Technical AdvanceBone biologyStem cells Open Access | 10.1172/jci.insight.173449

Samantha Donsante,1 Alice Pievani,2 Biagio Palmisano,1 Melissa Finamore,2 Grazia Fazio,2 Alessandro Corsi,1 Andrea Biondi,3,4 Shunji Tomatsu,5 Rocco Piazza,4,6 Marta Serafini,2,4 and Mara Riminucci1

1Department of Molecular Medicine, Sapienza University of Rome, Rome, Italy.

2Tettamanti Center, Fondazione IRCCS San Gerardo dei Tintori, Monza, Italy.

3Pediatrics, Fondazione IRCCS San Gerardo dei Tintori, Monza, Italy.

4School of Medicine and Surgery, University of Milano-Bicocca, Monza, Italy.

5Department of Biomedical Research, Nemours Children’s Health, Wilmington, Delaware, USA.

6Hematology, Fondazione IRCCS San Gerardo dei Tintori, Monza, Italy.

Address correspondence to: Mara Riminucci, Department of Molecular Medicine, Sapienza University of Rome, Viale Regina Elena 324, 00161 Rome, Italy. Phone: 39.06.4457069; Email: mara.riminucci@uniroma1.it. Or to: Marta Serafini, Tettamanti Center, Fondazione IRCCS San Gerardo dei Tintori, via Pergolesi 33, 20900 Monza, Italy. Phone: 39.039.2332232; Email: serafinim72@gmail.com.

Authorship note: SD and AP are co–first authors. MS and MR are co–senior authors.

Find articles by Donsante, S. in: JCI | PubMed | Google Scholar

1Department of Molecular Medicine, Sapienza University of Rome, Rome, Italy.

2Tettamanti Center, Fondazione IRCCS San Gerardo dei Tintori, Monza, Italy.

3Pediatrics, Fondazione IRCCS San Gerardo dei Tintori, Monza, Italy.

4School of Medicine and Surgery, University of Milano-Bicocca, Monza, Italy.

5Department of Biomedical Research, Nemours Children’s Health, Wilmington, Delaware, USA.

6Hematology, Fondazione IRCCS San Gerardo dei Tintori, Monza, Italy.

Address correspondence to: Mara Riminucci, Department of Molecular Medicine, Sapienza University of Rome, Viale Regina Elena 324, 00161 Rome, Italy. Phone: 39.06.4457069; Email: mara.riminucci@uniroma1.it. Or to: Marta Serafini, Tettamanti Center, Fondazione IRCCS San Gerardo dei Tintori, via Pergolesi 33, 20900 Monza, Italy. Phone: 39.039.2332232; Email: serafinim72@gmail.com.

Authorship note: SD and AP are co–first authors. MS and MR are co–senior authors.

Find articles by Pievani, A. in: JCI | PubMed | Google Scholar

1Department of Molecular Medicine, Sapienza University of Rome, Rome, Italy.

2Tettamanti Center, Fondazione IRCCS San Gerardo dei Tintori, Monza, Italy.

3Pediatrics, Fondazione IRCCS San Gerardo dei Tintori, Monza, Italy.

4School of Medicine and Surgery, University of Milano-Bicocca, Monza, Italy.

5Department of Biomedical Research, Nemours Children’s Health, Wilmington, Delaware, USA.

6Hematology, Fondazione IRCCS San Gerardo dei Tintori, Monza, Italy.

Address correspondence to: Mara Riminucci, Department of Molecular Medicine, Sapienza University of Rome, Viale Regina Elena 324, 00161 Rome, Italy. Phone: 39.06.4457069; Email: mara.riminucci@uniroma1.it. Or to: Marta Serafini, Tettamanti Center, Fondazione IRCCS San Gerardo dei Tintori, via Pergolesi 33, 20900 Monza, Italy. Phone: 39.039.2332232; Email: serafinim72@gmail.com.

Authorship note: SD and AP are co–first authors. MS and MR are co–senior authors.

Find articles by Palmisano, B. in: JCI | PubMed | Google Scholar |

1Department of Molecular Medicine, Sapienza University of Rome, Rome, Italy.

2Tettamanti Center, Fondazione IRCCS San Gerardo dei Tintori, Monza, Italy.

3Pediatrics, Fondazione IRCCS San Gerardo dei Tintori, Monza, Italy.

4School of Medicine and Surgery, University of Milano-Bicocca, Monza, Italy.

5Department of Biomedical Research, Nemours Children’s Health, Wilmington, Delaware, USA.

6Hematology, Fondazione IRCCS San Gerardo dei Tintori, Monza, Italy.

Address correspondence to: Mara Riminucci, Department of Molecular Medicine, Sapienza University of Rome, Viale Regina Elena 324, 00161 Rome, Italy. Phone: 39.06.4457069; Email: mara.riminucci@uniroma1.it. Or to: Marta Serafini, Tettamanti Center, Fondazione IRCCS San Gerardo dei Tintori, via Pergolesi 33, 20900 Monza, Italy. Phone: 39.039.2332232; Email: serafinim72@gmail.com.

