α-Gal nanoparticles were prepared as previously described (24, 25, 27) from rabbit red blood cell (RBC) membranes because these RBC present the highest concentration of α-gal epitopes in comparison to other mammalian RBC and a large proportion of these α-gal epitopes is presented on glycolipids (34). Rabbit RBC (1 L) are lysed in water and washed for hemoglobin removal. Washed RBC membranes are mixed with 800 ml chloroform and 800 ml methanol for 2 h, then 800 ml methanol are added and stirred overnight, resulting in the extraction of the phospholipids, cholesterol and glycolipids into the chloroform:methanol solution (24, 25, 40). Residual RBC membranes and proteins are precipitated by the chloroform:methanol solution. The precipitate is removed by filtration through Whatman paper. The extract is dried in a rotary evaporator, weighed, and sonicated in saline, in a sonication bath to generate a suspension of 100 mg/ml liposomes comprised of phospholipids, cholesterol and glycolipids. Residual debris is pelleted at low speed centrifugation (800 rpm) and removed. Liposomes are further sonicated on ice with a sonication probe to break the liposomes into submicroscopic liposomes (50–300 nm), referred to as “α-gal nanoparticles,” which are sterilized by filtration through a 0.45 μm filter (Millipore). These nanoparticles present ~10¹⁵ α-gal epitopes per mg nanoparticles (25). Nanoparticles lacking α-gal epitopes were produced by the same method from RBC of knockout pigs for the α1,3galactosyltransferas gene (GT-KO pigs) that cannot synthesize α-gal epitopes (42, 43).

The experimental protocol was approved by the Institutional Animal Care and Use Committee at Rush University Medical Center (Chicago, IL) and performed in accordance with AAALAC guidelines. In order to simulate a human-like immune environment, a previously established, α1,3galactosyltransferase (α1,3GT) knockout mouse (GT-KO) was used (38). Like humans, these knockout mice do not synthesize the α-gal epitope and therefore can produce the anti-Gal antibody with post-natal exposure to this epitope, such as immunization with PKM homogenate (24, 25, 39, 40). PKM were empirically found to be more effective than rabbit RBC in immunization of GT-KO mice for the production of the anti-Gal antibody (unpublished observations). GT-KO mice used (males and females, age 12–16 weeks) were of C57BL/6 × BALB/c genetic background (38). Mice were induced to produce the anti-Gal antibody at titers comparable to those in humans by 5 weekly immunizations with 50 mg PKM homogenate (200 mg/ml). These homogenates were used for immunization that elicits anti-Gal production because pig kidney membranes present a large number of α-gal epitopes (41). In the absence of such immunization, GT-KO mice do not produce anti-Gal because of lack of immunizing gastrointestinal bacteria (39).

MI induction with subsequent coronary reperfusion was performed in mice based on modification of previously published protocols (44, 45). Mice were anesthetized with ketamine (100 mg/kg)/xylazine (5 mg/kg) injected intraperitoneal, intubated and anesthesia maintained by inhaled 1.0% isoflurane. A left thoracotomy was performed via the fourth intercostal space and lungs were retracted to expose the heart. The mid-LAD coronary artery was ligated with a 7-0 silk suture. Ligation was confirmed as 100% occlusive by the appearance of pallor of the anterior wall of the left ventricle (LV). After 30 min occlusion, the ligature was removed, permitting reperfusion of the LAD territory. This was confirmed by noting a change in the color of the anterior wall of the LV from pallor to deep red, as observed prior to coronary occlusion. One minute after reperfusion, a 32-gauge needle was used to inject 10 μl of a 10 mg/ml α-gal nanoparticles suspension in saline [dose optimized in refs. (24, 25)] into the LV anterior wall, ~1 mm distal to the LAD ligation site. Treated mice received two injections, 1–2 mm apart (100 μg nanoparticles per site). Control mice underwent the same procedure but received two injections of 10 μl normal saline. Subsequently, the chest wall, subcutaneous tissue, and skin were sutured. Among treated animals, 20 underwent post-mortem studies on day 28 post-MI, and 2 animals per day at 4, 7, and 14 days post-MI. Among controls, 10 mice underwent post-mortem studies on day 28 post-MI, and 2 animals per day, 4, 7, and 14 days post-MI. Mice were euthanized by CO₂ inhalation followed by cervical translocation.

