Short-term cultured, interleukin-15 differentiated dendritic cells have potent immunostimulatory properties

Since their discovery by Steinman and Cohn in 1973, dendritic cells (DCs) have been recognized as the strategic orchestrators of the innate and adaptive immune system [1‐3]. Although our knowledge of DC biology is still expanding, several concepts are yet well established [3, 4]. Immature DCs are known to be the vigilant sentinels of the human immune system; they relentlessly screen the environment for the presence of antigen and are highly capable of antigen uptake [4, 5]. Mature DCs are able to present the processed antigens via major histocompatibility complexes (MHC) to T cells after their migration to secondary lymphoid organs. This process of DC-mediated migration is regulated by multiple factors, but expression of the chemokine receptor CCR7 is recognized to play a pivotal role [6]. In the lymph nodes, three signals are required for the formation of an optimal immunological synapse between DCs and T cells and for the induction of desired T helper type 1 (T_h1) immune response: (1) recognition of MHC-presented antigens by T cell receptors, (2) delivery of costimulatory signals via the CD80/CD86-CD28 pathway, and (3) secretion of interleukin (IL)-12p70 by DCs after CD40/CD40 ligand signalling [5].

Since DCs are key regulators of the human immune system, their use under the form of a cellular vaccine is an attractive strategy for the treatment of cancer and infectious diseases [3]. Since the results of the first clinical DC vaccine trial were published in 1996 [7], the field of DC-based immunotherapy has been increasingly translated into clinical practice, as evidenced by the growing number of clinical studies. To date, more than 100 trials have been performed or are currently ongoing to evaluate the effect of DC vaccines in a wide variety of disease states, with a main focus on the treatment of cancer [4].

While CD34⁺ bone marrow progenitor cells and circulating blood myeloid DCs have been applied as DC precursors in some clinical studies, the vast majority of DCs used for vaccination purposes are derived from autologous peripheral blood monocytes [8]. The classic strategy for the ex vivo generation of monocyte-derived DCs consists of a two-step culture protocol, in which monocytes are differentiated towards immature DCs, followed by the induction of DC maturation. The total in vitro culture duration lasts one week, 5-6 days for DC differentiation and 1-2 days for subsequent DC maturation [5, 9, 10]. However, there is an increasing body of evidence that mature monocyte-derived DCs can be generated even after short-term cell culture for 2-3 days [9, 11‐15]. As compared to the traditional 7-day approach, rapid expansion of DCs is associated with several advantages; it simplifies the laborious and time-consuming process of DC manufacturing and it reduces the actual risk of microbial contamination related to in vitro culture [10, 15]. Moreover, short-term cultured DCs exhibit equal or superior functional DC characteristics compared to their conventional long-term counterparts [13, 14]. Previous work has already demonstrated the feasibility of short-term culture of monocyte-derived DCs differentiated in the presence of granulocyte macrophage colony-stimulating factor (GM-CSF) and IL-4 (IL-4 DCs) [12‐16].

Alternative differentiation of monocyte-derived DCs using a combination of GM-CSF and IL-15 has recently gained increasing interest. Interleukin-15 is a pleiotropic cytokine that plays a pivotal role in the generation of antigen-specific CD8⁺ T lymphocytes [17‐19], the induction of memory CD8⁺ T cell immunity [20] and natural killer (NK) cell activation [21]. Interleukin-15 differentiated DCs (IL-15 DCs) have been previously described to exhibit a distinct Langerhans cell(LC)-like phenotype and to possess unique immunostimulatory properties [22, 23]. This was more recently supported by the demonstration that IL-15 DCs are endowed with a superior capacity to induce antigen-specific cellular immune responses in an in vitro melanoma tumor model [24, 25]. These promising results make IL-15 DCs a qualified candidate for application in DC-based tumor immunotherapy [4, 26].

