Agrin is a major heparan sulfate proteoglycan accumulating in Alzheimer's disease brain

Abstract

Heparan sulfate proteoglycans (HSPGs) have been suggested to play an important role in the formation and persistence of senile plaques and neurofibrillary tangles in dementia of the Alzheimer's type (DAT). We performed a comparative immunohistochemical analysis of the expression of the HSPGs agrin, perlecan, glypican-1, and syndecans 1-3 in the lesions of DAT brain neocortex and hippocampus. Using a panel of specific antibodies directed against the protein backbone of the various HSPG species and against the glycosaminoglycan (GAG) side-chains, we demonstrated the following. The basement membrane-associated HSPG, agrin, is widely expressed in senile plaques, neurofibrillary tangles and cerebral blood vessels, whereas the expression of the other basement membrane-associated HSPG, perlecan, is lacking in senile plaques and neurofibrillary tangles and is restricted to the cerebral vasculature. Glypican and three different syndecans, all cell membrane-associated HSPG species, are also expressed in senile plaques and neurofibrillary tangles, albeit at a lower frequency than agrin. Heparan sulfate GAG side chains are also associated with both senile plaques and neurofibrillary tangles. Our results suggest that glycosaminoglycan side chains of the HSPGs agrin, syndecan, and glypican, but not perlecan, may play an important role in the formation of both senile plaques and neurofibrillary tangles. In addition, we speculate that agrin, because it contains nine protease-inhibiting domains, may protect the protein aggregates in senile plaques and neurofibrillary tangles against extracellular proteolytic degradation, leading to the persistence of these deposits.

Figure 1.

Immunohistochemical staining of senile plaques in serial sections (a-f) from DAT hippocampus for Aβ (mAb 6C6) and agrin (mAb JM72). Overview of senile plaques that are diffusely stained for both Aβ (a) and agrin (b). Occasionally, an additional granular staining of senile plaques (c, mAb 6C6) for agrin (d) is observed. Also, small Aβ deposits (e) are immunopositive for agrin (f). Higher magnification of a senile plaque stained by JM72 (g). Original magnifications: a-f: ×125; g: ×250.

Figure 2.

Immunohistochemical staining of neurofibrillary tangles in DAT hippocampus for agrin (mAb JM72). Intracellular tangles are stained by JM72 (a). Ghost tangles are also stained for agrin (b). Cerebral blood vessels are also strongly stained by JM72 (a and b; examples indicated by arrowheads). In c, staining of ApoE in ghost tangles (examples indicated by arrows) and senile plaques (arrowheads) is shown. Original magnification, ×125.

Figure 3.

Immunohistochemical staining of DAT hippocampus with anti-perlecan antibodies. Cerebral blood vessels are stained by mAb 95J10 (a). The anti-perlecan pAb EY-90 stains coarse granules in both senile plaques (b) and neurofibrillary tangles (arrows) in DAT hippocampus (c). Blood vessels are also strongly stained by EY-90. These coarse granules in senile plaques remain distinct from dystrophic neurites that are stained for tau (d) and ubiquitin (e). Many scattered neuropil threads are detected with anti-tau and only a few by anti-ubiquitin. Original magnification, ×100.

Figure 4.

Immunohistochemical expression of glypican-1 with mAb S1 in DAT hippocampus (a) and cortex (b). S1 stains senile plaques and tangles (a). In the cortex the cores of amyloid plaques are accentuated by S1 (b). Original magnification, ×100.

Figure 5.

Immunohistochemical staining of syndecans in DAT brain tissue. Shown are staining of senile plaques (a), tangles (b), and ghost tangles (c) by the anti-syndecan-2 mAb 10H4; staining of senile plaques (d), tangles (e), and ghost tangles (f) by the anti-syndecan-3 mAb 1C7; and staining of senile plaques (g) and tangles (h) by the anti-syndecan-1 and -3 mAb 2E9. mAb 1C7 also stains reactive astrocytes associated with senile plaques (i). In b and g, tissue sections from DAT neocortex are shown; all others are hippocampal sections. Original magnification, ×100.

Figure 6.

Immunohistochemical staining of HS GAG side chains in DAT brain hippocampus (a-e, g, and h) and neocortex (f and i). The anti-HS stub mAb 3G10 stains senile plaques (a), tangles (a and b), and ghost tangles (c). HS GAG side chains are identified in senile plaques by the mAbs JM403 (d), 10E4 (e), and JM13 (f and i). HS GAG side-chain expression in tangles is demonstrated by staining with JM403 (g), 10E4 (h), and JM13 (i). In f, a uniform staining of white matter by JM 13 is shown in addition to staining of senile plaques. All four anti-HS GAG antibodies label cerebral blood vessels. Original magnification, ×100.

Figure 7.

Double immunofluorescence staining for HS GAG (mAb 3G10) and tau of an intracellular tangle and a ghost tangle. Both the intracellular tangle and the ghost tangle are stained by 3G10 (a), whereas only the intracellular tangle is stained by the anti-tau antibodies (b). Original magnification, ×1000.

Figure 8.

Immunohistochemical staining of senile plaques in serial sections from control hippocampus. Diffuse senile plaques in control brain tissue are stained for both Aβ (a, mAb 6C6) and agrin (b, mAb JM72). Original magnification, ×125.