Iron oxide-coupled CRISPR-nCas9-based genome editing assessment in mucopolysaccharidosis IVA mice

Double nCas9-dependent nicking leads to the knock-in of an expression cassette-carrying GALNS cDNA into the ROSA26 locus

Two single guide RNAs (sgRNAs) targeting the ROSA26 locus were designed and cloned into the CRISPR-nCas9 plasmid. Plasmids were transfected into NIH/3T3 mouse fibroblasts using lipofectamine LTX (LP-LTX) (Figures 1A and 1B). After transfection, genomic DNA was used for PCR amplification using primers flanking the expected ROSA26 on-target region, and a ∼0.9-kb PCR product was obtained (Figure 1C). The PCR products were then treated with T7 enzyme to detect mismatch alignments. As observed in Figure 1D, by transfecting the CRISPR-nCas9 plasmid containing either AS- or S-sgRNA, no Cas9 cut was observed. However, when both sgRNAs were cloned into the CRISPR-nCas9 plasmid, 56% on-target cut was obtained on fluorescence-sorted NIH/3T3 cells (Figure 1D). These results were further confirmed by Sanger sequencing (Figure 1E). Insertions and deletions (Indels) were mainly represented by deletions (Figure S1).

Off-target evaluation of the top ten predicted sequences by Sanger sequencing showed no significant changes between untreated and treated NIH/3T3 fibroblasts (Figure S2). In contrast, homologous recombination assays showed an expected band of about 1.5 kb in NIH/3T3 cells transfected with both CRISPR-nCas9 and donor ROSA26:GALNS plasmids, suggesting the insertion of Donor ROSA26:GALNS into ROSA26 locus (Figure S3). Sanger sequencing confirmed the knock-in of the donor template (Figure S3). Overall, these results suggest a highly efficient Cas9-dependent on-target cutting that leads to knock-in of donor ROSA26:GALNS with undetectable off-target effects.

The IONPs-coupled CRISPR-nCas9 system leads to phenotype recovery in long-term-treated MTOL fibroblasts

Before conducting the in vitro evaluation of the CRISPR-nCas9 system, a characterization of the physical properties of the IONPs was carried out. Empty IONPs and the IONPs-coupled CRISPR-Cas9 system (CRISPR-nCas9 and donor ROSA26:GALNS plasmids) showed a mean size of 174.4 nm and 191 nm, respectively, as well as a Z-potential of +44.9 and +39.5 mV, respectively (Figure 2A). Moreover, IONPs prevented the enzymatic degradation of CRISPR-nCas9-associated plasmids in the DpnI assays (Figure 2B), suggesting that electrostatic interactions between positively charged IONPs and negative pDNA may lead to efficient transport and delivery of the CRISPR-nCas9 system in both in vivo and in vitro assays.

MTOL fibroblasts were transfected with IONPs or LP-LTX, followed up to 1 month after treatment, and analyzed for GALNS activity and lysosomal mass. Untreated MTOL fibroblasts showed undetectable intra- and extracellular GALNS activity. After treatment with the CRISPR-nCas9 system using LP-LTX (Figure 2C), the extracellular GALNS activity increased up to WT levels (∼0.0009 U/mL) compared with untreated fibroblast (p ≤ 0.0001). Transfection of the CRISPR-nCas9 system using IONPs led to a GALNS activity increase of 77% of WT levels (Figure 2C). Consistently, intracellular GALNS activity increased up to 63.4% and 61.8% of WT levels after transfection of the CRISPR-nCas9 system by using LP-LTX and IONPs, respectively (Figure 2D). Although an increase in GALNS activity was observed in MTOL fibroblasts transfected only with donor ROSA26:GALNS plasmid, those levels remained as low as 5% WT levels, suggesting that donor plasmid was lost as a consequence of the cell proliferation.

