Introduction
Cancer remains one of the leading causes of death worldwide with approximately 14 million new cases in 2012 and 8.8 million related deaths recorded in 2015 [
1]. Breast cancer represents 14% of the total global cancer-related deaths in females [
2]. Breast cancer staging was defined according to standard guidelines [
3]. Stage 1 was defined as a tumour < 20 mm in size that was confined to one breast only. Stage 2 was defined as a tumour < 50 mm in size with or without malignant cell invasion of auxiliary lymph nodes and/or lymph nodes near the breastbone. Stage 3 was defined as a tumour > 5 cm which had spread to auxiliary lymph nodes and/or lymph nodes near the breastbone or any size tumour that has spread to other areas within the breast. Stage 4 was defined as breast cancer that has metastasised beyond the breast to the lungs, lymph nodes, skin, bones, liver or brain.
Surgery for the most part is an effective treatment method, but its success is limited to the early stages of the disease before breast cancer has metastasised. Other forms of therapy, including chemotherapy, are partially effective but are associated with substantial and severe adverse events. Thus, new therapeutic options are urgently required.
DC are potent antigen presenting cells, which prime and activate T-cells during microbial or viral infection [
4]. DCs offer an attractive immunotherapeutic option because they can be primed with different antigens in vitro to target different diseases in vivo. Various TLR agonists (e.g. TLR-3 [Ampligen® and Poly I:C] and TLR-7/8 [R848]) have been used to mature DCs in vitro for use as immunotherapeutic agents against malignant melanoma [
5], prostate cancer [
6], malignant glioma [
7] and renal cancer [
6]. These DCs have the ability to express bioactive IL-12p70, IFN-α, IFN-γ, and TNF-α [
8‐
10], indicating that they can support anti-tumour Th1 responses. By contrast, earlier DC vaccines could cross-present tumour antigens but lacked either co-stimulatory ability or lymph node homing capacity, or they produced low levels of IL12-p70, which is essential for Th1 polarising immunity [
11]. The ability of DCs to produce IL-12p70 has been shown to directly translate to clinical benefits in vivo [
12‐
14].
Over the last 5 years’ clinical trials have been conducted involving different cancers using different DC vaccines, which support the efficacy of DCs as immunotherapeutic agents [
15‐
17]. Notably, these studies evaluated vaccines developed using cancer cell lines. However, in contradistinction to cell lines there is considerable antigenic variability amongst tumours from different individuals with the same type of cancer. For example, the commonly used MCF-7 breast cancer cell line does not expresses some antigens that are highly expressed in 75–80% of breast cancers encountered in clinical practice [
18‐
21]. This may result in poor vaccine efficacy because of tumour antigen heterogeneity [
4,
16]. To address this issue, we tested DC vaccine efficacy to the patient’s own tumour cells in vitro (and to our knowledge the first study to do so) by recruiting female patients with stage 1, 2 and 3 breast cancer.
We show that we can optimally mature patient-derived DCs in vitro with tumour-specific lysate, Ampligen®, an IFN-containing cocktail (IFN-α, IFN-γ, IL-1β, CD40L) and R848. We cultured and used patient-derived primary breast cancer cells as “targets” to test the efficacy of the DC vaccine in vitro. The mature DCs had the ability to prime effector cells, which resulted in Th1 cytotoxic CTL-mediated killing of the patient’s own breast cancer cells in vitro. We further show that the mature DCs were sterile, endotoxin/mycoplasma free, and they maintained their mature phenotype and high viability 2 months’ post-cryopreservation.
Methods
Study site and population
Women undergoing surgery as the standard of care at Groote Schuur Hospital in Cape Town, South Africa were identified as potential participants. Patients over the age of 18 and diagnosed with stage 1, 2, or 3 breast cancer were recruited to the study and written informed consent was obtained. A clinical research form was completed for every patient recruited, which indicated age, reproductive status and medication status. Exclusion criteria included (i) patients undergoing immunotherapy, (ii) patients receiving immunosuppressive medication (iii) patients on hormonal treatment for breast cancer, (iii) active second malignancies, i.e. any malignancy not treated with curative intent within the last 5 years, (iv) patients with auto-immune disease, (v) any substance abuse. All participants agreed to donate a piece of malignant breast tissue and to undergo a leukapheresis procedure at a later date.
