Programmable DNA-Augmented Hydrogels for Controlled Activation of Human Lymphocytes

Abstract

Contractile forces within the planar interface between T cell and antigen-presenting surface mechanically stimulate T cell receptors (TCR) in the mature immune synapses. However, the origin of mechanical stimulation during the initial, i.e., presynaptic, microvilli-based TCR activation in the course of immune surveillance remains unknown and new tools to help address this problem are needed. In this work, we develop nucleic acid nanoassembly (NAN)-based technology for functionalization of hydrogels using isothermal toehold-mediated reassociation of RNA/DNA heteroduplexes. Resulting platform allows for regulation with NAN linkers of 3D force momentum along the TCR mechanical axis, whereas hydrogels contribute to modulation of 2D shear modulus. By utilizing different lengths of NAN linkers conjugated to polyacrylamide gels of different shear moduli, we demonstrate an efficient capture of human T lymphocytes and tunable activation of TCR, as confirmed by T-cell spreading and pY foci.

Graphical Abstract

We enhance activation of primary human T cells by increasing sensitivity of TCR complex to the mechanical signals provided by biomimetic antigen. We achieve this effect by using nucleic acid nanoassemblies for tethering antigen to the presenting surface. Our results highlight the important role for nanoscale configuration of mechanical forces within the TCR-pMHC complex during antigen presentation.

INTRODUCTION

T cells play a central role in allergies^1–3, autoimmune diseases^4,5, infections⁶, and cancers^7–11 orchestrating immune response with the molecular machinery consolidated around the mechanosensitive T-cell receptor (TCR)^12–15. TCR activates formation of the immune synapse (IS) upon recognition of cognate peptide-loaded major histocompatibility complex (pMHC)^15–18 on the surface of antigen-presenting cell (APC). Most studies of mechanosensitive TCR activation use a simplified model of the IS with the T cell-APC interactions are approximated to a 2D plane that features tangential actomyosin-driven mechanostimulatory tractions^19–21, non-physiological levels of cell stimulation with cytokines (e.g., IL-2)²², and nominal contributions of non-actomyosin cytoskeleton subsystems²³.

In the contractility-centered IS model, TCR activation requires prerequisite coalescence of solid state²⁴ signalosomes^25–27 that appear through F-actin microcluster-enriched foci²⁸. These TCR foci initiate the antigen-induced signaling at the tips of nanoscale membranous protrusions, called microvilli, that are responsible for the initial T-cell-APC engagement^29,30 during the IS development (Figure 1A–1) and later for signaling in the mature IS^28,31. The TCR-pMHC complex receives mechanosensitive^12,15,32 stimulation from the mature IS cytoskeletal centripetal forces, facilitated by non-muscle myosin 2A contractility^20,33 and F-actin polymerization³⁴ (Figure 1A–2,3). The same set of forces triggers downstream CasL, Zap70, and Nck-Lck signaling^20,33 (Figure 1A–3) resulting in the activation of T cells.

Figure 1. Artificial immune synapse with an introduced DNA-based mechanics-controlling module.

(A) T-cell’s cognate TCR-pMHC axis: (1) Human primary T cells engage with their cognate antigen peptides, exposed on antigen-presenting cells (APC) via TCR-decorated microvilli^29,30. The antigen-engaged microvilli secure early cognate T-cell-APC interactions and facilitate the formation of the mature immune synapse. (2) TCR-pMHC interactions within the IS form a 13 nm-long molecular axis between APC and T-cell surfaces. (3) T-cell cytoskeletal centripetal flow and actomyosin contractility within the IS exert shear stresses onto the TCR that results in the mechanical 3D force momentum (torque) and bending of the TCR-pMHC axis. 3D force momentum (torque), along with the shear and possibly other forces, facilitate the mechanosensory activation of the TCR downstream signaling via the Nck-Lck-Zap70 pathway. Note that the interaction energy between T-cell and antigen is modulated by the „catch-bond’ mechanism within the TCR-pMHC axis in addition to TCR downstream mechanosensing signaling. (B) CD3ε TCR complex subunit is a relay for the mechanical forces from the extracellular TCR-pMHC axis to the intracellular TCR signaling apparatus. Bypassing the „catch-bond’-modulated variability within the TCR-pMHC axis via the CD3ε subunit targeted by the OKT3 mAb allows for standardized studying of the TCR’s downstream torque mechanosensing. Note the introduction of the additional controlling module as the bundled dsDNA between the elastic polyacrylamide gel surface and the TCR complex.

