Development and characterization of a model of mucopolysaccharidosis type IVA for evaluating therapies targeting bone disease

TL;DR

A new mouse model for MPSIVA was developed to study and test therapies for skeletal disorders caused by this rare genetic disease.

Contribution

The creation of a Galns−/− mouse model that accurately replicates key metabolic and skeletal features of MPSIVA.

Findings

The Galns−/− mice accumulate CS and KS in urine, plasma, and tissues.

Mice exhibit skeletal abnormalities and increased inflammatory and oxidative markers.

The model supports preclinical testing of therapies like gene therapy for skeletal dysplasia.

Abstract

Mucopolysaccharidosis type IVA (MPSIVA) is a lysosomal storage disease (LSD) caused by deficiency of N-acetylgalactosamine-6-sulfate sulfatase (GALNS), which causes the accumulation of keratan sulphate (KS) and chondroitin sulphate (CS). Patients with MPSIVA typically present with severe skeletal and joint disorders, which are not addressed by conventional therapies. Currently, no animal model accurately replicates the human disease, hindering the development of novel therapeutic interventions. To overcome this limitation, we established, by CRISPR-Cas9 technology, a Galns−/− mouse model that expresses a non-functional enzyme and accumulates CS and KS in the urine, plasma and distinct tissues, and glycosaminoglycans in the spleen. The mice exhibit shortened long bones, trabecular bone alterations and skeletal abnormalities in the growth plate. Additionally, we observed increased levels of inflammatory and oxidative markers in visceral organs and plasma. Our newly developed model of MPSIVA demonstrates clear and quantifiable signs of skeletal alterations, providing novel means of assessment of the safety and efficacy of innovative therapies, including hematopoietic stem and progenitor cell gene therapy, which has recently been shown to provide a beneficial effect on skeletal alterations in Hurler syndrome.

Summary: We developed a novel mouse model of mucopolysaccharidosis type IVA, showing clear and measurable signs of metabolic and skeletal alterations, to support the preclinical advancement of innovative therapies for skeletal dysplasia in patients.

Figures

Genetic and metabolic characterization of Galns−/− strains generated by CRISPR-Cas9 technology. (A) Graphic representation of CRISPR-Cas9-based strategy used to induce Galns gene deletion. Two gRNAs targeting intron 1 (in yellow) and exon 2 (in blue) of the murine Galns gene were used and reported in the targeted genomic region. (B) Representative images of immunofluorescence analysis of mesenchymal stromal cells (MSCs) isolated from the bone marrow of a pool of three age- and sex-matched knockout (KO)7 and wild-type (WT) mice. We evaluated the colocalization of LAMP1 (green) with GALNS enzyme (red). Nuclei are stained in blue with Hoechst. (C) Western blot analysis GALNS expression in liver biopsies from representative WT and KO mice for each strain. GALNS was detected in all KO strains. Calnexin (CLNX) was used as sample normalizer. The WT lane was derived from the same blot as the KO samples, and the line indicates the removal of an incorrectly loaded lane. (D,E) GALNS activity measured in protein extracts from liver biopsies (n=6; D) and peripheral blood mononuclear cells (PBMNCs) (n=11; E) isolated from 200 μl blood from WT and KO mice at 8 weeks of age. Each dot represents a mouse. Filled dots, male mice; empty dots, female mice. Data are represented as median±interquartile range.

Histopathological evaluation of the knee joint and surrounding structures. (A) Representative images of Hematoxylin and Eosin (H&E)-stained sections of mouse growth plate (left column) and meniscus (right column) used for the histopathological evaluation of skeletal alterations by certified pathologists. Samples from six male and six female MPSIVA and WT mice were analyzed at each time point. Vacuolation (black arrows) in the growth plate of Galns−/− mice and inflammatory cells infiltrates (red arrow) in the synovia of KO7 mice were evident in skeletal biopsies. (B,C) Percentage of damage evaluated in the growth plate, articular cartilage, synovia and meniscus for each of the KO groups at 5 months of age (n≥3). (D) Percentage of skeletal alterations evaluated in WT and KO7 mice [selected mouse model of mucopolysaccharidosis type IVA (MPSIVA)] at 8 months of age. The degree of skeletal alterations was scored and summarized in the graphs. NAD, no abnormality detected.

Body weight and bone size characterization in KO6 and KO7 mice. (A) Body weight monitoring of KO6, KO7 and WT mice of both sexes (left). A trend for reduced body weight was observed in KO7 compared to WT mice at 5 months of age (right). Body weights are represented as median weight±interquartile range. P-values were determined by two-way ANOVA in the left graph, and two-tailed Mann–Whitney test with Bonferroni correction was employed in the right graph. (B-E) Lengths of femur (B), tibia (C), humerus (D) and forelimb (E) measured in males (filled dots) and females (empty dots) of KO6 and KO7 strains compared to WT littermates at 3 and 5 months of age. KO7 male mice display significantly shorter tibiae than those of WT mice at all the time points analyzed. Reduced femur length was observed in KO7 mice compared to that in WT mice at 5 months of age, whereas shorter humeri were observed in KO7 mice at the earliest time point. (F,G) Cranium diameter (F) and cheekbone area (G) were also determined. A reduced cranium diameter was observed in KO7 compared to WT mice at 3 months of age. In B-G, data are represented as median of single values±interquartile range. P-values were determined by two-tailed Mann–Whitney test (*P<0.05; P<0.01). In all graphs, each dot represents a mouse. For all analyses, the sample size was ≥10. Filled dots, male mice; empty dots, female mice.

Dissecting the enhanced skeletal phenotype in KO7 mice. (A) Representative images of LysoTracker Deep Red staining in MSCs isolated from the bone marrow of a pool of three KO7 and WT mice. CellTracker Green was used as counterstaining. Nuclei are stained in blue with Hoechst. (B,C) Quantification of lysosome number (B) and LysoTracker mean intensity (C) per cell in WT (light blue) and MPSIVA (orange) MSCs. Each dot represents a cell. Results are expressed as median values±interquartile range. P-values were determined by two-tailed Mann–Whitney test (P<0.01). (D) Representative electron microscopy images of growth plate chondrocytes (GPCs) and osteoblasts from tibial bones of 2-month-old WT (n=2) and KO7 (n=2) mice showing lysosomal-like vacuoles in skeletal cells from KO7 mice. White arrows indicate the presence of storage vacuoles. Note that chondrocytes were more affected than osteoblasts. Scale bars are indicated for each image. B, bone; BM, bone marrow; ER, endoplasmic reticulum (black arrows); N, nucleus; Os, osteoid. (E) Boxplots showing the distribution of the frequency (%) of somatic variants colored by the corresponding impacts (i.e. high, moderate, low and modifier) in all KO strains and in WT controls. Five summary statistics (the median, two hinges and two whiskers) are visualized, and all 'outlying' points are plotted individually. The lower and upper hinges correspond to the first and third quartiles (the 25th and 75th percentiles). The upper whisker extends from the hinge to the largest value no further than 1.5× interquartile range from the hinge, while the lower whisker extends from the hinge to the smallest value at most 1.5× interquartile range from the hinge. Data beyond the end of the whiskers are called 'outlying' points and are plotted individually.

Substrate accumulation in MPSIVA mouse model. (A,B) Accumulation of chondroitin sulphate (CS; A) and keratan sulphate (KS; B) was evaluated by mass spectrometry analysis of urine samples from MPSIVA mice compared to WT littermates at 5 and 8 months of age. Results (µg/ml) were normalized to the amount of urinary creatinine (mg/ml) determined by enzyme-linked immunosorbent assay (ELISA) of the 24-h urine samples. At 5 months of age, MPSIVA mice showed significant accumulation of both substrates. (C) ELISA measuring the CS levels in bone samples from WT and MPSIVA mice at 5 months of age. (D,E) Levels of CS (D) and KS (E), measured by ELISA and mass spectrometry, respectively, in the plasma of MPSIVA mice and WT controls. Results are normalized to the volume (ml) of analyzed plasma. (F) Glycosaminoglycan (GAG) assay performed on the protein extracts from spleen biopsies of MPSIVA mice compared to those of WT mice at 5 and 8 months of age. (G) KS levels, determined by mass spectrometry, in livers from MPSIVA mice at 8 months of age compared to those from age-matched WT controls. Results (mg) were normalized to organ weight (g). For all analyses, the sample size was ≥6. In all graphs, each dot represents an experimental mouse. Data are represented as median of single values±interquartile range. P-values were determined by two-tailed Mann–Whitney test. (*P<0.05; P<0.01).

Inflammatory profiling of MPSIVA mice. (A) Reverse transcription quantitative PCR analysis of inflammatory genes in spleen samples of 5-month-old MPSIVA and WT littermates. Results are expressed as median fold change in comparison to WT averaged as ΔCt±interquartile range. (B,C) Expression of inflammatory markers in heart (B) and bone (C) samples from MPSIVA mice at 8 months of age compared to age- and sex-matched WT controls. Results are expressed as median fold change in comparison to WT±interquartile range. (D,E) Detection of murine IL-18 (D) and IL-1B (E) in the plasma of MPSIVA and WT controls by ELISA. (F) Quantification of RANKL in the plasma of MPSIVA mice compared to that in plasma of WT mice by Luminex assay. For all analyses, the sample size was ≥6. Results are expressed as median of single values±interquartile range. For all the experiments, each dot represents a mouse. P-values were determined by two-tailed Mann–Whitney test (*P<0.05; P<0.01).