Autologous breast cancer cell culture
Approximately, ten 10 mm × 2 mm biopsy specimens (mean weight = 244 mg; Table
1) were obtained from the core of each tumour post-surgery (mean size = 22 mm × 21 mm [w × d]; Table
1) and the tissue was cut into 1 mm by 1 mm pieces and separated into two equal portions; for autologous breast cancer cell culture and for the generation of a tumour lysate. The autologous primary cells were isolated from the biopsy sample using Collagenase II according to the manufacturers specifications (Ambion, USA). The cells were washed and seeded in the appropriate culture vessel at 100% confluency in DMEM/F12 medium containing 10% human A/B serum (Western Province Blood Transfusion Services, South Africa), 100 IU penicillin/streptomycin, 0.1 mM sodium pyruvate (Lonza, Switzerland), 10 µg/ml insulin, 10 µg/ml transferrin, 10 µM ethanolamine, 10 ng/ml selenium (DMEM/F12-10; Sigma–Aldrich, Germany) and 100 nM estradiol (Sigma–Aldrich, Germany). After 2 days incubation at 37 °C the medium was replaced without estradiol, but with 100 nM cortisol (Sigma–Aldrich, Germany) to prevent fibroblast growth [
22,
23]. The cells were continually cultured until 100% confluency. They were lifted with trypsin/EDTA (Lonza, Switzerland) and cultured in larger culture vessels until the cells were confluent (~ 2 × 10
7 cells in total) in a T175 tissue culture flask (Greiner, Germany). The cells were cultured in DMEM/F12-10 without cortisol one week prior to co-culture with the effector T-cells. We demonstrated that we had the ability to culture the primary breast cancer cells for several weeks. Each culture was cryopreserved in DMEM/F12 with 40% human A/B serum and 10% DMSO as indicated below.
Table 1
Demographic data of the cohorts used in study and phenotypic characterisation of the primary breast cancer cells
PC001 | 44 | African | 3 | 21×20 | 253 | Yes | No | Yes | Yes | + | +++ | +++ | +++ | A30, A68 |
PC003 | 58 | Mixed | 2 | 25×20 | 404 | No | No | No | IC | ++ | + | +++ | ++++ | A02, A30 |
PC004 | 71 | Mixed | 3 | 20×15 | 186 | No | No | Yes | Yes | +++ | ++++ | +++ | ++++ | A30, A33 |
PC007 | 58 | Mixed | 3 | 20×15 | 220 | No | No | Yes | Yes | ++ | +++ | +++ | ++++ | A03, A11 |
PC009 | 39 | Mixed | 2 | 20×15 | 345 | No | Yes | Yes | Yes | + | ++ | +++ | ++++ | A01, A03 |
PC010 | 44 | Mixed | 2 | 30×30 | 192 | No | Yes | Yes | Yes | + | + | +++ | ++++ | A02, A66 |
PC011 | 42 | Mixed | 1 | 20×20 | 116 | No | Yes | Yes | Yes | ++ | + | +++ | ++++ | A02, A24 |
PC012 | 48 | Mixed | 2 | 35×30 | 224 | No | Yes | Yes | Yes | ND | ND | ND | ND | A02, A11 |
PC013 | 41 | African | 3 | 40×45 | 165 | Yes | No | No | Yes | +++ | ++ | +++ | ++++ | A02 |
PC015 | 38 | Mixed | 2 | 11×12 | 216 | Yes | No | No | Yes | +++ | ++ | +++ | ++++ | A02, A03 |
PC016 | 44 | Mixed | 3 | 7×16 | 185 | Yes | Yes | Yes | Yes | +++ | ++ | +++ | ++++ | A02, A26 |
PC021 | 41 | Mixed | 3 | 20×16 | 320 | No | Yes | Yes | Yes | +++ | +++ | ++++ | ++++ | A02, A24 |
Preparation of tumour lysate
For the generation of a tumour lysate, the tumour tissue was homogenised on ice with a tissue ruptor (Qiagen, Germany). The homogenate was subjected to 5 freeze thaw cycles, which involved snap freezing in liquid nitrogen followed by incubation at 37 °C for 5 min. Total protein was determined using a standard Bradford assay (BioRad, USA) as per the manufacturer’s instruction.
