*Article* **Participation of L-Lactate and Its Receptor HCAR1/GPR81 in Neurovisual Development**

**Samuel Laroche <sup>1</sup> , Aurélie Stil <sup>1</sup> , Philippe Germain <sup>1</sup> , Hosni Cherif <sup>1</sup> , Sylvain Chemtob 2,3 and Jean-François Bouchard 1,\***


**Abstract:** During the development of the retina and the nervous system, high levels of energy are required by the axons of retinal ganglion cells (RGCs) to grow towards their brain targets. This energy demand leads to an increase of glycolysis and L-lactate concentrations in the retina. L-lactate is known to be the endogenous ligand of the GPR81 receptor. However, the role of L-lactate and its receptor in the development of the nervous system has not been studied in depth. In the present study, we used immunohistochemistry to show that GPR81 is localized in different retinal layers during development, but is predominantly expressed in the RGC of the adult rodent. Treatment of retinal explants with L-lactate or the exogenous GPR81 agonist 3,5-DHBA altered RGC growth cone (GC) morphology (increasing in size and number of filopodia) and promoted RGC axon growth. These GPR81-mediated modifications of GC morphology and axon growth were mediated by protein kinases A and C, but were absent in explants from *gpr81*−*/*<sup>−</sup> transgenic mice. Living *gpr81*−*/*<sup>−</sup> mice showed a decrease in ipsilateral projections of RGCs to the dorsal lateral geniculate nucleus (dLGN). In conclusion, present results suggest that L-lactate and its receptor GPR81 play an important role in the development of the visual nervous system.

**Keywords:** lactate; GPR81; HCAR1; retinal ganglion cells; growth cone; dLGN; retina; axon; 3,5-DHBA

#### **1. Introduction**

The physiological significance of lactic acid and its conjugate base lactate have been a major source of controversy since their discovery in biological tissues. Lactic acid was long considered to be simply the waste product of anaerobic glycolysis. Mainly occurring as the L-enantiomer in physiological conditions, lactate is now known to have multiple effects on cell homeostasis, serving as a metabolic fuel and buffering agent, while also acting as a signaling molecule, also known as "Lactormone". This signaling action is obtained via the hydroxycarboxylic acid receptor 1 (HCAR1) [1,2]. Also known as GPR81, this is a G-protein-coupled receptor activated by L-lactate and the exogenous agonist 3,5 dihydroxybenzoic acid (3,5-DHBA) [3]. GPR81 is expressed in diverse organs, including adipose tissues, skeletal muscle, liver, kidney, brain, and retina [4,5].

The retina and central nervous system have an inherently high energy demand due to the continuous depolarization of neuronal membranes [6]. During the development of the visual system, the axons of the retinal ganglion cells (RGCs) that form the optic nerves follow chemotropic molecules to grow toward and across the optic chiasma [7]. In mammals, most of the axons cross at the optic chiasm to reach the contralateral side of the optic tract while only axons from RGCs on the ventro-temporal side of the retina

**Citation:** Laroche, S.; Stil, A.; Germain, P.; Cherif, H.; Chemtob, S.; Bouchard, J.-F. Participation of L-Lactate and Its Receptor HCAR1/GPR81 in Neurovisual Development. *Cells* **2021**, *10*, 1640. https://doi.org/10.3390/cells10071640

Academic Editor: Stefan Liebau

Received: 22 May 2021 Accepted: 22 June 2021 Published: 30 June 2021

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

**Copyright:** © 2021 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/).

do not cross at the optic chiasm. Ventro-temporal RGCs specifically express the ephrin receptor (EphB1) on their GC. When the GCs come into contact with the Ephrin-B2 ligand secreted by radial glial cells in the midline of the optic chiasm, a chemorepulsive effect redirects them and the axons project to the ipsilateral optic tract [8]. Following the optic tract, some of RGC axons join the dorso-lateral geniculate nucleus (dLGN) in the thalamus, making synapses with dLGN neurons. These neurons project their axons to the layer 4 of the primary visual cortex [9].

The consequent high metabolic demand for ATP production is met by glycolysis, resulting in L-lactate generation, despite a rich endowment of mitochondria for aerobic respiration. ATP generation by oxidative phosphorylation is limited and high energy demand leads to an increase of glycolysis filling the ATP equilibrium [2,10]. It has been previously reported that the lactate dehydrogenase (LDH-1) subunit, which preferentially catalyzes the conversion of L-lactate to pyruvate, is found in neurons and astrocytes [11,12]. However, LDH-5, which is more highly expressed in astrocytes, preferentially favors the conversion of pyruvate back to L-lactate [11]; LDH-5 is also present in glycolytic tissues such as skeletal muscles [13,14]. This cell-type distribution of LDH subunits indicates a close link with characteristic features of L-lactate metabolism, which has implications for neuronal development, growth, and survival [6,11].

GPR81 has been reported to mediate effects of L-lactate in diverse processes, including wound healing, angiogenesis, neuroprotection, cancer cell survival, attenuation of inflammation, and antilipolytic effects [5,15–20]. Recent studies have revealed the presence of GPR81 in Müller cells and retinal ganglion cells (RGCs) [21]. Its activation in these cells regulates angiogenic Wnt ligands and Norrin, which together participate in intra-retinal vascularization. Receptors specific for other metabolic intermediates have also been shown to govern angiogenesis and neuronal growth in the visual system. Thus, receptors for the Krebs cycle intermediates succinate (GPR91) and α-ketoglutarate (GPR99) control vascular and axonal growth [22,23]. Hence, there is a close link between metabolism and cell signaling in regulating the development of vascular and neuronal systems. However, the role of L-lactate and its cell membrane target GPR81 during the development of the central nervous system (CNS) is yet to be explored. In this study, we built upon established findings in the CNS to investigate the role of GPR81 during development of the visual system. More specifically, we investigated the effects of L-lactate/GPR81 on axonal growth guidance in the developing mouse neuro-visual system.

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

#### *2.1. Ethics Statement and Animals*

All animal procedures were performed in accordance with the Animal Care Committee of the University of Montreal following the guidelines from the Canadian Council on Animal Care. The *gpr81*−/<sup>−</sup> mice were purchased from Lexicon Pharmaceuticals (The Woodlands, TX, USA). These mice were developed by the insertion of a 4-kb IRESlacZ-neo cassette in the trans-membrane domain 2 coding region (100 base pairs) of the *gpr81* gene on C57BL/6J mice [20], with control studies in C57BL/6J WT control mice. Other experiments were undertaken in golden Syrian hamsters (Charles River Laboratories, Saint-Constant, QC, Canada). All animals were maintained in an environmentally controlled room held at 21 ± 2 ◦C, and with 12 h dark/light circle. Mice and hamsters of both sexes were used in this study. Food and water were provided ad libitum.

#### *2.2. Genotypic Screening*

Mice were genotyped by PCR using the amputated tail tip for DNA extraction. Tail samples were immersed in 50 mM NaOH, incubated for 20 min at 95 ◦C, vortexed, and neutralized with 1 M Tris-HCl, pH 8.0. The samples were then revortexed and centrifuged for eight minutes at 13,000 rpm in a Fisher ScientificTM accuSpinTM Micro R centrifuge. The supernatant was taken for DNA amplification and added to the PCR reagent mix containing PCR Buffer, MgCl2, dNTP mix, Taq DNA polymerase, forward and reverse

primers. PCR cycle conditions were: 5 min at 95.0 ◦C, 30 cycles of three steps (1 min at 50.0 ◦C, 1 min at 72.0 ◦C and 1 min at 95.0 ◦C). Genotype of the *Gpr81*−*/*<sup>−</sup> mice was confirmed using specific primers for wild-type (WT), with normal mouse DNA as a control. The primer pairs w (forward: 5′ -CATCTTGTTCTGCTCGGTCA–3′ and reverse: 5′ - GAGGAAGTAGAGCCTAGCCA-3′ ) were used to amplify a 160 bp fragment present in the *gpr81*+/+ genome but absent in the *gpr81*−/<sup>−</sup> mice.

#### *2.3. Reagents*

L-(+)-lactic acid, bovine serum albumin (BSA), ciliary neurotrophic factor (CNTF), DNase, forskolin, insulin, laminin, poly-D-lysine, progesterone, selenium, putrescine, gelatin from porcine skin, chromium(III) potassium sulfate dodecahydrate, sucrose, sodium chloride (NaCl), PCR reagent mix, potassium chloride (KCl), hydrochloric acid (HCl), disodium hydrogen phosphate (Na2HPO4), potassium phosphate monobasic (KH2PO4), *gpr81* primers, mouse anti-Brn-3a (MAB1585), rabbit anti-GPR81-S296 (SAB1300790), mouse anti-MAP2 (M9942), transferrin, trypsin, and triiodothyronine were purchased from Sigma Aldrich (Oakville, ON, Canada). B27, N2, fetal bovine serum (FBS), neurobasal medium (NB), penicillin-streptomycin, Minimum Essential Medium Eagle medium, Spinner Modification (S-MEM), and sodium pyruvate were purchased from Life Technologies (Burlington, ON, Canada). The standard goat serum, Peroxidase-AffiniPure Donkey anti-rabbit IgG, and Peroxidase-AffiniPure Donkey anti-mouse IgG were from Jackson ImmunoResearch (West Grove, PA, USA). N-acetyl cysteine (NAC) was acquired from EMD technologies (Saint-Eustache, QC, Canada). 3,5-DHBA was purchased from Tocris (Oakville, ON, Canada). Cholera toxin subunit B (CTb) recombinant conjugate with Alexa Fluor 555 (C34776) and 647 (C34778), GlutaMAX™ Supplement, green Neg-50 Frozen section medium, Hank's 1X Balanced Salt Solutions (HBSS), Tween 20, Triton X-100, paraformaldehyde (PFA), sucrose, TEMED, sodium hydroxide (NaOH), Tris base, Taq DNA Polymerase, Shandon Immu-Mount, AlexaFluor donkey anti-mouse 488, Alexa Fluor 488 goat anti-rabbit IgG (H+L), AlexaFluor goat anti-rabbit 546, Alexa Fluor 546 goat anti-mouse, and Alexa Fluor 546 phalloidin were obtained from Fisher Scientific (Ottawa, ON, Canada). Rabbit anti-PKA C-α (#4782), rabbit anti-Phospho-PKA C (Thr197) (#4781), rabbit anti-PKCα (#2056), and rabbit anti-Phospho-PKC (pan) (βII Ser660) (#9371) were acquired from Cell Signalling Technology (Whitby, ON, Canada). Heparin and sterile saline solution (0.9%) were purchased from CDMV (St-Hyacinthe, QC, Canada).

#### *2.4. Tissue Preparation for Immunohistochemistry*

Adult mice and postnatal day 5 (P5) golden Syrian hamsters were euthanized by an isoflurane overdose. A transcardial perfusion was conducted with 10 U/mL of heparin in 60 mL of phosphate-buffered 0.9% saline (PBS; 0.1 M, 4 ◦C, pH 7.4), followed by 60 mL phosphate-buffered 4% paraformaldehyde 4 ◦C (PFA). Following the harvesting of mouse embryos E14-16 extraction by Cesarian, they were deeply anesthetized by hypothermia. The orbits were removed and two small holes were made in the cornea prior to immersion fixation in 4% PFA 4 ◦C for 60 min. The eyecups were washed in PBS, cryoprotected in 30% sucrose overnight, embedded in Neg 50 medium, flash-frozen, and kept at −80 ◦C until processing. Sections 14 µm-thick were cut with a cryostat (Leica Microsystems, Exton, PA, USA) and mounted on slides coated with gelatin/chromium (double-frosted microscope slides, Fisher Scientific, Ottawa, ON, Canada).

#### *2.5. Immunohistochemistry*

Frozen sections were thawed, washed three times for five minutes each time in 10 mM PBS with 0.05% Tween 20(PBST), and then blocked in 1% BSA, gelatin, and 0.5% Triton X-100 in PBS for one hour. The sections were then co-incubated overnight with rabbit anti-GPR81 and mouse anti-Brn3a (a specific marker for RGCs). The next morning, sections were washed three times for five each time minutes in PBST and incubated for one h with secondary antibodies: AlexaFluor donkey anti-mouse 488 and AlexaFluor goat antirabbit 546. After three washes in PBST, the sections were slide mounted with Shandon Immu-Mount.