Extended skeletal characterization of MPSIVA mice. (A) Body weight analysis of MPSIVA mice compared to age- and sex-matched WT controls at 8 months of age. MPSIVA model mice appear to maintain a trend for lower body weight compared to that of the control mice. (B) Computed tomography scan analysis of femur (left) and forelimb (right) length revealed shorter bones in MSPIVA mice than those in WT mice at 8 months of age. (C-H) Characterization of trabecular and cortical bone compartments by peripheral quantitative computed tomography analysis performed on the right tibia of MPSIVA mice at 3, 5 and 8 months of age. Trabecular (Tb) and cortical (Ct) bone parameters are reported. (I) Histomorphometry analysis of bone cellular components performed on the tibiae of MPSIVA mice and WT littermates at 4 months of age. For all analyses, the sample size was ≥6. Results are expressed as median of single values±interquartile range. Each dot represents a mouse. P-values were determined by two-tailed Mann–Whitney test (*P<0.05; P<0.01). BMD, bone mineral density; BS, bone surface; BV/TV, bone volume over tissue volume; OB, osteoblast; OC, osteoclast; OT, osteocyte; SSI, stress strain index.

Patient-related defects in MPSIVA mice. (A) Total pathway and velocity are normal during the open field test in MPSIVA and WT mice at 5 and 8 months. (B) Average grams of force of the six trials during the grip test, showing a deficit in MPSIVA mice compared to WT mice at both ages. Data are presented as median±interquartile range. P-values are determined by Mann–Whitney test (*P<0.05; P<0.01). (C) Representative images showing histopathological changes related to disease in the spinal column from sex-matched MPSIVA and WT mice at 5 months of age. Top row, H&E 10×, scale bars: 200 µm; bottom row, H&E 15×, scale bars: 100 µm. AF, annulus fibrosus; GP, growth plate; NP, nucleus pulposus. (D) Higher-magnification images of the GP and AF showing prominent chondrocyte and fibrochondrocyte swelling, accompanied by GP column disorganization. H&E 40×, scale bars: 50 µm. For all analyses, the sample size was ≥6.

Tables

Table 1.. Description of KO7-specific moderate-impact variants

Chr	Position	Ref	Alt	DP	AF	AD	AD_Ref	AD_Alt	Allele	Annotation	Impact	Gene	Gene_ID	Feature_type	Biotype	Variant	Sample	Variant_ID
chr1	81314413	C	T	69	0.548	32.37	32	37	T	missense_variant	Moderate	Nyap2	ENSMUSG00000054976.15	transcript	protein_coding	chr1-81314413-C-T	KO7 liver	chr1_81314413_C_T
chr3	96761873	A	G	52	0.625	21.31	21	31	G	missense_variant	Moderate	Pdzk1	ENSMUSG00000038298.15	transcript	protein_coding	chr3-96761873-A-G	KO7 liver	chr3_96761873_A_G
chr4	137948984	C	T	79	0.8	17.62	17	62	T	missense_variant	Moderate	Hp1bp3	ENSMUSG00000028759.14	transcript	protein_coding	chr4-137948984-C-T	KO7 liver	chr4_137948984_C_T
chr5	62906330	C	A	60	0.75	13.47	13	47	A	missense_variant	Moderate	Arap2	ENSMUSG00000037999.14	transcript	protein_coding	chr5-62906330-C-A	KO7 liver	chr5_62906330_C_A
chr7	3271211	T	C	67	0.738	18.49	18	49	C	missense_variant	Moderate	Nlrp12	ENSMUSG00000078817.5	transcript	protein_coding	chr7-3271211-T-C	KO7 liver	chr7_3271211_T_C
chr7	44642430	C	T	117	0.677	39.78	39	78	T	missense_variant	Moderate	Bcl2l12	ENSMUSG00000003190.13	transcript	protein_coding	chr7-44642430-C-T	KO7 liver	chr7_44642430_C_T
chr7	55781040	G	A	61	0.585	26.35	26	35	A	missense_variant	Moderate	Herc2	ENSMUSG00000030451.17	transcript	protein_coding	chr7-55781040-G-A	KO7 liver	chr7_55781040_G_A
chr8	123616920	C	G	92	0.982	0.92	0	92	G	missense_variant	Moderate	Ankrd11	ENSMUSG00000035569.18	transcript	protein_coding	chr8-123616920-C-G	KO7 liver	chr8_123616920_C_G
chr11	6214013	G	A	92	0.596	36.56	36	56	A	missense_variant	Moderate	Ddx56	ENSMUSG00000004393.11	transcript	protein_coding	chr11-6214013-G-A	KO7 liver	chr11_6214013_G_A
chr11	103391258	C	T	63	0.526	28.35	28	35	T	missense_variant	Moderate	Lrrc37a	ENSMUSG00000078632.5	transcript	protein_coding	chr11-103391258-C-T	KO7 liver	chr11_103391258_C_T
chr15	98750941	C	T	65	0.537	30.35	30	35	T	missense_variant	Moderate	Kmt2d	ENSMUSG00000048154.18	transcript	protein_coding	chr15-98750941-C-T	KO7 liver	chr15_98750941_C_T
chrX	71313682	AGGCTCCGGCTCC	A	56	0.811	8.48	8	48	A	disruptive_inframe_deletion	Moderate	Gabre	ENSMUSG00000031340.9	transcript	protein_coding	chrX-71313682-AGGCTCCGGCTCC-A	KO7 liver	chrX_71313682_AGGCTCCGGCTCC_A

Full text

INTRODUCTION

Lysosomal storage diseases (LSDs) include ∼70 genetic disorders, characterized by impaired lysosomal function, with a combined incidence of 1 every 5000-8000 livebirths (Schultz et al., 2011). Among these, mucopolysaccharidoses (MPSs) result from deficiencies in enzymes responsible for degrading glycosaminoglycans (GAGs), key regulators of extracellular matrix (ECM) assembly and cell signaling (Caterson and Melrose, 2018; Gaffke et al., 2021; Garantziotis and Savani, 2019; Karamanos et al., 2018; Nagpal et al., 2022). MPSs affect 1 in 25,000 livebirths, with variable incidence depending on the MPS type (I-XI) and geographical location (Celik et al., 2021; Karamanos et al., 2018; Lowry et al., 1990; Tomatsu et al., 2019).

Pathogenic variants primarily affect surface or core residues of lysosomal proteins, causing misfolding and disrupted structural integrity, which, in turn, lead to protein deficiency and substrate accumulation (Rivera-Colón et al., 2012). Although the mechanisms linking abnormal substrate storage to progressive cellular impairment are still largely unclear, increased oxidative stress and activation of inflammatory pathways have been implicated in the pathogenesis of MPSs (Abdulkhaleq et al., 2018; Simonaro et al., 2005; Yoshihara et al., 2000). Among them, mucopolysaccharidosis type IVA (MPSIVA), also known as Morquio syndrome type A [Online Mendelian Inheritance in Man (OMIM) #253000], results from variants in the N-acetylgalactosamine-6-sulfatase (GALNS) gene located on chromosome 16q24. Pathogenic variants in the lysosomal hydrolase GALNS impair the removal of sulphate groups from mucopolysaccharides, causing keratan sulphate (KS) and chondroitin sulphate (CS) to accumulate primarily in bones, cartilage, corneas and heart valves (Khan et al., 2017; Peracha et al., 2018; Tomatsu et al., 2016; Zhou et al., 2020). KS, a key component of proteoglycans, contributes to regulate tissue hydration, structure and cell-signaling interactions (Caterson and Melrose, 2018; Wong et al., 2002). CS, abundant in the articular cartilage, skin, blood vessels, tendons and ligaments (Lauder, 2009; Zhang et al., 2006), is essential for ECM organization, growth plate formation and chondrocyte differentiation (Schwartz and Domowicz, 2022). MPSIVA prevalence varies worldwide from 1 per 76,000 to 1 per 1,179,000 births (Leadley et al., 2014), with an estimated incidence of 1 in 200,000-300,000 individuals.

In MPSIVA, undegraded KS and CS accumulate in chondrocytes and ECM, disrupting cartilage and bone development (Tomatsu et al., 2014). Symptoms appear in early childhood and include severe skeletal dysplasia, short stature, organomegaly, cardiac valve defects, corneal clouding and respiratory complications (Tomatsu et al., 2011). Disease progression greatly reduces quality of life and mobility, and severe cases lead to death before the age of 30, mainly from airway obstruction and heart valve failure (Lavery and Hendriksz, 2014; Montaño et al., 2016).

MPSIVA is diagnosed by detecting KS in urine and confirmed through GALNS molecular analysis, with early detection being critical for timely intervention. Enzyme replacement therapy with recombinant GALNS (elosulfatase alpha) is currently the only approved treatment, although correction of skeletal lesions is limited, and immune reaction to the recombinant enzyme can further reduced its effectiveness on visceral organs (Taylor et al., 2019; Do Cao et al., 2016; Doherty et al., 2019; Tomatsu et al., 2015a,b). An alternative treatment is allogenic hematopoietic stem and progenitor cell transplantation (HSCT), which supplies the missing enzyme through transplanted healthy hematopoietic stem progenitor cells (HSPCs) (Yabe et al., 2016). However, the use of HSCT is limited by the need for human leukocyte antigen-matched donors and transplant-related risks, including conditioning regimen side effects and potential graft versus host disease development (Noh and Lee, 2014). Furthermore, the risk–benefit profile for this treatment is still unclear, with limited data supporting the use of HSCT for MPSIVA and a small number of cases reported (Chinen et al., 2014; Taylor et al., 2019; Wang et al., 2016; Yabe et al., 2016).

In the past two decades, ex vivo HSPC gene therapy (HSPC-GT) has emerged as a developing treatment for LSDs, such as metachromatic leukodystrophy (Biffi et al., 2013; Sessa et al., 2016) and MPSI (Hurler syndrome) (Consiglieri et al., 2024; Gentner et al., 2021). In HSPC-GT strategies for LSD, autologous cells are genetically modified to constitutively express supra-physiological levels of therapeutic enzymes and become a more effective source of functional enzyme than healthy donor cells (Consiglieri et al., 2024; Gentner et al., 2021). Preclinical and clinical data have shown superior safety and efficacy of HSPC-GT to metabolically correct neurological and skeletal defects in LSDs (Consiglieri et al., 2024; Gentner et al., 2021; Visigalli et al., 2010), supporting its potential applicability to other LSDs, including MPSIVA.

