We have studied the influence of wasp venom immunotherapy (VIT) on cellular recruitment and cytokine mRNA expression during allergen-induced cutaneous late-phase responses (LPR). Nine subjectswith a history of wasp sting anaphylaxis, and specific IgE in their sera underwent wasp VIT. Skin biopsies were taken 24 h after intradermal diluent and allergen before and after 3 months VIT. Pre-immunotherapy, there were significant allergen-induced increases in EG2+ eosinophils, elastase+ neutrophils, CD68+ macrophages and IL–10 protein+ cells, and increased expression of mRNA for IL–4, IL–13, IFN–γ, IL–12, IL–10, TGF-β, RANTES and eotaxin. When these allergen-induced changes in cytokine mRNA and cellular profiles were compared with those obtained after 3 months VIT there was a significant reduction in IL–4 mRNA (p=0.012) and increase in IL–10 protein+ cells (p=0.004) with a trend to an increase in IL–10 mRNA (p=0.054). There were also significant reductions in eosinophils (p<0.004) and the size of the cutaneous LPR (p<0.01) but no change in mRNA to IFN–γ, IL–13 or IL–12. Therefore, VIT is associated with a significant increase in cells positive for IL–10 protein but not IL–12 or IFN–γ. These results suggest that induction of IL–10 may be important in VIT and occur independently of the switch to a Th1 phenotype. IL–10 generation may down-regulate IL–4 expression and eosinophil recruitment.
Immunotherapy is highly effective in patients severely allergic to wasp stings. A number of hypotheses have been advanced to explain the mechanism. Specific IgG levels, particularly subclassesIgG1 and IgG4, increase during immunotherapy with both inhalant and venom allergens 1–6 and have been postulated to competitively inhibit allergen-specific IgE. However, this theory found little support from a number of studies demonstrating a lack of correlation between total venom-specific IgG and a reduced sensitivity to stings 5–9. With a greater understanding of the Th1-Th2 paradigm it became established that allergy was associated with Th2-type cytokines and the generation of IL–4 and IL–13 leading to IgE isotype selection by B lymphocytes and generation of IL–5 resulting in eosinophilia 10–12. By contrast, in the normal immune response to allergen, the Th1-type cytokine IFN–γ induces an IgG and particularly an IgG4 isotype switch by memory B cells 3. We and others have shown that successful venom immunotherapy is associated with a Th2 to Th1-type cytokine switch 13, 14. Down-regulation of IgE is an important component to any hypothesis explaining the mechanism of immunotherapy. However, absolute levels of specific IgE cannot entirely explain desensitization during immunotherapy because, although there is a long-term reduction, there is also an initial elevation in specific IgE levels ata time when clinical tolerance has been achieved 4. Therefore, an IL–4 to IFN–γ switch does not fully explain the mechanism of IT particularly in the early stages, although achange of balance between IgE and IgG may be important.
The late-phase response (LPR) to allergen is a well studied model of allergic inflammation and there is substantial evidence that the efficacy of immunotherapy is related to its ability to inhibit allergen-induced late-phase responses in the skin [15–17] and in the lung [18–20]. This in-vivo model has given us the opportunity to study the global allergen-induced cellular and cytokine transcriptional changes within the dermal milieu, without the need to study ex vivo peripheral blood cells stimulated in isolation. In the present study we describe the effect of allergen on the cutaneous LPR during venom immunotherapy in wasp allergic patients by examining skin biopsies taken before and after 3 months VIT at a time when the subjects were tolerant to wasp venom. We report that after 3 months VIT there is an increase in both the transcription and translation of the regulatory cytokine IL–10 which is associated with a reduction in size of the LPR to allergen, a reduction in the recruitment of eosinophils and the transcription of pro-inflammatory cytokines.