Regarding lysosomal mass, untreated MTOL fibroblasts showed a 2.5-fold increase in the Lysotracker signal compared with WT fibroblasts (Figure 3A). The lysosomal mass was significantly reduced after long-term CRISPR-nCas9 genome editing, with no significant differences compared to WT levels, suggesting normalization of this parameter (Figure 3A). Similar findings were observed in fluorescence microscopy (Figure 3B). Overall, these results indicate that inserting a human GALNS cDNA expression cassette into the ROSA26 locus may result in long-term GALNS expression and lysosomal mass normalization.

IONPs-coupled CRISPR-nCas9 failed to induce chronic toxicity in MTOL mice

Before long-term CRISPR-nCas9-based genome editing in vivo experiments, a biodistribution assessment was performed in 8-week-old animals using DiIC-labeled IONPs due to their convenience for autopsies compared with newborn mice. As observed in Figure 4A, ex vivo imagining analysis revealed that IONPs were distributed mainly to the liver, followed by the lung, muscles, spleen, and kidney. No radiance fluorescence was observed in the brain and heart.

To evaluate the potential chronic toxic effects of IONPs, alone or carrying the CRISPR-nCas9 and donor plasmids, blood samples collected at autopsy were assessed for hepatic and renal biomarkers, as well as iron homeostasis (Figures 4B–4E). After 12 weeks of treatment, no significant changes were observed in the hepatic (Figure 4B) or renal (Figure 4C) biomarkers tested. Consistently, hematoxylin and eosin staining revealed no noticeable histopathological changes (Figure 4D). Iron homeostasis was also unaffected (Figure 4E). Collectively, these results suggest that 2.5 μg IONPs carrying 3.8 μg pDNA are non-toxic in newborn MTOL mice.

IONPS-coupled CRISPR-nCas9 system led to long-term GALNS expression in MTOL mice

Although body weight changes in MTOL mice had not been reported as a significant feature in this MPS IVA model,³² we observed a significant difference in body weight between WT (32.7 ± 0.77 g) and MTOL (28.8 ± 0.83 g) untreated mice after 12 weeks of follow up (p = 0.0006) (Figure 5A). These differences were not evident before 3 weeks after birth, although they increased after weaning. Even though the body weight was not affected upon administration of IONPs carrying Donor ROSA26:GALNS plasmid, at 12 weeks post-treatment, the body weight from MTOL mice increased up to 31.9 ± 1.4 g (p = 0.0197) and 30 ± 0.5 g (p = 0.049) when treated with the naked plasmids and the IONPs-coupled CRISPR-nCas9 system, respectively, compared with untreated MTOL mice (Figure 5A).

As previously reported,^17,32 untreated MTOL mice showed no GALNS activity on plasma or tissues (Figures 5B–5D). After treatment with the IONPs-coupled CRISPR-nCas9 system, the plasma GALNS activity initially increased from 0.5 ± 0.046 U/mL 5 weeks post-treatment up to 1.36 ± 0.206 U/mL at 12 weeks post-treatment, which correspond with 10.2% and 27.7% of WT levels (4.9 ± 0.253 U/mL), respectively (Figure 5B). Although IONPs-coupled donor ROSA26:GALNS did not arise on any detectable GALNS activity in plasma, the administration of the naked CRISPR-nCas9 system resulted in a significant GALNS activity increase (0.2 ± 0.088 U/mL, 4.1% WT levels), compared with untreated animals (Figure 5B), suggesting that plasmids could be taken up by cells in the absence of the IONPs. No changes in plasma GALNS activity were observed in MTOL treated with empty IONPs.

In contrast, the liver (75.4% WT levels), muscle (67.4% WT levels), spleen (34.8% WT levels), heart (57.2% WT levels), kidney (55.4% WT levels), and lung (47.8% WT levels) from MTOL mice treated with IONPs-coupled CRISPR-nCas9 system showed a significant increase in the GALNS activity at 12 weeks post-treatment compared with untreated MTOL mice (Figure 5C). A significant GALNS activity increase was also noticed in the liver (13.3% WT levels) and kidney (9.7% WT levels) of MTOL mice injected with the naked CRISPR-nCas9 system, compared with untreated ones.

