Differential response of BDCA-1⁺ and BDCA-3⁺myeloid dendritic cells to respiratory syncytial virus infection

Respiratory Syncytial Virus (RSV), a ssRNA pneumovirus of the paramyxoviridae family, is the leading cause of acute lower respiratory tract infections in children, elderly, and immunocompromised patients. More than 95% of children become infected with RSV by 2 years of age, and approximately 50% of children infected in the first year of life are reinfected during the second year [1]. Reinfection continues throughout life, re-emerging as a major cause of community-acquired pneumonia in the elderly [2]. RSV infection is associated with short- and long-term morbidity including pneumonia, recurrent wheezing, and abnormal pulmonary function. Moreover, RSV bronchiolitis has been identified as the highest independent risk factor for the development of asthma and allergic sensitization later in life [3].

Primary RSV infection is controlled by T cell responses which are crucial in viral clearance and limiting disease severity. Weak antiviral cytotoxic responses and decreased numbers of CD4⁺ and CD8⁺ T cells in the lower respiratory tract of infants with severe infection [4, 5] along with the susceptibility to repeat infections with RSV strongly suggest that the failure to develop adaptive T cell responses is critical in the pathophysiology of RSV-mediated disease. Additionally, the balance between Th1 and Th2 cytokines produced by T cells plays a crucial role in determining the outcome of infection, and severe RSV infection may be associated with skewing the Th1/Th2 balance of the virus-specific response towards Th2 [6, 7]. However, the mechanisms by which RSV is able to manipulate host immunity are not fully understood.

Dendritic cells (DCs) are an important first line of defense against invading pathogens and are regarded as the most potent antigen presenting cell (APC) type. Upon antigen recognition, DCs produce cytokines and costimulatory signals needed to guide T cell differentiation, ultimately determining the quality and quantity of the resulting immune response. Human DC populations are defined based on their lineage and expression of unique Blood Dendritic Cell Antigens (BDCA), and depending on their derivation, direct immune responses through either the initiation or modulation of immunity. Plasmacytoid DCs (pDCs), characterized by expression of BDCA-2, mediate antiviral immunity via the production of IFN-α [8]. Myeloid DCs (mDCs) are particularly efficient in the uptake, processing and presentation of antigens and can be further subdivided into distinct subsets identified by differential expression of either BDCA-1 (BDCA-1⁺) or BDCA-3 (BDCA-3⁺).

The lung contains an elaborate network of dendritic cells that regulate the local environment. During respiratory infections, both tissue-resident DCs, as well as recruited DCs, are activated as part of the host immune response [9]. Although present at much lower frequencies than BDCA-1⁺ mDCs in peripheral blood, the DC network in the airway mucosa is comprised predominately of BDCA-3⁺ mDCs [10]. Studies of mDC subsets isolated from human lungs and blood suggest that each subset exhibits distinct roles in coordinating the immune response against pathogens [11‐14]. Although the relationship between mDCs found in the lung and peripheral blood is not clear, parallel phenotypic analysis and transcriptome mapping provides evidence that lung and other non-lymphoid tissue BDCA-3⁺ mDCs are potentially related to blood BDCA-3⁺ mDCs [15]. The relationship of blood to tissue BDCA-1⁺ mDCs was not as clearly defined, but did show clustering of their gene signatures [15]. Thus, even though it is likely that lung DCs may in some ways be functionally distinct from blood DCs, these studies demonstrate that blood mDC subsets are potentially related to mDC subsets found in tissues.

As key regulators of immunity, DCs are an ideal target for the virus to exert its immune altering mechanisms. In this regard, a possible role for mDCs in RSV infection is evidenced by observations of increased numbers of mDCs in the airways, and decreased numbers of mDCs in the blood of children with acute RSV infection, suggesting that DCs are recruited out of the blood to the site of infection [16]. Correlations between the number of mDCs and the level of pro-inflammatory cytokines in the nasal cavity of RSV-infected patients indicates that mDCs may represent an important source of inflammatory cytokines during the host immune response against infection [16]. Additionally, RSV infection inhibits the capacity of monocyte derived DCs (Mo-DCs) to stimulate T-cell proliferation in vitro[17, 18] and impairs the function of pulmonary DCs in RSV-infected mice [19]. Taken together, these results suggest that infection with RSV impairs the ability of host immune response to adequately clear viral pathogens by altering DC mediated immune responses.

However, current RSV-mDC studies often utilize murine DCs or mDCs derived in vitro from human monocytes, and even though Mo-DCs have many characteristics similar to myeloid blood DCs [20‐22], studies have not shown direct functional correlations between in vitro derived Mo-DCs and individual mDC subsets isolated from lymph nodes or blood [14, 23, 24]. Thus, studies using Mo-DCs may not adequately recapitulate the function of primary mDCs during RSV infection. Few studies have examined the interactions between viral pathogens and primary human mDC subsets, especially BDCA-3⁺ mDCs. To identify cell-specific responses of primary mDC subsets to RSV infection, we examined the expression of costimulatory markers and production of cytokines by BDCA-1⁺ and BDCA-3⁺ mDCs after exposure to RSV. The functional response after infection was evaluated by examining their ability to stimulate T cell proliferation in a mixed lymphocyte reaction. To our knowledge, this is one of the first studies to examine the functional response of highly purified BDCA-1⁺ and BDCA-3⁺ mDCs to RSV.

