Introduction
Understanding of anti-tumor immunity has increased significantly over the past decade, including fundamental insights into the role of innate and adaptive immune responses in targeting and eliminating tumors [
1,
2]. In both animal studies and, more recently, clinical studies, CD8+ cytotoxic T cells (CTLs) have been shown to be major effector cells involved in the eradication of tumor cells [
3‐
5]. A requisite step for the targeting of tumors by CTL is the binding of T-cell receptors (TCRs) to peptides derived from the proteome of tumors and presented on the cell surface in the context of MHC Class I molecules (pMHCs) [
6]. A major limitation for the effective targeting of established tumors by T cells results from central tolerance, whereby T cells with strong affinity for self-antigens are deleted in the thymus [
7]. In addition, the natural selection process inherent in the growth and establishment of tumors in immune-competent hosts leads to the selection of tumors with decreased antigen and MHC expression and alterations in antigen processing and presentation, reducing the surface density of pMHC complexes [
8]. Finally, peripheral tolerance mediated through host and tumor-driven immunosuppressive mechanisms contributes to the ultimate inability to mount an effective T-cell immune response to target tumors [
9‐
11].
One approach to overcome the lack of potent anti-tumor T-cell immunity is the ex vivo genetic modification of T cells to target tumors through the use of affinity-enhanced receptors generated from either T cell or antibody-derived receptors; recent clinical data have provided evidence that such approaches hold significant clinical promise [
12,
13]. A complementary approach that does not require ex vivo manipulation of T cells involves the use of fusion proteins that combine tumor recognition and T cell engaging domains to redirect T cells to target tumors [
14]. Among this class of reagents are several monoclonal antibody-anti CD3 scFv fusion reagents, some of which have undergone clinical evaluation with strong evidence of efficacy [
14‐
17]. The use of antibody-based T-cell targeting agents is restricted to targeting of antigens expressed on the cell surface. To overcome this limitation, we have recently developed a class of bi-specific fusion protein reagents termed ImmTACs (Immune-mobilizing monoclonal TCRs against Cancer), which target proteosomally processed epitopes derived from intracellular, surface bound, and secreted antigens [
18]. ImmTACs are comprised of soluble, high-affinity (pico-Molar) monoclonal TCRs (mTCRs) specific for antigen-derived pMHC complexes, combined with a humanized anti-CD3 scFv domain, which activates a potent anti-tumor T-cell response. The first clinical candidate ImmTAC product, IMCgp100, specific for gp100
280–288 is currently undergoing Phase I clinical testing in melanoma patients (trial number NCT01211262).
NY-ESO-1 (CT6, also known as LAGE-2) and LAGE-1 are among the prototypic cancer testis antigens [
19,
20]. Although little is known about their functions, normal tissue expression of both of these antigens is principally limited to immune privileged sites [
20‐
22]. Like other CT antigens, both NY-ESO-1 and LAGE-1 are expressed during development as well as by a wide range of tumors including myeloma [
23,
24] and a variety of solid tumors such as ovarian [
25], non-small-cell lung cancer (NSCLC) [
26,
27], and melanoma [
28,
29]. NY-ESO-1 has been targeted extensively in clinical trials using a variety of approaches including vaccines and gene-modified T cells [
13,
30]. Multiple studies have demonstrated that NY-ESO-1 is antigenic in human patients [
21,
31], although vaccination approaches alone rarely lead to a clinically significant response [
32,
33].
We have previously described both a high-affinity mTCR and ImmTAC reagent (ImmTAC-NYE) specific for the immunogenic HLA-A2 presented peptide NY-ESO-1
157–165 (SLLMWITQC) [
18,
34]. Notably, since the same epitope is processed from the LAGE-1 antigen [
20,
35], ImmTAC-NYE can be used to redirect T cells against HLA-A*0201-positive tumors that express either NY-ESO-1 or LAGE-1 antigens. In this report, we demonstrate that ImmTAC-NYE redirects potent antigen-restricted T-cell activity against NY-ESO-1- and/or LAGE-1-positive tumors, including established cell lines shown previously to present low levels of cell surface epitope [
34] and primary lung cancer cells. We show in vivo targeting of high-affinity NYE TCR to tumors presenting the NY-ESO epitope via time-domain optical imaging of fluorescently labelled mTCR. Importantly, we demonstrate ImmTAC-NYE-mediated inhibition of tumor growth in vivo, in a model co-xenografted with both human peripheral blood lymphocytes and cancer cells. The results provide rationale to support the further preclinical and clinical development of ImmTAC-NYE as a bi-specific immunotherapeutic agent to treat HLA-A*0201-positive and NY-ESO-1- and/or LAGE-1-expressing solid cancers.