However, TCR activation in the presynaptic phase happens through the glycocalyx coat between antigen-presenting surfaces²⁹ or through the cell surface with invadosomes³⁵, allowing for coalescence of antigen chemo- and mechanosensing that is necessary for the stimulation of T cells^28,35–37. The F-actin microcluster forces are similar to focal 3D deformation forces of F-actin “comet tails”³⁸ that generate derivative momentums^39–42 and membrane deformations^40,43–46 but not in-plane shear traction forces^31,47–50. Moreover, the activation of TCR in microvilli happens in the presence of the actomyosin contractility inhibitor Blebbistatin^51,52, thus suggesting that F-actin microclusters generate enough signals for TCR activation via the contractility-independent focal polymerization of actin.

We hypothesize that actomyosin contractility is less important during the initial microvilli-based TCR activation^29,30,53. Instead, activation can be provided by the applied 3D force momentum of polymerizing actin (also “torque” or “TCR bending”) that is generated during microcluster deformation. The hypothesis of TCR activation by focal polymerization of 3D actin (e.g., by the “comet tail” F-actin microclusters) could explain the early stages of IS formation^20,28,54, presynaptic TCR activation by antigen-decorated plastic beads^47,49,50,55 and spatially constrained adhesions^36,37 that do not support formation of the conventional planar IS. TCR activation by focal actin polymerization could also explain the extreme lymphocyte activation by complex 3D objects, such as carbon nanotubes, that induce much higher focal stresses within polymerizing 3D F-actin networks compared to the 2D shear modulus forces induced by the planar IS^22,56–58.

No currently available methods - elastic biomimetic gels¹², micro/nanopillars^59–61, atomic force microscopy¹⁵, DNA^25,62,63, and red blood cell probes³⁶, can control the mechanical resistance at the level of individual foci. To test the role of 3D force momentum in force-dependent TCR activation^{12,28,31,64–67}, we set out to develop a new approach for the functionalization of hydrogels with programmable nucleic acid nanoassemblies (NANs) that can be facilitated as linkers for regulated TCR binding and activation. The predictable and regulated base pairing, together with an expanding library of structural and interacting motifs and widely accessible computational tools allow for engineering nucleic acid nanodevices and nanoassemblies with controlled structural features, physicochemical properties, and biological profiles^68–71. Using NANs, we decorate polyacrylamide gels (PAG, G’=50 kPa, shear modulus⁷²) with OKT3⁷³, a mouse monoclonal antibody specific to the CD3ε subunit of the human TCR, that binds the CD3ε subunit and subsequently activates the TCR under the mechanical load¹⁵. Using OKT3-NANs-PAG surfaces, we demonstrate that NAN linkers of various lengths modulate the reaction of T cells to mechanical stresses transmitted via the PAG-NANs-OKT3-CD3ε axis. We also show that this system activates the TCR more effectively if both 3D force momentum and 2D shear force signals are present compared to the 2D shear force signals of conventional elastic substrates.

METHODS

All methods are detailed in Supporting Information.

Cell culture:

Human CD4+ T cells were isolated from commercially available whole human blood (STEMCELL Technologies Inc., USA) with EasySep Human CD4+ T Cell Isolation Kit (STEMCELL Technologies Inc., USA). Unless otherwise indicated, cells were cultured, activated and expanded in ImmunoCult-XF T Cell Expansion Medium (STEMCELL Technologies Inc., USA) with the addition of ImmunoCult Human CD3/CD28/CD2 T Cell Activator and Human Recombinant Interleukin 2 (IL-2, STEMCELL Technologies Inc., USA) as per STEMCELL Technologies Inc. commercial protocol.

