Erythromycin reduces nasal inflammation by inhibiting immunoglobulin production, attenuating mucus secretion, and modulating cytokine expression

Allergic rhinitis (AR) and chronic rhinosinusitis (CRS) share some similar pathological mechanisms. In current study, we intend to investigate the impact of AR on CRS. In addition, we explored the efficacy of erythromycin (EM) treatment on CRS mice with or without AR (CRSwoAR, CRSwAR). Study subjects were divided into control, CRSwoAR, and CRSwAR groups. Experimental mice were divided similarly into control, CRSwoAR, and CRSwAR groups. In addition, CRS mice were treated with EM at 0.75, 7.5, or 75 mg/kg or with dexamethasone (Dex) at 1 mg/kg. In our results, allergy exacerbates inflammation that was evident in nasal histology and cytokine expression both in patients and in mice with CRS. Dex 1 mg/kg, EM 7.5 or 75 mg/kg treatments significantly inhibited serum IgE and IgG2a in CRS mice. EM-treated CRS mice had significantly elevated IL-10 levels and had a reversal of Th-1/Th-2 cytokine expression in nasal-associated lymphoid tissue. MUC5AC expressions were significantly reduced in the 7.5 or 75 mg/kg EM-treated mice compared with untreated mice. EM showed inhibitions on immunoglobulin production and mucus secretion stronger than Dex. We concluded that comorbid AR enhanced inflammation of CRS. EM and Dex treatments showed similar anti-inflammatory effects on CRS but through partly different mechanisms.

www.nature.com/scientificreports/ pathophysiologic finding of AR 10,11 . Defects in the sinonasal epithelial barrier, increased exposures to pathogens, dysregulation of the host immune system and mucociliary clearance all of which are considered important in the pathophysiology of chronic rhinosinusitis 12,13 . Underlying inflammatory mechanisms are no doubt the main etiology of CRS 14,15 . Recently, researchers recommended new classification of CRS patients based on inflammatory patterns (endotype) rather than the clinical presentation (phenotypes) 16,17 . Bachert et al. proposed the CRS classification according to Th-cell populations and Th-related cytokines 17,18 . Update guidelines for rhinosinusitis recommend utilizing the Th-1 (IFN-γ, TNF-α), Th-2 (IL-4, -5, 13, ECP), and Th-3 (IL-17A) biomarkers to defining subtypes of CRS 1,2 . Patients of CRS comorbid with allergy display immune responses comparable to those underlied by Th-2 inflammation.
Nasal steroid and long-term macrolide are current standard treatment for CRS 1 . Steroid is well-known for its broad-spectrum anti-inflammatory effects but the effects of macrolides on CRS remained uncertain. Macrolides treatment experiences come from the dramatic lifesaving outcomes from panbronchiolitis 19 , then extend their use to many chronic airway diseases. Despite the wide use of macrolides in treating CRS, updated rhinosinusitis guidelines recommended more studies are needed to clarify their effects 1,2 . The efficacy of macrolides might come from both antibiotic and non-antibiotic effects. Mechanisms proposed on the non-antibiotic effects of macrolides on CRS include the inhibition of biofilm formation, enhancement of mucociliary clearance, modulation of cytokine production, and promotion of neutrophil apoptosis 20,21 . Systematic reviews and meta-analyses reported macrolides are beneficial only on some but not all CRS patients 5,22 . Seresirikachorn et al. 22 concluded that macrolides are beneficial in treating patients with CRS without nasal polyps as opposed to CRS with nasal polyps. In the study by Perić et al. 23 nasal polyposis patients with or without AR received clarithromycin for 8 weeks. They found that immunomodulatory effects on cytokines were different between the two groups. The purpose of this study was to investigate the impact of AR on pathophysiology of CRS. In addition, we determined the efficacy of the erythromycin (EM), a 14-membered macrolide, on CRS mouse models with or without AR.

