Gelsolin amyloidosis: genetics, biochemistry, pathology and possible strategies for therapeutic intervention

Abstract

Protein misassembly into aggregate structures, including cross-β-sheet amyloid fibrils, is linked to diseases characterized by the degeneration of post-mitotic tissue. While amyloid fibril deposition in the extracellular space certainly disrupts cellular and tissue architecture late in the course of amyloid diseases, strong genetic, pathological and pharmacologic evidence suggests that the process of amyloid fibril formation itself, known as amyloidogenesis, likely causes these maladies. It seems that the formation of oligomeric aggregates during the amyloidogenesis process causes the proteotoxicity and cytotoxicity characteristic of these disorders. Herein, we review what is known about the genetics, biochemistry and pathology of familial amyloidosis of Finnish type (FAF) or gelsolin amyloidosis. Briefly, autosomal dominant D187N or D187Y mutations compromise Ca(2+) binding in domain 2 of gelsolin, allowing domain 2 to sample unfolded conformations. When domain 2 is unfolded, gelsolin is subject to aberrant furin endoproteolysis as it passes through the Golgi on its way to the extracellular space. The resulting C-terminal 68 kDa fragment (C68) is susceptible to extracellular endoproteolytic events, possibly mediated by a matrix metalloprotease, affording 8 and 5 kDa amyloidogenic fragments of gelsolin. These amyloidogenic fragments deposit systemically, causing a variety of symptoms including corneal lattice dystrophy and neurodegeneration. The first murine model of the disease recapitulates the aberrant processing of mutant plasma gelsolin, amyloid deposition, and the degenerative phenotype. We use what we have learned from our biochemical studies, as well as insight from mouse and human pathology to propose therapeutic strategies that may halt the progression of FAF.

Figure 1

Scheme outlining the D187N/Y plasma gelsolin pathogenic proteolytic cascade. The 6 domains of gelsolin are depicted by rectangles with the D187N/Y mutation highlighted in red. A portion of the mutant gelsolin is cleaved within domain 2 by furin in the trans-Golgi network to generate a 68 kDa protein (C68). Upon secretion, the C68 is further cleaved by proteinases in the extracellular matrix, such as matrix metalloproteases, to produce the 8- and 5-kDa amyloidogenic gelsolin fragments that deposit systemically.

Figure 2

Gelsolin structure and activation by calcium binding. (A) Structure of calcium-free human gelsolin (PDB: 3FFN). Domains 1, 2, and 3 are shown in red, green, and blue respectively, and domains 4, 5, and 6 are shown in light red, light green, and light blue respectively. All of the domains are arranged in a compact fashion around domain 6, which has a C-terminal helix that interacts with both domains 2 and 4. (B) Structure of domains 1–3 of calcium-activated human gelsolin with G1–G3 bound to actin (shown in grey) (PDB:3FFK). Domains 1, 2, and 3 are once again shown in red, green, and blue respectively. The binding of calcium by domain 2 of gelsolin alters its conformation to allow new interactions with domain 3 (cf. the orientation of the red, blue and green domains in A vs. B). The cleavage sites of furin and the β-gelsolinase(s) are indicated. The residues that coordinate calcium binding in domain 2 are shown in ball-and-stick format (see Figure 3A for an expanded view). The structures depicted are based on the crystal structures as described by (Nag et al., 2009).

Figure 3

D187N/Y mutation compromises Ca⁺² binding in domain 2, altering its conformation to become protease sensitive. (A) In domain 2 of plasma gelsolin, a calcium ion (depicted by the black sphere) is coordinated by the backbone carbonyl of G186, and the carboxylates from the side chains of D187, E209, and D259. The D187N/Y mutation removes one of these Ca⁺²-binding side chains, making calcium unable to bind at this site at physiological concentrations. (B) As D187N/Y plasma gelsolin travels through the secretory pathway to be secreted into the bloodstream, domain 2 samples intermediate conformations between the unfolded (grey circle) and folded states (grey square), with unfolded and partially folded states being susceptible to furin cleavage in the trans Golgi. Once cleaved by furin, the secreted C68 fragment is susceptible to metalloprotease cleavage or endoproteolysis by other related proteases in the extracellular matrix, generating the 8 and 5 kDa amyloidogenic fragments. Wild type gelsolin retains its ability to bind calcium and is protected from furin cleavage.