Materials and methods
Human tissues, cells, and cell lines
Primary tumor tissue from lung and ovarian cancer patients was collected at the University of Pennsylvania under an IRB approved protocol. Pathologic review of biopsies confirmed diagnosis in each case. Tumor biopsies were processed as described [
36] to obtain single-cell suspensions which were viably frozen.
Cell lines J82 (urothelial), SK-Mel-37, Mel526, Mel624, A375 (melanoma), IM9 (EBV B), and U266 (multiple myeloma) were obtained from American Type Culture Collection or as described previously [
34] and cultured in R-10 medium. Cell lines were acquired between 2004 and 2010; HLA genotype and/or phenotype and expression of various cancer-associated antigens were confirmed by DNA sequencing, RT-PCR, and antibody staining, and lines displayed phenotypic characteristics of the tumor of origin. HEP2, HA2, and CM12 (normal human cells grown in culture) were purchased from ScienCell and NHEM10 cells from PromoCell.
Construction of J82-NY-ESO
157–165 and SK-Mel-37-NY-ESO
157–165 minigene-transfected cell lines has been described previously (as SLLMWITQC-minigene transfectants) [
34].
J82-NY-ESO
157–165 cells stably expressing enhanced green fluorescent protein (EGFP) (Clontech Laboratories, Inc., Mountain View, CA) were engineered using the pCGFP retroviral vector [
37]. Production of infectious retroviral vector particles in Phoenix A cells and infection of cells were carried out as described [
38]. This procedure was repeated twice, producing cells expressing high levels of GFP as determined by fluorescence microscopy (Leica Microsystems, GmbH, Wetzlar, Germany). The 5 % brightest cell population was isolated by a modified FACSAria II flow cytometer (BD BioSciences).
Peripheral blood lymphocytes (PBLs) were obtained from human healthy volunteers following Ficoll-Hypaque density gradient separation (Lymphoprep, Nycomed Pharma AS, Oslo, Norway).
Magnetic bead immunodepletion (Dynal) removed CD19+, CD4+, and CD14+ cells to obtain CD8+ ve T-cell populations.
Quantitative RT-PCR on normal and neoplastic human cells
RNA was isolated from viably frozen specimens or cell lines growing in logarithmic phase using RNAqueous-4 PCR kits (Ambion corp, AM1914). cDNA was synthesized using iScript cDNA synthesis kits (Biorad Corp, 170-8891). qRT-PCR analysis was performed using standard Taqman—MGB technology and amplification conditions using an ABI 7500 FAST instrument (ABI-Life technologies), and the following ABI inventoried primer-probe sets: NY-ESO-1: HS00265824_m1; LAGE-1: HS00535628_m1; Gus-B: HS99999908_m1; β-Actin: HS99999903_m1. Amplifications were performed in triplicate; individual Ct values were determined (minimum of 2/3 replicates with % CV < 15 %) and average values reported. RQ (relative quantification) values for NY-ESO-1 and LAGE-1 transcripts were determined according to the formula RQ = 2
−ΔΔCt, with ΔΔCt = ΔCt
sample − ΔCt
reference, with ΔCt
sample = Ct
sample − Ct
sample normalizer and ΔCt
reference = Ct
reference − Ct
reference normalizer [
39]. For all analyses, the melanoma cell line A375 (positive for NY-ESO-1 and LAGE-1) served as a reference; either Gus-B or β-Actin (for a subset of analyses) was used as the normalizer “housekeeping” gene.
Engineering high-affinity TCRs and bi-specific ImmTACs
TCR isolation and engineering to produce ImmTAC reagents has been described previously [
18,
40].
Cytotoxicity and cytokine release assays
Cytotoxicity (LDH release) and IFNγ ELISpot assays were performed as previously described [
18]. Multiple cytokine release (TNFα, IFNγ, and IL-2) was measured by Meso Scale Discovery (MSD) immunoassay following overnight incubation of CD8+ effector T cells and targets at a 1:1 ratio in the presence of 1 nM ImmTAC-NYE, in accordance with manufacturer’s instructions.
