Fucosylation Enhances the Efficacy of Adoptively Transferred Antigen-specific Cytotoxic T Lymphocytes

Abstract

Purpose:

Inefficient homing of adoptively transferred cytotoxic T lymphocytes (CTLs) to tumors is a major limitation to the efficacy of adoptive cellular therapy (ACT) for cancer. However, through fucosylation, a process whereby fucosyltransferases (FTs) add fucose groups to cell surface glycoproteins, this challenge may be overcome. Endogenously fucosylated CTLs and ex vivo fucosylated cord blood stem cells and regulatory T cells were shown to preferentially home to inflamed tissues and marrow. Here we show a novel approach to enhance CTL homing to leukemic marrow and tumor tissue.

Experimental Design:

Using the enzyme FT-VII, we fucosylated CTLs that target the HLA-A2-restricted leukemia antigens CG1 and PR1, the HER2-derived breast cancer antigen E75, and the melanoma antigen gp-100. We performed in vitro homing assays to study the effects of fucosylation on CTL homing and target killing. We used in vivo mouse models to demonstrate the effects of ex vivo fucosylation on CTL anti-tumor activities against leukemia, breast cancer and melanoma.

Results:

Our data show that fucosylation increases in vitro homing and cytotoxicity of antigen specific CTLs. Furthermore, fucosylation enhances in vivo CTL homing to leukemic bone marrow, breast cancer and melanoma tissue in NOD/SCID gamma (NSG) and immunocompetent mice, ultimately boosting the anti-tumor activity of the antigen-specific CTLs. Importantly, our work demonstrates that fucosylation does not interfere with CTL specificity.

Conclusion:

Together, our data establish ex vivo CTL fucosylation as a novel approach to improving the efficacy of ACT, which may be of great value for the future of ACT for cancer.

Introduction

The success of adoptive cellular therapy (ACT) for cancer has been impeded by the paucity of T cells homing to tumor tissue (1,2). A number of molecules, including T cell adhesion molecules, chemokine receptors, and selectins, have been shown to play important roles in the homing of T cells to malignant tissues (3-5). Of these, selectins appear to be the major molecules involved in T cell homing into tumors and inflamed tissues (6-10). T cell expression of selectin ligands is critical for T cell efficacy in providing long-term protection against tumor recurrence. Selectins are a family of three C-type lectins (P-selectin, E-selectin, and L-selectin) expressed by hematopoietic, immune, and endothelial cells.

Selectins have been shown to play a role in T cell function, including trafficking (6,7). They bind to the tetrasaccharide sialyl-Lewis X (sLeX), which decorates certain proteins that are expressed on T cells and tumor endothelial cells. sLeX is the moiety found on selectin ligands such as PSGL-1 and CD44, which are the critical binding units that mediate the interaction between circulating cells with E- and P-selectins expressed on vascular endothelial cells and inflamed tissues (11,12). This cooperative binding leads to the attachment, rolling, and adherence of cytotoxic T lymphocytes (CTLs) to vascular endothelial cells, supporting the first critical step in the homing process.

sLeX formation is controlled in part by the expression of fucosyltransferases (FTs) that attach a terminal fucose to trisaccharide acceptor molecules. In T cells, this activity is regulated by FT-IV and FT-VII (13). Fucosylation of T cell surface molecules has been shown to facilitate the migration of T cells into tissues (13-18). In contrast, defects in fucosylation contribute to decreased T cell homing (19-21).

Currently available data highlight the importance of endogenous fucosylation of selectin ligands in T cell homing and trafficking; however, to date, ex vivo fucosylation has been studied only in the setting of allogeneic stem cell transplantation (allo-SCT) (22-24). After validating the effects of ex vivo fucosylation in animal models, one study showed that ex vivo fucosylation of cord blood hematopoietic stem cells shortened time to engraftment following allo-SCT in 22 patients (24). Moreover, Parmar et al. showed that ex vivo fucosylation of regulatory T cells (T-regs) enhances homing into inflamed tissues affected by graft-versus-host disease (GvHD) in a xenograft mouse model (25). In these studies, fucosylation was achieved by a simple reaction involving a short incubation of cells with the substrate guanosine diphosphate-fucose (GDP-fucose) and FT-VI (TZ-101: FT-VI + GDP fucose). Since FT-VII fucosylates CTLs more efficiently than FT-VI, we used FT-VII (TZ102: FT-VII + GDP fucose) to fucosylate CTLs ex vivo in this study (22-25). Incubating cells with TZ102 results in an enzymatically mediated, site- and stereo-specific addition of fucose to form the tetrasaccharide sLeX.

