Share

August 01, 1995; 45 (8) ARTICLES

Altered expression of transforming growth factor-beta in Alzheimer's disease

K. C. Flanders, C. F. Lippa, T. W. Smith, D. A. Pollen, M. B. Sporn
First published August 1, 1995, DOI: https://doi.org/10.1212/WNL.45.8.1561
Full PDF
Citation
Permissions

Make Comment

See Comments

Downloads
53

Share

Abstract

We compared immunohistochemical expression of the transforming growth factor-beta s (TGF-beta 1, TGF-beta 2, and TGF-beta 3) using brain tissue from patients with nondominantly inherited Alzheimer's disease (NDAD) (n equals 9), autosomal dominantly inherited Alzheimer's disease with linkage to 14q24.3 (FAD-14) (n equals 4), and cognitively normal controls (n equals 10) to determine whether their pathologic changes are associated with an altered distribution of the TGF-beta s. We found increased expression of TGF-beta 2 in large, tangle-bearing neurons with widespread staining of glia in NDAD and FAD-14 patients compared with control cases. This result was confirmed with sandwich ELISA assays of brain tissue, which showed TGF-beta 2 levels in AD and NDAD to average 3.2 times the average level of control cases. Despite proximity of TGF-beta 1 and TGF-beta 3 to the sites of susceptibility loci on chromosomes 19 and 14, we did not find that TGF-beta 1 and TGF-beta 3 were selectively altered in any AD subtypes. However, selective induction of TGF-beta 2 may occur in NDAD and FAD-14.

The transforming growth factor-beta s (TGF-beta s) regulate cell growth and differentiation in common with other neurotrophic factors, and also play a major role in tissue repair following injury. [1] In the brain, TGF-beta can exert a neurotropic effect by working synergistically with nerve growth factor and by increasing nerve growth factor production. [2] TGF-beta has also been shown to improve neuronal survival in vitro. [2] Additionally, TGF-beta 1 expression is increased in acute CNS injury [3-5] and it may protect the brain against neuronal degeneration from hypoxic injury and glutamate neurotoxicity. [6]

Genes for two of the three TGF-beta s are close to Alzheimer's disease (AD) susceptibility genes. TGF-beta 3 is located on chromosome 14q24.3, [7] the same site linked to the majority of familial AD (FAD) lineages. [8,9] TGF-beta 1 is located [10] on chromosome 19q13.1-3 near a susceptibility gene for late-onset AD. [11]

Moreover, there are physiologic reasons to suspect that the TGF-beta s might influence the course of AD. Brains of AD patients contain beta-amyloid (A beta) in the form of plaques in the neuropil and in cerebral blood vessels. Several point mutations within the A beta portion of the amyloid precursor protein gene (beta PP) on chromosome 21 result in AD. [12-18] A six-fold increase in the production of beta PP (with a shift from the benign beta PP 695 isoform to the more toxic beta PP 751 and beta PP 771 isoforms) is seen after treatment of cultured astrocytes with TGF-beta 1. [19] Furthermore, a stable complex is formed between TGF-beta 2 and soluble A beta, [20] and one study [21] has reported the immunohistochemical localization of TGF-beta in plaques of AD and Down's syndrome patients. To evaluate the expression of TGF-beta in the brains from patients with nondominantly inherited Alzheimer's disease (NDAD) and FAD, we compared the immunohistochemical staining patterns of the three human TGF-beta s in NDAD and FAD patients with those of individuals without dementia.

Methods.

Patients.

Control patients.

We obtained the brains from 10 adult individuals of various ages who were cognitively intact and had no other neurologic disorders prior to death (mean age 64.0 years at death) Table 1. Histories were obtained by medical record review. All control patients had clear documentation of a normal mental status up to the time of their death.

View this table:

Table 1. Clinical features of Alzheimer's disease and control patients

Nondominantly inherited Alzheimer's disease patients.

We obtained brains from nine patients who met the National Institute of Neurological and Communicative Disorders and Stroke and the Alzheimer's Disease and Related Disorders Association (NINCDS/ADRDA) clinical criteria for ``probable Alzheimer's disease'' [22] (mean age 67.4 years) who had no affected first-degree relatives. All patients except one (case 1) were in nursing homes with end-stage disease, requiring total care for all activities at the time of death. Patient 1 had mild dementia, which was well documented. He was living at home and still able to play golf when he died of an acute myocardial infarction. All brains from this group met the Consortium to Establish a Registry for Alzheimer's Disease (CERAD) pathologic criteria for AD. [23]

The gene for apolipoprotein E (APOE), the major known genetic risk factor for AD, is located near the TGF-beta 1 gene on chromosome 19q13. [11] There is no evidence that abnormalities of the TGF-beta 1 gene are involved in any form of AD. Despite this, APOE genotype might differentially affect the staining for TGF-beta 1 in a manner similar to APOE's reported effect on the intensity of A beta staining. [24] We therefore determined whether there is evidence of a relationship between APOE genotype and TGF-beta 1 immunostaining by examining the nine AD cases described above whose APOE genotype was known Table 1 and three additional AD cases with known APOE genotypes.

Alzheimer's disease patients with chromosome 14 mutations.

We obtained the brains from four individuals with early-onset dementia (mean age 60.5 years at death) who belong to an FAD lineage with a genetic abnormality on chromosome 14q24.3 (FAD-14). [8] Two patients (cases 10 and 11) were in the process of evaluation for nursing home placement at the time of death; the other two patients (cases 12 and 13) had advanced dementia and required skilled nursing home care for several years prior to their death.

Tissue processing.

The brains were removed after an average postmortem delay of 7.3 hours in NDAD, 4.7 hours in FAD-14, and 10.3 hours in controls. Immediately following removal, the right hemisphere was cut in the coronal plane at 1-cm intervals and fixed either in 10 to 15% formalin for 2 weeks or in a modified paraformaldehyde solution. [25] After fixation, two tissue blocks from the medial temporal lobe (one at the level of the amygdala and the other at the rostral body of the hippocampus) and one tissue block each from the frontal cortex and from the cerebellar cortex were embedded in paraffin and sectioned at a thickness of 5 micro meter. In one NDAD case (case 7), 16 representative regions throughout the brain were prepared. Sections from the left hemispheres were stained with hematoxylin and eosin (HE) and Campbell's silver stain [26] for independent diagnostic confirmation by a neuropathologist.

