Method for Detecting and Treating Alzheimer's Disease

It has been appreciated for some time that Alzheimer's disease has a complex etiology. At least 15 percent of the cases appear to be due to the inheritance of an autosomal-dominant mutation, but the majority are "sporadic", showing no clear association with any identifiable genetic or environmental factor. Feldman, R. G., et al., Neurology, 13:811-824 1963; Heston, L. L., et al., Arch Gen. Psychiat., 38:1084-1090 (1981); Terry, R. D., Aging, 7:11-14 (1978); Jarvik, L. F. and Matsuyama, S. S., "The Biological Substrates of Alzheimer's Disease", Academic Press, pp. 17-20 (1986). Even identical twins can show a large discordance in the age of onset of the disease. Nee, L. E., et al., Neurology. 37:359-363 (1987). Yet despite this variation, Alzheimer's disease shows a uniform set of clinical and pathological features--progressive loss of memory and other intellectual functions beginning in middle to late life, coupled with neuronal cell loss in the higher centers of the brain. Price, D. L., Ann. Rev. Neurosci., 9:489-512 (1986).

When examined by histochemica) stains, Alzheimer's disease brains, particularly the hippocampus, neocortex, and amygdala, exhibit certain neuropathological protein deposits that serve as the defining characteristic of the disease. One such deposit, termed the neurofibrillary tangle, occurs inside neurons and is composed of "paired helical" protein filaments (PHF). Because they can be found in other neurodegenerative diseases, paired helical filaments are likely to be a common feature of dying neurons. The more definitive lesion of Alzheimer's disease is the "neuritic or senile plaque", which consists of a spherical, extracellular core of filamentous protein material surrounded by a halo of degenerating nerve cell processes. Extracellular protein filaments similar to those seen in the cores of neuritic plaques also accumulate in the walls of meningeal and intracortical blood vessels. The deposits of protein filaments in the cores of neuritic plaques and in blood vessels are referred to by the generic term "amyloid".

The first identical constituent of Alzheimer amyloid deposits was purified from meningeal blood vessels and its sequence determined by Glenner, G. G. and Wong, C. W., Biochem. Biophys. Res. Commun., 122:1131-1135 (1984). This protein, termed .beta. or A4, a is .about.42 amino-acid-long fragment of a larger protein that is a normal constituent of the brain and other tissues. A second protein component of Alzheimer amyloid deposits was identified as the serine protease inhibitor .alpha..sub.1 -antichymotrypsin (ACT).

While much has been learned about the biochemistry and expression of the aberrant protein deposits that characterize Alzheimer's disease, progress toward the development of methods for the diagnosis and treatment of the disease has been slow. This is due, at least in part, to the fact that the molecular basis for the disease pathology has remained obscure.

The subject invention relates to a method for detecting Alzheimer's disease comprising testing an individual for the presence of a mosaic population of cells. The mosaic population of cells includes normal cells having two copies of chromosome 21 and abnormal cells which contain 3 copies of chromosome 21. The mosaicism can be detected, for example, by in situ hybridization or by detecting increased sensitivity to cholinergic agonists or antagonists.

The invention enables detection of early signs of Alzheimer's disease prior to the onset of dementia. Also disclosed are methods for preventing the onset of dementia in Alzheimer's disease.

Applicant's invention was made possible by the development of a mechanistic theory which provides cohesiveness to apparently disparate experimental results in Alzheimer's Disease research. A brief review of the literature is essential to an appreciation of the invention.

Perhaps the most interesting clue to the cause of Alzheimer's disease is the fact that Down syndrome patients who live beyond the age of 30 or 40 develop dementia and neuropathology essentially indistinguishable from classic Alzheimer's disease. Olson, M. I. and Shaw, C. M., Brain. 92:147-156 (1969); Glenner, G. G. and Wong, C. W., Biochem. Biophys. Res. Commun., 122:1131-1135 (1984); Wisniewski, H. M. and Terry, R. D., In: Ford DH (ed) Progress in Brain Research, 40:1108-1109 (1988). The implication of this finding is that trisomy for chromosome 21--the pathogenetic cause of Down syndrome--is also capable of causing Alzheimer's disease, possibly through the overexpression of a gene residing on chromosome 21. Schweber, M., Ann. N.Y. Acad. Sci., 450:223-238 (1985).

