Nanofibrous MultiDomain Peptide Hydrogels Provide T Cells a 3D, Cytocompatible Environment for Cell Expansion and Antigen-Specific Killing

Abstract

T cells have the ability to recognize and kill specific target cells, giving therapies based on their potential for treating infection, diabetes, cancer, and other diseases. However, the advancement of T cell-based treatments has been hindered by difficulties in their ex vivo activation and expansion, the number of cells required for sustained in vivo levels, and preferential localization following systemic delivery. Biomaterials may help to overcome many of these challenges by providing a combined means of proliferation, antigen presentation, and cell localization upon delivery. In this work, we studied self-assembling Multidomain Peptides (MDPs) as scaffolds for T cell culture, activation, and expansion. We evaluated the effect of different MDP chemistries on their biocompatibility with T cells and the maintenance of antigen specificity for T cells cultured in the hydrogels. We also examined the potential application of MDPs as scaffolds for T cell activation and expansion and the effect of MDP encapsulation on T cell phenotype. We found high cell viability when T cells were encapsulated in noncationic MDPs, O₅ and D₂, and superior retention of antigen specificity and tumor-reactivity were preserved in the anionic MDP, D₂. Maintenance of antigen recognition by T cells in D₂ hydrogels was confirmed by quantifying immune synapses of T Cells engaged with antigen-presenting cancer cells. When 3D cultured in anionic MDP D₂ coloaded with anti-CD3, anti-CD28, IL2, IL7, and IL15, we observed successful T cell proliferation evidenced by upregulation of CD27 and CD107a. This study is the first to investigate the potential of self-assembling peptide-based hydrogels as 3D scaffolds for human T cell applications and demonstrates that MDP hydrogels are a viable platform for enabling T cell in vitro activation, expansion, and maintenance of antigen specificity and therefore a promising tool for future T cell-based therapies.

Graphical Abstract:

INTRODUCTION

Cell therapies are promising emerging treatments for diseases such as cardiovascular disorders, neurological pathologies, and cancer.^1–3 They provide many benefits over other biologics due to their ability to respond dynamically to tissue microenvironmental cues.¹ Among different cell therapies under development worldwide, T cell-based therapies comprise 45% of current clinical trials.⁴ As the field of immunotherapy continues to expand alongside our understanding of T cell-mediated immune response, the application of T cell therapies has continued to grow for their potential to suppress undesirable autoimmune responses (as in diabetes and Crohn’s Disease) and their ability to direct antigen-specific effector responses, against infected or cancerous cells.^5–9 Most cell-based therapies currently in development involve adoptive T cell therapy (ATCT). This process requires the isolation, activation, expansion, and reinfusion of tumor-specific T cells.¹⁰ However, because of the hostile tumor microenvironment that can impair the survival and function of T cells and the requirement for a large number of cells, sustainable clinical responses with T cell therapies remain challenging.^4,11–16 In addition to the often labor-intensive ex vivo processing required for the production of ATCT, cellular heterogeneity in function and toxicity associated with T cells are additional limitations for clinical application.⁴ Furthermore, with systemic administration of ATCT, tumor-specific T cells may circulate in the bloodstream or localize in lymphatic tissue. Hence, infusing large numbers of effector T cells may be required to ensure the localization of adequate antitumor activity in the tumor microenvironment.^4,17 Although alternative delivery approaches have been developed to substitute systemic bloodstream infusion, improve bioavailability, and mitigate systemic toxicities,⁴ the use of biomaterials to deliver T cells remains relatively unexplored.

Biomaterials can be used to create 3D scaffolds, structures with spatiotemporal control of biological cues, and platforms for improved cellular localization and delivery, making them useful tools for ATCT.^18–26 The need for better expansion and activation methods for implementing ATCT has been tackled by developing scaffolds that present all the necessary stimulatory signals.^27–29 Biomaterials such as chitosan, alginate, and alginate-PEG conjugates for enhanced T cell delivery have also been explored, achieving improved antitumor activity, enhanced T cell persistence, or increased proliferation.^30–35 In the aforementioned cases, improved antitumor activity, enhanced T cell persistence, or increased proliferation compared to controls were achieved to varying degrees, which suggests the potential importance of gel-based approaches. However, further research is needed to elucidate the mechanism of action of many of these platforms and to provide materials that possess more than one of these improvements.

Self-assembling peptide-based scaffolds have been widely studied for their use in cancer treatment, drug delivery applications, and regenerative medicine.³⁶ However, the study of their efficacy in T cell therapies for T cell delivery and activation remains limited. Although there are several examples of the use of peptides as part of more complex systems for immunotherapy and T cell therapies,^{20,23,31,33,37} self-assembling peptide-based hydrogels have only been studied as vaccines or vaccine adjuvants for harnessing the local immune response and the improvement of host T cell function using systems such as RADA16, poly-L-Lysine (PLL) based hydrogels, and naphthylacetic acid modified tetrapeptide of GFFY-based hydrogels.^22,25,26 Because of this, further studies on the use of self-assembling peptide-based scaffolds for the improvement of T cell therapies would be beneficial for developing potential novel approaches for ATCT, especially with well-characterized materials with suitable mechanical properties and adequate biocompatibility. In this study, we assessed the effectiveness of self-assembling multidomain peptide (MDP)-based scaffolds for T cell therapy applications.

