Increase in secreted airway mucins and partial Muc5b STAT6/FoxA2 regulation during Pneumocystis primary infection

Airway mucus responses to subclinical infections may explain variations in progression of chronic lung diseases and differences in clinical expression of respiratory infections across individuals. Pneumocystis associates to more severe Chronic Obstructive Pulmonary Disease (COPD), asthma, respiratory distress of premature newborns, and is a consistent subclinical infection between 2 and 5 months of age when hospitalizations for respiratory cause and infant mortality are higher. This atypical fungus associates to increased mucin 5AC (MUC5AC), a central effector of Th2-type allergic inflammation, in infant lungs. However, mucus progression, expression of MUC5B essential for airway defense, and potential for pharmacologic modulation of mucus during Pneumocystis infection remain unknown. We measured MUC5B and Pneumocystis in infant lungs, and progression of mucin levels and effect of inhibition of the STAT6/FoxA2 mucus pathway using Kaempferol, a JAK/STAT6 inhibitor, in immunocompetent rats during Pneumocystis primary infection. Pneumocystis associated to increased MUC5B in infant lungs. Muc5b increased earlier and more abundantly than Muc5ac during experimental primary infection suggesting an acute defensive response against Pneumocystis as described against bacteria, while increased Muc5ac levels supports an ongoing allergic, Th2 lymphocyte-type response during primary Pneumocystis infection. Kaempferol partly reversed Muc5b stimulation suggesting limited potential for pharmacological modulation via the STAT6-FoxA2 pathway.

been documented 14 . MUC5B has an essential role in defense against bacterial pneumonia, and lack of this mucin severely affected infection-related survival in animal models 8,10 .
Pneumocystis, a fungus well-known by the severe pneumonia of immunocompromised individuals, associates to increased severity of Chronic Obstructive Pulmonary Disease (COPD) 15 , to respiratory distress syndrome of newborns 16 , and is likely the most frequent and consistent infection of early infancy 17 . This primary Pneumocystis infection of immunocompetent infants goes undetected and peaks between two and five months of age [18][19][20][21] providing a particular epidemiologic context that coincides with the highest prevalence of infant hospitalizations for respiratory cause 22,23 . We have reported increased MUC5AC and CLCA1 associated to Pneumocystis primary infection in lungs of infants dying in the community 19,24 . Understanding the effect of Pneumocystis on mucin production and its controlling pathways is underscored by the epidemiological context of this fungal infection [18][19][20][21] and by the demonstration that Pneumocystis primary infection induces a Th2 environment in the healthy lung 25,26 . MUC5AC has been recently described as an essential effector of the epithelial response to allergic inflammation 7 . Increased MUC5AC is consistent with the intense Th2 (allergic type) airway immune response 25,26 and STAT6 pathway activation 27 plus induction of mucus-related genes such as Muc5ac and Clca3 28 associated to mucus hypersecretion documented in animal models of Pneumocystis infection 26,27,29 . Of interest, anti-Muc5ac immune staining is able to recognize only a minimal fraction of the mucus that stains with Alcian blue, suggesting additional mucins are involved during Pneumocystis infection 26 . No studies of MUC5B expression in infant lungs or of the murine homolog gene Muc5b during Pneumocystis infection are available.
This work shows for the first time, that Pneumocystis associates to increased levels of MUC5B in infants, and replicates this finding in an experimental animal model of naturally acquired primary infection that resembles the mode of contagion and course of the primary infection in humans. We also show that pulmonary Muc5b occurs earlier and is more abundant that Muc5ac, that the mechanism of Muc5b hypersecretion partly depends on STAT6 stimulation by Pneumocystis -that inhibits FoxA2 repressor-, and that this mechanism can be reversed by pharmacological de-repression of FoxA2 using Kaempferol, a specific inhibitor of JAK3 that activates STAT6 30 .

MUC5B determinations in infant lungs.