Formalin fixed hearts were cut from mid-level (at the level of the LAD occlusion) into 5 equal double sections (5 μm-thick) 300 μm apart. Sections at each level were stained with either hematoxylin-eosin (H&E) (to assess macrophage infiltration), or Trichrome (collagen/fibrosis stain blue; viable cardiomyocytes stain red; debrided cardiomyocyte areas stain gray, RBCs stain brown). The LV section demonstrating the greatest extent of macrophage infiltration or the greatest infarct size (out of the sections prepared at 5 levels), was scanned for planimetry measurement (46) using AperioImageScope planimetry program. The scanned histology slides were measured for infarct size in a blinded manner according to a code number which subsequently was compared with a key list which indicated the treatment of each heart. The infarct size for each scanned section was expressed as percentage, relative to the total area of uninjured LV (46). Sections were also stained with antibody to F4/80 antigen for identification of macrophages (25) and with antibodies to proliferating cell nuclear antigen (PCNA) (Abcam, Cambridge, MA) for detection of proliferating cardiomyocytes. The staining was performed following antigen retrieval with Tris-EDTA buffer (10 mM Tris Base, 1 mM EDTA solution, 0.05% Tween 20, pH 9.0, kept for 10 min in boiling solution) and completed with a fluorescein coupled secondary antibody. After washes the slides were counterstained with DAPI.

High resolution transthoracic echocardiography (Vevo 770, VisualSonics) was performed in mice (n = 4 treated, n = 4 controls) that were sedated with 1% isoflurane, before LAD ligation, and post-MI (at 7 and 28 days). Parasternal short-axis views at the mid-ventricular level were obtained for M-mode analysis of the LV internal diameter at end diastole (LVDD) and end systole (LVDS). Fractional shortening (FS) was calculated as: %FS = [(LVDD – LVDS)/LVDD] × 100.

Mice were injected intraperitoneally with 1.0 ml of a solution containing 1 mg BrdU (Sigma-Aldrich, St. Louis, MO). The injection was performed on Day 9 post-MI or on Day 11 post-MI. The mice were euthanized after 24 and 48 h, heart and intestine were harvested fixed in formalin and embedded in paraffin. The intestine was used as positive control for each of the hearts studied. Paraffin sections were deparaffinized and subjected to BrdU nuclear staining by the use of “BrdU in situ Detection Kit” (BD Pharmingen cat. No. 550803), according to the manufacturer instruction. The kit includes biotinylated anti-BrdU antibody and avidin coupled peroxidase. The color reaction is achieved by diaminobenzidine (DAB) precipitation. Counter staining was performed with hematoxylin.

The data were presented as mean ± SE. Comparisons between 2 groups were made using Student t-test. Multiple group comparisons were made using One-way ANOVA, followed by Tukey's multiple comparison test.

Some of the outcomes of anti-Gal interaction with α-gal nanoparticles, hypothesized in Stages 1 and 2 in Figure 1 are illustrated in Figure 2. Anti-Gal within the serum of GT-KO mice producing this antibody readily binds to α-gal epitopes on α-gal nanoparticles, as shown by flow cytometry in Figure 2A. Nanoparticles lacking α-gal epitopes (produced from RBC of GT-KO pigs that lack these epitopes) do not bind anti-Gal produced by GT-KO mice. The binding of anti-Gal to α-gal nanoparticles was previously shown to result in activation of the complement system and production of complement cleavage chemotactic peptides that effectively induce macrophage recruitment to the nanoparticles injection sites in the uninjured skin of GT-KO mice (25). A similar recruitment of macrophages could be demonstrated by two injections (each of 10 μl of the α-gal nanoparticles) into uninjured LV myocardium of healthy hearts, not subjected to LAD occlusion. This is shown in a representative heart evaluated 4 days post-injection (Figure 2B). High power magnification inspection of the section in Figure 2B and of the other 3 uninjured normal hearts that received the same injection technique revealed macrophages and no polymorphonuclear cells or fibroblasts among the infiltrating cells. Similar infiltration of only macrophages (indicated as F4/80 stained cells) was previously observed in mouse dermis injected with α-gal nanoparticles (25, 27). Injection of saline into uninjured hearts resulted in no infiltration of macrophages or other cells at any time point (not shown). The similarity in recruitment of macrophages in uninjured skins (25) and hearts injected with α-gal nanoparticles (Figure 2B), strongly suggests that complement cleavage chemotactic peptides produced as a result of anti-Gal/α-gal nanoparticles interaction, shown to mediate recruitment of macrophages in the skin (25) have a similar chemotactic effect in post-MI reperfused hearts injected with these nanoparticles, as illustrated in Stage 1 of Figure 1.

