JNK2 Gene Silencing for Elastic Matrix Regenerative Repair

Abstract

Elastic fibers do not naturally regenerate in many proteolytic disorders, such as in abdominal aortic aneurysms, and prevent restoration of tissue homeostasis. We have shown drug-based attenuation of the stress-activated protein kinase, JNK-2 to stimulate elastic matrix neoassembly and to attenuate cellular proteolytic activity. We now investigate if JNK2 gene knockdown with small interfering RNA (siRNA) provides greater specificity of action and improved regenerative/antiproteolytic outcomes in a proteolytic injury culture model of rat aneurysmal smooth muscle cells (EaRASMCs). A siRNA dose of 12.5 nM delivered with a transfection reagent significantly enhanced downstream elastic fiber assembly and maturation versus untreated EaRASMC cultures. The optimal siRNA dose was also delivered as a complex with a polymeric transfection vector, polyethyleneimine (PEI) in preparation for future in vivo delivery. Linear 25 kDa PEI-siRNA (5:1 molar ratio of amine to phosphate) and linear 40 kDa PEI-siRNA (2.5:1 ratio) were effective in downregulating the JNK2 gene, and significantly increasing expression of elastic fiber assembly proteins, and decreases in elastolytic matrix metalloprotease-2 versus treatment controls to significantly increase mature elastic fiber assembly. The current work has identified siRNA dosing and siRNA-PEI complexing conditions that are safe and efficient in stimulating processes contributing to improved elastic matrix neoassembly via JNK2 gene knockdown. The results represent a mechanistic basis of a broader therapeutic approach to reverse elastic matrix pathophysiology in tissue disorders involving aberrations of elastic matrix homeostasis, such as in aortic aneurysms.

Impact statement

The elastic matrix and elastic fibers are key components of the structural extracellular matrix of elastic tissues and are essential to their stretch and recoil and to maintain healthy cell phenotype. Regeneration and repair of elastic matrix is naturally poor and impaired and is an unresolved challenge in tissue engineering. In this work, we investigate a novel gene silencing approach based on inhibiting the JNK2 gene, which provides significant downstream benefits to elastic fiber assembly and maturation. Combined with novel delivery strategies such as nanoparticles, we expect our approach to effect in situ elastic matrix repair in the future.

Introduction

Proteolytic disorders are characterized by chronic inflammation of tissues leading to progressive enzymatic breakdown of the structural extracellular matrix (ECM).^1,2 In a prototypic proteolytic disorder, abdominal aortic aneurysms (AAAs), the wall elastic matrix, which provides stretch and recoil, is broken down by matrix metalloproteinases (MMP)^3,4 chronically overexpressed by diseased smooth muscle cells (SMCs) at the AAA site.^5,6 Despite their multifactorial causes, elastolysis is a hallmark of AAA development⁷ due to the elastic matrix comprising up to 50% dry vessel weight.⁸ Inherently, poor elastic matrix regeneration and repair by adult and diseased cells represent a critical challenge to reversing the pathophysiology of proteolytic disorders. To enable this, a sustained stimulus to elastogenesis and deterrent to elastolysis must be provided locally in the AAA wall toward restoring elastin homeostasis and thus stabilizing or regressing AAA growth.⁷

We have shown that MMP-inhibitor, doxycycline (DOX), when released at low doses from polymer nanoparticles, stimulates elastic matrix neoassembly in aneurysmal rat SMC cultures,⁸ an effect that was attributed to its inhibition of the protein kinase, c-Jun N-terminal kinase (JNK2), which is upregulated in AAA tissue^9,10 as in other proteolytic disorders (e.g., chronic obstructive pulmonary disease). JNK2 attenuation is thus a promising approach to enhance elastic matrix regenerative repair in AAAs.¹¹ To provide greater specificity and hence potency of JNK inhibition, we presently investigate effects of small interfering RNA (siRNA) on rat aneurysmal SMC cultures. Complexing anionic siRNA with a polymeric transfection vector, polyethyleneimine (PEI), can enhance its penetration of the cell membrane for cytoplasmic uptake gene silencing function.

In this study, we compared the safety and gene knockdown effects of PEI-siRNA conjugates as a function of siRNA dose, PEI type, and PEI/siRNA and importantly the downstream effects on elastic fiber neoassembly and anti-MMP outcomes in rat aneurysmal SMC cultures.

Materials and Methods

AAA induction and isolation and culture of rat aneurysmal SMCs

Animal procedures were approved by the Cleveland Clinic's Institutional Animal Care and Use Committee. Elastase injury aneurysmal rat aortic smooth muscle cells (EaRASMCs) were isolated from AAAs induced in young male Sprague-Dawley rats by elastase infusion, as previously published.^12,13 Primary cells were propagated over 2 weeks in 10% v/v fetal bovine serum (FBS; PAA Laboratories, Etobicoke, Ontario) containing DMEM-F12 medium (ThermoFisher, Waltham, MA) and 1% v/v PenStrep (ThermoFisher). Cells were harvested and pooled (n = 3 rats/pool) before culture at passages 2–6. Healthy rat aortic SMCs (RASMCs) were similarly isolated from aortae from separate animals. Phenotypic characterization of the cells was performed as previously published.¹³

siRNA treatment for JNK-2 gene silencing

siRNA and commercial transfection reagent

RASMCs (healthy controls) were cultured at 30,000/well in six-well plates for 7 days before harvesting and analysis. At 5 days of culture, EaRASMCs were serum starved (3 h), then stimulated with cytokines, interleukin-1 beta (IL1-β; 10 ng/mL; Bio Basic Inc., Ontario, Canada), and tumor necrosis factor-alpha (TNF-α; 10 ng/mL; PeproTech, Rocky Hill, NJ) in serum-free DMEM-F12 medium (30 min). EaRASMCs were treated with scrambled or anti-JNK2 siRNA (ON-TARGETplus Non-targeting pool and Rat Mapk9 SMARTpool siRNA; Dharmacon, Lafayette, CO) in DMEM-F12 medium containing 2% v/v FBS. siRNA was delivered in presence of DharmaFECT 2 TR (Dharmacon) for intracellular uptake of siRNA over 8 h.