2.1 Subjects and the late-phase cutaneous response
All subjects gave a history of a severe life-threatening reaction to a wasp sting and had positive wasp venom-specific serum IgE (Table 1). All subjects exhibited a positive early cutaneous reaction to intradermal wasp venom at 20 min. The size of the late skin responses after intradermal injection of wasp venom was recorded before and after 3 months VIT (Fig. 1A). Eight out of the nine subjects exhibited a positive cutaneous LPR to intradermal wasp venom. There was a decrease in the size of the LPR (calculated as the mean of two diameters) in all eight of these subjects from a mean diameter of 143 (19.9) mm before VIT to 50.4 (17.6) mm after 3 months VIT (p<0.01; Fig. 1A). Subject 9 did not have a visible late-phase dermal response either before or after 3 months VIT but was included in the subsequent analysis. There was no difference in the magnitude of reduction of the LPR when the three atopic subjects were compared with the six non-atopic subjects.
All nine subjects continued with wasp VIT and completed the 3 year program of desensitization. Seven individuals were able to tolerate a dose equivalent to 2 stings at the time of the second biopsy at 3 months from the start of VIT and were maintained on this dose. Subjects 1 and 2 developed side effects (hypotension and laryngeal edema, respectively) and so the dose schedule was modified so that by 3 months they were able to tolerate the equivalent of 1.5 stings.
|Subject||Sex||Age||Atopy||Sp. IgE||LPR (mm)||RAST-venom (KU/l)||Symptoms|
|1||F||45||A, R||Dp, Cat||88||4.6||Hypotension, LE|
|2||F||30||A, SR||Dp, GP||108||45||Hypotension|
|8||M||40||–||Dp, GP||155||> 100||Hypotension|
2.2 Immunohistochemistry of cutaneous late-phase responses
At baseline and after 3 months VIT, the immunohistology of the cutaneous LPR was examined in all nine subjects. Cell counts in the dermis 24 h after intradermal provocation with wasp venom or with allergen diluent are presented in Table 2.
|Baseline (Pre-VIT) mean (SEM)||After 3 months VIT mean (SEM)||Pre-vs. post-VIT|
|Monoclonal antibody||Control||Wasp venom||Mean difference||Control||Wasp venom||Mean difference||Mean difference||p value|
|Mast cells||26.7 (5.4)||45.9 (7.8)||19.2||31.1 (4.8)||42.3 (6.9)||11.2||− 8.0||0.496|
|Eosinophils||2.4 (1.7)||144 (25.1)||141.7**||6.1 (4.2)||75.6 (15.9)||69.2**||− 72.5||0.004|
|Neutrophils||4.4 (2.9)||160 (37.6)||155.9**||35 (24.3)||228 (44.4)||193.3||37.4||1.0|
|CD68+||31.3 (9.5)||108 (14.2)||76.4**||65.2 (20.1)||136 (18.7)||70.3||− 6.1||0.82|
|CD3||9.1 (2.6)||42.8 (4)||33.7**||10.3 (2.8)||32.4 (5.1)||22.1**||− 11.6||0.097|
|CD4||7.5 (2.2)||28.3 (2.9)||20.8**||8.1 (2.6)||26.9 (4.4)||18.8**||− 2.0||1.0|
|CD8||2.4 (0.9)||14.8 (2.3)||12.4**||3.6 (1.3)||11.5 (2.9)||7.9**||− 4.5||0.25|
|CD25||6.5 (2.3)||28.7 (4.1)||22.2**||6.1 (1.4)||14.7 (3.5)||8.6*||− 13.6||0.05|
|IL-10 protein||1.1 (0.3)||14.6 (3.4)||13.5**||2.4 (1.1)||37.6 (7.2)||35.2**||21.7||0.004|
2.3 Diluent control sites
A comparison of cell counts at the diluent control sites before and after 3 months VIT showed that numbers of mast cells, T lymphocytes (CD3+), CD4+ and CD8+ cells and cells with the activation marker CD25+ were almost identical. Although, there were increases in mean numbers of eosinophils, neutrophils, macrophages (CD68+), and IL–10 protein+ cells after 3 months VIT at the control site these were not significantly different.
