Studies of the role of basophils in aspirin-exacerbated respiratory disease pathogenesis

Abstract

Background:

Aspirin-exacerbated respiratory disease (AERD) is characterized by the triad of chronic rhinosinusitis with nasal polyps (CRSwNP), asthma, and intolerance to cyclooxygenase-1 enzyme inhibitors. The underlying mechanisms contributing to AERD pathogenesis are not fully understood, but AERD is characterized by an enhanced type 2 inflammatory phenotype. Basophils are potent type 2 effector cells, but their involvement in AERD pathophysiology remains unclear.

Objective:

We sought to characterize the systemic and local basophil responses in patients with AERD compared with patients with CRSwNP.

Methods:

Sinonasal tissues including inferior turbinate and/or nasal polyps (NPs) and peripheral blood were collected from controls, patients with AERD, and patients with CRSwNP. Expression of cell surface (CD45, FcεRI, CD203c), activation (CD63), and intracellular (2D7) markers associated with basophils was characterized using flow cytometry. Clinical data including Lund-Mackay scores and pulmonary function were obtained.

Results:

The mean number of basophils (CD45⁺CD203c⁺FcεRI⁺CD117⁻) detected in AERD NPs (147 ± 28 cells/mg tissue) was significantly elevated compared with that detected in CRSwNP NPs (69 ± 20 cells/mg tissue; P = .01). The number of circulating basophils was significantly elevated in patients with AERD (P = .04). Basophils in NPs had significantly higher CD203c and CD63 mean fluorescence intensity compared with blood in both conditions (P < .01). Basophils from AERD NPs had lower expression of the granule content marker 2D7 compared with those from matched blood (P < .01) or NPs of patients with CRSwNP (P = .06), suggesting ongoing degranulation. Basophil 2D7 mean fluorescence intensity significantly correlated with pulmonary function (r = 0.62; P = .02) and inversely correlated with sinonasal inflammation (r = −0.56; P = .004).

Conclusions:

Increased basophil numbers and extent of ongoing degranulation in NPs of patients with AERD compared with patients with CRSwNP may contribute to the exaggerated disease pathogenesis and severity unique to AERD.

GRAPHICAL ABSTRACT

Aspirin-exacerbated respiratory disease (AERD) is a chronic inflammatory condition of the upper and lower respiratory tracts. Clinically, AERD is composed of the triad of chronic rhinosinusitis with nasal polyps (CRSwNP), asthma, and intolerance to medications that inhibit the cyclooxygenase-1 enzyme. AERD is estimated to occur in 8% to 16% of patients with CRSwNP and 15% of patients with severe asthma.^1,2 On average, patients with AERD have more severe sinonasal disease, undergo significantly more sinus surgeries, and are more likely to require daily oral corticosteroids than patients with CRSwNP alone or with asthma but lacking sensitivity to aspirin-like drugs.² The underlying cellular and molecular mechanisms contributing to the enhanced severity and unique phenotype of AERD are not fully understood.

In Western countries, nasal polyps (NPs) are predominantly characterized by type 2 inflammation.^3,4 Levels of the type 2 cytokines IL-5 and IL-13 are elevated in NPs when compared with levels in healthy controls^5,6 although no significant difference was observed between patients with AERD and patients with CRSwNP.² Cells commonly associated with type 2 inflammation, including type 2 CD4⁺ T cells and group 2 innate lymphoid cells, are elevated in NPs in the West.^7–9 Eosinophils are also significantly elevated in NPs compared with healthy controls, and their associated granule proteins are even higher in NPs of patients with AERD compared with those with only CRSwNP.^2,5,6,10 Elevations in the number of circulating eosinophils are clinically associated with an increased risk of disease recurrence in patients with CRSwNP and AERD.¹¹ Mast cells are elevated and also suspected to play a role in NP pathogenesis, especially in AERD in which increased production of mast cell mediators including prostaglandin D₂ and cysteinyl leukotrienes has been demonstrated and implicated in reactions to cyclooxygenase-1 inhibitors.^12–16

Basophils are potent granulocytes that have long been associated with type 2 inflammation. These cells can secrete various proinflammatory mediators, including histamine, cysteinyl leukotrienes, IL-4, and IL-13, that can all contribute to the type 2 allergic inflammatory response.^17,18 Clinical studies found that basophils were significantly elevated in postmortem lung sections of patients with asthma who died in status asthmaticus compared with lung sections of patients with asthma who died from nonasthmatic causes.^19,20 Activation of blood basophils has been used as a biomarker of disease in food allergy, allergic rhinitis, and asthma.