PEI-siRNA transfection

As in siRNA and Commercial Transfection Reagent section, at 5 days of culture EaRASMCs were serum starved (3 h), then stimulated with IL1-β (10 ng/mL) and TNF-α (10 ng/mL) prepared in serum-free DMEM-F12 medium (30 min). PEI (25 kDa linear transfection grade, Polysciences, Warrington, PA; 25 kDa branched, Millipore Sigma, Burlington, MA; 40 kDa linear, Polysciences) was complexed with JNK2 siRNA (12.5 nM; Dharmacon) at amine (N) to phosphate (P) molar ratios of 2.5:1, 5:1, 10:1, 20:1, and 30:1. EaRASMCs were transfected (8 h) with the complexes dissolved in DMEM-F12 with 2% v/v FBS.

Experimental design

To assess JNK-2 gene knockdown effects on downstream elastogenesis, EaRASMCs treated with siRNA with transfection reagent (TR; for optimizing siRNA dose) and complexed PEI-siRNA (at the optimized siRNA dose) were evaluated at 7 and 21 days postseeding. The 7 day cultures were assessed (polymerase chain reaction [PCR], western blots, LIVE/DEAD imaging, enzyme-linked immunosorbent assay (ELISA), gel zymography, immunofluorescence [IF] for proteolytic proteins) for JNK-2 gene knockdown and proteolytic activity. The 21-day cultures were assessed for elastic matrix assembly (desmosine, DNA/Fastin assays, IF for elastic matrix proteins). All cultures were transfected weekly on the fifth day of each week of culture.

LIVE/DEAD assay to assess cytotoxicity of PEI-siRNA

To assess their cytotoxicity, PEI-siRNA were administered to cytokine-treated EaRASMCs at 5 days of culture and stained with LIVE/DEAD assay (Invitrogen) reagents. The cells [n = 3 cultures/group; six regions of interests (ROIs) per cell layer] were imaged using an Olympus IX51 fluorescence microscope. ImageJ^® software quantified the % fraction of all cells that were live (green) as measure of viability.

Real time polymerase chain reaction for verification of gene silencing

Cultures were harvested in RLT buffer with 1% w.v beta-mercaptoethanol (Qiagen, Hilden, Germany) and diluted in 70% v/v ethanol (n = 6 cultures/group; 30,000/well in six-well plates). RNA was isolated from cultures using an RNeasy mini kit (Qiagen, Valencia, CA) and quantified using a Quant-iT™ RiboGreen^® RNA Kit (Invitrogen). iScript cDNA synthesis kit (Bio-Rad, Hercules, CA) was used to synthesize cDNA using 1 μg RNA, and reverse transcription was performed per manufacturer's instructions. Real-time PCR was performed on an Applied Biosystems 7500 Real-Time PCR system with Power SYBR^® Green Master Mix (Applied Biosystems). Gene expression was quantified against housekeeping gene, 18s. Primer sequences (Real-Time Primers, LLC, Melrose Park, PA) details are detailed in Table 1. Data were analyzed using LinReg PCR program as published¹⁴ and used to calculate gene expression ratios.¹⁵

Table 1.

List of Primer Sequences Used for Polymerase Chain Reaction Analysis of Gene Expression Changes Upon Small Interfering RNA/Polyethyleneimine-Small Interfering RNA Treatment

Gene	Protein	Catalog no.	Manufacturer	Forward Primer (5′→3′)	Reverse Primer (3′→5′)
18S	18s	Custom	Real Time Primers		CGGACAGGATTGACAGATTG	ACGCCACTTGTCCCTCTAAG
MAPK9	JNK-2	VRPS-3524	Real Time Primers		CGATTGAAGAGTGGAAAGAAC	GGATGAGATGTCATTGATGG
ELN	Elastin	VRPS-1839	Real Time Primers		CCTGGTGGTGTTACTGGTATTGG	CCGCCTTAGCAGCAGATTTGG
LOX	LOX	VRPS-3359	Real Time Primers		AGACGATTTGCCTGTACTGC	ATAGGCGTGATGTCCTGTGT
FBN1	Fibrillin-1	186806917-1/2	Applied Biosystems		ATAAATGAATGTGCCCAGAATCCC	ACTCATCCTCATCTTTACACATCC
EFEMP2	Fibulin-4	4694957-4/5	Applied Biosystems		GGCTCTGCCAAGACATTGTA	GACACTTGGACATAGGGCTC
FBLN5	Fibulin-5	VRPS-2035	Real Time Primers		CGAGGGTCGAGAGTTCTACA	CAGAACGGATACTGGGACAC
MMP2	MMP-2	VRPS-3674	Real Time Primers		GGAGCGACGTAACTCCACTA	AAGTGAGAATCTCCCCCAAC
MMP9	MMP-9	VRPS-3682	Real Time Primers		ACTTCTGGCGTGTGAGTTTC	TGTATCCGGCAAACTAGCTC
TIMP1	TIMP-1	VRPS-6297	Real Time Primers		CATGGAGAGCCTCTGTGGAT	ATGGCTGAACAGGGAAACAC
MAGP1	MAGP-1	VRPS-3611	Real Time Primers		GTCCAACAGGAAGTCATCCCAG	CCTGTGTATGGAGTAGAGGCGG
TIMP2	TIMP-2	Custom	Real Time Primers		ACCCTCTGTGACTTCATCGTGC	GGAGATGTAGCACGGGATCATG
LOXL1	LOXL1	VRPS-3360	Real Time Primers		CGACTATGACCTCCGAGTGCTA	GTAGTGGCTGAACTCGTCCATG

Western blots for expression of signaling and ECM proteins

Western blot analysis (n = 3 cultures/group; 30k/well in six-well plates) was performed as published.⁸ Cell layers were harvested in RIPA buffer with protease inhibitor cocktail (Thermo Scientific, Waltham, MA). Protein amounts were quantified using a bicinchoninic acid (BCA) assay kit (Thermo Scientific). Equal protein amounts (2 to 5 μg/well) were mixed with loading buffer, reduced, then loaded with prestained MW ladder (HiMark^® or SeeBlue^®; Invitrogen), onto NuPAGE^® Novex^® 3–8% tris-acetate (for MW <500 kDa), 4–12% bis-tris (for MW <60 kDa), or 10% bis-tris polyacrylamide gels (for MW >60 kDa; Thermo Scientific). Antibodies used are listed in Table 2.

