A Preclinical protocol exploring mRNA-transfected allogeneic monocytes as a potential cancer vaccine
Li-Jun Mu1,2, Gunnar Kvalheim2, Stein Sæbøe-Larssen1, AnnaCarin Wallgren3, Bengt Andersson3, Alex Karlsson-Parra3,Gustav Gaudernack1
1Section for Immunotherapy, 2Lab of Cellular Therapy, The Norwegian Radium Hospital, University of Oslo, Norway; 3 Department of Clinical Immunology, Sahlgrenska University Hospital, Göteborg, Sweden
Abstract
Aiming to target antigen-presenting cells in vivo, we developed a clinical grade protocol using allogeneic monocytes as a combined antigen-vehicle and adjuvant. Enriched monocytes were obtained from leukapheresis product by ElutraTM. Fresh or frozen/thawed monocytes were transfected with EGFP or hTERT mRNA by square wave electroporation and transferred to Teflon bags containing X-vivo 20 medium supplemented with GM-CSF 1000 U/ml. After culturing for 24 hours, cells were treated with Vibrio Cholerae neuraminidase 0.025U/ml for 30 minutes, then concentrated and frozen. Fresh and frozen/thawed transfected monocytes gave a similar cell yield based on the initial number of loaded cells. Electroporation was highly efficient in transfecting mRNA into monocytes as tested by flow cytometry of EGFP-mRNA transfected cells, or by a TRAP assay after transfection with hTERT mRNA. To prime T-cell responses, PBMCs (containing T cells as well as antigen-presenting cells) were incubated with thawed allogeneic transfected monocytes for 7-10 days in vitro. Re-challenge with transfected autologous monocytes was found to generate a significant T-cell response specific for transfected mRNA (as measured by proliferation and IFN-gamma ELISPOT assays).Our results indicate that antigen-loaded allogeneic monocytes may act as antigen-vehicles and adjuvant for efficient cross-presentation of transfected antigens by antigen-presenting cells in the “recipient” PBMC-population and thus provide the basis for “on the shelf” cellular vaccines made from spared products of ordinary blood banking.
Introduction
Dendritic cells (DCs) in tumour vaccine therapy can be applied in two approaches. One is to treat patients with DCs that have been isolated and manipulated in vitro. The other strategy is to target DCs in vivo (Biragyn et al., 2000). Immature DCs with high capacity to capture antigens are located in the peripheral tissues but DC-precursors may also be actively recruited from the circulating pool of phagocytic monocytes (Radolph et al, 1999, Immunity,11,753). By receiving adequate maturation stimuli, recruited DCs become matured and can migrate to secondary lymph nodes to prime CTL responses. Amicroenvironmentenriched by proinflammatory components is thus a prerequisite not only for activation of DC’s and induction of apotent immune response but also for efficient recruitment of immature DCs and DC-precursors in situ.
T cells recognizing allogeneic MHC molecules by the direct pathway of allorecognition are present at very high frequencies in the T-cell repertoire, approximately 1–10% of an individual's T lymphocytes will respond to intact foreign MHC molecules expressed on antigen-presenting cells (APCs) from of another, allogeneic, individual(Wang et al., 1999).
By performing conventional allogeneic mixed leukocyte reactions (MLRs) in vitro we recently showed that primary, and particularly secondary MLR-supernatants, contain high levels of monocyte/immature DC-recruiting CC-chemokines and pro-inflammatory cytokines (Wallgren et al, 2005, Scand J Imm. 62, 234 ). Exposure of immature DCs to primary or secondary MLR-supernatants was found to upregulate CD40-expression and further enhanced LPS-induced interleukin-12 p70 production. Secondary MLR-supernatants additionally induced upregulation of CD86 and deviated allogeneic T cells-responses towards Th1 (enhanced IFN-gamma production without concomitant induction of detectable IL-4 or IL-10 production) (Wallgren et al, 2005 ). Taken together, these previous findings predict that the inflammatory process induced during direct allorecognition in vivo may have the potential to provide a milieu rich in DC-recruiting , DC-maturating and Th1-deviating cytokines and chemokines without the addition of other immunostimulants or prior clonal expansion of specific cells. Once recognised by the host T cells, the allogeneic cells are killed and taken up by resident autologous phagocytic cells, including DCs (Inaba 1998). This opens for the following novel strategy: To use allogeneic APCs loaded with tumour antigens as a combined antigen-vehicle and adjuvant to induce potent Th-1 deviated responses against cancer.
