Microinjection of synthetic amyloid beta-protein in monkey cerebral ...

3 downloads 0 Views 2MB Size Report
A5P 40-1. VVGGVMLGIIAGINSGVDEAFFVLKQHHVEYGSDHRFEAD. B. E e c-i. C. CSF vehicle ..... ford F, Fidani L, Giuffra L, Haynes A, Irving N, James L,. Mant R, Newton ... Saunders RD, Luttman CA, Keith PT, Little SP: Beta- amyloid proteinĀ ...
American journal of Pathology, Vol. 142, No. 1, January 1993 Copyight C) American Society for Investigatiwe Pathology

Short Communication Microin jection of Synthetic Amyloid (3-Protein in Monkey Cerebral Cortex Fails to Produce Acute Neurotoxicity

Marcia B. Podlisny,*t Diane T. Stephenson,A Matthew P. Frosch,Ā§ Dean R. Tolan,t Ivan Lieberburg,11 James A. Clemens,$ and Dennis J. Selkoe*

and that control and experimental AaP peptides produced indistinguishable effects. (Am JPathol 1993, 142:17-24)

From the Departmentts of Neurology* anid Pathologl,5

Hanrard Mledical School and CenterforNeurologic Diseases, Brigham and Women s Hospital, Boston. Massachusetts; Department of Biologv.t Boston University, Bostont, Massachusetts; tLilly, Research Laboratonies, Eli Lilly anid Company, Inidianiapolis, Intdiantapolis; anid I"Athenza Neurosciences, Souith Sani Franicisco, Californiia

The cerebral deposition of amyloid ,B protein

(A/3P) is an early pathogenetic event in Alzheimer's disease (AD). Recent studies suggest both neurotoxic and neurotrophic effects ofA 1P in vitro. Because progressive A PP deposition and surrounding neuritic dystrophy occur spontaneously in primates, we evaluated the in vivo effects of synthetic AfJP in monkey cortex. Experimental and control (reverse or substituted) peptides were stereotacticaly injected into multiple neocortical sites of adult rhesus monkeys in a vehicle of either artificial cerebrospinal fluid or acetonitrile. After 2 weeks, all injection sites were identified and characterized. A fP antibodies specificaly detected the injected Af3P,40 peptide. Serial sections stained with silver and antineurofilament protein demonstrated comparable degrees of degenerating neurons, dystrophic neurites, and axonal spheroids associated with both experimental and control peptide injections. Alz 50 staining was sparse or absent in aU sites. Similar results were obtained in an animal kiled 3 months after injection. We conclude that specific cellular changes closely resembling the pathology of Alzheimer's disease were not detected in these acute experiments,

Cerebral deposition of amyloid n-protein (A,BP) is an invariant feature of Alzheimer's disease (AD). Qualitatively similar A,BP deposition occurs in the cortex of aged humans, primates,12 and dogs,23 but not rodents. Recent evidence from numerous laboratories supports a seminal role for At3P deposition in the pathogenesis of AD. DNA sequence analyses of the 3-amyloid precursor protein (f3APP) have revealed six distinct missense mutations within or just outside the APP region in several families exhibiting autosomal dominant transmission of AD.4-9 Several laboratories have observed neurotrophic10-12 or neurotoxic12-15 effects of synthetic A,BP in vitro. Yankner et al12 reported that residues 25-35 of AfP could mediate both of these effects. Kowall et al16 described neuronal toxicity in vivo by microinjecting synthetic AP3P1 40 in a vehicle of 35% acetonitrile/0.1% trifluoroacetic acid into rat hippocampus and neocortex and concluded that AfP is directly and acutely toxic to neurons in a manner that resembles AD cytopathology. However, other laboratories have found that AI3P shows little direct neurotoxicity in vitro but can potentiate the toxicity of other agents, such as glutamate13,15 or hydrogen peroxide. 1 7 Whereas aged rodents do not spontaneously accumulate APP in brain, aged monkeys provide a Supported by National Institutes of Health grants AG07911 (LEAD Award) and AG06173 (to DJS) and by grants from Athena Neurosciences and the Foundation for Neurologic Diseases. Accepted for publication September 22, 1992. Address reprint requests to M.B. Podlisny, Ph.D., Center for Neurologic Diseases, 221 Longwood Ave., Boston, MA 02115.