We hypothesized that fucosylation of antigen-specific CTL in the setting of leukemia and breast cancer enhances their homing into tumor tissues and their anti-tumor activities. Using CTL that target the human leukemia antigens PR1 and CG1 (PR1- and CG1-CTL)(26-30), the human breast cancer antigen E75 (E75-CTL)(31,32), and the mouse melanoma antigen gp-100 (pmel-1 CD8⁺ T cell)(33), we show that fucosylation of CTLs results in: (A) increased migration and cytotoxicity of antigen-specific CTLs following fucosylation using in vitro assays; (B) favorable changes in the expression of CTL adhesion molecules, co-stimulatory receptors, CTL cytolytic granules and CTL:target synapse formation; (C) enhanced killing of leukemia, breast cancer and melanoma by CG1-CTL, PR1-CTL, E75-CTL and pmel-1 CD8+ T cells in vivo; and that (D) fucosylation does not alter the specificity of the antigen-specific CTL for their targets and does not increase homing of CTL to normal mouse tissue.

Materials and methods

Cells and cell culture

The U937 (histiocytic leukemia), SKBR3 (breast cancer), and T2 (HLA-A2⁺ T/B cell hybridoma) cell lines were obtained from American Type Culture Collection (ATCC, Manassas, VA, USA). U937 and T2 cell lines were grown in RPMI-1640 supplemented with 10% fetal bovine serum (FBS) (Gemini Bio-Products, Sacramento, CA, USA), 100 U/mL penicillin, and 100 μg/mL streptomycin (Cellgro, Manassas, VA, USA); SKBR3 breast cancer cell line was grown in Dulbecco’s Modified Eagle’s Medium (DMEM) supplemented with the same. All cell lines were cultured and maintained in 5% CO₂ at 37°C. The HLA-A2⁻ cell line U937 and the HLA-A2^low SKBR3 (SKBR3 is genotypically one allele HLA-A2 positive, but phenotypically HLA-A2 low) were transduced with HLA-A*0201 (i.e. HLA-A2) using lentiviral transduction as previously described (28,34). Cell lines were validated using short tandem repeat DNA fingerprinting by our institutional sequencing facility. After thawing, cells were used for experiments within 5 passages. Mycoplasma testing was performed quarterly using the MycoAlert Kit (Lonza Inc., Allendale, NJ, USA). Patient peripheral blood and apheresis samples were obtained through an Institutional Review Board approved protocol.

Peptide-specific CTL generation

CG1-CTLs, PR1-CTLs, and E75-CTLs were generated using the modified dendritic cell (DC)-based CTL expansion method (35). HLA-A2⁺ peripheral blood mononuclear cells (PBMCs) from healthy donors were isolated by Histopaque (Sigma Aldrich, St. Louis, MO) density gradient cell separation. An enriched population of CD14 monocytes was prepared by magnetic bead depletion of CD2, CD19, and CD8-expressing cells (Dynal-Invitrogen, Carlsbad, CA). Enriched monocytes were then cultured in Macrophage Serum-Free Media (Gibco-Invitrogen, Carlsbad, CA) with 100 ng/mL GM-CSF and 100 ng/mL IL4 (R&D Systems, Minneapolis, MN). TNFα (R&D Systems) was added at 10 ng/mL after 48 hours to mature the DC population. Antigen-specific CD8+ T cells were prepared from PBMCs by supplementing complete media with 10 ng/ml of IL7 (R&D Systems) and pulsing the cells with 40 μg/ml of peptide (CG1-FLLPTGAEA/ PR1-VLQELNVTV/ E75-KIFGSLAFL, Biosynthesis, Lewisville, TX). DCs were harvested on day 5 and pulsed with the above peptides for 2 hours. Once pulsed, the DCs were then co-cultured with autologous CD8+ T cells in complete media containing 10 ng/ml of IL7 and 50 IU/mL of IL2. All cultures were maintained in a humidified incubator in 5% CO₂ at 37°C. On day 14, CTLs were harvested and used in in vitro assays and in vivo studies. Prior to use, CTLs were passed through a negative selection column (MACS Miltenyi Biotec- CD8+ T Cell Isolation Kit, Auburn, CA).

Fucosylation of CTLs

Ex vivo expanded T cells were incubated in fucosylation solution: 20 μg/mL of FT-VII in 1 mM GDP Fucose in phosphate-buffered saline (PBS) with 1% human serum albumin (Targazyme Inc, Carlsbad, California) at room temperature for 30 minutes, as previously described (25). FT-VII was used since it fucosylates CTLs at a much higher efficiency than FT-VI. Cells were then re-suspended in PBS. Fucosylation was confirmed using flow cytometry (LSR Fortessa; BD Biosciences, San Jose, CA) after the cells were stained with the FITC-conjugated HECA-452 antibody (BD Biosciences), which targets cutaneous lymphocyte antigen (CLA), shown to be sLeX on PSGL-1 (14).

CTL Migration Assay

CTL migration was assessed using a CytoSelect Leukocyte Transmigration assay (Cell Biolabs, Inc., San Diego, CA). Human umbilical vein endothelial cells (HUVECs) (1 × 10⁵) were cultured in each of 24 trans-well inserts for 24 hours. Antigen-specific CTLs labeled with LeukoTracker dye were then placed into each inner well, in contact with full serum media below. Cells that had migrated through the membrane and into the media were lysed with specific lysis buffer, and the fluorescence was measured with a plate reader at 480/520 nm (BioTek Cytation3, Winooski, VT).