Despite evidence that basophils are important contributors to type 2 respiratory diseases, there are few studies evaluating basophils in NPs. In a previous study, our group used immunohistochemistry to detect 2D7⁺ basophils and found elevations of these cells in CRSwNP tissue, but paradoxically not in AERD.¹⁰ That study had some weaknesses because it did not quantify these cells rigorously, it did not assess the level of basophil activation or allow for changes in 2D7 during activation, and it did not relate basophils or basophil activation to markers of disease severity. In the current study, we significantly advanced our previous work on basophils by using a flow cytometry–based approach to characterize, quantify, and compare basophils in NPs of patients with AERD and CRSwNP. We found that basophils were significantly elevated and activated in NPs of patients with CRSwNP and in NPs of patients with AERD. We also found that levels of a marker of basophil degranulation correlated with enhanced clinical disease severity.

METHODS

Study population

All subjects signed informed consent and were recruited from the Division of Allergy and Immunology or the Department of Otolaryngology clinics within Northwestern Medicine. Additional information is provided in this article’s Online Repository at www.jacionline.org, with demographic information listed in Table I. The Institutional Review Board of Northwestern University Feinberg School of Medicine approved this study.

TABLE I.

Subjects’ demographic characteristics

Characteristic	Control (n = 18)	CRSwNP (n = 22)	AERD (n = 22)	P value
Age (y), mean ± SD	37.8 ± 14.6	49.3 ± 10.9	44.3 ± 12.6	.01
Sex: female, n (%)	4 (22)	8 (36)	11 (50)	.19
Caucasian race, n (%)*	14 (78)	19 (90)	13 (62)	.05
Asthma, n (%)	2 (11)	12 (55)	22 (100)	<.001
Atopy, n (%)†	6 (75)	6 (66)	12 (100)	.11
Never smoker, n (%)	14 (78)	13 (59)	16 (73)	.40
FEV1 (L), ± SEM‡	NA	3.27 ± 0.23	2.41 ± 0.13	.002
LM score, ± SEM§	NA	16.6 ± 0.9	19.9 ± 0.8	.02
Perioperative corticosteroid use, n (%)
Intranasal	4 (22)	5 (22)	13 (59)	.02
Inhaled	0 (0)	5 (23)	13 (59)	<.001
Oral	0 (0)	10 (45)	14 (64)	<.001

NA, Not available.

^*Two controls, 1 patient with CRSwNP, and 1 patient with AERD declined to answer.

^†Atopy status not available for 10 controls, 13 patients with CRSwNP, and 10 patients with AERD.

^‡FEV₁ (L) not available for all controls, 12 patients with CRSwNP, and 5 patients with AERD.

^§The LM score was not available for all controls and 2 patients with AERD.

Isolation of cells

Cells were isolated from NPs and peripheral blood as previously described,⁸ with additional information provided in this article’s Online Repository at www.jacionline.org.

Flow cytometric analysis

Protocols and reagents used for flow cytometric analyses of NPs and peripheral blood are provided in this article’s Online Repository at www.jacionline.org. Representative gating strategies are shown for eosinophils (see Fig E1 in this article’s Online Repository at www.jacionline.org) and mast cells and basophils (see Fig E2 in this article’s Online Repository at www.jacionline.org).

Statistics

Data were analyzed using a chi-square test, Mann-Whitney test, 1-way ANOVA Kruskal-Wallis test with Dunn’s test for multiple comparisons, and Spearman rank as appropriate. A P value of less than .05 was considered significant, with further details provided in this article’s Online Repository at www.jacionline.org.

