Portal de Periódicos da CAPES

Generation of Amyloidogenic C-terminal Fragments during Rapid Axonal Transport in Vivo of β-Amyloid Precursor Protein in the Optic Nerve

1995; Elsevier BV; Volume: 270; Issue: 29 Linguagem: Inglês

10.1074/jbc.270.29.17268

1083-351X

Anil Amaratunga, Richard E. Fine,

Amyloidosis: Diagnosis, Treatment, Outcomes

The amyloid β-protein (Aβ) is a major component of extracellular deposits that are characteristic features of Alzheimer's disease. Aβ is derived from the large transmembrane β-amyloid precursor protein (βAPP). In the rabbit optic nerve/optic tract (ON), βAPP is synthesized in vivo in retinal ganglion cell perikarya, rapidly transported into the ON axons in small transport vesicles and is subsequently transferred to the axonal plasma membrane as well as to the presynaptic nerve terminals (Morin, P. J., Abraham, C. R., Amaratunga, A., Johnson, R. J., Huber, G., Sandell, J. H., and Fine, R. E.(1993) J. Neurochem. 61, 464-473). Present results indicate that there is rapid processing of βAPP in the ON to generate a 14-kDa C-terminal membrane-associated fragment that contains the Aβ sequence. By using equilibrium sucrose density gradient fractionation, this fragment, as well as non-amyloidogenic C-terminal fragments and intact βAPP, are detected in at least two classes of transport vesicles destined for the plasma membrane and the presynaptic nerve terminal. The two classes of transported vesicles are distinguished by labeling kinetics as well as by density. In contrast to the ON, only non-amyloidogenic C-terminal fragments are generated in the retina, which contains the perikarya of retinal ganglion cells and glial (Muller) cells which also produce βAPP. The amyloid β-protein (Aβ) is a major component of extracellular deposits that are characteristic features of Alzheimer's disease. Aβ is derived from the large transmembrane β-amyloid precursor protein (βAPP). In the rabbit optic nerve/optic tract (ON), βAPP is synthesized in vivo in retinal ganglion cell perikarya, rapidly transported into the ON axons in small transport vesicles and is subsequently transferred to the axonal plasma membrane as well as to the presynaptic nerve terminals (Morin, P. J., Abraham, C. R., Amaratunga, A., Johnson, R. J., Huber, G., Sandell, J. H., and Fine, R. E.(1993) J. Neurochem. 61, 464-473). Present results indicate that there is rapid processing of βAPP in the ON to generate a 14-kDa C-terminal membrane-associated fragment that contains the Aβ sequence. By using equilibrium sucrose density gradient fractionation, this fragment, as well as non-amyloidogenic C-terminal fragments and intact βAPP, are detected in at least two classes of transport vesicles destined for the plasma membrane and the presynaptic nerve terminal. The two classes of transported vesicles are distinguished by labeling kinetics as well as by density. In contrast to the ON, only non-amyloidogenic C-terminal fragments are generated in the retina, which contains the perikarya of retinal ganglion cells and glial (Muller) cells which also produce βAPP. Alzheimer's disease is characteristic of neuronal loss, neurofibrillary tangles, and senile plaques in selective brain regions. Mature senile plaques contain a central core of extracellular amyloid, the major component being the 4-kDa amyloid β-protein (Aβ)1 1The abbreviations used are: Aβamyloid β-proteinβAPPβ-amyloid precursor proteinONoptic nerve/optic tractPBSphosphate-buffered salineTricineN-[2-hydroxy-1,1-bis(hydroxymethyl)ethyl]glycine. 1The abbreviations used are: Aβamyloid β-proteinβAPPβ-amyloid precursor proteinONoptic nerve/optic tractPBSphosphate-buffered salineTricineN-[2-hydroxy-1,1-bis(hydroxymethyl)ethyl]glycine.(reviewed in 26Price D.L. Annu. Rev. Neurosci. 1986; 9: 489-512Crossref PubMed Google Scholar, 24Muller-Hill B. Beyreuther K. Annu. Rev. Biochem. 1989; 58: 287-307Crossref PubMed Scopus (190) Google Scholar and 36Sisodia S.S. Koo E.H. Beyreuther K. Unterbeck A. Price D.L. Science. 1990; 248: 492-495Crossref PubMed Scopus (745) Google Scholar, and 32Selkoe D.J. Annu. Rev. Neurosci. 