Generation of Large Numbers of Antigen-Expressing Human Dendritic Cells Using CD14-ML Technology

We previously reported a method to expand human monocytes through lentivirus-mediated introduction of cMYC and BMI1, and we named the monocyte-derived proliferating cells, CD14-ML. CD14-ML differentiated into functional DC (CD14-ML-DC) upon addition of IL-4, resulting in the generation of a large number of DC. One drawback of this method was the extensive donor-dependent variation in proliferation efficiency. In the current study, we found that introduction of BCL2 or LYL1 along with cMYC and BMI1 was beneficial. Using the improved method, we obtained CD14-ML from all samples, regardless of whether the donors were healthy individuals or cancer patients. In vitro stimulation of peripheral blood T cells with CD14-ML-DC that were loaded with cancer antigen-derived peptides led to the establishment of CD4+ and CD8+ T cell lines that recognized the peptides. Since CD14-ML was propagated for more than 1 month, we could readily conduct genetic modification experiments. To generate CD14-ML-DC that expressed antigenic proteins, we introduced lentiviral antigen-expression vectors and subjected the cells to 2 weeks of culture for drug-selection and expansion. The resulting antigen-expressing CD14-ML-DC successfully induced CD8+ T cell lines that were reactive to CMVpp65 or MART1/MelanA, suggesting an application in vaccination therapy. Thus, this improved method enables the generation of a sufficient number of DC for vaccination therapy from a small amount of peripheral blood from cancer patients. Information on T cell epitopes is not necessary in vaccination with cancer antigen-expressing CD14-ML-DC; therefore, all patients, irrespective of HLA type, will benefit from anti-cancer therapy based on this technology.


Introduction
Vaccination therapies that use antigenic peptides, for example, those emulsified in adjuvant or loaded onto dendritic cells (DC), have been broadly used to treat cancer. During the last two decades, considerable effort has been devoted to identifying cancer antigen-derived CTL epitopes that are restricted to the common alleles of HLA class I, such as HLA-A Ã 02:01 [1][2][3][4]. As a result, a vast amount of information has been accumulated on epitopes that are presented by major alleles of HLA class I [5][6][7][8]. On the other hand, relatively few epitopes have been identified for low-frequency HLA alleles. Thus, cancer patients who are negative for common types of HLA class I are excluded from most of the currently conducted vaccination therapies. Although HLA-A Ã 02:01 is the most common class I allele worldwide, gene frequency of HLA-A Ã 02:01 is at most 30% in most ethnic groups. Thus, a considerable number of patients cannot benefit from current vaccination therapies [1][2][3][4]. In addition, HLA-B-restricted epitopes have hardly been identified, probably due to the absence of particularly dominant alleles in the HLA-B locus. However, there should be many useful HLA-B-restricted epitopes, including already known cancer antigens. If HLA-B-restricted CTLs could also be stimulated, the efficacy of anti-cancer vaccination therapies would be improved substantially.
As a possible means to overcome the restrictions associated with synthetic peptide-based vaccination therapies, gene-based vaccinations, such as plasmid DNA vaccinations or vaccinations using recombinant viruses, may be considered [9,10]. However, plasmid-based DNA vaccines are not efficient enough to induce anti-cancer immunity [9]. As for therapies using recombinant viruses, the potential risk caused by the administration of infectious virus into patients may be problematic. Vaccination with genetically modified DC expressing cancer antigens may be more efficient and safer [11][12][13].
DC that are used for anti-cancer therapy are usually generated by in vitro differentiation of monocytes in peripheral blood samples, because DC exist in very few numbers in human blood [14,15]. Genetic modification of monocytes using viral vectors has been reported as a means to generate cancer-antigen-expressing DC [11][12][13]. Nevertheless, monocytes cannot be propagated, and the selection and expansion of transgenic cells is not feasible. Methods to propagate DC or the precursor monocytes are desirable for a more efficient generation of antigen-expressing DC.
We previously found that lentivirus-mediated transduction of cMYC along with BMI1 induced proliferation of CD14 + monocytes [16]. This observation led to the first established method for the amplification of human monocytes, and we named the monocyte-derived proliferating cells CD14-ML. The expanded CD14-ML differentiated into functional DC (CD14-ML-DC) upon the addition of IL-4 to the culture. One drawback to this method was the donor-dependent variation in proliferation induction, and the amplification of CD14 + monocytes was unsuccessful in 3 out of 12 blood donors in the previous study [16]. In the current study, we found a way to improve efficiency. In addition, we established a procedure to generate a large number of genetically modified DC expressing antigenic proteins. This method exploits the proliferating capability of CD14-ML. The capacity of CD14-ML-DC to induce vigorous T cell proliferation and cancer antigen-specific T cells demonstrated in this study indicates a potential value in vaccination therapy. Anti-cancer vaccination without information of T cell epitopes will become feasible, based on genetically modified CD14-ML-DC expressing antigenic proteins.