The clinical characteristics of infants whose autopsy lung biopsies were selected for this study were previously described 19 . Eight Pneumocystis-positive and eight control samples were randomly selected for analysis from a pool of 39 Pneumocystis-positive and 20 Pneumocystis-negative lung specimens. MUC5B protein levels increased in Pneumocystis-infected compared to non-infected infants (P < 0.001) (Fig. 1).

Muc5b and Muc5ac mRNA and protein levels in lungs of
Pneumocystis primary infection experimental and control animals. mRNA levels of Muc5b were determined by qRT-PCR and compared with Muc5ac levels at each sacrifice point. mRNA levels of both mucins increased from day 60 and reached significance on day 75 ( Fig. 2A,C). Of interest, the mRNA levels of Muc5b in the Pneumocystis-positive group increased over 10-fold than those of Muc5ac (Fig. 2E). Immunoblotting showed that protein expression levels of Muc5ac increased on day 75 (Fig. 2B), and Muc5b increased from day 60 reaching significance on days 60 and 75 (Fig. 2D). Muc5b protein levels were over 2-fold higher than those of Muc5ac in Pneumocystis-positive animals (Fig. 2F). Pneumocystis lung burden in these animals, estimated by qPCR of the Dihydrofolate Reductase (dhfr) gene of Pneumocystis, was characterized in previously published experiments 26 . Immunohistochemistry for Muc5b and Muc5ac. Airways were classified according to diameter in bronchi (>250 μm) and bronchioles (<250 μm) 31 . Terminal bronchioles (<50 μm) were not observed in these microscopy slide sections. Muc5b was highly detected in bronchi of infected animals on days 45 and 75 of age compared to those of controls (Fig. 3A,B). Muc5b determinations on bronchioles of control and infected animals were not different (Fig. 3C). Signal for Muc5ac was not visualized in sections adjacent to those positive for Muc5b (Fig. 3D).
Effect of Kaempferol, a specific inhibitor of the JAK/STAT6 pathway, on Pneumocystis-induced STAT6 phosphorylation. The phosphorylation state of transcription factor STAT6 was evaluated in lung tissue using anti-Phospho STAT6 (P-STAT6) specific antibodies and Western Blot. P-STAT6 protein levels were increased in Pneumocystis-infected Kaempferol-untreated animals respect to control rats treated with saline. The Pneumocystis burden reached similar levels than in lungs where muc5ac 26 and muc5b mucins were quantitated (Fig. 4A). The P-STAT6 stimulation effect of Pneumocystis was attenuated by Kaempferol as documented by the Effect of Pneumocystis in transcriptional regulation of Muc5b promoter. JAK/STAT6 pathway normally regulates mucin transcription through the repressor FoxA2 which binds to the respective mucin promoter to control transcription and consequently mucin mRNA levels. A highly conserved FoxA2 binding site was identified by comparison of DNA sequences of the human, mouse and rat Muc5b promoter regions. The presence of FoxA2 in this Muc5b promoter region was first confirmed using chromatin immunoprecipitation (ChIP). Binding of FoxA2 decreased in 0,5-fold in Pneumocystis-infected animals compared to controls (Pc− v/s Pc+) (Fig. 6A). Kaempferol partly reversed the effect of Pneumocystis (Pc+v/s Pc + K). No differences were detected in FoxA2 mRNA and protein levels analyzed from total RNA and protein extracts, after Kaempferol treatment ( Fig. 6B,C). However, a significant decrease in FoxA2 nuclear protein levels was detected in animals infected with Pneumocystis receiving saline only (Fig. 6D,E). This effect of Pneumocystis in lowering nuclei FoxA2 levels was attenuated completely by Kaempferol (Fig. 6D,E).

