Model fluorescent nanoparticles target the heart if co-administered with PLGA

PLGA polymers are attractive drug delivery vehicles, especially in the form of nanoparticles, because of their excellent safety profile and ability to spontaneously degrade in biological systems into lactic acid and glycolic acid (Fig. 1a) [22]. We were interested in utilizing PLGA nanoparticles not as drug delivery vehicles but as a cargo-less aid to existing therapies. A previous study showed co-delivery of polymer-based “nanoprimers” lead to reduced clearance of therapeutic LNPs. [20] In a proof-of-concept experiment, we first investigated biodistribution of co-delivered various non-PLGA nanoparticles and nanoparticles made of PLGA. We observed an intriguing heart-biased biodistribution pattern of model polystyrene latex fluorospheres that were co-injected with PLGA nanoparticles. To gain insight into this phenomenon, we synthesized different cargo-less PLGA nanoparticle formulations of varying polymer molecular weight. We co-injected these particles with 25 nm fluorescent latex beads into C57BL/6 mice i.v.. Hearts were excised 1 h post injection (p.i.) and fluorescence was recorded using whole organ imaging following heart perfusion (Additional file 1: Fig. S1a). Larger PLGA polymers (120 kDa) enhanced the uptake of the latex beads in the heart (Additional file 1: Fig. S1b,c). Based on these initial findings, 120 kDa poly(L-lactic-co-glycolic acid) (PLLGA) with a lactic to glycolic ratio of 65:35 was chosen for future experiments. For simplicity, we refer to the nanoparticles made from this polymer as ePL (enhancer polymer). ePL was synthesized by means of standard flash nanoprecipitation from acetonitrile to polyvinyl alcohol (PVA) surfactant in water (Fig. 1b). ePL was characterized by dynamic light scattering (DLS) to measure the particle size and zeta potential (Fig. 1c). The diameter of the particles was determined to be 234.7 ± 15.3 nm, while the zeta potential was 0.12 ± 0.18 mV. Using transmission electron microscopy (TEM), we determined that the particles had a spherical shape (Fig. 1d). Next, we studied the biodistribution of ePL using a fluorescently-labeled ePL. We conjugated PLGA to a europium (Eu) fluorescent dye and blended 5% of Eu-labeled PLGA with unlabeled ePL. After the injection at 10 mg/kg in C57BL/6 mice, followed by perfusion, organ excision, and imaging (Fig. 1e), we found that ~ 90% of ePL accumulated in the liver and ~ 5% in the spleen. No ePL deposition in the heart, brain, or muscle was observed. The clearance of ePL from the systemic circulation was very rapid (T1/2 = 5.3 ± 2.6 min), suggesting that ePL has a high liver affinity. We did not expect significant toxicities associated with ePL because it is based on poly-lactide, which is highly biodegradable and considered safe [22, 23]. Examination of histological staining of heart and liver sections from mice bolus-injected with 30 mg/kg ePL revealed that these tissues appeared all normal following H&E staining, with no obvious sign of fibrosis or cell infiltrations (Additional file 1: Fig. S1d). Because ePL is highly biodegradable, a hepatobiliary route of elimination was observed (Additional file 1: Fig. S1e). Finally, because of significant ePL accumulation in the spleen (Fig. 1e), we tested for the possibility of splenic injury, immune cell infiltration and a splenic infarct. However, we noticed none of these effects on spleen gross pathology (Additional file 1: Fig. S1f) even at high concentrations of ePL. ePL was free from endotoxin contamination (Additional file 1: Fig. S1g).