Table 2.

List of Proteins Used for Immunofluorescence Imaging and Western Blot Analysis of Changes in Expression Upon Small Interfering RNA/Polyethyleneimine-Small Interfering RNA Treatment

Protein	Duration of culture (days)	Immuno-fluorescence dilution (v/v)	Western blot dilution (v/v)	Catalog no.	Source
JNK-2	7	1:250	1:1000	Ab179461	Abcam
pJNK	7	1:20	1:1000	AF1205	R&D Systems
MMP-2	7	1:200	1:500	Ab37150	Abcam
MMP-9	7	NA	1:500	Ab19016	Abcam
TGF-β	7	1:400	NA	Ab119558	Abcam
Elastin	21	1:200	NA	Ab2039	Abcam
Fibrillin-1	21	1:200	1:200	Bs-1157R	Bioss Antibodies
MAGP-1	21	1:500	NA	ABT39	Millipore Sigma
Fibulin-4	21	1:250	1:2500	NBP1-84724	Novus
Fibulin-5	21	1:200	1:500	Ab202977	Abcam

NA.

ELISA for transforming growth factor-beta 1 expression

An ELISA (Abcam) was used to analyze spent medium from siRNA-treated and control cultures (n = 3 cultures/group; 30k/well in six-well plates, 7 days) for secreted mature transforming growth factor-beta (TGF-β).

Gel zymography for MMP activity

Gel zymography quantified MMP-2 and -9 enzyme activities (n = 3 cultures/group; 30,000/well in six-well plates, 7 days). A BCA assay (Thermo Scientific) quantified protein content. Protein (2 μg protein/sample) was loaded onto a 10% zymogen gel (Invitrogen) with 10 ng of MMP2 and -9 protein standards (EMD Millipore, Billerica, MA) and a prestained protein ladder (SeeBlue; Invitrogen). Gels were electrophoresed at 125 V (2 h, washed in 2.5% v/v Triton-X-100 (30 min), and incubated in Novex Zymogen Developing Buffer (48 h, 37°C; Thermo Scientific). Gels were washed and stained with SimplyBlue™ SafeStain (Thermo Scientific) for 1 h before destaining gels until clear bands were visible (48 h). Gels were scanned using a flatbed scanner with ImageJ to quantify band intensities.

DNA assay for cell proliferation and fastin assay for elastic matrix amounts

DNA content (n = 3 cultures/group; 30,000 cells/well in six-well plates, 21 days) was quantified using a fluorometric assay to deduce cell proliferation.¹⁶ Cells were harvested in NaCl-Pi buffer and evaluated for DNA content. Cell count was estimated based on 6 pg of DNA/cell. Elastic matrix and soluble elastin precursor, tropoelastin, were quantified using a Fastin assay (Accurate Scientific and Chemical, Westbury, NY), as published.^17,18 Cell number-normalized elastin amounts were compared between groups.

Desmosine assay

Cultures were harvested in phosphate-buffered saline (PBS), centrifuged (500 g, 5 min), hydrolyzed with 6N HCl (48 h, 105°C), evaporated to dryness, and reconstituted in water (n = 6 cultures/group. 30k/well in six-well plates, 21 days). Samples were filtered through 0.45 um filters and desmosine levels determined using a competitive ELISA assay.¹⁹ Total protein/sample was measured using a ninhydrin assay.²⁰

IF labeling to confirm siRNA uptake

IF was used to visualize cellular PEI-siRNA uptake by EaRASMCs (n = 3 cultures/group, 30,000/well in chamber slides, 7 days), versus treatment controls. siRNA was labeled using a Silencer Cy3 Labeling Kit (ThermoFisher), then dosed to cultures 5 days postseeding (8 h). Immediately following transfection, cell layers were fixed in 4% w/v paraformaldehyde (25°C, 15 min; Fisher Scientific). The actin cytoskeleton was visualized with Alexa Fluor 488 phalloidin (1:40 v/v dilution, 25°C, 20 min; Invitrogen) and cell nuclei with aqueous 4′,6-diamidino-2-phenylindole (DAPI, 1:10,000 v/v dilution, 25°C, 10 min; Vector Laboratories, Burlingame, MA). Cell layers were imaged on a Leica SP8 confocal microscope (Leica Microsystems, Inc., Buffalo Grove, IL) and the ratio of corrected total cell fluorescence to number of nuclei within field of view determined using ImageJ^® (n = 6 ROIs/call layer).

IF to visualize expression of signaling and ECM proteins

Cell layers (n = 3 cultures/group, 30k/well in chamber slides, 7 days (JNK2-related) or 21 days (ECM related) were fixed in 4% w/v paraformaldehyde (25°C, 15 min), blocked with 5% v/v goat serum in PBS (20 min; ThermoFisher Scientific), and proteins detected using a rabbit anti-rat polyclonal antibodies (4°C, 24 h), and visualized with a Alexa Fluor 633 fluorophore-conjugated goat anti-rabbit secondary antibody (1:1000 v/v dilution, 25°C, 1 h; Invitrogen). Actin cytoskeleton and DAPI were visualized as described in 2.10. siRNA-untreated EaRASMCs and siRNA-treated, but primary Ab-unexposed EaRASMCs, served as treatment and imaging controls, respectively. All images were obtained using a Leica SP8 confocal microscope. Fluorescence due to expressed protein was normalized to number of nuclei within field of view on ImageJ software and values compared between groups (n = 3 ROIs/cell layer).

