To establish whether CSPG sulfation occurs after MI caused by I/R, we used antibodies specific to 4S and 6S CS-GAGs (Yi et al., 2012) to compare CSPG sulfation in mouse unoperated left ventricle to day 14 post-I/R scar tissue. Day 14 represents a time point when scar tissue is relatively stable and sympathetic nerves are excluded from entering the scar. I/R led to increased 4,6-sulfation of CSPGs in the cardiac scar (Figure 1B–D). In contrast, CSPG levels are low in remote myocardium (Gardner and Habecker, 2013), and 4S is not altered after I/R (Figure 1—figure supplement 1).

To test whether 4S of CS-GAGs prevented neurite outgrowth across CSPGs, we enzymatically removed 4S from purified CSPGs using the 4-sulfatase, ARSB. An effective ARSB dose was identified in pilot studies to remove 4S while leaving 6S intact (Figure 2—figure supplement 1). Dissociated sympathetic neurons from neonatal rats were then grown on CSPGs treated with vehicle or ARSB. Sympathetic neurite extension across untreated CSPGs was suppressed 40 hr post-plating compared to laminin alone, while ARSB treatment restored neurite outgrowth in a dose-dependent manner (Figure 2A and B). These results indicate that 4S of CS-GAGs inhibits sympathetic neurite extension.

We treated mouse myocardial I/R explants with ARSB to test whether removing 4S from CS-GAGs in cardiac scar tissue could restore sympathetic axon outgrowth in explant co-cultures (Figure 3A). Superior cervical ganglion (SCG) explants from neonatal mice were co-cultured with left ventricle tissue. Although SCG contain a relatively small number of cardiac sympathetic neurons, they provide a well-characterized model for axon outgrowth studies (Gardner and Habecker, 2013). Ganglion explants co-cultured with unoperated left ventricle displayed uniform axon outgrowth in all directions while explants cultured alongside cardiac scar tissue (10–14 days post-I/R) exhibited shorter axons growing in the direction of the scar tissue (Figure 3B–D). Interestingly, ARSB treatment (0.6 µg/mL) of scar tissue fully restored axon outgrowth towards the scar (Figure 3D), while growth away from the scar tissue was normal in all conditions (Figure 3E). ARSB treatment of control explants had no effect on axon outgrowth, suggesting its effects are through altering the cardiac scar rather than the ganglion. Together, these data suggest that reducing 4S of CS-GAGs in the cardiac scar is sufficient to enable sympathetic axon regeneration.

To understand the mechanisms by which CS-GAG sulfation increases after I/R in mice, we examined protein levels of three critical CSPG sulfation enzymes, comparing post-MI expression to unoperated left ventricle. We examined chondroitin-4-sulfotransferase CHST11, 4S-dependent chondroitin-6-sulfotransferase CHST15 (4,6-tandem sulfation enzyme), and 4-sulfatase ARSB (Figure 1A). CHST11 levels decreased significantly within 24 hr, persisting until day 7 (Figure 4A and D), while CHST15 was increased significantly at days 7 and 14 post-MI (Figure 4B and D). ARSB levels decreased significantly by day 3 and remained low through day 14 (Figure 4C and D).

In light of protein changes in CSPG sulfation enzymes after I/R, we examined CS-GAG sulfation in the mouse infarct at these same time points. 4S and 6S CS-GAGs are significantly increased in the infarct by day 7 post-MI extending through day 14 (Figure 5A, B and E). We asked if the CSPG core protein NG2 (also called CSPG4) was also altered after I/R, and found that NG2 was increased in the infarct by 7 days after MI (Figure 5C and E). Thus, core proteins and sulfation of CS-GAG side chains are both in greater abundance. We chose to look at NG2 because it was prominently expressed in a recent glycoproteomics analysis of cardiac scar tissue whereas another commonly studied core proteoglycan, Aggrecan, was not detected (Tian et al., 2014). In an effort to better understand how this process may be coordinated, we examined the expression of a key wound healing regulator, Galectin-3, which has also been connected to CSPG sulfation and ARSB expression (Bhattacharyya et al., 2020; Bhattacharyya et al., 2017; Bhattacharyya et al., 2014; Bhattacharyya et al., 2015). Galectin-3 was increased significantly by 24 hr after I/R, which may contribute to altered expression of CSPG sulfation enzymes at later time points (Figure 5D and E). These data suggest that increased CHST15 and decreased ARSB expression contribute to increased 4,6-sulfation of CS-GAGs observed by day 7 post-MI.