To evaluate TGF-beta expression in frozen tissue, we studied tissue from six AD cases (cases 6, 8, and four additional cases meeting NINCDS/ADRDA clinical and CERAD pathologic criteria for AD) using frontal cortex blocks in liquid nitrogen after fixation in a modified paraformaldehyde solution and cryoprotection in graded sucrose solutions. After sectioning at 6 to 7 micro meter, we stored the tissue at minus 70 degrees C pending immunostaining. Immunocytochemistry on these sections was performed using similar procedures with anti-TGF-beta 1LC (see ``Immunohistochemical staining'' below) at a concentration of 1 mu g IgG/ml.

We froze one unfixed frontal cortex block at minus 70 degrees C for the sandwich ELISA (SELISA) assay from each of six of the AD cases in the table (cases 2, 8 to 11, and 13), control cases 16 to 23, and three additional cases meeting NINCDS/ADRDA clinical and CERAD pathologic criteria for AD.

Immunohistochemical staining.

Immunohistochemical staining with rabbit polyclonal antibodies raised to synthetic peptides corresponding to regions of the mature form of TGF-beta 1 (developed in the Laboratory of Chemoprevention [LC] and in the Collagen Corporation [CC]), TGF-beta 2, or TGF-beta 3 was used for determination of neuronal staining. The TGF-beta 1 (LC) and TGF-beta 1CC antibodies were raised to different preparations of a peptide corresponding to the N-terminal 30 amino acids of TGF-beta 1, and they recognize different epitopes of this peptide. [27] The LC antibody usually stains intracellular TGF-beta 1 while the CC antibody stains extracellular TGF-beta 1. Anti-TGF-beta 2 was raised to peptides corresponding to amino acids 50 to 75 of TGF-beta 2, and anti-TGF-beta 3 was raised to peptides corresponding to amino acids 50 to 60 of TGF-beta 3. [28,29] These antibodies were affinity-purified against either protein A-Sepharose (anti-TGF-beta 1CC), TGF-beta 1 (anti-TGF-beta 1LC), or the immunizing synthetic peptides (anti-TGF-beta 2 and anti-TGF-beta 3) and have been assayed for specificity by Western blot analysis. [27-31] All antibodies reacted with the appropriate TGF-beta isoform except anti-TGF-beta 1CC, which showed some cross-reactivity with TGF-beta 3 on Western blots, so staining of TGF-beta 3 in tissue sections cannot be ruled out. These antibodies have previously been shown to give specific staining in the mouse and rat nervous system. [29,31]

Following deparaffinization, sections were treated with H2 O2 (to block endogenous peroxidase), hyaluronidase (to improve antibody penetration), and normal goat serum (to block nonspecific binding). After initially incubating sections with anti-TGF-beta antibodies (IgG concentrations of 1 mu g/ml for anti-TGF-beta 1LC, 2 mu g/ml for anti-TGF-beta 1CC, 2 mu g/ml for anti-TGF-beta 2, and 1 mu g/ml for anti-TGF-beta 3) overnight at 4 degrees C, the adherent antibodies were reacted with biotinylated goat anti-rabbit secondary antibodies bridged to biotinylated horseradish peroxidase with avidin (Peroxidase ABC Elite Kit, Vector Labs). We visualized the TGF-beta-antibody complexes using 3,3 prime-diaminobenzidine as a chromogen. Optimal dilution and incubation times for each antiserum have been previously established. [27] An antigen retrieval technique [32] was used to enhance antigenicity for TGF-beta 2.

For each histologic section in the study, we ran a successive section in parallel using normal rabbit IgG at 3 mu g/ml in place of the primary antibody. Control tissue was also run in which the primary antibody was preincubated with a 50-fold molar excess of immunizing peptide before being applied to the section. To evaluate gliosis, sections from the medial temporal lobe block containing the entorhinal cortex and amygdala were stained with monoclonal antibodies for glial fibrillary acidic protein (GFAP) (Biogenex, StrAviGen, 1:1,200). To estimate the percent of neurofibrillary tangle (NFT)-bearing neurons that stained for TGF-beta, we stained a section containing frontal cortex with mouse monoclonal antibodies directed against tau-2 (Sigma, StrAviGen, 1:750) for each of our cases.

Sandwich ELISA assays.

A SELISA assay for TGF-beta 2 was performed on acid-ethanol extracts of frozen brain tissue. Each gram of tissue was homogenized in 4 ml of cold ETOH:H2 O:HCl (47.5:2.5:1) and extracted overnight. Following extensive dialysis against 4 mM HCl, samples were lyophilized, resuspended in 1/10 volume 4 mM HCl containing 1% Triton X-100, and tested in a SELISA [33] that specifically detects TGF-beta 2. Briefly, Nunc Maxisorp microtiter ELISA samples were coated with affinity-purified rabbit anti-TGF-beta 2 IgG overnight. Wells were blocked with 1% bovine serum albumin for 1 hour and washed. Then 100-mu l samples were added to wells and serially diluted in binding buffer in the wells. Samples were assayed in duplicate at four dilutions. Following 1-hour incubation at room temperature and washing, wells were coated with 100 mu l of affinity-purified turkey anti-TGF-beta 2 and incubated for an additional hour. After washing, phosphataselinked goat anti-turkey IgG was added, and following a 1-hour incubation and additional washes, color was developed with p-nitrophenylphosphate in ethanolamine buffer for 16 hours at 4 degrees C. For quantitation of unknown concentrations of TGF-beta 2, data points (difference in absorbance at 410 nm and 450 nm) were transferred to a log-logit function and fitted to a quadratic or cubic regression Equation basedon a TGF-beta standard curve. Results are expressed as ng TGF-beta 2/g tissue plus minus SEM.

Morphometry.

When neuronal immunostaining with any of our four TGF-beta antibodies was present, quantitative data were obtained in runs of 10 consecutive fields from two different premarked sites in the frontal cortex, one premarked site in the entorhinal cortex and CA1 sector of the hippocampus, and one run of consecutive fields across the lateral-to-medial extent of the amygdala at its point of widest girth. All quantitative data were obtained at a magnification of 200 times (0.92-mm2 fields). The final data were expressed as counts per mm2.

When glia and blood vessels showed TGF-beta staining, such staining was graded on a 0 to 3 semiquantitative scale where 0 indicates no staining, 1 indicates trace staining, 2 indicates moderate staining, and 3 indicates intense TGF-beta immunostaining.

