Th2-skewed T cells correlate with B cell response to α-Gal and tick antigens in α-Gal syndrome

In total, 50 patients with AGS and 19 individuals without AGS (controls) were enrolled in the study. All patients had IgE to α-Gal (median 18.5 kUA/L [range 0.76 to greater than 100 kUA/L]), and all except 1 patient reported being tick bitten. Most of the patients (41/49) also had IgE to I. ricinus TE (median 1.1 kUA/L [range < 0.1 – 14 kUA/L]). Of the 19 controls, 9 reported being tick bitten, but all were IgE negative to α-Gal and TE. There was no difference in the proportions of men and women between patients with AGS and controls (P = 0.43) or in which season the samples were collected (winter/spring or summer/autumn, P = 0.29), but the patients with AGS were significantly older than the controls (P = 0.008). For detailed characteristics of the patients, see Table 1 and of the controls see Table 2.

Table 1

Characteristics of patients with AGS

Table 2

Characteristics of healthy controls

CD4+ T cells from patients with AGS proliferate more in response to TE than CD4+ T cells from healthy controls. Peripheral blood mononuclear cells (PBMCs) labeled with carboxyfluorescein succinimidyl ester (CFSE) were cultured for 7 days in the presence of different stimulants for analysis of T cell proliferation. The gating strategy for proliferating T cells is depicted in Figure 1A. Analysis of CFSE dilution revealed that T cell proliferation in patients with AGS was dose-dependent, where 10 μg/mL TE gave the strongest response and was significantly higher than at 0.1 μg/mL (Figure 1B, P = 0.007, n = 13). T cells from patients with AGS and controls proliferated in response to TE (Figure 1C, P < 0.001, n = 35 [AGS]; P = 0.001, n = 13 [controls]), but there was significantly higher proliferation in patients with AGS compared with controls (Figure 1D, P = 0.043, n = 35 [AGS] and n = 13 [control]). Interestingly, there was no difference in the T cell proliferation between tick-bitten and nonbitten controls (Supplemental Figure 1; supplemental material available online with this article; https://doi.org/10.1172/JCI158357DS1). To test the α-Gal dependence of the T cell proliferation, the PBMCs were also stimulated with deglycosylated TE, in which the α-Gal epitope had been enzymatically removed by α-galactosidase treatment. There was no difference in proliferation of T cells to deglycosylated TE compared with unaltered TE (Figure 1E, P = 0.28, n = 10 [AGS]; P = 0.85, n = 13 [controls]). However, in 5 patients with AGS and 3 controls, the proliferation was reduced by more than 22%, whereas only 1 patient and 1 control showed an increased proliferation of more than 22%. Furthermore, there was a statistically significant decrease in proliferation of T cells from patients with AGS when stimulated with a nontick α-Gal–containing protein compared with unstimulated cells (Figure 1F, P = 0.04, n = 14). Still, the difference was less than 3% for all individuals, and the levels of proliferation were less than 6% for all data points. In addition, patients with AGS’s T cell proliferation did not correlate with IgE levels to α-Gal or TE (ρ = –0.187, P = 0.28, n = 35, and ρ = –0.257, P = 0.14, n = 35, respectively).

T cell proliferation measured by dilution of CFSE.Figure 1

T cell proliferation measured by dilution of CFSE. (A) Gating strategy for proliferation of CD4+ T helper cells. (B) T cell proliferation in response to different doses of TE in patients with AGS, Friedman test with Dunn’s multiple comparisons test, **P < 0.01, n = 13. (C) T cell proliferation to TE compared with unstimulated cells in patients with AGS (left, n = 35) and healthy controls (right, n = 16), Wilcoxon matched-pairs signed rank test, **P < 0.01, ***P < 0.001. (D) Comparison of patients with AGS and healthy controls, Mann-Whitney U test, *P < 0.05, n = 35 (AGS) and n = 13 (controls). (E) T cell proliferation after removal of α-Gal from the TE in patients with AGS (left, n = 10) and healthy controls (right, n = 13), Wilcoxon matched-pairs signed rank test. (F) T cell proliferation to an α-Gal containing nontick protein in patients with AGS, Wilcoxon matched-pairs signed rank test, *P < 0.05, n = 14. Each point within the box plot represents 1 individual. Box plots represent IQR and median, whiskers extend to the farthest data points.