For the tibia and trachea, GALNS enzyme activity from the IONPS-coupled CRISPR-nCas9 system MTOL-treated mice increased to 6.8% and 12.6% of WT levels, respectively. In contrast, no activity increase was observed in untreated MTOL mice or treated with the naked plasmids or the IONPs-coupled donor ROSA26:GALNS plasmid (Figure 5D). Finally, no detectable GALNS activity was observed in the brain of any treated mouse (data not shown).

CRISPR-nCas9-dependent long-term GALNS expression led to a significant decrease in the mono-sulfated KS accumulation

A significant mono-sulfated KS increase was observed in plasma, liver, and bone from untreated MTOL animals compared with WT mice (Figure 6), suggesting substrate accumulation due to the absence of GALNS activity.^17,32 Consistent with GALNS activity levels, a significant mono-sulfated KS reduction was also observed in plasma (Figure 6A), liver (Figure 6B), and humerus (Figure 6C) of MTOL mice treated with the IONPs-coupled CRISPR-nCas9 system (Figure 6) compared with untreated MTOL mice, suggesting a positive impact of the CRISPR-nCas9-based GT. Even though mono-sulfated KS levels were not reduced significantly in plasma samples from naked CRISPR-nCas9-treated mice, a significant reduction of mono-sulfated KS levels was detected in the liver of naked CRISPR-nCas9-treated mice, consistent with the GALNS levels observed in the liver of those animals. This behavior was not observed for the humerus of naked CRISPR-nCas9-treated mice.

IONPs-coupled CRISPR-nCas9 system positively impacted pathological findings in the tibia

As previously reported,¹⁷ pathology analysis of the tibia from untreated MTOL mice showed a 1.7-, 1.5-, and 2.0-fold increase in chondrocyte size placed on the growth plate, articular cartilage, and meniscus, compared with WT mice, respectively (Figure 7). Bone regions were positively impacted in MTOL mice treated with the IONPs-coupled CRISPR-nCas9 system. Chondrocyte size improvement in the growth plate and meniscus did not reach WT levels. In contrast, chondrocyte size in the articular cartilage was not different from that observed in WT mice (Figure 7B). Likewise, a vacuolization decrease was also observed in chondrocytes in the growth plate, articular cartilage, and meniscus (Figure 7). Furthermore, vacuolization and columnar structure in pathological findings were scored in the knee joint as previously described.³³ In the tibia, chondrocyte vacuolization was significantly improved in IONPs-coupled CRISPR-nCas9 system-treated MTOL mice compared with untreated ones in the growth plate and articular cartilage. At the same time, a slight non-significant recovery of the column structure was observed in these regions (Table S4). No significant improvement was detected in the femur, ligament, or meniscus (Table S4). Overall, these results suggested that the IONPs-coupled CRISPR-nCas9 system positively impacts key pathological findings in the MPS IVA mouse model.

No anti-GALNS nor anti-nCas9 antibodies response was achieved upon nCas9 treatment

To determine the potential impact of the immune response on the outcome of the CRISPR-nCas9 system, anti-GALNS and anti-nCas9 antibodies were evaluated. As expected, no significant increase in anti-GALNS IgG was observed after 12 weeks of treatment in MTOL mice (Figure 8A), supporting the immunotolerance to human GALNS of this MPS IVA model.³² Under the experimental conditions tested in this study, no increase in anti-nCas9 antibodies was observed in MTOL mice treated with naked plasmids or the IONPS-coupled CRISPR-nCas9 (Figure 8B) regardless of the persistence of nCas9 plasmid (0.003 ± 0.0009 copies per diploid cell) in the liver of MTOL animals after 12 weeks of treatment (Figure S4). We did not observe copies of the donor template or nCas9 plasmid in the testes of the MTOL mice treated after 12 weeks post-treatment (data not shown).