Statistical analysis.

Statistical analysis for density data (TGF-beta-positive neurons) was performed using a one-way ANOVA. All data significant at the 0.05 level or greater were further evaluated using the Tukey-Kramer post hoc analysis to determine which groups were significantly different from each other. This method takes into account multiple comparisons to minimize the likelihood of type 1 error. Nonlinear (semiquantitative grading of glial and vascular staining on a 0 to 3 scale) data were analyzed using the Kruskal-Wallis one-way ANOVA by ranks. When data were significant at the 0.05 level or greater, post hoc pairwise comparisons between groups were made using the Kruskal-Wallis method to determine which groups differed. This method of analysis also accounts for multiple comparisons.

Results.

Microscopic examination.

The microscopic examinations of H-E- and silver-stained sections from all patients with NDAD and FAD-14 revealed numerous A beta plaques, NFT, and neuropil threads throughout the cerebral cortex and limbic structures. A beta plaque densities well exceeded the CERAD criteria for the diagnosis of AD in all NDAD and FAD-14 cases. [23] No control brains had neuritic plaques. There were no Lewy bodies, Pick bodies, or other pathologic features suggesting concurrent diagnoses except for marked amyloid angiopathy, which was given as an additional diagnosis in cases 2, 4, 5, 10, 11, and 13. Except where mentioned below, postmortem delay and fixative did not appreciably change the tissue staining pattern. Tissue stained with normal rabbit serum as a control and tissue stained with antibody plus blocking peptide showed staining of some Hirano bodies but no significant background or other cellular staining. All control brains were histologically normal.

Qualitative and quantitative analysis.

TGF-beta 1CC. Control subjects.

The most striking feature of TGF-beta 1CC immunostaining was perineuronal staining of large pyramidal neurons in layers III and V of the frontal cortex and of cerebellar dentate neurons; this pattern of staining was occasionally seen in large and medium-sized neurons in other brain regions. The intensity of rim staining in control brains was highly variable, but all control cases showed at least one neuron with ``rim'' staining. The staining appeared to be either within or adjacent to the cell membrane and it sometimes extended along proximal cell processes. It was reminiscent of synaptophysin staining of synaptic terminals surrounding cell bodies and proximal neuronal processes in gangliogliomas. [34] There was no intracytoplasmic or nuclear staining of affected neurons. There was no clear relationship between perineuronal staining and the patients' age at death, postmortem delay, or choice of fixative.

Anti-TGF-beta 1CC also showed fine granular staining in the neuropil and the white matter in control cases. Subpial regions occasionally stained with anti-TGF-beta 1CC, and in aged control cases the corpora amylacea stained intensely. There was no TGF-beta 1CC glial staining, and vascular staining, when present, was minimal.

Dementia subjects.

The perineuronal staining in control cases was seen in all NDAD and FAD-14 patients. Rim staining involved some neuronal populations affected in AD, including large pyramidal neurons in layers II, III, and V of the frontal cortex Figure 1 and large and medium-sized neurons of the insula. Rim staining was not associated with tangle-bearing neurons. Large and medium-sized neurons in other layers in the frontal cortex were involved to a lesser degree. In most AD cases there was no staining of remaining perforant pathway afferents in layer II of the entorhinal cortex, amygdaloid neurons, and hippocampal CA1 neurons, all of which are selectively vulnerable to the AD process. [35-39]

Figure

Figure 1. Parietal cortex of case 7 after immunohistochemical staining for TGF-beta 1CC. (A) Staining is present in large pyramidal neurons in cortical layers III and V (scale bar equals 100 micro meter). (B) High-power magnification of the rim staining in pyramidal neurons in layer V (scale bar equals 10 micro meter).

In the NDAD case with 16 brain regions sampled (case 7), widespread ``rim'' staining of neurons was seen most prominently in large and mediumsized neurons of the neocortex, cingulate gyrus, and some subregions of the hypothalamus. Rim staining also involved some regions that are relatively spared in AD, including cranial nerve nuclei and cerebellar dentate neurons. There was little staining of the large neurons in the nucleus basalis of Meynert and median raphe and no staining of neurons in the periaqueductal gray region, locus ceruleus, substantia nigra, or inferior olive.

The overall density of neurons with perineuronal staining was greater in controls than in the averaged NDAD and FAD-14 patients (mean equals 1.11 plus minus 1.23 neurons with rim staining/mm2 in controls and 0.31 plus minus 0.39 in AD; t equals minus 2.210, p equals 0.05). There were no differences in perineuronal staining in the specific brain regions between AD (averaged NDAD and FAD-14) and control groups. To compare the effect of APOE genotype on TGF-beta 1 expression, we used nine cases (cases 3 to 5 and 8 to 13). We also examined the brains from three additional AD cases with the APOE-element 4 allele using the same clinical and pathologic selection criteria described in the Methods section above. TGF-beta was compared in a total of four patients with the APOE-element 4 allele (one homozygous, three heterozygous) and seven patients with other genotypes (six of seven with the APOE-element 3,3 genotype, one with the APOE-element 2,3 genotype). Overall, there was no significant difference in staining in cases with the APOE-element 4 allele compared with those cases with other genotypes (mean equals 1.1 neurons with rim staining/mm2 for the APOE-element 4 allele and 1.68 for other genotypes; t equals 0.601, p equals 0.567).

There was no intracytoplasmic staining within neurons or glia with anti-TGF-beta 1CC in NDAD or FAD-14 cases. The granular staining in the neuropil and white matter present in control cases was also seen in dementia cases. Vascular staining and subpial staining were rare in dementia cases. There was no staining of A beta plaques, NFT, or neuropil threads or glia in any AD cases.

TGF-beta 1LC. Control subjects.

There was no TGF-beta 1LC neuronal or glial staining, no background staining, and only occasional mild subpial staining or staining in the media of medium and large arteries. There was intense staining of corpora amylacea in all cases where these structures were present.

Dementia subjects.

TGF-beta 1LC showed staining identical to control cases. It did not stain NFT or neuropil threads. Most cases had no staining of A beta plaques; however, fine punctate staining of occasional A beta plaques in superficial layers of the frontal and entorhinal cortex was present in two cases (cases 6 and 10). There was occasionally mild staining of the subpial region and of arterial walls. In case 7, there were rare pyramidal neurons with punctate staining in the cingulate gyrus, but none of the other 15 brain regions sampled showed neuronal staining. No other neuronal or glial staining was seen. Staining of 6- to 7-micro meter cryostat sections with this antibody gave similar results; there was no immunostaining of neurons. In one case we observed astrocytic immunostaining lightly silhouetting occasional plaques, which also was present diffusely in subpial regions of that case.