To determine if CSPG sulfation suppresses sympathetic nerve regeneration after I/R in mice, we decreased 4,6-tandem sulfated CS-GAGs in vivo by reducing the expression of the Chst15 gene using silencing RNA (siRNA). After identifying siRNA that decreased Chst15 mRNA in myoblast-like C2C12 cells (Figure 6—figure supplement 1A,B), we tested the efficacy of our Chst15 siRNA in vivo. Intravenous delivery of 100 µg siChst15 reduced CHST15 protein compared to non-targeting control siRNA (Figure 6—figure supplement 1C). Mice were then treated with 100 µg siChst15 on days 3, 5, and 7 post-MI, leading to a significant reduction in 4S,6S CS-GAGs. Specifically, siChst15 reduced the amount of 6S present 10 days post-MI compared to non-targeting siRNA control (Figure 6A). By 10 days after MI, CHST15 protein levels had returned to normal suggesting that transient knockdown of CHST15 is sufficient to alter 4,6-sulfation of CS-GAGs (Figure 6—figure supplement 1D,E). Interestingly, siChst15 knockdown led to increased NG2 core protein expression compared to non-targeting siRNA control (Figure 6C). The sympathetic neuron marker Tyrosine Hydroxylase (TH) was increased significantly in cardiac scar tissue from siChst15-treated animals compared to non-targeting siRNA control animals (Figure 6D), suggesting successful sympathetic reinnervation of the cardiac scar.

To ensure that reinnervation of the cardiac scar occurred after Chst15 knockdown, we examined the infarct (labeled by fibrinogen) for TH-positive sympathetic nerve fibers by immunohistochemistry (IHC). IHC analysis in animals treated with a non-targeting siRNA control showed clear denervation of the infarct (Figure 7A) compared to a peri-infarct region adjacent to the scar (Figure 7B). Animals treated with an siRNA targeting Chst15 had restored TH-positive fibers in the infarct (Figure 7C) at the same innervation density as the peri-infarct region (Figure 7D). TH innervation density was significantly higher in the infarct of Chst15-targeted animals compared to non-targeting siRNA-treated animals (Figure 7E). Cardiac scar size was examined between the two groups and no significant difference existed (Figure 7G and H). To examine whether restoration of nerves reduced arrhythmia susceptibility, we examined arrhythmias by ECG after administration of the β-agonist isoproterenol and caffeine (Gardner et al., 2015; Wang et al., 2014). Restoration of nerves with Chst15 siRNA treatment reduced arrhythmias when compared to non-targeting siRNA control-treated animals (Figure 7F). These results indicate that reducing 4,6-tandem sulfated CS-GAGs in the cardiac scar after MI promotes sympathetic nerve regeneration back into the cardiac scar and reduces arrhythmia susceptibility after MI.

C57BL/6J mice obtained from Jackson Laboratories West (Sacramento, CA) were used for all experiments. All mice were kept on a 12 hr:12 hr light-dark cycle with ad libitum access to food and water. Age and gender-matched male and female mice 12–18 weeks of age were used for myocardial-ischemia reperfusion surgeries, as described below. Pregnant Sprague-Dawley rats were purchased from Charles River Laboratories (Wilmington, MA) for experiments with dissociated sympathetic neurons. SCG from male and female neonatal rats (P0-P2) were used to generate primary sympathetic neuron cultures. All rats were kept on a 12 hr:12 hr light-dark cycle with ad libitum access to food and water. All procedures were approved by the OHSU Institutional Animal Care and Use Committee and comply with the Guide for the Care and Use of Laboratory Animals published by the National Academies Press (8th edition).

