Protein-Like Polymers Targeting Keap1/Nrf2 as Therapeutics for Myocardial Infarction

Abstract

Myocardial infarction (MI) results in oxidative stress to the myocardium and frequently leads to heart failure (HF). There is an unmet clinical need to develop therapeutics that address the inflammatory stress response and prevent negative left ventricular remodeling. Here, the Keap1/Nrf2 protein-protein interaction is specifically targeted, as Nrf2 activation is known to mitigate the inflammatory response following MI. This is achieved using a Nrf2-mimetic protein-like polymer (PLP) to inhibit the Keap1-Nrf2 interaction. The PLP platform technology provides stability in vivo, potent intracellular bioactivity, and multivalency leading to high avidity Keap1 binding. In vitro and in vivo assays to probe cellular activity and MI therapeutic utility are employed. These Keap1-inhibiting PLPs (Keap1i-PLPs) impart cytoprotection from oxidative stress via Nrf2 activation at sub-nanomolar concentrations in primary cardiomyocytes. Single-digit mg kg^-1, single-dose, intravenous PLP administration significantly improves cardiac function in rats post-MI through immunomodulatory, anti-apoptotic, and angiogenic mechanisms. Thus Keap1i-PLPs disrupt key intracellular protein-protein interactions following intravenous, systemic administration in vivo. These results have broad implications not only for MI but also for other oxidative stress-driven diseases and conditions.

Keywords: biomaterial; drug delivery; myocardial infarction; polymers.

Figure 1

Protein‐like polymers to inhibit the Keap1/Nrf2 protein–protein interaction in myocardial infarction. a) Keap1/Nrf2 pathway wherein oxidative stress or therapeutic inhibition of Keap1 can lead to Nrf2 accumulation and subsequent activation of the nuclear antioxidant response element (ARE), resulting in upregulation of cytoprotective genes and mitigation of inflammation. b) Protein‐like polymers (PLPs) incorporating Keap1‐inhibiting and scrambled control peptides synthesized via graft‐through ROMP. c) Molecular dynamics simulation showing the globular structure of PLP with the hydrophobic backbone (purple) surrounded by the hydrophilic peptide side chains (gray). The ETGE motif residues are highlighted in orange. d) Overview of studies done, where acute PLP administration was shown to elicit in vitro rescue from oxidative stress and lead to a pro‐reparative response in an MI model, with the figure panel created in BioRender.

Figure 2

Keap1i‐PLPs rescue cardiomyocytes following peroxide exposure via activation of the Nrf2 pathway. a) Kelch TR‐FRET binding assay showing pm inhibition for PLPs (n = 5 per concentration) of the Keap1 Kelch domain and lack of inhibition, as expected, for the scrambled control PLP (n = 3–5 per concentration). The Scramble‐PLP and Keap1i‐PLP were compared via a two‐tailed unpaired t‐test at each concentration. NOTE: A TR_FRET ratio of 100% is observed for vehicle control, representing no displacement, giving a signal equivalent to Scramble‐PLP. b) ARE‐Luciferase HepG2 reporter assay, showing ARE activation following Keap1i‐PLP treatment (n = 3–4 per concentration). The assay presented is relative to activation of the vehicle controls, normalized to 100% luminescence. See Fig S6B. c) Parallel MTS Cell Viability Assay in ARE‐Luciferase HepG2 cells (n = 3–4 per concentration). The assay presented is relative to viability of the vehicle controls, normalized to 100% viability. See Fig S6C. d) Cytocompatibility in rat neonatal cardiomyocytes (n = 5 per condition). 24 h after treatment with Keap1i‐PLPs as quantified by Alamar Blue. Cell viability is reported relative to the untreated control and compared to the positive cytotoxic control, 0.1% zinc diethyldithiocarbamate (ZDEC), with significance calculated using a one‐way ANOVA with Dunnett's post‐hoc test. e–h) Primary cells were pretreated with tert‐butyl hydroperoxide. PLP treatments were subsequently given 6 h after peroxide exposure. Rescue, determined via Alamar Blue cell viability readout, was conducted 24 h after PLP treatment, with diagram created in BioRender e). f) Keap1i‐PLPs rescue primary cardiomyocytes from peroxide‐induced stress (n = 3 for all conditions). g) Murine bone marrow‐derived macrophages were pre‐treated with tert‐butyl hydrogen peroxide (70 µM) and rescued via Keap1i‐PLP (n = 6 for all conditions). h) Rat cardiac endothelial cells were pre‐treated with tert‐butyl hydrogen peroxide (125 µM) and rescued via Keap1i‐PLP (n = 3 for all conditions). Significance was calculated using a one‐way ANOVA with Dunnett's post‐hoc test. i,j) qPCR of primary cardiomyocytes, pretreated with peroxide with subsequent administration of Keap1i‐PLP (0.1 nM and 3 µM) and Scramble‐PLPs at 3 µM. Keap1i‐PLP treatment groups show increased Hmox1 i) and Nqo1 j) expression as compared to the vehicle and Scramble‐PLP controls (n = 2–3 for all conditions). Significance was calculated using a one‐way ANOVA with Tukey's post‐hoc test. k,l) Gene expression in healthy (non‐peroxide treated) cardiomyocytes at a low (0.1 nM) and high (3 µM) (n = 2–3 for all conditions) derived from survival experiments performed in f). Compared to the vehicle control, Keap1i‐PLPs show significant and dose‐dependent activation of  two Nrf2‐mediated genes: Hmox1 and Nqo1. Significance was calculated using a one ‐way ANOVA with Tukey's post‐hoc test. Data are mean ± SEM. * p < 0.05, ** p < 0.01, *** p < 0.001, **** p < 0.0001.