On the other hand, almost all aged humans (and monkeys) develop some amyloid deposits which, by several criteria, appear to be identical to those that accumulate in much larger numbers and at an earlier time in Alzheimer's disease and Down syndrome (Wisniewski, H. M. and Terry, R. D., In: Ford DH (ed) Progress in Brain Research, 40:1108-1109 (1973); Selkoe, D. J., et al., Science, 235:873-877 (1987); Abraham, C. R. and Potter, H., Biotechnology, 7:147-153 (1989). Thus, any hypothesis for the pathogenesis of Alzheimer's disease should be able to explain not only the relation between the familial and sporadic forms of the disease but also how these are related to Down syndrome and to the "normal" process of aging.

The association between Alzheimer's disease and chromosome 21 has been reinforced by a number of recent clinical and experimental findings. These, and earlier results on the genetics, epidemiology, and cell biology of Alzheimer's disease, have been considered. In particular, its association with Down syndrome has been assessed.

A molecular basis for Alzheimer's disease, which is consistent with both the genetic and sporadic forms of the disease can be explained as arising from the accumulation of chromosome 21 trisomy cells during the life of the individual. That is, trisomy 21 cells, developing over time by unequal chromosome segregation during mitosis, may ultimately lead to Alzheimer's disease through the same (as yet unknown, and perhaps multistep) mechanism by which Down syndrome patients acquire the disease, but at a later age due to the modulating effect of the mosaicism.

The first specific model linking Alzheimer's disease to Down syndrome arose when the gene for the amyloid .beta.-protein was cloned and found to be located on chromosome 21. Goldgaber, D., et al., Science, 235:877 (1987); Kang, J., et al., Nature, 325:733 (1987); Tanzi, R. E., et al., Science, 235:880 (1987); Robakis, N. K., et al., Proc. Natl. Sci. U.S.A., 84:4190 (1987). The implication of these results seemed that accumulation of amyloid in Alzheimer's disease was caused by the overexpression of a routant .beta.-protein gene or by a duplication of the .beta.-protein gene on chromosome 21 that mimicked the gene-dosage effect of Down syndrome. The fact that some Alzheimer's disease families could be shown to harbor their autosomal dominantly-inherited mutation on chromosome 21 (St. George-Hyslop, P. H., et al., Science, 235:885-889 (1987)) and that the .beta.-protein precursor gene was apparently overexpressed in Down syndrome (Tanzi, R. E., et al., Science 235:880 (1987); Neve, et al., 1988) further implicated the .beta.-protein gene as a potential site for the disease locus.

Very recently, a variant form of the .beta.-protein precursor gene encoding a mutant .beta.-protein has been found in families with Hereditary Cerebral Hemorrhage with Amyloidosis of Dutch Origin, suggesting that this mutation may be the inherited defect in this disease (Van Broeckhoven, C., et al., Science, 248:1120-1122 (1990); Levy, E., et al., Science, 248:1124-1126 (1990). However, an early study suggesting that the .beta.-protein gene existed in three copies in Alzheimer's disease patients was not confirmed. Also the pattern of expression of the .beta.-protein gene was subtly altered in Alzheimer brain but not simply overexpressed (see Tanzi, R. E., et ., Science 235:880 (1987); Neve, R. L., et al., Neuron, 1:669-677 (1988); Palmert, M. R. et al., Science. 241:1080-1084 (1988); Higgins, G. A., et al., Proc. Natl. Acad. Sci., 85:1297-1301 (1988); and Golde, T. E., et al., Neuron 4:253-267 (1990)) for data and discussion of Alzheimer- and Down syndrome-specific changes in the expression of the several .beta.-protein precursor mRNAs). Finally, the actual location of a potential Alzheimer's disease mutation on chromosome 21 in some families was soon shown to be far from the .beta.-protein gene itself and closer to the centromere (Tanzi, et al., Nature 329:156-157 (1987); Van Broeckhoven, et al., Nature 329:153-155 (1987); Goate, et al., Lancet February 18:352 (1989)).