RESULTS

Identification of type 2 inflammatory granulocytes in NPs

Previous work has shown that cells staining with the basophil-specific marker 2D7 were present and elevated in NPs (see Fig E3 in this article’s Online Repository at www.jacionline.org).¹⁰ The reliability of 2D7 is variable because it is found in the granules of basophils, and degranulation of the cells can theoretically diminish staining of activated cells. We therefore undertook a challenging study in which we used flow cytometry of dispersed tissue cells to identify and characterize basophils along with other inflammatory cells in NPs of patients with AERD or CRSwNP. The average number of CD45⁺ cells (leukocytes) was elevated in NPs from patients with AERD (5606 ± 1150 cells/mg tissue) compared with NPs from patients with CRSwNP (3504 ± 649 cells/mg tissue), but this difference did not reach statistical significance (P = .09, Mann-Whitney test) (Fig 1, A). For reference, inferior turbinate of control patients without chronic rhinosinusitis had significantly fewer leukocytes (1733 ± 182 cells/mg tissue) than did the NP subgroups (P <.007, Kruskal-Wallis test) (Fig 1, A). These findings demonstrate that NPs have an exaggerated cellular inflammatory response compared with healthy sinonasal tissue.

FIG 1.

Basophils are elevated in NPs of patients with AERD. Flow cytometry was used to quantify the number of CD45⁺ leukocytes (A), CD45⁺ Siglec-8⁺ FcεRI⁻ eosinophils (B), CD45⁺ CD117⁺ FcεRI⁺ mast cells (C), and CD45⁺ CD203c⁺ FcεRI⁺ CD117⁻ basophils (D) in NPs of patients with CRSwNP or AERD. For comparison, the number of leukocytes (Fig 1, A) and basophils (Fig 1, D) was quantified in inferior turbinate tissue of healthy controls. The number of tissue basophils strongly correlated with the number of tissue eosinophils (E) and tissue mast cells (F) in NPs. IT, Inferior turbinate; n.d., not determined. Dot plots illustrate individual data points, with solid lines representing the median and interquartile range. Statistical significance (P < .05) was determined by Mann-Whitney U test comparing AERD and CRSwNP (Fig 1, A–D) or Spearman rank correlation (Fig 1, E and F). *P < .05.

To further characterize the inflammatory milieu within NPs, we next evaluated various type 2–associated inflammatory granulocytes using flow cytometry. Eosinophils were identified as CD45⁺Siglec-8⁺FcεRI⁻ cells, whereas mast cells were defined as CD45⁺CD117⁺FcεRI⁺ cells. The average number of eosinophils was increased in AERD NPs (6277 ± 2548 cells/mg tissue) compared with CRSwNP (4161 ± 1221 cells/mg tissue), but the difference was not significant (P = .54) (Fig 1, B). Similarly, there was a nonsignificant trend (P = .15) toward an increase in the average number of NP mast cells in AERD (191 ± 46 cells/mg tissue) compared with CRSwNP (124 ± 36 cells/mg tissue) (Fig 1, C). However, inferior turbinate of control patients had significantly fewer mast cells (15 ± 9) than either polyp subgroup (P < .001, Kruskal-Wallis test) (Fig 1, C).

Basophils, defined as CD45⁺CD203c⁺FcεRI⁺CD117⁻ cells, were significantly elevated in NPs of patients with AERD (147 ± 28 cells/mg tissue) compared with patients with CRSwNP (69 ± 20 cells/mg tissue; P = .01) (Fig 1, D). Basophils also trended to represent a larger proportion of total NP leukocytes in AERD (3.2%) compared with CRSwNP (2.2%; P = .09) (data not shown). For reference, inferior turbinate of control patients without chronic rhinosinusitis had significantly fewer average total numbers (4 ± 2 cells/mg tissue) and percentage (0.25%) of basophils than either NP subgroup (Fig 1, D, and data not shown). However, we found no significant difference in the number of NP basophils based on the presence or absence of atopy (data not shown). Among all patients with NPs (both CRSwNP and AERD), the number of basophils significantly correlated with the number of eosinophils (r = .64; P < .001) (Fig 1, E) and the number of mast cells (r = 0.56; P < .001) (Fig 1, F). These findings suggest that NPs in both AERD and CRSwNP are characterized by a type 2 inflammatory environment and that the mechanisms of expansion of these 3 cell types may be overlapping.

