• Open Access

P-glycoprotein is a marker of tissue eosinophilia and radiographic inflammation in chronic rhinosinusitis without nasal polyps


  • Rachel E. Feldman BA,

    1. Department of Otology and Laryngology, Massachusetts Eye and Ear Infirmary, Harvard Medical School, Boston, MA
    Search for more papers by this author
  • Allen C. Lam MD,

    1. Department of Otology and Laryngology, Massachusetts Eye and Ear Infirmary, Harvard Medical School, Boston, MA
    Search for more papers by this author
  • Peter M. Sadow MD, PhD,

    1. Department of Pathology, Massachusetts General Hospital, Harvard Medical School, Boston, MA
    Search for more papers by this author
  • Benjamin S. Bleier MD

    Corresponding author
    1. Department of Otology and Laryngology, Massachusetts Eye and Ear Infirmary, Harvard Medical School, Boston, MA
    • Correspondence to: Benjamin S. Bleier, MD, Harvard Medical School, Massachusetts Eye and Ear Infirmary, Department of Otology and Laryngology, 243 Charles Street, Boston, MA 02114; e-mail: bleierb@gmail.com

    Search for more papers by this author

  • Potential conflict of interest: B.S.B. is an inventor on a provisional patent regarding modulation of P-glycoprotein in the treatment of chronic rhinosinusitis.



P-glycoprotein (P-gp) is a membrane-bound efflux pump that is upregulated in chronic rhinosinusitis (CRS) with nasal polyps (CRSwNP) and participates in epithelial cytokine secretion. Eosinophilic CRS (ECRS) shares a similar cytokine profile with CRSwNP and is associated with significant inflammation and poor surgical outcomes. The goal of this study is to determine if P-gp expression is associated with degree of eosinophilia and severity of radiographic inflammation in patients with CRS without polyps (CRSsNP).


An institutional review board (IRB)-approved study using sinus tissue in 39 steroid-naive patients with CRS. P-gp expression was calculated using quantitative fluorescent immunohistochemistry (Q-FIHC) to generate an epithelial to background staining ratio. Patients were stratified into low and high epithelial expression groups (<3 and ≥3, respectively). Average eosinophils per high powered field (hpf) and Lund-Mackay scores were calculated and compared with P-gp staining ratios using a 2-tailed Student t test.


Among the 39 patients, 7 (17.95%) had high P-gp expression ratios (mean ± SD, 4.86 ± 1.33) while 32 (82.05%) had low expression ratios (1.91 ± 0.45). The number of eosinophils/hpf were significantly greater in the high P-gp expression group as compared to the low expression group (62.38 ± 83.69 vs 5.11 ± 10.12, p = 0.0003). The Lund-Mackay scores were significantly greater in the high P-gp expression group as compared to the low expression group (11.86 ± 2.79 vs 6.84 ± 4.19, p = 0.005).


P-gp is known to be overexpressed in CRSwNP. This study suggests that among patients with CRSsNP, P-gp is similarly overexpressed in those with high tissue eosinophilia and correlates with severity of radiographic inflammation.

Chronic rhinosinusitis (CRS) represents a heterogeneous group of diseases with a variety of pathophysiologic mechanisms. A broad division between CRS with and without nasal polyps (CRSwNP and CRSsNP, respectively) serves as a widely accepted distinction secondary to phenotypic differences evident clinically. Our evolving understanding of the immunologic profiles underlying these disease states[1] suggests that a broader spectrum exists with a predominantly eosinophilic profile subtending not only patients with CRSwNP but a subset of those with CRSsNP as well.[2] The presence of eosinophilic CRS (ECRS) is clinically relevant as these patients share not only the immunologic profile of those with CRSwNP but also the propensity for greater symptom severity[3-5] and worse outcomes.[6, 7] We have previously reported on the overexpression P-glycoprotein (P-gp), an adenosine triphosphate (ATP)-dependent efflux pump, in patients with CRSwNP. Given that P-gp is capable of modulating cytokine secretion,[8] thereby potentially contributing to the pathogenesis of CRSwNP, we hypothesize that P-gp should be similarly elevated in patients with ECRS. The purpose of this study is to examine the association of P-gp expression with mucosal eosinophilia and radiographic inflammation in patients with CRSsNP.

