Introduction

Worldwide, 10-15% of couples (or 70 million) face infertility [1]. Infertility is thus a major public health issue presenting significant medical, scientific and economic challenges (a multibillion € annual market) [2]. A significant proportion of infertility is due to altered gametogenesis, where the sperm and eggs produced are incompatible with fertilization and/or embryonic development. Approximately 40% of cases of infertility involve a male factor, either exclusively, or associated with a female deficiency.

Male gametogenesis, or spermatogenesis, is a highly complex physiological process which can be split into three successive steps: proliferation (mitosis of spermatogonia), reduction of the number of chromosomes (meiosis of spermatocytes), and spermatid differentiation (spermiogenesis). The whole process involves several thousands of genes, of which a large part are expressed almost exclusively in the testes [3]. Because of this multiplicity of genes, spermatogenesis is strongly affected by genetic factors [3], with most severe disorders likely to be of genetic origin.

For patients with severe infertility, only assisted reproductive technologies (ARTs) – such as Intra Cytoplasmic Sperm Injection (ICSI) – can efficiently overcome the problems faced. However, there are some concerns surrounding the risks associated with this technique; statistical analyses of birth records has shown a real risk of birth defects associated with ICSI [4]. To overcome these drawbacks, a number of experimental strategies have been proposed to bypass ARTs and restore spermatogenesis, including gene therapy [5–8].

Gene therapy consists of introducing a DNA sequence into the genome to compensate for a defective gene. It can thus rescue production of a missing protein, and is now widely applied both in research and for the treatment of human diseases. Given the genetic basis of male infertility, the first strategy tested to overcome spermatogenic failure associated with monogenic diseases was delivery of an intact gene to deficient germ cells. Gene therapy works efficiently in germ cells, and numerous publications have shown that conventional plasmids can be transferred into spermatogonia in several species with a high success rate, allowing their transcription in all cells of the germinal lineage [5–8]. Most experiments were performed in mouse models, delivering DNA constructs into living mouse germ cells by testis electroporation after microinjection of a DNA-containing solution into the seminiferous tubules. Using this method, it was possible to rescue meiosis and fertility in mouse models of infertility [5, 8]. However, the genetic changes induced are transmitted to any descendants. Consequently, gene therapy cannot be used to treat infertility in humans, as for both ethical reasons and to avoid any eugenic deviations, transmissible changes are illegal in humans in most countries. Moreover, genetic modification of germ cell lines presents multiple biological risks, including that of inducing cancer [9, 10]. Gene therapies have thus raised both ethical controversy and long-term safety issues.

For these reasons, we decided to test an alternative strategy to DNA transfection based on the use of naked mRNA. Thanks to this small change, there should be no risk of genomic insertion, and thus there is no question of heritable alterations. The first part of this study presents a complete characterization of the protein expression patterns obtained following transfection of naked mRNA coding for reporter genes into the testes of mice.

Among infertility disorders, oligo-astheno-teratozoospermia (OAT) is the most frequent; it is likely to be of genetic origin. Spermatocytograms for OAT patients show a decrease in sperm concentration, multiple morphological defects and defective motility. Because of these combined defects, patients are infertile and can only conceive by ICSI. The aim of this work was 1/ to perform a full characterization of the efficiency of naked mRNA to express proteins, using reporter genes, and 2/ to apply the protocol to a preclinical mouse model of OAT. Patients and mice carrying mutations in the ARMC2 gene present a canonical OAT phenotype and are completely infertile. The preclinical Armc2 deficient (Armc2 KO) mouse model is therefore a valuable model to assess whether in vivo injection of naked mRNA combined with electroporation can restore sperm spermatogenesis. We chose this model for several reasons: first, it is an easy model to implement. Armc2 KO mice are sterile and all sperm exhibit short, thick or coiled flagella [13]. As a result, 100% of sperm are completely immobile, thus it should be easy to determine the efficacy of the technique by measuring sperm motility with a CASA system. Second, the Armc2 gene codes for an 867-amino acid protein and this large size represents a challenge for expression in the testis following electroporation.

To determine the efficacy of naked mRNA transfection as a method to achieve protein expression in the testes, we assessed the level of transcription of reporter proteins after mRNA injection compared to the injection of a non-integrating plasmid, the Enhanced Episomal Vector (EEV). EEV is a vector derived from Epstein-Barr virus, it includes an EBV Ori and the EBNA1 protein. It does not integrate or modify the host genome. Replication of this vector is synchronized with host cell division [11, 12].

In the present in vivo work, we injected and electroporated three distinct mRNAs coding for the reporter proteins green fluorescent protein (GFP), luciferase (Luc) and mCherry, and an EEV episome vector containing the sequences coding for both GFP and luciferase reporter proteins. We first characterized and validated the method of injection via the rete testis followed by electroporation in adult males. Next, we compared the kinetics and patterns of expression of the electroporated mRNAs and EEV by several methods, including whole testis imaging, in vivo bioluminescence imaging and tissue clearing.

Then we used a preclinical mouse model of OAT to test the mRNA transfection protocol, with the aim of restoring fertility. Three- and 5-weeks after injection and electroporation of Armc2-mRNA into the testes of Armc2 KO mice, we identified normal motile sperm cells in the epididymis.

Materiel and methods

Animals

All procedures involving animals were performed in line with the French guidelines for the use of live animals in scientific investigations. The study protocol was approved by the local ethics committee (ComEth Grenoble #318) and received governmental authorization (ministerial agreement #38109-2022072716142778).

The animals used were B6D2 F1 hybrid (♀ C57BL/6JRj crossed with ♂ DBA/2, Janvier laboratories;france) adult male mice aged between 8 and 25 weeks, and the Armc2 KO mouse strain obtained by CRISPR-Cas9, as described in Coutton et al. 2019. Experiments were carried out on wild type (WT) or Armc2 KO adult male mice aged between 8 and 15 weeks.