1989; 12: 463-490Crossref PubMed Scopus (209) Google Scholar). Aβ is derived from the large transmembrane β-amyloid precursor protein (βAPP) which is found in many cell types. There are three major alternatively spliced transcripts of the βAPP gene that are 695, 751, and 770 amino acids long; the two larger forms contain a Kunitz-type protease inhibitor domain. A variety of post-translational modifications of βAPP have been characterized in cultured cells (reviewed in 33Selkoe D.J. Annu. Rev. Neurosci. 1994; 17: 489-517Crossref PubMed Scopus (828) Google Scholar)). amyloid β-protein β-amyloid precursor protein optic nerve/optic tract phosphate-buffered saline N-[2-hydroxy-1,1-bis(hydroxymethyl)ethyl]glycine. amyloid β-protein β-amyloid precursor protein optic nerve/optic tract phosphate-buffered saline N-[2-hydroxy-1,1-bis(hydroxymethyl)ethyl]glycine. Mature βAPP can undergo proteolytic cleavage within the Aβ domain and this results in the secretion of soluble βAPP and a 9-12-kDa non-amyloidogenic fragment which remains membrane-bound (Selkoe et al., 1988; 29Schubert D. LaCorbiere M. Saitoh T. Cole G. Proc. Natl. Acad. Sci. U. S. A. 1989; 86: 2066-2069Crossref PubMed Scopus (167) Google Scholar; 40Weidemann A. Konig G. Bunke D. Fisher P. Salbaum J.M. Masters C.L. Beyreuther K. Cell. 1989; 57: 115-126Abstract Full Text PDF PubMed Scopus (1037) Google Scholar; 8Esch F.S. Keim P. Beattie E. Blacher R. Culwell A. Oltersdorf T. McClure D. Ward P. Science. 1990; 248: 1122-1124Crossref PubMed Scopus (1202) Google Scholar; 19Koo E.H. Sisodia S.S. Archer D.A. Martin L.J. Weidemann A. Beyreuther K. Fisher P. Masters C.L. Price D.L. Proc. Natl. Acad. Sci. U. S. A. 1990; 87: 1561-1565Crossref PubMed Scopus (635) Google Scholar). βAPP is also processed and degraded within an endosome/lysosome pathway, following reinternalization of full-length βAPP from the cell surface via clathrin-coated vesicles (35Shoji M. Golde T.E. Ghiso J. Cheung T.T. Estus S. Shaffer L.M. Cai X.-D. McKay D.M. Tintner R. Frangione B. Younkin S.G. Science. 1992; 258: 126-129Crossref PubMed Scopus (1320) Google Scholar; 12Haass C. Koo E. Mellon A. Hung A. Selkoe D. Nature. 1992; 357: 500-503Crossref PubMed Scopus (770) Google Scholar). C-terminal fragments containing the complete Aβ sequence of βAPP (of ∼11 kDa and larger) have been detected in endosomes/lysosomes in cultured cells and brain tissue (25Nordstedt C. Gandy S.E. Alafuzoff I. Caporaso G.L. Iverfeldt K. Grebb J.A. Winblad B. Greengard P. Proc. Natl. Acad. Sci. U. S. A. 1991; 88: 8910-8914Crossref PubMed Scopus (99) Google Scholar; 9Estus S. Golde T.E. Kunishita T. Blades D. Lowery D. Eisen M. Usiak M. Qu X. Tabira T. Greenberg B.D. Younkin S.G. Science. 1992; 255: 726-728Crossref PubMed Scopus (345) Google Scholar; 39Tamaoka A. Kalaria R.N. Lieberburg I. Selkoe D.J. Proc. Natl. Acad. Sci. U. S. A. 1992; 89: 1345-1349Crossref PubMed Scopus (141) Google Scholar) and could thus serve as potential intermediates for Aβ formation. Both neurons and astrocytes in cell culture also generate and secrete soluble Aβ (13Haass C. Schlossmacher M. Hung A. Vigo-Pelfrey C. Mellon A. Ostaszewski B. Lieberburg I. Koo E. Schenk D. Teplow D. Selkoe D. Nature. 1992; 359: 322-325Crossref PubMed Scopus (1757) Google Scholar; 34Seubert P. Vigo-Pelfrey C. Esch F. Lee M. Dovey H. Davis D. Sinha S. Schlossmacher M. Whaley J. Swindlehurst C. McCormack R. Wolfert R. Selkoe D. Lieberburg I. Schenk D. Nature. 1992; 359: 325-327Crossref PubMed Scopus (1591) Google Scholar; 35Shoji M. Golde T.E. Ghiso J. Cheung T.T. Estus S. Shaffer L.M. Cai X.-D. McKay D.M. Tintner R. Frangione B. Younkin S.G. Science. 1992; 258: 126-129Crossref PubMed Scopus (1320) Google Scholar), which is also found in the cerebral spinal fluid of normal individuals (34Seubert P. Vigo-Pelfrey C. Esch F. Lee M. Dovey H. Davis D. Sinha S. Schlossmacher M. Whaley J. Swindlehurst C. McCormack R. Wolfert R. Selkoe D. Lieberburg I. Schenk D. Nature. 1992; 359: 325-327Crossref PubMed Scopus (1591) Google Scholar). Anterograde axonal transport of neuronal βAPP in the central nervous system is likely to be a pathway by which Aβ is released to extracellular regions of brain, following its generation from βAPP. βAPP695 was shown to be the predominant form that is expressed in rat sensory ganglia and transported along the sciatic nerve (37Sisodia S.S. Koo E.H. Hoffman P.N. Perry G. Price D.L. J. Neurosci. 