The finding in some families showing no linkage to any marker on chromosome 21 suggests that the inherited form of Alzheimer's disease is genetically heterogeneous (Schellenberg, et al., Science 241:1507-1510 (1988); St. George-Hyslop, et al., Nature 347:194-197 (1990)). Chromosome 19 (Roses, et al., Adv. Neurol. 15:185-196 (1990)) and possibly the region -of chromosome 14 now known to be close to the ACT gene (Weitkamp, et al., Amer. J. Hum. Genet. 35:443-453 (1983); Rabin, et al., Somatic Cell Mol. Genet. 12:209-214 (1986)) have been proposed as candidate locations for the disease locus. These results indicate that an aberrant biochemical pathway leading to the Alzheimer neuropathology can be initiated by mutation in a number of genes, including one on chromosome 21, but not the .beta.-protein precursor gene itself.

Chromosome 21 was further implicated in the etiology of Alzheimer's disease by the discovery that some families in which Alzheimer's disease is inherited as an autosomal dominant mutation produce a significantly higher-than-normal number of Down syndrome children (Heston and Mastri, Arch. Gen. Psychiatry 34:976-981 (1977); Heston, et al., Arch. Gen. Psychiat. 38:1084-1090 (1981); Heyman, et al., Ann. Neurol. 14:507-515 (1983)). In the first study, the total number of Down's cases was 11 out of 3,044 Alzheimer's disease relatives. The mothers' ages at the birth of their children were given as 21, 26, 30, 30, 35, 39, 40, 41 and 46, plus two unknown ages. The average maternal age (34) is higher than the average maternal age for all births, which is approximately 29. However, the results would still be significant if, for instance, the children of the two or three oldest mothers were not considered (the frequency of Down syndrome is 1.3 per thousand live births to mothers of all ages). The number of relatives analyzed in the second study was 1278, including 4 Down syndrome individuals conceived to mothers of ages 26, 31, 33, and 38. These numbers are small and the average maternal age is a little high, but even if the one case of age 38 were artificially removed, the results are also statistically significant. In contrast, other researchers have failed to confirm the increased incidence of Down syndrome in families with inherited Alzheimer's disease, but they report that the number of relatives they analyzed was too few for the lack of Down syndrome to be statistically significant (Whalley, et al., Brit. J. Psychiat. 140:249-256 (1982); Amaducci, et al., Neurology 36:922-931 (1986); Chandra, et al., Neurology 37:1295-1300 (1987)).

Recently, mouse chromosome 16, which is partially homologous to human chromosome 21, including the .beta.-protein gene, has been shown to result, when trisomic, in neurodegeneration somewhat like that seen in Alzheimer's disease (Richards et al., EMBO J. 10:297 (1991)). Because mouse chromosome 16 is much larger and contains many more genes than does human chromosome 21, trisomy 16 mice suffer many developmental abnormalities and do not survive to term. However, the specific effect of this trisomy on the nervous system can be tested by transplanting embryonic brain tissue from a trisomy 16 embryo into the brain of a normal adult. When the brains of such host mice with their trisomy 16 grafts were examined, it was found that some of the neurons in the graft had accumulated aberrant immunoreactivity similar to that found in and around degenerating neurons in Alzheimer's disease.

Specifically, thioflavin S, a histological marker for amyloid, showed positive staining within a few cells and around some blood vessels. In addition, antisera to the .beta.-protein precursor, to .beta.-protein itself, to ACT, to PHF, and to phosphorylated epitopes of tau labeled a few percent of cells in the trisomy 16 grafts. There was also some extracelular staining for ACT and .beta.-protein precursor. When dissociated cells from trisomy 16 embryos were transplanted, the effects were not observed, suggesting that cell-cell interaction or cell degeneration in the bulk trisomy 16 tissue grafts used by Richards and her colleagues may be necessary for the neuropathology to develop. Interestingly, although .beta.-protein precursor RNA was overexpressed approximately two-fold in trisomy 16 fetal mouse brains, it was overexpressed five-fold in the brains of chimerio (mosaic) mice having 40-50% trisomy 16 cells, again suggesting that a complex cell-cell interaction, affects the expression of this gene (Holtzman, et al., Soc. Neurosci. Abs. 469.6 (1990)).