Chemicals and reagents

Fast Green (FG) (F7258 – 25 g), phosphate buffered saline (PBS, D853 7-500 mL), hematoxylin (GH5316 – 500 mL), eosin (HT110216 – 500 mL), terbutanol (471712 – 100 mL), Histodenz (D2158 – 100 g), sorbitol (S1876 – 100 g) and urea (U5128 – 500 g) were purchased from Sigma Aldrich (Saint Louis, MI, USA). Schorr staining solution was obtained from Merck (XXX). Mfel HF (R35895) and RNase-free DNase I (M03035) were obtained from New England Biolabs (Ipswich, MA, USA). Paraformaldehyde (PFA, 15710) was obtained from Electron Microscopy Science (Hatfield, PA, USA). Ketamine and xylazine were obtained from Centravet (Dinan, France). Fluorescent i-particles NIRFiP-180 were obtained from Adjuvatis (Lyon, France). CleanCap AG, CleanCAp EGFP-mRNA and CleanCap Luciferase-mRNA were obtained from Tebubio (Le Perray en Yvelines, france).

Plasmids

All plasmids were amplified by bacterial transformation (E. coli, EC0112; Thermo Fisher, Waltham, MA, USA). After expansion, plasmids were purified with a NucleoBond Xtra Midi kit (740410-50; Macherey-Nagel, Düren, Germany). DNA pellets were solubilized in (DNAse- and RNAse-free) milliQ water.

Non-integrative episome EEV CAGs-GFP-T2A-Luciferase

The EEV CAGs-GFP-T2A-Luciferase (EEV604A-2) used for in vivo testis microinjection and electroporation was obtained from System Biosciences (Palo Alto, CA, USA). This episome contains the cDNA sequences of Green Fluorescent Protein (GFP) and luciferase, under the control of a CAGs promoter (supp Fig 1). After purification, the EEV CAGs-GFP-T2A-Luciferase plasmid concentration was adjusted to 9 μg μL^-1. Prior to injection, 3.3 μL of this plasmid solution was mixed with 1 μL 0.5% Fast Green and 5.7 µL sterile PBS to obtain a final EEV concentration of 3 µg mL^-1.

mCherry plasmid

The was obtained from Dr. Conti’s lab at UCSF. This plasmid contains the cDNA sequence of mCherry under the control of CMV and T7 promoters (supp Fig 1). After amplification and purification, the final plasmid concentration was adjusted to 9 μg μL^-1.

Plasmid EEV-Armc2-GFP

The EEV-Armc2-GFP plasmid used for in vivo testes microinjection and electroporation was obtained from Trilink (San Diego, USA). This non-integrative episome contains the cDNA sequences of Armc2 and the Green Fluorescent Protein (GFP) genes under the control of a strong CAGs promoter (Fig 1 supplementary).

Figure 1:

After amplification and purification, the final plasmid concentration was adjusted to 9 μg μL^-1. Prior to injection, 3.3 μL of this plasmid solution was mixed with 1 μL of 0.5% Fast Green and 5.7 µL of sterile PBS to obtain a final EEV concentration of 3 µg µL^-1.

Armc2-mRNA

Armc2-mRNA used for in vivo testes microinjection and electroporation was obtained from Trilink (San Diego, USA). The commercial Armc2-mRNA has an AG CleanCap, a poly A tail of 120 adenosines and 3 HA FLAG. The main challenge with mRNA-based therapies is mRNA stability. To improve mRNA stability in vivo and avoid its degradation by ribonucleases, optimization techniques were implemented. Thus, the used mRNA had codon optimization, a polyA tail and a CleanCap. To verify germ the efficiency of cell transfection, an EGFP-mRNA was injected together with the Armc2-mRNA. During the injection, the concentration of EGFP-mRNA and Armc2-mRNA were 300 ng µL^-1 each.

mcherry-mRNA transcription in vitro

The circular mCherry plasmid was linearized using the restriction enzyme Mfel HF at 37 °C for 1 h. The pm-mCherry was then extracted and purified with the DNA Clean and Concentrator-25 ^TM kit (D4033; Zymo Research, Irvine, CA, USA). The pm-mCherry was subsequently transcribed in vitro using the T7-FlashScribe kit (C-ASF3507; Tebubio, Le Perray en Yvelines, France). The mRNA was capped with a clean cap (CleanCap AG, Tebubio), and a poly A tail (C-PAP5104H; Tebubio) was added before purification using the NucleoSpin RNA Clean Up kit (740948-50; Macherey-Nagel). Prior to injection, mcherry-mRNA was mixed with Fast Green to obtain a final concentration of 300 ng mL^-1 (0.05% FG, PBS).

Agarose gel electrophoresis of the Episomal Vector EEV and mRNA-mCherry

Before loading on a pre-stained (Gel Green) 1.5% agarose gel, the EEV-plasmid and mRNA were mixed with a loading dye (Promega Blue Orange G1904). An aliquot of each sample (500 ng) was loaded into each well and electrophoresis was performed for 30 min at 100 V at room temperature (RT). A DNA size marker (Gene ruler, Thermo Scientific, SM1331) was used to assess molecular weight. Gel images were acquired using a ChemiDoc XRS+ (BIORAD).

In vivo microinjection and electroporation of testes

Electroporation was conducted as previously described [7]. Briefly, male mice B6D2, Armc2 +/+ or Armc2 -/-, depending the experimental conditions, were anesthetized with ketamine/xylazine solution (100mg mL^-1 and 10 mg mL^-1 respectively). The testes were pulled out of the abdominal cavity or scrotum. Under a binocular microscope and using microcapillaries pipettes (FemtoTip II ®, Eppendorf, Montesson, France), 10 mL DNA (3 mg mL^{- 1} -0.05% FG) or mRNA (300 ng mL^-1 -0.05% FG) was injected into the seminiferous tubules through the rete testis applying constant pressure (microinjector, Femto Jet 4i). Two series of 8 square electric pulses (25 V for 50 ms) were applied to the testis using tweezer-type electrodes linked to an electroporator (Gemini, BTX, Holliston, USA). The testes were then replaced in the abdominal cavity, and the abdominal wall and skin were sutured. For each animal, the left testis was injected and electroporated with the different nucleic acids (mRNA, EEV), whereas the right testis was injected with a control solution (PBS, 0.5% FG) as a control. Both testes were electroporated.