17

18

Podlisny et al

AJPJanuary 1993, Vol. 142, No. 1

close animal model for ,B-amyloidogenesis. Monkey AJ3P deposits are highly similar to those in AD: both diffuse18 and compacted (neuritic) plaques1' 2 are seen, and these are recognized by antibodies to native or synthetic human A,3P.2,18 Indeed, the amino acid sequences of human and monkey A,3P and f3APP695 are identical, and immunoblot analyses suggest highly similar post-translational processing of 3APP in both species.18 Some fibrillar APP deposits in monkeys are surrounded by dystrophic neurites containing epitopes of phosphorylated neurofilament proteins,2 ubiquitin,19 and 3APP.19'20 However, paired helical filaments composed of altered tau proteins, which are present in some dystrophic neurites surrounding fibrillar APP in AD cortex, have not been detected in the aged monkey.2 18 To elucidate the possible cellular effects of Aj3P in primates, we microinjected synthetic APP and appropriate control peptides into the cortex of rhesus monkeys.

Materials and Methods Four male rhesus monkeys (Macaca mulata, 4 to 5 years old, from Charles River Laboratories, fed Purina certified primate chow 5048) were injected with synthetic peptides (0.5 nmol) in either 1 pl of 35% acetonitrile/0.1% trifluoroacetic acid, pH -2 (AN), or 2 pl of artificial cerebrospinal fluid (CSF) (150 mmol/L NaCI, 3 mmol/L KCI, 1.7 mmol/L CaCI2, 0.9 mmol/L MgCI2, pH 5, sterile filtered). Peptides appeared to be soluble under these conditions as assayed by centrifugation at 16,000g for 5 minutes. Peptides were synthesized by standard FMOC chemistry, purified by reverse-phase high-performance liquid chromatography, and confirmed by amino acid analysis, mass spectrometry, and (for APP1 40 only) sequencing. Monkeys were deeply anesthetized with isoflurane. The skull, supported in a stereotactic apparatus, was exposed and - 1 cm2 bone flaps were cut into the anterior medial, posterior medial, and posterior lateral regions overlying each cerebral convexity. The dura was incised and reflected. Experimental and control peptides were injected into opposite hemispheres. Four injections were made in each of the six burr holes at a depth of 2.2 mm using a Kopf microinjector (24 injection sites per monkey). Samples were slowly and evenly injected at a rate of 0.5 PI/min; the needle was left undisturbed for 2 minutes before removal. The dura was replaced, the burr holes were covered by bone wax and dental cement, and the scalp was sutured. After either 2 weeks or 3 months, two animals were perfused transcardially with saline followed by

periodate/paraformaldehyde/lysine (PLP) fixative. Tissue of interest was postfixed in PLP for 4 hours, embedded in paraffin, and cut into 8-pm coronal sections. Lesions were identified by hematoxylin and eosin (H&E) and by immunostaining with antibodies to Af3P1 40 (R1280, 1:1500),21 and glial fibrillary acidic protein (GFAP, 1:500).22 The extent and morphology of the lesions were initially characterized by H&E. Serial sections adjacent to the approximate center of each injection site were stained by modified Bielschowsky silver (postfixed in 10% buffered formalin for 4 hours before staining) and immunocytochemically with antibodies to A,3P1_40, phosphorylated neurofilament 200-kd protein (SMI-34, 1:5000) (Sternberger Monoclonals, Inc., Baltimore, MD), GFAP, tau protein (5E2, 1:100),23 and abnormally phosphorylated tau protein (Alz 50, 1:100).24 Positive control sections of PLP-fixed AD cortex were included in all staining runs; formalin-fixed sections of aged rhesus monkey cortex were included in some runs. Lesion width was determined using a calibrated eyepiece graticule and represents the maximal extent of cellular abnormality observed through the lesion by H&E staining at x50 magnification. Neuronal cell counts were done on H&E-stained sections at x50 magnification and were independently confirmed by two additional investigators. All neurons with visible nucleoli found within a 1-mm2 box aligned to include the injection tract were counted.

