Americanjournal oJ Palhtologt. lol. 142. No. 6,jitimz 199_3CGoptright c Americanz ScLietitJfr InvestigattieP'athologt

f31A4 Proteinlike Immunoreactive GranularStructures in the Brain ofSenescence-Accelerated Mouse

Manabu Takemura,* Shinichi Nakamura,*Ichiro Akiguchi,* Masaki Ueno,*Nobuyuki Oka,* Shintaro Ishikawa,tAtsuyoshi Shimada,t Jun Kimura,* andToshio TakedatFrom the Department of NVe1roloqgt. Faculty ofMeIuIedicin'e*and Department (o/Senescence Biology. Chest DiseaseResearch Institute.t Kyoto l rniversity, K1 oto, japan

The immunohistochemical localization of amy-loid 3/A4 protein in the senescence-acceleratedmouse brain was studied using six different an-tisera against human amyloidprecursorproteinpeptides. f4/A4 proteinlike immunoreactivity wasobserved in theform ofgranular structures (,¢LIGS) in various regions, including the medialseptum, cerebral cortex, hippocampus, cerebel-lunm and some cranial nerve roots. 13-LIGS were1.5 to 2.5

.in diameter and irregularly shaped.

They increased significantly in number with ag-ing predominantly in animals with a phenotypeofage-related deterioration ofmemory and learn-ing abilities. Congo red and thioflavine S did notstain the granules. On immunoblots, the main im-munoreactive bands were observed at 14 to 18 kd.The staining intensities of these bands also in-creased with advancing age. We consider that,¢LIGS are not only a new morphological mani-festation of senescence in mice, but also a perti-nent clue in understanding the mechanisms ofamyloid deposition. (Am J Pathol 1993, 142:1887-1897)

Amyloid ,B/A4 protein, a 4.2-kd polypeptide,' 2 is theintegral component of senile plaques and cere-brovascular amyloid fibrils in individuals with Alzhe-imer's disease,3-5 Down's syndrome,6 and normal ag-ing.78 Amyloid 3/A4 protein seems to originate fromamyloid precursor protein (APP), which has severalspecies generated by alternative splicing.9-14 Amy-loid ,B/A4 protein deposition seen in subjects withAlzheimer's disease has been found only in higher

mammalians such as primates, dogs, and polarbears.15-22 Recently, two groups reported that trans-genic mice and transplants of trisomy 16 mice couldserve as experimental models for amyloid P/A4 pro-tein deposition in the brain.23'24

The senescence-accelerated mouse (SAM), a mu-rine model of accelerated senescence, has been es-tablished by Takeda and colleagues.25'26 There arenow eight accelerated senescence prone lines(SAM-P) and three accelerated senescence resistantlines (SAM-R), the latter with normal characteristics ofaging. Each line has a relatively strain-specific patho-logical phenotype such as systemic senile amyloid-osis,27-30 senile cataract,31 senile osteoporosis,32 ordegenerative joint disease.33 SAM-P/8 shows earlyonset and rapid advancement of senescence re-vealed by analysis of aging dynamics such as Go-mpertz function, survivorship curve, and gradingscore system34 and exhibits a significant age-relateddeterioration of memory and learning abilities.35 36

To our knowledge, there has been no report of aspontaneous development of 3/A4 immunoreactivityin the mouse brain. In the present study, we examinedimmunohistochemically and immunochemically theSAM-P/8 brain and the SAM-R/1 brain, using six dif-ferent antisera against synthetic or recombinant hu-man APP peptides. We present here evidence indi-cating that ,B/A4-like immunoreactive structures occurin the SAM brain and show a marked age-related in-crease.

Materials and MethodsAnimals

We examined five young (2-month-old), five middle-aged (9-month-old), and four old (1 2-month-old)SAM-P/8 and SAM-R/1 male mice, respectively, and

Accepted for publication January 11, 1993.

Address reprint requests to Dr. Toshio Takeda, Department ofSenescence Biology, Chest Disease Research Institute, Kyoto Uni-versity, 53 Shogoin Kawahara-cho, Sakyo-ku, Kyoto 606, Japan.

1887

1888 Takemura et alA/P.June 1993, lol. 142, Ao. 6

two very old (22-month-old) SAM-R/1 male mice.The mice were reared in our laboratory under con-ventional conditions at constant temperature (24 +2 C) and were maintained on a 7 A.M. to 7 P.M. light/7P.M. to 7 A.M. dark schedule and on a commercialdiet (CE-2, Nihon CLEA) and tap water ad libitum.

Tissues

The mice were deeply anesthesized with an intra-peritoneal injection of 0.07 ml of 5% sodium pento-barbital and perfused transcardially with 50 ml ice-cold 0.01 mol/L phosphate-buffered saline (PBS, pH7.4), followed by a fixative of 150 ml of 4%paraformaldehyde and 0.2% picric acid in 0.1 mol/Lphosphate buffer (pH 7.4). The brains were re-moved immediately and soaked in a postfixativeconsisting of 4% paraformaldehyde and 0.2% picricacid in 0.1 mol/L phosphate buffer (pH 7.4). Follow-ing cryoprotection with 20% sucrose in 0.1 mol/Lphosphate buffer (pH 7.4) overnight at 4 C, blockswere frozen and cut serially into 15-p-thick sectionson a freezing microtome.