sgRNA design and cloning

Two sgRNA (Table S1) were designed to target the ROSA26 locus (NCBI: NC_000072.7). On- and off-target sequences were predicted using Benchling, Cas Off-finder, and Off-Spotter software (Table S2).³⁴ sgRNAs were obtained from Macrogen Inc. as single oligonucleotides and cloned into the AIO-mCherry plasmid (Addgene: #74120), which carries a mutant version of Cas9 at the RuvC domain (D10A), using the Golden Gate strategy as described previously.³¹ The resulting vector (AIO-mCherry:sgRNA, hereafter CRISPR-nCas9 plasmid) was transformed into Escherichia coli DH5α, purified, sequenced through Sanger, and analyzed by Chromas Pro and MEGA X software. Plasmids used in vitro and in vivo were produced by VectorBuilder Inc.

Donor ROSA26 plasmid

Two homologous recombination arms were designed according to the sgRNA sequences. Left and right arms flanking cytomegalovirus immediate-early enhancer/promoter, Kozak sequence, P2A, enhanced green fluorescent protein (EGFP), and bGH poly(A) signal sequence were synthesized by Gene Art (Thermo Fisher Scientific) (Figure S3A). Two enzyme restriction sequences for MfeI and MluI were included between Kozak and P2A sequences for cloning GALNS cDNA. GALNS cDNA was obtained from a previous donor vector tested for the AAVS1 locus.³¹ The resulting donor was verified through Sanger sequencing. Plasmids used in vitro and in vivo were produced by VectorBuilder Inc.

On- and off-target evaluations

A mismatch-based screening was used to detect the double nicking by nCas9 on the ROSA26 locus. For this, confluent NIH/3T3 cells were transfected or not with CRISPR-nCas9, according to previous reports.^30,64 The cells were harvested 48 h post-transfection to isolate genomic DNA (gDNA) using the Monarch Genomic DNA Purification Kit (New England Biolabs). The resulting gDNA was used for high-fidelity PCR with primers flanking targeting the ROSA26 locus predicted site (Table S3). The mismatches were recognized by T7 endo-nuclease-mediated PCR product digestion (EnGen Mutation Detection Kit, New England Biolabs) following the manufacturer’s instructions. The digested PCR products were electrophoresed on 2% w/v agarose gel (Lonza) and documented with a Geldoc XR + system (Biorad). Indel percentage was calculated using the GelAnalyzer 19.1 software.

A Sanger-based sequencing was performed to confirm the screening performed via T7 assay and the off-target effects of the sgRNAs designed. Briefly, NIH/3T3 cells were sorted by fluorescence-activated cell sorting (FACS) after transfection with the CRISPR-nCas9 plasmids using the mCherry fluorescence (Exc/Emi: 587/610), which is a reporter present in the AIO-mCherry plasmid. FACS was done in a (BD FACSAria) III Cell Sorter (Becton Dickinson). After 8 days post-sorting, gDNA was extracted for high-fidelity PCR using specific primers for on- and off-target sequences (Table S3). An in silico PCR was done to determine the theoretical amplicon sizes using online software from the Genomic Institute of the University of California, Santa Cruz (https://genome.ucsc.edu/cgi-bin/hgPcr). PCR products were then sequenced by Sanger using the same primer for PCR amplification. Electropherograms were used to detect sequence changes by trace decomposition through ICE Analysis from Synthego v2.0.⁷³

Homologous recombination assays

NIH/3T3 cells were transfected with Donor ROSA26:GALNS and CRISPR-Cas9 plasmids for 12 h. Then, gDNA was isolated, and a PCR was conducted using a forward primer flanking the bGH sequence placed in the Donor ROSA26:GALNS and a reverse primer flanking a genomic ROSA26 region absent in the homologous recombination right arm of the donor ROSA26:GALNS (Table S3). The resulting PCR product was sequenced by Sanger.