Flash Freezing and Fluorescence Analysis of testes

Depending on the experimental condition, 1-, 3- or 7-days post injection, the testes were collected and washed for 5 min in PBS. Then, they were embedded in Peel-A-Way Cryomolds filled with OCT (OCT Mounting media VWR BDH CHEMICALS). The samples were flash frozen in a 100% isopentane solution (Carlo ERBA), pre-cooled with liquid nitrogen. Once frozen, they were cut into 20-µm sections using a cryostat. The cryostat sections were then fixed with 4% PFA-PBS for 10 min at 4 °C and counterstained with 1.8 mM DAPI (nuclear stain) before observation using an Axioscan Z1 slide scanner or a confocal AxioObserver Z1 multiparameter microscope LSM710 NLO – LIVE7 – Confocor3 . The fluorescence of the different reporter proteins was detected using appropriate filters for DAPI, GFP, Texas Red (for mCherry), and Cy7 (for the Fluorescent i-particles NIRFiP-180).

Tissue collection and histological analysis

For histological analysis, treated and control B6D2 testes were fixed by immersion in 4% paraformaldehyde (PFA) for 14 h. They were then embedded in paraffin before cutting into 5-µm sections using a microtome (Leica biosystems). After deparaffination, the sections were stained with hematoxylin and eosin. Stained sections were digitized at ×20 magnification using an axioscan Z1-slide scanner (Zeiss, Germany). Spermatogenesis was assessed by measuring the area of seminiferous tubules (μm²) and the cross sections of round tubules ((n > 35 tubules per testis section; n= 5 testis sections per condition). Statistical significance of differences was determined using a Student t-test.

Ex vivo Fluorescence Analysis

To analyze the expression of the reporter proteins GFP and mCherry in seminiferous tubules, whole testes were examined under an inverted microscope (Olympus, CKX53). Exogenous fluorescence was detected using filters for GFP and Texas Red.

Harris-Shorr sperm Analysis

Sperm cells were collected from the caudae epididymides of mice (Control, injected with EEV-GFP, GFP-mRNA, or Armc2-mRNA). They were allowed to swim for 10 min at 37 °C in 1 mL M2 media. Sperm cells were centrifuged at 500 x g, washed once with PBS, and fixed in PBS/4% paraformaldehyde for 1 min at RT. After washing with 1 mL acetate ammonia (100 mM), the sperm cell suspensions were spotted onto 0.1% poly L-lysine precoated slides (Thermo Scientific) and left to dry. Harris–Schorr staining was then performed according to the WHO protocol [14], and at least 100 sperm cells were analyzed per animal.

Whole Testis Optical clearing and 3D image reconstructions

Optical clearing (adapted from uDISCO and Fast 3D clear protocols)

Adult mice were euthanized by cervical dislocation and transcardial perfusion with 1X PBS (Sigma-Aldrich). The testes were extracted and fixed for two days at 4 °C in 4% PFA (Electron Microscopy Sciences). Samples were then washed with PBS for at least two days. Mouse testes were subsequently dehydrated in graded series of tert-butanol solutions (Sigma-Aldrich) at 35 °C as follows: 30% overnight (O/N), 50% for 24 h, 70% for 10 h, 80% O/N, 90% for 10 h, 96% O/N, and 100% for 10 h. The testes were cleared in clearing solution (96 % Histodenz,2% Sorbitol, 20 % Ure) for 2 days. Then, nuclei were stained with 3.6 mM DAPI (20% DMSO; 2% Triton, 1% BSA) for 2 days. Finally, the testes were then conserved in the clearing solution at 4°C until the microscopy observation. All these steps were carried out under agitation and protected from light.

3D tissue Imaging

The optically-cleared mouse testes were imaged on a lightsheet fluorescence microscope (Blaze, Miltenyi Biotec, Germany), using a 4x NA 0.35 MI PLAN objective protected by a dipping cap for organic solvents, with an overall working distance of 15 mm. Acquisitions on the horizontal plane were obtained with a fixed lightsheet thickness of 3.9 µm at both 488 nm and 561 nm with no overlap between horizontal planes. Voxel resolution x = 1.21; y = 1.21; z = 2 μm. 3D reconstructions were created using Imaris software (Oxford Instruments plc, Tubney Woods, Abingdon, Oxon OX13 5QX, UK).

Cellular image analysis

The optically-cleared mouse testes were imaged using a ‘ConfoBright’ system which is a unique adaptive confocal microscope (Nikon A1R MP, Nikon Europe B.V., The Netherlands) equipped with a deformable mirror module (AOS-micro, AlpAO, Montbonnot, France) for the correction of geometrical aberrations. Indeed, the deep confocal imaging of the cleared 3D sample requires long distance objectives and immersion media of different refractive index, resulting in optical aberrations. Images were acquired using Adaptive Optics optimization with an apo LWD 40x/1.15 water immersion objective (WD 600 µm) and a 10x/0.45 objective. FIJI software was used to process and analyze images and Imaris software for the 3D reconstructions.

Bioluminescence imaging

In vivo Bioluminescence imaging was performed at several time points after in vivo Luciferase -mRNA or EEV-GFP-luciferase injection and electroporation (n = 5 mice per condition). Ten minutes before imaging, mice received an intraperitoneal injection of 150 µg g^-1 of D-luciferin (Promega), and were then anesthetized (isoflurane 4% for induction and 1.5% for maintenance) before placing in the optical imaging system (IVIS Lumina III, PerkinElmer). In vivo bioluminescence signals were measured in selected regions of interest (injected testes) and were expressed as mean photons per second (mean ± SEM). Background bioluminescence was measured on images from control mice. When the in vivo bioluminescence signal was no longer detectable, testes were collected and immersed in a luciferin solution for a few minutes before performing ex vivo imaging to confirm the absence of signal.