#### *2.6. L-Lactate Solution*

Daily fresh solution of 100 mM L-lactate was made using 0.1 g of L-(+)-lactic acid (MW = 90.08 g/mol) dissolved in 10 mL of NB (vehicle). The pH was adjusted to 7.4 ± 0.1 with NaOH, the volume reached 11 mL with NB, and the solution was filtered with Corning™ PES 0.20 µm pore Syringe Filters (09-754-29). The prepared L-lactate solution was equilibrated for at least one h at 37 ◦C in a 5% CO<sup>2</sup> incubator before use in experiments in vitro.

#### *2.7. Retinal Explant Culture*

The retina were isolated from E14-E16 mouse embryos, dissected into small segments in HBSS, and plated on 12 mm diameter glass coverslips previously coated with poly-D-Lysine (20 µg/mL) and laminin (5 µg/mL) placed in 24-well plates. The explants were cultured in NB medium supplemented with 100 U/mL penicillin/streptomycin, 5 µg/mL NAC, 1% B27, 40 ng/mL selenium, 16 µg/mL putrescine, 0.04 ng/mL triiodothyronine, 100 µg/mL transferrin, 60 ng/mL progesterone, 100 µg/mL BSA, 1 mM sodium pyruvate, 2 mM glutamine (glutaMAXTM), 10 ng/mL CNTF, 5 µg/mL insulin, and 10 µM forskolin at 37 ◦C and 5% CO2. At 0 DIV, starting 1 h following plating, the explants were treated for 15 h for projection analysis or for 1 h at one day in vitro (DIV) for the growth cone morphology analysis. Photomicrographs were taken with a Olympus IX71 microscope (Olympus, Markham, ON, Canada) and the axonal projection and growth cone measurement analysis were made using ImageJ software. These analyses were performed by operators blind to the experimental condition.

#### *2.8. Growth Cone Behavior Assay*

Similarly, embryonic retinal explants were cultured in Thermo Scientific™ Nunc™ Lab-Tek™ Chambered Coverglass with a borosilicate glass bottom (Lab-Tek; Rochester, NY, USA). After one or two DIV, explants were transferred to an incubator (Live cell chamber) at 37 ◦C and 5% CO2, mounted on an inverted Olympus IX71 microscope (Olympus, Markham, ON, Canada). A micropipette was positioned at a 45◦ angle about 100 µm from the growth cone of interest, as previously described [24–26]. A micro-injector (Harvard Apparatus, St-Laurent, QC, Canada) was used to deliver NB vehicle or L-lactate 20 mM (pH 7.4) at a rate of 0.1 µL/min of in the NB. Measurements were performed with ImageJ software at baseline and after 60 min of these treatments.

#### *2.9. Primary Neuron Culture*

Primary cortical neurons were used in this study because of the ease in culturing and harvesting sufficient numbers for biochemical assays, which is technically difficult for RGCs. C57BL/6J WT pregnant mice were used to obtain E14-16 embryo brains. The superior layer of each cerebral cortex was isolated and transferred to 2 mL S-MEM containing 2.5% trypsin and 2 mg/mL DNase and incubated at 37 ◦C for 15 min. After trituration, the pellet was transferred to 10 mL S-MEM with 10% FBS and stored at 4 ◦C. The pellet was again transferred in 2 mL S-MEM supplemented with 10% FBS and triturated three or four times. The supernatant was transferred to 10 mL NB medium. Dissociated neurons were counted under a microscope and plated at a density of 100,000 cells per well on 12 mm glass coverslips previously coated with poly-D-lysine (20 µg/mL) for immunocytochemistry, or at 250,000 cells per 35 mm Petri dish for western blot analysis. Neurons were cultured for two-four days in NB medium supplemented with 1% B-27, 100 U/mL penicillin/streptomycin, 0.25% N2, and 0.5 mM of glutaMAXTM. They were then treated with the GPR81 agonist L-lactate for 1 h to study acute effects on growth cone morphology, or for various other intervals for identifying the activation of signaling pathways by western blot analysis.

#### *2.10. Immunocytochemistry*

After treatment, retinal explants and primary cortical neuron cultures were washed with PBS (pH 7.4), fixed in 4% PFA (pH 7.4), and blocked with 2% normal goat serum (NGS) and 2% BSA in PBS containing 0.1% Tween 20 (pH 7.4) for 30 min at room temperature. The samples were then incubated overnight at 4 ◦C in a blocking solution containing anti-GPR81, anti-MAP2. The following day, the samples were washed and labeled with Alexa Fluor 488 and 546 secondary antibodies against the host species of the primary antibodies, Hoechst 33258, or AlexaFluor Phalloidin 546. The coverslips were mounted with Shandon Immu-Mount and photomicrographs were taken with an Olympus IX71 microscope (Olympus, Markham, ON, Canada) for quantitative analysis on ImageJ software.

#### *2.11. Western Blot Analysis*

Following L-lactate treatment at various time points, the primary cortical neurons were washed with cold 4 ◦C PBS (pH 7.4) and then lysed with 125 µL of laemmli sample buffer pre-warmed to 100 ◦C. The samples were then frozen at −20 ◦C until the day of Western blot analysis. On that day, the samples were thawed at 4 ◦C, placed in a 100 ◦C dry bath for 10 minutes, quickly vortexed and then centrifuged at 13,000 rpm at 4 ◦C. Twenty microliter samples were then resolved on a 10% SDS-polyacrylamide gel along with the trihalo compound from TGX Stain-Free Technology (Bio Rad, Mississauga, ON, Canada). During the electrophoresis, the trihalo compound covalently modifies tryptophan residues in the proteins to impart a fluorescence signal. Visualization of this signal was obtained by UV excitation in a Chemidoc imaging system (Bio-Rad). After this activation, gels were transferred onto a PVDF membrane with a TransBlot Turbo (Bio-Rad), and blots were imaged on the Chemidoc to reveal the total protein transferred on the membrane, which was used for later normalization of antibody signals to total protein. The blots were then blocked with 2% BSA in TBST (Tris 10 mM and NaCl 150 mM saline with 0.1% Tween 20) for 1 h and incubated overnight with antibodies against Phospho-PKA, PKA, Phospho-PKC, and PKC. They were then exposed to the species-appropriate HRPcoupled secondary antibodies for two h in blocking buffer, and dected using the Chemidoc with Clarity Max ECL substrate from Bio-Rad. The target protein expressions were then analyzed on the Image Lab v.6.0.1 software. All procedures were completed according to Bio-Rad protocols [27].

#### *2.12. Eye Specific Segregation*

In this in vivo experiment, *gpr81+/+* and *gpr81*−*/*<sup>−</sup> adult mice under anesthesia received an intraocular injection of 2 µL of 1% (mg/mL) CTb in 0.9% sterile saline conjugated to AlexaFluor 555 into the left vitrious humor eye, and with CTb conjugated to AlexaFluor 647 into the right vitrious humor eye. Four days after the injections, the animals were anesthetized and perfused transcardially, as described in the tissue preparation section above. The brains were removed, post-fixed in PFA 4% overnight, and then cryoprotected by immersion in gradient sucrose solutions of 10%, 20%, and 30% until the brains sank, followed by storage at −80 ◦C. Coronal sections 40 µm thick were cut on glass slides in a cryostat, air-dried, and mounted with Shandon Immu-Mount. Fluorescent images of entire brain sections were taken using 561 and 640 nm laser emissions and a 4× objective on an Fluoview FV3000 Olympus Confocal microscope to identify the three sections containing the largest ipsilateral projection. The dorsal lateral geniculate nucleus (dLGN) was scanned in these sections with a 20× objective and 2× amplification (total magnification of 40×). Zstacks of 19 images with both lasers were taken from top to bottom of the signal emissions. The colocalization of both channels with a multi-threshold analysis was performed on CellSens Dimension software. The percentage of ipsilateral signal overlapping with the contralateral signal was measured for each stack. The mean for the 19 stacks in each section was reported for each threshold, and differences evaluated by two-way ANOVA with Tukey post hoc testing [28,29].

#### *2.13. Statistical Analysis*

Data were imported in GraphPad Prism 8 software. Tests for normal distribution were performed by an Anderson–Darling test (α = 0.05). Depending on parametric or nonparametric distribution, the appropriate statistical analysis was computed. Values were reported as the mean ± SEM.

#### **3. Results**

#### *3.1. GPR81 Is Expressed in the Retina*

We used rabbit anti-GPR81-S296 (SAB1300790) immunohistochemistry to identify the retinal laminar distribution of GPR81. Results in retina from P5 Syrian gold hamster pups, E16 mouse embryos, and adult mice all showed retinal GPR81 expression. GPR81 protein immunoreactivity was consistently detected in the Outer Nuclear Layer (ONL), Inner Nuclear Layer (INL), Inner Plexiform Layer (IPL), RGCs, and the RGC fiber layer of hamsters (Figure 1A–C). GPR81 immunoreactivity was detected in all layers of the embryonic and adult mouse retina (Figure 1D-I). The colocalization of GPR81 with the specific RGC marker Brn-3a showed that GPR81 is predominantly expressed in RGCs and in the RGC fiber layer. From this result, we inferred possible involvement of GPR81 in retinal development and in projection navigation towards brain innervation targets.

μ **Figure 1.** GPR81 is Expressed in the Retina. Expression of GPR81 in RGCs of retinal sections from P5 hamster pups (**A**–**C**) E16 WT mouse embryo (**D**–**F**), and adult mouse (**G**–**I**). ONL: outer nuclear layer, RGC: retinal ganglion cells, OPL: outer plexiform layer, INL: inner nuclear layer, NBL: neuroblast layer, IPL: inner plexiform layer. The white arrows indicate colocalization of GPR81 and Brn-3a. Scale bars: 50 µm.

Primary cortical neurons and retinal explant cultures in vitro also expressed the lactate receptor; GPR81 fluorescence signal was visualized on the neuronal soma, neurites, GC, and filopodia (Figure 2A, Figure S1).

*− − − −* μ μ **Figure 2.** L-lactate and 3-5-DHBA influence GC dynamics and increase axon growth via GPR81 agonism. (**A**) GPR81 is expressed in GCs in vitro. (**B**) Microscopic photography of GCs following 1 h treatment with GPR81 agonists. Explants were marked with phalloidin conjugated with Alexa fluor 546. (**C**) Analysis of GC surface area and filopodia number (*n* = 48–147 GCs per condition). Values are presented as the means ± SEM. (**D**) Microscopic photography of retinal explants following 15 h treatment with the GPR81 agonists. Explants were marked with phalloidin conjugated with Alexa fluor 546. (**E**) Analysis of RGC axon projections( *n* = 278–2572 axons per condition). Values are presented as the means ± SEM. # Indicates significant changes compared to the control group. \* Indicates significant changes *p* < 0.05 between *gpr81+/+* and *gpr81*−*/*<sup>−</sup> genotype. \*\* Indicates significant changes *p* < 0.01 between *gpr81+/+* and *gpr81*−*/*<sup>−</sup> genotype. Statistical test used is the Krustal–Wallis nonparametric ANOVA. White scale bar: 100 µm. Yellow scale bar 25 µm.

#### *3.2. GPR81 Influences GC Morphology and Axon Growth*

μ A number of GPCRs exert effects on axon guidance [22,24–26,30,31]. To determine the implication of GPR81 in regulating GC morphology, we cultivated embryonic retinal explants during 1 DIV. Afterwards, treatment for one h with the GPR81 agonist L-lactate (10 mM) significantly increased the RGC GC surface area by 44.2 ± 9.6%, whereas 3,5- DHBA (300 µM) increased surface area by 56.1 ± 12.0% compared to the vehicle control. These same treatments increased filopodia numbers of GC by 38.1 ± 8.1% and 31.0 ± 6.5%, respectively on retinal explants obtained from *gpr81*+/+ mice. The corresponding changes

μ

− −

induced by GPR81 agonists were abrogated in similar retinal explants obtained from *gpr81*−/<sup>−</sup> mouse embryos (Figure 2B–C).