Transmission electron microscopy for matrix ultrastructure

At 21 days of culture (50k/well on Permanox chamber slides; ThermoFisher), cell layers were fixed in a solution of 4% w/v paraformaldehyde, 2.5% w/v glutaraldehyde, and 0.1 M sodium cacodylate, incubated (overnight, 4°C), and postfixed in 1% w/v osmium tetroxide (1 h). Samples were dehydrated in a graded ethanol series (50–100% v/v), embedded in Epon 812, sectioned, and mounted on copper grids where they were stained with uranyl acetate and lead citrate before imaging on a Hitachi TEM H7600T (High Technologies, Pleasanton, CA).

Statistical analysis

Experiments were performed with n = 6 replicate cultures per cell type with the following exceptions: TGF-β ELISA; n = 3, Zymogram for exogenously delivered siRNA; n = 3, PEI-siRNA LIVE/DEAD; n = 3, PEI-siRNA IF; n = 3. Results are reported as mean ± SE. Outlying data were eliminated if outside 1.5 × interquartile range. A Student's t-test or one-way analysis of variance was used to analyze significant differences in group-wise outcomes (p < 0.05).

Results

Effects of exogenous JNK-2 siRNA+TR on EaRASMCs

Exogenous JNK-2 siRNA with TR were used to treat EaRASMCs in various doses to best optimize JNK-2 gene knockdown. Real-time PCR (Supplementary Fig. S1) indicated lower expression of JNK-2 gene (p < 0.0031) versus treatment control (TC) for all doses (12.5–100 nM) of exogenous JNK-2 siRNA. Western blot (Supplementary Fig. S2A) showed JNK-2 protein synthesis to be lower versus TC at all the siRNA doses (12.5 nM: p = 0.004, 25 nM: p = 0.004, 50 nM: p = 0.042, 100 nM: p = 0.019). Progressively lower MMP-2 enzyme activity versus TC cultures was noted with increasing siRNA dose (Supplementary Fig. S2B). ELISA showed only the 50 nM siRNA-treated cultures to exhibit higher TGF-β expression versus TC cultures (Supplementary Fig. S2C; p = 0.037).

Further experiments were conducted using only the 12.5 nM dose since (1) JNK2 gene knockdown was significant and not enhanced any further at higher siRNA doses (Supplementary Fig. S1) and because (2) it was the only siRNA dose that increased expression of proelastogenic TGF-β1 (Supplementary Fig. S2C). Cell viability in 12.5 nM JNK2 siRNA-treated cultures was nearly 100%.

Protein content normalized desmosine amounts were higher in the 12.5 nM siRNA-treated cultures versus TC (Supplementary Fig. S2D; p = 0.009). IF images in Supplementary Figure S3 indicated lower JNK-2 expression in 12.5 nM siRNA-treated cultures versus TC and scrambled siRNA-treated cultures (p = 0.001 vs. TC, p = 0.009 vs. scrambled siRNA), and significantly lower expression of MMP-2 in the 12.5 nM JNK siRNA-treated cultures versus TC and scrambled siRNA-treated cultures. TGF-b expression in the 12.5 nM JNK siRNA-treated cultures was higher in magnitude than the control cultures, but the differences were not deemed statistically significant. Transmission electron microscopy (TEM) micrographs (Supplementary Fig. S4) showed presence of tropoelastin coacervating into mature elastic fibers of typical sizes (1–3 μm in diameter) in 12.5 nM JNK-2 siRNA-treated cultures (Supplementary Fig. S4A) with only sporadic elastin deposits seen in untreated EaRASMC cultures (Supplementary Fig. S4B).

Effect of PEI on cell viability and siRNA binding

Exogenous JNK-2 siRNA was coupled with PEI to treat determine effects on viability of EaRASMCs in various doses to best optimize cell survivability and siRNA binding to PEI. Results in Figure 1A and B indicate near 100% viability of EaRASMCs for most PEI-siRNA preparations using a 12.5 nM siRNA dose. However, exceptions include 25 kDa branched PEI at a 2.5:1 N:P ratio, and 40 kDa linear PEI for all N:P ratios >5:1. Linear PEI (25 and 40 kDa) at a 30:1 N:P ratio reduced cell viability to below that in the TC (p = 0.0004 for 25 kDa PEI and p < 0.0001 for 40 kDa PEI). While no significant difference was found between the TC and 40 kDa 10:1 and 20:1 N:P ratios, the clear drop in cytotoxicity shown in Figure 1A and B posed too great a risk to continue on in the study.

FIG. 1.

PEI-siRNA viability and complexing. LIVE/DEAD assay (A) images (20 × ) indicated the predominance of live cells in all PEI-siRNA-treated cultures. Plot in (B) shows quantification of live cells as a % of all cells in a field of view (n = 6 FOVs/culture/condition) following PEI-siRNA dosing. *, ** indicate statistical significance of differences versus TC deemed for p < 0.05, p < 0.01, respectively. (C) Gel electrophoresis image of PEI-siRNA complexes. When fluorescently tagged siRNA effectively complexes with PEI, it cannot migrate through the gel. FOV, field of view; PEI, polyethyleneimine; siRNA, small interfering RNA; TC, treatment control. Color images are available online.

Results of gel electrophoresis (Fig. 1C) indicated fluorescent, migrating bands in lanes loaded with noncomplexed siRNA controls, and siRNA complexed with 25 kDa branched PEI at N:P molar ratios of 2.5:1 and 5:1 with a very faint band at 10:1. Therefore, only 25 kDa linear PEI at all doses and 40 kDa linear PEI at 2.5:1 and 5:1 doses were investigated further.

