Cytosolic delivery of membrane-penetrating QDs into T cell lymphocytes: implications in immunotherapy and drug delivery

Abstract

We report single-particle characterization of membrane-penetrating semiconductor quantum dots (QDs) in T cell lymphocytes. We functionalized water-soluble CdSe/CdZnS QDs with a cell-penetrating peptide composed of an Asp-Ser-Ser (DSS) repeat sequence. DSS and peptide-free control QDs displayed concentration-dependent internalization. Intensity profiles from single-particle imaging revealed a propensity of DSS-QDs to maintain a monomeric state in the T cell cytosol, whereas control QDs formed pronounced clusters. Single-particle tracking showed a positive correlation between individual QD clusters’ mobility and aggregation state. A significant portion of control QDs colocalized with the endosome marker inside the T cells, and the percentage dropped to 9% for DSS-QDs. Endocytosis inhibition abrogated the internalization of control QDs, while DSS-QD internalization only mildly decreased, suggesting an alternative cell-entry mechanism. Using 3D single-particle tracking, we captured the rapid membrane-penetrating activity of a DSS-QD. The ability to characterize membrane penetrating activities in live T cells creates inroads for the optimization of gene therapy and drug delivery through the use of novel nanomaterials.

Graphical Abstract

Single-particle studies of intracellular delivery of DSS-QDs into T cell lymphocytes.

Introduction

Biotechnology and biomedical applications of nanomaterials have flourished in the last decade, especially in the fields of biomaterial separation, immunoassays, diagnostics, and drug delivery systems^1–7. To enhance their diagnostic and therapeutic efficacy, novel nanoassemblies must be engineered to function in biologically relevant environments and have multivalent loading capacity to facilitate detection and effective drug delivery. Reproducibly accessing the intracellular space with precision delivery has remained a highly desirable goal. In particular, harnessing the cellular machinery to suppress⁸ or enhance⁹ the cellular function for treatment using theranostic nanoparticles is a highly promising strategy.

T cell lymphocytes have been engineered as “programmable and living drugs” for the treatment of various types of refractory cancers¹⁰. To this end, engineering nanoparticle carriers for drug and gene delivery to targeted T cell populations represents a promising strategy for increasing the response rate and durability of immunotherapies. The unique physical, photophysical, and photochemical properties of engineered nanomaterials have enabled a broad range of finely controlled stimulation of T cells. For instance, magnetic nanoparticles have been used to enhance T cell activation via forced spatial clustering of T cell receptors and co-stimulatory molecules using an external magnetic force¹¹. Photothermal and photodynamic therapy using nanoparticles have shown efficacy towards augmenting the anti-tumor response and T cell infiltration^12–14. Upconverting nanoparticles enable the delivery of visible light into deep tissues and remote-controlled immunomodulation using ultraviolet light-activatable immunostimulatory agents based on CpG oligonucleotides¹⁵. In addition, targeted delivery of small molecule inhibitors to specific T cell subpopulations effectively reduces the toxicity of systematic administration and improves antitumor immunity¹⁶.

Transmembrane receptors with stimulatory or inhibitory functions, such as the mechanisms targeted by the immune checkpoint inhibitors, represent a major class of targets for engineered nanoparticles. A complementary approach involves modulating the intracellular signaling pathways by delivering small molecule drugs^{11, 17} or genetic materials such as siRNAs^18–21. This requires the delivery of engineered nanoparticles into the intracellular space. However, the delivery and localization of nanoparticles within cells have been mainly characterized using transmission electron micrographs (TEM) of fixed samples¹¹ or confocal imaging of particle ensembles²². Compared to fluorescence imaging, TEM has limited capability in multiplexed labelling and the detection of biomolecules. Moreover, systematic investigations of how design parameters affect the nanoparticle-T-cell interactions and subsequent cell entry require high-resolution live-imaging assays. However, confocal imaging lacks the necessary spatial resolution, and single-molecule imaging is difficult to perform with suspension T cells.

