Testicular mRNA‐LNP Delivery: A Novel Therapy for Genetic Spermatogenic Disorders

All animal care and experiments complied with the guidelines of the National Institutes of Health and were approved by the Animal Care Committee of Shanghai General Hospital (2022AWS0287). The Maps KO mouse in a C57BL/6 genetic background was generated by Cyagen Biosciences (Suzhou, China), and a 1.8 kb genomic DNA fragment spanning exon 3 was deleted using the CRISPR/Cas9‐mediated genome‐editing system, the strategy was according to the previous study [30]. The targeted gRNAs and Cas9 protein were co‐injected into the zygotes from C57BL/6 mice to generate targeted knock‐out offspring (gRNA: gRNA‐A1: TCACCAGTACATTTTCGTCTTGG; gRNA‐A2: GTTTAATTCATACCGATGATTGG). The F0 founder mice were identified by PCR and DNA sequencing analysis and were bred with wild‐type mice to obtain F1 heterozygous mutant (Maps^+/−) mice, and the homozygous mice (Maps^−/−) were obtained from breeding pairs of Maps^+/− male and female mice. DNA isolated from tail was used for genotyping by PCR and Sanger sequencing as described previously [8, 30], the primers listed in Table S1. All mice were maintained under controlled environmental conditions (temperature: 24 ± 1°C; relative humidity: 50%–60%) with a standard 12‐h light‐and‐dark cycle. The HEK‐293T (RRID: CVCL_0063), TM3 (RRID: CVCL_4326), and TM4 (RRID: CVCL_4327) cells were obtained from American Type Culture Collection, and cultured with Dulbecco's Modified Eagle Medium (DMEM, Gibco, Waltham, MA, USA) supplemented with 10% fetal bovine serum (FBS, HyClone, Marlborough, MA, USA) under standard culture conditions (37°C, 5% CO₂) without mycoplasma contamination.

LNPs were prepared by mixing lipids in an ethanol phase with mRNA in an aqueous phase in two syringe pumps. In brief, the ethanol phase was prepared by solubilizing a mixture of ionizable lipid, cholesterol (AVT Pharmaceutical Tech, Shanghai, China), 1,2‐distearoyl‐sn‐glycero‐3‐phosphocholine (DSPC, AVT Pharmaceutical Tech, Shanghai, China), and 1,2‐dimyristoyl‐rac‐glycero‐3‐methoxypolyethylene glycol‐2000 (DMG‐PEG 2000, AVT Pharmaceutical Tech, Shanghai, China) at a predetermined molar ratio in ethanol (50: 38.5:10:1.5) (Figure 1a). The aqueous phase was prepared in 10 mM citrate buffer (pH 4.0) with enhanced green fluorescent protein mRNA (EGFP mRNA, CATUG Biotechnology, Suzhou, China), Msh5, and Maps mRNA. Syringe pumps were used to mix the aqueous and ethanol phases at a ratio of 3:1. The resulting LNPs were dialyzed against 20 mM Tris in a 100 000 MWCO dialysis tube at 4°C for 4 h, and refreshed in 20 mM Tris buffer, following dialysis at 4°C overnight. A RiboGreen RNA assay (Invitrogen) was used to calculate the nucleic acid encapsulation, and a dynamic light scattering (DLS, Zetasizer, Malvern, UK) was used to measure the size and polydispersity index (PDI) of the LNPs (Figure 1b, Table S2).

Pool1‐LNP3 encapsulating luciferase mRNA was administered by rete testis microinjection. A 4 µg luciferase mRNA‐LNP 3 was injected into the left testis of each wildtype mouse. On days 1, 3, 5, 7, and 9 after injection, fluorescence imaging was captured using Living Image software on In vivo Imaging System (IVIS, PerkinElmer, Waltham, MA, USA). Ten minutes before imaging, d‐luciferin potassium salt (Thermo Fisher Scientific, Waltham, MA, USA) was injected intraperitoneal (i.p.) to mice at a dose of 150 mg kg⁻¹, exposure time: 30 s; Binning: Medium; F/Stop: f/1.