Discussion
This work documents increased expression of MUC5B associated to Pneumocystis primary infection in autopsy lung samples from immunocompetent infants dying in the community (Fig. 1), and the progression of gel-forming mucins (Muc5b and Muc5ac), and potential for pharmacological reduction of mucus through the STAT6/FoxA2 pathway, in an animal model of primary infection by Pneumocystis. We have previously documented increased MUC5AC associated to Pneumocystis in infant lung autopsy samples and that this increase was unrelated to common respiratory viruses 19,24 . This association has proven consistent 19,24 . This work ads to show that Pneumocystis strongly stimulates both main gel-forming mucins suggesting an altered mucus physiology during the primary infection 1,19 . Overproduction of MUC5B has also been associated with chronic obstructive pulmonary disease (COPD), and to the development of interstitial pulmonary fibrosis (IPF) in adults, and overproduction of MUC5AC with asthma 2 . The role and duration of Pneumocystis associated mucus changes in these medical conditions associated to Pneumocystis, remain to be determined.
Muc5b levels increased 10 to 15 times, and this increase occurred earlier than that of Muc5ac in animals in this study suggesting a faster Muc5b than Muc5ac response during Pneumocystis primary infection ( Fig. 2A-D), and that Muc5b may play a defensive role against Pneumocystis as recently described for bacterial pathogens 8 . Of interest, secreted MUC5AC increased later in the infection course ( Fig. 2A,B). This progression is consistent with the findings that the primary infection by Pneumocystis can drive the development and progression of asymptomatic asthmatic-type airway pathology as described 25,26 , and is also consistent with the role of Muc5ac in the development of airway hyperreactivity 7 adding to the possibility of altered airway responses during the course of this primary infection 19,26 . Muc5B was documented histologically in bronchi but not in bronchioles (Fig. 3). Goblet, mucus producing, cells are abundant in bronchi and rare in bronchioles and findings suggest stimulation of existing goblet cells rather than metaplasia during the time course of Pneumocystis infection evaluated in these experiments. This animal model reproduces the natural route of contagion and progression of this fungal infection in humans 32 , and documents an ongoing immune response throughout the course of the Pneumocystis primary infection 26 . Translating these findings to the human situation suggests that Pneumocystis may prime the lung to altered responses to co-infecting respiratory pathogens and support the hypothesis that subclinical, unnoticed, Pneumocystis may contribute to the higher incidence of respiratory morbidity and increased rate of hospitalizations for respiratory cause that occurs in infants between 2 and 5 months of age 22,23 . Respiratory viruses affect infants at any age. Pneumocystis however, is highly frequent between 2 and 5 months of age and is rare after the age of 6 months 18-21 . This possible, yet-unproven, role as a sub-clinical fungal co-pathogen is suggested by the extremely high prevalence of Pneumocystis in infants 19,21 , that peaks coincidently at this age window [18][19][20][21] and by the pulmonary pathology documented in animal models [25][26][27][28] . An important mechanistic result of this work is the documentation that Muc5b production is to certain extent dependent on STAT6/FoxA2 pathway and that pharmacological intervention of the STAT6/FOXA2 pathway, as has been previously proposed 33 , can serve as a mechanism for modulation of mucus during Pneumocystis infection. STAT6 is an important pathway in the innate immune response to Pneumocystis as suggested in our previous studies related to the association of Pneumocystis and MUC5AC increase in infant lungs and documented in STAT6 knockout mice 24,27 . STAT6 is activated by IL13 that favors mucin expression through binding inhibition of transcriptional repressor FoxA2 to the mucin promoters 9,34 . FoxA2 reduces the expression of mucins and, therefore, binding-inhibition of FOXA2 is necessary to induce enhanced expression of MUC5B 35 . Paradoxically, increased methylation of FoxA2 binding site in the mucin promoter has been shown to enhance overexpression of MUC5B 14 , indicating that regulation mucins expression is complex. Kaempferol, was selected as a drug to reverse the FoxA2 binding-inhibition of Pneumocystis because