CTL Phenotypic Analysis

CTL (1.5 × 10⁶) were stained for molecules that modulate T cell trafficking, including CD49d (clone 9F10; BioLegend, San Diego, CA, USA), CD162 (PSGL-1; clone KPL-1; BioLegend), CD183 (CXCR3; clone 1C6/CXCR3; BD Biosciences), and CD195 (CCR5; clone 2D7/CCR; BD), as well as molecules involved in co-stimulation/inhibition, including CD137 (41BB; clone 5F4; BioLegend), CD279 (PD1; clone EH12.2H7; BioLegend), and CD357 (GITR; eBioAITR; eBioscience, San Diego, CA), within 2 hours after fucosylation. The LIVE/DEAD™ Fixable Aqua Dead Cell Stain Kit (Life Technologies, Eugene, OR) was used to assess cell viability. Flow cytometry was done on live cells using a BD LSR Fortessa, and data was analyzed using FlowJo software (FlowJo, Ashland, OR).

CTL Cytotoxicity Assay

Antigen-specific CTL were harvested on days 12-14 and passed through a magnetic separation column by human CD8⁺ T cell Isolation kit (Miltenyi Biotec). Peptide-specific cytotoxicity was assessed with a standard 4-hour calcein-AM release assay as previously described (27,36,37). Target cells included peptide-pulsed T2, SKBR3-A2 breast cancer cells, primary patient acute myeloid leukemia (AML) cells, or U937-A2 leukemia cell line. Target cells (1 × 10³) were fluorescently labeled with calcein-AM (Invitrogen) for 15 min at 37°C and washed with RMPI-1640 to remove free calcein-AM. These cells were seeded into 60-well Tarasaki plates at indicated effector-to-target (E:T) ratios for 4 hours at 37°C. The reaction was quenched with trypan blue, and fluorescence was measured using a microplate fluorescence reader (BioTek Cytation3, Winooski, VT). The percent specific cytotoxicity was calculated using the following formula:

$$\begin{array}{cc}\hfill & \left(\right[1-fluoresenc{e}_{target+effector}-fluoresenc{e}_{media}]∕[fluoresenc{e}_{targetalone}-\hfill \\ \hfill & fluoresenc{e}_{media}\left]\right)\times 100\hfill \end{array}$$

In vivo Mouse Methods

NOD/SCID gamma (NSG) and NSG-HLA-A2 transgenic mice (NSG-A2), 4-6-week-old female mice (28,38), were purchased from Jackson Laboratory (Bar Harbor, ME). For AML experiments, HLA-A*0201 primary AML samples with a high leukemia burden or U937-A2 (U937-A2⁺/GFP⁺) cells were administered via tail vein at doses of 1 × 10⁶ to 1 × 10⁷ cells to sublethally irradiated (250 cGy) mice. Mice were monitored 2 times/week for AML engraftment by peripheral blood analysis and GVHD by clinical assessment. Fucosylated or non-fucosylated CG1- or PR1-CTLs (0.5 × 10⁵) (28,39) were administered to mice intravenously (IV) via tail vein 2 days after leukemia injection. Mice were sacrificed on day 3 to evaluate CTL homing and on week 2 to evaluate for tumor killing. Bone marrow was stained with human (h)CD13, hCD33, hCD3, hCD8 (BD Biosciences), hCD45 (BioLegend) and mouse (m)CD45 antibodies (eBioscience), and assessed for GFP to quantify xenograft disease and CTL homing into tumor (28,40). Residual AML was classified as hCD33⁺/hCD45⁺/mCD45⁻ and antigen-specific CTL were classified as hCD33⁻/hCD45⁺/mCD45⁻/hCD13⁻/hCD3⁺/hCD8⁺.

For breast cancer, 1.5 × 10⁷ luciferase-expressing SKBR3-A2 breast cancer cells were injected into NSG mice orthotopically on day 0. On days 7 and 10, fucosylated or non-fucosylated E75-CTLs (2 × 10⁶) were injected IV via tail vein. Mice were followed for weight loss, tumor size, survival, and tumor bioluminescence every three days. Mice were sacrificed on week 5, and tumor-infiltrating lymphocytes (TILs) were extracted from primary tumors using a tumor dissociation kit (Miltenyl Biotec, Auburn, CA). TILs were stained with the same panel used for bone marrow samples in AML mice, and E75-CTL were classified as mCD45⁻/hCD45⁺/ hCD3⁺/ hCD8⁺.

