Erythrocyte-mediated delivery of bioorthogonal nanozymes for selective targeting of bacterial infections

Abstract

Bioorthogonal transformation of imaging and therapeutic substrates using transition metal catalysts (TMCs) provides a toolkit with diverse applications in biomedicine. Controlled localization of bioorthogonal catalysis is key for enhancing their therapeutic efficacy by minimizing off-target effects. Red blood cells (RBCs) are highly biocompatible and are susceptible to hemolysis by bacterial toxins, providing them with intrinsic targeting to bacterial infections. A hitchhiking strategy using RBCs is reported, that activates bioorthogonal catalysis at infection sites. A library of nanoparticles embedded with TMCs (nanozymes) featuring diverse functional groups with different binding ability to RBCs is generated. These engineered nanozymes bind to RBCs and subsequently release upon hemolysis by bacterial toxins, resulting in selective accumulation at the site of bacterial infections. The antimicrobial action is specific: catalytic activation of pro-antibiotics eradicated pathogenic biofilms without harming non-virulent bacterial species.

Graphical Abstract

TOC : Hitchhiking bioorthogonal nanozymes on red blood cells for selective killing of pathogenic bacterial infections

Bioorthogonal catalysis offers a strategy for chemical transformations that complements natural bioprocesses and has become a useful tool in biochemistry and medical sciences.^1,2,3 Nanoparticles embedded with transition metal catalysts (nanozymes) can effect useful transformations with high efficiency and selectivity.^4,5 These bioorthogonal nanozymes can activate prodyes and prodrugs to their active forms in situ, providing ‘nanofactories’ at sites of interest.^6,7 Controlled localization of bioorthogonal nanozymes to targeted disease sites is a key strategy in enhancing therapeutic efficacy by minimizing off-target effects.^8,9 One approach to control spatiotemporal localization of biorthogonal catalysis utilizes tuning the size of carrier.^10,11 Alternatively, bioorthogonal catalysts can be functionalized with different ligands, peptides or biomolecules to target the diseased site.^7,8 However, these synthetic carrier-based approaches are susceptible to non-specific uptake and accumulation at non-targeted site, compromising the efficacy of the therapy.^12,13

Red blood cells (RBCs) feature biocompatibility, long circulation time and low immunogenicity, making them versatile carriers for cell-based drug delivery systems.^14,15,16 Recent studies have demonstrated that RBC-hitchhiking of nanoparticles (NPs) enhanced the delivery efficacy to target organs with minimal non-specific uptake by the reticuloendothelial system.^17,18 The surface functionality of NPs dictates their interaction with RBCs and is key to the generation of RBC “super-carriers” for drug delivery systems.¹⁹ Cationic NPs can bind to the anionic glycocalyx on RBC surface,²⁰ while NPs can bind to hydrophobic domains present on plasma membranes irrespective of NP surface charge.²¹ Significantly, tuning hydrophilic and hydrophobic moieties on NP surfaces can dramatically alter their hemolytic behavior.²²

RBCs are rapidly hemolyzed by bacterial toxins, providing RBC carriers with an intrinsic targeting ability towards pathogenic bacteria.^23,24,25 We hypothesized that integration of bioorthogonal nanozymes into RBCs with would offer an effective route to combat bacterial infections through selective biorthogonal catalysis at the infection site. We report here the fabrication of a series of nanozymes (NZs) that use nanoparticle surface structure to modulate their interactions with RBCs. Structure-activity studies revealed that hydrophilic cationic NZs stably attach onto RBC surfaces without significant hemolysis, yielding RBC-hitchhiked nanozymes (RBC-NZs). These RBCs were hemolyzed by bacterial toxins, resulting in accumulation of nanozymes at the infection site. The released NZs were highly catalytically active, uncaging pro-antibiotics to effectively eradicate biofilms formed by uropathogenic bacteria. In contrast, minimal catalysis was observed in non-virulent bacterial strains and macrophage cells, owing to negligible non-specific localization of RBC-hitchhiked biorthogonal catalysts.