Authorship note: SD and AP are co–first authors. MS and MR are co–senior authors.

Find articles by Finamore, M. in: JCI | PubMed | Google Scholar

1Department of Molecular Medicine, Sapienza University of Rome, Rome, Italy.

2Tettamanti Center, Fondazione IRCCS San Gerardo dei Tintori, Monza, Italy.

3Pediatrics, Fondazione IRCCS San Gerardo dei Tintori, Monza, Italy.

4School of Medicine and Surgery, University of Milano-Bicocca, Monza, Italy.

5Department of Biomedical Research, Nemours Children’s Health, Wilmington, Delaware, USA.

6Hematology, Fondazione IRCCS San Gerardo dei Tintori, Monza, Italy.

Address correspondence to: Mara Riminucci, Department of Molecular Medicine, Sapienza University of Rome, Viale Regina Elena 324, 00161 Rome, Italy. Phone: 39.06.4457069; Email: mara.riminucci@uniroma1.it. Or to: Marta Serafini, Tettamanti Center, Fondazione IRCCS San Gerardo dei Tintori, via Pergolesi 33, 20900 Monza, Italy. Phone: 39.039.2332232; Email: serafinim72@gmail.com.

Authorship note: SD and AP are co–first authors. MS and MR are co–senior authors.

Find articles by Fazio, G. in: JCI | PubMed | Google Scholar

1Department of Molecular Medicine, Sapienza University of Rome, Rome, Italy.

2Tettamanti Center, Fondazione IRCCS San Gerardo dei Tintori, Monza, Italy.

3Pediatrics, Fondazione IRCCS San Gerardo dei Tintori, Monza, Italy.

4School of Medicine and Surgery, University of Milano-Bicocca, Monza, Italy.

5Department of Biomedical Research, Nemours Children’s Health, Wilmington, Delaware, USA.

6Hematology, Fondazione IRCCS San Gerardo dei Tintori, Monza, Italy.

Address correspondence to: Mara Riminucci, Department of Molecular Medicine, Sapienza University of Rome, Viale Regina Elena 324, 00161 Rome, Italy. Phone: 39.06.4457069; Email: mara.riminucci@uniroma1.it. Or to: Marta Serafini, Tettamanti Center, Fondazione IRCCS San Gerardo dei Tintori, via Pergolesi 33, 20900 Monza, Italy. Phone: 39.039.2332232; Email: serafinim72@gmail.com.

Authorship note: SD and AP are co–first authors. MS and MR are co–senior authors.

Find articles by Corsi, A. in: JCI | PubMed | Google Scholar

1Department of Molecular Medicine, Sapienza University of Rome, Rome, Italy.

2Tettamanti Center, Fondazione IRCCS San Gerardo dei Tintori, Monza, Italy.

3Pediatrics, Fondazione IRCCS San Gerardo dei Tintori, Monza, Italy.

4School of Medicine and Surgery, University of Milano-Bicocca, Monza, Italy.

5Department of Biomedical Research, Nemours Children’s Health, Wilmington, Delaware, USA.

6Hematology, Fondazione IRCCS San Gerardo dei Tintori, Monza, Italy.

Address correspondence to: Mara Riminucci, Department of Molecular Medicine, Sapienza University of Rome, Viale Regina Elena 324, 00161 Rome, Italy. Phone: 39.06.4457069; Email: mara.riminucci@uniroma1.it. Or to: Marta Serafini, Tettamanti Center, Fondazione IRCCS San Gerardo dei Tintori, via Pergolesi 33, 20900 Monza, Italy. Phone: 39.039.2332232; Email: serafinim72@gmail.com.

Authorship note: SD and AP are co–first authors. MS and MR are co–senior authors.

Find articles by Biondi, A. in: JCI | PubMed | Google Scholar

1Department of Molecular Medicine, Sapienza University of Rome, Rome, Italy.

2Tettamanti Center, Fondazione IRCCS San Gerardo dei Tintori, Monza, Italy.

3Pediatrics, Fondazione IRCCS San Gerardo dei Tintori, Monza, Italy.

4School of Medicine and Surgery, University of Milano-Bicocca, Monza, Italy.

5Department of Biomedical Research, Nemours Children’s Health, Wilmington, Delaware, USA.

6Hematology, Fondazione IRCCS San Gerardo dei Tintori, Monza, Italy.

Address correspondence to: Mara Riminucci, Department of Molecular Medicine, Sapienza University of Rome, Viale Regina Elena 324, 00161 Rome, Italy. Phone: 39.06.4457069; Email: mara.riminucci@uniroma1.it. Or to: Marta Serafini, Tettamanti Center, Fondazione IRCCS San Gerardo dei Tintori, via Pergolesi 33, 20900 Monza, Italy. Phone: 39.039.2332232; Email: serafinim72@gmail.com.