MDPs are self-assembling peptides with an ABA motif, where B represents an amphiphilic domain of alternating hydrophilic (serine, S) and hydrophobic (leucine, L) amino acids and A represents the terminal domains that control peptide self-assembly. When in aqueous solution, two MDP chains can form a dimer, as the hydrophobic and hydrophilic side chains orient themselves in opposite directions, forming a hydrophobic core. These dimers then stack together, forming an antiparallel β-sheet with a hydrogen bond network between the peptide backbones. The A domain, by one of two mechanisms, can control the length of the fibers resulting from the stacking of the dimers. When the A domain is composed of charged amino acids, the addition of multivalent counterions is necessary to overcome electrostatic repulsion and molecular frustration between the peptide chains. This allows the fibers to elongate and bundle together to form a nanofibrous hydrogel with about 99% water content.^38,39 Neutral hydroxyproline (Hyp) at a physiological pH (7.2) can also mediate the formation of MDP fibers. Oligohydroxyproline terminal domains control the self-assembly of the fibers by steric effects between the hydroxyprolines, which also allow for increased solubility. O₅(SL)₆O₅ was found to have optimal properties regarding hydrogelation and self-assembly.⁴⁰

MDPs have previously been studied for nerve regeneration,⁴¹ vaccine adjuvants,⁴² diabetic wound healing models,⁴³ angiogenesis and tissue regeneration,⁴⁴ and small molecule delivery.^45,46 The host response has been studied in murine subcutaneous models by Lopez-Silva and collaborators, showing that different chemistries present in the hydrogels govern the elicited immune response.⁴⁷ In addition to their chemical tunability, the resulting hydrogels from both charged and neutral MDPs have a relatively low storage modulus and can be considered to be soft biomaterials. It has previously been reported that soft substrates are more successful at activating T cells than stiff substrates when combined with stimulatory ligands.¹⁷ Because of this and the ease with which MDPs can be used to physically encapsulate different biologics, MDPs constitute a promising material for developing scaffolds for T cell applications.

In this study, we assessed the effectiveness of self-assembling multidomain peptide (MDP)-based scaffolds for T cell therapy applications, specifically evaluating their cytocompatibility, in situ T cell activation in the presence of the corresponding biological cues, maintenance of cytotoxic T lymphocyte (CTL) activity, and persistence and differentiation of T cells over time. Figure 1 shows a schematic of the general approach for developing MDP scaffolds for T cell applications. To the best of our knowledge, this is the first study on peptide-based hydrogels used as 3D culture scaffolds for T cells that evaluates not only the cytocompatibility of the material but also their potential for in vitro activation, T cell proliferation, persistence, and T cell differentiation over time.

Figure 1.

Schematic diagram of the development of MDP-based scaffolds for T cell therapies. (a) Hydrogels are formed from the self-assembly of short synthetic peptides. (b) Charge on the peptide and therefore the entire gel scaffold can be selected as positive, negative, or neutral depending on the choice of amino acids. Cell behavior and cytocompatibility are dramatically affected by this choice. (c) T cells, cancer cells, and a wide variety of proteins can be mixed and colocalized in the MDP hydrogel to evaluate cytocompatibility, proliferation, and antigen-specific cell killing.

EXPERIMENTAL METHODS

Multidomain Peptide (MDP) Synthesis.

The peptides were synthesized following the methodology previously reported by the Hartgerink Lab using solid-phase peptide synthesis with Fmoc chemistry. For simplicity, the terms O₅(SL)₆O₅ will be referred to as O₅, D₂(SL)₆D₂ will be referred to as D₂, and K₂(SL)₆K₂ will be referred to as K₂ in this manuscript. O₅ and D₂ were manually synthesized on a low-loading Rink Amide 4-Methylbenzhydrylamine (MHBA) resin (0.37 mmol, EMD Millipore, Burlington, MA) at a 0.15 mmol scale. A 1:1 solvent mixture of N,N-dimethylformamide (DMF) (Fisher Scientific, Hampton, NH), and dimethyl sulfoxide (DMSO) (Fisher Scientific, Hampton, NH) was used to dissolve all the reagents except diisopropylethylamine (Fisher Scientific, Hampton, NH), which was diluted in DMF. For each coupling, 4 equiv of amino acids (Novabiochem, San Diego, CA), 4 equiv of 2-(7-aza-1-H-benzotriazol-1-yl)-1,1,3,3-tetramethylaminium hexafluorophosphate (HATU) (P3 Bio Systems, Louisville, KY), and 6 equiv of N,N-diisopropylethylamine (DiEA) (Fisher Scientific, Hampton, NH) were added and allowed to react for 25–45 min. Fmoc deprotections were carried out with a solution of 20% piperidine (Sigma-Aldrich, Saint Louis, MO) in 1:1 DMF-DMSO. After deprotecting the last amino acid, the N-terminus was acetylated by adding 5 equiv of acetic anhydride (Fischer Scientific, Hampton, NH) and 6 equiv of DiEA in dichloromethane (Fisher Scientific, Hampton, NH), allowing it to react for 45 min and then repeating the acetylation process. The peptide was cleaved from the resin using a cleavage cocktail with a ratio of 36:1:1:1 trifluoroacetic acid (TFA)/triisopropylsilane/anisole/ethane dithiol/water (Sigma-Aldrich, St. Louis, MO), shaking for 3 h at room temperature. TFA was evaporated from the resulting peptide solution, and the peptide was precipitated using cold diethyl ether (Fisher Scientific, Hampton, NH), drying it overnight after repeated washing and centrifuging.

MDP Purification.