IncuCyte and real-time quantification of cell killing
Target tumor cells were incubated with effector CD8+ T cells (5:1 E:T) in the presence or absence of ImmTAC-NYE. Images were taken every 10 min and the number of apoptotic cells per mm2 was quantified using the CellPlayer 96-well Kinetic Caspase 3/7 reagent and the IncuCyte-FLR-Platform (Essen Biosciences). The reagent is cleaved by activated Caspase 3/7 upon target cell apoptosis resulting in the release of the dye and green fluorescent staining of nuclear DNA. The cocktail of apoptotic drugs used as a positive control contains 10 μM stausporine, 2 μg/ml anisomycin, and 10 μM etoposide.
TCR fluorescence labelling
TCR-NYE-wt, TCR-NYE-(29 nM), and TCR-NYE-(0.048 nM) were labelled with Alexa
® Fluor 680 carboxylic acid, succinimidyl ester as previously described [
41]. Labelling was restricted to 2 fluors per molecule. Protein concentration and degree of labelling were determined by spectrophotometric analysis using a NanoDrop spectrophotometer (Thermo Scientific, Rockford, USA).
Animal studies
Animal experiments were approved by The Norwegian Animal Research Authority and performed in accordance with the European Convention for the Protection of Vertebrates Used for Scientific Purposes. NOD-scid and NOD-scid IL2rγnull (NSG) mice (6–8 weeks old) were originally a gift of Prof. L. Shultz, Jackson Laboratories, Bar Harbour, USA, and were bred at the Gades Institute, University of Bergen. 5 × 106 tumor cells were injected subcutaneously (s.c.) in a solution of PBS/Matrigel (1:1) bilaterally in flanks of mice (n = 2 per mouse). Tumor volumes were measured by digital calliper measurements using the formula: Volume = π (length × width × height)/6. Blood samples (100 μL) were acquired by submandibular bleeding.
TCR imaging studies
When J82-NY-ESO157–165 tumors reached an average volume of 100–150 mm3, background fluorescence images of the mice (n = 4 per group) were acquired prior to intravenous (i.v.) injection with Alexa680 labelled TCRs (0.6 mg/kg) at 0 h and the mice imaged for TCR fluorescence at 4, 8, 24, 48, and 96 h.
Tumor protection study in mice harboring human PBL xenografts
NSG mice were inoculated i.v. with 20 × 10
6 human PBLs and randomized into study groups (equally distributed per percentage hu-CD45
+CD3
+). When mice demonstrated at least 10 % hu-CD3
+ cells by flow cytometry (approx. week 3), SK-Mel-37-NY-ESO
157–165 cells [
34] were injected s.c. and mice treated i.v. with either ImmTAC-NYE (0.04 mg/kg, Q.D.) or PBS (equivalent volume) on days 1–5.
ImmTAC efficacy studies in mice harboring cancer and human PBL xenografts
When J82-NY-ESO
157–165
GFP
tumors reached 50 mm3, 20 × 106 human PBL per mouse were injected i.v. After mice demonstrated engraftment of human lymphocytes (at least 10 % hu-CD3+ cells) and tumor volumes reached 100–150 mm3, they were imaged for GFP fluorescence and randomized into study groups. Mice were then treated with ImmTAC-NYE, control ImmTAC-GAG (0.04 mg/kg, Q.D., i.v.), or PBS (equivalent volume, i.v.) for days 1–5 and tumors monitored weekly by time-domain optical imaging. The study was terminated 25 days following the first ImmTAC treatment.
Optical imaging
Time-domain optical imaging was performed with Optix
® MX2 (ART Inc., Saint-Laurent, QC, Canada) as described previously [
42]. Fluorescence images were acquired and analyzed with Optix
® Optiview™ (version 2.00; ART Inc.) and Optix
® Optiview™ (version 2.02.00, ART Inc.).
Flow cytometric analysis of blood
Peripheral blood cells were analyzed by FACSCanto II (BD) using mouse anti-huCD3 (PE conjugate; clone SK7) and hu CD45 (FITC conjugate; clone 2D1) monoclonal antibodies (BD Immunocytometry). Human T-cell engraftment in peripheral blood was defined as the percentage of huCD45+CD3+ mononuclear cells (MNCs) in murine peripheral blood.
Discussion
The current study was performed to investigate the utility of a bi-specific ImmTAC reagent targeting an epitope common to both NY-ESO-1 and LAGE-1 for cancer immunotherapy. The expression patterns of both NY-ESO-1 and LAGE-1 in tumors identify these proteins as suitable targets for the treatment of a wide variety of malignancies. NY-ESO-1 and LAGE-1 are members of the cancer testis class of tumor-associated antigens, which has the “cleanest” expression profile in normal tissue. Since these antigens are intracellular, they are not accessible to conventional therapeutic antibodies and targeting of MHC-presented peptide epitopes provides an opportunity for tumor-directed immunotherapy [
35,
44].