Effect of PEI on siRNA uptake and localization within cells

Exogenous JNK-2 siRNA with PEI were used to treat EaRASMCs in various doses to best optimize PEI-siRNA uptake into the cell cytoplasm. Figure 2A shows uptake and localization of fluorescently tagged siRNA in cells (n = 3 ROIs/cell layer, n = 3 cell layers/group). Figure 2B compares this fluorescence between culture groups (Fig. 2B). Significantly higher fluorescence intensities were seen in cultures receiving siRNA complexed with 25 kDa PEI at a 5:1 molar ratio of N:P (Fig. 2C; p = 0.006) and 40 kDa PEI at a 2.5:1 molar ratio of N:P (Fig. 2D; p = 0.003) compared to the background fluorescence in the TC cultures. No differences were found between PEI-siRNA-treated groups. XYZ planar images showed the siRNA to mostly localize within the cell cytoplasm, indicating successful delivery in both cases.

FIG. 2.

PEI-siRNA localization. (A) Fluorescence micrographs of 21 day cultures of EaRASMCs either untreated or treated with 25 kDa PEI-siRNA (5:1 N to P ratio) and 40 kDa PEI-siRNA (2.5:1 N to P ratio). Seen are DAPI-stained nuclei (blue), cell membrane (red), and Cy3-tagged siRNA (green). For each case, two columns of images are shown. The left column contains all three channels (red, blue, green), while the right column contains only blue and green channels to improve visibility of siRNA for easier comparison among cases. Images taken at 20 × . (B) Quantification of fluorescence intensity associated with Cy3-tagged siRNA internalized within cells. The fluorescence intensity is presented normalized to cell count in the field of view. ** indicates statistical significance of differences versus TC, deemed for p < 0.01. (C, D) XYZ planar immunofluorescence image (40 × ) showing 25 kDa PEI-Cy3-siRNA complexes (5:1 N to P ratio), (C) and 40 kDa PEI-Cy3-siRNA (2.5:1 N to P ratio) (D) (green). DAPI-stained nuclei appear blue and the cell membrane, red. EaRASMCs, selastase injury aneurysmal rat aortic smooth muscle cells. Color images are available online.

Effects of PEI-siRNA on gene expression in EaRASMCs

JNK-2 targeting PEI-siRNA were used to treat EaRASMCs in various doses to best optimize JNK-2 gene knockdown and downstream proelastogenic and antiproteolytic gene activity. Real-time PCR (Fig. 3) indicated lower JNK2 gene expression in cultures treated with siRNA+TR versus TC (p = 0.003). All other groups were found to be not different compared to TC. Elastin gene expression was found to be largely unaffected upon treatment. It was significantly upregulated in the siRNA+TR cultures (p < 0.0001) and in the 25 kDa linear PEI-siRNA (20:1 ratio)-treated cultures versus TC (p = 0.001). MMP2 gene expression was increased in cultures treated with 25 kDa linear PEI-siRNA (20:1 ratio) versus TC (p = 0.023).

FIG. 3.

Cellular gene expression following PEI-siRNA treatment. Gene expression profiles in PEI-siRNA-treated and TC at 7 days of culture, as measured using RT-PCR. In each plot, results are shown for cultures-treated exogenous siRNA+TR (left group), with 25 kDa linear PEI (middle group; N:P ratios increasing from 2.5:1 to 30:1 left to right), and with 40 kDa (right group; N:P ratios of 2.5:1 and 5:1 left to right). Gene expression is normalized to housekeeping gene, 18s. *, **, ***, **** indicate statistical significance deemed p < 0.05, p < 0.01, p < 0.001, and p < 0.0001, respectively. RT-PCR, real time polymerase chain reaction; TR, transfection reagent. Color images are available online.

FBN1 gene expression was increased in cultures treated with siRNA+TR (p < 0.0001) and 25 kDa PEI-siRNA (20:1 and 30:1 ratios; p = 0.04 and 0.003 respectively). FBLN4 gene expression was higher versus TC in cultures treated with siRNA+TR (p < 0.001) and 25 kDa PEI-siRNA (5:1 and 30:1 ratios; p = 0.0002 and 0.025 respectively). FBLN5 expression was higher versus TC in cultures treated with siRNA+TR only (p < 0.0001).

LOX and LOXL1 expression was increased upon treatment with siRNA+TR (p < 0.0001 for both proteins) versus TC, and there were no differences between PEI-siRNA-treated groups and TC. Relative to the TC, MAGP1 expression was upregulated in siRNA+TR cultures (p < 0.0001), but was not different (and at the limit of detection) in PEI-siRNA-treated cultures. TIMP1 and TIMP2 expression was higher in siRNA+TR-treated cultures versus TC (p < 0.0001 for both proteins), but no differences were seen between TC and PEI-siRNA-treated cultures.

Effect of PEI-siRNA on proteolytic and elastic matrix-associated proteins

JNK-2 targeting PEI-siRNA were used to treat EaRASMCs in various doses to best increase downstream proelastogenic and decrease downstream antiproteolytic protein expression. Western blots (Fig. 4A) showed no significant differences in pJNK, MMP-9, LOXL1, or fibulin-4 protein expression in siRNA-treated cultures versus TC. Gel zymography (Fig. 4B) did not detect expression of the zymogen or inactive form of MMP-2 protein (72 kDa) in PEI-siRNA-treated cultures. Accordingly, the plot only shows the active MMP-2 form (62 kDa).

FIG. 4.