For melanoma, immunocompetent mouse model, provided by Dr. Willem Overwijk, pmel-1 T cell receptor (TCR) transgenic mice on C57BL/6 background (The Jackson Laboratory) were crossed with CD90.1 congenic mice to yield pmel-1^+/+CD90.1^+/+ mice (referred to as pmel-1 mice)(33). Splenocytes from pmel-1 mice were extracted and then expanded using standard protocol (33). Concurrently, 0.3 × 10⁶ B16-F10 melanoma cells (ATCC) were inoculated subcutaneously into C57BL/6 mice (Charles River Laboratories, Wilmington, MA). After one week of pmel-1 CD8⁺ T cell ex vivo expansion, tumor bearing mice received IV tail vein injection of 5 × 10⁶ fucosylated or non-fucosylated Pmel-1 T cells (day 0). Mice underwent intraperitoneal injections of IL-2 (0.1 × 10⁶ IU/mice; Prometheus Therapeutics & Diagnostics, San Diego, CA) twice daily for two days. Mice were observed daily for tumor size, weight loss and survival. Mice were sacrificed on Day 9, and TILs were enriched from tumors, stained with mCD3, mCD4, mCD8 (BioLegend), mCD45, CD90.1 (eBioscience), and Ghost Dye Violet 510 (Tonbo Biosciences, San Diego, CA), and were analyzed by flow cytometry. Pmel-1 CD8⁺ T cells were identified as mCD3⁺, mCD8⁺,mCD45⁺ and CD90.1⁺.

The experiments were executed in compliance with institutional guidelines and regulations and after approval from the MD Anderson Cancer Center Institutional Animal Care and Use Committee.

Activation and Proliferation Assays

CTL activation was analyzed by measuring the expression of Fas ligand (FasL) perforin and granzyme B, as previously described (41,42).Fucosylated and non-fucosylated CTLs were co-cultured with T2 cells that were pulsed with E75, CG1, or PR1 (40 ug/mL) at a ratio of 1:1 overnight at 37°C. At the end of the incubation period, the cells were stained with fluorescently-conjugated antibodies targeting CD3, CD8, FasL (BioLegend), CLA, and Ghost Dye Violet 510. After staining for cell surface markers, cells were permeabilized and stained with fluorescently-conjugated antibodies targeting granzyme B and perforin (BioLegend). Staining was analyzed using flow cytometry (BD LSR Fortessa).

To test proliferation, we performed standard CFSE assays. Briefly, fucosylated and non-fucosylated CTLs were labeled with CFDA SE Cell tracer reagent (Invitrogen) following manufacturer’s protocols. Cells were cultured in 96-well round bottom plates at a density of 0.3 × 10⁶ cells in RPMI + 10% FBS in the presence or absence of anti-CD3 and CD28 antibodies (BD Biosciences) to stimulate proliferation. Cells were harvested on days 2, 4 and 5 and stained with anti-CD8, anti-CD4 (Biolegend) and Ghost Dye Violet 510. Live, CD8⁺/CD4⁻ cells were gated on CFSE-positivity, and percent proliferation was calculated by gating on CD8⁺ cells and comparing the CFSE⁺ populations between samples.

Colony Forming Unit (CFU) Assay

Fucosylated CTLs were co-cultured with HLA-A2⁺ healthy donor bone marrow to determine the effects of fucosylation on CTL specificity. We used CG1- and PR1-CTLs as we have an established in vitro model using CFU assays with normal donor bone marrow hematopoietic stem cells (HSC) to test the specificity of these CTLs (28). Normal HSC were thawed and cultured for 24 hours in RPMI with 1% Penicillin/Streptomycin (P/S). Bone marrow cells and CTLs were co-cultured at a 1:5 ratio in RPMI (10% FBS + 1% P/S) for 4 hours in a 48-well plate. The incubated cultures were resuspended in Iscove’s Modified Dulbecco’s Media (IMDM) + 2% FBS and added into settled MammoCultTM H4034 Optimum Human Medium (STEMCELL Technologies, Cambridge, MA). These mixes were transferred to 6-well plates to incubate for 10-14 days. CFUs were imaged (Leica DMi8) and counted on Day 14. To confirm cell phenotype, wells were incubated at 4°C for 2-hours and washed with IMDM and PBS to prepare for antibody staining. Cells were stained with: fluorescently conjugated CD3, CD4, CD8, CD14, CD16, CD19, CD33, CD34 and live/dead Aqua (BD Biosciences).

Histology

To determine the effects of fucosylation on the specificity of antigen-specific CTLs, mice were sublethally irradiated and on the following day were treated with 2.5 × 10⁶ antigen-specific CTLs, and twice weekly. After 4 weeks, mouse tissues including spleen, liver, kidney, bone marrow, intestine, brain, heart, and lung were harvested and fixed in 10% formalin. The fixed tissue samples were embedded in paraffin, sectioned, and stained with hematoxylin and eosin prior to histological examination.