Baseline (pre-VIT) cell counts 24 h after intradermal allergen were compared with diluent sites. There were significant allergen-induced increases in eosinophils (p<0.01), neutrophils (p<0.01), CD68+ macrophages (p<0.01), T lymphocytes (CD3+) (p<0.01) and in CD4+ (p<0.01) and CD8+ (p<0.01) T cell subsets. There were also significant increases in cells staining positive for the activation marker CD25 and for cells staining positive for IL–10 protein (p<0.01). Tryptase+ mast cells were the only cell type examined that did not increase significantly during the late-phase dermal response to intradermal allergen when compared to diluent.
2.5 After 3 months VIT
After 3 months of VIT, intradermal allergen again caused an increase in cellular infiltrate for the majority of cell types compared to diluent. There were significant allergen-induced increases in eosinophils (p<0.01), T lymphocytes (CD3+) (p<0.01) and in CD4+ (p<0.01) and CD8+ (p<0.01) T cell subsets. There was also a smaller but significant increase in cells staining positive for the activation marker CD25 (p<0.05) but a larger increase in cells staining positive for IL–10 protein (p<0.01). Neutrophils (elastase+), CD68+ macrophages and tryptase+ mast cells did not increase significantly during the late-phase allergen-induced dermal response when compared to diluent after 3 months VIT.
2.6 Before and after 3 months VIT comparisons
The magnitude of the net increases in cell counts (wasp venom minus control) was compared between the two time points (before and after 3 months VIT). After 3 months VIT there was a significant reduction in allergen-induced recruitment of eosinophils (p<0.01) (Fig. 1B) and a trend for reduction in cells positive for the activation marker CD25 (p=0.05) and CD3+ cells (p=0.097). Cells immunostaining positive for IL–10 protein showed a highly significant allergen-induced net increase after 3 months VIT (p=0.004) (Fig. 1C). There were no significant differences between the two time points in the cellular response to allergen with regard to tryptase+ mast cells, elastase+ neutrophils, CD68+ macrophages or for CD4+ and CD8+ T cell subsets. After 3 months VIT there was no difference in the magnitude of change in allergen-induced profile for any cell type when the 3 atopic subjects were compared with the six non-atopic subjects.
2.7 In situ hybridization of cutaneous late-phase responses
At baseline and after 3 months VIT, in situ hybridization of the cutaneous LPR was used to analyze mRNA expression. At control sites, there were only small numbers of cells positive for hybridization signals both before and after 3 months VIT and there were no significant differences between these two points. At baseline and after 3 months VIT, cells positive for hybridization signals 24 h after intradermal allergen were compared with diluent sites. At both time points compared to diluent there were significant allergen-induced increases of mRNA for IL–4 (p<0.01), IL–10 (p<0.01), IFN–γ (p<0.01), IL–12 (p<0.01), IL–13 (p<0.01), TGF-β (p<0.01), eotaxin (p<0.01) and RANTES (p<0.01). Mean numbers of cells in the dermis expressing positive hybridization signals for each cytokine 24 h after intradermal provocation with wasp venom or with allergen diluent are presented in Table 3.
2.8 Before and after 3 months VIT comparisons
The magnitude of the net increases in cells expressing mRNA (allergen minus control) were compared between the two time points (before and after 3 months VIT). After 3 months VIT there was a significant reduction in allergen-induced mRNA of the Th2-type cytokine IL–4 (p=0.012) (Fig. 1D) but this was not associated with an increase in signal to the Th1-type cytokine IFN–γ (p=0.73). Similarly, intradermal allergen lead to a significant decrease in cells positive for mRNA to eotaxin (p=0.039) consistent with the reduction in eosinophil recruitment seen with immunohistochemistry. In addition there was a reduction in positive signal to RANTES mRNA (p=0.023). There was a compelling trend for an increase in cells positive for hybridization signals to IL–10 mRNA (p=0.054) consistent with our data from immunohistochemistry demonstrating a significant increase in numbers of IL–10 protein+ cells. No significant changes were seen for IL–12 (p=0.91), IL–13 (p=0.3) or TGF-β (p=0.36). There was a significant positive correlation between the net change in IL–10 protein+ cells and cells expressing signal for IL–10 mRNA (Pearsons r=0.7224; p=0.028) comparing before and after 3 months VIT. After 3 months VIT there was no difference in the magnitude of change in allergen-induced mRNA signal for any cytokine when the three atopic subjects were compared with the six non-atopic subjects.