Notably, RBC-NZs showed high immune- and biocompatibility. RBC-hitchhiked nanozymes passively target bacterial infections triggered by bacterial toxins, resulting in controlled bioorthogonal catalysis, and minimizing off-target antibiotic activation.

RESULTS AND DISCUSSIONS

Synthesis and fabrication of functionalized nanozymes

Nanozymes were fabricated using AuNPs with ~ 2 nm core diameter as scaffolds. AuNPs were functionalized with ligands featuring three main components: (1) a hydrophobic alkyl chain interior enabling encapsulation of hydrophobic catalysts, (2) tetra ethylene glycol spacer providing biocompatibility and (3) terminal groups dictating NP binding with red blood cells.²⁶ Aqueous solution of AuNPs was mixed with catalysts (iron (III) tetraphenyl porphyrin (FeTPP)) dissolved in organic solvents in 1:1 ratio (% v/v). The mixture was slowly evaporated to remove organic solvents, resulting in encapsulation catalysts in the surface monolayer of AuNPs to form nanozymes (NZs). Excess catalyst was filtered out by using molecular-weight cut-off filter, centrifugal filter and dialysis (experimental details are provided in methods section).²⁷ The chemical functionality of NZ-surface ligands plays a critical role in determining their compatibility with RBCs, in-turn dictating ability of NZs to hitchhike on RBC surface.^20,23 We synthesized a family of NPs (1–9) with varying surface charge, hydrophobicity and aromatic properties using ligand place exchange reactions with a pentanethiol-capped 2-nm Au core (Figure 1a, detailed description for synthesis of NPs is described in Supporting Information). NPs (1–9) were used to fabricate NZs (1–9) and characterized using dynamic light scattering (DLS) (SI Figure S1, Table S1) and transmission electron microscopy (TEM, SI Figure S2). TEM and DLS results showed no signs of aggregation of NZs upon catalyst encapsulation. The number of catalysts encapsulated was quantified using inductively coupled mass spectrometry (ICP- MS, SI Table S2), indicating that 30±6 catalyst molecules were encapsulated per AuNP for NZ 1–9.

Figure 1.

a) Molecular structures of the nanozyme ligands used in the RBC-adsorption study. b) Structures of the substrates Resorufin and moxicillin derivative (Pro-Res, Pro-Mox) and products (Resorufin, Moxifloxacin) after cleavage by TMC. c) Schematic representation showing hitchhiking of NZs on red blood cells, selective targeting of biofilm infections due to lysis of RBCs in presence of bacterial toxins, and intrabiofilm generation of antibiotics by transition metal catalysts (TMCs) embedded in the nanoparticle monolayers.

Attachment of engineered nanozymes on RBCs

Our initial focus was to adsorb NZs on RBCs without compromising the stability of the cell membrane; as such, we screened the library of NZs for hemolytic activity against RBCs both in PBS and serum-containing media. NZs (1–9) were incubated with RBCs for 30 minutes and the absorbance of released hemoglobin was measured at 570 nm.²⁸ We observed that cationic hydrophilic NZs (NZ 1–2) showed minimal hemolysis as compared to their hydrophobic counterparts. Similarly, anionic and zwitterionic NZs (NZ 7–9) showed minimal hemolysis of RBCs (Figure 2a, SI Figure S3, Table S1), consistent with previously reported studies.²⁷ Next, we studied the adsorption of non-hemolytic NZs (NZ1–2, NZ7–9) on RBCs to determine their suitability for RBC hitchhiking. NZs were incubated with RBCs for 30 minutes and washed to remove excess NZs from the RBC-NZ product. The RBCs were then analyzed using ICP-MS to quantify gold content on the cells.

