Introduction
Patients with chronic rhinosinusitis with nasal polyposis (CRSwNP) exhibit within the same inflammatory microenvironment both nasal polyps (NP) and areas with healthy-looking or at least non-polypoid-looking mucosa. The histological differences between polyp and adjacent mucosal tissues have been well-identified.Reference Lacroix, Zheng, Goytom, Landis, Szalay-Quinodoz and Malis1, Reference Ardehali, Amali, Bakhshaee and Madani2 However, the differences between the molecular/inflammatory profiles of neighboring polypoid and non-polypoid tissues have scarcely been addressed, and a relevant review is currently lacking. Having the valuable advantage of being bias-free – since all underlying patient or environment-related factors are identical – one may expect that such comparisons may probably reveal useful new information on NP pathophysiology.
The present study’s first target was to identify factors that are different between NP and non-polypoid neighboring mucosa areas (MS) and, as such, probably essential for mature NP preservation in a CRS setting. A second goal was to find out which of the molecules that are significantly different between NP and healthy controls (cls) are actually not responsible for mature NP preservation, since they are comparable between NP and MS samples. A third challenge was to inquire into the molecular/inflammatory differences (if any) within the nasal microenvironment which is known to present significant differences in the occurrence of NP, with the anterior ethmoid complex being affected the most and inferior turbinates being affected very rarely.Reference Andrews, Bryson and Rowe-Jones3 Finally, this review highlights the need for further research on the molecular mapping of the inflammatory microenvironment of CRSwNP patients, which may hold valuable clinical and therapeutic implications.
Methods
Eligibility criteria
This review adhered to the recommendations of the PRISMA (Preferred Reporting Items of Systematic Reviews and Meta-analysis) statement.Reference Moher, Shamseer, Clarke, Ghersi, Liberati and Petticrew4 All studies comparing NP tissue samples with any specified adjacent non-polypoid nasal mucosa, (such as inferior turbinate [IT], middle turbinate [MT], uncinate process [UP] or ethmoid mucosa [EM]) were eligible for inclusion (Figure 1). Gene expression analyses were beyond the scope of this review.
The study selection flowchart.
Information sources and search
A literature review was conducted via the PubMed database of the US National Library of Medicine (www.pubmed.org) and The CENTRAL of Cochrane Library. The search terms “turbinate and polyp” as well as “uncinate and polyp” limited to the past 20 years and filtered for humans, English, French or German languages attributed 432 and 47 results, respectively, jn PubMed, in March 2024. In Cochrane Library, the same search criteria attributed 71 and 3 results, respectively, with no similar reviews.
Study selection
After duplicates were removed, the studies were screened for eligibility according to the criteria listed in Figure 1. Full texts were assessed for 116 studies. Studies which presented their results after endotype categorisation into eosinophilic (e-) or non-eCRSwNP were subjected to further qualitative analysis. Studies which referred to mixed CRSwNP without any endotype specification were only screened for the authors’ choice of molecules and adjacent mucosa sampling sites as well as comparative results. These studies were excluded from qualitative analysis in keeping with previous reports demonstrating a strong correlation between tissue eosinophilia and most of the studied molecules, as well as a considerable variability in the prevalence of eosinophilic/non-eosinophilic (e-/non-e) CRSwNP endotypes among different study populations.Reference Alanzi, AlBlaies, Elnadif and Riga5–Reference Nakayama and Haruna10
Data collection process
Two investigators (M.R. and O.A.) independently decided on the eligibility of the studies and extracted data from full-text manuscripts. Minor discrepancies were discussed and resolved.
Data items
Comparisons between NP and any specified adjacent MS, regarding any protein, or mRNA expression were considered as primary outcomes. Data on any specified adjacent MS, such as UP, EM, IT or MT as well as any comparisons between them were eligible for inclusion. Comparisons of NP and MS samples between CRSwNP patients and cls, as well as comparisons between eNP and non-eNP populations, were also considered as primary outcomes. Data regarding the consistency between mRNA and protein measurements were considered as secondary outcomes.
Risk of bias in individual studies
The Quality Assessment of Diagnostic Accuracy Studies (QUADAS-2) tool was used to evaluate the included studies. High risk for selection bias was attributed, if more than or equal to three of the following bias factors were unknown or significantly different among study groups; asthma, non-steroidal anti-inflammatory drug exacerbated respiratory disease (NERD), allergic fungal rhinosinusitis, atopy, radiological disease severity and local or systemic treatment with steroids, immunomodulating drugs and antibiotics (Table S1). However, due to the variety of the molecules included in this review, applicability of the aforementioned bias factors was examined separately for each molecule. If there was no research to support association of the molecule in question and any of the selected bias factor, the entry was not taken into consideration to the final risk of bias presented in the QUADAS-2 tool (Table S2).