Culture conditions to obtain mature DCs
Each patient underwent a leukapheresis procedure using the Colbe Spectra Optia® Apheresis System (Terumo BCT, USA). Following leukapheresis the monocytes (~ 2 × 107 cells) were purified by plastic adherence and differentiated into immature DCs with CellGenix DC medium (CellGenix, Germany) containing 100 µg/mL IL-4 and GM-CSF (Prospec Bio, Israel) for 5 days at 37 °C. After 5 days, immature DCs were pulsed with or without 100 µg/ml of tumour-specific lysate for 6 h at 37 °C and then matured with or without or with different combinations of 100 µg/mL Ampligen® (Hemispherx Biopharma, USA), an IFN-containing cocktail (25 ng/mL IFN-γ, 10 ng/mL IFN-α, 10 ng/mL IL1-β, 1 µg/mL CD40L; Prospec Bio, Israel) and 2.5 µg/mL R848 (InvivoGen, USA) for 42 h at 37 °C. Supernatants derived from the mature DCs were stored at − 80 °C for IL12-p70 analysis by the ELISA.
Phenotypic assessment of the mature DCs using flow cytometry
Immature and mature DCs were stained with HLA-DR PerCP/Cy5.5, CD40 FITC, CCR7 PE, CD80 PE/CY7, CD86 PE-Dazzle 594 and CD83 APC (Biolegend, USA). The cells were acquired using a LSRII flow cytometer (Beckton Dickinson, USA) and analysed using FloJo software (version 10.1; Treestar, USA). Dead cells were gated out of the scatter plots prior to analysis and negative gates were set using mean fluorescence one (MFO) controls.
Confocal microscopy
Monocytes, immature DCs and mature DCs were prepared as indicated previously. The cells were allowed to adhere to 3-aminopropyltriethoxysilane (APES; Sigma, Germany) coated slides overnight at 37 °C. The next day the cells were stained with or without or in combination with CD14 PE/Cy7, CD40 FITC and or CD83 APC (Becton Dickinson, USA) and the slides were mounted in Mowiol (Calbiochem, USA) containing n-propyl gallate (Sigma–Aldrich, Germany) as anti-fading agent. Confocal microscopy was performed with a Zeiss Axiovert 200M LSM 510 Meta NLO Confocal Microscope using the 40X water immersion objective and the 63X oil-immersion objective.
Cytospin, haematoxylin, eosin staining and light microscopy
Monocytes, immature DCs and mature DCs were concentrated onto glass slides using cytospin (Cytospin 3, Shandon, UK) and stained with haematoxylin and eosin (Merck, Germany) using a standard technique. The slides were viewed using a Nikon light microscope with the 100x oil-immersion objective.
Immunohistochemistry of the breast cancer biopsies
Immunohistochemistry of the biopsy samples using antibodies directed to the estrogen receptor (ER), progesterone receptor (PR) and human epidermal growth factor receptor (HER-2) were performed by the National Health Laboratory Services (NHLS) at Groote Schuur Hospital, Cape Town, South Africa.
Phenotypic characterisation of the autologous breast cancer cells using flow cytometry
The autologous breast cancer cells were stained with HER-2 PE, epithelial cell adhesion molecule (Ep-CAM) PE-Dazzle 594, mucin-1 (MUC-1) PE-Cy7 and integrin alpha 6 (CD49f) APC (Biolegend, USA) as recommended by the manufacturer. The cells were acquired on the LSRII flow cytometer and the data were analysed as indicated previously.
IL12-p70 ELISA
The expression of IL12-p70 was determined using a standard ELISA technique from the culture supernatants obtained above according to the manufacturer’s specifications (Mabtech, Sweden).
Generation of effector cells
Mature DCs prepared as previously described, were co-cultured with PBMCs as described by Koido et al. [
24]. Briefly, mature DCs were co-cultured with PBMCs at a ratio of 1:10 in RPMI (Lonza, Switzerland) medium supplemented with 10% human A/B serum (Western Province Blood Transfusion Services, South Africa), 2 mM L-glutamine, 25 mM HEPES, 0.1 mg/mL sodium pyruvate, 100 IU/ml penicillin and 100 mg/ml streptomycin (R-10; Sigma, Germany). After 3 days of culture the medium was replaced with fresh medium containing 10 U/ml IL-2 (Roche, Switzerland). The cells were then cultured for an additional 4 days at 37 °C to generate the effector cells.