The predicted Fc/Fc receptor interaction between anti-Gal coated α-gal nanoparticles and recruited macrophages (Stage 2 in Figure 1) is shown in Figure 2C with macrophages incubated for 2 h at 24°C with 10 mg/ml of such nanoparticles. The representative macrophage is covered with the bound α-gal nanoparticles that are coated with anti-Gal. The binding of anti-Gal coated α-gal nanoparticles also causes the contraction of the macrophage into a more spherical shape because of the multiple Fc/Fc receptor interactions. When the nanoparticles concentration is 100-fold less (i.e., 0.1 mg/ml) fewer nanoparticles bind to the macrophage, thus the macrophage maintains a normal flattened morphology (Figure 2D). In the absence of anti-Gal, no nanoparticles were bound to the macrophages (not shown).

The hypothesized Stage 3 in Figure 1 suggests that following the Fc/Fc receptor interactions between the recruited macrophages and anti-Gal coated α-gal nanoparticles in injured myocardium, the macrophages polarize into pro-reparative macrophages that secrete cytokines which, similar to their effects in wound healing, enhance repair and restore the normal structure and function of the post-MI myocardium. These hypothesized effects were studied in GT-KO mouse hearts assessed 28 days after the 30 min of mid-LAD occlusion and reperfusion in mice treated with α-gal nanoparticles, and saline-treated controls. The normal mouse heart (not subjected to LAD occlusion) (Figure 3A) demonstrates the expected uniform Trichrome red staining of uninjured cardiomyocytes and no fibrosis. Dashed circles mark epicardial and endocardial surfaces of the LV. The area between the circles represents 100% uninjured LV. The right ventricle (when visible) is on the left, and the LV on the right of each figure. The size of myocardial infarct (Trichrome staining collagen/fibrosis blue) was calculated as percentage, relative to the area of uninjured LV, based on planimetry measurements (46) of the injured and non-injured LV (Figure 3B). The histological sections of hearts providing the data for Figure 3B are presented in Figures 3C,D and in Supplementary Figure 1.

In saline injected controls (Figure 3C; Supplementary Figure 1A), mid-LAD occlusion for 30 min followed by reperfusion for 28 days resulted in large transmural infarcts with extensive fibrosis, scar formation and wall thinning. The infarct size, extent of fibrosis and patterns of scar formation are highly variable among the 10 control animals, in which the % mean ± SE infarct size of the LV was 19 ± 2.3%, with a range of 5–35% - (Figure 3B). This variability may result from small differences in the site of LAD occlusion and amount of myocardium supplied distal to the occlusion. Similar distribution of infarct size following 30 min occlusion/reperfusion in mice also was observed by other investigators (47–49). Twenty mice treated with α-gal nanoparticles demonstrated dramatically smaller infarcts, with greatly less fibrosis and thinning of the LV wall (Figure 3D; Supplementary Figure 1B). The mean myocardial infarct size in treated animals was 2.2 ± 1.2%, with a range of 0.1–5.0% (Figure 3B). The tissue stained red surrounding the residual blue fibrotic tissue in α-gal nanoparticles treated hearts (Figure 3E) is comprised of normal appearing cardiomyocytes that display (under high magnification of H&E stain) characteristic sarcomeric striation (Figure 3F), with minimal fibrosis.