Semiconductor quantum dots (QDs) have excellent characteristics, such as size-tunable optical properties and photochemical stability, for single-particle visualization in live T cells²³. QDs exhibit high molar extinction coefficients over a broad excitation region and narrow emission spectra, enabling multiplexed detection from the blue to near-infrared region²⁴. The development of methods for robust and reproducible cytosolic delivery of QDs into live cells has been challenging for a variety of reasons²⁵. Generally, the highest quality nanomaterials are prepared in a hydrophobic solvent and require surface modification to impart water solubility. Surface functionalities are often composed of amines, carboxylic acids, or PEG grafted onto polymers, liposomes or small molecule “caps” to name a few^26–30. Unfortunately, without further functionalization, these nanomaterials have minimal interactions with cells or may at best become trapped inside of endosomes^{31, 32}. To date, the majority of reports on successful cytosolic delivery use protein functionalized QDs to either circumvent endosomal trapping or escape once sequestered^32–35. Other strategies involve mechanical delivery, such as microinjection⁸, which may be difficult to implement on T cells. Recently, cell-penetrating poly(disulfide)s have been reported to be effective in drosophila cells³⁶. However, the number of reports on cytosolic delivery of fluorescent QDs remains few and far between, and it remains unclear what surface modification parameters can be optimized for maximal cargo delivery, and how such nanoparticles may interact with human T cells.

Here, the cytosolic delivery of biocompatible QDs into T cells was investigated. Cell uptake of CdSe/CdZnS QDs was enhanced by surface coating with a cell penetrating peptide (CPP) composed of a repeated Asp-Ser-Ser (DSS) sequence. The DSS repeats are derived from a motif found in dentin phosphophoryn, one of nature’s most acidic proteins, which was previously demonstrated by us to deliver quantum dots into live cells²². Furthermore, the DSS motif was used to coat lignin nanoparticles, facilitating their efficacy as a drug delivery agent for several types of cancer cells³⁷. To enable stable single-particle imaging, QD-laden T cells were immobilized on activating surfaces. Control and DSS-QDs were found to be distinctively distributed between the monomeric and clustered state, with the majority of DSS-QDs in the single-particle state. In addition, DSS-QDs displayed significantly higher mobilities compared to control QDs. This study creates in-roads for the use of DSS-coated nanoparticles for in vivo drug delivery and gene therapy.

Results

Our previous study demonstrated that dentin phosphophoryn (DPP), an acidic, phosphorylated protein that is a ubiquitous component of the dentin extracellular matrix, is internalized by several cell types via a non-conventional endocytic process^{22, 38}. Furthermore, as DPP contains Asp-Ser-Ser (DSS)_n repeats distributed throughout the protein, it was demonstrated that (DSS)_n facilitates cellular uptake and can function as a cell-penetrating peptide to deliver proteins for therapeutic applications^{22, 37}. It has been observed that DSS can internalize cargo such as CdSe/CdZnS quantum dots in several cell types. QDs conjugated with the chimeric protein DSS and the osteoblast-specific transcription factor Runx2 (DSS-Runx2) favored nuclear translocation²². Incubation of the functionalized nanoparticles with MC3T3 osteoblast precursor cells resulted in passive delivery into the cytoplasm and trafficking into the nucleus. In the present study, a linkable (DSS)₁₀K₄ polypeptide was synthesized and conjugated to water-soluble CdSe/CdZnS QDs to evaluate their membrane penetration activities with T cells. The DSS peptide carries an overall negative charge at physiological pH due to the repeated aspartate. Previous studies indicate that the surface charge affects the internalization and the cytotoxicity of the QD^39–44. Charged nanoparticle internalization is facilitated by utilizing interactions with various charged proteins on the cell membrane⁷. Specifically, negatively charged nanoparticles were found to internalize into cells more effectively compared to neutral nanoparticles. In addition, negatively-charged QDs have been shown to be relatively less cytotoxic compared to positively charged QDs⁴⁵. In alignment with these results, the accumulation of negatively-charged DSS peptide provides a more biocompatible environment and enhances its delivery through charged interactions between the peptide and membrane. It remains a challenge to quantify the number of peptides per QD. Commonly used methods rely on the weight of the lyophilized powder, absorbance of ultraviolet (UV) light or amino acid analysis. Peptides that do not contain tryptophan and tyrosine, as with the DSS peptide, would be difficult to quantify using the UV absorbance at 280 nm^{46, 47}. Furthermore, the broad excitation range of QDs could affect the absorbance characterization⁴⁸. NMR characterizations confirmed functionalization of the DSS peptide with the water-solubilizing polymer (Supplementary Fig. 1). While our NMR spectra demonstrate the presence of DSS on the QD, the measurements were obtained from solid-state materials and the quantification of peptides per QD is a subject of future investigation. To determine the hydrodynamic size (Dh) of the quantum dots, we analyzed control and the DSS-conjugated QDs using Dynamic Light Scattering. The average Dh was determined from three replicates and the value was found to be 89 nm for the control and 100 nm for the DSS-QDs (Supplementary Fig. 2). Of note the Dh of our QDs was on the large side compared to cell-penetrating peptides (CPP)-QDs reported in the literature⁴⁹. Despite the size, the small difference indicates that it is unlikely that hydrodynamic sizes played a significant role in the different behaviors of the QDs we observed.

