Dual Affinity to RBCs and Target Cells (DART) Enhances Both Organ- and Cell Type-Targeting of Intravascular Nanocarriers

Abstract

A long-standing goal of nanomedicine is to improve a drug's benefit by loading it into a nanocarrier that homes solely to a specific target cell and organ. Unfortunately, nanocarriers usually end up with only a small percentage of the injected dose (% ID) in the target organ, due largely to clearance by the liver and spleen. Further, cell-type-specific targeting is rarely achieved without reducing target organ accumulation. To solve these problems, we introduce DART (dual affinity to RBCs and target cells), in which nanocarriers are conjugated to two affinity ligands, one binding red blood cells and one binding a target cell (here, pulmonary endothelial cells). DART nanocarriers first bind red blood cells and then transfer to the target organ's endothelial cells as the bound red blood cells squeeze through capillaries. We show that within minutes after intravascular injection in mice nearly 70% ID of DART nanocarriers accumulate in the target organ (lungs), more than doubling the % ID ceiling achieved by a multitude of prior technologies, finally achieving a majority % ID in a target organ. Humanized DART nanocarriers in ex vivo perfused human lungs recapitulate this phenomenon. Furthermore, DART enhances the selectivity of delivery to target endothelial cells over local phagocytes within the target organ by 6-fold. DART's marked improvement in both organ- and cell-type targeting may thus be helpful in localizing drugs for a multitude of medical applications.

Keywords: RBC hitchhiking; click chemistry; dual targeting; human lung delivery; liposomes; nanocarriers; vascular targeting.

Figure 1.. DART more than doubles the efficiency of organ-targeting compared to targeting via affinity-ligands-only and RBC-hitchhiking.

A. The goal mechanism of DART. DART liposomes possess two types of antibodies, one targeting RBCs (red) and one targeting endothelial cells (blue). In the first step (left panel), DART liposomes bind to RBCs, via DART liposomes’ RBC-targeting antibodies. Next (center panel), RBCs transit to the first downstream capillary bed, which for IV injections is the lungs. There the RBCs squeeze through narrow capillaries, increasing the probability of interaction between DART liposomes’ endothelial-targeting antibodies and endothelial epitopes. DART liposomes are designed to have many more endothelial-targeting antibodies than RBC-targeting antibodies, so the DART liposomes stay with the endothelium while the RBCs flow past (right panel). B. DART liposome components, including the two radiolabeling methods (DTPA-¹¹¹In & IgG-¹²⁵I); not to scale. C. Nomenclature for DART, predicate technologies, and controls. Liposomes can have 3 antibody combinations: RT = RBC-targeted antibody; ET = endothelial targeted antibody; DT = dual-targeted, which contain both RBC- and endothelial-targeted antibodies. There are 2 protocols of injection: “free” liposomes (e.g. DT liposome or RT or ET liposome) are injected without being exposed to RBCs; RBC-hitchhiking (RH) liposomes are first adsorbed onto RBCs, and the RBC-liposome complexes are then injected intravascularly. D. In vivo lung localization of the above liposomes and controls, measured by % injected dose (%ID) in the lungs at 30 min post IV-injection in mice. DART (DT-RH) liposomes achieved 650-fold higher levels than free drug (here, free DTPA-¹¹¹In), and >2x higher than a simple combination of ET + RH (ET-RH). Error bars are standard deviation, n>= 4 for all samples. Statistical differences exist at p<0.005 between DT-RH and all groups, and none between RT-L and RT-RH by 1-way Anova.

Figure 2.. Characterization of liposome to RBC binding by RBC/EC antibody conjugation ratio, coating density, and liposome concentration.