Immunohistochemistry

The immunohistochemical staining was done usingthe avidin-biotin complex (Vector Laboratories, Bur-lingame, CA) method. Sections were kept in 0.3%H202 in 0.1 mol/L PBS for 30 minutes at room tem-perature to inhibit endogenous peroxidase activity.After rinsing with 0.1 mol/L PBS containing 0.3% Tri-ton X-100 (PBST), they were incubated in a primaryantibody diluted in PBST for 24 hours at 4 C. A rab-bit polyclonal antibody raised against a synthetichuman f3/A41 24 peptide39 (1:5,000), a rabbit poly-clonal antibody against a human f/A41 42 peptide(Boehringer Mannheim Biochemica, Mannheim,Germany, 30 pg/ml), a rabbit polyclonal antibodyagainst a synthetic human f/A41 15 (1:1,000), a rab-bit polyclonal antibody against a recombinant hu-man APP1 592, which corresponds to 1-592 ofAPP695 (1:5,000), a rabbit polyclonal antibodyagainst a synthetic human APP671 695, which corre-sponds to C-terminal 25 amino acid residues of hu-man APP695 (1:4,000), and a rabbit polyclonal anti-body against a synthetic human APP666 695(1:1,000) were used as primary antibodies. Afterrinsing with PBST, sections were incubated in a bi-otinylated anti-rabbit IgG diluted (1:200) in PBST for1 hour at room temperature. After washing withPBST, they were placed in an avidin-biotin peroxi-dase complex diluted (1:400) in PBST for 1 hour at

room temperature. After washing thoroughly with0.05 mol/L Tris-HCI buffer (pH 7.6), they were col-ored with 0.02% 3-3' diaminobenzidine tetrahydro-chloride, 0.45% nickel ammonium sulfate, and0.006% H202 in 0.05 mol/L Tris-HCI (pH 7.6) for 10minutes. These stained sections were then mountedonto gelatin-coated slides, air-dried, dehydratedwithin a graded series of ethanol, cleared by xy-lene, and coverslipped with Entellan new (Merck,Darmstadt, Germany). The immunohistochemicalabsorption experiment was done by adding 0.1 to10 pmol/L ,3/A41 24 peptide to the working solution.As a control experiment, we performed the identicalimmunohistochemical procedure but with omissionof the primary antibody or using a normal rabbit se-rum instead of the primary antibody. Some sectionsfrom all the animals were stained with hematoxylinand eosin, neutral red, Congo red, thioflavine S,and periodic acid-Schiff (PAS).

Immunoblotting Analysis

We examined SAM-P/8 and SAM-R/1 male mice ofthree age groups (2 months old, 9 months old, and12 months old). We studied three mice for each agegroup. These mice were anesthesized and per-fused transcardially with 50 ml ice-cold 0.01 mol/LPBS (pH 7.4). Each brain was homogenized at 4 Cin 10 volumes of 0.01 mol/L PBS (pH 7.4). Homoge-nized samples were centrifuged at 30,000g, andpellets were resuspended in 3 to 5 volumes of 2%sodium dodecyl sulfate containing 8 mol/L urea.These resuspended samples were solubilized in 50-fold sodium dodecyl sulfate sample buffer as de-scribed by Laemmli,41 incubated for 1 hour at 37 C,then sonicated. Diluted samples (15 to 20 pi) wereelectrophoresed on 8% or 20% polyacrylamide gelsand transferred to polyviniliden difluoride filters (Im-mobilon, Nihon Millipore Co., Ltd., Tokyo, Japan).The transferred membranes were incubated in ablocking solution of 3% bovine serum albumin in0.01 mol/L PBS (pH 7.2) and then in a primary anti-body for 24 hours at 4 C. An anti-f3/A41 24 antibody(1:3,000), an anti-f3/A41 42 antibody (30 pg/ml), ananti-f3/A41 15 antibody (1:1,000), and an anti-APP671 695 antibody (1:2,000) were used as primaryantibodies. After washing with 0.01 mol/L PBS (pH7.4), the membranes were incubated in a biotiny-lated anti-rabbit IgG diluted (1:1,000) in 0.01 mol/LPBS (pH 7.4) for 1 hour at room temperature,washed again with 0.01 mol/L PBS (pH 7.4), andthen reacted with an avidin-biotin peroxidase com-plex diluted (1:200) in 0.01 mol/L PBS (pH 7.4) for

3/A4-Like Immunoreactivity in SAM Brain 1889A/P Jfune 1993, V'ol. 142, Ao. 6

30 minutes at room temperature. They were im-mersed in 0.005% 3-3' diaminobenzidine tetrahy-drochloride and 0.006% H202 in 0.05 mol/L Tris-HCI(pH 7.4).

Quantitative Analysis

Two sections of the cerebral cortex at the level of2.6 to 2.8 mm rostral to the frontal zero plane, thevertical plane running through the interaural line,and two sections of the trigeminal nerve root at thelevel of 0.2 to 0.4 mm caudal to the frontal zeroplane were collected from j3/A41 24 immunostainedsections of each animal. Four representative micro-scopic fields under a x20 objective lens were arbi-trarily selected bilaterally from each section. The im-munoreactive granules in each field (3.72 x 104 p2)were counted using a computerized image ana-lyzer, LUZEX3U (Nikon, Tokyo, Japan). The numeri-cal densities (number/ip2) were calculated. A statis-tical analysis was done using paired t-test toevaluate differences in the numerical densities ofimmunoreactive granules among respective agegroups of both SAM-P/8 and SAM-R/1. Differencesin the numerical densities of granules between age-matched SAM-P/8 and SAM-R/1 were also statisti-cally analyzed by paired t-test.

Homology Search

To investigate the possibility of cross-reactivity ofthe present 3/A4 antiserum with other homologous

proteins, a homology search was performed using a

computerized homology search system, IntegratedDatabase and Extended Analysis System for Nu-cleic Acids and Proteins. Proteins with amino acidsequence homologous to the human ,B/A41 24 pep-tide were searched for among already known pro-teins of mammalians, including mouse, and the per-cent match was calculated.