Two additional types of controls demonstrate that the reparative effects of α-gal nanoparticles are dependent on the specific interaction between the anti-Gal antibody and α-gal epitopes on the nanoparticles, as hypothesized in Figure 1. Injection of GT-KO pig nanoparticles (i.e., nanoparticles lacking α-gal epitopes) into the reperfused post-MI myocardium resulted in fibrosis and scar formation that is similar to mice treated with saline (Figure 3B). The individual hearts receiving this treatment are shown in Supplementary Figure 2A. Furthermore, injection of α-gal nanoparticles into the post-MI heart of GT-KO mice lacking anti-Gal (i.e., not immunized with pig kidney membranes homogenate) resulted in fibrosis and scar formation, as well (Figure 3B; Supplementary Figure 2B), rather than in extensive myocardial restoration of structure, observed in anti-Gal producing mice. Thus, both anti-Gal production and α-gal epitopes on the nanoparticles are required for the induction of the reparative response in injured myocardium of adult mice.

Echocardiography was used to assess LV function in four saline treated controls (Figure 4A) and four α-gal nanoparticles treated mice (Figure 4B). Mice in both groups had normal fractional shortening (FS) before LAD occlusion (Figures 4C,D). By Day 7 post-MI, both groups demonstrated a marked reduction in FS, associated with LV chamber dilation, emphasizing the severity of the ischemic injury (Figures 4C–E). These Day 7 echocardiography results indicate that the post-ischemia injuries in both saline and α-gal nanoparticles treated hearts were severe and correlate with the histopathology results demonstrated below in Figure 5.

However, by Day 28, α-gal nanoparticles treated mice demonstrated a dramatic recovery of FS and return of LV chamber dimension to normal size, whereas control mice continued to display impaired FS and chamber dilation (Figures 4C–E). It is of note that the mice included in Figure 4 also provided histopathology results at Day 28. These results are included among the 10 saline treated mice which demonstrated fibrosis and scar formation and among 20 α-gal nanoparticles treated mice demonstrating repair and restoration of structure (Figure 3; Supplementary Figure 1) and are consistent with the functional results in Figure 4.

The near-complete restoration of structure and function in the post-MI myocardium treated with α-gal nanoparticles, compared to the extensive fibrosis and scar formation in saline injected control hearts, raised the question whether there are observable differences in the time-course of the inflammatory response in the two groups. This was assessed by histopathology at 4, 7, and 14 days post-MI (Figure 5). Myocardial histopathology in one of two mice studied at each time point is shown in pairs of which the upper is stained with H&E and lower with Trichrome. In these sections, H&E staining identifies the infiltrating macrophages which are much smaller than the cardiomyocytes and lack the elongated shape and sarcomere striation, whereas Trichrome staining identifies areas debrided of injured cardiomyocytes (stained gray) and areas containing healthy cardiomyocytes (stained red). Planimetry measurements of the debrided areas in two mice at each time point are presented in Figure 5C and the number of observed macrophages in Figure 5D. The identity of the infiltrating cells as macrophages is demonstrated in Figures 5E,F by the staining with the macrophage specific F4/80 antibody.