Figure 2.

a) Dose-dependent hemolytic activity of NZ 1–NZ 9 in the absence of plasma proteins. Hemolysis was calculated using water as the positive control, Error bars represent standard deviations (n = 3). b) NZ adsorption on red blood cells after incubation for 30 minutes, and c) after multiple cycles of centrifugation (500 nM of NZ 1), as measured using ICP-MS. d) Dose-dependent NZ adsorption for NZ 1 on red blood cells. e) Dose-dependent catalysis of free NZ 1 (Bare-NZ) and RBC-NZs (100 – 1000 nM) in PBS for 1 h at 37 °C, f) Confocal image of RBCs loaded with NZ1 catalyzing Pro-Res to generate fluorescence.

Cationic NZs showed significant adsorption on RBCs as compared to the anionic and zwitterionic NZs (Figure 2b, Table 1), attributed to electrostatic interactions between NZs and RBCs. NZ surface chemistry does not significantly affect RBC biocompatibility as evidenced by unaltered expression of CD47 self-markers on engineered RBCs.²⁹ Minimally hemolytic NZs (NZ 1–2 and 7–9) do not affect the expression of CD47 biomarkers on RBCs (SI Figure S9), whereas RBCs engineered with NZ (3–6) significantly reduce the expression of CD47 on RBCs. CD47, “markers of self” are critical for immunocompatibility of RBCs and prevents RBC phagocytosis by the cells of reticuloendothelial system.³⁰

Table 1.

showing the NZ size, zeta potential, hydrophobicity indices of functional group, number of NZs per RBC and hemolytic effect caused by NZs at 500 nM concentration to RBCs.

Nanozyme	Hydrodynamic diameter (nm)	Zeta Potential (mV)	Relative hydrophobicity indices of R groups (log P)	NZs per RBCs	% Hemolysis (NZ concentration = 500 nM)
NZ1	9.1 ± 2.1	18.3 ± 6.5	0.34	~ 5960	5.6 ± 0.15
NZ2	9.8 ± 2.3	18 ± 5.9	0.72	~ 5840	13.5 ± 0.23
NZ3	8.7 ± 3.5	19.9 ± 4.8	2.79	~ 4150	58.9 ± 4.2
NZ4	8.7 ± 0.9	24.8 ± 5.2	4.81	~ 2550	78.1 ± 6.3
NZ5	8.1 ± 1.9	20.2 ± 5.7	1.94	~ 3760	52.3 ± 4.6
NZ6	8.1 ± 2.4	19.6 ± 8.8	2.04	~ 3600	40.9 ± 3.2
NZ7	8.1 ± 3.7	−3.20 ± 5.3	−3.63	~ 660	2.83 ± 0.83
NZ8	8.4 ± 4.1	2.69 ± 4.9	−5.50	~ 780	2.7 ± 0.15
NZ9	9.4 ± 2.9	−27.7 ± 5.4	0.82	~ 1510	2.55 ± 0.12

Catalytic activity of RBC-hitchhiked nanozymes

Nanoparticles frequently detach from RBCs due to shear force and lose their targeting ability.²⁶ Hence, we further investigated the stability of NZs hitchhiked on RBCs by subjecting these RBC-NZs to multiple washing and centrifugation cycles. No significant difference in Au content was observed even after 5 centrifugation cycles, indicating that NZs remain attached to RBCs (Figure 2c). Adsorption of NZ 1 on RBCs was further investigated at a reduced incubation time of 30 minutes with varied concentrations (Figure 2d). It was observed that the nanozyme adsorption on RBC saturates at the concentration of 500 nM, with minimal increase in Au content on RBCs upon incubating with 1000 nM NZs. The catalytic activity of RBC-NZs was assessed by fluorometric measurement of resorufin molecule fragmented from the non-fluorescent pro-Res (Figure 1b) due to azide reduction by FeTTP catalyst in presence of glutathione (1 mM).³¹ The observed linear increase in fluorescence indicates that NZs retain their catalytic activity even after adsorption on RBCs. The rate of fluorescence increase was similar for free NZs as compared to RBC-NZs (500 nM). Moreover, the catalytic rate of RBC-NZs increased linearly as the concentration of RBC-NZs was increased (50–1000 nM), as seen in Figure 2e and SI Figure S11. RBC-NZs incubated with pro-Res and glutathione were visualized via confocal microscopy, further indicating that NZs retain their catalytic activity after adsorption.