1993; 13: 3136-3142Crossref PubMed Google Scholar; 37Sisodia S.S. Koo E.H. Hoffman P.N. Perry G. Price D.L. J. Neurosci. 1993; 13: 3136-3142Crossref PubMed Google Scholar). Employing a central nervous system neuron of rabbit, we demonstrated that βAPP695 as well as βAPP751/770 are synthesized in the retina; βAPP695 is rapidly transported in vivo into the optic nerve/optic tract (ON) in small vesicles and is subsequently transferred to a membrane fraction containing glucose transporters and other axonal plasma membrane markers, as well as to the presynaptic nerve terminals in the lateral geniculate (22Morin P.J. Abraham C.R. Amaratunga A. Johnson R.J. Huber G. Sandell J.H. Fine R.E. J. Neurochem. 1993; 61: 464-473Crossref PubMed Scopus (93) Google Scholar). Axonal transport of βAPP was also reported in the developing hamster ON where βAPP751/770 was the prominent form(s) detected in the lateral geniculate (23Moya K.L. Confaloni A. Allinquant B. J. Neurochem. 1994; 63: 1971-1974Crossref PubMed Scopus (15) Google Scholar). Anterograde axonal transport of βAPP may occur via a kinesin based motor, as shown in cultured neurons (10Ferreira A. Caceres A. Kosik K.S. J. Neurosci. 1993; 13: 3112-3123Crossref PubMed Google Scholar) and in the ON in vivo (2Amaratunga A. Leeman S.E. Kosik K.S. Fine R.E. J. Neurochem. 1995; 64: 2374-2376Crossref PubMed Scopus (38) Google Scholar). Following transport to the plasma membrane, βAPP is degraded with a t of less than 5 h. Protease activity that can potentially generate amyloidogenic βAPP-derived peptides is also present in the same ON membrane fractions as βAPP (22Morin P.J. Abraham C.R. Amaratunga A. Johnson R.J. Huber G. Sandell J.H. Fine R.E. J. Neurochem. 1993; 61: 464-473Crossref PubMed Scopus (93) Google Scholar). In order to examine the generation of potentially amyloidogenic fragments of βAPP in the ON in vivo, we carried out a series of studies employing gel electrophoresis conditions capable of detecting small molecular weight fragments. In this report, we describe results indicating that there is rapid processing of βAPP695 in the ON to generate a C-terminal membrane-associated fragment that contains the Aβ sequence. This fragment, as well as non-amyloidogenic fragments and intact βAPP, are seen in regions of an equilibrium sucrose density gradient that contain axonal plasma membrane and at least two classes of transport vesicles, destined for the plasma membrane and axon terminal, respectively. This finding suggests that there is significant processing of βAPP to an amyloidogenic C-terminal-containing species before it reaches the axonal plasma membrane and the nerve terminal. We also report that in the retina, where βAPP751 and βAPP770 are also present, βAPP is processed differently from that which occurs in the ON, generating only non-amyloidogenic C-terminal fragments. Adult male albino rabbits (6 lb) were anesthetized with 1 ml of 5% sodium pentobarbital (intravenous), and two drops of 0.5% proparacaine HCL (topical). U-100 insulin syringes with 28-gauge needles are used to introduce 0.5 mCi (50 μl) of [35S]methionine/cysteine (DuPont NEN) into the vitreous chamber of each eye. All animal use procedures were in strict accordance with the National Institutes of Health Guide for the Care and Use of Laboratory Animals and were approved by the local Animal Care Committee. Animals were sacrificed 3, 6, 9, or 12 h following radiolabel injection and retinas and ONs were dissected out. Tissues were homogenized in ice-cold phosphate-buffered saline (PBS) (containing 30 μg/ml phenylmethylsulfonyl fluoride, 0.5 μg/ml leupeptin, and 0.5 μg/ml aprotinin) in a motor-driven Teflon glass homogenizer and centrifuged at 45,000 rpm for 1 h in a Ti 70 rotor. Pellets were resuspended in PBS and diluted 1:1 by adding 2 × immunomix (1.0% (v/v) Triton X-100, 0.5% (v/v) deoxycholate, and 0.1% (w/v) sodium dodecyl sulfate (SDS) in PBS) for solubilization, followed by centrifuging at 25,000 rpm in a Ti 70 rotor for 10 min for clarification. The resulting solubilized membrane protein preparations were used for immune precipitations. Three- or 6-h post-35S-labeled ONs were dissected out (described above). Density gradient separations were carried out as described earlier (20Morin P.J. Liu N. Johnson R.J. Leeman S.E. Fine R.E. J. Neurochem. 