Generation of a Dendritic Cell-based Vaccine in Chronic Lymphocytic Leukaemia Using CliniMACS Platform for Large-scale Production

Authors


H. Mellstedt, Department of Oncology, CCK, Karolinska University Hospital Solna, 171 76 Stockholm, Sweden.
E-mail: hakan.mellstedt@karolinska.se

Abstract

We previously demonstrated that dendritic cells (DC) that have endocytosed apoptotic bodies of autologous leukemic cells (Apo-DC) can boost antileukemic T-cell responses. In this study, we report a description of the production procedure and product specification of the Apo-DC vaccine preparations for clinical use. Enriched populations of CD14+ monocytic precursors and CD19+ leukaemic cells were obtained using CliniMACS technology from a single leukapheresis product. Apoptotic bodies were obtained by irradiating (5 Gy) CD19+ selected B cells. DC were generated ex vivo by culturing monocytes with granulocyte macrophage colony-stimulating factor and interleukin-4. Following coculture with apoptotic bodies, DCs were matured with tumour necrosis factor-α. The mean percentage of CD14+ cells in the peripheral blood as well as in the leukapheresis product of the patients (n = 10) was approximately 2% (range, 0.8–3.3). Immunomagnetic selection using the CD14 reagent yielded a CD14+ population that was 91 ± 2.2% (mean ± SEM) pure. Immunomagnetic selection of CD19 expressing cells yielded a population that was 100 ± 0.03% pure. Cell viability immediately after selection was 97% and 98% after 7 days of culture. The Apo-DC cellular vaccine product showed a mature phenotype, with a high rate of endocytosis (84%) of apoptotic leukemic B-cells. In conclusion, despite significant variability in the circulating monocyte frequency of the chronic lymphocytic leukaemia patients, our method permitted the production of a DC vaccine with high reproducibility and conforming with recommended quality standards.

Introduction

Chronic lymphocytic leukaemia (CLL) of the B cell type is the most common adult leukaemia and is generally considered to be incurable. The clinical course is heterogeneous, with some patients remaining in a stable phase for a long time without the need for therapy while others experience a rapid disease progression and death despite therapy.

Treatment with chemotherapeutic agents, such as purine analogs and alkylanting agents, alone or in combination with monoclonal antibodies, induce a high response rate and prolongation of progression-free survival, but not a substantial increase in overall survival [1, 2]. Naturally occurring T cells recognizing leukemic cells have been previously described in CLL and play a role in controlling CLL [3, 4]. Allogeneic T-cells in haematopoietic stem cells transplants or administered as donor lymphocyte infusions are known to mediate ‘graft-versus-leukaemia’ effect which can effectively eradicate leukaemic cells [5–8].

Our previous studies [9–11] as well as those of other investigators [12–14] indicate that it is feasible to expand antileukaemic, cytolytic T cells ex vivo. A few target antigens that can be recognized by T cells have also been described in CLL [15].

Substantial efforts have been directed during the last decades to immunize patients with cancer antigens to induce/expand in vivo T cells capable of recognizing and lysing tumour cells. Use of autologous tumour cells, rather than a single, defined protein or peptide as the immunogen provides the benefit of potentially presenting the entire repertoire of tumour-associated antigens to the patient’s immune system. Ex vivo matured dendritic cells (DC) loaded with antigens express the whole array of co-stimulatory and adhesion molecules required for the activation of the innate and adaptive immune system to induce tumour-specific CD4 and CD8 T cells [16]. In previous studies [17, 18], a direct comparison of tumour lysate, apoptotic bodies, tumour RNA and fusion hybrids, revealed that apoptotic bodies were the best approach to loading whole tumour antigens into DC. Similar conclusions were also reached in a study by Jarnjak-Jankovic et al. [19]. The slow, smouldering nature of CLL, observed in the majority of the patients, makes it an ideal candidate for immunotherapy as the effect of this approach is typically seen after a delay following initiation of therapy. Development of a DC-based therapeutic approach in CLL has, however, practical constraints as the frequency of monocytes in the peripheral blood is very low amongst a huge number of leukemic cells.