However, testing novel therapies for MPSIVA has been hindered by the lack of representative animal models, which exhibit milder disease phenotypes than those of patients with the disorder (Tomatsu et al., 2003, 2005, 2007). Recently, a rat model was also developed, showing no GALNS activity, elevated KS and skeletal abnormalities, more resemblant of patients with MPSIVA (Bertolin et al., 2021). However, constraints related to housing and the prevalence of established protocols of HSPC-GT for mouse studies (Biffi et al., 2004; Visigalli et al., 2010) led us to develop a novel mouse model of MPSIVA, deemed more suitable for testing new HSPC-GT strategies. We selected the C57BL/6J strain to knock out the Galns gene using CRISPR-Cas9 technology with two guide RNAs (gRNAs). Four Galns^−/−^ genotype [knockout (KO)] mice were generated and characterized to identify the model that best reproduces patients' disease manifestations. A 15-nucleotide deletion in Galns exon 2 was associated with the absence of a functional enzyme, accumulation of metabolites, growth plate and cartilage alterations, shorter long bones and trabecular bone anomalies. Increased expression of inflammatory markers and oxidative stress-related proteins were also observed, concordant with MPSIVA symptomatology. This study provides a valuable model of MPSIVA with quantifiable signs of skeletal, metabolic and systemic alterations for testing the safety and efficacy of novel therapies, including HSPC-GT.

RESULTS

Generation of a Galns−/− mouse model using the CRISPR-Cas9 system

We selected the C57BL/6J strain to delete a region between intron 1 and exon 2 of the Galns gene (chr8:123,304,981-123,338,202 GRCm39/mm39) using CRISPR-Cas9 technology. Two gRNAs were designed to target the genomic region included between the Galns intron 1 (mGalns-in, AGACCGGAUGGCUGCAGAA) and exon 2 (mGalns-ex2_0054F, ACAGAGGUUGAUCUCAGCU) (Fig. 1A). We focused on this region because CRISPR-Cas9-induced frameshifts in early exons are known to trigger nonsense-mediated decay (Cartegni et al., 2003), causing the elimination of mRNAs containing premature termination codons. Additionally, previous studies have identified that the C76 residue encoded by exon 2 is crucial for GALNS substrate recognition (Tomatsu et al., 2005). The Cas9-ribonucleoprotein complex was assembled using 4 μM Cas9 protein and 4 μM of each single-guide RNA and directly electroporated into C57BL/6J embryos, as detailed in the Materials and Methods section. We obtained four viable offspring with genomic deletions of different base-pair length (Fig. S1A, Dataset 1) that were named KO3, KO6, KO7 and KO9. These mice were evaluated to determine the impact of CRISPR-Cas9 editing on Galns coding DNA sequence and enzyme expression (Datasets 2 and 3). We used genomic DNA extracted from tail biopsies to amplify the genomic region containing the gRNA-targeted sequences, using the following primers: intron 1 forward, 5′-ACTCATGTAGTCAGGGAAC-3′; exon 2 reverse, 5′-ATGGCAGCCAGAAGACTC-3′. We found a 15-nucleotide deletion in exon 2 of KO7 mice (Dataset 2), leading to the expression of a non-functional enzyme characterized by a deletion of five amino acid residues (AAEGM) in the catalytic domain and a predicted molecular mass of 57.22 kDa (Dataset 3). In addition, KO7 mice presented a 70 bp deletion in intron 1 encompassing an enhancer sequence (chr8:123,332,206-123,332,275) (Table 1; Fig. S1B). In KO7, the mutated enzyme was predicted to be targeted to the lysosomal compartment by both PrediSI (score, 0.9278) and SignalP (score, 0.9997). To experimentally validate this prediction, we examined GALNS localization in murine mesenchymal stromal cells (MSCs) isolated from wild-type (WT) and KO7 mice by immunofluorescence, confirming the in silico prediction that GALNS colocalizes with the endolysosomal marker LAMP1 in both cell types (Fig. 1B).

Genetic and metabolic characterization of Galns−/− strains generated by CRISPR-Cas9 technology. (A) Graphic representation of CRISPR-Cas9-based strategy used to induce Galns gene deletion. Two gRNAs targeting intron 1 (in yellow) and exon 2 (in blue) of the murine Galns gene were used and reported in the targeted genomic region. (B) Representative images of immunofluorescence analysis of mesenchymal stromal cells (MSCs) isolated from the bone marrow of a pool of three age- and sex-matched knockout (KO)7 and wild-type (WT) mice. We evaluated the colocalization of LAMP1 (green) with GALNS enzyme (red). Nuclei are stained in blue with Hoechst. (C) Western blot analysis GALNS expression in liver biopsies from representative WT and KO mice for each strain. GALNS was detected in all KO strains. Calnexin (CLNX) was used as sample normalizer. The WT lane was derived from the same blot as the KO samples, and the line indicates the removal of an incorrectly loaded lane. (D,E) GALNS activity measured in protein extracts from liver biopsies (n=6; D) and peripheral blood mononuclear cells (PBMNCs) (n=11; E) isolated from 200 μl blood from WT and KO mice at 8 weeks of age. Each dot represents a mouse. Filled dots, male mice; empty dots, female mice. Data are represented as median±interquartile range.

In the other strains, the various genomic alterations give rise to a stop codon or, eventually, to the activation of an alternative reading frame producing an enzyme lacking the signal peptide (MAACTAAQQLLLVLSALGLLAAG) and the first 63 amino acid residues of the GALNS chain (24-76), as predicted by Expasy. As a result, the enzyme may be synthesized and localized outside the endoplasmic reticulum (ER)–Golgi–lysosome pathway owing to the absence of a functional signal peptide (Kornfeld, 1987; Meraş et al., 2022) and lose its enzymatic activity. Among these strains, KO3 exhibited the largest gene deletion, spanning from intron 1 to exon 2 (chr8:123,331,927-123,332,198). In KO9, only exon 2 was affected, with a 58 bp deletion (chr8:123,331,928-123,331,985). Notably, the exon 2 deletion includes cysteine 76, which corresponds to cysteine 79 in exon 2 of the human GALNS gene – a catalytic site conserved among all known sulfatases (Schmidt et al., 1995). KO6 was characterized by a 143 bp deletion in intron 1 (chr8:123,332,207–123,332,349) and a single T insertion in exon 2 between chr8:123,331,932 and chr8:123,331,933.

These results explained the expression of GALNS with similar molecular mass by western blotting in all KO strains performed with an antibody recognizing the GALNS C-terminus (Fig. 1C). The levels of GALNS expression were lower in KO3, KO6 and KO7 mice than those in WT controls.

Additionally, we found that the genomic deletion involved an enhancer region located within intron 1 in all KO mouse models, except for KO9, in which the 58-nucleotide deletion was limited to exon 2 (Fig. S1B, Dataset 1). This intronic deletion may contribute to the reduced expression of GALNS enzyme in association with potential poor stability of the aberrant Galns mRNA (Fig. S1C).

Importantly, in all KO mice, the mutated enzymes were not functional. Indeed, when we measured GALNS activity in protein extracts from liver biopsies and peripheral blood mononuclear cells (PBMNCs) collected from WT and KO mice at 2 months of age, we found that GALNS activity was undetectable in KO mice compared to WT mice (Fig. 1D,E). The median GALNS activity in PBMNCs was 99.94 nmol/mg/17 h (median value, 137.2 nmol/mg/17 h), while in livers it was 184.4 nmol/mg/17 h (median value, 176 nmol/ml/17 h).

Histopathological evaluation and computed tomography (CT) scanning of Galns−/− models

To select the Galns^−/−^ strain that best mimics the skeletal disease observed in patients, we first conducted a histopathological evaluation of the growth plate proximal to the knee joint, the site of endochondral ossification reportedly altered in several LSDs (Bartolomeo et al., 2017), including MPSIVA (Bertolin et al., 2021; Jiang et al., 2020). The analysis was performed at different time points (3, 5 and 8 months of age). The main pathological changes were observed at 5 months of age, a time point balancing the dynamic bone formation and accumulation of skeletal defects (Jilka, 2013). For each time point, we analyzed n=6 males and females, comparing their results to those of age- and sex-matched WT littermates. The tibial growth plate was evaluated and scored by two experienced pathologists, after Hematoxylin and Eosin (H&E) staining and RGB trichrome staining (Fig. 2A,B; Fig. S1D,E). Overall, all Galns^−/−^ mice displayed a generalized growth plate disorganization with evidence of chondrocyte swelling (Fig. 2A, black arrows, Fig. 2B; Fig. S1D,E).

Histopathological evaluation of the knee joint and surrounding structures. (A) Representative images of Hematoxylin and Eosin (H&E)-stained sections of mouse growth plate (left column) and meniscus (right column) used for the histopathological evaluation of skeletal alterations by certified pathologists. Samples from six male and six female MPSIVA and WT mice were analyzed at each time point. Vacuolation (black arrows) in the growth plate of Galns−/− mice and inflammatory cells infiltrates (red arrow) in the synovia of KO7 mice were evident in skeletal biopsies. (B,C) Percentage of damage evaluated in the growth plate, articular cartilage, synovia and meniscus for each of the KO groups at 5 months of age (n≥3). (D) Percentage of skeletal alterations evaluated in WT and KO7 mice [selected mouse model of mucopolysaccharidosis type IVA (MPSIVA)] at 8 months of age. The degree of skeletal alterations was scored and summarized in the graphs. NAD, no abnormality detected.