Identification of type 2 inflammatory granulocytes in peripheral blood

We next examined the peripheral blood of control subjects and patients with AERD or CRSwNP to determine whether there were similar elevations in circulating inflammatory cells. Interestingly, patients with AERD on average had significantly more peripheral leukocytes (4257 ± 364 cells/μL blood) than did patients with CRSwNP (2777 ± 455 cells/μL blood; P < .01, Mann-Whitney) and controls (2967 ± 410 cells/μL blood; P = .01, Kruskal-Wallis) (Fig 2, A). In contrast, the number of circulating leukocytes of patients with CRSwNP was not different from that of controls (Fig 2, A). The average number of peripheral eosinophils was also higher in AERD (242 ± 57 cells/μL blood) than in CRSwNP (155 ± 37 cells/μL blood), but this did not reach statistical significance (P = .36) (Fig 2, B).

FIG 2.

Basophils are elevated in the peripheral blood of patients with AERD. Flow cytometry was used to quantify the number of CD45⁺ leukocytes (A), CD45⁺ Siglec-8⁺ FcεRI⁻ eosinophils (B), and CD45⁺ CD203c⁺ FcεRI⁺ CD117⁻ basophils (C) in the peripheral blood of patients with CRSwNP or AERD or healthy controls. The number of peripheral blood basophils strongly correlated with the number of circulating eosinophils (D) in patients with CRSwNP and AERD combined. Dot plots illustrate individual data points, with solid lines representing the median and interquartile range. Statistical significance (P < .05) was determined by Mann-Whitney U test comparing AERD and CRSwNP (Fig 2, A–C) or Spearman rank correlation (Fig 2, D). *P < .05. **P < .01.

Basophils were significantly elevated in the circulation of patients with AERD (38 ± 8 cells/μL blood) compared with patients with CRSwNP (17 ± 3 cells/μL blood; P = .04) (Fig 2, C). When data for the number of basophils and the number of eosinophils in the peripheral blood of patients with AERD and patients with CRSwNP were combined and compared, a strong correlation was observed between these 2 cell types (r = 0.64; P = .0001) (Fig 2, D).

Because tissue basophils are recruited from the periphery, we next assessed whether there was a significant correlation between the number of basophils in the peripheral blood and the number of basophils in NPs in matched individuals with CRSwNP or AERD. When assessing both conditions together, we found a significant correlation between basophils in NPs and blood (r = 0.44; P =.02) (Fig 3, A). However, this correlation was driven primarily by patients with AERD because we found no significant correlation between tissue basophils and peripheral basophils in patients with just CRSwNP (r = 0.32; P = .29) (Fig 3, B). In contrast, a stronger correlation was observed when assessing only patients with AERD (r = 0.52; P = .07) (Fig 3, C).

FIG 3.

Basophil numbers strongly correlated between the peripheral blood and NPs of patients with AERD but not in those with CRSwNP. The number of CD45⁺ CD203c⁺ FcεRI⁺ CD117⁻ basophils in matched peripheral blood and NPs was compared when patients with AERD or CRSwNP were combined (A) or when patients with CRSwNP (B) and patients with AERD (C) were analyzed separately. Dot plots illustrate individual data points, with blue representing a patient with CRSwNP and black representing a patient with AERD. Statistical significance (P < .05) was determined by Spearman rank correlation.

Assessment of basophil activation

We next assessed the activation status of basophils in CRSwNP and AERD by evaluating the surface expression of traditional basophil activation markers, CD203c and CD63, using flow cytometry. We found no difference in the geometric mean fluorescence intensity (MFI) of CD203c on circulating basophils of patients with AERD (939 ± 141), patients with CRSwNP (821 ± 103), or controls (1279 ± 216) (Fig 4, A). In contrast, CD203c MFI was significantly upregulated on basophils derived from NPs compared with those from peripheral blood in both AERD (2290 ± 294 vs 939 ± 141; P = .001) and CRSwNP (1805 ± 123 vs 821 ± 103; P = .001) (Fig 4, A). There was no significant difference in CD203c MFI on basophils from NPs of AERD versus CRSwNP (P > .99).

FIG 4.

