Epithelial Contributions to CRS

The past several decades have seen a gradual but definitive shift in our view of the sinonasal epithelium from a passive barrier to an active immunologic organ with both innate and adaptive components (Figure 1). In parallel, multiple investigations have demonstrated that the inflammatory profiles associated with CRS are often predicated upon the dysfunctional regulation of these complex mechanisms with both downstream and reciprocal sequelae.

Open in a separate window

Figure 1.

Epithelial Contributions to CRS

Extrinsic pathogens and irritants interact with multiple epithelial receptors to induce both innate and adaptive immune responses. Protease exposure induces tight junction (TJ) dysfunction and secretion of endogenous protease inhibitors (EPIs) which neutralize extrinsic proteases and stimulate type 2 inflammation. Similarly, pathogen-associated molecular patterns (PAMPs) interact with Toll-like receptors (TLRs) leading to apical anti-microbial peptide (AMP) secretion as well as epithelial derived cytokine secretion. P-glycoprotein (P-gp) functions to clear the cytoplasm of environmental toxins while reinforcing epithelial cytokine release. Epithelial derived exosomes support these innate immune responses by shuttling AMPs directly to mucus borne pathogens while promoting the inter-epithelial transfer of P-gp. Bacterial derived quorum sensing molecules, such as acyl-homoserine lactone (AHL) compounds, also interact with bitter taste receptors such as T2R38 to induce bacteriocidal nitric oxide (NO) release, enhanced mucociliary clearance (MCC), and cytokine/chemokine secretion.

Inflammatory endotypes in CRS

Recent studies suggest that inflammation in CRS is highly heterogeneous and can be separated by three main inflammatory endotypes: the T1 endotype by the elevation of T1 cytokine IFN-γ, the T2 endotype by eosinophilia and the elevation of T2 cytokines, and the T3 endotype by neutrophilia and elevation of T3 cytokines including IL-17A. A recent study elucidated associations between endotypes and clinical presentations and found some associations in that the T2 endotype is associated with smell loss, asthma comorbidity and nasal polyposis, whereas the T3 endotype is associated with presence of intraoperative pus suggesting a link with infection.⁹ However, most endotype associated studies only separated two main endotypes, eosinophilic (T2) and non-eosinophilic (non-T2) that contains several sub-endotypes including T1 and T3 (neutrophilia), until recently. Indeed, earlier studies compared these two groups in CRS.^88–92 Based on this limitation, we mainly describe T2- and non-T2-mmmediated inflammation in CRS.

Inflammation in CRSwNP

The T2 endotype is a common endotype in CRSwNP in Western countries.^{3, 7–9} In addition, the prevalence of T2 CRSwNP has been increasing over the past decades in Asia similar to other allergic diseases such as asthma and allergic rhinitis.^{3, 7, 93–96} There is the geographic variability even within a single country and indeed the presence of T2 CRSwNP was very different between Beijing (60%) and Chengdu (20%) in China.⁷ This implies that inflammatory endotype is controlled by microenvironments more than genetic factors.

