing PN [48]. The primary symptoms are hand and foot numbness besides pain caused by Paclitaxel accumulation in the DRG. DRG neurons are highly susceptible to Paclitaxel accumulation presumably due to a more permeable blood nerve barrier [49]. In the current study, the toxicity of Paclitaxel on viability of neurons was apparent at only 72 h post-treatment in comparison to the control group, while, at all-time windows studied, different Paclitaxel concentrations resulted in a significant reduction in neurite length of DRG neurons. These findings were in line with previous research on Paclitaxel-induced peripheral neuropathy with axonal sensory neuropathy that was length-dependent [50]. A significant reduction in neurite length was reported when DRG neurons were exposed to 10 µM Paclitaxel for 24 h [17]. In addition, Paclitaxel's toxic effects on neurons resulted in the enlargement of neuronal cell bodies obviously at 24, 48, and 72 h post-treatment. These findings are in agreement with previous findings, in which Paclitaxel treatment caused a significant increase in DRG neuron soma size after 24 h of treatment [15,16] and induced a significant enlargement of DRG nucleolus size [51].

Our data obviously demonstrated that Paclitaxel neurotoxicity on neurite outgrowth and soma size is time- and dose-dependent. Similar earlier studies reported on a dose- and infusion time-dependent-induced neurotoxicity that could be exacerbated by underlying conditions or co-application with other drugs [52,53]. The differences in toxic effects of Paclitaxel on the viability and morphology of neurons might possibly be due to higher susceptibility or vulnerability of neurites to toxins than neuronal somata [15,16]. As a result, the toxic effects of Paclitaxel were more rapid, with a significant reduction in neurite length of neurons after 24 h post-treatment. The data are in agreement with a previous work reporting on reduction of axon length after Paclitaxel treatment. Therefore, Paclitaxel seems to act directly on axons and causes axonal degeneration probably through local mechanisms [54]. Additionally, Paclitaxel disrupted intracellular microtubules and bindings with beta-tubulin inside cell soma, resulting in accumulation of non-functional beta-tubulin units of microtubules, vacuolization of mitochondria and cytoplasm in neuronal cell bodies, and cell enlargement [55]. Paclitaxel increased cell size, and, after 72 h of treatment, neurons may explode and die due to a non-apoptotic effect. Taken together, Paclitaxel targets the nerve fibers and causes local axonopathy in still viable neurons with increased soma sizes.

PEA is a bioactive lipid that is used as an anti-nociceptive agent in different animal models of neuropathic pain, including spinal cord injury [56] and diabetes-induced peripheral neuropathy [57]. In humans, PEA accumulates in painful tissues, as observed in the trapezius muscle of women suffering from chronic neck pain [58]. Moreover, PEA protected nerve tissue in neuropathic conditions [37] and prevented neurotoxicity and neurodegeneration [59,60]. Furthermore, PEA also alleviated painful diabetic neuropathy, chemotherapy neuropathy, idiopathic axonal neuropathy, nonspecific neuropathy, and sciatic and lumbosacral spine disease pain [61]. Here, we demonstrated that PEA partially counteracted the toxicity of Paclitaxel on DRG neurons. Regardless of PEA concentration, combining PEA with Paclitaxel significantly increased neuron cell viability compared to treatment with Paclitaxel at 72 h after treatment. These results were consistent with a previous study that showed a reduction of positive propidium iodide (PI) neuronal nuclei after the application of PEA to N-methyl D-aspartate (NMDA)-treated organotypic hippocampal slice cultures [36]. The positive effects on cell viability seem not to be confined to neurons. In astrocytes, different PEA concentrations increased the cell viability from 30 min to 18 h [62]. The data imply for Paclitaxel the need for a longer interaction with damaged cells. Paclitaxel's slow action might provide a good opportunity for PEA to exert protective effects and reverse the toxic effects of Paclitaxel, resulting in increasing the viability of DRG neurons.

In the present study, PEA plus Paclitaxel groups showed a significant increase in neurite length and a strongly decreased soma size of DRG neurons at all studied time points when compared to individual Paclitaxel treatment groups. Interestingly, at 24 h after treatment, PEA produced a protective effect on neurite length and size of cell bodies of neurons against toxicity of Paclitaxel, independent of PEA concentration. This phenomenon might be attributed to the short period of exposure of DRG neurons to Paclitaxel and PEA treatment, allowing PEA to mask and alleviate the toxicity of Paclitaxel. In line with previous evidence in a rat model of Oxaliplatin-induced neurotoxicity, acute intraperitoneal administration of PEA (30 mg kg-1) substantially relieved pain 30 min after administration [63].