E6–1 Jurkat T cells were incubated with the growth medium containing DSS-QD solution for 24 h. Before imaging, T cells were washed three times in HBSS and resuspended in phenol red-free HBSS containing 1% FCS. The T cells were immobilized on an 8-well chamber slide coated with anti-CD3 antibodies (clone OKT3). QD-laden T cells were subsequently imaged using TIRF with a Nikon Eclipse Ti-E2 inverted microscope (Fig. 1a, Supplementary Fig. 3)⁵⁰ To evaluate whether the internalization of DSS-QDs affected T cell immobilization and activation process, Jurkat T cells were examined by brightfield and fluorescence imaging. Fig. 1b shows a QD-laden T cell initially interacted with the OKT3-coated surface and then maintained a stable contact with the surface (Supplementary Video 1). On the activating surface, brightfield imaging revealed no morphological differences between DSS-QD-laden T cells and control T cells without prior QD exposure. To evaluate the cytotoxicity of the QDs terminated with DSS peptide on Jurkat T cells, a CellTiter-Glo^® Luminescent Cell Viability Assay was conducted on cells with DSS-QDs for 24h. Fig. 1c shows that the cell viability remained above 80% over the applied QD concentration range from 3.16 pM to 10 nM. The IC50 of DSS-QD was found to be 23.2 nM. These results reveal that DSS-QDs have no considerable cytotoxicity under this concentration range and support the suitability of DSS-QDs for studying nanoparticle internalization into T cells.

Fig. 1.

Single-particle visualization of QD delivery into T cell lymphocytes. a Schematic representation of the T cell immobilization and visualization on the activating surface coated with antibodies. b Selected imaging frames showing the initial interaction between a T cell and the activating surface coated with anti-CD3 antibodies. c Representative IC50 results for incubating Jurkat T cell for 24h with DSS-QDs in the culture medium. Each data point represents the mean ± standard deviation of wells performed in quadruplicate. d Quantification of the number of internalized DSS-QDs per cell characterized by TIRF imaging. Statistical significance was evaluated using an unpaired Student’s t-test. *** represents p<0.001 and n.s. denotes not significant. Scale bars: 5 μm.

A brief examination of the control and DSS-QD quantification in the cells revealed that the internalization of nanoparticles was concentration-dependent(Fig 1d). An ImageJ plugin, ThunderSTORM, with multiple-emitter-fitting capability was used to quantify this observation^{51, 52}. We counted the number of QDs from a static image using ThunderSTORM; ThunderSTORM enables automated localization, analysis, and visualization of single-molecule events. Utilizing its ability to gate the threshold for single QD intensity and quantify spatially overlapping QDs in a cluster, we obtained the total number of QDs observed per cell⁵². A similar approach using ThunderSTORM has been previously reported⁵³. Determining the number of QDs in a cluster is based on the premise that QDs of the same composition and structure have relatively the same intensity^{54, 55}. The number of QDs in a cluster correlates with the intensity emitted by the QD cluster. ThunderSTORM is able to determine the number of QDs in a cluster by enabling the Multiple-emitter fitting analysis tool. While the blinking dynamics and QD motility may lead to an inaccurate count of absolute numbers, our observations were made on a comparative basis (control vs. DSS-QDs) under the same experimental design and quantitative analysis procedure.