T2 CRSwNP

Type 2 cytokines including IL-4, IL-5 and IL-13 produced by Th2 cells, ILC2s and mast cells play important roles in T2 CRSwNP (Figure 2). IL-5 primarily is involved in eosinophilia and also participates in activation of plasma cells in NPs. Interleukin-4 and IL-13 are key factors that control many important factors in NPs including IgE responses, epithelial barrier dysfunction, mucus production, remodeling and fibrin deposition.³ Treatment with an anti-IL-4Rα antibody (dupilumab) that inhibits both IL-4 and IL-13 signaling reduced NP size, improved quality of life and symptoms including nasal congestion and anosmia in severe CRSwNP patients.⁹⁷ Several groups performed transcriptomic approaches on CRSwNP in Western countries (mostly T2) and eosinophilic (T2) CRSwNP in Asia, identified up-regulated genes in T2 NPs and predicted immunological and inflammatory mechanisms in CRSwNP more globally.^{91, 92, 98, 99} The up-regulated genes in T2 NPs included cell markers and chemokines for eosinophils (e.g. CLC, CCR3, CCL13 and CCL26), macrophages (especially M2 phenotype)/myeloid DCs (mDCs) (e.g. CCR1, CD163, F13A1, CCL2 and CCL18), Th2/ILC2 (e.g. IL1RL1, IL17RB and PTGDR2), mast cells (e.g. CPA3 and TPSAB1), epidermal differentiation (e.g. KRT6A and SPRR1B) and extracellular matrix (e.g. FN1 and FGF1), and other common IL-4/IL-13-induced genes (e.g. CST1 [cystatin SN] and POSTN [periostin]).^{91, 92, 98, 99} Nakayama et al. also directly compared NPs from Asian (residing in Japan) and Caucasian (residing in the US) CRSwNP patients by RNA-Sequencing and found that the gene expression profile in T2 NPs from both groups showed near-complete overlap.⁹⁹ This suggests that key molecular mechanisms of T2 NPs in Asian and Western countries are probably the same. Since immune and inflammatory cells and their factors play central roles in T2 CRSwNP, we will further introduce the presence and role of each cell type in this section.

Open in a separate window

Figure 2.

Potential mechanisms in T2 CRSwNP

Pathogens and environmental factors activate nasal epithelial cells inducing an innate cytokine TSLP. Some countries also found elevation of IL-25 and IL-33. TSLP and PAMPs stimulate DCs to induce naive CD4⁺ T cell differentiation into Th2 cells. These epithelial derived innate cytokines also activate ILC2s, pathogenic Th2 cells and mast cells to produce T2 cytokine. Th2 cells also activate ILC2s to induce T2 cytokines via a RANK-L dependent pathway. IL-4 and IL-13 activate epithelial cells, endothelial cells, macrophages and B cells to induce barrier dysfunction, mucus response, eosinophil recruitment, and IgE-mediated reaction. Epithelial dysfunction is also controlled by activation of eosinophils, neutrophils and complement. Antigen/IgE/IgER complexes on mast cells and basophils induce degranulation and release prestored molecules including histamine and enzymes that induce vascular leak and tissue damage. Plasma leak triggers fibrin deposition in the NPs via crosslinking by FXIIIA released from M2 macrophages and reduction of tPA in epithelial cells. Accumulation of IgG and autoantibodies induce the activation of complements and neutrophils.

ILC2s and Th2 cells

Antigen-independent T2 inflammation is mainly controlled by ILC2s. ILC2s are known to be highly elevated in Western T2 NPs.^100–102 In Asia, ILC2s are elevated in T2 NPs but not in non-T2 NPs.¹⁰³ Poposki et al. found that FACS sorted NP ILC2s but not peripheral blood ILC2s spontaneously released IL-5 and IL-13, suggesting that ILC2s in NPs are highly activated and release T2 cytokines in vivo.¹⁰¹ ILC2s express receptors responsive to epithelial cell derived innate T2 inducers, TSLP, IL-25 and IL-33. In the case of CRS, elevated TSLP in T2 NP tissue was found across the world.^{104, 105} TSLP is induced by viruses, protease containing allergens and T2 cytokines IL-4 and IL-13 in epithelial cells, and the combination of these stimuli synergistically or additively enhanced TSLP production.^{69, 104, 106} TSLP activity in NP tissue was enhanced by the post-translational modifications from tissue proteases, and these active TSLP metabolites stimulate ILC2s as well as DCs and mast cells much more potently than mature TSLP.^{104, 107} Importantly, anti-TSLP (tezepelumab) has been approved for treatment with asthma and is in clinical trial for CRSwNP.¹⁰⁸ In contrast, elevation of other epithelial derived innate T2 inducers, IL-25 and IL-33, is controversial. Elevation of IL-25 in T2 NPs was found mainly in Asia and a US study also demonstrated that IL-25 was detected in a minor subset of epithelial cells called solitary chemosensory cells and was elevated in NPs.^{105, 109} In contrast, the other US study showed almost undetectable levels of IL-25 in NPs.¹⁰⁵ This may suggest that IL-25 plays a role in a limited area of NPs or on an early event in NP pathogenesis. At IL-33, mixed results were reported in T2 NPs throughout the world.¹⁰⁵ However, since IL-33 requires two steps to obtain the cytokine activity: release from the nucleus and cleavage by enzymes, we cannot draw conclusions based only on the expression of IL-33.^{110, 111} Detailed studies examining whether IL-25 and IL-33 play a role in the pathogenesis of T2 CRSwNP are still required.