Anesthesia was induced with 4% isoflurane and maintained with 2% isoflurane. Mice were restrained supine, intubated, and mechanically ventilated. Core body temperature was monitored by a rectal probe and maintained at 37°C throughout the surgery. The left anterior descending coronary artery (LAD) was reversibly ligated for 40 min and then reperfused by release of the ligature. LAD occlusion was verified by a persistent S-T wave elevation, region-specific cyanosis, and wall motion abnormalities. Reperfusion was confirmed by return of S-T wave to baseline level and re-coloration of ventricle region distal to occlusion (Gardner and Habecker, 2013; Parrish et al., 2010).

Cardiac scar tissue from the left ventricle was dissected at 24 hr, 3 days, 7 days, and 14 days following MI, and control left ventricle tissue from unoperated animals. Heart tissue was pulverized in a glass douncer in NP40 lysis buffer (50 mM Tris [pH 8.0], 150 mM NaCl, 2 mM EDTA, 10 mM NaF, 10% glycerol, and 1% NP-40) containing complete protease inhibitor cocktail (Roche), phosphatase inhibitor cocktails 2 and 3 (Sigma). Lysates sat on ice for 30 min with intermittent vortexing. Lysates were centrifuged (13 k rpm, 10 min, 4°C) and resolved on 4–12% Bis-Tris gradient gel (3–8% Tris-Acetate gel for Figure 5 sulfation studies) by SDS/PAGE, transferred to nitrocellulose membrane (GE Life Sciences), blocked in 5% nonfat milk, probed with either CHST11 (1:500; Invitrogen: PA5-68129), CHST15 (1:1000; Proteintech: 14298-1-AP), ARSB (1:500; Proteintech: 13227-1AP), NG2/CSPG4 (1:1000; Millipore: AB5320), TH (1:1000; Millipore: AB1542), or Galectin-3 (1:500; Abcam: AB2785) then probed with goat anti-rabbit (or mouse) HRP-conjugated secondary antibody (1:10,000; Thermo), and detected by chemiluminescence (Thermo Scientific). To detect sulfation of chondroitin-sulfate proteoglycans, lysates were treated according to the detailed deglycosylation protocol from Mariano Viapiano PhD lab website (Massey et al., 2008). Briefly, 50–100 µg of protein lysate was treated for 6–8 hr with chABC (100 µU/mL; R&D systems) before being subjected to traditional western blotting, probed with anti-chondroitin-4-sulfate (1:1000; Millipore: MAB2030) or anti-chondroitin-6-sulfate (1:1000; Millipore: MAB2035). Protein expression was quantified with ImageJ densitometry and normalized to total protein as measured by Ponceau staining.

Cultures of dissociated sympathetic neurons were prepared from SCG of newborn rats as described (Dziennis and Habecker, 2003). Cells were pre-plated for 1 hr to remove non-neuronal cells, and then 5000 neurons/well were plated onto a 96-well plate (TPP) coated with (poly-L-lysine, 0.01%, Sigma-Aldrich) and either laminin (1 μg/mL, RND systems), laminin and CSPGs (2 μg/mL; Millipore), or laminin and CSPGs pre-treated with ARSB (0.3 or 0.6 µg/mL; R&D systems) for 6–8 hr. Neurons were cultured in serum free C2 medium (Pellegrino et al., 2011) supplemented with 10 ng/mL NGF (Alomone Labs), 100 U/mL penicillin G, and 100 μg/mL streptomycin sulfate (Invitrogen). Live cell imaging was carried out using an Incucyte Zoom microscope (Essen BioScience), with 20× phase images acquired every 2 hr over a 40 hr period. Neurite length was measured using Cell Player Neurotrack software (Essen BioScience) and was used to calculate the neurite growth rate.