The most recent link between Alzheimer's disease and chromosome 21 is evidenced by reports of two women whose lymphocytes were found to be mosaic for trisomy 21 and who, though not mentally retarded, had developed Alzheimer-like dementia by age 40 (Schapiro, et al., Neurology 39 Suppl. 1:169 (1989); Rowe, et al ., Lancet July 22:229 (1989); for discussion, see Hardy, et al., Lancet September 23:743 (1989)). In one case the woman also had a Down syndrome child. An unusual family with an inherited aberrant chromosome 22-derived marker chromosome was found by Percy M. E., et al., Am. J. Med. Genet., (1991) to also have a high frequency of Alzheimer's disease. The two living affected members of the family carried the marker chromosome and one was also found to be mosaic for trisomy 21; only lymphocytes were analyzed from the other patient (see relevant discussion below). The two patients reported by Schapiro and Rowe and their colleagues and possibly the mosaic individual reported by Percy and her colleagues demonstrate that it is not necessary for every cell of an individual to be trisomy 21 for the aberrant effects of this chromosome imbalance to result in early Alzheimer dementia. The later onset dementia of classic Alzheimer's disease could thus result from an even smaller percentage of trisomy 21 cells that may go undetected.

Our proposal, that Alzheimer's disease and Down syndrome result from unequal chromosome 21 segregation in somatic and germ cells respectively, reconciles a seemingly diverse body of literature. For instance, one immediate implication is that any genetic or environmental factor that increases the chances of forming chromosome 21 trisomic cells should increase the likelihood of developing Alzheimer's disease. Thus, in the families in which the disease is apparently inherited as an autosomal dominant mutation near the centromere of chromosome 21, the mutation probably resides in the centromere itself so as to cause an increased frequency of nondisjunction of chromosome 21. During mitosis, such nondisjunction would build up trisomy 21 somatic cells, eventually leading to Alzheimer's disease pathology, while during meiosis it would generate trisomy 21 germ cells and Down syndrome offspring, as consistent with the epidemiological evidence. Indeed, there are centromere mutations known in yeast that result in a 100-fold increase in chromosome nondisjunction (Gaudet and Fitzgerald-Hayes, Mol. Cell. Biol. 7:68-75 (1987)).

Of course chromosome segregation is a complex process under the control of many gene products (for review, see Murray and Szostak, Annu. Rev. Cell. Biol. 1:289-315 (1985)), and an inherited disorder of chromosome segregation could be caused by mutations at a number of loci. In this light, the fact that familial Alzheimer's disease appears to be genetically heterogeneous is not surprising, since any one of several mutations could lead to the development of trisomy 21 cells, both somatic and germline, with the consequent development of Alzheimer's disease in the individual and an increased frequency of Down syndrome offspring. Several researchers have suggested that a specific microtubule defect could lead directly to the neuronal pathology and indirectly to the increase in Down's offspring in Alzheimer's disease through chromosome nondisjunction (Heston and Mastri, Arch Gen. Psychiatry 34:976-981 (1977); Nordenson, et al., Lancet March 1,:481-482 (1980); Matsuyama and Jarvik, Proc Natl. Acad. Sci. 86:8152-8156 (1989).

Although improper chromosome segregation can result from a genetic mutation, it can also be caused by environmental agents. Of the many exogenous factors that influence chromosome segregation, microtubule-disrupting agents such as colchicine and low doses of radiation are perhaps the best studied (see, for example, Uchida, et al., Am. J. Hum. Genet. 27:419-429 (1975)). Aluminum, the consumption of which shows a weak, but significant association with the development of Alzheimer's disease (see, for example, Martyn, et al., Lancet 14 January 59-62 (1989)), also binds to microtubules and, in the form of aluminum silicate, causes chromosome nondisjunction in cultured cells (Paleker, et al., Carcinogen 8:553-560 (1987); for discussion see Ganrot, Environ. Health. Persp. 65:363-441 (1986)). Thus, the large proportion of Alzheimer's disease cases that arise in a sporadic manner not directly attributable to the inheritance of a genetic mutation can also be understood in the light of the chromosome 21 trisomy model.