We next measured the effect of GPR81 agonists on axon growth of RGCs on retinal explants. Interestingly, RGCs from *gpr81+/+* mice incubated with GPR81 agonists for 15 h increased axon outgrowth by 40.7 ± 2.2% (10 mM L-lactate) and 38.7 ± 4.2% (300 µM 3,5- DHBA) (Figure 2D–E). As expected, RGCs obtained from *gpr81*−/<sup>−</sup> mice did not respond to the agonist treatment (−4.2 ± 3.7% and 4.8 ± 5.0%, respectively) (Figure 2D–E). This impact of L-lactate on GC morphology and axon growth was confirmed with time-lapse live cell imaging in the GC behavior assay. Here, we exposed a GC of a retinal explant to a microgradient of L-lactate concentration to visualize any axon growth or guidance effects. In this experiment, we measured the GC surface area, filopodia numbers, axon length, and the angle between the GC and the micropipette at baseline (T0) and at 60 min of drug exposure (T60). The time-dependent differences were obtained by subtraction. At 60 min, the vehicle-treated explants (NB medium) showed negligible changes in the GC (−0.42 ± 3.03 µm<sup>2</sup> area), 0.5 ± 0.5 filopodia, and −1.39 ± 2.78 µm axon growth; in contrast, L-lactate-treated explants displayed mean increases of 22.8 ± 11.1 µm<sup>2</sup> GC area, 4.0 ± 1.8 in number of filopodia, and 4.31 ± 2.14 µm in axon growth. However, we did not observe any significant changes in the angle of orientation after the addition of lactate (−2.1 ± 2.0◦ for NB medium versus 2.5 ± 3.5◦ for 20 mM L-lactate gradient; Figure 3). − − − − μ − μ μ μ −

μ μ μ **Figure 3.** L-lactate influences GC morphology and axon growth, but not turning. (**A**) Time-lapse microscopy time of a DIV1-2 mouse RGC growth cone before and at 60 min after the addition of L-lactate. The black arrows indicate the microgradient direction created by a micropipette with a tip of 2 µm diameter and positioned at a 45◦ angle and around 100 µm distal to the growth cone. The blue arrows indicate the tips of the growth cone before the application of L-lactate or vehicle (T0). The white arrows indicate the position of the growth cone 60 min later (T60). (**B**) Analysis of growth cone dynamics. The values of the growth cone area, filopodia number on the growth cone, axon length, and the angle of the growth within the L-lactate microgradient were measured before (T0) and at 60 min (T60) after the addition of L-lactate or vehicle. Bar graphs show the means ± SEM of the effects induced by the treatments (*n* = 6–8 cells per condition). \* Indicates significant changes compared to the vehicle group, compared by parametric unpaired *t*-test (*p* < 0.05). Scale bar = 15 µm.

#### *3.3. Lactate Increases PKC and PKA Phosphorylation*

GPR81 is coupled to G<sup>i</sup> , such that agonism induces Ca2+ release in CHO-K1 adipocytes, resulting in decreased lipolysis [32]. To determine the signaling pathway resulting in GPR81 activation, we treated primary neurons obtained from mouse embryos with a GPR81 agonist (10 mM L-lactate). Interestingly, there was an increase of PKC phosphorylation (2.1 ± 0.4 fold) after five minutes and an increase of PKA phosphorylation (1.9 ± 0.2 fold) at 15 min after the addition of 10 mM L-lactate (Figure 4A). These data show that GPR81 agonism stimulates the PKC and the PKA signaling pathways in primary neurons.

**Figure 4.** L-lactate increases PKC and PKA phosphorylation (**A**,**B**) Protein expression levels of P-PKC, PKC, P-PKA, PKA, and to-tal proteins in primary neuron cultures incubated with L-lactate (10 mM) at 37 ◦C for five min (for PKC) and 15 min (for PKA). The total protein Western blots confirmed equal loading in all lanes with TGX Stain-Free Technology. Values of relative density to vehicule are presented as the means ± SEM (**C**) for P-PKC and (**D**) for P-PKA. \* Indicates significant changes compared to the control group, as compared by the nonparametric ANOVA Krustal Wallis (*p* < 0.05) with Dunn's post hoc correction.

#### *3.4. GPR81 Affects Retinothalamic Projections In Vivo*

*− − − −* We next studied the contribution of GPR81 to the development of retinothalamic projections in living transgenic *gpr81*−*/*<sup>−</sup> mice. Left and right RGC projections were visualized with the neuron tracer CTb conjugated to Alexa fluorescent molecules. RGC projections normally segregate through various brain structures to join dLGN [33]. Interestingly, adult *gpr81*−*/*<sup>−</sup> mice had significantly fewer ipsilateral projections in the dLGN, irrespective of the intensity threshold used for the analysis (Figure 5). This observation demonstrates, for the first time, an essential role of GPR81 in the normal development of the rodent retinothalamic pathway. *− − − −*

*− −* **Figure 5.** GPR81 Affects Retinothalamic Projections in vivo. (**A**) Micrographs of the dorsal lateral geniculate nucleus (dLGN) following the injection of CTb-Alexa 555 in the left eye and CTb-Alexa 647 in the right eye of *gpr81+/+* and *gpr81*−*/*<sup>−</sup> mice. Controlateral and ipsilateral projections from both

eyes RGC segregate in the dLGN. RGC projections from both eyes are shown in the merged images. (**B**) Line graph shows the percentage of ipsilateral projections overlapping in the contralateral projections expressed as the mean ± SEM. Percentage of overlapping signals is significantly lower in *gpr81*−*/*<sup>−</sup> mice irrespective of the intensity analysis threshold. \*\* Indicates significant changes between *gpr81+/+* (*n* = 18) and *gpr81*−*/*<sup>−</sup> (*n* = 12) genotypes, according to two-way ANOVA with Tukey post hoc test (*p* < 0.01). Scale bars: 100 µm.

#### **4. Discussion**

Lactic acid and L-lactate have drawn considerable attention since the early days of metabolic research. Indeed, the indisputable importance of lactate in the glycolytic pathway may have led to persisting misconceptions about the bodily functions of L-lactate [1]. The introduction of the astrocyte-neuron Lactate shuttle hypothesis (ANLSH) [34] led to a more mature understanding of the importance of L-lactate metabolism in the CNS. According to this model, glucose is predominantly taken up by astrocytes during neuronal activation. The astrocytes produce and release L-lactate, which serves as the primary metabolic fuel for activated neurons [35]. Furthermore, glucose consumption by the outer retina forms L-lactate by aerobic glycolysis [36,37], in metabolic support of RGCs. The recent discovery of receptors for metabolic intermediates has brought new physiologic perspectives on the relevance of L-lactate concentrations in tissue. We undertook in this study to explore the localization and neural functions of GPR81 in the developing retina and its CNS projections. We show for the first time expression of the lactate receptor GPR81 in the retina of hamsters and mice, noting its high abundance in the RGC layer and in the RGC nerve fiber layer. This result was confirmed by studies in vitro using retinal explants and cortical neurons of embryonic mice, which both proved to express GPR81 on their soma, axons, GCs, and GC filopodia. We demonstrated that specific activation of the GPR81 receptor by its agonists modulates GC size, filopodia numbers, and axon growth of RGCs. In vitro, treatment with L-lactate at a physiological concentration increased phosphorylation of PKC and PKA, which are key mediators of the pathways that likely promote neurite growth effects [38]. Importantly, genetic absence of GPR81 curtailed formation of RGC projections to the dLGN, as evident in the *gpr81*−*/*<sup>−</sup> mouse in vivo. Altogether, these results suggest an essential involvement of L-lactate and its receptor GPR81 in the development of RGCs and their axon projections in the CNS.

Previous findings for GPR81 impart an extra layer of complexity with regard to its involvement in CNS development, especially for retinal projections. The colocalization of GPR81 with glutamine synthetase, a specific marker of Müller cells, was previously established in early postnatal mice (P8-17). GPR81 activation in Müller cells regulates Norrin and Wnt ligands, which in turn promote angiogenesis and produce stimulate production of neuroprotective growth factors in the retina [21]. In our present embryonic model (E15), Müller cells have not yet been differentiated in the developing retina. Accordingly, the mechanism responsible for axon growth at this development stage likely differs from the retinal angiogenic effects produced by Müller cells in postnatal rodents [39].

L-lactate was long considered to be a metabolic waste of anaerobic glycolysis, but it is now known to be a metabolic fuel of aerobic glycolysis, especially in neurons [1,11,40]. In this capacity, L-lactate is converted by LDH to pyruvate, which then enters into the Krebs cycle for ATP generation in mitochondria [41]. In order to confirm that the present findings of L-lactate effect are not due trivially to an increase in the energy metabolite pool, we used 3,5-DHBA as a non-metabolized specific GPR81 agonist [3]. Activation of GPR81 by L-lactate and 3,5-DHBA of RGCs in retinal explants from embryonic mice positively modulated GC morphology and increased axon projection length. Interestingly, these effects of GPR81 agonist treatment were absent in explants from *gpr81*−*/*<sup>−</sup> mice. These findings support our claim that L-lactate and 3,5-DHBA act on RGCs through GPR81 agonism.

RGCs in the intact organism project their axons from retina to transmit visual information to CNS targets. GCs must detect and respond to a complex combination of chemotactic

signals to direct their axons through the visual pathway. Our time-lapse microscopy experiments confirmed that L-lactate concentration gradients modify GC morphology and increase axon growth without having a significant effect on axon guidance *per se*. However, L-lactate was without effect in explants from retinal mice genetically depleted of GPR81, thus suggesting a crucial role for GPR81 in the retinothalamic pathway formation.

We show in this study that L-lactate activates PKC and PKA through agonism at GPR81, as was reported previously. Indeed, PKC activation is well-known to participate in synaptic remodeling, presynaptic plasticity, and neuronal repair [38,42,43]. Some PKC isoforms are known to influence neurite adhesion and outgrowth in various neuronal cell types, including RGCs [40,42–44]. Activation of these second messenger pathways results in a reorganization of the growth cone cytoskeleton, which manifest to microscopic examination of the filopodia, lamellipodia, and axonal growth in vitro [45–47]. Our present results are thus also in accordance with studies showing that PKA activation increases GC area and filopodia numbers to promote RGC axon extension during development [48].

In conclusion, as demonstrated in this study for the first time, the GPR81 L-lactate receptors contribute critically to neuro-visual development in the rodent retina and CNS. This study thus provides a foundation for the ongoing investigation of these actors in broader aspects of central and peripheral nervous system development. Present results support the formulation of new hypotheses for elaborating effective therapies targeting the development and regeneration of the nervous system. In fact, our results suggest a possible therapeutic role of L-lactate and other GPR81 agonists in the neuroregenerative and neuroreparative field. First of all, L-lactate can act by itself as an efficient energy substrate [49]. It can also protect and attenuate neuronal death induces by oxygen and glucose deprivation following cerebral ischemia [50,51]. In traumatic brain injury model, preconditioning with L-lactate promotes plasticity-related expression helping to reduce neurological deficits effect via the GPR81 signaling pathway [52]. In accordance with our findings, the utilization of L-lactate or a GPR81 agonist could stimulate axon regeneration and prevent neuronal degeneration.

**Supplementary Materials:** The following are available online at https://www.mdpi.com/article/10 .3390/cells10071640/s1, Figure S1: GPR81 is expressed in DIV2 E15 cortical neuron.

**Author Contributions:** Conceptualization, S.L., A.S., and J.-F.B.; methodology, S.L., A.S., H.C., and J.-F.B.; software S.L. and P.G.; validation, A.S. and J.-F.B.; formal analysis, S.L.; investigation, S.L. and J.-F.B.; resources, J.-F.B. and S.C.; writing—original draft preparation, Samuel Laroche; writing review and editing, S.L., S.C., H.C., and J.-F.B.; supervision, J.-F.B.; project administration, J.-F.B.; funding acquisition, J.-F.B. and S.C. All authors have read and agreed to the published version of the manuscript.

**Funding:** This research was funded by the National Sciences and Engineering Research Council of Canada (NSERC, RGPIN-2020-05739 and the Canadian Institutes of Health Research (CIHR, PJT-156029) grants to Jean-François Bouchard.

**Institutional Review Board Statement:** The study was conducted according to the guidelines of the Institutional Review Board of Comité de déontologie de l'expérimentation sur les animaux ((CDEA) (20-052) date of approval: May 2020).

**Informed Consent Statement:** Not applicable.

**Data Availability Statement:** All relevant data are within the paper and its Supporting Information files.

**Acknowledgments:** We thank the Sylvain Chemtob laboratory members Xiaojuan Yang, Xin Hou, Sonja L'Espérance, and Christiane Quiniou for their precious help to acquire *gpr81*−*/*<sup>−</sup> mice. We thank Bruno Cécyre for his technical support and Paul Cumming for critical reading of the manuscript.

