Huntington's disease induced cardiac amyloidosis is reversed by modulating protein folding and oxidative stress pathways in the Drosophila heart

Abstract

Amyloid-like inclusions have been associated with Huntington's disease (HD), which is caused by expanded polyglutamine repeats in the Huntingtin protein. HD patients exhibit a high incidence of cardiovascular events, presumably as a result of accumulation of toxic amyloid-like inclusions. We have generated a Drosophila model of cardiac amyloidosis that exhibits accumulation of PolyQ aggregates and oxidative stress in myocardial cells, upon heart-specific expression of Huntingtin protein fragments (Htt-PolyQ) with disease-causing poly-glutamine repeats (PolyQ-46, PolyQ-72, and PolyQ-102). Cardiac expression of GFP-tagged Htt-PolyQs resulted in PolyQ length-dependent functional defects that included increased incidence of arrhythmias and extreme cardiac dilation, accompanied by a significant decrease in contractility. Structural and ultrastructural analysis of the myocardial cells revealed reduced myofibrillar content, myofibrillar disorganization, mitochondrial defects and the presence of PolyQ-GFP positive aggregates. Cardiac-specific expression of disease causing Poly-Q also shortens lifespan of flies dramatically. To further confirm the involvement of oxidative stress or protein unfolding and to understand the mechanism of PolyQ induced cardiomyopathy, we co-expressed expanded PolyQ-72 with the antioxidant superoxide dismutase (SOD) or the myosin chaperone UNC-45. Co-expression of SOD suppressed PolyQ-72 induced mitochondrial defects and partially suppressed aggregation as well as myofibrillar disorganization. However, co-expression of UNC-45 dramatically suppressed PolyQ-72 induced aggregation and partially suppressed myofibrillar disorganization. Moreover, co-expression of both UNC-45 and SOD more efficiently suppressed GFP-positive aggregates, myofibrillar disorganization and physiological cardiac defects induced by PolyQ-72 than did either treatment alone. Our results demonstrate that mutant-PolyQ induces aggregates, disrupts the sarcomeric organization of contractile proteins, leads to mitochondrial dysfunction and increases oxidative stress in cardiomyocytes leading to abnormal cardiac function. We conclude that modulation of both protein unfolding and oxidative stress pathways in the Drosophila heart model can ameliorate the detrimental PolyQ effects, thus providing unique insights into the genetic mechanisms underlying amyloid-induced cardiac failure in HD patients.

Figure 1. Cardiac defects are associated with expression of mutant Poly-Q.

(A) Images of semi-intact fly heart preparations during systole and corresponding M-mode records obtained from high-speed movies of 1 week old control (PolyQ-25, top and mutant PolyQ-72, bottom) hearts. M-mode records (6 sec) show heart wall movements (at the location indicated by the blue line) over time . Double-headed arrows in the M-mode records indicate the location of the heart walls and the heart diameter during diastole and systole. Compared to PolyQ-25, hearts expressing PolyQ-72 showed significant dilation in the conical chamber and in the third abdominal segment of the cardiac tube. Cardiac arrhythmias are also apparent in the M-mode record from the PolyQ-72 expressing heart. (B) M-mode records from 3-week old control (Hand/+, PolyQ-25) and mutant (PolyQ-46, PolyQ-72, PolyQ-102) flies. Compared to Hand/+ or PolyQ-25, cardiac-specific expression of PolyQ-46, PolyQ-72 and PolyQ-102 showed increasingly arrhythmic beating patterns. Expression of PolyQ-72 and PolyQ-102 show frequent asystolic (non-beating) events. (C) Summary of the qualitative cardiac defects from 3-week old control (Hand/+, PolyQ-25) and mutant (PolyQ-46, PolyQ-72, PolyQ-102) flies showing the percent of flies exhibiting defective ostia, one or more non-contractile regions, and non-beating hearts. (D) Cardiac-specific expression of mutant PolyQ-72 and PolyQ-102 resulted in a significant reduction in lifespan compared to flies expressing non-disease causing PolyQ-25 and wild-type Hand/+ controls (p<0.001). Cardiac-specific expression of PolyQ-46 caused a small but significant decrease in lifespan (p<0.05). Graph plots % survival (n = 150–200 for each group) vs. time post-eclosion.

Figure 2. Disease-causing Poly-Q yields cardiac defects in 3 weeks.

(A) Systolic and (B) diastolic diameters of the hearts from PolyQ-46, PolyQ-72 and PolyQ-102 expressing flies were significantly higher than those of age-matched controls (Hand/+ or PolyQ-25). (C) Cardiac contractility was quantified as % Fractional Shortening; hearts expressing PolyQ-46, PolyQ-72 and PolyQ-102 showed significantly reduced contractility compared to control hearts. (D) Systolic intervals were prolonged in hearts expressing PolyQ-46, PolyQ-72 and PolyQ-102 as well as (E) diastolic intervals compared to control hearts (Hand/+ or PolyQ-25). (F) Cardiac arrhythmicity (AI) was significantly increased in hearts expressing PolyQ-46, PolyQ-72 and PolyQ-102 compared to controls. There was no statistical difference between Hand/+ and PolyQ-25 hearts in any of the cardiac function parameters. All data are shown as means ± SE; statistical significance was determined using one-way ANOVA and Dunnett's post hoc test. Significant differences were assumed for p<0.05. (* = p<0.05, ** = p<0.01, *** = p<0.001).

