Background
Given the high mortality rate and recurrence risk in ovarian cancer (OC) patients, the current challenge is to develop new therapeutic strategies to prevent disease relapse and progression [
1]. One of the most promising approaches to accomplish these goals is to induce the patient’s immune system to direct a specific and durable anti-tumor response. The potential of immune surveillance in OC is supported by the observation that the presence of tumor-infiltrating lymphocytes (TILs) correlates with a favorable prognosis [
2]. TIL evaluation has consistently proven useful for surgical decision-making in OC treatment [
3]. Unfortunately, spontaneous immune reactions against tumor cells are ineffective due to multiple immunological evasion mechanisms associated with OC [
4].
Cellular immunotherapy aims to induce tumor-specific helper and cytotoxic T-cells capable of efficiently targeting and eradicating OC. This has been attempted by in vitro expansion of TIL [
5], and engineered T-cells [
6]. Nonetheless, the most efficient physiological process for T-cell priming requires full dendritic cell (DC) activation and antigen presentation. One of the major obstacles to the development of effective DC-based immunotherapy in OC is the circumvention of tumor-associated immunosuppressive mechanisms, the most notable being the accumulation of tumor-infiltrating Treg cells, which have been associated with increased mortality in patients [
7]. Thus, it is necessary to “program” DC towards activation of effector T-cell polarizing profiles and avoidance of Treg expansion.
Regulation of the p38 and ERK signal transduction pathways in DC plays a central role in determining the expression of certain cytokines by DC. As an example, inhibition of MEK 1/2 and MAPK promotes IL-12 production and Th1-polarizing responses, whereas inhibition of p38 MAPK blocks IL-12 production [
8]. Because IL-12 facilitates the differentiation of type 1 T-helper cells, the inhibition of IL-12 secretion in DC is expected to negatively affect DC-driven anti-tumor T-cell responses [
9]. Interestingly, p38 inhibition promotes differentiation and survival of monocyte-derived DC [
10], and p38 inhibition or ERK activation restores deficiencies in DC function in myeloma patients [
11], suggesting that treatment of DCs with pharmacological p38 inhibitors may be therapeutically useful. Furthermore, blockade of the p38 pathway can attenuate regulatory T cell induction by DC [
12], whereas blockade of the ERK pathway suppresses DC-driven Th17 responses [
13], suggesting that p38 blockade (which enhances ERK phosphorylation) may favor a switch from Treg induction to Th17 differentiation. This could be of critical importance for DC vaccination against OC since Treg expansion and infiltration is known to correlate with increased morbidity and mortality, [
7] whereas Th17 infiltration is strongly associated with prolonged patient survival [
14].
Cancer/testis antigens (CTA) are a novel class of tumor-associated antigens displaying potent immunogenicity and high expression levels in tumor cells but negligible expression in normal tissues. We have successfully exploited the cancer/testis antigen mSP17 for adoptive immunotherapy in OC [
15] and as a biomarker to track OC progression in animal models [
16]. The suitability of mSP17 as a target for immunotherapy is supported by the finding that it is expressed by primary and metastatic OC lesions in up to 70% of patients [
17]. Further, the relevance of mSP17 expression in OC has been confirmed by the demonstration that it is correlated with malignancy, chemo-resistance and tumor cell motility [
17]. We have also shown that recombinant adeno-associated virus vectors (rAAV) can successfully deliver tumor antigens in DC and are superior to protein loading techniques for MHC class I-restricted cellular mediated immunotherapy in multiple myeloma [
18], cervical [
19] and ovarian cancers [
20].
In this study, we combine pharmacological p38 MAPK inhibition with rAAV-based mSP17 antigen expression for autologous DC vaccination in the ID8 mouse model of epithelial OC, a system that faithfully reproduces the peritoneal dissemination seen advanced human OC [
21]. We present evidence that mSP17-targeted DC vaccination, following inhibition of p38 MAPK, results in a dramatic enhancement of the therapeutic efficacy of DC vaccination. Finally, we validated our results by demonstrating that p38 blockade in the presence of rAAV Sp17 transduction is superior to rAAV Sp17 transduction alone, in terms of activation of human DC and their ability to generate effector lymphocytes.
