he neu­ropathol­o­gy of Alzheimer’s dis­ease (AD) is char­ac­ter­ized by hy­per­phos­pho­ry­lat­ed tau neu­rofib­ril­lary tan­gles (NFTs) and amy­loid-be­ta (Aβ) plaques. Aβ plaques are hy­poth­e­sized to fol­low a de­vel­op­ment se­quence start­ing with dif­fuse plaques, which evolve in­to more com­pact plaques and fi­nal­ly ma­ture in­to the clas­sic cored plaque type. A bet­ter mol­e­c­u­lar un­der­stand­ing of Aβ pathol­o­gy is cru­cial, as the role of Aβ plaques in AD patho­gen­e­sis is un­der de­bate. Here, we stud­ied the de­po­si­tion and fib­ril­la­tion of Aβ in dif­fer­ent plaque types with la­bel-free in­frared and Ra­man imag­ing. Fouri­er-trans­form in­frared (FTIR) and Ra­man imag­ing was per­formed on na­tive snap-frozen brain tis­sue sec­tions from AD cas­es and non-de­ment­ed con­trol cas­es. Sub­se­quent­ly, the scanned tis­sue was stained against Aβ and an­no­tat­ed for the dif­fer­ent plaque types by an AD neu­ropathol­o­gy ex­pert. In to­tal, 160 plaques (68 dif­fuse, 32 com­pact, and 60 clas­sic cored plaques) were im­aged with FTIR and the re­sults of se­lect­ed plaques were ver­i­fied with Ra­man imag­ing. In dif­fuse plaques, we de­tect ev­i­dence of short an­tipar­al­lel β‑sheets, sug­gest­ing the pres­ence of Aβ oligomers. Aβ fib­ril­la­tion sig­nif­i­cant­ly in­creas­es along­side the pro­posed plaque de­vel­op­ment se­quence. In clas­sic cored plaques, we spa­tial­ly re­solve cores con­tain­ing pre­dom­i­nant­ly large par­al­lel β‑sheets, in­di­cat­ing Aβ fib­rils. Com­bin­ing la­bel-free vi­bra­tional imag­ing and im­muno­his­to­chem­istry on brain tis­sue sam­ples of AD and non-de­ment­ed cas­es pro­vides nov­el in­sight in­to the spa­tial dis­tri­b­u­tion of the Aβ con­for­ma­tions in dif­fer­ent plaque types. This way, we re­con­struct the de­vel­op­ment process of Aβ plaques in hu­man brain tis­sue, pro­vide in­sight in­to Aβ fib­ril­la­tion in the brain, and sup­port the plaque de­vel­op­ment hypothesis.

Alzheimer’s dis­ease (AD) is the most com­mon neu­rode­gen­er­a­tive dis­ease and is patho­log­i­cal­ly char­ac­ter­ized by hy­per­phos­pho­ry­lat­ed tau neu­rofib­ril­lary tan­gles (NFT) and amy­loid-be­ta (Aβ) plaques. Aβ orig­i­nates from the cleav­age of the amy­loid pre­cur­sor pro­tein (APP) and is se­cret­ed to the ex­tra­cel­lu­lar space. The most ac­cept­ed hy­poth­e­sis for AD patho­gen­e­sis is the amy­loid cas­cade hy­poth­e­sis. Ac­cord­ing to this hy­poth­e­sis, Aβ ag­gre­gates in the neu­ropil as plaques, due to an im­bal­ance of Aβ pro­duc­tion and clear­ance. The Aβ monomers mis­fold and form β‑sheet-rich oligomers, which then form protofib­rils that stack in­to high­ly or­ga­nized amy­loid fib­rils. The ag­gre­ga­tion of Aβ caus­es synap­tic stress and in­duces an in­flam­ma­to­ry re­sponse. Si­mul­ta­ne­ous­ly, synap­tic and neu­ronal in­jury leads to the hy­per­phos­pho­ry­la­tion of tau, which ag­gre­gates with­in neu­rons as NFTs that fi­nal­ly cause neu­ronal death. As the dis­ease spreads and pro­gress­es, there is ex­ten­sive neu­ronal death through­out the brain, which ul­ti­mate­ly leads to de­men­tia. The amy­loid cas­cade hy­poth­e­sis is cur­rent­ly un­der de­bate. While it is pro­posed that Aβ is the ini­tial trig­ger of patho­log­i­cal process­es, NFTs are con­sid­ered to be the pro­gres­sive force of the dis­ease. The dis­cus­sion is fu­eled by sev­er­al failed clin­i­cal stud­ies of Aβ-tar­get­ing an­ti­bod­ies, as well as en­cour­ag­ing re­sults of most re­cent an­ti-Aβ drug studies.