Results

Histological StudyWe found amyloid ,B/A4 proteinlike immunoreactivityin the form of a granular structure in various regionsof the SAM brains, using an anti-f3/A41 24 antibody(Figure 1). The ,B/A4-like immunoreactive granularstructures (,B-LIGS) were 1.5 to 2.5 p in diameterand irregularly shaped. The topographic distributionof ,B-LIGS is shown in Figure 2. They were dissemi-nated in both gray and white matter such as the lat-eral olfactory tract, medial septum, cerebral cortex,thalamus, hippocampus, caudate putamen, pyrami-dal tract, cerebellar peduncle, cerebellum, andsome cranial nerve roots and nuclei. In the cerebralcortex, the granules were mainly observed in thedeep cortical layers (Figure 1, A and B). In the hip-pocampus, two types of immunoreactive granuleswere observed (Figure 3A): f3-LIGS and granules ofa larger size. Larger granules were round to ovoidin shape. They were usually clustered and stained

Figure 1. Inifiniohi.%tochfe,istrp oJ therhrainsoJ SAMl1-1P/c8 ninsig aiitiserinni (1:5.00() again%sta snthvetic hbnian PIA4,_4 AA:1' rjeuj-JIcGS (arrol"s) cate oh.serived ill the/fiontopari-et'1l coliex J ci 2-ionlth-old SA f11-P8 miouse(X 360). B: (-LIGS in th fi-ontopariietal Cortex0] a 12- miontlb-olcd SAM-P/8 m e.mouNumerousgi.coiieIIi'S are oh.ser'edcl mainyi' in tbhedeepcorti-cal lalers (X 360) I%selt sbho's a higher powervitwo' ,'-IIG.S (,x 1 000); 1' /fih cottical lai'er1. sixth corlical la'er: IW:/ wbhitle matter c:3-IIGS (arrous) in the trigeminiial nerie root ofa 2-mnilth-old SAMI-P/,8nose (X 360). D:,3-IIG.S in the tfrigeinial ierier-oot of a 12-month old SAIlI-P/18 non%e. Granldes ar-e mark-ecdl/i increa.sed in mnohber ad are miorestronglI imniimnol/abeled comtpared it/h thosein a 2-mionth-old SAM-P8 mniouse (X 36)0).

.....-

Siim1.:

1890 Takemura et alA/P June 1993. Vol. 142. No. 6

FrPa 0 00

0'00

0

1-I

Xo 1)Po MD 0 0

I p cm . " "

00

*10 0DA 0

o DMi

Figure 2. Topog,-apbic distribution fJ f3-LIGS in the brain ota 12- month-old SA-P f8. O-LIGS car-e clisttibutecd in the various bracinl r-egiouls. inlcid-inig the nitedial septlou, cerebral cortex. bippocantipi.s aiid somie cranial nerie roots. krPa. firontoparietal cortex. CPl: caillidateplitallnel, CC corpics cal/losmnn. M:S. medial septal nucleus, LS lateral septal nuclels, AC anterior commnissore, 7'T: o/Jactorn tiubercle, DHC dorsal bippocanipalconinhissiure, Hi. hippocantipis. Fi: flnhtibia .IID: mediodlorsal thb/anicic nucleIts. (ClIt cenitral medial thalcamiic niuclelts, ID: late)orlot-sal thalcamuiicnuicleius. LP: latercal posterior thalcamiiic nniiclels. Po: posterior thclCatiic nIlucleils, 171. iventral posterior thu/COnliic nuclelts. DA: dorsal bhpothalaniicarea. DWII clorsomiedial bhipothalamic nIuclells, 13111 ventroniedial hipotbalanic incleus. SiuG. sipemficic l graY laYer qfsuperior colliclnl/s, INY(ointercollicilar niyiclents, CC. central grai', PPTg: pedunculopontine tegmuental niuclelus. XSCP. cleciussation oj sioperior cerebellar peduncle. P'nO.poniitne reticular nucleus, TFP. transierse fibers /polns, LL liateral lemninscus, BP. brachilmn pJontis. IICP.P mni/cIle cerebellarpednclc'. S5 sensoryroot q/ trigemninal/nere, sCP. superior cerebellar pecluncle. VSC ventraltlspuniocerebellar- tracict. J/o5: muotor trigenminal uucleuis. Pr5: principal sen-son, trigemiunal nucleins 7n: /CCiCal nerve. S: i'estibilocochlear nerve. Tz. trapezoid bodi', PI' puramuidal tract.

with PAS, findings compatible with previously re-ported PAS-positive granular structures (PGS).42PGS were also stained with various other antiseraand even with normal rabbit serum (1:2,000).3-LIGS were stained only when we used frozen sec-tions from perfused brains, in a free-floating condi-tion. We found no immunoreactive structures informalin-fixed paraffin-embedded sections, evenwith formic acid pretreatment.43 Counterstainingwith neutral red revealed that f-LIGS were presentin neuropils rather than in the cell body (Figure 3B).

f-LIGS were also stained with an anti-f3/A41 15antibody, although they were smaller in number,weaker in immunopositivity and in limited brainareas compared to those stained with an anti-,8/A41 24 antibody. 3-LIGS stained with an anti-p/A41 15 antibody appeared similar in terms of theirshape and size to those stained with an anti-13/A41 24 antibody. They were found in some brainareas including the cingulate cortex, hippocampus,cerebellar nuclei, cerebellar peduncles, and trigem-inal and vestibular nuclei. There was no regionwhere granular structures were stained only with ananti-t/A41 15 antibody. In the hippocampus, both3-LIGS and PGS were stained with an anti-f/A41 15antibody. An anti-APP1 592 stained a subset of neu-

rons, as described previously,44 45 although f3-LIGSwere not stained (Figure 3C). An anti-APP671 695, ananti-APP666 695 and an anti-4/A41 42 antibodies didnot label the 3-LIGS despite formic acid pretreat-ment with various concentrations.4346

Control staining was made by omission of the pri-mary antibody or exchanging the primary antibodywith normal rabbit serum. f-LIGS were never ob-served under these conditions. When sections wereincubated in the primary antibody preabsorbed with0.1 to 10 pmol/L synthetic human f/A41 24, no spe-cific staining was observed (Figure 3D). The identi-cal procedure also abolished senile plaque stainingin the brain of a subject with Alzheimer's disease(Figure 3, E and F). Congo red or thioflavine Smethods did not stain similar granular structures.PAS stained PGS but not ,B-LIGS.

