Life-long Overexpression of S100β in Down’s Syndrome: Implications for Alzheimer Pathogenesis

Abstract

Chronic overexpression of the neurite growth-promoting factor S100β has been implicated in the pathogenesis of neuritic plaques in Alzheimer’s disease. Such plaques are virtually universal in middle-aged Down’s syndrome, making Down’s a natural model of Alzheimer’s disease. We determined numbers of astrocytes overexpressing S100β, and of neurons overexpressing β-amyloid precursor protein (β-APP), and assayed for neurofibrillary tangles in neocortex of 20 Down’s syndrome patients (17 weeks gestation to 68 years). Compared to controls, there were twice as many S100β-immunoreactive (S100β⁺) astrocytes in Down’s patients at all ages: fetal, young, and adult (p = 0.01, or better, in each age group). These were activated (i.e., enlarged), and intensely immunoreactive, even in the fetal group. There were no neurofibrillary changes in fetal or young Down’s patients. The numbers of S100β⁺ astrocytes in young and adult Down’s patients correlated with the numbers of neurons overexpressing β-APP (p < 0.05). Our findings are consistent with the idea that conditions—including Down’s syndrome—that promote chronic overexpression of S100β may confer increased risk for later development of Alzheimer’s disease.

Adult Down’s syndrome patients develop a progressive mental deterioration, beyond the mental retardation present from birth (20,22). In addition, Alzheimer-like neuropathological changes are universal among middle-aged patients with Down’s syndrome (15,22), and Down’s syndrome may thus be considered a natural model for investigations of the pathogenesis and progression of Alzheimer’s disease (4,9). Investigations of Down’s syndrome, aimed at elucidating the pathogenesis of Alzheimer’s disease, offer the opportunity to define alterations in young and even fetal patients that precede by decades the characteristic appearance of the diagnostic neuropathological changes of Alzheimer’s disease.

S100β is an astrocyte-derived neurite growth-promoting factor (10) encoded by a gene mapped to the Down’s region of chromosome 21 (1). Biologically active S100β is markedly elevated in Alzheimer brain, and this elevation is largely attributable to plaque-associated, activated astrocytes overexpressing S100β (11). S100β has been implicated in dystrophic neurite formation (7,17) and in plaque evolution (13), as well as in the evolution of neurofibrillary tangle changes in Alzheimer’s disease (19). In normal human aging, S100β expression increases progressively, and this increase has been invoked to explain in part the age-related incidence of Alzheimer’s disease (18).

To evaluate temporal patterns of S100β expression and their relationship to the appearance of Alzheimer-type neuropathological changes in Down’s syndrome, we used immunohistochemical analyses of neocortical brain tissue from 20 Down’s patients representing a spectrum of ages from 17 weeks gestation to 68 years. We find increased numbers of S100β-immunoreactive (S100β⁺) astrocytes in Down’s patients of all ages. This overexpression of S100β, which precedes by decades the appearance of characteristic Alzheimer-type neuropathological changes in Down’s syndrome, supports a role for chronic overexpression of this neurotrophic cytokine in the pathogenesis of Alzheimer’s disease.

METHODS

Patients

For these studies, tissues were collected postmortem from a group of 20 Down’s patients, ranging in age from 17 weeks gestation to 68 years, classified as fetal (7 cases, gestational ages 17, 18, 18, 21, 22, 27, and 35 weeks), young (5 cases, ages 8 months; and 4, 7, 7, and 9 years), and adult (8 cases, 22, 29, 48, 55, 61, 61, 63, 68 years). A group of 16 patients of similar age ranges (5 fetal: 20, 22, 22, 24, 34 weeks gestation; 4 young: 3 months, 6 months, 2 years, and 9 years; and 7 adult: 37, 47, 58, 59, 66, 68, 68 years), without neurological disease, was used as controls. For Down’s cases, sections from neocortex of temporal lobe (17 cases), frontal lobe (2 cases, age 22 and 48 years), or occipital lobe (1 case, age 29 years) were used. Sections from control patients were obtained from mesial temporal lobe to include parahippocampal gyrus at the level of the lateral geniculate nucleus. Routine histological evaluation of sections from control cases—stained either with hematoxylin and eosin or with silver techniques for neuritic plaques—showed sparse diffuse plaques (but no neuritic plaques) in one patient, aged 59 years, sparse neuritic plaques in the hippocampus of one patient, aged 68 years, and sparse neocortical neuritic plaques in one patient, aged 58 years. For Down’s syndrome cases, neocortical neuritic plaques were identified in all patients aged 48 years or greater.