It is implicated that activation of four distinct transcriptional pathways (NF-κB, NFAT, STAT5 and STAT6) is needed for the most robust generation of T2 cytokines in ILC2s.¹¹² Published studies suggested that activators of NF-κB (receptor activator of NF-κB ligand [RANK-L], IL-25 in Asia and IL-33 in several studies), NFAT (prostaglandin D2 [PGD2], cysteinyl leukotriene [LT] C4 and LTD4), STAT5 (TSLP) and STAT6 (IL-13) are all elevated in NPs from T2 CRSwNP.^{105, 112–114} This suggests that NP tissue has a potent microenvironment that activates ILC2s to robustly produce T2 cytokines.

Antigen-dependent T2 inflammation is mainly controlled by Th2 cells. Th2 cells are differentiated from naive CD4⁺ T cells through activation of T cell receptor-mediated signaling, co-stimulation and STAT6 activation by cytokine signals (IL-4 and IL-13), and GATA3 is a key transcription factor that controls differentiation of Th2 cells.^{115, 116} TSLP also participates in DC-mediated Th2 differentiation via the induction of OX40L on DCs and OX40L⁺ DCs are elevated in T2 NPs.^{117, 118} Several groups identified the elevation of Th2 cells in T2 NPs.^{4, 80, 118–120}. Th2 cells in NPs also express RANK-L that activates ILC2s to produce T2 cytokines.¹¹³ Lam et al. and Ma et al. characterized Th2 cells in T2 NPs by bulk and single cell RNA-Seq, respectively, and found that these Th2 cells expressed GATA3, IL-17RB, IL-1RL1 (also known as ST2), HPGDS, and T2 cytokines.^{120, 121} This suggests that elevated Th2 cells in T2 NPs display a pathogenic Th2 cell phenotype.^122–124 In addition, scRNA-Seq analysis also found heterogeneity of Th2 cells in NPs and showed that the other Th2 subpopulation expressed CD109 but lacked CRTH2. Interestingly, CD109⁺CRTH2⁻ Th2 cells released not only classical T2 cytokines IL-4, IL-5 and IL-13 but also produced immunosuppressive IL-10 by the having a transcription factor FOXP3 in addition to GATA3.¹²¹ They suggested that CD109⁺CRTH2⁻ Th2 cells might be Tr1 cells that have been primed by T2 microenvironments. Future study will be required to understand the role of CD109⁺CRTH2⁻ Th2 cells in NPs.