The D₂ peptide was purified via reverse-phase high-performance liquid chromatography (HPLC), using an XBridge Protein BEH C4 OBD Prep Column C4 column (Waters, Milford, MA). Before performing the injection, the crude peptide was solubilized in a 70:30 mixture of HPLC grade water and acetonitrile with 5 mM NH₃OH and 4 mM AcOH and then centrifuged for 2 min before filtering the solution through a 0.2 μm nylon membrane filter. The mobile phase was a 3% gradient of the same solvent system, HPLC grade water, and acetonitrile with 5 mM NH₃OH and 4 mM AcOH. After rotoevaporating the acetonitrile from the purified solution, the sample was frozen at −20 °C for 24 h and −80 °C for at least one h and lyophilized. The identity of the D₂ peptide was confirmed using an Autoflex MALDITOF MS (Bruker Instruments, Billerica, MA), using the reflective negative mode. The peptide was then tested for endotoxin content using a ToxinSensor Gel Clot Endotoxin Assay (GenScript, Piscataway, NJ). The O₅ peptide was purified via dialysis and ion exchange chromatography (IEC). The MDP was dialyzed by solubilizing the peptide after cleavage in deionized water and adjusting to physiological pH (7–7.4) by using 0.1 M NaOH or 0.1 M HCl. The pH-adjusted peptide solution was then transferred to dialysis tubing of 100–500 Da mesh size (Spectrum Laboratories, Rockleigh, NJ). Dialysis was carried out for three to five days in deionized water, replacing it three times a day. The dialyzed O₅ peptide solution was frozen and lyophilized, and the resulting peptide was dissolved in 10 mM potassium phosphate buffer with 3 M guanidine HCl (GuHCl). The solution was then filtered through a 0.2 μm nylon membrane filter and ran through a Sartobind Q75 IEC column (Sartorius, Goettingen, DE) at a pH of 6.31 to remove the endotoxins from the sample. The O₅ solution in buffer and GuHCl was then redialyzed, this time using heat-treated glassware and endotoxin-free water, under sterile conditions.

Circular Dichroism (CD) Spectroscopy.

The MDPs were analyzed with a CD Jasco J-810 spectropolarimeter (Jasco Inc., Easton, MD) to determine their secondary structures. First, 2% peptide solutions were prepared in endotoxin-free water, and then, an equal volume of 2× HBSS (Corning, Corning, NY) was added for a final concentration of 1% MDP and 1× HBSS. D₂ and K₂ peptide solutions formed hydrogels upon the addition of HBSS with counterions, whereas O₅ peptide solutions had to be ultrasonicated to trigger hydrogelation. The hydrogels were placed in a 0.01 cm quartz cuvette for analysis, with data from five scans at room temperature from 180 to 250 nm. The results were presented in terms of molar residual ellipticity (MRE) values.

Attenuated Total Reflectance Fourier Transform Infrared (ATR-FTIR) Spectroscopy.

The Nicolet iS20 FT/IR spectrometer (Thermo Scientific, Waltham, MA) was purged with nitrogen for one hour prior to spectra acquisition, and then a background scan was collected to account for water contribution. Using a pressure tower, powder peptide samples were pressed down onto a Golden Gate diamond window on an ATR stage, and the IR spectra were collected under constant nitrogen flow, with 32 scans accumulation and 4 cm^–1 resolution. Data were processed and background-corrected using the OMNIC spectroscopy software.

Hydrogelation of MDPs.

The peptides were dissolved in endotoxin-free water to a 2% w/v concentration. Then, 1× Hank’s Balanced Salt Solution (HBSS) (Corning, Corning, NY) was added to the stock solution to double the volume to a final concentration of 1% w/v of MDP. In the case of D₂ and K₂, the addition of HBSS triggered hydrogelation due to the presence of counterions. In the case of O₅, hydrogelation was induced following the procedure reported in previous studies in the Hartgerink Lab.⁴⁰ The solution of O₅ was ultrasonicated using a Microson Ultrasonic Cell Disruptor, introducing the 2 mm microprobe into the MDP solution in HBSS. The frequency used was 22.5 kHz with a power output of 10 W (RMS), sonicating during 10 cycles of 10 pulses each and allowing 1 min relaxation times in between.

Oscillatory Rheology.

Hydrogels prepared as described in the previous section were evaluated for their viscoelastic properties using an AR-G2 rheometer (TA Instruments, New Castle, DE). Around 200 μL of MDP hydrogels were prepared 24 h before the analysis. Following the previously reported procedure by the Hartgerink Lab,⁴⁶ 150 μL of MDP hydrogel were transferred into a cut syringe and allowed to rest overnight, then set on the rheometer under a 12 mm stainless steel plate with a 1000 μm gap. This plate was used for both the strain sweep experiments (analyzing the storage and loss modulus from 0.1 to 200% strain at a frequency of 1 rad/s) and shear recovery experiments (evaluating the shear recovery of the sample after keeping it at 1% strain for 30 min, then 200% strain for 1 min, then once more at 1% strain for 30 min).

Cell Culture.