We recently initiated a phase I/II clinical trial in which CLL patients with steadily increasing leukaemic cell count (25–99% increase) during the last 6 months but no current or expected (within six months) need of cytostatic treatment, received 1.5 × 107 Apo-DC for a minimum of five immunizations. One of the primary objectives of this trial was to determine the feasibility of generating DC loaded with apoptotic bodies (Apo-DC) from CLL patients [20]. In the present report we describe our experience with preparing Apo-DC vaccine from the first 10 patients accrued in the trial. The detailed procedure and product specification can be utilized as a prototypical method of enriching scarce monocyte precursors from leukaemia patients and generating DC of adequate quality criteria, purity and numbers for clinical use.

Materials and methods

Patients.  The study was performed in keeping with the Helsinki declaration on research with human subjects and the protocol was approved by the institutional ethics committee and the Medical Product Agency (Svenska Läkemedelsverket). Table 1 shows patients’ characteristics at leukapheresis. The study included 10 patients, 6 males and 4 females. Six patients were Rai stage I and four stage II. Mean age of the patients was 68 years. Mean white blood cell (WBC) count of the CLL patients (n = 10) at leukapheresis was 40 × 109/l (range, 8–77 × 109/l) and mean percentage of lymphocytes was 84% (range, 68–94%). Four patients had previously received chemotherapy (chlorambucil), while the remaining patients had not received prior therapy. None of the previously treated patients had received any treatment for their leukaemia during the last 6 months.

Table 1.   Patient characteristic at leukapheresis, DC evaluation.
Pts characteristicsDC evaluation
Pt # Age/genderaRai stageWBC count (×109/l) Lymphocytes (%) Monocytes (%) DC yield (106 DC)b Apo-DC/vial (106 Apo-DC)cDC viability before banking (mean %)dDC viability after thawing (mean %)eDC recovery at thawing (mean %)f CD83+/CD86+/HLA-DR+
  1. CLL, Chronic lymphocytic leukaemia; DC, dendritic cell.

  2. aM: male; F: female.

  3. bFinal number of Apo-DC produced.

  4. cNumber of DC per vaccine vial.

  5. dMean viability of Apo-DC before freezing.

  6. eMean viability of Apo-DC (5 vials) after thawing.

  7. fMean recovery of Apo-DC (5 vials) after thawing.

CLL I-164/M262.0941.38817.5979610167/89/98
CLL I-279/F177.0940.8501098966959/84/100
CLL I-373/M224.2741.5601299959847/78/79
CLL I-463/M139.8852.115817.594927723/76/92
CLL I-565/F234.8881.315817.599988576/93/100
CLL II-172/F141.0910.87515979511791/98/95
CLL II-251/F172.0920.928017.5989811569/84/100
CLL II-376/M113.9752.78116.2989610063/97/83
CLL II-467/M27.8682.721017.597969179/95/99
CLL II-566/M127.8813.340317.5989610993/99/98
Mean68 40.084.21.7156.315.898969667/89/94

Leukapheresis.  A schematic presentation of the procedure for the production of the cell-based vaccine is shown in Fig. 1. The leukapheresis procedure was performed with a Cobe Spectra blood cell separator (Gambro BCT Inc., Lakewood, CO, USA) using the program for collection of mononuclear cells (version 7.1) with manual online adjustments of the plasma pump [21].

Figure 1.

 Schematic representation of the Apo-dendritic cell (DC) vaccine production platform. Isolation of CD14+ and CD19+ cells from the leukapheresis product was performed by CliniMACS immunomagnetic separation. CD14+ cells were cultured in medium supplemented with granulocyte macrophage colony-stimulating factor and interleukin-4 to obtain immature DC (imDC). CD19+ cells were cultured without cytokines and irradiated (5 Gy) on day 3 to induce apoptosis. On day 4 imDC and apoptotic B cells were mixed and further cultured for 3 days. On day 5 tumour necrosis factor-α was added. On day 7 Apo-DC was aliquoted in vials and cryopreserved at −150 °C until they were used for vaccination.