B cells and immunoglobulins

Accumulation of B lineage cells (B cells, plasmablasts and plasma cells) and elevated local production of immunoglobulins (IgA, IgG and IgE) are now believed to play a pathogenic role in CRSwNP.^{4, 90, 125, 126} Accumulation of B-lineage cells in NPs was associated with local production of BAFF.¹²⁷ Zhang et al. compared immunoglobulin isotypes between T2 and non-T2 NPs in China and found that only IgE was significantly elevated in T2 NPs compared to non-T2 NPs.⁹⁰ Elevated IgE plays a pathogenic role in T2 CRSwNP via the activation of mast cells and basophils and treatment with an anti-IgE antibody (omalizumab) reduced polyp size and improved symptoms in patients with CRSwNP.¹²⁸ IgE antibodies against enterotoxins from Staphylococcus aureus were well discussed regarding the pathogenic roles and biomarkers in severe CRSwNP.¹²⁹ High tissue IgE levels also associated with NP recurrence.¹³⁰ Although IgG4 was similarly elevated in T2 and non-T2 NPs,⁹⁰ IgG4 was more elevated in severe T2 CRS.¹³¹ Buchheit et al. found that IgG4 and IgE were further elevated in NPs from patients with AERD (aspirin-exacerbated respiratory disease) which is a respiratory syndrome consisting of severe T2 CRSwNP, asthma and an intolerance to inhibitors of cyclooxygenase 1 compared to NPs from patients with aspirin-tolerant CRSwNP in the US.¹³⁰ Although whether IgG4 acts on pathogenic or protective roles in NPs is still unclear, IgG4 may be used as a biomarker to predict severe T2 CRSwNP. Interestingly, Buchheit et al. also found IL-5Ra expression on plasma cells in AERD NPs and IL-5 activated plasma cells.¹³⁰ This suggests that biologics targeting IL-5 and IL-5Ra not only inhibit eosinophils but also suppress antibody secreting plasma cells to improve patient outcomes in AERD. Local accumulation of several autoreactive IgG and IgA in NPs such as those targeting double stranded DNA and basement membrane components has also been reported.^132–134 Levels of autoreactive antibodies in NPs were associated with eosinophil levels and disease recurrence indicating that elevation of autoreactive antibodies is associated with the T2 endotype. Finally, local accumulation of IgG induces activation of the classical complement pathway in NPs.¹³³ A recent study by Wang et al. showed that PD1^hiCXCR5⁻CD4⁺ cells are atypical T cells found in NPs in China, and that these cells purified from T2 NPs potently activate B cell production of all immunoglobulin isotypes in vitro.¹³⁵

Macrophages

Macrophages are shifted to the M2 phenotype and are elevated in T2 NPs.^{136, 137} Elevated M2 macrophages in NPs demonstrated impaired phagocytic activity against Staphylococcus aureus and this may link staph colonization that is frequently found in T2 CRSwNP.¹³⁶ M2 macrophages are a very important producer of factor XIII-A that can induce fibrin deposition by cross-linking fibrin that is involved in NP formation (see below).¹³⁸ M2 macrophages also produce CCL13 and CCL18, chemokines that recruit eosinophils and immature mDCs, respectively.^{99 , 137}

Mast cells and basophils

Mast cells and basophils play key roles in host defense and T2 allergic inflammation.¹³⁹ The primary stimulus for mast cells and basophils in T2 inflammation is activation through the high-affinity IgE receptor, FcεRI. Upon cross-linking of FcεRI via IgE-antigen, they release several pre-stored mediators, including histamine and proteases, and synthesize lipid mediators including CysLTs and PGD2.¹³⁹ They also produce T2 cytokines; in general mast cells produce IL-5 and IL-13 whereas basophils produce mainly IL-4. Mast cells also respond to cytokines including epithelial-derived IL-33 and TSLP and macrophage/neutrophil-derived IL-1β to produce IL-5 and IL-13.^{3, 140} Mast cells are frequently divided into two major phenotypes: mast cell-tryptase (MC_T) and mast cell-tryptase/chymase (MC_TC). Takabayashi et al. found that MC_T were elevated in mucosal epithelium and MC_TC were elevated in glandular epithelium in the US.¹⁴¹ Cao et al. found that activated mast cells were elevated in NPs from T2 CRSwNP in China.¹⁴² Dwyer et al. performed scRNA-Seq in NP mast cells and identified two more mast cell subsets: intermediate MC and proliferative MC (Ki67⁺). Within these 4 mast cell populations, MC_TC and intermediate MC showed enhanced expression of chemokines, proinflammatory cytokines, the T2 cytokine IL-13 and a key lipid mediator biosynthetic enzyme PTGS2, suggesting that these mast cell populations are activated and play proinflammatory roles in T2 NPs.¹⁴³ It is also important to note that MC_TC in NPs are phenotypically different from MC_TC in skin and lung. For example, although dermal MC_TC express C5AR1 and Mas-related G-protein coupled receptor X2 (MRGPRX2) and respond to C5a and neuropeptides, respectively, NP MC_TC do not have these receptors.¹⁴³ Although elevated MC_T in NPs seem to be activated by IL-4 and/or IL-13, their direct role in NP pathogenesis is still unclear.

