Dendritic cell vaccines containing lymphocytes produce improved immunogenicity in patients with cancer

Two separate DC vaccination studies were conducted for prostate cancer patients, the first using apoptotic LNCaP cells, as reported previously [11], and the second using apoptotic PC3 cells. Patients in the first study were vaccinated with DCs pulsed with LNCaP (DC/LNCaP) and DCs pulsed with LNCaP transfected with influenza M1 protein (DC/LNCaP-M1) [11]. In the second study patients were vaccinated with DCs pulsed with PC3 (DC/PC3) and DCs pulsed with PC3 transfected with influenza M1 (DC/PC3-M1). The ratio of DCs to apoptotic LNCaP or PC3 tumor cells was 1:1. In both studies, the acceptable dose range was 1-10⁶ DCs of each type at each time point, regardless of the method of DC preparation. Patients in both studies were also given DCs pulsed with KLH (DC/KLH) as a control antigen. Initial vaccination was followed by 3 booster vaccine immunizations, each 2 weeks apart, administered subcutaneously. In both studies, leukocytes for immunomonitoring were collected by leukapheresis at baseline and again 6 weeks after the last booster.

In a third study, patients with primary brain tumors were vaccinated with DCs pulsed with autologous apoptotic tumor cells and DC/KLH. In this study, both the number of boosters and the timing of the post-vaccination leukapheresis were different from the first 2 studies. Patients were vaccinated with either 2 or 3 doses every 3 weeks intradermally and leukapheresed 2 to 3 weeks after the 2^nd dose. Here, we report only the patients’ responses to DC/KLH as relevant and not the responses to the DC vaccine to brain tumor. In all 3 studies, the first dose of DC administered was “fresh” and all subsequent booster doses were thawed doses. All studies were conducted at Rockefeller University Hospital after Institutional Review Board approval. Written consent was obtained from all patients. Study identifiers on clinicaltrials.gov were: NCT00289341, NCT00345293, and NCT00893945.

DCs were prepared as previously described [11]. Briefly, leukapheresates were placed over lymphocyte separation media and the buffy layer was collected and washed. PBMCs were then plated in RPMI-1640 supplemented with 1% autologous plasma and allowed to adhere at 37°C. After 1 hour, the non-adherent cells were removed. The adherent cells were differentiated in RPMI-1640 supplemented with 1% autologous plasma, GM-CSF (Genzyme) and IL-4 (R & D Systems) over 6 days, at which point they are non-adherent and considered immature DCs. LNCaP and PC3 cells were obtained directly from American Type Cell Culture (CRL-1740 and CRL-1435) and cell banks were established as described [11]. LNCaP or PC3 cells were UV irradiated and cultured with immature DCs at a 1:1 ratio with PGE₂ (Sigma) and TNFα (R & D Systems, Miltenyi) over 36-48 hours. A subset of immature DCs was cultured with KLH (biosyn). The cells were harvested on the 8^th day, washed and resuspended in 5% dimethyl sulfoxide and 10% human serum albumin (HSA, Grifols) in normal saline for freezing or administration.

Leukapheresates were washed with PBS/EDTA supplemented with 2% HSA, incubated with CD14 MicroBeads (Miltenyi Biotec) for 15 minutes then washed. CD14+ cells were isolated using the CliniMACS System (Miltenyi Biotec). Positively selected cells were washed and plated in RPMI-1640 supplemented with 1% autologous plasma, GM-CSF (Genzyme) and IL-4 (R & D Systems) over 6 days. Cells were pulsed with tumor cells, matured and harvested in the same manner as Adherence DCs.

Lymphocyte proliferation responses, pre- and post-vaccination, were measured by ³H-thymidine incorporation assays as previously described [12] with the following changes. APCs were either pre-vaccination CD14+ cells or DCs. ³H-thymidine was added for the last 20 hours of culture. Results are presented as average count per minute (CPM) of 3 to 6 replicate wells.

