The high density of DCs in splenic MZ bridging channels has been appreciated for 30 years (Witmer and Steinman, 1984
) yet the factors controlling this localization have been unknown. The above findings establish a crucial role for EBI2 and its oxysterol ligand, 7α,25-OHC, in positioning CD4+
DCs in MZ bridging channels and in LN interfollicular regions. EBI2 is also required for the maintenance of a CD4+
DC compartment of normal size. These roles of EBI2 in DCs are necessary for supporting normal CD4 T cell and B cell proliferative responses, and early plasma cell and germinal center responses to particulate blood-borne antigens. These observations are in close agreement with a study that appeared online at the time this work was submitted (Gatto et al., 2013
). Our study additionally provides evidence that EBI2-mediated positioning of CD4+
DCs in splenic MZ bridging channels promotes two important processes: (1) encounter with B cell-derived LTα1β2 that engages the DC LTβR, delivering a signal necessary for maintaining the homeostasis of the population; and, (2) efficient exposure to blood-borne particulate antigens and an ability to promptly access the T-B zone interface to stimulate T-dependent B cell responses ().
Model for the role of EBI2 in mediating marginal zone (MZ) bridging channel positioning of CD4+ DCs in the spleen.
Our finding of a similar disruption in CD4+
DC numbers and bridging channel positioning in mice incapable of making 7α,25-OHC (Ch25h- and Cyp7b1-deficient) and in mice unable to properly degrade 7α,25-OHC (Hsd3b7-deficient) (Russell, 2003
; Yi et al., 2012
) provides strong evidence that the critical function of EBI2 in CD4+
DCs is to mediate their correct positioning. The alternative possibility that EBI2 is providing a maintenance signal that is independent of its positioning role seems unlikely as Hsd3b7−/−
mice with elevated 7α,25-OHC would not be expected to have the same defect in EBI2 signaling as mice that lack 7α,25-OHC. Cyp7b1 and Ch25h are abundantly expressed in the outer follicle and in fibroblastic reticular cells (Yi et al., 2012
) and our present findings suggest the necessary source of 7α,25-OHC for bridging channel positioning and maintenance of CD4+
DCs is stromal. Further experiments will be needed to determine the nature of the stromal cells expressing these enzymes in MZ bridging channels and LN interfollicular regions.
In previous work that established a role for LTα1β2 in maintaining normal numbers of CD4+
DCs within the spleen, a slight reduction in the rate of Ltbr−/−
DC turnover was observed (Kabashima et al., 2005
; Wang et al., 2005
). The basis for our inability to observe a statistically significant reduction in DC turnover in EBI2-deficient mice is not clear but may indicate that the cells continue to receive low level LTβR engagement. Precisely how a reduction in LTβR engagement leads to a decrease in CD4+
DC numbers needs further investigation. The sufficiency of increased EBI2 expression to lead to greater numbers of CD4+
DCs may be a consequence of increased LTβR signaling since increased LTβR engagement is adequate to increase CD4 DC numbers (Kabashima et al., 2005
; Wang et al., 2005
). To account for our inability to observe changes in the CD4+
DC turnover rate or rescuing effects of antagonizing apoptosis when EBI2 function was lacking, we suggest that CD4+
DCs are lost in EBI2-deficient mice at a similar rate irrespective of how long it has been since they were generated from pre-DCs or last proliferated. It seems possible that the cells suffer from the combined effect of mispositioning due to EBI2-deficiency and lack of expression of molecules downstream of LTβR signaling that control interactions with other cell types, with the outcome that the CD4+
DCs are more frequently engulfed by phagocytes or become caught in blood flow and lost from the spleen. In this regard, it is notable that mice lacking CD47 or its partner protein SIRPα, have a similar deficiency in splenic CD4+
DCs (Hagnerud et al., 2006
; Van et al., 2006
; Saito et al., 2010
). Although we did not observe alterations in CD47 or SIRPα expression by DCs in EBI2-deficient mice (T. Yi and J. Cyster, unpublished observation), it is possible that the function of these molecules in mediating cell–cell interactions or in regulating engulfment by phagocytic cells is somehow altered in EBI2-deficient DCs.