Total RNA was isolated from mouse testicular tissue using TRIzol reagent (Invitrogen, Waltham, MA, USA) following the manufacturer's protocol. RNA integrity and purity were verified using a NanoDrop (Thermo Fisher Scientific, Waltham, MA, USA). Reverse transcription was performed using 1 µg RNA with a First Strand cDNA Synthesis Kit (Thermo Fisher Scientific, Waltham, MA, USA).

For mouse Msh5 and Maps mRNA production, the double‐stranded DNA in vitro transcription (IVT) template was synthesized and cloned into a modified pUC19 plasmid vector containing a T7 RNA polymerase promoter sequence. The plasmids were linearized by BspQI (10 units per 1 µg DNA, 37°C overnight) (R0712S, New England BioLabs, Ipswich, MA, USA) [31] and purified using Monarch PCR & DNA Cleanup spin columns (New England BioLabs, Ipswich, MA, USA) [32]. IVT was performed using a HiScribe T7 mRNA Kit with CleanCap Reagent AG (New England BioLabs, Ipswich, MA, USA) following the manufacturer's protocol, and UTP was all replaced by N1‐methyl‐pseudouridine. RNA was purified using Monarch RNA Cleanup spin columns (New England BioLabs, Ipswich, MA, USA). mRNA product integrity was validated using native agarose gel electrophoresis and the mRNA was stored frozen at −80°C for later use.

DNA ploidy assay was performed as previously described with some modifications [33]. In brief, after the removal of tunica albuginea, testicular tissues were enzymatically digested with 1 mg mL⁻¹ collagenase type IV (diluted in DMEM) at 37°C for 5 min. Following centrifugation for 1 min at 100 g, the supernatant was removed. Subsequently, 0.6 mg mL⁻¹ trypsin was added to digest the tubules at 37°C for 10 min, and FBS was added for stopping digestion. The suspension was filtered through a 40 µm cell strainer and centrifuged for 10 min at 500 g. Finally, cells were suspended in 1× phosphatic buffer solution (PBS) and labeled with propidium iodide (PI) by 1 µg/1 × 10⁵ cells for 5 min on ice. FACS measurements were performed on BD FACSMelody Cell Sorter. Gate was set as previously reported [33].

Briefly, all mice were anesthetized via i.p. administration of avertin (250 mg kg⁻¹) and positioned in dorsal recumbency. Following aseptic preparation, a single incision was made approximately 1.5 cm up to the genitals. The testes were pulled out by holding the fat pad. mRNA‐LNP or PBS was injected into the rete testis using a glass capillary under a stereomicroscope(Figure 1c). The harvested testes were fixed in 4% paraformaldehyde (PFA) for 24 h at 4°C, paraffin‐embedded, and sectioned at 5 µm thickness using a rotary microtome (RM2255, Leica, Wetzlar, Germany). Major organs, including the heart, liver, spleen, lungs, kidney and brain of wildtype mice were collected 24 h post injection and major organs of Msh5^D486Y/D486Y and Maps KO mice were collected on day 21 post injection. All the tissues were fixed in PFA for 24 h and paraffin‐embedded, 5 µm sections were generated using a rotary microtome (RM2255, Leica, Wetzlar, Germany).

