Intraneuronal Aβ-Amyloid Precedes Development of Amyloid Plaques in Down Syndrome Open Access

Patients with Down syndrome (DS) who live beyond middle age invariably develop cerebral amyloid (Aβ) plaques and neurofibrillary tangles (NFTs), the characteristic lesions of Alzheimer disease (AD).1–11 Some of these patients, many with intelligence quotients in the moderate to severe mental retardation range, go on to develop further cognitive and functional deficits when they develop the neuropathologic changes of AD.12 Thus, the neuropathologic examination of young subjects with DS provides an opportunity to study the early stages of these lesions and their development. Several investigations have examined the development of AD changes in the brains of DS patients,7,9–11,13–15 but few have included very young DS patients. Recent studies have shown that the Aβ peptide ending at amino acid 42 (Aβ42) is the earliest form of this protein deposited in DS brains and may be seen in subjects as young as 12 years of age,15 and that soluble Aβ can be detected in the brains of DS subjects as early as 21 gestational weeks of age, well preceding the formation of Aβ plaques.16 However, the source of Aβ in both AD and DS remains controversial, and both neurons and microglia have been proposed as possible sources.17–19 

In the present study, we investigated the neuropathologic changes of AD in 14 patients with DS, ranging in age from 11 months to 56 years, 9 of whom were younger than 30 years. In particular, we focused on the presence and distribution of both intracellular and extracellular Aβ protein, NFTs, activated microglia, and reactive astrocytes in the hippocampus and cerebral cortex.

Brains for histologic examinations included 13 DS cases investigated by the Office of the Chief Medical Examiner of the State of Maryland (Baltimore, Md) and 1 case autopsied at The Johns Hopkins Hospital (Baltimore, Md) (Table). All of the brains demonstrated gross features consistent with DS, including decreased weight, abnormalities of the superior temporal gyrus, and shortening of the frontal-occipital length.9 In addition, brains from 3 nontrisomic subjects, including a 7-month-old infant who died of sudden infant death syndrome, an 8-year-old mentally retarded child, and a neurologically normal 18-year-old young adult who died of heart disease, were examined as young controls. Control tissues were free of significant autolysis and showed no signs of hypoxic or ischemic injury. These tissues were obtained from the Brain and Tissue Bank for Developmental Disorders at the University of Maryland (Baltimore, Md). As middle-aged controls, we also examined brain tissues from 10 women and 10 men, aged 40 to 52 years, who died within 24 hours of an accidental injury. Collection of these tissues was approved by The Johns Hopkins Medical Institutions Joint Committee on Clinical Investigations and by the Institutional Review Board of the University of Maryland.

Paraffin-embedded tissue sections from the hippocampus and entorhinal cortex or the adjacent inferior temporal cortex were examined in all cases; sections from frontal and temporal regions were examined in 10 DS cases and in the 3 young control cases. In the middle-aged controls, we examined inferior temporal cortex. Cerebellar sections were examined in 2 DS cases. Tissue sections were stained with Hirano silver20 and immunostained for Aβ (Aβ1–28 antibody, raised against peptide 1–28 and recognizes all forms of Aβ; dilution 1:200; a gift from Athena Neurosciences, South San Francisco, Calif), tau (1:1000; Sigma Chemical Company, St Louis, Mo), HLA-DR (1:100; Dako Corporation, Carpinteria, Calif) for detection of activated microglial cells, and glial fibrillary acidic protein (GFAP) (1:400; Dako). Sections were also immunostained with antibodies Aβ40 (R163) and Aβ42 (R165) (both a gift from Dr P. Mehta, Staten Island, NY), which recognize shorter (1–40) and longer (1–42) forms of Aβ, respectively.21 Immunostaining controls included sections from confirmed cases of AD (Consortium to Establish a Registry for Alzheimer's Disease [CERAD] criteria)22 and experiments in which the primary antibody was omitted.