TGF-beta 2. Control subjects.

TGF-beta 2 staining was present in the cytoplasm of rare neurons in the neocortex, entorhinal cortex, and hippocampus in four of 10 control cases. There was rare neuronal cytoplasmic staining for TGF-beta 2. This was enhanced using the antigen retrieval technique. There was no staining of A beta plaques and only rare staining of NFT in the medial temporal lobe in elderly control cases.

We observed mild staining of the glia in subpial regions, which did not extend through the cortex into the underlying white matter Figure 2. Blood vessels and corpora amylacea did not stain for TGF-beta 2. Purkinje cells of the cerebellum showed variable staining for TGF-beta 2 in all groups.

Figure

Figure 2. Immunohistochemical staining for TGF-beta 2 of: (A) Frontal cortex of 47-year-old AD patient (case 11) showing marked glial expression of TGF-beta 2 in association with occasional staining of cellular processes in the neuropil (scale bar equals 10 micro meter). (B) A 62-year-old control (case 18) showing no staining for TGF-beta 2 in the frontal lobe (scale bar equals 10 micro meter). (C) A 79-year-old AD patient (case 6) showing a neurofibrillary tangle staining for TGF-beta 2 (scale bar equals 100 micro meter). (D) Hippocampal CA4 region of a 65-year-old AD patient (case 5) showing a TGF-beta 2-negative amyloid core within a senile plaque (arrow). Note the intense staining of glial cells that surround the plaque (scale bar equals 100 micro meter).

Dementia subjects.

In NDAD and FAD-14 subjects, there was widespread staining of astrocytes for TGF-beta 2 in the gray and white matter, which was not seen in control cases. Gliosis was also seen in the frontal lobe and in the cores of the cerebellar folia. These glial cells often had the morphologic appearance of reactive astrocytes. The observation of gliosis was supported by pronounced immunoreactivity of GFAP in all our AD cases. Additionally there was an overall higher level of immunostaining in AD cases than there was in controls. Cases with paraformaldehyde fixation showed slightly less intense staining of glia. In case 7, all brain regions where astrocytes showed fibrillar changes demonstrated TGF-beta 2 staining including all neocortical regions, cingulate, nucleus basalis of Meynert, substantia nigra, and locus ceruleus.

The average (median grade) of glial gray matter staining was 0.3 in controls and 1.8 in the combined AD groups (KWTS equals 22.50, p equals 0.005). The median grade of glial white matter staining was 0.0 in controls and 1.8 in AD (KWTS equals 21.70, p equals 0.006). Subpial staining was also different between controls (median 0.8) and AD (median 2.8; KWTS equals 19.86, p equals 0.047). Vascular staining for TGF-beta 2 was not observed.

Intraneuronal staining was more conspicuous in AD brains than in control subjects. The TGF-beta 2 staining pattern did not differ between the NFT in FAD-14 and the NFT in NDAD cases. Using the antigen retrieval technique, immunostaining was enhanced and appeared to label intracellular NFT Figure 2 C. TGF-beta 2 staining was most prominent in the frontal cortex but was also seen in other regions where NFTs occur. In case 7, there was staining of NFTs in neurons of the neocortex in all lobes, nucleus basalis of Meynert, dorsomedial nucleus of the thalamus, and median raphe nuclei. The non-NFT-bearing cerebellar Purkinje cells also showed variable cytoplasmic staining for TGF-beta 2 in AD cases and controls.

The overall (mean) density of TGF-beta 2-stained neurons averaged from the frontal cortex, entorhinal cortex, amygdala, and CA1 was 1.26 plus minus 1.86 neurons/mm2 in AD and 0.05 plus minus 0.07 neurons/mm2 in controls (t equals minus 2.048, p equals 0.05). When specific brain regions were examined, differences in the frontal lobe between AD cases 1 to 13 and controls were significant (t equals minus 2.768; 1.65 plus minus 1.79 for AD and 0.07 plus minus 0.14 for controls; p equals 0.01). The early AD case showed an average of 0.2 TGF-beta 2-positive neurons/mm (2). The number of TGF-beta 2-positive neurons with NFTs in frontal lobe cortex was approximately 14% of all tau-positive frontal cortex neurons. There were no tau-positive NFT-bearing neurons in frontal cortex of any control cases. TGF-beta 2 staining was present in cerebellar dentate neurons only in case 9.

Using anti-TGF-beta 2 antibodies we observed rare staining of the periphery of plaques from ingrowth of immunoreactive astrocytic processes around A beta cores Figure 2 D. We confirmed the astrocytic derivation of these nonneuronal cells by demonstrating that they were strongly GFAP-positive in adjacent sections. We observed little microglial staining for TGF-beta 2 around plaques. There was no staining of the A beta cores of plaques, and in cases with amyloid angiopathy there was no increased staining in the perivascular spaces. Vascular staining was not present. The increased immunostaining for TGF-beta 2 in astrocytes and neurons in AD cases as compared with controls was supported by SELISA results. Extracts from control brains contained 0.55 plus minus 0.44 ng TGF-beta 2/g tissue compared with 1.76 plus minus 0.68 ng/g tissue in extracts from AD brains (t equals minus 4.055, p less than 0.001), a 3.2-fold increase in TGF-beta 2 levels in AD.

TGF-beta 3. Control subjects.

There was cytoplasmic staining of some neurons in the frontal lobe, medial temporal lobe, and cerebellum for TGF-beta 3 in young and elderly control cases Figure 3 A. Although there was a predominance of large-cell staining, occasionally medium and small neurons were involved.

Figure

Figure 3. Immunohistochemical staining for TGF-beta 3 of large neurons was present in familial Alzheimer's disease (FAD), nondominantly inherited Alzheimer's disease (NDAD), and controls. (A) A 35-year-old man (control, case 20) showing staining of pyramidal neurons in CA1 (scale bar equals 10 micro meter). (B) A 72-year-old man with NDAD (case 1) showing cortical pyramidal neurons staining for TGF-beta 3 (scale bar equals 10 micro meter). (C) A 47-year-old patient with FAD (case 11) with staining of cerebellar Purkinje cells (scale bar equals 100 micro meter).

Dementia subjects.

