The autophagic pathway is a critical player in the pathogenesis of Pompe disease and other lysosomal storage diseases, recommending it as a potential site for therapy. This additional therapy is needed because even at very high dosages of the drug the current ERT has a variable effect in skeletal muscle. The rationale for the suppression of autophagy in Pompe skeletal muscle is two-fold: first, to prevent the accumulation of a disruptive autophagic buildup in myofibers and second, to reduce the glycogen load.
Muscle-specific suppression of autophagy in Pompe mice reduced glycogen levels in muscle, but the degree of this reduction was significantly different in the MLCcre:Atg7F/F
strains (with higher glycogen levels in HSAcre:Atg5F/F
mice). It is not clear why these two autophagy-deficient strains have different amounts of glycogen. Assuming that autophagy is involved in glycogen trafficking to the lysosomes it appears counterintuitive that muscles from the HSAcre:Atg5F/F
in which LC3-II was not present19
accumulate more glycogen than muscles from the MLCcre:Atg7F/F
in which LC3-II is still detectable. One possible explanation is that, in addition to their role in the formation of autophagosomes,34–38
Atg5 and Atg7 may fulfill other functions in skeletal muscle; for example, it has been recently shown in tumor cells that post-translational cleavage of Atg5 by calpain produces a truncated protein with pro-apoptotic function.20
Another possibility is that the lack of Atg5 in muscles from HSAcre:Atg5F/F
stimulates microautophagy or an Atg5/Atg7-independent alternative autophagy pathway. This latter, nonconventional autophagy has been recently discovered,39
and it may transport glycogen to the lysosomes. In the MLCcre:Atg7F/F
, autophagy is not completely suppressed and therefore, the alternative pathway may not be triggered. This issue, however, is beyond the scope of this paper.
Despite the difference in glycogen levels, both autophagy-deficient Pompe strains responded remarkably well to ERT, suggesting that the removal of autophagic buildup is an important factor that permits this therapy to be so effective. The successful clearance of lysosomal glycogen with rhGAA reported here is unprecedented—it has never been observed in Pompe mice treated with the recombinant enzyme alone.4,40–43
The finding in the GAA−/−
mice of an increased phosphorylation (inactivation) of glycogen synthase, the enzyme that controls bulk glycogen synthesis, suggests a feedback mechanism used by muscle cells to control the level of cytoplasmic glycogen. The major kinase that phosphorylates GS and reduces its activity is, as indicated by its name, glycogen synthase kinase-3 (GSK-3).44
The activity of the kinase itself is inhibited by phosphorylation. Glycogen depletion has been observed in skeletal muscle of GSK-3β transgenic mice as well as in muscle cells expressing constitutively active GSK-3β.45,46
In addition to its role in glycogen metabolism, GSK-3β has been implicated in a large number of pathways and in numerous disorders47
including lysosomal storage diseases.48,49
Autophagy has recently joined the long list of pathways affected by the status of GSK-3β. This kinase is a positive regulator of autophagy, as shown by the suppression of autophagy in GSK-3β−/−
mouse embryonic fibroblasts (MEFs) as well as in COS-7 cells treated with a GSK-3β inhibitor.50
A direct correlation between the GSK-3β activity and the induction of autophagy has also been shown in cadmium-treated mouse kidney mesangial cells.51
We hypothesize that excess glycogen accumulation in GAA−/−
muscle triggers a mechanism by which cytoplasmic glycogen synthesis is suppressed. A relatively slow rate of glycogen accumulation in fast GAA−/−
muscle attests to this hypothesis.40,52
The suppression of glycogen synthesis by activation of GSK-3β may have an unintended consequence in GAA−/−
muscle by contributing to the induction of autophagy. This suggests that direct modulation of GS rather than modulation through GSK-3β might be a better therapeutic approach because it can reduce the glycogen load without activation of GSK-3β and thus without induction of autophagy. In fact, the beneficial effect of suppression of glycogen synthesis in skeletal muscle has been recently demonstrated in Pompe mice.53
Thus, the two different therapeutic approaches, suppression of autophagy or suppression of glycogen synthesis, may have a very similar outcome—lessening the glycogen load and reducing autophagic accumulation.
The potential benefit of combining ERT with suppression of either glycogen synthesis or autophagy depends upon the balance between alleviating the symptoms of Pompe disease by reducing glycogen accumulation and risking the creation of new underlying conditions—a deficiency of glycogen synthase or a deficiency of autophagy in skeletal muscle. The risk of creating glycogen synthase deficiency is discussed in Douillard-Guilloux et al.53
As for the suppression of autophagy, we and others did not observe any gross phenotypical abnormalities in muscle-specific, autophagy-deficient, wild-type mice.28,54
However, the impairment of autophagy has been shown to lead to the accumulation of dysfunctional mitochondria and oxidative stress,54
as well as to muscle atrophy and an age-dependent decrease in force in gastrocnemius muscle.28
The relatively mild negative effects of inactivation of autophagy that we observe in normal skeletal muscle pale by comparison to the benefits this approach provides in Pompe skeletal muscle. The experiments reported here, we believe, are the first successful attempt to use suppression of autophagy to return cells towards a near normal state.
Bringing the observations reported here to patients with Pompe disease will require developing new approaches that permit a controllable suppression of autophagy directed to skeletal muscle. Oligonucleotides such as shRNA and morpholinos given locally or systemically and pharmaceuticals targeted to the autophagic pathway are sensible ways to consider.
The disturbances of autophagy are involved in the pathogenesis of other lysosomal storage diseases and in other families of diseases and in each, the disturbance is different;10,14
like Tolstoy's unhappy families, each is unhappy in its own way.55
Up to the present, however, the approaches to therapy—primarily in neurodegenerative diseases56
and drug-resistant malignancies57
—have relied on a limited number of chemicals, established drugs and genes. The promising results so far point to the need for new, more precise and more practical methods of autophagy modulation.