Support fabrication:

Polyacrylamide gels (PAG) with micropatterned streptavidin lanes are manufactured as described elsewhere^72,74. Briefly, a clean glass surface is microprinted with fluorescently labeled streptavidin-acrylamide conjugates via microcontact printing. The use of a hard PDMS (hPDMS⁷⁵) composite stamp prevents patterns’ collapse onto the printed surface^37,76. PAG premix with a desired shear modulus (G’)⁷² is polymerized in a “sandwich” manner. The polymerized “sandwich” is then hypotonically treated and the “intermediate” glass surface is gently peeled off the PAG layer, leaving the fluorescent pattern crosslinked to the PAG surface.

Preparation of NANs:

All sequences used in this project have been previously designed for the conditional activation of split functionalities intracellularly^77,78 and their biotinylated analogs can be purchased (IDT). All sequences and protocols for heteroduplex formations are listed in the Supporting Information.

PAG conjugation with NAN-linkers:

The streptavidin-micropatterned PAG are incubated with various biotin-NAN conjugates for further OKT3 adsorption. Similarly, for more direct OKT3 crosslinking to the PAG surface via biotin-streptavidin or direct PAG crosslinking, the molecularly conjugated fluorescent OKT3 is printed on the glass surfaces, then cross-linked to the polymerizing PAG in a “sandwich” manner^64,74.

Imaging:

Cell fixation, Alexa Fluor® 488 Anti-Phosphotyrosine, total myosin IIa and phalloidin-iFluor 488 or 647 labelling are performed according to the commercial protocol. Cell immunofluorescent labeling and confocal 3D imaging performed as described in the previous report³⁷. Cell samples are mounted in 90% Glycerol (Sigma-Aldrich) in PBS (KDMedical) with 1 μg/mL Hoechst 33258 and 0.01% sodium azide preservative (Sigma-Aldrich) to prevent bacterial contamination and degradation of samples over time. The instant structured illumination microscopy (iSIM) is performed using a custom-build microscope (VisiTech).

RESULTS

Identification of 3D force momentum using different linkers for CD3ε-mediated TCR activation

Early antigen sensing and TCR activation at the nanoscale microvilli tips starts before the formation of the planar tensile IS. It indicates that traction force-independent non-contractile mechanisms provide the initial mechanical stimulation of TCR during the microvilli-based antigen surveillance and TCR engagement (Figure 2A–1,2,3). To test this hypothesis, we presented freshly isolated CD4+ T cells to the OKT3 dots. Both contractile (+DMSO, Figure 2A–4) and non-contractile (+Blebbistatin, 50 μM, Figure 2A–5) human T cells quickly engage multiple distant OKT3 dots via microvilli which is accompanied structural thickening of microvilli after 10 minutes. On T-cell microvilli tips, growing TCR microclusters can accumulate innate deformational stresses and 3D bending forces, that provide the mechanical stimulation for the TCR complex (Figure 2B–1) whereas mature immune synapse can also feature 2D tangential shear stresses driven by actomyosin contractility and lamellipodial polymerization-depolymerization turnover dynamics (Figure 2B–2). However, due to the spatial and time constraints, OKT3 dots do not support the formation of a mature tensile IS.

Figure 2. Presynaptic antigen surveillance and engagement of T cells via microvilli TCR foci.

(A) TCR foci-decorated membranous linkers, i.e., microvilli, facilitate antigen surveillance of T cells with a larger pMHC scanning radius, but require non-contractile pre-synaptic force generation and pMHC mechanosensing mechanisms. Laser confocal microscopy. Freshly isolated, human naïve (resting) CD4+ T-cells (green, myosin IIa) are presented to the OKT3 dots micropattern (red), printed on the plasma treated glass (1–3). Contractile (+DMSO) and non-contractile (+Blebbistatin, 50 μM) spherical human T cells (4–5) quickly engage multiple OKT3 dots (1 μm each with linear distances between dot centers equal 5 μm). T cell engagement happens without prerequisite T-cell spreading and retraction (short term, 10 minutes), and without external mechanical stimuli (no vibration, no hydrodynamic flow). Quick engagement of the multiple distant OKT3 dots occurs via microvilli, accompanied with their immediate (10 minutes) structural thickening. (B) For T-cell microvilli tips, growing TCR microclusters accumulate innate deformational stresses and 3D bending forces, that provide the mechanical stimulation for TCR complex (1). Mature immune synapse features both TCR foci-driven torques and actomyosin contractility- and lamellipodial polymerization-depolymerization turnover dynamics-driven 2D tangential shear stresses (2). However, due to the experimental constraints, OKT3 dots do not support the formation of a planar tensile IS.