There was staining in the same neuronal populations in NDAD and FAD-14 cases as in controls. This included staining of some neuronal populations susceptible in AD (pyramidal neurons in the frontal cortex Figure 3 B and hippocampus) as well as neuronal populations not vulnerable in AD (Purkinje cells of the cerebellum Figure 3 C). In case 7, staining was seen in neurons in the insula, claustrum, thalamus, hypothalamus, nucleus basalis of Meynert and locus ceruleus, Betz cells, cranial nerve nuclei (oculomotor and facial), inferior olive, basis pontis, and cerebellar dentate cells. In case 7, there was also staining of oligodendroglia in white matter regions.

Staining of neurons for TGF-beta 3 was similar for NDAD, FAD-14, and control cases (overall means equals 0.58 plus minus 0.65 per mm2 for NDAD, 0.30 plus minus 0.36 for FAD-14, and 0.30 plus minus 0.29 for control cases; F equals 0.913, p equals 0.418). There were no significant differences in any subregions. Although there was staining for TGF-beta 3 in rare NFT-bearing neurons, this staining was limited to the cytoplasm and did not stain NFT.

Vascular staining of TGF-beta 3 did not differ between groups (average grade equals 1.5 in controls, 0.9 in dementia cases; KWTS equals 9.350, p less than 0.499). Similarly, subpial staining for TGF-beta 3 did not differ between groups (average grade equals 0.8 in controls and 1.35 in AD; KWTS equals 10.323, p less than 0.502).

Discussion.

Our study indicates that there is differential TGF-beta expression in AD. TGF-beta 2 is quantitatively increased in NDAD and FAD-14 brains and is localized to reactive glia and to abnormal neurons containing NFT. TGF-beta 3 expression was similar in AD and control brains, and differences between TGF-beta 1 perineuronal staining were marginal between control and AD patients. The pattern of TGF-beta 1 staining does not differ between APOE-element 4 cases and non-APOE-element 4 cases. While a variety of systemic diseases could alter growth factor levels, this is unlikely to account for such a significant increase in TGF-beta 2 expression in the brains of all AD patients as compared with similarage controls who would be likely to suffer from similar chronic conditions. Acute preterminal systemic disease such as hypoxia or pneumonia with sepsis (endotoxemia) could also theoretically have an effect on the expression of growth factors in either our AD group or our control group; however, by chart review there were no large, consistent differences in the occurrence of terminal disease states in our AD or aged normal populations.

NFTs are the principal marker of affected neurons in AD. NFTs are also important in AD because the severity of dementia correlates closely with the degree of neuritic degeneration (NFT and neuropil thread formation). [40,41] Our study suggests that TGF-beta 2 immunostaining is selective for NFT-bearing neurons in AD; however, the nature of this association is unknown. One intraneuronal event that precedes the formation of NFT is hyperphosphorylation of the microtubule-associated protein tau. [42-44] It is not known if TGF-beta 2 affects the regulation of the kinases and/or phosphorylases that influence the phosphorylation of tau. Elucidation of the temporal expression of TGF-beta 2 and NFTs in these regions might suggest a function for TGF-beta 2 in tangle formation.

The underlying cause of TGF-beta 2 induction in NFT in AD is unknown. It may be a homeostatic response to normalize the intraneuronal environment either by repairing the cytoskeleton, normalizing the factors that initiated the process of NFT formation, or helping to develop compensatory mechanisms to maintain cellular integrity. It could also be a nonspecific response to the presence of NFT or to some other factor in the process of neuronal degeneration. The increased expression of TGF-beta 2 by reactive astrocytes in our AD cases could be a response to normalize the extracellular environment. Studies of TGF-beta expression in other tangle-bearing and nontangle-bearing degenerative diseases as well as in nondegenerative diseases associated with NFT formation (like gangliogliomas) might help determine whether glial TGF-beta 2 production is a response occurring in all degenerative diseases, conditions associated only with NFT formation, or a response specific to AD.

The relationship between TGF-beta expression and A beta deposition is unknown. In AD cases with severe amyloid angiopathy we observed no TGF-beta staining of blood vessels, not even of vessels with ``double-barrelled'' lumina. Similarly, perivascular spaces and plaques did not react with TGF-beta 2. It is possible that the co-localization of TGF-beta 2 with plaques was inapparent because the A beta deposits in AD are largely insoluble [45] whereas the TGF-beta 2 binding to A beta that Bodmer et al [20] reported involved soluble A beta deposits. The only plaque staining by anti-TGF-beta 2 in our AD cases was from secondary changes in the composition of the neuropil underlying plaques or immunostaining of enlarged (reactive) glial processes within the plaque. Similarly, cases with intense TGF-beta 2 expression were not the cases with the most plaques. Previous immunohistochemical studies reported localization of TGF-beta 1 in plaque-like structures in paraformaldehyde-fixed, nonembedded 50-micro meter sections. [21] Our attempts to repeat this observation with our TGF-beta 1 antibodies in nonembedded sections were largely unsuccessful. This discrepancy may result from the use of different TGF-beta 1 antibodies; the antibodies used by van der Wal et al [21] were total IgG fractions raised to the entire TGF-beta 1 molecule, while antibodies in our study were raised to short peptide sequences in TGF-beta s to generate isoform-specific reagents. Our antibodies also were affinity-purified against the immunizing peptide to decrease the chance of nonspecific binding. The reported binding of TGF-beta 1 to the A beta plaque [21] may expose an epitope that consists of nonconsecutive amino acids that are only found in the entire TGF-beta molecule and not in any short peptide sequence and so would not be recognized by our peptide antibodies. Additionally, our immunohistochemical protocol involves pretreatment of sections with hyaluronidase, which we have found to maximize the sensitivity of our antibodies. This pretreatment could disrupt the interaction between TGF-beta and the A beta plaque.

The gene for TGF-beta 1 and the gene for the susceptibility locus associated with increased risk for late-onset FAD both reside [46] on chromosome 19q13 while the genes for TGF-beta 3 and the mutation present in a majority of FAD cases are on chromosome 14q24.3. Genetic linkage studies have not yet pointed to a direct role for TGF-beta 1 and TGF-beta 3 in AD. Similarly, we observed no evidence of a role for these TGF-beta s in AD in our immunohistochemical studies. For example, TGF-beta 3 neuronal staining occurred even in young controls, suggesting it is probably not specifically related to AD or to the aging process. However, a structural abnormality of the TGF-beta 3 gene could be present in FAD-14 without affecting the overall level of expression of TGF-beta 3 in the brain. Alternatively, the function of TGF-beta 3 might be reduced in patients with a genetic abnormality on chromosome 14.