Fig. 1d shows that the number of observed QDs per cell monotonically decreased when the concentration of the QDs was reduced. Between the two groups, DSS-QDs showed significantly more internalization when the T cells were incubated with a high concentration of 5 nM DSS-QDs in growth medium. On average, the number of internalized QDs per cell was 62 ± 9 (mean ± SEM, n = 36 cells) for DSS-QDs and 38 ± 6 (mean ± SEM, n = 28 cells) for control QDs. This difference was diminished as the concentration decreased (Fig. 1d). The lack of significant increase in the DSS-QD uptake at sub-5 nM concentrations may be related to how QDs were incubated with suspension T cells. The number of QDs encountering a T cell is related to the QD concentration and the total T cells in the medium. Importantly, QDs do not precipitate and immobilize onto the cell membrane as in adherent cell cultures. The internalization depends on the QD’s immobilization onto the plasma membrane of a freely mobile T cell and subsequent nano-bio interactions that bring the QD into the intracellular space. At very low concentrations, the internalization may become a low-probability event, and the difference between control and DSS-QDs may diminish. Despite the similar number of particles per cell, we observed that control QDs formed more pronounced and isolated clusters. Furthermore, control and DSS-QDs revealed distinct particle distributions in the form of large and small clusters.

Notably, DSS-QDs were visible inside the T cells through TIRF imaging (Fig. 2a). Fig. 2b, c are typical images of activated T cells after 24 h incubation with DSS and the control QDs. At 500 pM, the random distribution of DSS-QDs is sparse and manifests as small dots inside the cell, while the control-QDs displayed a more localized distribution and pronounced puncta. Incubation with a higher QD concentration, 5000 pM, increased the number of internalized DSS-QDs, while pronounced clusters of control QD were still visible. To investigate whether the variation in the brightness resulted from different optical behavior of control vs. DSS-QDs, single-particle imaging was performed on a coverglass coated with monodispersed QDs (Fig. 2d). Fig. 2e and 2f demonstrates a background-corrected average brightness value of approximately 40–60 a.u. for DSS and control QDs at the single-particle level, respectively. Background subtraction was performed by encompassing a region outside the cell in a rectangle and using ImageJ’s Plot Profile Tool to find the average intensity of the region. The background value was then subtracted from the intensity of the QDs. The same background value was used for all images. Fig. 2g plots the intensity profile of internalized QD clusters. The majority of DSS-QDs registered intensity levels below 80 a.u., indicating that DSS-QDs were mostly monomeric in the T cell cytosol. In contrast, the intensity profile of control QDs was much more heterogeneous; the intensity distribution displayed a minor peak between 160 and 480 a.u. and a major peak greater than 800 a.u. in the cell; this suggests that upon cell entry, control QDs are more likely to form aggregates.

Fig. 2.

Intensity analyses indicate distinct aggregation states of DSS-QDs and control-QDs in the cytosol of T cells. a Brightfield and fluorescence images of a Jurkat T cell on the activating surface incubated with QDs. b Fluorescent images of internalized DSS-QDs and control-QDs after 24 h incubation with 500 pM QDs. c Fluorescent images of internalized control-QDs and DSS-QDs after 24 h incubation with 5000 pM QDs. d Schematic figure of single QDs coated on coverglass surface and corresponding QD intensity distribution histogram. e,f Intensity profiles of individual DSS-QD (e) and control-QD (f) dispersed on a coverglass. g Schematic figure of QDs inside of Jurkat T cells and corresponding intensity distribution of QDs inside T cells after 24 h incubation with 5000pM QDs h Intensity profiles of control-QD clusters inside T cells. i Intensity profiles of small DSS-QD clusters inside T cells. j Intensity profiles of single DSS-QD inside T cells. Images were acquired under identical experimental conditions. Scale bars: 5 μm.

Due to the fluorescence intermittency of CdSe QDs, a random switching takes place between bright fluorescence periods (“on” state) and dark, non-emissive ones (“off” state). A distinct transition between the “on” and “off” state is visible from time trace plots of single QDs on the cover glass (Fig. 2e, f). Clusters of QDs (Fig. 2h, i) exhibited fluctuations of emission at much higher intensities. Small dots of DSS-QDs in the T cell cytosol (Fig. 2j) displayed identical blinking properties compared to single QDs on the cover glass, further confirming their monomeric state.