Protein expression following PEI-siRNA treatment. (A) Results of western blot analysis of the expression of key proteins involved in elastic fiber homeostasis in PEI-siRNA-treated EaRASMC cultures. Plot show results for 25 kDa 5:1 (N:P ratio; left bars)- and 40 kDa 2.5:1 (right bars)-treated cultures. Beta actin was analyzed as a housekeeping protein. All results are normalized to TC. (B) MMP-2 and -9 activity determined using gel zymography and normalized to TC. (C) TGF-β expression determined using ELISA assays. (D) Elastin and desmosine content quantified using a Fastin assay and ELISA, respectively. *, **, ***, **** indicate statistical significance of differences compared to TC and deemed for p < 0.05, p < 0.01, p < 0.001, and p < 0.0001, respectively. ELISA, enzyme-linked immunosorbent assay; MMP, matrix metalloproteinase; TGF-β, transforming growth factor-beta. Color images are available online.

In 25 kDa PEI-siRNA-treated cultures, JNK-2 protein synthesis was lower (p = 0.037) and expression of active MMP-2 isoform (p = 0.044) and TIMP-2 (p = 0.023) was higher versus TC (Fig. 4A). ELISA (Fig. 4C) showed TGF-β expression to be higher in 25 kDa 5:1 PEI-siRNA-treated cultures versus TC (p = 0.020). Protein-normalized desmosine amounts (Fig. 4D) were higher in 25 kDa 5:1 PEI-siRNA-treated cultures versus TC (p = 0.0001).

In 40 kDa 2.5:1 PEI-siRNA-treated cultures, JNK-2 and LOX protein expression were lower (JNK-2: p < 0.0001; LOX: p = 0.006) versus TC, while expression of the active MMP-2 (p = 0.008), MMP2/TIMP2 ratio (p = 0.018), and fibulin-5 (p = 0.039) expression were higher compared to TC (Fig. 4A). TIMP-2 and LOX protein expression in 25 kDa PEI-siRNA-treated cultures were higher (p = 0.004 and p = 0.026, respectively) than in cultures treated with 40 kDa PEI-siRNA. MMP-2 enzyme activity was lower and MMP-9 enzyme activity higher in 40 kDa PEI-siRNA-treated cultures versus TC (p = 0.031, 0.0004 respectively). TGF-β expression (Fig. 4C) and desmosine amounts normalized to protein content (Fig. 4D) were higher in 25 kDa PEI-siRNA-treated cultures versus TC (p = 0.020, p = 0.0001, respectively) but not in cultures treated with 40 kDa PEI-siRNA (p = 0.658, 0.514, respectively, versus TC).

Figure 4D shows higher elastin production per cell in 40 kDa PEI-siRNA-treated cultures versus linear 25 kDa PEI-siRNA-treated cultures (p = 0.022) and TC (p = 0.013). Cell count following 21 days of culture for TC, 25 kDa, and 40 kDa PEI-siRNA-treated cultures was found to increase by 6.97-fold, 6.74-fold, and 6.92-fold, respectively, compared to time control culture counted 1-day postseeding.

Figure 5A and B show lower cell number-normalized fluorescence for JNK2 expression in 25 kDa PEI-siRNA 5:1 (p = 0.0385) and 40 kDa PEI-siRNA 2.5:1 (p = 0.044)-treated cultures versus TC. pJNK expression was attenuated only in the latter case (p = 0.0005). TGF- β expression was significantly greater in 25 kDa 5:1 PEI-siRNA-treated cultures compared to TC cultures (p = 0.017), but not different in 40 kDa PEI-siRNA-treated cultures relative to other culture groups. MMP-2 expression was not different between culture groups. IF imaging for elastic matrix proteins (Fig. 5C, D) showed elastin and fibulin-5 expression to be significantly higher in PEI-siRNA-treated culture groups versus TC (elastin: p < 0.0001 for 25 kDa PEI-siRNA and p = 0.031 for 40 kDa PEI-siRNA; p = 0.010 for 25 kDa and p = 0.005 for 40 kDa PEI-siRNA for fibulin-5).

FIG. 5.

Immunofluorescence detection and quantification of key proteins influencing elastic fiber homeostasis in PEI-siRNA-treated cultures. Cultures were treated with PEI-siRNA for 30 min at either 7 days (B) or 21 days (D) of culture of EaRASMCs. In the IF micrographs, DAPI-stained nuclei appear blue, actin cytoskeleton appear green, and the protein of interest is red. In (B, D), all images were taken at 20 × . In (D), for each culture group, two rows of images are shown. The top row contains all three channels (red, blue, green), while the bottom row contains only blue and red channels to improve visibility of siRNA for easier comparison among cases. Panels (A, C) indicate mean ± SE of the fluorescence intensity of expressed target proteins in the cultures shown in (B, D), respectively, normalized to cell numbers in a field of view, and based on analysis of n = 3 regions of interest/culture group. *, **, ***, **** indicate statistical significance of differences compared to TC deemed for p < 0.05, p < 0.01, p < 0.001, and p < 0.0001, respectively. IF, immunofluorescence. Color images are available online.

Fibrillin-1 expression was significantly higher only in 25 kDa PEI-siRN-treated cultures (p = 0.021). Fibulin-4 was significantly higher only in 40 kDa PEI-siRN-treated cultures (p = 0.003) with no significant difference in expression of any protein between 25 and 40 kDa PEI-siRNA-treated groups.

TEM micrographs showed that unlike the lack of elastic fibers in TC cultures (Fig. 6 left), well-developed elastic fibers were numerous in both 25 kDa (Fig. 6 middle) and 40 kDa (Fig. 6 right) PEI-siRNA-treated cultures. Fiber diameters of ∼2 μm diameter were found in both cases, with instances of thinner fibers aggregating into larger, fibers.

FIG. 6.

TEM of elastic fibers following PEI-siRNA treatment. TEM images showing elastic fibers in treatment control, 25 kDa PEI-siRNA (5:1), and 40 kDa PEI-siRNA (2.5:1)-treated cultures. Boxes on upper images show area of focus for higher magnification image, below. White arrows used to indicate elastic fibers. All cell layers are shown at 21 days of culture. TEM, transmission electron microscopy.