Patients’ cytokine secreting cells are Th2-skewed after activation by TE. PBMCs were cultured for 40 hours together with TE in FluoroSpot plates to detect cells secreting specific cytokines. An example of IL-3, IL-4, IL-13, IL-31, IFN-γ, IL-10, and IL-22 secreting cells detected by FluoroSpot with unstimulated cells, TE- or phytohemagglutinin-stimulated (PHA-stimulated) cells from an AGS patient is depicted in Figure 2A. Overall, cytokine production was detected in most patients, whereas it was detected in only a few of the healthy controls. Patients (n = 24 unless otherwise stated) had a Th2 profile of secreted cytokines, where the number of IL-3, IL-4, and IL-31 secreting cells were significantly increased compared with controls (n = 8; Figure 2B, P = 0.001, n = 23; Figure 2C, P = 0.0027; and Figure 2E, P = 0.0026, respectively). In fact, IL-31 was not detectable in any of the controls. The number of IL-13 secreting cells was also increased in patients compared with controls, but not significantly (Figure 2D, P = 0.061). In contrast with what was observed for the Th2 cytokines, there was no difference between patients with AGS and healthy controls in the number of cells secreting the Th1 cytokine IFN-γ (Figure 2F, P = 0.91), the Th22 cytokine IL-22 (Figure 2G, P = 0.24, n = 21), and the Treg cytokine IL-10 (Figure 2H, P = 0.21). IL-5 and IL-17 secreting cells were not detectable in most individuals after TE stimulation. When the PBMCs were cultured in the presence of PHA as positive controls, the numbers of IL-3, IL-4, and IL-31 secreting cells were higher in patients with AGS than in healthy individuals (P = 0.05 [n = 23 for patients], P = 0.02 and P = 0.002, respectively).

Cytokine expression by PBMCs in response to TE from patients with AGS and hFigure 2

Cytokine expression by PBMCs in response to TE from patients with AGS and healthy controls. (A) Representative photos of FluoroSpot wells from a patient with AGS for unstimulated cells, TE-stimulated cells, and PHA-stimulated cells. (B) IL-3, (C) IL-4, (D) IL-13, (E) IL-31, (F) IFN-γ, (G) IL-22, and (H) IL-10. Mann-Whitney U test, **P < 0.01, n = 24 (AGS, n = 23 for IL-3) and n = 8 (control). Each point within the box plot represents 1 individual. Box plots represent IQR and median, whiskers extend to the farthest data points.

To investigate the effect of α-Gal on cytokine secretion, AGS patients’ cells were also stimulated with deglycosylated TE in the FluoroSpot assay (n = 14 unless otherwise stated). For IL-3 and IL-31 the number of cytokine secreting cells was unchanged for deglycosylated TE compared with unaltered TE (Figure 3A, P = 0.68, and Figure 3D, P = 0.20, respectively), whereas for IL-4 and IL-13 there was a significant decrease in the number of cytokine secreting cells (Figure 3B, P = 0.02, and Figure 3C, P = 0.008, respectively). When analyzing IFN-γ, IL-22, and IL-10, a substantial reduction in the number of cytokine secreting cells was noted for all of them (Figure 3E, P = 0.02, Figure 3F, P = 0.03, n = 11, and Figure 3G, P = 0.04, respectively).

Cytokine secreting cells in PBMCs from patients with AGS after stimulationFigure 3

Cytokine secreting cells in PBMCs from patients with AGS after stimulation with TE and deglycosylated TE. (A) IL-3, (B) IL-4, (C) IL-13, (D) IL-31, (E) IFN-γ, (F) IL-22, and (G) IL-10. Wilcoxon matched-pairs signed rank test. *P < 0.05, and **P < 0.01. n = 14 for all except IL-22, where n = 11. Each point within the box plot represents 1 individual. Box plots represent IQR and median, whiskers extend to the farthest data points.