Authorship note: SD and AP are co–first authors. MS and MR are co–senior authors.

Find articles by Tomatsu, S. in: JCI | PubMed | Google Scholar

1Department of Molecular Medicine, Sapienza University of Rome, Rome, Italy.

2Tettamanti Center, Fondazione IRCCS San Gerardo dei Tintori, Monza, Italy.

3Pediatrics, Fondazione IRCCS San Gerardo dei Tintori, Monza, Italy.

4School of Medicine and Surgery, University of Milano-Bicocca, Monza, Italy.

5Department of Biomedical Research, Nemours Children’s Health, Wilmington, Delaware, USA.

6Hematology, Fondazione IRCCS San Gerardo dei Tintori, Monza, Italy.

Address correspondence to: Mara Riminucci, Department of Molecular Medicine, Sapienza University of Rome, Viale Regina Elena 324, 00161 Rome, Italy. Phone: 39.06.4457069; Email: mara.riminucci@uniroma1.it. Or to: Marta Serafini, Tettamanti Center, Fondazione IRCCS San Gerardo dei Tintori, via Pergolesi 33, 20900 Monza, Italy. Phone: 39.039.2332232; Email: serafinim72@gmail.com.

Authorship note: SD and AP are co–first authors. MS and MR are co–senior authors.

Find articles by Piazza, R. in: JCI | PubMed | Google Scholar |

1Department of Molecular Medicine, Sapienza University of Rome, Rome, Italy.

2Tettamanti Center, Fondazione IRCCS San Gerardo dei Tintori, Monza, Italy.

3Pediatrics, Fondazione IRCCS San Gerardo dei Tintori, Monza, Italy.

4School of Medicine and Surgery, University of Milano-Bicocca, Monza, Italy.

5Department of Biomedical Research, Nemours Children’s Health, Wilmington, Delaware, USA.

6Hematology, Fondazione IRCCS San Gerardo dei Tintori, Monza, Italy.

Address correspondence to: Mara Riminucci, Department of Molecular Medicine, Sapienza University of Rome, Viale Regina Elena 324, 00161 Rome, Italy. Phone: 39.06.4457069; Email: mara.riminucci@uniroma1.it. Or to: Marta Serafini, Tettamanti Center, Fondazione IRCCS San Gerardo dei Tintori, via Pergolesi 33, 20900 Monza, Italy. Phone: 39.039.2332232; Email: serafinim72@gmail.com.

Authorship note: SD and AP are co–first authors. MS and MR are co–senior authors.

Find articles by Serafini, M. in: JCI | PubMed | Google Scholar

1Department of Molecular Medicine, Sapienza University of Rome, Rome, Italy.

2Tettamanti Center, Fondazione IRCCS San Gerardo dei Tintori, Monza, Italy.

3Pediatrics, Fondazione IRCCS San Gerardo dei Tintori, Monza, Italy.

4School of Medicine and Surgery, University of Milano-Bicocca, Monza, Italy.

5Department of Biomedical Research, Nemours Children’s Health, Wilmington, Delaware, USA.

6Hematology, Fondazione IRCCS San Gerardo dei Tintori, Monza, Italy.

Address correspondence to: Mara Riminucci, Department of Molecular Medicine, Sapienza University of Rome, Viale Regina Elena 324, 00161 Rome, Italy. Phone: 39.06.4457069; Email: mara.riminucci@uniroma1.it. Or to: Marta Serafini, Tettamanti Center, Fondazione IRCCS San Gerardo dei Tintori, via Pergolesi 33, 20900 Monza, Italy. Phone: 39.039.2332232; Email: serafinim72@gmail.com.

Authorship note: SD and AP are co–first authors. MS and MR are co–senior authors.

Find articles by Riminucci, M. in: JCI | PubMed | Google Scholar |

Authorship note: SD and AP are co–first authors. MS and MR are co–senior authors.

Published March 8, 2024 - More info

Published in Volume 9, Issue 5 on March 8, 2024
JCI Insight. 2024;9(5):e173449. https://doi.org/10.1172/jci.insight.173449.
© 2024 Donsante et al. This work is licensed under the Creative Commons Attribution 4.0 International License. To view a copy of this license, visit http://creativecommons.org/licenses/by/4.0/. Published March 8, 2024 - Version history
Received: June 26, 2023; Accepted: January 25, 2024 View PDF Abstract