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

### **References**


### *Article* **Autophagy Involvement in the Postnatal Development of the Rat Retina**

**Noemi Anna Pesce 1,2, Alessio Canovai <sup>2</sup> , Emma Lardner <sup>1</sup> , Maurizio Cammalleri <sup>2</sup> , Anders Kvanta <sup>1</sup> , Helder André 1,\* ,† and Massimo Dal Monte 2,†**


**Abstract:** During retinal development, a physiologic hypoxia stimulates endothelial cell proliferation. The hypoxic milieu warrants retina vascularization and promotes the activation of several mechanisms aimed to ensure homeostasis and energy balance of both endothelial and retinal cells. Autophagy is an evolutionarily conserved catabolic system that contributes to cellular adaptation to a variety of environmental changes and stresses. In association with the physiologic hypoxia, autophagy plays a crucial role during development. Autophagy expression profile was evaluated in the developing retina from birth to post-natal day 18 of rat pups, using qPCR, western blotting and immunostaining methodologies. The rat post-partum developing retina displayed increased active autophagy during the first postnatal days, correlating to the hypoxic phase. In latter stages of development, rat retinal autophagy decreases, reaching a normalization between post-natal days 14-18, when the retina is fully vascularized and mature. Collectively, the present study elaborates on the link between hypoxia and autophagy, and contributes to further elucidate the role of autophagy during retinal development.

**Keywords:** eye; retina; development; vascularization; hypoxia; autophagy

#### **1. Introduction**

The mature retina is considered one of the highest oxygen-demanding tissues in the body, with a considerable metabolic activity [1,2]. The heightened metabolic demand of the retina is supplied by a structured vascular systems, including retinal vessels and the choriocapillaris, which provide nutrients and oxygen to the inner and the outer layers of the retina respectively [3,4]. During development of the mammalian eye, the retinal vasculature undergoes considerable changes and reorganization [5]. In the early stages of embryogenesis, the interior of the eye is metabolically supplied by a transient embryonic circulatory network in the vitreous, referred to as the hyaloid system [6]. In the latter stages of development, the hyaloid vasculature regresses and concurrently is replaced by the retinal vasculature [7]. The physiologic hypoxia in uterus (O<sup>2</sup> levels < 5%) drives the proliferation of retinal blood vessels from the optic nerve to the periphery [8], through vascular endothelial growth factor (VEGF)-mediated angiogenesis [9,10].

At this level, the developing retinal vasculature lacks a functional barrier, necessary to maintain homeostasis into the retina and controlling vascular permeability [11,12]. Thus, the retinal capillary endothelial cells interact with each other to create a complex network, composed of tight junctions between transmembrane and peripheral membrane proteins. In this manner, the retinal endothelial cells form an inner blood retinal barrier (BRB),

**Citation:** Pesce, N.A.; Canovai, A.; Lardner, E.; Cammalleri, M.; Kvanta, A.; André, H.; Dal Monte, M. Autophagy Involvement in the Postnatal Development of the Rat Retina. *Cells* **2021**, *10*, 177. https://doi.org/10.3390/cells10010177

Received: 18 December 2020 Accepted: 14 January 2021 Published: 17 January 2021

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

**Copyright:** © 2021 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/).

which contributes to preserve neuronal environment regulating the entry of molecules from the blood into the retina [13,14]. At these critical times in developmental events, both retinal and endothelial cells (ECs) endure morphological changes and reorganization; as consequence, they require mechanisms for the degradation and recycling of obsolete cellular components [15].

Autophagy is an essential process in maintaining the normal cellular homeostasis under physiological conditions [16]; and it plays an important role in the turnover of damaged organelles, such as peroxisomes and endoplasmic reticulum, as well as in removing unnecessary aggregated or misfolded proteins [17,18]. Previous findings indicate that vascular remodeling in ocular development can be regulated by autophagy [19]; the blood vessels need autophagy to balance their bioenergetic dynamic mechanisms [20]. Moreover, autophagic mechanisms seem to have a critical role in anatomical involution of the hyaloid blood vessels [19]. Several studies have demonstrated that autophagy can be induced by physiological hypoxia, the key stimulus for retinal angiogenesis [21–23]. At the molecular level, hypoxia stimulates several molecules involved in different signaling pathways, including hypoxia-inducible factors (HIFs) that induce angiogenesis through VEGF; and adenosine monophosphate-activated protein kinase (AMPK), a positive regulator of autophagy [24,25]. In conditions where nutrients are scarce, such as during development, AMPK is activated by a decreased ATP/AMP ratio and leads to phosphorylation of several molecules, including Unc-51-like autophagy activating kinase (ULK)1 [26,27]. Activated ULK1 is involved in the formation of multiple protein complexes that are responsible for the initiation of autophagic mechanisms that lead to the formation of autophagosome [28,29]. Elongation and maturation of the autophagosome involve the microtubule-associated protein I light chain 3 (LC3 I) system. LC3 I is conjugated to phosphatidylethanolamine, converted to LC3 II and inserted into the autophagosome membrane [30]. The synthesis and processing of LC3 II is increased during autophagy, thus acting as a key marker of levels of autophagy in cells [31]. The cargo is selected by targeted ubiquitination and carried to the autophagosome through the binding of LC3 II with sequestosome-1 (SQSTM-1), also known as ubiquitin-binding protein p62 [32,33]. The p62 is degraded by autophagy and a decrease in its protein levels correlates with an active autophagic flux [34]. Autophagy ends with the fusion of the autophagosome with the lysosome, where the inner cargo is degraded by lysosomal hydrolases.

Considering the myriad of autophagic mechanisms, the aim of the present study was to examine the changes of expression of autophagy markers in the developing retina in postnatal rats. Due to its postnatal development and accessibility, the rat retinal vasculature warrants a bonafide model to assess vascular developmental autophagy mechanisms from birth through postnatal day (P) 18 when retinal vasculature has attained its adult pattern.

#### **2. Material and Methods**

#### *2.1. Animals and Ethics Statements*

After birth, 84 Wistar rat pups were maintained with their nursing mothers through the experimental times P7, P14 and P18 in a regulated environment (24 ± 1 ◦C, 50 ± 5% humidity), with a 12 h light/dark cycle and provided with food and water. Rat pups were euthanized with an intraperitoneal injection of 30 mg/kg of pentobarbital. All animal protocols were in accordance with the Statement for the Use of Animals in Ophthalmic and Vision Research (ARVO), the Italian regulation for animal care (DL 116/92), and the European Communities Council Directive (86/609/EEC). Animal procedures were authorized by the Ethical Committee in Animal Experiments of the University of Pisa (permit number: 133/2019-PR, 14 February 2019).

#### *2.2. Vascular Labeling*

A total of 24 rat pups of different ages (birth, P7, P14 and P18; six rats for each time point) were used to prepare whole-mount and retina sections. Isolated retinas were fixed in 4% paraformaldehyde in 0.1 M phosphate buffer, pH 7.4 (PB), at room temperature for

3 h. Subsequently, retinas were washed three times (5 min per wash) in PB and incubated for 1 h at room temperature (RT) in blocking buffer (PB containing 10% donkey serum and 0.5% Triton X-100; Sigma-Aldrich, St. Louis, MO, USA) to prevent non-specific labeling. Sequentially, retinas were incubated with fluorescein-labelled isolectin B4 (1:200; Vector Laboratories, Burlingame, CA, USA) in blocking solution, at 4 ◦C overnight (ON). Finally, after three washes with PB, retinas were placed onto a slide, mounted and covered with a coverslip. Immunostaining was observed by a digital fluorescence microscope (Ni-E; Nikon-Europe, Amsterdam, The Netherlands) and immunofluorescent images of the retinal vasculature were acquired using a digital camera (DS-Fi1c; Nikon-Europe). The vascular area and the total area were measured using ImageJ freeware. Vascular area was reported as percentage of the total area.

#### *2.3. Western Blot Analysis*

Proteins were extracted from retinas of 24 rat pups at different ages (birth, P7, P14 and P18; six rats for each time point) using RIPA lysis buffer (Santa Cruz Biotechnology, Dallas, TX, USA), supplemented with phosphatase inhibitor (Sigma-Aldrich) and protease inhibitor (Roche, Mannheim, Germany) cocktails. Protein extracts were quantified by the microBCA method (Thermo Fisher Scientific, Waltam, MA, USA) and 15 µg of total proteins were separated by SDS-PAGE and transferred onto polyvinylidene difluoride (PVDF) membranes (Bio-Rad Laboratories, Hercules, CA, United States). Membranes were blocked either with 5% of skim milk in Tris-buffered saline (TBS-T; Bio-Rad Laboratories, containing 0.05% Tween-20; Sigma-Aldrich) or with 4% of Bovine Serum Albumin (BSA; Sigma-Aldrich) in TBS-T, for 1 hour at RT. Subsequently, the membranes were incubated at 4 ◦C ON with primary antibodies: anti-HIF-1α (1:500, rabbit polyclonal, cat. no. NB100479; Novus Biologicals, Centennial, Colorado, USA); anti-pAMPKα (Thr172, 1:500, rabbit monoclonal, cat. no. 2535S; Cell Signaling Technology); anti-AMPKα (1:1000, rabbit monoclonal, cat. no. 5832S; Cell Signaling Technology); anti-pULK1 (Ser555; 1:500, rabbit monoclonal, cat. no. 5869S; Cell Signaling Technology, Danvers, MA, USA); anti-ULK1 (1:1000, rabbit monoclonal, cat. no. 8054S; Cell Signaling Technology); anti-LC3 I and II (1:1000, rabbit polyclonal, cat. no. 4108S; Cell Signaling Technology); anti-p62 (1:1000, rabbit polyclonal, cat. no. ab-91526; Abcam, Cambridge, UK); and anti-β-actin (1:5000, rabbit monoclonal, cat. no. SAB5600204; Sigma-Aldrich). Secondary antibody anti-rabbit-IgG conjugated to horseradish peroxidase (1:10,000, cat. no. P044801-2; Dako, Carpinteria, CA, USA) was incubated for 1 h at RT. Following the incubation with both primary and secondary antibodies the membranes were washed with TBS-T, three times for 5 min. Finally, the protein of interest was visualized using the Clarity Western ECL substrate with a ChemiDoc XPS<sup>+</sup> imaging system (Bio-Rad Laboratories, Hercules CA, USA). The optical density (OD) of the bands was determined with the Image Lab 3.0 software (Bio-Rad Laboratories). Protein levels were corrected to the β-actin loading control or nonphosphorylated proteins.

#### *2.4. Quantitative PCR*

Total RNA was extracted and purified from retinas of 24 rat pups at different ages (birth, P7, P14 and P18; six rats for each time point), using the Trizol® reagent (Invitrogen, Waltam MA, USA), resuspended in RNAse-free water, and quantified by spectrophotometry (BioSpectrometer basic; Eppendorf AG, Hamburg, Germany). First-strand cDNA was generated from 1 µg of total RNA (QuantiTect Reverse Transcription Kit; Qiagen, Hilden, Germany). Quantitative PCR was performed with a kit (SsoAdvanced Universal SYBR Green Supermix; Bio-Rad Laboratories) on a CFX96 Real Time PCR Detection System (equipped with the CFX manager software (Bio-Rad Laboratories). Forward and reverse sequence of primers were chosen to hybridize to unique region of the appropriate gene sequence: occludin-1 (Forward: 5'-TTTCATGCCTTGGGGATTGAG-3'/Reverse: 5'-GACTTCCCAGAGTGCAGAGT-3'; Invitrogen ); zonula-occludens-1 (ZO-1; Forward: 5'- AGTCTCGGAAAAGTGCCAGG-3'/Reverse: 5'-GGGCACCATACCAACCATCA-3′ ; Invit-

rogen); VEGF-A (Forward: 5'-CAATGATGAAGCCCTGGAGTG-3'/Reverse: 5'-AGGTTTG ATCCGCATGATCTG-3; Invitrogen) and Ribosomal Protein L13a (Rpl13a; Forward: 5′ - GGATCCCTCCACCCTATGACA-3′/Reverse: 5′ -CTGGTACTTCCACCCGACCTC-3′ ; Invitrogen). Samples were compared by the threshold cycle analysis (Ct) and absolute expression values were calculated using the 2−∆∆Ct formula, with Rpl13a as the housekeeping gene.