Methods
Rhinosinusitis subjects. Sinonasal tissues of CRS were obtained from patients after endoscopic sinus surgery. Control tissues were from patients without rhinosinusitis but had undergone septum or turbinate surgery. The CRS diagnosis was based on typical symptoms over 12 weeks and the results of endoscopy and computed tomography (CT), in line with the criteria of European Position Paper on Rhinosinusitis and Nasal Polyps 2020 1 . Enrolled subjects underwent sinus surgery due to medical treatment failure. All CRS patients received CT before surgery. CT images were graded according to the Lund and Mackay staging system with a total score range from 0 to 24 24 . Patients were diagnosed with AR according to specific IgE tests and clinical history for at least 2 years on having allergic symptoms when contacted with allergens and irritants. Enrolled subjects were divided into three groups: control, CRS without allergic rhinitis (CRSwoAR), and CRS with allergic rhinitis (CRSwAR) groups. In this study, we did not recruit asthmatic patients.
Histologic characteristics of human sinonasal tissues. Tissues were fixed with buffered formalin, embedded in paraffin, and cut into slices. Hematoxylin and eosin (H&E) and immunochemistry (IHC) stainings were performed to evaluate histological features, changes in protein expression at tight and adherens junctions, and mucus production. The primary antibodies for evaluating the epithelium inter-cellular junction were claudin-1(Catalog No. 51-9000, polyclonal antibody, Thermo Fisher, IL, USA) and e-cadherin (catalog No. GTX100443, GeneTex, Irvine, CA, USA). Periodic acid-Schiff (PAS) staining (ScyTek Laboratories, Inc., Logan, UT, USA) was used to evaluate mucus production.
Allergic rhinitis and chronic rhinosinusitis mouse models. Female BALB/c mice at 6-8 weeks of age were divided into 3 groups: control, CRSwoAR, and CRSwAR. On days 0, 7, and 14, AR mice were sensitized each with an intraperitoneal injection of 4 μg house dust mite (HDM, Indoor Biotechnologies Ltd, Cardiff, UK) mixed with 40 μg aluminum hydroxide gel adjuvant (Invitrogen, San Diego, CA). From days 22 to 26, animals received intranasal challenges with 4 μg HDM. Non-AR mice received sham sensitization. Mice underwent nasal surgery to induce chronic rhinosinusitis. In brief, under anesthesia, a 1-cm incision was made over the mouse head, and a 3 mm hole was drilled on the skull to reach nasal cavity on one side. Gelfoam pledgets (Johnson www.nature.com/scientificreports/ & Johnson, Gargrave, Skipton, UK) were inserted into the nasal cavity to induce local inflammation. The scalp wounds were approximated with sutures at the end of the procedure. With the protocols, experimental AR and CRS could be reliably induced with evidence of histology and biomarkers in our previous study 25 . Mice were sacrificed on day 27. Mouse rhinosinusitis lasting > 4 weeks was regarded, in the literature, as a chronic disease model 26 .

Statistical analyses.
Chi-Square test was used to compare the categorical data. Mann-Whitney U test was used to compare continuous data between two groups. Kruskal-Wallis test was used to compare 3 or more groups. Then Dunn's multiple comparisons test was used for post hoc examination of between-group differences. Statistically significance was set at P < 0.05. Statistical analyses were performed using GraphPad Prism 9.2.0 for Mac (GraphPad Software, San Diego, CA, USA) and IBM SPSS 20.0 (IBM Corp. Armonk, NY, USA).
Ethics. The Institutional Review Board of Taichung Veterans General Hospital approved the present study on human subjects (CE17143A). The research was performed in accordance with relevant human study guidelines and regulations. Inform consent was obtained from each study subject.
The animal use protocol and experimental procedures have been reviewed and approved by the Institutional Animal Care and Use Committee of Taichung Veterans General Hospital (La-1071579). All experiments were performed following the recommendations of guidelines. Current study is reported in accordance with ARRIVE guidelines 28 .