Dysostosis multiplex is a major cause of morbidity in Hurler syndrome (mucopolysaccharidosis type IH [MPS IH], OMIM #607014) because currently available therapies have limited success in its prevention and reversion. Unfortunately, the elucidation of skeletal pathogenesis in MPS IH is limited by difficulties in obtaining bone specimens from pediatric patients and poor reproducibility in animal models. Thus, the application of experimental systems that can be used to dissect cellular and molecular mechanisms underlying the skeletal phenotype of MPS IH patients and to identify effective therapies is highly needed. Here, we adopted in vitro/in vivo systems based on patient-derived bone marrow stromal cells to generate cartilaginous pellets and bone rudiments. Interestingly, we observed that heparan sulphate accumulation compromised the remodeling of MPS IH cartilage into other skeletal tissues and other critical aspects of the endochondral ossification process. We also noticed that MPS IH hypertrophic cartilage was characterized by dysregulation of signaling pathways controlling cartilage hypertrophy and fate, extracellular matrix organization, and glycosaminoglycan metabolism. Our study demonstrates that the cartilaginous pellet–based system is a valuable tool to study MPS IH dysostosis and to develop new therapeutic approaches for this hard-to-treat aspect of the disease. Finally, our approach may be applied for modeling other genetic skeletal disorders.

Graphical Abstract Introduction

Mucopolysaccharidoses (MPSs) are rare, inherited lysosomal storage diseases caused by loss-of-function mutations of genes encoding glycosaminoglycan-degrading (GAG-degrading) enzymes (1, 2). MPS type I results from the lack or reduced activity of the α-L-iduronidase (IDUA) enzyme, with consequent accumulation of 2 highly abundant GAGs, heparan sulfate (HS) and dermatan sulfate (DS) (3). HS and DS storage interferes with the normal development and/or function of multiple tissues and organs, although the severity of the clinical phenotype varies according to the residual enzymatic activity. Hurler syndrome (MPS IH, OMIM #607014) is the most severe phenotype of MPS type I, reflecting absence or extremely low levels of IDUA enzymatic activity, associated with insertions/deletions, nonsense, splice variants, and missense variants in the IDUA gene (4). In MPS IH patients, symptoms appear shortly after birth and progress with a widespread organ involvement that, in the absence of treatment, may lead to early death. The clinical picture is dominated by progressive skeletal abnormalities with impaired musculoskeletal function, coarse facial features, hepatosplenomegaly, valvular heart disease, vision and hearing loss, upper airway obstruction, and delayed mental development with a maximal functional age of 2–4 years (5).

Dysostosis multiplex is a characterizing skeletal component of the MPS IH phenotype and a major cause of disease morbidity, as it is less likely to be prevented or resolved compared with other features by current therapies, consisting of enzyme replacement therapy and hematopoietic stem cell transplantation (6, 7). It develops in late infancy and childhood, resulting in short stature, facial dysmorphism, and several chest and limb bone deformities (8, 9). Experimental work on MPS IH transgenic mice and other spontaneously occurring dog and cat models supports the hypothesis that dysostosis multiplex may result from early developmental changes that become clinically apparent after birth (10, 11). In particular, many studies identified inappropriate retention of cartilage at sites of bone formation as a recurrent pathological feature of the disease (10, 12–14). This suggests that HS and DS accumulation during skeletal development and growth interferes with the replacement of cartilage anlage with bone and bone marrow (BM), a process that relies on endochondral ossification and, as recently shown, on the direct phenotypic switch of the cartilaginous tissue (15). However, pathological changes and mechanisms observed in MPS IH animal models could not be confirmed in human patients because performing bone biopsies at growing cartilaginous skeletal sites in pediatric patients is not advisable. Thus, there is an urgent need to develop relevant disease models that can be used to identify and study the pathophysiological mechanisms that underlie the skeletal phenotype and to adequately test new therapeutic strategies. In particular, suitable human-cell-based experimental systems that reproduce the different phases of cartilage formation as well as the different processes that ensure cartilage replacement with other skeletal tissues are highly required.

In this study, we used BM stromal cell (BMSC) populations, which include skeletal stem/progenitor cells (16, 17), isolated from pediatric MPS IH patients and performed in vitro and in vivo assays to investigate the effect of IDUA mutations on human cartilage features and fate. We observed that early stages of cell condensation and cartilaginous differentiation steps occurred and progressed normally in MPS IH BMSC pellet cultures. However, in the late phases of maturation, structural and molecular changes appeared and impaired the ability of MPS IH cartilage to complete its maturation in vitro and to remodel into bone and BM in vivo.

Results

Early stages of differentiation are preserved in MPS IH cartilage. BMSCs isolated from the BM aspirates of MPS IH patients and age-matched healthy donors (HDs) were grown to confluence and then incubated in high-density pellet cultures in the presence of TGF-β1–supplemented chondrogenic differentiation medium (CDM) (Figure 1A). After 3 weeks, MPS IH skeletal progenitor cells formed 3-dimensional cartilaginous structures similar to those formed by BMSCs isolated from HDs in terms of overall size (Supplemental Figure 1A; supplemental material available online with this article; https://doi.org/10.1172/jci.insight.173449DS1), morphology, and histochemical features, such as metachromatic staining with toluidine blue and alcianophilia (Figure 1B). Histological analysis revealed only focal morphological abnormalities for MPS IH pellets in their inner core, consisting of slender-shaped and/or vacuolated chondrocytes hosted in poorly defined lacunae (Figure 1B). Similar patterns of immunolocalization for SRY-box transcription factor 9 (SOX9) and the 2 major components of cartilage matrix, collagen type II α1 chain (COL2A1) and aggrecan (ACAN), were observed in the MPS IH and HD groups (Figure 1C and Supplemental Figure 1B). In addition, gene expression analysis confirmed the absence of significant differences between the 2 groups in the expression of SOX9, COL2A1, ACAN, and other cartilage genes such as VCAN (versican) and HSPG2 (perlecan) (Supplemental Figure 1C).