Risk of bias across studies
To evaluate any risk of bias regarding CRSwNP as a reference standard, we abstracted data on the criteria applied for the classification into eosinophilic/non-eosinophilic (Table S1). In case the classification criteria were unconventional or not specifically reported, the study was considered of high risk of bias as far as CRSwNP diagnosis is concerned and classified as raising concerns regarding the applicability of the results.Reference Chen, Westhofen and Lorenzen11–Reference Linke, Pries, Könnecke, Bruchhage, Böscke and Gebhard13
Summary measures and synthesis of results
A summary of the molecules that are expressed with significant differences between NP and MS, was first conducted, seeking an insight into the biochemical changes than ensure mature NP preservation (either eNP or non-eNP). A summary of the molecules that are found to be comparable between NP and MS, though significantly different to cls, aimed at outlining the surrounding inflammatory processes. These molecules were considered less likely to account at least for the preservation of NP, although their contribution to early NP pathogenesis seems possible. Finally, a molecular mapping of different nasal anatomical subsites targeted to a better understanding of the known differences in the occurrence of polypoid changes between nasal structures.Reference Andrews, Bryson and Rowe-Jones3 Within individual studies statistical significance was attributed to two-sided p less than 0.05.
Results
Study selection
The total number of records identified through database searching (n = 553) was reduced into 10 studies following the procedure presented as a flowchart in Figure 1.Reference Chen, Westhofen and Lorenzen11–Reference Takemoto, Lomude, Takeno, Kawasume, Okamoto and Hamamoto20 Seventeen studies referred to mixed CRSwNP without any endotype specification,Reference Chen, Arab, Westhofen and Lorenzen21–Reference Jeruzal-Świątecka, Borkowska, Łaszczych, Nowicka and Pietruszewska35 4 of which attempted a molecular mapping of the nasal cavity by including in their comparisons samples harvested from more than or equal to two anatomic regions.Reference Chen, Arab, Westhofen and Lorenzen21, Reference Zaravinos, Soufla, Bizakis and Spandidos24, Reference Vinokur, Berenshtein, Chevion and Eliashar27, Reference Homma, Kato, Sakashita, Takabayashi, Norton and Suh31
Study characteristics
Few studies have attempted comparisons between different anatomical sites within the same CRSwNP patients. In a noteworthy majority of 204 studies, researchers chose to compare NP tissue samples with external cls. Endotype-oriented studies were also a minority in the published literature (Figure 1) with heterogenous criteria applied for the diagnosis of eCRSwNP (Table S1). In the 10 studies included in our qualitative analysis, a large variety of 68 molecules were investigated, while literature on the related bias factors is, for most of them, currently lacking (Tables S1 and S2).Reference Chen, Westhofen and Lorenzen11–Reference Takemoto, Lomude, Takeno, Kawasume, Okamoto and Hamamoto20 Only 4/10 studies conducted a localisation analysis within mucosal layersReference Chen, Westhofen and Lorenzen11, Reference Linke, Pries, Könnecke, Bruchhage, Böscke and Gebhard13, Reference Könnecke, Böscke, Waldmann, Bruchhage, Linke and Pries15, Reference Linke, Pries, Könnecke, Bruchhage, Böscke and Gebhard16 with all of them reporting comparable distribution of the studied molecules between NP and IT for the studied endothelial structures/layers. None of the studies proceeded with a histopathological analysis (e/non-e) of CRS without nasal polyposis (CRSsNP) tissue samples. Surprisingly, in the 17 studies on mixed e-/none-CRSwNP populations, a totally different large variety of molecules were investigated.Reference Chen, Arab, Westhofen and Lorenzen21–Reference Gomes, Delemarre, Holtappels, Van, Derycke and Bonne37 Due to this heterogeneity, data from mixed populations were added in Table S3 for only six molecules. Among all 27 studies screened (either endotype oriented or mixed) only 5 molecules (IL-1β, IL-5, IL-6, IL-17A and vascular cell adhesion molecule-1 [VCAM-1]) were investigated by more than 2 research teams in the present literature (Table S3), although the differences in the anatomical sites chosen for sample harvesting actually, and quite surprisingly, made all results unique.Reference Chen, Westhofen and Lorenzen11–Reference Gomes, Delemarre, Holtappels, Van, Derycke and Bonne37
Most of the studies were prospective studies on NP samples taken intra-operatively during functional endoscopic sinus surgery, with UP being the adjacent site of choice for most of the comparisons. Out of the 68 comparisons for eNP, 42 refer to UP, 22 to IT, 2 to MT and 2 to EM samples. Similarly, out of the 37 comparisons for non-eNP, 35 refer to UP and 2 to IT (Table S3).