Figure 4

Cleavage of gelsolin domain 2 by furin in vitro. Wild type gelsolin domain 2, as well as the D187N and E209Q Ca⁺²-binding-compromised variants of domain 2 were incubated without (−) or with (+) one unit of furin. Cleavage at the 172–173 furin recognition site was observed with D187N and E209Q (an engineered mutant), but not with wild type gelsolin domain 2. Figure reproduced from Huff et al. 2003b.

Figure 5

Atomic force microscopy images of 8 kDa gelsolin amyloidogenesis reactions. The 8 kDa gelsolin fragment was agitated via overhead rotation in the absence (A, C and E) or presence of heparin (10 µg/mL) (B, D and F), and the amyloidogenesis reaction was followed as a function of time. A and B reflect images recorded early in the amyloidogenesis time course (after 40 min of agitation). C and D reflect images recorded about halfway through the amyloidogenesis time course (after 100 min. for C; after 80 min. for D). E and F represent images at the completion of amyloidogenesis (after 180 min. of agitation). 8 kDa gelsolin fragment amyloidogenesis is accelerated in the presence of heparin. Figure reproduced from Solomon et al. 2011.

Figure 6

Gelsolin forms amyloid fibrils by way of a nucleated polymerization mechanism. This mechanism is characterized by a lag phase, during which a high energy oligomeric nucleus is formed, followed by a spontaneous growth phase, where amyloid fibrils grow rapidly, presumably by monomer addition to the nucleus or a growing fibril in a thermodynamically favorable manner. Amyloidogenicity can be monitored by thioflavin T binding-associated fluorescence. Thioflavin T is a dye that increases its fluorescence quantum yield upon binding to amyloid fibrils.

Figure 7

Heparin accelerates 8 kDa fragment gelsolin amyloidogenesis by binding and orienting gelsolin fragment cross-β-sheet oligomers. (A) Aggregation of the 8 kDa fragment of gelsolin is faster in the presence of heparin (1 µg/mL) (red trace) than in its absence (black trace). (B) Cross-β-sheet oligomers of gelsolin (assembly of green circles) formed by a nucleated polymerization from monomers (blue squares) accumulate post-nucleation and then bind to heparin (assembly of red hexagons). 8 kDa gelsolin oligomer sequestration by heparin accelerates the fibril extension phase of the amyloidogenesis reaction by binding, concentrating, aligning, and allowing fusion of oligomers into amyloid fibrils. Adapted from Solomon et al. 2011.

Figure 8

8- and, to a lesser extent, 5-kDa gelsolin amyloid deposition in a transgenic mouse model of familial amyloidosis of Finnish type (FAF). (A) Electron microscopy of amyloid isolated from 18-month old D187N gelsolin mouse. Fibrils are labeled with anti-FAF antibodies attached to 10 nm gold particles (indicated by arrowheads). (B) Schematic of a cross section of a muscle fiber with the same orientation as C–J. (C–J) Cross sections of muscle fibers were analyzed in D187N mice at 3 months (C,G), at 9 months (D,H) and at 18 months (E,I), and in 18-month wild-type controls (F,J). In 3-month mice, Congo red fluorescent deposits were localized exclusively around endomysial capillaries (C, arrows). By 9 months, Congo red fluorescence was present around endomysial capillaries but also extended into the endomysium surrounding muscle fibers (D, arrows). At 18 months, Congo red fluoresecence surrounded all myofibers with several fibers also showing internal sarcoplasmic deposits (E, arrow). By comparison, no Congo red positivity (F) was identified in 18-month wild-type mice. Electron micrographs confirmed fibrillar deposits in a pericapillary localization in 3-month D187N mice (G, arrow). At 9 months, more extensive fibrillar deposits surrounded capillaries (H, arrows) and also extended into the endomysium. Thick pericapillary and endomysial fibrillar deposits (I, black arrow) were found in 18-month D187N muscle. The large fibrillar deposits were between the sarcolemma (I, white arrows) of adjacent cells and the connective tissue of normal endomysium (I, white arrowhead) and were absent in wild-type siblings (J, sarcolemma (arrows); endomysium (arrowhead)). Capillaries indicated with *. The high magnification inset (I) shows the fibrillar nature of the deposits. Bar = 40 100nm for A, 100 µm for C–F, 3.3 µm for G, 2.2 µm for H, 2.0 µm for I, 0.3 µm for (I insert), 2.0 µm for J. Figure reproduced from Page et al. 2009.