Figure 1

Chondrogenic differentiation of HD and MPS IH BMSCs after 3 weeks of micro-mass culture. (A) Schematic representation of the experimental model for cartilage differentiation. (B) Representative histological images of consecutive sections stained with toluidine blue and Alcian blue/PAS demonstrating similar staining for pellets generated from HD and MPS IH BMSCs. Higher-magnification images highlight the morphology of MPS IH chondrocytes, with elongated shape and indistinct lacunae. Scale bars: 25 μm. (C) Representative confocal images of pellet sections immunostained for SOX9 demonstrating a similar expression pattern in HD and MPS IH. Scale bars: 200 μm (left column) and 50 μm (right 3 columns).

MPS IH cartilaginous pellets show defective remodeling into bone in vitro. To investigate the subsequent maturation of MPS IH cartilage in vitro, pellets grown in standard CDM for 3 weeks were exposed to mineralization-inductive conditions for 2 additional weeks (Figure 2A). At the end of the incubation time, 2 zones with distinct morphology and histochemical properties were recognized in HD pellets, which showed a peripheral collar composed of a dense, toluidine blue–negative and periodic acid–Schiff–positive (PAS-positive) tissue matrix and an inner region composed of large hypertrophic chondrocytes (Figure 2B). In contrast, 5-week MPS IH pellets were almost entirely composed of hypertrophic chondrocytes, often with heterogeneous size and vacuolated cytoplasm as shown by toluidine blue and Alcian blue/PAS staining (Figure 2B). At this time, SOX9 immunoreactivity showed a different distribution in HD pellets, in which it was restricted to the inner core, compared with MPS IH samples, in which it was more abundant and found throughout the whole cross-sectional area (Figure 2C).

Figure 2

Histological and immunohistochemical analysis of 5-week HD and MPS IH pellets. (A) Schematic representation of the experimental model for mineralization of cartilage matrix. (B) Representative histological images showing the presence of the toluidine blue–negative/PAS-positive peripheral collar in HD samples (dotted lines) and its absence in MPS IH pellets. Higher-magnification bottom images show hypertrophic chondrocytes with enlarged and vacuolated cytoplasm in MPS IH compared with HD. Scale bars: 200 μm (top) and 25 μm (bottom). (C) SOX9 immunofluorescence demonstrating the differential expression pattern between the inner, proteoglycan-rich (toluidine blue–positive) matrix and the peripheral proteoglycan-poor (toluidine blue–negative, dotted lines) zone in HD pellets and the widespread staining in MPS IH pellets. Scale bars: 50 μm. All the toluidine blue images are derived from the same sections of the HD and MPS IH pellets represented in B and C. Therefore, different fields of the same pellet are shown in multiple panels.

The peripheral collar of HD pellets was highly reminiscent of bone-like matrix. Indeed, double staining with Alcian blue and Sirius red performed on sections from formalin-fixed, decalcified, paraffin-embedded (FFDPE) pellets clearly distinguished the type I collagen–enriched Sirius red–stained peripheral region from the central alcianophilic region in which hypertrophic chondrocytes were dispersed (Figure 3A). This zonal pattern was poorly delineated in MPS IH samples, in which the outer Sirius red–positive border was absent or very thin and discontinuous (Figure 3A). Furthermore, von Kossa stain on undecalcified, resin-embedded pellets highlighted peripheral calcium deposits in HD but not in MPS IH pellets, in which the fraction of the mineralized extracellular matrix (ECM) was significantly reduced (Figure 3B). Consistent with these findings, an enriched immunoreactivity for COL1A1 and other proteins typically expressed during late cartilage hypertrophy and ECM mineralization, such as alkaline phosphatase (ALPL), matrix metalloproteinase 13 (MMP13), and osteocalcin (OCN), was observed in the outer layer of HD pellets but not in MPS IH samples (Figure 3C).