#### *2.5. Immunohistofluorescence*

In total, 12 rat pups at different ages (birth, P7, P14 and P18; three rats for each time point) were fixed in 4% paraformaldehyde in PB at RT for 5 days prior to immunofluorescence staining. Sections of 4 micrometers were processed for immunohistofluorescence in a Bond III robotic system (Leica Biosystems, Newcastle, UK), as previously described [35]. According to the manufacturer's instructions, antigen retrieval was performed in 10 mM citrate buffer pH6 and sections were incubated with primary antibody step with anti-LC3 I and II (1:100) and isolectin-biotin (1:1000; cat. no. I21414; Thermo Fisher Scientific), while secondary antibody steps included Streptavidin-Alexa 488 (1:500; cat. no. S11223, Thermo Fisher Scientific), anti-rat-Alexa 546 (1:500; cat. no. A11081, Invitrogen) and anti-goat-Alexa 647 (1:500; cat. no. SAB4600175, Sigma-Aldrich). Sections were mounted by using a vector Vectashield with DAPI mounting medium (Vector Laboratories, CA, Burlingame, USA), and visualized with an Axioscope 2 plus with the AxioVision software (Zeiss, Gottingen, Germany).

#### *2.6. Statistical Analysis*

The statistical analyses were performed using Prism software (GraphPad software, Inc., San Diego, CA, USA), applying one-way ANOVA with Bonferroni's multiple comparisons posttest. Data were presented as mean ± standard error of mean (SEM) of *n* = 6; *p* values < 0.05 were considered statistically significant.

#### **3. Results**

#### *3.1. The Rat Retina Is Partially Avascular at Birth*

Fluorescein-labeled isolectin B4 was used to assess the progress of retinal vascularization in the rat. As previously described, rat retinal vascularization is almost completed around P13-P16 [36]. As depicted in Figure 1A, at birth the hyaloid vasculature was still present and the retina remained partially avascular, as determined by measuring the vascular area as less than 50% of total area of the retina (Figure 1B). Quantitative analysis demonstrated a significant difference at P7, where the retina was 90% vascularized (*p* < 0.001) and in the late stages of ocular blood vessel development, P14-P18 (*p* < 0.001), where the retinas were fully vascularized (100%) as compared to birth.

#### *3.2. Different Expression of Blood-Retina Barrier Genes during Rat Retina Development*

In retinal blood vessels, ECs present tight junctions that function as a part of the BRB, fundamental to maintain retinal homeostasis and to mediate selective diffusion of molecules from the circulation to the retinal tissue [14,37]. Consequently, analysis of the expression of BRB genes in the rat retinas at birth, P7, P14 and P18 was performed by qPCR (Figure 2). Transcript expression levels of occludin-1 and ZO-1 were upregulated at P14 and P18, compared to birth and P7 (*p* < 0.01 both). These results confirmed that at birth and P7 the rat retinal vasculature was still under development, presenting an incomplete vascular network.

Scale bar = 500 μm. followed by Bonferroni's multiple comparisons test ( **Figure 1.** Development of the vascular network in rat retina. (**A**) Visualization of blood vessels by isolectin B4 staining of rat pup retinas at birth, postnatal day (P)7, P14 and P18. Dashed circle delineates the developing retinal vasculature (inner of the dashed circle) with the hyaloid vasculature (outer of the dashed circle). Scale bar = 500 µm. (**B**) Quantitative analysis of vascular area of the retina. Data is presented as mean ± SEM. One-way ANOVA was used as statistical analysis, followed by Bonferroni's multiple comparisons test (*n* = 6; \*\*\* *p* < 0.001 vs. birth).

#### *3.3. At P7 the Rat Retina Is Hypoxic*

During retinal development, physiologic hypoxia induces the activation of HIF-1α, which promotes the transcription of *VEGF-A* gene. This process is pivotal to induce endothelial cell proliferation and migration to form the vasculature network [38]. In this context, Western blotting was performed to analyze HIF-1α protein levels in the rat retina (Figure 3A) at birth, P7, P14 and P18. Densitometric analysis showed an increase of HIF-1α protein levels at P7 (*p* < 0.001) compared to birth, followed by a decrease at P14 (*p* < 0.01 vs. birth; *p* < 0.01 vs. P7) and at P18 (*p* < 0.01 vs. P7; Figure 3B).

Since HIF-1α promotes the transcription of *VEGF-A* gene, expression analysis was performed by qPCR in rat retinas at birth, P7, P14 and P18. As depicted in Figure 3C, VEGF-A was upregulated at P7, (*p* < 0.001 vs. birth) while at P14 and P18 its levels were comparable to those at birth, in agreement with the reduced levels of HIF-1α at the latter postnatal days.

1α protein levels at P7 (

cal analysis, followed by Bonferroni's multiple comparisons test ( **Figure 2.** Expression of blood-retina barrier genes during retina rat development. mRNA expression of occludin-1 (**A**) and zonula occludens (ZO-1) (**B**) genes was evaluated by qPCR in rat retinas at birth, P7, P14 and P18. Data is presented as mean ± SEM. One-way ANOVA was used as statistical analysis, followed by Bonferroni's multiple comparisons test (*n* = 6; \*\* *p* < 0.01 vs. birth, §§ *p* < 0.01 vs. P7).

1α and β 1α. followed by Bonferroni's multiple comparisons test was used as statistical analysis of mean ± SEM **Figure 3.** Protein levels of hypoxia-inducible factor (HIF) change during rat retina development. (**A**) Western blots illustrate representative immunoreactive bands of HIF-1α and β-actin (loading control) in the retina of rat pups from birth to P18. (**B**) Quantitative analysis of optical density of the immunoreactive bands of HIF-1α. (**C**) mRNA expression of vascular endothelial growth factor (*VEGF)-A* gene evaluated with qPCR in rat retinas at birth, P7, P14 and P18. One-way ANOVA followed by Bonferroni's multiple comparisons test was used as statistical analysis of mean ± SEM datasets (*n* = 6; \* *p* < 0.05, \*\* *p* < 0.01, \*\*\* *p* < 0.001 vs. birth, §§ *p* < 0.01 vs. P7).

α

#### *3.4. Autophagic Mechanisms Increased during the Hypoxic Phase*

During development, autophagic mechanisms support cells to adapt and respond to several processes, including proliferation, differentiation and migration [39]. In this context, Western blotting analysis was performed to evaluate the levels of proteins involved in autophagy during retina development (Figure 4A). Densitometric analysis demonstrated a significant increase of phosphorylated levels of AMPKα (*p* < 0.001 vs. birth; Figure 4C) and ULK1 (*p* < 0.01 vs. birth; Figure 4B) as well as LC3 II (*p* < 0.01 vs. birth; Figure 4D) at P7. At P14 and P18, the levels of these autophagic markers decreased then at levels comparable to those at birth.

AMPKα / AMPKα ; LC3 II / β and p62 / β way ANOVA followed by Bonferroni's multiple comparisons test was used as sta-**Figure 4.** Protein levels of autophagic markers during rat retinal development. (**A**) Western blots depict representative immunoreactive bands of proteins involved in autophagic mechanisms in rat retinas, from birth to P18. Quantitative analysis of optical density of the ratio of immunoreactive bands between pSer 555 -Ulk1 / Ulk1 (**B**), *p*-AMPKα / AMPKα (**C**); LC3 II / β-actin (**D**) and p62 / β-actin (**E**). One-way ANOVA followed by Bonferroni's multiple comparisons test was used as statistical analysis of mean ± SEM datasets (*n* = 6; \*\* *p* < 0.01, \*\*\* *p* < 0.001 vs. birth, § *p* < 0.05, §§§ *p* < 0.001 vs. P7).

On the contrary, a trend to a decrease of SQSTM1/p62 protein levels, yet not statistically significative, was observed at P7, followed by a substantial increase at both P14 (*p* < 0.001 vs. birth; *p* < 0.001 vs. P7) and P18 (*p* < 0.001 vs. birth; *p* < 0.001 vs. p7).

#### *3.5. High Expression of Autophagic Marker at P7 in Rat Retina*

In rat, developing retinal cells are already organized into layers from P7, giving the tissue its stratified feature [40,41]. The variation of autophagic flux was evaluated relative to the retinal layers using immunostained retinal sections to visualize autophagy and vasculature, with LC3 as an autophagic marker and isolectin B4 as a marker of endothelial cells. The immunofluorescence demonstrated a clear variation of autophagy during retinal development, indicating a predominant expression of the autophagic marker LC3 at P7 (Figure 5). At this specific time point, a substantial expression of LC3 was observed in both the inner plexiform and outer plexiform layers (IPL; OPL). At birth and on latter stages of retinal developmental, LC3 was predominantly expressed in the IPL and almost undetectable in the OPL. In addition, heightened colocalized expression of the autophagy and vascular markers was observed at P7, as compared to birth, P14 and P18, which could be related to the hypoxia associated with the involution of the hyaloid blood vessels. Albeit a positive staining for LC3 was detected in ganglion cell layer (GCL) and the photoreceptor outer segments (OS), no changes were observed in the studied times of development.

ONL, outer nuclear layer; OS, outer segments of photoreceptors. Scale bar = 50 μm. **Figure 5.** Expression pattern of LC3 in the developing rat retina. Representative immunohistofluorescence analysis of LC3 (red) and isolectin B4 (IB4; green) and Hoechst (Hst; blue) in retina sections of rat pups at birth, P7, P14 and P18. Dashed squares indicate the magnification area of GCL and IPL layers. Arrows represent colocalization of LC3 with IB4 in retinal vasculature. GCL, ganglion cell layer; IPL, inner plexiform layer; INL, inner nuclear layer; OPL, outer plexiform layer; ONL, outer nuclear layer; OS, outer segments of photoreceptors. Scale bar = 50 µm.

#### **4. Discussion**

During mammalian development, cells go through proliferation, cell death and differentiation, culminating in an adult organism formation. During these stages, autophagy assures cell adaptation, by promoting rapid changes in cytosolic composition and accelerating organelle and protein turnover [39,42]. The present study elaborates on the influence of autophagy mechanisms in the development of rat's retina from birth to P18.

In rats, the retinal vasculature develops postnatally. At birth, the retina surface is still covered by the hyaloid vessels, and the retinal vessels have merely begun to raise from the optic disc [43]. Previous studies have shown that the hyaloid network persists until P7, by which time the retinal vessels have almost propagated to the periphery of the retina [44,45]. Interestingly, in newborn rat retinas a large area lacking retinal capillary coverage is observed, indicating the presence of hyaloid vessels, while at P7 the rat retinas display nearly full vascularization. At this time point, the developing retina vascular network is still immature and the BRB is not formed [11]. To form the BRB, endothelial cells require tight junctions, essential to regulate the movement of solutes and nutrients from the outer to the inner retinal layers [46]. Tight junctions comprise several proteins, including occludin-1 and ZO-1, responsible for anchoring the junctional complex to the cytoskeleton [47]. As demonstrated here, at birth and P7 the retina of rat pups presents low transcript levels of both occludin-1 and ZO-1 genes. Reversely, mRNA levels of these genes are increased at P14 and P18, when the retinal vasculature is covering the total surface of retina. This suggests that retina vascularization is completed and mature around P14-P18, in alignment with the presence of tight junctions in retinal endothelial cells.

From embryonic stages to birth, a physiological hypoxia is paramount to drive retinal neovascularization, through the upregulation of HIF-1α and subsequently VEGF-A [9,48]. During the early postpartum retinal developmental stage in rats, the involution of the hyaloid vasculature to the retinal vascular network results in a partially avascular and ischemic retina, correlating to increased oxygen demand and resulting in a hypoxic stimulus [10]. In the present study, a peak of HIF-1α and VEGF-A expression is determined at P7, which decreases during latter stages when the retina is fully vascularized. The hypoxic environment contributes to activate essential mechanisms in adaptation and survival, ensuring cellular homeostasis during angiogenesis [49,50]. In fact, the developing retina exposed to changing environment and metabolic stress requires autophagy to adjust its bioenergetic and biosynthetic demands [20,49]. In this respect, both hypoxia and energy deprivation can promote AMPK activation, a known inducer of autophagy [22,25]. In agreement, an increase of p-AMPKα, p-ULK1 and LC3 II protein levels are demonstrated at P7, with a trend of decreased p62 protein levels, indicating an active autophagic flux at this developmental stage. At P14 and P18, with the presence of a fully vascularized retina, a decrease in p-ULK1 and LC3 II protein levels is determined concomitantly with an increase of p62. The observed accumulation of p62 with a decrease of LC3 II protein levels at P14 could be associated with a transition from autophagy-dependent to -independent mechanisms of retinal homeostasis, as previously suggested in retinal pigment epithelium cells [51,52].