Methods
Construction of rAAV-mSP17 virus
The AAV-mSP17 genome was constructed as a plasmid as previously described [
18,
20]. The mSP17 cDNA was inserted into the rAAV vector, dl6-95 [
22]. The mSP17 gene was expressed from the rAAV p5 promoter, which is well described in the literature to be active in DC [
22]. We evaluated their ability to infect HEK293 cells using our rAAV-mSP17 construct. The human rAAV-SP17 vector was produced following the same scheme described for mSP17. Human full length SP17 coding sequence was obtained by gene synthesis from GenScript.
Titration of virus stocks
DNA was extracted from virus crude lysates, and the titer of virus stocks was determined by real-time PCR. Briefly, we used serial dilutions of the corresponding rAAV vector for construction of a standard curve. The real-time PCR was performed on an ABI Prism 7000 instrument (Applied Biosystems, Darmstadt, Germany) in a 50 μl reaction volume.
Generation of dendritic cells (DC)
Murine DC were generated from splenocytes. We infected adherent monocytes with rAAV as previously described [
18,
19,
23]. 10 µM p38 MAPK inhibitor (catalog #ML3403, EMD Chemicals, La Jolla, CA) was added at days 0, 3 and 5. Human mature DC were differentiated from PBMCs of five ovarian cancer patients [
24,
25]. rAAV-transduced human DC were differentiated from rAAV-infected monocytes as we have previously described [
18,
19]. 10
7 rAAV-SP17 encapsidated genomes (eg) were used to transduce adherent monocytes in 6-well plates [
26]. To block p38, we used the same schedule and concentration detailed above for murine cells.
Mice and ID8 cell line
Six-week-old female C57BL/6 mice were obtained from the Jackson Laboratory (Bar Harbor, ME, USA). Approval for the study was obtained from the local Institutional Animal Care & Use Committee. The ID8 cell line was kindly provided by Dr. Roby (University of Kansas). Cells were cultured in RPMI 1640 medium supplemented with 10% fetal bovine serum in 5% CO2 at 37 ℃.
Immunization and tumor challenge
Female C57BL/6 mice (6 weeks old) were challenged i.p. with 1 × 10
6 ID8 cells [
1]. 30 days after tumor challenge, mice were i.m. injected once a month for 10 months with rAAV or rAAV-mSP17 transduced DC treated or not with p38 MAPK inhibitor. Each mouse was injected with 10
6 DC.
ELISpot assay
Cytokine expression by splenocytes challenged in vitro with 1 ug/mL recombinant SP17 protein or irrelevant protein (BSA) was evaluated using an ELISPOT assay (U-CyTech, Utrecht, Netherlands), according to the manufacturer’s directions. Positive control for cellular activation was Con-A (5 µg/mL), and background wells contained RPMI 1640 medium only. Spot counts were performed with the AID ELISPOT Reader System (Cell Technology, Inc., MD). Results represent the average number of spots in each condition obtained with SP17-challenged splenocytes negative the number of spots obtained with BSA-challenged solenocytes (no spots were detected in the background control).
Cytotoxicity assay
We performed a EUROPIUM-based cytotoxicity assay using the DELFIA
® EuTDA system according to the manufacturer’s instructions [
27] with DC-co-cultured splenocytes. To assay the memory response, we performed an identical assay using unprimed splenocytes taken from mice at the time of necropsy.
Migration assay
The bottom chambers of polycarbonate Corning® Transwell™ Permeable Supports (5 μm pore size, Cole-Parmer, Vernon Hills, Illinois) were coated with ID8 cells. 200,000 splenocytes were added in the upper chamber in complete medium. After 4 h, the density of migrated cells in the bottom wells was determined. The assay was performed in triplicate and mean ± SEM were determined.