A375 and MEL526 melanoma cell lines, as well as normal donor-derived CD4⁺ T cells, CD8⁺ T cells, and HLA-A2-restricted gp100 specific CTLs, were provided by the Yee Laboratory at The UT MD Anderson Cancer Center (Houston, TX). The gp100-specific CTLs were generated from normal donor PBMC by using autologous dendritic cells pulsed with the A2-restricted peptide epitope of gp100 (G154: KTWGQYWQV) to stimulate PBMC; gp100-specific CD8 T cells were sorted using peptide-MHC tetramers.^48–50 The phenotype of cells was determined using flow cytometry, and the antitumor function of T cells was determined by chromium release assay. A375 (HLA-A2+, gp100-negative) and MEL526 (HLA-A2+, gp100-positive) cell lines were cultured using Roswell Park Memorial Institute (RPMI) 1640 medium (Gibco, Billings, MT), supplemented with 10% fetal bovine serum (FBS) and 2% l-glutamine (200 mM). Cells were cultured in T75 culture flasks in an incubator at 37 °C and 5% CO2 with humidity control and passaged at 80–90% confluency. For each passage, cells were plated with a cell density of 5000 cells per cm² of surface area, and frozen in 10% DMSO and 90% complete RPMI 1640 medium at a cell density of 1 million cells/mL. Normal donor-derived CD4⁺ and CD8⁺ T cells and gp100-specific CTLs were cultured in suspension in RPMI 1640 medium with 10% FBS, 2% l-glutamine (200 mM), and supplemented with IL-2 (50 U/mL). CellTrace Violet (Thermo Scientific, Waltham, MA) staining was performed according to the protocol recommended by the supplier.

Cell Encapsulation in MDP Hydrogels.

Cell Encapsulation in D₂ and K₂ Hydrogels.

CD4⁺ T cells, CD8⁺ T cells, gp100 specific CTLs, and/or MEL526/A375 cells were encapsulated in D₂ and K₂ by preparing a cell suspension in 2× HBSS with calcium and magnesium (Corning, Corning, NY) and adding it to a 2% MDP solution in endotoxin-free water, triggering the hydrogelation of the material and encapsulating the cells at a concentration of 1,000,000 cells/mL in a 1% total MDP hydrogel.

Cell Encapsulation in O₅ Hydrogels.

CD4⁺ T cells, CD8⁺ T cells, gp100 specific CTLs, and/or MEL526/A375 cells were encapsulated in O₅ by first preparing a cell suspension of 10,000,000 cells/mL in 1× HBSS with calcium and magnesium. Next, the peptide was dissolved in 1× HBSS with calcium and magnesium and ultrasonicated as described in Section 2.5 to trigger the hydrogelation using only 90% of the volume of buffer necessary for a 1% MDP solution. The remaining 10% of the HBSS volume needed for the final 1% MDP concentration was then added with the cells for a final concentration of 1,000,000 cells/mL and resuspended in the hydrogel.

Hydrogel Plating for 3D Culture and Imaging.

A total of 70 μL of cell-loaded MDP hydrogel per well was placed in Lab-Tek 16-well glass chambers (0.4 mm², Thermo Scientific, Waltham, MA) and 200 μL of the IL-2 supplemented medium was placed on top of the hydrogel without disrupting it. The glass chambers with the MDP hydrogels were kept in an incubator at 37 °C and 5% CO2 with a humidity control.

Cytocompatibility Studies and CTL Activity Assay.

Cell viability of cells encapsulated in MDP hydrogels was evaluated using a live/dead viability kit (Invitrogen, Thermo Scientific, Waltham, MA). Seven groups were encapsulated in O₅, D₂ and K₂ as described in the previous section: (1) CD4⁺ T cells; (2) CD8⁺ T cells; (3) anti-gp100 CTLs; (4) A375 melanoma cells; (5) MEL526 melanoma cells, (6) gp100-specific CTLs with A375 melanoma cells (gp100–) at effector:target ratios of 1:1 and 5:1; and (7). gp100-specific CTLs with MEL526 melanoma cells (gp100+) at an effector/target ratio of 1:1 and 5:1. Cells were encapsulated for 1, 3, or 5 days, replacing the cell media every 48 h without disrupting the hydrogel. At the corresponding time point, the media was removed and replaced with 100 μL of a 2 μM Calcein AM and 4 μM ethidium homodimer-1 solution in Dulbecco’s Phosphate-Buffered Saline (DPBS). The plates with the MDP hydrogels and the staining solution were placed in the incubator for 20 min, to then remove the staining solution and flip the hydrogels onto a glass slide, to be imaged using a Nikon A1 RSI fluorescent confocal microscope (Nikon Instruments, Melville, NY) using the 488, 561, and 405 nm lasers. Three z-stack images (comprised of 10 individual images) of 100 μm depth were obtained for each of the two biological replicates. They were rendered using the NIS Element software and processed using the FIJI software application.⁵¹ The z-stack images were additionally processed using Imaris 3D/4D Image Processing software (Bitplane, Concord, MA) to generate three-dimensional projections and quantify the cell viability and cell density of the samples. Colocalization of spots was used to distinguish different cell populations in cocultures using the CellTrace Violet stain to distinguish Anti-GP100 CTLs from melanoma cells.

Quantification of CTLs Engaging in Immune Synapses.

Immune synapses present in cocultures of CTLs and melanoma cells were quantified using Imaris 3D/4D Image Processing software (Bitplane, Concord, MA), by generating spots for each population of cells and quantifying the blue spots (CellTrace Violet positive cells) within a 20 μm range from the surface of green stained cells (live cells). To account for Calcein AM stained CTLs and ensure only immune synapses (CTL-target cell contacts) were quantified, the colocalized spots between the blue and green populations were also quantified and subtracted from the total number of events under the distance threshold.

Antibody and Cytokine-Mediated Activation of CTLs in MDP Hydrogels.