Cell samples and donors
This study was conducted with the approval of the institutional ethics review board of Kumamoto University, Graduate School of Medical Sciences (Approval numbers: 118 and 499).
PBMC samples were collected from healthy donors after obtaining written informed consent. CD14 + monocytes of healthy donors purchased from Takara Bio (Otsu, Shiga, Japan) were also used. Blood samples were also collected from 2 head-and-neck cancer (HNC) patients enrolled in a peptide vaccine trial, after obtaining written informed consent. The phase I/II clinical trial of cancer immunotherapy using three HLA-A24-binding short peptides, CDCA1 56-64 , IMP-3 508-516 and LY6K 177-186 , were approved by the Institutional Review Board of Kumamoto University, Kumamoto, Japan (Approval number: 841), and registered in University Hospital Medical Information Network Clinical Trials Registry (UMIN-CTR; UMIN000008379). UMIN-CTR is approved by ICMJE (International Committee of Medical Journal Editors). In total, monocytes derived from 10 healthy donors and 2 cancer patients were used ( Table 1).

Generation of recombinant lentiviruses
A cDNA fragment of human cMYC was obtained by PCR and cloned into the pENTR-TOPO vector (Invitrogen, Carlsbad, CA), and cDNAs for BMI1 and BCL2 were provided by the RIKEN BioResource Center (Tsukuba, Japan). cDNAs for CMVpp65 and MART1/MelanA were generated by gene synthesis (GenScript, Piscataway, NJ). These cDNAs were introduced into a lentivirus vector, pCSII-EF or pCSIIEF-IRES-PuroR, by using the LR Clonase system (Invitrogen). pCSII-EF and the plasmids for lentiviral vector packaging [17], pCMV-VSV-G-RSV-Rev and pCAG-HIVgp, were kindly provided by Dr. H. Miyoshi (Keio University). Plasmid constructs were introduced into 293T cells by lipofection (Lipofectamine 2000, Invitrogen) and 3 days later, the recombinant lentiviruses were recovered from the culture supernatant by centrifugation (50,000 ×g for 2 h), as described previously [18].

Quantitation of IL-12p70 production
To analyze the production of IL-12p70 by mo-DC and CD14-ML-DC, the cells were cultured in 96-well flat-bottomed culture plates (1 × 10 5 cells/200 μl medium/well) in the presence of OK432 (10 μg/ml). After 60 h of culture, the supernatant was collected and the concentration of IL-12p70 was measured using an ELISA kit (eBioscience).

Improved method for generation of CD14-ML
We previously reported induction of proliferating CD14 + monocytes by lentivirus-mediated introduction of cMYC and BMI1, resulting in the generation of CD14-ML [16]. Nonetheless, there was variation in proliferation efficiency among blood donors, and we could not induce the proliferation of monocytes derived from 3 out of the 12 blood donors in the previous study [16]. In the current study, we tried to improve the efficiency in generating CD14-ML. For this purpose, we searched for another factor to add to the 2 genetic factors already in place. We  Table). Four to five weeks after introducing cMYC, BMI1, plus BCL2, we observed proliferation of cells. Monocyte samples from all the donors exhibited significant proliferation, provided that they were freshly prepared but not freeze-thawed. The samples used in the current study included samples from that we could not establish CD14-ML with the introduction of cMYC and BMI1 only. Although the duration, rate, and the magnitude of proliferation still varied among the donors, there was at least a 100-fold increase in cell number attained with the current method (Table 1). Similar to the CD14-ML generated with the previous method, the proliferation of CD14-ML using the current method was dependent on the presence of both M-CSF and GM-CSF [16]. As shown in Fig 1A, CD14-ML were larger than monocytes and possessed round nuclei instead of monocyte-like, irregularly shaped nuclei.
DC differentiation of CD14-ML that were generated using the improved procedure In order to let CD14-ML differentiate into DC, we added IL-4. After 3 days, we added OK432 (penicillin-killed Streptococcus pyogenes), a potent maturation inducer for human monocytederived DC (mo-DC) [27][28][29]. Similar to the CD14-ML generated by the previous method, CD14-ML generated by the current procedure also differentiated into CD14-ML-DC with typical DC-like morphology and protrusions (Fig 1B). Expression of CD80, CD86, CD83, CD40, HLA class I, and HLA class II was detected in CD14-ML-DC and their expression was enhanced by stimulation with OK432 (Fig 2A). The levels of expression for these molecules were almost equal to that of mo-DC (Fig 2A). Under the stimulation with OK432, CD14-ML-DC produced higher amounts of IL-12p70 than mo-DC ( Fig 2B). Collectively, CD14-ML that was generated by the current method differentiated into CD14-ML-DC and possessed the phenotype of fully mature DC.