In the case of basophils, Stevens et al. found that basophils were significantly elevated in NPs from CRSwNP and were further elevated in AERD NPs.¹⁴⁴ Importantly, basophils in AERD NPs were more activated and degranulated and levels of these markers in basophils correlated with sinonasal and pulmonary disease severity.¹⁴⁴ This suggests that basophils also play a role in the pathogenesis of T2 CRSwNP, especially in severe cases.

Eosinophils and neutrophils

Eosinophilia is the central feature in T2 CRSwNP around the world and biologics targeting eosinophils by anti-IL-5 (Mepolizumab and Reslizumab) and anti-IL-5 receptor (Benralizumab) have been approved or are in clinical trials for treatment of severe T2 CRSwNP patients.¹⁴⁵ Accumulation of eosinophils in T2 CRSwNP, especially in NPs, is controlled by local elevation of adhesion molecules (VCAM-1 and ICAM-1) induced by IL-4 and IL-13, recruitment factors (PGD2, cysteinyl LTs and chemokines including CCL13 [MCP4] and CCL26 [eotaxin-3]), as well as survival and activation factors (IL-5 and GM-CSF).¹⁴⁵ Eosinophils in NPs presented increased expression of CD69, which is a marker of activation.¹⁴⁶ Upon activation, eosinophils release pre-stored cytotoxic granule proteins including major basic protein (MBP), eosinophil cationic protein (ECP), eosinophil-derived neurotoxin (EDN) and eosinophil peroxidase (EPX) that contribute to tissue damage and remodeling.^{145, 147} Activated eosinophils can also exhibit extracellular trap cell death and release EETs. EETs are formed with DNA, granule proteins and Charcot-Leyden crystals (CLCs). CLCs induce proinflammatory cytokines including IL-6, TNF and GM-CSF as well as recruitment of neutrophils.^{148, 149} CLCs also act as an adjuvant to enhance adaptive T2 immunity.¹⁴⁸ In addition, activated eosinophils synthesize and release cysteinyl LTs. Importantly, NP eosinophils have elevated levels of LTC4 synthase and γ-glutamyl transferase-5 (converts to LTD4 from LTC4) and produce more LTD4 upon activation compared to peripheral blood eosinophils.¹⁴⁶ Released cysteinyl LTs promote further eosinophil recruitment, mucus secretion, vascular permeability and activation of ILC2s. Finally, eosinophils are a key producer of a chemokine CCL23 that recruits CCR1⁺ DCs, monocytes and macrophages after undergoing post-translational modifications in NP tissue.¹⁵⁰

Although neutrophils play a central role in non-T2 CRSwNP, recent studies suggest that neutrophils also contribute to pathogenesis in T2 CRSwNP. Pothoven et al. found that the epithelial barrier-disrupting cytokine OSM was elevated in T2 NPs and that OSM was detected in neutrophils in NP tissue.^{20, 21} Delemarre et al. reported the elevation of neutrophils in T2 NPs and found that deposition of CLCs is associated with neutrophil infiltration in Belgium.¹⁵¹ Subsequently, Poposki et al. showed the elevation of activated neutrophils in T2 NPs in the US, and found that these neutrophils contributed to IL-1β production and were associated with NP recurrence.¹⁵² Succar et al. found that mixed eosinophilic and neutrophilic CRSwNP patients showed the highest disease symptom scores in the US.¹⁵³ Since the disease severity and recurrence rates are lower in the purely neutrophilic non-T2 NPs than in T2 NPs,⁹⁵ neutrophils may have different pathogenic roles between T2 and non-T2 CRSwNP.