Results
Human study. Their clinical characteristics of CRS patients are listed in Table 1. Sinonasal tissues from the CRSwAR group showed higher mucosal eosinophilia and globlet cell hyperplasia when compared with the other two groups. In addition, in nasal epithelia of the CRSwAR group immunochemistry stainings to of claudin-1 and e-cadherin were less compared with the other two groups. Representative images of H&E and IHC stainings are shown in Fig. 1.
Effect of erythromycin or dexamethasone treatment on experimental mice. After treatments with 7.5 or 75 mg/kg EM, or with Dex, we found reduced cellular infiltrations in nasal histology of CRS mice (with or without AR). Marked reduction in mucus production was especially noted in CRS mice receiving EM treatments. In addition, we found stronger e-cadherin stainings in nasal tissues of EM-treated mice when compared with untreated mice (Fig. 3A,B). CRSwoAR mice treated with 75 mg/kg EM had significantly reduced mucosal eosinophilia as compared to untreated CRSwoAR mice (P = 0.0439). Dex-treated CRSwAR mice had significantly reduced mucosal eosinophilia than untreated CRSwAR mice as well (P = 0.0105) (Supplementary Tables S3 & S4; Fig. 4A).
CRSwoAR mice had significantly lower serum levels of IgG2a after treatment with 7.5 or 75 mg/kg EM (both P < 0.0001) or after 1 mg/kg Dex (P = 0.0229) when compared with untreated mice (Supplementary Table S7). CRSwAR mice treated with 75 mg/kg EM also significantly lowered their IgG2a levels when compared with untreated mice (P = 0.0272). In addition, 7.5 and 75 mg/kg EM-treated CRSwAR mice showed significantly lower serum levels of IgG2a compared with Dex-treated CRSwAR mice (P = 0.0167 and = 0.0075, respectively) (Supplementary Table S8; Fig. 4B-D). CRSwAR mice treated with 7.5 mg/kg EM or 1 mg/kg Dex significantly lowered their serum levels of total IgE when compared with the untreated CRSwAR mice (Supplementary  Table S8, P = 0.0141 and = 0.0433, respectively).
The western blotting results are shown in Fig. 6 and Supplementary Tables S11 & S12. MUC5AC protein expressions were significantly reduced in the 7.5 or 75 mg/kg EM-treated CRSwoAR mice compared with untreated mice (P = 0.0003 and = 0.0032, respectively). Similarly, CRSwAR mice receiving 7.5 or 75 mg/kg EM treatment had significantly lower nasal MUC5AC protein expressions when compared with untreated mice (P < 0.0001 and = 0.0134, respectively). In addition, we found significant differences in MUC5AC protein expressions between 7.5 mg/kg EM-treated and Dex-treated CRSwAR mice (P = 0.0190). Both CRSwoAR and CRSwAR mice showed higher e-cadherin expressions in nasal tissues after treatments with 7.5 or 75 mg/kg EM. No significant difference was found between EM-treated and untreated groups, but EM-treated CRS mice had significantly higher e-cadherin expression than Dex-treated mice (CRSwoAR: EM 75 mg/kg vs. Dex, P = 0.0293; CRSwAR, EM 75 mg/kg vs. Dex, P = 0.0105). Regarding claudin-1, there were no significant differences among the experimental groups. In brief, both erythromycin and dexamethasone treatment reduced nasal inflammation in treated mice, but through partly different mechanisms.

Discussion
Our CRS subjects, with or without AR, showed similar CT grading scores and incidence of polyposis. Nevertheless, CRSwAR patients showed in their nasal tissues, signs of stronger mucosal inflammation, higher mucus production, higher Th-2 cytokines expression, and poorer epithelial tight junction protein expression when compared with those without AR. Xing et al. 29 conducted a study investigating the effect of AR on nasal mucosa remodeling with nasal tissues taken from CRSwoAR and CRSwAR patients. They concluded that AR could enhances tissue remodeling process in CRS. Our results are consistent with CRS endotype being a better predictor than phenotype on disease severity. Based on the underlying inflammatory patterns, the classification of CRS patients with endotype could choose precise therapeutic strategy better for patients, and hence a better prediction of treatment outcome 30 .