As part of release criteria testing, DCs were counted and stained with HLA-DR, CD14, CD83 antibody (Becton Dickinson) and propidium iodide (Serologicals). For additional phenotyping, cells were stained with CD86, CD40, CCR7, CD4, CD8, CD69, and CD19 antibody (Becton Dickinson). Data was acquired on the FACSCalibur (Becton Dickinson) or MACSQuant VYB (Miltenyi Biotec) and data were analyzed using FlowJo software.

Immature Selected or Adherence DCs were stained with HLA-DR-APC and cultured with apoptotic PKH26 stained allogeneic lymphocytes with or without EDTA for 24 hours. Analysis was restricted to HLA-DR+ cells and the phagocytosis of apoptotic cells by DC groups were identified by positive PKH26 dye staining.

As previously described [12], DCs were cultured with the non-adherent fraction of allogeneic PBMCs.

Supernatants of Adherence and Selected DCs were assessed for the presence of cytokines using the Meso Scale Discovery Human ProInflammatory 9-Plex plate and read on the SECTOR Imager (Meso Scale Discovery). DCs were prepared using each of the 2 methods from 3 donors.

The IFNγ ELISPOT assay was done as previously described [12] with the following modifications. Influenza-infected DCs (DCF) were used to elicit influenza-specific IFNγ responses. Adherence DC vaccine preparations were harvested on day 8 and re-plated on an IFNγ ELISPOT coated plate at a density of 25,000 lymphocytes/well.

PBMCs or Adherence DC preparations from A0201 donors were co-cultured with apoptotic 3T3 (DC/ctrl) or apoptotic influenza-infected 3T3 cells (DC/flu) for 5 hours in GolgiStop (BD Bioscience) before staining with APC-conjugated A0201 Influenza-M1 dextramer (GILGFVFTL, Immudex) followed by CD8-FITC (BD Bioscience). Cells were fixed and permeabilized (Cytofix/Cytoperm, BD Bioscience) as per manufacturer’s protocol. Intracellular cytokine staining was performed using IFNγ − PE (BD Bioscience) followed by flow cytometry analysis.

To evaluate the difference between proliferation responses between the patients given Selected DCs and Adherence DCs, the Mann-Whitney test was used. This test was also used to describe differences in the number and phenotype of the 2 types of DCs given. The paired t-test was used to describe differences in all other assays. P-values less than 0.05 were considered statistically significant.

Seven patients received Selected DCs and three received Adherence DC vaccine preparations, all pulsed with apoptotic PC3 cells. We compared the post- minus pre-vaccination proliferation responses of these patients. There was a significantly higher proliferation response to PC3 in patients that received Adherence DCs as compared to those who received Selected DCs (Figure 1A, p = 0.033; Additional file 1: Figure S1). Similarly, there was a significantly higher proliferation response in the Adherence versus the Selected DC vaccines to LNCaP, another prostate cancer cell line with which this cohort was not vaccinated (p = 0.033). Background post-vaccination proliferation responses to APCs not pulsed with antigen did not differ between the patients receiving Adherence or Selected DC/PC3 vaccines (No Ag, p = 0.867). This data indicated that the DC/PC3 vaccines prepared by Adherence method elicited higher antigen specific immune responses than those prepared by the Selection method.

This observed difference in immunogenicity was not anticipated. Therefore, we confirmed this by pooling data from other DC vaccine trials conducted by our laboratory, including the previously reported study in prostate cancer patients using DC/LNCaP and an additional DC vaccine study of patients with primary brain tumors. Since the tumor type was different between these studies, tumor specific responses could not be reasonably compared. However, all patients were vaccinated with DC/KLH made with either Selected or Adherence DCs, therefore, proliferation responses to KLH were examined. There were a total of 10 evaluable patients who received Selected DC/KLH and 31 patients who received Adherence DC/KLH. The vaccination induced significantly higher lymphocyte proliferation responses to KLH in patients that received Adherence DCs than those that received Selected DCs (Figure 1B; p = 0.023). No difference was observed in the proliferative lymphocyte responses to negative control APCs (not pulsed with antigen, p = 0.576). This result indicates that Adherence DCs consistently elicit higher immunogenicity even in different populations of cancer patients.