To elaborate on the role of autophagy in the different cell layers during rat retinal development, a predominant expression of LC3 is confirmed in the GCL, IPL, OPL and OS, in agreement with previous studies in rodent retinas [53–55]. At P7, an increase of LC3 staining is denoted in the IPL and OPL, with a noticeable reduction in the autophagy marker at P14 and P18. During the early postpartum developmental stage, the rat retinal cells are affected by ischemia, which is correlated to the peak of expression of HIF-1α and VEGF-A, concomitantly with LC3. Moreover, an expression of LC3 is observed in endothelial cells at P7, suggesting that autophagy may contribute directly to the formation of the retinal vascular network. These findings indicate that during the physiologic hypoxia in the rat retina, HIF-mediated signaling induces the increase in VEGF-A to promote endothelial cell proliferation, and an upregulation of autophagy markers to sustain cellular homeostasis and cellular quality control in retinal cells.

#### **5. Conclusions**

The present study indicates that increased autophagy is intrinsically associated with the hypoxic phase of retinal development and critically contributes to the physiologic development of the different cell layers of the retina during the transition from the hyaloid to the retinal vasculature, thus allowing the normal development of the retina.

**Author Contributions:** Conceptualization, H.A. and M.D.M.; formal analysis, N.A.P.; investigation, N.A.P., A.C., H.A. and M.D.M.; methodology, N.A.P., A.C. and E.L.; resources, M.C. and A.K.; supervision, H.A. and M.D.M.; writing—original draft, N.A.P., H.A. and M.D.M.; writing—review and editing, N.A.P., H.A. and M.D.M. All authors have read and agreed to the published version of the manuscript.

**Funding:** This research was funded by the grants from the Karolinska Institutet Foundations and The Swedish Eye Foundation.

**Institutional Review Board Statement:** Animal protocols were conducted in accordance with the Statement for the Use of Animals in Ophthalmic and Vision Research (ARVO), the Italian regulation for animal care (DL 116/92), and the European Communities Council Directive (86/609/EEC). Animal procedures were authorized by the Ethical Committee in Animal Experiments of the University of Pisa (permit number: 133/2019-PR, 14 February 2019).

**Informed Consent Statement:** Not applicable.

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

**Acknowledgments:** The authors thank Flavia Plastino, Maria Grazia Rossino and Filippo Locri for technical support.

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

#### **References**


### *Article* **Analysis of Programmed Cell Death and Senescence Markers in the Developing Retina of an Altricial Bird Species**

**Guadalupe Álvarez-Hernán 1 , José Antonio de Mera-Rodríguez <sup>1</sup> , Ismael Hernández-Núñez <sup>1</sup> , Alfonso Marzal <sup>2</sup> , Yolanda Gañán 3 , Gervasio Martín-Partido <sup>1</sup> , Joaquín Rodríguez-León 3,\* ,† and Javier Francisco-Morcillo 1,\***


**Abstract:** This study shows the distribution patterns of apoptotic cells and biomarkers of cellular senescence during the ontogeny of the retina in the zebra finch (*T. guttata*). Neurogenesis in this altricial bird species is intense in the retina at perinatal and post-hatching stages, as opposed to precocial bird species in which retinogenesis occurs entirely during the embryonic period. Various phases of programmed cell death (PCD) were distinguishable in the *T. guttata* visual system. These included areas of PCD in the central region of the neuroretina at the stages of optic cup morphogenesis, and in the sub-optic necrotic centers (St15–St20). A small focus of early neural PCD was detected in the neuroblastic layer, dorsal to the optic nerve head, coinciding with the appearance of the first differentiated neuroblasts (St24–St25). There were sparse pyknotic bodies in the non-laminated retina between St26 and St37. An intense wave of neurotrophic PCD was detected in the laminated retina between St42 and P8, the last post-hatching stage included in the present study. PCD was absent from the photoreceptor layer. Phagocytic activity was also detected in Müller cells during the wave of neurotrophic PCD. With regard to the chronotopographical staining patterns of senescence biomarkers, there was strong parallelism between the SA-*β*-GAL signal and p21 immunoreactivity in both the undifferentiated and the laminated retina, coinciding in the cell body of differentiated neurons. In contrast, no correlation was found between SA-*β*-GAL activity and the distribution of TUNEL-positive cells in the developing tissue.

**Keywords:** programmed cell death; cellular senescence; retinogenesis; altricial bird species; precocial bird species; senescence-associated galactosidase activity

### **1. Introduction**

Programmed cell death (PCD) and cellular senescence during vertebrate embryogenesis are transient phenomena that contribute mainly to tissue remodeling [1–3] through the degeneration of temporary structures in the embryo. Indeed, it has been described that PCD processes are accompanied by cell senescence in interdigital regression [4–6], heart morphogenesis [7], pronephros and mesonephros degeneration [8–11], and degeneration of structures in the developing otic vesicle [12–14].

The vertebrate visual system constitutes an excellent model for investigating the mechanisms involved in cell degeneration and the phases of PCD that affect different structures (for a review, see [3]). Areas of intense PCD have been described in the developing visual

**Citation:** Álvarez-Hernán, G.; de Mera-Rodríguez, J.A.; Hernández-Núñez, I.; Marzal, A.; Gañán, Y.; Martín-Partido, G.; Rodríguez-León, J.; Francisco-Morcillo, J. Analysis of Programmed Cell Death and Senescence Markers in the Developing Retina of an Altricial Bird Species. *Cells* **2021**, *10*, 504. https://doi.org/10.3390/ cells10030504

Academic Editor: Maurizio Sorice

Received: 29 January 2021 Accepted: 23 February 2021 Published: 26 February 2021

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

**Copyright:** © 2021 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/).

system in fish [15–19], amphibians [20–22], reptiles [23–25], and mammals [26–31]. With respect to birds, similar studies have been conducted in the chicken [32–38] and in the quail [39], two precocial bird species. In these species, PCD during visual system morphogenesis and retinogenesis is completely restricted to the embryonic period. During early stages of avian eye morphogenesis, two pyknotic zones have been described in the central region of the retinal neuroepithelium and in the dorsal rim of the optic cup [35]. Furthermore, two areas of intense PCD appear in the neuroepithelium located laterally to the optic chiasm, in the so-called sub-optic necrotic centers (SONCs) [40,41]. With the onset of neurogenesis in the neural retina, PCD also affects neuroepithelial cells and newborn ganglion cell neuroblasts [33,36–38]. At later stages, coinciding with the synaptogenesis between retinal neurons, PCD affects those neurons that are unable to successfully innervate their targets [34,37,39]. The last wave of cell death follows different gradients that resemble the spatiotemporal patterns of cell differentiation [34,39].

With regard to developmental cellular senescence, several markers are currently employed to identify the distribution of senescent cells in vertebrate embryos. One of the most commonly used is the histochemical technique that detects the presence of β-galactosidase enzymatic activity at pH 6.0 (senescence-associated β-galactosidase, SA-β-GAL), different from that normally observed at pH 4.0 within lysosomes [42]. Increased expression of intracellular proteins such as p21, p16, p63, and p73 and the Btg/Tob tumor suppressor gene family also identifies cell senescence in several regions of the developing embryo [6]. These markers have been described in different embryonic tissues, but little is known about their distribution in the developing visual system. In this sense, we have recently described that some of these senescence markers are detected not only in several subpopulations of neurons in the developing retina, but also in the retinal pigment epithelium [43,44].

Although the ontogenetic mechanisms involved in visual system development and the basic structure of the retina are similar across bird species, the developmental rate and the acquisition of retinal structures are highly variable. Visual system morphogenesis and retinogenesis occur early in embryogenesis in precocial bird species [45,46], while these ontogenetic processes are delayed in altricial birds [47–49]. This delay can reach the stage of hatching and the first week of life, in which intense postnatal neurogenesis has been detected in the altricial retina [50]. The timing of histogenesis and cell differentiation and the state of retinal maturation at hatching thus differ significantly between precocial and altricial bird species.

All these data suggest that it is necessary to study visual system development across a broad range of avian species to conduct interspecific comparisons that can clarify the ontogenetic patterns. In the present study, we use classical histological, histochemical, and immunohistochemical methods (i) to describe the chronotopographical patterns of cell death and cell senescence markers in the developing visual system of an altricial bird species, the zebra finch (*Taeniopygia guttata*, Vieillot 1817), (ii) to study whether the distribution of senescence markers correlates with the progression of cell death in the *Taeniopygia guttata* retinal tissue, and (iii) to compare these results with those described in other precocial bird species, such as *Gallus gallus* or *Coturnix japonica*, and in the rest of the vertebrates.

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

#### *2.1. Animal and Tissue Processing*

All animals were treated according to the regulations and laws of the European Union (EU Directive 2010/63/EU) and Spain (Royal Decree 53/2013). A total of twenty-seven *T. guttata* embryos and twelve hatchlings were used in the present study (Table 1). Embryos were obtained by incubating eggs in a rotating egg incubator (Masallés S.A., Spain) that was maintained at 37.5 ± 1 ◦C, 80–90% humidity. The degree of development of the embryos and hatchlings (Figure 1) was determined in accordance with the stages (St) established by by [51]. Embryos and hatchlings were fixed with paraformaldehyde (PFA) 4% in phosphate-buffered solution (PBS) (0.1 M, pH 7.4) overnight at 4 ◦C. For histological analysis with toluidine blue staining, some fixed embryos were dehydrated in a graded series of acetone and propylene oxide and embedded in Spurr's resin. Serial frontal 3 µm sections were cut in a Reichert Jung microtome.

For the histochemical and immunohistochemical procedures, embryos and hatchlings were immersed overnight in a cryoprotective solution (15% sucrose in PBS) at 4 ◦C, soaked in embedding medium, and frozen. Cryosections of 20 µm were obtained in a cryostat microtome (Leica CM 1900, Charleston, SC, USA), thaw mounted on SuperFrost Plus slides, air dried, and stored at 20 ◦C.

**Figure 1.** Stereomicroscope images of some embryos and postnatal specimens of *Taeniopygia guttata* showing the external morphological changes of the eye. The embryos were staged in accordance with the developmental stages (St) established by [51]. The optic cup was distinguishable between St15 and St23 (**A**–**D**). Pigmentation in the RPE was observed at St25 (**E**). At St37, the eye was completely pigmented (**F**). From St42 until perinatal stages, the eyelids progressively covered the eye (**G**–**J**). Eyelids were closed at P5 (**K**), but slightly open at P8 (**L**). Scale bars: 2 mm (**A**,**B**); 3 mm (**C**–**E**); 6 mm (**F**); 7 mm (**G**); 10 mm (**H**–**L**).

℃

*β β*

*−* ℃

*β*



#### *2.2. Toluidine Blue Staining*

Morphological analysis of development of cell death was conducted on resin sections stained with toluidine blue 0.5% and sodium tetraborate 0.5% solution. For this purpose, slides were put in the colorant at 90 ◦C for 45 s and then rinsed with distilled water. Sections were mounted with Eukitt (Kindler, Freiburg, Germany).

#### *2.3. Detection of β-Galactosidase Activity*

We followed the protocol described by [52]. Cryosections were incubated in 450 µL of chromogenic SA-*β*-GAL substrate X-gal (5-bromo-4-chloro-3-indolyl-*β*-D-galactopyranoside) in *PBS*−*MgCl*2 at pH 6.0 at 37.5 ◦C for 24 h. A blue-green precipitate was developed by *SA*−*β*− *GAL*-positive cells. Then, sections were washed in *PBS* −*MgCl*2 acid buffer for 10 min. After histochemical reaction, some of the sections were counterstained with DAPI (Sigma-Aldrich, Madrid, Spain, Ref. D9542) and others were used to perform immunohistochemical analyses. Slides were rinsed in PBS and mounted with Mowiol (Polyvinyl alcohol 40–88, Fluka, Madrid, Spain, Ref. 81386).

#### *2.4. Immunohistochemistry*

After histochemical analyses to detect *β*-galactosidase activity, slides were subjected to an antigen retrieval process with citrate buffer (pH 6) at 90 ◦C or 30 min. Sections were chilled at RT for 20 min. Slides were washed several times in 0.1% Triton-X-100 in PBS (PBS-T) and pre-blocked in 0.2% gelatin, 0.25% Triton-X-100, and Lys 0.1M in PBS (PBS-G-T-L) for 1 h.

Sections were incubated with mouse anti-p21 monoclonal antibody (1:200, Abcam, Madrid, Spain, ab109199) overnight at RT in a humidified chamber. The day after, slides were washed several times in PBS-T and PBS-G-T and incubated with Alexa Fluor 488 goat anti-mouse IgG antibody (1:200, Molecular Probes, Eugene, OR, USA, A11029) for 2 h at RT in a humidified chamber in darkness. Sections were washed several times in PBS-T and PBS-G-T in darkness and incubated for 10 min with DAPI at RT, followed by two washes in PBS. Slides were mounted with Mowiol.