Human monocytes constitute about 10-20% of peripheral blood mononuclear cells and express both HLA class-I and -II antigens. They are capable of initiating allogeneic mixed leukocyte responses in culture, particularly when interacting with allogeneic memory T cells ( Thomas et al, J Immunol, 1993, 151, 6840) and can be loaded with antigens by a variety of mechanisms. The allogenicity of APCs has further been shown to become enhanced by removing negatively charged surface sialic acid by enzymatic treatment with bacterial neuraminidases (Taira and Nariuchi, J Immunol, 1988,141,440; Hirayama et al, J Exp Med, 1988, 168,1443; Fanales-Belasi, J Immunol, 1997,159,2203). Treatment of human monocytes with neuraminidase has further been shown to activate the extracellular signal-related kinases ERK 1/2 which result in enhanced production of monocyte/immatureDC-recruiting chemokines, including macrophage inflammatory protein (MIP)- 1 alpha and MIP-1 beta (Stamatos, et al. 2004). Since monocytes can be loaded with tumour antigens in the form of mRNA by electroporation, this enables the production of ready made, “on the shelf” allogeneic vaccines for any cancer where one or more antigens are available .
We here present data demonstrating that neuraminidase-treated allogeneic monocytes loaded with the “universal” tumour antigen telomerase reversed transcriptase (hTERT) (to a level similar to that found in cancer cell lines) are able to elicit T-cell responses against autologous monocytes loaded with the same antigen. Aiming to apply this strategy in clinical settings, we have developed a clinical grade cancer vaccine protocol.
Material and methods
Enrichment of monocytes
Peripheral blood mononuclear cells (PBMCs) were collected by leukapheresis from patients who were enrolled in our approved ongoing protocol for DC-based vaccination. After informed consent, excess PBMC were frozen at -80ºC for further use in pre-clinical protocol development.Monocytes were enriched directly from the leukapheresis products using a cell separator (Elutra™, Gambro BCT). Cells were washed once and resuspended in X-vivo 20 medium. Samples were taken for enumeration and viability test. The purity of monocytes was determined by flow cytometry analysis. These monocytes were either cultured directly or frozen for later use.
Transfection of EGFP-mRNA or hTERT-mRNA and culture of monocytes
Bulk preparations of mRNA for EGFP (Enhanced Green Fluorescence Protein) and hTERT (Human Telomerase Reverse Transcriptase) were performed as described earlier (Sæbøe-Larssen et al., 2002). The quality of all mRNA-preparations was controlled by electrophoresis on agarose gels stained by Gelstar (Cambrex Bio Science, Verviers, Belgium). The RNA samples were also evaluated on an Agilent Bioanalyser instrument (Agilent Technologies, Palo Alto, CA, USA), as previously described (Kyte et al., 2005). The mRNA content was estimated according to the manufacturers protocol.Fresh or frozen/thawed monocytes were washed and suspended in culture medium, 0.4 ml (107–108) cells were mixed with mRNA (50–100 μg/ml), transferred to a 4-mm-gap cuvette and pulsed with a BTX ECM 830 square-wave electroporator (Genetronics, San Diego, CA) using different parameter settings. Cells were then immediately transferred into Teflon bags containing X-vivo20 medium supplemented with GM-CSF (1000U/ml). Mock-transfected monocytes were cultured following the same electroporation procedure without mRNA. After 24 hours, cells were treated in medium with neuraminidase (Sigma) 0.025U/ml for 30 minutes. Afterwards, cells were washed and frozen in culture medium with 50% human albumin and 10% DMSO at -80ºC until use. For monocytes transfected with EGFP mRNA, we tested the transfection efficacy and cell viability with the FACSCalibur flow cytometer. For monocytes transfected with hTERT mRNA, expression of hTERT was analysed by TRAP assay (Telomeric Repeat Amplification Protocol), as described previously (Sæbøe-Larssen et al., 2002).
