Insulin-degrading Enzyme Regulates Extracellular Levels of Amyloid β-Protein by Degradation*

      Excessive cerebral accumulation of the 42-residue amyloid β-protein (Aβ) is an early and invariant step in the pathogenesis of Alzheimer's disease. Many studies have examined the cellular production of Aβ from its membrane-bound precursor, including the role of the presenilin proteins therein, but almost nothing is known about how Aβ is degraded and cleared following its secretion. We previously screened neuronal and nonneuronal cell lines for the production of proteases capable of degrading naturally secreted Aβ under biologically relevant conditions and concentrations. The major such protease identified was a metalloprotease released particularly by a microglial cell line, BV-2. We have now purified and characterized the protease and find that it is indistinguishable from insulin-degrading enzyme (IDE), a thiol metalloendopeptidase that degrades small peptides such as insulin, glucagon, and atrial natriuretic peptide. Degradation of both endogenous and synthetic Aβ at picomolar to nanomolar concentrations was completely inhibited by the competitive IDE substrate, insulin, and by two other IDE inhibitors. Immunodepletion of conditioned medium with an IDE antibody removed its Aβ-degrading activity. IDE was present in BV-2 cytosol, as expected, but was also released into the medium by intact, healthy cells. To confirm the extracellular occurrence of IDE in vivo, we identified intact IDE in human cerebrospinal fluid of both normal and Alzheimer subjects. In addition to its ability to degrade Aβ, IDE activity was unexpectedly found be associated with a time-dependent oligomerization of synthetic Aβ at physiological levels in the conditioned media of cultured cells; this process, which may be initiated by IDE-generated proteolytic fragments of Aβ, was prevented by three different IDE inhibitors. We conclude that a principal protease capable of down-regulating the levels of secreted Aβ extracellularly is IDE.
      amyloid β-protein
      IAβ
      125I-Aβ
      AD
      Alzheimer's disease
      APP
      β-amyloid precursor protein
      IDE
      insulin-degrading enzyme
      CSF
      cerebrospinal fluid
      FBS
      fetal bovine serum
      CM
      conditioned media
      CHO
      Chinese hamster ovary
      PAGE
      polyacrylamide gel electrophoresis
      Tricine
      N-[2-hydroxy-1,1-bis(hydroxymethyl)ethyl]glycine
      NEM
      N-ethylmaleimide.
      Converging lines of evidence support the hypothesis that progressive cerebral accumulation of the 40–42-residue amyloid β-proteins (Aβs)1 is an early, invariant, and necessary step in the pathogenesis of Alzheimer's disease (AD). As a result, there is growing interest in decreasing cerebral Aβ levels as a therapeutic and preventative approach to the disease. Aβ is generated by endoproteolysis of the β-amyloid precursor protein (APP) and secreted constitutively by most mammalian cells throughout life. Whereas many studies have examined the proteolytic processing of APP and the mechanisms of Aβ production, almost nothing is known about how Aβ peptides are normally degraded and cleared following their secretion. We recently screened the conditioned media of several different cell lines for Aβ-degrading activity and found that the principal such activity was conferred by a nonmatrix metalloprotease that was released by microglial cells and other cells and efficiently degraded both endogenous and synthetic Aβ (
      • Qiu W.Q.
      • Ye Z.
      • Kholodenko D.
      • Seubert P.
      • Selkoe D.J.
      ). The release of the protease from microglial cells was augmented by activating the cells with lipopolysaccharide, suggesting that physiological and pathological stimuli may regulate the degree of Aβ degradation in extracellular fluids (
      • Qiu W.Q.
      • Ye Z.
      • Kholodenko D.
      • Seubert P.
      • Selkoe D.J.
      ).
      The properties of the Aβ-degrading metalloprotease we described are reminiscent of those of insulin-degrading enzyme (IDE), a metalloendopeptidase that can cleave a variety of small peptides, including insulin, glucagon, atrial natriuretic factor, and insulin-like growth factors I and II (reviewed in Refs.
      • Becker A.B.
      • Roth R.A.
      and
      • Authier F.
      • Posner B.I.
      • Bergeron J.J.M.
      ). IDE was recently identified as the protease responsible for the conversion of β-endorphin to γ-endorphin (
      • Safavi A.
      • Miller B.C.
      • Cottam L.
      • Hersh L.B.
      ). The protease was also recently reported to be present in human brain tissue, where it was shown to bind and cleave synthetic Aβ peptides at neutral pH (
      • Kurochkin I.V.
      • Goto S.
      ,
      • McDermott J.R.
      • Gibson A.M.
      ). However, the latter studies relied on homogenized brain, so that IDE released from the cytosol could have conferred the Aβ-degrading activity nonphysiologically. Indeed, the two major reported locations of IDE in cell types studied to date, the cytosol and peroxisomes (
      • Authier F.
      • Posner B.I.
      • Bergeron J.J.M.
      ,
      • Kuo W.-L.
      • Gehm B.D.
      • Rosner M.R.
      • Li W.
      • Keller G.
      ,
      • Chesneau V.
      • Perlman R.K.
      • Li W.
      • Keller G.-A.
      • Rosner M.R.
      ), raise the question of whether and how this protease could function to degrade endogenous Aβ, considering that the peptide has not been found in these two subcellular sites.
      In this study, we have compared the properties of the extracellular, Aβ-degrading metalloprotease, which we previously described (
      • Qiu W.Q.
      • Ye Z.
      • Kholodenko D.
      • Seubert P.
      • Selkoe D.J.
      ), with those of IDE and find that the two are indistinguishable. Partial purification of the protease from the conditioned medium of a microglial cell line demonstrates that the Aβ-degrading activity is IDE. This conclusion is confirmed by immunodepletion of IDE from conditioned medium. We further report experiments that suggest that the enzyme is released into the extracellular fluid by intact microglial cells that show no abnormal permeability. That such release occursin vivo in humans is confirmed by the immunochemical detection of authentic, 110-kDa IDE in freshly obtained lumbar cerebrospinal fluid (CSF). In addition to a role for IDE in degrading secreted Aβ, we find that a time-dependent oligomerization of synthetic Aβ in the conditioned media of cultured cells is completely blocked by three different inhibitors of IDE activity. This finding suggests that IDE is capable of regulating the level of monomeric Aβ by both degradation and oligomerization. The implications of these findings for the fate of Aβ monomers and the role of Aβ accumulation in AD are discussed.

      EXPERIMENTAL PROCEDURES

       Cell Culture

      Mouse BV-2 microglial cells were routinely cultured in RPMI 1640 with 10% fetal bovine serum (FBS). To characterize Aβ-degrading activity in their conditioned media (CM), the cells were conditioned in either fresh RPMI 1640/10% FBS or the serum-free medium, N2, for 16–18 h, and the CM was passed through a 0.22-μm filter to remove floating cells. Chinese hamster ovary (CHO) cells stably transfected with βAPP770 cDNA containing the V717F FAD mutation (7PA2 cells) (
      • Qiu W.Q.
      • Ye Z.
      • Kholodenko D.
      • Seubert P.
      • Selkoe D.J.
      ) were cultured in Dulbecco's modified Eagle's medium with 10% FBS and G418 (200 μg/ml).