Figure 3

Mineralization and ossification of 5-week HD and MPS IH pellets. (A) Representative histological images and quantification (right graph) of pellet areas stained with Alcian blue and Sirius red, respectively, showing the COL1A1-enriched (Sirius red–stained) collar in HD samples and its reduction in MPS IH samples (pellet area stained with Sirius red, mean ± SEM from 5 HDs and 4 MPS IH): HD, 23.53% ± 3.81%; MPS IH, 10.57% ± 0.84%. *P < 0.05 by unpaired, 2-sided t test. (B) Different amounts of von Kossa–stained calcium deposition in the 2 types of pellets. Histomorphometry results (right graph) show a significant reduction in the fraction of mineralized matrix in MPS IH samples compared with HDs (pellet area stained with von Kossa, mean ± SEM from 5 HDs and 9 MPS IH): HD, 20.66% ± 4.90%; MPS IH, 7.17% ± 3.40%. *P < 0.05 by unpaired, 2-sided t test. Scale bars (A and B): 200 μm (left) and 100 μm (right). (C) Immunohistochemical staining for COL1A1, ALPL, MMP13, and OCN of 5-week HD and MPS IH pellets. Scale bars: 100 μm.

Temporospatial pattern of GAG accumulation in MPS IH cartilage. Total GAG content was assessed at different time points during in vitro chondrogenic differentiation. Comparison between the 2 experimental groups demonstrated significantly higher levels of GAGs in MPS IH compared with HD pellets at 1 week of chondrogenic differentiation (Figure 4A), thus indicating that GAG accumulation in MPS IH can be seen during cell condensation and early differentiation stages.

Figure 4

Analysis of spatial distribution of heparan sulphate and total GAG content in HD and MPS IH pellets at different stages of cartilage maturation. (A–C) Confocal images showing the expression of heparan sulphate after 1 week (A), 3 weeks (B), and 5 weeks (C) of chondrogenic differentiation in HD and MPS IH samples. The whole pellets and higher magnification of both central and peripheral zone are shown. The right graphs show the total GAG content expressed as μg GAG/μg DNA in each sample, at the same time points. Each dot represents 1 pellet. Data are represented as mean ± SEM. *P < 0.05 by unpaired, 2-sided t test. Scale bars (A–C): 200 μm (left) and 50 μm (right).

At 3 and 5 weeks, GAG levels were similar in HD and MPS IH pellets (Figure 4, B and C). However, immunolocalization of HS revealed a peripheral enrichment in MPS IH samples starting from week 3 as opposed to HD samples, which showed an overall even spatial distribution of the proteoglycan in the ECM at all times (Figure 4, A–C).

Abnormal molecular profile of hypertrophic MPS IH cartilage. To explore the molecular mechanisms involved in the delayed maturation and mineralization of the cartilage matrix of MPS IH pellets, we performed RNA-seq analysis with next-generation sequencing (NGS) on 5-week pellets. This showed 122 differentially expressed genes (DEGs), of which 87 were upregulated and 35 downregulated in MPS IH compared with HD pellets.

The upregulated transcripts in MPS IH pellets included different genes involved in ECM organization and composition, such as OGN (osteoglycin, a small keratan sulfate proteoglycan, 15.7-fold higher), SDC1 (syndecan-1, a transmembrane HS proteoglycan, 4.2-fold higher), MMP7 (matrix metalloproteinase 7, a proteolytic enzyme responsible for the syndecan-1 shedding, 2.2-fold higher), ANGPTL1 (angiopoietin-like protein 1, a gene that is known to regulate angiogenesis but it is also involved in connective tissue and cartilage development, 4.9-fold higher), and FGFR2 (fibroblast growth factor receptor 2, a receptor involved in cartilage and bone development, 3.3-fold higher). In contrast, PAPPA2 (pappalysin 2, a protease that cleaves insulin-like growth factor binding protein-5 [IGFBP-5] and regulates IGF signaling, 9.8-fold lower), BIRC3 (baculoviral IAP repeat containing 3, an inhibitor of apoptosis, 24.9-fold lower), PGF (placental growth factor, a member of the vascular endothelial growth factor [VEGF] family, 5.3-fold lower), and PTGS2 (prostaglandin endoperoxide synthase 2, an enzyme involved in chondrocyte hypertrophy, 14.9-fold lower) were included among the downregulated genes (Figure 5A). A Reactome pathway enrichment analysis showed an overrepresentation of genes related to GAG degradation and ECM degradation in MPS IH–derived pellets compared with controls (Figure 5B).

Figure 5

Whole-transcriptome analysis by next-generation sequencing of 5-week HD and MPS IH pellets. (A) Volcano plot showing the differentially expressed genes (DEGs) in MPS IH samples; each dot represents 1 gene and the log2(fold change) indicates the mean expression level for each gene. All genes with Benjamini-Hochberg–adjusted P values of less than 0.10 were considered differentially expressed. (B) KEGG pathway enrichment analysis showing significantly enriched pathways for DEGs in MPS IH pellets. (C) Heatmap representation of the top genes of different gene sets. (D) GSEA of transcriptome data demonstrating an underrepresentation of genes involved in VEGFR1, HOXA5, TGF-β1, and IGF1 pathways, and an overrepresentation of genes involved in the GAG pathway. GSEA gene sets with a Benjamini-Hochberg–adjusted P value of less than 0.25 were considered to be significantly enriched. HD, n = 3; MPS IH, n = 5.