Since we target the CD3ε subunit of the TCR with an anti-CD3ε monoclonal antibody (OKT3, Figure 1B–1,2), the CD3ε subunit acts as the TCR complex’s mechanotransducer relay module that transmits the applied forces from the TCR-pMHC axis to the intracellular TCR signalosome^79–81. We hypothesized that the mechanical stimuli applied to CD3ε via OKT3 binding would allow us to study the T-cell’s perception of mechanical stimuli under precisely controlled conditions (Figure 1B–2), e.g., bypassing highly variable TCR-pMHC interactions^12,36,82.

Previous studies indicate that along with pulling-tension forces^15,62, the TCR detects force momentum along each axis of the individual TCR-pMHC complexes^14,47,55. Indeed, IS shear stresses induce the TCR drag that translates into secondary shear-induced force momentum (Figure 1B–3). However, a common simplification of the IS interface to a singular plane reduces the TCR-pMHC axis to the material point with shear stresses inducing only a symmetric reaction without the resulting momentum^12,65. In this approximation, the force momentum can only be induced directly by the IS off-plane deformation forces. Conversely, a detailed T-cell-APC interaction model stipulates a shear-induced TCR-pMHC axis force momentum (Figure 1B–3) along with the momentum that is induced by the IS deformation forces⁵², effectively rendering force momentum as one of the most common mechanostimulatory signals during T-cell activation.

To test this hypothesis, we assess the effects of various linker lengths (Figure 3A) on the spreading of T cells along micrometer-wide OKT3 lanes. For “long” linkers, we utilize biotinylated OKT3 and streptavidin-PAG conjugates that consist of streptavidin (~5 nm), biotin-NMS (~1.4 nm), and OKT3 (~15 nm)^83,84 accruing to the final length of ~22nm (Figure 3A–1). The “long” OKT3-biotin-streptavidin-acrylamide configuration results in poor T-cell spreading (Figure 3B,C, long linker), in agreement with previous reports⁶⁵. Alternatively, “short” molecular linkers (~1 nm), made with OKT3 directly conjugated to PAG (Figure 3A–2), induce much stronger T-cell spreading (Figure 3B,C, short linker). The difference in T-cell responses is likely caused by either steric TCR-igniting exclusion of the CD45 molecule from the site of TCR-agonist cognate interaction^27,85–88, or by the variation in the bending resistance of the OKT3-PAG axis, as both “long” and “short” linkers are attached to PAG of the same shear modulus and with comparable surface density of OKT3, as assessed by the relative brightness of all lanes in Figure 3C. Strikingly, when OKT3 antibodies are immobilized on the plasma-activated glass (Figure 3A–3), the strongest T-cell spreading (Figure 3B,C, Glass-OKT3) is observed, thus confirming the importance of linker length and origin of the support.

Figure 3. 1D spreading of human CD4+ T-cells depends on lengths of OKT3-PAG linkers.

(A) Schematic views of OKT3 linkage to the PAG substrate. (B) Comparison of 1D T-cell spreading lengths during the course of one hour along the OKT3 attached via long linker, short linker, or direct adsorption onto a plasma-treated activated glass surface (C) Fluorescent micrographs (iSIM) of the hCD4+ T-cells (blue/green) spreading along OKT3 lanes (red).

Designing a platform with tunable force resistance

To further assess the contribution of different antigen immobilization methods on force-dependent TCR activation, we utilize this newly developed platform with tunable force resistance. We take advantage of double-biotinylated dsDNAs as modular linkers between the streptavidin-micropatterned PAG and OKT3 antibodies. We compare three possible protocols of double-biotinylated DNA duplex formation: (i) one-step introduction of pre-made double-biotinylated dsDNAs, (ii) step-wise hybridization between single-stranded biotinylated DNAs, and (iii) step-wise toehold-mediated re-association of RNA/DNA heteroduplexes⁷⁷ (Figure 4 and Figures S1–2).

Figure 4. Multi-step strategy for PAG surface micropatterning with NANs.