       Assaying Degradation of Endogenous Aβ by Immunoprecipitation

      Confluent monolayers of 7PA2 cells in 10-cm dishes were preincubated for 30 min in methionine- and serum-free medium and labeled for 4 h with 300 μCi of [35S]methionine. The labeled CM were centrifuged at 2100 × g for 30 min and used immediately or stored at −70 °C. Two ml of labeled CM were mixed with 2 ml of unconditioned N2 medium (control) or BV-2 CM or fractions of the latter obtained during IDE purification (see below), either in the absence or presence of N-ethylmaleimide (NEM) or 1,10-phenanthroline or human insulin (Sigma) and incubated at 37 °C for 16 h. The amount of labeled Aβ remaining in each sample was assessed by immunoprecipitation with the high affinity Aβ polyclonal antibody R1282 followed by 10–20% Tris/Tricine SDS-PAGE and gel fluorography (
      • Haass C.
      • Schlossmacher M.G.
      • Hung A.Y.
      • Vigo-Pelfrey C.
      • Mellon A.
      • Ostaszewski B.L.
      • Lieberburg I.
      • Koo E.H.
      • Schenk D.
      • Teplow D.B.
      • Selkoe D.J.
      ).

       Preparation of Cell Cytosol and Human Brain Extracts

      Confluent cultures of BV-2 or human kidney 293 cells were detached from 15-cm dishes with 20 mm EDTA in ice-cold phosphate-buffered saline and pelleted. These cells or aliquots of frozen human brain tissue were suspended in homogenization buffer (10 mm HEPES, pH 7.4, 1 mm EDTA, 0.25 msucrose, supplemented with a protease inhibitor mixture) and disrupted using 10 strokes in a Dounce homogenizer followed by four passages through a 25-gauge needle. Nuclei and unbroken cells were pelleted by centrifugation at 3000 × g for 10 min. The pellets were resuspended in 1.5 ml of homogenization buffer and centrifuged at 3000 × g for 10 min. The postnuclear supernatants from both centrifugation steps were combined and centrifuged at 80,000 × g for 1 h to separate the cytosol and membrane fractions.

       Western Blotting of Insulin-degrading Enzyme

      Equal volumes of (a) cytosols prepared from BV-2 cells or human kidney 293 cells, (b) BV-2 CM (concentrated to a volume equal to that of the BV-2 cytosol by NH4SO4 precipitation), (c) human CSF, or (d) human CSF concentrated 10-fold by a Centricon 30 (Amicon) filter were electrophoresed on 10% Tris/glycine SDS-PAGE gels and transferred to polyvinylidine difluoride membrane (
      • Qiu W.Q.
      • Borth W.
      • Ye Z.
      • Haass C.
      • Teplow D.B.
      • Selkoe D.J.
      ). The SuperSignal HisProbe Western blotting kit (Pierce) was used for immunoblotting IDE according to the manufacturer's instructions. Polyclonal antibody 2BS, raised specifically against IDE (
      • Kuo W.-L.
      • Gehm B.D.
      • Rosner M.R.
      • Li W.
      • Keller G.
      ), was used at 1:10,000 dilution.

       Assaying Degradation of 125I-Aβ by Gel Fluorography or Trichloroacetic Acid Precipitation

      125I-Aβ (IAβ) (specific activity, ∼2000 Ci/mmol) was purchased from Amersham Pharmacia Biotech, dissolved in H2O, aliquoted, and stored at −20o until use to avoid freeze/thawing. 10,000 cpm were added per ml of medium to whole cultures or their collected CM. After increasing intervals of incubation at 37 °C, aliquots of the media were removed, centrifuged, and examined by 10–20% Tris/Tricine SDS-PAGE and gel fluorography to observe any degradation or oligomerization of Aβ. At the same time, 90 μl of each aliquot was mixed with 110 μl of 15% trichloroacetic acid and incubated on ice for 15 min to precipitate undegraded IAβ. The trichloroacetic acid-precipitated samples were centrifuged (16,000 × g, 10 min), and the amounts of label in the supernatant (representing degraded products) and the pellet (representing intact Aβ) were counted. Because partially degraded peptide products of Aβ may still be precipitated by trichloroacetic acid, this assay underestimates the absolute level of substrate degradation compared with assaying by SDS-PAGE/gel fluorography.

       Characterization of Cell Viability

      BV-2 cells were routinely grown at 37 °C in RPMI, 10% FBS. At increasing intervals up to 24 h, CM were collected and filtered through 0.22-μm cellulose acetate, and the amount of lactate dehydrogenase in each sample was measured by an in vitro toxicology assay (Sigma) according to instructions. In addition, BV-2 cells were grown in the presence or absence of 10% FBS at 37 °C for 24 h followed by the LIVE/DEAD viability/cytotoxicity assay (Molecular Probes, Inc., Eugene, OR) according to the manufacturer's instructions. In this assay, polyanionic calcein is retained within live cells, producing an intense uniform green fluorescence (excitation/emission = 495/515 nm). Ethidium bromide, normally excluded, enters cells having damaged membranes and undergoes a 40-fold enhancement of fluorescence upon binding to nucleic acids, thereby producing a bright red fluorescence in permeable cells (excitation/emission = 495/635 nm). As a positive control in these experiments, BV-2 cells were briefly treated (5 min) with digitonin (0.05%) to permeabilize them.

       Purification of the Aβ-degrading Activity from Conditioned Media

      BV-2 cells were cultured in RPMI 1640, 10% FBS in T100 flasks for 3 days. The cells were collected, washed, changed to serum-free N2 medium, and further cultured for 16 h at 37 °C. Pooled CM (∼1 liter) were filtered through 0.22-μm cellulose acetate to remove floating cells. The CM were precipitated with 40% NH4SO4, and the resultant supernatant was further precipitated with 60% NH4SO4. The latter precipitate was dissolved in ∼100 ml of 50 mmTris-HCl, pH 7.5, and dialyzed in the same buffer overnight. The sample was assayed for Aβ-degrading activity by immunoprecipitation of endogenous Aβ with R1282 as described above and then applied to a Bio-Scale Q (Bio-Rad) anion exchange column, which was preequilibrated with 50 mm Tris-HCl, pH 7.5. The column was eluted with a linear NaCl gradient from 10 mm to 1 m run at a flow rate of 0.5 ml/min. Aliquots of the resultant fractions (1.5 ml) were assayed for Aβ-degrading activity either by immunoprecipitation of [35S]Met-labeled endogenous Aβ or by trichloroacetic acid precipitation of synthetic IAβ, as described above. Fractions containing the most Aβ-degrading activity were pooled and concentrated to 200 μl by Centricon 30, followed by chromatography on a Superose 12 (Amersham Pharmacia Biotech) gel filtration column. This column was equilibrated with 50 mm phosphate buffer, pH 7.4, and the sample was eluted in the same buffer at a flow rate of 0.3 ml/min. Fractions (0.9 ml) were monitored for protein by UV absorbance at 280 nm, for Aβ-degrading activity and Aβ oligomer formation as described above, and for protein composition by SDS-PAGE followed by Coomassie staining. The fractions were also assayed for IDE by Western blotting with antibody 2BS.