Non-T2 CRSwNP

Although there is heterogeneity in the endotypes of non-T2 CRSwNP such as T1, T3 and neutrophilic, most mechanistic studies did not separate each sub-endotype. Therefore we are only able to summarize general non-T2 mechanisms in this review (Figure 3). T cells, B cells, macrophages and DCs were similarly elevated in both non-T2 and T2 NPs compared to control sinus mucosa; however the phenotype of these cells were different between two endotypes. Unlike T2 NPs, non-T2 NPs presented elevated levels of IFN-γ, IL-1β, IL-6, IL-8 and IL-17 and accumulation of Th1 and Th17 cells.^{5, 118} Although activated DCs were elevated in both endotypes, OX40L and PD-L1 were only elevated on DCs from T2 NPs. Macrophages are also shifted to M1 phenotype and produce IL-1β and ROS via the activation of NLRP3 inflammasome.¹⁵⁴ Transcriptome analyses also identified elevated expression of IFN-γ-induced chemokines (CXCL9, CXCL10 and CXCL11), IL-17-induced molecules (SAA, CHI3L1, CXCL5 and CCL20 [recruits DCs and Th17]) and neutrophil activators and chemokines (CSF3, CXCL5 and CXCL6)^{91, 92} suggesting that the mixed T1 and T3 endotype with neutrophilia is the common endotype in non-T2 CRSwNP. Neutrophils play an important role in the inflammation associated with non-T2 CRSwNP.¹⁵⁵ Activated neutrophils release prestored molecules including myeloperoxidase, elastase, cathepsin G, proteinase 3 and antimicrobial peptides that protect from infected bacteria and induce tissue damage and remodeling. Activated neutrophils also synthesize and produce lipid mediators (mainly LTB4), ROS and several cytokines and chemokines including OSM, IL-1β and IL-8 and contribute to inflammation and further recruitment of neutrophils. Neutrophils also induce NETosis and release neutrophil extracellular traps (NETs) and the presence of NETs and positive correlations with neutrophil counts were reported in Chinese NPs.¹⁵⁶

Open in a separate window

Figure 3.

Potential mechanisms in non-T2 CRSwNP

T1 mediated inflammation is mainly triggered by the accumulation and activation of Th1 cells, CD8⁺ T cells and NK cells and IFN-γ is a key cytokine that controls further recruitment of T1 immune cells, epithelial barrier dysfunction, fibrin deposition and development of M1 macrophages. M1 macrophages produce IL-1β and ROS to induce inflammation. T3 mediated inflammation is triggered by the accumulation and activation of Th17 cells. IL-17 induces recruitment of Th17 cells and neutrophils. Activated neutrophils release OSM, NETs, IL-1β, LTB4 enzymes and ROS to induce epithelial barrier dysfunction, inflammation and tissue damage. IgG antibodies are also elevated in non-T2 CRSwNP and may play a role on the activation of neutrophils and complement pathways.

Deposition of fibrin as the main mechanism of hypertrophy in polyps and sinuses

Work by Tetsuji Takabayashi and collaborators revealed widespread deposition of fibrin in NP tissue.^{65, 74} Earlier work by Shimizu et al. had provided evidence for the presence of active thrombin in nasal lavage fluids from patients with CRS, suggesting activation of the coagulation system.¹⁵⁷ Coagulation is activated both by the intrinsic pathway via surface contact, and the extrinsic pathway via tissue factor; both pathways are likely operative in CRSwNP tissue, and lead to activation of factor X, followed by formation of thrombin and subsequently fibrin.^{158, 159} While it can be expected that any condition or disease that causes vascular leak will promote attendant activation of coagulation and formation of extravascular fibrin clot, ordinarily tissue plasminogen activator (tPA), which is primarily produced by epithelial cells in the airways, activates plasmin and degrades the tissue fibrin clot. We now know that the extensive and tenacious deposition of fibrin in CRSwNP is partly attributable to suppression of tPA production and release, resulting in reduced expression of this major fibrinolytic pathway and impeding or preventing clot degradation.^{65, 74} Compounding the impact of impaired fibrinolytic pathways, later studies by Takabayashi et al. showed strong induction of factor XIII-A, an important enzyme that crosslinks and stabilizes fibrin. This factor is expressed by so-called alternatively activated macrophages that are found in high numbers in the tissue of NP patients.¹³⁸