Fig. 1

ePL synthesis and characterization. a Schematic showing PLGA composition and degradation in vivo. b Schematic of ePL particle synthesis by nanoprecipitation. c Size distribution and zeta potential of ePL by DLS. d Transmission electron microscopy (TEM) examination of ePL shows that ePL particles are spherical in shape. e Biodistribution of ePL labeled with fluorescent europium cryptate 1 h after i.v. injection, showing significant accumulation in the liver (n = 3 animals per group)

ePL enhances the heart tropism of AAV and virus-like nanoparticles

To determine whether the ePL heart-redirecting effect may impact both nanoparticle and AAV delivery, we used latex fluorospheres incorporating far red fluorophore (NPs) and virus-like nanoparticles (VLNPs), which we engineered and synthesized with glycosylated surface that mimicked an AAV capsid. Both NPs and VLNPs had an average diameter of ~ 25 nm, thus closely resembling the average size of AAV capsids. Upon optimization of the injection schedule (Fig. 2a), we discovered that administration of ePL 15 min before VLNPs significantly enhanced VLNP deposition in the heart and detargeted the liver (Fig. 2b, Additional file 1: Figure S2a). Other organs of the reticuloendothelial system (RES) were also detargeted. The effect was particularly striking when glycosylated VLNPs were used (Fig. 2b). Specifically, an injection of ePL before the injection of VLNP resulted in an order of magnitude increased accumulation of VLNP in the heart vs. the liver as evidenced by heart-to-liver ratio 0.41 ± 0.20 for VLNP vs. 4.16 ± 1.55 for ePL + VLNP (p = 0.017) (Fig. 2b). The VLNPs were confirmed to accumulate within cardiomyocytes using immunofluorescence in heart sections (Fig. 2c). Collectively, these data demonstrate that ePL is able to robustly redirect nanoparticles to the heart and detarget the liver and that this effect was dependent on surface glycosylation engineered for VLNPs.

Fig. 2

ePL facilitates heart targeting of virus-like nanoparticles (VLNP), AAVrh.74, and AAV1. a Optimized injection schedule to test VLNP distribution with ePL. b Biodistribution imaging in various organs of fluorescently-labeled VLNP 2 h after i.v. administration. Fluorescent intensity is quantified and normalized to organ fluorescence from non-injected animals. c Fluorescence microscopy of heart sections from mice injected as indicated. The sections were stained with anti-cardiac troponin (CT3) antibodies (green) co-registered with endogenous NP fluorescence (red, arrow-heads). Nuclei were visualized after staining with DAPI (blue). BAT brown adipose tissue, WAT white adipose tissue. n = 5–6 mice/group. d Immunofluorescence analysis of GFP protein expression in hearts and livers of ePL- or vehicle-injected mice 30 days after injection of a single dose of 5e11 vg/kg AAVrh74.CMV-eGFP. GFP expression was visualized and quantified after the staining with anti-GFP antibodies. e Western blot analysis of GFP protein in heart and liver lysates of AAVrh74.CMV-eGFP-injected animals. CBB  Coomassie brilliant blue. f Densitometry quantification of blots in e. g qPCR analysis of vector copy numbers (VCN) in various organs from AAVrh74-injected animals with and without ePL. § data point below range of plot (8.37E3 VCN) h Western blot analysis of GFP protein in heart and liver lysates of 5e11 vg/kg AAV1.CMV-GFP-injected animals. i Densitometry quantification of (h) from AAV1-injected animal heart and liver lysates, as well as the ratio of those organs. j qPCR analysis in animals injected with 5e11 vg/kg AAV1.CMV-eGFP. n = 3–6 mice/group

Encouraged by the preliminary results on ePL-enhanced VLNP delivery (Fig. 2b, c) and safety (Additional file 1: Fig. S1), we set out to test the effect of ePL on AAV. In our pilot studies, we used the AAVrh.74 serotype injected i.v. because it has been clinically validated and exhibits cardiac muscle tropism. [24]