T cell stimulation and antigen presentation by CD14-ML-DC
We examined the ability of CD14-ML-DC to stimulate naive T cells with a primary allogeneic T cell stimulation assay. Compared to CD14-ML and mo-DC, CD14-ML-DC induced a higher magnitude of proliferation in the allogeneic T cells (Fig 3A). Treatment with OK432 further enhanced the T cell stimulation capacity of CD14-ML-DC.
Next, we examined whether CD14-ML-DC were able to present an antigenic peptide in the context of HLA class II. We used a previously established clone of GAD65 111-131 -specific DR53 (DRB4 Ã 01:03)-restricted human CD4 + T cells as the responder. CD14-ML-DC derived from a DR53-positive donor were pulsed with a GAD65 peptide and cultured with the T cell clone (Fig 3B). The T cells showed specific proliferative responses, indicating that CD14-ML-DC were able to present an antigenic peptide in the context of HLA class II to stimulate the T cells.
We examined the processing of protein antigens. CD14-ML-DC were cultured in the presence of recombinant GST-fused GAD65 protein or GST protein for 16 h before the addition of GAD65-specific T cells. The GAD65-specific T cells proliferated in a dose-dependent manner, indicating that CD14-ML-DC were able to process a protein antigen and present an antigenic peptide (Fig 3C).

Induction of antigen-specific CD8 + T cells from healthy donors by CD14-ML-DC
We tried to prime cancer antigen-specific CD8 + T cells by in vitro stimulation with antigenic peptide-loaded CD14-ML-DC, according to the culture protocol shown in Fig [26][27][28][29][30][31][32][33][34][35] , and IMP3 515-523 were induced from peripheral blood CD8 + T cells from the healthy donor 2 that was positive for HLA-A Ã 02:01 (Fig 4D and 4E). These results suggest that HLA class I-restricted peptide-loaded CD14-ML-DC were able to stimulate specific CD8 + T cells in autologous CD8 + T cell populations and induce their expansion in the same way as mo-DC.

T cell stimulation by CD14-ML-DC established from cancer patients
Monocytes obtained from cancer patients are often less viable than those from healthy donors. In such cases, the generation of DC from monocytes has been reported to be very difficult [30]. Thus, it would be very useful if CD14-ML could be established not only from healthy donors but also from cancer patients. We tried to establish CD14-ML from monocytes obtained from HNC patients. As a result, we successfully established CD14-ML from monocytes obtained from the 2 cancer patients (cancer patient 1 and 2), and the CD14-ML were able to differentiate into CD14-ML-DC after stimulation with IL-4.
We examined the capacity of CD14-ML-DC to induce antigen-specific CD8 + T cell lines. Stimulation of CD8 + T cells by peptide antigen-loaded autologous CD14-ML-DC in induction cultures was carried out using the same protocols as those that were used for cell samples from healthy donors. To analyze the induction of antigen-specific T cells, resulting CD8 + T lines were subjected to IFN-γ ELISPOT analysis and stained with an HLA/peptide tetramer. The response of CD8 + T cells to CDCA1 56-64 , LY6K 177-186 , and KIF20A 66-75 was observed for cancer patients (Fig 6A-6D).
Similarly, we tried to induce of cancer antigen-specific CD4 + T cell lines by using CD14-ML-DC derived from cancer patients. The cancer patient 1-derived CD4 + T cell lines  (Fig 7A). The cancer patient 2-derived CD4 + T cell lines that were harvested from the induction culture responded to the CDCA1 39-764 , CDCA1 55-78 , and LY6K 172-191 peptides in an HLA-DQ or HLA-DR-restricted manner (Fig 7B). , and subsequently mixed with autologous CD4 + T cells in AIM-V with 5% human decomplemented plasma. On day 7, the T cells were restimulated with the autologous peptide-pulsed CD14-ML-DC and supplemented with rIL-7 (5 ng/ml). After two days, these cultures were supplemented with rIL-2 (10 IU/ml). CD14-ML-DC were added with only IL-4 (did not add OK432). On day 14, the stimulated CD4 + T cells in each well were analyzed for specificity in IFN-γ ELISPOT assays. The T cells showing a specific response to the cognate peptide were transferred to 24-well plates and restimulated with the autologous peptide-pulsed CD14-ML-DC, and subsequently supplemented with rhIL-7 (5 ng/ml) and rhIL-2 (20 IU/ml). On day 21, the T cells were restimulated with the autologous peptide-pulsed CD14-ML-DC and supplemented with rhIL-7 and rhIL-2. IFN-γ ELISPOT assays were performed after 6 or 7 days from the fourth round of peptide stimulation. (B, C) After the stimulation (more than three times), the number of CD4 + T cells reacting to each peptide was analyzed with an IFN-γ ELISPOT assay (Day 28). The results of the T cells before stimulation culture are shown (Day 0). Dimethyl sulfoxide was used as a control. The results for the healthy donor 3 (B) and donor 2 (C) are shown.  Collectively, the generation of CD14-ML-DC derived from monocytes of cancer patients was feasible, and the cancer patient-derived CD14-ML-DC were sufficiently potent to induce antigen-specific T cell responses.