Discussion

JNK2 is upregulated during cell stress response¹⁰ in proteolytic disorders, including AAAs,^9,21–24 and causes downstream increases in elastolytic MMPs.^25–27 Elastin breakdown in such disorders is naturally irreversible due to poor elastogenicity of most adult cell types and in the case of AAAs, diseased vascular SMCs.^28–30 The novelty of this work is the investigation, in an in vitro cell culture model of disease, of a siRNA-based gene silencing approach to stimulate new elastic matrix assembly and attenuate proteolysis. We have shown the MMP inhibitor, DOX, at low micromolar doses, to stimulate elastogenesis and attenuate MMPs by inhibiting JNK2. To more efficiently inhibit JNK, we investigated JNK2 gene silencing for this purpose. In identifying a safe JNK2 siRNA dose, we initially used commercial transfection agent for siRNA delivery due to its low-cost, pH independence, and stability.³¹

Since TR use is not possible in vivo, intracellular delivery of PEI-siRNA complexes was studied. The goal of this study was to specifically identify PEI properties (MW, linear or branched structure), and siRNA-PEI complexing ratios that are noncytotoxic, most conducive to efficient JNK2 siRNA uptake into aneurysmal SMCs, and that can most impressively attenuate JNK gene expression to significantly (1) increase elastin expression (primary outcome 1), (2) stimulate possible increases in other proteins important to elastic fiber assembly (Secondary effect) or at the least not attenuate expression of these proteins (as a secondary effect), and (3) attenuate proteolytic activity (primary outcome 2). The useful complexes identified in this study will, in the future, released within the AAA wall, from actively targeted, polymer nanoparticle -based drug delivery platforms that are being separately developed.

Our study showed a 12.5 nM siRNA dose to be ideal for use since it inhibited the JNK2 gene, and caused secondary increases in expression of genes for elastin (ELN) and fiber assembly proteins, including LOX and LOXL1 (elastin crosslinking enzymes), FBN1 (glycoprotein prescaffold), FBLN4, and FBLN5 (chaperone proteins for fiber assembly). siRNA treatment (12.5 nM) stimulated increases in TGF-β1 protein versus TC. We have previously shown TGF-β1 to increase tropoelastin production and elastic matrix deposition by aneurysmal SMCs to healthy levels.^32,33

The increases in desmosine (elastin crosslinker) content in JNK2 siRNA (12.5 nM)+TR-treated cultures versus the TC was consistent with increases in LOX, which catalyzes the crosslinking and increase in elastic matrix amounts. A mild antiproteolytic effect was also seen with slightly lower MMP2 and 9 expression versus TC, and increases in TIMP-2. Although anti-MMP effects were more prominent at higher siRNA doses, enhanced elastin deposition was only seen with the 12.5 nM dose. Consistent with these outcomes, TEM showed increased mature elastic fiber content in cultures with 12.5 nM siRNA versus TC. In the former, fibers exhibited sizes (1–2 microns in diameter) and appearance (coalescing amorphous elastin deposits with closely apposing glycoprotein microfibrils) that indicated maturity.^10,20 Based on these proelastogenic and anti-MMP effects (Supplementary Figs. S1–S4), a 12.5 nM JNK2-siRNA dose was deemed beneficial to elastin regeneration.³⁴

Both linear and branched PEIs have been found useful as transfection agents.³⁵ The transfection efficiency of PEI depends on its physical properties. In general, transfection efficiency tends to increase with PEI MW³⁶ due to increased density of cationic charges that enhances the ability of PEI to condense siRNA into a complex. As compact colloidal particles, these complexes can then be efficiently taken up by cells via an endocytotic process. A high enough charge density on the PEI is also necessary to prevent premature siRNA release, and to protect it from degradation within endosomes and lysosomes through tight condensation. PEI with higher charge density also provides buffering capacity in the endosomal-lysosomal compartments, termed as “proton sponge effect,” which is responsible for osmotic swelling and lysis of the endosomes to release the siRNA preventing their degradation by lysosomal enzymes. This has implications to improving gene knockdown activity of the siRNA. Conversely, high cationic charge density associated with longer PEI chains can cause an increase in cell cytotoxicity due by disrupting integrity of the anionic plasma membrane and through increased mitochondrial and lysosomal damage.

Branched PEI has also been reported to be less tolerated versus linear PEI,³⁷ and due to high charge density associated with its branches, it is able to ensure greater nuclei acid integrity and delivery to the cytoplasm without degradation; this also depends strongly on complexing ratios of PEI to siRNA.³⁸ Despite these advantages, the multiple branches of branched PEI structures can enhance PEI interaction with nuclear processes to cause DNA damage and cell death.³⁹ The gene knockdown activity of siRNA complexed with PEI, and the cytotoxicity and secondary effects of these complexes on target cells tend to be also majorly influenced by target cell type. This provides a strong argument to characterize PEI properties in the context of introducing JNK siRNA for gene knockdown specifically in our target cell type, EaRASMCs toward safely and effectively stimulating elastic matrix assembly and attenuating proteolytic activity, our primary outcomes.

In our study, we found only 25 kDa branched PEI at high N:P molar ratios (i.e., high PEI to siRNA molar ratios) to reduce EaRASMC viability. Gel electrophoresis showed poor association between siRNA and 25 kDa branched PEI (several N:P ratios) and 40 kDa linear PEI (ratios >5:1) as evidenced by migration of fluorescently tagged free siRNA down the gel. Free PEI can cause cytotoxicity via membrane destabilization and induce a stress response in the cell.^36,40 Thus, cytotoxicity observed in our study may have been due stress effects of excess, siRNA-unbound PEI and/or from PEI remaining in the cell following transfection. The above PEI-siRNA complexes were therefore eliminated from further investigation.