|Baseline (Pre-VIT) mean (SEM)||After 3 months VIT mean (SEM)||Pre-vs. post-VIT|
|ISH to RNA for||Control||Wasp venom||Mean difference||Control||Wasp venom||Mean difference||Mean difference||p value|
|IL-4||0.9 (0.3)||51.2 (9.4)||50.3**||1.5 (0.4)||29.6 (3.7)||28.1**||− 22.2||0.012|
|IL-10||1.3 (0.6)||22.6 (4.0)||21.3**||2.7 (0.6)||34.4 (5.1)||31.7**||10.4||0.054|
|IFN-γ||0.2 (0.1)||13.7 (2.8)||13.5**||1.7 (0.7)||16.2 (3.4)||14.5**||1.0||0.73|
|IL-12||2.9 (1.5)||41.2 (5.5)||38.3**||3.3 (1.0)||39.9 (7.1)||36.6**||− 1.7||0.91|
|IL-13||3.9 (1.4)||57.4 (10.1)||53.5**||3.4 (1.1)||50.7 (7.9)||47.3**||− 6.2||0.3|
|TGF-β||6.0 (0.9)||48.3 (6.6)||42.3**||8.7 (1.1)||62.3 (7.1)||53.6**||11.3||0.36|
|Eotaxin||6.3 (1.7)||80.5 (8.6)||74.2**||8.4 (1.8)||63.7 (8.6)||55.3**||− 18.9||0.039|
|RANTES||4.2 (1.6)||65.9 (9.0)||61.7**||5.4 (1.4)||49.5 (8.6)||44.1**||− 17.6||0.023|
We have used the allergen-induced cutaneous late-phase model to study the cellular and cytokine transcriptional changes during venom immunotherapy. Although previous studies of successful VIT have shown a cytokine switch from a Th2 to Th1-type profile 13, 14 the present study is the first to investigate this in vivo. This is the most appropriate model of VIT available as it allows an in-vivo study of the changing response to allergen within the dermal milieu without the requirement for ex vivo stimulation of isolated cells. This model is especially suitable because venom immunotherapy is conducted via the cutaneous route and local changes in the skin are likely to mirror the effectiveness of systemic desensitization to an insect which imparts allergen by stinging through the skin. There is evidence that immunotherapy is associated with inhibition of allergen-induced late-responses, although this has not previously been demonstrated in VIT. Immunotherapy with grass pollen resulted in a significant reduction in the cutaneous LPR and was associated with an improvement in clinical scores 16. In allergic asthmatic children, desensitization for 1 year with Dermatophagoides pterynissinus resulted in either substantial improvement or complete resolution of the late-asthmatic response to inhaled allergen 19 and correlated with clinical improvement 20.
In the present study we have found an increase in transcription and expression of the regulatory cytokine IL–10 in skin biopsies taken during the late-phase allergic response after induction of venom immunotherapy. In addition, there was a significant reduction in allergen-induced hybridization signal to the Th2-type cytokine IL–4 and this was accompanied by a reduction in the size of the allergen-induced cutaneous late-phase response, recruitment of eosinophils, transcription of eotaxin and RANTES and by a trend for a decrease in CD25+ cells. Our results showed no overall reduction in numbers of allergen-induced granulocytes after 3 months of VIT. This is in accord with our data on hybridization signal for mRNA to the neutrophil chemoattractant TGF-β that also did not change. However, pre-immunotherapy, our data demonstrates a significant increase in numbers of allergen-induced neutrophils but there was no such increase after 3 months VIT perhaps as a result of down-regulation by IL–10 on neutrophil recruitment and chemoattractant cytokines such as IL–8 and TNF–α 21, 22. There was however, an overall increase in numbers of neutrophils after 3 months VIT and this may reflect the repetitive dermal stimulation by venom during the 3-month period of immunotherapy. We did not look for IL–8 and other specific neutrophil chemoattractants in this in vivo system because we had to limit the size of the skin biopsies for ethical reasons to target only those cells and cytokines which had the greatest likelihood of demonstrating a change after immunotherapy.