kaempferol is a specific inhibitor of JAK3, which phosphorylates and activates STAT6 5,10,30 . We first documented increasing P-STAT6 levels by Pneumocystis activation and reversal by Kaempferol (Fig. 4), and then, the stimulatory effect of Pneumocystis in Muc5b secretion and reversal by Kaempferol (Fig. 5). However, the limited potential for Muc5b pharmacological modulation via the STAT6/FoxA2 pathway was confirmed in these animals by the partial reversal of the binding of FoxA2 to Muc5b promoter associated to Pneumocystis by the addition of Kaempferol (Fig. 6). Therefore, documenting that Pneumocystis stimulation of Muc5b via the STAT6/FoxA2 pathway, occurs by inhibition of the FoxA2 repressor and, can be partly reverted by Kaempferol. Other potential pharmacological intervention that has been suggested to modulate the mucus hypersecretion associated to Pneumocystis in infants and patients with COPD, is to increase methylation of histone 3 which is known to suppress STAT6 35 suggesting epigenetic modifications also deserve exploration 35,36 . Modulation of mucus is complex and additional mucogenic pathways like EGFR and NFkB may be involved. For example, STAT6 and NFkB pathways that may interact via transcription factors P-STAT6 and NFkB subunits p50 and p65 37 or through a non-canonical pathway where IL13 stimulation has been documented to activate NFkB 38 . In summary, this work documents a relevant mucus response to Pneumocystis primary infection in the immunocompetent host. The progression of individual gel forming mucins over time showed an earlier and higher increase of Muc5b as part of the innate defensive response, and a slower increase of Muc5ac that suggests an ongoing Th2-type immune response that is more related to airway sensitization against this fungus occurring during the primary infection. Pneumocystis sensitizes the airway to strikingly potent constrictive responses 39 . Results show also, that the STAT6/FOXA2 pathway is involved in the positive regulation of mucus during Pneumocystis primary infection, and that pharmacologic modulation of this pathway may be a suitable strategy to limit Pneumocystis-induced mucus overproduction.

Methods
Infant lung samples. Infant autopsies were legally required as per Chilean law for all infants dying in the community, and conducted at the Servicio Médico Legal de Santiago, which is the coroner's office institution for the Metropolitan Area of Chile. Autopsy lung samples from infants without an ascertainable cause of death and categorized as Sudden Unexpected Infant Deaths (SUID) were selected using a randomization protocol available at www.randomizer.org from a pool of 39 Pneumocystis-positive and 20 Pneumocystis-negative samples stored at -80 °C in our laboratory without possible identifiers to link their identity 19 . Samples were numbered 1 to 39 in Pneumocystis-positive group and 1 to 20 in Pneumocystis-negative group and eight positive and eight negative samples were selected.
Rat model of Pneumocystis Primary Infection. Sprague Dawley rats from same colony were used for each individual experiment. Timed-pregnant rats were randomly assigned to be exposed or non-exposed to Pneumocystis as described 26 and received tylosin tartrate (93 mg/l) during 3 weeks prior to delivery to prevent bacterial infections. Control pregnant rats received cotrimoxazole (TMS) in addition to secure prevention of Pneumocystis. Dams and pups of the exposed groups were co-housed at birth with seeding rats with Pneumocystis pneumonia as described 26 . Female pups were selected at weaning. However, the effect of gender in Pneumocystis primary infection has not been studied. They were sacrificed at 45, 60 and 75 days from birth under deep anesthesia with Xylazine 10 mg/kg and Ketamine 100 mg/kg. Half of animals at each sacrifice were exsanguinated and their lungs kept at -80 °C until processed to extract RNA and proteins. The other half were perfused via cava vein with 3.7% PBS-buffered formalin and removed after 24 hours as described 26 . Administration and evaluation of Kaempferol effect. The same model 26 was reproduced to evaluate pharmacologic modulation of STAT6/FoxA2 pathway with Kaempferol at day 70. Pneumocystis-positive and Pneumocystis-negative rat groups were each separated in two groups to start Kaempferol 7 mg/Kg or saline daily intra-peritoneal injections for two weeks from days 55-70. All animals were sacrificed at day 70 for mucus and STAT6/FoxA2 pathway markers determinations.