Before staining, tissue sections were dewaxed in xylene, rehydrated using a gradient series of ethanol solutions, and washed in distilled water. After rehydration, sections were processed for antigen retrieval with 10% sodium citrate (pH 6.0) at 115 °C for 15 min. Non‐specific binding was blocked with 5% normal donkey serum containing 0.3% Triton X‐100 for 2 h at room temperature. Primary antibodies (Table) were applied overnight at 4°C, followed by species‐matched Alexa Fluor‐conjugated secondary antibodies (1:1000, Invitrogen, Waltham, MA, USA) for 2 h. Nuclear counterstaining utilized Hoechst 33342 (1:1000, Roche, Basel, Switzerland), with imaging performed on a Leica SP8 confocal system (Leica, Wetzlar, Germany) equipped with LAS X software. S3

Tissue sections (including testis, epididymis, heart, liver, spleen, lungs, kidneys, and brain) were dewaxed in xylene and rehydrated through a graded ethanol series followed by distilled water rinses. Sections were immersed in hematoxylin solution for 6 min, briefly differentiated in acid alcohol, and blued in tap water. Subsequently, slides were counterstained in acidified eosin solution for 45 s, dehydrated through ascending ethanol gradients, cleared in xylene, and mounted. Whole‐slide digital images were acquired using a Panoramic SCAN slide scanner (3D Histech). Histopathological evaluation for tissue lesions was performed by the experienced pathologists. For Hematoxylin‐Eosin‐Saffron (HES) Staining: tissue sections were mounted on glass slides, deparaffinized in xylene, and rehydrated using graded alcohols. Nuclei were stained by immersion in hematoxylin solution for one min, followed by rinsing in distilled water. Cytoplasmic differentiation was achieved via a 5 s incubation in 1% acid alcohol. Sections were then blued in running tap water for two min. Cytoplasmic counterstaining was performed by incubating slides in eosin solution for five min, followed by a brief distilled water rinse. For saffron counterstaining, slides were immersed for 15 min in saffron (HY‐D0215, MedChemExpress, Monmouth Junction, NJ, USA) solution (warmed to 35°C–40°C prior to use). Then sections were microscopically examined as recommended following a rapid distilled water rinse. Final dehydration was accomplished by two sequential five min incubations in anhydrous ethanol. Slides were cleared briefly in xylene and cover slipped using neutral gum.

Caput spermatozoa chromatin maturity was assessed using the aniline blue staining method (R30404, Yuanye biotechnology, Shanghai, China). Spermatozoa were smeared onto an adhesive‐coated glass slide and air‐dried. The smears were fixed with 4% PFA for 30 min at room temperature, followed by rinsing with distilled water. The smears were then stained with aniline blue staining solution for 5 min at room temperature, rinsed with distilled water, dehydrated through a graded series of ethanol, cleared with xylene, and mounted with neutral gum. Spermatozoa with immature chromatin appeared purple‐blue, while mature sperm appeared light blue or unstained.

Caput spermatozoa were fixed in 4% PFA, uniformly spread onto glass microscope slides, and overlaid with 100 µL of CMA3 staining solution (0.25 mg mL⁻¹ Chromomycin A3 in McIlvaine buffer; (MedChemExpress, Monmouth Junction, NJ, USA) [34]. The McIlvaine buffer (pH 7.0) was prepared with 7 mL of 0.1 M citric acid and 33 mL of 0.2 M Na₂HPO₄, supplemented with 10 mM MgCl₂. Slides were protected from light during a 20 min incubation period, followed by three sequential washes in McIlvaine buffer. Fluorescent signals within individual sperm nuclei were identified using Leica SP8 confocal system (Leica, Wetzlar, Germany) equipped with LAS X software, with 405 nm excitation and brightfield overlay. Spermatozoa were classified into two groups based on fluorescence patterns.

Caput spermatozoa morphology was evaluated using Shorr's staining method (R20357, Yuanye biotechnology, Shanghai, China). Sperm smears were prepared on a glass slide and immediately fixed in 95% ethanol/glacial acetic acid (3:1) fixative solution for 10–15 min. The fixed slides were rehydrated through a graded series of ethanol and rinsed with distilled water. Slides were sequentially stained with hematoxylin, G–6 orange, and EA–50 green, with intermediate rinses and a differentiation step with 1% acetic acid. After staining, the slides were dehydrated, cleared with xylene, and mounted for microscopic examination.

