Immunotherapy of cancer must promote antitumor effector cells for tumor eradication as well as counteract immunoregulatory mechanisms which inhibit effectors. Immunologic therapies of cancer are showing promise, including dendritic cell-(DC-) based strategies. DC are highly malleable antigen-presenting cells which can promote potent antitumor immunity as well as tolerance, depending on the environmental signals received. Previously, we tested a peptide-pulsed DC vaccine to promote Alpha-fetoprotein (AFP-) specific anti-tumor immunity in patients with hepatocellular carcinoma (HCC), and reported on the CD8+ T cell responses induced by this vaccine and the clinical trial results. Here, we show that the peptide-loaded DC enhanced NK cell activation and decreased regulatory T cells (Treg) frequencies in vaccinated HCC patients. We also extend these data by testing several forms of DC vaccines
Hepatocellular carcinoma (HCC) is the third leading cause of cancer mortality worldwide [
Alpha-fetoprotein (AFP) is an oncofetal antigen that is expressed by more than half of HCC tumors and detectable at elevated levels in the blood and tumor microenvironment in these HCC patients [
Dendritic cell (DC) vaccines are promising vehicles for activating antitumor specific T cells and NK cells for tumor immunotherapy. They are immunologic sentinels which can induce antigen-specific immunity or tolerance [
While DC are critical to induction of immunity, other immune cells are important as effectors and regulators in cancer immunity. CD56+CD16+/- natural killer (NK) cells are the effectors of the innate immune system that are able to directly kill tumor or virally infected cells with reduced levels of MHC class I molecules or that overexpress stress-induced activating cell surface molecules (e.g., MICA/B recognition via NKG2D), and that otherwise may escape immune detection. Hepatic lymphocytes are enriched (up to 30%) in NK cells that may play a role in antitumor defense [
We previously tested an AFP peptide-pulsed DC vaccine in a phase I clinical trial. The vaccine was found to be safe and immunogenic in late-stage HCC patients [
It has been demonstrated that DC and NK cells are capable of interacting with and activating each other [
PBMC were obtained from healthy volunteers (HD) and from HCC patients enrolled in a peptide-pulsed DC vaccine (UCLA IRB #00-01-026, IND BB9395; UPCI #04-001 and #04-111; informed consent was obtained from all patients and donors). The clinical trial was previously published in detail [
HCC patient demographics.
Pt. | Dose DC | Risk Factor | Stage | Previous treatments1 | Pre-AFP (ng/mL) | Post-AFP | Response2 | PFS3 | OS4 |
---|---|---|---|---|---|---|---|---|---|
A1 | 1 × 106 | ? | IVb | Chemoembo, CDDP, Adriamycin, 5-FU, Xeloda, Thalidomide | 2.811 | 2.748 (+28) | PD | 0 | 4 |
A2 | 1 × 106 | HBV | IVa | Chemoembo | 4.740 | 5.770 | PD | 0 | 20 |
A35 | 1 × 106 | EtOH | IVa | RFA | 3.080 | (no DC) | (no DC) | 0 | 2 |
A46 | 1 × 106 | HCV | IVb | — | 10,800 | 10.650 | (1 vaccine) | 0 | — |
B2 | 1 × 106 | ? | IVa | Surgery | 5.100 | 7.650 | PD | 0 | 4 |
B37 | 5 × 106 | HCV | IVa | Chemoembo RFA | 102 | 65 | NE | 0 | 35 |
B5 | 5 × 106 | HBV | IVa | Chemoembo, CarboTaxol, Xeloda | 1.630 | 2.515 | PD | 0 | 3+ |
B8 | 5 × 106 | HCV | IVb | Chemoembo | 96.7 | 134 | PD | 0 | 5+ |
1Previous treatments received (chemoembo, chemoembolization; CDDP, cis-platin; 5-FU, 5-flouro-uracil; Xeloda, capecitabine; RFA, radiofrequency ablation; carbo, carboplatin; XRT, radiation therapy).
