|Home | About | Journals | Submit | Contact Us | Français|
Beta-amyloid (Aβ) degrading endopeptidases are thought to protect against Alzheimer disease (AD) and are potentially therapeutic. Of particular interest are endopeptidases that are blocked by thiorphan and phosphoramidon (T/P), as these inhibitors rapidly induce Aβ deposition in rodents. Neprilysin is the best known target of T/P; however neprilysin knockout results in only modest Aβ increases insufficient to induce deposition. Therefore, other endopeptidases targeted by T/P must be critical for Aß catabolism. Another candidate is the T/P sensitive membrane metallo-endopeptidase-like protein (MMEL), a close homolog of neprilysin.
The endopeptidase properties of β and γ splice forms of human MMEL were determined in HEK293T cells transduced with the human cDNAs for the two splice forms; this showed degradation of both Aβ42 and Aβ40 by hMMEL-β but not hMMEL-γ. hMMEL-β activity was found at the extracellular surface with no significant secreted activity. hMMEL-γ was not expressed at the extracellular surface. Finally, it was found that hMMEL cleaves Aβ near the α-secretase site (producing Aβ1-17 >> Aβ1-16). These data establish hMMEL as a mediator of Aβ catabolism and raises the possibility of its involvement in the etiology of Alzheimer's disease and as a target for intervention.
Previous studies have identified metalloendopeptidases as mediators of Aβ catabolism (Hersh, 2003). One key enzyme is neprilysin, a cell-surface associated endopeptidase (Howell et al., 1995, Iwata et al., 2000). It has been reported that NEP localization in the brain is reduced in AD and in aging, particularly in brain regions vulnerable to plaque formation in humans and rodents (Akiyama et al., 2001, Reilly, 2001, Yasojima et al., 2001a, Yasojima et al., 2001b, Iwata et al., 2002, Apelt et al., 2003, Caccamo et al., 2005, Maruyama et al., 2005, Wang et al., 2005). Infusion of the NEP inhibitor, thiorphan, has been shown to induce a dramatic increase in Aβ levels (~30-fold) sufficient to produce plaque deposition after only 1 month in wild-type rats and mice (Iwata et al., 2000, Dolev and Michaelson, 2004). Another NEP inhibitor, phosphoramidon, produced similar results in mice (personal communication from D. M. Michaelson). This implicates that enzymes targeted by these inhibitors are critical for maintaining normal Aβ levels in vivo.
The murine NEP homolog termed membrane metallo-endopeptidase like protein −1/2 (MMEL) has a membrane bound (mMMEL-α), and a secreted form (mMMEL-β) as a result of alternative splicing (Fig. 1) (Ikeda et al., 1999, Ghaddar et al., 2000). MMEL (also known as SEP, NL1, NEPLP, and NEP2) was shown to be able to degrade vasoactive peptides, and was determined to be sensitive to inhibition by thiorphan and phosphoramidon (Ikeda et al., 1999). The α splice-form of MMEL was shown to degrade Aβ (Shirotani et al., 2001). The human ortholog of MMEL has been identified, and was found to be expressed in the brain (Bonvouloir et al., 2001). Figure 1 compares alternate splice forms of human MMEL (β, γ, and δ) to known splice forms of the murine gene (α and β). The murine β form retains an alternate exon that contains a furin-like cleavage site (arrow) resulting in its secretion, and loss of activity (Ikeda et al., 1999, Shirotani et al., 2001). There is 77% identity between the mouse and human β homologs (Bonvouloir et al., 2001) and all three hMMEL isoforms retain the homologous furin-site containing exon. However, the human β (accession # BC101027) and δ (accession # BC101029) forms contain a large alternate exon excluded from the human γ (accession # BC101028) form, and the δ form utilizes an alternate splice acceptor site resulting in a frameshift mutation and truncation of the protein, eliminating the zinc binding motif critical for activity. Very little is known regarding the properties of hMMEL. Therefore, we set out to test the Aβ degrading activity and molecular properties of these splice forms in a HEK293T cell culture system.
Plasmids containing the human MMEL cDNAs were purchased from Open Biosystems (Huntsville, AL). To express the human β and γ isoforms of hMMEL, we constructed lentiviral gene transfer vectors expressing these cDNAs. A lentiviral vector expressing murine MMEL-β was also constructed for testing. All vectors expressed the desired transgene driven by the human cytomegalovirus promoter in a ‘third generation’ lentiviral vector (Tiscornia et al., 2006). Control vectors included: lenti-NEP (positive control) and lenti-GFP (negative control) (Marr et al., 2003). To achieve the production of Aβ in our HEK293T cell cultures, a lentiviral vector (lenti-APP) directing the expression of mutant human APP (Swedish and London mutations-K670M/N671L/V717I) and PS-1 (Δ9) was used (Singer et al., 2005).
