Here we show that IL-17-producing T cells are involved in the formation of ectopic lymphoid follicles in the lung, and that IL-17 appears to act by triggering the expression of CXCL13 and CCL19 independently of LT signaling. Thus, we propose a new model of ectopic follicle formation in which the initial step is the production of IL-17 by non-LTi cells, such as T cells. In turn, IL-17 promotes the expression of CXCL13 and CCL19, which recruits B and T cells and maintains the structure of FDC networks. Once the initial inflammatory response resolves, the expression of CXCL13 and CCL19 as well as the overall architecture of iBALT is maintained by the homeostatic interaction of LT-expressing lymphocytes with LTβR-expressing stromal cells or vascular endothelial cells. This model is also consistent with the requirement of T cells, but not LTi cells, for the formation of ectopic follicles in transgenic mice expressing CCL21 in the thyroid31
. Moreover, these data suggest that simultaneously targeting both IL-17 and LT signaling pathways may be beneficial in diseases such as Sjogren's syndrome or multiple sclerosis, in which the development of ectopic follicles is linked to local pathology.
A striking observation from these studies is that iBALT forms much more easily in neonates than in adults, a fact that we initially attributed to the higher frequency of LTi cells in neonates. However, our results with Rorc−/−
mice suggest that LTi cells are not required for the development of iBALT - a scenario similar to the RORγ- and Id2-independent development of tear duct-associated lymphoid tissue (TALT)32
and the milky spots of the omentum33
as well as the RORγ-independent (but Id2-dependent) development of nasal-associated lymphoid tissue (NALT)34,35
. All of these tissues develop in the neonatal period and may be dependent on different types of LTi cells or perhaps IL-17-producing T cells. A caveat to this explanation is that, in addition to the development of LTi cells, RORγ is important for the production of IL-17 by T cells - both αβT cells36
as well as γδT cells (not shown). Thus, the formation of ectopic lymphoid tissues in Rorc−/−
mice may rely on compensatory mechanisms as suggested by studies looking at the development of ectopic follicles in the colons of Rorc−/−
. Alternatively, despite reduced IL-17 production by Rorc−/−
T cells, the residual IL-17 produced by T cells may trigger low expression of CXCL13 in the lung, which in mice lacking conventional secondary lymphoid organs, may be sufficient to recruit lymphocytes and initiate lymphoid organogenesis.
Assuming that iBALT formation is not dependent on LTi cells, but is instead dependent on IL-17-producing T cells, a different hypothesis is required to explain the preferential development of iBALT in neonates. For example, γδT cells first emerge from fetal progenitors prior to αβT cells38
, suggesting they are the first T cells available during the neonatal period. γδT cells also express IL-17 in response to microbial products39
and many γδT cells are located in the lung, where they would be positioned to promote iBALT development. Moreover, Tregs are not exported to the periphery during the first week after birth and as a result, neonatal thymectomy leads to autoimmunity40
. Given that the lack of proper Treg homing in CCR7-deficient mice appears to promote spontaneous iBALT formation41
, it is possible that the relative paucity of Tregs and relative abundance of IL-17-producing γδT cells in neonates may ease the formation of iBALT following pulmonary inflammation. The preferential formation of iBALT in neonates is also consistent with observations in humans showing that iBALT is infrequently found in healthy adults42,43
, but is more frequently found in the lungs of healthy children 44
. Moreover, the incidence of iBALT approaches 100% in the lungs of neonates with pulmonary infections45
Of course, the formation of iBALT is not exclusively restricted to the neonatal period, as iBALT can be observed in adult mice following pulmonary viral4,6
and bacterial infections46,47
as well as the pulmonary administration of cigarette smoke48
, particulates or protein nanoparticles49
. It is also observed in patients with a variety of inflammatory conditions, such as COPD9
, and rheumatoid arthritis7
. Although these conditions often develop in adults, it is unknown whether the patients with iBALT experienced pulmonary inflammation as neonates, which may have conditioned their lungs to form iBALT as adults. Nevertheless, we would suggest that conditions that lead to the recruitment and activation of any IL-17-producing cell type, including γδT cells, Th17 cells, NK cells and LTi cells, is likely to lead to the development of ectopic structures, particularly if the stimulus is chronic, such as in rheumatoid arthritis50
and multiple sclerosis51
or upon chronic infection.
Given the association of Th17 cells with bacterial infections52
, it is not surprising that the frequency of Th17 cells in a particular individual is dependent on their history of exposure to pathogens and even commensal bacteria in the gut. Similar observations have been made in mice. In particular, those mice exposed to Segmented Filamentous Bacteria (SFB) in the gut have high frequencies of Th17 cells53
. Since mice from different commercial vendors and institutional colonies have widely different gut flora, including SFB, each colony has different frequencies of Th17 cells, which leads to different biological outcomes in terms of inflammatory diseases, including autoimmunity54
. The mice used in our study were imported to the Trudeau Institute, where they were rederived by embryo transfer into germ-free foster mothers that were reconstituted with altered Schaedler's flora, a mix of eight species that does not include SFB55
. These animals were then maintained in a barrier facility at Trudeau Institute and then imported to a barrier facility at the University of Rochester. Therefore, we suspect that the limited diversity of gut flora in our mice is one reason that we do not observe the development of iBALT as adults when exposed to influenza infection or intranasally administered LPS. In contrast, other investigators using genetically identical mice are likely to obtain different results depending of the gut flora in their colonies.
In summary, our results establish a new paradigm of how ectopic lymphoid follicles form in non-lymphoid tissues and show that the initial step in ectopic follicle formation is dependent on IL-17, which triggers the maturation of the FDC network and promotes the expression of CXCL13 and CCL19 independently of LT. In turn, these chemokines recruit lymphocytes, which form organized B and T cell domains that are maintained via LT-dependent homeostatic mechanisms once inflammation is resolved.