      RESULTS

       The Aβ-degrading Metalloprotease Released into Microglial Conditioned Medium Has the Properties of Insulin-degrading Enzyme

      We recently reported that the clearance of secreted Aβ peptides from the media of several neural and nonneural cell lines was principally mediated by a nonmatrix metalloprotease released by these cells (
      • Qiu W.Q.
      • Ye Z.
      • Kholodenko D.
      • Seubert P.
      • Selkoe D.J.
      ). Because the properties of the enzyme we described were similar to those of insulin-degrading enzyme, we further examined the metalloprotease with several reagents known to inhibit IDE. Using BV-2 microglial cells, which release higher levels of the Aβ-degrading metalloprotease than other cell types we screened (
      • Qiu W.Q.
      • Ye Z.
      • Kholodenko D.
      • Seubert P.
      • Selkoe D.J.
      ), we examined the effects of the sulfhydryl-modifying reagent NEM. Confluent BV-2 cells were washed and changed to the serum-free medium, N2, for conditioning at 37 °C for 16 h. The CM were then filtered and collected. CHO cells stably transfected with APP770 cDNA containing the V717F mutation were metabolically labeled with [35S]methionine as described (
      • Qiu W.Q.
      • Ye Z.
      • Kholodenko D.
      • Seubert P.
      • Selkoe D.J.
      ), and the resultant medium, containing abundant labeled Aβ and p3 peptides, was incubated with the filtered BV-2 CM (or with unconditioned N2 medium as a control) for 24 h at 37 °C in the presence or absence of NEM. (p3 is a peptide comprising residues 17–40 (or 17–42) of Aβ that results from constitutive proteolysis of APP by α- and then γ-secretases, whereas Aβ (residues 1–40 or 1–42) results from constitutive proteolysis of APP by β- and then γ-secretases.) Subsequent immunoprecipitation with an Aβ antibody (R1282) and gel fluorography revealed the expected marked decrease of Aβ in the BV-2 CM, and this was completely inhibited by 100 μm NEM (Fig. 1 A). This result, together with the previously observed inhibition by 1,10-phenanthroline (
      • Qiu W.Q.
      • Ye Z.
      • Kholodenko D.
      • Seubert P.
      • Selkoe D.J.
      ), suggests that Aβ degradation in BV-2 CM is specifically mediated by a thiol metalloprotease such as IDE. In parallel, we examined the degradation of iodinated insulin (10 pm) in the BV-2 CM and observed a time-dependent loss of intact insulin by SDS-PAGE that was similar to the degradation of iodinated Aβ1–40 at the same concentration. The insulin degradation was likewise inhibited by 1,10-phenanthroline and NEM (data not shown).
      Figure thumbnail gr1
      Figure 1Loss of secreted Aβ is mediated by insulin-degrading enzyme present in the extracellular fluid of cultured microglial cells. A, aliquots of CM of 7PA2 (APP-transfected CHO) cells labeled with [35S]Met for 4 h in serum-free medium were incubated (37 °C, 16 h) with 3 ml of either unconditioned N2 medium (lane 1) or BV-2 CM (lanes 2–4) in the absence (lanes 1 and 2) or presence of 10 μm (lane 3) of 100 μm(lane 4) NEM. The incubated samples were then immunoprecipitated with Aβ antibody R1282 and electrophoresed on Tris/Tricine gels followed by autoradiography. Mass markers (kDa), Aβ, and p3 are all indicated. B, aliquots of labeled 7PA2 CM were incubated with 3 ml of unconditioned N2 medium (lane 1) or BV-2 CM (lanes 2–7) in the presence of increasing amounts of human insulin as indicated (lanes 3–6) or in 1 mm1,10-phenanthroline (lane 7) at 37 °C for 16 h and assayed as in A. C, equal volumes of human kidney 293 cell cytosol (lane 1), BV-2 cell cytosol (lane 2), and BV-2 CM concentrated to the same volume as BV-2 cytosol by 60% NH4SO4 precipitation and dialysis (lane 3) were electrophoresed on 10% Tris-glycine gels and blotted with the IDE antibody 2BS. ×, ∼110-kDa insulin-degrading enzyme.
      To further characterize the protease released by the microglial cells, we added the principal substrate of IDE, insulin, to the reaction mixture to determine whether insulin can compete out the Aβ-degrading activity. Degradation of endogenous Aβ in BV-2 CM was found to be progressively inhibited by increasing amounts of insulin between 100 nm and 10 μm (Fig. 1 B). One μm insulin inhibited degradation of Aβ by ∼50%, and 10 μm completely prevented the loss of Aβ. At the latter dose, insulin was as effective in this assay as 1 mm1,10-phenanthroline (Fig. 1 B), which we previously found to inhibit the Aβ-degrading metalloprotease completely (
      • Qiu W.Q.
      • Ye Z.
      • Kholodenko D.
      • Seubert P.
      • Selkoe D.J.
      ). To support the conclusion that it is IDE that degrades Aβ in BV-2 CM, we added other known substrates of IDE such as glucagon, transforming growth factor-α, or brain natriuretic peptide to the reaction mixture and showed that each of these inhibited the Aβ-degrading activity, but to a lesser extent than insulin did on a molar basis (data not shown). The p3 peptide is relatively resistant to degradation by the protease (Fig. 1, A and B), as reported previously (
      • Qiu W.Q.
      • Ye Z.
      • Kholodenko D.
      • Seubert P.
      • Selkoe D.J.
      ). In this regard, the cleavage sites in synthetic Aβ1–40 generated by IDE purified from the cytosol of CHO cells have recently been localized to residues 4–5, 13–14, 14–15, and 20–21,
      A. Safavi and L. Hersh, unpublished data.
      potentially explaining the relative resistance of the p3 peptide (Aβ 17–40/42) to IDE-mediated degradation.
      To demonstrate directly the presence of IDE in the BV-2 CM, we used an IDE-specific polyclonal antibody, 2BS (
      • Kuo W.L.
      • Montag A.G.
      • Rosner M.R.
      ), to immunoblot the CM. Aliquots of human 293 cytosol (which contains abundant IDE and serves as a positive control), BV-2 cytosol, and BV-2 CM (concentrated by NH4SO4 precipitation to a volume equal to that of the BV-2 cytosol sample) were electrophoresed on 10% SDS-PAGE gels and immunoblotted with 2BS (Fig. 1 C). The mobilities of human IDE (lane 1) and mouse IDE (lanes 2 and 3) are slightly different. In all three samples, a characteristic 110-kDa protein was specifically immunolabeled, suggesting that BV-2 cells release intact IDE into their medium under our culture conditions.