Monocyte characterization by flow cytometry
Monocytes were phenotyped before incubation, and before and after treatment with neuraminidase by the use of the following panel of FITC-, PE-, PerP- and APC- conjugated antibodies: CD1a, CD14, HLA-DR (Dako Cytomation, Glostrup, Denmark), CD40, CD83, CD86, and CCR7 (Immunotech, Marseilles, France). Irrelevant, matched antibodies were used as negative control. Cells were analyzed by flow cytometry using a FACSscan (Becton Dickinson).
Generation of T-cell responses in vitro
Frozen PBMCs and monocytes were thawed in a 37ºC water bath and washed once. After enumeration, PBMCs were plated in 24-well plate (3x106/well) and stimulated with irradiated (30Gy) transfected monocytes, either autologous or allogeneic (3x105/well). After 7-10 days culture in X-vivo 20 medium at 37ºC in a 5% CO2 incubator, cells were collected and washed once for further testing. Culture supernatant was harvested at 6 hours, 12 hours, 24 hours, 48 hours, 5 days and 7 days and frozen for later cytokine measurement.
T-cell proliferation assay
PBMCs stimulated by allogeneic monocytes and autologous monocytes were seeded in 96-well U-bottomed microtiter plates as responder cells (Tallo, Tauto). Irradiated (30Gy) autologous transfected and mock- transfected monocytes (tMo, mMo) were used as stimulators. Responder cells 1x105/well were cultured in triplicate with various numbers of stimulators for 2-3 days at 37˚C, 5% CO2 in X-vivo 20 medium. After labelling with 3.7 x 104 Bq [3H] thymidine (Laborel, Oslo, Norway) for 18 hours, cells were harvested. Radioactivity incorporated into proliferating cells was determined using a Packard TopCount microplate scintillation counter. Data are reported as counts per minute (CPM). Medium only, responder cells only and stimulator cells only were used as negative controls.
INF γ ELISPOT assay
The same responder and stimulator cells as described above were used. Monocytes as stimulators were plated with a cell concentration of 5x103/well, while the responder cells were added in a titration from 1x105 to 2.5x104 cells/well in duplicate. Details have been described previously (Mu, et al, 2003). Spots were enumerated using an automated ELISPOT counter (Carl Zeiss Vision, Oberkochen, Germany). Results were recorded as spots per 105 T cells.
Determination of cytokine levels By Bioplex assay
Supernatants were thawed and analysed by Bioplex cytokine assays (Bio-Rad Laboratories, Hercules, Ca, USA) according to the manufacturers protocol. We analysed the supernatants for the presence of IL-2, IL-4, IL-5, IL-10, IL-12, IL-13, TNF-α, INF-γ, GM-CSF, macrophage inflammatory protein 1β (MIP-1β), and monocyte chemoattractant protein 1 (MCP-1). Standard kits for all cytokines/chemokines were obtained from Bio-Rad Laboratories (Hercules, Ca, USA).
Statistical analysis
Student’s T test for independent samples comparing CPM in proliferation test and spots in ELISPOT assay for transfected and non-transfected monocytes was used. Results are displayed as mean ±standard deviation (SD).Statistical significance was determined at P< 0.05.