Figure 3. Mutant PolyQ causes structural defects and GFP-positive aggregates in the fly heart.

(A) Hearts from 3-week old PolyQ-25 controls show typical circumferential arrangements of actin-containing myofibrils within the cardiomyocytes and (B) homogeneously distributed GFP in the cardiomyocytes. (C) In hearts expressing PolyQ-72 the myofibrillar organization is severely disrupted (region indicated by dashed lines) with a loss of actin-containing myofibrils. (D) GFP-positive aggregates of various sizes are distributed throughout the cardiomyocytes in hearts expressing PolyQ-72. (E) Heart from 3-week old fly expressing PolyQ-25 shows myosin-containing myofibrils. This control heart shows typical circumferential arrangement of myosin-containing myofibrils (thick arrow). (F) The circumferential arrangement of myosin-containing myofibrils is nearly lost upon expression of mutant PolyQ-72 (F, dashed boxes). Thin arrows in E and F indicate myofibrils in the non-cardiac longitudinal fibers. (G). Dot blot protein assay (top) showing similar overall expression of Htt-PolyQ-25 GFP (control) and Htt-PolyQ-72 GFP protein relative to histone H2B (middle, loading control). Filter trap assay (bottom) shows more GFP positive protein aggregates present in the Htt-PolyQ-72 GFP expressing hearts compared to control hearts.

Figure 4. Mutant PolyQ causes myofibrillar and mitochondrial ultrastructural defects in the fly heart.

(A, A') Electron micrograph of a cross section through the dorsal vessel of 4-week old PolyQ-25 controls reveals a layer of contractile myocardial cells which form the heart tube and a non-contractile supporting layer of ventral-longitudinal fibers (VL). Note that myofibrils (MF) are intact and mitochondria (MT) contain densely packed cristae. (B, B') In contrast, the myocardial layer of 4-week old PolyQ-46 hearts contains gaps indicating some myofibrillar degeneration (arrow) and severely fragmented mitochondria (asterisks). Bars, 500 nm.

Figure 5. Mutant PolyQ induces oxidative stress.

(A–C) Immunofluorescence micrographs showing GFP expression and DHE staining in hearts expressing PolyQ-25. (D–F) GFP expression and DHE staining in hearts expressing PolyQ-46 and (G–I) GFP expression and DHE staining in hearts expressing PolyQ-72. Expression of mutant PolyQ correlates with enhanced DHE staining and more aggregation compared to age-matched control (PolyQ-25), with moderate length PolyQ (PolyQ-46) expressing hearts showing less staining than PolyQ-72 hearts. Note that many of the GFP-positive aggregates co-localize with areas of strong DHE staining in the merged images for both PolyQ-46 and 72 (arrows).

Figure 6. Cardiac function of hearts expressing control PolyQ-25 and mutant PolyQ-46 affected differently by oxidant.

(A & B) The already large systolic and diastolic heart diameters were further increased in the presence of the oxidant H₂O₂ in the hearts expressing PolyQ-46. (C) Cardiac contractility (% FS) was further decreased in the presence of the oxidant in hearts expressing PolyQ-46. (D & E) The already prolonged systolic and diastolic intervals in hearts expressing PolyQ-46 were further increased by feeding flies the oxidant H₂O₂. (F) The incidence of cardiac arrhythmias was further increased in the presence of oxidant compared to age-matched PolyQ-46 expressing hearts without oxidant (A–F). Hearts expressing PolyQ-25 in the presence of oxidant showed cardiac defects without cardiac dilations (A–F). Data are shown as means ± SE. (G) Actin staining and GFP aggregates in heart expressing PolyQ-46 without oxidant. (H) PolyQ-46 expressing heart from fly fed oxidant showed more myofibrillar defects and contained more GFP-positive aggregates compared to controls. (I) Number of aggregates/unit area is expressed as percentage relative to control. Flies fed oxidant had 30% more aggregates than did hearts from age-matched PolyQ-46 without oxidant. For all data statistical significance was determined using one-way ANOVA and Dennett's post-hoc test. Significant differences were assumed for p<0.05. (* = p<0.05, ** = p<0.01, *** = p<0.001).

Figure 7. Over-expression of SOD or treatment with resveratrol or over-expression of UNC-45 all rescue PolyQ-72 induced cardiac dysfunction to some extent.

