Non-neuronal induction of immunoproteasome subunits in an ALS model: Possible mediation by cytokines

Abstract

Protein aggregation is a pathologic hallmark of familial amyotrophic lateral sclerosis caused by mutations in the Cu, Zn superoxide dismutase gene. Although SOD1-positive aggregates can be cleared by proteasomes, aggregates have been hypothesized to interfere with proteasome activity, leading to a vicious cycle that further enhances aggregate accumulation.

To address this issue, we measured proteasome activity in transgenic mice expressing a G93A SOD1 mutation. We find that proteasome activity is induced in the spinal cord of such mice compared to controls but is not altered in uninvolved organs such as liver or spleen. This induction within spinal cord is not related to an overall increase in the total number of proteasome subunits, as evidenced by the steady expression levels of constitutive α7 and β5 subunits. In contrast, we found a marked increase of inducible beta proteasome subunits, LMP2, MECL-1 and LMP7. This induction of immunoproteasome subunits does not occur in all spinal cord cell types but appears limited to astrocytes and microglia. The induction of immunoproteasome subunits in G93A spinal cord organotypic slices treated with TNF-α and interferon-γ suggest that certain cytokines may mediate such responses in vivo.

Our results indicate that there is an overall increase in proteasome function in the spinal cords of G93A SOD1 mice that correlates with an induction of immunoproteasomes subunits and a shift toward immunoproteasome composition. These results suggest that increased, rather than decreased, proteasome function is a response of certain cell types to mutant SOD1-induced disease within spinal cord.

Introduction

Mutations in the Cu, Zn superoxide dismutase (SOD1) gene cause one form of familial amyotrophic lateral sclerosis (FALS) via a gain in function that may be related to abnormal SOD1 protein processing and aggregation (Rosen et al., 1993). Indeed, the presence of SOD1 aggregates within the CNS is a pathologic hallmark of disease both in humans dying from SOD1-related FALS and in transgenic mice expressing various SOD1 mutations (Gurney et al., 1994, Watanabe et al., 2001, Shibata et al., 1996, Bruijn et al., 1998, Wang et al., 2002a, Wong et al., 1995). The accumulation of SOD1 aggregates begins early in the lifespan of mutant SOD1 transgenic mice, clearly preceding the onset of motor dysfunction and consistent with having a primary role in the disease process (Wang et al., 2002b). Moreover, the tissue distribution of SOD1 aggregates in vivo is strikingly limited to those regions clearly affected by the disease, including brain and spinal cord, but is absent from ALS-resistant organs such as liver or kidney, despite their having overall higher mutant SOD1 expression levels (Wang et al., 2002b, Puttaparthi et al., 2003). SOD1 aggregates can readily be visualized either as cytoplasmic inclusions or as high molecular weight protein complexes (HMWPCs) on Western blots (Wang et al., 2002b, Johnston et al., 2000, Kato et al., 1997, Bruijn et al., 1997, Wang et al., 2003).

Several studies have clearly demonstrated that proteasome-mediated degradation pathways are responsible for clearing SOD1 aggregates from a variety of differing cell types including intact murine spinal cord (Puttaparthi et al., 2003, Johnston et al., 2000, Urushitani et al., 2002, Hyun et al., 2003). The 26S proteasome is composed of one 20S proteolytic complex and two axially positioned 19S (PA700) regulatory complexes (DeMartino and Slaughter, 1999). The 20S complex is itself composed of 2 copies each of 7 α and β type subunits, each encoded by distinct genes. Each β ring contains three differing proteolytic sites that differ in its specificity including a chymotrypsin-like site that cleaves after hydrophobic residues, a trypsin-like site that cleaves after basic residues and a post-glutamyl peptide hydrolase or caspase-like activity that cuts after the acidic residues glutamate and aspartate (Kisselev et al., 2003). Certain β subunits of the 20S core appear to be inducible and convey differing proteolytic function on the proteasome. Inducible β subunits such as LMP2, MECL-1 and LMP7 can replace the normal constitutive β1, β2 and β5 subunits and have recently been shown to be increased in Huntington's disease (Diaz-Hernandez et al., 2003, Fabunmi et al., 2001, Rivett and Hearn, 2004). It is unknown whether inducible proteasomes, also known as immunoproteasomes, possess a capacity to degrade disease related aggregates that is distinct from constitutive proteasomes.

Pre-formed SOD1 aggregates within mouse spinal cord can be cleared by enhancing proteasomal function, suggesting that the accumulation of SOD1 aggregates within the CNS is a reversible phenomenon potentially dependent on equilibrium between aggregate formation and clearance via proteasomes (Puttaparthi et al., 2003). In certain experimental conditions, protein aggregates themselves may directly impair proteasome function leading to a vicious cycle that favors even greater aggregate accumulation (Bence et al., 2001). Alternatively, potentially representing a cellular attempt to improve protein degradation capability, proteasome induction has been observed within certain human degenerative diseases characterized by prominent aggregates including Huntington's and inclusion body myopathies (Diaz-Hernandez et al., 2003, Ferrer et al., 2004). Therefore, we sought to determine whether proteasome function is altered within the spinal cord of mutant SOD1 transgenic mice as aggregates accumulate, and if so what is the potential mechanism of that change?