#### *2.5. TUNEL Technique*

The TUNEL technique (Tdt-mediated dUTP Nick End Labeling, Sigma-Aldrich, Madrid, Spain, Cat. No. 11 684 795 910), described by [53], is the histochemical technique commonly used to detect apoptotic nuclei. Cryosections were washed in PBS for 15 min at RT and incubated in 10 µg/mL of proteinase K in PBS for 10 min at 37 ◦C. The slides were then washed in PBS and incubated in blocking solution (3% H2O<sup>2</sup> in PBS) for 15 min. Subsequently, sections were washed several times in PBS and then incubated for 60 min at 37 ◦C with TUNEL reaction mixture, consisting of the enzyme terminal

deoxynucleotidyl transferase (TdT) and fluorescein-conjugated nucleotides in a reaction buffer. After rinsing in PBS, sections were incubated in blocking solution (PBS-G-T-L) and covered with the HRP-conjugated anti-fluorescein antibody solution. The apoptotic nuclei were visualized using DAB as a chromogen. The sections were then washed thrice in PBS, dehydrated, and mounted with Eukitt® (Kindler, Freiburg, Germany) for observation. In control sections in which the enzyme TdT was absent from the reaction solution, no stained nuclei were observed.

#### *2.6. Quantification of TUNEL-Positive Nuclei*

Quantification was performed by counting all TUNEL-positive nuclei in micrographs of the central region of the retina. The surface area of the retina in digital microphotographs was measured using the ImageJ free open-source software package (http://rsb.info.nih. gov/ij/ accessed on 28 January 2021). The density profiles were expressed as the mean ± sem of the number of apoptotic nuclei per square millimeter (an/mm<sup>2</sup> ). Similar procedures have been described in the literature [23,34,36]. Statistical analyses were performed using Student's two-tailed *t*-test. Differences between groups were considered as significant (\*) when *p* < 0.05 and (\*\*) when *p* < 0.01.

#### *2.7. Image Acquisition and Processing*

Toluidine blue-stained, TUNEL, and *SA–β–GAL* and immunofluorescence sections were observed with a bright-field and epifluorescence Nikon Eclipse 80i microscope and photographed using an ultra-high definition Nikon DXM1200F digital camera. Images were processed with Adobe Photoshop CS4.

#### **3. Results**

#### *3.1. Programmed Cell Death in the Developing T. guttata Visual System*

In order to identify dying cells in the developing *T. guttata* visual system, we used some of the methods for detecting PCD in embryonic tissues [3]. Light microscopy observation of toluidine blue-stained semi-thin sections revealed pyknotic bodies in the ganglion cell layer (GCL) and in the inner nuclear layer (INL) of the retinal tissue at the hatching day (P0) (Figure 2A–D). Cryosections labeled with DAPI staining identified nuclear condensation in the laminated retina (Figure 2E,E'). Abundant TUNEL-positive nuclei were observed both in the GCL and in the INL (Figure 2F), but also in other eye tissues, such as the lens (Figure 2G) where DNA of cells of the equatorial zone breaks down due to nuclear endodeoxyribonuclease activity [54]. Therefore, PCD was intense and clearly detected in the developing *T. guttata* visual system.

The distribution of pyknotic nuclei and TUNEL-positive bodies was carefully examined from stage 11 (St11), coinciding with the formation of the optic vesicle [48,51], to postnatal day 8 (P8), the last postnatal stage considered in the present study. Pyknotic bodies were absent from the optic anlage from St11 to St14 (not shown). At St15, when the lateral wall of the optic vesicle invaginates to form the optic cup, abundant pyknotic bodies were found in the central undifferentiated neural retina (Figure 3A,B). Moreover, dead cell fragments were observed in two groups of neuroepithelial cells located on either side of the presumptive optic chiasm (Figure 3A,C). Similar areas of cell degeneration have been described in the chicken embryo, the so-called sub-optic necrotic centers (SONCs) [40,41]. The distribution of PCD was similar at St16 in the neuroretina (Figure 3D–G), but the presence of pyknotic bodies in the SONCs (Figure 3E–G) increased notably. Furthermore, pyknotic bodies were also detected in the anterior wall of the lens anlage (Figure 3D).

' **Figure 2.** Programmed cell death in the *T. guttata* retina detected by using various sensitive methods. (**A**–**D**) Transversal semi-thin section of the P0 retina showing pyknotic bodies with morphological features typical of apoptosis after toluidine blue staining. (**E**,**E**') Identification of neuronal cell death in the ganglion cell layer (GCL) and in the inner nuclear layer (INL) (arrowheads) in cryosections of *T. guttata* retinas at P0 stained with DAPI. (**F**,**G**) Eye cryosections of a P0 *T. guttata* hatchling showing intense abundant TUNEL-positive bodies in the GCL and INL (arrowheads in (**C**)) and in the equatorial region of the lens (arrowheads in (**D**)). Abbreviations: GCL, ganglion cell layer; INL, inner nuclear layer; IPL, inner plexiform layer; ONL, outer nuclear layer; OPL, outer plexiform layer. Scale bars: 50 µm (**A**,**E**–**G**), 7 µm (**B**–**D**,**E'**).

**Figure 3.** Pyknotic fragments during visual system development in *T. guttata*. Toluidine blue-stained semi-thin sections were obtained from the heads of embryos at different stages of development. Pyknotic bodies were mainly located in the central neural retina ((**A**), arrows in (**B**)) and in the sub-optic necrotic centers (SONCs) (arrowhead in (**A**), arrows in (**C**)) at St15 in the early optic cup. At St16, pyknotic fragments were restricted to the central neural retina, to the anterior wall of the lens vesicle (arrows in (**D**)), and to the SONCs (arrowheads in (**E**), arrows in (**F**,**G**)). Abbreviations: LP, lens placode; LV, lens vesicle; NE, neuroepithelium. Scale bars: 50 µm.

At St19, sparse pyknotic bodies were detected in the anterior wall of the lens vesicle (Figure 4A). Pyknotic bodies were still detected in the SONCs (Figure 4B,C). The first differentiating retinal neuroblasts in *T. guttata* appeared by St24 [48,49]. At this stage, pyknotic bodies were concentrated in the NbL in a region located dorsally to the optic nerve head (Figure 4D,E). PCD was also detected in the presumptive retinal pigment epithelium (pRPE), adjacent to the region of the distal optic nerve (Figure 4F,G). At St25, pyknotic bodies were concentrated at the level of the distal optic nerve (Figure 4H,I). From St26 (not shown) to St36, pyknotic bodies were sparsely observed, randomly localized throughout the NbL (Figure 4J–L).

**Figure 4.** Pyknotic bodies during visual system development in *T. guttata*. Toluidine blue-stained semi-thin sections were obtained from the heads of embryos at different stages of development. At E19, pyknotic fragments were mainly detected in the anterior wall of the lens vesicle (arrows in (**A**)), but also in the SONCs (arrowhead in (**B**), arrows in (**C**)). At St24 (**D**–**G**), pyknotic bodies were concentrated in retinal regions located dorsally to the optic nerve head (arrows in (**E**)) and in the presumptive pigment epithelium located surrounding the optic nerve head (arrows in (**G**)). At St25 (**H**,**I**), abundant pyknotic fragments were detected in the dorsal region of the distal optic nerve (arrows in (**I**)). Pyknotic bodies were sparse and dispersed throughout the neuroblastic layer (NbL) by St32 (arrow in (**J**)), St34 (arrows in (**K**)), and St 36 (arrow in (**L**)). Abbreviations: LV, lens vesicle; NbL, neuroblastic layer; NE, neuroepithelium; pRPE, presumptive retinal pigment epithelium. Scale bars: 50 µm.

At St37, scattered TUNEL-positive nuclei were found dispersed throughout the NbL (Figure 5A), similar to the distribution of pyknotic nuclei described from St26 to St36. At St42, retinal stratification was evident, and a few TUNEL-positive nuclei were observed in the GCL and in the INL (Figures 5B and 6). The incidence of cell death rose significantly in the GCL between St42 and St44 (Figures 5C and 6) (2 days before hatching), reaching the highest values in this layer by this stage (Figure 6). At P0, the density of TUNEL-positive nuclei in the GCL diminished (Figures 5D and 6), but increased significantly in the INL, reaching a peak at P5 (Figures 5E and 6). At P8, the last stage analyzed, there was a high incidence of cell death in the INL (Figures 5F and 6), but TUNEL-positive nuclei almost disappeared from the GCL (Figure 5F), reaching values close to 0 in this layer (Figure 6).

**Figure 5.** Spatial distribution of TUNEL-positive nuclei in the developing retina of *T. guttata*. Retinal cryosections of embryos and postnatal specimens were treated in accordance with this histochemical technique. Sparse randomly distributed TUNEL-positive nuclei were detected in the NbL at St37 (double arrowheads in (**A**)). At St42, sparse TUNEL-positive nuclei were detected both in the GCL (arrowhead in (**B**)) and in the INL (double arrowheads in (**B**)). TUNEL-positive nuclei were mainly detected in the GCL at St44 (arrowheads in (**B**)), but also in the INL (double arrowhead in (**C**)). TUNEL-positive nuclei progressively diminished from P0 to P8 in the GCL (arrowheads in (**D**,**E**)), but they increased markedly from P0 to P5 in the INL (double arrowheads in (**D**,**E**)). At P8, TUNELpositive nuclei in the INL were less abundant than observed at previous stages (double arrowheads in (**F**)). Abbreviations: GCL, ganglion cell layer; INL, inner nuclear layer; IPL, inner plexiform layer; NbL, neuroblastic layer; ONL, outer nuclear layer; OPL, outer plexiform layer. Scale bars: 50 µm.

**Figure 6.** Quantitative analysis of the density of TUNEL-positive nuclei during *T. guttata* retinal histogenesis. The incidence of cell death rose significantly in the GCL between St42 and St44, reaching the highest values in this layer by this latter stage. From P0 onwards, the density of TUNEL-positive nuclei in the GCL diminished progressively. The INL contained a low density of TUNEL-positive nuclei between St42 and St44. A significant increase in the density of TUNEL-positive nuclei was observed at P0, reaching a maximum at P5. At P8, there was a high incidence of cell death in the INL. Abbreviations: an/mm<sup>2</sup> , apoptotic nuclei per square millimeter; GCL, ganglion cell layer; INL, inner nuclear layer. Asterisks correspond to *p* values; \* *p* < 0.05 and \*\* *p* < 0.01.

At late embryonic stages (St44) (Figure 7A,B) and at P0 (Figure 7C–F), TUNEL-labeling was occasionally detected in the cell somata and in fine processes of radially oriented cells with an apparent intact healthy morphology (Figure 7A–D). Some of the vitreal TUNELpositive processes form endfeet that seemed to be anchored to the inner limiting membrane ILM (Figure 7A,B). In semi-thin sections, pyknotic bodies were found radially aligned in the cytoplasm of cell processes (Figure 7E,F).

Finally, it is important to note that cell death was completely absent from the ONL during all the embryonic stages and postnatal ages analyzed. Furthermore, the chronotopographical distribution of TUNEL-positive nuclei in the developing *T. guttata* retinal tissue from St42 onwards followed central-to-peripheral and vitreal-to-scleral gradients, in concordance with the gradients of cell differentiation described in this altricial bird species [48,55].

**Figure 7.** Spatial distribution of TUNEL-positive elements in the developing retina of *T. guttata*. Retinal cryosections of embryos (St44: (**A**,**B**)) and newly hatched chicks (P0: (**C**,**D**)) were treated in accordance with this histochemical technique. Elongated cell somata located in the INL (arrowheads in (**A**,**C**,**D**)) and fine processes (arrows in (**B**–**E**)) of radially oriented cells were diffusely labeled with this technique. Occasionally, TUNEL-positive Müller cell endfeet were labeled in the vitreal surface of the retina (double arrowheads in (**A**,**B**)). Semi-thin sections treated according to the toluidine blue technique revealed small pyknotic bodies within the cytoplasm of Müller cells (arrows in (**E**,**F**)). Abbreviations: GCL, ganglion cell layer; INL, inner nuclear layer; IPL, inner plexiform layer; ONL, outer nuclear layer; OPL, outer plexiform layer. Scale bars: 50 µm in (**A**,**C**); 10 µm in (**B**,**D**–**F**).

