Making Dendritic Cells that Turn Immune Responses Off

R eplacement therapies for inherited protein deficiencies such as hemophilia or lysosomal storage disorders are plagued with antibody-mediated rejection of the therapeutic protein. This serious problem is well described in protein/enzyme replacement therapies and is likely to be a hurdle for gene therapies as well, especially when performed in subjects with severe mutations such as gene deletions, which are more likely to result in a lack of tolerance to the functional protein (Cao et al., 2009). Perhaps the most extensive clinical experience with confronting this problem comes from treatment of the X-linked bleeding disorder hemophilia A, caused by mutations in the gene for coagulation factor VIII (F.VIII). Approximately 30% of patients with severe hemophilia A form neutralizing antibodies (“inhibitors”) during factor replacement therapy (which is based on intravenous infusion of plasma-derived or recombinant F.VIII). Above a certain titer, these antibodies render therapy ineffective, necessitating use of more complicated and even more expensive bypass reagents and increasing risks of morbidity and mortality. With a success rate of 60–80%, frequent high-dose F.VIII infusions eradicate the inhibitor over the course of months, and sometimes 1 to 2 years (Di Michele, 2011). No prophylactic protocols for prevention of the inhibitor response have been explored yet. In contrast, immune tolerance regimens are increasingly being performed early or even before the onset of enzyme replacement therapy (ERT) for the deadly infantile form of the lysosomal storage disorder Pompe disease, caused by mutations in acid α-glucosidase (Byrne et al., 2011). It is hoped that prevention of antibody formation increases the effectiveness of ERT and eliminates immunotoxicities. However, these protocols thus far rely on general immune suppression. Clearly, more refined antigen-specific approaches are desirable. Indeed, multiple avenues toward novel immune tolerance protocols for treatment of genetic diseases are being explored in preclinical studies. These include introduction of the antigen along with blockers of costimulation, transient blockade of T cell function, in vivo expansion of regulatory CD4⁺ T cells (Tregs) with interleukin (IL)-2/IL-2 receptor complexes, deletion of effector T cells coupled with expansion of Tregs using the mTOR inhibitor rapamycin, mucosal antigen administration (such as oral tolerance), and transient manipulation of B cell function (Verma et al., 2010; Liu et al., 2011; Miao, 2011; Moghimi et al., 2011; Zhang et al., 2011; Sack et al., 2012). Furthermore, certain routes of gene transfer, such as ex vivo into hematopoietic stem cells or in vivo into hepatocytes, can induce antigen-specific tolerance (Cerullo et al., 2007; Breous et al., 2009; Cooper et al., 2009; LoDuca et al., 2009; Goudy et al., 2011; Zakas et al., 2011).

An alternative avenue toward immune tolerance is to harness the power of professional antigen-presenting cells (APCs) such as dendritic cells (DCs), whose discovery by the late Dr. Ralph Steinman was honored with the 2011 Nobel Prize in Medicine (Steinman and Cohn, 1973). Among other functions, DCs survey tissues to detect pathogens, take up and process pathogen-derived antigens, and initiate primary immune responses by presentation to T cells along with costimulation. Consequently, DC-based anticancer therapies have been developed and are being extensively tested (Boudreau et al., 2011; Palucka and Banchereau, 2012). However, as with many immune cells, DCs come in different subsets and have multiple functions. For example, plasmacytoid DCs produce large amounts of type I interferon (IFN) in antiviral responses but have also been shown to play an important role in immune tolerance. Building on the concept that the ability of DCs to effectively present antigen to T cells can be exploited to induce immune tolerance, methods have emerged for generation of tolerogenic DCs (tDCs, which express only low levels of costimulatory molecules and lack inflammatory cytokine expression). A typical protocol is based on ex vivo generation of DCs from bone marrow precursors, using a modified cytokine cocktail that includes IL-10 and/or an immune suppressive neuropeptide (vasoactive intestinal peptide, VIP). Tolerogenic DCs can also be generated by gene transfer, such as by expression of IL-4 or other immunosuppressive molecules. Once transplanted, such tDCs can capture antigen, suppress inflammation, and present antigen in a tolerogenic fashion at distant sites via release of exosomes (Kim et al., 2007). Tolerance induction is in part mediated by induction of Tregs with immune-suppressive properties. Use of tDCs has been extensively studied in various models of autoimmune diseases. However, the literature becomes much more scarce for tolerance induction in genetic disease.