Materials and methods

Tissue procurement

Procurement of the sinus mucosal biopsy samples was approved by the Massachusetts Eye and Ear Infirmary Institutional Review Board. Established consensus diagnostic criteria[9] were used to define patients as having CRS. Tissue was harvested from the anterior ethmoid sinus in each patient. Exclusion criteria included the following: use of oral steroids or immunotherapy within the preceding 4 weeks, aspirin sensitivity (ASA triad), ciliary dysfunction, autoimmune disease, cystic fibrosis, or any known immunodeficiency. Patient with focal etiologies for sinusitis including mucoceles, odontogenic sources, and fungal balls were similarly excluded from the study.

Tissue staining

Following mucosal sampling, the tissue was stained as described.[10] Briefly, following blocking, the primary antibody (monoclonal anti-P-gp clone F4, 1:250; Sigma Aldrich, St. Louis, MO) was applied for 24 hours at 4°C. The tissue was then rinsed followed by application of the secondary antibody (anti-mouse immunoglobulin G [IgG] [Fc specific] F[ab′]2 fragment-fluorescein isothiocyanate [FITC]; 1:160; Sigma Aldrich) for 30 minutes at room temperature. Slides were then rinsed and mounted in Vectashield (Vector Labs, Burlingame, CA) containing propidium iodide (PI) for nuclear counterstaining. Negative control slides were considered those in which the primary antibody was omitted from the staining procedure.

Quantitative fluorescent immunohistochemistry

Fluorescent staining intensity was quantified using a modification of previously described methods.[10] Briefly, image capture was performed with an upright epifluorescent microscope following a standard 1000-ms exposure. Images were then exported into Image J (v1.45s; NIH, Bethesda, MD). The nuclear stain was used to select both the epithelium and a non–tissue-bearing background region, generating a staining intensity ratio. Samples were excluded if the epithelial staining intensity in the negative control slide exceeded that of the adjacent stroma. An epithelial/background staining ratio ≥3 was defined as high P-gp expression.

This cut point was derived from pilot data demonstrating P-gp expression ratios of <3 in nasal septal mucosa, a region with a previously described low level of basal P-gp epithelial expression.[8]

Determination of eosinophilia and radiographic inflammation

For each patient, a representative hematoxylin and eosin (H&E) slide generated as part of their routine pathologic analysis at surgery was selected. The number of eosinophils per five 400× high powered fields (hpf) were recorded by 2 independent and blinded observers (R.E.F. and A.C.L.) as described.[6] The values were averaged to generate a mean eosinophil/hpf score for each patient. Radiographic inflammation was quantified by a single blinded observer (A.C.L.) using the Lund-Mackay staging system.[11]

Statistical analysis

P-gp expression ratios, tissue eosinophilia, and radiographic scores between the patient groups were compared with a 2-tailed Student t test using R (v2.15.2, 2012; R Project for Statistical Computing; http://www.r-project.org). Values of p <0.05 were considered statistically significant.


P-gp expression

Among the 39 patients included in study, the epithelial/background ratio of the high expression group (mean ± SD, 4.86 ± 1.33; n = 7; 17.95%) was significantly greater than that of the low expression group (1.91 ± 0.45; n = 32; 82.05%; p < 0.001) (Fig. 1A). While there was a relative predominance of females in the high P-gp expression group, there were no significant differences between the 2 groups with respect to patient age or race (Table 1).

Table 1. Patient demographics
 Low expressionHigh expression
Age (years)47.8 ± 16.744.3 ± 14.8
Female (%)5085.7
Male (%)5014.3
Caucasian (%)87.585.7
Minority (%)12.514.3
Figure 1.

(A) Bar graph demonstrating the mean P-gp epithelial/background staining ratios between the low and high P-gp–expressing patient groups. (B) Box and whisker plot demonstrating the distribution of eosinophils/hpf between the low and high P-gp–expressing patient groups. (C) Box and whisker plot demonstrating the distribution of Lund-Mackay scores between the low and high P-gp–expressing patient groups. hpf = high power field; P-gp = P-glycoprotein.

Tissue eosinophilia

In the high P-gp expression group, all patients demonstrated >10 eosinophils/hpf with a mean of 62.38 (range, 10.0–240.6). The low P-gp expression group demonstrated a mean of 5.11 eosinophils/hpf (range, 0.0–41.4) which was significantly lower than that of the high expression group (p = 0.0003) (Fig. 1B).