(A) Fluorescently labeled streptavidin-acrylamide conjugate micropattern is imprinted into polymerized bulk PAG from intermediate glass surfaces. (1) Cured gel is then incubated with biotin-5’-DNA/RNA heteroduplex. (2) Passivation of remaining streptavidin reactivity: open streptavidin biotin-binding sites are blocked with three-fold excess free biotin, then (3) rinsed from non-reacted biotin. (4–5) Isothermic reassociation of PAG-biotin-dsDNA-biotin linker by addition of cognate biotin-5’-DNA/RNA heteroduplex, followed by rinsing of the dsRNA byproducts. Resulting dsDNA linker exposes a free and available biotinylated distal end, without self-locking onto the PAG-streptavidin surface. (6–7) dsDNA-biotin linker functionalized with fluorescently labeled streptavidin-OKT3. (8) Biotinylated dsDNA and streptavidin-OKT3 conjugates self-assemble into the tent-like molecular nanopillars - a rigid multipod nanostructure that facilitates T-cell adhesion via CD3ε-OKT3 interaction. (B) (Top left) - Micrographic views of fluorescent streptavidin lanes, imprinted into the PAG surface. (Bottom left) - Streptavidin-conjugated, fluorescent OKT3 (Alexa Fluor® 555), adsorbed onto streptavidin-biotin-dsDNA-biotin PAG lanes (G’=2.3 kPa). (Right) - Human CD4+ T cell spread along the resulting OKT3 lane (G’=50 kPa).

Prior to PAG functionalization tests, we assess the relative kinetics of dsDNA linker formations using streptavidin-decorated quantum dots (QDs). Upon the formation of double-biotinylated duplexes using any of the assembly strategies, the crosslinking of QDs results in QD lattice formations⁸⁹ which alter the electrophoretic mobility of the samples and present as aggregation when visualized via electrophoretic mobility shift assays. While the use of double-biotinylated dsDNAs has the fastest kinetics of QD lattice formation, its direct introduction to streptavidin-coated PAG passivates all available biotins (Figure S1D), thus preventing any further conjugation for streptavidin-OKT3. Among the other two protocols, the reassocication of RNA/DNA heteroduplexes results in a complete QD lattice formation after 30 minutes of incubation, while the ssDNA approach still shows some incomplete assemblies even after 2 hours of incubation at 37 °C. This difference can be explained by the potential formation of secondary structures within ssDNAs that delay their annealing to complementary strands, while this possibility is absent in the case of RNA/DNA heteroduplexes. Therefore, we establish the toehold-mediated reassociation of RNA/DNA heteroduplexes to be an optimal strategy for all future studies with OKT3-NAN-PAG systems. Importantly, the same strategy can be applied for longer RNA/DNA heteroduplexes with either continuous or segmented RNAs in their composition (Figure S2).

The protocol of isothermal reassociation of DNA/RNA heteroduplexes demonstrates best results (Figure 4B) and allows for a tighter control over the OKT3-NAN-PAG system by following four simple steps (Figure 4A): (i) conjugation of biotinylated RNA/DNA heteroduplexes to streptavidin-coated PAG surfaces (step 1), (ii) passivation of unoccupied streptavidin with free biotin (steps 2 and 3), (iii) introduction of the distal biotin group via cognate DNA/RNA heteroduplexes’ reassociation (steps 4 and 5), and (iv) conjugation of OKT3 antibodies (steps 6 and 7). Passivation of the unoccupied streptavidin with free biotin on the PAG surface prior to the introduction of the distal biotin group (steps 2 and 3) prevents its locking onto the PAG-streptavidin surface in the “pull handle” configuration (Figure S1D–1). We verify this notion by comparing the efficiencies of all protocols through measurements of the fluorescent OKT3-streptavidin adsorption onto microlanes (Figure S1). The results indicate a significant drop in [OKT3-streptavidin] - [biotin-dsDNA-biotin] - [streptavidin-PAG] formation when assembled via a single-step protocol, hence confirming a significant loss of accessible distal biotin due to its locking onto the streptavidin-PAG surfaces.