The frequency of amyloid plaques and NFTs was assessed as absent, sparse, moderate, or frequent utilizing the CERAD criteria22 on the Aβ1–28- and tau-immunostained sections, respectively. Plaques were also classified as diffuse if they lacked thickened neurites, or as neuritic if they possessed these structures on the silver stains. We also evaluated for intracellular Aβ; the type of labeled cells was determined based on morphologic criteria. Reactive astrocytes were assessed on GFAP immunostains and rated as absent, mild, moderate, or severe. The relationship of reactive astrocytes with blood vessels and/or amyloid plaques was documented. Staining for HLA-DR was determined to be present or absent; if present, its relationship to structures such as plaques, NFTs, or vascular elements was documented.

The neuropathologic findings in DS patients are summarized in the Table. Intracellular Aβ immunostaining of hippocampal and cortical neurons, detected with the Aβ1–28 and Aβ40 (R163) antibodies, was present in all DS cases, including subjects as young as 1 year. This immunostaining had a distinct vesicular pattern and was strongest in the perikaryon (Figure 1), but was also present in neuronal processes. Intracellular immunostaining of neurons was detected both in the brains of young DS subjects (0–10 years of age) who had no Aβ plaques and in older subjects (17 years of age and older) who exhibited Aβ plaques (Figure 2). The proportion of neurons labeled with Aβ1–28 and Aβ40 antibodies was variable, but increased with age. Of the 2 antibodies, Aβ40 (R163) antibody revealed the strongest labeling. By contrast, intraneuronal immunostaining with Aβ42 (R165) was rarely observed in younger subjects, but was present in all 3 subjects aged 44 and older. No Aβ immunostaining was present in cerebellar neurons. In all DS cases, astrocytes displayed Aβ1–28 and Aβ40 immunostaining in their cell bodies and in processes that extended to the vicinity of blood vessels. In general, in the hippocampus and entorhinal cortex, large pyramidal neurons exhibited the greatest degree of intracellular staining. In the neocortical sections, however, astrocytes were more often stained than neurons. None of the tissues in young controls displayed intracellular Aβ immunoreactivity in either neuronal or glial cells. However, some staining with Aβ40 was noted in vascular smooth muscle cells in both DS cases and young controls.

Among the 20 middle-aged control subjects, 15 had no extracellular or intracellular Aβ deposits. We detected extracellular diffuse Aβ deposits in 4 brains, and in 3 of these brains there were subtle punctate Aβ deposits in the neuronal perikaryon. A single brain showed subtle Aβ punctate aggregates in the neuronal perikarya in the absence of extracellular deposits of Aβ. In all of these instances, the intraneuronal Aβ was subtle and less conspicuous than in DS brains.

Extracellular Aβ, occurring as perineuronal deposits or diffuse plaques (Figure 2), was first detectable at age 17 years with silver stain and all Aβ antibodies, and was present in all subjects older than 17 years. These perineuronal Aβ deposits were free of neuritic abnormality and glial or microglial activation, similar to changes reported in early stages of AD.23 The earliest neuritic plaques were identified in the hippocampus and entorhinal cortex of a 23-year-old patient and were prominent in the hippocampus and neocortex in the 3 oldest patients (>44 years) (Figure 3), all of whom also had frequent diffuse plaques. The distribution of Aβ plaques was variable from case to case. However, in general, these lesions were present earlier and in greater numbers in the hippocampus and entorhinal cortex than in neocortical areas. Sections stained with antibodies to Aβ40 (R163) and Aβ42 (R165) showed a similar distribution of lesions, but Aβ42-labeled plaques were more numerous (Figure 3, D and E). Aβ immunostaining was not present in any of the young control subjects.

Neurofibrillary tangles (Figure 3, C) were present in the hippocampus and neocortex in the 2 oldest DS patients (ages 51 and 56 years) and in the hippocampus and entorhinal cortex of the same 23-year-old patient who showed neuritic plaques. No NFTs were present in the controls.

HLA-DR immunostains showed microglia in the senile plaques of 3 DS patients. Interestingly, these were the same 3 cases that had NFTs and represented 3 of the 4 cases with neuritic plaques. HLA-DR–positive microglia were seen mainly in association with Aβ neuritic plaques (Figure 3, B) and around blood vessels.