Flow cytometry analysis.  Flow cytometry analysis (FCM) analysis was performed using EPICS XL (Beckman-Coulter, Miami, FL, USA) with Coulter System II software. The distribution of monocytes, lymphocytes and granulocytes was analyzed using CD45-fluorescein isothiocyanate (FITC)/CD14-PE and CD33-FITC/CD15-PE. Monocytes and mature Apo-DC were assayed for phenotype using a panel of 10 specific cell surface marker consisting of CD14-PE, CD20-FITC, CD45-FITC, CD80-FITC, CD83-PE, CD86-PE, CD-1a-PC5, ILT3-PC5, DC-Sign-FITC and human leucocyte antigen (HLA)-DR-FITC. All the fluorochrome-conjugated primary antibodies and the corresponding isotype Ig controls were purchased from Immunotech Coulter (Marseille, France), except for DC-Sign-FITC, which was purchased from R&D Systems (Minneapolis, MN, USA). 7AAD-positive (non-viable) cells were excluded from FCM analysis.

CliniMACS CD14+ and CD19+ isolation.  The isolation of the CD14+ and CD19+ cells from the leukapheresis product was performed by immunomagnetic separation using the CliniMACS affinity-based technology (Miltenyi Biotec GmbH, Bergisch Gladbach, Germany) according to the manufacturer’s recommendations. Reagents, tubing sets and buffers were purchased from Miltenyi Biotec.

Cell culture. Preparation of apoptotic B-cells. B-cells were cultured in VueLife teflon bags (CellGenix, Freiburg, Germany) at a cell density of 5 × 106 cells/ml without cytokines and irradiated (5 Gy) on day 3 using a GAMMA CELL 2000 (Molsgaard Medical, Horsholm, Denmark) and cultured in the same medium for 4 days. The frequency of apoptotic B-cells was assayed by FCM using the Annexin V FITC apoptosis kit (BD Pharmingen™, Franklin Lakes, NJ, USA) and propidium iodide (BD Pharmingen™).

Generation of Apo-DC. CD14+ cells were cultured in bags at a cell density of 1–2 × 106 cells/ml in CellGro serum-free DC-medium (CellGenix) supplemented with GM-CSF (100 ng/ml, Leukine, BERLEX®, Seattle, WA, USA) and interleukin-4 (IL-4, 20 ng/ml) (CellGenix). One volume of fresh medium with double the concentration of GM-CSF (200 ng/ml) and IL-4 (40 ng/ml) was added to the bags after 2 days of cell culture. On day 4 immature DC (imDC) and apoptotic B cells were mixed at a ratio of 2:1 and cultured at a density of 3 × 106 cells/ml in fresh medium supplemented with cytokines. TNF-α (20 ng/ml) (CellGenix) was added on day 5 and the cells cultured for an additional 48 h. Finally, Apo-DC were suspended in 90% autologous, heat-inactivated plasma and 10% dimethyl sulfoxide (Sigma-Aldrich, St. Louis, MO, USA), aliquoted at a concentration of 10–20 × 106 cells/ml, in vials (Nunc, Roskilde, Denmark) and cryopreserved using a gradient freezer (Planer). Apo-DC were mantained at −150 °C until used for vaccination.

Cell viability.  Cell viability was determined by staining with 7AAD/CD45 (Immunotech Coulter) and FCM or after Trypan Blue staining using a Bürker chamber (Paul Marienfeld GmbH & Co. KG, Lauda-Königshofen, Germany) and microscope (Leica, Wetzlar, Germany).

Endocytotic capacity of imDC.  On day 4, a sample was drawn from the culture of imDC and the cells were cocultured for 24 h with irradiated B-cells labelled with PKH67 fluorescent dye (Sigma-Aldrich). DC were stained with HLA-DR-PE (Immunotech Coulter) and dual-positive cells were analyzed by flow cytometry to assess the rate of apoptotic tumour cell uptake.