The tunica albuginea of the mouse testes was removed using microsurgical forceps, then the testes were transferred into 6‐cm dish and washed by PBS. All the tissues were meticulously isolated into single seminiferous tubule using microsurgical forceps and subsequently suspended in hypotonic buffer (30 mM Tris–HCl, 50 mM sucrose, 17 mM trisodium citrate dehydrate, 5 mM EDTA, 2.5 mM dithiothreitol, and 1 mM phenylmethylsulfonyl fluoride) for 25 min. The seminiferous tubules were torn into small pieces and pipetted 15 µL cell suspension spread on a glass slide in a thin layer of PFA solution containing 0.15% Triton X‐100. Slides were placed in a humid chamber and kept for at least 3 h at room temperature. After air drying, the slides were stored at −80°C for further experiments. The slide was then blocked with 5% Bovine Serum Albumin (BSA) and incubated with primary antibodies (Table) overnight at 4°C. After washing, the cells were incubated with the secondary antibody (Table) for 1 h at room temperature. Images were captured with Leica SP8 confocal system (Leica, Wetzlar, Germany). S3 S3

Serum samples of wildtype mice were collected 24 h after EGFP mRNA‐LNP3 delivery. Mouse IL‐1 beta ELISA Kit (Servicebio, Wuhan, China) was used to evaluate the IL‐1β level in the blood serum following the instructions of the manufacturer.

Testes and other major organs (heart, liver, spleen, lungs, kidney, and brain) were collected from wildtype mice 24 h post‐injection of PBS, 2 µg EGFP mRNA‐LNP 3 and 4 µg EGFP mRNA‐LNP 3, biological replicates of each group were 3. Tissues were lysed with RIPA Lysis and Extraction Buffer (Thermo Fisher Scientific, Waltham, MA, USA) with 1% protease inhibitor mixture (Thermo Fisher Scientific, Waltham, MA, USA) in accordance with the manufacturer's guidelines. Protein extracts were acquired from the supernatants and were boiled for 10 min, and equal amounts from each sample were loaded onto 10% SDS–PAGE gels for gel electrophoresis (80 V in the stacking gel and 120 V in the resolving gel). After electrophoresis, the proteins in the gels were transferred to polyvinylidene difluoride membranes (Bio‐Rad, Hercules, CA, USA) and blocked using a 5% milk solution for 1 h at room temperature. These membranes were then incubated with primary antibodies (Table S3) at 4°C overnight and subsequently with secondary antibodies according to the protocol provided by the manufacturer. β‐actin was used as the loading control.

Female B6D2F1 mice were super‐ovulated by the injection of 5 IU equine chorionic gonadotropin (San‐Sheng Pharmaceutical, Shenyang, China) and followed by 5 IU human chorionic gonadotropin (hCG, San‐Sheng Pharmaceutical, Shenyang, China) 48 h later. Cumulus‐oocyte complexes were collected from oviducts at 14–16 h after hCG injection, and they were placed in HEPES‐buffered CZB medium and treated with 0.1% hyaluronidase (Sigma‐Aldrich, St. Louis, MO, USA) to disperse cumulus cells. Sperm were collected from the testis of Msh5^{D486Y/ D486Y} male mice (11 weeks old) at 3 weeks after mRNA‐LNP injection and cultured in HEPES‐buffered CZB medium for 15 min. The sperm head was separated from the tail by the application of several Piezo pulses, and the head was then injected into the oocyte according to the method described by Kimura and Yanagimachi [35]. Following micromanipulation, all injected oocytes were cultured in G1‐plus (Vitrolife, Gothenburg, Sweden) under standard culture conditions. Pronucleus formation was checked at 6 h after ICSI, and outcomes were scored up to the blastocyst stage. For embryo transfer, two‐cell embryos were surgically transferred into the oviduct of each 8‐week‐old pseudo‐pregnant female. In this study, pseudo‐pregnant females were generated by mating them with sterile males that had undergone bilateral vasectomy [36]. The surgical procedure for generating vasectomized males was carried out as follows: Mature male mice were anesthetized and placed on a warming pad. After aseptic preparation of the surgical site, a small midline incision (∼0.5–1.0 cm) was made in the lower abdomen, just superior to the scrotum. Through this incision, the vas deferens on each side was sequentially identified and carefully separated from the accompanying blood vessels by blunt dissection. Each vas deferens was then permanently occluded by double ligation with silk sutures, followed by transection between the ligatures. After confirming hemostasis, the testes were returned to the abdominal cavity, and the skin incision was closed with sutures. Postoperatively, the males were allowed to recover for at least two weeks to ensure complete clearance of residual sperm before being used to induce pseudo‐pregnancy. Sterility was confirmed by their ability to form vaginal plugs in females without resulting in pregnancy.