We have also demonstrated increased transcription and translation of IL–10 during induction of VIT. Our results suggest that IL–10 is pivotal to the mechanism of VIT and there is emerging evidence for this view if the actions of IL–10 are considered. The term Cytokine Synthesis Inhibiting Factor (CSIF) was originally coined for IL–10 when first discovered and found to inhibit the generation of Th1-type cytokines 23. It has since been found to suppress not only Th1-type cytokine release but to also regulate cytokine synthesis in a number of antigen-dependent systems. IL–10 was not able to inhibit human T lymphocytes in isolated culture but only suppressed antigen-specific proliferation when monocytes were used as antigen-presenting cells (APC) 24 or in in vivo systems [25–27]. This further strengthens the case for the use of an in-vivo model such as the late-phase cutaneous response before any conclusions can be reached on the observed T cell cytokine changes. The mechanism for the APC-dependence was subsequently explained when it was discovered that IL–10 blocks the CD28-B7.1 interaction and consequent costimulatory signaling pathways in T lymphocytes 28. This confirmed that IL–10 does not act directly on T cell activation but has an indirect function by inhibiting APC and by inference may be considered a regulator of antigen-dependent T cell proliferation and cytokine release.
There is a substantial body of evidence demonstrating that IL–10 acts as an inhibitor of Th-2 cell responses which are key to allergic inflammation. In vitro, IL–10 inhibits antigen-specific proliferation of peripheral blood T cells and all CD4+ T cells including Th0, Th1 and Th2 subsets 24, 29. IL–10 inhibits IL–5 generation from human T cells 28 and has been shown to suppress antigen-induced IL–5 generation and eosinophil recruitment in murine models of allergic inflammation 26, 27. In addition, IL–10 down-regulates CD-40 mRNA expression and granulocyte-macrophage colony-stimulating factor production by activated eosinophils and promotes eosinophil apoptosis 30, 31. In a murine model, administration of IL–10 before epicutaneous allergen application, induced antigen-specific T cell tolerance 32. More recently, antigen-specific airway inflammation in mice was found to be abrogated by concurrent gene expression of IL–10 with a reduction in numbers of eosinophils, and decreased levels of IL–4, IL–5 and TNF–α. This effect of IL–10 could be reproduced in a mouse IFN–γ knockout model, suggesting that the actions of IL–10 are independent of IFN–γ expression 33 and contribute further momentum to the concept of a regulatory role for IL–10 in immunotherapy. We found an antigen-induced reduction in dermal eosinophils and postulate that this resulted from the up-regulation of IL–10 found concurrently in the skin biopsies from our patients undergoing VIT. Although, the cellular source of IL–10 was not studied it is known that this cytokine, like TGF–β is produced bymany cell types including T cells, macrophages and eosinophils. Similarly, as we have previously demonstrated by double-staining, eosinophils and mast cells as well as CD3+ T lymphocytes transcribe mRNA for IL–4, IL–5 and IL–13 in allergen-induced late-phase skin responses 34, 35. Previous investigators have also reported that monocytes and B lymphocytes generate IL–10 after 28 days VIT 36. In accord with our data, Akdis et al. 36, 37 have demonstrated suppression of allergen-specific proliferativeand cytokine, including Th1-type responses, after 7 days rush bee venom SIT. There was a concurrent increase in IL–10 production from isolated PBMC stimulated in vitro. The authors concluded that IL–10 had induced T cell anergy because neutralization of IL–10 fully reconstituted both the specific proliferative and cytokine responses. Furthermore, addition of IL–10 to IL–4-stimulated PBMC or purified B cells, inhibited antigen-specific and total IgE and enhanced IgG4 formation providing further support for the role of IL–10 as a regulatory cytokine in VIT.