Additionally, we examined the intra-articular cavity including the articular cartilage, meniscus fibrocartilage and synovium. Similar pathological alterations, such as swelling of chondrocytes and fibrochondrocytes, and synovial cell vacuolation were found in the articular cavity of KO6 and KO7 mice, together with occasional inflammatory infiltrates in the meniscus region, whereas KO3 and KO9 mice showed only very mild abnormalities and no sign of inflammation (Fig. 2A, red arrow; Fig. 2B). Mineral deposits were detected in the epiphyseal region more frequently in KO mice than in WT mice (Fig. 2B). No increased occurrence of fracture or phenotypic involvement of the tendon were observed (Fig. 2C). Overall, all KO mice showed minimal to moderate signs of growth plate disorganization, with KO6 and KO7 mice showing the most evident alterations (Fig. 2A,B; Fig. S1D,E). These alterations remained evident at 8 months of age in KO7 compared to WT mice (Fig. 2D).

Based on these findings, we focused on the Galns^−/−^ KO6 and KO7 strains to identify additional signs of skeletal alterations. We monitored the overall growth of KO6 and KO7 mice in comparison to that of WT controls at 5 months of age; however, we observed only a trend for reduced body weight in KO7 mice compared to WT mice, with a median body weight of 28.2 g for WT and 23.6 g for KO7 mice at 5 months of age (Fig. 3A).

*Body weight and bone size characterization in KO6 and KO7 mice. (A) Body weight monitoring of KO6, KO7 and WT mice of both sexes (left). A trend for reduced body weight was observed in KO7 compared to WT mice at 5 months of age (right). Body weights are represented as median weight±interquartile range. P-values were determined by two-way ANOVA in the left graph, and two-tailed Mann–Whitney test with Bonferroni correction was employed in the right graph. (B-E) Lengths of femur (B), tibia (C), humerus (D) and forelimb (E) measured in males (filled dots) and females (empty dots) of KO6 and KO7 strains compared to WT littermates at 3 and 5 months of age. KO7 male mice display significantly shorter tibiae than those of WT mice at all the time points analyzed. Reduced femur length was observed in KO7 mice compared to that in WT mice at 5 months of age, whereas shorter humeri were observed in KO7 mice at the earliest time point. (F,G) Cranium diameter (F) and cheekbone area (G) were also determined. A reduced cranium diameter was observed in KO7 compared to WT mice at 3 months of age. In B-G, data are represented as median of single values±interquartile range. P-values were determined by two-tailed Mann–Whitney test (*P<0.05; *P<0.01). In all graphs, each dot represents a mouse. For all analyses, the sample size was ≥10. Filled dots, male mice; empty dots, female mice.

We then focused on differences in bone size by analyzing CT scans performed at 3 and 5 months of age. We found a significant reduction in the size of long bones (femur, tibia and humerus) only in KO7 mice compared to controls (Fig. 3B-D). A smaller cranium diameter was also measured in KO7 mice compared to that in WT littermates (Fig. 3F). No differences were detected in the forelimb length or cheekbone area (Fig. 3E,G). The results were not statistically significant for KO6 mice (Fig. 3A-G). When we considered the mice based on their sex, the alterations in long bone size appeared more pronounced in males (Fig. S2A-C). The median femur length in KO7 males was 12.94±0.47 mm (±s.d.), compared to 13.85±0.33 mm in WT mice at 3 months of age. This difference remained statistically significant at 5 months, with KO7 male femurs measuring 14.7±0.62 mm and WT femurs measuring 15.5±0.34 mm (Fig. S2A). Similarly, the median tibia length was shorter in KO7 males than in controls at 3 months (KO7, 15.6±0.44 mm; WT, 16.6±0.53 mm) (Fig. S2B). These differences in tibia length in KO7 females compared to sex-matched WT controls were significant only at 5 months (Fig. S2B). The humerus length was significantly reduced only in KO7 males compared to sex-matched WT controls at 3 months of age (KO7, 9.2±0.31 mm; WT, 10.3±0.66 mm) (Fig. S2C). No significant alterations were observed at 5 months of age, and at 3 and 5 months in females compared to sex-matched WT controls (Fig. S2C). Based on these findings, we concluded that KO7 exhibited the most pronounced skeletal phenotype among all strains.

To investigate the basis of the severe skeletal abnormalities observed in KO7, we examined cellular differences in MSCs. We detected increased LAMP1 staining in Galns^−/−^ compared with WT MSCs, suggesting that Galns^−/−^ cells compensate for inefficient cellular degradation by increasing the number of degradative organelles (Platt et al., 2012) (Fig. 1B). To specifically analyze the lysosomal compartment and its functionality, we stained WT and Galns^−/−^ MSCs with LysoTracker combined with high-content imaging (Xu et al., 2014). We found that, in WT MSCs, LysoTracker-positive lysosomes were uniformly distributed throughout the cytoplasm, whereas in MPSIVA cells they accumulated in the perinuclear region (Fig. 4A), consistent with previous reports associating perinuclear clustering of lysosomes with LSDs (Arévalo et al., 2022; Oyarzún et al., 2019). Moreover, Galns^−/−^ cells displayed a reduced number of LysoTracker-positive lysosomes and decreased mean fluorescence intensity, potentially reflecting altered acidification, which may impair lysosomal function, exacerbating the storage phenotype (Kuk et al., 2021) (Fig. 4B,C).

*Dissecting the enhanced skeletal phenotype in KO7 mice. (A) Representative images of LysoTracker Deep Red staining in MSCs isolated from the bone marrow of a pool of three KO7 and WT mice. CellTracker Green was used as counterstaining. Nuclei are stained in blue with Hoechst. (B,C) Quantification of lysosome number (B) and LysoTracker mean intensity (C) per cell in WT (light blue) and MPSIVA (orange) MSCs. Each dot represents a cell. Results are expressed as median values±interquartile range. P-values were determined by two-tailed Mann–Whitney test (*P<0.01). (D) Representative electron microscopy images of growth plate chondrocytes (GPCs) and osteoblasts from tibial bones of 2-month-old WT (n=2) and KO7 (n=2) mice showing lysosomal-like vacuoles in skeletal cells from KO7 mice. White arrows indicate the presence of storage vacuoles. Note that chondrocytes were more affected than osteoblasts. Scale bars are indicated for each image. B, bone; BM, bone marrow; ER, endoplasmic reticulum (black arrows); N, nucleus; Os, osteoid. (E) Boxplots showing the distribution of the frequency (%) of somatic variants colored by the corresponding impacts (i.e. high, moderate, low and modifier) in all KO strains and in WT controls. Five summary statistics (the median, two hinges and two whiskers) are visualized, and all 'outlying' points are plotted individually. The lower and upper hinges correspond to the first and third quartiles (the 25th and 75th percentiles). The upper whisker extends from the hinge to the largest value no further than 1.5× interquartile range from the hinge, while the lower whisker extends from the hinge to the smallest value at most 1.5× interquartile range from the hinge. Data beyond the end of the whiskers are called 'outlying' points and are plotted individually.

These findings were corroborated in vivo by analysis of the different cellular components of the tibial bones from KO7 and WT mice at 2 months of age (Fig. 4D). An increased number of enlarged degradative organelles was observed by electron microscopy in columnar and hypertrophic growth plate chondrocytes, as well as in osteoblasts. This phenotype is commonly associated with impaired lysosomal function, in which lysosomes fail to efficiently degrade their contents (Arévalo et al., 2022; Platt et al., 2012). Based on these observations, we hypothesized that the presence of the defective protein within the lysosomes exerts a negative effect on lysosome function, thereby exacerbating disease severity in KO7 mice compared with that in mice with the other mutant enzymes.

To exclude a contribution of additional genetic alterations, we performed whole-exome sequencing (WES) on liver-derived DNA samples from all KO strains to assess the presence of somatic mutations. In KO7 mice, no high-impact variants with a frequency ≥50% were identified (Fig. 4E, Table 1). Among the KO7-specific moderate-impact variants, we found missense mutations in the Nyap2 and Kmt2d genes. NYAP2 is primarily involved in actin dynamics neuronal development (Yokoyama et al., 2011), while Kmt2d encodes a histone lysine methyltransferase associated with Kabuki syndrome type 1 (Mertens et al., 2025). Mice bearing missense mutations in Kmt2d display sensory and psychomotor impairments, including hypotonia and increased general activity, as well as mild fine motor coordination dysfunction, without clear evidence of skeletal involvement (Kirsten et al., 2022; Yamamoto et al., 2019).

Collectively, these results indicated that somatic mutations are unlikely to contribute to the pronounced skeletal phenotype observed in KO7 mice.

Given the consistent skeletal changes observed in KO7 male mice through histopathological analyses and CT scans, together with clear evidence of lysosomal dysfunction underlying the disease, paralleling observations in patients with LSDs, we selected this strain as the mouse model of MPSIVA for further characterization. Specifically, we conducted a metabolic assessment along with an inflammatory profile analysis in MPSIVA mice and evaluated skeletal alterations up to 8 months of age compared to WT littermates.

Metabolic assessment of MPSIVA mice

We first measured substrate accumulation in 24-h urine samples from MPSIVA mice housed in metabolic cages. The levels of KS were determined by high-performance liquid chromatography–mass spectrometry (HPLC-MS), and CS content was quantified by capillary electrophoresis–laser-induced fluorescence and normalized to creatinine levels (mg/ml) determined by enzyme-linked immunosorbent assay (ELISA). Our findings revealed a significant accumulation of CS in the urine of MPSIVA mice at 5 months of age (Fig. 5A). We measured a median amount of CS of 42.05 µg/mg creatinine in WT mice (range, 10.41-73.40 µg/mg) and of 53.98 µg/mg creatinine in MPSIVA mice (range, 21.88-78.2 µg/mg). At 8 months, we only observed a trend for CS accumulation in the urine of MPSIVA compared to WT mice (Fig. 5A). Similar results were obtained for KS, which was more abundantly secreted in the urine of MPSIVA mice (median amount of KS, 0.3066 µg/mg creatinine; range, 0.1458-0.4161 µg/mg) compared to WT mice (median amount of KS, 0.178 µg/mg creatinine; range, 0.075-0.239 µg/mg) at 5 months (Fig. 5B). The absence of CS and KS accumulation in urine samples from aged mice aligned with previous studies showing that, in patients with MPSIVA, the levels of KS nearly normalize once ossification has occurred (Khan et al., 2017; Tomatsu et al., 2015a). Moreover, we observed that KS levels remained considerably lower than CS levels in MPSIVA mice at all the analyzed time points (Fig. 5A,B). This could be attributed to species-specific differences in cartilage composition and metabolite biosynthesis. As previously reported, mice synthesize less KS than do humans (Barry et al., 1994; Khan et al., 2017). Moreover, murine aggrecan, one of the main cartilage components, shows very few binding sites for KS compared to that of other mammals but is linked to CS (Barry et al., 1994). In contrast, human aggrecan contains both CS and KS chains attached to its core protein (Hayes and Melrose, 2020).