Lund-Mackay scoring

The mean Lund-Mackay score was significantly greater among the high P-gp expression group than that of the low P-gp expression group (11.86 ± 2.79 vs 6.84 ± 4.19; p = 0.005) (Figs. 1C and 2).

Figure 2.

Vertical columns depict findings from 2 distinct patients derived from the low (A, C, E) and high (B, D, F) P-gp expression groups. (A,B) Fluorescent immunohistochemical images of mucosa depicting representative low (A) and high (B) expression of epithelial P-gp (bar = 50 μM, lower right inset represents negative control in which the primary antibody was omitted). (C,D) Matched high-powered (400×) H&E stromal images depicting the absence (C) and presence (D) of mucosal eosinophilia (black arrows denote individual eosinophils). Note the thickened basement membrane in (C) consistent with CRSsNP. (E,F) Coronal CT scans demonstrating increased radiographic inflammation in the patient with high P-gp expression (F) relative to the patient with low P-gp expression (E). CRSsNP = chronic rhinosinusitis without nasal polyps; CT = computed tomography; H&E = hematoxylin and eosin; P-gp = P-glycoprotein.


ECRS represents a histologic diagnosis consisting of mucosal eosinophilia evident in biopsy specimens. While the precise definition of ECRS may be debated, Soler et al.[6] demonstrated that a cut point of >10 eosinophils/hpf provided the best correlation with patient outcomes. ECRS has also been shown in multiple studies to correlate with worse symptoms,[3-5] lower airway hyperactivity,[12, 13] and poor surgical outcomes.[6] These findings may be understood in the context of a shared T-helper 2 (Th2)-skewed immunologic profile as patients with frank nasal polyps.[14] A parsimonious interpretation of these findings suggests that ECRSsNP and CRSwNP may reflect different phenotypic manifestations of the same etiologic process.

P-gp is a 170-kDa glycoprotein encoded by the multidrug resistance polypeptide 1 (MDR1; ABCB1) gene located on chromosome 7q21.12 (Ref. [8]). P-gp has been previously demonstrated to participate in noncanonical cytokine secretion in T cells[15] and our group has demonstrated a similar function in primary sinonasal epithelial cells.[8] The known overexpression of P-gp in CRSwNP,[8] coupled with its ability to promote cytokine secretion, suggests that it may play a role in the pathogenesis of the eosinophilic inflammation seen in nasal polyps. Given our evolving understanding of the relationship between CRSwNP and ECRS we would therefore also expect to see a concomitant overexpression of P-gp in patients with ECRSsNP as compared to those with CRSsNP.

Our findings confirm our hypothesis demonstrating that among patients with CRSsNP, P-gp overexpression predicts tissue eosinophilia. While the mechanistic relationship between epithelial P-gp and eosinophilic inflammation remains unclear, the common finding of upregulation in both CRSwNP and ECRS lends support to the idea that it may play an etiopathologic role. While the clinical utility of Lund-Mackay score may be debated, one of its strengths is that it provides an objective reflection of the degree of global inflammation present in the patient. As with CRSwNP, the intraluminal disease burden seen in patients with ECRS tends to be more severe involving most if not all of the sinuses.[6] In non-eosinophilic patients, disease may be more isolated to a specific region leading to a lower overall Lund-Mackay score even if the local inflammation is quite severe. Consequently, we chose to use the Lund-Mackay score as an additional surrogate marker of inflammation that served to support our histologic findings.

An obvious weakness of our study is the relatively small cohort of patients who were found to have high P-gp expression by quantitative fluorescent immunohistochemistry (Q-FIHC). Despite this, our data was still quite statistically significant, which may represent an attestation to strength of the proinflammatory effect conferred by the presence of P-gp overexpression. However, given the relatively nascent stage of current investigations into the sinonasal function of P-gp it is important to avoid prematurely attributing causality to these findings.


ECRS shares a similar clinical and histologic profile with CRSwNP suggesting the 2 may represent different manifestations of the same underlying process. P-gp is known to be overexpressed in CRSwNP and is capable of modulating epithelial cytokine secretion. Our data suggest that P-gp is similarly overexpressed in ECRS and is associated with radiologic evidence of increased inflammation. These findings further strengthen the link between ECRS and CRSwNP and suggest the potential for an etiopathologic role for P-gp in both diseases.