(A) PolyQ-72 induced cardiac systolic and (B) diastolic diameters are reduced nearly to wild-type (Hand/+) levels. (C) Cardiac contractility (% FS) is improved upon over-expression of SOD, treatments with resveratrol or over-expression of UNC-45. (D) Systolic and (E) diastolic intervals and (F) cardiac arrhythmias from hearts expressing PolyQ-72 and expressing SOD-1 or fed resveratrol or expressing UNC-45 were reduced approaching towards wild-type values compared to hearts expressing PolyQ-72 alone. Hearts were from 3-week old flies. (A–F) Data shown as mean ± SE. (G) Myofibrillar organization and GFP-positive aggregates in PolyQ-72 expressing hearts and (H) in PolyQ-72 hearts over-expressing SOD or (I) flies fed resveratrol or (J) hearts over-expressing UNC-45. (K) Relative GFP-aggregate size was reduced in response to SOD over-expression or feeding resveratrol, however, over-expression of UNC-45 did not affect the average aggregate size. (L) Quantification of the number of aggregates per unit area shows that all treatments significantly reduced the density of aggregates formed in response to PolyQ-72 expression. K and L shows averaged data from 4 to 6 hearts, expressed as a percent relative to the mutant PolyQ. (M) Quantification of relative GFP-aggregate density with the filter trap assay. An example of one filter trap blot is shown (top), and the averaged densitometric scan results expressed as a percent relative to the mutant PolyQ (bottom). Compared to Htt-PolyQ-72, treatments with resveratrol or over-expression of SOD or UNC-45 reduced PolyQ-induced aggregation to different extents. For all panels significance was determined using a one-way ANOVA and Dunnett's post hoc test. Significant differences were assumed for p<0.05. (*** = p<0.001, ** = p<0.01, * = p<0.05, NS = no statistical difference, p>0.05). Red * and NS indicate differences relative to the Hand/+ control; black text indicates differences relative to mutant PolyQ-72. Scale bar in G is 20 µm.

Figure 8. Mitochondrial ultrastructural defects in mutant Poly-Q expressing hearts are suppressed with SOD over-expression.

(A, A') Cardiac expression of mutant PolyQ-46 resulted in myofibrillar degeneration (arrow), along with mitochondrial fragmentation (asterisks) in cardiomyocytes. (B, B') Over-expression of SOD in PolyQ-46 expressing hearts improved mitochondrial ultrastructure. MT indicates normally shaped mitochondria with densely packed cristae. Myofibrillar (arrow) degeneration was reduced to some extent in SOD-overexpressing hearts. MF refers to myofibrils. VL refers to non-cardiac ventral-longitudinal fibers. Scale bar is 500 nm.

Figure 9. Combined over-expression of UNC-45 and SOD-1 or over-expression of UNC-45 and treatment with antioxidant resveratrol are required for nearly complete suppression of PolyQ-induced cardiac defects and aggregation.

(A) Systolic diameters and (B) diastolic diameters of hearts from 3-week old flies expressing PolyQ-72 and either over-expressing both UNC-45 and SOD-1 or over-expressing UNC-45 with resveratrol treatment. Both treatments reduce cardiac dilation to wild-type values. (C) % FS is increased in response to both sets of manipulations to near wild-type levels. (D) Systolic intervals and (E) diastolic intervals of hearts from 3-week old flies; both treatments reverse the prolonged systolic and diastolic intervals observed in response to PolyQ-72 expression alone. (F) Cardiac arrhythmias in response to PolyQ-72 expression are reduced by both treatments to wild-type levels. Data shown as mean values ± SE; statistical significance was determined using one-way ANOVA and Dunnett's post hoc test. Significant differences were assumed for p<0.05. (*** = p<0.001). Red * and NS indicate differences relative to the Hand/+ control; black text indicates differences relative to mutant PolyQ-72. (G–O) Micrographs of hearts from 3-week old flies stained to show F-actin (phalloidin, red). (G, J, M) Hearts showed more organized myofibrils when co-overexpressing UNC-45 and SOD (compare G and J) or overexpressing UNC-45 and being treated with resveratrol (compare G and M). (H, K, N) PolyQ-72-induced GFP-tagged aggregates (green) were significantly reduced with both treatments (compare K and N to H). (I, L, O) Merged images. (P) Quantification of relative GFP-aggregate density (number of aggregates/unit area). Averaged data from 4 to 6 hearts is expressed as a percent relative to the mutant PolyQ. Over-expression of UNC-45 and SOD or over-expression of UNC-45 in the presence of resveratrol reduced the number of PolyQ-induced aggregates significantly. (Q) Quantification of relative GFP-aggregate content with a filter trap assay. Filter trap spot densities are shown at top, with densitometry scan results from averages of duplicate samples shown below; data are expressed as a percent relative to the mutant PolyQ. Compared to Htt-PolyQ-72, both over-expression of UNC-45 with SOD and over-expression of UNC-45 in the presence of resveratrol reduced PolyQ-induced aggregation significantly.