Sections immunostained with GFAP demonstrated increased numbers of reactive astrocytes in all DS cases aged 17 years or older (Table). In the hippocampal/entorhinal cortical sections, these cells were most prominent in the CA4 region, as well as in the subpial zone and in cortical layer V of the entorhinal cortex. In neocortical regions, GFAP-positive astrocytes were most prominent subpially, in white matter, and around blood vessels. In brains with Aβ plaques, astrocytes were associated with both diffuse and neuritic plaques.

The principal observation in this study is the presence of intraneuronal Aβ immunostaining in the brains of DS patients many years before the development of senile plaques, a finding consistent with the report that soluble Aβ is present in the brains of DS patients as early as 21 gestational weeks, but not in controls.16 Our observations suggest the following sequence of events in the development of the neuropathologic changes of AD in DS. The earliest abnormality is the intracellular accumulation of Aβ in neurons and astrocytes, noted as early as 1 year of age. Subsequently, diffuse Aβ plaques appear in the late teenage years, and neuritic plaques follow a few years later in the early 20s. Neurofibrillary tangles can appear in the early 20s in the hippocampus and entorhinal cortex, but develop later in the neocortex, usually after 40 years of age. The appearance of microglia in senile plaques is a late event in DS, and it coincides with the development of neuritic changes in Aβ plaques and of NFTs.

Our findings on the development of AD lesions in DS are in accordance with those of previous reports,2–5,7,9,10 confirming that these lesions affect the hippocampus and entorhinal cortex before the neocortex, that Aβ plaques appear before NFTs,2,4 and that significant numbers of Aβ plaques may be found in occasional individuals with DS younger than 20 years.7,9–11 The principal and novel finding of our study is the detection of intracellular Aβ in neurons and astrocytes in DS brain, preceding by years or decades the deposits of Aβ in the neuropil. Intraneuronal amyloid has been reported in the brains of old monkeys24 and in late-stage AD,25 but not in the brains of young individuals. Furthermore, our findings suggest that neurons are a source of soluble Aβ. The detection of Aβ in astrocytes may reflect either increased synthesis of Aβ by these cells or enhanced clearance of the peptide by glial cells. The pattern of immunostaining of intracellular Aβ suggests that this peptide is contained in vesicular compartments, such as endoplasmic reticulum, Golgi, endosomes, or lysosomes, all of which have been implicated in the generation of the Aβ peptides.26,27 An important question is whether intracellular Aβ accumulation may occur in nontrisomic individuals in the early stages of AD. Such accumulations have been reported in the brains of subjects with mild cognitive impairment, suggesting that intracellular Aβ42 accumulation is an early event in neuronal dysfunction in AD.28 From our observations on the brains of the middle-aged controls, it appears that intracellular Aβ accumulation may also occur in nontrisomic individuals. Furthermore, we have observed intraneuronal Aβ in the brain of a 47-year-old nondemented woman whose mother had autopsy-confirmed AD22 (J.C.T. et al, unpublished observations, December 1998). In concert, these observations suggest that intracellular Aβ accumulation is not limited to DS, but is also present in sporadic AD. The enrichment of intracellular compartments with Aβ peptides may be the consequence of overexpression of Aβ precursor protein, whose gene is on chromosome 21.7 

The present observations in DS are relevant to the pathogenesis of AD in several ways. First, they support the notion that the generation of Aβ is, at least in part, an intraneuronal process. Second, they suggest that the brain possesses clearance mechanisms that can compensate for the abnormal generation of Aβ for many years before this peptide becomes deposited in the neuropil. Finally, it appears that once Aβ is deposited in the neuropil, it triggers a cascade of events that include microglial activation, neuritic abnormalities, and neurofibrillary tangles.

The authors acknowledge the technical assistance of G. Rudow, M. Peper, and P. Boylan, and the editorial assistance of E. Mosmiller. Robin-Anne V. Ferris, MFS, provided photographic support at the Armed Forces Institute of Pathology, Washington, DC. L. Martin, PhD, and B. Crain, MD, PhD, offered helpful discussions of the manuscript. Tissues were obtained from the Brain and Tissue Bank for Developmental Disorders at the University of Maryland, Baltimore, Md (National Institute of Child Health and Human Development contract N01-HD8–3283). This investigation was supported by the Johns Hopkins University Alzheimer's Disease Research Center (National Institutes of Health AG 05146).