*Substrate accumulation in MPSIVA mouse model. (A,B) Accumulation of chondroitin sulphate (CS; A) and keratan sulphate (KS; B) was evaluated by mass spectrometry analysis of urine samples from MPSIVA mice compared to WT littermates at 5 and 8 months of age. Results (µg/ml) were normalized to the amount of urinary creatinine (mg/ml) determined by enzyme-linked immunosorbent assay (ELISA) of the 24-h urine samples. At 5 months of age, MPSIVA mice showed significant accumulation of both substrates. (C) ELISA measuring the CS levels in bone samples from WT and MPSIVA mice at 5 months of age. (D,E) Levels of CS (D) and KS (E), measured by ELISA and mass spectrometry, respectively, in the plasma of MPSIVA mice and WT controls. Results are normalized to the volume (ml) of analyzed plasma. (F) Glycosaminoglycan (GAG) assay performed on the protein extracts from spleen biopsies of MPSIVA mice compared to those of WT mice at 5 and 8 months of age. (G) KS levels, determined by mass spectrometry, in livers from MPSIVA mice at 8 months of age compared to those from age-matched WT controls. Results (mg) were normalized to organ weight (g). For all analyses, the sample size was ≥6. In all graphs, each dot represents an experimental mouse. Data are represented as median of single values±interquartile range. P-values were determined by two-tailed Mann–Whitney test. (*P<0.05; *P<0.01).

Indeed, although KS was undetectable by ELISA in bone extracts from both MPSIVA and WT mice (Fig. S2D), we observed a trend toward increased CS accumulation in the femurs of MPSIVA mice compared to those of controls (Fig. 5C).

In addition, using ELISA, we detected increased levels of CS in the plasma of MPSIVA mice compared to that in plasma of WT mice at 5 months of age (Fig. 5D), as well as elevated levels of KS in the plasma as determined by HPLC-MS (Fig. 5E). Finally, we analyzed spleen and liver samples, representative visceral organs, to assess GAG accumulation. Increased levels of GAGs were observed only in the spleens of MPSIVA mice compared to those of the control group at 5 and 8 months of age (Fig. 5F). No significant differences were found in liver samples (Fig. S3A). However, when we specifically measured KS by HPLC-MS, we found significantly higher levels of KS in the livers from MPSIVA mice than in those from WT mice (Fig. 5G). We concluded that metabolic parameters are altered in our newly developed MPSIVA mouse model and can potentially be monitored to assess the metabolic correction achieved by new therapeutic interventions for MPSIVA.

Inflammatory and proteomic profiling of MPSIVA mice

Given the recent important insights into the relationship between lysosomal GAG storage and inflammation in patients with MPSs (Çopur et al., 2024; van den Broek et al., 2021), we investigated the expression of pro-inflammatory markers in MPSIVA mice by reverse transcription quantitative PCR (RT-qPCR) analysis of various tissues and by ELISA of plasma samples (Fig. 6). We focused on inflammatory mediators previously reported to be upregulated in patients with MPS (van den Broek et al., 2021). In the spleens of MPSIVA mice, we observed significant overexpression of inflammatory genes compared to WT controls (Fig. 6A). In heart samples from MPSIVA mice, we observed a trend toward higher Il1a levels and a slight increase in Ccl4 expression compared with age- and sex-matched WT controls (Fig. 6B). Similarly, bone samples from MPSIVA mice showed a trend toward increased expression of Il1a, together with Il1b and Mmp1, compared with those from controls, highlighting an altered inflammatory state in tissues that are also primarily affected in patients (Fig. 6C). We further quantified IL-18 and IL-1B protein levels in the plasma at two different time points. At both 5 and 8 months of age, IL-18 levels were significantly elevated in MPSIVA mice compared to those in WT mice (Fig. 6D). Notably, IL-18 is also produced by osteoblasts and has been associated with abnormal bone formation when persistently expressed (Cornish et al., 2003; Nozaki et al., 2019). In contrast, at 5 months of age, IL-1B levels in the plasma did not differ between MPSIVA and WT mice (Fig. 6E). Moreover, we also quantified plasma levels of RANKL, a key regulator of bone remodeling and found significantly reduced levels in samples from MPSIVA mice compared to those from control mice (Fig. 6F). These findings suggest inflammation-driven alterations in the bone remodeling process of MPSIVA mice.

*Inflammatory profiling of MPSIVA mice. (A) Reverse transcription quantitative PCR analysis of inflammatory genes in spleen samples of 5-month-old MPSIVA and WT littermates. Results are expressed as median fold change in comparison to WT averaged as ΔCt±interquartile range. (B,C) Expression of inflammatory markers in heart (B) and bone (C) samples from MPSIVA mice at 8 months of age compared to age- and sex-matched WT controls. Results are expressed as median fold change in comparison to WT±interquartile range. (D,E) Detection of murine IL-18 (D) and IL-1B (E) in the plasma of MPSIVA and WT controls by ELISA. (F) Quantification of RANKL in the plasma of MPSIVA mice compared to that in plasma of WT mice by Luminex assay. For all analyses, the sample size was ≥6. Results are expressed as median of single values±interquartile range. For all the experiments, each dot represents a mouse. P-values were determined by two-tailed Mann–Whitney test (*P<0.05; *P<0.01).

To further explore systemic changes, we performed a full proteome analysis of plasma samples, revealing the upregulation of enzymes involved in the oxidative stress response (Fig. S4A). These results aligned with recent studies highlighting the role of oxidative damage in the pathophysiology of LSDs (Cacciapuoti et al., 2024; Donida et al., 2017; Lieberman et al., 2012; Saffari et al., 2017; Seranova et al., 2017). Together with IL18 and RANKL, these plasma proteins could potentially serve as biomarkers for assessing disease severity and evaluating the therapeutic efficacy of new treatments.

Multiparametric analysis of patient-related defects in MPSIVA mice

We extended to 8 months the evaluation of skeletal disease in MPSIVA mice compared to WT controls. We found that MPSIVA mice displayed a trend for reduced body weight up to 8 months of age, with a median body weight of 39.9 g (range, 31.7-40 g) for WT and 36.7 g (range, 27.5-42.8 g) for MPSIVA (Fig. 7A). We also analyzed the CT scans performed on MPSIVA and control mice at 8 months of age. We found that the length of the femurs was significantly reduced in MPSIVA mice compared to those in WT mice (Fig. 7B, left). Additionally, we highlighted a significant reduction in the length of the forelimbs in these mice (Fig. 7B, right). In contrast, we only observed a trend for reduced humerus length (WT median length, 11,12 mm; MPSIVA median length, 10.89 mm) and cheekbone area (WT median area, 0.33 cm^2^; MPSIVA median area, 0.31 cm^2^) (Fig. S3B). No differences in the cranium diameter and tibia length were observed (Fig. S3B). These changes at later time points suggested involvement of a different bone, possibly due to age- and bone-specific changes in GAGs levels in the bone matrix (Grzesik et al., 2002; Wang et al., 2018). Next, we investigated the bone density and structure by peripheral quantitative computed tomography (pQCT) performed on the right tibiae of MPSIVA and WT mice at three different ages (3, 5 and 8 months). At the proximal metaphysis, we observed age-dependent reduction in trabecular bone mineral density (Tb BMD) in MPSIVA and control mice, without any significant difference between them (Fig. 7C). Additionally, the measurements revealed a stable trabecular area in WT mice, whereas MPSIVA mice exhibited a continuous decrease, with a significant reduction observed at 8 months of age (Fig. 7D). The diaphysis analysis did not reveal significant differences in the cortical bone mineral density (Ct BMD) or thickness (Fig. 7E-G). However, we did observe a trend towards a reduced cortical area at the diaphysis level in MPSIVA mice compared to that in WT mice at 8 months of age (Fig. 7F). Additionally, the stress strain index assessed by pQCT indicated a decrease in MPSIVA mice compared to that in WT mice at 5 months of age (Fig. 7H). Overall, these pQCT data suggest a mild alteration in bone structure. We selected the tibiae collected from MPSIVA and WT mice at 5 months of age to perform histomorphometry evaluation of static bone parameters. We observed a downward trend in bone volume over tissue volume (BV/TV), along with a decrease trend in the number of osteoblasts and osteocytes per bone surface when comparing MPSIVA to WT mice. No differences were noted in the osteoclast count per bone surface (Fig. 7I; Fig. S3C).

*Extended skeletal characterization of MPSIVA mice. (A) Body weight analysis of MPSIVA mice compared to age- and sex-matched WT controls at 8 months of age. MPSIVA model mice appear to maintain a trend for lower body weight compared to that of the control mice. (B) Computed tomography scan analysis of femur (left) and forelimb (right) length revealed shorter bones in MSPIVA mice than those in WT mice at 8 months of age. (C-H) Characterization of trabecular and cortical bone compartments by peripheral quantitative computed tomography analysis performed on the right tibia of MPSIVA mice at 3, 5 and 8 months of age. Trabecular (Tb) and cortical (Ct) bone parameters are reported. (I) Histomorphometry analysis of bone cellular components performed on the tibiae of MPSIVA mice and WT littermates at 4 months of age. For all analyses, the sample size was ≥6. Results are expressed as median of single values±interquartile range. Each dot represents a mouse. P-values were determined by two-tailed Mann–Whitney test (*P<0.05; *P<0.01). BMD, bone mineral density; BS, bone surface; BV/TV, bone volume over tissue volume; OB, osteoblast; OC, osteoclast; OT, osteocyte; SSI, stress strain index.