In patients receiving specific grass-pollen immunotherapy there was a decrease in the size of the cutaneous LPR, accompanied by clinical improvement and a trend for a decline in EG2+ eosinophils. IL–10 expression was not examined in the biopsies but by contrast with our results, there was a significant increase in IFN–γ mRNA, and a reduction in numbers of infiltrating CD3+ and CD4+ cells 16. There were no changes in IL–4 mRNA levels but an unexpected increase in CD25+ cells. Similar results were obtained when the effect of grass-pollen immunotherapy on late nasal responses was examined after 12 months treatment in a placebo-controlled study 38. Although, the effect on IL–10 was not reported and no change in hybridization signal to IL–4 was found, there was an increase in IFN–γ mRNA and this was associated with inhibition of eosinophil and CD4+ cellular infiltration. In a more recently reported study, however, and in agreement with our results, there was significant suppression of allergen-induced IL–4 mRNA expression in skin biopsies taken during the late-phase response to grass-pollen after 3–4 years treatment 39. We found no increase in IFN–γ, but substantial transcriptional suppression of the Th2-type cytokine IL–4 after just 3 months VIT. This suggests distinct mechanisms for immunotherapy with differing allergens. Although IL–10 expression was not examined in the studies reported above, we speculate that the increase in allergen-induced levels of transcription and expression of IL–10 that were found during the LPR to venom may have inhibited transcription of IFN–γ 40. Furthermore, because the release of IL–10 isinhibited by IL–4 25 a primary suppression of IL–4 during VIT immunotherapy may act to lift the brake on IL–10 production. By contrast, grass pollen immunotherapy may result in aprimary increase in IFN–γ which then leads to an anti-inflammatory cascade.
Bellinghausen et al. 41 studied rush VIT in both bee- and wasp-allergic subjects and demonstrated reduced IL–4 and increased IFN–γ and IL–10 in isolated, maximally stimulated PBMC after treatment for 1 week. Interestingly, in this study, as in others 42 addition of blocking anti-IL–10 to the ex vivo system further up-regulated IFN–γrelease but had no effect on IL–4. This may further explain why we had no increase in IFN–γ in our in vivo system in which we were able to measure the net effect of complex transcellular interactions on cytokine transcription within the dermal microenvironment.
We have therefore reported up-regulation of IL–10 expression in a model of VIT that takes into account complex in vivo cytokine interactions. From the available data we suggest that induction of IL–10 is likely to be a pivotal event in the mechanism of VIT by inhibiting Th2 responses and resulting in the observed reduction of IL–4 transcription and eosinophil recruitment.
4 Materials and methods
Nine subjects (five male) of mean age 39 (range 26–51) years were studied. All gave a good history of severe systemic reactions to wasp stings with the development of hypotension in eight subjects and laryngeal edema in the remaining subject as the principal symptom. Three subjects with hypotension also developed laryngeal edema. Sensitivity was confirmed by an intradermal skin test withwasp venom and measurement of serum specific IgE antibodies by an enzyme immuno-assay (ImmunoCAP, Pharmacia & Upjohn, Milton Keynes, GB). The study was approved by the Addenbrooke's Hospital Ethics Committee and all participants gave informed written consent.
4.2 Venom immunotherapy
Five patients underwent wasp-VIT using a conventional regime and four subjects underwent an ultra-rush regime. The ultra-rush regime was used solely to minimize the number of outpatient visitsin these four patients and no particular subgroup was chosen for this treatment. The wasp venom used for the study for both intradermal injection and immunotherapy was our own and the collection, purification and standardization has been described previously 5. The protocol for conventional VIT consisted of weekly incremental subcutaneous injections of venom given over 11 weeks with a planned top dose equivalent to 2 stings at the end of this period. In ultra-rush VIT incremental injections of venom were given subcutaneously at 15–30 min intervals over a few hours. Byday 2 all patients had reached the equivalent of between one and two stings and had received a cumulative induction dose equivalent to 2 to 2.5 stings. Thereafter, weekly injections were given for 3 to 5 weeks (incremental to 2 stings if this dose had not been reached by day 2). After the induction phase for both rush and conventional VIT, desensitization was continued using monthly injections at the maintenance dose for three months followed by injections every 3 months.