B cells highly express the activation marker CD23 after stimulation with TE. The gating strategy for B cells expressing CD23 after PBMCs were cultured for 20 hours in the presence of different stimulants is depicted in Figure 4A. We found that B cells expressed higher levels of CD23 after PBMCs were stimulated with TE compared with unstimulated cells (Figure 4B, P < 0.001, n = 30 [AGS]; P < 0.001, n = 18 [control]), but the CD23 expression was significantly higher in patients with AGS than in healthy controls (Figure 4C, p =0.028, n = 30 [AGS] and n = 18 [control]). Further investigation of the CD23 response in patients with AGS showed that upregulation of CD23 by TE could be inhibited with anti-CD40L and anti-IL-4 antibodies (Figure 4D, P = 0.002 and P < 0.001, respectively, n = 19), but not with an isotype-matched antibody control (Supplemental Figure 2A, P = 0.81, n = 19). Stimulation with deglycosylated TE led to significantly increased CD23 expression compared with TE stimulation in patients with AGS, whereas CD23 expression did not change for the controls (Figure 4E, P = 0.0078, n = 9 [AGS]; P = 0.40, n = 13 [control]). Furthermore, stimulation with a nontick related protein carrying the α-Gal epitope did not affect CD23 expression compared with unstimulated cells (Figure 4F, P = 0.89, n = 21 [AGS]; P = 0.96, n = 15 [control]). The CD23 expression was significantly higher from naive cells compared with memory cells in patients with AGS, defined as CD27+ (memory) and CD27-IgD+ (naive), after stimulation with TE (Figure 4G, P < 0.001, n = 17). Moreover, the CD23 expression strongly correlated with T cell proliferation in patients with AGS (Figure 4H, ρ = 0.792, P < 0.001, n = 19). However, the CD23 expression did not correlate to the α-Gal or TE-specific IgE levels (ρ = 0.132, P = 0.49 and ρ = 0.010, P = 0.96, respectively, n = 30).

B cell expression of CD23.Figure 4

B cell expression of CD23. (A) Gating strategy for CD23-expressing B cells. (B) CD23 expression in unstimulated compared with TE stimulated B cells in patients with AGS (left, n = 30) and healthy controls (right, n = 18), Wilcoxon matched-pairs signed rank test, ***P < 0.001. (C) Comparison of patients with AGS and healthy controls, Mann-Whitney U test, *P < 0.05, n = 30 (AGS) and n = 18 (controls). (D) Effect of inhibition with anti-CD40L and anti-IL-4 antibodies in patients with AGS, Friedman test with Dunn’s multiple comparisons test, **P < 0.01, ***P < 0.001, n = 19. (E) Effect of removing α-Gal from the TE in patients with AGS (left, n = 9), and healthy controls (right, n = 13), Wilcoxon matched-pairs signed rank test, **P < 0.01. (F) Comparison of CD23 expression in unstimulated B cells compared with B cells stimulated with an α-Gal containing nontick protein in patients with AGS (left, n = 21) and healthy controls (right, n = 15). (G) Comparison of CD23 expression by naive (CD27-IgD+) and memory (CD27+) B cells after TE stimulation in patients with AGS, Wilcoxon matched-pairs signed rank test, ***P < 0.001, n = 17. Each point within the box plot represents 1 individual. Box plots represent IQR and median, whiskers extend to the farthest data points. (H) Correlation of T cell proliferation and CD23 expression in response to TE in patients with AGS, Spearman’s rank correlation, ρ = 0.792, P < 0.001, n = 19. Each point represents 1 individual.

B cell proliferation in response to TE stimulation is partly α-Gal specific in patients with AGS. Similar to T cells, CFSE dilution in B cells was analyzed after 5 days of culturing PBMCs in the presence of various stimulants. The gating strategy for proliferating B cells is depicted in Figure 5A. TE induced a significant proliferation in B cells from both patients with AGS and controls (Figure 5B, P < 0.001, n = 28, and P = 0.006, n = 15, respectively). However, the proliferation was significantly higher in patients with AGS compared with controls (Figure 5C, P < 0.001, n = 28 (AGS) and n = 15 (control)). B cells also proliferated in patients that did not show high CD23 expression. Furthermore, B cell proliferation in patients with AGS was significantly lower in response to deglycosylated TE compared with stimulation with unaltered TE (Figure 5D, P = 0.0137, n = 10), which was not observed in the controls (Figure 5D, P = 0.19, n = 13). B cell proliferation showed a similar dose dependency in patients with AGS as T cells did, with significantly reduced proliferation at the lower doses of TE (Figure 5E, P = 0.008 and P < 0.001 for 1 μg/mL and 0.1 μg/mL, respectively, n = 22). Proliferation in response to TE stimulation was inhibited by blocking with anti-CD40L, but not anti-IL-4 antibodies (Figure 5F, P = 0.002 and P > 0.99, respectively, n = 16) or with an isotype-matched antibody control (Supplemental Figure 2B, P = 0.11, n = 12). B cells from patients with AGS also showed a significant proliferative response to a nontick protein containing α-Gal, but the increase was less than 4 % (Figure 5G, P = 0.04, n = 20), whereas no proliferation was seen for the controls (Figure 5G, P = 0.75, n = 10). There was no difference in the proliferation of naive and memory B cells in response to TE in patients with AGS (Figure 5H, P = 0.94, n = 28). Furthermore, the B cell proliferation did not correlate with the IgE levels to α-Gal or TE (ρ = 0.306, P = 0.11, n = 28 and ρ = 0.177, P = 0.37, n = 28, respectively), or with T cell proliferation or CD23 expression (Supplemental Figure 3, ρ = 0.05, P = 0.84, n = 18 and ρ = 0.143, P = 0.48, n = 27, respectively).