Detection and quantification of Pneumocystis. Total genomic DNA was isolated from fresh-frozen homogenized lung tissue 0.3 g aliquots using QIAmp DNA mini kit (Qiagen). Pneumocystis was identified by nPCR using P. carinii specific oligonucleotide primers as described 40 . Pneumocystis burden was determined by quantification dhfr gene using qPCR with primers as described 26 . Copies of the dhfr gene were determined by interpolation in a calibration curve. PCR reactions conditions utilized an initial denaturation step of 5 minutes at 94 °C, 45 cycles of 20 seconds at 94 °C, 20 seconds at 57 °C, and 20 seconds at 72 °C.

Mucus determinations.
Paraffin embedded lungs sections (3 μm) were stained with Alcian Blue (1% AB in 3% glacial acetic acid, pH = 2.5) and periodic acid-Schiff 's reagent (AB/PAS) using standard protocols. All images were acquired using an Olympus BX60 microscope connected to a Q-IMAGING Micropublisher 3.3 RTV camera (QImaging, Burnaby, BC, Canada). Immunohistochemical staining was done as per manufacturer protocols sc-2018 using anti-MUC5B primary antibody sc-135508 (Santa Cruz Biotechnology, USA) and anti-Muc5ac ab 64259 kit (Abcam, UK). Muc5ac and Muc5b immunohistochemical quantifications were done by estimating the percentage of airway epithelium area stained by each respective mucin antibody using Image Pro Plus software version 5.1.0 (Media Cybernetics Inc., Rockville, MD, USA) in an average of 16 optical fields (3-35 range) per animal using a 10x objective (100x magnification). Each mucin stained area per animal represents the mean percent airway epithelial stained area value of the individual airways measured in that animal. Means per animal were then used to calculate the group mean. Airways were classified according to diameter in bronchi (>250 μm), bronchioles (50 to 250 μm) and terminal bronchioles (<50 μm) 31 .
Muc5ac and Muc5b qRT-PCR determinations in rat lungs. Lungs were placed in RNA later (QIAGEN, USA) immediately after sacrifice, kept for one night at 4 °C and then frozen at -80 °C until total RNA extraction. Total RNA extraction was performed according QIAGEN RNeasy mini spin columns method, starting with 30 mg of tissue. Total RNA (2 μg) was reversely transcribed into cDNA using random hexamer primers and SuperScript IV (Invitrogen, USA). The resulting one-stranded cDNA was diluted 1:2 and 1 μl used for PCR reactions. Amplification of Muc5b and Muc5ac was performed using the SensiMix SYBR Hi-ROX Kit (Bioline, UK) using a RotorGene 6000 Series equipment (Corbett Research, Canada). Primer sets used for amplification were: 20 seconds at 60 °C, 20 seconds at 72 °C. Actin was used as internal amplification control 26 . mRNA quantifications were normalized by actin. Results were compared using the corresponding control determination at each sacrifice point as the reference value.
Western blotting determinations in human and rat lungs. 30  Chromatin immunoprecipitation in rat lungs. 30  Statistics. Sample size for animal experiments was estimated considering a minimum of 60% difference in means per group with 95% potency for all comparisons using online tool available at http://oep.umh. es/calculo-del-tamano-muestral/. GraphPad Prism 6.0 software (San Diego, CA, USA) was used to analyze the means ± Standard Deviations (SD) of qRT-PCR and immunoblotting determinations by non-parametric Mann-Whitney test, assuming non-normal distribution and give sample size < 10 individual determinations. Two-way ANOVA was used for comparisons in sequential determinations of mucins by qPCR and protein levels by western blot, with multiple comparisons using Bonferroni analysis. Differences were considered significant when P < 0.05. Original data is provided as supplementary material.