Immune synapse studies

T2 cells were pulsed with E75, PR1 and CG1 peptides (40 ug/mL), were washed and labeled with CellTracker Deep Red Dye (Life Technologies, Thermo Fisher Scientific, Waltham, MA), and were co-cultured with fucosylated and non-fucosylated CTLs at a 1:1 ratio at 37°C. Cells were then fixed in 2% paraformaldehyde, washed, and stained with anti-CD8-FITC (Biolegend) to identify CTLs. Following surface staining, cells were washed, permeabilized using 0.1% Triton-X (Sigma, St. Louis, MO), and stained with phalloidin AlexaFlour 594 (Life Technologies, Thermo Fisher) to visualize actin filaments at the synapse between CTLs and T2 cells. Data were collected on an imaging flow cytometer (ImageStreamXMarkII, Millipore Sigma, Burlington, MA) and analyzed using Ideas Software (Millipore Sigma).

Statistical analysis

Statistical analyses were performed using GraphPad Prism 7.0 software. P values less than 0.05 were considered significant.

Results

Fucosylated CTLs demonstrate enhanced endothelial transmigration in vitro

Fucosylation of antigen specific-CTLs with FT-VII was verified by staining cells with HECA-452 antibody, which targets CLA (Fig. 1A and 1B) (14). Non-fucosylated groups showed 0% CLA⁺ CG1- and PR1-CTLs and 1.1% E75-CTLs, whereas CTLs incubated with FT-VII showed 96.7%, 94.7%, and 99.5 CLA⁺ CG1-, PR1-, and E75-CTLs, respectively.

Figure 1. Fucosylation enhances CTL migration.

CD8 enriched antigen specific CTL were fucosylated with FT-VII. Results show representative HECA-452 staining by flow cytometry of A, leukemia specific CTL and B, breast cancer specific CTL. C, Fucosylated or non-fucosylated antigen specific CTL were passed through a HUVEC barrier to assess functional alterations to transmigration. LeukoTracker fluorescence was measured in the post-barrier well. Relative fluorescence units (RFU) are indicative of three independent experiments. Statistical testing was performed using unpaired t-test (*p < 0.05).

Using a trans-well assay, we showed that fucosylated CG1-, PR1-, and E75-CTLs had increased ability to migrate through the HUVEC barrier, as indicated by the increased cell count and relative fluorescence in the lower chamber (Fig. 1C). Fucosylated CTLs specific for CG1, PR1, and E75 had 1.7, 1.3, and 1.5 fold-increased migration, respectively, in comparison to non-fucosylated CTLs (all p < 0.05).

Fucosylation alters the expression of CTL surface markers

Differences in cell surface marker expression post-fucosylation were determined by cell surface staining and flow cytometry (Fig. 2). Fucosylated antigen-specific CTLs showed increased surface expression of the trafficking molecule CD162/PSGL-1 and CD183 (CXCR3), and the co-regulatory molecule CD137 (41BB) (Fig. 2).

Figure 2. Fucosylation alters expression of select CTL surface markers.

Fucosylated antigen specific CTL were assessed for known trafficking and regulatory cell surface molecules using flow cytometry. Fucosylation of antigen-specific CTL increased the cell surface expression of a number of trafficking molecules. Cell surface expression of CD162 (PSGL-1) (1.7-fold) and CD137 (41BB) (3-fold) was significantly increased in comparison to non-fucosylated CTL. Results are representative of 6 independent experiments. Statistical testing was performed using ANOVA (*p < 0.05; ****p < 0.0001).

Fucosylation enhances CTL cytotoxicity in vitro

To study the efficacy of fucosylated CTL-induced killing of targets expressing cognate peptide-MHC, we utilized calcein-AM assays at multiple E:T ratios. Fucosylated antigen-specific CTLs demonstrated higher cytotoxicity against corresponding target cells than non-fucosylated CTLs (Fig. 3). CG1-CTLs showed increased killing of CG1 pulsed T2 cells and primary patient AML samples (Fig. 3A; Supplementary Table S1). Similarly, fucosylated PR1-CTLs had increased killing of PR1 peptide-pulsed T2 cells and the U937-A2 AML cell line in comparison to non-fucosylated PR1-CTLs (Fig. 3B). Likewise, fucosylation of E75-CTLs was associated with significantly increased killing of E75 pulsed T2 cells and SKBR3-A2 breast cancer cells (Fig. 3C).

Figure 3. Fucosylation enhances targeted cytotoxicity in vitro.

Antigen-specific CTL were co-cultured with cognate peptide-MHC expressing target cells at varying E (effector):T (target) ratios in a calcein-AM release cytotoxicity assay. Results are depicted as % specific lysis of target cells. A, CG1-CTL and B, PR1-CTL were co-cultured with T2 cells pulsed with corresponding peptide, patient AML (UPN#2) or U937-A2 cells. Similarly, C, E75-CTL were co-cultured with E75 peptide-pulsed T2 cells or SKBR3-A2 cells at varying E:T ratios. Results are representative of 3 independent experiments. Statistics were obtained using unpaired t-test (*p < 0.05; **p < 0.005; ***p < 0.001).