Aβ plaques show dif­fer­ent mor­pholo­gies. Here, we con­sid­er (i) the dif­fuse type, (ii) the com­pact (or prim­i­tive) type, and (iii) the clas­sic cored type. It is pro­posed that these dif­fer­ent mor­pholo­gies rep­re­sent the pro­gres­sive stages of Aβ fib­ril­la­tion. Plaque for­ma­tion is pro­posed to start as dif­fuse amor­phous struc­tures that main­ly con­sist of ag­gre­gat­ed Aβ oligomers and protofib­rils. Then, with the pro­gres­sion of Aβ fib­ril­la­tion, the plaque shows an in­creas­ing­ly com­pact mor­phol­o­gy with a more clear­ly de­fined out­line. An in­flam­ma­to­ry re­sponse, dri­ven main­ly by mi­croglia, is strong­ly as­so­ci­at­ed with the ear­ly stages of Aβ plaque for­ma­tion and even con­sid­ered to dri­ve the con­tin­u­ing build-up of amy­loid fib­rils and the ac­com­pa­nied neu­ro­tox­ic ef­fects. The fi­nal fib­ril­la­tion stage is reached when Aβ is con­densed to a core that con­tains most­ly Aβ fibrils.

Here, we ap­plied Fouri­er trans­form in­frared (FTIR) and Ra­man imag­ing to snap-frozen thin sec­tions of hu­man brain tis­sue. These la­bel-free meth­ods are much less in­va­sive to­wards the sam­ple than stain­ing meth­ods be­cause the tis­sue is ex­am­ined with­out chem­i­cal al­ter­ations. The vi­bra­tional mi­crospec­troscopy ap­proach pro­vides spa­tial­ly re­solved spec­tra that re­flect the bio­chem­i­cal fin­ger­print of an­a­lyzed sam­ples, in­clud­ing the pro­tein sec­ondary struc­ture. Ra­man is a com­ple­men­tary spec­tro­scop­ic tech­nique to FTIR and is used here to ver­i­fy the FTIR re­sults. The ma­jor con­stituents of brain tis­sue are pro­teins and lipids [53]. The sec­ondary struc­ture of pro­teins can be de­ter­mined by an­a­lyz­ing the Amide I ab­sorbance band (C=O stretch­ing vi­bra­tion of the pro­tein back­bone). The Amide I ab­sorbance band is in­dica­tive for the sec­ondary struc­ture. It con­sist of sev­er­al bands, each as­so­ci­at­ed with dis­tinct sec­ondary struc­tures. The po­si­tion of the main β‑sheet-band around 1630 cm⁻¹ shifts to­wards low­er wavenum­bers, when the strands be­come arranged in par­al­lel β‑sheets. Ac­cord­ing­ly, amy­loid fib­rils of­ten ab­sorb at a low­er wavenum­ber than na­tive β‑sheet pro­teins. For in­stance, oligomer­ic Aβ with typ­i­cal­ly an­tipar­al­lel β‑sheet struc­ture ab­sorbs around 1630 cm, where­as a shift to low­er wavenum­bers has been re­port­ed for Aß fib­rils. Fur­ther­more, an­tipar­al­lel β‑sheets dis­play a char­ac­ter­is­tic band around 1693 cm⁻¹. This band is n⁻¹ot ob­served in Aβ fib­rils with pre­dom­i­nant­ly par­al­lel β‑sheets. Here, the ac­cu­mu­la­tion of β‑sheet-rich Aβ oligomers and fib­rils in plaques is stud­ied, an­a­lyz­ing the Amide I band. Apart from that, lipids con­sti­tute about 40% of the gray mat­ter dry weight [56] and show char­ac­ter­is­tic ab­sorbance bands as well. The fat­ty acids in lipids con­sist most­ly of meth­yl­ene and methyl (CH₂ and CH₃ groups, which al­so oc­cur in pro­tein side chains) that gen­er­ate stretch­ing vi­bra­tion bands in the re­gion 3000–2800 cm⁻¹. The head groups of most phos­pho­lipids, which make up ~ 70% of the lipid con­tent, con­tain es­ter groups that gen­er­ate the lipid-as­so­ci­at­ed band (es­ter C=O stretch­ing vi­bra­tion) around 1738 cm⁻¹.

In this study, the pro­gres­sion of Aβ fib­ril­la­tion, along­side the pro­posed de­vel­op­ment se­quence of Aβ plaques (dif­fuse, com­pact, clas­sic cored) in AD is stud­ied with spa­tial and mol­e­c­u­lar res­o­lu­tion, us­ing la­bel-free imag­ing. Post-mortem sec­tions from fresh-frozen brain tis­sue were an­a­lyzed by FTIR and Ra­man imag­ing with­out chem­i­cal tis­sue treat­ment to stay as close to the brain’s con­di­tions as pos­si­ble. Par­tic­u­lar­ly, the sec­ondary struc­ture-sen­si­tive Amide I band was an­a­lyzed spa­tial­ly re­solved in dif­fer­ent Aβ plaque types. Plaques in the an­a­lyzed re­gion were sub­se­quent­ly con­firmed by an­ti-Aβ im­muno­his­to­chem­istry (Aβ-IHC) on the same tis­sue sec­tion. We ob­served in­creased Aβ fib­ril con­tents along­side the as­cend­ing plaque stages. The spec­tral im­age analy­sis pro­vides in­sight in­to the spa­tial dis­tri­b­u­tion of Aβ struc­ture in dif­fer­ent plaque types, con­tribut­ing ev­i­dence for the cur­rent hy­pothe­ses on plaque development.