In addition, the Toll-like receptor (TLR) signal transduction pathway has recently emerged as an attractive alternative for the induction of DC maturation [10, 27, 28]. Toll-like receptors recognize pathogen-derived signals, such as microbial constituents (viral or bacterial-derived proteins, RNA or DNA) [29, 30]. Monocyte-derived DCs are known to express a series of TLRs, either on their cell surface (TLR2, TLR4) or intracellularly (TLR3, TLR7, TLR8 and TLR 9) [27]. Recent studies have suggested that DC maturation using TLR3 or TLR7/8 ligands in association with prostaglandin E₂ (PGE₂) results in the generation of DCs that, besides migratory properties, possess the desired capacity to produce T_h1-polarizing cytokines such as IL-12p70 [31‐33]. In the context of cancer immunotherapy, T_h1 polarization is considered conditio sine qua non for the induction of anti-tumor cytotoxic immune responses. However, despite the discovery of TLR ligands as powerful DC maturation agents, a non-TLR ligand-based maturation cocktail is currently regarded as the 'gold standard' for the induction of DC maturation in clinical trials. This widely adopted maturation cocktail was first described by Jonuleit et al. and is composed of the pro-inflammatory cytokines tumor necrosis factor (TNF)-α, IL-1β, IL-6 and PGE₂[34]. Prostaglandin E₂ is generally believed to be indispensable for potentiating the migratory potential of DCs [35, 36], but hampered IL-12p70 production is considered to be its main drawback [37‐39].

In view of the consideration that short-term DC culture, differentiation with IL-15 and TLR-induced maturation are proposed as separate attractive strategies to optimize the immunogenicity of clinical DC vaccination, we sought to determine whether an integration of these approaches is feasible and results in the generation of potent immunostimulatory DCs. We therefore examined the effect of culture duration on IL-15 DC phenotype and function by comparing short-term and long-term culture protocols. In addition, we evaluated the effect of two different maturation procedures on IL-15 DCs, juxtaposing the traditional pro-inflammatory cytokine combination with a clinical grade available maturation cocktail that includes a TLR7/8 ligand (resiquimod; R-848).

The current standard to generate DCs for use in clinical trials consists of a one-week, two-step culture protocol in which monocyte-derived DCs are first differentiated in the presence of GM-CSF and IL-4, and subsequently matured with a combination of pro-inflammatory cytokines [8, 34]. In the present study, we established a novel protocol for the generation of monocyte-derived DCs by implementing the following modifications: (1) short-term culture for 2-3 days instead of 7 days, (2) alternative differentiation in the presence of GM-CSF and IL-15 (IL-15 DCs) and (3) alternative maturation induction through engagement of the TLR7/8 signalling pathway.

The modified protocol described here proved feasible for rapidly generating stimulatory and migratory DCs without any detrimental effects on cell viability and function. Neither differentiation with IL-15 nor TLR7/8 triggering had a negative influence on cell viability (data not shown). IL-15 DCs displayed a typical DC morphology already after 2-3 days of in vitro culture. As compared to standard IL-4 DCs, however, we observed that IL-15 DCs still retained some CD14 on their cell surface which, together with the lower expression levels of CD1a and CD209 (DC-SIGN), points to a less differentiated DC phenotype. This observation seemed unrelated to the duration of IL-15 DC culture, since long-term cultured IL-15 DCs expressed even higher levels of CD14 as compared to their short-term cultured counterparts. Previous studies have shown that replacement of IL-4 by IL-15 switches the differentiation of monocytes from 'genuine' monocyte-derived DCs to cells with a complex LC-like phenotype [22, 23, 40, 41]. This finding has fuelled the interest in alternative differentiation of monocyte-derived DCs by IL-15, since LC-like DCs have been advocated as ideal cellular vaccine vehicles in view of their potent antigen-presenting capacity [4, 42]. In the present study, we confirmed the Langerin (CD207)-positivity of IL-15 DCs [22, 23] and showed that CD207 upregulation is already maximal after short-term culture.

Another intriguing phenotypic finding was that a fraction of IL-15 DCs expresses CD56, a marker with predominant expression on natural killer (NK) and NK-T cells [43]. In this regard, IL-15 DCs bear phenotypic similarity with monocyte-derived DCs generated in the presence of GM-CSF and IFN-γ (IFN-DCs). A subset of IFN-DCs was recently identified as being CD56-positive and endowed with endogenous cytotoxic activity, mediated by TNF-α-related apoptosis-inducing ligand (TRAIL) [43‐45]. It is not completely speculative to draw a parallel between these IFN-DCs and IL-15 DCs, since there is evidence that type I interferons regulate IL-15 expression, suggesting a close relationship between both cytokines. In view of these data, it might be of particular interest to further elaborate on the phenotypic and potential functional resemblance of IFN-DCs and IL-15 DCs.

While it induces full phenotypic maturation in conventional IL-4 DCs, we observed that IL-15 DCs exhibit a suboptimal phenotype upon maturation induction with the widely adopted pro-inflammatory cytokine cocktail (cc-mDC). This was exemplified by the lower expression levels of the DC maturation marker CD83 and of vital costimulatory molecules such as CD80 and CD86. These phenotypic differences indicate that the results obtained with the classic maturation cocktail in IL-4 DCs cannot necessarily be extrapolated to IL-15 DCs.