Notably, 10 µM PEA did not show any significant protective effect on neurite outgrowth at 48 h post-treatment, although it enhanced neurite extension at 72 h after treatment. The results might be interpreted as an attempt of damaged neurons to develop a survival pressure to resist death caused by Paclitaxel toxicity. They retract and aggregate short neurites. Therefore, neurite extension might become a secondary process at 48 h post-treatment and, as a result, PEA remains unable to express any protective effects. These data are agreed with previous studies on PEA effects on preserving myelin sheet thickness and axonal diameter and preventing myelin degeneration [37]. PEA reduced myelin loss caused by sciatic nerve injury, maintained neuron cell diameters, reduced nerve edema, and restored nerve function, all of which were associated with decreased hypersensitivity [64].

In summary, PEA induced strong neuroprotective actions against Paclitaxel toxicity in DRG neurons and improved their viability and morphology.

#### **5. Conclusions**

Our findings showed the ability of PEA to attenuate the toxicity of Paclitaxel on DRG neurons. The effects of Paclitaxel on neuronal viability alone were apparent at 72 h posttreatment only. Furthermore, treatment with Paclitaxel led to a strong reduction in neurite length and enlargement of neuronal cell bodies at all investigated time windows. PEA showed neuroprotective effects by partially reversing the toxic effects of Paclitaxel, including increasing cell viability, enhancing DRG neuron neurite outgrowth, and decreasing swelling of neuronal soma. These findings contribute to our understanding of Paclitaxel's site and mode of action on the peripheral nervous system and highlight the critical need for novel peripheral neuropathy protective strategies. More research will be needed to elucidate the signaling pathways underlying PEA's neuroprotective effects against Paclitaxel neurotoxicity. With these results, PEA might be a promising therapeutic option for cancer patients suffering from CIPN.

**Supplementary Materials:** The following supporting information can be downloaded at: https:// www.mdpi.com/article/10.3390/biom12121873/s1. Figure S1: Representative example for tracing neurites by ImageJ program; Figure S2: Effects of different PEA concentrations on neuronal cell viability (%) (mean ± SEM) at 72 h post-treatment; Figure S3: Effects of different concentrations of PEA on neurite lengths and soma sizes (mean ± SEM) of DRG neurons at 24, 48, and 72-h posttreatment; Figure S4: Effects of different PEA concentrations (0.1, 1, and 10 µM) combined with 0.01 µM Paclitaxel on neurite length of DRG neurons at 24, 48, and 72-h post-treatment; Figure S5: Effects of different PEA concentrations (0.1, 1, and 10 µM) combined with 0.1 µM Paclitaxel at 24, 48, and 72-h post-treatment; Figure S6: The effects of different PEA concentrations (0.1, 1, and 10 µM) co-applied with 0.01 µM Paclitaxel on soma sizes of DRG neurons at 24, 48, and 72-h post-treatment.

**Author Contributions:** Conceptualization, F.D. and A.E.; methodology, A.E.; formal analysis, A.E.; investigation, A.E.; resources, F.D.; data curation, A.E.; writing—original draft preparation, A.E. and F.D.; writing—review and editing, F.D. and A.E.; visualization, A.E.; supervision, F.D.; project administration, F.D. All authors have read and agreed to the published version of the manuscript.

**Funding:** This research received no external funding.

**Institutional Review Board Statement:** All animal experiments were carried out in accordance with the policy on ethics and the policy on the use of animals in neuroscience research, as specified in directive 2010/63/EU of the European Parliament and of the Council of the European union on the protection of animals used for scientific purposes and were approved by local authorities for laboratory animal care and use (State of Saxony-Anhalt, Germany, permission number: I11M27).

**Informed Consent Statement:** Not applicable.

**Data Availability Statement:** The data presented in this study are available on request from the authors.

**Acknowledgments:** We acknowledge Katholischer akademischer Ausländer-Dienst (KAAD) for the Ph.D. scholarship to A.E. Authors would like to thank Chalid Ghadban and Candy Rothgänger-Strube for excellent technical assistance and Tim Hohmann, Marc Richard Kolbe, Urszula Hohmann, Johanna Marie Dela Cruz, Derek C Molliver, and Steven Fagan for helpful discussions. We acknowledge the financial support of the Open Access Publication Fund of the Martin Luther University Halle-Wittenberg.

**Conflicts of Interest:** The authors declare no conflict of interest.

#### **References**