2PD: progressive disease, NE: no evidence of disease.
3PFS: progression free survival.
4OS: overall survival.
5No DC: no DC vaccines could be generated which passed clinical protocol release criteria.
61 vaccine: patient progressed early and did not receive the 3 DC vaccinations.
7 NE: patient B3 responded to chemoembolization and RFA and was vaccinated shortly thereafter, and had 35 mo. OS.
Cells were stained according to manufacturer recommendations, fixed in 0.5% paraformaldehyde, and analyzed on an CyAn high-speed analyzer (Dako, Carpinteria, Calif) (UPCI Flow Cytometry Facility) and the Summit v4.3 software within four days. NK cell phenotype was investigated using: CD8 PE, CD16 ECD, CD3 APC (Beckman Coulter), granzyme B FITC, CD25 PE-Cy7, CD56 PE-Cy5, and CD69 APC-Cy7 (BD Pharmingen). Treg were investigated using: CD4 FITC, FOXP3 PE, and CD25 APC (eBioscience) and reported as either the FOXP3 positive percentage or the MFI of FOXP3 expression in the CD4+CD25+ cells.
CD69 and CD25 expression on CD56loCD16+ and CD56hiCD16− NK cells. Phenotyping of NK cells from patients who received the AFP pep/DC vaccine, showing longitudinal changes. “Pre” denotes PBMC from time point 0. DC vaccines were delivered (after blood draws) on days 0, 14, and 28. “Post” denotes PBMC from postvaccine administration at time points available in the remaining batched PBMC. Patient A1 tested at days 35 and 56 (7 days and 28 days after the third vaccine); pt B2 at d28, 56, and 112; pt B5 at d14 and 28; pt B8 and d14 and 112, pt A4 at d14. CD69 and CD25 markers (by MFI) are shown for both NK cell subsets. Percent positivity is shown in Supplementary Figure 2. One-tail
CD56+ NK cells and CD4+ T cells were isolated from PBMC (Miltenyi Biotech) according to the manufacturer’s directions (CD56 beads, NK isolation kit, CD4+ T cell isolation kit). Change in MFI was considered “positive” if the increase was ≥25% of the baseline MFI. Percent positivity was considered positive if ≥5% greater than baseline.
Monocytes were isolated from PBMC using adherence to T75 flasks (Costar). They were cultured for 6-7 days in RPM1640/5% human AB serum/PennStrep medium with 500 U/mL IL-4 and 800 U/mL GM-CSF (Schering-Plough, Kenilworth, NJ; Amgen, Thousand Oakes, Calif) to promote differentiation to myeloid DC.
After culture, DC were harvested, counted (Trypan Blue Stain; BioWhittaker, Walkersville, MD), and cultured as described below. DC were subsequently cocultured with NK cells or T cells isolated from the autologous donor and incubated 24 hr (NK and CD4) to 5 days (CD4) at ratios of 1 DC to 1–10 NK or T cells. After the coculture, cells were harvested, supernatant was collected and stored at −80°C, and cells were analyzed by flow cytometry as described above.