Determination of cytotoxicity and CTL–induced cell death of autologous breast cancer cells
The autologous breast cancer cells were washed then detached with Accumax (Innovative Cell technologies, USA) as indicated by the manufacturer. The autologous breast cancer cells were then co-cultured with the effector cells (generated as indicated) at various ratios of 2:1, 5:1 and 10:1 (effector cells : autologous breast cancer cells). Autologous cells alone served as a negative control. After 4 h of incubation at 37 °C, cytotoxicity was determined using the LDH assay (Cytotoxicity Detection KitPlus LDH; Roche, Germany) and cell death was measured using 7-aminoactinomycin D (7-AAD; Becton Dickinson, USA) by flow cytometry.
Tetramer assay
The MHC-1-specific tetramers used in the current study were HLA-02 positive, therefore, only matched patient samples were analysed for the recognition of HER-2 and MUC-1 antigens by the TCRs of CD8 + T-cells. Effector cells were stained with MUC-1 PE tetramer, HER-2 APC tetramer (MBL, USA), CD8 FITC (Becton Dickinson, USA) and Zombie NIR (Biolegend, USA) as recommended by the manufacturer then acquired by flow cytometry and analysed as previously indicated.
Cryopreservation, sterility and endotoxin/mycoplasma determination
Mature DCs were cryopreserved in R-10 containing 10% DMSO (Sigma, Germany) and 40% human A/B serum at a concentration of 1 × 107/ml at − 80 °C. After 2 months’ cryopreservation, the viability was assessed using trypan blue staining and the maturation phenotype by flow cytometry.
Routine bacterial and mycological sterility testing was conducted on every batch of mature DCs by the NHLS at Groote Schuur Hospital, Cape Town, South Africa. The levels of endotoxin and mycoplasma was determined using the Endpoint Chromogenic Limulus Amoebocyte Lysate (LAL) Assay (ThermoFisher, Scientific, USA) or the MycoAlert™ detection kit (Lonza, Germany) according to the manufacturer’s specifications, respectively.
Statistics
Data were analysed for statistical significance by one-way Anova with Dunnets post-test or a Wilcoxon signed rank paired t test using GraphPad Prism software (version 6.0; GraphPad Software, USA), where *, **, ***, **** indicate p < 0.05, p < 0.01, p < 0.005, p < 0.0001, respectively.
Discussion
We have developed a Th1-polarising DC vaccine that has high efficacy against patient-derived breast cancer cells in vitro. We show that we can optimally mature DCs in vitro with autologous tumour-specific lysate and a cocktail containing cytokines and TLR agonists. The mature DCs produced high levels of the Th1 effector cytokine IL12-p70. In addition, the TCRs of the mature DC-primed CD8 + T-cells could recognise HER-2 and MUC-1 antigens using a tetramer assay. We further show that these mature DCs could prime effector cells, which resulted in cytotoxic killing of patient-specific autologous breast cancer cells in vitro. To our knowledge this is the first DC vaccine preclinical cancer study that has tested the efficacy of the vaccine against the patient’s own tumour cells in vitro. This is critical to measure vaccine efficacy as breast cancer antigen heterogeneity is high relative to that in cancer cell lines [
18,
19,
21].
A major finding was that the IL-12p70-producing mature DCs were proficient in co-stimulating CD8 + antigen-specific tumoricidal responses. This was only observed when Ampligen® and R848 were included during maturation together with tumour-specific lysate and the IFN-containing cocktail. Although the use of DCs as an adoptive cell-mediated therapy for cancer has been widely used [
28], our study differs considerably from others as we used autologous breast cancer cells as “target” cells in vitro (and to our knowledge the first to do so). The levels of toxicity reported here are comparable to other studies where the investigators utilised cell lines to test vaccine efficacy [
10,
12,
29]. However, the precise levels of cell line-specific cytotoxicity are difficult to measure because of tissue mismatch and induction of an allogenic immune response occurring in tandem, thus underestimating the incremental efficacy of our vaccine. For example the commonly used MCF-7 cell line express very low to undetectable levels of HER-2 [
21]. In contrast HER-2 is expressed in some breast cancers that present at the clinic [
18,
19]. Therefore, vaccines directed to cell lines may not truly represent the antigenic phenotype of autologous tumours. In this study, all the tumour cells expressed high levels of HER-2, which further highlights the limitations of using cell lines as a model system to test vaccine efficacy.