1991; 56: 415-427Crossref PubMed Scopus (31) Google Scholar, 22Morin P.J. Abraham C.R. Amaratunga A. Johnson R.J. Huber G. Sandell J.H. Fine R.E. J. Neurochem. 1993; 61: 464-473Crossref PubMed Scopus (93) Google Scholar). Briefly, ONs were homogenized in 7 ml of ice-cold homogenization (H) buffer (1 mM triethanolamine, 320 mM sucrose), containing 30 μg/ml phenylmethylsulfonyl fluoride, 0.5 μg/ml leupeptin, and 0.5 μg/ml aprotinin). Homogenates were diluted to 40 ml with H buffer and centrifuged at 1,200 × g for 7 min. The resulting supernatant was then centrifuged at 100,000 × g for 60 min. The resulting pellet was resuspended in 2 ml of H buffer and loaded onto a discontinuous gradient consisting of 2 ml each of 20/26/31/37/45% sucrose (w/w) in H buffer. These gradients were centrifuged at 150,000 × g in a SW 40 rotor (Beckman) for 16 h. Twenty-four 0.5-ml fractions were collected from each gradient, and each fraction was assayed for protein and radioactivity to determine specific activity. Fractions containing the highest specific activities were pooled, diluted to 5 ml with PBS, and pelleted at 150,000 × g in an SW 50.1 rotor for 60 min. The pellets were resuspended in PBS, solubilized with immunomix and used for immune precipitations. Antibodies were added at appropriate dilutions to aliquots of equal volume and incubated overnight at 4°C (22Morin P.J. Abraham C.R. Amaratunga A. Johnson R.J. Huber G. Sandell J.H. Fine R.E. J. Neurochem. 1993; 61: 464-473Crossref PubMed Scopus (93) Google Scholar). Sepharose-linked anti-rabbit or anti-mouse immunoglobulin G beads (Organon Teknika) were then added to samples (15 μl of beads/1 μl of primary antibody) and incubated for 3 h at 4°C with agitation. Beads were separated in an Eppendorf centrifuge and washed five times using immunomix. Washed beads were suspended in SDS sample buffer and analyzed by 10-20% Tris-Tricine polyacrylamide gel electrophoresis. Gels were incubated in an enhancing solution for 1 h, dried under vacuum for 2 h, and exposed to films. Rabbit sera made against a synthetic peptide corresponding to the C-terminal 20 amino acids of all isoforms of APP, C8 (31Selkoe D.J. Podlisny M.B. Joachim C.L. Vickers E.A. Lee G. Fritz L.C. Oltersdorf T. Proc. Natl. Acad. Sci. U. S. A. 1988; 85: 7341-7345Crossref PubMed Scopus (539) Google Scholar), and corresponding to 1-40 amino acids of Aβ, R1280 (39Tamaoka A. Kalaria R.N. Lieberburg I. Selkoe D.J. Proc. Natl. Acad. Sci. U. S. A. 1992; 89: 1345-1349Crossref PubMed Scopus (141) Google Scholar) (gifts of Dr. D. Selkoe), were used at dilutions of 1:1000 and 1:200, respectively. R1280 recognizes Aβ (4 kDa) peptide as well as a 3-kDa peptide starting at residue 16 of the Aβ near the secretory cleavage site of βAPP that ends near residue 40 (13Haass C. Schlossmacher M. Hung A. Vigo-Pelfrey C. Mellon A. Ostaszewski B. Lieberburg I. Koo E. Schenk D. Teplow D. Selkoe D. Nature. 1992; 359: 322-325Crossref PubMed Scopus (1757) Google Scholar). A monoclonal antibody specific to the N-terminal 17 amino acids of Aβ, 6E10 (18Kim K.S. Miller D.L. Sapienze V.J. Chang C.J. Grundke-Iqbal I. Currue J.R. Wisniewski H.M. Neurosci. Res. Commun. 1988; 2: 121-130Google Scholar), was used at a dilution of 1:300. Control experiments to determine nonspecific immune precipitations were carried out by using normal rabbit serum (Rockland Co.) and a monoclonal antibody against insulin-regulatable glucose transporter, 1F8, which is not found in brain (15James D.E. Lederman L. Pilch P.F. J. Biol. Chem. 1987; 262: 11817-11824Abstract Full Text PDF PubMed Google Scholar). Scanning of autoradiograms was performed by volume integration of protein bands using a densitometer (Molecular Dynamics). Direct determination of radioactive counts in dried gels was performed by using an instant imager (Packard). We have utilized the retinal ganglion cell/ON preparation combined with 10-20% Tris-Tricine gradient gels to detect Aβ containing βAPP-derived fragments in vivo of the [35S]methionine/cysteine-labeled ON. ON membranes were taken at various times after 35S injection, and aliquots were subjected to immune precipitation with either a C-terminal-recognizing antibody, C8, or with a polyclonal antibody