At Day 4, the myocardium of control animals displayed extensive cardiomyocytes elimination (gray staining with Trichrome), coinciding with extensive infiltration of macrophages, similar to previous reports (4–8, 50–52). Hearts treated with α-gal nanoparticles displayed distinctly less macrophage infiltration at Day 4 (Figures 5B–D), suggesting an attenuated early inflammatory response. In accord with previous observations (52), the number of macrophages in control hearts greatly decreased within the injured myocardium at Day 7, whereas at Day 14, there was thinning and fibrosis of the LV wall and very low macrophage content (Figures 5A,C,D). In contrast to control mice, α-gal nanoparticles treated hearts demonstrated at Day 7 a marked increase in macrophage infiltration in the area debrided of cardiomyocytes. The areas in which cardiomyocytes died and were debrided in these mice are stained gray in Trichrome staining and comprise 19–24% of the LV myocardium (Figures 5B,C). The infiltrating cells on Day 7 were confirmed to be macrophages by positive in situ staining of the infiltrating mononuclear cells (Figure 5E, H&E) with the macrophage specific peroxidase-linked F4/80 antibody (Figure 5F). The areas in which macrophages infiltrated at Day 7 in the α-gal nanoparticles treated hearts, fully overlapped the areas devoid of cardiomyocytes (Figure 5B). This is further demonstrated in additional sequential sections at different planes of Day 7 α-gal nanoparticles treated heart (Figure 5G, middle section also shown in Figure 5B). Histological sections above and below the middle section (left and right figures, respectively) display a different distribution of infiltrating macrophages, suggesting that the injured area is of a substantial size with three-dimensional irregular shape. The overlap in the shape of the macrophage infiltrate (H&E) and the area devoid of cardiomyocytes (gray color in Trichrome staining) in Figure 5G, further suggests that a high proportion of the cardiomyocytes died following the ischemia and that one of the activities of macrophages recruited by α-gal nanoparticles is debriding areas of necrotic cardiomyocytes that were injured by the 30 min of LAD occlusion. Similar to control mice, very few macrophages were detected at Day 14 in α-gal nanoparticles treated hearts (Figures 5B,D). However, in clear contrast to control hearts which displayed at Day 14 thinning of the LV wall and fibrosis, α-gal nanoparticles treated hearts at Day 14 displayed near-complete restoration of myocardium structure and only minor residual fibrosis (Figures 5B,C). These observations strongly suggest that the areas debrided of cardiomyocytes at Day 7 in α-gal nanoparticles treated hearts were repopulated with healthy cardiomyocytes, as shown in hearts studied both at Days 14 and 28. It is of note that although the kinetics of macrophage infiltration differs between saline treated hearts and α-gal nanoparticles treated hearts, the damage caused to the myocardium (i.e., % infarct size) is not significantly different in both groups on Day 7. This observation is supported by the echocardiography studies displaying similar impaired contractility in the two groups on Day 7 (Figure 4). However, by Day 28 the contractility is restored to normal in the α-gal nanoparticles treated mice vs. continuation in impaired contractility in saline treated mice.

The histological studies on post-MI myocardial repair presented in Figure 5 imply that the repopulation by cardiomyocytes following injection of α-gal nanoparticles occurs in the second week post-treatment. This raised the question whether the repopulation of the injured myocardium with healthy cardiomyocytes is the result of proliferation of pre-existing cardiomyocytes, or of resident stem/progenitor cells. This was studied on Days 9 and 11 post-MI in mice treated with α-gal nanoparticles by evaluating BrdU (thymidine analog) uptake into nuclei of proliferating cells in the heart 24 and 48 h after intraperitoneal injection of 1 mg BrdU. The sections were stained overnight with the biotinylated anti-BrdU antibody followed by avidin-peroxidase and DAB precipitation. Figure 6 shows the results in a representative mouse studied 24 h post-injection of BrdU on Day 9 (out of 5 mice with similar results). BrdU was effectively taken up by nuclei of proliferating cells at the base (crypt) of intestinal villi of the mice, which served as positive controls of proliferating epithelium (Figure 6A). However, in the heart, the staining of nuclei was sparse and mostly of the stained nuclei that did not correspond to distinct cardiomyocytes (Figures 6B,C). In view of distinct areas of injured cardiomyocytes debrided by macrophages (Figure 5), it would be expected that repopulation of these debrided areas by cells proliferating within the heart will result in appearance of groups of cardiomyocytes displaying uptake of BrdU. Similarly, no such BrdU staining pattern has been observed in any of the hearts in mice injected with BrdU on Day 9 and evaluated 48 h later, or in mice injected on Day 11 and evaluated after 24 or 48 h (not shown). In addition, no distinct nuclear BrdU staining was detected in cardiomyocytes in regions adjacent to areas containing infiltrating macrophages (Figure 6D).

A second method for detecting proliferating cardiomyocytes was performed by anti-PCNA antibody staining of sections of treated hearts harvested on Days 9 and 11. This antibody binds to the proliferating cell nuclear antigen (PCNA) which is a DNA clamp found in nuclei at the stage of DNA synthesis. As shown in Figure 7A, nuclei of proliferating cells are readily stained by anti-PCNA at the base (crypts) of intestinal villi. In contrast, no distinct staining of cardiomyocyte nuclei was observed in the hearts treated with α-gal nanoparticles and harvested 9 or 11 days post-MI (Figures 7B,C). However, in view of the staining of a small number of nuclei observed in treated hearts both by BrdU uptake (Figures 6B,C) and by staining for PCNA (Figures 7B,C), we cannot regard these observations as conclusive evidence that there is a complete absence of proliferating cells in post-MI hearts treated with α-gal nanoparticles.