#### *3.2. Senescence Markers in the Developing T. guttata Visual System*

Retinal cryosections of zebra finch embryos and hatchlings were stained with SA-*β*-GAL histochemistry and examined for the appearance of positively stained cells. At St34, the vitreal-most region and the scleral surface of the central NbL appeared faintly stained

with SA-*β*-GAL histochemistry (Figure 8A,B). In contrast, SA-*β*-GAL staining was mainly detected in the scleral region of the peripheral rim of the retina (Figure 8C,D). The staining pattern of SA-*β*-GAL changed with the appearance of plexiform layers. At St43, SA-*β*-GAL labeling was mainly detected in the GCL, amacrine cell layer, and horizontal cell layer (Figure 8E,F). Double labeling with antibodies against p21 (inhibitor of cyclin-dependent kinases), which has been demonstrated to be overexpressed in senescent cells during embryonic development [1,2,4], showed a strong parallelism between the SA-*β*-GAL signal and p21 immunoreactivity (Figure 8E–G). The same staining patterns were detected in the retina of *T. guttata* hatchlings (Figure 8H–J). *β*

*β β*

*β β*

*β*

*β β β β* **Figure 8.** Distribution of SA-*β*-GAL labeling and p21 immunoreactivity in retinal cryosections of embryos (**A**–**G**) and post-hatched specimens (**H**–**J**) of *T. guttata*. All sections were counterstained with DAPI. DAPI staining showed that at St34, the retinal tissue comprised an NbL (**A**,**C**). SA-*β*-GAL activity presented two bands of labeling located in the vitreal and scleral regions of the NbL in the central retina (**B**), but in a single band located sclerally in the peripheral retina (**D**). At St43 and P7, DAPI staining revealed the central retina to present a multi-laminated structure (**E**,**H**). SA-*β*-GAL activity was mainly detected in the GCL, amacrine cell layer, horizontal cell layer, and photoreceptor cell layer (**F**,**I**). The p21 immunosignal was highly coincident with SA-*β*-GAL staining (**G**,**J**). Abbreviations: acl, amacrine cell layer; GCL, ganglion cell layer; hcl, horizontal cell layer; INL, inner nuclear layer; IPL, inner plexiform layer; NbL, neuroblastic layer; ONL, outer nuclear layer; OPL, outer plexiform layer. Scale bars: 50 µm.

These staining patterns of cell senescence markers were homogeneous throughout the GCL, amacrine, and horizontal cell layers. Furthermore, TUNEL-positive bodies in the horizontal cell layer were almost absent. Therefore, PCD and senescence markers did not correlate in the developing bird retina.

#### **4. Discussion**

We have presented details of the distribution of pyknotic bodies and TUNEL-positive nuclei during development of the visual system in the altricial bird species *T. guttata*. Previous work in our laboratory has shown that these are effective methods for the detection of dying cells in the developing visual system of vertebrates (for a review, see [3]). To the best of our knowledge, the present study provides the first description of the spatiotemporal distribution of dying cells in an altricial bird species. Furthermore, in order to find any possible coincidence between apoptotic and senescent cells in the developing visual system, we also labeled retinal cryosections with SA-*β*-GAL histochemistry and p21 immunohistochemistry. All the results will be discussed below.

#### *4.1. Cell Death during Early Visual System Morphogenesis in T. guttata*

During optic cup stages, abundant pyknotic bodies were found in the central region of the neural retina, coinciding with previous results described in the chicken [32,33,35] and in the mouse [30,31,35]. This wave of PCD may be involved in shaping the optic cup [3].

With respect to the *T. guttata* lens vesicle, pyknotic bodies appeared during detachment of this structure from the head ectoderm, coinciding with results described in all vertebrates studied [19,27,31,56,57]. In this case, cell death seems to be involved in eliminating cells in the interface between the ectoderm and lens tissue, facilitating the separation of the lens vesicle.

Finally, abundant pyknotic nuclei were detected in the SONCs, areas of intense cell degeneration located laterally to the ventral midline of the diencephalon in the chicken [40,41] and in the mouse [31]. SONCs were detected between St15 and St20. This wave of cell death preceded the arrival of ganglion cell axons at the presumptive optic chiasm and therefore seems to be involved in the invasion of pioneer axons in this region of the visual system.

#### *4.2. Cell Death during the Period of Cell Differentiation in the T. guttata Retina*

At St24 (E4.5), coinciding with the appearance of the first differentiated neuroblasts in the *T. guttata* retinal tissue [49], pyknotic nuclei were found in the central retina, dorsally to the optic nerve head. At this stage, cell death affects mainly some proliferating neuroepithelial cells and recent newborn neuroblasts, coinciding with the emergence of the pioneer ganglion cell axons [33,36–38]. This wave of cell death (known as "early neural cell death") could be involved in the creation of extracellular channels that facilitate axonal guidance during early stages of ganglion cell differentiation (for reviews, see [3,58]).

An area of cell death was also detected by St25 in the distal optic nerve, at the junction of this structure with the rudiment of the eye. A similar area of degeneration has been described in the small-spotted catshark, *Scyliorhynus canicula* [19], at stages prior to the invasion of the ganglion cell axons. Neurotrophic cell death affected differentiated neurons in the layered *T. guttata* retina. The emergence of the plexiform layers occurred between St38 (E8.5) and St39 (E9) [48], but the presence of TUNEL-positive bodies was sparse until St42 (E11). At St44 (E13), the incidence of cell death in the GCL increased abruptly, reaching a peak by this stage. In contrast, the maximum of cell death density in the INL was reached at P5, indicating a vitreal-to-scleral progression of cell death, similar to the vitreal-to-scleral wave of cell differentiation described in this bird species [48,49].

These results also reveal marked differences in the timing of visual system maturation between altricial and precocial bird species (Figure 9). Neurotrophic cell death in the GCL occurs in the quail in the period E8–E14, peaking at E10 [39], while in the chicken, it takes place in the period E8–E15, also peaking at E10 [34]. In contrast, dying ganglion cells are detected in *T. guttata* from embryonic stages (St42–E10.5) to a post-hatching period (P8), peaking at St44 (E12). In the case of the INL, cell death extends from E8 to P1 in the quail, peaking at E12 [39], and from E8 to E19 in the chicken, peaking at E11 [34]. In the present study, we have shown that cell death in the INL is detected from St42 (E10.5) to at least P8, the last stage analyzed in the present study, peaking at P5. Therefore, the highest incidence of cell death in the *T. guttata* INL occurred in the post-hatching period, suggesting that most of the synapses established between retinal cells located in this nuclear layer occur during the first week of life. This is a very interesting finding which suggests that, during early post-hatching life, the retinal tissue is still immature and is unable to process the light information it receives.

**Figure 9.** Schematic summary of the chronological patterns and the intensity of neurotrophic cell death in the developing retina of *G. gallus* [34], *C. coturnix* [39], and *T. guttata* (present study). Neurotrophic cell death occurred in the altricial bird at perinatal stages and extended through the first week of life. In contrast, it was restricted to the embryonic period in both of the precocial species. Color codes: white (absence of cell death); light gray (low levels of cell death density) (+); gray (moderate levels of cell death density) (++); dark gray (high levels of cell death density) (+++).

Previous studies in our laboratory have shown that mitotic activity is intense during the first postnatal week in the retina of this altricial species [50], reinforcing the idea of the immature state of this tissue during early life. Indeed, *T. guttata* hatchlings open their eyes at P7 [59], coinciding with a decrease in the incidence of cell death in the retina.

These differences in the timing of ontogenetic cell death between altricial and precocial species have been found in all vertebrates studied. The main waves of cell death occur during the embryonic period in precocial fish [16,19], reptiles [23–25], and birds [33,34,39,40]. In contrast, cell death takes place mainly after hatching/birth in altricial fish [18], birds (present study), and most of the mammals studied [26,29,31,55].

#### *4.3. TUNEL Labeling in the Cytoplasm of Radially Oriented Cells*

Diffuse TUNEL-labeling was also found in the cytoplasm of cells that have a bipolar morphology in the radial plane. Their somas were located at the center of the INL, from which radially oriented processes emerge to span the thickness of the neuroretina. Similar results have been described in the developing retina of fish [19], reptiles [23], birds [14], and mammals [60]. Similar staining following retinal injury has also been described in the retina of fish [61–63] and mammals [64,65]. These radially oriented TUNEL-positive cells were also GS-immunoreactive [62,63,65,66]. The morphology and immunochemical profiles of these labeled cells coincided with those described for Müller cells [67]. Müller glia possess phagocytic activity to remove degenerating cells during development or under experimental conditions (reviewed in [66]). This cytoplasmic labeling is due to the engulfment of TUNEL-positive cell debris by the phagocytic Müller cells.

#### *4.4. Senescence Markers in the Developing Retina of T. guttata*

Cellular senescence occurs in different embryonic tissues during restricted time windows, in most cases contributing to degeneration of the interdigital mesoderm [4,6], pronephros [9], mesonephros [1], and developing heart [7] or inner ear [12–14] structures. SA-*β*-GAL histochemistry is widely used as a biomarker of cellular senescence in vivo and in vitro [42], even in whole-mount embryos [1,2,4,7,9,11]. Most of these works report that SA-*β*-GAL labeling strongly correlates with areas of cell death. The developing visual system of vertebrates is also affected by several waves of cell death (for a review, see [3]), which we also detected in the *T. guttata* visual system (see above). However, we

found no correlation of the labeling pattern of SA-*β*-GAL activity with the TUNEL-positive nuclei detected in the developing retina, in concordance with previous results obtained in the developing chicken retina [43,44]. In this sense, we clearly demonstrated that SA-*β*-GAL activity was restricted to several subpopulations of differentiated neurons (ganglion, amacrine, and horizontal cells) in the embryonic *T. guttata* retina.

Furthermore, the establishment of the state of cell senescence in embryos is associated with the expression of anti-proliferative mediators, such as p21 that seems to act independently of p53 [1,2]. It has been described that p21 expression in mouse embryos strongly correlates with known locations of developmental senescence [68]. In the present study, p21 immunoreactivity faithfully correlates with SA-*β*-GAL labeling, similar to results described in the developing chicken eye [43,44]. Therefore, the present work has clearly shown that the expression of typical senescence markers, including SA-*β*-GAL and p21, in the developing bird retina is up-regulated in subpopulations of differentiated neurons. Notably, both markers have been found to be highly expressed by the first differentiating retinal neurons in the chicken [43,44]. These data indicate that senescence is not the only developmental event that can increase SA-*β*-GAL activity and p21 expression in embryonic tissues. Senescent cells and differentiated retinal neurons share a common biological feature—they are in a characteristic non-proliferative state. Therefore, SA-*β*-GAL activity and p21 could be involved in distinct biological phenomena such as cell senescence and terminal cell differentiation of neurons. In this sense, typical senescence markers have been found to be associated with cell differentiation in the developing tendons [6] and the maturing ventricular myocardium of embryonic mice [7]. However, the possible relationship between the mechanistic events involved in cell senescence and terminal cell differentiation remains to be clarified.

#### **5. Conclusions**

Relative to precocial bird species, in altricial species, some aspects of brain maturation such as telencephalic neurogenesis are delayed into the post-hatching period [69–73]. Retinal neurogenesis is intense in altricial birds at hatching [48,49] and during the first week of life [50]. Furthermore, it has been demonstrated [74] that the formation of some retinal structures, the foveal pit in particular, is delayed until the second week of life (P10–P14). In the present study, we have demonstrated that there is intense ontogenetic cell death in the retina of the hatched animals. Thus, *T. guttata* constitutes an excellent model in which to study retinal development events during the first weeks of life.

**Author Contributions:** G.M.-P., J.R.-L. and J.F.-M. designed research; A.M. and Y.G. analyzed data; G.Á.-H., J.A.d.M.-R. and I.H.-N. performed research; J.R.-L. and J.F.-M. wrote the paper. All authors have read and agreed to the published version of the manuscript.

**Funding:** G.A.-H. was a recipient of a pre-doctoral studentship from the Universidad de Extremadura. This work was supported by grants from Dirección General de Investigación del Ministerio de Educación y Ciencia (BFU2017-85547-P), and Junta de Extremadura, Fondo Europeo de Desarrollo Regional, "Una manera de hacer Europa" (GR15158, GR18114, IB18113).

**Institutional Review Board Statement:** The study was conducted according to the guidelines of the Declaration of Helsinki, and approved by the Ethics Committee of the University of Extremadura (protocol code 264/2019, 29 June 2020).

**Informed Consent Statement:** Not applicable.

**Data Availability Statement:** Some or all data used during the study are available from the corresponding author by request.

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

#### **References**