Cooperation of 3D Force Momentum and 2D Shear Force in the Activation of T cells

The demonstrated effect of the linker length on TCR activation (Figure 3) can result from sterically driven kinetic segregation (KS) between the TCR and CD45 receptors^27,85–88. In the KS model, an elongated and rigid extracellular domain of the CD45 receptor (ECD-CD45) is sterically segregated from areas of tight contact between the T-cell and APC. The 13 nm-long axis of TCR-pMHC adhesion results in an intercellular gap clearance that sterically excludes 21 nm-long ECD-CD45. This spatial separation between receptor complexes modulates TCR activation⁸⁶. However, the KS model has been largely unverified for early TCR activation⁹⁰. To assess the potential contribution of KS, we evaluate the final clearance for all our experimentally tested linkers. In our systems, we bypass the TCR receptor subunit and directly target the 3 nm-long extracellular portion of the CD3ε subunit²⁷ with the 15 nm-long OKT3 antibody, resulting in the minimal clearance of 18 nm (Figure 5A). The final steric clearances after the addition of NANs (39, 66 and 93 bp) and biotin-streptavidin linkers for OKT3 and PAG accrue to lengths that exceed that of ECD-CD45 (Figure 5A), confirming that TCR activation effects are independent from KS. Moreover, the strong agonistic effect of the anti-CD3ε OKT3 antibody activates TCR irrespectively to TCR-CD45 segregation^91,92.

Figure 5. 1D spreading of human CD4+ T cells recorded as a function of the TCR-OKT3 molecular torque resistance.

(A) Comparison of OKT3-PAG linker-facilitated clearance for 21 nm-long CD45 extracellular domain length (ECD-CD45). All linker components are shown to scale: streptavidin - 5 nm, OKT3 antibody molecule - 15 nm, 39 bp dsDNA - 13.3 nm, 66 bp dsDNA - 22.5 nm, 93 bp dsDNA - 31.6 nm, biotin and acryloyl links - below 1 nm. The OKT3-acryloyl linker system accrues to the length of ~15 nm, which when linked to the 3 nm-long extracellular CD3ε domain, results in ~18 nm clearance (below ECD-CD45 length). All other OKT3 linker configurations exceed the ECD-CD45 length, allowing for the CD45 clearance, and thus independence from TCR-CD45 kinetic segregation (KS). (B) Three PAG surfaces with G’ of 2.3, 8.6, and 50 kPa are tested and each G’ value features either long linker (Fig. 2A, low torque resistance), or short linker (Fig. 2A, intermediate torque resistance), or multiple lengths dsDNA linkers (high torque-resistance). Graphs show T-cell spreading lengths measured along the 1 μm-wide OKT3 lanes and corresponding brightness of OKT3 lanes proportional to the relative OKT3 surface concentrations. High torque-resistance 39 bp dsDNA nanopillar within the OKT3-PAG system increases TCR activation at G’=8.6 kPa when compared to the lower torque resistance linkers for the G’. Overall early TCR activation-induced 1D T-cell spreading is reduced with increased dsDNA linker lengths and overall molecular flexibility. (C) 3D micrographic views of the human CD4+ T cells spreading along OKT3-decorated PAG lanes with various G’ and different molecular linker systems. (D) T-cell 1D spreading along PAG lanes with 66, 81, and 93 bp dsDNA linkers decorated with OKT3.

Considering the dsDNAs as elastically flexible rods⁹³, the NANs’ lengths are inversely proportional to the mechanical bending resistance (Figure 1B). Therefore, we examine the effects of NAN length (by testing 39, 66, 81, and 93 bp linkers) and PAG shear modulus (by testing 2.3, 8.6, and 50 kPa PAGs) on T-cell adhesion and spreading along OKT3-NAN-PAG microlanes. We also compare NANs to polymer linkers (short and long, Figure 3) attached to the same shear moduli PAG. When 2.3 kPa PAG is used, the T-cell spreading is comparable between all types of linkers. For 50 kPa PAG support, the T-cell spreading is equally higher for “short” and dsDNA linkers when compared to “long” linkers. However, we identify that by keeping bearing PAG at a shear modulus of G’=8.6 kPa, the significant difference in T-cell responses as a function of the dsDNA length is clearly observed (Figure 5B,C). These results indicate that, despite the identical mechanical rigidity of gels and OKT3 density, T-cells develop significantly stronger responses (assessed by 1D spreading/early activation) on the microlanes decorated with 39 bp dsDNAs (Figure 5 B,C). For the same linker used in combination with different shear moduli PAGs, the spreading increases with a higher shear modulus. Thus, we conclude that along with in-plane shear stress resistance, TCR-antigen sensing is strongly modulated by the CD3ε-OKT3-NAN-PAG axis bending resistance to 3D force momentum.