RT-PCR studies have detected NY-ESO-1 or LAGE-1 transcripts in the testis, placenta, and ovary; other normal human tissues completely lacked expression, with the exception of one qRT-PCR study which also identified low levels of NY-ESO-1 mRNA in liver and pancreas [
20,
22,
45]. In broad agreement with previous data, qRT-PCR analyses performed in the current study did not detect NY-ESO-1 or LAGE-1 transcripts in several normal human tissues including hepatocytes, melanocytes, and a range of fibroblasts and endothelial cells derived from different organs. In contrast, NY-ESO-1 and/or LAGE-1 transcript expression was detected not only in cell lines derived from a variety of tumor types but also in primary tumors of ovarian cancer and NSCLC. In agreement with other studies [
25,
35,
46,
47], there was general overlap in NY-ESO-1 and LAGE-1 mRNA expression, although a subset of samples was identified which expressed only one of the antigens (Table
1). Particularly for the ovarian cancer specimens, LAGE-1 expression was detected at higher RNA levels and at a greater frequency than NY-ESO-1, suggesting that LAGE-1 may be a superior target antigen for cancer immunotherapy in at least some tumor types. Andrade et al. [
24] made a similar observation in multiple myeloma (49 % LAGE-1 vs. 33 % NY-ESO). A previous study detected presentation of the NY-ESO-1
157–165 epitope by the SK-Mel-37 and Mel 624 cell lines [
34], which express far higher levels of RNA for LAGE-1 than for NY-ESO-1 (Table
1), confirming that this peptide can be processed efficiently from LAGE-1.
We have engineered three soluble mTCRs specific for the NY-ESO-1
157–165 epitope, TCR-NYE-wt (the wild-type affinity TCR), and two affinity matured variants TCR-NYE-(29 nM) and TCR-NYE-(0.048 nM). Since previous studies have shown the importance of mTCR affinity for ImmTAC potency [
18], the highest affinity ImmTAC-NYE was evaluated in vitro and in vivo. ImmTAC-NYE was able to activate and redirect normal CD8+ T cells to a range of cell lines derived from melanoma (A375, Mel624, and SK-Mel-37), myeloma (U266), and EBV transformed B cells (IM9), in addition to freshly isolated lung cancer cells. The demonstration that the redirected cell lysis was limited to those cells expressing endogenous NY-ESO-1 and/or LAGE-1 proteins [
34] was essential to exclude the possibility of non-specific toxicity against normal tissues in a clinical setting. While significant lysis of the majority of the tumor cell lines occurred during the first 24 h, the IncuCyte real-time imaging system demonstrated that some cell lines required longer periods for caspase 3/7-dependent apoptosis to occur. One possible explanation for the observed variation in cell lysis is the difference in density of target epitopes on the surface of each cell line; previous studies using TCR-NYE-(0.048 nM) reported the presence of 25 and 45 epitopes/cell on the SK-Mel-37 and Mel624 cell lines respectively [
34], and U266 cells have 5–10 epitopes per cell (unpublished data). Other factors which may influence the rate of cell killing are firstly susceptibility of individual tumor cell lines to lysis and secondly the presence of cofactors such as CD80 and CD86 on the tumor cells which may augment killing.
An essential element for the efficacy of targeted TCR therapy is the effective delivery of engineered reagents to the tumor site. Such targeting of biologics has previously been demonstrated in vivo employing molecular imaging in both clinical and preclinical settings [
48]. In particular, fluorescence reflectance molecular imaging has typically been employed preclinically owing to its high sensitivity [
49], low cost, and potential for rapid in vivo screening. A major challenge for this strategy is to achieve sufficient target-to-background ratios to evaluate specific targeting over endogenous background auto-fluorescence and non-specific binding. However, the density of specific MHC/peptide complexes on the cell surface is usually low (between 10 and 150 epitopes per cell) ([
34] and unpublished data), thwarting the use of TCRs in conventional fluorescence imaging.