Immunoblotting Analysis

For immunoblotting analysis, we used two gels withpolyacrylamide concentrations of 8% and 20% toclearly separate the high molecular weight region,including full-length APP (120 to 130 kd), and thelow molecular weight region, including f/A4 protein

,B/A4-Like Immunoreactivity in SAM Brain 1891AJPJune 1993, Vol. 142, No. 6

gure 3. Imnritoohistochemnistry of the brain of a 12-month-old SA I-P18 (A to D). and imunoiohstochemnistry of the hlmnani braini of Alzbeimerssease as a positive conitrol (E antd F). A: In the hippocamnpi.s ininiliiolabeled by aniti-B/A4,12 antisernum (1:5.000). two types of inimluniloreactiueannII/Ces are oh.seed.- the larger PGS (long arrows) and the sniialler 13-LIGS (short arrows) (X 360). Or oriens laer. Pp pyramnicdal ce-ll lover, Rad/atium radiatiotn, Alol/o niolecular layer. B: Counterstaiming with neiutr^al red suggests that graniules are main1nb localized in neuropils (X-360). C.the trigeminal nerve root iuiiiinillolabelced by antiserumin (1: 5,000) against a recombiblanlt API',-592 n0o grcalular striLctiires are stainiecd (X 360).The trigeminal nerve root immuninolabeled by antit-(3A4,-24 antiserumn after preab.soiption niib I ,uumiollL synithetic humrlani (3 A4,_2 3,-LIGS

tlining is abolished fX 360). E. Humelani temiporal cortvx uninniinolabeled by, anti-P3IA4,2 antiserumu (1.8,000, X 360). F: Human tem*lporcal cor-tc-wuitinolabeled by, aniti-,(3A4,1. antisemru after preabsopption with 1 Iymicol L synthetic (3 A4,_2. Senile plaque staininlg is abolished (X 60).

.-2 kd), respectively. With the 20% polyacrylamide-I, robust 14- to 18-kd immunoreactive bandsere identified (Figure 4A). The predicted 4.2-kdand that corresponds to the human f/A4 proteinas not detected. Using the 8% polyacrylamide

gel, a robust band was observed at the end of thegel (Figure 4B). Thus, the molecular weight of themain immunoreactive protein was lower than that ofthe gel end, <30 kd, which was separated into 14-to 18-kd bands with the 20% gel (Figure 4A). A faint

1892 Takemura et alAJPJune 1993. lVol. 142. No. 6

Figure 4. Western inimmunioblots (of the brainis of SAM, A Immtnioblot of a 12-monith-old SAl-PP8 usinzg a 2000 polVacrylamide gel. Lane 1 anti-PXA41 -28 antiserum (1 3.000) lane 2 anti- BA4, 24 antiserum after ahsorption with 1 motnol L spbthetic , A4, 24 14- to 18-kd hands are stainied.Ihes staininlg is abolished hbi preahsorption with the antigen. B immunlohlot o?f a 12- niouthb-old SAil -P18 utsing a 8%yo polacvrvlamide gel A roblustbanod at gel enid (< 30 kd) and a faint hband ait 120 kd (arrowi, head) are labeled TIhe labeling is abolished hi preabsotptionl u ith the antigen Lane1 anti-BA4, -2 antiserum (1 3.000). lane 2 anti-BA4, 2j antiscrumn after ahsotption with 1 imolll svnbthetic PA4, 2, C- Inimunoblhts u/fa 12-nionth-old SA.lt-1 '8 using a 20'YO polacnlamide gels Lane 1 anti-PIA4 2, anitiseruiii (1 3.000) lane 2 anlti-PA4, , anltihbod (30 yg/nis,) lane3 anti-APP(, 1-695 antiserum (I 2,000) All antibodies utsed here shoued a sintiilar, imniiiobhlotting pattern. D: niiunobhlots of 2-, 9- anid 12-mnioth-oldc SAM-P 8 and SAM-RR I using an anti-,A41 2, antiserum (1 3,000). Lanes 1 and 4 2-month-o(ld, lanes 2 anid 5. 9 -niouzth-old, lanes 3anOd 6 12-monith-old SAIA1tP8 (lanes 1. 2. anid 3) and SAM-R1 lanes 4, 5, anid 6) Staining intensities o?f inminmuoreac tive bands seeni to inCreasenith aginig in both SANI P18 andSAlt1-R'1 Inicrements of staining intensities occut in SAJI-PP8 at a i'ounger age than in SAMt-RI1 25 jg (A to C)or 15 ug (D) ofJ total protein as applied to i'ach lane. Bars inldicate moleciular eights 20 1. 144. and 6 2 kl. respectiveli' (C anid D)

120-kd band that seemed to be a full-length APPwas also observed. Staining of these bands wasabolished by preabsorption of the primary antibodywith 1 to 10 pmol/L synthetic human P/A41 24 pep-

tide (Figure 4, A, lane 2, and B, lane 2). The 14- to18-kd bands were also recognized using an anti-f3/A41 42 antibody, an anti-APP671 695 antibody (Fig-ure 4C), and an anti-f/A41 15 antibody as primary

3/A4-Like Immunoreactivity in SAM Brain 1893AJPJune 1993, Vol. 142, No. 6

antibodies. Staining intensities of immunoreactivebands increased with aging in both SAM-P/8 andSAM-R/1, and increases in staining intensitiesseemed to occur in the earlier life period in SAM-P/8rather than in SAM-R/1 (Figure 4D).

