The Inside Story of Dendritic Cell Immunotherapy

Dendritic cells (DCs), initially discovered by Ralph Steinman in 1973, are broadly divided in mice and humans into conventional (cDCs) and plasmacytoid (pDCs) DCs. Tissue-resident DCs loaded with antigen migrate to lymph nodes where they can present antigen to T cells while free antigen can be picked up and presented by lymph node-resident DCs. Depending on the DC subset, maturation status and cytokine milieu, DCs may activate T cells to eliminate tumors and pathogens, or tolerize T cells, or generate Tregs. These properties of DCs have been harnessed therapeutically for anti-tumor immunity or induction or restoration of tolerance (1). Initial approaches targeted endogenous DCs randomly following injection of peptides, DNA or viral vectors. More recently, antigens or siRNAs have been directed to DC subsets in vivo via coupling to antibodies against surface molecules, such as CD205. However, these methods rely on endogenous DC subsets being present and functional and such assumption may not be fulfilled in all clinical situations. Therefore, another approach is to bypass resident DCs altogether generating in vitro the exact type of cell desired, by differentiating DCs from stem cells or monocytes, or sorting mature DCs, and exposing them to cocktails of factors or siRNAs to stabilize their maturation or fix a protolerogenic state. In transplantation, donor-specific immunosuppression has been pursued using donor DCs infused a few days before transplantation, clinically applicable for living-related transplants, or with recipient DCs pulsed with alloantigen, perhaps more widely translatable. Morelli and colleagues had previously shown that injected DCs die rapidly transferring their antigens to host DCs, but the relative requirement of donor versus host DCs for immunosuppression was not known.

In the current paper, Wang and colleagues demonstrate convincingly that prolongation of cardiac allograft survival by donor immunosuppressive (IS) DCs is completely dependent on host DCs (2). The authors utilize an elegant system of bone marrow chimeric mice in which cDCs can be deleted transiently following administration of diphtheria toxin (DT). They show that the abortive proliferation of indirect CD4⁺ and direct CD8⁺ T cells induced by donor ISDCs in the presence or absence of rapamycin, or by alloantigen-pulsed recipient ISDCs, absolutely depends on the presence of host cDCs. Surprisingly, endogenous DCs were not only required for inhibition of indirect alloreactivity by CD4⁺ T cells, as expected after donor antigen is phagocytosed and cross-presented, but also for reduction of direct alloreactivity by CD8⁺ T cells. This suggests that help from indirect CD4⁺ T cells is required for optimal function of direct CD8⁺ T cells, aborted if indirect CD4⁺ T cells are tolerized by recipient cDCs. Alternatively, intact MHC/peptide from the infused DCs may be transferred to host APCs for tolerization of CD8⁺ T cells by the semi-direct pathway. In either case, the requirement of endogenous DCs for T cell inhibition is consistent with prior observations that constitutive ablation of all CD11c⁺ DCs results in fatal autoimmunity under steady state conditions (3). The importance of host cDCs but not pDCs for immunosuppression in the model by Wang et al, as pDCs are not depleted following DT injection in this model, contrasts with the tolerogenic role ascribed to pDCs in transplant models of costimulation blockade. The role of each DC subset may depend on which cell is better at picking up antigen from different donor cells and/or which is preferentially targeted by costimulation blockade therapies. The data are consistent with another model of tolerance in which antigen-coupled, ethylene carbodiimide (ECDI)-fixed splenocytes that lack MHC class I and II can suppress autoimmune responses indicating a role for recipient APCs. This immunosuppressive effect was more potent when infused cells expressed MHC, suggesting that direct or semi-direct antigen presentation may play a partial role (4). Although the requirement for host DCs appears unequivocal from this work, it is important to keep in mind that the properties of the cells infused may still dictate the phenotypic outcome of the immune response. For instance phagocytosis by host DCs of apoptotic cells that are infected may trigger Th17 differentiation rather than tolerance (5).

The implication of these results for future pursuit of DC immunotherapies is profound, as obtaining the best DCs in vitro may not yield the desired outcome after infusion if endogenous DCs are defective. This is especially important as clinical trials are ongoing to assess the safety and tolerogenic efficacy of autologous dendritic cells in diabetes and rheumatoid arthritis (clinicaltrials.gov). Ex vivo testing of DCs generated for immunotherapy, be it for their ability to stimulate T cells, resistance to maturation stimuli, or morphological or functional phenotype may not predict their effects in vivo, although it remains possible that other routes or regimens of administration may empower a more direct role of infused DCs. Rather, the phenotype and function of endogenous DCs will have to be studied in different disease states as well as in response to medications, as, for instance, corticosteroids and Campath-1 can deplete peripheral pDCs and rapamycin and calcineurin inhibitors can alter the ratio of pDC/cDC.

Abbreviations

DC

dendritic cell

pDC

plasmacytoid DC

cDC

conventional DC

DT

diphtheria toxin

ECDI

ethylenecarbodiimide