In addition, we studied the CTL cytolytic machinery after fucosylation by analyzing the intracellular expression of granzyme B and perforin, and surface expression of FasL, after co-culturing CTLs with peptide-pulsed targets. Data demonstrate an increase in the percentage of CTLs expressing all three markers after fucosylation (Supplementary Fig. S1). We also investigated whether fucosylation enhances the binding of CTLs to target cells by analyzing synapse formation between CTLs and peptide-pulsed target cells (Supplementary Fig. S2). Our data show that fucosylation enhances the frequency of CTLs forming conjugates with target cells compared with non-fucosylated CTLs.

Fucosylation enhances the efficacy and homing of antigen-specific CTLs in vivo

The efficacy of antigen specific CTLs was tested in vivo using NSG xenograft mouse models for AML and breast cancer. Treatment of patient AML- or SKBR3-A2-engrafted NSG mice with the corresponding fucosylated antigen-specific CTLs resulted in a decreased disease burden in comparison with non-fucosylated antigen specific CTLs (Fig. 4). Specifically, mice that received fucosylated CG1-CTLs (n=12) showed significantly greater inhibition of leukemia than mice that received non-fucosylated CG1-CTLs (n=13) (p<0.05) (Fig. 4A; Supplementary Table S1). Similarly, treatment of the same patient primary AML with fucosylated PR1-CTLs (n=13) resulted in better disease inhibition when compared to treatment with non-fucosylated PR1-CTLs (n=13) (p<0.01) (Fig. 4B). Similar results were also obtained in the setting of breast cancer, where fucosylated E75-CTLs (n=6) demonstrated significantly improved disease inhibition in SKBR3-A2 engrafted mice compared with non-fucosylated E75-CTLs (n=6) (p<0.05) (Fig. 4C). Likewise, in the immunocompetent C57BL/6 mouse model, mice treated with fucosylated pmel-1 CD8⁺ T cells (n=9) (Supplementary Fig. S3) demonstrated a significant decrease in tumor volume growth (p<0.05), in comparison with mice treated with non-fucosylated pmel-1 CD8⁺ T cells (n=8) and untreated mice (n=9) (Fig. 4D).

Figure 4. Fucosylation enhances the efficacy of CG1-, PR1- and E75-CTL in vivo.

Irradiated NSG mice were engrafted with primary patient AML UPN#1 (1 × 10⁶ to 1 × 10⁷ cells) on day 0. Fucosylated or non-fucosylated A, CG1- or B, PR1- CTL were injected on day 2 (0.5 × 10⁶). Mice were sacrificed and bone marrow was harvested on week 2 and analyzed by flow cytometry. Results are expressed as percent hCD33⁺/hCD45⁺/mCD45⁻ cells of total live cells in bone marrow. C, For the breast cancer model, irradiated NSG mice were engrafted with SKBR3-A2 (1.5 × 10⁷ cells) in the mammary fat pad on day 0. Fucosylated or non-fucosylated E75-CTL (2 × 10⁶ cells) were injected IV on days 7 and 10. Mice were sacrificed at week 5 and tumor sizes were measured using bioluminescence imaging. Results are expressed as percent tumor size following treatment with CTL in comparison with pre-treatment tumor size. D, C57BL/6 mice were inoculated with B16-F10 melanoma cells (0.3 × 10⁶). Fucosylated or non-fucosylated expanded pmel-1 CD8⁺ T cells (5 × 10⁶) were injected IV on day 0. Mice tumor volumes were measured daily for 9 days. Results are expressed as mean tumor size on day 9 from two separate experiments. Statistical testing was performed using ANOVA (*p < 0.05; **p < 0.005).

These findings were corroborated by evidence of increased homing of fucosylated antigen specific CTLs into the leukemic bone marrow and tumor in comparison with non-fucosylated antigen-specific CTLs (Fig. 5). Frequencies of total live human CD3+ CG1-CTLs (n=11) in mouse bone marrow were significantly higher than for non-fucosylated CG1-CTLs (n=10) (p<0.001) (Fig. 5A). Fucosylated E75-CTLs (n=15) showed similar trends in homing to SKBR3-A2 primary tumor when compared with non-fucosylated E75-CTLs (n=13) (p=0.05) (Fig. 5B). These findings were also confirmed in the immunocompetent C57BL/6 mice. Mice treated with fucosylated pmel-1 CD8⁺ T cells demonstrated higher frequencies of pmel-CD8⁺ T cells (live, mCD3⁺ mCD8⁺, CD90.1⁺) compared with mice that were treated with non-fucosylated pmel-1 CD8+ T cells (p<0.05) (Fig. 5C).

Figure 5. Fucosylation enhances CTL homing in vivo.

Irradiated tumor-bearing NSG mice treated with fucosylated or non-fucosylated CTL were analyzed for CTL homing into tumor using flow cytometry. A, Results show percent of CG1-CTL (hCD33⁻/hCD45⁺/mCD45⁻/hCD13⁻/hCD3⁺/hCD8⁺) of total live cells in bone marrow harvested from leukemia-engrafted mice. B, Breast cancer engrafted mice were sacrificed at week 5, and TILs were enriched from the tumors. Results show percentage of E75-CTL (hCD45⁺/ mCD45⁻/hCD3⁺/ hCD8⁺) of total live cells from TIL extraction. C, Melanoma-engrafted C57BL/6 mice were sacrificed on Day 9, and TILs were extracted from the primary tumor. Results show percent of pmel-1 T cells (mCD45⁺/mCD3⁺/mCD8⁺/CD90.1⁺) of total live cells. Statistical testing was performed using unpaired t-test (*p < 0.05; ***p < 0.001).