An important prediction of this model is that it is the dividing cells in an individual that are most likely to develop chromosome 21 trisomy and lead to Alzheimer's disease. Extensive analysis by Rakic, Science 227:1054-1056 (1985) has shown that the only dividing cells in the brains of adult monkeys exposed to .sup.3 H-thymidine are glial cells and the endothelial cells lining blood vessels, while neurons, the cells most apparently affected by Alzheimer's disease, do not divide. The labeled glia were seen primarily in the hippocampus and the cerebral cortex. Thus cell division in the brains of adult primates occurs in those general regions that develop neuropathology in Alzheimer's disease, Down syndrome, and normal aging. Interestingly, astroglia in the hippocampus and cortex of Alzheimer's disease brain overexpress ACT, and astrocytes can be induced by kainic acid lesions to overexpress the B-protein precursor (Pasternack, et al., Am. J. Pathol. 135:827-834 (1989); Siman, et al., Neuron 3:275-285 (1989); for discussion of how overexpression of ACT or B-protein precursor can lead to amyloid formation, see Abraham and Potter, Biotechnology :147-153 (1989)).

Recently, two rapidly dividing peripheral tissues (skin and intestinal mucosa) have been reported to contain pre-amyloid deposits of .beta.-protein in sporadic Alzheimer patients and some aged, normal subjects (Joachim, et al., Nature 341:226-230 (1989)). Another region of active cell division, which has been shown to exhibit pathological changes in Alzheimer's disease, is the olfactory epithelium (Talamo et al., Nature 337:736-739 (1983)). Thus there seems to be a rough correlation between regions of cell division and areas where Alzheimer pathology can develop. Of course, mitotic nondisjunction could also occur early enough in embryogenesis to generate trisomy 21 in nondividing adult cells such as neurons.

Although it would seem reasonable that amyloid should develop in the regions, immediately surrounding aberrant cells (for instance trisomy 21 cells), the precedent provided by other amyloidoses suggests that this need not be the case. For instance, the autosomal-dominantly inherited diseases Familial Amyloidotic Polyneuropathy and Hereditary Cerebral Hemorrhage with Amyloidosis of both the Dutch and Icelandic types have very specific regions of amyloid deposition despite the fact that all cells in the body carry the point mutation in the affected amyloid gene (transthyretin, cystatin C, or .beta.-protein precursor, respectively), and that these genes are expressed in many parts of the body where the amyloid does not deposit (for review see Castano and Frangione, Lab. Invest. 58:122-132 (1988)). Thus, by analogy, the trisomy 21 cells that are relevant for the formation of amyloid pathology in Down syndrome (and, according to the hypothesis, Alzheimer's disease) need not reside in the brain at all. Indeed, some researchers believe that the .beta.-protein is transported to the brain by the circulation, having been generated elsewhere (see, for example, Selkoe, Neurobiol. Aging 10:873-877 (1989) for recent data and discussion).

In sum, both genetic and sporadic forms of Alzheimer's disease can be explained as arising from the effects of trisomy 21 cells accumulating during the life of the individual. A propensity to develop such cells can be genetic in origin (either due to an aberrant chromosome 21 centromere or to a mutation elsewhere in the genome affecting all chromosome segregation), or it can be caused by environmental factors. A combination of genetic and environmental influences on the formation of trisomic 21 cells is responsible for the observed variation in the age of onset of Alzheimer's disease in identical twins and in Alzheimer's disease families. In addition, the fact that almost 50% of the population over the age of 85 show some symptoms of Alzheimer's disease dementia (Evans, et al., JAMA 262:2551-2556 (1989)), and an even larger proportion show some of the same neuropathological lesions, indicates that all individuals may, to some degree, be subject to stochastic events that lead to aberrant chromosome segregation with increasing age. The possibility that further biochemical or genetic events may be required before full Alzheimer neuropathology arises is indicated by the mature age (20's to 30's) that Down syndrome patients begin to accumulate amyloid deposits.