Localized catalysis by nanozymes at the infection site

RBCs lose their prolonged circulation upon damage to their plasma membrane such as that caused by pore-forming toxins (PFTs).^16,19 Infections caused by bacteria often involve secretion of PFTs as a virulence mechanism.^32,33 These toxins disrupt the host cell membrane for pathogenesis, causing hemolysis of RBCs. For example, the α-hemolysin toxin released by S. aureus and E. coli is one of the key virulence factors of the invading strains.^17,34,35 We investigated the hemolysis of RBCs caused by uropathogenic clinical isolates (E. coli, methicillin-resistant S. aureus (MRSA)) and non-pathogenic laboratory strains (P. aeruginosa, B. sub). We observed that uropathogenic strains caused complete hemolysis of the RBCs within 30 minutes of incubation with RBCs, whereas the non-pathogenic strains caused minimal hemolysis (Figure 3a). Having established that cationic hydrophilic NZs can hitchhike onto RBCs and that these RBCs were hemolyzed in presence of bacterial infections, we set out to determine whether hemolysis of RBCs could result in selective catalytic activity of RBC-NZs against pathogenic bacterial infections.

Figure 3.

a) Hemolysis of red blood cells by bacterial biofilms. b) Confocal images of biofilms incubated with RBC-NZs (1 h) followed by incubation with Pro-Res (1 h, 10 μM). c) Nanozyme diffusion of Au (ng/well) in different bacterial biofilms including pathogenic (methicillin-resistant S. aureus, MRSA, E. coli, S. epidermidis and En. cloacae) and non-pathogenic (P. aeruginosa ATCC 17660, B. Sub FD6b, Micrococcus luteus, Burkholderia cepacia, GFP exp E. coli) biofilms after incubation for 1 day with RBC-NZ (10⁷ cell/mL, 100 nM NZ), as measured by ICP-MS. Cellular uptake of Au (ng/well) in macrophage (RAW 264.7) (20,000 cells/well) after incubation for 1 day with RBC-NZ (10⁷ cell/mL, 100 nM NZ), as measured by ICP-MS. d) Quantification of Au (ng/ml) on RBCs-nanozymes incubated in PBS /Triton-X.

Conventional antibiotic-based strategies to combat bacterial infections often disrupt the ecology of the human microbiome by killing helpful bacteria species inhabiting the host.^36,37 Hence, it is critical to develop strategies with increased specificity to target pathogenic infections.^38,39 We investigated the selectivity of RBC-NZs towards virulent biofilms through imaging studies using confocal microscopy. Studies for imaging biofilms were based on generation of fluorophore (Resorufin) through aryl reduction of non-fluorescent precursor (Pro-Res) as shown in Figure 1b.^26,29 RBC-NZs were incubated with toxin-secreting (green fluorescent protein (GFP) expressing methicillin-resistant Staphylococcus aureus, MRSA) and non-virulent (GFP-expressing E. coli) bacterial biofilms for 24 hours. Biofilms were then washed multiple times, followed by 1-hour incubation with substrate and subsequent washings. Uropathogenic biofilms showed bright red fluorescence when observed via confocal imaging, with minimal fluorescence observed in non-pathogenic biofilms (Figure 3b). These studies demonstrate that RBC-NZs showed dramatically increased catalytic activity of NZs in pathogenic biofilms as compared to their non-pathogenic counterparts. Moreover, Z-stack confocal images demonstrate that NZs can completely penetrate the EPS matrix of biofilms, indicating their potential as an effective therapeutic strategy.⁴⁰