A. Agglutination of RBCs by %RBC Ab coating and # liposomes added. Round-bottom well assay demonstrates the effect of %RBC Ab on liposomes and their concentration in which aggregated RBCs appear diffuse and non-aggregated cells settle into a tight red dot. Image data demonstrate the effect of both the RBC Ab coating density (top left to right, 0-100%) and increased liposome numbers bound (left side top to bottom) affect aggregation of RBCs bound. Human RBCs are tested as a control. RBC samples within the black box define the Ab coating ratios and liposome binding concentration benign to RBC viability with respect to agglutination. B. Complement activation in vitro, as measured by C3a ELISA. To serum, liposomes +/− RBCs were added, and C3 was measured 10 minutes later. Both DT- RH and free DT liposomes had statistically equivalent complement activation to naïve serum. Cobra venom factor (CVF) is the standard positive control for C3 activation. N=2 biological replicates and 2 technical replicates. Comparisons were done with two-way ANOVA followed by Tukey’s post-test using Prism. P values: **< 0.01, ****< 0.0001. Error bars = SEM. C. ET/DT/RT liposome immunoreactivity shows liposome binding efficiency against a vast excess of RBC binding sites. The binding of 125I labeled liposomes to mouse RBCs was measured against % of mouse RBC Ab on the liposome surface (with the balance of Ab against ICAM). Binding efficiency increases nearly linearly until about 10% RBC mouse Ab, after which binding asymptotically approaches completion. Control binding against human RBCs (gray line) with the same particles demonstrates maximum potential adsorption of non-RBC-targeted liposomes at a given Ab coating. Mouse data N=3, Error = st. dev. D & E. RBC binding to ET/DT/RT ¹²⁵I labeled liposomes conjugated with EC Ab against ICAM (D) or PECAM (E). Liposome binding to RBC in vitro was measured against the ratio of RBC-to-EC targeting Ab on the liposome surface, with 200 total Ab/ liposome. Graph labels refer to ET= 100% EC targeting, DT= dual targeting at 10%/90% or 25%/75% RT/ET (D), 2.5%/97.5% or 10%/90% RT/ET (E), and RT= 100% RBC targeting. Error bars = st. dev., N= 3 F &G. Flow cytometry of RBC loaded with DT-RH liposomes (10%/90%, ICAM targeting) and ET-RH liposomes (100% ICAM targeting) and compared to control RBC. F. Flow cytometry was performed on RBC loaded with DT-RH liposomes (99.7% of RBC population binds liposomes) and ET-RH liposomes (73.8% of RBC population binds liposomes) and compared to control RBC. Insets. Fluorescence microscopy of RBC loaded with liposomes. G. Mean fluorescent intensity (MFI) quantification of the peaks shown in F indicate 43-fold increase of liposome signal in DT-RH vs ET-RH (ET-ICAM).

Figure 3.. DART liposomes rapidly localize to their lung targets and safely release the carrier RBCs.

A & B. Biodistribution of ¹²⁵I-liposomes (A) and their ⁵¹Cr carrier RBCs (B), using the endothelial-targeting antibody PECAM. Here we approximate concentration of the isotopes in the organ by plotting % ID per gram of tissue (% ID / g), which permits values >100% if an organ is < 1 g. A. DT liposomes (DART) achieved > 2x the lung uptake of ET liposomes. DT liposomes add just 5 RBC-targeting antibodies per liposome, keeping 195 PECAM-targeting antibodies (compared to ET liposomes that have 200 PECAM-targeting antibodies), p<0.001 by Student T-test. B. ⁵¹Cr-RBC of RT, ET, and DT liposomes all circulate equally (no statistically significant difference by Student t-test in ⁵¹Cr blood concentration) and show no statistically significant difference in lung retention. Inset in B displays the “transfer ratio,” defined as (liposome-to-RBC ratio in lung) ÷ (liposome-to-RBC ratio in blood), measured by their respective isotopes. The transfer ratio describes numerically the transfer of ¹²⁵I-liposomes from ⁵¹Cr RBCs to the target organ (lungs). C. Kinetics of DT-RH (DART) biodistribution of ¹²⁵I liposomes and ⁵¹Cr RBC (inset) at 2-20 min after IV injection. DT liposomes’ conjugated antibodies are at a ratio of 2.5% anti-RBC to 97.5% anti-PECAM (total 5 and 195 antibodies, respectively), as was used in A & B. D. Evaluation of DART targeting when the ratio of anti-RBC to anti-PECAM antibodies is increased from 2.5%/97.5% to 10%/90% (total 20 and 180 antibodies, respectively). The increase of the RT antibody from 2.5% (dark purple bars) to 10% (light purple bars) results in higher lung localization of liposomes (¹²⁵I). However, the 10%/90% liposomes (light purple) massively increase the number of carrier RBCs (⁵¹Cr) in the lungs. This excessive RBC trapping in the lungs is further quantified by the transfer ratios in the inset. Right: Transfer ratio of 2.5%/97.5% to 10%/90% is 610 to 120. Error bars = st. dev, n=4.