Basophils expressed higher levels of activation markers CD63 and CD203c in NPs compared with peripheral blood in both AERD and CRSwNP. Flow cytometry was used to quantify the MFI of CD203c (A) and CD63 (B) on basophils isolated from the peripheral blood and from NPs of patients with AERD or CRSwNP. Peripheral blood from healthy controls was used for comparison. The number of CD63^hi basophils was quantified in NPs (C) and peripheral blood (D) from the diseased and healthy conditions. IT, Inferior turbinate. Dot plots illustrate individual data points, with solid lines representing the median and interquartile range. Statistical significance (P < .05) was determined by Kruskal-Wallis test with post hoc analysis using the Dunn’s test for multiple comparison (Fig 4, A and B) and by Mann-Whitney U test comparing AERD and CRSwNP (Fig 4, C and D). **P < .01. ***P < .001.

We observed similar findings when measuring another basophil activation marker, CD63. The average levels of CD63 were comparable on circulating basophils in controls (989 ± 200), patients with CRSwNP (1202 ± 221), and patients with AERD (1121 ± 189) (Fig 4, B). However, CD63 was significantly upregulated on basophils in NPs compared with the circulation in both AERD (P < .001) and CRSwNP (P = .003). Levels of CD63 on NP basophils were similar between CRSwNP and AERD (P > .99) (Fig 4, B). We also quantified the absolute number of basophils that highly expressed CD63. In NPs, there were significantly more CD63^hi basophils in AERD (25 ± 8 cells/mg tissue) than in CRSwNP (6 ± 1 cells/mg tissue; P = .002) (Fig 4, C). Basophil activation appears to be restricted to the tissue, because there were minimal CD63^hi basophils detected in the blood, and no differences seen between controls, patients with CRSwNP, and patients with AERD (Fig 4, D). Taken together, these findings suggest that basophils become activated similarly in CRSwNP and AERD only once they enter NPs. Although the level of activation is equivalent (as measured by CD63 and CD203c MFI) in both disease conditions, the number of highly activated basophils in NP tissue is higher in AERD than in CRSwNP.

Assessment of basophil degranulation

The antibody 2D7 is well established and specifically labels basophil secretory granules.²¹ As mentioned above, we previously observed that 2D7 staining in NPs via immunohistochemistry was lower in AERD compared with CRSwNP.¹⁰ We hypothesized that basophils in AERD had degranulated to a greater extent in AERD, leading to lower levels of 2D7 and reduced detection by immunohistochemistry. To test this hypothesis, we used flow cytometry to measure intracellular levels of 2D7 in basophils from patients with AERD and CRSwNP.

The MFI of 2D7 was the highest in circulating basophils, and there were no significant differences in 2D7 observed between controls (5690 ± 1263) and patients with CRSwNP (9668 ± 1652; P = .74) or AERD (7728 ± 1570) (Fig 5, A). In contrast, 2D7 MFI levels in NP basophils were reduced in AERD (1522 ± 384) compared with CRSwNP (4742 ± 1073; P =.06) (Fig 5, A). Levels of 2D7 were lower in basophils isolated from NPs compared with the peripheral blood in patients with AERD (P < .001) but not in patients with CRSwNP (P = .28), suggesting that basophil degranulation may be occurring to a greater extent in NPs of AERD versus CRSwNP (Fig 5, A).

FIG 5.

Basophils have reduced levels of 2D7 expression in NPs of patients with AERD compared with those with CRSwNP. Flow cytometry was used to quantify the MFI of intracellular 2D7 expression (A) within basophils isolated from the peripheral blood and from NPs of patients with AERD or CRSwNP. Peripheral blood from healthy controls was used for comparison. The percentage of basophils expressing 2D7⁺ in peripheral blood (B) and NPs (C) was also determined. IT, Inferior turbinate. Dot plots illustrate individual data points, with solid lines representing the median and interquartile range. Statistical significance (P < .05) was determined by Kruskal-Wallis test with post hoc analysis using the Dunn’s test for multiple comparison (Fig 5, A) and by Mann-Whitney U test comparing AERD and CRSwNP (Fig 5, B and C). *P < .05. ***P < .001.