The potential for in situ activation of CTLs in MDP hydrogels was tested by encapsulating CTLs (prestained with CellTrace Violet) in MDPs along with CD3 and CD28. 70 μL of hydrogel was plated with 200 μL of RPMI 1640 medium supplemented with 10% FBS and 2% l-glutamine (200 mM). Additionally, different combinations of IL-2 (50 U/mL), IL-7, and IL-15 were added to the medium. The resulting proliferation of CTLs was first assessed using flow cytometry, by evaluating the fluorescence intensity of CellTrace Violet positive cells and comparing different time points. We also quantified the expansion index (total fold increase from day 0) of the CTL cultures using confocal imaging and Imaris 3D/4D Image Processing software (Bitplane, Concord, MA) and compared the cell densities at the beginning of the experiment and at day 5 postencapsulation. In order to extract the cells from the hydrogels after confocal imaging, the MDP hydrogels were disrupted using 1× PBS buffer (5 mL) and centrifuged, the cells were washed twice with 1× HBSS and then fixed with a 4% paraformaldehyde (PFA) solution.

Phenotype Analysis.

T cells were recovered from the hydrogels using the approach described previously and were stained with a cocktail of antibodies including antihuman CD3 (Pacific Blue, 1:50), CD8 (BV570), CD27 (BV615), CD28 (PE-Cy7), CD137 (APC), and CD107a (FITC) (Biolegend, San Diego, CA). Stained samples were acquired on a Novocyte flow cytometer (Agilent, Santa Clara, CA) and data were analyzed using FlowJo (version 10.7.1) software (BD Biosciences, Ashland, OR).

Statistical Analysis.

The experimental values, in all cases, were averaged and displayed with error bars representing standard deviation. Statistical comparisons between groups were performed using unpaired t tests with Welch correction in GraphPad Prism v. 9.2.0. P-values <0.05 were considered significant for the experiments presented in this paper.

RESULTS AND DISCUSSION

MDP Synthesis, Characterization of Secondary Structure, and Hydrogelation.

Peptides were synthesized using solid-phase peptide synthesis and purified by either HPLC (K₂ and D₂) or dialysis (O₅). Their molecular weight was confirmed using MALDI-TOF mass spectrometry, and the resulting spectra, as well as the chromatograms from purification, are shown in Supporting Information. Peptides were then tested for endotoxin content using a ToxinSensor Gel Clot Endotoxin Assay. HPLC purification protocol was sufficient to reduce the endotoxin levels to <25 EU/mL in the case of 1% w/v D₂ and 1% w/v K₂. In contrast, the purification of O₅ required ion exchange chromatography (IEC) followed by dialysis under endotoxin-free conditions.

The secondary structure of the resulting endotoxin-free MDPs was characterized using circular dichroism (CD) and attenuated total reflectance Fourier-transform infrared spectroscopy (ATR-FTIR) to ensure batch consistency for these MDP formulations. As expected, the CD data for charged MDPs showed canonical β-sheet signatures, with a maximum at 197 nm and a minimum at 218 nm.⁵² The CD spectra for O₅ presented a maximum at 195 nm and a minimum at 216 nm, respectively. This is consistent with previously reported values for O₅ in water at pH 7 and is also indicative of a β-sheet conformation.⁴⁷ The ATR-FTIR spectra of the MDPs confirmed their β-sheet character through the peaks in the amide I band, in the range between 1618 and 1635 cm⁻¹, and the antiparallel orientation of the β-sheet can be observed by the presence of a shoulder at 1695 cm ¹.³⁸ Taking both CD and ATR-FTIR into consideration, the predominant secondary structure of MDPs prepared in 1× HBSS was found to be an antiparallel β-sheet. Similarly, the rheology measurements for all tested MDPs were consistent with previously published G′ of 204 Pa, 90, and 23.3 Pa for K₂, D₂, and O₅, respectively.^40,41 The characterized endotoxin-free MDP hydrogels proved to be shear-thinning, injectable materials with good shear recovery postshearing, suitable for biological applications.

Impact of MDP Chemistry on Cytocompatibility with T Cells in 3D Culture.

To evaluate the cytocompatibility of differently charged MDPs for T cell culture, CD8⁺ and CD4⁺ T cells were encapsulated in K₂, O_5, and D₂ hydrogels at a cell density of 1 million cells per milliliter. The resulting cell-laden hydrogels were plated in a 16-well glass chamber (0.4 mm²) and 3D cultured with 10% FBS supplemented RPMI media with IL-2 (50 IU/mL), for 1 to 5 days, replacing the media every other day. At the indicated time point, the cell media was removed, and live/dead staining was performed to image the cells using confocal microscopy. A representative sample of the resulting z-stacks for CD8⁺ T cells, formed by 10 individual images across a 100 μm depth of the cell-laden hydrogels, is presented in Figure 2a, and the corresponding images for CD4⁺ T cells are found in Figure S5. K₂ showed predominantly dead cells stained red by ethidium homodimer, as early as day 1 and continuing through day 5. In contrast, both the ligands O₅ and D₂ show predominantly live cells stained green by Calcein AM, and the cells remain viable through the entirety of the experiment. D₂ showed higher cell viability compared to that of O₅ on day 3 and day 5, and a significantly lower cell viability was observed for CD8⁺ cells (compared to CD4⁺ cells) by day 5 in the O₅ hydrogel. Percent viability graphs for CD4⁺ and CD8⁺ T cells encapsulated from Day 1 to 5 in the three evaluated MDPs are shown in Figure 2b. With these results, it is possible to conclude that among the three MDPs evaluated, only the neutral and anionic peptide-based hydrogels (O₅ and D₂) are suitable for T cell culture applications, as the cationic peptide (K₂) showed high cytotoxicity when encapsulating human T cells. This adverse effect on cell viability may be due to the impact of the cationic peptides on the cell membrane,⁵³ where the charged lysine residues may interact with lipids causing membrane disruption. Poly-l-lysine is also known to promote molecular weight-dependent, mitochondria-mediated apoptosis which could possibly be involved here as well.⁵⁴ Other cell viability patterns could be due to the effect of different chemistries presented by the terminal domains of the MDPs, which have been previously reported to govern the immune response in vivo.⁴⁷ Based on these results, only the O₅ and D₂ were used for the remainder of the study.

