Morphing low-affinity ligands into high-avidity nanoparticles by thermally triggered self-assembly of a genetically encoded polymer

Abstract

Multivalency is the increase in avidity resulting from the simultaneous interaction of multiple ligands with multiple receptors. This phenomenon, seen in antibody-antigen and virus-cell membrane interactions, is useful in designing bioinspired materials for targeted delivery of drugs or imaging agents. While increased avidity offered by multivalent targeting is attractive, it can also promote nonspecific receptor interaction in nontarget tissues, reducing the effectiveness of multivalent targeting. Here, we present a thermal targeting strategy--dynamic affinity modulation (DAM)--using elastin-like polypeptide diblock copolymers (ELP(BC)s) that self-assemble from a low-affinity to high-avidity state by a tunable thermal "switch", thereby restricting activity to the desired site of action. We used an in vitro cell binding assay to investigate the effect of the thermally triggered self-assembly of these ELP(BC)s on their receptor-mediated binding and cellular uptake. The data presented herein show that (1) ligand presentation does not disrupt ELP(BC) self-assembly; (2) both multivalent ligand presentation and upregulated receptor expression are needed for receptor-mediated interaction; (3) increased size of the hydrophobic segment of the block copolymer promotes multivalent interaction with membrane receptors, potentially due to changes in the nanoscale architecture of the micelle; and (4) nanoscale presentation of the ligand is important, as presentation of the ligand by micrometer-sized aggregates of an ELP showed a low level of binding/uptake by receptor-positive cells compared to its presentation on the corona of a micelle. These data validate the concept of thermally triggered DAM and provide rational design parameters for future applications of this technology for targeted drug delivery.

Figure 1

(a) Schema of DAM via temperature triggered self-assembly of an ELP_BC. At T < CMT, ELP_BC exist as soluble unimers and lead to monovalent ligand presentation. At T > CMT, the ELP_BC unimers self-assemble into micelles following desolvation and collapse of the hydrophobic block. This leads to multivalent ligand presentation in the corona of the micelle.(b) The ELP_BCs incorporate an RGD peptide ligand at the hydrophilic terminus and a cysteine residue for conjugation of fluorophores (or drugs) at the hydrophobic terminus. The ligand-negative, control ELP_BC does not contain the terminal RGD ligand but includes the C-terminal cysteine.(c) SDS-PAGE of purified RGD-ELP_BCs (left) and parent ELP_BCs (right) yields a thick band corresponding to monodisperse purified protein, showing that ELP_BC can be purified by ITC.

Figure 2

Dynamic, thermally triggered self-assembly of RGD-ELP_BCs and ELP_BCs. Hydrodynamic radius (R_h) and molecular weight (MW) of the ELP_BCs were measured as a function of temperature by DLS.(a) Both ELP_BCs exhibit distinct and stable unimer and micelle regions as a function of solution temperature. The temperature at which the unimer to micelle transition occurs is defined as the CMT (dashed line). (b) Presentation of the RGD peptide ligand on the hydrophilic terminus of the ELP_BC altered the self-assembly properties of the ELP-64/90 less than the ELP-96/60 or ELP-64/60 constructs. The resulting terminus density was nearly identical for all ligand and non-ligand constructs.

Figure 3

Flow cytometry analysis of K562 and K562/αvβ3 cells following incubation with 10 µM RGD-ELP-64/90 or ELP-64/90 below (dashed line) and above (solid line) the CMT. There was no significant difference in the histograms of any of the cell populations incubated below the CMT. Neither cell line showed enhanced binding/uptake of the ELP_BCs above their CMT, as seen by their similar flow cytometry histograms (panels IV–VI; X–XII), and the RGD-ELP_BCs did not show appreciable interaction with K562 cells above the CMT (panels VII–IX). There was a slight increase in binding/uptake of RGD-ELP-96/60 above its CMT by K562/α_vβ₃ cells, but there was a dramatic increase in the fraction of K562/α_vβ₃ cells that take up RGD-ELP-64/60 and RGD-ELP-64/90 above their CMT (panels I–III). This of this second peak increasingly shifted right with increasing SR of each ELP_BC indicating greater levels of interaction per cell.

Figure 4

Analysis of segment ratio (SR) on cellular binding/uptake.(a) Approximately 60% of K562/α_vβ₃ were positive for the α_vβ₃ receptor as seen by the binding of the LM609 antibody that is specific for the α_vβ₃ integrin. The percentage of AlexaFluor488⁺ cells increased to 50%–60% relative to unheated controls when RGD-ELP-64/90 and RGD-ELP-64/60 were incubated with K562/α_vβ₃, similar to the % of α_vβ₃ cells. There was no significant increase in % AlexaFluor488+ cells with heating for any other combination of construct and cells. (b) The fold-increase in median fluorescence of AlexaFluor488+ cells was measured for each cell/construct combination. There was a small increase in fluorescence of both cell lines incubated with the parent ELP_BCs and of K562 cells that were incubated with RGD-ELP_BCs. The median fluorescence of α_vβ₃increased with RGD-ELP-64/60 and RGD-ELP-64/90 above their CMT, while there was a slight increase in binding of RGD-ELP-150 and RGD-ELP-96/60 by α_vβ₃. The median fluorescence also increased with increasing SR of the RGD-ELP_BCs.

Figure 5

Confocal fluorescence images of K562/α_vβ₃ cells following incubation with 10 µM of various ELP_BCs (green).(a) There was no visible binding of any of the ELP_BC by K562/α_vβ₃ cells when they were incubated below and above the CMT with the ligand-negative ELP_BC controls, demonstrating that micelle formation alone does not promote nonspecific interaction. (b) There was also minimal visible binding/uptake of all three RGD-ELP_BC below or above the CMT when incubated with K562 cells, showing that overexpression of the receptor is necessary for enhanced interaction. (c) There was no visible binding/uptake of RGD-ELP-64/60 or RGD-ELP-64/90 below the CMT by K562/α_vβ₃ cells, but there was significant binding/uptake of RGD-ELP-64/60 or RGD-ELP-64/90 above their CMT. A binary population of highly fluorescent and non-fluorescent cells in the field of view was observed by fluorescence microscopy, corresponding to the bimodal distribution seen in the flow cytometry histogram in panels II and III in Fig. 3. Size bar = 50 µM.