against Aβ, R1280. As early as 3 h after 35S injection, a fragment with an apparent molecular mass of 14 kDa is specifically precipitated with both antibodies, whereas several other fragments are only detected with C8 (Fig. 1). The 14-kDa fragment precipitated by both antibodies is maximally labeled at 3 and 6 h following 35S injection and then disappears gradually at 9 and 12 h after injection. Of the C8-precipitated proteins, fragments of ∼12-16 kDa are detected at 3 h after 35S injection, and smaller fragments of ∼10-14 kDa are detected at 6 h, which also gradually disappear at later time points (Fig. 1). We also observe the full-length βAPP695 (molecular mass of ∼110 kDa) at all these times but a considerably lesser amount at 12 h. It is expected that the cleaved N-terminal fragments are present in the soluble fraction of our ON preparation and therefore should only be recognized by antibodies against the βAPP N terminus. Consistent with this expectation, we have not observed any C-terminal fragments released into the soluble fraction of ON by using C8 and R1280 (data not shown). We have previously demonstrated that in the rabbit ON in vivo, there are at least three types of anterograde rapidly transported vesicles, two types of synaptic vesicle precursors, one for classic transmitters (small, light vesicles) and one for peptides (large, dense vesicles), and a small, light vesicle that carries plasma membrane proteins to the axonal and presynaptic nerve terminal plasma membrane (20Morin P.J. Liu N. Johnson R.J. Leeman S.E. Fine R.E. J. Neurochem. 1991; 56: 415-427Crossref PubMed Scopus (31) Google Scholar). In the rabbit ON in vivo, βAPP695 was shown to be anterogradely axonally transported in small vesicles of the plasma membrane type (22Morin P.J. Abraham C.R. Amaratunga A. Johnson R.J. Huber G. Sandell J.H. Fine R.E. J. Neurochem. 1993; 61: 464-473Crossref PubMed Scopus (93) Google Scholar). We have subfractionated the 35S-labeled membranes into the three major membrane containing peaks as described earlier (20Morin P.J. Liu N. Johnson R.J. Leeman S.E. Fine R.E. J. Neurochem. 1991; 56: 415-427Crossref PubMed Scopus (31) Google Scholar, 1Amaratunga A. Morin P.J. Kosik K.S. Fine R.E. J. Biol. Chem. 1993; 268: 17427-17430Abstract Full Text PDF PubMed Google Scholar). At 3-h post-injected ON, we detect 10-14-kDa fragments in peaks 1 (light fraction) and 2 (intermediate fraction) in the gradient, which correspond to the plasma membrane, and small transport vesicles destined for the plasma membrane and the nerve terminal, respectively (Fig. 2). According to previous results, peak 3 (dense fraction) shows delayed transport kinetics and contains substance P/K containing dense transport vesicles destined for the nerve terminal (21Morin P.J. Johnson R.J. Shachar I. Leeman S.E. Fine R.E. Ann. N. Y. Acad. Sci. 1991; 632: 442-443Crossref PubMed Scopus (2) Google Scholar, 23Moya K.L. Confaloni A. Allinquant B. J. Neurochem. 1994; 63: 1971-1974Crossref PubMed Scopus (15) Google Scholar). In 6-h post-injected ON, in addition to peaks 1 and 2, peak 3 also contains C-terminal fragments (Fig. 2). The full-length βAPP695 is present in all three peaks at 3 h as well as at 6 h. We also carried out immune precipitation of βAPP and its C-terminal fragments from the 35S-labeled retina. 35S-Labeled amino acids injected into the vitreous of the eye can be utilized by retinal ganglion cells as well as by Muller glial cells, which send their processes to the ganglion cell layer (23Moya K.L. Confaloni A. Allinquant B. J. Neurochem. 1994; 63: 1971-1974Crossref PubMed Scopus (15) Google Scholar). Previous results indicate that the synthesis of βAPP695 (molecular mass, ∼110 kDa) as well as βAPP isoforms containing the Kunitz-type protease inhibitor domain occurs in the retina (23Moya K.L. Confaloni A. Allinquant B. J. Neurochem. 