First, we tested the hypothesis that ePL would further increase the AAVrh.74 affinity to the heart. A single i.v. injection of 5e11 vg/kg (vector genomes per kilogram) single-stranded AAVrh74 carrying the eGFP transgene under the control of the CMV promoter (AAVrh74.CMV-eGFP, referred to as AAVrh.74 further in the text for simplicity) was administered following the injection schedule shown in Fig. 2a. Because the expression of the AAV-carrying transgene (eGFP in this case) becomes appreciable at four-to-six weeks following the delivery [1, 25], the eGFP protein expression was examined in the heart and the liver 30 days after the ePL and AAV injections. A 5.1 ± 0.3 fold increase (p < 0.001) in the eGFP expression in the heart was observed with ePL + AAVrh.74 as compared to animals that received AAVrh.74 alone (Fig. 2d, quantified in Additional file 1: Fig. S2b, c), as demonstrated by immunofluorescence studies using antibodies against the eGFP transgene. This targeting pattern was additionally confirmed by quantitative Western blotting using an antibody against eGFP and total protein staining (Fig. 2e, f) which demonstrated heart-to-liver ratios of 5.5 ± 1.0 for ePL + AAVrh.74 group vs. 0.9 ± 0.1 1.2 for AAVrh.74 alone (p = 0.031) (Fig. 2f). Further biodistribution studies 2 h p.i. were conducted by qPCR in major organs to detect AAVrh.74 DNA using primers and probes specific for CMV (Fig. 2g). The data demonstrated robust heart targeting and liver de-targeting with ePL and is suggestive of redirecting AAVrh.74 capsids from the liver to the heart, rather than acting on the eGFP expression in the heart. In these experiments, AAVrh.74 displayed a 6.5 ± 1.9-fold increase (p = 0.045) in heart targeting and a 6.8 ± 0.05-fold decrease in liver uptake (p = 0.0027) with ePL as compared to vehicle control (Fig. 2g). Interestingly, the data also demonstrated a trend towards lung detargeting. Finally, skeletal muscle uptake of AAVrh.74 was examined separately by Southern blotting (Additional file 1: Fig. S2d) due to the levels of viral DNA in the muscle being too low for traditional qPCR detection. The data demonstrated a significant 2.5 ± 0.8-fold increase (p = 0.041) in AAVrh.74 accumulation in the skeletal muscle with ePL as compared to the vehicle injection.

Next, we wanted to assess whether ePL-driven heart targeting is still possible with an AAV serotype known for its poor heart tropism in i.v.-injected mice. We selected AAV1 for this purpose [1, 26]. We administered AAV1.CMV.eGFP (AAV1) with and without ePL in the same manner as above, analyzing the protein expression 30 days p.i. (Fig. 2h, i) and AAV1 DNA levels 2 h p.i. in the heart and liver (Fig. 2j). Notably, on the eGFP protein expression level, ePL significantly improved the heart-to-liver ratio was 1.31 ± 0.14 for the ePL + AAV1 group vs. 0.89 ± 0.10 for the AAV1-only group (p = 0.032) (Fig. 2i). Strikingly, a 14.1 ± 0.4-fold increase (p = 0.034) in heart delivery with ePL (as compared to the AAV1-only control) was observed on the DNA level (Fig. 2j). In contrast, 3 of 5 injected animals in the AAV1-only group did not show any viral DNA in the heart, whereas high liver accumulation was detected in the same animals. Collectively, this data suggests that ePL pre-injection enables significant enhancement of AAV accumulation in the cardiac muscle for AAV serotypes with or without inherent cardiac tropism. We hypothesized that this heart targeting mechanism likely relies on the systemic distribution of ePL and AAV. This is suggested by the fact that we have not observed any evidence of ePL accumulation in the heart (Fig. 2b), and direct interactions of ePL with the viral particles are also unlikely due to the time delay between the ePL and AAV injections.

ePL accumulates in Kupffer cells and produces serum-derived factors enhancing AAV uptake in cardiac/muscle cells

The next step was to uncover the driving forces behind heart targeting and liver detargeting with ePL. Given that more than 90% of the injected dose of ePL accumulates in the liver (Fig. 1), we examined the composition of liver cells targeted by the ePL after a bolus injection in C57BL/6 mice (Fig. 3a). Flow cytometric evaluation of liver cells after the injection of Atto647-labeled ePL indicated that a subpopulation of Kupffer cells (KCs), identified by F4/80+CD11b+, represented the majority of the ePL uptake in the liver (Fig. 3b, gating strategy in Additional file 1: Fig. S3a). We tested intraperitoneal (i.p.) and intravenous (i.v.) routes of administration because our earlier studies showed that i.p. injections of ePL do not produce heart targeting of nanoparticles and AAVs (data not shown). Other cell types tested (T regulatory cells, F4/80+CD68+ KCs, hepatocytes, LSECs, and neutrophils, Fig. 3a) [27] were only marginally targeted by ePL. F4/80+CD11b+ KCs are known to be liver-resident macrophages that secrete various acute response factors including cytokines [28, 29]; therefore, we tested plasma cytokine levels in mice injected with ePL (Fig. 3c). Interestingly, profiling major cytokines using a Luminex 36-cytokine panel showed that ePL did not induce the release of either immunosuppressing or inflammatory cytokines 4 h after i.v. injection (Fig. 3c).