4.3 Intradermal allergen provocation
The late-phase skin response to intradermal allergen was measured in all subjects before and 3 months after the start of immunotherapy. Intradermal challenges were performed at the same time of day, 15 cm apart, with 0.1 ml wasp venom (equivalent to about 1/12th of a sting) and 0.1 ml of allergen diluent into the flexor aspect of each arm. The size of the late-phase (24 h) response was rcorded as the mean of the longest and perpendicular diameters.
4.4 Skin biopsies
Skin biopsies were taken before and 3 months after the start of immunotherapy using a 4 mm disposable punch. After local anesthesia with 1% lignocaine, two skin biopsies were taken from the center of allergen and diluent control sites and this was repeated three months after the start of wasp VIT. Skin biopsies were divided into two and one segment (for in situ hybridization andT cell immunohistochemistry) immediately snap frozen using isopentane (Sigma Chemical Co. Ltd, Poole Dorset, GB), cooled down in liquid nitrogen and mounted in OCT (Bayer, Basingstoke, GB). The samples were stored in aluminum foil in an airtight container at –80oC until further analysis. The second fragment (for immunohistochemistry) was immediately placed in ice-cooled acetone containing the protease inhibitors iodoacetamide (20 mM) and phenylmethyl sulfonyl fluoride (2 mM), stored at –20oC for 24 h and then processed into water soluble glycol methacrylate (GMA) resin according to the method of Britten et al. 43. The biopsies were coded and analyzed by an investigator blinded to the status of the biopsies.
4.5 Immunohistochemistry (IHC)
Immunohistochemistry for mast cells, eosinophils, neutrophils and macrophages/dendritic cells was undertaken in GMA-embedded tissue as described in the method of Britten et al. 43. Briefly, 2-μm sections were cut using a microtome, floated on 0.2% ammonia solution in water for 1 min, collected onto slides coated with 3-aminopropyltriethoxy-silane (Sigma) and the sections were then dried at room temperature for 1 to 4 h. Sections were incubated with a mixture of 0.1% sodium azide and 0.3% hydrogen peroxide for 30 min to abolish endogenous peroxidase activity, washed in Tris-buffered saline (TBS; 5 mM Tris-HCl, 150 mM NaCl, pH 7.4) (2×5 min) then treated for 30 min with blocking medium consisting of Dulbecco's MEM (Sigma), 10% FCS (PAA Biological Ltd, Consett, GB) and 1%BSA (Sigma). The sections were then incubated overnight with the following mouse IgG1 mAb at previously titrated optimal dilutions: AA1 to mast cell tryptase (DAKO Ltd, High Wycombe, GB); EG2 (Pharmacia, Milton Keynes, GB) to the cleaved form of eosinophil cationic protein in eosinophils; NP57 to neutrophil elastase (DAKO); EBM11 to CD68 on human monocytes, macrophages and dendritic cells (DAKO). After washing (3×5 min), bound antibodies were labeled with a biotinylated secondary antibody (DAKO) for 2 h, washed (3×5 min) and detected using the streptavidin-biotin-peroxidase system (DAKO). After further washing aminoethyl-carbazole was applied as the chromagen giving a red reaction product and the sections were counter-stained with Mayer's hematoxylin (Sigma).