Figure 4.. DART liposome proof-of-concept in fresh, perfused, ex vivo human lungs.

A. Binding curves and agglutination tests of human RBCs and humanized DART liposomes. Liposomes were functionalized with (1) an endothelial-targeting (ET) antibody binding to human PECAM (the same target protein used in the above mouse studies) and (2) an RBC-targeting (RT) antibody. We compared two RBC surface targets, either binding to human GPA (left panel; the same target protein used in the above mouse studies) or Rh (right panel). The binding curves are shown for 5 different ratios of RT-to-ET antibodies (legend for the curves is top center). The insets show RBC agglutination assay results, varying the % RT antibody on the surface of the liposomes, and the # of liposomes per RBC. As with mice, there is a window of safety for these parameters at which no agglutination occurs. Error bars represent st. dev, N=2. B. Schematic of ex vivo lung perfusion (EVLP). The pulmonary artery is cannulated and perfused using a solution similar to that used in clinical-grade EVLP. The radiolabeled liposomes and RBCs are then injected into the pulmonary artery cannula, allowing a single-pass through the pulmonary capillaries, and then perfusion is continued for 10 minutes, with the perfusate (and radiolabeled material) collected via pulmonary vein efflux. C. Fresh human lung prepared for EVLP. Both right upper lobe (RUL) and right middle lobe (RML) bronchi were cannulated for inflation and oxygenation. The pulmonary artery was cannulated for perfusion. Green tissue dye was perfused to confirm adequate cannulation and perfusion through the vasculature with efflux seen leaving the pulmonary vein. D. Ex vivo human lungs were perfused using humanized DART liposomes. Liposomes were traced with ¹²⁵I and RBCs were traced with ⁵¹Cr. Of the initial injected dose, 27.5% remained in the lung tissue after perfusion compared to only 15.4% of the carrier RBCs.

Figure 5.. DART works with alternative targeting epitopes to efficiently and rapidly transfer nanocarriers to the target organ.

A & B, Liposomes were constructed similarly to Figure 3A, except the endothelial-targeted (ET) antibody employed was anti-ICAM here, instead of anti-PECAM. Liposomes were either RBC-targeted (RT; red), endothelial-targeted (ET; blue), or dual-targeted (DT; light and medium purple), with the DT liposomes containing either 10% or 25% RT antibody, and 90% or 75% anti-ICAM. These ¹²⁵I-labeled liposomes were loaded onto ⁵¹Cr-labeled RBCs via an RBC-hitchhiking (RH) protocol, IV-injected into mice, and 30 minutes later the mice were sacrificed for biodistribution analysis. (A) shows ¹²⁵I (liposomes), while (B) shows ⁵¹Cr (carrier RBCs). (A) Liposome accumulation in the target organ (lungs) is 2x higher with DT-RH (DART) liposomes (light purple striped) than ET liposomes. (B) Carrier RBCs are retained in circulation, not in the lungs. C. Delivery of “free” liposomes (non-RBC bound) identical to ET/DT/RT-RH shown in A. D. Pharmacokinetics of DART and related controls when the ET antibody targets ICAM and not PECAM. This compares DART condition, (DT, row 1), with controls ET (row 2) and RT (row 3). It also compares delivery by RBC-hitchhiking (RH; columns 1 & 2) versus direct injection of each “free” liposome (column 3). The inset table on the left side of D describes the % of each antibody used, with the “EC antibody” being anti-ICAM. Y-axis is (%ID/g) and is the same scale for all plots. Most notably, the top left plot shows DT-RH (DART) liposomes are rapidly transferred to the target organ and remain there, while the carrier RBCs (top middle plot) leave the lung overtime. N= 4, error bars are st. dev.