In this study, we examined the effect of RSV infection on primary human mDC subsets isolated from peripheral blood. To ensure purity of the specific mDC populations among donors, BDCA-1⁺ and BDCA-3⁺ mDCs were isolated using multicolor FACs sorting with parameters similar to those previously published [13]. This method allowed us to consistently obtain sufficient numbers of both populations with purity ≥95% for all donors. Similar to Mo-DCs [17, 18], we demonstrate that primary mDC subsets are susceptible to infection with RSV. Consistent with previous reports [25], BDCA-1⁺ and BDCA-3⁺ mDCs show similar rates of RSV infection. While mDC activation is not dependent on viral replication, we demonstrate that RSV infection induces a distinct pattern of costimulatory molecule expression and cytokine production by BDCA-1⁺ and BDCA-3⁺ mDCs, and impairs their ability to stimulate T cell proliferation.

Enhanced expression of CD86, CD80, and PD-L1 on RSV-infected mDCs coincides with the costimulatory molecule expression on RSV-infected Mo-DCs and murine lung mDCs which have an impaired capacity to stimulate T cell proliferation [18, 19]. These findings suggest that the expression of inhibitory molecules by infected mDCs may contribute to their decreased capacity to stimulate T cell activation. As CD80 is a highly effective ligand for CTLA-4 on activated T cells [37], upregulation of CD80 on BDCA-1⁺ and BDCA-3⁺ mDCs provides a possible mechanism by which RSV could inhibit T cell proliferation. CD86 seems to be more important in CD28 interactions, and in Tregs, CD28 activation provides critical signals needed for proliferation and survival [38, 39], thus providing an additional mechanism for the suppression of reactive T cell responses. Our findings of significant levels of CD86 induction by RSV-infected BDCA-3⁺ mDCs may then indicate a differential role for BDCA-1⁺ and BDCA-3⁺ mDCs in Treg homeostasis during RSV infection. However, it is likely the relative expression levels of CD80 and CD86 on mDCs dictates the balance between stimulatory and inhibitory outcomes and work in conjunction to regulate T cell proliferation. The differential expression of PD-L1 by RSV-infected BDCA-1⁺ and BDCA-3⁺ mDCs suggests another mechanism by which RSV infection could impair the ability of mDCs to stimulate T cell proliferation as seen in mice [19].

The cytokine milieu present at the time of TCR activation is crucial in determining the outcome of T cell responses, and DC-derived cytokines and chemokines are at the center of a complex network of cells that regulate these signals. We found that RSV-infected BDCA-1⁺ mDC produced a profile of cytokines and chemokines associated with pro-inflammatory responses, whereas BDCA-3⁺ mDCs did not. Whether the differences in the cytokine profiles produced by RSV-infected BDCA-1⁺ and BDCA-3⁺ mDCs are due to inherent functional differences purported to exist between the subsets, or represents differences in the way RSV interacts with each mDC subtype is unclear. Activation with poly I:C has been shown to induce the production of pro-inflammatory cytokines from both BDCA-1⁺ and BDCA-3⁺ mDCs, as well as the increased production of IL-12(p70) by BDCA-3⁺ mDCs when activated with a cocktail of poly I:C and additional activation signals [13]. However, in contrast to purified toll like receptor agonists, viral pathogens likely activate multiple signaling pathways that lead to alternative patterns of cytokine expression by each cell type. Our findings of IL-12(p70) production by infected BDCA-1⁺ and not BDCA-3⁺ mDCs, may then represent responses specific to infection with RSV. Findings that as compared to activation with poly I;C, RSV infected BDCA-1⁺ mDCs produced greater amounts of proinflammatory and immunoregulatory cytokines, whereas RSV-infected BDCA-3⁺ mDCs produced decreased amounts of proinflammatory cytokines and increased amounts of immunoregulatory cytokines, provides additional evidence for RSV-specific responses. However, the TLRs governing RSV recognition, and their subsequent responses, in primary mDC subsets have yet to be defined. Thus, we cannot exclude the possibility that the differences in cytokine expression between RSV and poly I:C treated DCs are related to differences in the strength or quality of TLR activation by RSV, rather than from a virus-specific effect on cell activation. Further work is needed to determine the cytokine profiles produced by BDCA-1⁺ and BDCA-3⁺ mDCs in response to infection with other viral pathogens.