We observe a strong modulation of T-cell adhesion and spreading on the PAG-NANs-OKT3 platforms with gel-to-cell surface clearance gaps that are well beyond the physical length (~21 nm) of the CD45 extracellular domain (Figure 5A)^27,88. In our system, the narrowest T-cell-PAG surface clearance for PAG-NANs-OKT3 linkers is ~39 nm, where L_{(NAN 39 bp)>}L_(ECD-CD45), which supports efficient T-cell spreading (Figure 4B, Figure 5B,C). Moreover, 39 bp-long NAN linkers induce T-cell spreading on the soft G’=8.6 kPa PAG surfaces with a shear modulus value that does not induce T-cell spreading via the short OKT3-acryl linker (Figure 5C, 8.6 kPa, short linker). This is despite the fact that the latter linker-facilitated clearance is ~18 nm, which is shorter than the ~21 nm-long CD45 extracellular domain L_(OKT3-acryl)<L_(ECD-CD45), (Figure 5A). Instead, we observe a sharp decrease of T-cell spreading activity on the platform with the next available T-cell-PAG clearance width of ~48 nm (L_{(NAN 66 bp)}>L_(ECD-CD45), Figure 5B). Thus, the CD45-driven KS effect is not responsible for early TCR-driven T-cell adhesion and spreading, but rather the resistance to mechanical bending along the CD3ε-OKT3-dsDNA-PAG axis. One of the alternative mechanosensing models, known as “catch bond,” suggests that the energy of TCR-pMHC interactions increases with the applied force⁹⁴ acting as a TCR-pMHC proofreading mechanical latch aiding to differentiate between stimulatory and non-stimulatory TCR-pMHC interactions^95,96. However, the dynamics of this model is not always captured in non-cellular systems⁹⁷, indicating that the proposed mechanism may be an effect of other cellular responses, and does not represent an intrinsic TCR-pMHC bonding property. In particular, 3D forces⁵² within the T cell immune synapse and our results indicate that the 2D tangential forces, which are responsible for sensing the shear modulus of the substrate, do not fully represent the mechanical configuration of all forces during the TCR mechanosensation.

Signal transduction through the TCR initiates a cascade of signaling events which induce the phosphorylation of CD3 tyrosine-based activation motifs which are not specifically required for late T-cell activation, but rather which contribute quantitatively to TCR signaling⁹⁸. To assess the role of NANs in TCR signaling, we analyze the pY foci response (Figure 6) comparing the gels with OKT3 immobilized by the short linkers and the 39 bp dsDNAs (Figure 6A–4). Immobilization of antibodies with short linkers provides up to 40x more OKT3 on the surface compared to dsDNAs (Figure 6A–1, lanes 1 and 4, lanes 2 and 5). The surface density of OKT3 translates into increased pY foci responses, up to 2x, both for gels of 8.6 kPa and 50 kPa shear moduli (Figure 6A–2, lanes 1 and 4, lanes 2 and 5; Figure 6B, lanes 1 and 4, lanes 2 and 5). However, OKT3 lanes, G’ of 50 kPa, with a normalized density of OKT3 (diluted, short linker), provide a negligible pY foci response (Figure 6A–2, lane 3; Figure 6B, lane 3) and no cell spreading (Figure 6A–3, lane 3) compared to 39 bp dsDNA-immobilized OKT3 lanes of similar density and shear modulus (Figure 6A–2, lane 5; Figure 6A–3, lane 5; Figure 6B, lane 5). In addition, the 39 bp dsDNA-immobilized OKT3 lanes show an increase in pY response in T cells following the increase of gel shear force modulus from 8.6 kPa gel to 50 kPa (Figure 6A–2, lanes 4 and 5; Figure 6B, lanes 4 and 5).

Figure 6. Formation of pY foci in human CD4+ T cells in response to linker force momentum resistance and antigen density.

