Autophagy in astrocytes

Abstract

Neurodegeneration is a prominent feature of lysosomal storage disorders (LSDs). Emerging data identify autophagy dysfunction in neurons as a major component of the phenotype. However, the autophagy pathway in the CNS has been studied predominantly in neurons, whereas in other cell types it has been largely unexplored. We studied the lysosome-autophagic pathway in astrocytes from a murine model of multiple sulfatase deficiency (MSD), a severe form of LSD. Similar to what was observed in neurons, we found that lysosomal storage in astrocytes impairs autophagosome maturation and this, in turn, has an impact upon the survival of cortical neurons and accounts for some of the neurological features found in MSD. Thus, our data indicate that lysosomal/autophagic dysfunction in astrocytes is an important component of neurodegeneration in LSDs.

LSDs comprise nearly 60 different inherited disorders, caused by a genetic defect that leads to the inability of the lysosomal system to degrade specific metabolites, resulting in abnormal storage/accumulation within the lysosome. As a consequence, many tissues and organs are affected, with early onset central nervous system (CNS) dysfunction predominating. However, the mechanism by which the lysosomal defect causes cellular and organ dysfunction is not clear. An emerging view is that lysosomal storage hampers basic cellular processes such as intracellular trafficking and autophagy, ultimately leading to cellular dysfunction.

An impairment of autophagy, which leads to a secondary accumulation of autophagy substrates such as polyubiquitinated proteins and dysfunctional mitochondria, has been reported in many LSDs. This cellular phenotype seems to be particularly severe in neurons and is associated with significant neurodegeneration. These observations suggest that LSDs can be considered as “autophagy disorders.” Interestingly, autophagy impairment is a common feature shared by LSDs and more common types of neurodegenerative diseases, such as Alzheimer, Parkinson and Huntington diseases, in which autophagy dysfunction and intraneuronal accumulation of autophagy substrates has been documented. To date, investigations have mainly focused on neurons, but the contribution of other cell types to the neurodegeneration remains largely unexplored.

Astrocytes are the most abundant cell type in the mammalian brain. They are responsible for the maintenance of brain homeostasis, and account for several critical functions such as neurotransmitter trafficking and recycling, nutrient and ion metabolism, and defense against oxidative stress. Consistent with the functional cooperation that exists between astrocytes and neurons, the impairment of astrocyte function has been recently found to contribute to neuronal dysfunction in several neurodegenerative diseases.

We investigated the contribution of astrocyte dysfunction to neurodegeneration in a severe type of LSD, multiple sulfatase deficiency (MSD). MSD is caused by mutations in the Sulfatase Modifying Factor 1 (SUMF1) gene, whose protein product is responsible for the post-translational activation of sulfatases, a family of enzymes required for the turnover and degradation of sulfated compounds mainly in lysosomes. Lack of SUMF1 causes sulfatases to be inactive and in turn leads to a progressive storage of sulfated substrates in lysosomes. We studied the consequences of loss of Sumf1 in astrocytes in vivo using different mouse models in which the gene was deleted either in the whole body (Sumf1^−/−), only in neurons and astrocytes (Sumf1^flox/flox; Nestin-Cre), or only in astrocytes (Sumf1^flox/flox; GFAP-Cre).

In all of the analyzed models we observed the presence of a severe cytosolic vacuolization in astrocytes. Electron microscopy clearly identified autophagosomes with a characteristic double membrane containing portions of cytoplasm, autolysosomes with a single limiting membrane containing partially degraded cytosolic components and membranes, and lysosomes containing clear and amorphous material, likely undigested glycosaminoglycans. Furthermore, cortical astrocytes isolated from Sumf1^−/− mice harboring a transgene that expresses a GFP-tagged LC3 protein (well-established marker for autophagosomes) show cytoplasmic accumulation of large GFP–positive autophagosomes. Inhibition of autophagy causes a build-up of cytosolic poly-ubiquitinated protein aggregates. We found that in Sumf1^−/− brain the ubiquitin positive aggregates colocalize not only with the neuronal marker RBFOX3/NeuN but also with the astrocyte-specific marker GFAP. Furthermore, the SQSTM1/p62 receptor, required for both the formation and degradation of polyubiquitinated proteins by autophagy also accumulates in the cytoplasm of Sumf1^−/− astrocytes. These observations suggest a block of the autophagy pathway in Sumf1^−/− astrocytes.