       IDE Purified from Microglial Cell Medium Mediates the Degradation of Aβ

      Because the above data suggested that IDE or an IDE-like protease can efficiently degrade Aβ, we substantially purified and further characterized the activity from ∼1 liter of CM of BV-2 cells conditioned in serum-free N2 medium at 37 °C for 16 h. The CM was subjected to sequential fractionation by ammonium sulfate precipitation, anion exchange chromatography on a Bio-Scale Q column, and size exclusion chromatography on a Superose 12 column (TableI). Aβ-degrading activity was monitored at each step by immunoprecipitation/gel fluorography of [35S]Met-labeled 7PA2 medium (for analyzing endogenous Aβ) or by the degradation of synthetic IAβ followed by trichloroacetic acid precipitation. Any apparent oligomers of the IAβ (bands migrating above the 4-kDa monomer position) were simultaneously visualized by SDS-PAGE (
      • Podlisny M.B.
      • Walsh D.M.
      • Amarante P.
      • Ostaszewski B.L.
      • Stimson E.R.
      • Maggio J.E.
      • Teplow D.B.
      • Selkoe D.J.
      ). The final fractions obtained from the size exclusion chromatography (Fig. 2 A) were incubated with IAβ, and those fractions that caused degradation in the absence but not the presence of 1,10-phenanthroline were identified by trichloroacetic acid precipitation (Fig. 2 B). Comparison of this proteolytic activity curve with the size standards showed that the Aβ-degrading moiety eluted from the column with a molecular mass of ∼100–150 kDa (Fig. 2, A and B). The Aβ-degrading activity of these fractions was further confirmed using the Aβ immunoprecipitation assay on endogenous([35S]Met-labeled) Aβ (Fig. 2 C). Western blotting with the 2BS antibody specifically identified intact 110-kDa IDE in the active fractions (Fig. 2 D), and its amount in each active fraction correlated closely with the relative Aβ-degrading activity of the fractions. Coomassie staining of the active fractions revealed the 110-kDa IDE band plus three or four other light bands (not shown), indicating marked enrichment but not full purification of IDE by this method. However, all of the other bands were equally or more abundant in fractions that had low or no Aβ-degrading activity. When the active column fractions (fractions 9–15) were incubated with IAβ followed by SDS-PAGE/fluorography, we observed the decrease in the IAβ monomer as well as an apparent SDS-stable oligomerization of some of the peptide (Fig. 2 E) (see below). Taken together, the results in Fig. 2 establish that the amount of IDE present in column fractions during purification correlates very closely with the degree of Aβ degradation.
      Table ISummary of partial purification of insulin-degrading enzyme from BV-2 CM
      Purification StepVolumeProteinTotal proteinYieldAβ-degrading activity
      mlmg/mlmg%
      BV-2CM10000.20200100+
      Ammonium sulfate800.1084+
      Bio-Scale Q150.081.20.6+
      Superose 122.70.110.30.15+
      Figure thumbnail gr2
      Figure 2Purification of IDE from BV-2 conditioned medium. Aβ-degrading activity was substantially purified from BV-2 CM as described under “Experimental Procedures” and Table .A, protein content per fraction (0.9 ml) from the final size exclusion column was determined by UV spectrophotometry at 280 nm. The approximate elution positions of marker proteins thyroglobulin (570 kDa), ovalbumin (158 kDa), myoglobin (44 kDa), γ-globulin (17 kDa), and vitamin B12 (1.3 kDa) are indicated. B, Aβ-degrading activity was assayed by incubating IAβ in each fraction for 16 h at 37 °C in the absence (open squares) or presence (closed squares) of 1,10-phenanthroline followed by trichloroacetic acid precipitation; total cpm in the trichloroacetic acid pellets is graphed. C, aliquots (0.5 ml) of fractions 9–15 were assayed by incubating with 3 ml of [35S]Met-labeled 7PA2 CM for 16 h at 37 °C and immunoprecipitating with Aβ antibody R1282. D, fractions 9–15 were examined by immunoblotting with IDE antibody 2BS. The position of the 97-kDa marker and the characteristic position of IDE are indicated. E, fractions 9–15 were incubated with IAβ for 16 h at 37 °C followed by 10–20% Tris/Tricine SDS-PAGE and autoradiography.
      To prove unequivocally that IDE is the protease in CM responsible for the degradation of Aβ, we performed immunodepletion of CM with an IDE monoclonal antibody, 9B12 (
      • Kurochkin I.V.
      • Goto S.
      ); this completely removed the Aβ-degrading activity from the medium of CHO cells (TableII). This antibody cannot be used with BV-2 CM, because it reacts very poorly with murine IDE, but we reported previously that CHO CM has an Aβ-degrading protease that is qualitatively indistinguishable from that in BV2 CM (
      • Qiu W.Q.
      • Ye Z.
      • Kholodenko D.
      • Seubert P.
      • Selkoe D.J.
      ). In parallel with the removal of Aβ-degrading activity, immunodepletion of IDE simultaneously removed insulin-degrading activity (Table II). Immunodepletion performed in the absence of the IDE antibody resulted in no significant loss of Aβ-degrading activity.
      Table IISimultaneous immunodepletion of Aβ- and insulin-degrading activities by anti-IDE antibody 9B12
      ConditionAβ degradationInsulin degradation
      %%
      CHO CM100100
      CHO CM + protein A-Sepharose8993
      CHO CM + 9B12 + protein A-Sepharose00

       IDE Mediates Aβ Degradation Extracellularly in the Absence of Detectable Cellular Injury

      We previously reported that the degradation of Aβ in the CM of BV-2 cells was significantly lower when the cells were conditioned in 10% FBS than in serum-free medium (
      • Qiu W.Q.
      • Ye Z.
      • Kholodenko D.
      • Seubert P.
      • Selkoe D.J.
      ). One explanation for this finding could be that the cells are more prone to injury in the absence of serum and thus release more protease, particularly in light of the general assumption that IDE is localized to cytosol and not normally released by intact cells. To address this issue, we used a highly sensitive assay to detect Aβ degradation in the presence of 10% FBS; 125I-labeled synthetic Aβ (∼20 pm) served as the substrate and was incubated in whole BV-2 cultures or in their CM alone. The loss of intact IAβ during a 24-h incubation was assayed by quantitating trichloroacetic acid-precipitable counts, a method widely used to study the degradation of other iodinated substrates of IDE (e.g. Refs.
      • McDermott J.R.
      • Gibson A.M.
      and
      • Garcia J.V.
      • Stoppelli M.P.
      • Decker S.J.
      • Rosner M.R.
      ). Radioactivity in both the trichloroacetic acid-insoluble pellets (representing intact peptide) and the supernatant (representing the products) was measured. Because partially degraded fragments of Aβ may still be precipitated by trichloroacetic acid, this assay underestimates the absolute level of substrate cleavage. Whole BV-2 cultures degraded ∼50% of the extracellular IAβ during an 18-h incubation (Fig. 3 A), and this was completely inhibited by 10 μm insulin (Fig. 3 B). We also performed this experiment on cultures of M17 human neuroblastoma cells and rat PC12 cells. Substantially less degradation of IAβ occurred in M17 then in BV-2 cultures, and the PC12 cultures showed very little or no detectable degradation under the same conditions (data not shown). Because both M17 and PC12 cells have similar amounts of IDE in their cytoplasm as do BV-2 cells (not shown), the secretion of IDE activity capable of mediating extracellular Aβ degradation is cell type-dependent and does not simply reflect nonspecific release in cell culture.
      Figure thumbnail gr3
      Figure 3Degradation of Aβ in BV-2 cultures is mediated by IDE activity in the conditioned medium. Aand B, whole BV-2 cultures were incubated with IAβ in the absence (A) or presence (B) of 10 μm human insulin, and aliquots of the CM were removed at the indicated times. C–F, BV-2 cells were conditioned in the presence of 10% FBS for 24 h at 37 °C, and the CM were collected. CM were incubated with IAβ in the absence of inhibitors (C) or in the presence of 1 mm1,10-phenanthroline (D), 100 μm NEM (E), or 10 μm human insulin (F). Aliquots were removed at the indicated times and assayed by trichloroacetic acid precipitation (see “Experimental Procedures”). Pellets (open squares) and supernatants (filled squares) were assayed for cpm (means ± S.D. values for three experiments).
      Because the loss of intact IAβ in whole cultures could be due to a mixture of cell-based and secreted activities, we asked whether just the CM of BV-2 cells is sufficient to mediate the degradation of Aβ seen in the above experiments. IAβ was incubated for 24 h at 37 °C solely in the CM of BV-2 cells conditioned in the presence of 10% FBS. The CM degraded ∼50% of the IAβ during the incubation (Fig. 3 C), similar to the whole cultures, and this was inhibited by 1,10-phenanthroline (Fig. 3 D), NEM (Fig. 3 E), and insulin (Fig. 3 F). These results clearly indicate that Aβ degradation is largely mediated by extracellular IDE activity in BV-2 cultures.
      To confirm that this release of IDE into the culture medium occurs while BV-2 cell membranes are intact, we simultaneously assayed lactate dehydrogenase activity in the CM of cultures conditioned in 10% FBS and found no significant release of this cytoplasmic enzyme at time points up to 24 h (Fig. 4 F). To prove further the intactness of the plasma membrane, the LIVE/DEAD viability/cytotoxicity assay was performed on BV-2 cells cultured in the presence (Fig. 4,A and B) or absence (Fig. 4, C and D) of 10% FBS for up to 24 h. We observed no significant cell death (by calcein AM signal and ethidium bromide staining) even in the absence of serum, whereas sister cultures permeabilized briefly with digitonin (0.05%) were clearly positive in this assay (Fig. 4 E). Taken together, these several assays provide no evidence that the release of proteolytically active IDE from the BV-2 microglial cells occurs as a secondary result of cell injury.
      Figure thumbnail gr4
      Figure 4Viability and intactness of cultured BV-2 cells. A–E, the LIVE/DEAD viability/cytotoxicity assay was performed according to the manufacturer's instructions on BV-2 cells cultured at 37 °C for 24 h in the presence (Aand B) or the absence (C and D) of 10% serum. As a positive control, BV-2 cells were briefly permeabilized with digitonin (0.05%) (E). The polyanionic calcein is well retained within live cells, producing an intense uniform green fluorescence (A and C). Ethidium bromide enters cells with damaged membranes and undergoes a 40-fold enhancement of its fluorescence upon binding to nucleic acids, thereby producing a bright red fluorescence in leaky cells (B, D, and E). F, BV-2 cells were conditioned in 10% FBS at 37 °C, and aliquots of CM were removed at the indicated times. Cytoplasmic lactate dehydrogenase present in the CM at each time point was measured.