We showed that the DCs which were pulsed with tumour-specific lysate and matured with full cocktail expressed high levels of CCR7. The high expression levels of co-stimulatory molecules together with CCR7 expression indicate that the DCs not only have optimal T- and natural killer (NK) cell co-stimulatory capacity [
30] but also optimal lymph node homing ability [
31]. The infiltration of DCs into primary tumour lesions has been associated with significantly prolonged patient survival [
32]. A meta-analysis of clinical trials involving DC-based immunotherapy favoured administration of vaccines closest to lymph nodes [
6] as only 4–5% of the DCs reach the draining lymph nodes [
33]. CCR7 is the dominant receptor involved in the migration of DCs to the draining lymph node, and thus the upregulation of the homing cytokine, CCR7, in our study further supports the use of our DC vaccine as a candidate for therapy.
The individual components included to induce maturation in the current study were chosen to favour type-1 polarisation. Both IFN-γ and CD40L drive high levels of IL-12p70 expression [
34] and IL-12p70 and IFN-γ are important for CD8 + T-cell memory development [
27]. The TLR agonists, Ampligen® and R848 have been shown to enhance the expression of IFN-γ and IL-12p70 from DCs [
9,
35]. Interestingly, R848 induces myeloid-derived suppressor cell (MDSC) differentiation into macrophages and DCs [
36]. It is thus an attractive candidate for enhancing the effects of cancer immunotherapy as cells differentiated from MDSCs by the action of R848 exert higher proliferation-inducing activity on antigen-primed T cells compared to untreated MDSCs [
36].
We initially pulsed the immature DCs with a tumour lysate prepared from biopsies of breast cancer patients. A meta-analysis from 3444 cancer patients has shown that patients treated with tumour lysate-matured DCs have a more favourable outcome than patients treated with peptide-matured DCs [
37]. Electroporation of patient-specific tumour mRNA has been reported to be a more efficient method to enhance MHC class I-mediated antitumor immunity, which mediates a cytotoxic T-cell response without functional deterioration of the DCs [
38]. However, in our extensive optimisation studies we found that electroporation of the DCs resulted in suboptimal viability and decreased co-stimulatory molecule expression on the mature DCs (data not shown).
We show that the tumour-specific lysate and full cocktail-matured DCs produced high levels (1.2 ng/1×10
6/ml/ml) of IL12-p70. A number of human in vitro DC vaccine preclinical trials indicate that IL-12p70 expression is an important predictor of how effective a vaccine can be in an in vivo clinical setting [
12,
13] and IL-12p70 has been shown to be indispensable in regulating T-cell effector function [
39‐
42] and NK-induced antitumor responses [
43]. In addition mature DCs that produce high levels of IL-12p70 have increased antigen presentation capacity [
39] as well as an increased capacity to induce CTL responses to tumour cells [
44].
There are limitations to the current study. It was only conducted at one site, so the efficacy of the vaccine was not tested in different clinical settings. However, this was an in vitro preclinical trial and not a phase II or III clinical study. The flow cytometry cell death data may not represent a true reflection of the actual levels of cell lysis and/or death over the 4-h incubation period. The CTL assay is more representative of actual cytotoxicity levels because the assay measures cell membrane lysis over the entire incubation period, while flow cytometry would only measure whole intact dead cells. As a result, the flow cytometric assay would not measure cells that have already lysed or are in the process of lysing due to cytotoxicity. Finally, we were unable to recruit patients with stage 4 breast cancer. However, the immunomodulatory capacity of stage 3 and 4 breast cancer patients would be expected to be similar.
In conclusion, we have developed a DC vaccine to breast cancer, which had potent Th1 polarising ability that is tumoricidal to autologous breast cancer cells in vitro. This has not been reported before and the techniques and methodology used in this preclinical trial will be applied in a phase I safety study.