The differences in immunogenicity were not a result of differences in DC dose, phenotype, such as CD83 or CD14 expression or a result of differences in DC viability. The total number of DC/PC3, DC/PC3-M1, and DC/KLH administered also did not differ between Adherence and Selected vaccine preparations (Table 1). Within the second study of DC loaded with apoptotic PC3, the number of cells administered, the cell surface expression of CD83 and CD14, and viability of DC/PC3 and DC/PC3-M1 were similar. Further, in all three studies (DC/LNCAP, DC/PC3 and DC/autologous brain tumor), these same parameters are also similar in the DC/KLH that were administered. Given such similarities between Adherence and Selected DCs, DC phenotype, viability, and dose do not explain the difference in immunogenicity.

To test whether Adherence or Selected DCs express different levels of markers of maturation, DCs were prepared using both methods from the same donor’s PBMC. Adherence and Selected DCs were assessed for the surface markers CD14, HLA-DR, CD83, CD86, CD40 and CCR7, but no differences were found (Figure 2A). We next evaluated the in vitro function of the two types of DCs. First, we assessed their ability to phagocytose apoptotic cells, which could affect antigen presentation. Selected or Adherence immature iDCs were cultured with PKH26 stained apoptotic lymphocytes. Cells positive for both HLA-DR staining and PKH26 dye indicated DC phagocytosis of apoptotic cells. 66.4% of Selected and 67% of Adherence iDCs phagocytosed apoptotic cells, indicating no difference in their phagocytic activity (Figure 2B). This process was calcium dependent and inhibited in the presence of EDTA, indicating that apoptotic cells were phagocytosed and not positive for both markers simply by adhering to one another or other non-specific mechanisms.

We next assayed the functional ability of both DC types in an allo-MLR assay. Allogeneic lymphocytes were cultured with Adherence or Selected DCs prepared from PBMC of the same donor. The proliferative response of allogeneic lymphocytes to Adherence DC was not different than their response to Selected DC (Figure 2C, p = 0.483). We then tested the ability of the DC groups to stimulate lymphocyte proliferation in an antigen specific manner. Selected or Adherence DCs were co-cultured with either apoptotic 3T3 cells infected with influenza (DC/flu) or uninfected apoptotic 3T3 cells (DC/ctrl). DC/flu or DC/ctrl were then cultured with syngeneic lymphocytes. Both the Selected and Adherence DCs strongly stimulated influenza-specific lymphocyte proliferation (Figure 2D, p = 0.082 at peak proliferation), again pointing to a similar profile of DC function. Lastly, we assessed the differences in DC ability to induce an IFNγ response in T cells in an antigen specific manner. DC (DC) or influenza infected DC (DCF) were co-cultured with either CD4 or CD8 T cells in an ELISPOT assay. Again, both groups of DCs were able to induce a comparable CD4 and CD8 T cell IFNγ response to influenza (Figure 2E, p = 0.250 and p = 0.822 respectively). In summary, these assays indicate that there is no detectable difference in DC number, phenotype or function that could account for the observation that Adherence DCs produced increased immunogenicity in patients with cancer.

We next looked beyond the DCs for an explanation for why the two vaccine preparations differed in their ability to stimulate T cell proliferation. Upon examination of the 10 most recent Adherence and Selected DC vaccines administered, we found that Adherence DC vaccines contained a median of 10.8% lymphocytes (range 4.7 to 25.0%, Figure 3A and data not shown), determined by forward and side scatter on flow cytometry. In contrast, the Selected DC vaccines contained very few lymphocytes (median 1.0%, range 0.3 to 1.8%). Thus, we turned our attention to assessment of this population of administered cells. The lymphocytes in the Adherence DC preparation contained a mixture of CD4, CD8 and CD19 cells (Figure 3A). These lymphocyte populations were resting at the time of initial harvest from patients’ blood (Day 0), but were highly activated after 8 days of culture within the DC vaccine preparation. Both CD4 and CD8 T cells up-regulated CD69 and HLA-DR expression and CD19 B cells up-regulated CD40, HLA-DR, and CD83 expression compared to baseline PBMC expression at Day 0 (Figure 3B).