B cell proliferation measured by CFSE dilution.Figure 5

B cell proliferation measured by CFSE dilution. (A) Gating strategy for proliferation of CD3-CD19+ B cells. (B) Proliferation of unstimulated compared with TE stimulated B cells in patients with AGS (left, n = 28) and healthy controls (right, n = 15), Wilcoxon matched-pairs signed rank test, **P < 0.01 and ***P < 0.001. (C) Comparison of patients with AGS and individuals acting as healthy controls, Mann-Whitney U test, ***P < 0.001, n = 28 (AGS) and n = 15 (controls). (D) Effect of removing α-Gal from the TE in patients with AGS (left, n = 10) and individuals acting as healthy controls (right, n = 13), Wilcoxon matched-pairs signed rank test, *P < 0.05. (E) B cell proliferation in response to different doses of TE in patients with AGS, Friedman test with Dunn’s multiple comparisons test, **P < 0.01, ***P < 0.001, n = 22. (F) Effect of inhibition with anti-CD40L and anti-IL-4 antibodies in patients with AGS, Friedman test with Dunn’s multiple comparisons test, **P < 0.01, n = 16. (G) Comparison of B cell proliferation in unstimulated cells and cells stimulated with an α-Gal containing nontick protein in patients with AGS (left, n = 20) and individuals acting as healthy controls (right, n = 10), Wilcoxon matched-pairs signed rank test, *P < 0.05. (H) Comparison of proliferation in naive (CD27-IgD+) and memory (CD27+) B cells in patients with AGS, Wilcoxon matched-pairs signed rank test, n = 28). Each point within the box plot represents 1 individual. Box plots represent IQR and median, whiskers extend to the farthest data points.

Patients with AGS’s antibody reactivity to TE is dominated by antibodies against α-Gal. To assess to what extent the patients with AGS’s IgE reactivity to TE was directed toward α-Gal, Western blots were performed to detect IgE binding to TE and deglycosylated TE. IgE antibodies from a pool of AGS patient sera bound strongly to TE, whereas a much weaker binding to deglycosylated TE was observed (Figure 6A). The serum of a healthy control did not react with TE or deglycosylated TE (Supplemental Figure 4).

Patient antibody responses to TE.Figure 6

Patient antibody responses to TE. (A) Binding of a serum pool sample from a patient with AGS (n = 8) in a Western blot. MM = molecular marker, TE = tick extract, deglycoTE = deglycosylated TE, Enz. = α-galactosidase from green coffee bean. (B) Inhibition of IgE and IgG1 antibody binding to tick extract by α-Gal in patients with AGS. Wilcoxon matched-pairs signed-rank test, ***P < 0.001, for IgE n = 14, for IgG1 n = 23. Each pair of points connected by a line represents 1 individual. (C) Correlation of the inhibition of IgE and IgG1 antibody binding to TE by α-Gal. Spearman’s correlation coefficient ρ = 0.609, P < 0.05, n = 14. Each point represents 1 individual.

We also investigated the proportion of the IgG1 and IgE reactivity to TE that could be blocked by α-Gal in an inhibition ELISA. IgE binding to TE was detected in 14 of 24 patients with AGS tested, and the binding was inhibited up to 80% (median 38.1 %, range 0.0–81.5 %, Figure 6B, P < 0.001) by α-Gal disaccharide. IgG1 binding to TE was detected in 23 of 24 patients with AGS tested and the α-Gal disaccharide inhibited the response up to 100% (median 70.2 %, range 0.0–100.0 %, Fig. 6B, P < 0.001). Two of 9 tested controls had detectable IgG1 levels to TE, but these were lower than the levels detected in patients and were inhibited 6.3 % and 58.9 %, respectively, by α-Gal disaccharide. IgE levels to TE could not be detected in the controls, in line with the inclusion criteria. There was a moderate correlation between the inhibition of IgG1 and IgE binding in patients with AGS (Figure 6C, ρ = 0.609, P = 0.02, n = 14).

留言 (0)

沒有登入
gif