To investigate the dynamics of internalized QDs and how particle dynamics correlate with the cluster size, two-dimensional single-particle tracking was performed. Intensity profiling was utilized to categorize the formation state of each cluster into a single particle (approximately 40–60 a.u.) and aggregates (intensity greater than 200 a.u.). A similar approach has been previously used to quantify membrane receptors via calibrated QD intensities⁵⁶. Figs. 3a–c show that single-particle positions can be tracked with image frames 50-ms apart. These observations prompted a changeover in illumination to a highly inclined and laminated optical sheet (HILO) configuration. Figs. 3a–c illustrates representative single-particle images, tracks and mean square displacement of control and DSS-QDs after 24-h incubation with 5000 pM QDs. For control-QDs, the aggregates displayed significantly slower diffusion (red track). Most DSS-QDs were dispersed in the cytosol as monomers with the occasional emergence of small aggregates (black track) at higher QD concentrations. Monomeric QDs were mobile in the cytosol (green track). Averaged two-dimensional mean squared displacement (MSD) showed similar characteristics (Fig. 3d). The diffusion coefficient of single DSS-QDs was found to be 7.5 ± 1.9⋅10⁻² μm²/s (N=18). The linear MSD curves indicate normal diffusion observed within the imaging duration of approximately two seconds.

Fig. 3.

Two-dimensional single-particle tracking and fluorescence imaging of aggregate control QDs, aggregate DSS-QDs, and monomeric DSS-QDs a,b,c Single-particle dynamics of a. control QD aggregates b. small aggregates of DSS-QDs, and c. single particles of DSS-QDs imaged using TIRF. Upper panel: Single-particle images, middle panel: Representative single-particle tracks of a control QD aggregates (red), b. small aggregates (black), and c. single particles of DSS-QDs (green), lower panel: mean squared displacement analysis. Scale bar: 300 nm. d Averaged MSD measurements with the error bars indicating the standard error of mean.

We further confirmed these results by co-incubating T cells with QDs and the small-molecule dye, Acridine Orange (AO) to label acidic endosomes to investigate potential endosomal trapping of the nanoparticles through the dual-color imaging. Fig. 4a shows that, while several pronounced puncta of control QDs overlapped with the endosome vesicles (arrows in Fig. 4a), little to no overlap between the DSS-QDs and endosomes was observed. We evaluated the percentage of QDs that overlapped with the endosome marker. Quantitative colocalization analysis revealed that 57% of the control QDs overlapped with the endosome marker, while the percentage was much lower at 9% for the DSS-QD (Fig 4b). The small amount of overlap indicates that some DSS-QDs could have entered T cells through endocytosis. Some of the DSS-QDs may have rapidly passed through the endosomes into the cytosol, while others remained in the endosome at the time of imaging. Through single-particle imaging, the diffusion coefficient of endosomes was found to be 3.7 ± 2.2×10⁻⁵ μm²/s. These data are consistent with Supplementary Fig. 4 that demonstrates slower endosomal diffusion compared to monomeric DSS-QDs in T cells. Studies by others reported a higher diffusion coefficient for actively trafficking endosomes. These values range from 2.1 ± 0.1×10⁻³ μm²/s for kinesin-directed endosome diffusion during cell division⁵⁷ to 2.4 ± 0.7×10⁻² μm²/s in the motile population during early endosome trafficking⁵⁸. Of note is that the diffusion coefficient of DSS-QD is still higher than these values. The rapid diffusion suggests that DSS-QDs are free from endosomes whereas large aggregates of control QDs are likely sequestered within.

Fig. 4.

Fluorescence imaging using an endosomal marker reveals that DSS-QDs were significantly less entrapped in endosomes. a Fluorescence images of an T cell labelled with QDs and Acridine Orange (AO). White and yellow arrows mark the colocalization of control-QDs and endosomes. b The percentage of QDs overlapped with endosomes. c Fluorescence images of representative single Jurkat T cell incubated with 500 pM QDs for 1h at 4°C vs 37°C. d Quantification of the number of internalized QDs per cell characterized by images and the corresponding fold change. e Fluorescence images of 80mM Dynasore-treated vs 2 % DMSO-treated Jurkat T cells incubated with QDs for 1h. f Quantification of the number of internalized DSS-QDs per cell and the corresponding fold change. The mean is represented by “+” and the horizontal line indicates the median. Scale bar: 5 μm.