IONPs characterization

IONPs were purchased from OZ Biosciences in a formulation denominated in vivo DogtorMag, composed of magnetite (Fe₃O₄) and lipid-based nanoparticles positively charged. Plasmid loading was performed according to the manufacturer’s instructions. To determine the maximum pDNA loading capacity of these IONPs, ratios ranging from 1:1 to 1:10 were tested. After IONPs loading, 5 μL loading buffer (New England Biolabs) was added, and 15 μL were loaded on 1% w/v agarose gel stained with GelRed Nucleic Acid stain (Sigma-Aldrich) and resolved at 100 V for 90 min. Images were acquired using a FluorChem E instrument (Bio-Techne) and processed with ImageJ software version 1.53t (NIH). These experiments revealed a maximum ratio of 1:1.5 IONPs:pDNA (Figure S5), which agreed with those reported by Park et al., 2022.⁴³ IONPs:pDNA complexes were further assessed to hydrodynamic diameter and Z potential using dynamic light scattering in a Zeta-Sizer Nano-ZS (Malvern Panalytical).

DpnI restriction assay

To test if IONPs protect the loaded pDNA against endonucleases-mediated digestion, a DpnI digestion was carried out according to previous protocols.²⁷ Briefly, IONPs carrying pDNA in a ratio 1:1.5 IONPs:pDNA in glucose 5% were incubated with 20 U DpnI for 1 h at 37°C. Then, 8 μL loading buffer (New England Biolabs) was added, and the whole reaction volume was resolved at 100 V for 90 min. Images were acquired and processed as detailed previously.

Long-term editing in MTOL fibroblasts

Long-term CRISPR-nCas9-based genome editing was performed in MTOL fibroblasts by transfecting them with CRISPR-nCas9 and donor ROSA26:GALNS plasmids using both LP-LTX (Invitrogen) and IONPs, according to manufacturer’s instructions. Briefly, 40,000 cells/mL MTOL fibroblasts were seeded on 24-well plates (Falcon) 24 h before assays. The media was removed and replaced by a fresh one 1 h before transfection. MTOL fibroblasts were transfected with 0.5 μg CRISPR-nCas9 and 0.5 μg donor plasmids. For IONPs, the plates were placed on a magnetic plate (DogtorMag) for 30 min at room temperature (RT), following the manufacturer’s recommendations. The media was replaced by fresh ones after 24 and 48 h post-transfection with IONPs and LP-LTX, respectively, and followed up to 1 month. Culture media was changed every four days and stored at −20°C until its use in the GALNS assessment.

GALNS activity assays

Intra- and extracellular GALNS activity was measured using 4-MU-Gal-6S at 22 mM (Toronto Chemicals Research) according to previous protocols.^31,74 Briefly, 2 μL of the sample was incubated with 9 μL 4-MU-Gal-6S and 19 μL Milli-Q ultrapure water (Sigma). Samples were incubated for 17 h and then set with 2 μL β-galactosidase (10 mg/mL, Sigma-Aldrich) for 1 h. The reaction was stopped by adding 968 μL glycine-carbonate pH 9.8 buffer. Then, 200 μL were transferred to 96-well black flat-bottom microplates (Corning). Finally, samples were read in a FLUOstar Omega microplate reader (Exc/Emi: 365/450 nm, BMG LabTech). One unit of enzyme activity (U) was defined as the amount of enzyme hydrolyzing of 1 nmol substrate per hour. The standard curve was done with 4-methylumbelliferone (Sigma). Specific enzyme activity was expressed as units per milligram protein, which was determined using a BCA Protein Assay Kit (Thermo Fisher Scientific).