Since active coagulation is occurring in the tissue of CRS patients, it is not surprising that further studies by Imoto et al. have found increased levels of two pathway members, thrombin-activatable fibrinolysis inhibitor (TAFI) and thrombin/anti-thrombin complex (TATc), in the tissue and nasal fluids of patients with CRSwNP.¹⁶⁰ Interestingly, levels of both TATc and TAFI were higher in CRSwNP patients with comorbid asthma, compared to CRSwNP patients without asthma, suggesting that activation of coagulation is correlated to severity of disease as indicated by asthma comorbidity.¹⁶⁰ Studies by Takabayashi and collaborators in Fukui demonstrated that culture of NP tissue overnight with Nattokinase, a fibrinolytic enzyme, led to large scale shrinkage of polyp tissue but not normal tissue and led them to conclude that the majority of the wet weight of the polyp tissue (approaching 90%) was attributable to trapping of water by deposition of fibrin mesh. This important study strongly suggests that polyp formation itself is to a large extent driven by deposition of large quantities of fibrin. Interestingly, in the same study, they showed that experimental degradation of fibrin in mucus led to nearly a complete loss of mucus viscosity, suggesting that crosslinked fibrin is also found in nasal mucus and endows mucus with many of its elastic and adhesive properties.¹⁶¹ The initial studies referred to above demonstrated fibrin deposition utilizing immunohistochemistry or immunofluorescence.⁷⁴ More recently, unpublished studies at Northwestern using a biochemical assay to quantitate fibrin deposition have demonstrated a 10 to 20 fold increase in fibrin protein in NP tissue compared to normal ethmoid or uncinate tissue. Interestingly, hypertrophic tissue from the ethmoid sinuses of patients with CRSsNP were also found to contain fibrin mesh elevated to nearly the same level as seen in NP tissue.

Mechanisms by which the coagulation pathway mediates fibrin deposition and subsequent formation and maintenance of polyp and hyperplastic tissues have been discussed elsewhere.⁶⁵ Several studies suggested that T2 inflammation drives the fibrin deposition in polyps. For example, the extent of fibrin deposition correlates with T2 biomarkers; the formation of alternatively activated macrophages in the tissue, and their production of the fibrin crosslinking molecule Factor XIII-A also correlated with T2 biomarkers; in vitro studies with epithelial cells demonstrated that exposure to IL-4 or IL-13 suppressed the expression of tPA; levels of secondary markers of coagulation also correlated with T2 biomarkers; expression of L-plastin in polyp eosinophils correlated with T2 responses.^{65, 74, 138, 160} These findings, while implicating T2 cytokines in activation of coagulation and fibrin deposition, do not explain why there are many patients in China with non-eosinophilic disease along with formation of NPs. To study this, Zheng Liu and colleagues evaluated these questions in China and found that fibrin was increased and tPA was decreased in the tissue of both eosinophilic polyps and non-eosinophilic polyps. TATc was also increased in both endotypes of patients. Finally, in vitro studies showed that T1 and T3 cytokines could also decrease tPA.¹⁶² These studies suggest that activation and prolongation of coagulation and fibrin deposition are not events related strictly to T2 inflammation.