As the size of QDs increases, the chance for energy-dependent endocytosis increases^59–61. To support the notion that the transfer of DSS-QD through the plasma membrane is not solely dependent on the mechanism of endocytosis, we performed two sets of experiments to observe the role that endocytosis plays in QD internalization. The common notion is that QDs utilize the Clathrin-Mediated-Endocytosis (CME) mechanism to enter the cell⁶². In the first experiment, cells were incubated with QDs at 4 °C. This condition blocked fluid‐phase endocytosis and receptor-mediated endocytosis (also known as CME)⁶³. Fig. 4c shows that while the internalization of control QDs vanished, DSS-QDs still entered the T cell at 4°C. The delivery efficiency was reduced nearly five-fold for control QDs, while the temperature had a negligible effect on the DSS-QDs (Fig. 4d). The strong response to the inhibition of endocytosis at 4°C suggests that endocytosis was likely the main internalization mechanism for the control QDs, while the DSS-QD internalization was less dependent on it.

In a separate experiment, we treated T cells with dynasore, a specific inhibitor of atypical GTPase dynamin that represents a key component of clathrin-coated pit constriction⁶⁴. After treating T cells with 80 mM dynasore for 30 min, we incubated the cells with QDs in the presence of the dynasore. Fig. 4e shows that T cells internalized DSS-QDs under endocytosis-restricted conditions. Moreover, we excluded the effect from the 2% DMSO incubation solution as dynasore was dissolved in DMSO (Fig. 4f). The delivery efficiency was reduced slightly over one-fold under these inhibitory conditions in comparison to normal conditions, suggesting that the DSS-QDs utilized both the typical CME mechanism and an alternative mechanism for internalization. In contrast, a limited amount of measurable control-QDs were observed after incubation with dynasore, and the delivery efficiency was reduced more than three-fold (Fig. 4f). These findings combined with the fluorescent images shown above strongly suggest that internalization of the DSS-QD primarily occurs through an alternative mechanism, such as translocation through the membrane, while some levels of endocytosis-driven uptake may still play a role.

The endosomal bypass pathway of DSS-QDs indicates a possible alternative cell-entry mechanism, i.e., through direct membrane penetration due to the surface conjugation of DSS-enriched peptide. However, this process has not been directly observed. To probe the evidence of this cell entry mechanism, we incubated T cells with 5000 pM DSS-QDs for 10 min, activated T cells on the glass substrate, and performed astigmatism-based 3D single-particle tracking⁶⁵. To obtain the 3D calibration data, we first immobilized TetraSpeck^™ Microspheres on a clean coverglass. Next we stepped the coverglass from −350 nm to +350 nm around the focus using a 10-nm step size. The width (Wx) and height (Wy) of single molecule images within the field of view was calculated and averaged at each position. A representative calibration curve collected at 600 nm emission was shown in Supplementary Fig. 5. From there, curve fitting was applied to interpret the relationship between (Wx-Wy) and z. A proprietary cylindrical lens was placed in front of the camera (Nikon N-STORM). The calibration was performed using the 100x/1.49 objective. The relatively short incubation period allowed us to capture the early point of cell entry. Fig. 5a illustrated the x, y, and z components of a DSS-QD trajectory attached to a cell membrane. The DSS-QD initially exhibited confined movements up to approximately 90–100 s. Some oscillatory behaviors were observed between 30 and 70 s. The z position reached its peak at 133 s (Fig. 5b), which correlated with a drastic increase of mobility near the end of the track (Supplementary Fig. 6, Supplementary Video 2). We employed a Lowess local linear regression fit to construct a 3D surface based on the first 2,000 coordinates (up to 100 s in Fig. 5a). The surface topology represents the local membrane area the DSS-QD has migrated over 100 s (Fig. 5c). We then plotted the last 30 coordinates (the last 1.5 s) of the DSS-QD in red circles. These positions showed an abrupt increase in mobility, indicating the particle was no longer attached to and confined by the membrane. In addition, the inward direction of migration indicates the particle has reached the cytosolic side of the membrane within this short period of time (Fig. 5c). Due to the thin cytoplasmic region of Jurkat T cells, after penetrating the membrane, the particle stayed relatively close to the membrane before it eventually moved outside the tracking range in z.

Fig. 5.