Immunohistochemistry for S100β was performed on all of these cases. Immunohistochemistry for β-APP was successfully performed on a subset of these cases: 18 Down’s cases (all fetuses; all young; and 6 adults: ages 22, 29, 48, 55, 61, and 68 years) and 13 controls (all fetuses; 3 young: ages 3 months, 6 months, and 9 years; and 5 adults: ages 47, 58, 59, 66, and 68 years). Immunohistochemistry for Tau2 was performed on all fetal cases; on all young cases except for the 2-year control; and on 4 adult Down’s cases (ages 22, 55, 61, 68 years) and on 4 adult controls (ages 47, 59, 66, 68 years).

Tissue Preparation and Immunohistochemistry

Tissue was fixed in 20% formalin (8% formaldehyde) and processed for paraffin embedding. Paraffin blocks were serially sectioned at a thickness of 10 µm, and immunoreaction of rabbit-anti-bovine S100β (a gift from Dr. Linda Van Eldik (21)) and mouse-anti-human β-APP (clone 22C11, Boehringer-Mannheim) antibodies was performed on glass slide-mounted tissue sections as previously described (13). Sections from Down’s syndrome and control patients were simultaneously immunoreacted. Briefly, S100β and β-APP immunoreactions were carried out overnight at room temperature at dilutions of 1:300 and 1:50, respectively.

For each case, five microscopic fields (0.25 mm² each) of neocortex were analyzed in serial sections immunoreacted for either S100β, β-APP, or Tau2. Where possible, fields rich in neuritic plaques were chosen. For controls, fields of parahippocampal gyrus were chosen from sections obtained at the level of the lateral geniculate nucleus. The distribution and numerical densities of S100β⁺ astrocytes and of β-APP⁺ and Tau2⁺ neurons were determined in these sections.

The significance of differences between numbers of S100β⁺ astrocytes and of β-APP⁺ neurons in Down’s and control patients was assessed by using ANOVA followed by Fisher’s test.

RESULTS

S100β

S100β expression in cerebral cortex of Down’s patients was higher than controls for all age groups (Figs. 1 and 2). For fetal (17–35 weeks gestational age) Down’s patients, the number of S100β⁺ astrocytes was 1.7 times that found in age-matched controls (p < 0.001); for young patients with Down’s syndrome (8 months to 9 years of age), this value was 2.0 times that of controls (p < 0.05); and for adult patients with Down’s syndrome this value was 1.9 times that of controls (p < 0.005). For fetal Down’s patients, the increased numbers of S100β⁺ astrocytes were diffusely distributed in the parahippocampal cortical areas, while increases in white matter were less striking (Fig. 3). The S100β⁺ astrocytes found in Down’s syndrome were activated [i.e., enlarged, as defined by da Cunha et al. (5)], even in Down’s fetuses, but generally not overtly gemistiocytic. They had prominent, tortuous processes and were more intensely immunoreactive than those of controls (Fig. 1). In contrast, S100β⁺ astrocytes in controls were relatively small, non-activated, and weakly immunoreactive.

FIG. 1.

Photomicrographs showing S100β⁺ astrocytes in neocortex of fetal (F) young (Y), and adult (A) Down’s (DS) and control (C) patients. Bars = 15 µm.

FIG. 2.

Numerical density of S100β⁺ astrocytes in neocortex of Down’s (●) and control (○) patients. The mean values were significantly different for fetal (17–35 weeks gestation, p < 0.001), for young (8 months to 9 years, p < 0.05), and for adult (22 years to 68 years, p < 0.005) age ranges.

FIG. 3.

Whole-mount photomicrographs of hippocampus and adjacent neocortical regions from control (A, 24 weeks gestation) and Down’s (B, 18 weeks gestation) fetuses showing the distributions of S100β⁺ astrocytes. Magnification: 6X.

β-Amyloid Precursor Protein

The numerical density of β-APP⁺ neurons in young and adult patients with Down’s syndrome was significantly greater than that of controls (Figs. 4 and 5). In fetal Down’s patients, the numerical density of β-APP⁺ neurons was also higher than that found in controls, but a high degree of variation among the results precluded statistical significance. For young patients with Down’s syndrome, the number of β-APP⁺ neurons was 5 times (p < 0.02), and for adults with Down’s syndrome 12 times (p < 0.0001), that of controls.