Sperm motility

The cauda epididymis was dilacerated in 1 mL of M2 medium (Sigma-Aldrich, l’Isle d’Abeau, France) and spermatozoa were allowed to swim out for 10 min at 37 °C. Motility of the spermatozoa was evaluated with an Olympus microscope and Computer Aided Sperm Analysis (CASA) (CEROS II apparatus; Hamilton Thorne).

Western blot

Western blotting was performed on HEK-293T cells (ATCC, Manassas, VA, USA) transfected with EEV-Armc2 or Armc2-mRNA. Cells were transfected using JetPrime (Polyplus, 101000027) for DNA and JetMessenger (Polyplus, 101000056) for mRNA vectors, both according to the supplier’s recommendations. After 48 h, the cells were washed with PBS and scraped off before centrifuging at 4 °C, 1500 RPM for 5 min. The cell pellet was then resuspended in lysis buffer (Thermo Scientific, 87787) supplemented with an EDTA-free cocktail of protease inhibitors (Roche, 11836170001). The suspension was stirred at 4 °C for 2 h, and then centrifuged at 16,000 x g at 4 °C for 10 min. The protein content of the supernatant was estimated (QuantiPro™ BCA Assay kit; Sigma-Aldrich QPBCA) before adding 5X Laemmli + 5% β-mercaptoethanol and heating at 95 °C for 10 min. For the Western blot, 30 μg of proteins were deposited on a ready-made Bis-Tris gel 12 % (ThermoFisher). After transfer, the PVDF membrane was blocked with 5% milk in Tris-Buffered Saline solution containing 0.1% Tween 20 (TTBS) before immunodetection. The anti HA antibody (Cat. No. 11867423001, Sigma Aldrich) was diluted in TTBS at 1/5000. After incubation with secondary antibodies (AP136P, Sigma Aldrich, 1:10,000 in TTBS), binding was revealed with an enhanced chemiluminescence detection kit (ECL plus, Amercontrol Biosciences). Membranes were imaged on a ChemiDoc™ system (Bio-Rad).

Immunofluorescence

Immunofluorescence analysis of dissociated testicular cells was performed as follows. After collection, the tunica albuginea was removed from the testes. Then the tissue was mechanically digested with 18G needles. Once washed with PBS, the testicular cells were placed in a dissociation medium (25 mL RPMI and 25 mg collagenase type V) for 20 min at 37 °C. After filtration (100 mM filter) and centrifugation (5 min at 200 x g) the pellets were resuspended in PBS before centrifugation once again. The pellet was then fixed in 1 mL PFA 4% for 5 min. Then 10 mL of ammonium acetate (0.1 M) was added. Finally, 2 mL of medium was spread on a slide. Testicular cells were permeabilized with 0.1% PBS-Triton X-100 for 20 min at room temperature. Slides were then blocked in 10% BSA with PBS-Tween 0.1% for 30 min before incubating overnight at 4 °C with primary antibodies anti-rabbit ARMC2 (HPA053696, Sigma Aldrich, 1:50) and anti-guinea pig tubulin (AA345, Swiss Institute of Bioinformatics, 1:100) diluted in PBS-Tween 0.1% – 5% BSA. Slides were washed with PBS before incubating for 1 h with secondary antibodies (for tubulin: A-11073, Thermo Fischer Scientific, 1/500; for ARMC2: A-11036, Thermo Fischer Scientific, 1:1000). Samples were counterstained with DAPI and mounted with DAKO mounting media (Life Technology). Fluorescence images were acquired under a confocal microscope (Zeiss) fitted with a 63× oil immersion objective. Images were analyzed with ZEN lite software (Zeiss).

Statistical analyses

Statistical analyses were performed using SigmaPlot (version 10; Systat Software, Inc., San Jose, CA). To account for sample variability between animals, a paired t-test and Wilcoxon test were used. Data are displayed as mean ± SEM. P values of *≤0.05, **≤0.01, or ***≤0.001 were considered to represent statistically significant differences.

Results

1. In vivo microinjection and electroporation of mouse testes

Two routes have been described for microinjection of DNA into the testes: direct injection through the tunica albuginea, or injection into the lumen of the seminiferous tubules via the rete testis. We chose the rete testis route and evaluated the efficacy of the microinjection protocol, in particular as we wished to better characterize the diffusion of the injected solution in the volume of the testis, as we were unable to find any information on this parameter in the literature. The efficacy of microinjection via rete testis was assessed using fluorescent i-particles NIRFiP-180, and by measuring their diffusion in testis cross sections examined by microscopy 3 days post-injection (Fig 1). To avoid lesions due to overfilling, the injection was controlled by measuring the expansion of the staining of the peripheral seminiferous tubules during the injection. Injections were stopped when the testes were filled to 2/3 of their capacity (fig 1 A-B). In testis cross sections, the fluorescent i-particles NIRFiP-180 were heterogeneously distributed, and mainly observed in seminiferous tubules located in the peripheral region of the testes, with fewer particles present in the center of the testes (Fig 1 C-D). Moreover, no fluorescent i-particles NIRFiP-180 were visible in the peritubular space. These results indicated that microinjection through the rete testis did not produce a homogenous distribution of the particles throughout the seminiferous tubules. Nevertheless, the seminiferous tubules remained intact, as no signal was observed in the peritubular space (Fig 1C-D).

Next, we assessed the overall safety of the rete testis microinjection and electroporation of mRNA and EEV into testes. The safety of the protocol was evaluated by comparing macroscopic and microscopic anatomies of control (injected with control solution, PBS, 0.05% FG), and treated testes (injected either with EEV-GFP (PBS, 0.05% FG) or GFP-mRNA (PBS, 0.05 % FG)). Three days post-injection, the testes were first observed under a binocular microscope to identify possible macroscopic degeneration of the seminiferous tubules (Fig 2A1 and 2B1). Degenerations appear as pearly white lesions at the surface of the testis as illustrated in Supp Fig 1 following over electroporation. With the protocol used, no such lesions were observed. Next, the testes were measured and weighted. No differences were observed between control and treated testes (Fig 2 A2, A3, B2, B3), Then, microscopic differences were sought by histological analysis of 5-µm sections (Fig 2C). No difference was observed between the control condition and EEV-GFP or GFP-mRNA on the full cross sections (Fig 2 C1, C2, C3). Next, we observed all the different testicular cells, including spermatogonia, spermatocytes, Sertoli cells, round spermatids and epidydimal sperm cells (Fig 2D1, D2, D3) in each condition. The layered structure of germ cells was respected in all cases. Analysis of the histological sections revealed no differences in the tubules area of the testes injected either with EEV-GFP or GFP-mRNA (Fig 2E). Finally, Harris-Shorr staining of the epididymal sperm cells showed no morphological defects (Fig 2F). Taken together, these results suggest that in vivo microinjection and electroporation of EEV or mRNA is safe.