Explants were generated as previously described (Gardner and Habecker, 2013). Briefly, SCG were dissected from neonatal mice and placed into pre-marked tissue culture wells so that ganglia were approximately 1 mm from left ventricle tissue. LV tissue was taken from unoperated control animals or from hearts collected 10–14 days after I/R. Cardiac tissue from a single animal was split in half and cultured with the left or right ganglion from a single animal, this enabled tissue from the same mice to be treated with either vehicle (5% DMSO) or ARSB to remove 4S-CS GAGs. The ganglia and the cardiac tissue were co-cultured inside a Matrigel bubble surrounded by C2 media supplemented with 2 ng/mL NGF to ensure neuron survival and stimulate a basal level of axon outgrowth in all directions. ARSB was added directly to the media (0.6 µg/mL) at the time of plating and again 24 hr later. At 48 hr the explants were imaged by phase microscopy with a Keyence BZ-X microscope. Neurite length from the edge of the ganglia to the most distal tip of visible neurites was measured using ImageJ.

A pool of four siRNAs targeting the Chst15 gene were purchased from Horizon Discovery (formerly GE Dharmacon) and were tested for their efficacy in reducing gene expression in C2C12 myoblasts. C2C12 myoblasts were plated on 12-well plates coated with collagen and were transfected with the various targeting and control siRNAs using the Dharmafect transfection reagent (3 µL of Dharmafect reagent per well with 120 nM siRNA). The next day wells were split using Versene (Gibco), one-third of the cells were collected for a 24 hr time point and the other two-thirds were split into two wells for 48 and 72 hr knockdown time points. Cells were processed using an RNA Mini-kit (Qiagen) to purify RNA. One µg of RNA was loaded into the cDNA reaction with the iScript cDNA synthesis kit. Gene expression was examined with multiplexed Taqman probes targeting Chst15 and Gapdh using 2 µL of cDNA template and Taqman reagents and measured with an ABI7500 Thermocycler. The delta delta ct method was used to calculate knockdown efficiency. Once an effective siRNA against Chst15 was identified, knockdown efficiency was measured at the protein level with C2C12 myoblasts 48 hr post-knockdown. CHST15 expression was assessed via western blot with a CHST15 antibody (1:1000; Proteintech: 14298-1-AP) normalized to GAPDH (1:1000; Thermo Scientific: MA1-16757); siRNA Chst15-2 was selected for larger scale production and used for in vivo studies.

After MI surgery mice were treated with 100 µg of siRNA targeting Chst15 or a non-targeting control. siRNA was delivered systemically via tail-vein injection on days 3, 5, and 7 following MI. Tissue was collected at day 10 post-MI for western blot analysis of CSPG sulfation and the sympathetic neuron marker TH. siRNA used for this experiment was custom synthesized by Horizon Discovery; siAccell in vivo formulation.

Norepinephrine (NE) levels in heart tissue were measured by high-performance liquid chromatography (HPLC) with electrochemical detectionLi, Knowlton et al. 2004. Frozen, pulverized tissue was weighed and homogenized in 300 µL of 0.1 M perchloric acid (PCA) containing 0.5 µM dihydroxybenzylamine (internal standard). The homogenate was centrifuged at 14,000 rpm for 4 min and NE in 100 µL of the supernatant was adsorbed onto 15 mg alumina, followed by 15 min of tumbling. The alumina was washed twice with ddH₂O and the catechols desorbed with 150 µL of 0.1 M PCA. The catechols were separated by reversed-phase HPLC on a C18 column (Agilent Microsorb, 150 × 4.6 mm, 5 μm) and measured by an electrochemical detector (Coulochem III; ESA, Bedford, MA) with the electrode potential set at +180 mV as described previously (Parrish et al., 2010). The mobile phase used consisted of 75 mM sodium phosphate (pH 3.0), 1.7 mM sodium octane sulfonate, 3.0% acetonitrile. NE standards (0.5 µM) were processed in parallel with the tissue samples.