In addition to measuring the intensity of 2D7 staining (MFI), we assessed how many basophils expressed 2D7 out of the total basophil population (the percent positive). In the circulation of all subject groups, nearly all basophils expressed 2D7 (Fig 5, B). In contrast, significantly fewer NP basophils expressed 2D7 in AERD (33%) compared with CRSwNP (65%; P = .005) (Fig 5, C). These findings suggest that basophils have degranulated and released 2D7 to a greater extent in AERD NPs and provide a mechanistic explanation for our previous observation of reduced levels of 2D7-positive cells in AERD polyps.¹⁰

Association with clinical disease severity

To assess the relevance of basophil recruitment, activation, and degranulation to sinus disease across the spectrum of nasal polyposis disease severity, we evaluated whether markers of basophil activation and degranulation were associated with clinical disease severity. The Lund-Mackay (LM) score is a radiographic assessment of sinonasal inflammation visible on sinus computed tomography scan, with a higher score representing more severe disease. We found a significant correlation between the number of CD63^hi basophils detected in NPs and the LM score (r = .39; P = .02) (Fig 6, A). In addition, the number of CD63^hi basophils inversely correlated with FEV₁ (r = −0.42; P = .04), that is to say that patients with more activated basophils had lower lung function and more severe lower airway disease (Fig 6, B). Furthermore, there was a significant inverse correlation between basophil 2D7 MFI and the LM score (r = −0.56; P = .004) (Fig 6, C) and a significant positive correlation between 2D7 MFI and FEV₁ (r = 0.62; P = .02) (Fig 6, D). This suggests that basophil degranulation (and activation) is associated with more severe upper and lower respiratory clinical disease as measured by radiography and pulmonary function testing, respectively.

FIG 6.

Markers of basophil activation (CD63) and degranulation (2D7 MFI) correlated with sinonasal and pulmonary disease severity. The number of CD63^hi basophils significantly correlated with the LM score (A) and inversely correlated with FEV₁ (B) in patients with either CRSwNP or AERD. In addition, 2D7 MFI inversely correlated with the LM score (C) and positively correlated with FEV₁ (D) in both CRSwNP and AERD. Dot plots illustrate individual data points, with blue representing a patient with CRSwNP and black representing a patient with AERD. Statistical significance (P < .05) was determined by Spearman rank correlation.

DISCUSSION

To our knowledge, this is the first study to functionally characterize basophils in NPs of patients with AERD and CRSwNP using flow cytometry or other methods. Basophils were more numerous in NPs compared with healthy sinonasal tissue, and the total number of basophils and the proportion of activated (CD63^hi) basophils were highest in AERD compared with CRSwNP. This is also the first study to quantitatively evaluate the expression of the basophil granule marker 2D7 in diseased tissue. We found a lower percentage of 2D7⁺ basophils and reduced 2D7 intensity on basophils in NPs of patients with AERD compared with patients with CRSwNP. On the basis of 2D7 levels in peripheral blood basophils and total basophil counts in tissues, we hypothesize that the reduction in 2D7 likely represents degranulation of basophils within the NP tissue. The degree of basophil degranulation, as determined by loss of 2D7, significantly correlated with enhanced sinus and asthma disease severity, measured by the LM score and FEV₁, respectively. Our findings support the hypothesis that basophils are recruited to NPs in AERD and are subsequently activated to release granular contents, suggesting that they may contribute to clinical disease.

In addition to increased basophil numbers in NPs, we found elevated basophils in the blood of patients with AERD. This novel observation may be secondary to our assessment of peripheral basophils using flow cytometry. In contrast, previous studies examining peripheral basophils used automated cell counters available in clinical laboratories.^22,23 Unfortunately, we are unable to directly compare the number of basophils assessed by flow cytometry with automated counts because routine blood cell counts with differentials were not drawn on the day of sinus surgery. Nonetheless, absolute basophil counts were available from some of our study patients before their sinus surgery. Comparison of this limited number of tests suggested no significant difference in circulating basophil numbers between patients with AERD or CRSwNP by automated count (P > .99). We suspect the higher sensitivity of the flow cytometry–based approach explains the difference in results.