       Detection of Intact IDE in Fresh Human Cerebrospinal Fluid

      Although there have been reports of the release of IDE from cells (
      • Roth R.A.
      • Mesirow M.L.
      • Cassell D.J.
      • Yokono K.
      • Baba S.
      ,
      • Semple J.W.
      • Lang Y.
      • Speck E.R.
      • Delovitch T.L.
      ) and the presence of intact, 110-kDa IDE at the cell surface (
      • Seta K.A.
      • Roth R.A.
      ), it has generally been assumed that IDE is restricted to the cytosol and peroxisomes and is not normally released from cells (
      • Kuo W.-L.
      • Gehm B.D.
      • Rosner M.R.
      • Li W.
      • Keller G.
      ,
      • Authier F.
      • Bergeron J.J.
      • Ou W.J.
      • Rachubinski R.A.
      • Posner B.I.
      • Walton P.A.
      ). However, our results in the previous section show that readily detectable amounts of enzymatically active IDE are found in the medium of intact BV-2 cells under normal culture conditions, including in serum-containing cultures. Another way to confirm that a protein is released from cells under entirely physiological circumstances in vivo is to examine normal, essentially acellular biological fluids such as fresh human CSF for the presence of the protein. To this end, we probed both fresh and frozen lumbar CSF samples collected from living nondemented subjects (n = 11), patients with AD (n = 6), and one patient with a non-AD dementia by Western blotting with a well characterized IDE antibody, 2BS. In all CSF samples examined (exemplified in Fig. 5 A), a full-length 110-kDa IDE protein was specifically detected, and it co-migrated with the characteristic IDE band found in the cytosol of human 293 cells (Fig. 5 A, lane 7). These results were confirmed using another human IDE antibody, 9B12 (not shown). The amounts varied somewhat from sample to sample but were not obviously different in the small number of AD CSF samples available to us to date. Next, we probed soluble fractions prepared from human cortex of one non-AD and three AD subjects. Western blotting showed that 110-kDa IDE was present in all four samples and comigrated with IDE in 293 cytosol (Fig. 5 B). We conclude that intact IDE protein is normally present in human CSF and brain.
      Figure thumbnail gr5
      Figure 5Detection of intact IDE in fresh human cerebrospinal fluid and human brain tissue. A, lumbar CSF samples from patients with AD (lanes 1–3) and nondemented control subjects (lanes 4–6) were concentrated 10-fold by Centricon, separated by 10% Tris/glycine SDS-PAGE, and immunoblotted with the IDE antibody 2BS. The cytosol of human 293 cells was used as a positive control (lane 7). The upper band, but not IDE, was recognized by the secondary antibody alone (not shown), indicating that it is nonspecific. B, postmortem samples of frozen human cortex from one nondemented control subject (lane 1) and three AD cases (lanes 2–4) were extracted. Equal amounts of protein (60 μg) from each sample were electrophoresed on 10% Tris/glycine SDS-PAGE and immunoblotted with the IDE antibody 2BS. Human 293 cell cytosol served as a positive control (lane 5).

       IDE Activity in Microglial Conditioned Medium Is Associated with the Formation of Apparent Oligomers of Aβ

      To visualize Aβ degradation directly in the above incubations of IAβ in whole BV-2 cultures or their CM, these samples were analyzed by gel fluorography after 10–20% Tris/Tricine SDS-PAGE (Fig. 6, A–C). As the incubation interval at 37 °C increased, the amount of the monomeric Aβ band gradually decreased by ∼50% (Fig. 6 B, lanes 1–5), consistent with the results of trichloroacetic acid precipitation (Fig. 3). We also observed the progressive formation over time of small amounts of SDS-stable IAβ species migrating at 6–20 kDa, as described previously in CHO cell cultures under similar conditions (
      • Podlisny M.B.
      • Walsh D.M.
      • Amarante P.
      • Ostaszewski B.L.
      • Stimson E.R.
      • Maggio J.E.
      • Teplow D.B.
      • Selkoe D.J.
      ,
      • Podlisny M.B.
      • Ostaszewski B.
      • Haass C.H.
      • Selkoe D.J.
      ). In the presence of insulin, the loss of the 4-kDa Aβ band during the 37 °C incubation was abolished, consistent with all of the previous data implicating IDE in the degradation. Unexpectedly, the formation of the apparent Aβ oligomers was also completely prevented by insulin (Fig. 6 B, comparelanes 6–10 with lanes 1–5), suggesting that both of these events might be mediated by IDE. To prove that IDE is the same factor that mediates the degradation and the apparent oligomerization of Aβ, IAβ was incubated in BV-2 CM alone in the absence or presence of insulin or of the protease inhibitor 1,10-phenanthroline or NEM and then assayed by SDS-PAGE/gel fluorography (Fig. 6 C). The two protease inhibitors and insulin, which efficiently inhibited IDE from degrading Aβ monomer, simultaneously prevented the apparent oligomerization. The experiment was repeated five times with the same result. Plain, unconditioned medium (RPMI, 10% FBS) incubated identically with IAβ for up to 24 h produced no change in the Aβ monomer (Fig. 6 A). Finally, when IDE substantially purified from BV-2 CM was used, the characteristic degradation of the IAβ as well as its apparent oligomerization were again observed by SDS-PAGE (Fig. 2 E).
      Figure thumbnail gr6
      Figure 6Simultaneous inhibition of both the degradation and the apparent oligomerization of Aβ mediated by IDE in BV-2 cultures and their media. A, IAβ was incubated in plain, unconditioned N2 medium (RPMI/10% FBS) for 3 h (lane 1) or 5 h (lane 2) or in BV-2 CM in the absence (lane 3) or presence (lane 4) of 1 mm 1,10-phenanthroline, followed by 10–20% Tris/Tricine SDS-PAGE and autoradiography. B, whole BV-2 cultures in 10% FBS were incubated with IAβ in the absence (lanes 1–5) or presence (lanes 6–10) of 10 μm human insulin for 0 h (lanes 1 and 6), 0.5 h (lanes 2 and 7), 3 h (lanes 3and 8), 18 h (lanes 4 and 9), or 24 h (lanes 5 and 10). Aliquots of the CM were removed and examined by 10–20% Tris/Tricine SDS-PAGE and autoradiography. Note the decrease in the intensity of the IAβ monomer band over time (lanes 1–5) and the corresponding appearance of apparent SDS-stable oligomers and some radioactive degradation products.C, BV-2 cells were cultured in 10% FBS for 24 h, and the CM were collected and incubated with IAβ in the absence (lanes 1–7) or presence of 1 mm1,10-phenanthroline (lanes 8–14), 100 μm NEM (lanes 15–21), or 10 μm human insulin (lanes 22–28). Aliquots of the reaction mixtures were removed at 0 h (lanes 1, 8, 15, and 22), 0.5 h (lanes 2, 9, 16, and 23), 3 h (lanes 3, 10, 17, and 24), 6 h (lanes 4, 11, 18, and 25), 18 h (lanes 5, 12, 19, and 26), 24 h (lanes 6, 13, 20, and 27), or 48 h (lanes 7, 14, 21, and 28) and examined by 10–20% Tris/Tricine SDS-PAGE and autoradiography. Labelsare as in B.