Genotyping of single sperm and single blastocyst was performed using a Multiple Annealing and Looping‐Based Amplification Cycles (MALBAC) WGA kit (XK‐028‐24, Yikon, Suzhou, China) following the manufacturer's protocol. Briefly, individual sperm or blastocysts were lysed in a 5 µL reaction mixture containing 4.5 µL lysis buffer and 0.5 µL lysis enzyme. The samples were incubated at 50°C for 20 min, followed by enzyme inactivation at 80°C for 10 min. To avoid sample loss, cell suspensions were not vortexed after transfer and were briefly centrifuged. For amplification, 62 µL of reaction mix (60 µL processing buffer and 2 µL processing enzyme per sample) was added to the lysate. Thermal cycling was performed under the following conditions: initial denaturation at 94°C for 3 min; 8 cycles of 20 s at 10°C, 30 s at 30°C, 40 s at 50°C, 2 min at 70°C, and 20 s at 95°C; followed by 17–21 cycles (17 cycles for single blastocysts, 19–21 cycles for single sperm cells) of 20 sec at 94°C, 15 s at 58°C, and 2 min at 72°C; and a final extension at 72°C for 5 min. Cycle numbers were optimized based on starting material quantity, as recommended for low‐input samples. Amplified products were stored at 4°C until downstream genotyping analysis.

For genotyping, thermal cycling was performed under the following conditions: initial denaturation at 94°C for 3 min; 35 cycles of 30 s at 94°C, 30 s at 60°C, 30 s at 72°C, and a final extension at 72°C for 5 min. Genotyping strategy was described in the method ‘mice models’ as described above.

Statistical analyses were performed using SPSS Statistics 24. Mann‐Whitney U test was used for the results in Figure 2e. For the results in Figures 2, 3, 4, 6, 7d, Figures S1d–f and S2e, intergroup comparisons among the three groups were conducted using the non‐parametric Kruskal–Wallis test. When the overall test indicated statistical significance, pairwise post hoc comparisons were conducted with Bonferroni‐adjusted p values to control for type I error. Data were represented as median and interquartile range (IQR), n ≥ 3 biologically independent replicates per group. Statistical significance was defined as ^*p < 0.05, ^**p < 0.01, and ns indicating no significant difference.

Currently, testicular mRNA LNP delivery lacks efficient, safe, and germ cell‐specific vehicles. To address these challenges, we constructed a combinatorial LNP library comprising 30 distinct formulations with different ionizable lipid architectures. Each LNP formulation was systematically prepared with four essential components: ionizable lipid, cholesterol, DSPC, and DMG‐PEG‐2000 with 50: 38.5: 10: 1.5 molar ratio, encapsulating enhanced green fluorescent protein (EGFP)‐encoding mRNA (Figure 1a). A comprehensive analysis was performed to characterize physicochemical properties of the LNP, including size (nm) and PDI (Figure 1b).