Fig. 3

Intravenous administration of ePL targets cytokine-releasing KCs in the liver. a Mice were administered i.v. or intraperitoneally (i.p.) with fluorescently-labeled ePL at 70 mg/kg (n = 3 per group). b Livers were digested to obtain single cell suspension which was then analyzed using flow cytometry. Cell markers: regulatory T cells (Tregs): CD3+CD4+CD25+; Neutrophils: CD11b+F4/80−Ly6G+; Cytokine-releasing Kupffer Cells (crKCs): CD11b+F4/80+; Phagocytic Kupffer Cells (pKCs): CD11b+F4/80+CD68+; Hepatocytes: CD11b−F4/80−CD146−CD206−; liver sinusoidal endothelial cells (LSECs): CD11b−F4/80−CD146+CD206+. c Cytokine levels in plasma from ePL treated mice (n = 3/group) 4 h after i.v. injection. d Scheme of in vivo-in vitro serum testing assay. Cells were supplemented with mouse serum extracted from ePL- or PBS-treated mice. e Heatmap obtained from luminescence readout of various transduced cells as indicated. f mRNA transcripts expression of immune defense and anti-viral genes in cultured primary macrophages after treatment with 1 mg/mL ePL in vitro. Nanostring nCounter gene expression counts are shown (n = 2/gr). g ePL activated STAT1 in macrophages as shown by increased phosphorylation at Tyr701. N.d.  not detected

Because F4/80+CD11b+ KCs are known to secrete factors other than cytokines [30], we hypothesized that the presence of non-cytokine factors in the serum could be attributed to heightened AAV uptake as shown above. To test this hypothesis, we performed in vivo-in vitro serum screening assays as depicted in Fig. 3d. Here, C57BL/6 mice were injected with ePL or PBS (normal control) and serum was collected 1 h p.i.. Next, we cultured 5 different cell lines and primary cardiomyocytes obtained from C57BL/6 mice. The cultured cells were treated with collected serum at different doses in the presence of AAV1 carrying a luciferase transgene. Luciferase luminescence was then analyzed 24 h after the treatment and normalized to the total cellular protein. Strikingly, human embryonic kidney (HEK293) and cervical cancer HeLa cell lines (non-muscle, non-cardiac) did not show any significant AAV1 uptake enhancement with ePL serum vs. control normal serum, while cardiomyoblast H9C2 cells and L6 rat skeletal muscle myocytes and myotubes demonstrated robust dose-dependent increases in transduction (Fig. 3e, Additional file 1: Fig. S3b-g). Primary cardiomyocytes showed preferential AAV1 accumulation when treated with the ePL serum, however, only at the lowest dose of the ePL serum. This is likely due to the disruption of the normal cardiomyocyte phenotype by high serum concentrations in culture, as was reported previously. [31, 32]

In order to further elucidate the actions of ePL in macrophages, we incubated ePL with mouse primary bone marrow derived macrophages (BMDMs) in vitro. Fluorescent Atto647-labeled ePL was rapidly engulfed by BMDMs as seen from fluorescence microscopy imaging (Additional file 1: Fig. S3h). In contrast, hepatocyte-like HepG2 cells showed almost no accumulation. Further, ePL drastically increased a number of viral defense genes in cultured BMDMs, including those responsible for viral replication and translational initiation (Ifit, Stat), indicating an anti-viral uptake phenotype. This was registered 24 h after ePL incubation in culture (Fig. 3f). Most notably, ePL rapidly increased STAT1 activation manifested in the increase in phosphorylation at Tyrosine 701 [33]. This occurred as fast as 30 min after ePL incubation with BMDMs (Fig. 3g) and persisted for at least 2 h. STAT1 and its activation through phosphorylation are essential for the host immune defense, especially in the context of viral infections [34]. Therefore, it is possible that ePL “primes” the macrophages, KCs, and possibly other cells, to be less permissive to infection by AAV, thus causing the delayed AAV clearance by the immune system.