In contrast to conventional imaging, fluorescence lifetime or time-domain imaging can discriminate between fluorophores with different lifetime decay rates [
50]. Time-domain optical imaging has been employed with great success in distinguishing fluorescently labelled monoclonal antibodies bound to target antigens from both auto-fluorescence [
41] and background fluorescence from unbound antibodies [
51]. In the current study, the administration of three near-infrared-labelled TCR-NYEs of varying affinities, combined with the use of time-domain optical imaging techniques, confirmed targeting of the mTCRs to tumors presenting the NY-ESO-1
157–165 epitope.
To evaluate in vivo efficacy of ImmTAC-NYE, we utilized NSG mice engrafted with human PBL in both engraftment and established tumor models. We demonstrated that a short course (5-day administration) of ImmTAC-NYE substantially delayed engraftment of SK-Mel-37-NY-ESO157–165 tumor cells, with tumors just detectable in treated mice 50 days post-implantation, while tumors had reached an average volume of 368 mm3 in the control treated mice. In an established tumor model using J82 tumor cells engineered to express NY-ESO157–165 and GFP, we observed that five daily doses of ImmTAC-NYE resulted in a significant reduction in GFP fluorescence 7 days post-initiation of therapy compared to control treated mice, indicating a substantial reduction in tumor cell viability. An issue with this particular cell line in vivo was its inherent slow growth kinetics, which made differentiation of therapeutic effect in this pilot study impossible by calliper measurements. The significant reduction in fluorescence demonstrated with only five doses of ImmTAC-NYE by time-domain optical imaging endorses both application of this model and fluorescence lifetime gating in further efficacy strategies.
In xenografted animal models which have investigated ImmTAC efficacy (this study and [
18]), inhibition of tumor growth continues for long periods of time (over 5 weeks) after the final dose of reagent. ImmTAC-activated T cells release soluble factors including interferon-γ, IL-2 and TNFα (Fig.
1b), which can not only attract additional T cells and other effector immune cells to the tumor site but also promote components of the death receptor pathway in the tumor cells [
52], which can result in long-term anti-tumor activity. In cancer patients, it is probable that mechanisms such as epitope spreading (activation of endogenous T cells specific for other tumor epitopes) will occur, following cross-presentation of antigens from lysed tumor cells by dendritic cells. Indeed, we have demonstrated cross-presentation induced by ImmTAC redirected tumor cell killing in vitro (manuscript in preparation). These mechanisms have the potential to stimulate a prolonged anti-tumor immune response capable of breaking tumor tolerance. In addition, we have previously demonstrated serial killing of several tumor cells by a single ImmTAC-activated T cell, which may partly account for the high potency of these reagents [
18]. It is not known whether the PBMC-engrafted mice have functioning dendritic cells and therefore whether cross-presentation and epitope spreading can occur in this model system.
ImmTACs have been shown to detect low densities of cell surface epitope (as few as 10 per cell), in contrast to naturally occurring tumor-specific T cells which require higher levels of target epitope to become activated [
18]. Thus, ImmTACs have the potential to kill tumors which are not recognized by low-avidity T cells in patients, for example responses induced by cancer vaccines, since higher avidity CTL specific for self-peptides have been removed by negative thymic selection or inactivated by peripheral tolerance mechanisms. The ability of ImmTACs to detect a low density of pMHCs and to activate immune cells independently of co-receptors and other regulatory cells could circumvent immune tolerance in the tumor microenvironment and overcome inhibitory mechanisms such as MHC down-regulation or the presence of Tregs [
8,
9]. Decreased levels of MHC (relative to normal tissue cells) are frequently observed in tumors, whereas complete loss of both chromosomal loci encoding MHC and associated processing machinery is a relatively rare event (unpublished observations and [
53,
54]). Only patients with tumors positive for both NY-ESO-1/LAGE-1 and HLA-A*0201 are eligible for treatment with ImmTAC-NYE; we are currently isolating NY-ESO-specific TCRs with different HLA restrictions, in order to expand the target patient population. To conclude, we have demonstrated specific and potent ImmTAC-NYE redirected killing of NY-ESO-1 and LAGE-1 tumor cells both in vitro and in vivo, supporting the clinical development of this bi-specific reagent.
Acknowledgments
We thank M. Popa, L. Vikebø, M. Boge, and K. Jacobsen for expert assistance in all preclinical work and M. Enger for flow cytometry assistance. All imaging and flow cytometry was performed at the Molecular Imaging Center (MIC), the Department of Biomedicine, University of Bergen, Norway. This study has been supported by the Norwegian Cancer Society, the Western Health Board of Norway, and the Bergen Research Foundation.