All 30 LNPs were divided into three combinatorial pools (Pool 1–3) and transfected into three kinds of cell lines (human embryonic kidney cell line (HEK293T), mouse Sertoli cell line (TM4), and mouse Leydig cell line (TM3) to confirm their expression of EGFP (Figure S1a–c). EGFP was expressed in 3 cell lines after delivery of three LNP pools, but with different efficiency (Figure S1d–f). To evaluate the delivery efficiency in vivo, LNP of three pools were separately injected to the seminiferous tubules of adult wildtype male mice via rete testis microinjection (Figure 1c). Mice were euthanized after 24 h of injection, and testicular wholemount fluorescence imaging confirmed EGFP expression across seminiferous tubules in all three LNP pools (Figure 1d). To delineate the cellular specificity of LNP‐mediated mRNA delivery, testicular tissues underwent immunofluorescence staining with germ cell specific markers DEAD‐Box Helicase 4 (DDX4), which was highly expressed in the cytoplasm of spermatocytes [37, 38]. The results showed that LNP of Pool1 emerged as the only candidate presenting preferential mRNA delivery to germ cells (Figure 1e). To further identify the spermatocyte tropic LNP, we injected all 10 LNPs of Pool1 individually to the seminiferous tubules of wildtype male mice. Immunofluorescence staining confirmed that Pool1‐LNP3 (hereafter referred to as LNP3) exclusively targeted spermatocytes (Figure 2a,b, Figure S2e), and it was shown that EGFP was highly expressed after delivery of EGFP mRNA‐Pool1‐LNP3 in 293T cells (Figure S2a). The chemical structure of the ionizable lipid used for the preparation of LNP3 is shown in Figure S3. However, EGFP was mainly expressed in Sertoli cells after injection of other LNPs of Pool1 (Figure S2b–d). To further assess the duration of intracellular mRNA expression following mRNA delivery, we encapsulated luciferase‐encoding mRNA within LNP3 and injected it into the seminiferous tubules via rete testis microinjection. Images were recorded on days 1, 3, 5, 7, and 9 after injection by an in vivo imaging system (Figure 2c). It was showed sustained luciferase activity from day 1 to day 7 after injection (Figure 2d,e). Altogether, these results demonstrated that LNP3 were superior vehicles for mRNA delivery into spermatocytes and the sustained expression could be effective for the following rescue studies.

To evaluate whether spermatogenic disorders could be rescued via the novel mRNA delivery strategy, Msh5^D486Y/D486Y male mice with meiotic arrest as described as previously was chosen in the current study [8]. Msh5 mRNA and EGFP mRNA were encapsulated in LNP3 with 1:1 mass ratio and injected to Msh5^D486Y/D486Y male mice (Figure S4a). Testicular wholemount fluorescence imaging on days 7, 14, and 21 after injection revealed the expression of EGFP sustained in the seminiferous tubules for more than 14 days (Figure S4b). Hematoxylin and eosin staining revealed that while spermatogenesis was arrested at the spermatocyte stage in untreated Msh5^{D486Y/ D486Y} testis, elongated spermatids occurred in the seminiferous tubules by day 21 and disappeared on day 28 after delivery (Figure 3a). Furthermore, we performed immunofluorescence staining for germ cell markers of each stage, including SYCP3 (spermatocyte marker), AKAP3 (spermatid marker), TNP1 (elongating spermatid marker), and peanut agglutinin (PNA)‐lectin to label the acrosome of the spermatid. The results confirmed the presence of PNA‐positive round spermatids as early as day 7 post‐treatment, albeit in low numbers (Figure S4c). By day 14 AKAP3‐positive spermatids had increased significantly in the seminiferous tubules, and by day 21, most spermatids in the tubules had completed spermiogenesis (Figure 3b,c, Figure S4d). However, elongated spermatids were absent in the testes collected on days 28 and 35 after injection, indicating transient expression of the mRNA within spermatocytes, sufficient to facilitate meiotic progression (Figure 3b,c).