Two sets of experiments were performed aimed at transducing HEK293T cells with lentiviral vectors expressing our endopeptidases of interest and APP. In the first set of experiments (co-infection), HEK293T cells were transduced with the lenti-APP vector and then these same cells were expanded and then co-infected with a lentiviral vector expressing the endopeptidases or GFP control. In the second set of experiments (co-culture), the lenti-APP transduced HEK293T cells were expanded and co-cultured with HEK293T cells transduced separately with the lentiviral vectors expressing the endopeptidases or GFP control. In both experimental sets, the same lenti-APP transduced cells were used across the test and control groups ensuring consistent levels of Aβ production. Our results indicate that hMMEL-β was able to degrade cell-produced Aβ42 and Aβ40 in both co-infection (Fig. 2A, B) and co-culture (Fig. 2C, D) experiments (hMMEL-γ was inactive). The catabolic activity of hMMEL-β was found to be sensitive to thiorphan and phosphoramidon similar to reports for the murine α homolog (data not shown). Murine MMEL-α has previously been shown to degrade Aβ (Shirotani et al., 2001). However, the reported activity was highly selective for Aβ40 rather than Aβ42. This has relevance to AD research as it is thought that Aβ42 is more amyloidogenic and has been implicated as the primary mediator of the disease. Our findings clearly suggest the human homolog of MMEL has similar activities against both forms of Aβ (Fig. 2).
MMEL endopeptidases have preferred substrate cleavage sites. To further characterize hMMEL, synthetic Aβ1-40 or Aβ1-42 was digested with purified recombinant hMMEL-β and the products analyzed by mass spectrometry (Fig. 3). The primary substrate cleavage for both forms of Aβ was Aβ 1-17 (hydrolyzing between Leu17 and Val18) and to a lesser extent Aβ 1-16 (hydrolysis at the α-secretase cut site between Lys16 and Leu17). This preference is within a subset of sites also cut by ECE-1, a more distantly related homologue of MMEL (Eckman et al., 2001).
We next set out to determine if hMMEL-β is a secreted or cell surface associated enzyme. Western blot analysis of transfected HEK293T cells indicated that both hMMEL-β and hMMEL-γ are detected in the cell lysate (Fig. 4A). The γ isoform showed the predicted shift in electrophoretic mobility due to the absence of an alternate exon. hMMEL-β and not γ was detected in the cell supernatant. hMMEL-β, was immunodetected on the surface of transfected and unpermeabilized HEK293T cells by flow cytometry (Fig. 4B). The majority of the cell population transfected with hMMEL-β showed more intense green fluorescence producing a shift to the right in the population histogram. However, very little hMMEL-γ was detected at the cell surface as there was not a substantial shift in the transfected cell population. Aβ degrading activity was assayed in the cell culture supernatants of these transfected cells. Supernatants expressing hMMEL-β, hNEP (also cell surface associated), and GFP were mixed with Aβ-containing medium and later assayed for remaining Aβ levels. An expression vector producing a secreted-active ectodomain of hNEP (secNEP) (Marr et al., 2004) was used as a positive control in this experiment. As expected, only cell culture supernatant from secNEP transfected cells showed reduced Aβ. This indicated a lack of soluble enzyme activity for hMMEL-β and NEP (Fig. 4C). Combined with the data above (Fig. 2) these results show that hMMEL-β, in its active form, is a cell-surface associated protein with an extracellular active site, similar to NEP. However, it was also detected as a secreted higher molecular weight form that did not possess significant Aβ degrading activity (Fig. 4A). This observation is similar to what was reported for murine MMEL, as the β form (SEP-β) was shown to be secreted as an inactive higher molecular weight protein (Ikeda et al., 1999). The increased mass of the processed / secreted from was shown to be due to glycosylation. Therefore, our results differ from the reported murine data in that the human β splice form was also found to be active on the cell surface. The hMMEL-γ splice variant was not detected at the cell surface in significant amounts but was found in the cell lysate. This suggests this isoform has altered protein trafficking patterns.
MMEL expression studies have been performed in mice and rats (Carpentier et al., 2003, Facchinetti et al., 2003). It has been suggested that MMEL and NEP exhibit a complementary expression pattern in the brain. For example, both enzymes were found to be expressed in the hippocampal formation in rats. However, NEP was found primarily in the granule cell layer of the dentate gyrus while MMEL was found throughout the pyramidal cell layers (Facchinetti et al., 2003). MMEL and NEP have been reported to degrade many of the same smaller neuropeptide substrates including substance P, enkephalin, and bradykinin (Ghaddar et al., 2000, Carpentier et al., 2003). This suggests a redundancy in the functions performed by these enzymes in the brain, similar to the related enzymes thimet oligopeptidase and neurolysin (Massarelli et al., 1999, Tullai et al., 2000). Also, both enzymes have been reported to be expressed in the human brain (Akiyama et al., 2001, Bonvouloir et al., 2001) and our results demonstrate degradation of Aβ42 and Aβ40 by both. Therefore, the shared functions between MMEL and NEP could include the control of Aβ peptide levels in humans. The potential use of NEP augmentation to treat AD has been extensively explored in rodent models (Leissring et al., 2003, Marr et al., 2003, Hong et al., 2006, Hemming et al., 2007). Our data support the potential use of MMEL augmentation for the treatment of Alzheimer's disease and indicate the possibility that MMEL dysfunction could be involved in progression of neurodegeneration.
R.A.M is supported by funds from the Schweppe Foundation. M.J.G. is supported by funds from the National Institutes of Health NS39892 and RR19325. Flow cytometry was performed at the Flow Cytometry Core Facility at Rosalind Franklin University. We would like to thank Dr. Luc Desgroseillers for providing us with the mMMEL-β cDNA. We also would like to thank Arun George Paul and Soyoung Kim for technical assistance.
Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.