Figure 3

Biodistribution of Keap1i‐PLPs. a) MI was induced in rats on Day 0, followed by IV tail vein administration of Cy5.5‐Keap1i‐PLPs (n = 4) or Cy5.5‐Scramble‐PLPs (n = 4) on Day 1. The heart and satellite organs were harvested for whole organ scanning via LiCoR Odyssey on Day 5. b,c) Both Cy5.5‐Keap1i‐PLPs b) and Cy5.5‐Scramble‐PLPs c) (dose 7.5 mg kg⁻¹, IV bolus tail vein injection) exhibited significant distribution to the LV‐infarct zone as compared to the right ventricle (RV). Significance was determined via a two‐way unpaired student's t‐test. d,e) Both the Keap1i‐PLP d) and the Scramble‐PLP e) show biodistribution patterns in satellite organs wherein the kidney is the primary site of distribution, followed by the liver and spleen. f,g) LiCoR imaging showing biodistribution of the Cy5.5‐Keap1i‐PLPs f) and Cy5.5‐Scramble‐PLP g) at 7.5 mg kg⁻¹ dose. h,i) Histology revealed PLPs (red) distributed within the LV‐infarct zone h) for both PLPs, with little to no PLP present in the remote myocardium i). Scale bars are 100 µm. Data are mean ± SEM. * p < 0.05, ** p < 0.01. *** p < 0.001, **** p < 0.0001.

Figure 4

Keap1i‐PLPs activate downstream inflammatory, angiogenic, and protective pathways in vivo following MI. a) To further probe the mechanisms of pro‐repair induced by Keap1i‐PLP treatments following MI, cardiac tissue was assessed for immunomodulatory, apoptotic, and angiogenetic metrics. b) Volcano plot of differentially expressed genes with upregulated genes relative to 3 mg kg⁻¹ Keap1i‐PLP treatment as positive fold changes with downregulated genes as negative fold changes (n = 5 for 3 mg kg⁻¹ Keap1i‐PLP, n = 4 for saline). c) Volcano plot of differentially expressed genes with upregulated genes relative to 7.5 mg kg⁻¹ Keap1i‐PLP treatment as positive fold changes with downregulated genes as negative fold changes (n = 5 for 7.5 mg kg⁻¹ Keap1i‐PLP, n = 4 for saline). Differential gene expression was determined via p < 0.05, and −1 < logFC < 1. d) Quantification of cleaved caspase 3 (apoptosis, red) of cardiomyocytes (α‐actinin, green) with saline, Scramble‐PLP, 3 mg kg⁻¹ Keap1i‐PLP, and 7.5 mg kg⁻¹ Keap1i‐PLP. Scale bar: 50 µm. e) Arteriole staining via colocalization of α‐SMA (red) and isolectin (green) was measured. Scale bar: 50 µm. f) Quantification of Cd68+ cells in the infarct. Scale bar: 20 µm. g) Quantification of anti‐inflammatory macrophages (Cd163+). Scale bar: 20 µm. Data in d‐g: n = 3 for Scramble‐PLP, n = 4 for saline, n = 5 for 3 and 7.5 mg kg⁻¹ Keap1i‐PLP, with significance determined via a one way ANOVA with Dunnett's post‐hoc test. Data are mean ± SEM. * p < 0.05, ** p < 0.01, *** p < 0.001.

Figure 5

Keap1i‐PLPs improve cardiac function and long‐term pathological changes due to MI. a) Schematic showing experimental design for functional assessment. Rats underwent surgeries to induce MI at Day 0. On Day 1, animals were arbitrarily assigned an IV injection of saline (n = 10) or 7.5 mg kg⁻¹ Keap1i‐PLP (n = 10), given after obtaining a baseline MRI. On Day 35, animals were reimaged with MRI prior to harvesting tissue. b,c) Absolute changes in ejection fraction b) and in infarct wall thickness c). Significance was determined via a two‐tailed unpaired t‐test. d–g) Comparison of healthy controls (n = 4) to saline and Keap1i‐PLP treated animals at the 5‐week timepoint post‐MI, particularly measuring EF d), ESV e), EDV f), and wall thickness g). Significance was determined via a one‐way ANOVA with Tukey's post‐hoc test. h) Cardiomyocyte area was measured by staining for wheat germ agglutinin (green). Scale bar: 400 µm. Significance was determined via a two‐tailed unpaired t‐test. i,j) Trichrome staining was used to quantify infarct fibrosis i) and interstitial fibrosis j) with Keap1i‐PLP administration. Scale bar in i = 2 mm. Scale bar in j = 50 µm. Significance was determined via a two‐tailed unpaired t‐test. k) Higher arteriole staining via colocalization of alpha‐SMA (red) and isolectin (green) was measured. Scale bar: 50 µm. Significance was determined via a two‐tailed unpaired t‐test. m). Fewer myofibroblasts were measured with Keap1i‐PLP administration compared to saline. Scale bar: 20 µm. Significance was determined via a two‐tailed unpaired t‐test. Data are mean ± SEM. * p < 0.05, ** p < 0.01, *** p < 0.001.