To assess lymphocyte proliferation within the DC preparations, both Adherence and Selected DCs were pulsed with ³H-thymidine. Adherence DC preparations were found to have incorporated more ³H-thymidine than Selected DCs, an indication that the lymphocytes were proliferating in these cultures (Figure 4A, p = 0.03, p = 0.003, and p = 0.021 respectively at 48, 72, and 96 hours). In addition, when we examined the Adherence and Selected DC cultures for the presence of cytokines after 6 days in culture, we found that the Adherence DC supernatants contained a mean of 4.42 pg/ml IL-2 while the levels in the Selected DC culture supernatants were below the level of detection (<0.68 pg/ml), a statistically significantly increase (Figure 4B, p = 0.012). This supports the observation that lymphocytes present in the Adherence DC culture were proliferating. There was a similar trend after 2 more days of culture in fresh media but this difference was not statistically significant (p = 0.070). There were no differences in the concentrations of IL-8, IL-12p70, IL-1β, IFNγ, IL-6, and IL-10 found in the supernatants of Adherence or Selected DC cultures. In conclusion, Adherence DC vaccines contain both DCs as well as CD4+, CD8+ and CD19+ lymphocytes that are activated in vitro, are proliferating, and are producing IL-2.

Since patient vaccines were prepared by culture of immature (Day 6) DCs with apoptotic tumor cells, we hypothesized that the activated lymphocytes in Adherence DC cultures may expand in an antigen specific manner to apoptotic tumor cells. In this scenario, Adherence DC vaccines could contain an expanded population of tumor cell-specific, activated lymphocytes in addition to DC loaded with tumor cell antigen. Since we did not have enough cancer patient derived specimens to test this directly, we tested the more general hypothesis that lymphocytes contained in Adherence DC preparations proliferate in an antigen specific manner to antigens added to DC cultures on Day 6.

Adherence DCs from healthy donors were co-cultured with either apoptotic influenza infected cell lines (DC/flu) or uninfected control cells (DC/ctrl) and tested for proliferation. There was increased proliferation (Figure 5A, p = 0.009 at 96 hours) in Adherence DC/flu cultures compared to Adherence DC/ctrl cultures. To test whether there were antigen specific CD8+ IFNγ secreting T cells, analogous to those which would be critical for the generation of tumor immunity, we evaluated the cultures for expansion of influenza M1 specific dextramer positive cells. On day 10 of Adherence DC/flu culture, we found an increase in the percentage of influenza M1+ CD8+ cells (0.16%) compared to DC/ctrl cultures (0.0623%) and baseline PBMCs (0.047%, Figure 5B, upper panels). In addition to an increase in the frequency of antigen specific cells, we also found evidence of their activation; on day 10, 75% of the M1-specific CD8 T cells within the DC/flu culture produced IFNγ (Figure 5B, lower panels). Similar results were obtained when this experiment was repeated with a second donor (52% of M1-specific CD8+ T cells produced IFNγ on day 10). Earlier, when we assayed the supernatant for cytokines after 6 and 8 days of culture, no IFNγ was detected due to the lack of antigen. However, when DCs are co-cultured with antigen such as influenza as it was in this culture, not only are we able to detect antigen specific proliferation but we are now able to detect IFNγ within that subset. In conclusion, Adherence DC cultures contain a population of proliferating, activated lymphocytes not present in Selected DC cultures and these can expand in an antigen specific manner.