The underlying pathophysiological mechanisms leading to the elevated number of basophils in both the circulation and NPs in AERD are unknown. Greater production, longer survival, and/or reduced exit from the circulation could all theoretically result in increased total numbers of basophils in the circulation of patients with AERD. Reduced exit seems less likely based on the increased numbers of basophils we observed in NPs. However, more basophil progenitors could be produced in the bone marrow and/or basophil progenitors in the bone marrow could have a longer productive life and produce greater numbers of basophils in AERD. The sinonasal tissue could also generate basophil-stimulating factors that work locally or activate the marrow remotely to generate more cells.^24,25 IL-3 is an important cytokine for basophil production, and studies evaluating levels in the circulation or bone marrow in AERD would be of interest.

Differential expression of various chemotactic factors may also contribute to the elevated numbers of NP basophils observed in AERD. Eotaxin-1, eotaxin-2, and eotaxin-3 are well-established basophil chemoattractant factors.²⁶ Compared with healthy sinonasal tissue, these chemokines are significantly elevated in NPs, although no difference was observed between AERD and CRSwNP.²⁷ Although eotaxins likely contribute to some degree of baseline recruitment, we hypothesize that other mediators may be involved in the selective recruitment of basophils in AERD. Because basophils express the prostaglandin D₂ receptor CRTH2 (chemoattractant receptor-homologous molecule expressed on T_H2 cells), it is possible that basophils are recruited by elevated levels of prostaglandin D₂, known to selectively occur in AERD.^12,13 It is also possible that the quantity of basophils recruited in AERD is similar to that recruited in CRSwNP but that basophils survive longer in AERD NPs as a result of local survival factors.

In addition to measuring the total number of basophils, we evaluated their activation status. CD63 is a marker of degranulation because it is exposed on the cell surface upon the fusion of the basophil granular and surface membranes. In our study, CD63 MFI was significantly upregulated on NP basophils in both CRSwNP and AERD compared with peripheral blood. In contrast, CD203c is a surface marker associated with basophil activation but not necessarily degranulation.^28,29 We found that CD203c MFI was elevated on NP basophils in both CRSwNP and AERD compared with peripheral blood. MacGlashan³⁰ has previously reported that upregulation of CD203c and CD63 expression on basophils may vary on the basis of type of stimulus and mode of degranulation (piecemeal vs anaphylactic). However, the mechanism by which basophils are activated in NPs remains unclear.

IgE remains one of the most well-known activators of basophils. However, in contrast to atopic conditions such as allergic rhinitis or asthma, neither CRSwNP nor AERD is thought to be due to conventional antigen-driven IgE-mediated hypersensitivity reactions. However, previous work from our group and others found elevated local production of IgE in NPs compared with the circulation of matched subjects or with healthy sinonasal tissue.^31,32 It is thus possible that local IgE is responsible in part for activating basophils in NPs.

Alternatively, basophils could be activated through non–IgE-mediated pathways. Various chemokines (eg, monocyte chemotactic proteins 1–3, eotaxins 1–3, and RANTES), proteases (which activate basophils through protease-activated [PAR] receptors), formyl methionine peptides (which activate basophils through specific G protein–coupled receptors), and complement split products (C3a, C5a) should be evaluated as other potential activating factors. C3a and C5a are especially intriguing given that we and others have reported elevated levels of these products in chronic rhinosinusitis.^33,34 It may also be that basophils are more selectively primed by cytokines such as IL-3 in AERD and thus are more easily activated by the same level of stimuli. Future studies investigating the mechanisms of production, survival, recruitment, priming, and activation of basophils in AERD (and CRSwNP) are needed.

We also assessed the intracellular expression with 2D7, an antibody that labels a secretory granule protein found in basophils.²¹ Previously, we reported a reduction in the number of 2D7⁺ basophils in NPs of AERD compared with CRSwNP by immunhistochemistry.¹⁰ This observation was confirmed by our current studies showing that basophils expressed less 2D7 (both as a percentage and as an average relative intensity) in NPs from patients with AERD compared with either matched blood or NPs from patients with CRSwNP. We hypothesize that basophils in AERD NPs lose their 2D7 staining on degranulation, thus making basophils more difficult to enumerate. We had initially expected that 2D7 MFI would inversely mirror that of CD63 MFI (because both are associated with degranulation). However, unlike 2D7 MFI, we did not see any difference in CD63 MFI between NP basophils from patients with CRSwNP and AERD, nor was there any significant correlation between CD63 and 2D7 MFI (data not shown). This dissociation of the 2 biomarkers may be due to differences in dose response or time course for 2D7 loss versus CD63 upregulation; alternatively, it is possible that CD63, once transported to the cell surface, may subsequently undergo further processing (eg, shedding, internalization, and cleavage). Further work is thus needed to understand the molecular basis for the dissociation of 2D7 downregulation and CD63 upregulation associated with basophil degranulation.