In line with these results, gene set enrichment analysis (GSEA) highlighted that several molecular pathways responsible for cartilage maturation and remodeling such as VEGFR1 (VEGF receptor 1), HOXA5 (homeobox A5), TGFB1 (transforming growth factor β1), and IGF1 (insulin growth factor 1) signaling were significantly downregulated in MPS IH compared with HD pellets (Figure 5, C and D). On the other hand, GAG pathway and genes related to ECM organization were upregulated in MPS IH compared with HD (Figure 5, C and D, and Supplemental Figure 2A).

To further validate our system, we compared our data with a published whole-transcriptome data set of human fibroblasts from patients with different MPS types (18). We initially generated an MPS-specific gene set by isolating all the genes consistently upregulated in MPS in the published cohort and we used it to carry out a GSEA using our gene data as ranked input. We found a statistically significant, positive enrichment (Benjamini-Hochberg–corrected P = 0.0069; normalized enrichment score = 1.62) between the 2, confirming that genes overexpressed in fibroblasts from different MPSs were also expressed at high levels in our data set (Supplemental Figure 2B). Subsequently, a Pearson’s correlation analysis carried out between the log2(fold change) of DEGs in the 2 data sets revealed a strong, positive correlation of our data, especially with MPS IH (Pearson’s r = 0.71) as expected, and MPS IIIC (Pearson’s r = 0.55).

MPS IH cartilage shows defective remodeling into bone and BM in vivo. To assess the fate of MPS IH cartilage in vivo, after 3 weeks of incubation in standard CDM, pellets were heterotopically transplanted into immunocompromised mice for ossicle generation (Figure 6A). Eight weeks after transplantation, hematoxylin and eosin (H&E) staining of FFDPE ossicles showed replacement of HD pellets with a thin layer of cortical-like bone and a well-structured marrow cavity hosting donor-derived adipocytes and murine hematopoiesis. In contrast, at the same time point, MPS IH ossicles were extensively composed of residual cartilage (Figure 6B). The peripheral ring of bone was not clearly defined, whereas the area of the marrow cavity was significantly smaller compared with HD and was occupied mainly by large adipocytes, with a minimal amount of murine hematopoiesis (Figure 6B and Supplemental Figure 3). Interestingly, in MPS IH ossicles, tartrate-resistant acid phosphatase–positive (TRAP-positive) osteoclasts were significantly increased compared with controls (Figure 6C). In contrast with the cortical-like bony collar observed in HD ossicles, the tissue encasing the poorly developed marrow space in MPS IH ossicles had histochemical and structural features of both bone and cartilage, as shown by double staining with Sirius red and Alcian blue (Figure 6D). Polarized light view of the same microscopic sections highlighted the absence of lamellar bone in MPS IH ossicles (Figure 6D). Moreover, von Kossa stain on undecalcified, plastic-embedded sections revealed variable levels of mineralization within the cartilaginous matrix of the MPS IH ossicles, while mineralization in the ossicles from HDs involved the cortical-like bone collar circumferentially (Figure 7). Interestingly, in MPS IH ossicles, immunolocalization of osterix (OSX) and SOX9 on adjacent tissue sections revealed cells with topographic distribution and morphological features typical of osteoblasts coexpressing both transcription factors in the nuclei (Figure 8, A and B). In the same cells, immunostaining for COL2A1 demonstrated a cytoplasmic reaction that was not observed in HD osteoblasts (Figure 8C).

Figure 6

Conversion of HD and MPS IH pellets into heterotopic ossicles in vivo. (A) Schematic representation of the experimental model for cartilage remodeling into bone and BM in vivo. (B) H&E-stained sections and marrow area histomorphometry showing the reduced marrow cavity of MPS IH samples compared with HD (Ma.Ar/TA, mean ± SD from 4 HDs and 9 MPS IH): HD, 68.24% ± 7.10%; MPS IH, 42.35% ± 19.33%. *P < 0.05 by unpaired, 2-sided t test. Scale bars: 200 μm. (C) TRAP staining and histomorphometry showing the increased number of osteoclasts per bone perimeter in MPS IH ossicles compared with HD (N.Oc/B.Pm, mean ± SD from 4 HDs and 4 MPS IH): HD 4.32 ± 0.46/mm; MPS IH, 5.88 ± 0.24/mm. ***P < 0.001 by unpaired, 2-sided t test. Scale bars: 50 μm. (D) Alcian blue– and Sirius red–stained sections analyzed by transmitted (left panels) and polarized (right panels) light microscopy demonstrating retention of cartilaginous tissue (blue stain) and an overall disorganized structure in MPS IH ossicles. Scale bars: 200 μm (top) and 100 μm (higher magnification of the boxed areas, bottom).

