To the Editor:
Asthma–chronic obstructive pulmonary disease (COPD) overlap (ACO) is
recognised as a distinct clinical disorder that can be differentiated
from asthma or COPD alone, owing to the rapid disease progression,
frequent exacerbations, and increased comorbidities.1A major impediment to the elucidation of mechanisms underlying ACO
pathogenesis has been the lack of development and detailed validation of
appropriate animal models. Administration of papain, a cysteine protease
from the fruit of the plant species Carica papaya, was shown to
induce acute/subacute allergic airway inflammation (AAI) via type 2
responses.2 Results from previous mouse model
experiments also indicated that short-term exposure to papain induces
emphysema.3 We hypothesised that weekly intratracheal
aerosol administration of papain could induce a clinically-feasible
murine ACO model that features concurrent emphysema, persistent AAI, and
airway hyper-responsiveness. Further, we verified its validity with
multifaceted analysis including comprehensive cytokine analyses,
neutrophil gelatinase-associated lipocalin (NGAL) measurement, and
long-term analyses.
We first validated the model with respect to pulmonary emphysema.
Porcine pancreatic elastase (PPE) has been used to induce murine models
of COPD.4 We intratracheally administered PPE or
papain using a microsprayer (Figure 1A). The changes in body weight
detected among the mice in the PPE- or papain-treated groups were not
significantly different from the controls (Figure S1). Both PPE- and
papain-treated groups showed increased inspiratory capacity/weight and
dynamic compliance (Figure 1B); representative lung histology revealed
PPE- and papain-induced emphysema (Figure 1C). The mean linear intercept
was significantly increased in the PPE- and papain-treated groups
(Figure 1D). As such, we concluded that both PPE and papain induced
emphysema. Furthermore, goblet cell hyperplasia was detected exclusively
among mice in the papain-treated group (Figure 1E). Consistent with this
finding, whole lung RNA preparations from the papain- (but not PPE-)
treated mice exhibited increased expression of transcript encoding
Muc5ac (Figure S2).
We then evaluated asthmatic features associated with papain
administration. Representatives of bronchoalveolar lavage fluid (BALF)
cytospin preparations are shown in Figure 1F. The total number of cells,
macrophages, and eosinophils were significantly increased in the
papain-treated group (Figure 1G). The papain-treated group exhibited
higher airway responsiveness to increasing doses of methacholine
compared to other groups (Figure 1H); no differences in baseline airway
resistance were observed (Figure 1B). Whole lung homogenates of the
papain-treated mice revealed increased expression levels of Eotaxin1 and
Eotaxin2 (Figure S2). Thus, our papain-treated model had both
COPD-associated and asthmatic features and was considered to be an
appropriate model for ACO.
For more detailed analyses, multiple cytokines/chemokines in the BALF
were quantitatively detected (Figure 2A; cytokine array data are shown
in Table S1). Regarding macrophage/neutrophil-related
cytokines/chemokines, macrophage colony-stimulating factor (M-CSF),
keratinocyte-derived chemokine (KC), and IL-6 were detected in BALF from
the papain-treated mice only. With regard to type 2
inflammation-associated cytokines, IL-4 levels were not significantly
different among the groups while IL-5 and IL-13 levels were detectable
only in the papain-treated mice. Collectively, our ACO model reproduced
the characteristics of both asthmatic and COPD-associated airway
inflammation. IL-33 concentrations were not significantly different
among the groups (Figure S3). The levels of total IgE, a serum marker
used to define asthma endotypes,5 were significantly
higher in the papain-treated mice (Figure 2B). Significantly higher
levels of NGAL were detected in the BALF of papain-treated mice; NGAL
has been identified as a promising biomarker for clinical evaluation of
ACO6 (Figure 2C).
We also provide a more long-term (8 weeks; day 56) assessment of the
papain-treated mice (Figure 2D). BALF eosinophils remained significantly
elevated, a finding that indicates prolonged eosinophilic inflammation
in response to papain (Figure 2E). Inspiratory capacity/weight and lung
compliance also remained elevated at the 8-week time point (Figure 2F),
as did emphysema-associated lung pathology (Figure 2G). ELISA was
notable for elevated serum levels of total and papain-specific IgE at
both 4 and 8 weeks (Figure 2H). Taken together, our papain-induced model
of ACO included prolonged eosinophilic type 2 inflammation along with
emphysema. The papain-treated group at both 4 and 8 weeks similarly
exhibited higher levels of NGAL in the BALF (Figure 2I).
Polyinosinic:polycytidylic acid [poly(I:C)] is a form of dsRNA that
acts as a TLR3 agonist and is used to mimic viral infection-induced
exacerbations of airway inflammation.7 In our
poly(I:C)-induced exacerbation models (Figure S4A),
neutrophil/macrophage-dominated inflammation was triggered while AAI was
sustained in the ACO model (Figure S4B-E). We also found administration
of poly(I:C) resulted in increased levels of NGAL both in serum and BALF
(Figure S4F); of note, serum NGAL levels were increased only slightly
above baseline by papain administration alone (Figure 2I). Since NGAL is
expressed not only in the neutrophils but also in the respiratory
epithelial cells,8 we hypothesised that
poly(I:C)-induced elevations in NGAL were largely related to the influx
of neutrophils; by contrast, NGAL resulting from papain administration
alone may be result from inflammation-activated and/or damaged
epithelial cells.
Paediatric asthma is a major predictor for the development of clinical
ACO after middle age.9 As our model features
administration of papain to mice at 5–6 weeks of age, our model would
correspond to patients diagnosed with ACO associated with early-onset
asthma. Specifically, our ACO model reflects early-onset, eosinophilic,
high IgE, and Th2-dominant asthma phenotypes/endotypes. We recognise
that emphysema in this model develops rapidly and at an earlier time
point than what might be anticipated clinically and is not associated
with smoking; these are both potential limitations of this model.
However, this model is simpler and easier to use than most smoking
models as it does not require several months of daily procedures and
special equipment.4
To the best of our knowledge, this is the first established murine model
of ACO that includes both prolonged emphysema and asthmatic features.
Additionally, this is the first model of ACO that evaluated elevated
levels of NGAL and as such parallels human clinical findings. The strong
validation of this model includes results of comprehensive cytokine and
NGAL measurements that reflect findings from clinical reports focused on
ACO.6 Our murine model will be useful for elucidating
the pathological processes associated with ACO and for identifying new
diagnostic markers and therapeutic targets.