TH (sympathetic nerve fibers) and fibrinogen (Fib; infarct/scar) staining was carried out as described previously (Gardner et al., 2015). Tissue was collected 10 days after surgery, fixed in 4% paraformaldehyde, frozen and 12 µm sections generated. To reduce autofluorescence sections were rinse 3 × 10 min in 10 mg/mL sodium borohydride and rinsed for 3 × 10 min in PBS. Slides were placed in 2% BSA, 0.3% Triton X-100 in PBS for 1 hr, and then incubated with rabbit anti-TH (1:1000; Millipore: AB1542) and sheep anti-fibrinogen (1:300; BioRad: 4400-8004) overnight. The following day the slides were incubated with Alexa-Fluor IgG-specific antibodies (Molecular Probes, 1:1000) for 1.5 hr and rinsed 3 × 10 min in PBS. Background autofluorescence was reduced further with a 30 min incubation in 10 mM CuSO₄ (diluted in 50 mM ammonium acetate). Following this, slides were rinsed 3 × 10 min in PBS before mounting in 1:1 glycerol:PBS and visualized by fluorescence microscopy. Threshold image analysis of TH staining has been described previously (Gardner et al., 2015) but briefly, the threshold function in ImageJ was used to generate black and white images discriminating TH + nerves for six sections spanning 200 µm of the infarct/scar or peri-infarct region from each heart. Percent area TH + fiber density (20× field of view) was quantified within the infarct and the area immediately adjacent to the infarct (peri-infarct). Images were acquired with a Keyence BZ-X 800 microscope.

Infarct size was determined 10 days after myocardial I/R injury and tissue was prepared for IHC as previously mentioned, but omitting the steps to reduce autofluorescence. Autofluorescence was used to image the infarct size. Scar tissue is notably lacking autofluorescence enabling easy identification of the infarct. Images were acquired with a Keyence BZ-X 800 microscope at 2× magnification and analyzed using Image J freehand selection tool. Left ventricle (LV) and infarct was outlined, measured, and the percent area of cardiac scar was determined by (infarct area/LV area) × 100. The scar was imaged in six sections across 200 µm of the infarct as previously described (Gardner et al., 2015).

Anesthesia was induced with 4% isoflurane and maintained with 2% isoflurane in day 10 post-MI animals treated with siRNA. ECG leads were connected to monitor arrhythmias and animals were maintained at 37°C throughout the analysis. All parameters were monitored with Powerlab LabChart software (AD Instruments). A 30 min baseline was used to assess spontaneous arrhythmia susceptibility prior to administration of β-agonist isoproterenol (50 µg) and caffeine (3 mg) as described previously (Wang et al., 2014). Arrhythmias were measured for 30 min following drug administration and scored according to the modified Lambeth conventions (Curtis et al., 2013) on a scale of 0–4. Individual animals received a single score based on the most severe arrhythmia observed. 0 indicates no arrhythmia. 1 indicates 1–2 premature ventricular contractions (PVCs) followed by normal sinus rhythm of at least 2 beats. 2 indicates bigeminy (1 PVC followed by one normal sinus beat, repeating for 4 or more continuous cycles) or salvo (3–5 PVCs in a row). 3 indicates non-sustained ventricular tachycardia (nsVT) defined as 6 or more PVCs in a row lasting less than 30 s. 4 indicates sustained VT (>30 s) or Torsades de Pointes.

Student’s t-test was used for comparisons of just two samples. Data with more than two groups were analyzed by one-way ANOVA using the Tukey’s post hoc test to compare all conditions or the Dunnet’s post-test when comparing to a single control group. Data with multiple variables was analyzed by two-way ANOVA. All statistical analyses were carried out using Prism 9.