CFSE assays showed that fucosylation does not affect CTL proliferation (Supplementary Fig. S4); however, we acknowledge that the in vitro proliferation assay does not conclusively exclude an effect of fucosylation on CTL proliferation in vivo.

Fucosylated CTL do not impair healthy HSC growth in vitro

CFU assays were used to verify that fucosylation does not interfere with CTL specificity. We have an established normal hematopoiesis model that tests the specificity of CG1- and PR1-CTL (26,28). This model takes advantage of the fact that CG1 and PR1 are both processed from azurophil granule proteases normally expressed by progenitor cells, although at lower levels than what is found in leukemia (26-28,40). We show that human HLA-A2⁺ normal bone marrow (BM) was not affected when co-cultured with fucosylated or non-fucosylated CG1- (Fig. 6A) or PR1-CTLs (Supplementary Fig. S5). Non-fucosylated and fucosylated CG1-CTLs co-cultured with BM resulted in an average of 388 CFUs and 448 CFUs, respectively (not significant). The same conditions with PR1-CTLs resulted in 140 and 137 CFUs, respectively (not significant). BM co-cultured with 1mM cytarabine was used as a positive control. BM co-cultured with leukemia specific CTLs also did not result in any significant differences in CFUs when compared to BM co-cultured with HIV-CTLs, an irrelevant control for antigen specificity.

Figure 6. Fucosylation does not alter CTL specificity or homing to normal tissue.

A, Healthy donor bone marrow was co-cultured with fucosylated or non-fucosylated CG1-CTL at a 1:5 ratio of BM:CTL. Cultures were grown in Mammocult medium in a 6-well plate for 14 days before CFU counts were taken. Results show CFU on day 14 for CG1-CTL. BM co-cultured with cytarabine (1 mM) served as a positive killing control. Data is representative of 3 independent experiments. B, Fucosylated (fuco) and non-fucosylated (non-fuco) antigen-specific CTL were administered to mice twice weekly IV via tail vein. After 4 weeks, mice were sacrificed and tissues were fixed, paraffin-embedded, sectioned, and stained with hematoxylin and eosin prior to histological examination. Data show equal lymphocytic infiltration in normal mouse tissue between fuco CTL and non-fuco CTL treated groups. Statistical testing was performed using ANOVA. BM, bone marrow; NS, not significant.

Fucosylation does not affect CTL homing into normal tissues

Fucosylation is believed to enhance the interaction of cell surface expressed selectin ligands with selectins, which are expressed on tumor and endothelial tissues and are unregulated during inflammatory states (6-10). To confirm that fucosylation does not increase homing of CTLs to normal (i.e. non-inflamed) mouse tissues, we administered CG1-, PR1-, and E75-CTLs to normal NSG and NSG-A2 transgenic mice. Histologic evaluation of various tissues confirmed that fucosylation does not increase the homing of antigen-specific CTLs to normal tissue (Fig. 6B).

Discussion

We have shown that fucosylation enhances the homing of antigen specific CTLs to malignant niches, resulting in increased anti-tumor efficacy. Importantly, fucosylation does not increase CTL homing to normal tissues. Furthermore, fucosylation enhances CTL efficacy, in part by altering CTL trafficking, cytolytic machinery, synapse formation, and expression of distinct activating surface molecules. Taken together, our data illustrate a novel, robust, and simple approach to enhance CTL homing to hematologic and solid tumor malignancies. Additionally, our data indicate ex vivo CTL fucosylation is an effective method to enhance T cell efficacy against tumor cells.

T cell homing to tumor sites following adoptive transfer is critical to achieve desired anti-tumor immune responses (43). Leukocytes migrate to inflamed tissues via interactions between T cell ligands and specific molecules expressed by the tumor and the endothelia. Most immunotherapeutic approaches have focused on the expression of chemokines within the tumor tissues and their cognate receptors by T cells. Numerous studies have employed methodologies to virally transduce chemokine receptors into effector cells to enhance homing. In one study by Garetto et al., the C-C chemokine receptor type 2 (CCR2) was transduced into CD8⁺ T cells that were specific for the SV40 large T antigen (SV40Tag) in a murine prostate cancer mouse model that expressed the chemokine (C-C motif) ligand 2 (CCL2) (44). Increased expression of CCR2 resulted in enhanced T cell homing and significantly improved anti-tumor activity. These data were corroborated by Craddock et al. in a neuroblastoma model using GD2-specific chimeric antigen receptor (CAR)-T cells retrovirally-transduced with CCR2b (45). A similar approach was employed using the chemokine (C-X-C motif) ligand 1 (CXCL1) and its receptor CXCR2 (46). In that study, CXCR2 virally-transduced gp100 TCR transgenic T cells were adoptively transferred into CXCL1-expressing melanoma-bearing mice. CXCR2-transduced T cells showed increased migration and anti-tumor activity following adoptive transfer.