Collectively, this data suggests that the ePL serum contains yet unidentified factor(s) responsible for paracrine signaling that leads to the drastic increase of AAV and nanoparticle targeting of the heart, and that ePL induces an antiviral-uptake phenotype in macrophages.

ePL allows for delayed blood clearance of AAV and nanoparticles

The findings shown above only partially explain the enhanced AAV and nanoparticle heart uptake and liver detargeting when injected with ePL. Therefore we examined various factors that may contribute to AAV/nanoparticle targeting to the heart.

Previous studies on AAV9, a serotype with one of the best-in-class heart transduction efficiencies [19] and the longest systemic circulation [35], have proposed that increased circulation half-life may help overcome slow transvascular AAV transport through the tightly sealed capillary endothelium in the heart, thus efficiently transducing cardiomyocytes by virtue of a substantially longer circulation time [35]. Pharmacokinetic studies using ePL co-administered with 1e10 vg/kg AAV9 demonstrated a 4.0 ± 0.6-fold (p = 0.011) increase in the circulation half-life with ePL as compared to vehicle-injected mice (Fig. 4a).

Fig. 4

ePL injection changes pharmacokinetics of AAV9 and nanoparticles and acts on neutralizing antibodies. a qPCR analysis of blood from AAV9.CMV-GFP injected mouse over 30 h with and without ePL. b Ear imaging IVM setup. Intravital microscopy (IVM) enabled blood clearance quantification of model latex nanoparticles with and without ePL pre-treatment. c ePL or vehicle PBS was administered i.v. to C57BL/6 mice 15 min prior to injection of latex fluorospheres (green) via catheter. Background staining was accomplished after injection of rhodamine dextran (red). A representative blood vessel imaging is depicted. A white rectangle is shown as a typical region of interest in which signal over time was quantified. d Quantification of the fluorescence signal over time from IVM experiments. e Lec2 cell transduction assays in the presence of AAV9-injected mouse serum with and without ePL injection. AAV9.CMV.Luc was used to transduce the cells and luminescence values were normalized to total protein content. f Lec2 transduction assays as in e, but with ePL added directly into cell culture media

To determine if the findings on AAV9 pharmacokinetics extend to nanoparticles as well, we performed blood clearance measurements of FluoSpheres latex polystyrene nanoparticles using intravital microscopy (IVM) of C57BL/6 mouse ear microcirculation (Fig. 4b). IVM has been previously used to measure clearance of various nanoparticles in several mouse models with success [36]. This method is non-invasive and allows for blood clearance recording in real time. FluoSpheres (1 mg/kg) were injected just after the bolus ePL at 30 mg/kg i.v. and the fluorescence was measured over 1 h. The decay of fluorescence intensity in the mouse ear vein was recorded and plotted against time (Fig. 4c, d). The data indicate that ePL injection significantly delayed the blood clearance of FluoSpheres, especially very early (< 1 min) after the injection.