Conceptually, one would generate autologous tDCs ex vivo and pulse with the same therapeutic protein used for the replacement therapy or introduce a vector that expresses that particular antigen. For example, success with suppression of inhibitor formation against F.VIII has been reported after gene transfer with a foamy virus vector into tDCs generated with IL-10 and VIP (Su et al., 2011). A potential problem with this approach is the risk of insertional mutagenesis when using a nonspecifically integrating vector. In this issue of Human Gene Therapy, Sule and colleagues describe the beginnings of an elegant solution (Sule et al., 2012). The authors optimized the potency of tDCs pulsed with protein antigen by using a combination of IL-10 and transforming growth factor (TGF)-β₁ in their DC differentiation cocktail. Although IL-10 was sufficient to generate tDCs capable of suppressing antibody formation to the model antigen ovalbumin in T cell receptor transgenic mice, the combination of the two cytokines was required to control the antibody response to F.VIII in hemophilia A mice (knockout for F.VIII). These animals form extraordinarily robust antibody titers to human F.VIII and therefore represent a stringent test for immune tolerance protocols that aim to prevent B cell responses. In the various scenarios using ovalbumin or F.VIII protein administrations, the study demonstrates a 4- to 10-fold reduction in antibody titers, using two rounds of tDC transfer to either prevent or reverse antibody formation. Importantly, the animals responded normally to a different protein antigen, indicating that the induced unresponsiveness was in fact specific to the antigen with which the tDCs had been pulsed. Future studies need to show how long-lived the suppression of antibody formation is (or whether tDC transplantation may have to be periodically repeated) and what the effect on the functional antibody titer is for F.VIII (the so-called Bethesda titer).

A less clear aspect of the study is the mechanism by which tDCs mediated tolerance induction, which is critical to moving this area of research forward. Some evidence for a reduction of effector CD4⁺ T cell numbers and activation (which normally provide T cell help for B cells) was shown. There was no evidence for a CD4⁺CD25⁺FoxP3⁺ Treg response, which had been implicated in the foamy virus study by adoptive T cell transfer (Su et al., 2011). It is possible that other subsets of regulatory cells such as IL-10-producing type 1 regulatory T (Tr1) cells were induced, or that alternative techniques will still reveal a role for CD4⁺CD25⁺FoxP3⁺ Tregs, which had been found crucial for tolerance induction to coagulation factors in many protocols. Regardless, Sule and colleagues demonstrate a synergistic effect of the cytokines IL-10 and TGF-β on the generation and potency of tDCs, which represents a potentially superior alternative to prior methods (Sule et al., 2012). These two key cytokines repeatedly appear in the literature as mediators of immune suppression. For example, TGF-β is required for peripheral induction of CD4⁺CD25⁺FoxP3⁺ Tregs from naive CD4⁺ T cells (which occurs in the absence of IL-6 and is enhanced by retinoic acid). Tr1 cells and, depending on the tissue context, CD4⁺CD25⁺FoxP3⁺ Tregs, exert immune suppression via IL-10 secretion. Tolerogenic B cells, alternative tolerogenic APCs generated by retroviral gene transfer to primary B cells, have been described to induce CD4⁺CD25⁺FoxP3⁺ Tregs and Tr1 cells (Ahangarani et al., 2009; Skupsky et al., 2010). TGF-β (also expressed by regulatory helper T type 3 [Th3] cells) and IL-10 have been implicated as critical components of immune regulation in the gut (including in oral tolerance) as well as for prevention of deleterious CD8⁺ T cell responses in hepatic gene transfer (Breous et al., 2009; Hoffman et al., 2011). Transferred tDCs may themselves express these cytokines or induce their expression in T cells. For example, it is known that maturation of tDCs in the presence of IL-10 induces these tDCs to produce IL-10 themselves. Furthermore, one could speculate that Tregs and other immune-regulatory cells may create an environment in which endogenous DCs acquire the phenotype of tDCs in vivo on exposure to IL-10 and TGF-β, thereby amplifying the tolerance effect.

The experiments described in this issue of Human Gene Therapy by Sule and colleagues provide a concrete stepping stone for future studies. The paper also offers a tantalizing glimpse into what might lay ahead for controlling immune responses in the treatment of genetic diseases. It has long been hoped that the body's own immune system could be harnessed to treat illnesses such as cancer by directing the immune system to attack and destroy tumors. Now this study, and others like it, point to harnessing the potential power of the suppressive arm of the immune system to target responses to specific antigens. By producing tDCs in an antigen-specific manner, we may one day create an off-switch for unwanted immune responses, which would greatly enhance therapy for a variety of currently difficult to treat genetic diseases. Existing clinical experience with DC therapies for cancer should facilitate the production of clinical-grade tDCs and translation of the approach.