Quantitative Analysis

f-LIGS seemed to be more numerous in the olderanimals in all regions where P-LIGS were observed,most prominent in the cerebral cortex, medial sep-tum, thalamus, and trigeminal nerve root. We per-formed a quantitative analysis in the cerebral cortexand trigeminal nerve root using a computerized im-age analyzer, LUZEX3U (Figure 5). The mean nu-merical densities of ,-LIGS in 2-, 4-, 9-, 12-, and 22-month-old SAM-R/1 were 0.13, 4.93, 33.5, 40.1, and53.6 (x 140-/p2) in the cerebral cortex and 33.7,46.5, 91.7, 97.3, and 107.6 (x10-4/p2) in the trigem-inal nerve root, respectively. The mean numericaldensities of f3-LIGS in 2-, 4-, 9-, and 12-month-oldSAM-P/8 were 0.24, 5.38, 59.0, and 75.1 (x10-4/p2)in the cerebral cortex and 34.9, 51.0, 111.0, and126.3 (x10-4/p2) in the trigeminal nerve root, re-spectively. Differences in numerical densities be-tween 2-month-old SAM-R/1 versus 9-, 12-, and 22-month-old SAM-R/1 were statistically significant at Pvalues of 0.0532 (t = -4.159), 0.0082 (t =

-10.974), and 0.0591 (t = -10.744) in the cerebralcortex and at P values of 0.0005 (t = -10.465),0.0001 (t = -18.621), and 0.0235 (t = -27.109) inthe trigeminal nerve root. Differences in numericaldensities between 2-month-old SAM-P/8 versus 4-,9-, and 12-month-old SAM-P/8 were also statisticallysignificant at P values of 0.0163 (t = -7.732),0.0001 (t = -25.34), and 0.0027 (t = -9.236) in thecerebral cortex and at P values of 0.0143 (t =-4.144), 0.0006 (t = -7.713), and 0.0009 (t =

-8.776) in the trigeminal nerve root, respectively. Inaddition, in the cerebral cortex, differences in nu-merical densities of j3-LIGS were statistically signifi-cant at a P value of 0.0386 (t = 3.533) for 9-month-old SAM-P/8 versus SAM-R/1 and at a P value of0.0029 (t = 6.478) for 12-month-old SAM-P/8 versus12-month-old SAM-R/1. In the trigeminal nerve root,the difference was statistically significant at a Pvalue of 0.0354 (t = 3.124) for 12-month-old SAM-P/8 versus 12-month-old SAM-R/1.

Homology Search

The Integrated Database and Extended AnalysisSystem for Nucleic Acids and Proteins revealed that

Acerebral cortex

(No

6

0S

a4

'.x1E 2Sa

2 4 9 1 2 2 2 (Month)Age

B

(No./Jtm2

stn_.

trigeminal nerve root

D R/I

.; 100AOi9

.l't2Slat

2 4 9 12 22 (Month)

Age

Figure 5. N\umerical densities of 13-LIGS in the cerebral cortex (A)ald the trigeminal n7ene root (B). Nuimerical densities of P-LIGS in-crea.se significantl/ uwith advancing age in both SAM-R11 and SAt-P18 mice brains. Comparced uwith 2-month-old SAM-RI], increases innumerical densities of(-LIGS in 9-. 12-, anzd 22-montth-old SAM-R/1are statisticall significant at P values of 0.0532, 0.0082, and 0.0591in the cerebral cortex, anld at P v'alues of0.0005. 0.0001, and 0.0235in the trigeminlal neene root, respectivelv. Similarly, compared with2-month -old SAM-P/8, increases in 4-, 9-, and 12-month-old SAM-P/8are sign7ificant at P valuees of 0.0163, 0.0001, and 0.0027 in the ce-rebral cortex, anid at P values of 0.0143, 0.0006, and 0.0009 in thetrigeminal nenve root. In addition, inicreases in numerical densitiesof,-LIGS are more niarked in SAM-P18 compared unith SAMl-RI (t: I'< 0.05, **: P < 0.01). Each statistical anlalvsis was dotne using pairedtest. Each bar representts mean + SF.

the mouse P3/A4 region was 83.3% homologous tothe human P/A41-24. No established mouse protein,except for the mouse ,B/A4 region, was found to behomologous to the human 3/A41-24

DiscussionTo our knowledge, this is the first report of sponta-neous development of 3/A4-like immunoreactivity inthe mouse brain. ,B-LIGS were stained only when weused frozen sections from perfused brains, in a

1894 Takemura et alAJPJune 1993, Vol. 142, No. 6

free-floating condition. We found no f/A4-like immu-noreactive structures in formalin-fixed, paraffin-embedded sections, despite a formic acid pretreat-ment.43 Based on the clear absorption of the ,B/A4immunostaining of f-LIGS by a low concentration ofthe corresponding synthetic peptide, we considerthat the staining is a specific immunoreaction. Simi-lar to findings with another anti-3/A4 antibody,44 thepresent anti-f/A41 24 antibody also recognized a120-kd band that may correspond to a full-lengthAPP. However, the staining intensity of a 120-kdband was much lower than that of 14- to 18-kdbands. In addition, the staining intensities of 14- to18-kd bands increased with advancing age, whichis compatible with the age-related increase in nu-merical densities of ,-LIGS. Thus, the main immu-noreactivity of ,-LIGS is considered to reflect rec-ognition of the 14- to 18-kd proteins rather than full-length APR It is also possible to suppose that theanti-3/A41 24 antibody recognizes APP-uncon-cerned proteins that have amino acid sequenceshomologous to the human 1/A41 24. For elucidation,we used the computerized homology search sys-tem, Integrated Database and Extended AnalysisSystem. This system revealed that no alreadyknown mouse protein, except for the mouse f3/A4region, was homologous to the human f3/A41 24,which suggests that no known protein is likely to ac-count for a cross-reaction. Furthermore, f-LIGSwere stained not only with an anti-3/A41 24 antibodybut also with an anti-13/A41 15 antibody, and bothantibodies recognized the same 14- to 18-kd bandson immunoblotting lanes. Thus, we consider that thepossibility of a cross-reaction is unlikely, whereaswe could not still exclude some possibility of an im-munochemical or immunohistochemical cross-reaction with unknown protein.An anti-f3/A41 42 antibody and an anti-APP671 695

antibody did not stain 3-LIGS immunohistochemi-cally, whereas they recognized the 14- to 18-kdbands on immunoblotting lanes. This discrepancymight be attributed to differences of epitopes rec-ognized by each antibody. In tissue sections,epitopes recognized by an anti-3/A41 42 antibodyor an anti-APP671 695 antibody might be hidden dueto conformations of the target protein or surround-ing structures, whereas epitopes recognized by ananti-4/A41 24 antibody or an anti-f3/A41 15 antibodywere exposed using the present immunohistochem-ical procedure. In contrast, when we homogenizedsamples, these architectures might be broken downand epitopes exposed to be recognized by eachantibody.