Figure 2.

3D culture of CD8⁺ and CD4⁺ T cell populations in MDP hydrogels. (a) Cell viability of CD8⁺ T cells encapsulated in O₅, K₂, and D₂ with IL-2 supplementation (50 IU/mL) at days 1, 3, and 5, by staining with Calcein AM and ethidium homodimer (viable cells shown in green, dead cells in red). Scale bar: 100 μm. (b) Cell viability quantification at day 1, 3, and 5 for in O₅, K₂, and D₂ postencapsulation for CD4⁺ (blue) and CD8⁺ (orange) T cells. Error bars correspond to standard deviation (unpaired t tests with Welch correction, *p < 0.05, **p < 0.01, ****p < 0.0001).

MDP Hydrogels Cytocompatibility with CTLs and Target Cells.

Because gp100 is a well-characterized melanoma protein that is highly expressed in melanoma tumors and is being targeted using immunotherapy approaches including vaccines and T cell therapy,⁵⁵ we used gp100-specific cytotoxic T lymphocytes (CTLs) and melanoma cells presenting the gp100 antigen as the model for this in vitro study. To demonstrate the antigen-specific function of T cells in the MDPs, we cocultured them with melanoma cell line MEL526, expressing cognate antigen gp100, and A375 cells, which do not express the antigen.

First, we determined the cytocompatibility of the hydrogels with CTLs and both melanoma cell lines. gp100-specific CTLs, MEL526, and A375 were encapsulated separately in O₅ and D₂ hydrogels at a cell density of 1 million cells per mL and cultured with IL-2 supplemented media for 1–5 days. The cells were imaged at different time points by using confocal microscopy. The resulting images are shown in Supplementary Figures S6 and S7. Neutral and anionic MDPs alone proved to be nontoxic to gp100-specific CTLs, especially in the case of D₂, but the T cells do not actively proliferate when encapsulated in MDP scaffolds alone. None of the MDP hydrogels demonstrated cytotoxic activity against melanoma cell lines A375 and MEL526, which were both highly viable and proliferated rapidly. Future investigations could potentially benefit from their examination as a scaffold for the proliferation and formation of spheroid models.⁵⁶

CTL Anticancer Activity in 3D MDP Scaffolds.

Since cytocompatibility of MDPs for gp100-specific CTLs and melanoma cells was established separately and no cytotoxic effect of MDPs was observed, the next step was to evaluate the maintenance of antigen selectivity of CTLs in MDPs. CTLs were encapsulated in MDP scaffolds with each cancer cell line at different effector/target cell (E/T) ratios, and antigen-specific cytolytic activity was assessed. The coculture studies were carried out by encapsulating (in either O₅ and D₂) gp100-specific CTLs with MEL526 (gp100+) and A375 (gp100−) cells at an E:T ratio of 1:1 and 5:1, as summarized in Table 1.

Table 1.

Experimental Groups of Anti-Gp100 CTLs Cocultured with Cancer Cells in O₅ and D₂, and the Corresponding Results

MDP	cancer cell line	E:T ratio	group number	result
D2	A375(gp100−)	1:1	1	no cell killing
		5:1	2
	MEL526 (gp100+)	1:1	3	specific cell killing
		5:1	4
O5	A375(gp100−)	1:1	5	nonspecific cell killing
		5:1	6
	MEL526 (gp100+)	1:1	7	specific cell killing

Cells were encapsulated, cultured, and imaged as previously described, and the cell density was quantified. The results are presented in Figures 3 and 4. In D₂, there was an increase in the density of A375 melanoma cells for Group 1 and Group 2, and no effect of T cells cytolytic activity was observed (Figure 3a). In the case of Group 3 and Group 4, there was a measurable reduction in MEL526 cell density when cocultured with gp100 targeting CTLs (Figure 3b). In terms of cancer cell density in D₂, Group 1 and Group 2 showed significant differences when compared to Group 3 and Group 4, at time points day 3 and day 5 (Figure 3c). On the other hand, in O₅ the cell density of A375 either remained stable or decreased for Group 5 and Group 6, respectively (Figure 4a), whereas MEL526 melanoma cells showed a measurable reduction in cell density for Group 7 and Group 8, as observed with D₂ (Figure 4b). Group 7 cells showed a significantly higher cell density compared to group 5 at day 1, and that behavior was inverted at day 3. However, the proliferation by day 5 showed no significant differences for any of the O5 groups (Figure 4c). For both D₂ and O₅, imaging data also revealed the formation of immune synapses characterized by aggregation of T cells around the antigen-matched cells (MEL526) that were not as visibly present in the antigen-matched cells. Considering that CTLs require antigen recognition for contact killing,⁵⁷ this is an important observation and was further evaluated as described in the next section. From these results, it can be concluded that the antigen selectivity of the encapsulated CTLs was better maintained by the D₂ hydrogels, in which the CTLs were able to lyse the antigen-expressing melanoma cells (MEL526) but did not engage in the nonspecific killing of the antigen mismatched tumor cells (A375), something we did observe in the case of O₅ encapsulated CTLs.