Results of individual studies
The results of the included studies are summarised in Table S3.
(a) Non-endotype-oriented studies should be interpreted with cautiousness: Significant differences between NP and adjacent non-polypoid mucosa have been reported more frequently (chi-squared test, p = 0.006) for eCRSwNP (43/68 comparisons) than non-eCRSwNP patients (13/37 comparisons). Significant differences between non-polypoid CRSwNP mucosa and cls were comparably common (chi-squared test, p = 0.76) among eCRSwNP (20/51 comparisons) and non-eCRSwNP (19/45 comparisons) (Table S3).
(b) Comparisons of different anatomical sites within the same patients are currently lacking: None of the existing endotype-oriented studies extended their measurements so as to compare different non-polypoid samples within the same patients. Table S4 presents such comparisons in mixed e-/non-eCRSwNP populations. Only IL-5 mRNA was found to be expressed significantly different within the nasal microenvironment by one research team.Reference Chen, Arab, Westhofen and Lorenzen21
Synthesis of the results
The quantitative synthesis of our results was prohibited by the very limited number of studies conducted for each molecule, as well as the lack of standardisation of laboratory samples processing for most of the reported molecules. The qualitative synthesis attributed the following primary outcomes.
1. Associations in eosinophilic populations
(a) eNP ≠ MS ≠ cls: The following proteins were found to be significantly different between eNP, MS and respective cls: β-chemokine ligand-11 (CCL-11), CCL-24, chemokine ligand (CXCL)-1, IL-2Rα, IL-5, ratio of metalloproteinase-1 (MMP1) to tissue inhibitor of metalloproteinase-1 (TIMP1) (M1T1), anti-human eosinophil major basic protein (MBP), monocyte chemoattractant protein-4 (MCP-4), mitogen-activated protein/ extracellular signal-regulated kinases1/2 (MEK1/2), MMP-1, regulated upon activation normal T cell expressed and presumably secreted (RANTES) and tryptase.Reference Kim, Park, Chang, Eun, Shin and Mo14–Reference Kim, Eun, Kim, Cho, Han and Han18 All comparisons refer to UP samples, except MEK1/2 which refers to IT. No similar mRNA analyses are available in the present literature.
(b) eNP ≠ MS ≈ cls: The following proteins were found to be significantly different between eNP and non-polypoid adjacent tissues, while the latter were comparable to cls: phosphorylated extracellular signal-regulated kinases1/2 (pERK1/2), phosphorylated glycogen synthase kinases3α,β (pGSK-3α,β), interferon (IFN)γ, IL-1α, IL-1β, IL-5, IL-6, IL-23, M3T1, M9T1, pMEK1/2, MMP-3, MMP-9, phosphorylated signal transducer and activator of transcription-5α/β (pSTAT-5α/β) and VCAM-1 [12,13,16-18]. All comparisons refer to UP or MT samples, except pERK1/2, pGSK-3α,β, pMEK1/2 and pSTAT-5α/β which refer to IT. Similar associations were found for the mRNA of bitter taste receptors T2R14 and T2R38.Reference Takemoto, Lomude, Takeno, Kawasume, Okamoto and Hamamoto20
(c) eNP ≠ cls and eNP ≈ MS: Such associations were reported for the following proteins: GSK-3α,β, anti-human neutrophil elastase (HNE), inter-cellular adhesion molecule-1 (ICAM-1), IL-4, IL-13 and STAT5b.Reference Kim, Park, Chang, Eun, Shin and Mo14, Reference Linke, Pries, Könnecke, Bruchhage, Böscke and Gebhard17, Reference Kim, Eun, Kim, Cho, Han and Han18 All comparisons refer to UP samples, except STAT5b which refers to IT samples.
2. Associations in non-eosinophilic populations
(a) non-eNP ≠ MS ≠ cls: The following proteins were found to be significantly different between non-eNP, MS and respective cls tissue samples: IL-2Rα, IL-5, M9T1, MMP-9 and RANTES.Reference Kim, Eun, Kim, Cho, Han and Han18 All comparisons refer to UP samples.