Another approach to enhance T cell homing focused on the extracellular matrix (ECM) components of the tumor microenvironment, which can pose an impediment to the efficacy of the anti-tumor immune response (47). Caruana et al. virally transduced GD2-specific CAR T cells with the enzyme heparanase (HPSE), which degrades heparan sulfate proteoglycans, a main component of the subendothelial basement membrane (48). In this neuroblastoma model, the investigators demonstrated enhanced tumor infiltration and anti-tumor activity by the HPSE-transduced CAR T cells. Collectively, these studies have validated the potential to improve immunotherapy by modulating chemokine receptor- and ECM-degrading protease-expression by effector T cells. However, both approaches mandate prior knowledge of the tumor microenvironment, including the chemokine milieu of the tumor, the ECM components, and the expression of cognate receptors and proteases by the effector T cells. Furthermore, these approaches require a viral transduction step that could interfere with T cell functions and adds expense and technical challenges.

Other methodologies to enhance the homing of CTLs to tumor tissues have relied on the use of reagents that modulate the tumor vasculature (49,50). In a study by Calcinotto et al., a TNF-fusion protein, NGR-TNF, was utilized to increase T cell homing in mouse prostate and melanoma tumor models (49). Administration of NGR-TNF altered the endothelial layer of the tumor microvasculature, increased adhesion molecule expression by endothelial cells, and increased cytokine and chemokine release within the tumor microenvironment. Together, these modifications enhanced the extravasation of immune cells to tumors, resulting in reduced tumor burden. A similar result was shown using angiogenesis inhibitors and paclitaxel, where treatment of tumor-bearing mice with angiogenesis inhibitors increased the leukocyte-endothelium interactions and expression of endothelial cell adhesion molecules, and enhanced CD8⁺ CTL infiltration of the tumor (50). However, one significant drawback of treatment with angiogenesis inhibitors is that it requires an intact host adaptive cellular immune system, including endogenous antigen-specific CTLs that are capable of tumor elimination. It is possible that the co-administration of angiogenesis inhibitors with fucosylated antigen-specific CTLs could synergistically enhance T cell infiltration of tumor and further boost the anti-tumor immune response.

Our study validates the novel approach of fucosylation to enhance CTL homing and anti-tumor efficacy. Although the changes in in vivo antitumor effects were small in some of our experiments, this study provides a proof of principal for the potential of fucosylation to enhance the efficacy of ACT for cancer. This approach has broad implications for the field of cellular immunotherapy for cancer. Vaccines are effective in the setting of minimal tumor burden, however, for patients with metastatic disease, the efficacy of vaccines is limited by the number of antigen-specific CTLs that vaccination can generate, the TCR avidity of vaccine-induced CTLs for the target antigen, the activation status of vaccine-induced CTLs, and the ability of CTLs to home to tumor tissue (51). ACT approaches can address many of these shortcomings, including CTL number, TCR avidity, and CTL activation status. However, methodologies to engineer and expand CTLs are oftentimes hindered by the inadequate quantity of resulting effector cells and by the scant tumor infiltration of the adoptively transferred CTLs (45,48). Furthermore, methodologies that drive T cells to acquire potent cytotoxic functions in vitro can lead to the downregulation of lymphoid homing molecules (52). Our approach, therefore, is highly clinically relevant as it allows for the generation and infusion of fewer antigen-specific CTLs with improved homing capabilities. Additionally, our simple fucosylation strategy involves a short incubation of effector T cells with an enzyme/substrate cocktail, which can be applied rapidly to expanded/engineered cells. This approach was shown to be safe in the setting of SCT and therefore can be rapidly translated to the clinical setting for the treatment of hematologic as well as solid tumor malignancies, and possibly to CAR-T cell therapies.

Statement of Translational Relevance:

Adoptive cellular therapy (ACT), including tumor-specific cytotoxic T lymphocytes (CTLs) and chimeric antigen receptor (CAR)-T cells, has shown promise in the treatment of a number of malignancies. One limitation to ACT is the inadequacy of immune cell homing to the tumor site. To overcome this obstacle, we investigated the role of ex vivo fucosylation of antigen-specific CTLs as a mechanism to boost CTL efficacy against solid tumors and leukemia. Our results show that ex vivo fucosylation of CTLs enhances their homing and cytolytic machinery in the setting of breast cancer and leukemia. These findings are highly valuable from a clinical perspective because they identify a novel and simple approach to improve the efficacy of ACT, circumventing the standard labor-intensive and costly engineering methodologies to improve the potency and homing of therapeutic T cells. Clinically, fucosylation could prove to be a valuable tool to improve ACT for cancer.