The marked defect in T cell proliferative responses in EBI2-deficient mice following injection with antigen-conjugated to SRBCs, but intact response to soluble Ova, observed here and in the recent study of Brink and coworkers (Gatto et al., 2013
), suggests that EBI2 function in splenic DCs is most important for responses against particulate antigens. We provide evidence that one reason for this differing dependence on EBI2 is that particulate antigen is not able to freely access DCs already positioned within the T zone. This finding is consistent with other studies showing an inverse relationship between the size of molecules and their ability to diffusively access the white pup (Nolte et al., 2003
). The intact response of EBI2-deficient mice to 33D1-Ova was unexpected but suggests that the defective response to SRBC-antigen is not solely a consequence of reduced CD4+
DC numbers. Instead, these data suggest that appropriate positioning in MZ bridging channels, and thus in close proximity to the marginal sinus and blood-rich MZ, is critical for supporting the response against particulate antigens. This requirement might reflect both the improved efficiency with which DCs in this region can capture blood-borne particles and the ability to be triggered for rapid movement into the T zone (). Although 33D1 coupling ensured efficient delivery of antigen to CD4+
DCs it did not promote their movement into the T zone (Chappell et al., 2012
and data not shown).
The reduction in CD4 T cell responses in mice lacking EBI2+
DCs described here and by Gatto et al. (2013)
might be sufficient to account for the reduced plasma cell and germinal center responses observed in both studies. However, a recent report showed that when antigen was targeted to 33D1+
DCs as an antibody conjugate, it promoted some B cell activation events in a manner that did not depend on T cell help (Chappell et al., 2012
). A number of studies have shown that DCs can directly augment B cell responses (MacPherson et al., 1999
; Balazs et al., 2002
; Jego et al., 2005
; Qi et al., 2006
). Our finding that 33D1+
DCs move to the B-T zone interface following particulate antigen immunization also seems consistent with the possibility that these DCs interact with activated B cells as well as T cells. Future studies will be needed to determine whether EBI2-dependent DC–B cell interactions contribute to driving early B cell activation events during responses to particulate antigens.
The surface markers ICOS and PD1 are highly expressed by T follicular helper (Tfh) cells (Vinuesa and Cyster, 2011
). The less effective induction of ICOS and PD1 on T cells responding to SRBC-Ova in EBI2-deficient hosts suggests that antigen presentation by CD4+
DCs may be important in favoring induction of an early B-helper phenotype in CD4 T cells. In addition to efficiently presenting antigen in the context of MHC class II (Pooley et al., 2001
; Dudziak et al., 2007
) it is possible that appropriately activated CD4+
DCs bias T cell differentiation in a manner that favors provision of B cell help. Consistent with early induction of Tfh cell properties in activated T cells augmenting B cell responses, when T cells lack the ‘master regulator’ of Tfh cell differentiation, Bcl6, they are poorly able to support extrafollicular plasmablast responses (Lee et al., 2011
). The reduced induction of Tfh-phenotype cells likely also contributes to the reduced germinal center response.
Our findings indicate that CD4+
DC positioning is controlled by a balance of EBI2 and CCR7 expression and ligand responsiveness. Under homeostatic conditions, EBI2 has a dominant influence and promotes positioning in MZ bridging channels. However, following activation by antigen exposure the balance shifts in favor of CCR7 and the cells move promptly into the T zone (). While we suggest that direct exposure to antigen triggers DC movement, we do not exclude the possibility that CCR7 upregulation is promoted indirectly as a consequence of cytokine production by other antigen-exposed cells. The basis for the antigen activated DCs favoring the B-T zone interface compared to the central T zone is not yet clear but this does not seem to depend on EBI2 expression. These data add to a series of findings showing how differential responsiveness to chemoattractants emanating from adjacent zones determines cell position (Reif et al., 2002
; Cyster, 2005
; Bromley et al., 2008
). Although LPS causes CCR7 upregulation on splenic DCs and promotes their movement to the T zone (De Becker et al., 2000
; De Trez et al., 2005
; Seul et al., 2012
) we found that LPS-mediated repositioning happened more slowly and with a more uniform distribution of DCs through the T zone than occurred following SRBC immunization. The DC sensor triggered following SRBC immunization is not clear but in preliminary experiments we do not find a requirement for Myd88 (unpublished observation). SRBCs rapidly become opsonized by complement and presumably also by natural antibody and it is possible that these factors contribute to SRBC capture by CD4+
DCs and to triggering CCR7 upregulation. It seems possible that detection of damaged endogenous RBCs may be a trigger for DC maturation as a number of important pathogens, most notably the malaria parasite, propagate inside RBCs, insert foreign proteins in their membranes and alter their functional properties.