To evaluate the therapeutic rescue efficiency of mRNA‐LNP3 delivery, we quantified seminiferous tubules with PNA‐positive spermatids at 21 days after injection. The results revealed that 31.91 ± 2.97% of tubules cross‐sections exhibited PNA‐positive spermatids within their luminal compartments, indicating restored spermatogenesis in these seminiferous tubules (Figure 3d,e). Flow cytometric quantification of testicular cell suspensions showed 55.9% of the testicular cells were haploid germ cells after mRNA delivery (1N population), demonstrating significant rescue efficiency compared with untreated Msh5^{D486Y/ D486Y} testis (Figure 3f). Furthermore, cytologic examinations revealed that few spermatozoa in the caput epididymis at 28 days after injection showed non‐progressive motility (Movie S1). Also, CMA3 staining, aniline blue staining, Shorr's staining, and PNA staining revealed normal head condensation, intact acrosomal structures, and typical flagellar architecture (Figure 3g, Figure S5).

In order to delineate the critical window for presence of elongated spermatids, testis and caput epididymidis were collected on days 14, 16, 18, 20, 22, 24, 26, and 28 after mRNA delivery. H&E, immunofluorescence and HES staining analysis indicated the presence of elongated spermatids in the seminiferous tubules by day 18 and were no longer detectable by day 22 (Figure 4a,c,d, Figures S6 and S7), whereas H&E staining analysis revealed that there were no spermatozoa until day 28 after injection in the epididymis (Figure 4b). The percentages of PNA‐positive seminiferous tubules at 14, 16, and 20 days after injection were 32.89 ± 1.44%, 32.91 ± 3.57%, and 33.47 ± 2.94%, respectively (Figure 4e–g).

To evaluate whether the delivered mRNA integrated into rescued sperm genome, we collected the sperms from the testis on day 20 after injection and performed genotyping on 10 randomly selected sperm (Figure 5a), Sanger sequencing revealed that all 10 sperm samples exhibited a mutation genotype (Msh5 p.D486Y) (Figure 5b,c). To further evaluate the fertility of the mice after Msh5 mRNA‐LNP3 delivery, we microinjected 53 sperms into wildtype female mouse oocytes and observed 48 embryos developed to the 2‐cell stage on day 2 after ICSI. On day 5–6 after ICSI, we observed 37 embryos proceeded to morula stage and 30 embryos proceeded to blastocysts (Figure 5d). Sanger sequencing revealed that all 10 randomly chosen blastocysts exhibited heterozygous genotypes (Msh5 p.D486Y) (Figure 5e,f). After transplantation of the embryos, a live offspring was successfully delivered (Figure 5g). This genetic pattern provides conclusive evidence that the therapeutic mRNA delivery restored the full development of spermatogenesis in Msh5^D486Y/D486Y male mice without genomic editing, highlighting the therapeutic potential of mRNA‐based approaches in the treatment of spermatogenic disorders with genetic defects.

To decipher the mechanism of restoration of spermatogenesis in Msh5^D486Y/D486Y mice after Msh5 mRNA LNP3 delivery, meiotic chromosomal spread analysis was performed to evaluate the meiotic progression and recombination. In wildtype DMC1, foci were abundant during early prophase I and progressively diminished in pachytene, indicating timely strand invasion and homologous recombination progression. In contrast, DMC1 foci of Msh5^D486Y/D486Y spermatocytes persisted into pachytene, suggesting defects in strand invasion, intermediate resolution, and a prolonged recombination process. However, after injection of mRNA‐LNP3, the rescued spermatocytes showed normal DSB formation and repair process by timely diminished DMC1 foci in pachytene (Figure S8a). Furthermore, the result of chromosomal spread analysis by SYCP1 and SYCP3 indicated both the Msh5^D486Y/D486Y and the rescued spermatocytes could form a normal synaptonemal complex (Figure S8b).