1994; 63: 1971-1974Crossref PubMed Scopus (15) Google Scholar). A sizable amount of a ∼12-kDa C-terminal fragment is produced in the membrane fraction of retina at 3 h (Fig. 3) and also at 6 h (data not shown) following 35S injection. This fragment does not appear to be amyloidogenic as indicated by the inability of R1280 to precipitate this component (Fig. 3). We examined the possibility that the recognition of C-terminal fragments by the antibody C8 may be due to the metabolism of APPLP2, a recently cloned βAPP-like protein. APPLP2 is highly homologous to βAPP at the C terminus, but does not contain the extracellular Aβ sequence (38Slunt H.H. Thinakaran G. Van Koch G. Lo A.C.Y. Tanzi R.E. Sisodia S.S. J. Biol. Chem. 1994; 269: 2637-2644Abstract Full Text PDF PubMed Google Scholar). When a monoclonal antibody specific for the N-terminal 17 amino acids of Aβ, 6E10, was used in immune precipitations, a 14-kDa fragment was detected only in the ON and not in the retina (Fig. 4). This confirms that the 14-kDa fragment is derived from βAPP as well as containing the entire Aβ sequence. We compared the C8 and 6E10 immune-precipitated protein bands shown in Fig. 4 by quantitating radioactive counts in the gel of the 110-kDa band and by densitometric analysis of the 14-kDa band. With respect to the 110-kDa species, 6E10 immune-precipitated 86% of that precipitated by C8 (Fig. 4). Although at present no data are available on the expression of APPLP2 in the rabbit retinal ganglion cell, this result may suggest that βAPP is the predominant APP form that is axonally transported in the ON. With respect to the 14-kDa fragment, 6E10 precipitated 77% of that precipitated by C8 (Fig. 4), thereby indicating that this fragment is produced predominantly by the cleavage of βAPP to generate a fragment containing the entire Aβ sequence. The results presented have demonstrated that there is a rapid (<3 h) metabolism of the retinal ganglion cell produced βAPP in the ON to generate a C-terminal 14-kDa fragment containing the Aβ sequence. This fragment is a likely precursor to Aβ (7Dyrks T. Dyrks E. Monning U. Urmoneit B. Turner J. Beyreuther K. FEBS Lett. 1993; 335: 89-93Crossref PubMed Scopus (95) Google Scholar), although we have no evidence as yet that Aβ is produced in this system in vivo. It is possible that in this system the incorporation of 35S into Aβ, which has only one methionine residue and no cysteine, might be insufficient for detecting Aβ. The cell type which produces Aβ in the central nervous system is not known, and it is also possible that the retinal ganglion cells do not produce Aβ. However, the amyloidogenic fragment(s) is metabolized fairly rapidly, since it disappears by 12 h after labeling (Fig. 1). This is to be expected, since there is evidence that overproduction of a C-terminal Aβ-containing fragment in transfected cells can produce neurotoxicity when these cells are differentiated into neuron-like cells (17Kammesheidt A. Boyce F.M. Spanogannis A.F. Cummings B.J. Ortegon M. Cotman C. Vaught J.L. Neve R.L. Proc. Natl. Acad. Sci. U. S. A. 1992; 89: 10857-10861Crossref PubMed Scopus (140) Google Scholar). The 12-kDa C-terminal fragment recognized only by C8 may be produced by an α-secretase that cleaves within the Aβ sequence to generate a C-terminal fragment containing the transmembrane portion and cytoplasmic domain (31Selkoe D.J. Podlisny M.B. Joachim C.L. Vickers E.A. Lee G. Fritz L.C. Oltersdorf T. Proc. Natl. Acad. Sci. U. S. A. 1988; 85: 7341-7345Crossref PubMed Scopus (539) Google Scholar). It is possible that in this system R1280 precipitates only the 14-kDa amyloidogenic C-terminal fragment. The nondetection of a 12-kDa fragment by R1280 may be due to the fact that R1280 appears to be a weak immune-precipitating antibody consistent with peptide antibodies against the Aβ region of βAPP (39Tamaoka A. Kalaria R.N. Lieberburg I. Selkoe D.J. Proc. Natl. Acad. Sci. U. S. A. 1992; 89: 1345-1349Crossref PubMed Scopus (141) Google Scholar), compared with C8 and 6E10. When simultaneously used in equal aliquots of an ON membrane preparation, R1280 precipitated only 14% of the 110-kDa species precipitated by C8 (Fig. 1). Since over 80% of the C8-precipitated 110-kDa species is βAPP (Fig. 4), this shows that R1280 precipitates only a minor fraction of intact βAPP. Therefore, it is possible that the detection of the 12-kDa fragment by R1280 was beyond the limit of sensitivity. We also simultaneously detect other C-terminal fragments that are not recognized by R1280 (Fig. 1) and 6E10 (Fig. 4). The cleavage of APPLP2 could generate these fragments, which are recognized by C8 (38Slunt H.H. Thinakaran