For peptide-pulsed DC (pep/DC), DC were pulsed with 1 or 2 specific AFP peptides (AFP158 FMNKFIYEI and AFP542 GVALQTMKQ; synthesized at the University Pittsburgh Peptide Synthesis Facility) at 10
For AdV-transduced DC (AdV/DC), DC were transduced for 2 hr at 37°C in serum-free media (IMDM) at MOI = 1,000 with an AdV encoding full length AFP protein (AdVhAFP) [
Cell-free supernatants were collected from cultures and frozen at −80°C. They were subsequently thawed and simultaneously analyzed with the multiplex Luminex assay (Invitrogen) per manufacturer’s protocol in a BioRad reader (UPCI Immunologic Monitoring Laboratory). The following analytes were tested: GM-CSF, IFN-
One-tail
Based on our previous study (17), we hypothesized that HCC patients vaccinated with immature DC pulsed with AFP peptides (pep/DC) would not impact activation of circulating NK cells. We assessed this by evaluating upregulation of CD69 or CD25 activation markers on CD56hi/CD16− and/or CD56lo/CD16+ NK cell subsets. We also wished to determine whether Treg frequency (as determined by a change in FOXP3-expressing CD4+CD25hi T cells) was modulated by vaccination, which might also include changes in CTLA-4 [
We tested banked PBMC samples from five HCC patients, isolated at different time points during vaccination with AFP pep/DC. Cells were stained immediately after thawing to assess phenotype by flow cytometry (see analysis strategy shown in Supplemental Figure 1 in Supplementary Material available online at doi:10.1155/2011/249281). Contrary to our hypothesis, both regulatory CD56hiCD16− and cytotoxic CD56loCD16+ NK cells demonstrated activation post-pep/DC vaccination, compared to baseline. Activation was determined by both an increase in population MFI (Figure
To examine Treg cell frequencies, CD3+CD4+ T cells were gated on CD25hi or total CD25+ and intracellularly stained for FOXP3. The Treg lymphocyte frequencies were then assessed by flow cytometry. FOXP3 expression in the CD3+CD4+CD25hi T cells showed a consistent change, decreasing overall in 4/5 of the patients tested, by both percent positivity and MFI (Figure
FOXP3 expression in CD4+CD3+CD25hi (Treg) cells. Phenotyping of Treg cells from patients who received the pep/DC vaccine, showing longitudinal changes. “Pre” denotes PBMC from time point 0. “Post” denotes PBMC from postvaccine administration at time points available. Patient A1 tested at days 35 and 56; pt B2 at d28, 56, and 112; pt B5 at d14 and 28; pt B8 and d14 and 112, pt A4 at d14. (a) FOXP3 assessed intracellularly in CD3+CD4+CD25hi cells, showing MFI, or (b) % positivity. One-tail
The AFP peptide-pulsed DC did not undergo a specific maturation step during vaccine preparation. Maturation cocktails can impact surface levels of MHC class I and II, costimulatory molecule levels, and cytokine production. We hypothesized that different DC antigen-loading strategies, some of which impact DC maturation, would result in unique phenotypic changes in the DC that would impact activation and frequencies of other immune cells (like NK cells and Treg) they interacted with. We previously tested AdV-mediated genetic engineering of DC to enable expression of full length antigens in DC [
We hypothesized that NK cells would be activated (as measured by increased CD69 and CD25 expression) and that Treg frequencies might be reduced (decreased FOXP3 expression) after interactions with DC that were at least partially matured and that these trends would be observed after coculture with the AdV/DC. Monocyte-derived DC were antigen-loaded as described and cocultured with autologous NK or CD4+ cells for 24 hr or 6 days, respectively. The cells were then harvested and assessed by FACS for phenotypic changes in specific subsets.
We first assessed different AFP antigen loading modalities (peptide, protein and AdV) in HD cells. Peptide pulsing and protein-loading do not include any maturation agents and have been observed not to alter DC phenotype
CD69 expression on healthy donor CD56hi and CD56loCD16+ NK cells. Phenotyping of NK cells from HD PBMC after 48 hr coculture with different DC groups, showing CD69 upregulation on the (a) CD56loCD16+ and (b) CD56hiCD16− subsets for three HD.
We then tested the impact of differentially antigen loaded and matured DC on NK cell activation of HCC patients. HCC patient CD56loCD16+ NK cells showed increased level of activation (CD69 MFI and percent positivity) after coincubation with DC (3/4 patients, particularly with AdVhAFP/DC, Figure
CD69 expression on HCC patient NK cells. Phenotyping of NK cells from HCC patients after 48 hr coculture with different groups of DC, showing CD69 upregulation on CD56hiCD16− and CD56loCD16+ subtypes, by MFI (a) and percent positivity (b).