Genetic modification of CD14-ML to generate CD14-ML-DC expressing CMVpp65
We demonstrated the establishment of antigen-specific CD4 + or CD8 + T cell lines by in vitro stimulation with CD14-ML-DC that were pre-loaded with synthetic peptides (Figs 4 and 5). We next generated genetically modified CD14-ML-DC expressing antigenic proteins by taking advantage of the proliferative capabilities of CD14-ML.
We transduced CD14-ML derived from an HLA-A Ã 24:02-positive donor with a lentivirus vector for CMVpp65, including an IRES (internal ribosomal entry site)-puromycinacetyltransferase cassette (Fig 8A). The cells were cultured for 2 weeks in the presence of puromycin to select and expand cells expressing the protein. After that, the expression of CMVpp65 in the resultant CD14-ML (CD14-ML/CMV) was confirmed by flow cytometric analysis (Fig 8B). We added IL-4 to CD14-ML/CMV-derived from an HLA-A Ã 24:02-positive donor to generate CD14-ML-DC/CMV, and used them as a stimulator for the induction of CMV-specific CD8 + T cell lines from the autologous peripheral blood CD8 + T cells. To examine CMV-specific responses, we examined the IFN-γ-producing cells after overnight co-culture with C1R cells that were pulsed with HLA-A Ã 24:02-restricted a CMVpp65-derived epitope (CMVpp65 341-349 ). The results shown in Fig 8C indicate the successful induction of CD8 + T cell lines reactive to CMVpp65 341-349 . CMVpp65 341-349 -reactive CD8 + T cells were also detected by staining with the tetramer (Fig 8D).
Next, we tried to generate a CMV-pp65-reactive CD8 + T cell line from another blood donor who is homozygous for HLA-A Ã 26:01 (healthy donor 4). No information on CMV-derived epitopes is available for this HLA-A allele. Generation of CMVpp65-expressing CD14-ML-DC and co-culture with autologous CD8 + T cells were done by the same procedure described above. To detect reactivity to CMVpp65, the resulting T cells were co-cultured with autologous CD14-ML with or without the transgene for CMVpp65. As shown in Fig 8E, the response CD14-ML-DC Expressing Antigenic Protein to pp65 protein was observed in this analysis. The T cells stimulated with CD14-ML-DC/ CMVpp65 were further activated, and compared to T cells before stimulation culture. The response was blocked by the addition of an anti-HLA class I blocking antibody, indicating that the response was restricted by HLA class I. Although we have not identified the epitope for this donor, it was experimentally demonstrated that stimulation of antigen-specific CD8 + T cells is feasible regardless of whether information on the epitope is available or not.

CD14-ML-DC expressing MART1/Melan-A
We next generated CD14-ML-DC expressing MART1/Melan-A by a similar method as above. Expression of the MART1 protein in the resulting CD14-ML-DC (CD14-ML-DC/MART1) was confirmed by staining with a specific antibody and subsequent flow cytometric analysis (Fig 8F).
CD14-ML-DC/MART1 were generated from 2 blood donors, one (healthy donor 2) who was positive for HLA-A Ã 02:01 and another (healthy donor 4) who was negative for HLA-A Ã 02:01. Both CD14-ML-DC/MART1 were co-cultured with autologous peripheral blood CD8 + T cells. The CD8 + T cells were stimulated with the CD14-ML-DC/MART1 once a week, and the culture was continued for 3 weeks. On day 21, the cultured T cells were analyzed for their reactivity to MART1.
Collectively, we can readily generate functional DC expressing antigenic proteins based on the CD14-ML technology. More importantly, by using the antigen-expressing DC, we can induce antigen-specific T cells even if information on the HLA-restricted epitopes is not available.