Three-dimensional single-molecule tracking demonstrating the membrane-penetrating activity of a DSS-QD into a T cell. a Single-particle positions for each Cartesian coordinate axis as a function of time. b Single-particle images recorded at four time-points (vertical black lines in a). c A three-dimensional track of a DSS-QD: the surface was interpolated with 2,000 single-particle positions (blue) up to 100 s; the red circles correspond to the last 30 single-particle positions before the QD moved out of the tracking range. The red circles correspond to red lines in a. PM: plasma membrane. Scale bar: 1μm.

Discussion

Our results establish a methodology to evaluate cell penetrating behaviors of individual nanocarriers using high resolution microscopy. We observed internalization of both control and DSS-QDs into T cells. The observation is aligned with the notion that T cell lymphocytes may exhibit some basal levels of phagocytic activities. For instance, human ɣδ T cells have been reported to be capable of professional phagocytosis⁶⁶. T cells may also internalize QDs through micropinocytosis facilitated by the microvilli structures on the plasma membrane⁶⁷. Importantly, our data provide the evidence of QD internalization in the low-concentration range, which may be below the detection threshold of standard imaging techniques, such as confocal microscopy. Supporting data (not shown) also indicate that DSS-QDs internalize more readily in both Dendritic (DC 2.4) Cells and U2OS Cells. The use of DSS-QDs for efficiency subcellular delivery remains an active subject of investigation by us and others.

Our imaging study demonstrates the ability to monitor individual nanoparticles within suspension T cells on an activating surface. In vitro activation facilitates the immobilization of T cells for stable imaging. Our TIRF and HILO illumination yielded relatively high signal-to-noise ratio in regions near the contact side on the cover glass (Fig. 2b,c). Single-particle volumetric imaging, such using the lattice light-sheet system^{68, 69}, will further advance our understanding by providing long-range 3D tracking in the entire cytosolic space. In addition to the particle concentration, the temperature, pH, and incubation duration may also affect the QD characteristics in the cytosol and are subjects of future investigations.

Importantly, the single-particle tracking capability enables the study of sub-cellular trafficking of multifunctional nanocarriers delivering gene products, small molecule drugs, and other immune modulatory agents. The membrane penetrating mechanism enabled by DSS-enriched peptides bypasses standard receptor-mediated endocytosis, thereby enabling sub-cellular targeting and access to sub-cellular machineries. Multiplexed DSS-QDs can be used as the base structure to investigate how different design formulations affect the cell entry, sub-cellular trafficking, and distribution of engineered nanocarriers. The in vitro data provide valuable insight to guide the optimal design of in vivo experiments.

Figure 5 represents the direct visualization of a single DSS-QD’s internalization into human T cells. While the visualization provides a means to evaluate the behaviors of individual QDs, our technique is low-throughput. The unique insight, including the local nanoscale landscape constructed by the 3D diffusion on the plasma membrane of a T cell, the oscillatory behavior of the QD, and the rapid time scale for T cell entry, i.e., on the order of a second, will complement the ensemble-level characterizations to advance the development of nanocarriers for next-generation immunoengineering.

In summary, we investigated the cytosolic delivery of DSS-QDs in suspension T cells. Functionalized CdSe/CdZnS QDs were synthesized and water solubilized with amphiphilic poly(acrylic acid) functionalized with a peptide cell delivery vector comprised of a repeat sequence of aspartic acid and serine. DSS-QDs were found to be monomeric or in small clusters. In comparison, control QDs without the DSS peptide were generally aggregated. Through single-particle tracking, monomeric DSS-QDs showed substantially higher mobility compared to the small-cluster counterparts and the larger clusters of control QDs. Our imaging data indicated the colocalization of endosomes and control-QD clusters in the T cells and absence of colocalization for DSS-QDs. 3D single-particle tracking captured the evidence of membrane penetration of a DSS-QD. Taken together, the imaging platform enables systematic investigations of design parameters that affect nanoparticle-T-cell interactions at the level of single particles in vivo, thereby advancing the understanding of nanomaterials dynamic in T cell cytoplasm and their use for enhancing cancer immunotherapy and drug delivery.