In order to determine any impact of DC antigen loading on Treg expansion
Treg cell responses to DC coculture. HD (a) and HCC patient (b, c, and d) CD4+ T cells were cocultured with differentially treated DC. The FOXP3 expression in CD3+/CD4+/CD25hi cells is shown in (a) as MFI (left), percent positive (middle). The right group is the frequency of total activated (CD25+) CD4+ T cells. The FOXP3 MFI in Treg in HCC patients is shown in (b). The percent CD3+/CD4+/CD25hi/FOXP3+ Treg in HCC patients is shown in (c). The overall frequency of activated CD4+ T cells is shown in (d) (% CD3+/CD4+/CD25+ cells).
In order to characterize the DC-lymphocyte (NK cell or CD4+ T cell) interaction environment, cell-free supernatants were collected from the different HD and HCC patient cell coculture experiments. Supernatants were tested by multiplex Luminex assay to simultaneously assess the levels of cytokines, chemokines and growth factors, including: GM-CSF, IFN-
Luminex results: production of chemokines and cytokines. Graphs are grouped according to scale of cytokine production and function. (a) IFN-
Coculture of NK cells with the differently antigen-loaded DC groups yielded minimal levels of IFN-
Immunotherapy holds potential for treatment of hepatocellular carcinoma, as few effective treatments are available, and immunotherapy vaccine strategies have largely shown immunogenicity and less toxicity than current chemotherapy [
An effective vaccine against HCC would activate not only tumor antigen-specific adaptive immune responses, but also innate NK cell effectors to crosstalk with DC, promote type I responses, and potentially also directly kill HCC cells. In addition, downregulating Treg cells would help to minimize immunosuppression and potentially allow enhanced antitumor effector function. By testing PBMC from the peripheral blood of patients treated with the AFP pep/DC vaccine, we found evidence for activation of NK cells in most patients, as shown by increase in CD69 and CD25 expression. We also found evidence for downregulation of Treg cells in most patients, as shown by decreased FOXP3 expression in those CD4+CD25+ T cells. These results illustrate the possibility of rationally modulating the immune system with DC to increase anti-HCC immunity. While additional functional assays of NK cell killing and Treg suppression would have strengthened our report, there were insufficient banked PBMC remaining for such assays.
In this data set, A1, A4, and B2 received 106 DC/vaccine, and B5 and B8 received 5 × 106 DC/vaccine, and all were stage IV (Table
By testing
Cytokine production in response to DC vaccine co-culture is a functional measure of activation of NK and Treg cells. It was interesting to note that cytokines and chemokines tested for by Luminex were produced more abundantly by HD-derived cells than HCC-derived cells. This again highlights the difficulty of inducing an antitumor immune response in HCC patients, and suggests that additional immune stimulatory and immune suppression reducing efforts may be required to promote the desired antitumor immunity
We have performed additional preliminary studies comparing AdVLacZ and AdVhAFP in HD DC, and we find a reduction in DC surface transmembrane TNF expression (but similar
In conclusion, we find that DC-based vaccines can modulate not only antigen-specific T-cell responses, but also innate effectors and counter-regulatory mechanisms. Optimal antigen loading of DC and maturation signaling may allow for development of DC vaccines which will subsequently deliver specific signals to the broad array of tumor-reactive cells they encounter
hepatocellular carcinoma
dendritic cells
Alpha-fetoprotein
adenovirus
mean fluorescence intensity
regulatory T cells
peripheral blood mononuclear cells.
L. H. Butterfield is coinventor of patents covering aspects of AFP as a target for T cell-mediated anti-HCC immunity.
This study was supported by the University of Pittsburgh Cancer Institute and NCI RO1 CA 104524 (LHB); T35 DK065521 T. Kleyman (for SMB). The authors thank the UPCI Flow Cytometry Facility (A. Donnenberg), the University of Pittsburgh Vector Core Facility (A. Gambotto), and the UPCI Immunologic Monitoring Laboratory (for Luminex assays, for which we gratefully acknowledge Sharon Sember).