There has been no report on granular structure inthe mouse brain similar to ,-LIGS. PGS, describedby Akiyama et al,42 are granules of a larger size,measuring up to 5 p, are usually clustered, andstain with PAS. In addition, immunohistochemically,PGS stained not only with the present anti-f/A41 24antibody, but also with various others, includinganti-APP1 592 antibody (1:3,000) and even with nor-mal rabbit serum (1:2,000). Thus, we conclude that,-LIGS and PGS are dissimilar.

Numerical densities of f3-LIGS increased signifi-cantly with advancing age in brains of both SAM-P/8 and SAM-R/1, characteristics of importancebecause only neuroaxonal dystrophies and eosino-philic thalamic intraneuronal inclusions,47 lipofuscinaccumulation,48 and PGS42 have been found asmorphological parameters of aging in the mousebrain. Furthermore, numerical densities of ,B-LIGSincreased more markedly in brains of SAM-P/8 com-pared with those of SAM-R/1. Increases in numeri-cal densities of ,B-LIGS were most rapid at agesfrom 4 to 9 months. Studies on memory and learn-ing abilities using both active and passive avoid-ance tasks showed that the deterioration of abilitiesprogressed from age 4 to age 8/9 months35-38 inSAM-P/8. Thus, the rapid and accelerated accumu-lation of j3-LIGS might be associated with the age-related deterioration of memory and learning abili-ties seen in SAM-P/8.As described above, the immunoreactivity of

f-LIGS may be associated with the 14- to 18-kdproteins. Because the 14- to 18-kd proteins werespecifically recognized with anti-f3/A4 antibody, the14- to 18-kd proteins seem to be the f/A4 region-containing APP fragments generated by normal oraberrant processing associated with aging. More-over, certain parts of these fragments may harborthe C-terminal region of APP, according to the resultof immunoblot using an anti-APP671 695 antibody.As the identical anti-f3/A41 24 antibody used in thepresent study recognized a 4.2-kd robust band inthe brain of a subject with Alzheimer's disease,39processing of APP may occur in a different mannerbetween mouse and human brains. Differences inthe primary structure of APP21l49,50 or the process-ing enzyme may relate to the differences in the pro-cessed proteins, 4.2-kd f3/A4 protein in the humanbrain and 14- to 18-kd APP fragments in the mousebrain. The differences in the processed proteinscould affect their polymerization into the p,-pleatedsheet,51 leading to the development of characteris-tic senile plaques classically with Congo red bire-fringence in the human brain and to the develop-ment of granular deposits in the absence of Congo

,B/A4-Like Immunoreactivity in SAM Brain 1895AJPJune 1993, Vol. 142, No. 6

red birefringence in the mouse brain. The character-ization of the 14- to 18-kd proteins in the mousebrain is necessary to address this assumption.

In the present study, we obtained evidence indi-cating that 3-LIGS, which may have 14- to 18-kdAPP fragments as a component, develop spontane-ously in the SAM brain and that they increase innumber with advancing age, predominantly in ani-mals with a phenotype of age-related deteriorationof memory and learning abilities. We suggest thatf3-LIGS are not only a new morphological manifesta-tion of senescence in mice brains, but also a perti-nent clue to determine mechanisms involved inamyloid deposition.

AcknowledgmentsWe thank Drs. N. Kitaguchi and Y. Tokushima (Bio-science Laboratory, Life Science Research Labora-tories, Asahi Chemical Industry Co., Ltd., Japan) forgenerous gifts of anti-4/A41.24 and anti-APP, 592antibodies. We are grateful to Dr. K. Yoshikawa(Tokyo Metropolitan Institute for Neuroscience,Japan) for provision of anti-APP671,695 antibody andto Dr. H. Yamaguchi (Gunma University school ofmedicine) for anti-4/A4115 and anti-APP666-695 anti-bodies with pertinent advice. We thank Drs. J.Tateishi, T. Kitamoto (Kyushu University), K. Higuchi,and Y. Ashida (Kyoto University) for pertinent adviceand Miss Y. Kitasaka for technical assistance.

References1. Bobin SA, Currie JR, Merz PA, Miller DL, Styles J,

Walker WA, Wen GY, Wisniewski HM: The comparativeimmunoreactivities of brain amyloids in Alzheimer'sdisease and scrapie. Acta Neuropathol 1987, 74:313-323

2. Castano EM, Ghiso J, Prelli F, Gorevic PD, Migheli A,Frangione B: In vitro formation of amyloid fibrils fromtwo synthetic peptides of different lengths homolo-gous to Alzheimer's disease ,B-protein. Biochem Bio-phys Res Commun 1986, 141 :782-789

3. Glenner GG, Wong CW: Alzheimer's disease: initial re-port of the purification and characterization of a novelcerebrovascular amyloid protein. Biochem BiophysRes Commun 1984, 120:885-890

4. Masters CL, Simms G, Weinman NA, Multhaup G, Mc-Donald BL, Beyreuther K: Amyloid plaque core proteinin Alzheimer disease and Down syndrome. Proc NatlAcad Sci USA 1985, 82:4245-4249

5. Selkoe DJ, Abraham CR, Podlisny MB, Duffy LK: Isola-tion of low-molecular-weight proteins from amyloidplaque fibers in Alzheimer's disease. J Neurochem1986, 46:1820-1834

6. Glenner GG, Wong CW: Alzheimer's disease andDown's syndrome: sharing of a unique cerebrovascu-lar amyloid fibril protein. Biochem Biophys Res Com-mun 1984, 122:1131-1135