FIG. 4.

Photomicrographs showing β-APP⁺ neurons in neocortex of young (Y) and adult (A) Down’s (DS) and control (C) patients (top four panels). Lower two panels show Tau2⁺ neurons in adult Down’s (A-DS) but not control (A-C) patients. Bars = 15 µm.

FIG. 5.

Numerical density of β-APP⁺ neurons (A) and of Tau2⁺ neurons (B) in neocortex of Down’s and control patients. For β-APP⁺ neurons, the mean values were significantly different for young (* p < 0.02), and for adult (** p < 0.0001), but not for fetal age ranges. For Tau2⁺ neurons the mean values were significantly different for adults (** p < 0.0001).

Correlation Between Numbers of S100β⁺ Astrocytes and of β-APP⁺ Neurons in Down’s Syndrome

A dual plot of numerical densities of S100β⁺ astrocytes vs. β-APP⁺ neurons showed a separation of control and Down’s patients for young and adult age groups (Fig. 6). For these Down’s patients, there was a significant correlation between the numbers of S100β⁺ astrocytes and the numbers of β-APP⁺ neurons (r = 0.75, p < 0.05). In contrast, these two parameters were not significantly correlated for control patients.

FIG. 6.

Dual plots of the numerical density of β-APP⁺ neurons vs. numerical density of S100β⁺ astrocytes for young and adult Down’s (●) and control (○) patients show a separation of the two patient populations. A regression line is shown for the Down’s syndrome values only. Values were significantly correlated for Down’s (r = 0.75, p < 0.05) but not for control patients.

Tau2⁺ Neurons

No Tau2⁺ neurons were found in any of the control patients examined. In contrast, Tau2⁺ neurons were found in all adult Down’s syndrome patients examined, even in a 22-year-old patient (Figs. 4 and 5).

DISCUSSION

We show elevated expression of S100β in Down’s syndrome patients of all ages from 17 weeks gestation to 68 years of age. The number of astrocytes expressing S100β in Down’s patients was approximately twice that of controls at all ages. There were concomitant increases in the numbers of neurons expressing β-APP that were statistically significant for young and adult Down’s syndrome patients, but not for Down’s fetuses, and these increases were correlated significantly with the increases in the numbers of S100β⁺ astrocytes in these Down’s patients.

S100β appears to be an important neurotrophic agent during normal fetal brain development, with effects on neuroblasts and glia during this period (3,16). S100β, synthesized and secreted by astrocytes, increases intraneuronal free calcium levels (2) and stimulates the growth of neuronal processes (10). Our results, showing increased S100β expression in brains of Down’s fetuses, suggests that S100β may also contribute to the pathogenesis of the dendritic abnormalities (6,14) and the mental retardation characteristic of Down’s syndrome.

A virtually inevitable neuropathological consequence of Down’s syndrome is the development of Alzheimer-type neuropathological changes in middle-aged patients (15,22). Overexpression of S100β by activated astrocytes has been implicated in the pathogenesis of these changes [for review see (8)]. In this study we found neurons expressing the neurofibrillary tangle antigen tau in all adult Down’s patients examined, but not in any fetuses or young patients with Down’s syndrome, or in control patients of any age. Thus, the overexpression of S100β shown here precedes by decades the appearance of Alzheimer-type neurofibrillary changes in Down’s syndrome.

Our findings are consistent with the ideas that S100β plays an important role in Alzheimer pathogenesis (8,12), and that chronic overexpression of S100β and of β-APP, over the course of decades, contributes to the Alzheimer-like neuropathological changes seen in Down’s patients (7). It is of particular note that the cerebral cortical expression of S100β is increased in elderly control patient populations (18). This increased S100β expression, which is most marked in control patients over the age of 60, is however much smaller than that shown here for Down’s patients. The normal age-related increase in S100β expression has been invoked to explain, in part, the age-related incidence of Alzheimer’s disease (18). In analogous fashion, chronic, life-long overexpression of S100β may contribute to the development of Alzheimer-type pathology at middle age in Down’s syndrome patients. These ideas might be directly tested using transgenic animal models that overexpress both human S100β and human βAPP.