Results
Cell yield and Viability
The purity of monocytes following elutriation was 85.7 ± 4.2% (n=3). After transfection and 24 hours incubation, the average cell recovery from fresh monocytes was 34.4±5.1% with a mean viability of 95.7±4.9%, while from frozen/thawed monocytes the yield was 36.7±7.9% with a viability of 93.3±3.8% (Table1). The monocytes cultured freshly following elutriation tended to adhere together and to form large cell aggregates, which were difficult to separate. Since cell counting is based on single cells in suspension, this accounted for much of the recorded cell loss in this protocol. The actual cell number present in the cell suspensions may thus be considerably higher. In contrast, this phenomenon was much less pronounced after freezing and thawing. The finding that the total cell recovery from fresh and frozen monocytes was similar indicated that cell loss during freezing and thawing was roughly equal to cell “loss” due to aggregation.
Transfection efficacy and hTERT-mRNA evaluation
Efficiency of RNA transfection was evaluated by transfection with EGFP mRNA using 4-mm cuvettes. Several parameters with regard to voltages and time of exposure were tested and the optimal protocol was found to be a single pulse, 500V at 4ms. By using these parameters, a transfection efficacy of more than 95% CD14+ cells can be achieved and the mean fluorescence level using EGFP was increased on an average to 74-fold above background. The survival rate of transfected monocytes tested by flow cytometry using propidium iodide staining was similar to that of non-transfected cells. A representative EGFP-transfection experiment is shown in Figure1. In other experiments, a common tumor antigen, represented by synthetic mRNA for hTERT, was used for transfection. We analysed the telomerase enzymatic activity of the transfected monocytes after freezing and thawing, as a measure of hTERT protein expression, using buffer as a negative control and the tumor cell line K562 as a positive control. A clear enzymatic activity was detected (Figure2) in mRNA transfected monocytes compared with mock-transfected cells, showing that mRNA transfection resulted in telomerase reverse transcriptase activity levels comparable to that found in a tumor cell line expressing very high levels of the enzyme.
Flow cytometry evaluation of monocytes
Generally, the phenotypic analysis showed little changes in the expression of the HLA-DR, CD1a, CD40 and CD86 during culturing and treatment of the monocytes. After transfection and treatment with neuraminidase, cells showed some up regulation of CD83 and CCR7 when compared with monocytes before culturing (Figure3). This might indicate the presence of GM-CSF during the 24 hr culture period may to some extent have driven the differentiation of the monocytes towards dendritic cells. Freezing and thawing of the monocytes had no apparent effect on their phenotype (data not shown).
T-cell monitoring and functional study
To prime T-cell responses in vitro, PBMCs were incubated with thawed autologous or allogeneic transfected monocytes (mRNA either from EGFP or from hTERT) in three individual experiments for each antigen. In this experimental setting, autologous tMo will elicit T-cell responses specific for antigens encoded by the transfected mRNA, presented by autologous HLA molecules. Allogeneic tMo on the other hand, will engage at least two different sets of reactions: 1) viable tMo (stimulator cells) will elicit a conventional mixed lymphocyte reaction (MLR) by stimulating allo-specific T cells directly recognizing allogeneic MHC molecules on stimulator cells, 2)stimulator cells that subsequently become killed by alloreactive CTLs within the responder population will be phagocytosed by responder APC’s and their antigens, including epitopes from hTERT or EGFP, will be presented by autologous HLA molecules (cross presentation). Only the latter set of T-cell reactivity will be measured if the primed T cells are re-challenged with autologous tMo. After 7-10 days, specific responses of primed PBMCs were therefore tested by ELISPOT and proliferation assay using transfected autologous monocytes. Data from the proliferation assays are shown in Figure 4. A single priming of PBMCs by autologous or allogeneic transfected monocytes could generate a significant T-cell response specific for transfected mRNA. As demonstrated in the upper panel, clearly both allogeneic and autologous EGFP transfected monocytes give rise to specific T-cell responses (4A and B). There was no clear trend towards auto being better than allo or vice versa. Similar results were seen when hTERT mRNA was used (Fig. 4C and D). Cells prepared from fresh or frozen monocytes were similar in stimulating T-cell response (data not shown). The extent of autologous MLR observed was generally less than seen when autologous mRNA transfected DC were used to prime T cells (Mu et al., 2003). The background observed with the allo-stimulated PBMC may also be due to an auto MLR, or may partly represent continued proliferation by some of the MLR reactive T cells in the culture.