Lysosomal mass determination

One month after treatment, MTOL fibroblasts were stained with LysoTracker Deep Red (Exc/Emi: 647/668 nm; Thermo Fisher Scientific) according to the previous protocols.^31,75 Briefly, the fibroblasts were incubated for 1 h with 50 nM diluted in supplemented DMEM for flow cytometry experiments. The fibroblasts were harvested by trypsinization for flow cytometry experiments and washed twice with 1× PBS. Finally, the fibroblasts were homogenized in 1 X Hanks' Balanced Salt solution plus 2% FBS. We recorded 50,000 ungated events in the NovoCyte 3000 cytometer (Agilent Technologies). Propidium iodide was used for identifying viable fibroblasts. Viable single cells were analyzed in the FlowJo software.

For epifluorescence microscopy, the fibroblasts, seeded on Poly-D lysine-covered coverslips, were incubated for 1 h with 70 nM LysoTracker diluted in supplemented DMEM. The cells were washed once with 1× PBS and fixed with 4% paraformaldehyde in PBS (4% PFA) at RT for 15 min. Three washes with 1× PBS were done and then transferred to slides containing a drop of Prolong Glass Antifade Mountant with NucBlue Stain (Thermo Fisher). Finally, the MTOL fibroblasts were observed under fluorescence microscopy (Axio Observer Z1 microscope, Zeiss). The images were analyzed with the aid of the ImageJ software version 1.53t (NIH).⁷⁶

In vivo assessment of the IONPs-coupled CRISPR-nCas9 system in MTOL mice MPS IVA mice model

Previously, we developed an MPS IVA mice tolerant to human GALN protein (MTOL, GALNS^{tm(hC79S.mC76S)slu}) in the C57BL/6 background.³² MTOL mice contain a transgene expressing mutant hGALNS (GALNS^hC79S) in intron 1 of the mouse genome and an active site mutation in mouse GALNS (GALNS^mC76S) adjacent to exon 2. At birth, homozygotes MTOL male mice were injected into the superficial temporal vein with 2.5 μg IONPs carrying 4 μg CRISPR-nCas9-related plasmids. Untreated MTOL and unaffected C57BL/6 littermates were injected with glucose 5% and included as controls. A total of 100 μL were injected into each experimental group. The dose was determined after dose-response experiments in which IONPs: CRISPR-nCas9 doses ranging between 0 and 20 μg were evaluated (Figure S6). Treated and control animals were weaned for 21 days and raised independently. Mice were euthanized at 12 weeks post-treatment with CO₂ gas. All animal care and experiments were approved by the Research and Ethics Committee of the Faculty of Science at Pontificia Universidad Javeriana (InvestigarPUJ 20289) and the Institutional Animal Care and Use Committee of Nemours Children’s Health (Protocol ID: RSP21-12482-002).

Ex vivo biodistribution analysis

To determine the biodistribution of the IONPs carrying the CRISPR-nCas9 system, an ex vivo analysis was performed according to the protocol described by Park et al.⁴³ Briefly, 20 μg loaded IONPs were fluorescently labeled with DiIC18(5) solid (1,1′-dioctadecyl-3,3,3′,3′-tetramethylindodicarbocyanine, 4-chlorobenzenesulfonate Salt) (ThermoFisher Scientific) for 10 min at RT. Then, IONPs were washed three times in an amicon Ultra-0.5 10K centrifugal device (DID, UFC501008, Millipore) with glucose 5%. DiIC-labeled IONPs were injected into 8-week-old animals via tail injection. Unlabeled IONPs were injected in mice as autofluorescence control. One hour later, the animals were euthanized and autopsied with CO₂. Tissues were collected and visualized in an IVIS Lumina LT Series instrument using a Cy5 filter (Exc/Emi, 644/665 nm), and images were processed with Aura Imaging software (https://www.spectralinvivo.com/imaging/software).