The mismatch repair protein MLH1 was used to mark the formation of crossovers during meiotic recombination [39]. In mid‐pachytene spermatocytes, MLH1 foci appeared as ∼22 discrete foci per nucleus, with at least one per homolog pair, evenly distributed along the synapsed axes [40]. The result of chromosomal spread analysis showed almost no MLH1 foci in Msh5^D486Y/D486Y pachytene spermatocytes, which is consistent with previous studies [8], indicating the elimination of crossovers. In contrast, the rescued group displayed a significantly increased number of MLH1 foci, suggesting successful restoration of crossover formation (Figure S8c). Altogether, these results suggested that delivery of Msh5 mRNA LNP3 could rescue the DSB combination defects and meiotic arrest in Msh5^D486Y/D486Y pachytene spermatocytes.

To confirm the safety of Msh5 mRNA‐LNP delivery in vivo, we tested the safety of Pool1‐LNP3 and Msh5 mRNA in vivo, respectively. First, we delivered PBS, 2 and 4 µg EGFP mRNA‐LNP3 into seminiferous tubules of wildtype mice via rete testis microinjection to explore the safety of LNP. And multiple tissues throughout the body were collected, including the testis, blood, heart, liver, spleen, lungs, kidneys, and brain after 24 h. Gross photograph showed no significant abnormalities in any organs assessed (Figure 6a). Also, immunofluorescence analysis revealed that EGFP mRNA did not leak into other major organs throughout the body after rete testis injection (Figure 6b). Western blot revealed that EGFP was expressed in testis after mRNA‐LNP delivery via concentration‐dependent manner, and no expressions of EGFP except testis were detected in multiple tissues throughout the body (Figure 6c,d). Furthermore, H&E staining revealed no significant abnormalities in any organs assessed (Figure 6e). ELISA assay showed there were no significant differences of the expression of pro‐inflammatory cytokines (IL‐1β) among the three groups (PBS, 2 and 4 µg, respectively) (Figure 6f).

Subsequently, we evaluated the safety of Msh5 mRNA in vivo via rete testis microinjection of a mixture of Msh5 mRNA and EGFP mRNA at 2 µg in Msh5^{D486Y/ D486Y} male mice. Multiple tissues throughout the body on day 21 after injection were collected, including heart, liver, spleen, lungs, kidney and brain. And the histological analysis showed that there were no significant abnormalities in any organs assessed compared with untreated Msh5^D486Y/D486Y male mice (Figure S9). Collectively, these results illustrated that Msh5 mRNA‐LNP3 delivery via rete testis microinjection in vivo showed no obvious toxicity in multiple tissues throughout the body.

To explore the broad applicability of mRNA‐LNP technology, Maps (C3orf62) KO mice via CRISPR/Cas9 mediated genome editing were generated as described previously [30]. Maps KO mice faithfully recapitulated the human infertility phenotype as we reported previously, showing complete spermatogenic arrest at the spermatocyte stage. Intriguingly, sequential restoration of spermatogenesis was observed after Maps mRNA+EGFP mRNA LNP3 (2 µg per testis) delivery via rete testis microinjection. Specifically, seminiferous tubules exhibited the occurrence of round spermatids on day 14 after injection and elongated spermatids on day 21 after injection (Figure 7a). The results of immunofluorescence staining also revealed that the PNA‐positive round spermatids and AKAP3‐positive elongating spermatids could be detected on day 14 post injection and TP1‐positive elongated spermatids emerged on day 21 (Figure 7b,c, Figure S10). The percentage of PNA‐positive seminiferous tubules at 21 days after injection was about 29.47 ± 3.2% (Figure 7d). Also, a comprehensive safety evaluation was conducted in Maps KO mice after mRNA delivery through histopathological analysis of multiple tissues throughout the body. Tissue specimens including heart, liver, spleen, lungs, kidney, and brain were systematically collected from both untreated Maps KO controls and rescued mice at 21 days after injection. Rigorous histological examination revealed no significant pathological alterations observed in the rescued tissues compared with Maps KO mice (Figure 7e). These findings demonstrated no obvious systemic toxicity associated with mRNA delivery in these mouse models. The successful rescue of meiotic arrest in these models validates the platform's potential for broad applicability in NOA‐affected patients with genetic disorders.