Discussion
The present study is based on our previous observation that lentiviral vector-mediated introduction of cMYC and BMI1 induced proliferation of human peripheral blood monocytes [16]. In that study, we did not succeed in generating CD14-ML from monocytes derived from some of the donors. One of the purposes of the current study was to find a way to improve the efficiency of the CD14-ML induction. We assumed that differences in the expression of some factor related to cell proliferation or anti-senescence would explain the variation among the monocytes from different donors, and that we could improve the efficiency of the method by adding additional factors.
In order to find factors to add, we tested several genes related to cell-proliferation, anti-apoptosis, or anti-senescence, including BCL2, LYL1, MEIS1 and other various genes. As a result, we found that the introduction of BCL2 or LYL1 along with cMYC and BMI1 significantly improve efficiency. By the currently improved procedure, we could induce the proliferation of monocytes derived from all healthy individuals and cancer patients, provided that the monocytes were freshly prepared.
We comparatively analyzed the expansion of monocytes subject to old and new methods for CD14-ML generation. In the experiments, we used monocyte samples obtained from two donors ( Figure B in S2 Fig). For one sample, vigorous proliferation was induced by the both method and the magnitude of amplification induced by the new method was somehow greater than that by the old one. For another cell sample, significant cell amplification was achieved only by the new method. Features of CD14-ML generated by the current method is indistinguishable from those generated by the original method in terms of morphology, cell surface molecules, dependency on GM-CSF and M-CSF, and capacity to differentiate into DC (Figs 1  and 2).
Critical issue may be the risk of malignancy accompanying clinical use of cells transduced with cMYC, BMI1 plus BCL2 by lentivirus. Although CD14-ML cease to proliferate in several days after addition of IL-4, we cannot totally rule out the risk of development of malignancy. As shown in Fig 3A, proliferative capacity of CD14-ML-DC was completely inactivated by irradiation at 45 Gy. The data suggest that we can avoid the risk of malignancy development by irradiation of CD14-ML-DC before the administration to patients.
Another purpose of the current study is to evaluate the possibility of generating genetically modified CD14-ML-DC expressing antigenic proteins and to examine their ability to induce antigen-specific T cells in culture. In order to generate DC expressing whole antigenic proteins, we introduced lentiviral vectors for the antigenic proteins, CMVpp65 or MART1/MelanA, into CD14-ML. The vectors were designed to drive the expression of the antigenic proteins by the EF1 promoter and contained an IRES-puromycinacetyltransferase in the downstream region of the cDNA for the antigenic protein. We can efficiently select CD14-ML expressing the antigen by culture in the presence of puromycin. Indeed, the selected cells were mostly positive for the expression of the transgene-derived proteins (Fig 8B and 8F). Subsequently, the CD14-ML were subjected to differentiation culture in order to generate CD14-ML-DC. As shown in Fig  8, in vitro stimulation of peripheral blood T cells with antigen-expressing CD14-ML-DC generated by this procedure successfully induced CD8 + T cell lines that were reactive to CMVpp65 or MART1/Melan-A.
Anti-cancer vaccination therapies with antigenic peptides are currently conducted at many institutions. The synthetic peptide-based vaccination therapy requires information on the sequence of the epitopes for the cancer antigens. One of the drawbacks of the peptide vaccine is that so far, the identified epitopes are limited to those restricted by common HLA class I alleles, such as HLA-A Ã 02:01 [1][2][3][4]. Thus, a considerable number of patients that are negative for major alleles are currently excluded from vaccination therapies. Another problem may be that HLA-B-restricted epitopes have been hard to identify.
The generation of antigen-expressing CD14-ML-DC should be feasible from any donor. Efficient T cell stimulation and the induction of antigen-reactive T cell lines demonstrated in the current study, imply that CD14-ML-DC can efficiently work in vaccination therapy. We consider that it is possible that the technology based on the CD14-ML will be able to change the current situation of anti-cancer vaccination therapy. Determination of cancer antigen epitopes may become unnecessary, and all cancer patients, irrespective of HLA type, will benefit from vaccination therapy.  Table. Fold increase of cell number at 6 weeks after introduction of various factors along with cMYC plus BMI1 (DOCX)