The balance of pro-inflammatory cytokines, in particular IL-12, produced by RSV-infected BDCA-1⁺ mDCs suggests that BDCA-1⁺ mDCs may promote Th1 responses, whereas RSV-infected BDCA-3⁺ mDCs, which produce IL-10 in the setting of low IL-12, may skew naïve T cells away from Th1 with possible polarization towards Th2 responses. Although mDCs were not found to produce IL-4, RSV-infected BDCA-1⁺ mDCs produced a number of chemokines implicated in the recruitment of cells known to produce IL-4 [40]. The potentially divergent roles of RSV-infected BDCA-1⁺ and BDCA-3⁺ mDCs in the polarization of T cell responses may explain why despite findings of Th2 responses in vivo[7, 41], in vitro studies have not demonstrated the ability of RSV-infected Mo-DCs to bias naïve T cells towards Th2 responses [6, 17, 42]. Of note, our findings of IL-12 production by BDCA-1⁺ mDCs and the lack of significant production of G-CSF, MCP-1 or MIP-1α by BDCA-3⁺ mDCs are in contrast to a recently published report by Johnson et al. examining the impact of RSV infection on the maturation and cytokine response of primary human mDCs isolated using sequential immunomagnetic selection alone [25]. These differences may be related to their use of a recombinant RSV strain, as well as differences in the methodologies used to isolate and culture mDC subsets. In prior reports, IL-12 and IFN-γ production by Mo-DCs [18] and cord-blood derived DCs [30, 31] infected with non-recombinant RSV strains has been demonstrated.

Interestingly, both BDCA-1⁺ and BDCA-3⁺ mDCs produced IL-10 in response to RSV. DC-derived IL-10 is shown to act in an autocrine fashion to inhibit pro-inflammatory cytokine production and impair the capacity of DCs to foster Th1 responses [43]. IL-10 produced by RSV-infected BDCA-1⁺ and BDCA-3⁺ mDCs may then also contribute to the impaired T cell responses seen in vivo. In prior reports using RSV-infected Mo-DCs, adding IL-10 blocking antibody did not restore T cell proliferation or function [17]. However, in vitro models of E. coli infection demonstrating that infected BDCA-1⁺ mDCs, but not Mo-DCs, impair T cell proliferation in an IL-10 dependent manner [44], provide further evidence that findings in Mo-DCs may not reflect the function of primary mDCs. IL-10 has also been used in vitro to generate tolerogenic DCs that are functionally anergic, and capable of generating and expanding Tregs [45, 46]. In vivo activation of pathways by pathogens that promote IL-10 production by DCs could therefore induce functionally tolerogenic DCs [47]. Tolerogenic DCs are characterized by decreased production of pro-inflammatory cytokines and increased production of IL-10 [46]. Similar findings in RSV-infected BDCA-3⁺ mDCs suggest that they may be functionally tolerogenic. Although RSV-infected BDCA-1⁺ mDCs also produced IL-10, the concomitant induction of IL-12 and other pro-inflammatory cytokines suggests a non-tolerogenic phenotype, which may explain why RSV-infected Mo-DCs were not found to generate T cells with suppressive properties [17]. Studies in murine models of acute RSV infection [48, 49] indicate that Tregs play an important role in determining the balance between effective antiviral immunity and controlling harmful immunopathology in the host response against RSV. Our findings suggest a possible mechanism for homeostatic regulation of the Treg compartment by BDCA-3⁺ mDCs during RSV infection.

Similar to studies using Mo-DCs and murine DCs [17‐19], RSV infection of BDCA-1⁺ and BDCA-3⁺ mDCs impairs their ability to stimulate CD4⁺ T cell proliferation in a mixed lymphocyte reaction. The lack of inhibition mDCs exposed to UV-inactivated RSV suggests that viral replication is required to trigger these events. The mechanism underlying this inhibitory effect is unknown; however, our findings of differential expression PD-L1 and IL-10 by RSV-infected mDCs suggest that RSV modulates the function of BDCA-1⁺ and BDCA-3⁺ mDCs. Although we did consider that the loss of stimulatory function may be due to reduced viability of infected cells, there was no difference in cell viability between uninfected or RSV-infected mDCs.

To our knowledge, this is one of the first studies to examine the functional responses of highly purified mDC subsets to infection with RSV. We recognize that using a reductionist in vitro model does not represent the interactions between the complex network of mDC subsets in the lungs and blood during natural infection and a suitable human model to further study these questions is needed. Furthermore, although RSV primarily affects young children, mDC subsets were isolated from the peripheral blood of healthy adults due to the feasibility of obtaining sufficient volumes of blood from pediatric populations. It is not well known whether age dependent differences in mDC subset function exist [50]. However, despite these limitations, this study provides novel evidence for a virus-specific and subset specific-effect of RSV infection on the functional response of BDCA-1⁺ and BDCA-3⁺ mDCs.

It is not known what type of mDCs are present in the lungs during natural RSV infection, or how host genetic and environmental factors affect the downstream function of BDCA-1⁺ and BDCA-3⁺ mDCs interacting with virus. However, findings of increased frequencies of BDCA-3⁺ mDCs in the peripheral blood of atopic asthmatics [51] suggest that the balance and function of individual mDC subsets during RSV infection plays a crucial role in determining the short- and long-term sequela of disease. Defining the mechanisms by which RSV is able to modulate virus-specific immune responses will provide important new insights into the immunopathogenesis of RSV infection and the role mDC subsets play in the host immune response against pathogens as a whole.