(b) non-eNP ≠ MS ≈ cls: The following proteins were found to be significantly different between non-eNP and non-polypoid adjacent tissues, while the latter were comparable to cls: CCL-11, IL-23, M3T1, MMP-3 and VCAM-1.Reference Kim, Eun, Kim, Cho, Han and Han18 All comparisons refer to UP samples.
(c) Non-eNP ≠ cls and non-eNP ≈ MS: Such associations were reported for the following proteins: CCL-24, CXCL-1, CXCL-2, CXCL-8, ICAM-1, IFN-γ, IL-1α, IL-1β, IL-4, IL-6, IL-13, IL-17A, M1T1, M2T1, MCP-4, MMP-1 and TIMP-1.Reference Kim, Park, Chang, Eun, Shin and Mo14, Reference Kim, Eun, Kim, Cho, Han and Han18 All comparisons refer to UP samples.
Secondary outcomes
Protein and mRNA measurements’ consistency: Protein and m-RNA analyses were compatible with each other for CCL-11,Reference Takemoto, Lomude, Takeno, Kawasume, Okamoto and Hamamoto20, Reference Chen, Arab, Westhofen and Lorenzen21 ERK1/2,Reference Linke, Pries, Könnecke, Bruchhage, Böscke and Gebhard16 hTERT,Reference Chen, Westhofen and Lorenzen11 ICAM-1Reference Könnecke, Böscke, Waldmann, Bruchhage, Linke and Pries15, Reference Kim, Eun, Kim, Cho, Han and Han18 and E-selectin,Reference Könnecke, Böscke, Waldmann, Bruchhage, Linke and Pries15 while they were non-compatible for MEK1/2,Reference Linke, Pries, Könnecke, Bruchhage, Böscke and Gebhard16 P-selectinReference Könnecke, Böscke, Waldmann, Bruchhage, Linke and Pries15 and VCAM-1Reference Könnecke, Böscke, Waldmann, Bruchhage, Linke and Pries15, Reference Kim, Eun, Kim, Cho, Han and Han18 (Table S3). The limited number of available reports, however, calls for a skeptical consideration of these results.
Discussion
Study rationale
Although CRSwNP has been the subject of intensive research over the past decades, the pathophysiology underlying NP formation and preservation remains elusive. The great number and variety of molecules that have been found to express significantly different in NP when compared to cls, as well as the variety of overlapping inflammatory endotypes still make CRSwNP pathophysiology incomprehensive and perplex. The rationale behind the present review is to study the changes in the molecular level by combining data from three tissue sources: (1) the NP, which is the most troublesome sequela of the chronic inflammation as the main contributor to nasal obstruction and extranasal complications; (2) the neighboring healthy-looking mucosa, which of course is subject to the same inflammatory process, co-morbidities, genetic and environmental factors, but does not exhibit the edematous stroma and the hyperplasia of goblet cells in the overlying epithelium that characterise a NP;Reference Ardehali, Amali, Bakhshaee and Madani2 and (3) an external cls without any chronic inflammation. The results of this combined analysis, ideally further enriched with comparisons between CRSwNP and CRSsNP and a biochemical mapping of the areas with substantially different prevalence of NP, might help to shorten the long list of CRSwNP-related molecules down to those that actually play a key role to NP formation and preservation.
Interpretation of the results
To date, research on the changes detected within the inflammatory microenvironment of CRSwNP patients is very limited, involves a surprising variety of molecules and refers to substantially different study designs in terms of (1) populations, (2) origin of non-polypoid samples and (3) laboratory methods (Tables S1 and S3). It is however of interest that, known Th2 mediators, tissue remodeling mediators and pro-inflammatory cytokines that have been found to be significantly different between NP&cls, actually present all three models of comparative associations between NP, MS and cls.
Out of the 58 molecules or molecular ratios tested, only 15 were found to differ significantly between NP, MS and cls (Table S3), compatible with a logical escalation of chronic inflammation with NP, CRS inflammation and healthy mucosa. Only IL-5 and IL-2Ra have been found to differ significantly between NP, MS and cls in both eosinophilic and non-eosinophilic samples. A larger variety of molecules were found either comparable between NP and MS (possibly because of their minor role in polyp preservation) or comparable between MS and cls (possibly because their persistence is related to NP persistence). The limited literature does not yet allow for safe categorisations yet.