Figure 2:

2. Analysis of EEV-GFP and GFP-mRNA testicular expression by whole testis imaging

After validating the injection method, we compared the kinetics of GFP expression for mRNA and EEV. To do so, we injected and electroporated one testis of129 adult B6D2 mice with EEV-GFP and of 65 mice with GFP-mRNA. At 0, 1, 7, 15, 21-, 28-, 35-42 and 49-days post-injection, the whole testes were observed under an inverted microscope. The exogenous fluorescence was directly visible at the surface of the testes when illuminated with light at the appropriate wavelength (Fig 3 and 4). These observations revealed no testicular lesions at any post-injection time (Fig 3.A1-H1 and 4A1-F1). In addition, both GFP-mRNA and EEV-GFP induced GFP expression in the testes (Fig 3A2-H2 and 4A2-F2). It is worth noting that both vectors induced GFP expression from the very first day. However, the durations of expression were different. For EEV, GFP fluorescence was still observable on day 42 for 100% of samples, and 56% of samples were positive on day 49, indicating that expression lasted around 1.5 months (Fig 4G). In contrast, for mRNA, 100% of testes were labelled on day 21, but none showed any fluorescence on day 28 (Fig 4G). Thus, EEV transfection allowed a considerably longer duration of expression than mRNA. (Fig 3 and 4). In addition to differences in duration of expression, the GFP expression patterns were clearly different: mRNA produced a large, diffuse pattern, highlighting the shape of the seminiferous tubules; EEV-GFP produced a punctiform pattern (Fig 3B and 4B.) These results suggest that GFP- mRNA and EEV-GFP targeted different testicular cell types, or that there were differences in terms of efficiency of transfection.

Figure 3:

Figure 4:

3. Kinetics of EEV and mRNA expression assessed by in vivo imaging

To further assess and compare the kinetics of expression of the two vectors, we expressed exogenous luciferase in the testis using EEV or mRNA and observed the level of luciferase by in vivo bioluminescence imaging. For EEV, we took advantage of the fact that the EEV plasmids contains the DNA sequence of the luciferase protein (CAGs-GFP-T2A-Luciferase) in addition to the DNA sequence of the GFP fluorescent protein (Fig supp 2). For mRNA, we injected commercial naked luciferase-mRNA into the rete testis according to the same protocol as used for GFP-mRNA. For this set of experiments, we injected the EEV-GFP-Luc and luciferase-mRNA into the testes of 6 adult mice on day 0. Importantly, we injected a similar number of copies of mRNA and EEV. The testes were imaged in vivo to detect bioluminescence expression, on day 1 and weekly until disappearance of the signal, no more than 120 days post-injection (Fig 5). For EEV-GFP-Luciferase, the bioluminescence induced by transfection was detected from day 1. After a rapid decrease in signal intensity over the first 3 weeks, a weak but constant signal remained detectable for 3 months, then faded away (Fig 5 A1-A2). For Luciferase-mRNA, expression was also detected from day 1. The bioluminescence decreased gradually over 3 weeks, becoming undetectable thereafter (Fig 5B1-B2). These results are consistent with our previous results (Fig 3) and confirm that EEV allows a longer expression of exogenous protein within the testis. Overall, our results indicated stronger expression for mRNA than for EEV, but with expression decreasing rapidly over-time, and almost no remaining signal after 3 weeks. In contrast, residual expression was detectable for several months with EEV.

Figure 5:

4. Assessing testicular and cellular GFP-mRNA expression using whole testicle optical clearing, light sheet microscopy, and 3D-microscopic reconstructions

To better characterize the spatial distribution of GFP-mRNA expression in the testis, we performed whole testicular optical clearing. On day 0, we injected and electroporated 6 testes with GFP-mRNA. On day 1, we harvested the testes, cleared them with a modified optical clearing protocol (see Methods). After complete clearing (Fig 6 A), we observed the testes under a light sheet microscope and performed 3D reconstruction (video 1, 2 and Fig 6 BC). From this 3D reconstruction, we determined the volume of testis stained with GFP. The testis was sliced into 500 stacks, and the volume occupied by GFP staining was estimated by measuring the GFP-positive area in each stack and multiplying it by the thickness of the stack (10 µm). Measurements were possible only for the first 250 stacks due to optical problems with the last 250 stacks. To overcome this problem, and acquire information for the whole testis, the sample was turned by 180° and the measures repeated. A total GFP-stained surface of 0.51 mm³ and 0.23 mm³ was determined for face 1 and face 2, respectively. The corresponding total volume for the 250 stacks was measured as 60 mm³, therefore 0.81% and 0.24% of the testis volume was transfected for face 1 and face 2 respectively (Fig 6 BC).

Figure 6.

5. Assessing GFP cellular expression using whole testicle optical clearing and confocal microscopy

Because the GFP fluorescence patterns were different for the two nucleic vectors when observed under the inverted microscope (Fig 3), we wondered whether the same cell types were targeted in both cases. To address this question, the whole optical cleared testes from EEV-GFP and GFP-mRNA-transfected mice were observed under a confocal microscope. The different cell types were identified based on their positions within the seminiferous tubule, their cellular shape and their nuclear morphology – revealed by nuclear staining. For instance, Sertoli cells have an oval to elongated nucleus and the cytoplasm presents a complex shape (“tombstone” pattern) along the basement membrane [15]. Round spermatids have small, round and compact nuclei with a nucleole and are localized between the spermatocytes and elongated spermatids [16]. For EEV-GFP, on day 1 post injection and electroporation, a strong punctiform green fluorescence was visible inside the seminiferous tubules (Fig 7 A). Based on the different morphological criteria, this fluorescent signal was detected in spermatocytes, round spermatids and Sertoli cells (Fig 7 BC). On day 7, the GFP signal induced by EEV-GFP was reduced and only isolated signals in a few seminiferous tubules (1 per 11 tubules) were observed (Fig 7 D). These signals were associated only with Sertoli cells (Fig 7 E).