To explore the possible mechanism behind such delayed blood clearance of AAV9 and FluoSphere nanoparticles, we looked into potential involvement of neutralizing antibodies (nAbs). Pre-existing nAbs impair AAV persistence in circulation by opsonization [37, 38]. To investigate whether ePL has the ability to reduce or prevent nAbs production, we designed an in vivo-in vitro assay that leveraged inherent immunogenicity of AAV9 to produce high levels of nAbs after a single injection [39, 40]. Two separate experiments were conducted. First, a bolus i.v. injection of AAV9 at 1e10 vg/kg in C57BL/6 mice with or without 30 mg/kg of ePL, followed by serum collection 30 days later produced AAV9-specific nAbs-enriched serums. These serums were mixed with AAV9 at MOI 50,000 in culture medium followed by the incubation with Lec2 cells, a cell line that expresses sialylated glycans on the surface, which are known targets for the AAV9 intracellular entry. The results show that co-injection of AAV9 with ePL produced significantly less nAbs as evidenced by the increased AAV9 infectivity of Lec2 cells when the corresponding serum was added (6.7E-3 ± 8E-4 for AAV9 + ePL serum vs. 4.0E-3 ± 5E-4 for AAV9 serum, p = 0.0096, Fig. 4e). Second, using the serum from AAV9-only injected mice, we incubated the Lec2 cells in the presence of ePL, spiked directly in the culture medium. Strikingly, this similarly increased AAV9 infectivity (1.0E-2 ± 2E-3 for ePL-spiked cells vs 4.0E-3 ± 5E-4 for AAV9 serum only, p = 0.014, Fig. 4f), suggesting that ePL may have the ability to directly prevent AAV9 neutralization by nAbs.

Collectively, these data suggest that ePL improves the blood residence time of AAV and nanoparticles by acting on nAbs as at least one of the modes of its action.

ePL paracrine signaling in cardiomyocytes enables heart targeting

Because ePL is sequestered rapidly by macrophages in vitro and KCs in vivo and the serum from mice injected with ePL aids in the AAV uptake in muscle/cardiac cells, we hypothesized that one of the ways ePL exerts its actions in the heart could be through communication between macrophage secretome and heart cardiomyocytes. To test this, we performed a series of experiments to probe paracrine signaling in the heart induced by ePL (Fig. 5). We isolated primary mouse cardiomyocytes from C57BL/6 mice and conditioned medium (C/M) from BMDMs cultured in serum-free medium and treated with ePL, PBS vehicle control, and lactate, which is a product of ePL biodegradation. Next, using cardiomyocytes pre-loaded with fluorescence indicators, we conducted a series of treatments using various C/Ms while simultaneously recording calcium levels (Ca2+) and the cell length of single cardiomyocytes through quantitative video microscopy. Notably, ePL C/M significantly decreased the amplitude of cytosolic Ca2+ transients without significant changes in cardiomyocyte contraction (Fig. 5a, Additional file 1: Fig. S4a, b). Direct treatment of cardiomyocytes with ePL or lactate did not have the same effect, or its magnitude was substantially lower relative to untreated BMDM C/M (Additional file 1: Fig. S4a, b).

Fig. 5

ePL induces paracrine signaling in the heart. a Contractility and intracellular Ca2+ in primary mouse cardiomyocytes in response to ePL or conditioned media (C/M) from bone marrow-derived macrophages (BMDM) pre-incubated with ePL. b Schematic demonstrating ePL C/M effect on calcium uptake and the subsequent signaling cascades resulting in increased glucose uptake. c GLUT4 translocation assays in rat L6 myotubes expressing Myc-tagged GLUT4. d Western blot of heart lysates from PBS- or ePL-treated mice probing for expression of pPDH Ser293. e Quantification of immunoblot in d. f NBDG (fluorescently-labeled glucose) uptake in primary cardiomyocytes using ePL alone or conditioned media (C/M) from ePL-treated BMDMs. g Western blot analysis of heart lysates from ePL-treated mice, showing higher expression of paucimannose, SLC35A1, and glycoconjugates detected by lectin MALII. h Quantification of immunoblots in g, with additional data on liver expression (see Additional file 1: figures). n = 3 animals per group for in vivo experiments and n = 6 for in vitro experiments in isolated cardiomyocytes

It has been previously proposed that the decrease in Ca2+ entry in cardiomyocytes may be a driver of increased glycolysis and glucose uptake in the heart (Fig. 5b) [41]. Indeed, cardiomyocyte in vitro glucose uptake was increased with ePL C/M treatment (Fig. 5f). Glucose uptake is dependent on the expression of GLUT transporters in all cells. We tested whether muscle-specific GLUT4 transporter expression is altered after ePL C/M treatment. To accomplish this task, we chose a well validated, genetically engineered L6 cell line constitutively expressing myc-tagged GLUT4. Immunostaining with anti-myc antibodies in these cells allows for sensitive tracking of the GLUT4 expression. As expected, GLUT4 was highly upregulated in these cultured muscle L6 myotubes upon ePL C/M treatment (Fig. 5c). The expression of GLUT4 was time-, ePL C/M dose- and insulin-dependent (Additional file 1: Fig. S4c, d).