Based on the results of the present study, the differences between e- and non-eCRSwNP endotypes become even more evident when the parameter of the adjacent mucosa is included in the analysis and therefore studies on mixed populations should be interpreted with cautiousness. In the present literature, 27/40 proteins were initially found to be comparable between eNP and non-eNP. However, when the parameter of MS was added to the comparisons comparable results were maintained for only 8/40 molecules (ICAM-1, IL-2Rα, IL-10, IL-13, IL-23, M3T1, MMP-2 and tumor necrosis factor-α [TNFα]) (Table S3). Further studies are needed to verify the results and test the correlation coefficients of those associations, which may actually differ.
Identification of the molecular differences between neighboring nasal areas was another primary goal of this study. These comparisons, being free from all patient or environment-related bias factors, targeted to inquire into the molecular/inflammatory differences that, may reflect the clear preference of NP occurrence in certain parts of the nasal microenvironment such as the anterior ethmoid complex. However, the relevant literature is currently lacking. Few studies have attempted a biochemical mapping of the nasal cavities, on mixed e-/non-eCRSwNP populations (Table S4). Quite interestingly, Chen et al. (2005)Reference Chen, Arab, Westhofen and Lorenzen21 have reported that in their subpopulation of eCRSwNP patients, anterior EM samples exhibited the highest relative expression of IL-5 mRNA followed by samples taken from the lateral portion of the MT. The lowest amount of IL-5 mRNA expression was found in the medial portion of the MT and IT21. However, this question has not been addressed again in the literature.
Limitation of the evidence
The present study targeted to identify the differences in the molecular level between NP and adjacent MS in an attempt to thus eliminate demographic, genetic, immunologic, environmental and disease-severity-related bias. It is, however, of note that comparisons between NP, MS and IT samples also enclose different bias factors. A neuro-immune interaction was hypothesised by Coffey et al. (2009)Reference Coffey, Mulligan and Schlosser25 based on their findings that in patients with CRSwNP, paranasal sinuses exhibited higher levels of nerve growth factor (NGF) than IT, while in cls, a reverse association was noted. The nasal airflow and mucosal temperature, which are also presumed to play a role in epithelium cells response, have been found to be different between IT and middle meatus areas.Reference Xiong, Zhan, Jiang, Li, Rong and Xu38–Reference Wang de, Lee and Gordon40 The nasal FeNO/NO coefficient has been reported to be markedly higher in the IT than the middle meatus area in patients with allergic rhinitis.Reference Takeno, Yoshimura, Kubota, Taruya, Ishino and Hirakawa41 The submucosal glands and veins, as well as the interstitial hydrostatic pressure increase following saline injection have been shown to differ significantly between IT and MT in healthy subjects.Reference Berger, Finkelstein, Ophir and Landsberg42–Reference Pezato, Voegels, Stamm and Gregório44 Further studies are needed to determine whether the above-mentioned differences represent strong bias for comparisons between NP and MS to adjacent or cls IT samples.
• There are three models of comparative associations between nasal polyp tissue (NP), non-polypoid adjacent mucosa (MS) and healthy controls (cls). In the first they all differ from each other, in the second MS is comparable to cls and in the third NP is comparable to MS
• Direct comparisons between nasal polyp samples and healthy inferior turbinate cls should therefore be interpreted with cautiousness, at least as far as polyp pathogenesis is concerned
• All three models of comparative associations have been identified in most major cytokine groups, such as Th2 mediators, tissue remodeling mediators and pro-inflammatory cytokines
• Among the numerous molecules studied only interleukin (IL) -5 and IL-2Ra have been found to differ significantly between NP, MS and cls in both eosinophilic and non-eosinophilic samples
• Further studies are needed in order to map the inflammatory microenvironment of patients with chronic rhinosinusitis with nasal polyps (CRSwNP) and evaluate the possible clinical and therapeutical implications of this multi-dimensional approach
Conclusions
Highlighting the need for further studies, that will refresh the current perspective of the pathophysiology of CRSwNP by comparisons within the inflammatory microenvironment of CRSwNP patients, is one of the main functions of the present review. Although a meta-analysis was not possible due to the large heterogeneity of the studied molecules and the limited number of existing studies, the qualitative analysis of the present data revealed that several molecules present significant differences between NP and non-polypoid adjacent mucosa. Direct comparisons between nasal polyp samples and healthy IT cls should therefore be interpreted with cautiousness, at least as far as polyp pathogenesis is concerned.
Supplementary material
The supplementary material for this article can be found at https://doi.org/10.1017/S0022215125102946.
Funding statement
None/
Competing interests
No potential conflicts of interest for any of the authors.