Figure 7.

For the mRNA vector, on day 1 and day 7 post-injection and electroporation, an intense fluorescence was observed in all the germ cells in the seminiferous tubules (Fig 8). At the cellular level, this fluorescent signal was associated with Sertoli cells, spermatogonia, spermatocytes, round spermatids, elongated spermatids, and sperm cells to similar extents for both post-injection times (Fig 8 B, D). In contrast to what was observed with EEV on day 7, no reduction in the number of fluorescent seminiferous tubules was noted when using GFP-mRNA, with 8 out of 10 tubules stained (Fig 8C).

Figure 8.

6. Expression of naked mcherry-mRNA in testis following electroporation

Heterologous expression is well known to depend on the protein studied, we therefore tested the same reporter proteins to compare the mRNA and EEV vectors in the experiments presented above. Apart from the bioluminescence experiments with luciferase, we compared only GFP protein expression. To validate and confirm the capacity of naked mRNA to safely express proteins in the testes after injection and electroporation, we further challenged the method with another reporter protein – mCherry. We injected homemade naked mRNA coding for mCherry into the testes (Fig supp 2BC). As previously with GFP- mRNA, no testicular lesions were observed (Fig supp 3 A1, B1, C1, D1, E1, F1). Assessment of the remanence of testicular expression of the mCherry fluorescent protein on days 0, 1, 7, 15, 21 and 28 post-injections revealed a strong red fluorescence from day 1 that remained detectable 15 days post-injection (Fig supp3 B2, C2 and D2). At the cellular level, the fluorescent signal was detected in germ cells, including spermatogonia, spermatocytes, round spermatids, testicular sperm cells, and Sertoli cells on days 1 and 7 post-injection (Fig supp 4).

Finally, we compared the kinetics and levels of expression of the three different mRNA molecules, coding for mCherry, GFP and luciferase. By comparing the number of mice expressing mCherry-mRNA, GFP-mRNA and luciferase-mRNA fluorescence/luminescence over 21 days, we observed first that expression was detectable for all mRNAs on day 1, and second that the duration of expression was slightly different for the individual mRNAs. For instance, on day 15, 100%, 80% and 60% of mice injected with GFP-mRNA, mcherry-mRNA, and mRNA-luciferase, respectively presented fluorescence/bioluminescence and on day 21 100% of mice expressed GFP, whereas no signal was observed for mCherry or Luciferase (Fig supp 5).

7. Endogenous expression of ARMC2 in germ cells

Before attempting to rescue expression, we felt it was important to better characterize Armc2 expression in healthy germ cells, and in particular to study the timing of expression. By IF, we determined when ARMC2 protein translation started during spermatogenesis. For these experiments, dissociated cells from testes were observed to detect the presence of ARMC2 on different spermatogenic cells. ARMC2 was present only in the flagellum of elongated spermatids (Fig 9A). The specificity of the signal was validated using testicular cells from Armc2 KO mice, where no signal was observed on all spermatogenic cells (Fig 9B). In WT cells, no evidence of ARMC2 was found in earlier germ cell types like spermatogonia or spermatocytes (data not shown). These results suggest that the ARMC2 protein is expressed late during spermatogenesis.

Figure 9.

8. Co-injection of Armc2-mRNA and eGFP-mRNA followed by electroporation is safe and induces green fluorescence in the seminiferous tubules

We next tested whether the injection and electroporation of two RNA molecules had any deleterious effects on testis anatomy and seminiferous tubule structure. We first verified the quality of Armc2-mRNA synthesis by transfecting HEK cells and performing Western blot (Fig supp 2). After this validation, we co-injected Armc2-mRNA and eGFP-mRNA into the left testes of mice, using the right testes as untreated controls. eGFP-mRNA was co-injected to verify and monitor transfection efficiency. The testes were observed under a binocular microscope at different times (3, 6, 10, 15, 21, 28 and 35 days) after electroporation to identify possible macroscopic degeneration of the seminiferous tubules. No morphological defects were observed in the testes co-injected with Armc2-mRNA and eGFP-mRNA. An example of control and injected testes from day 15 is presented in Fig 10 A1, B1. The testes were also weighed at different times post-injection, and the weight ratios of injected testes to non-injected control testes were determined. For all time points, this ratio was close to 1 (Fig 10 C), confirming that the method and the mRNAs did not cause any injury at the organ level. Next, under blue light, the efficiency of the transfection was assessed by observing the GFP fluorescence at the surface of the testes. Obvious GFP fluorescence was observed on testes injected with Armc2-mRNA and eGFP-mRNA 2 weeks after injection (Fig 10 B2), indicating that the naked mRNA was successfully transfected into testicular cells and remained present up to 2 weeks after surgery.

Figure 10.

9. Motile sperm cells detected in Armc2 KO mice following Armc2-mRNA injection and electroporation into testes

We then assessed whether the injection of Armc2-mRNA into the testes in Armc2 KO mice restored sperm cell motility. We examined motility of sperm cells present in the caudal part of the epididymis at different times post-injection (3 to 35 days post-injection). For each condition, between 3 and 5 KO mice were used.

The Armc2 KO model used is known to produce sperm cells with short and irregular flagella that are therefore immotile on day 0. No motile sperm were observed on days 3, 6, 10, 15 or 28 after surgery (Fig 11 A). However, motile sperm cells were found in the epididymis of some Armc2 KO mice at 21- and 35-days post-treatment (Fig 11 A). Indeed, 1 in 3 mice had motile sperm cells at 21 days post-surgery, rising to 3/5 mice at 35 days post-injection. Nevertheless, the number of motile sperm cells observed remained low: 4.5% after 21 days and 6.15% after 35 days post-injection (Fig 11 A). Videos showing sperm motility in different conditions are available in the online material associated with this article (Videos 3 to 6).