Both EM and Dex effectively reduced serum levels of IgE in allergic mice. Furthermore, treatments both reduced serum IgG2a production in either CRSwoAR or CRSwAR mice. Only two randomized controlled studies   32 did not found such advantage in their similar patients. Eosinophilic inflammation is a major hallmark of AR and CRS with nasal polyps. In CRSwoAR mice, EM treatment at a dose of 75 mg/kg demonstrated an inhibitory effect of tissue eosinophilia. Consistent with our present results, macrolides could have a role in treating IgE-medicated allergic airway diseases despite reported studies favored benefits on low-IgE CRS subjects. We had conducted a study comparing the effects of intranasal steroid or EM towards rhinosinusitis patients 33 . Results of aforementioned study demonstrated that the improvement in mucociliary function was higher in the EM group than in the intranasal steroid group. In current study, reduction of mucus production after treatment of EM was significantly stronger than Dex. Disruption of epithelial barrier junction can enhance the infiltration of pathogens or allergens into the submucosal area, stimulating airway inflammation 10 35 . In our current study, we found a tendency of reduced e-cadherin expression in CRS subjects and mice. However, the effect of EM on restoring nasal epithelial barrier dysfunction was not evident in our CRS mice.
Cytokines and chemokines are key regulators of the inflammatory responses. Previous studies of macrolides on CRS patients revealed no consistent effects on cytokine productions. Wallwork et al. 31 conducted a trial of roxithromycin on CRS patients and found a reduction of IL-8 in nasal lavage after treatment. The anti-inflammatory function of macrolides is believed coming from both antibiotic and non-antibacterial effect. Macrolide treatment efficacy has shown in patients infected with pathogens insensitive to macrolides like pseudomonas aeruginosa 8 . Sadamatsu et al. 36 treated asthmatic mice with a macrolide EM900 (a derivative of EM without antibacterial effects) and found significantly lower levels of IL-5, IL-13, RANTES, and IL-17A in their bronchoalveolar lavage. EM 900 also demonstrated inhibitory effect on IL-8 on culture human nasal epithelial cells in another in vitro study 37 . Pukhyalsky et al. 38 studied patients with cystic fibrosis after prolonged treatments of clarithromycin. They found significant improvements of lung function and a switch from Th-2 to Th-1 cytokines in their blood and sputum. Interestingly, Park et al. 39 found in patients with panbronchiolitis, a shift from Th-1 to Th-2 cytokine www.nature.com/scientificreports/ production in bronchoalveolar lavage fluid after long-term treatment with EM. The Th-1/Th-2 cytokines switches observed in the two human studies could reflect an immune-modulatory effect of macrolides. In current study, we also found the immunomodulatory effect in the EM-treated mice. Clinical studies of macrolides on CRS involved application of various drugs and at different dosages. But the optimal drug, dosage, or duration are not currently known. Subgroup analyses of Seresirikachorn's meta-analysis showed the results favor patients receiving macrolides at a half of regular antibiotics dose, and favor macrolide treatment for a duration of 24 weeks instead of shorter periods 22 . In our current study, we found that medium and high doses (7.5 and 75 mg/kg) worked well for treating of CRS. No dose-response could be with only 3 doses were used in our study. A meta-analysis reported that adding a macrolide to an intranasal steroid may achieve better results than using steroid alone to treat CRS 40 . Combined two drugs could be considered for patients unresponsive to monotherapy. In conclusion, we identified that comorbid AR exacerbates CRS severity. Our mice experiments showed Dex and EM treatments effects on CRS acted through somewhat different mechanisms.