We identified two viable PEI-siRNA complexing conditions, 25 kDa PEI-siRNA (N:P molar ratio of 5:1) and 40 kDa PEI-siRNA (N:P molar ratio of 2.5:1). siRNA intracellular uptake was comparable in both cases and limited to the cytoplasm where its function is fully activated. While our primary outcome is reduced JNK2 gene expression, we also assessed nonspecific effects on other elastic fiber assembly proteins/genes. The 25 kDa 5:1 PEI-siRNA beneficially decreased JNK2 and MMP2, and increased FBN1 and FBLN4, genes for proteins critical to elastic fiber formation. The 40 kDa 2.5:1 PEI-siRNA also attenuated JNK2, and stimulated FBN1 and FBLN4, but less so than the 25 kDa PEI-siRNA. Experimental groups were run with separate treatment controls requiring all results to be normalized to the housekeeping gene, 18s, and also the TC. During this normalization, many key differences between groups previously deemed statistically significant were lost. This included significant decreases in JNK2 gene expression for 25 kDa 5:1- and 40 kDa 2.5:1 PEI-siRNA-treated groups versus TC, which had rationalized further studies using only these two complexes.

TGF-β, desmosine, and elastin content in PEI-siRNA-treated cultures showed mixed results regarding benefits to elastogenesis. TGF-β is known to stimulate elastin deposition and crosslinking,^41–43 and in the context of treating AAAs, to slow AAA progression.^44,45 We observed higher TGF-β1 expression in 25 kDa 5:1 PEI-siRNA-treated cultures versus other cases, and higher than 40 kDa PEI-treated cultures (p = 0.055). This corresponded with higher levels of LOX synthesis and higher content of desmosine crosslinks, in 25 kDa 5:1 PEI-siRNA-treated cultures versus TC, but no such differences in 40 kDa 2.5:1 PEI-siRNA-treated cultures.

For maturation of elastic matrix, elastin protein must be crosslinked onto the microfibrillar prescaffold. Desmosine content is a useful metric of alkali-insoluble (mature) elastin in cell cultures.^46,47 Since the 25 kDa PEI-siRNA cultures showed higher desmosine content and TGF-β1 expression, we expected them to contain more crosslinked elastic matrix versus the 40 kDa PEI-siRNA-treated cultures. However, the Fastin assay indicated elastic matrix content to be higher in the latter cultures, suggesting that the elastic matrix in those cultures represent the less crosslinked and alkali-soluble elastin fraction. Together, the results suggest that elastic matrix in more mature in the 25 kDa PEI-siRNA-treated cultures and hence this treatment more effective from this perspective.

Antiproteolytic effect is another consideration in selecting a useful PEI-siRNA complex. The 25 kDa 5:1 and 40 kDa 2.5:1 PEI-siRNAs induced significant increases in MMP2 protein synthesis versus TC, while MMP9 expression was unaffected. Despite higher expression of TIMP2, natural inhibitor of MMP2, in the 25 kDa 5:1 PEI-siRNA-treated cultures versus TC the MMP2/TIMP2 protein ratio, a gauge of net proteolytic activity,^48–50 was not different; the ratio was adversely higher in the 40 kDa 2.5:1 PEI-siRNA-treated cultures. This outcome, together with the significantly higher MMP9 activity in the 40 kDa PEI-siRNA-treated cultures, motivates us to further pursue use of the 25 kDa 5:1 PEI-siRNA complex as a JNK2 gene silencing tool.

Quantitative results from IF images mostly agreed with our biochemical assessments, with the exception of elastic matrix amounts/cell, which were not different between the treated groups. This may be owed to the fact that cell numbers are not necessarily homogenous across the cell layers, and quantification of fluorescence is based on assessment of a limited number of discrete field of views (FOVs), and not on the entirety of the cell layer. While TEM indicated presence of mature elastic fibers in 25 and 40 kDa PEI-siRNA-treated cultures, fibers in the 25 kDa PEI-siRNA-treated cultures were visibly thicker, suggesting enhanced crosslinking of tropoelastin aggregates.

Conclusions

The inherent deficiency and impairment of natural processes for elastic fiber and matrix assembly is a critical limitation to restoration of elastin homeostasis in proteolytic disorders of soft connective tissues (e.g., AAAs), a requirement for reinstatement of healthy function and tissue mechanics. Specifically, in the context of reversing AAA pathophysiology, there is a need to achieve the dual primary functional outcomes of (1) stimulating elastin regeneration in light of the poor innate elastogenicity of diseased vascular cells, and (2) attenuating chronic elastolysis in the AAA wall.^51,52 In this work, we have demonstrated in an in vitro proteolytic injury model of cultured EaRASMCs, the feasibility of stimulating these matrix regenerative outcomes by inducing JNK2 knockdown using PEI-siRNA complexes.

For this cell type and, more specifically, this specific application of induced elastic matrix regeneration, we have identified the PEI properties and PEI/siRNA complexing conditions essential for safe and significant stimulation of elastic matrix assembly. The findings are promising toward further development of a gene silencing therapy to counter JNK upregulation, a key pathological mechanism that leads to naturally irreversible breakdown of elastic matrix structure in the aortic wall. While coupling PEI-siRNA enables protection from nucleolytic degradation and enhances cell uptake,⁵³ for the purpose of predictable and localized delivery of these complexes to the AAA wall in vivo, and to further protect them from degradation in circulation, our future work will integrate the useful and safe PEI-siRNA complexes identified in this work with actively targeted polymer nanoparticles we are developing separately.

Disclosure Statement

No competing financial interests exist.

Funding Information

The authors acknowledge research funding for this project from the National Institutes of Health (HL139662 to A.R., and HL53325 and HL105314 to R.M.), and American Heart Association (TPA 1934890029 to A.R. and 20PRE35120023 to S.C.). This work utilized the FEI Tecnai G2 Spirit transmission electron microscope that was purchased with funding from National Institutes of Health SIG grant 1S10RR031536-01.

Supplementary Material

Supplementary Figure S1

Supplementary Figure S2

Supplementary Figure S3

Supplementary Figure S4