Figure 11.

After verifying motility, we looked at the morphology of the spermatozoa present in the cauda epididymis. Six days after injection of Armc2-mRNA, the cells detected were mostly round cells and abnormal spermatozoa with a short or coiled flagellum measuring between 7 and 20 µm. The same cell types were observed at 3-, 10-, 15- and 28-days post-surgery.

In contrast, the motile sperm detected on days 21 and 35 had a normal morphology with a long flagellum (greater than 100 µm) and a hook-shaped head (Fig 11 B).

Discussion

The challenge of treating male infertility remains to be addressed. Current assisted reproduction techniques (ARTs) are unable to treat all patients, and alternative strategies need to be developed to meet the legitimate desire to be a father. The aim of this study was to evaluate the potential of naked mRNA as a means to induce expression of exogenous proteins in male germ cells in a preclinical adult mouse model. Based on previous studies using electroporation, we investigated whether the combination of the injection of naked mRNA and in vivo electroporation could lead to efficient protein expression in spermatogenic cells. We chose to first study the efficiency of capped and poly-A-tailed mRNA coding for reporter proteins and compared results to those obtained with a non-integrative enhanced episomal vector plasmid. No EEV plasmid has ever been tested in the context of infertility treatment before this study.

Using an adult mouse model, we optimized the micro-injection and electroporation method described by Michaelis et al [7]. We show that the microinjection through the rete testis did not provide a homogenous distribution of the particles throughout the seminiferous tubules. Nevertheless, the seminiferous tubules remained intact, with no signal detected in the peritubular space. The peripheral expression observed was not due to the close vicinity of cells to the electrodes, but rather to a peripheral dispersal of the injected solution, as shown by the distribution of the fluorescent i-particles NIRFiP-180. Our results also showed that the combination of injection and electroporation is safe when electric pulses are carefully controlled. Using such a protocol, we were able to induce the expression of 3 reporter proteins, GFP, mCherry and luciferase in the testis by mRNA injection/electroporation. Based on whole testes fluorescence and, for the first time, in vivo bioluminescence imaging of testes, we characterized the kinetics of mRNA expression. Expression was observed to decrease rapidly after electroporation, with the total duration ranging from 15 to 21 days, depending on the molecule being expressed. Using whole testicular optical clearing, the reporter proteins were synthetized from the injected mRNA in Sertoli cells, spermatogonia, spermatocytes, round spermatids, and spermatozoa from day 1, and were still detectable after 1 week.

When using the EEV vector, expression persisted for longer – up to 119 days – but with a lower yield of seminiferous tubule and cellular transfection. In particular, a lower level of transfection of germ cells was observed than with the mRNA, and after 1 week, only Sertoli cells remained transfected.

These results suggest that although EEV expression lasted longer, mRNAs, by targeting more efficiently male germ cells and allowing higher transfection yields of seminiferous tubules, could be a more effective and potent tool to express exogenous proteins in germ cells. By expressing a missing protein in the case of male infertility due to monogenic causes, it could be possible to restore failed spermatogenesis and thus to treat infertility.

ARMC2 is expressed in late spermatogenesis stages

Dissociated WT testicular cells contained ARMC2 in the flagellum of spermatids. No Armc2 expression was detected in earlier germ cell type lineages like spermatogonia. These results suggest that the ARMC2 protein is expressed late during spermatogenesis, which explains why motile sperm were only found in the cauda epididymis from 3 weeks after injection in our treated mice. Indeed, full spermiogenesis (from round cell to sperm) takes around 15 days [17], and the journey across the epididymis lasts around 8 days, making a total of 3 weeks. Our results also confirm those recently published by Lechtreck et al. [18] from their study of the role of ARMC2 in the Intra-Flagellar Transport (IFT) of radial spokes in Chlamydomonas. Proteins are supplied to the flagellum by IFT trains, with some proteins binding directly to IFT train subcomplexes. This is the case, for example, for tubulin, which binds to the N-terminal domains of the central IFT-B proteins IFT74 and IFT81 [19]. In other cases, protein binding is mediated by adapters. For example, the placement of the Outer Dynein Arms (ODA) requires an adapter protein ODA16 [20]. Lechtreck et al. [18] suggested that the transport of the radial spokes along the flagellum involves ARMC2 acting as an IFT adapter [18]. The presence of ARMC2 in the flagella of elongating spermatid supports this hypothesis.

Exogenous Armc2-mRNA expression rescued the motility of oligo-astheno-teratozoospermic sperm

This is the first demonstration that proteins can be expressed in the testis following electroporation with capped and poly-A-tailed mRNA.

Our objective was to develop a new targeted therapeutic approach for infertility associated with monogenic defects. In this preclinical study, we tested the ability of mRNA to restore flagellar motility in a mouse model deficient for Armc2 that displays an oligo-astheno-teratozoospermia .

Our results strongly suggest that the strategy is safe, as injection and electroporation of Armc2-mRNA or EEV-Armc2 had no effect on testicular morphology or weight. More importantly, the technique was effective, with motile sperm cells found in Armc2 KO epidydimal cauda 3 and 5 weeks after Armc2-mRNA injection. Naked mRNA injection/electroporation is therefore a promising method to treat infertility. In contrast, no motile spermatozoa were found after injection/electroporation of EEV-Armc2, confirming our previous results suggesting that this nucleic tool does not efficiently enter or transfect germ cells.

The method deployed here involved optimized, capped mRNA adapted for the expression of Armc2. The vector was injected via the rete testis into adult Armc2 KO mice. The testes were then electroporated. The results obtained showed that the injected and electroporated testes had a morphology and weight similar to the controls at all times after treatment, thus validating the safety of the method in this preclinical model of male infertility.