Cytokine array assay.  Supernatants from the final Apo-DC cultures were collected on day 7 and assayed for 15 different cytokines. These included IL-2, IL-4, IL-5, IL-6, IL-8, IL-10, IL-12p40, IL-12p70, IL-15, IL-17, GM-CSF, interferon-γ (IFN-γ), TNF-α, transforming growth factor-α and CD40L, and were assayed using LINCO-15 plex bead-array assay by Capio Diagnostik, Stockholm, Sweden. Cytokine concentrations (pg/ml) in culture supernatants were determined using standard curves plotted with pure recombinant cytokines.

Results

Monocytes, lymphocytes and granulocytes in peripheral blood and the leukapheresis product were assayed by flow cytometry assayed using CD14/CD45 and CD45 versus side scatter (SSC). The percentages of the various cell populations in CLL patients (n = 10) are provided in Fig. 2A,B. Monocytes, lymphocytes and granulocytes were comparable in peripheral blood and the leukapheresis product.

Figure 2.

 Frequency (%) of cells in (A) peripheral blood; (B) leukapheresis product and (C) CliniMACS CD14+ selected product of the 10 individual chronic lymphocytic leukaemia patients and mean value (mean ± SEM). (D) Frequency (%) of recovered monocytes (mean ± SEM) after CliniMACS-positive selection. SSC, side scatter.

In patients with ≤2% monocytes in the leukapheresis product, up to three separate CD14 selection columns and kits have been utilized to obtain adequate number of monocytes. Following positive selection with the CliniMACS beads, the purity of CD14+ cells was 91 ± 2.2% (mean ± SEM) with contaminating lymphocytes and granulocytes decreasing to 3 ± 0.6% and 4 ± 1.3% (mean ± SEM), respectively (Fig. 2C).

Monocytes recovered in the CD14+ fraction were calculated as percentage yield relative to the total number in the initial leukapheresis product. This was determined either on the basis of CD14/CD45 positivity or CD45+ versus SSC. Monocyte yield in the CD14+ fraction was 48 ± 5% (mean ± SEM) determined by CD14+/CD45+ positivity and 51 ± 5.4% determined by CD45-positive cells versus SSC (Fig. 2D).

For the first five patients, a single selection for CD19+ cells was performed with the CliniMACS system which increased the purity of CD19+ cells used for generating apoptotic bodies from 87 ± 3.1% (mean ± SEM) in the leukapheresis product to 100 ± 0.04% post-selection. For the remaining five patients, the selection step was omitted as high percentage of CD19+ cells in the initial leukapheresis product made further enrichment superfluous.

Cell viability

Monocyte viability was 97% immediately following CliniMACS selection. The viability of mature Apo-DC was 98% ± 0.5 (mean ± SEM) after 7 days of culture (Fig. 3). Viability and recovery of Apo-DC after thawing and immediately prior to vaccination was 96 ± 0.6% and 96 ± 4.7% (mean ± SEM) respectively. Data for individual patients are provided in Table 1.

Figure 3.

 Percentage of apoptotic B cells after 5 Gy irradiation, percentage of dendritic cell (DC) that have taken up at least one apoptotic body and percentage of viable cells in the final product before cryopreservation (7AAD staining) for 10 individual chronic lymphocytic leukaemia patients and mean values (mean ± SEM).

Endocytotic capacity of apoptotic B-cells

Five Gy irradiation followed by overnight culture resulted in apoptosis induction in 80 ± 4.7% (mean ± SEM) of the CD19+ cells. Following coculture, 84 ± 4.1% of the imDC had endocytosed apoptotic bodies as quantified by dual positivity for PKH67 and HLA-DR (Fig. 3).