      DISCUSSION

      Evidence from many laboratories suggests that decreasing Aβ levels in brain tissue is a rational approach to prevent or slow the development of AD. In the case of the three genes implicated to date in autosomal dominant forms of AD, excessive cellular production of Aβ, particularly Aβ42, appears to be the common mechanism by which they produce the AD phenotype (
      • Selkoe D.J.
      ). However, in the large number of AD cases in whom a genetic risk factor has not yet been identified or is not operative, the reasons for the excessive accumulation of Aβ in brain are unknown. Such cases, often referred to as “sporadic” AD, are not known to have an abnormality of Aβ production. It is therefore possible that decreases in the normal degradation and clearance of Aβ in brain rather than a rise in its production could underlie some or many cases of the disease. It has been reported that microglia can bind and internalize microaggregates of Aβ in vitro (
      • Paresce D.M.
      • Chung H.
      • Maxfield F.R.
      ). However, the protease(s) involved in the degradation and clearance of Aβ monomer have received very little attention, and the principal mechanisms that regulate the steady state levels of the peptide in normal brain remain undefined.
      In vitro biochemical studies examining the proteolysis of synthetic Aβ by certain known purified proteases have found evidence of variable degradation of the peptide by gelatinase A (
      • Gowing E.
      • Roher A.E.
      • Woods A.S.
      • Cotter R.J.
      • Chaney M.
      • Little S.P.
      • Ball M.J.
      ,
      • Walsh D.M.
      • Williams C.H.
      • Kennedy H.E.
      • Allsop D.
      • Murphy G.
      ); EC3.4.24.11 (
      • Howell S.
      • Nalbantoglu J.
      • Crine P.
      ); cathepsins B and D (
      • McDermott J.R.
      • Gibson A.M.
      ,
      • Marks N.
      • Berg M.J.
      • Chi L.M.
      • Choi J.
      • Durrie R.
      • Swistok J.
      • Makofske R.C.
      • Danho W.
      • Sapirstein V.S.
      ); collagenase, chymotrypsin, and trypsin (
      • Van Nostrand W.E.
      • Schmaier A.H.
      • Neiditch B.R.
      • Siegel R.S.
      • Raschke W.C.
      • Sisodia S.S.
      • Wagner S.L.
      ); a serine protease-α2-macroglobulin complex (
      • Qiu W.Q.
      • Borth W.
      • Ye Z.
      • Haass C.
      • Teplow D.B.
      • Selkoe D.J.
      ), and IDE (
      • Kurochkin I.V.
      • Goto S.
      ,
      • McDermott J.R.
      • Gibson A.M.
      ). Because many purified proteases are capable of cleaving a variety of pure substrates with poor specificity under in vitro conditions, we chose to screen several neural and nonneural cell lines for the secretion of endogenous proteases that efficiently degrade naturally secreted Aβ peptides under biologically relevant conditions and concentrations. Using this approach, we identified an Aβ-degrading metalloprotease released constitutively by the microglial cell line BV-2 and, to a lesser extent, by certain other cells (
      • Qiu W.Q.
      • Ye Z.
      • Kholodenko D.
      • Seubert P.
      • Selkoe D.J.
      ). We have now partially purified and characterized this protease and show that it is indistinguishable from the well known metalloendopeptidase, insulin-degrading enzyme, which is known to degrade several small, structurally diverse secreted peptides. Surprisingly, we find that IDE activity is also associated with an apparent oligomerization of synthetic Aβ in cell culture, suggesting that this protease, which we show is present in vivo in human brain and CSF, may play a key role in regulating the amount and fate of extracellular Aβ in brain tissue.

