Dear editor,
Asthma is a major burden of disease that occurs in children and adults.
Respiratory infections are a risk factor for the development and
exacerbations of asthma, as well as COPD. Various reports indicate that
allergen-specific IgA concentrations may be reduced in (allergic) asthma
patients1, 2. IgA antibodies are predominantly
produced in local tissues such as the respiratory tract. IgA can be
secreted into the lumen of the airways and is believed to contribute to
protection against a broad range of infections and asthma exacerbations
as a result of allergen exposure3. An important
cytokine that contributes to asthma is thymic stromal lymphopoetin
(TSLP), which is demonstrated in clinical trials where blocking TSLP
resulted in alleviation of asthma disease symptoms4.
TSLP is produced in a short (shTSLP) form under homeostatic conditions,
and as a long form (loTSLP) following viral infection and in asthma.
ShTSLP exerts anti-inflammatory activity, whereas loTSLP promotes
inflammatory responses5. Here we aimed to elucidate
whether TSLP, expressed in the respiratory epithelium at elevated levels
as the long form in asthma6, directly affects the
secretion of IgA from B cells.
To this end, purified CD19 B cells from healthy donors were stimulated
in a T cell-dependent manner with CD40L cells, CpG, IL-2, IL-10 and
IL-21 for eleven days (more details in Supplementary methods ).
All voluntary donors provided informed consent following the guidelines
set by the Dutch government. Because no personal data were processed
this study did not require evaluation by a medical ethical committee.
The addition of shTSLP or loTSLP to these cultures did not affect B cell
proliferation (Fig S1) . However, loTSLP but not shTSLP
significantly reduced the secretion of IgA (Fig 1A ). Similar to
T cell-dependent inhibition of IgA production, loTSLP inhibited the
production of IgA under T cell-independent conditions using APRIL or
BAFF and IL-10 (not significant, Fig 1B ). This condition was
met by using primary normal human bronchial epithelial cells that are
targeted by viruses such as respiratory syncytial virus (RSV), and
confirmed to produce APRIL and BAFF (Fig S2 ), which are key
cytokines involved in local mucosal IgA production and B cell viability
in the absence of T cell help.
As asthma is associated with a T helper type 2 response, B cells were
subsequently stimulated in a T cell-dependent manner in the presence of
the type 2 cytokines IL-13 and IL-4. IL-13 and IL-4 promoted the
production of IgG, regardless of the presence of TSLP (Fig 1C,
S3 ). IL-13 alone did not alter IgA production, whilst in the presence
of loTSLP or shTSLP, IL-13 did inhibit IgA production. IL-4 alone was
sufficient to inhibit IgA production. As shown in Figure S3, in addition
to IgG1, IL-4 enhanced the production of IgG2,3 and especially 4. IgG4,
normally the least produced IgG subclass was produced at levels similar
to IgG1 when stimulated with IL-4. The production of IgA in the presence
of shTSLP was inhibited when IL-13 or IL-4 were also present, indicating
that these cytokines dominate over shTSLP.
The inhibition of IgA production by IL-4 was not caused by impaired B
cell viability or proliferation, as the highest B cells numbers were
rather observed in the presence of IL-4 (Fig S1 ). Analysis of
viable B cells, prior to terminal plasma cell differentiation and loss
of membrane expression of Ig, at day 7 of stimulation indicate that the
proportion of IgG+ B cells was increased by IL-4 alone, and by
combinations of TSLP and IL-4 or IL-13 (p<0.01 for all these
conditions, Fig 1D-E ). The percentage of IgA+ B cells was not
affected by TSLP.
To investigate whether TSLP would affect naïve (CD27-IgD+) and memory
(CD27+IgD-) B cells differently, both populations were FACS-sorted from
PBMCs and then stimulated in the T cell-dependent protocol in the
absence or presence of TSLP. As expected, naïve B cells show the lowest
IgA production levels, while memory B cells largely determined the
overall production of IgA (Fig 2A ). Again, loTSLP, but not
shTSLP, suppressed production of IgA by memory B cells
Retinoic acid (RA) is a potent inducer of (IgA) antibody secreting cells
(CD38+ CD20-) and known to reduce allergic inflammation (Fig
2B-C and ref 7 9). TSLP, either short or long, did
not influence the differentiation of B cells into antibody secreting
cells. In line with our previous results7, RA
upregulated IgA production, regardless of the presence of TSLP
(Fig 2D ). Also, in memory cells RA reverted the inhibition of
IgA production by loTSLP (Fig S4 ). Retinoic acid together with
either shTSLP increased the production of IgA by naïve cells, but not
memory cells.
Here we show that loTSLP induced by viral (RSV) infection can
subsequently inhibit the production of IgA by B cells. Under specific
conditions, shTSLP increased the production of IgA by B cells. Retinoic
acid, implicated to benefit asthma patients, promotes the production of
IgA and may also be able to restore IgA production in asthma patients in
the presence of aberrant TSLP signalling8. We propose
that either B cells from asthma patients may have altered responses to
TSLP, or the altered expression of TSLP by the epithelium of asthma
patients alone may drive diminished IgA production in asthma patients.
As various studies indicate that IgA levels may be diminished in asthma
patients, or that aberrant IgA responses may even precede asthma
development, future studies could investigate whether restoration of IgA
production may help to protect asthma patients against viral infections
and associated exacerbations1, 9. In
conclusion, we show that TSLP regulates IgA production, which may help
explain the mechanisms behind the development and exacerbations of
asthma and support the effectiveness of therapeutic interventions
targeting aberrant TSLP production.
The authors acknowledge Maria Rescignio and Giulia Fornasa, HUNIMED,
Italy for kindly providing shTSLP. We thank Puck van Kasteren, RIVM, and
Xiao Yu, Amsterdam UMC, The Netherlands for providing supernatants of
RSV-infected primary bronchial epithelial cells. The authors acknowledge
Dr. Diana Wouters for critically reviewing the manuscript.