DISCUSSION
The eLTs can produce antibodies locally and therefore perpetuate
inflammation in NPs.8 However, the mechanisms
underlying the formation and expansion of eLTs in NPs remain unexplored.
In this study, for the first time, we identified a critical role for
IL-17A-induced stromal cell remodeling in the initiation, and crosstalk
between B cells and stromal cells via CXCL13 and
LTα1β2 in the enlargement and
maintenance of eLTs in NPs.
Expression of homeostatic chemokines is central to the initiating events
that lead to lymphoneogenesis.13-15, 39, 40 In NPs
without eLTs, we found that stromal cells were the major cellular source
of CXCL13, indicating that stromal cells are critical for the initiation
of B cell recruitment and compartmentalization in NPs. Evidence that
IL-17A may play an important role in eLT formation by inducing the
production of lymphoid chemokine CXCL12 and CXCL13 emerge from recent
animal studies of bronchial infection and autoimmune
encephalitis.18, 32 In this study, we found that
IL-17A expression correlated with CXCL13 expression in NPs. We further
discovered that nasal stromal cells had the expression of IL-17RA and
IL-17A induced CXCL13 production in nasal stromal cells. Although IL-17A
induced the production of CXCL12, another important B cell chemokine,
during bacteria-induced lung lymphoid neogenesis,18 we
failed to find an induction of CXCL12 in nasal stromal cells by IL-17A,
which is consistent with our previous finding of no association between
CXCL12 expression and eLT formation in NPs.8Therefore, distinct mechanisms may underline the effect of IL-17A in
promoting eLT development in different organs and pathological
conditions.
Stromal cells have a complex role at local microenvironments, which
induce immune cell migration, activation and survival, and support
lymphoid enlargement. FRCs provide homeostatic chemokines, and secrete
extracellular matrix proteins to form the structural framework for
immune cell interaction in SLOs.41 BECs and LECs
regulate lymphocyte entry into SLOs. DNs have recently been shown to
contain a novel subset of fibroblastic contractile
pericytes.42 Although the phenotype and function of
stromal cells are well documented in SLOs, little is known of their role
in eLT formation. In this study, we revealed an expansion of FRC
population of stromal cells in both NPs with and without eLTs, which was
likely induced by IL-17A and LTα1β2. The
expansion of FRCs in NPs without eLTs indicates a role of FRCs, but not
other types of stromal cells, in eLT formation in its infant stage given
to the findings that stromal cells were the main producer of CXCL13 in
NPs without eLTs and FRCs were the major source for CXCL13 in stromal
cells in NPs. LECs is also a fundamental compartment in controlling SLOs
organogenesis.43 Nevertheless, LECs were only expanded
in NPs with eLTs, and the expansion of LECs was induced by
LTα1β2 but not IL-17A. Since
LTα1β2 was only upregulated in NPs with
eLTs, LECs are more likely involved in the enlargement rather than the
initiation of eLTs in NPs by facilitating the entry of lymphocytes into
NPs.
In NPs with eLTs, we found that B cells were main producer of CXCL13.
After the B cell recruitment under the control of CXCL13 derived from
stromal cells, B cells themselves may provide a “second wave” of
supply of CXCL13. LTα3 and
LTα1β2are reported to be involved in eLT formation.44, 45Compared with control tissues, the mRNA expression of LTβ was only
upregulated in NPs with eLTs, suggesting an involvement of membrane form
of LT in the later stage of eLT formation in NPs. Previous studies show
that CXCL13 induces murine splenic B cells to upregulate membrane-bound
LT via Grb2, and CXCL13 expression induction is dependent on LTβR
pathway in SLOs.37, 38, 46 In this study, we
demonstrated a positive feedback loop between CXCL13 and
LTα1β2 on B cells in eLTs in NPs, which
obviously exaggerates the B cell recruitment and compartmentalization.
In addition to B cells, 36% reduction of CXCL13 expression was founded
in NPs with eLTs after depletion of stromal cells. Nasal stromal cells
also had LTβR expression. We found that
LTα1β2 reshaped stromal cells to FRC and
LEC type and promoted their CXCL13 production. Thus, the crosstalk
between stromal cells and B cells further perpetuate the eLT development
in NPs.
Using a murine model with high nasal type 17 inflammation, we confirmed
that IL-17A was able to induce eLT formation in nasal mucosa and this
process was dependent on CXCL13 and
LTα1β2 in vivo .In the animal study, we found that only high levels of IL-17A induced by
100 μg curdlan led to eLT formation. In contrast, the comparatively
lower IL-17A levels induced by curdlan at 20 μg only induced lymphocyte
clusters. This is in line with the finding in humans that the IL-17A
levels were more prominently elevated in NPs with eLTs than those
without eLTs.
There are several limitations in this study. We established a mouse
model by using curdlan, which elicits a high local IL-17A inflammation
but may not mirror the pathogenesis of NPs in humans. It is considered
that the type 17 response is less important for Caucasian patients with
NPs than for Chinese patients. Nevertheless, eLTs have also been
reported in Caucasian patients,47 indicating that
additional mechanisms may underlie eLT formation in Caucasian patients.
These comments notwithstanding, for the first time, we have established
a paradigm of how eLTs are formed in NPs in Chinese patients. We suspect
that targeting IL-17A, CXCL13 and LTα1β2may provide opportunities for the design of therapies to manipulate eLT
formation and alleviate inflammation in patients with NPs.