Flow cytometry
Flow-cytometric analyses were performed with a Beckman-Coulter FC500 flow cytometer and CXP Software. Cells were washed twice in PBS, incubated with 10% autologous serum in PBS for 30 min on ice to block Fc receptors, then incubated with phycoerythrin-conjugated anti-CD3 PE-CF-594 antibody (BD Biosciences) for 1 h on ice in staining buffer (0.5% BSA in PBS). After washing twice in 0.5 mL staining buffer, cells were fixed with 4% buffered paraformaldehyde at 4 C for 30 min in the dark. Then, cells were washed twice with staining buffer and permeabilized with 0.5% saponin in staining buffer (permeabilization buffer) for 10 min on ice, washed once with 0.5 mL permeabilization buffer and resuspended in 50 μL buffer/200,000 cells. Anti-Foxp3-FITC anti-IFNγ Alexafluor 647 (Bio Legend), anti-TNFα Alexafluor 700 (BD Biosciences), were added and allowed to incubate for 1 h on ice. Cells were washed twice with permeabilization buffer (200 µL) and resuspended in 400 µL staining buffer supplemented with 1% W/V paraformaldehyde and stored at 4 C in the dark until data acquisition.
Preparation of lymphocytes
Briefly, the spleen was meshed with a sterile filter, and then centrifuged at 800×g for 5 min at 4 °C. Splenocytes were then isolated by Lympholyte-M (Cedarlane Ltd, Burlington, NC). Human lymphocytes were recovered from the non-adherent fraction of purified PBMCs after 3 h incubation in 6-well plates.
Generation of human DC-primed lymphocytes
For experiments with human cells, non-adherent PBMCs from the same subject were washed in PBS and resuspended in serum-free DC-medium (CellGenix) at 10
6 cells/well in 6-well culture plates, with autologous DC (20:1 PBMCs to DC ratio) [
19,
24]. Cultures were supplemented with 800 U/mL GM-CSF and 10 U/mL IL-2 for 7 days prior to analysis.
Statistical analyses
Tumor growth, cytotoxicity assays, ELISPOT, migration assays and flow-cytometry were analyzed by a two-tailed, paired Student’s test and survival rates were analyzed by the log-rank test.
Discussion
OC has been referred to as a “silent killer”, since the lack of specific symptoms prevents detecting the disease in early, ovary-confined stage with favorable prognosis, when nearly 90% of patients could be saved [
23]. Therefore, OC remains a lethal disease [
28]. Cellular immunotherapy, based on adoptive T-cell or DC transfer, has the potential to provide long-term protection and prevent metastatic dissemination of the tumor [
6,
22,
29‐
31]. In this report, we describe the efficacy of an innovative therapeutic vaccine based on DC treated with a p38 MAP-kinase inhibitor, and transduced with an mSp17 expression rAAV vector.
We and others have reported SP17 as a useful tumor cell biomarker and ideal target for immunotherapy in a variety of malignancies, including multiple myeloma, nervous system tumors, esophageal cancer and OC [
15,
16,
21‐
23,
32‐
35]. We have shown that the widely used ID8 cell line model of OC expresses significant levels of SP17 protein, both at the cell surface and cytoplasm [
1], resembling the majority of primary and metastatic tumor samples obtained from epithelial OC patients [
15,
17] and supporting the rationale for SP17-targeted immunotherapy. Further characterization of ID8 cells through flow cytometry revealed that 95% of cells were MHC class I-positive, [
1] a relevant finding since antigen presentation within MHC class I complexes is a required step for cytotoxic T-cell target recognition.
Here we tested two different vaccine strategies based on autologous DC transduced with rAAV-mSP17, with or without exposure to a p38 MAPK inhibitor (rAAV-mSP17 DC ± p38i), and compared those with two controls, i.e. rAAV-transduced DC (without mSP17 coding sequence) or absence of vaccination. Our results clearly demonstrate benefits of this innovative DC vaccination. rAAV-mSP17 DC and rAAV-mSP17 DC + p38i vaccines resulted in increased survival rates, but long-term protection was achievable only when p38 MAPK signaling within the DC was blocked. Tumor mass and ascites production were markedly diminished, and survival was extended following rAAV-mSP17 DC administration compared with rAAV DC, but the best protection was provided by the rAAV-mSP17 DC + p38i vaccine, which prolonged survival for at least 300 days. We have previously shown that rAAV-based antigen transduction of DC could induce potent MHC class I-restricted in vitro immune responses against OC and multiple myeloma, overcoming immune tolerance [
18,
20]. Here, we extended our study to show that rAAV-transduced DC vaccination provides excellent therapeutic results in vivo. The efficacy of DC vaccination was clearly enhanced through p38 MAPK inhibition, allowing exceptional levels of protection against tumor growth in syngeneic recipients. These observations are highly significant, since long-term protection against OC progression is critical to control fatal disease relapse [
36].