       Insulin-degrading Enzyme Released from Intact Cells Can Degrade Aβ Peptides

      We report several lines of evidence that the Aβ-degrading enzyme in the BV-2 medium is IDE or a highly homologous, immunochemically cross-reactive protease of the same size. First, the Aβ degrading activity of BV-2 CM is fully inhibited by the sulfhydryl-modifying reagent NEM and the chelating agent 1,10-phenanthroline (Fig. 1 A), consistent with the inhibitor profile of IDE (
      • Becker A.B.
      • Roth R.A.
      ,
      • Authier F.
      • Posner B.I.
      • Bergeron J.J.M.
      ). Second, a specific substrate of IDE, insulin, competitively inhibits the degradation of endogenous Aβ by BV-2 CM (Fig. 1 B). Third, Western blot analysis with an IDE-specific antibody confirms the presence of an immunoreactive 110-kDa protein in BV-2 CM, and it comigrates with an immunoreactive band of this size in BV-2 and 293 cell cytosol (Fig. 1 C), a compartment in which IDE is principally found and has been well characterized (
      • Authier F.
      • Bergeron J.J.
      • Ou W.J.
      • Rachubinski R.A.
      • Posner B.I.
      • Walton P.A.
      ,
      • Stentz F.B.
      • Harris H.L.
      • Kitabchi A.E.
      ). Fourth, partial purification of our Aβ-degrading protease from BV-2 CM shows that chromatographic fractions that contain maximal Aβ-degrading activity directly parallel those with maximal amounts of immunoreactive 110-kDa IDE (Fig. 2). Fifth, immunodepletion with an IDE monoclonal antibody, 9B12(5), completely abolishes the Aβ-degrading activity in the medium of Chinese hamster ovary cells (Table II), which we previously showed release the same IDE-like Aβ-degrading thiol metalloprotease that BV-2 cells do (
      • Qiu W.Q.
      • Ye Z.
      • Kholodenko D.
      • Seubert P.
      • Selkoe D.J.
      ); this antibody cannot be used on BV-2 CM, because it reacts very poorly with murine IDE.
      Although our data cannot formally exclude the possibility that another protease is activated via cleavage by IDE and then mediates the Aβ degradation we observe, two experimental results make this highly unlikely. First, the fact that insulin inhibits Aβ degradation in CM alone (without cells present) rules against this, since the hypothetical “activated” protease should already be present in the CM and then insulin would not inhibit. Second, the hypothetical other protease would have had to co-purify with IDE throughout our purification (Fig. 2).
      IDE has been shown to play a major role in the degradation of insulin, and several other small peptide hormones can also serve as substrates, including atrial natriuretic peptide, transforming growth factor α, and insulin-like growth factor II (reviewed in Refs.
      • Becker A.B.
      • Roth R.A.
      and
      • Authier F.
      • Posner B.I.
      • Bergeron J.J.M.
      ). Previous studies of CHO cells have reported that IDE is localized to the cytosol and peroxisomes (
      • Authier F.
      • Bergeron J.J.
      • Ou W.J.
      • Rachubinski R.A.
      • Posner B.I.
      • Walton P.A.
      ), loci in which Aβ has not been found, raising the question of whether and how the protease could degrade endogenous Aβ. Although some studies have detected insulin-degrading activity in the conditioned media of cultured cells (
      • Roth R.A.
      • Mesirow M.L.
      • Cassell D.J.
      • Yokono K.
      • Baba S.
      ,
      • Semple J.W.
      • Lang Y.
      • Speck E.R.
      • Delovitch T.L.
      ) and more recently on the cell surface (
      • Seta K.A.
      • Roth R.A.
      ), the extent of cell permeability and thus possible release of IDE from leaky cells, especially when serum is absent from the culture medium, was not specifically assessed in these reports. In our study, extracellular Aβ degradation mediated by IDE was observed both in whole BV-2 cultures and in their CM alone (Fig. 3), suggesting that the enzyme can be released, at least by the microglial cell line we used. Because we grew the cells in the presence of 10% FBS in such experiments, the likelihood of cell damage during the brief conditioning is very low. This conclusion is supported by several important control experiments: (a) no significant rise in lactate dehydrogenase was detected during the conditioning of the medium at the same time that Aβ-degrading activity was clearly increasing (Fig. 4 F); (b) only a very few leaky BV-2 cells were detected in the calcein/ethidium bromide cell viability assay (Fig. 4, A–E); (c) no morphological alterations or overt cell death were seen in the serum-free cultures over the time course of our experiments (≤24 h) (Fig. 4, Cand D); and (d) no extracellular IDE activity was detected in certain other IDE cell types (e.g. PC12 cells) cultured under identical conditions (data not shown), indicating that IDE release is not a nonspecific consequence of culturing cells. Most importantly, we demonstrate for the first time the presence of intact, 110-kDa IDE in fresh lumbar CSF obtained from living patients (Fig. 5), indicating that this protease exists in normal extracellular fluid under in vivo conditions. Further experiments will be needed to determine whether the IDE present in CSF is normally inhibited under basal conditions (as would be expected) but can be shown to have proteolytic activity under other conditions, e.g. following microglial activation in the brain and consequent enhanced release of the enzyme. In this regard, the physiological nature of IDE release from microglial cells is supported by our earlier finding that stimulation of BV-2 cultures with lipopolysaccharide results in a reproducible increase in Aβ-degrading activity in the medium, and this is prevented by IDE inhibitors (
      • Qiu W.Q.
      • Ye Z.
      • Kholodenko D.
      • Seubert P.
      • Selkoe D.J.
      ). Finally, Mentlein et al. (
      • Mentlein R.
      • Ludwig R.
      • Martensen I.
      ) have recently reported the presence of a metalloprotease in the media of primary rat microglia, the properties of which strongly suggest that it is IDE.

       A Potential Role for IDE in the Oligomerization of Aβ at Physiological Concentrations

      A central unresolved question about the pathobiology of AD is how excessive accumulation of the soluble, secreted form of Aβ gradually leads to the development of insoluble Aβ fibrils in innumerable extracellular plaques. Since the discovery of normal Aβ secretion in 1992, many studies have shown that APP-expressing neural and nonneural cells produce soluble, monomeric Aβ at picomolar to low nanomolar levels under physiological conditions (reviewed in Ref.
      • Selkoe D.J.
      ). On the other hand, synthetic Aβ peptides studied at much higher concentrations (10–1000 μm) aggregate upon in vitro incubation at 37 °C into insoluble fibrils resembling those in AD brain. In separate work, we have shown that both endogenous and synthetic Aβ peptides at physiological (low nm) concentrations can form small amounts of SDS-stable apparent oligomers in cell culture (
      • Podlisny M.B.
      • Walsh D.M.
      • Amarante P.
      • Ostaszewski B.L.
      • Stimson E.R.
      • Maggio J.E.
      • Teplow D.B.
      • Selkoe D.J.
      ,
      • Podlisny M.B.
      • Ostaszewski B.L.
      • Squazzo S.L.
      • Koo E.H.
      • Rydel R.E.
      • Teplow D.B.
      • Selkoe D.J.
      ). In the current study, we unexpectedly found that the apparent oligomerization of synthetic Aβ in the medium of BV-2 cells was completely inhibited by the competitive IDE substrate, insulin (Fig. 6 A), and by two IDE inhibitors, 1,10-phenanthroline and NEM (Fig. 6 B). These findings suggest that IDE activity is capable of mediating the oligomerization of Aβ. Such a conclusion is strongly supported by similar results obtained with substantially purified IDE from BV-2 CM (Fig. 2). At longer incubation times (≥18 h), apparent IAβ oligomers disappeared (Fig. 6 B), while trichloroacetic acid supernatant counts increased (Fig. 3), suggesting that the oligomers may be degraded. Mechanistically, we hypothesize that some Aβ fragments, which are generated by IDE, can enhance the oligomerization of the IAβ peptide and/or can themselves oligomerize. In view of the fact that IDE purified from CHO cytosol has been shown to generate Aβ fragments by cleavage after residues 4, 13, 14, and 20,2 the resultant N-terminally truncated fragments could promote Aβ oligomerization. It has been reported that N-terminally truncated fragments of synthetic Aβ aggregate more rapidly than the full-length peptide (
      • Pike C.J.
      • Overman M.J.
      • Cotman C.W.
      ). We cannot exclude the possibility that some of the bands we observe at −6–20 kDa on gels could represent oligomers composed in part of truncated Aβ peptides or even anomalously high migration of the truncated monomers themselves, as has been suggested to occur with some truncated fragments of synthetic Aβ (
      • Burdick D.
      • Soreghan B.
      • Kwon M.
      • Kosmoski J.
      • Knauer M.
      • Henschen A.
      • Yates J.
      • Cotman C.
      • Glabe C.
      ). It is known that the presence of very small amounts of aggregated Aβ or its fragments can act as powerful seeds to enhance the subsequent polymerization of monomeric Aβ into fibrils (
      • Jarrett J.T.
      • Lansbury Jr., P.T.
      ,
      • Jarrett J.T.
      • Berger E.P.
      • Lansbury Jr., P.T.
      ). Although a proteolytic effect of IDE on Aβ is the most plausible explanation for its oligomer-promoting activity, IDE could mediate the degradation and the oligomerization of Aβ by independent actions. Further work is needed to clarify the mechanism by which IDE activity seems to promote Aβ oligomerization and to determine the fate of the soluble oligomers of Aβ formed in our culture system. We are currently attempting to determine whether IDE can be stably transfected into BV-2 and other cells serve to increase both the degradation and oligomerization of extracellular Aβ.
      Serum insulin levels have been reported to rise with age in humans (
      • Stolk R.P.
      • Breteler M.M.B.
      • Ott A.
      • Pols H.A.P.
      • Lamberts S.W.J.
      • Grobbee D.E.
      • Hofman A.
      ), and insulin-like growth factor-2 is significantly increased in AD CSF compared with controls (
      • Tham A.
      • Nordberg A.
      • Grissom F.E.
      • Carlsson-Skwirut C.
      • Viitanen M.
      • Sara V.R.
      ). These molecules can serve as substrates of IDE and could thus interfere competitively with the efficient degradation and clearance of Aβ in the brain during aging and in AD. In this regard, Aβ can accumulate progressively with age in the brains of normal humans as well as in lower primates, dogs, cats, and certain other mammals (
      • Selkoe D.J.
      • Bell D.
      • Podlisny M.B.
      • Cork L.C.
      • Price D.L.
      ). These observations, taken together with the evidence that IDE in brain tissue can bind and cleave Aβ (
      • Kurochkin I.V.
      • Goto S.
      ,
      • McDermott J.R.
      • Gibson A.M.
      ) and our findings that IDE is the principal protease released by microglial cells that degrades naturally secreted Aβ at physiological concentrations and is present in human brain and CSF, make it important to determine whether alterations in the activity or regulation of IDE and other Aβ-cleaving proteases could explain some of the many cases of “sporadic” AD in which Aβ accumulates excessively but its production is apparently normal.