Cytogenetic analysis of Alzheimer's disease patients has been carried out in a number of laboratories, with mixed reports of increased aneuploidy or other abnormalities as measured directly (Jarvik, et al., Arch. Gen. Psychiat. 30:186-190 (1974); Ward, et al., Am. J. Med. Genet. 3:137-144 (1979); Nordenson, et al., Lancet March 1:481-482 (1980); White, et al., Am. J. Med. Genet. 10:77-89 (1981); Buckton, et al., J. Med. Genet. 20:46-51 (1983); Moorhead and Heyman, Am. J. Med. Genet. 14:545-556 (1983)). Furthermore, premature centromere division (PCD), a correlate and potential cause of improper chromosome segregation in vitro and in vivo, was found to be positively correlated with age and to be increased in women with familial Alzheimer's disease (3.6% vs. 0.6% in age-matched controls), particularly affecting the X chromosome (Fitzgerald, et al., Ann. Hum. Genet. 38:417-428 (1975); Moorhead and Heyman, Am. J. Med. Genet. 14:545-556 (1983)). Trisomy 21, 18, and X occurred in the lymphocytes and fibroblasts of a woman apparently prone to PCD, who also had three trisomy conceptuses (Fitzgerald, et al., Hum. Genet. 72:58-62 (1986)). Patients with Roberts syndrome, a rare autosomal recessive disorder characterized by growth and mental retardation and craniofacial abnormalities, also shows PCD--there can be significant aneuploidy, usually involving chromosome loss rather than gain (except for trisomy 7) (see Romke, et al., Clin. Genet. 31:170-177 (1987), Jabs et al., Proc. Clin. Biol. Res. 318:111 (1989)). PCD can also be found that appears limited to the X chromosome and results, presumably by nondisjunction, in many cells with one or three X chromosomes (Fitzgerald, et al., Ann. Hum. Genet. 38:417-428 (1975)). Other Roberts syndrome families with extensive PCD have been found that exhibit normal karyotypes and phenotypes (Madan, et al., Am. Genet. 77:193-196 (1987)). Thus PCD need not result in nondisjunction, but when it does, severe developmental abnormalities can result if the autosomes are affected. The report by Fitzgerald and his colleagues (1986) is the only case of PCD in which trisomy of chromosomes other than the X chromosome were prevalent. Why Roberts syndrome generally results in chromosome loss rather than gain is not clear. Perhaps because of the severe mental retardation exhibited by these patients, neurological and pathological tests for Alzheimer's disease have not been reported.

The fact that the chromosomes that exhibit PCD in an individual do not necessarily correspond to those which ultimately are lost or gained to give an aberrant karyotype (there is a prevalence of trisomy 21, 18 and X in the general PCD case of Fitzgerald, et al., Hum. Genet. 72:58-62 (1986)), probably reflects differential cell viability. For instance, lymphocyte cultures from trisomy 21 mosaic individuals often show a lower proportion of trisomy cells than do, for instance, fibroblast cultures, and some patients with over 10% trisomy fibroblasts can show a normal karyotype in lymphocytes (out of, for example, 30 metaphases) (Pagon, et al., Am. J. Hum. Genet. 31:54-61 (1979); Ford, Trisomy 21, Springer, Berlin, Heidelberg and New York, pp 103-143 (1981)).

Thus, the fact that cytogenetic studies on Alzheimer patients have almost always relied on peripheral blood lymphocytes (for example Jarvik, et al., Arch. Gen. Psychiat. 30:186-190 (1974); Ward, et al., Am. J. Med. Genet. 3:137-144 (1979); Nordenson, et al., Lancet March 1:481-482 (1980); White, et al., Am. J. Med. Genet. 10:77-89 (1981); Buckton, et al., J. Med. Genet. 20:46-51 (1983); Moorhead and Heyman, Am. J. Med. Genet. 14:545-556 (1983)) may have prevented trisomy 21 mosaicism from being detected and linked to Alzheimer's disease. In these studies, fewer than 100 or even 50 lymphocyte metaphases per sample were examined, and the few percent with increased aneuploidy (generally a loss) for any chromosome in Alzheimer's disease was usually not significantly different from controls. The specific frequency of trisomy 21 was too low to be useful or was not stated.