Blood and tissue collection

Blood samples were collected biweekly in EDTA-containing tubes (BD Microtainer) by a venous puncture on the superficial temporal vein with a 22G needle (BD Eclipse Vacutainer), starting at 4 weeks post-injection. At the endpoint, blood was collected in EDTA and non-anticoagulant tubes (BD MicrotainerUSA) to obtain plasma and serum, respectively. Plasma samples were stored at −20°C, while serum samples were stored at −80°C until their processing. Additionally, at autopsy, transcardiac perfusion with NaCl 0.9% w/v was performed with a Masterflex L/S Easy-Load II system (Wavantor) using a 27G X ¾ scalp vein into the left ventricle. Then, the arms, brain, eyes, gonads, heart, kidneys, legs, liver, lungs, muscle, spine, spleen, and trachea were collected and preserved at −80°C or in PFA 4% accordingly.

IONPs-coupled CRISPR-Cas9-related toxicity

To determine the potential side effect of the IONPs empty or carrying the CRISPR-nCas9 system, alanine transaminase (EALT-100, BioAssay Systems), aspartate transaminase (EASTR-100, BioAssay Systems), blood urea (DIUR-100), blood urea nitrogen (calculated from urea values), and plasma creatinine (700460, Cayman Chemical), according with manufacturer’s instructions. Besides, serum iron (702230, Cayman Chemical), total iron-binding capacity (TIBC, 702230, Cayman Chemical), and percentage of saturation of transferrin (calculated from serum iron and TIBC) were conducted. All the reactions were performed on 96-well plates and read in a FLUOstar Omega microplate reader BMG LabTech) at the wavelength specified by the supplier.

GALNS activity

GALNS activity was measured on plasma samples following the protocol described for in vitro validation. For tissue samples, ∼50 mg of each tissue was transferred to homogenizer tubes (OMNI International) containing Tris buffer (25 mM, 1 mM PMSF, pH 7.2), and homogenized in a Bead Mill Homogenizer (OMNI International). The homogenized tissues were used for GALNS activity as described for in vitro validation.

Mono-KS levels

Plasma, liver, and humerus samples were analyzed for KS by liquid chromatography-tandem mass spectrometry (LC-MS/MS) as described previously.¹⁷ Disaccharides were isolated by digestion with chondroitinase B, heparitinase, and keratanase II. Samples were run in a 1290 Infinity LC system with a 6460 triple quad mass spectrometer (Agilent Technologies). Tissue homogenization and disaccharide analysis were performed as described in Methodology S1.

Hematoxylin and eosin staining

Liver and kidneys from treated MTOL mice were stained with hematoxylin and eosin. PFA-fixed samples were xylene clearing, paraffin embedding, and microtome cutting into 6-mm-thick sections. Tissues were analyzed under light microscopy to determine liver and kidney injury.

Toluidine blue staining

Cellular vacuolation was determined in 4% PFA preserved samples through toluidine blue staining in the knee joint to evaluate cell vacuolation as described previously.¹⁰ Briefly, the knee joint was decalcified for three weeks in a decalcification solution containing 10% EDTA in 0.1 M Tris HCl buffer (pH 7.0). The decalcification solution was changed twice per week by a fresh one. Then, samples were washed three times with 0.1 M sodium cacodylate buffer and later fixed for 2 h in 1% osmium tetroxide in 0.1 M sodium cacodylate buffer. Sample dehydration was performed by keeping samples on ascending ethanol solutions (25%–95%). An initial infiltration was done by incubating samples into 1 part 100% anhydrous ethanol:1 part propylene oxide for 1 h, followed by immersion in 100% for one additional hour. Samples were then embedded in Spur low viscosity resin (EMS) and polymerized at 60°C for 2 days. Finally, embedded blocks were sectioned to obtain 0.5-μL-thick sections with a Leica UC7 ultramicrotome instrument (Leica Biosystems) and then stained using toluidine blue (EMS). Images were acquired using an Olympus BX51 light microscopy and processed with ImageJ software version 1.53t (NIH).⁷⁶