Results Aliquots (0.5 nmol) of synthetic Af3P (Figure 1A) were stereotactically microinjected into adult rhesus monkey neocortex in either 1.0 pl AN, the vehicle utilized by Yankner and colleagues,12'16 or 2.0 p1 of artificial CSF, a physiological carrier. In monkeys receiving injections in the CSF vehicle, the experimental preparation was APP1_40; controls were CSF alone, reverse Aj3P (APP40-1), or control A,3P1P40 containing 15 amino acid substitutions (CA4). In monkeys receiving injections in the AN vehicle, the experimental preparations were APP1-40 or Af3P25-35; controls were AN vehicle alone, reverse Af3P40-1 peptide, and scrambled Af3P25 35 peptide. No neurological or behavioral deficits were observed. All injection sites in monkeys killed at 2 weeks were identified on H&E-stained sections, as indicated by linear tracts of reactive astrocytes and microglial cells. No consistent differences in lesion size were detectable between experimental and control injections in either vehicle (Figure 1, B and C). Multiple sections through all lesions were then

Amyloid 3-Protein Toxicity in Monkey

19

AJPJanuary 1993, Vol. 142, No. 1

A

Synthetic APP sequences Experimental:

ApPl1-40

DAEFRHDSGYEVHHQKLVFFAEDVGSNKGAI IGLMVGGVV

ApP25-35

GSNKGAIIGLM

Control: C25-35 CA4 A5P 40-1

IXLKGNGASIG

DARFEHVSGYEVKHQKLDSYAEDVGPNKGVAIGL*VGIAG

VVGGVMLGIIAGINSGVDEAFFVLKQHHVEYGSDHRFEAD

CSF vehicle B '.5

-

E

e

-ic

.0

-

.

05

CSP alone

ASPI 40

reverse APP40-1

substituted

AAPI-40

AN vehicle C 1.5

-i

!

.0o ,

.

..

.

.

._

I

AN alone

APPI-40

revcrse

APP40- 1

AP25- 35

scrambled

APP25-35

Figure 1. Sequences of synthetic A3PPpeptides usedfor microinljectionl in motnkey cerebral cortex anid scattergrams of miaximal lesioni uidth (mnm) of injection sites in monkess killed at 2 u'ees. A: Bold letters represenit amino acid suibstituitionis in the three control peptides. represents a norleucine residuie. Abbreviations: C25-35. scrambled A(3P,5_,5; CA4, substituted AP3PI .4 B: Samples inijected in the artificial CSF vehicle (2 pl) uere: 0.5 uimol (2.5 jig) of syInthetic peptides (Af3PP ,4o, reverse Af3PP4(,, or suibstittuted AJ3PPo40). C: Samples injected in the 35% acetonlitrile0. 1% trifluioroacetic acid (AN) vehicle (1 1d) were: 0.5 uimol synthetic peptides (A(PP 40, APP40_-, A(P2%35, anid scrambled A3PP2535) All injection sites uwere idetntified. Lesionl u idth represenits the maximum uidth encompassing all celluilar chatngces suirrountdinig ani inijectioni tract, as detected by staininlg uith hematoxylin anid eosini.

examined by modified Bielschowsky silver staining. Representative pairs of experimental and control lesions were analyzed in detail by immunostaining serial sections with antibodies to synthetic Af3P1 40, phosphorylated neurofilament protein, and abnormally phosphorylated tau protein (Alz 50). Light microscopic cell counts on representative pairs of APP1 40 and APP40-1 peptide injection sites matched for lesion width and cortical depth revealed almost identical numbers of neurons within a 1-mm2 area surrounding the injection sites. Experimental and control peptide injection sites that were matched for lesion size showed highly similar cellular responses independent of the peptide injected. H&E staining revealed reactive astrocytes and activated microglia as well as eosinophilic axonal spheroids (swellings) immediately surround-