Immunocompatibility of RBC-attached nanozymes

After establishing selective catalytic activity of RBC-NZs in pathogenic biofilms, we tested the accumulation of RBC-NZs using ICP-MS in pathogenic and non-pathogenic biofilms. We observed that RBC-NZs showed increased accumulation in toxin-secreting uropathogenic bacterial biofilms based on Au, whereas a minimal amount of Au was observed in non-hemolytic bacterial biofilms (Figure 3c). Free NZ 1 showed similar accumulation in pathogenic and non-pathogenic biofilms (SI Figure S4). These results are consistent with our observations indicating that pathogenic bacteria cause higher hemolysis of biofilms as compared to their non-pathogenic counterparts. In another experiment, we determined the amount of Au attached to the cell debris upon hemolysis of RBCs to understand the association of NZs with their carrier cells. It was determined that a significant number of NZs were released into the solution upon hemolysis of RBCs, whereas NZs remained attached to the cell surface in case of non- hemolyzed RBCs (Figure 3d). Erythrocytes lose significant negative surface charge upon hemolysis, hence this release could be attributed to the compromised electrostatic interaction between NZs and RBCs upon lysis of the erythrocyte cells.⁴¹

Selective activation of pro-antibiotics for treatment of bacterial infections

After establishing the ability of RBC-NZs to selectively target pathogenic bacterial biofilms, we investigated their immunocompatibility. We initially studied whether RBC-hitchhiking prevents non-specific uptake of NZs by macrophage cells that are involved in clearance by reticuloendothelial system. RAW 264.7 macrophage cells were incubated with RBC-NZs for 24 hours, followed by washing and addition of pro-Res for 24 hours. It was observed that macrophages incubated with RBC-NZs exhibited minimal fluorescence, whereas macrophages incubated with free NZ 1 (Bare-NZ) showed strong punctate fluorescence due to activation of pro-fluorophore by macrophage uptaken nanozymes (SI Figure S5). These results show that non-hitchhiked NZs were readily endocytosed by macrophages. Whereas RBC-hitchhiked NZs were able to evade macrophage uptake, owing to the immune evasive capabilities of RBCs. Moreover, we quantified the Au content in macrophages using ICP-MS and determined that Bare-NZ showed high uptake in macrophages whereas RBC-NZs showed minimal uptake (Figure 3c, SI Figure S6). Having established that RBC-NZs can avoid non-specific uptake by macrophages, we next investigated their immunocompatibility in relation to inflammatory cytokine responses from macrophage RAW 264.7 cells. RBC-NZs demonstrated no significant tumor necrosis factor alpha (TNF-α) cytokine expression (SI Figure S10), indicating in vitro immunocompatibility with mammalian immune cells. These results further suggest that RBC hitchhiking can be used to selectively target pathogenic biofilms while maintaining high biocompatibility.

After establishing the localization of NZs at the site of pathogenic bacteria, we investigated the ability of RBC-NZs to selectively activate antibiotic-precursors and eradicate pathogenic bacterial biofilms. For this study, aryl azide protected moxifloxacin (pro-Mox, Figure 4e) was chosen as a model pro-antibiotic due to the high clinical relevance of moxifloxacin in the treatment of MDR infections.⁴² The synthetic protection of the secondary amine group on moxifloxacin inhibits binding with target bacterial enzymes, inhibiting their antimicrobial activity prior to activation.⁴³ Alamar Blue assays were performed on biofilms treated with RBC-NZs and pro-Mox to determine biofilm viability.

Figure 4.