Figure 7

Matrix mineralization of HD and MPS IH ossicles. Von Kossa stain of undecalcified ossicles revealing the lack of a mineralized bony cortex and a variable degree of cartilage mineralization in MPS IH ossicles. Scale bars: 200 μm (top) and 100 μm (higher magnification of the boxed areas, bottom).

Figure 8

Bone-forming cells in HD and MPS IH ossicles. (A–C) Immunolocalization of OSX (A), SOX9 (B), and COL2A1 (C) on serial sections of HD and MPS IH ossicles showing the presence in the latter of cells with osteoblast topography and morphology coexpressing bone (OSX) and cartilage (SOX9 and COL2A1) markers. Scale bars (A–C): 50 μm.

Treatment with laronidase partially reverts abnormal MPS IH cartilage phenotype in vitro. To further support our results and to start to explore the utility of the pellet system in a therapeutic context, we evaluated the effect of the treatment with the recombinant human α-L-iduronidase (laronidase) on cartilage maturation and bone-like matrix deposition in MPS IH pellets. To address this point, MPS IH BMSCs were induced for cartilaginous differentiation for 5 weeks in differentiation and mineralization culture media supplemented with 5 μg/mL laronidase. As expected, laronidase-treated MPS IH pellets showed higher IDUA activity compared with untreated MPS IH pellets, with values comparable to those of HD samples (Figure 9A). The treatment with the recombinant enzyme was able to reduce the GAG content in MPS IH pellets compared with untreated samples (Figure 9B). Importantly, histological analysis revealed a thickening of the Sirius red–stained collagenic border in MPS IH–treated pellets, indicating an increased deposition of type I collagen–enriched matrix likely reflecting an enhanced osteogenic differentiation. However, calcium deposits at this zone remained poor, despite an overall increase in mineralization with respect to untreated pellets, as showed by von Kossa stain on undecalcified sections (Figure 9C).

Figure 9

MPS IH pellets treated with laronidase. (A) IDUA activity in 5-week MPS IH pellets untreated or treated with laronidase (mean ± SEM from 5 MPS IH and 5 MPS IH + rhIDUA pellets): MPS IH, 10.95 ± 4.08 nmol/mg/h; MPS IH + rhIDUA, 290.1 ± 38.72 nmol/mg/h. **P < 0.01 by paired, 2-sided t test. The red dotted line represents the HD mean. (B) Relative GAG content in 5-week MPS IH pellets treated with laronidase compared with untreated (mean ± SEM from 5 MPS IH + rhIDUA compared with 5 untreated pellets): MPS IH + rhIDUA, 0.36 ± 0.11. **P < 0.01 by paired, 2-sided t test. (C) Histological images of Alcian blue– and Sirius red–stained (upper panels) and von Kossa–stained (bottom panels) sections of MPS IH pellets from the same patient untreated or treated with laronidase. Scale bars: 100 μm.

Discussion

In this study, we used patient-derived models for reproducing and elucidating the cartilage pathology in MPS IH using skeletal progenitor cells. We applied these models to provide insight into the biological and transcriptional changes that are associated with endochondral ossification dysfunction in MPS IH patients, offering the first evidence to our knowledge that cartilaginous pellets and heterotopic ossicles, already employed in studies on hematological malignancies and infectious diseases (19, 20), could be used to reproduce the skeletal pathology features of an inherited disease.

The timely cartilage replacement with bone and BM at sites of endochondral ossification is required for normal skeletal development, growth, and shaping. Essential steps in the substitution process include apoptosis of hypertrophic chondrocytes, resorption of mineralized ECM, and homing of blood vessel–associated progenitor cells producing bone, marrow stroma, and marrow adipose tissue (16, 21). Accordingly, available studies on the mechanisms underlying the abnormal retention of cartilage observed in MPS IH endochondrally formed bone (12–14, 22–24) are mainly focused on the effects of GAG accumulation on this sequence of events (10, 14, 25). Recent evidence, however, demonstrates that endochondral ossification is a more complex phenomenon in which specific chondrocyte subsets escape death and undergo a phenotypic shift that leads to the direct remodeling of cartilage into other skeletal tissues (26–29). Therefore, we attempted to better understand the fate of MPS IH cartilage in the developing skeleton by recapitulating ex vivo and using human IDUA-mutated skeletal progenitor cells, the different processes, or specific aspects thereof, through which endochondral ossification unfolds.

We observed that condensation/differentiation of skeletal stem/progenitor cells into cartilaginous pellets as well as progression to hypertrophy of newly formed chondrocytes were not major targets of impaired IDUA activity, as also demonstrated in our previous work (30). However, subsequent processes leading to cartilage maturation, mineralization, and replacement with other skeletal tissues were greatly compromised/delayed in MPS IH. In particular, the ability of MPS IH cartilage to shift to a bone-like phenotype was reduced compared with normal cartilage. As previously reported (