To evaluate possible behavioral and motor deficits comparable to those observed in patients with MPSIVA (Chen et al., 2021; Sawamoto et al., 2020), two independent cohorts of 5- and 8-month-old MPSIVA and WT littermate adult male mice were subjected to a battery of behavioral tests to evaluate explorative behavior, motor coordination, and muscular fatigue and strength. In the open field test, locomotor activity, scored by the total pathway and velocity (Fig. 8A), was not significantly affected by age and mutation. Similarly, normal motor coordination and muscular fatigue, measured by the latency to fall off the rod, were observed in the rotarod test (Fig. S3D). Instead, we detected a significant impairment in MPSIVA mice in the averaged grams of force in the grip strength test, used to assess forepaw muscle strength (Fig. 8B).

*Patient-related defects in MPSIVA mice. (A) Total pathway and velocity are normal during the open field test in MPSIVA and WT mice at 5 and 8 months. (B) Average grams of force of the six trials during the grip test, showing a deficit in MPSIVA mice compared to WT mice at both ages. Data are presented as median±interquartile range. P-values are determined by Mann–Whitney test (*P<0.05; *P<0.01). (C) Representative images showing histopathological changes related to disease in the spinal column from sex-matched MPSIVA and WT mice at 5 months of age. Top row, H&E 10×, scale bars: 200 µm; bottom row, H&E 15×, scale bars: 100 µm. AF, annulus fibrosus; GP, growth plate; NP, nucleus pulposus. (D) Higher-magnification images of the GP and AF showing prominent chondrocyte and fibrochondrocyte swelling, accompanied by GP column disorganization. H&E 40×, scale bars: 50 µm. For all analyses, the sample size was ≥6.

To further exclude potential neurological changes, we examined the H&E staining of brain samples collected from 5-month-old mice, including both MPSIVA and WT mice. We found no morphological abnormalities or GAG accumulation in the neurons (Fig. S3E). This suggests that there is no evidence of neurological phenotype, similar to observations in patients with MPSIVA (Davison et al., 2013).

Additionally, we performed histopathological evaluation of heart, spine and eye tissues, as representative organs affected in patients with MPSIVA, collected from MPSIVA and WT mice at 5 months of age. As previously reported, we did not find significant deformity of the spine in MPSIVA mice compared with age- and sex-matched WT mice (Tomatsu et al., 2003). However, we found several histopathological alterations in the spinal vertebral bodies of MPSIVA mice, characterized by swelling and disorganization of chondrocytes of the growth plate and fibrochondrocytes of the annulus fibrosus (Fig. 8C,D). In the heart, eye and cornea, considered as target tissues of the disease, there were no significant morphological differences between WT and MPSIVA mice, and no pathological changes were observed in MPSIVA mice (Fig. S5A-C).

We concluded that our newly developed Galns^−/−^ mice display measurable skeletal abnormalities by histopathological, pQCT and CT scan analysis, although these are milder than those observed in patients. As expected in MPSIVA, no neurological alterations were found. These phenotypic findings, together with elevated inflammatory markers, oxidative stress and bone remodeling indicators, establish our Galns^−/−^ strain as a model for MPSIVA and support the development of novel therapeutic approaches.

Unfold protein response (UPR) and off-target analysis

Given that our model expresses a non-functional enzyme (Fig. 1C-E), we investigated the potential activation of the UPR due to improper folding of the mutated enzyme. This analysis aimed to determine whether the pathological phenotype observed in MPSIVA mice was linked to UPR activation. Specifically, we measured the expression of UPR-associated genes using RT-qPCR analysis. This analysis was conducted on biopsies of kidneys, in which GALNS is abundantly expressed as predicted for the NM_001193645.1 gene, from MPSIVA mice in comparison to those from WT mice. No significant differences in UPR-associated genes were observed in MPSIVA mice compared to controls. However, the relative ratio between spliced, unspliced and total X-box binding protein 1 (sXBP1, uXBP1 and tXBP1, respectively) showed a statistically significant difference between MPSIVA and WT mice at 5 months of age (Fig. S4B). This suggests that activation of the IRE1-mediated signaling pathway could result from GAG accumulation rather than from an unfolded aberrant protein, in accordance with previous studies (Martinon et al., 2010; Termeer et al., 2002).

Finally, we also excluded the presence of off-target mutations by performing Sanger sequencing on genomic regions containing putative off-targets predicted by CRISPOR software (Haeussler et al., 2016) for both gRNAs used to generate MPSIVA mice. We screened potential off-target pathogenic alterations that would target the exon with up to four mismatches, as well as intergenic mutations up to four mismatches, demonstrating the absence of genetic alterations in all genomic regions.

DISCUSSION

MPSIVA is a rare genetic disease caused by deficiency in GALNS, leading to KS and CS accumulation and resulting in severe skeletal symptoms. Affected children fail to achieve typical growth milestones, experience rapid skeletal deterioration and severe pain, and frequently require multiple surgical interventions (Chen et al., 2021). Current therapeutic strategies provide limited improvement of skeletal manifestations (Mynarek et al., 2012; Taylor et al., 2019), leaving a significant unmet clinical need for patients with MPSIVA. This need is exacerbated by poor understanding of MPSIVA pathophysiology and the lack of preclinical models fully replicating MPSIVA skeletal phenotypes. To address this, we invested in generating a novel Galns^−/−^ mouse model of MPSIVA, using CRISPR-Cas9 technology, targeting a genomic region included between intron 1 and exon 2 in C57BL/6J mice. We obtained four viable strains that expressed a non-functional GALNS enzyme and exhibited different levels of skeletal alterations at the histopathological level, with KO6 and KO7 showing more pronounced signs of growth plate disorganization, with evidence of chondrocyte swelling. However, only KO7 male males presented evident cartilage abnormalities, altered bone mineral density and reduced bone length. We excluded off-target effects, UPR stress response to the mutated enzyme or additional somatic mutations as being the cause of these defects. We provided evidence that the mutated enzyme, with deletion of five amino acidic residues (AAEGM) in the catalytic domain and a conserved signal peptide, may exert a negative effect within the lysosomes, where it could interfere with other components of the degradative machinery or the availability of additional lysosomal enzymes (Cosma et al., 2003), thereby exacerbating the disease phenotype compared with that caused by the other mutant enzymes.

Although sex differences have not been reported in patients with MPSIVA, in this study, male mice were more affected than female mice, consistent with several reports suggesting that sex-related factors influence the musculoskeletal system in mice (Ho-Pham et al., 2018; Nakayama et al., 2002; Ott, 2018; Uesaka et al., 2004; Wei et al., 1996). We also hypothesized that the discrepancy between our mouse model and patients reflects the earlier occurrence of disease onset in patients, before sex hormone production. In contrast, in mice, sexual maturity at 6-8 weeks of age coincides with the emergence of initial disease signs.

Moreover, although our Galns^−/−^ mouse model targets a genomic region similar to that of previously published studies (Tomatsu et al., 2003, 2005, 2007), it exhibits clear and multifaceted signs of skeletal disease. This may be due to the CRISPR-Cas9 protocol used to knock out Galns directly in the murine C57BL/6 background, a strain previously reported to exhibit exacerbated skeletal symptoms (Brommage and Ohlsson, 2020).

Additionally, although traditional methods based on homologous recombination in embryonic stem cells require further back-crossing of the chimeric mice with variable transmission-rate efficacy and potential attenuation of the disease phenotype, injection of Cas9-gRNAs into C57BL/6J zygotes can directly generate heterozygote animals for breeding with transmission rates often higher than in the case of chimeric mice (Liu et al., 2017).

Although Galns^−/−^ mice were generated using a different approach than in previously published models (Tomatsu et al., 2003, 2005, 2007), our findings similarly indicate that the type of causative mutation influences skeletal disease in mice, consistent with data in patients with MPSIVA (Montaño et al., 2007; Morrone et al., 2014; Sawamoto et al., 2020; Zanetti et al., 2019). Previously described Galns^−/−^ mice with a 600 bp deletion spanning intron 1 to exon 2 (Tomatsu et al., 2003), a deletion size broadly comparable to that in our KO3 and KO9 models, show clinical and pathological abnormalities in several visceral organs but no clear bone or cartilage defects (Tomatsu et al., 2003). In contrast, the Galns^−/−^ mice carrying a C76S point mutation in exon 2 exhibit clear bone pathology, although their symptoms are milder than those reported in patients with MPSIVA (Tomatsu et al., 2005). These findings align with our observations in KO6 and KO7 mice, which carry a single-nucleotide insertion and a limited 15-nucleotide deletion affecting exon 2 of the Galns gene, respectively.

We confirmed that the expression of a non-functional enzyme led to toxic metabolite accumulation in urine, plasma, liver, spleen and bone samples at 5 months, as reported in other models of MPSIVA. Except in the spleen, the levels of KS and CS tended to normalize, consistent with previous patient data showing that urine KS and GAG levels decrease with age (de Jong et al., 1989; Elliott and Gardner, 1979; Ellsworth et al., 2017; Kubaski et al., 2017; Martell et al., 2011; Piraud et al., 1993; Whitley et al., 1989). Overall, the levels of CS were higher than those of KS, possibly due to species-specific cartilage composition, as previously described (Hayes and Melrose, 2020; Caterson and Melrose, 2018; Venn and Mason, 1985), and they may explain the challenges in developing MPSIVA mouse models with severe skeletal symptoms.