7. Coria F, Castano EM, Frangione B: Brain amyloid innormal aging and cerebral amyloid angiopathy is anti-genically related to Alzheimer's disease f-protein. AmJ Pathol 1987, 129:422-428

8. Joachim CL, Duffy LK, Morris JH, Selkoe DJ: Proteinchemical and immunocytochemical studies of menin-govascular ,B-amyloid protein in Alzheimer's diseaseand normal aging. Brain Res 1988, 474:100-111

9. Kang J, Lemaire HG, Unterbeck A, Salbaum JM, Mas-ters CL, Grzeschik KH, Multhaup G, Beyreuther K,Mueller-Hill B: The precursor of Alzheimer's diseaseamyloid A4 protein resembles a cell-surface receptor.Nature 1987, 325:733-736

10. Ponte P, Gonzalez-DeWhitt P, Schilling J, Miller J, HsuD, Greenberg B, Davis K, Wallace W, Lieberburg I,Fuller F, Cordell B: A new A4 amyloid mRNA containsa domain homologous to serine proteinase inhibitors.Nature 1988, 331:525-527

11. Tanzi RE, McClatchey Al, Lamperti ED, Villa-KomaroffL, Gusella JF, Neve RL: Protease inhibitor domain en-coded by an amyloid protein precursor mRNA associ-ated with Alzheimer's disease. Nature 1988, 331:528-530

12. Kitaguchi N, Takahashi Y, Tokushima Y, Shiojiri S, ItoH: Novel precursor of Alzheimer's disease amyloidprotein shows protease inhibitory activity. Nature1988, 331 :530-532

13. De Sauvage F, Octave JN: A novel mRNA of the A4amyloid precursor gene coding for a possibly se-creted protein. Science 1989, 245:651-653

14. Golde TE, Estus S, Usiak M, Younkin LH, Younkin SG:Expression of ,B amyloid protein precursor mRNAs:recognition of a novel alternatively spliced form andquantitation in Alzheimer's disease using PCR. Neuron1990, 4:253-267

15. Wisniewski HM, Johnson AB, Raine CS, Kay WJ, TerryRD: Senile plaques and cerebral amyloidosis in ageddogs: a histochemical and ultrastructural study. LabInvest 1970, 23:287-296

16. Wisniewski HM, Ghetti B, Terry RD: Neuritic (senile)plaques and filamentous changes in aged rhesusmonkeys. J Neuropathol Exp Neurol 1973, 32:566-584

17. Struble RG, Price DL Jr, Cork LC, Price DL: Senileplaques in cortex of aged normal monkeys. Brain Res1985, 361 :267-275

18. Selkoe DJ, Bell DS, Podlisny MB, Price DL, Cork LC:Conservation of brain amyloid proteins in aged mam-mals and humans with Alzheimer's disease. Science1987, 235:873-877

19. Abraham CR, Selkoe DJ, Potter H, Price DL, Cork LC:a1-Antichymotrypsin is present together with the,B-protein in monkey brain amyloid deposits. Neuro-science 1989, 32:715-720

1896 Takemura et alA/P june 1993, Vol. 142. No. 6

20. Giaccone G, Verga L, Finazzi M, Pollo B, Tagliavini F,Frangione B, Bugiani 0: Cerebral preamyloid depositsand congophilic angiopathy in aged dogs. NeurosciLett 1990, 114:178-183

21. Podlisny MB, Tolan DR, Selkoe DJ: Homology of theamyloid beta protein precursor in monkey and humansupports a primate model for beta amyloidosis inAlzheimer's disease. Am J Pathol 1991, 138:1423-1435

22. Ishihara T, Gondo T, Takahashi M, Uchino F, Ikeda S,Allsop D, lmai K: Immunohistochemical and immuno-electron microscopical characterization of cerebrovas-cular and senile plaque amyloid in aged dogs' brains.Brain Res 1991, 548:196-205

23. Quon D, Wang Y, Catalano R, Scardina JM, MurakamiK, Cordell B: Formation of 3-amyloid protein depositsin brains of transgenic mice. Nature 1991, 352:239-241

24. Richards SJ, Waters JJ, Beyreuther K, Masters CL,Wischik CM, Sparkman DR, White CL ll, AbrahamCR, Dunnett SB: Transplants of mouse trisomy 16 hip-pocampus provide a model of Alzheimer's diseaseneuropathology. EMBO J 1991, 10:297-303

25. Takeda T, Hosokawa M, Takeshita S, Irino M, HiguchiK, Matsushita T, Tomita Y, Yasuhira K, Hamamoto H,Shimizu K, lshii M, Yamamuro T: A new murine modelof accelerated senescence. Mech Ageing Dev 1981,17:183-194

26. Takeda T, Hosokawa M, Higuchi K: Senescence-Accelerated Mouse (SAM): a novel murine model ofaccelerated senescence. J Am Geriatr Soc 1991, 39:911-919

27. Matsumura A, Higuchi K, Shimizu K, Hosokawa M,Hashimoto K, Yasuhira K, Takeda T: A novel amyloidfibril protein isolated from Senescence-AcceleratedMice. Lab Invest 1982, 47:270-275

28. Higuchi K, Yonezu T, Kogishi K, Matsumura A, Take-shita S, Higuchi K, Kohno A, Matsushita M, HosokawaM, Takeda T: Purification and characterization of a se-nile amyloid-related antigenic substance (apoSASSAM)from mouse serum. J Biol Chem 1986, 261:12834-12840

29. Yonezu T, Tsunasawa S, Higuchi K, Kogishi K, Naiki H,Hanada K, Sakiyama F, Takeda T: A molecular-pathologic approach to murine senile amyloidosis: se-rum precursor-apolipoprotein A-Il valiant (Pro5 -* Gin)presents only in the senile amyloidosis-prone SAM-P/1and SAM-P/2 mice. Lab Invest 1987, 57:65-70