There are no known families yet with linkage to chromosome 1, the genetic location of TGF-beta 2. There is at least one large Volga German FAD lineage and several Swedish families with early-onset FAD that do not show linkage to chromosomes 14, 19, or 21. [47,48] We had no cases from these lineages in our study. The possibility that TGF-beta 2 plays a primary role in the pathogenesis of dementia in these lineages or other unlinked lineages remains to be addressed.

We determined that there was no significant effect of APOE genotype on TGF-beta 1CC immunoreactivity in AD cases with the APOE-element 4 genotype compared with those lacking the APOE-element 4 allele. This is important because the APOE-element 4 allele is the only known major genetic risk factor for NDAD and late-onset FAD. [11,24,49-51] Differential rim staining for TGF-beta 1 in cases with different APOE genotypes would have associated TGF-beta 1 in the pathogenesis of late-onset AD. However, even our AD case homozygous for APOE-element 4,4 showed similar staining for TGF-beta 1 compared with individuals lacking the APOE-element 4 allele.

In summary, it appears that TGF-beta 2 has increased expression in AD. Neurotrophic factors including TGF-beta 2 may prove to be useful agents to modify the course of AD. Some neurotrophic factors are currently being investigated as possible therapy for neurodegenerative diseases because they can influence degeneration of injured CNS neurons. [52,53] Nerve growth factor is under particularly intensive investigation as a potential treatment for AD because cholinergic neurons are susceptible in AD. [54,55] Since neuronal loss in AD is not restricted to the cholinergic system, [56-58] neurotrophic agents that act on other neurotransmitter systems or agents with a broader site of action (such as TGF-beta) may also have potential to influence the course of neurodegeneration that underlies the dementia of AD patients.

Acknowledgments

We wish to thank David A. Drachman, MD, for his support. We also thank D. Danielpour, PhD, for providing antibodies for the SELISA, A. Bhandiwad for performing the SELISA, D. Pulaski-Salo for technical help, and M.S. Lucia, MD, for expert advice.