Our multiparametric characterization of skeletal disease in MPSIVA showed significant cartilage and bone alterations at 5 months of age that are partially attenuated at longer time points, possibly due to the bone aging process (Brennan et al., 2014; Dutta and Sengupta, 2016; Hoffseth et al., 2021). Specifically, we observed moderate structural disorganization in the growth plate of our Galns^−/−^ mice, with hypertrophic chondrocytes, synovial vacuoles and inflammatory infiltrates. Similar characteristics have previously been reported in MPSVI mice, a model of joint and bone disease (Simonaro et al., 2001). Similarly, patients with MPSIVA show distorted bone geometric shape, collagen deposits in ECM and abnormal bone remodeling on histopathological analysis (Kecskemethy et al., 2016; McClure et al., 1986). Moreover, MPSIVA mice exhibit shorter long bones, with some age-related differences. Although the tibia and humerus are significantly shorter only at early time points, the femur length is consistently reduced across all analyzed time points. In contrast, forelimbs are shorter only at 8 months of age. This characteristic resembled the skeletal dwarfism observed in patients with age as a result of the relatively short trunk and lower limbs (Różdżyńska-Świątkowska et al., 2020; Sawamoto et al., 2020). pQCT analysis also revealed mild alterations in the bone trabecular area and a reduced stress strain index, suggesting that MPSIVA bone might be more fragile in response to mechanical stressors (Ott, 2018).

The pQCT findings were corroborated by histomorphometry analysis, showing a downward trend in BV/TV, correlating with a trend for reduced osteoblast numbers at 5 months of age, without alterations in the resorptive compartment, suggesting a prevailing role for osteoblasts in driving the observed bone phenotype. Overall, these findings reveal clear alterations at multiple skeletal levels, consistent with patients' clinical features.

Different from patients, defects in bone structure were not associated with motor alterations. However, MPSIVA mice appeared significantly weaker than WT mice, possibly due to muscular defects related to the accumulation of GAGs. Similarly, MPSI mice showed altered performance in the grip test, possibly associated with muscle atrophy (Visigalli et al, 2010; Kim et al., 2015). No neurological alterations were found, in line with the clinical findings in patients with MPSIVA (Chen et al., 2021).

Previous studies have indicated that several inflammatory pathways are significantly activated in LSDs, contributing to exacerbate the pathological phenotype in patients (DiRosario et al., 2009; Killedar et al., 2010; Simonaro et al., 2005, 2010; Tessitore et al., 2009; van den Broek et al., 2021) and LSD models (Johnson et al., 2002; Simonaro et al., 2008). In addition, lysosome storage impairs autophagosome–lysosome fusion, causing progressive accumulation of ubiquitinated proteins, dysfunctional mitochondria and altered oxidative metabolism in different LSDs (Settembre et al., 2008; Raben et al., 2008; Hargreaves et al., 2005). Based on these data, we investigated the inflammatory profile of our model, revealing upregulation of several proinflammatory markers in the splenic tissue of MPSIVA mice. Similarly, these markers were found to be upregulated in patients with LSDs (van den Broek et al., 2021). We also observed upregulation of inflammatory mediators in other tissues primarily affected in patients, including the heart and bones. Specifically, in heart samples from MPSIVA mice, we detected a trend toward higher levels of Il1a, a local mediator of inflammation (Cavalli et al., 2021) released by damaged or stressed cardiomyocytes (Lugrin et al., 2015). Similarly, bone samples from MPSIVA mice showed a trend toward increased expression of Il1a, together with Il1b and Mmp1, compared to those from WT mice. The inflammasome and IL-1 family are known regulators of bone homeostasis, and dysregulation of their activity has been linked to bone pathologies (Tseng et al., 2022). Interestingly, MMP1 also modulates bone remodeling (Kulesza et al., 2023; Shao et al., 2020), and its overexpression is associated with bone loss in arthritis and osteoporosis (Azevedo et al., 2018; Yamada et al., 2002).

Furthermore, in the plasma of MPSIVA mice, we detected persistently higher levels of IL-18, a pro-inflammatory cytokine also produced by osteoblasts and chondrocytes, which contributes to regulation of bone homeostasis (Mundy, 2007; Romas et al., 2002). These findings link sustained systemic inflammation to abnormal bone development. Accordingly, IL-18 may represent a valuable marker for assessing the inflammatory status of the MPSIVA mouse model, potentially driven by the accumulation of unmetabolized substrates, as reported in patients with LSDs (Ago et al., 2024; van den Broek et al., 2021; Wiesinger et al., 2022). Consistent with clinical observations, inflammatory markers may serve as potential disease biomarkers to monitor treatment efficacy (van den Broek et al., 2021) and could be translated to our disease model. Similarly, we found RANKL to be significantly reduced in the plasma of MPSIVA models, an indicator of altered bone homeostasis (El-Masri et al., 2024). The reduced levels of RANKL correlate with the trend for reduced osteocyte counts in bone from MPSIVA mice. We also performed proteomic analysis of plasma samples, showing significant enrichment of proteins involved in oxidative metabolism and oxidative stress in MPSIVA mice, in line with altered mitochondrial metabolisms associated with impaired autophagy in LSDs.

Therefore, in this study we have generated a novel mouse model for MPSIVA and completed a comprehensive multiparametric characterization, revealing alterations at the metabolic, inflammatory and skeletal level, although attenuated by species-specific biological limitations. In addition, we have individuated a new set of quantifiable skeletal parameters for more rigorous monitoring of skeletal disease progression. Collectively, our results provide evidence of the suitability of this newly generated mouse model for preclinical development and testing of novel therapies, including HSPC-GT, which we aim to test for safety and efficacy in the near future.

MATERIALS AND METHODS

Galns−/− mouse model generation

Galns^−/−^ mice were generated by CRISPR-Cas9 technology, using two gRNAs targeting intron 1 and exon 2 of Galns: mGalns-intron 1, AGACCGGAUGGCUGCAGAA and mGalns-exon 2, ACAGAGGUUGAUCUCAGCU. This work was conducted with the approval of the Institute and the Institutional Animal Care and Use Committee (6EEAF.199-IACUC 1222), in accordance with the guidelines for the use and care of experimental animals established by the Italian Ministry of Health. The gRNAs (4 µM each) and Cas9 protein (4 µM) were electroporated into 50 C57BL/6J embryos using a Bio-Rad Gene Pulser XCell electroporator. Two square wave pulses were applied (30 V, 3 ms pulse duration, two pulses, 100 ms interval). These embryos were then transferred into CD1 foster mothers, and the pregnancies were monitored. Genomic DNA was extracted from tail biopsies of the newborn mice using a Tissue PCR Kit (Sigma-Aldrich, R4775). The following primers (20 µM) were used to amplify the genomic region targeted by the two gRNAs [forward, 5′-ACTCATGTAGTCAGGGAAC-3′ and reverse, 5′-GAGTCTTCTGGCTGCCA-3′], following the manufacturer's instructions for the REDExtract-N-AMP PCR ReadyMix (Sigma-Aldrich). The six Galns^−/−^ variants that were obtained were further characterized in terms of metabolic and skeletal alterations, in line with the ministerial protocol 6EEAF.199 (IACUC 1222).

Protein extraction

Protein extracts were obtained through homogenization of tissue biopsies and cell lysis in RIPA Buffer (Thermo Fisher Scientific, 89900), supplemented with protease inhibitor cocktail (Thermo Fisher Scientific, 78438) at 4°C for 20 min. Samples were centrifuged for 15 min at 10,000 g at 4°C, and protein extracts were collected for quantification. Protein concentration was determined by RC DC Protein Assay (Bio-Rad, 5000111) using a bovine serum albumin (BSA) standard curve (Bio-Rad, 5000206), following the provider's protocol. A total of 20 µg of protein extracts were mixed with 4× Loading Buffer (Bio-Rad, 1610747) supplemented with β-mercaptoethanol at a dilution of 1:100. The samples were then denatured at 95°C for 5 min prior to western blot analysis.

For enzymatic activity assay, protein extracts were obtained by homogenizing tissue in 50-150 µl cold water. For PBMNC samples, cell pellets from 200 µl whole blood was treated with 1× Red Blood Lysis Buffer (Thermo Fisher Scientific, 00-4333-57) according to the provider's instructions and lysed in 50 µl cold water. Protein extracts were then sonicated for 25 s on 0.5 amplitude using a Sonoreaktor UTR200 (Hielscher) and quantified following the manufacturer's instructions for the RC DC Protein Assay (Bio-Rad, 5000111) prior to enzymatic activity assay.

GALNS enzymatic activity

5 µg tissue protein extract diluted in 10 µl of 0.2% BSA was incubated with 20 µl of 10 mM 4-methylumbelliferyl-β-D-galactoside-6-sulphate (4MU-Gal-6S) for 17 h at 37°C. 10 µl protein extract obtained from 200 µl whole blood was used to limit the effect of hemoglobin on protein quantification in PBMNCs. After the addition of 5 µl stopping buffer 1 (0.9 M Na-phosphate, pH4.3), 10 µl β-galactosidase (10 U/ml) was added to each sample and incubated for 2 h at 37°C. At the end of the reaction, 200 µl stopping buffer 2 (0.5 M carbonate, pH 10.7, 0.025% Triton X-100) were added to each sample, and enzymatic activity was measured as fluorescence emission (450/10 nm) on a Multiskan SkyHigh Microplate Spectrophotometer (Thermo Fisher Scientific). The level of enzymatic activity was calculated using the fluorescence emission based on a known amount of 4-methylumbelliferone (4-MU) standards (nmol) and protein extract amount (mg) or volume (ml).