Although motile sperm cells were observed at 3- and 5-weeks after Armc2-mRNA injection into the epidydimal cauda of Armc2 KO mice, not all injected mice were efficiently treated. For instance, in a group of 5 mice littermates treated and euthanized at the same time (5 weeks post-injection), only 3 of the 5 had motile sperm cells. The absence of motile sperm at 5 weeks may be due to the type of spermatogenic cells transfected during the electroporation step. The transfected cells may differ between individuals, influenced by the injection and the position of the electrodes during electroporation. If the mRNA transfection takes place in a spermatogonia, it will take more than 6 weeks (including the time to cross the epididymis) before motile epididymal sperm cells emerge. This timeline could explain the absence of motile sperm in some littermates at 5 weeks.

Although we rescued the motility of oligo-astheno-teratozoospermic sperm, the transfection rate needs to be improved to obtain larger numbers of sperm cells. Indeed, our experiments did not provide enough motile sperm cells to carry out IVF or to allow natural mating. This low proportion could be explained by the mode of cell transfection. Nevertheless, ICSI with normal sperm could be implemented, increasing the overall safety of the technique.

Naked mRNA, a new therapeutic strategy to treat severe infertility

Non-obstructive Azoospermia (NOA) and severe oligozoospermia (SO) are the most severe disorders of spermatogenesis and are the most likely to be of genetic origin. NOA is defined by the complete absence of spermatozoa in the ejaculate. Approximately 10–15% of infertile men have azoospermia, and a further 15% have SO [21]. For patients with NOA, few clinical solutions are currently available. Generally, testicular sperm extraction is attempted to collect some spermatozoa from the seminiferous tubules, which can then be used for ICSI [22]. When no sperm are retrieved, intra-conjugal conception is impossible. Our results could also be applied to treat NOA, even if synthesis of the target protein does not persist throughout the whole spermatogenesis cycle. Indeed, it has been shown that spermatogenic proteins have a very low turnover [23] because synthesized proteins are stored in sperm organelles which remain stable until fertilization. Consequently, fertility rescue can probably be achieved even for proteins expressed earlier and involved in meiosis. Moreover, the presence of a few sperm in the ejaculate could be sufficient to obtain pups because with ICSI, only one sperm is required per oocyte. In other words, the persistent expression is not mandatory to treat infertility.

We used in this report electroporation to transfect the testicular cells. Although we injected 2/3 of the testes, the transfection is limited by the size of the electrodes which produce a restricted electric field. To increase the testicular transfection rate, encapsulation of mRNA into lipids nanoparticles could be used, as used for Covid vaccination [24]. During the writing of this manuscript, Dong team [25], used a self-amplifying RNA (saRNA) encapsulated in cholesterol-amino-phosphate derived lipid nanoparticle to restore spermatogenesis in infertile mice. They successfully restore the expression of the DNA Meiotic Recombinase 1 (DMC1) in Dmc1 KO infertile mice by injecting a self-amplifying RNA-Dmc1 in the testes. Dmc1 is 37 KDa protein implicated in the meiotic recombination during sperm meiosis [26–28]. Although this impressive article reported successful treatment of mouse infertility, a safety study should be carried out to validate such a method. Indeed, saRNA are genetically engineered replicons derived from self-replicating single-stranded RNA viruses [29]. They can be delivered encapsulated in lipid nanoparticles, or as a completely synthetic saRNA produced after the in vitro transcription. The saRNA contains the alphavirus replicase genes and encodes an RNA-dependent RNA polymerase (RdRP) complex which amplifies synthetic transcripts in situ and the target RNA sequence. The target RNA is expressed at high levels as a separate entity. As a result of their self-replicative activity, saRNAs can be delivered at lower concentrations than conventional mRNA to achieve comparable expression levels [30]. Moreover, saRNA constructs need to be condensed by a cationic carrier into a nanoparticle measuring ∼100 nm to enable their uptake into target cells and protect the saRNA from degradation [31]. Finally, saRNA will amplify the RNA without cellular regulation. For all these reasons, if such a strategy is to be pursued, a potential toxicity effect due to saRNA over expression must be investigated in the testes and progeny. Another difficulty with saRNA relates to the size of the molecular construct. saRNA sequences are large and complex. The length of the sequence RdRP is around 7 kilobases, which often makes the full length of saRNA more than 9 kilobases once the sequence for the protein of interest has been integrated [31]. Dong et al. [25] successfully used their saRNA construct to rescue spermatogenesis failure induced by the absence of the small protein Dmc1, but it may be more challenging with larger proteins such as the structural proteins involved in OAT, including the 98-kDa ARMC2 [13].

Our next step will be to assess whether encapsulating Armc2-mRNA in LNP-CAP could allow a larger number of germ cells to be transfected.

In conclusion, this paper presents the first in vivo testicular injection and electroporation of capped and poly A tailed mRNA, demonstrating that it is an efficient strategy to transfect male germ cells and that the duration of expression of the resulting proteins is compatible with restoring spermatogenesis. Our comprehensive study revealed mRNA to be more efficient than an episomal vector, despite longer-lived expression in male germ cells with EEV. The difference was linked to EEV achieving a lower rate of seminiferous tubule transfection and a shorter duration of expression in germ cells. Our results also showed for the first time that functional sperm motility can be restored in mice with an oligo-astheno-teratozoospermia phenotype using a technique combining injection and electroporation of capped and poly-A-tailed mRNA. The findings presented open new opportunities to develop efficient strategies to treat male infertility with monogenic causes.

Fig supp 1:

Fig supp 2:

Fig supp 3:

Fig supp 4:

Fig supp 5:

Videos 1 and 2: 3D-microscopic reconstructions of face 1 and 2 of a testis injected with GFP-mRNA.

Video 3: CASA recording of WT epidydimal sperm cells

Video 4: CASA recording of Armc2 KO epidydimal sperm cells

Videos 5 and6: CASA recordings of epidydimal sperm cells from Armc2 KO mice on day 21 and day 35, respectively, after injection/electroporation with Armc2-mRNA.

Acknowledgements

This work was supported by ANR-20-CE18-0007 grant to JE. This work was supported by the Fondation pour la Recherche Médicale, grant number « ECO202006011669 » to CV.