(A) T cell pY foci response is measured by the number of pY foci, for short linker (green) and 39 bp dsDNA linker (red, (4)). Direct immobilization of OKT3 antibodies increases (~40x) surface density of OKT3 compared to dsDNA linkers (1). Increased OKT3 density correlates with increased pY foci response compared to dsDNA lanes of similar G’= 8.6, or 50 kPa (2). The lane, G’ of 50 kPa, with normalized OKT3 density (diluted, short linker) provides negligible pY foci response (2) and no spreading (3) compared to dsDNA lanes. In addition, OKT3 lanes with 39 bp dsDNA linkers demonstrate increased pY response on gels with increased shear force modulus (2). (B) Micrographic views. T-cell pY foci development along PAG lanes with G’ of 8.6, and 50 kPa for OKT3 antibody molecule immobilized with short linkers (green), and 39 bp dsDNA linkers (red). The 39 bp dsDNA linker (G’ 50 kPa) provides more pY foci compared to short linker (G’ 50 kPa) with normalized OKT3 density (views 3 and 5).

DISCUSSION

Our results show that programmable hydrogel coating, i.e., NANs, is suitable for the ligand attachment and controlling its mechanical resistance. Using NANs, we show the complementary role of 3D force momentum and in-plane 2D shear forces in mechanical lymphocyte stimulation. We conclude that NANs will be useful for studying the complex signalosome assembled around the microvilli TCRs.

From potential biomedical perspectives, NANs would be (i) pyrogen-free⁹⁹, immunoquiescent^100,101, and safe for extracellular use with human immune cells⁷⁰; (ii) tunable for force resistance actuation through reconfiguration in isothermal conditions^{77,78,102–104}; and (iii) recomposable with cognate NANs with similar physicochemical properties, reassociation rates, and biological activities¹⁰⁵. NANs can be designed to vary in size¹⁰⁰, stiffness⁹³, functional composition¹⁰⁶, and encode thermodynamically-driven isothermal interactions that promote conformational changes, which, in turn, could alter the experimental outcomes for mechanosensory and functional properties of the initial NANs¹⁰⁰.

We envision this modular platform to become instrumental for exploring the behaviour of human immune cells ex vivo, thus expanding the possibilities of basic research and immunotherapies. In basic research, NAN-PAG conjugates could be used to study homing receptors and molecular machinery consolidated around the mechanosensitive TCR to elucidate co-stimulatory and co-inhibitory signals that regulate antigen search and TCR engagement strategies. In translational studies, NANs could be used for optimizing CD8+ T cell activation, e.g., during conditioning of CAR T constructs in non-clinical settings.

Supplementary Material

1

Supporting Figure S1. Alternative strategies for the stepwise formation of double-biotinylated DNA duplexes visualized through the assembly of QD lattices. (A) Schematic representation of three alternative protocols shows the underlying concept of QD lattice assembly. The incubation of streptavidin-decorated QDs with (1) double-biotinylated DNA duplexes, (2) biotinylated complementary ssDNAs, or (3) biotinylated cognate RNA/DNA heteroduplexes will result in the formation of QD lattices (B). The progress of QD lattice formation can be tracked by agarose gel electrophoresis (C). In all cases, the dsDNA-driven formation of QD lattices is confirmed by treatment with DNase. (D) Schematic representation of three different protocols tested with PAG (Top) and - OKT3 (Alexa Fluor® 555) adsorption and T-cell spreading results (Bottom). Note that only the RNA/DNA heteroduplex re-association strategy is optimal for fluorescent OKT3-to-streptavidin lanes adsorption (OKT3 brightness), and T-cell spreading.

Supporting Figure S2. Schematic representation and corresponding agarose gels of thermodynamically driven toehold-initiated reassociation of RNA/DNA heteroduplexes, (or hybrids 1 and 2) releasing the double-biotinylated dsDNA (39, 66, 81, 93 bp) and shorter dsRNAs. Note that for longer hybrids, either single longer RNA strands or multiple shorter RNAs can be used as 81 bp and 66 bp duplexes, respectively. The use of shorter RNAs may offer an additional advantage since no IVT is needed and short RNAs can be reused as in the case of 39 (one RNA), 66 (two RNAs), and 93 bp (three RNAs).