Anti-GALNS antibodies

To evaluate the anti-GALNS antibody response upon CRISPR-nCas9 treatment, plasma samples were used for ELISA experiments, according to the protocol reported by Sawamoto et al.¹⁷ Briefly, 100 μL 2 μg/mL elosulfase alfa (Vimizin, Biomarin Pharmaceutical Inc.) in 15 mM Na₂CO₃, 35 mM NaHCO₃, and 0.02% NaN₃, pH 9.6 were seeded on 96-well plates (Falcon) and incubate overnight at 4°C in a wet chamber. The next day, the wells were washed three times with 250 μL Tris-buffered saline (TBS)-T (10 mM Tris pH 7.5, 150 mM NaCl, 0.05% Tween 20) and wells were blocked by adding 200 μL 3% BSA in 1× PBS pH 7.2. The plate was then incubated for 1 h at room temperature, and later washes were done with 250 μL TBS-T. Diluted plasma samples (1:100 in TBS-T) were added in duplicate and incubated for 2.5 h at 37°C.

The plate was washed four times with 250 μL TBS-T, and 100 μL peroxidase-conjugated goat anti-mouse immunoglobulin G (HRP-IgG 1:1000 in TBS-T) was included. An additional incubation at RT was done for 1 h, and the plate was washed three times with 250 μL TBS-T and twice with 250 μL TBS. Finally, 100 μL peroxidase substrate (ABTS, Invitrogen, Thermo Fisher Scientific) was added and incubated for 30 min at RT. The reaction was stopped by adding 100 μL 1% SDS, and the plate was then read at 410 nm in a FLUOstar Omega microplate reader (BMG LabTech). Data were analyzed as optical density at 410 nm. Plasma samples from MKC mice under ERT with elosulfase alfa (Vimizin) were included as a positive control.

Anti-nCas9 antibodies

To evaluate the anti-nCas9 antibody response after CRISPR-nCas9 treatment, a protocol was standardized according to procedures reported by Chew et al.,⁷⁷ Simhadri et al.,⁶⁶ Charlesworth et al.,⁷⁸ and Sawamoto et al.¹⁷ Briefly, A 96-well plate (Falcon) was covered with 100 μg/well purified nCas9 protein (CAS9-D10A, Millipore) in 15 mM Na₂CO₃, 35 mM NaHCO₃, and 0.02% NaN₃, pH 9.6. After overnight incubation at 4°C, three washes were done using TBS-T, and then the plate was blocked using 3% BSA in 1× PBS pH 7.2 for 1 h at RT. Three washes were performed with TBS-T and then incubated with 100 μL mouse plasma (1:100 in TBS-T) or primary antibody dilutions (sc-517386) in duplicate. The following steps were performed as detailed for anti-GALNS antibodies. Anti-nCas9 antibodies were reported as ng/mL after comparing the OD obtained for each sample against a standard curve (Figure S7).

Vector copy number assays

Liver and testes were used to determine the presence of Donor ROSA26:GALNS and nCas9 after long-term assessment in MTOL mice using digital PCR in a QuantStudio 3D digital real-time PCR system (Thermo Fisher Scientific). The Tfrc gene was used as a mouse genome reference and detected through a TaqMan probe (4458367), while an EGFP TaqMan probe (Mr00660654) targeting the donor template and a customized TaqMan probe for nCas9 (57085_CDPRK2P).

Statistical analysis

Experimental data were analyzed using GraphPad Prism version 8.0.0 for Mac (GraphPad SoftwareUSA) and the statistical software R (version 4.3.0; R Core Team, R Foundation for Statistical Computing). The data are presented as mean ± standard error (SE). Shapiro-Wilk and Levene's tests were used to evaluate the normal distribution and homoscedasticity. Mean comparison between groups was made with Student’s t-test, Mann-Whitney U test, or ANOVA, according to the normality test findings; p < 0.05 was considered statistically significant.