ing the injection tracts (Figure 2). The sites in the AN vehicle, whether experimental or control, tended to contain a slightly larger (Figure 1, B and C) elliptical area of necrotic tissue surrounded by activated astrocytes and microglia than was seen in the CSF sites (Figure 2). Many neurons close to the injection tracts appeared cytologically normal (Figure 2, C and D, G and H). In both the AN and CSF vehicles, anti-A13P1 40 (R1280) detected the APP1 40 peptide in all injection sites (Figure 3, A and B); this reaction was abolished by peptide absorption. Anti-Af3P1 40 detected peptide present within the tract and in the thin rim of necrotic tissue immediately lining the tract (Figure 3, A and B), but not in the surrounding cortex, suggesting that the AI3P1 40 peptide, which is probably partially aggregated at physiological pH, did not diffuse into the parenchyma. Silver staining detected some axonal spheroids and dystrophic neurites to varying degrees in the area immediately surrounding virtually all injection sites; no differences were seen between experimental and control preparations (Figure 3, C to F). Antiphosphorylated neurofilament protein (SMI-34), which labels dystrophic neurites in AD and aged monkey2,18 cortex, detected the silver-positive spheroids and neurites as well as a population of degenerating neuronal cell bodies in cortical layer 11 of some experimental and control sites (Figure 3, H and 1). The occurrence of these immunoreactive neurons correlated with lesion size and not with the nature of the peptide. Such degenerating perikarya in layer 11 were also abundantly detected by SMI-34 in occasional foci of minor traumatic injury separate from the injection sites that were induced accidentally during surgery (Figure 3J); thus they represent a nonspecific neuronal response to insult. Alz 50, which like other tau antibodies detects the dystrophic neurites of plaques in human cortex24 but rarely labels the neurites of naturally occurring APP deposits in the monkey,2'19 revealed very rare immunoreactive fine neurites associated with -50% of both the experimental and control peptide injections in both vehicles (Figure 3G), although positive neuronal staining by Alz 50 was found around a traumatically induced focal cortical microinfarct (data not shown). The anti-tau monoclonal, 5E2, which detects neuritic plaques in human23 but not monkey brain,2 was negative at all sites (data not shown). Injections of correct and scrambled APP25-35 peptides likewise produced no consistent differences in local cellular reaction after 2 weeks (Figure 1C). Immunostaining with anti-GFAP showed a similar degree of reactive astrocytosis surrounding the experimental and control lesions (not

shown).

20

Podlisny et al

AJPJanuary 1993, Vol. 142, No. 1

4

,

.. I .

4 .i,

4 .'*4

*

4

,*:.,

Figure 2. Hematoxylin and eosin staininig of monkey cortex microinjected with experimental (A/3PP 40) or control (reverse Af3P40-) peptides in either AN (A to D) or artificial CSF (E to H). A: APP,c40 peptide in AN. B: Reverse A(3P4(,I peptide in AN. C, D: Higber magnification of the louwer left portion (open arrou') of the injection sites show'n in A and B, respectively. Either peptide (0.5 nmol) produces reactive astrocytes (arrows) anid activated microglia immediately suirrounditng the inijection tract as u'ell as axonial spheroids (arrowheads). Many nezurons close to the injection tracts appear cytologically niormal (asterisks). E. A,3PP 40 peptide in CSF. F. Reverse Af3P4.,o peptide in CSF. G, H. Higher magniificationi of the lower le(ft portion (open arrow) of the injectioni sites shoun in E and F, respectively. Arrous, arrowheads, and asterisks are as in C and D. Nezuronial cell counts/mm2 uwere: A, 365; B, 346; E, 373; F, 378. Bar = 200,um (A, B, E, F) or 50 jim (C, D, G, H).

Examination of a monkey killed 3 months after neocortical injection of synthetic peptides in the CSF vehicle also failed to reveal differences between experimental and control peptides. Virtually all injection sites (15 of 17) were located by immunostaining with anti-GFAP and were characterized by highly focal (