We showed that rAAV-mSP17 DC + p38i vaccination significantly increased the frequency of IFN-γ and TNF-α-producing T-cells compared to rAAV-mSP17 DC or AAV DC vaccines. This is an important finding since IFN-γ and TNF-α have been shown to provide protective effects in OC patients [
37‐
39]. These results suggest that p38 MAPK inhibition directs DC differentiation towards a Th1-polarizing profile [
40].
Our observations are in accordance with the superior anti-tumor cytotoxic activity displayed by the splenocytes of rAAV-mSP17 + p38 inhibitor treated mice, compared to all of the other groups. Of note, the cytotoxic response was achieved without re-stimulation of splenocytes with autologous dendritic cells in vitro, indicating that our vaccine strategy might have induced a memory response, a hypothesis that is supported by the long (300 days) persistency of the survival advantage seen in the rAAV-mSP17 + p38 inhibitor group.
It is self-evident that stimulation of strong anti-tumor effector T cell responses is of little benefit to the host unless the T cells have the capacity to migrate into the tumor microenvironment. Since lymphocytic tumor infiltration is considered a positive prognostic factor in OC patients, the observation that rAAV-mSP17 DC + p38i vaccine resulted in an increased ability of the host splenocytes to migrate towards ID8 OC cells highlights the potential clinical relevance of our results.
Since T-reg infiltration is associated with increased mortality in OC patients [
7], DC vaccination strategies that diminish T-reg activation and expansion and increase Th1 and CTL responses may result in strong anti-tumor immune reactivity and increased survival in OC patients [
14,
41]. Recent studies have pointed to p38 MAPK in DC-driven induction of T-reg responses. For example, Jarnicki and colleagues have shown that Toll-like receptor (TLR) activation promotes induction of T-reg through DC production of IL-10 dependent on p38 MAPK signaling [
12]. In addition, clinical studies in myeloma have shown that inhibition of p38 MAPK can correct defects in DC function, restoring their ability to activate alloreactive and tumor antigen-specific T cells [
11].
Accordingly to our previous report [
25], p38 inhibition in differentiated DC affords for a dramatic down-regulation of the T-cell inhibitory signaling molecule, B7-H1. Indeed, we found that p38 inhibition did not significantly alter CD80/83/86 expression levels in autologous human rAAV-Sp17 DC but reduced that of B7-H1. Consistently, when analyzing DC ability to induce activated T cells versus T-reg cells in vitro, we found that non-adherent PBMCs co-cultured with DC differentiated in the presence of p38i displayed both a reduced expression of the T-reg marker Foxp3 and up-regulation of the T cell activation marker, IFNγ. While it is known that rAAV transduction offers advantages over standard DC antigen pulsing methods, in terms of up-regulation of co-stimulatory molecules and consequently of T-cell activation [
19,
20] due to the known pro-inflammatory mechanisms triggered by AAV entry [
42], the molecular mechanism explaining the different outcome of p38 blockade on B7-H1 and CD80/83/86 remains to be elucidated. Indeed, we have previously demonstrated that p38 inhibition alone, without rAAV, results in a down-regulation of CD80/83/86 expression along with B7-H1 in DC [
25]. Our hypothesis is that CD80/83/86 expression in p38i-treated DC in the presence of rAAV is rescued by the virus ability to activate the NF-κB pathway [
43], which has been proven to increase the levels of co-stimulatory molecules in DC [
44].
Publisher's Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.