Lysosomal storage defects impair lysosome-autophagosome fusion, hence causing a block of the autophagy pathway at the level of autolysosome formation. In addition, lysosomal storage defects also cause an impaired ability of the lysosome to degrade the autophagic content, leading to cytosolic accumulation of large autolysosomes. In our study, electron microscopy data show the presence of both autophagosomes and large autolysosomes in the cytoplasm of Sumf1^−/− astrocytes supporting a model in which the autophagic block in astrocytes occurs at multiple levels. One possibility could be that the incomplete degradation of autolysosomes inhibits the lysosome reformation process through which the cell restores its pool of lysosomes. This block may lead to a lysosomal depletion that, in turn, impairs the fusion efficiency between newly generated autophagosomes and lysosomes.

We also tested whether the autophagic/lysosomal dysfunction in astrocytes contibuted to the neurological deterioration observed in MSD. At the histological level, Sumf1^flox/flox;GFAP-Cre mice present a significant and progressive decrease in the number of cortical neurons compared with control mice, suggesting that astrocytes lacking Sumf1 are unable to properly support neuronal functions. To test this hypothesis, we co-cultured wild-type cortical neurons onto a layer of cortical astrocytes isolated from either wild-type or Sumf1^−/− mice. Consistently, we found that the Sumf1^−/− astrocytes have impairments in their ability to support the differentiation and survival of the adjacent neurons.

We then extended our analysis to Purkinje cells in the cerebellum, as their degeneration was documented in several LSDs. Despite the deletion of Sumf1 in the astrocytes of the Bergmann glia, Purkinje cell survival is not affected in Sumf1^flox/flox;GFAP-Cre mice. In contrast, the deletion of Sumf1 in both astrocytes and neurons in Sumf1^flox/flox; Nestin-Cre mice, results in a massive Purkinje cell vacuolization and death, suggesting that Purkinje cell loss is a cell-autonomous process.

Neuroinflammation has been described in most types of LSDs with neurological involvement. It is still unclear whether microglial and astrocyte activation in LSDs is triggered by their intracellular storage or represents a response to neuronal damage. We demonstrated that Sumf1 deletion in both neurons and astrocytes induces astrogliosis and strong microglial activation even though Sumf1 was not deleted in microglia. In contrast, lysosomal storage defects only in astrocytes do not cause either astroglyosis or microglial activation. Thus, we concluded that lysosomal storage defects in neurons is the major cause of neuroinflammation observed in LSDs.

Finally, to address whether astrocyte dysfunction may contribute to the neurological phenotype of MSD, we performed a panel of behavioral tests in both Sumf1^flox/flox;GFAP-Cre mice and Sumf1^flox/flox;Nestin-Cre mice. Surprisingly, we observed some striking behavioral differences between the two mouse models. The deletion of the Sumf1 gene in astrocytes only caused hypoactivity and anxiety-like behavior, whereas its concomitant deletion in astrocytes and neurons induced hyperactivity and reduced motor learning ability. Interestingly, these neurological symptoms represent common features observed in different phases of disease manifestations in LSDs. Thus, we propose that some of the pathological manifestation of the LSDs can be associated with astrocyte dysfunction.

In conclusion, our data indicate a novel, non-cell autonomous mechanism of neurodegeneration in LSDs, triggered by autophagic/lysosomal dysfunction in astrocytes.

Di Malta C, Fryer JD, Settembre C, Ballabio A. Astrocyte dysfunction triggers neurodegeneration in a lysosomal storage disorder. Proc Natl Acad Sci U S A. 2012;109:E2334–42. doi: 10.1073/pnas.1209577109.