Self-eating in skeletal development

Abstract

Macroautophagy (a.k.a. autophagy) is a cellular process aimed at the recycling of proteins and organelles that is achieved when autophagosomes fuse with lysosomes. Accordingly, lysosomal dysfunctions are often associated with impaired autophagy. We demonstrated that inactivation of the sulfatase modifying factor 1 gene (Sumf1), a gene mutated in multiple sulfatase deficiency (MSD), causes glycosaminoglycans (GAGs) to accumulate in lysosomes, which in turn disrupts autophagy. We utilized a murine model of MSD to study how impairment of this process affects chondrocyte viability and thus skeletal development.

Mucopolysaccharidoses (MPS) are genetic disorders belonging to the family of lysosomal storage disorders (LSD).¹ They are caused by defective activity of lysosomal enzymes needed for the degradation of cellular macromolecules.¹ As a consequence of this deficiency, undigested substrates accumulate in lysosomes, ultimately resulting in cellular dysfunction and death.² Multiple Sulfatase Deficiency (MSD) is a rare disease caused by mutations in SUMF1, the activator of all sulfatases. Sulfatases are a class of hydrolases that remove sulfate groups from a plethora of molecules including GAG, the major component of proteoglycans.³ One of the major consequences of the lack of sulfatase activities is the accumulation of multiple substrates, notably GAGs, in lysosomes.⁴ We previously demonstrated through the analysis of Sumf1-deficient mice that the accumulation of GAGs impairs autophagic cargo delivery to lysosomes.⁵ In neurons, this results in a reduced protein turnover favoring the formation of toxic aggregates and neurodegeneration.⁵ These data indicate that GAGs accumulation leads to global lysosomal dysfunction, and suggest an impairment of macroautophagy (hereafter termed autophagy) as a possible cause of neurodegeneration in MSD.⁶ This prompted us to investigate whether the impairment of autophagy was causing additional phenotypic features known to occur in Sumf1^−/− mice or in MSD patients. In particular, Sumf1^−/− mice display a remarkable skeletal dysplasia, characterized by severe shortening of the axial and appendicular skeletal elements, suggesting that the accumulation of GAGs severely affects skeletal development.⁴

Bone formation mainly takes place via endochondral ossification (EO).⁷ It starts during embryonic development and proceeds to the end of puberty. This process is achieved by the coordinated activity of chondrocytes and osteoblasts.⁷ Chondrocytes regulate bone length⁷ through their proliferation and differentiation. They produce an abundant extracellular matrix, mainly formed of two components including the collagens and the proteoglycans (composed of a protein core decorated by GAGs) that protect the collagen fibers.⁸ The chondrocytes together with the extracellular matrix form a cartilagineous structure named the growth plate.⁸ The observation that the growth plate is an avascular and hypoxic structure⁹ raised questions about metabolic pathways used by chondrocytes to fulfill their functions. Since autophagy is a pathway allowing cells to survive in adverse metabolic conditions,¹⁰ such as nutrient starvation or oxygen shortage,¹¹ we asked whether its impairment may eventually account for some of the abnormalities observed in Sumf1^−/− chondrocytes.

Analysis of Sumf1^−/− growth plates revealed decreased chondrocytes cellularity compared to wild-type mice starting from embryonic day 16.5 (E16.5), whereas a decreased proliferation due to abnormal fibroblast growth factor signaling was observed only postnatally. This evidence suggested the presence of a cell-death mechanism in Sumf1^−/− chondrocytes during embryonic development. Moreover, when we rescued the proliferation defect by removing one allele of fgf18 (encoding a fibroblast growth factor) in Sumf1^−/− mice we did not fully normalize chondrocyte number. Thus we investigated defective autophagy as a possible mechanism of cell death in Sumf1^−/− chondrocytes.

Electron microscopy (EM) and western blot analysis of LC3-II, the main molecular marker of autophagic vesicles (AVs),¹² showed the presence of AVs in wild-type growth plate chondrocytes. To firmly establish the presence of autophagosomes in chondrocytes we analyzed growth plate sections of mice expressing the LC3 protein fused to GFP¹³ and observed GFP-positive vesicles principally in the cytoplasm of proliferating and prehypertrophic chondrocytes. Altogether, these data indicate that autophagy is a constitutive process in chondrocytes during EO, as previously hypothesized.¹⁴

When we analyzed Sumf1^−/− chondrocytes by EM we observed a severe lysosomal vacuolization at E16.5, presumably as a result of the block in GAGs desulfation. Accordingly, Sumf1^−/− chondrocytes displayed increased AV numbers compared to wild-type chondrocytes as determined by the LC3-II amount, EM analysis and GFP visualization. These data fully support the hypothesis that the accumulation of GAGs hampers the digestion of AVs by lysosome and leads to the accumulation of AVs.⁵

Subsequently, we performed a series of in vitro experiments in wild-type and Sumf1^−/− chondrocytes to determine the cellular consequences of defective autophagy in these cells. As a control, we stimulated wild-type chondrocytes with bafilomycin A₁ (BAF), a specific inhibitor Landes of the vacuolar type H(+)-ATPase present on lysosomal membranes.¹⁵ This inhibition affects lysosomal enzyme activity and results in the accumulation of AVs.¹⁵ This mechanism of action is comparable, albeit of stronger intensity, to the inhibitory effect exerted by GAGs accumulation in Sumf1^−/− lysosomes.

Autophagy allows energy production and cell survival in the absence of serum and nutrients.¹⁶ Accordingly, treatment of wild-type chondrocytes with BAF (24 h) decreased the cellular content of ATP almost three-fold, supporting the notion that autophagy allows cell metabolism to take place in chondrocytes. In Sumf1^−/− chondrocytes, the ATP level was already lower than in wild-type, but BAF addition decreased it to the same extent as wild-type + BAF, suggesting that BAF treatment may act on the same pathway as Sumf1, but in a stronger fashion. Moreover, when we measured cellular viability of wild-type, wild-type + BAF, wild-type + 3-methyladenine (an inhibitor of autophagosomes formation¹⁷) and Sumf1^−/− chondrocytes in serum- and glucose-free medium we observed that wild-type chondrocytes survive longer than both Sumf1^−/− and wild-type chondrocytes treated with drugs. Interestingly, the same experiment in standard medium (10% FBS-high glucose DMEM) gave similar results (Fig. 1), but at later time points, further supporting a constitutive role of autophagy in chondrocyte metabolism.

Figure 1.

Wild-type (WT) and Sumf1^−/− (KO) chondrocytes were cultured in the indicated conditions for 24 h and 96 h. WT chondrocytes were also treated with bafilomycin A₁ (baf) or 3-methyladenine (3MA). Cell viability was measured using CellTiter 96 Aqueous (Promega). Student's test (*) p < 0.05; (**) p < 0.01.

Together these in vitro and in vivo observations support the concept that autophagy is important for chondrocyte metabolism during endochondral ossification, and also indicate that its impairment may contribute to the development of skeletal abnormalities, such as those observed in MSD.