Although the biological significance of 2D7 remains to be fully elucidated, we found a significant correlation between loss of 2D7 intensity and upper and lower respiratory disease severity as measured by greater opacification on sinus computed tomography and reduced pulmonary function. This is the first reported association, to our knowledge, between 2D7 and clinical disease severity. However, our findings support the conclusions of Brescia et al,²³ who found that patients with CRSwNP who had peripheral basophil counts greater than 0.03 × 10⁹ cells/L had higher rates of sinonasal disease recurrence.

It is possible that the granule protein identified by 2D7 contributes to AERD pathogenesis in yet-to-be-discovered mechanisms. Activated basophils produce IL-4 and IL-13, which can stimulate type 2 inflammatory responses including mucus hypersecretion, induction of IgE class switching, and impairment in the epithelial barrier, all known to be dysregulated in NPs.³⁵ In addition, basophils are potent producers of cysteinyl leukotrienes, which are known to be elevated in AERD.^14,15 Interestingly, we recently identified a novel lipid mediator, 15-oxo-ETE, in AERD NPs that is generated through the conversion of arachidonic acid via the 15-lipoxygenase and hydroxyprostaglandin dehydrogenase enzymes.³⁶ Mast cells³⁶ and basophils³⁷ (http://www.proteinatlas.org) both highly express hydroxyprostaglandin dehydrogenase, and it is thus possible that basophils are an important source of 15-oxo-ETE.

A limitation of this study is that 45% and 64% of patients with CRSwNP and patients with AERD, respectively, were taking oral corticosteroids perioperatively, and this may have affected basophil numbers. To address this, we examined basophils numbers in both conditions according to steroid use. We found no difference in basophil numbers in NPs of either AERD or CRSwNP when stratified by history of intranasal, inhaled, or systemic corticosteroid use (see Fig E4 in this article’s Online Repository at www.jacionline.org). Furthermore, Lampl et al³⁸ reported in the last millennium that chronic treatment with steroids in patients with severe asthma or connective tissue disorders led to a resistance in the in vitro inhibitory effects of dexamethasone on basophils.

In this study, patients with AERD had not undergone an aspirin desensitization nor were they taking a daily high-dose aspirin regimen. In work by Cahill et al,³⁹ peripheral basophil numbers increased in those patients with AERD who had been successfully desensitized and maintained on high-dose aspirin for 8 weeks. This increase in peripheral basophils may represent a reduction in basophil recruitment to NPs and thus potentially reflect the underlying mechanism by which patients experience clinical improvement.

Conclusions

Basophils play an important role in promoting a type 2 inflammatory response in various allergic conditions. This is the first study to report significantly elevated numbers and activation of basophils in NPs compared with healthy sinonasal tissue. Patients with AERD had particularly high elevations of basophils not only in their NPs but also in their peripheral blood. Although basophils are activated in NPs of both CRSwNP and AERD, basophils appear to have degranulated to a greater extent in AERD. Levels of basophil recruitment and activation showed clear correlations with the physiological and pathological measures of disease severity, FEV₁, and the LM score.

We thank Dr Lawrence Schwartz for his generous gift of the 2D7 mAb used to identify basophils by immunohistochemistry in Fig E3.

Key messages.

• Basophils are elevated in NPs and peripheral blood of patients with AERD compared with patients with CRSwNP.

• Basophils are activated in NPs of both patients with CRSwNP and patients with AERD; however, basophils appear to have degranulated to a greater extent in patients with AERD.

• The extent of basophil degranulation significantly correlated with clinical markers of enhanced disease severity.

Abbreviations used

AERD

Aspirin-exacerbated respiratory disease

CRSwNP

Chronic rhinosinusitis with nasal polyps

LM

Lund-Mackay

MFI

Mean fluorescence intensity

NP

Nasal polyp