To further investigate the effects of reduced Ca2+ uptake by ePL, we looked into metabolic enzymes that regulate glucose metabolism and are Ca2+-dependent. Pyruvate dehydrogenase (PDH) is the enzyme that catalyzes the transformation of pyruvate to acetyl-CoA, however, phosphorylated PDH (pPDH), including at serine 293, inhibits this catalytic action [42]. Notably, pPDH dephosphorylation is catalyzed by a Ca2+-sensitive phosphatase PDP1 (Fig. 5b). It has been previously shown that Ca2+-mediated dephosphorylation of pPDH in cardiomyocytes decreases glucose oxidation and promotes glycolysis, without changes in cardiomyocyte contraction [41]. Given these facts and the results above, we tested whether pPDH levels change in the heart after injection of ePL in C57BL/6 mice. Indeed, the treatment with ePL resulted in significantly higher levels of pPDH as compared to vehicle PBS-injected control (1.54 ± 0.23 for ePL vs. 0.68 ± 0.6 for PBS, p = 0.02, Fig. 5d, e). This data is suggestive of cardiomyocyte “reprogramming” in response to ePL C/M treatment, causing cardiomyocytes to ramp up their glucose consumption, likely due to dampened dephosphorylation by PDP1.

Enhanced glucose uptake in cardiomyocytes could be partially responsible for glycosylated VLNPs uptake in the heart in vivo. However, even though many AAVs are highly surface-glycosylated [43, 44], AAV tissue uptake through GLUT glucose transporters is largely unknown. Rather, tissue protein glycosylation seems to have a significant role in predicting AAV organ tropism, which is governed by the expression of various N- and O-glycoprotein glycans on the cell surface, serving as receptors for the majority of the AAV serotypes [45,46,47]. Protein glycosylation is partially driven by glucose-derived mannose that is directly used for glycoconjugate synthesis. To test whether ePL plays a role in tissue glycoconjugate expression, we injected a bolus ePL at 30 mg/kg or PBS vehicle control in C57BL/6 mice and extracted whole hearts 1 h later. The heart-derived proteins were analyzed by Western blotting immunoprobing for key molecules involved in glycoconjugate synthesis (Fig. 5g). First, pauci-mannosylation was assessed using antibodies against paucimannose, which detect the posttranslational modification of proteins by simple mannose units. ePL injection significantly increased the levels of paucimannose in the heart, suggestive of glucose-to-mannose conversion in the first step of glycoconjugate synthesis [48]. Next, in the same heart lysates, we probed for an CMP-sialic acid transporter, also known as CMP-Neu5Ac, which is a key transporter protein in cellular sialylation (integral component of glycoproteins) [49]. Notably, SLC35A1 was also highly upregulated in the hearts of ePL-treated mice (Fig. 5g). Finally, we screened for diversity of glycoconjugates in the same heart lysates by probing with various lectins that detect specific oligomannose-rich glycoproteins (Fig. 5h, Additional file 1: Fig. S4e). The cardiac proteins from ePL-injected animals demonstrated statistically significant increase in binding to MALII lectin (Maackia Amurensis) that binds to sialic acid in an (α-2,3) linkage (14.69 ± 0.33 for ePL vs. 12.24 ± 0.11 for vehicle, p = 0.002) [50]. Similarly, RCA lectin (Ricinus Communis Agglutinin), which binds to galactose or N-acetylgalactosamine also demonstrated enhanced affinity to proteins from ePL-injected heart extracts (Fig. 5h, Additional file 1: Fig. S4e). Conversely, liver tissue lysates showed a significant reduction in binding to MALII (2.7 ± 0.2 for ePL vs. 4.4 ± 0.4 for vehicle, p = 0.2, Fig. 5h, Additional file 1: Fig. S4f), suggesting that such differential glycoprotein expression in the heart vs. the liver could be a consequence of heart targeting and liver detargeting by A