Figure 6.. DART improves cell-type-targeting.

Fluorescent dual-targeted (DT) liposomes were constructed that contained 10% anti-GPA and 90% anti-ICAM antibodies. They were injected either directly (“free” liposomes) or via the RBC-hitchhiking (RH) protocol, and mice were sacrificed 30 min later for flow cytometry on single cell suspensions of the lung (A, B, D, E) or lung histology (C). A. Dot-plot displaying how cell types were determined by CD31 and CD45 positivity. B. Liposome positivity among various cell types. C. Fluorescence micrographs indicating association of liposomes (red) with either endothelial cells (left) or leukocytes (right) in the lung after circulation for 30 minutes. D. Quantification of flow cytometry data by cell type and liposome positivity. The left column is endothelial cells, and the right is leukocytes. The top row is from mice that had DT-liposomes injected directly (“free” liposomes), and the bottom row is from mice that received DT-RH (DART) liposomes. The lighter colored wedge in each pie chart shows the fraction of liposome+ cells. DT-RH results in a 20-fold increase in endothelial cell targeting and near 4-fold increase in leukocyte targeting. E. Analysis of cell localization of DT-RH vs freely injected liposomes. Green bar indicates the increase in endothelial cell localization by fold increase from free liposome injection to DT-RH (DART), and the aqua bar an equivalent calculation of leukocyte localization. Thus, DART produces a 5.7-fold increase in selectivity for endothelial cells vs leukocytes. For flow cytometry, N= 2 biological replicates and 2 technical replicates.

Figure 7.. Evaluation of potential side effects of DT-RH in vivo.

A. In vivo evaluation of complement activation / opsonization after injection of DART liposomes or controls. Liposomes were IV-injected and 10 minutes later serum was drawn and measured for complement activation by C3a ELISA. Cobra venom factor (CVF) is a positive control, inducing C3 cleavage to release C3a. All the liposome formulations lacked significant difference from naive (no injection) controls, including DART liposomes (DT-RH). B. Cardiopulmonary physiology of mice 24 hours after IV-injection of DART liposomes and controls. DART liposomes caused no cardiopulmonary changes compared to naive mice, in any of the following: blood oxygenation (measured by oxygen saturation of blood, SpO2), breathing rate, and heart rate. One of the control liposomes (ET liposomes) displayed slightly increased heart rate. C. Measurement of RBC count, hemoglobin, and hematocrit at 24 hours after IV-injection of DARTs and controls. All liposome injections, including DART and control liposomes, led to small but statistically significant decreases in these parameters. D. White blood cell counts (WBC), measured at the same time as in C, showed none of the liposomes changed total WBC or any of the subsets of lymphocytes, monocytes, and neutrophils. E. Platelet counts were also unchanged. F, G. Mice IV-injected with DART liposomes or controls were sacrificed and their lungs removed for histology (H&E staining). As a positive control, separate mice underwent intra-tracheal acid-aspiration, since it is known to induce “RBC aggregates” that represent hemorrhages and clots in the lungs, and those were the two pathologies which most needed investigation for RBC-related nanocarrier delivery. G displays representative images, which show DART-liposome-injected mice had lung histology indistinguishable from naive mice. By contrast, the positive control (acid aspiration) shows multiple RBC aggregates. Blinded observers quantified RBC-aggregate frequency (number of fields in which these were detected), defined as localized collections of RBCs. As shown in F, DART liposomes (DT-RH) and controls all showed significantly less hemorrhages than the positive control, and DART was indistinguishable from the control liposomes. Statistics: A-C: N=3 biological replicates with 2 technical replicates. D, E: N=3 biological replicates. All comparisons were done with two-way ANOVA followed by Tukey’s post-test using Prism; error bars = SEM. * p < 0.05; ** p < 0.01; *** p < 0.001; **** p < 0.0001. F: N=3 biological replicates, 3 slides per replicate, 15 distinct fields analyzed per slide.