Figure 3.

Coculture of melanoma cells with gp100 CTLs. (a) 3D culture of gp100 CTLs with A375 cells and (b) 3D culture of gp100 CTLs with MEL526 cells in D₂ hydrogel with IL-2 supplementation (50 IU/mL) at days 1, 3 and 5, by staining with Calcein AM and ethidium homodimer (viable cells shown in green, dead cells in red). Scale bar = 100 μm, effector/target (E/T) ratios of 1:1 and 5:1. (c) Viable cell density of A375 (blue) and MEL526 (orange) cocultured with gp100 CTLs in D₂ at E/T ratios of 1:1 and 5:1. Error bars correspond to standard deviation (unpaired t tests with Welch correction, *p < 0.05, **p < 0.01, ****p < 0.0001).

Figure 4.

Coculture of melanoma cells with gp100 CTLs. (a) 3D culture of gp100 CTLs with A375 cells and (b) 3D culture of gp100 CTLs with MEL526 cells in an O₅ hydrogel with IL-2 supplementation (50 IU/mL) at days 1, 3, and 5, by staining with Calcein AM and ethidium homodimer (viable cells shown in green, dead cells in red). Scale bar = 100 μm, effector/target (E/T) ratio of 1:1 and 5:1. (c) Viable cell density of A375 (blue) and MEL526 (orange) cocultured with gp100 CTLs in O₅ at an E/T ratio of 1:1 and 5:1. Error bars correspond to standard deviation (unpaired t tests with Welch correction, *p < 0.05, **p < 0.01, ****p < 0.0001).

Quantification of Immune Synapses in 3D Cultures.

An immune synapse is a stable junction between a T lymphocyte (or other thymus-derived cells) and an antigen-presenting cell (such as the gp100 expressing MEL526 melanoma cells) and has an important role in target cell killing, cytokine secretion, and T cell differentiation.^58,59 In addition to determining the antitumor cytolytic function of CTLs encapsulated in MDP hydrogels, we also quantified the physical interaction of T cells with tumor cells by determining the number of immune synapses present in different coculture conditions. Considering the sizes and distances observed in one-to-one immune synapses, the quantification threshold was optimal at 20 μm from the modeled cell surfaces, as shown in Figure 5a. Some of the representative immune synapses in D₂ are shown in Figure 5b, while the representative immune synapses in O₅ are found in Supporting Information. The immune synapse formation ranged from engagement of a single CTL with a target cell to multiple CTLs associated with a cluster of target cells. The results are expressed as the percentage of Anti-Gp100 CTLs engaging in immune synapses for the different experimental conditions at day 1, 3, and 5 of coculture (Figure 5c). In all conditions evaluated, a higher percentage of immune synapse formation was observed for cocultures containing the matched antigen cells (MEL526) compared to those with mismatched antigen cells (A375) regardless of the E/T ratio. Based on the observation that CTLs predominantly form more immune synapses with MEL526 than with A375 cells for all of the groups, these results suggest the antigen recognition of the encapsulated CTLs was maintained in both D₂ and O₅ MDP formulations and that the nonspecific killing of A375 in the O5 formulation may be attributable to the induction of bystander apoptosis^57,60 (Figure S9).

Figure 5.

Immune synapses quantification. (a) Schematic of the quantification procedure. gp100 CTLs are prestained with CellTrace Violet (shown in blue), and the samples were live/dead stained with Calcein AM and ethidium homodimer (viable cells shown in green, dead cells in red). The amount of gp100 CTLs engaging in immune synapses is quantified as all CellTrace Violet stained cells within 20 μm of the surface of viable, Calcein AM stained target cells (CellTrace Violet negative). (b) Representative images of immune synapses, or lack thereof, for cocultures containing MEL526 or A375 in D₂. (c) Percent gp100 CTLs engaged in immune synapses at days 1, 3, and 5 in coculture with A375 (blue) or MEL526 (orange), at E/T ratios of 1:1 and 5:1. Error bars correspond to standard deviation (unpaired t tests with Welch correction, *p < 0.05, **p < 0.01, ****p < 0.0001).

In Situ Proliferation of Cytotoxic T Lymphocytes in MDP Hydrogels.