      ACKNOWLEDGEMENTS

      We thank Dr. Richard Roth (Stanford University) for helpful suggestions and discussions and for the generous gift of monoclonal antibody 9B12.

      REFERENCES

        • Qiu W.Q.
        • Ye Z.
        • Kholodenko D.
        • Seubert P.
        • Selkoe D.J.
        J. Biol. Chem. 1997; 272: 6641-6646
        • Becker A.B.
        • Roth R.A.
        Methods Enzymol. 1995; 248: 693-703
        • Authier F.
        • Posner B.I.
        • Bergeron J.J.M.
        Clin. Invest. Med. 1996; 19: 149-160
        • Safavi A.
        • Miller B.C.
        • Cottam L.
        • Hersh L.B.
        Biochemistry. 1996; 35: 14318-14325
        • Kurochkin I.V.
        • Goto S.
        FEBS Lett. 1994; 345: 33-37
        • McDermott J.R.
        • Gibson A.M.
        Neurochem. Res. 1997; 22: 49-56
        • Kuo W.-L.
        • Gehm B.D.
        • Rosner M.R.
        • Li W.
        • Keller G.
        J. Biol. Chem. 1994; 269: 22599-22606
        • Chesneau V.
        • Perlman R.K.
        • Li W.
        • Keller G.-A.
        • Rosner M.R.
        Endocrinology. 1997; 138: 3444-3451
        • Haass C.
        • Schlossmacher M.G.
        • Hung A.Y.
        • Vigo-Pelfrey C.
        • Mellon A.
        • Ostaszewski B.L.
        • Lieberburg I.
        • Koo E.H.
        • Schenk D.
        • Teplow D.B.
        • Selkoe D.J.
        Nature. 1992; 359: 322-325
        • Qiu W.Q.
        • Borth W.
        • Ye Z.
        • Haass C.
        • Teplow D.B.
        • Selkoe D.J.
        J. Biol. Chem. 1996; 271: 8443-8451
        • Kuo W.L.
        • Montag A.G.
        • Rosner M.R.
        Endocrinology. 1993; 132: 604-611
        • Podlisny M.B.
        • Walsh D.M.
        • Amarante P.
        • Ostaszewski B.L.
        • Stimson E.R.
        • Maggio J.E.
        • Teplow D.B.
        • Selkoe D.J.
        Biochem. 1998; 37: 3602-3611
        • Garcia J.V.
        • Stoppelli M.P.
        • Decker S.J.
        • Rosner M.R.
        J. Cell Biol. 1989; 108: 177-182
        • Roth R.A.
        • Mesirow M.L.
        • Cassell D.J.
        • Yokono K.
        • Baba S.
        Diabetes Res. Clin. Pract. 1985; 1: 31-39
        • Semple J.W.
        • Lang Y.
        • Speck E.R.
        • Delovitch T.L.
        Int. Immunol. 1992; 4: 1161-1167
        • Seta K.A.
        • Roth R.A.
        Biochem. Biophys. Res. Commun. 1997; 231: 167-171
        • Authier F.
        • Bergeron J.J.
        • Ou W.J.
        • Rachubinski R.A.
        • Posner B.I.
        • Walton P.A.
        Proc. Natl. Acad. Sci. U. S. A. 1995; 92: 3859-3863
        • Podlisny M.B.
        • Ostaszewski B.
        • Haass C.H.
        • Selkoe D.J.
        Soc. Neurosci. 1995; 21: 1281
        • Selkoe D.J.
        Science. 1997; 275: 630-631
        • Paresce D.M.
        • Chung H.
        • Maxfield F.R.
        J. Biol. Chem. 1997; 272: 29390-29397
        • Gowing E.
        • Roher A.E.
        • Woods A.S.
        • Cotter R.J.
        • Chaney M.
        • Little S.P.
        • Ball M.J.
        J. Biol. Chem. 1994; 269: 10987-10990
        • Walsh D.M.
        • Williams C.H.
        • Kennedy H.E.
        • Allsop D.
        • Murphy G.
        Nature. 1994; 367: 27-28
        • Howell S.
        • Nalbantoglu J.
        • Crine P.
        Peptides. 1995; 16: 647-652
        • Marks N.
        • Berg M.J.
        • Chi L.M.
        • Choi J.
        • Durrie R.
        • Swistok J.
        • Makofske R.C.
        • Danho W.
        • Sapirstein V.S.
        Peptides. 1994; 15: 175-182
        • Van Nostrand W.E.
        • Schmaier A.H.
        • Neiditch B.R.
        • Siegel R.S.
        • Raschke W.C.
        • Sisodia S.S.
        • Wagner S.L.
        Biochim. Biophys. Acta. 1994; 1209: 165-170
        • Stentz F.B.
        • Harris H.L.
        • Kitabchi A.E.
        Endocrinology. 1985; 116: 926-934
        • Mentlein R.
        • Ludwig R.
        • Martensen I.
        J. Neurochem. 1998; 70: 721-726
        • Selkoe D.J.
        Annu. Rev. Cell Biol. 1994; 10: 373-403
        • Podlisny M.B.
        • Ostaszewski B.L.
        • Squazzo S.L.
        • Koo E.H.
        • Rydel R.E.
        • Teplow D.B.
        • Selkoe D.J.
        J. Biol. Chem. 1995; 270: 9564-9570
        • Pike C.J.
        • Overman M.J.
        • Cotman C.W.
        J. Biol. Chem. 1995; 270: 23895-23898
        • Burdick D.
        • Soreghan B.
        • Kwon M.
        • Kosmoski J.
        • Knauer M.
        • Henschen A.
        • Yates J.
        • Cotman C.
        • Glabe C.
        J. Biol. Chem. 1992; 267: 546-554
        • Jarrett J.T.
        • Lansbury Jr., P.T.
        Cell. 1993; 73: 1055-1058
        • Jarrett J.T.
        • Berger E.P.
        • Lansbury Jr., P.T.
        Biochemistry. 1993; 32: 4693-4697
        • Stolk R.P.
        • Breteler M.M.B.
        • Ott A.
        • Pols H.A.P.
        • Lamberts S.W.J.
        • Grobbee D.E.
        • Hofman A.
        Diabetes Care. 1997; 20: 792-795
        • Tham A.
        • Nordberg A.
        • Grissom F.E.
        • Carlsson-Skwirut C.
        • Viitanen M.
        • Sara V.R.
        J. Neurol. Transm. 1993; 5: 165-176
        • Selkoe D.J.
        • Bell D.
        • Podlisny M.B.
        • Cork L.C.
        • Price D.L.
        Science. 1987; 235: 873-877