Deprotection of antimicrobials in biofilms using RBC-hitchhiked nanozymes. RBC-NZ was used for selective activation of antibiotic prodrugs that decrease biofilm viability. a) methicillin-resistant S. aureus (MRSA, toxin producing) biofilms and b) P. aeruginosa (non-virulent) biofilms treated with pro-Mox and RBC-NZ (red bars) at 37 °C. Biofilms treated only with pro-Mox (blue bars) or with Mox (grey bars) were used in all experiments as negative and positive controls, respectively. Biofilm viability of c) hemolytic (pathogenic) and d) non-pathogenic bacterial strains after treatment with RBC-NZ and pro-Mox and moxifloxacin antibiotic alone. Each experiment was replicated five times. e) Scheme of activation of pro-Mox Nanozymes. Error bars represent standard deviations of these measurements. *p < 0.05, ***p < 0.001

We chose methicillin-resistant S. aureus (MRSA) and E. coli for biofilm viability studies due to their high clinical relevance and pathogenicity caused by α-hemolysin toxin secreted by both species.^17,18,36 Non-pathogenic bacterial strains of P. aeruginosa, B. sub were used as models strains to study the effect of RBC-NZs on non-hemolytic strains. Biofilms were incubated with RBC-NZs (500 nM) for 24 hours, washed and subsequently incubated with different concentrations of pro-Mox for 24 hours (more details are provided in methods section). Cells incubated with only pro-Mox and moxifloxacin antibiotics were used as negative and positive controls respectively (Figure 4, SI Figure S7, S8). It was observed that pro-Mox did not reduce biofilm viability in both pathogenic and non-pathogenic conditions. However, pro-Mox incubated with RBC-NZs led to reduced biofilm viability of pathogenic biofilms, whereas no significant antimicrobial activity was observed against non-pathogenic biofilms. These results indicate that selective accumulation of NZs in pathogenic biofilms enabled catalytic activation of pro-antibiotics, thereby increasing the specificity of the therapy. Moreover, moxifloxacin reduced bacterial viability of both pathogenic and non-pathogenic species, indicating the non-selective targeting of antibiotic treatment.

Conclusions

We present a strategy that integrates biomimetic nanozymes with RBCs. This strategy employs non immunogenic RBC-hitchhiked nanozymes to selectively eradicate pathogenic bacterial infections. Hemolysis by toxins secreted by pathogenic bacteria results in selective accumulation of and activation of nanozymes at the site of infection. These accumulated nanozymes subsequently uncage antibiotics at the disease site, providing in situ ‘drug factories’ that eradicate pre-formed biofilms without harming non-virulent bacterial species. This strategy should be particularly useful in treating recurring bacterial infections, including chronic wounds and medical device associated infections. Notably, majority of the state-of-the art studies either exhibit broad spectrum antibacterial activity, selective targeting of Gram negative/positive bacteria (irrespective of their pathology), or stimuli-responsive specific prodrug activation.^44,45 Here, we have generated a system to specifically target pathogenic bacteria without the need for designing microbe-specific probes that can selectively target pathogenic/virulent pathogens. Moreover, our strategy offers a promising alternative for overcoming delivery barriers in a complex biological environment, while maintaining the inherent catalytic activity of biorthogonal nanozymes.^46,47 These studies can further encourage researchers in the field of bioorthogonal nanozymes to utilize cellular interactions for multiple biomedical applications.

Taken together, combining RBCs with engineered nanozymes provides a platform with controlled and potentially “unlimited” therapeutic loading while circumventing the limitations associated with delivery of nanovehicles. The modular nature of this approach makes it suitable for numerous imaging and therapeutic applications for a diverse range of diseases.

New Concepts.

Infections caused by multidrug-resistant (MDR) bacteria and biofilms pose a serious global burden of mortality, causing thousands of deaths each year. In this study, we integrate synthetic nanomaterials with a biological platform, incorporating bioorthogonal nanozymes on the surface of red blood cell (RBCs) These functionalized RBCs perform abiotic chemical reactions, with targeting to pathogenic bacteria provided by lysis of the RBC by bacterial toxins. This biomimetic nanozyme platform shows potential to combat bacterial infections without harming the ecology of human microbiome, as well as circumvent the issues associated with non-specific uptake of nanoparticles by the immune system