Optimal immunohistochemical staining for total and subtypes of human T lymphocytes (CD3, CD4, CD8, CD25) and for IL–10 was obtained on sections from cryopreserved skin as previously described 44, 45. Briefly, 6 μm thick sections were cut from frozen skin biopsies onto slides coated with 3-aminopropyltriethoxy-silane. The sections were then fixed inacetone and stored at –80oC until used for immunohistochemistry. An alkaline phosphatase anti-alkaline phosphatase (APAAP) technique was used to enumerate cells binding to monoclonal antibodies against total and subtypes of human T lymphocytes (CD3, CD4, CD8 and CD25, Becton Dickinson, Cowley, Oxford, GB). Optimal concentration of all antibodies used were determined in pilot experiments. The cryostat sections were incubated with the monoclonal antibodies against phenotypic markers for 30 min, washed in TBS, then incubated with rabbit anti-mouse immunoglobulin (Dako) for 30 min.After washing in TBS, the sections were incubated with soluble complexes of alkaline phosphatase and a mouse anti-alkaline phosphatase (APAAP, DAKO) for a further 30 min, and developed with Fast Red (Sigma) as chromagen for signal visualization. For immunostaining of IL–10, rat anti-human IL–10 monoclonal antibody (clone No: JES3–19F1) was purchased from Pharmingen (San Diego, CA). The slides were pre-treated with 0.1% saponin (Sigma)/PBS for 30 min, incubated with rat anti-IL–10 antibody (1:50) overnight at room temperature. After washing in TBS, the slides were treated with biotin-labeled-goat anti-rat IgG (1:100) (Vector Lab, Peterborough, GB) in TBS containing 0.1% saponin and 1% goat serum (Vector Lab) for 30 min. Following subsequent washes in TBS/saponin, the slides were incubatedwith avidin-biotin-alkaline phosphatase (ABC-AP-kit, Vector Lab). Color reaction was developed with Fast Red.
Positive cells stained red after development with Fast Red. No immunoreactivity was observed in sections stained by omission of the primary antibody or substitution of this antibody with an irrelevant antibody of the same species. Positive controls for each antibody were also included in each staining run using nasal polyp tissue obtained from patients undergoing routine polypectomy.
4.6 In situ hybridization (ISH)
All reagents were from Sigma Chemicals, unless otherwise indicated. The cDNA fragments encoding human IL–4, IL–10, IL–12 (p40), IL–13, TGF–β, eotaxin and RANTES were inserted into pGEM RNAexpression vectors (Promega, Southampton, GB). 35S-labeled riboprobes were prepared as previously described 34, 46. Briefly, riboprobes (antisense or sense) were synthesized in the presence of adenine triphosphate (ATP), guanosine triphosphate (GTP), cytosine triphosphate (CTP), and [35S]uridine triphosphate (UTP), and appropriate RNA polymerases (T7, SP6 or T3), respectively.
Slides were thawed for the experimental procedure. Permeabilization, prehybridization, and hybridization protocols were as described previously 34, 46. Incubation in N-ethyl maleimide, iodoacetamide and triethanolamine pre-hybridization reduced non-specific binding of the [35S]UTP-labeled probes. Furthermore, the experiments were performed under very high stringency conditions (hybridization at 50oC, and post-hybridization washing at 60oC, 0.1 SSC) to minimize nonspecific hybridization. Negative controls employed hybridization with the sense probe and pretreatment of slides with RNase A (Promega) prior to hybridization with antisense probe. For autoradiography, slides were dipped into K-5 emulsion (Ilford, Basildon, GB) and exposed at 4oC for 2 weeks in absolute darkness in a desiccated environment. The slides were developed (D-19 developing solution; Eastman Kodak Co., Rochester, NY), rinsed and counterstained with Harris hematoxylin. Dense deposits of silver grains on autoradiographs were present over cells expressing cytokine and chemokine mRNA 42, 43.
4.7 Quantitation and statistics
Whole sections from both IHC and ISH were counted in duplicate, blind to the patients' clinical status, using an eyepiece graticule as previously described 34, 46. Results were expressed as the total number of positive cells/mm2 of biopsy. The coefficient of variability of the duplicate counts obtained from all slides was less than 5%. All results are presented as mean ± SEM. Statistical analyses were performed using GraphPad Prism Software package (GraphPad Software Inc. San Diego, CA). The number of positively staining cells after intradermal venom was compared with values obtained after intradermal diluent both before and after 3 months VIT using the Wilcoxon rank sum test. A figure for the difference in cell numbers between intradermal venom and intradermal diluent was calculated for the two time points and the values compared using the Wilcoxon rank sum test. p values less than 0.05 were regarded as significant.