GAG quantification

The level of GAG accumulation was determined in spleen and bone samples using Blyscan Sulfated Glycosaminoglycan colorimetric assay (Biocolor Ltd) according to the manufacturer's instructions. A portion of spleen was placed in a solution of papain (Sigma-Aldrich) and sodium phosphate buffer (200 µl papain/1 ml sodium phosphate buffer) and incubated overnight at 65°C. Bone samples were depleted of bone marrow cells by flushing and then disaggregated in 1× PBS and stainless-steel beads, using a Qiagen TissueLyserII. Equal amounts of protein extracts from MPSIVA and WT tissue samples were incubated for 30 min at room temperature while shaking at 950 rpm, with Blyscan dye reagent containing Dimethylmethylene Blue (DMMB) dye in an optimized buffer. Precipitated, dye-labeled sulphated GAGs were collected by centrifugation at 15,000 g for 10 min. Following removal of unbound dye, the remaining bound dye is released from the complex by addition of a dye dissociation reagent. Released dye was quantified spectrophotometrically at 620 nm using the Multiskan SkyHigh Microplate Spectrophotometer (Thermo Fisher Scientific). GAG amounts were determined in mouse samples by comparison against a calibration curve (µg) of purified chondroitin-4-sulfate and normalized to loaded amounts (mg) of protein extracts.

ELISA and Luminex assay of plasma

Accumulated KS/CS in the plasma were measured with a Mouse KS ELISA Kit (AssayGenie, MOEB2495) and a Mouse Chondroitin Sulphate (CS) ELISA Kit (AssayGenie, MOEB2502). 50 μl of 1:2 diluted plasma was loaded on the pre-coated 96-well micro-ELISA plate, together with a blank sample (sample diluent alone) and standard curve. The specimens were processed according to the manufacturer's instructions, and the plate was measured at 450 nm on the SkyHigh Microplate Spectrophotometer (Thermo Fisher Scientific). Concentrations (ng/ml) of KS/CS were determined using a standard curve. Mouse IL-18 and IL-1B were measured in mouse plasma using a DuoSet (R&D Systems) following the manufacturer's instructions. Optical densities were determined using a FLUOStar Omega plate reader (BMG Labtech). The levels of RANKL in mouse plasma were determined by Mouse Luminex Discovery Assay (LXSAMSM, R&D Systems) using a Bio-Plex 200 Reader.

CT scan analysis

Total-body CT scan was performed using an X-RAD225Cx, SmART Precision X-Ray equipped with a flat panel detector. Mice were maintained under gaseous anesthesia (2-3% isoflurane and 0.5-1 l/min oxygen) during the entire imaging workflow. Correct positioning of mice was determined by an initial ‘scout scan’. Cone beam CT images were acquired using the following parameters: tube voltage, 60 kVp; current, 5 mA; voxel size, 0.1 mm^3^. Obtained images were used to measure the lengths of long bones (femurs, tibia, humerus, forelimb), mandibular area, and cranium diameter between the joints of the squamosal body to the zygomatic process of the squamosal body by 3D Slicer Software (5.0.3 r30893/7ea0f43).

pQCT and histomorphometry analysis

Tibiae from Galns^−/−^ mice were analyzed using a Stratec Research SA+ pQCT scanner (Stratec Medizintechnik GmbH, Pforzheim) with a voxel size of 70 µm and a scan speed of 3 mm/s. To orientate the long axes of the bones parallel to the image planes, excised bone specimens were fixed with manufacturer-made plastic holders. Correct longitudinal positioning was determined by an initial scout scan. The scans were performed at the proximal metaphysis and at the diaphysis of the tibiae and were analyzed with pQCT software 6.00B using contour mode 2 and peel mode 2, with a threshold of 500 mg/cm^3^ for the calculation of trabecular and total bone parameters and a threshold of 710 mg/cm^3^ for cortical bone parameters.

For histomorphometry analyses, tibiae were fixed in 4% paraformaldehyde (PFA) and decalcified in 0.5 M EDTA 4 at pH 7.4 overnight. After a 24-h incubation in 1× PBS solution containing 20% sucrose and 2% PVP-40, the tibiae were embedded in a 1× PBS solution containing 8% porcine skin gelatin, 20% sucrose and 2% PVP-40. Samples were cryo-sectioned into 5 µm sagittal sections and stained with Toluidine Blue to quantify BV/TV, osteoblast number and osteocyte number.

TRAP staining was used to quantify the osteoclasts (Merck). All images were acquired at 20× magnification on a Nikon Eclipse 50i and quantified with Fiji software. All analyses were performed in the secondary spongiosa of the bones. Employed methods were in line the American Society for Bone and Mineral Research Histomorphometry Nomenclature Committee standards.

Histopathological evaluation

The left hind limb, brain, heart, spine and eyes of our models were collected post-mortem and fixed in 10% neutral buffered formalin. After 24 h, the bone samples were transferred to a 4% EDTA solution for decalcification, which lasted for 2 weeks. For histopathological examination, all samples were embedded in paraffin wax, cut into 3 µm sections, mounted on glass slides, and stained with H&E and RGB. Histopathological bone lesions were graded on a scale of 1 to 5 as minimal (1), mild (2), moderate (3), marked (4) or severe (5); minimal referred to the least extent discernible and severe to the greatest extent possible. Histopathological analyses were performed by experienced pathologists of the SR-TIGET GLP Facility. Relevant pictures were acquired from examined sections.

Electron microscopy

Hindlimb long bones were fixed in 1% glutaraldehyde in 0.2 M HEPES buffer (pH 7.4) for 24 h at 4°C followed by demineralization in 15% EDTA (pH 7.4) for 7 days at 4°C. Bone epiphyses were post-fixed as described in Polishchuk et al. (2019), dehydrated, embedded in epoxy resin and polymerized at 60°C for 72 h. Thin 60-nm sections were cut on a Leica EM UC7 microtome. Electron microscopy images were acquired from thin sections using a Talos L120C electron microscope (Thermo Fisher Scientific).

Immunofluorescence analysis and LysoTracker assay

MSCs were isolated from the bone marrow of MPSIVA and WT age- and sex-matched mice, according to the manufacturer’s instructions (STEMCELL Technologies, 05513). MSCs were expanded in vitro using a mouse MesenCult™ Expansion Kit (STEMCELL Technologies, 05513) for five passages before being plated for immunofluorescence analysis and LysoTracker assay. For immunofluorescence analysis, cells were plated at a concentration of 0.2×10^5^/cm^2^, fixed in 4% PFA and permeabilized in PBS, 10% fetal bovine serum (FBS) and 0.1% Triton X-100 for 10 min at room temperature. After blocking (40 min in PBS+10% FBS and 1 h in goat serum), cells were stained overnight at 4°C with the following antibodies: rat anti-LAMP1 (1:1000, Santa Cruz Biotechnology, sc-19992) and mouse anti-GALNS (1:500, Santa Cruz Biotechnology, sc-390713). The following day, cells were washed five times in PBS and incubated with Alexa Fluor dye-conjugated secondary antibodies (Invitrogen, goat anti-mouse 546 and goat anti-rat 488). After five washing in PBS, nuclei were stained with Hoechst (1:10,000, Invitrogen, H3570), and coverslips were mounted using Fluoromount-G Mounting Medium (Invitrogen) for image acquisition. Images were acquired using a Leica TCS SP5 Laser Scanning Confocal and analyzed by Fiji. For LysoTracker assay, MSCs isolated from three distinct MPSIVA and WT mice were plated in a 96-well plate (Greiner Bio-One, 655090). The following day, cells were incubated with 50 nM LysoTracker^TM^ Deep Red (Invitrogen, L12492) for 2 h at 37°C. CellTracker^TM^ Green CMFDA (1:10,000, Invitrogen, C7025) and Hoechst (1:10,000, Invitrogen, H3570) were added 20 and 10 min before stopping the LysoTracker incubation, respectively. Images were acquired using an ImageXpress Micro Confocal High Content Imaging System (Molecular Devices) equipped with 4.2 Mpixel sCMOS Camera at 16-bit depth. Cells were maintained at 37°C and 5% CO_2_, and humidity was controlled throughout the entire acquisition. Image acquisition was performed as follows: 144 sites (image resolution 2048×2048 pixels) were acquired for each well using a Plan Apo Lambda 40×/0.95 NA dry objective with a 60-µm pinhole spinning disk inserted. The DAPI filter (excitation, 377/54 nm; emission, 447/60 nm) with 150 ms exposure time was used to image Hoechst; the FITC filter (excitation, 475/34 nm; emission, 536/40 nm) with 400 ms exposure time was used for CellTracker^TM^ Green CMFDA; and the Cy5 filter (excitation, 631/28 nm; emission, 692/40 nm) with 500 ms exposure time was used for LysoTracker^TM^ Deep Red, all at 100% illumination power. During acquisition, a z-stack of 4.5 µm thickness (0.5 µm step size) was collected, and a 2D maximum-projection image was saved.

Images were quantified using Custom Module Editor (CME) in the IN Carta Image Analysis Software. Nuclei stained with Hoechst and cytoplasm stained with CellTracker Green CMFDA were first identified using the Cell Scoring application module. The newly generated nucleus and cytoplasm masks for each cell were then combined with the Boolean operator ‘OR’ to reconstruct the mask of each entire cell. All cell masks touching the edge of the image were excluded from the analysis. To enhance lysosomes from the LysoTracker Deep Red channel, a Top-Hat filter with a 3-pixel size and a circular filter shape was applied, followed by a Gaussian filter with sigma equal to 1. The Boolean operator ‘AND’ between the lysosome masks and the cell masks were then used to retain only the puncta that were located inside each cell mask. The number of lysosomes was counted for each cell, and their mean intensity was quantified.

Statistical analysis

For both the in vitro and in vivo experiments, data were analyzed using non-parametric tests based on the small sample size (<30) and lack of information about sample normality. In particular, non-parametric one-way ANOVA with Kruskal–Wallis test was applied for multiple comparisons to a control group; non-parametric Mann–Whitney test was employed for comparison between two groups, with Bonferroni correction in the case of multiple comparisons. P<0.05 was considered statistically significant.

Supplementary Materials and Methods

Further information on experimental methods can be found in the Supplementary Materials and Methods.

Supplementary Material

10.1242/dmm.052540_sup1Supplementary information