REFERENCES

  1. 1.↵
    Roberts AB, Sporn MB. Physiological actions and clinical applications of transforming growth factor- beta. Growth Factors 1993;8:1-9.
    OpenUrlPubMed
  2. 2.↵
    Chalazonitis A, Kalberg AJ, Twardzik DR, Morrison RS, Kessler JA. Transforming growth factor- beta has neurotrophic actions on sensory neurons in vitro and is synergistic with nerve growth factor. Dev Biol 1992;152:121-132.
    OpenUrlCrossRefPubMed
  3. 3.↵
    Logan A, Frautschy SA, Gonzalez A, Sporn MB, Baird A. Enhanced expression of transforming growth factor beta 1 in the rat brain after a localized cerebral injury. Brain Res 1992;587:216-225.
    OpenUrl
  4. 4.
    Klempt ND, Sirimanne E, Gunn AJ, et al. Hypoxia-ischemia induces transforming growth factor beta 1 mRNA in the infant rat brain. Brain Res Mol Brain Res 1992;13:93-101.
    OpenUrl
  5. 5.
    Lindholm D, Castren E, Kiefer R, Zafra F, Thoenen H. Transforming growth factor- beta 1 in the rat brain: increase after injury and inhibition of astrocyte proliferation. J Cell Biol 1992;117:395-400.
    OpenUrl
  6. 6.↵
    Prehn JHM, Peruche B, Unsicker K, Krieglstein J. Transforming growth factor- beta 1 prevents degeneration of primary neuronal cultures induced by cytotoxic hypoxia or glutamate. J Neurochem 1993;60:1665-1672.
    OpenUrlCrossRefPubMed
  7. 7.↵
    ten Dijke P, Geurts van Kessel AHM, Foulkes JG, Le Beau MM. Transforming growth factor beta type 3 maps to human chromosome 14, region q23-q24. Oncogene 1988;3:721-724.
    OpenUrl
  8. 8.↵
    Schellenberg GD, Bird TD, Wijsman EM, et al. Genetic linkage evidence for a familial Alzheimer's disease locus on chromosome 14. Science 1992;258:668-671.
    OpenUrl
  9. 9.
    St. George-Hyslop P, Haines J, Rogaev E, et al. Genetic evidence for a novel familial Alzheimer's disease locus on chromosome 14. Nat Genet 1992;2:330-334.
    OpenUrlPubMed
  10. 10.↵
    Fugii D, Brissenden JE, Derynck R, Francke U. Transforming growth factor beta gene maps to human chromosome 19 long arm and to mouse chromosome 7. Somat Cell Genet 1986;12:281-288.
    OpenUrl
  11. 11.↵
    Strittmatter WJ, Saunders AM, Schmechel D, et al. Apolipoprotein E: high affinity binding to beta -amyloid increased frequency of type 4 allele in familial Alzheimer's disease. Proc Natl Acad Sci USA 1993;90:1977-1981.
    OpenUrlAbstract/FREE Full Text
  12. 12.↵
    Goate A, Chartier-Harlin M, Mullan M, et al. Segregation of a missense mutation of the amyloid precursor protein gene with familial Alzheimer's disease. Nature 1991;349:704-706.
    OpenUrl
  13. 13.
    Naruse S, Igarashi S, Kobayashi H, et al. Mis-sense mutation Val-Ile in exon 17 of amyloid precursor gene in Japanese familial Alzheimer's disease [letter]. Lancet 1991;337:978-979.
    OpenUrl
  14. 14.
    Murrell J, Farlow M, Ghetti B, Benson MD. A mutation in the amyloid precursor protein associated with hereditary Alzheimer's disease. Science 1991;254:97-99.
    OpenUrlFREE Full Text
  15. 15.
    Chartier-Harlin M, Crawford F, Houlden H, et al. Earlyonset Alzheimer's disease caused by mutations at codon 717 of the beta -amyloid precursor protein gene. Nature 1991;353:844-846.
    OpenUrlPubMed
  16. 16.
    Chartier-Harlin MC, Crawford F, Hamandi K, et al. Screening for the beta-amyloid precursor protein mutation (APP717: Val-Ile) in extended pedigrees with early onset Alzheimer's disease. Neurosci Lett 1991;129:134-135.
    OpenUrlCrossRefPubMed
  17. 17.
    Kennedy AM, Newman S, McCaddon A, et al. Familial Alzheimer's disease: a pedigree with a mis-sense mutation in the amyloid precursor protein gene (amyloid precursor protein 717 valine-glycine). Brain 1993;116:309-324.
    OpenUrl
  18. 18.
    Mullan M, Crawford F, Axelman K, et al. A new mutation in APP demonstrates that pathogenic mutations for probable Alzheimer's disease frame the beta -amyloid sequence. Nat Genet 1992;1:345-347.
    OpenUrlCrossRefPubMed
  19. 19.↵
    Gray CW, Patel AJ. Regulation of beta -amyloid precursor protein isoform mRNAs by transforming growth factor- beta 1 and interleukin-1 beta in astrocytes. Brain Res Mol Brain Res 1993;19:251-256.
    OpenUrl
  20. 20.↵
    Bodmer S, Podlinsky MB, Selkoe DJ, Heid I, Fontana A. Transforming growth factor- beta bound to soluble derivatives of the beta -amyloid precursor protein of Alzheimer's disease. Biochem Biophys Res Commun 1990;171:890-897.
    OpenUrl
  21. 21.↵
    van der Wal EA, Gomez-Pinilla F, Cotman CW. Transforming growth factor- beta 1 is in plaques in Alzheimer's and Down's pathologies. Neuroreport 1993;4:69-72.
    OpenUrl
  22. 22.↵
    McKhann G, Drachman D, Folstein M, Katzman R, Price D, Stadlan EM. Clinical diagnosis of Alzheimer's disease: report of the NINCDS-ADRDA Work Group under the auspices of Department of Health and Human Services Task Force on Alzheimer's Disease. Neurology 1984;34:939-944.
    OpenUrl
  23. 23.↵
    Mirra SS, Heyman A, McKeel D, et al. The Consortium to Establish a Registry for Alzheimer's Disease (CERAD). Part II. Standardization of the neuropathologic assessment of Alzheimer's disease. Neurology 1991;41:479-486.
    OpenUrl
  24. 24.↵
    Schmechel DE, Saunders AM, Strittmatter WJ, et al. Increased amyloid beta -peptide deposition as a consequence of apolipoprotein E genotype in late-onset Alzheimer's disease. Proc Natl Acad Sci USA 1993;90:9649-9653.
    OpenUrl
  25. 25.↵
    McLean IW, Nakane PK. Periodate-lysine-paraformaldehyde fixative. A new fixative for immunoelectron microscopy. J Histochem Cytochem 1974;22:1077-1083.
    OpenUrlPubMed
  26. 26.↵
    Campbell SK, Switzer RC, Martin TL. Alzheimer's plaques and tangles: a controlled and enhanced silver staining method [abstract]. Soc Neurosci Abstr 1987;3:678.
    OpenUrl
  27. 27.↵
    Flanders KC, Thompson NL, Cissel DS, et al. Transforming growth factor- beta 1: histochemical localization with antibodies to different epitopes. J Cell Biol 1989;108:653-660.
    OpenUrl
  28. 28.↵
    Flanders KC, Cissel DS, Mullen LT, Danielpour D, Sporn MB, Roberts AB. Antibodies to transforming growth factor- beta 2 peptides: specific detection of TGF- beta 2 in immunoassays. Growth Factors 1990;3:45-52.
    OpenUrl
  29. 29.↵
    Flanders KC, Ludecke G, Engels S, et al. Localization and actions of transforming growth factor- beta s in the embryonic nervous system. Development 1991;13:183-191.
    OpenUrlPubMed
  30. 30.
    Flanders KC, Roberts AB, Ling N, Fleurdelys BE, Sporn MB. Antibodies to peptide determinants of transforming growth factor- beta and their applications. Biochemistry 1988;27:739-746.
    OpenUrl
  31. 31.
    Unsicker K, Flanders KC, Cissel DS, Lafyatis R, Sporn MB. Transforming growth factor- beta isoforms in the rat central and peripheral nervous system. Neuroscience 1991;44:613-625.
    OpenUrl
  32. 32.↵
    Shi S, Key ME, Kalra KL. Antigen retrieval in formalinfixed, paraffin embedded tissues: an enhancement method for immunohistochemical staining based on microwave oven heating of tissue sections. J Histochem Cytochem 1991;39:741-748.
    OpenUrlAbstract/FREE Full Text
  33. 33.↵
    Danielpour D, Kim K-Y, Dart LL, Watanabe S, Roberts AB, Sporn MB. Sandwich enzyme-linked immunosorbent assays (SELISAs) quantitate and distinguish two forms of transforming growth factor- beta (TGF- beta 1 and TGF- beta 2) in complex biological fluids. Growth Factors 1989;2:61-71.
    OpenUrlPubMed
  34. 34.↵
    Miller DC, Koslow M, Budzilovich GN, Burstein DE. Synaptophysin: a sensitive and specific marker for ganglion cells in central nervous system neoplasms. Hum Pathol 1990;21:271-276.
    OpenUrl
  35. 35.↵
    Ball MJ. Neuronal loss, neurofibrillary tangles and granulovacuolar degeneration in the hippocampus with ageing and dementia--a quantitative study. Acta Neuropathol (Berl) 1977;37:111-118.
    OpenUrlCrossRefPubMed
  36. 36.
    Herzog AG, Kemper TL. Amygdaloid changes in aging and dementia. Arch Neurol 1980;37:625-629.
    OpenUrlCrossRefPubMed
  37. 37.
    Hyman BT, Van Hoesen GW, Kromer BA, Damasio AR. Perforant pathway changes and the memory impairment of Alzheimer's disease. Ann Neurol 1986;20:472-481.
    OpenUrl
  38. 38.
    Van Hoesen GW, Hyman BT, Damasio AR. Entorhinal cortex pathology in Alzheimer's disease. Hippocampus 1991;1:1-8.
    OpenUrlPubMed
  39. 39.
    Lippa CF, Hamos JE, Pulaski-Salo D, DeGennaro LJ, Drachman DA. Alzheimer's disease and aging: effects on perforant pathway perikarya and synapses. Neurobiol Aging 1992;13:405-411.
    OpenUrl
  40. 40.↵
    McKee AC, Kosik KS, Kowall NW. Neuritic pathology and dementia in Alzheimer's disease. Ann Neurol 1991;30:156-165.
    OpenUrl
  41. 41.
    Arriagada PV, Growdon JH, Hedley-Whyte ET, Hyman BT. Neurofibrillary tangles but not senile plaques parallel duration and severity of Alzheimer's disease. Neurology 1992;42:631-639.
    OpenUrlPubMed
  42. 42.↵
    Kosik KS, Joachim CL, Selkoe DJ. Microtubule-associated protein tau is a major antigenic component of paired helical filaments in Alzheimer's disease. Proc Natl Acad Sci USA 1986;83:4044-4048.
    OpenUrlFREE Full Text
  43. 43.
    Kosik KS, Orecchio LD, Binder L, Trojanowski JQ, Lee VM-Y, Lee G. Epitopes that span the tau molecule are shared with paired helical filaments. Neuron 1988;1:817-825.
    OpenUrl
  44. 44.
    Lee VM-Y, Balin BJ, Otvos L, Trojanowski JQ. A68: a major subunit of paired helical filaments and derivatized forms of normal tau. Science 1991;251:675-678.
    OpenUrl
  45. 45.↵
    Selkoe DJ, Podlinsky MB, Joachim CL, et al. beta -Amyloid precursor protein of Alzheimer's disease occurs as 110- to 135-kilodalton membrane-associated proteins in neural and nonneural tissues. Proc Natl Acad Sci USA 1988;85:7341-7345.
    OpenUrl
  46. 46.↵
    Roberts AB, Sporn MB. The transforming growth factors- beta. In: Sporn MB, Roberts AB, eds. Peptide growth factors and their receptors I. Springer-Verlag, 1990:419-472.
  47. 47.↵
    Bird TD, Lampe TH, Nemens EJ, et al. Familial Alzheimer's disease in American descendents of the Volga Germans: probable founder effect. Ann Neurol 1988;23:25-31.
    OpenUrl
  48. 48.
    Lannfelt L, Lilius L, Appelgren H, et al. No linkage to chromosome 14 in Swedish Alzheimer's disease families. Nat Genet 1993;4:218-219.
    OpenUrlCrossRef
  49. 49.
    Saunders AM, Strittmatter WJ, Schmechel D, et al. Association of apolipoprotein E allele element 4 with late-onset familial and sporadic Alzheimer's disease. Neurology 1993;43:1467-1472.
    OpenUrlAbstract/FREE Full Text
  50. 50.
    Strittmatter WJ, Weisgraber KH, Huang D, et al. Binding of human apolipoprotein E to beta/A4 peptide: isoform specific effects and implications for late-onset Alzheimer's disease. Proc Natl Acad Sci USA 1993;90:8096-8102.
    OpenUrl
  51. 51.
    Corder EH, Saunders AM, Strittmatter WJ, et al. Gene dose of apolipoprotein E type 4 allele and the risk of Alzheimer's disease in late onset families. Science 1993;261:921-923.
    OpenUrlFREE Full Text
  52. 52.↵
    Hefti F. Nerve growth factor (NGF) promotes survival of septal cholinergic neurons after fimbrial transection. J Neurosci 1986;6:2155-2162.
    OpenUrl
  53. 53.
    Williams LR, Varon S, Peterson GM, et al. Continuous infusion of nerve growth factor prevents basal forebrain neuronal death after fimbria-fornix transection. Proc Natl Acad Sci USA 1986;83:9231-9235.
    OpenUrlAbstract/FREE Full Text
  54. 54.↵
    Coyle JT, Price PH, Delong MR. Alzheimer's disease: a disorder of cortical cholinergic innervation. Science 1983;219:1184-1189.
    OpenUrl
  55. 55.
    Bartus R, Dean RL, Beer C, Lippa AS. The cholinergic hypothesis of cholinergic memory dysfunction. Science 1982;217:408-417.
    OpenUrl
  56. 56.↵
    D'Amato RJ, Zweig RM, Whitehouse PJ, et al. Aminergic systems in Alzheimer's disease and Parkinson's disease. Ann Neurol 1987;22:229-236.
    OpenUrlCrossRefPubMed
  57. 57.
    Ellison DW, Beal MF, Mazurek MF, Bird ED, Martin JD. A post mortem study of neurotransmitters in Alzheimer's disease. Ann Neurol 1986;20:616-621.
    OpenUrl
  58. 58.
    Palmer AM, Francis PT, Bowen DM, et al. Catecholinergic neurons assessed ante-mortem in Alzheimer's disease. Brain Res 1987;414:365-375.
    OpenUrl