Experimental

Materials and methods

Control QD and DSS QD synthesis

Control QDs were water solubilized using 40% octylamine-modified poly(acrylic acid)(average M_W 180, Aldrich) as per the protocol outlined in ref. 47. To prepare DSS-QD conjugates, initially water soluble QDs and DSS peptide were incubated with poly(ethylene glycol) conjugation reagent^{34, 70}. However, it was found that the resulting materials did not display significantly different behaviour compared to controls. The preparation was modified to enhance the yield of functionalized nanomaterials by conjugating DSS to the modified acrylic acid solubilizing polymer first, which was then processed by precipitation in acidic water and subsequently used to solubilize the CdSe/CdZnS nanomaterials as outlined below.

DSS QD conjugation

Our approach is to conjugate DSS peptide to the modified acrylic acid solubilizing polymer first, which was then processed by precipitation in acidic water and then used to solubilize the CdSe/CdZnS nanomaterials. To this end, 5 mg of (DSS)₁₀K₄ (1.46 mmol) was added to a solution of DMF with 8 mg poly(acrylic acid) (111 mmol) and 9 mg EDC (47 mmol, G-Biosciences). After stirring a few moments, 7.3 μL of octylamine (44.1 mmol, TCI) was slowly added. The sample was stirred overnight and was precipitated with the addition of water. After centrifugation, the supernatant was discarded. The functionalized polymer was dissolved in basic water, and was titrated with mildly acidic water to pH< 5, upon which the polymer precipitated. The supernatant was discarded, and the polymer was dried under vacuum. The final mass was 12 mg (64% yield). Approximately 6 mg was used to solubilize 1.8×10⁻⁸ moles CdSe/CdZnS QDs as previously reported⁷¹.

QD incubation with Jurkat cells

Jurkat E6-1 cells were obtained from ATCC. For the incubation, 500 μL of 5 μM QDs was added with ~90 k Jurkat cells in culture medium (RPMI, 10% FCS, Thermo Fisher Scientific) to a 12 wells plate. Jurkat cells were cultured with QD-containing medium at 37 °C and 5% CO₂ for 24 hours. Jurkat cells were then transferred into a 10 mL conical tube with 5 mL HBSS (Thermo Fisher Scientific) added into the medium and centrifuged at @1400 rpm for 3 min for three times. The cells were resuspended in the imaging buffer containing 1% bovine serum albumin, 0.5 mM Ca²⁺, 2 mM Mg²⁺, and HBSS pre-warmed to 37 °C.

Acridine Orange (AO) incubation with Jurkat cells

Before imaging, Jurkat cells were removed from a culture flask and placed in a 1.5 mL centrifuge tube. They were incubated in 0.1μg/mL Acridine Orange (AO) in culture medium for 15 min at 37 °C in the dark. The 1.5 mL centrifuge tubes were wrapped in aluminum foil to protect the dye from light exposure. After incubation, cells were washed with 1.5 mL Gibco Hank’s Balanced Salt Solution (HBSS) three times, and then were resuspended in DPBS.

Dynasore inhibition

To test the endocytosis path of the DSS-QD, the following experimental procedure was applied. Jurkat cells were treated with 80 μM dynasore (VWR International, which was dissolved in 2% DMSO within culture medium for 30 min. At the end of the treatment, cells were washed three times with Ca²⁺/Mg²⁺-free HBSS and the medium was replaced with the QD-containing culture medium. 1 hour following medium replacement, cells were treated with QDs. Treated cells were washed and were resuspended in the imaging buffer as above.

Temperature-dependent endocytosis inhibition

Jurkat cells were cultured with cold QD-containing medium at 4 °C for 1 hour. Then followed by washing the cells with cold PBS at 4 °C three times to remove all free QDs in the suspension.

Immobilization and imaging of T cells

To make the activating surface, 8-well chamber slides were cleaned with absolute ethanol and DI H₂O, then incubated overnight at room temperature. Coated OKT3 surface was produced by adding 200 μL OKT3 antibody at a concentration of 1 μg/mL in PBS per well. TIRF was performed on a Nikon N-Storm super resolution ECLIPSE Ti2-E microscope (TIRF 100 ×, 1.49 NA objective lens). QD-laden T cells were a 405 nm continuous wave laser and the emission was collected at 561 nm. The images were collected by a Photometrics Prime 95B sCMOS camera with a pixel size of 110 nm. Image analysis was performed by the ThunderSTORM ImageJ plug-in and a custom MATLAB code.