Upon TLR activation, however, IL-15 DCs undergo an efficient maturation program and reach acceptable levels of CD83, CD70, CD80 and CD86; their phenotype appears close to that of fully mature IL-4 DCs, despite a distinct expression of CD83. The functional relevance of CD70 expression on the cell surface of TLR7/8-matured IL-15 DCs should be stressed, since CD70⁺ DCs favour T_h1 immunity via the CD70-CD27 signalling pathway in an IL-12p70-independent fashion [46].

We next examined several functional endpoints to which IL-15 DCs must conform in order to be a valid immunotherapeutic vaccine candidate. Migration of DCs to secondary lymphoid organs is generally considered a conditio sine qua non for the success of DC-based immunotherapy [47]. The migratory potential of IL-15 DCs has been sparsely investigated until present, with only one prior report demonstrating their migratory responsiveness to the CCR6 ligand CCL20 [22]. However, acquisition of CCR7 upon maturation is one of the critical factors involved in effective DC migration to the draining lymph nodes [6, 48, 49].

As expected, immature DCs showed absent expression of CCR7 and correspondingly failed to migrate in the direction of CCL21 in a standard Transwell™ migration assay, hence validating our experimental set-up [48]. While the classical combination of pro-inflammatory cytokines induces a migratory phenotype in standard IL-4 DCs (this study and [36, 50]), IL-15 DCs are found to be refractory to this maturation cocktail. The low CCR7 expression and concomitant weak migratory potential of conventionally matured IL-15 DCs could be, at least in part, explained by their less mature phenotype, as reflected by the relative low expression of CD83. In contrast, TLR7/8 agonist-matured IL-15 DCs are capable of effective CCR7-mediated migration. This result is in line with recent studies, showing that the addition of PGE₂ to the maturation protocol reinstates the migratory program affected by TLR signalling [32, 33]. Our results clearly point to a superior migratory potential of short-term cultured TLR7/8-activated IL-15 DCs, which combine CCR7 expression with a migratory activity close to that of standard mature IL-4 DCs.

Besides possessing strong migratory properties, production of T_h1-polarizing and pro-inflammatory cytokines is considered to be another characteristic of immunostimulatory DCs. Dendritic cell-mediated production of IL-12p70 upon T cell encounter in the lymph nodes is regarded as a decision step in the induction of a desired T_h1 immune response [10]. Absent IL-12p70 release is a major barrier to effective immunotherapy, which could be circumvented by modifying the current in vitro DC maturation protocol [32, 51]. As mentioned previously, the combination of TNF-α, IL-1β, IL-6 and PGE₂ has been implemented as the standard maturation cocktail in most DC vaccine trials, despite its well-known drawback of hampering IL-12p70 production [38, 39]. Analogous to conventional IL-4 DCs, we observed no overt IL-12p70 release by IL-15 DCs matured with the pro-inflammatory cytokine cocktail. Conversely, TLR7/8-activated IL-15 DCs are able to produce detectable amounts of IL-12p70 after mimicking in vivo T cell encounter with CD40L-expressing fibroblasts. It should be noted that typical high IL-12p70 levels could not be attained after activation of the TLR7/8 pathway in IL-15 DCs. This may be due to an intrinsic inability of IL-15 DCs to produce IL-12, as had been previously suggested [23]. The inclusion of PGE₂ in our maturation cocktail provides another possible explanation. In contrast to recent studies [32, 33], we found that exposure to PGE₂clearly suppresses the IL-12p70 release by TLR7/8-matured IL-15 DCs (data not shown). However, the physiological relevance of this limited IL-12p70 production capacity is questionable for several reasons. First and foremost, it has been suggested that even minor amounts of IL-12p70 have a T_h1-skewing influence on the immune response [52, 53]. Thus one can hypothesize that the qualitative aspects (presence or absence of IL-12p70) are far more important than the quantitative. Within this context, it is also difficult to judge the significance of the more pronounced IL-12p70 release by long-term cultured TLR7/8-matured IL-15 DCs as opposed to their short-term counterparts. Secondly, IL-12p70 is an important but not exclusive signal for the induction of T_h1 responses. Effective cellular immune responses can occur in the absence of functional IL-12p70, as has been exemplified in the case of Langerhans cells [42]. Thirdly, we were able to show that TLR7/8-activated IL-15 DC preparations contain high amounts of IFN-γ, which are likely derived from expanded NK cells in the IL-15 DC cultures [25]. A recent study by Hardy et al. has pointed out the pivotal role of contaminating IFN-γ producing NK cells in the induction of T_h1 immunity by IL-15 DCs. As such, it might be speculated that NK cell-derived IFN-γ can partially replace IL-12p70 as a T_h1-polarizing cytokine, thereby providing another mechanism by which IL-15 DCs can induce cellular immunity in an IL-12-independent fashion [25]. Fourthly, TLR7/8-matured IL-15 DCs express CD70, which contributes to IL-12-independent T_h1 differentiation as described above [46]. Lastly, Dubsky et al. have recently elucidated that the enhanced potential of IL-15 DCs to induce cellular immune responses can be ascribed in part to membrane transpresentation of the T_h1-polarizing cytokine IL-15 [24].