Apo-DC phenotype

Flow cytometric analysis confirmed that the Apo-DC had a mature phenotype after 7 days of culture with abundant expression of the costimulatory molecules CD80 (95 ± 2.9) (mean ± SEM), CD86 (89 ± 2.7), as well as the maturation markers CD83 (67 ± 6.6) and DC-SIGN (84 ± 6.7). HLA-DR was expressed practically on all the cells (94 ± 2.3). CD14 was expressed on a minority of the cells (14 ± 8) while CD20 expression was not significant above background levels, indicating that no residual non-endocytosed B cells remained in the final product (Fig. 4A).

Figure 4.

 (A) Surface staining (mean ± SEM) of 7-day cultured apo-dendritic cell (DC) of 10 individual chronic lymphocytic leukaemia patients; (B) cytokine levels (pg/ml) in supernatants of 7-day cultured Apo-DC of 10 individual CLL patients (colour bars) and mean values (yellow bars). IL-12, interleukin-12; IFN-γ, interferon-γ; TGF-α, transforming growth factor.

Cytokines production by Apo-DC

Supernatants harvested from Apo-DC cultures on day 7 were analyzed for secreted cytokines by multiplex bead array technology. High levels of IL-8 was noted in all patients. IL-6 was also secreted at high levels by virtually all patients. IL-12p40 was produced in 100–400 pg/ml range (Fig. 4B). The other cytokines (IL-2, IL-10, IL-12p70, IL-15, IL-17, IFN-γ, sCD40L) were detected in very low concentrations.

Discussion

Two previous reports by the same group of investigators have described DC-based vaccination studies in CLL patients. In the first report, DC generated from peripheral blood mononuclear cells of normal healthy volunteers were loaded with lysates or apoptotic bodies of CLL cells and used to vaccinate patients [22]. Four to six vaccinations were given to the patients, each consisting of DC generated from a different HLA-mismatched allogeneic donor. Notwithstanding the HLA incompatibility, the vaccination stimulated antileukemic immune responses. In a more recent study the same investigators generated autologous DC from CLL patients [23]. Two methods were used to enrich precursor monocytes for generating DC. For the first two patients, precursor monocytes were enriched from a leukapheresis product by plastic adherence. For the other 10 patients, precursor monocytes were immunomagnetically enriched from 150 ml of peripheral blood. In the autologous study, lysate rather than apoptotic bodies were used as a source of CLL antigens.

The above studies served an important purpose as they demonstrated that it is feasible and safe to administer DC and whole cell antigen-based vaccines to slowly progressive CLL patients. There are, however, technical constraints in the methodologies used in these studies. The allogeneic DC generated from normal donors has limited utility as the HLA disparity would result in their rapid clearance after administration. With regard to the autologous DC generation, the plastic adherence method yielded enough precursor monocytes only for a single vaccine dose from a leukapheresis product. The problem was partly ameliorated through immunomagnetic enrichment using the MACS system, but even then the patient had to be repeatedly phlebotomized for the 150 ml of peripheral blood required per vaccine dose. Another issue with multiple batch preparation of the cellular vaccine is the variability in properties between the different lots of the vaccine.

Ex vivo generation of monocyte-derived DC is invariably contingent on enriching monocytes from peripheral blood cells. While discontinuous density gradients are commonly used in the research laboratory for this purpose, they cannot be utilized for generation of clinical products under good manifacturing practice conditions as they can rarely be performed in a completely closed system. One approach to enrich monocytes is using counterflow centrifugal elutriation. This technology is based on physical separation of cells by counterflow centrifugation, on the basis of sedimentation velocity depending on their size and density [24]. This method does not require a density gradient pre-enrichment step and can be performed in a closed system. For clinical applications, the elutriation technology has been utilized by us and others using the ELUTRA® (Caridian BCT, Lakewood, CO, USA) platform. Berger et al. [25] demonstrated that following elutriation, monocytes were enriched from a mean of 15.2% in the leukapheresis product to a mean of 83% in the monocyte-rich fraction. In total, 98.5% monocytes were recovered after elutriation. In our previous study [26], the purity of the monocyte fraction after counterflow elutriation depended primarily on the percentage of monocytes in the starting leukapheresis material. We noted that a minimum of 4% CD14+ cells in the starting material was needed to achieve a satisfactory enrichment of monocyte precursors from cancer patients. This constraint of 4% monocyte in the starting material is a major limitation of the elutriation technology when dealing with samples of CLL patients who usually have very few circulating monocytes in the peripheral blood.