Letters: Rapid online correspondence

No comments have been published for this article.
Comment

REQUIREMENTS

If you are uploading a letter concerning an article:
You must have updated your disclosures within six months: http://submit.neurology.org

Your co-authors must send a completed Publishing Agreement Form to Neurology Staff (not necessary for the lead/corresponding author as the form below will suffice) before you upload your comment.

If you are responding to a comment that was written about an article you originally authored:
You (and co-authors) do not need to fill out forms or check disclosures as author forms are still valid
and apply to letter.

Submission specifications:

  • Submissions must be < 200 words with < 5 references. Reference 1 must be the article on which you are commenting.
  • Submissions should not have more than 5 authors. (Exception: original author replies can include all original authors of the article)
  • Submit only on articles published within 6 months of issue date.
  • Do not be redundant. Read any comments already posted on the article prior to submission.
  • Submitted comments are subject to editing and editor review prior to posting.

More guidelines and information on Disputes & Debates

Compose Comment

More information about text formats

Plain text

  • No HTML tags allowed.
  • Web page addresses and e-mail addresses turn into links automatically.
  • Lines and paragraphs break automatically.
Author Information
NOTE: The first author must also be the corresponding author of the comment.
First or given name, e.g. 'Peter'.
Your last, or family, name, e.g. 'MacMoody'.
Your email address, e.g. higgs-boson@gmail.com
Your role and/or occupation, e.g. 'Orthopedic Surgeon'.
Your organization or institution (if applicable), e.g. 'Royal Free Hospital'.
Publishing Agreement
NOTE: All authors, besides the first/corresponding author, must complete a separate Publishing Agreement Form and provide via email to the editorial office before comments can be posted.
CAPTCHA
This question is for testing whether or not you are a human visitor and to prevent automated spam submissions.

Vertical Tabs

You May Also be Interested in

Back to top
Advertisement

Related Articles

  • No related articles found.

Alert Me