30. Naiki H, Higuchi K, Yonezu T, Hosokawa M, Takeda T:Metabolism of senile amyloid procursor and amyloido-genesis: age-related acceleration of apolipoproteinA-Il clearance in the Senescence Accelerated Mouse.Am J Pathol 1988, 130:579-587

31. Hosokawa M, Takeshita S, Higuchi K, Shimizu K, IrinoM, Toda K, Honma A, Matsumura A, Yasuhira K,Takeda T: Cataract and other ophthalmic lesions inSenescence Accelerated Mouse (SAM). Morphology

and incidence of senescence associated ophthalmicchanges in mice. Exp Eye Res 1984, 38:105-114

32. Matsushita M, Tsuboyama T, Kasai R, Okumura H,Yamamuro T, Higuchi K, Higuchi K, Kohno A, YonezuT, Utani A, Umezawa M, Takeda T: Age-relatedchanges in bone mass in the Senescence-Accel-erated Mouse (SAM): SAM-R/3 and SAM P/6 as newmurine models for senile osteoporosis. Am J Pathol1986, 125:276-283

33. Chen WH, Hosokawa M, Tsuboyama T, Ono T, lizukaT, Takeda T: Age-related changes in the temporoman-dibular joint of the Senescence Accelerated Mouse.Am J Pathol 1989, 135:379-385

34. Hosokawa M, Kasai R, Higuchi K, Takeshita S,Shimizu K, Hamamoto H, Honma A, Irino M, Toda K,Matsumura A, Matsushita M, Takeda T: Grading scoresystem: a method for evaluation of the degree of se-nescence in Senescence Accelerated Mouse (SAM).Mech Ageing Dev 1984, 26:91-102

35. Miyamoto M, Kiyota Y, Yamazaki N, Nagaoka A, Mat-suo T, Nagawa Y, Takeda T: Age-related changes inlearning and memory in the Senescence-AcceleratedMouse (SAM). Physiol Behav 1986, 38:399-406

36. Yagi H, Katoh S, Akiguchi I, Takeda T: Age-related de-terioration of ability of acquisition in memory andlearning in senescence accelerated mouse: SAM-P/8as an animal model of disturbances in recent memory.Brain Res 1988, 474:86-93

37. Yagi H, Irino M, Matsushita T, Katoh S, Umezawa M,Tsuboyama T, Hosokawa M, Akiguchi I, Tokunaga R,Takeda T: Spontaneous spongy degeneration of thebrain stem in SAM-P/8 mice, a newly developedmemory-deficient strain. J Neuropathol Exp Neurol1989, 48:577-590

38. Ohta A, Hirano T, Yagi H, Tanaka S, Hosokawa M,Takeda T: Behavioral characteristics of the SAM-P/8strain in Sidman active avoidance task. Brain Res1989, 498:195-198

39. Takahashi H, Kurashima C, Utsuyama M, Hirokawa K:Immunohistological study of senile brains by using amonoclonal antibody recognizing f amyloid precursorprotein: significance of granular deposits in relationwith senile plaques. Acta Neuropathol 1990, 80:260-265

40. Yamaguchi H, Ishiguro K, Shoji M, Yamazaki T, Naka-zato Y, Ihara Y, Hirai S: Amyloid ,B/A4 protein precursoris bound to neurofibrillary tangles in Alzheimer-typedementia. Brain Res 1990, 537:318-322

41. Laemmli UK: Cleavage of structural proteins duringthe assembly of the head of bacteriophage T4. Nature1970, 227:680-685

42. Akiyama H, Kameyama M, Akiguchi I, Sugiyama H,Kawamata T, Fukuyama H, Kimura H, Matsushita M,Takeda T: Periodic acid-Schiff (PAS)-positive, granularstructures increase in the brain of senescence accel-erated mouse (SAM). Acta Neuropathol 1986, 72:124-129

f3/A4-Like Immunoreactivity in SAM Brain 1897AJP.June 1993, Vol. 142, No. 6

43. Kitamoto T, Ogomori K, Tateishi J, Prusiner SB: Formicacid pretreatment enhances immunostaining of cere-bral and systemic amyloids. Lab Invest 1987, 57:230-236

44. Card JP, Meade RP, Davis LG: Immunocytochemicallocalization of the precursor protein for ,B-amyloid inthe rat central nervous system. Neuron 1988, 1:835-846

45. Kawarabayashi T, Shoji M, Harigaya Y, Yamaguchi H,Hirai S: Amyloid /3/A4 protein precursor is widely dis-tributed in both the central and peripheral nervoussystems of the mouse. Brain Res 1991, 552:1-7

46. Perry G, Cras P, Siedlak SL, Tabaton M, Kawai M: ,Bprotein immunoreactivity is found in the majority ofneurofibrillary tangles of Alzheimer's disease. Am JPathol 1992, 140:283-290

47. Fraser H: Eosinophilic bodies in some neurones in the

thalamus of ageing mice. J Pathol 1969, 98:201-20448. Nandy K, Bourne GH: Effect of centrophenoxine on

the lipofuscin pigments in the neurones of senileguinea-pigs. Nature 1966, 210:313-314

49. Yamada T, Sasaki H, Furuya H, Miyata T, Goto I,Sakaki Y: Complementary DNA for the mouse ho-molog of the human amyloid beta protein precursor.Biochem Biophys Res Commun 1987, 149:665-671

50. Yamada T, Sasaki H, Doh-ura K, Goto I, Sakaki Y:Structure and expression of the alternatively-splicedforms of mRNA for the mouse homolog of Alzheimer'sdisease amyloid beta protein precursor. Biochem Bio-phys Res Commun 1989, 158:906-912

51. Kirschner DA, Inouye H, Duffy LK, Sinclair A, Lind M,Selkoe DJ: Synthetic peptide homologous to 3 proteinfrom Alzheimer disease forms amyloid-like fibrils invitro. Proc Natl Acad Sci USA 1987, 84:6953-6957