G. Van Koch G. Lo A.C.Y. Tanzi R.E. Sisodia S.S. J. Biol. Chem. 1994; 269: 2637-2644Abstract Full Text PDF PubMed Google Scholar). If it is assumed that both βAPP and APPLP2 are expressed in the rabbit retinal ganglion cell, then APPLP2 represents less than one-fifth of the total that is axonally transported (Fig. 4). Therefore, it is unlikely that the C-terminal fragments recognized exclusively by C8 are generated entirely from APPLP2. Using an equilibrium sucrose gradient, we were able to fractionate βAPP-derived C-terminal fragments in the ON into three separate membrane compartments (Fig. 2) (22Morin P.J. Abraham C.R. Amaratunga A. Johnson R.J. Huber G. Sandell J.H. Fine R.E. J. Neurochem. 1993; 61: 464-473Crossref PubMed Scopus (93) Google Scholar). At 3 h following 35S injection, some βAPP has already reached the axonal plasma membrane (peak 1) and C-terminal fragments are also detected in this fraction (Fig. 2). The observation of C-terminal fragments (and βAPP) in peak 2 indicates that βAPP processing to generate C-terminal fragments occurs within transport vesicles bound for the axonal plasma membrane. The fact that protease activity specific for the generation of amyloidogenic fragments from βAPP has been co-localized in peak 2 vesicles (22Morin P.J. Abraham C.R. Amaratunga A. Johnson R.J. Huber G. Sandell J.H. Fine R.E. J. Neurochem. 1993; 61: 464-473Crossref PubMed Scopus (93) Google Scholar) further supports this possibility. Also in PC12 cells, βAPP cleavage was shown to occur intracellularly following glycosylation and sulfation within the vesicles that transport proteins to the plasma membrane (27Sambamurti K. Shioi J. Anderson J.P. Pappolla M.A. Robakis N.K. J. Neurosci. Res. 1992; 33: 319-329Crossref PubMed Scopus (134) Google Scholar). It is likely that βAPP-containing vesicles bound for the plasma membrane (peak 2) consist of a subpopulation of small, light vesicles, as was shown to be the case for glucose transporter-containing vesicles (Morin et al., 1991a). Also, immune staining experiments did not co-localize βAPP and synaptic vesicle markers in the same vesicles (Schubert et al., 1991; Ferreira et al., 1993; Caporaso et al., 1994) It is possible that fragments observed in peaks 1 and 2 of 3-h post-labeled ON (Fig. 2) could be derived from internalization of βAPP from the plasma membrane into an endosome-lysosome-like compartment (12Haass C. Koo E. Mellon A. Hung A. Selkoe D. Nature. 1992; 357: 500-503Crossref PubMed Scopus (770) Google Scholar; 11Golde T.E. Estus S. Younkin L. Selkoe D. Younkin S. Science. 1992; 255: 728-730Crossref PubMed Scopus (623) Google Scholar). This is supported by the fact that intact βAPP is detected in the plasma membrane fraction (peak 1) (Fig. 2). Immune staining in cultured neurons indicated βAPP to be associated with clathrin-coated vesicles and endosome-like vesicles, some of which may be involved in retrograde transport (10Ferreira A. Caceres A. Kosik K.S. J. Neurosci. 1993; 13: 3112-3123Crossref PubMed Google Scholar; 6Caporaso G.L. Takei K. Gandy S.E. Matteoli M. Mundigl O. de Camilli P. J. Neurosci. 1994; 14: 3122-3128Crossref PubMed Google Scholar). Also, βAPP has been localized within clathrin-coated vesicles in mammalian brain (28Sapirstein V.S. Durrie R. Berg M.J. Marks N. J. Neurosci. Res. 1994; 37: 348-355Crossref PubMed Scopus (23) Google Scholar). However, clathrin-coated vesicles are much too dense to be present in this region of the 20-45% gradient used in these experiments. In cultured cells, amyloidogenic fragments as well as Aβ appear to be generated within an acidic compartment in the secretory pathway (35Shoji M. Golde T.E. Ghiso J. Cheung T.T. Estus S. Shaffer L.M. Cai X.-D. McKay D.M. Tintner R. Frangione B. Younkin S.G. Science. 