It is worthwhile pointing out that suppression of coagulation by topical or systemic drugs has the potential to promote fibrinolysis and shrink NPs regardless of endotype. Imoto and collaborators have shown that short chain fatty acids (SCFA) such as propionic acid or butyric acid, as well as retinoic acid, are active in inducing the expression of tPA in epithelial cells, so it may be worthwhile to attempt local delivery of a topically bioavailable retinoic or other SCFA in a trial. (Short-chain fatty acids induce tissue plasminogen activator in airway epithelial cells via GPR41&43.¹⁶³ Trials of anticoagulants, either locally or systemically, may also be worthwhile to determine whether these drugs are capable of shrinking NP tissue. Studies suggest that Nattokinase, a fibrinolytic enzyme used as a food supplement and short chain fatty acids both have potential to be used in such an application.^{161, 163}

Inflammation in CRSsNP

Although CRSsNP was initially classified as T1 inflammation based on the elevation of IFN-γ,^{4, 6} recent studies could not confirm overall elevation of IFN-γ in CRSsNP.^7–9 There are several possibilities for this. One possibility is that there has been an inflammation shift toward T2 away from T1 as has occurred in Asian NPs. In addition, studies in CRSsNP have been complicated by the use of variable sinonasal biopsy sites having inherent tissue-specific molecular differences.¹⁶⁴ By the exclusive use of ethmoid sinus mucosa, it is now clear that inflammatory endotypes in CRSsNP are highly heterogeneous and the frequency of each endotype varies geographically.^7–9 The T2 endotype has become most common in Western countries and 30–55% of CRSsNP patients present the T2 endotype. In contrast, T1 is still a predominant endotype in Asian CRSsNP, although 30–40% of CRSsNP patients in Beijing (China) also have T2 and/or T3 endotypes.⁷ However, the majority of CRSsNP patients in Chengdu (China) did not show the elevation of any endotypic (T1, T2 or T3) markers. Currently, heterogeneity and geographical differences introduce difficulty studying the pathogenic mechanisms in CRSsNP.

Based on these limitations, Klingler et al. used pre-endotyped ethmoid sinus mucosa in CRSsNP and performed microarray analysis.¹⁶⁵ This study identified T1, T2 and T3 endotype-specific genes expressed in CRSsNP that having been useful in developing an understanding of the underlying molecular mechanisms of each endotype in CRSsNP. Based on the gene expression profile, T1 CRSsNP is characterized by the accumulation of Th1 cells, CD8⁺ cytotoxic T cells, NK cells and antigen presenting cells. The release of IFN-γ and cytotoxic molecules including granzymes (GZMA, GZMB and GZMH) may play key pathogenic roles in T1 CRSsNP. T2 CRSsNP is characterized by the elevation of T2 cytokines IL4, IL5 and IL13, markers of T2 immune cells including Th2 cells, ILC2s, eosinophils, mast cells, basophils, DCs and M2 macrophages, and other common T2 markers including CCL18, CCL26, POSTN and CST1.¹⁶⁵ Markers of Treg cells (FOXP3 and IL10) were also elevated in T2 CRSsNP and this may be due to the presence of conventional Treg cells as well as CD109⁺CRTH2⁻ Th2 cells that are found in T2 NPs.¹²¹ In addition, this study clarified that T2 CRSsNP-specific genes were also elevated in T2 CRSwNP, suggesting that the mechanisms of T2 inflammation in T2 CRSsNP and T2 CRSwNP are very similar. T3 CRSsNP seemed to be characterized by the accumulation of Th17 cells, B cells, DCs, M1 macrophages and neutrophils. Activation of neutrophils and complement pathways, and production of pro-inflammatory cytokines TNF and IL-1β from macrophages and neutrophils may play key pathogenic roles in T3 CRSsNP. Finally, it is important to note that a significant proportion of CRSsNP patients did not display a visible elevation of any T1, T2 or T3 endotypic markers in the ethmoid sinus mucosa across the world.^{7–9, 166} There are many potential explanations for this: these patients may have mild elevation of endotypic markers but do not permit classification based on the thresholds; they may have T1, T2 or T3 inflammation in other sinonasal mucosae; or they may have unrecognized endotypes besides T1, T2 and T3.¹⁶⁶ To understand pathogenic roles of CRSsNP, future studies will require larger study populations with consideration of these possibilities and their pathogenic mechanisms.

This research was supported in part by NIH grants, R01 AI137174, R01 NS108968 and P01 AI145818 and by a grant from the Ernest S. Bazley Foundation.