The mechanism proposed herein can be tested by analysis of dividing cells from affected areas of the brain--glia, endothelial cells of the meningeal and cortical vessels, and the olfactory epithelium and possibly of skin fibroblasts. Procedures have recently been developed (Lichter, et al., Proc. Natl. Acad. Sci. U.S.A. 85:9664-9668 (1988); Fuscoe, et al., Genomics 5:100-109 (1989)) that allow the number of chromosome 21s to be counted in both metaphase and interphase nuclei. The methodology is based on in situ hybridization with, for instance, biotin-labeled chromosome 21-specific probes, that are then visualized by fluoresceine-labeled streptavidin. The advantage of this approach over standard cytogenetics is that both dividing and non-dividing cells can be studied, and, more important, the number of chromosomes in interphase nuclei in tissue sections can be counted (Lichter, et al., Proc. Natl. Acad. Sci. U.S.A. 85:9664-9668 (1988)). Because the number of cells that might harbor three copies of chromosome 21 in Alzheimer's brain or peripheral tissue might be very small, the finding of a small cluster of trisomy cells would be far more significant than the same number of cells found one at a time among a population of thousands of other cells after they have been disaggregated and induced to divide in culture to yield metaphase chromosome spreads. The large number of "normal" aged individuals that show some symptoms of Alzheimer's disease and some Alzheimer pathology (neurofibrillary tangles and amyloid deposits) will make it necessary to carry out careful comparisons between Alzheimer's disease patients and age-matched controls. Such an analysis would, of course, be made easier by concentrating on earlier onset (often familial) Alzheimer's cases. Although initial studies would seem to be best directed at searching for trisomy 21 cells in the brain, the possibility (discussed above) that Alzheimer amyloid deposits may arise from aberrant cells in the periphery suggests that a similar in situ hybridiztion analysis should also be carried out on various other tissues, for instance skin and the intestinal mucosa.

If Alzheimer's disease patients are found to be mosaic for trisomy 21, then we might expect them to exhibit some other abnormalities of Down syndrome in addition to dementia--for example, hypersensitivity to acetylcholine agonists and antagonists (see, Berg et al., Lancet 2:441 (1959); Harris and Goodman, N. Eng. J. Med. 279:407 (1968); Sacks and Smith, J. Neurol. Neurosurg. Psychiatry 52:1294 (1989)). Such characteristics, together with in situ hybridization for chromosome 21, form the basis of a diagnostic test for Alzheimer's disease.

Diagnosis for Alzheimer's disease before symptoms of dementia arise can be accomplished by determining whether the individual is indeed mosaic for trisomy 21. This can be carried out either directly by in situ hybridization as described above, or indirectly by measuring certain characteristics known to be associated with Down syndrome that arise due to the trisomy 21 nature of this disorder, and assay them for their presence in a suspected Alzheimer patient. Examples of such Down's patient features include Brushfield spots on the irises (Grouchy and Turleau, Clinical Atlas of Human Chromosomes, 2nd edition, Wiley, New York (1984)), hypersensitivity of heart rate, pupil contraction, and sweat production to acetylcholine agonists and antagonists (Berg et al., Lancet 2:441 (1959); Harris and Goodman, et al., N. Eng. J. Med. 279:407 (1968); Symon et al., J. Mental Defic. Res. 29:257 (1985); Reyes et al., J. Neural. Neurosurg. Psychiatr. 50:113 (1987); Sacks and Smith, J. Neurol. Neurosurg. Psychiatry 52:1294 (1989) and Inzelberg et al., Clin. Neurpharmacol. 13:241 (1990)). Many methods for monitoring hypersensitivity to cholinergic agonists and antagonists are well known in the art. In short, any characteristic of Down syndrome may be used as a basis for designing a diagnostic test for Alzheimer's disease.

Because the data indicate that Alzheimer's disease is a mosaic form of Down syndrome due to nondisjunction (but perhaps during meiosis followed by nondisjunction early in development to yield normal cells), much Alzheimer's disease may be prevented by preventing the nondisjunction from occurring, including such approaches as avoiding environmental agents that cause translocation by inducing chromosome nondisjunction, treatment with agents that reduce spontaneous nondisjunction or that obviate the effects of environmental agents, and mitotic inhibitors such as colcemid or methyl benzimidazole-2-yl-carbamate. Such treatment might include, but is not limited to, heavy metal chelaters, antoxidants, and promoters of microtubule assembly. Drugs that improve chromosome segregation will include those that affect DNA toposiomerase II (Holm et al., Mol. Cel. Biol. 9:159 (1989)), or centromere binding proteins such as CBF1 (Cai and Davis, Nature 349:704 (1991)), or DYS1 (Rockmill and Fogel, Genetics 119:261 (1988)). An additional approach may be to treat patients with drugs to which trisomy 21 cells may be particularly sensitive in the expectation that they will be preferentially killed and thus no longer pose a threat to the patient. This may be accomplished by killing cells that have an excess of certain cell surface markers known to be increased in cells from Down syndrome patients due to the trisomy 21. These include the cell surface marker S14 and interferon-.alpha. receptor.