Very similar data were obtained in the ELISPOT assay. The data in Fig. 5 represent experiments with cells from five different donors. Clearly, in donor A similar responses to EGFP were obtained using either way of priming. When mRNA transfection was done with monocytes that had been frozen, a similar degree of priming was seen. Higher background was observed with donor B in the EGFP series and with all the donors in the hTERT series, indicating a considerable auto MLR reactivity, or alternatively reactivity against neuraminidase. The finding that the reactivity with non-treated/non-transfected monocytes (nMo) was much less than with mock transfected, neuraminidase treated monocytes, argue in favour of the latter explanation. Taken together, these experiments strongly indicate that antigens encoded by the transfected mRNA in allogeneic tMo are taken up and processed in autologous APC’s and presented to the T cells.
The presentation of tumor antigens in context of an allo MLR rests on two concepts: generation of an inflammatory response and cross presentation. We accordingly measured the cytokine concentrations in supernatant from PBMC during priming by autologous tMo and allogeneic tMo. As shown in Table 2, cytokine levels are much higher when PBMC are incubated with allogeneic tMo compared with those of autologous, indicating that a strong MLR reaction develops. No IL-4 could be detected and IL-10 was very low, while GM-CSF, INFγ, TNFα, MIP-1β and MCP were very high when tested on day 7. IL-12 was also low in the supernatant, but the peak of the secretion was observed during the first 12 hours to 48 hours of incubation (data not shown). This cytokine/chemokine profile is favourable to drive APC-maturation into a Th-1 polarizing direction. We also noticed that the concentration of IL-5 and IL-13 tested on day 7 was high. The secretion was very low during the first 24-48 hours but increased gradually after 2 days of the incubation.
Discussion
Aiming to target DCs in vivo, we have developed a clinical grade protocol for production of allogenic monocytes as vehicles to delivertumor antigens. The rationale is that such cells will stimulate a potent allo-response when injected into patient and subsequently be killed and taken up by resident or newly recruited DCs at the site of injection.
Monocytes obtained by Elutra™ are now being applied widely for routine generation of DCs. We, like others, demonstrate that elutriation of monocytes with Elutra™ provides highly purified “untouched” monocytes in large quantities within a closed system (Berger, et al., 2005). We show that these monocytes can be used as antigen carriers in vitro, resulting in priming of T cells specific for epitopes presented by autologous APCs. The efficiency of passive antigen loading of cells is greatly dependent on the cell’s capacity for antigen uptake and on the nature of the antigen used (protein, mRNA, DNA). Antigen loading can be done in many ways. In a “shotgun approach” using autologous tumor cells as a source of antigen to obtain a broad and “individualized” immune response, the antigen can be loaded in the form of tumor extracts/eluates, apoptotic cells or as tumor mRNA. In the context of GMP (Good Manufacturing Practice) production, the use of mRNA has several advantages. We have previously shown that square-wave electroporation is a highly efficient way to transfect mRNA into monocyte-derived DCs and CD34+-derived DCs (Sæbøe-Larssen et al., 2002; Mu, et al 2003; Mu et al., 2004), and that DCs loaded with tumor mRNA is equal to or better than DCs that have phagocytized apoptotic cells from the same tumor cell line, in priming T-cell responses in vitro (Jarnjak-Jankovic et al., 2005). In the model experiments presented here, we accordingly used this method for loading. We demonstrate here that this method could also be applied successfully to monocytes. By using EGFP mRNA as a reporter gene to study the optimal transfection conditions by flow cytometry, we could define the optimal parameters to balance between transfection efficacy and viability of cells. In addition, hTERT mRNA was used as a model mRNA for a tumor antigen to allow a direct measurement of telomerase activity in TRAP assays 24 h after transfection. Protein expression after mRNA transfection could be confirmed and was comparable to that found in a human cancer cell line. Based on these expression studies, the mRNA encoding the two antigens were used to study immune responses in vitro.