1992; 258: 126-129Crossref PubMed Scopus (1320) Google Scholar; 14Haass C. Hung A.Y. Schlossmacher M. Teplow D. Selkoe D. J. Biol. Chem. 1993; 268: 3021-3024Abstract Full Text PDF PubMed Google Scholar). Axonal endosomes appear not to be acidified until reaching the cell body, making it unlikely that proteolysis in endosomes occurs to any great extent in the axon (4Augenbraun E. Maxfield F.R. Robert S.J. Setlik W. Holtzman E. Eur. J. Cell Biol. 1993; 61: 34-43PubMed Google Scholar). It is, however, likely that the axonal secretory vesicles are acidified as are other secretory vesicles (3Anderson R. Pathak R.K. Cell. 1985; 40: 635-643Abstract Full Text PDF PubMed Scopus (218) Google Scholar). In 6-h post-labeled ON, C-terminal fragments are detected in both peaks 2 and 3 (Fig. 2), thereby indicating intracellular βAPP processing in two classes of vesicles of different densities bound for the nerve terminal. The peak 2 fraction at 6 h contains transport vesicles carrying plasma membrane proteins as well as small synaptic vesicle precursors (22Morin P.J. Abraham C.R. Amaratunga A. Johnson R.J. Huber G. Sandell J.H. Fine R.E. J. Neurochem. 1993; 61: 464-473Crossref PubMed Scopus (93) Google Scholar). The peak 3 fraction at 6 h contains dense neuropeptide (substance P/K) containing vesicles (20Morin P.J. Liu N. Johnson R.J. Leeman S.E. Fine R.E. J. Neurochem. 1991; 56: 415-427Crossref PubMed Scopus (31) Google Scholar, 21Morin P.J. Johnson R.J. Shachar I. Leeman S.E. Fine R.E. Ann. N. Y. Acad. Sci. 1991; 632: 442-443Crossref PubMed Scopus (2) Google Scholar). The possibility that βAPP and substance P/K are co-localized in these dense vesicles is suggested by immune localization results (30Schubert W. Prior R. Weidemann A. Dircksen H. Multhaup G. Masters C.L. Beyreuther K. Brain Res. 1991; 563: 184-194Crossref PubMed Scopus (237) Google Scholar), in which case they would be bound for the nerve terminal. It is also possible that βAPP is localized in a separate class of dense vesicle which is bound for the plasma membrane. In conclusion with regard to βAPP processing in the transport vesicles, the data strongly support the possibility that axonally transported βAPP695 is processed intracellularly to generate C-terminal fragments in at least two classes of vesicle, which are distinguished by labeling kinetics as well as by density. These vesicles are bound for both axonal plasma membrane and nerve terminal. In contrast with the ON, in which both amyloidogenic and non- amyloidogenic C-terminal fragments are produced, we only detect the production of a non-amyloidogenic C-terminal fragment in the retina (Figure 3:, Figure 4:). The two cell types in the retina which produce βAPP are the retinal ganglion cells that produce βAPP695 (and rapidly axonally transport it) and the Muller glial cells that produce the Kunitz-type protease inhibitor-containing βAPP isoforms, βAPP751 and βAPP770, as well (22Morin P.J. Abraham C.R. Amaratunga A. Johnson R.J. Huber G. Sandell J.H. Fine R.E. J. Neurochem. 1993; 61: 464-473Crossref PubMed Scopus (93) Google Scholar). Also, cultured rabbit Muller cells synthesize 120- and 140-kDa species of βAPP, corresponding to βAPP751 and βAPP770, respectively.2 2A. Amaratunga, C. R. Abraham, R. B. Edwards, J. H. Sandell, B. M. Shreiber and R. E. Fine, submitted for publication. Because of the rapid clearance of neuronally produced βAPP into the ON (22Morin P.J. Abraham C.R. Amaratunga A. Johnson R.J. Huber G. Sandell J.H. Fine R.E. J. Neurochem. 1993; 61: 464-473Crossref PubMed Scopus (93) Google Scholar), it is likely that the retinal non-amyloidogenic C-terminal metabolite is produced in the Muller cell. Therefore, this result may indicate differential processing of βAPP with respect to the central nervous system cell type and/or to the βAPP isoform. This result also suggests that in the normal central nervous system, only neurons are capable of generating amyloidogenic fragments. This is in contrast to the reported results that cultured astrocytes, which are similar in certain ways to Muller cells, produce more Aβ than cultured neurons (5Busciglio J. Gabuzda D.H. Matsudaira P. Yankner B.A. Proc. Natl. Acad. Sci. U. S. A. 1993; 90: 2092-2096Crossref PubMed Scopus (524) Google Scholar). The above results may help to elucidate the normal processing of βAPP in vivo in a central nervous system neuron. In contrast to rat and mouse, the Aβ sequence of the rabbit is identical to that of human (16Johnstone E. Chuney M. Norris F. Pascual R. Little S. Mol. Brain. Res. 1991; 10: 299-306Crossref PubMed Scopus (296) Google Scholar), making it a more appropriate system to study βAPP processing. We thank Dr. Peter Morin, Dr. Robin Johnson, and Hae Yung Pyun for their invaluable advice and assistance.