As showed by our previous study [27], our technique for enriching monocytes using the CliniMACS® system and methods for ex vivo generation and cryopreservation of Apo-DC, permitted five or more doses of vaccine, produced from a single leukapheresis product from CLL patients with >1% CD14+ monocytes. Also, each dose consisted of a minimum of 1 × 107 cells. The use of multiple immunomagnetic columns increased the efficiency of the process to the extent that adequate vaccine doses could be produced even in patients with barely discernable DC precursors in their circulation. Three significant advantages of this process are (1) minimum discomfort to the patients as they are subjected to only a single apheresis procedure, (2) elimination of batch-to-batch variations as the vaccine is prepared as a single lot, aliquoted and cryopreserved and (3) inclusion of patients with very low circulating monocytes extending this therapy option to a larger patient population.

A limitation, however, of the immunomagnetic enrichment method utilized for CLL patients in our application is the number of cells from the starting leukapheresis product that can be loaded onto the column. Indeed, the CD14+ CliniMACS column has a maximum loading capacity of 4 × 109 CD14+ cells with a total maximum load of 20 × 109 leukocytes. The number of cells to be retained by the columns is very low relative to the number of total cells passing through, which can overload the system. To overcome this problem, we divided the initial leukapheresis products into three fractions which were loaded on separate CD14+ CliniMACS columns. This permitted enrichment of adequate numbers of CD14+ monocytes with high purity even when they were present at very low frequency in the circulation of the patient.

Depletion of CD19+ cells is theoretically an alternative strategy for enriching monocytes. However, the preponderance of leukemic cells in the leukapheresis products makes this approach impractical for CLL patients.

Meticulous attention to quality assurance and quality control of the individual steps in the production of a cellular vaccine and proper product specification and documentation with regard to functional and phenotypic characteristics is indispensible for accurate interpretation of safety and efficacy results following vaccination [28].

Moreover, the information on the phenotypic differences of individual vaccine preparations should be as comprehensive as possible, in order to highlight possible correlations with the immunological and clinical responses observed in the different patients. An integrated component of our production process was detailed documentation of a number of phenotypic and functional characteristics of the cellular vaccine, including frequency of endocytosis of the apoptotic bodies and secretion of cytokines by the Apo-DC following maturation with TNF-α.

The cell viability throughout the whole production procedure beginning with the enriched precursor monocytes to the final cellular vaccine was high (≥97%). The rate of uptake of apoptotic B cells by DC was substantial (84 ± 13%) and the presence of B cells in the final Apo-DC product was negligible. Thus the purity of the final product was higher than the 80% indicated which has been previously suggested as the minimum acceptability criterion [28]. Thus in summary, our described method of enriching monocytes using a combination of leukapheresis and affinity-based technologies (CliniMACS) yielded a large number of clinical-grade, highly purified monocytes and allowed the production of a sufficient amount of DC vaccine that met accepted and established quality criteria. The technology yields a highly reproducible outcome despite considerable variability in the starting material from different patients which make this a highly suitable procedure for patients with very low frequency of monocytes in peripheral circulation.

Acknowledgment

Supported by grant from The Swedish Cancer Society, The Cancer Society in Stockholm, King Gustav V Jubilee Fund, The Cancer and Allergy Foundation, The Karolinska Institutet Foundation, The Stockholm County Council, Miltenyi Biotec GmbH, EU-DC-Thera and Felix Mindus Research Foundation. For excellent secretarial help we thank Ms. Leila Relander.

Ancillary