In addition to assessing the maintenance of antitumor function of CTLs encapsulated in MDPs, we evaluated the potential of MDP hydrogels as scaffolds for in situ T cell activation and proliferation. In addition to tumor cells, we used anti-CD3 and anti-CD28 to simulate TCR engagement.⁶¹ gp100-specific CTLs (prestained with CFSE alternative CellTrace Violet) were encapsulated in the anionic MDP D₂ along with anti-CD3 and anti-CD28 and cultured with cell media supplemented with the cytokines IL-2, IL-7, and/or IL-15. The cell-laden hydrogels were imaged on days 1, 3, and 5, as well as extracted from the MDP hydrogel and evaluated for CellTrace Violet intensity using flow cytometry. CellTrace Violet enables monitoring of proliferating cells since each dye is diluted by half in subsequent generations, and the intensity of the dye can be used to track the proliferation of a population. The resulting proliferation histograms and quantified expansion indexes (total fold increase from day 0) based on quantified cell density are shown in Figure 6. Figure 6a shows the histograms for CTLs in the presence of cytokines IL-2, IL-7, and IL-15 either alone or in combination, encapsulated with CD3 and CD28 and cocultured with melanoma cells (to monitor their cytotoxicity). The corresponding histograms for A375 containing cultures are found in Figure S10. The histograms were normalized to Unit Area to account for different cell amounts across populations (between 10,000 and 20,000). As the number of cytokines (IL-2, IL-7, and/or IL-15) in the cell media increased, the intensity of the CellTrace Violet peak began to decrease, demonstrating the cells were actively proliferating when in the presence of coactivator cytokines IL-15 and/or IL-7, whereas IL-2 had minimal effect on proliferation. This was further explored by quantifying the cell density of CTLs within the hydrogels using confocal microscopy, as described in previous sections. The expansion indices were calculated as previously described,⁶² and Figure 6b,c show the resulting expansion indices for CTLs cultured with different cytokine combinations. In the case of CTLs cocultured with MEL526, a statistically significant higher expansion was observed in all conditions when cytokines were included in the culture media compared to the negative control condition with media alone. Such an increase in the presence of IL-2, IL-7, or IL-15 alone in the media was not observed with cocultures containing A375. In cocultures with A375, a significant increase in expansion indices was only observed when two or more cytokines were present in the culture media. From the outcomes of the proliferation assay via flow cytometry and the expansion indexes calculated via confocal microscopy, in the presence of CD3, CD28, and the appropriate cytokines, it is possible to induce the proliferation of CTLs in anionic MDPs. The expansion indices obtained are in the order of magnitude of those reported by some published systems in the literature,^28,63 in some cases lower and in some cases higher. However, most published works do not include an in vitro evaluation of the preservation of T cell antitumor activity (and selectivity), which was previously evaluated for our system.

Figure 6.

In situ activation and expansion of CTLs in D₂ MDP hydrogel. (a) Histograms of CellTrace Violet Anti-Gp100 CTLs encapsulated with CD3 and CD28 and cocultured with MEL526. The cell medium was supplemented with IL2, IL7, and/or IL15, testing all the permutations (treatments A–H). Proliferation expansion index (calculated as the fold increase by Day 5) for gp100-specific CTLs encapsulated with CD3 and CD28 and cocultured with (b) A375 or (c) MEL526 (unpaired t tests with Welch correction, *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001).

Phenotype Analysis of CTLs Encapsulated in MDPs.

To evaluate the phenotype of the T cells cocultured with tumor cells in the different MDP hydrogels, T cells were extracted after 10 days of encapsulation (to assess longer-term T cell persistence) and stained with a panel of antibodies including antihuman CD8, CD27, CD28, CCR7, and CD62L. The activation status of the cells was also determined based on the expression of degranulation marker CD107a. The results are shown in Figure S11. CD27 expression was upregulated in T cells cocultured with MEL526 in anionic MDP D₂ compared to T cells cocultured with A375 in anionic MDP D₂, demonstrating increased antigen-mediated T cell activation for cells cocultured with MEL526.⁶⁴ The expression of CD107a on the T cells cocultured with MEL526 in D₂ was also elevated compared to cells cocultured with A375. CD107a is another marker that indicates T cell activation in the cocultured CTLs and is also known to protect cytotoxic lymphocytes from damage during degranulation, explaining the maintenance of CTL viability in the hydrogels.⁶⁵ Similar differences in the expression profile of CD27 and CD107a in T cells cocultured with MEL526 or A375 in O₅ were not observed. These results suggest that TCR-mediated activation in D₂, indicated by the upregulation of CD27 and CD107a, enables the persistence of T cells.

CONCLUSIONS

Human CD4⁺ and CD8⁺ T lymphocytes were encapsulated in positively charged, neutral, and negatively charged MDPs (K₂, O_5, and D₂), and showed high cell viability up to day 5 in noncationic materials with T cells demonstrating the cytocompatibility of O₅ and D₂. The maintenance of T cell cytolytic activity within the hydrogels was evaluated using a panel of coculture conditions comprising MEL526 cells (expressing gp100 antigen), A375 cells (lacking gp100 antigen), and gp100-specific cytotoxic T lymphocytes. gp100-specific cytotoxic T cells (CTLs) were encapsulated by themselves or in coculture with the melanoma cell lines to assess the maintenance of T cell antigen-specific activity in the hydrogel through cell viability, cell density, and immune synapse quantification. From day 1–day 5, the CTLs exhibited antigen-specific activity in D₂ hydrogels, characterized by increased cytotoxicity targeting MEL526 cells. Conversely, the CTLs were cytotoxic to both A375 and MEL526 cells in the O₅ hydrogels. In addition to this, a statistically greater proportion of immune synapse formation was observed for MEL526-containing cocultures in comparison to A375-containing cocultures. This indicates the maintenance of antigen recognition of the encapsulated CTLs for both MDPs, despite possible bystander nonspecific killing in the case of A375-containing cultures in O₅. Proliferation assays were performed via flow cytometry and confocal microscopy, encapsulating CTLs in the presence of either A375 or MEL526 cancer cells, as well as CD3 and CD28, and different combinations of IL-2, IL-7, and IL-15. Under these conditions, we show that CTLs can proliferate in D₂. Finally, to understand why these two MDPs obtain different responses and selectivity from CTLs, we studied their phenotype after 10 days of encapsulation in the presence of either A375 or MEL526. The results demonstrated that D₂ enables upregulation of CD27 and CD107a that supports proliferation, as well as the survival of antigen-specific T cells. This study provides the first approach to self-assembling peptide-based hydrogels used as 3D scaffolds for T cell culture, proliferation, and potential delivery for adoptive T cell therapy while also studying the effect of the material on the phenotype of the cells and their persistence in vitro.