Efficient antigen presentation is another prerequisite that DCs must fulfill in order to be considered for implementation in DC-based immunotherapy protocols. Previous studies convincingly showed that IL-15 DCs are highly capable of inducing antigen-specific T cell responses in both viral and tumor antigen models [22, 24, 41]. It has been put forward that IL-15 DCs have an optimal antigen-presenting capacity; a recent study by Dubsky et al. has emphasized their potent ability to prime and expand high-avidity tumor antigen-specific CD8⁺ cytotoxic T lymphocytes [24]. Our study extend these findings in several important respects. In the first place, activation of IL-15 DCs with a TLR7/8 stimulus appears to result in enhanced antigen-specific T cell responsiveness compared to maturation with a standard combination of pro-inflammatory cytokines. This observation should be interpreted together with the other effects of the two studied maturation cocktails on IL-15 DCs. On the basis of their phenotypic profile and their impaired ability for CCR7-driven migration and cytokine production, it can be hypothesized that IL-15 DCs are relatively inert to classical maturation stimuli, thereby providing an explanation for their inferior capacity to induce antigen-specific T cell responses. Since fully mature, immunostimulatory DCs are required for successful cancer immunotherapy, the clinical use of cytokine cocktail-matured IL-15 DCs cannot be recommended. Conversely, TLR7/8-matured IL-15 DCs showed a strong T cell stimulation capacity. This is particularly true for short-term cultured IL-15 DCs, which demonstrated a clear superiority over their long-term cultured counterparts and over conventional IL-4 DCs with regard to the induction of antigen-specific CD8⁺ T cell responses.

Taken together, short-term cultured and TLR7/8-matured IL-15 DCs best meet the imposed requirements to be recognized as immunostimulatory DCs. Not only do they possess an accurate phenotype as described above, they also combine strong migratory properties with some degree of IL-12p70 production and, most importantly, with a potent ability to promote antigen-specific T cell responses. Despite a more pronounced secondary production of IL-12p70, their long-term cultured counterparts behaved inferior with respect to migration and T cell stimulation capacity. In addition, reducing the time of DC culture facilitates the still arduous and costly process of ex vivo DC generation. Since short-term culture and TLR7/8-induced maturation of IL-15 DCs was considered as the "best-fit" approach to generate immunostimulatory DCs, the last part of our study was dedicated to the mRNA electroporability of this DC subset. Electroporation of mRNA is being increasingly applied in clinical vaccination trials as an elegant strategy for antigen loading of DCs [54, 55]. The attractiveness of this technique is based on the fact that it overcomes several drawbacks of other antigen delivery methods, such as the biosafety issues posed by viral gene delivery or the need for genome integration and the related risk of insertional mutagenesis associated with DNA transfection. In contrast to exogenous peptide pulsing, mRNA electroporation does not require prior knowledge of the HLA restriction characteristics of the antigen epitopes nor the need for HLA-matched donor DCs [55‐57]. Here we demonstrate for the first time the mRNA transfectability of IL-15 DCs. Short-term cultured TLR7/8-matured IL-15 DCs show accurate transgene expression after eGFP mRNA electroporation, consistent with prior studies on TLR7/8-mediated DC maturation [31, 58]. Moreover, the observation that M1 mRNA-electroporated IL-15 DCs are capable to induce influenza matrix protein M1-specific T cells further proves the feasibility and applicability of this method.

In conclusion, we propose a novel approach for the generation of DCs, based on a combined strategy of (1) short-term culture of monocyte-derived DCs, (2) differentiation in the presence of IL-15 and (3) maturation using a TLR7/8 ligand-based cocktail. This integrative approach results in the generation of DCs that meet the phenotypic and functional endpoints for implementation in clinical vaccination trials.