Specific serum IgE levels and FcɛRIβ genetic polymorphism in patients with penicillins allergy

Authors


Prof. Hai-Ling Qiao
Department of Clinical Pharmacology
School of Medicine
Zhengzhou University
Zhengzhou 450052
China

Abstract

Background:  Numerous studies have suggested that both genetic and environmental influences are involved in the pathogenesis of allergic disease and atopy. The objective of this investigation is to elucidate the underlying mechanism of penicillins allergy and improve the diagnostic methods.

Methods:  Radioallergosorbent test was used to examine eight kinds of specific IgE antibodies, which included four kinds of major and minor antigenic determinants, respectively, in the sera of 448 patients with penicillins allergy and 101 healthy subjects. A restriction endonuclease fragment length polymorphism of a polymerase chain reaction product was used for analysis of the FcɛRIβ polymorphism.

Results:  The positive rate of specific IgE in 448 patients was 58.26% (261), in which 37.28% (167) patients had positive IgE to major antigenic determinants and 47.09% (211) patients had positive IgE to minor antigenic determinants. Of the 179 patients with allergic history, 70.83% (17/24) patients had positive antibodies within 30 days, while 45.28% (24/53) had positive antibodies after 5 years. The positive reaction degree of skin test was absolutely correlated with specific IgE (P = 0.047). Among patients with positive specific IgE, significant differences of E237G genotype were observed between patients with positive benzylpenicillanyl (BPA)-, phenoxomethylpenicilloyl (PVO)- or ampicilloyl (APO)-IgE and control group (P = 0.015, 0.015, and 0.008, respectively). There were significant differences in E237G genotype between positive and negative BPA-, PVO- as well as APO-IgE patients (P = 0.014, 0.02, and 0.011, respectively).

Conclusions:  The patients with penicillins allergy have positive specific IgE not only to major antigenic determinants but also to minor antigenic determinants. The E237G variant of the FcɛRIβ gene is involved in the development of penicillins allergy through the process for the production of specific IgE antibodies.

Penicillins are widely used, inexpensive, low toxicity, efficacious, and often the drug of choice in many clinical situations. Allergy to β-lactam drugs is commonly reported, however, especially penicillins allergy (1). Fear of anaphylaxis, the most serious allergic response, has led many clinicians to over-diagnose penicillins allergy based solely on patient history, thus avoiding the use of penicillins in clinically appropriate situations.

Penicillin skin testing is a well-established method to evaluate current penicillin allergic status. The advantage of the method is quick, sensitive, reproducible, inexpensive, and can be performed by almost anybody. Its disadvantages include difficulties related to standardization of the method and the solutions used for testing, and variations that may be seen in results from different operators as well as risks in patients with hypersensitivity reaction. There is a need for in vitro tests, which can replace the skin tests. The radioallergosorbent test (RAST) is informative even at the time of the allergic reaction or in the first 15 days afterwards and seems to be very valuable for early diagnosis of penicillins allergy especially in cases when many drugs have been given. The great advantage of the method is high sensitivity and specificity. RAST with major antigenic determinants has been described previously (2, 3) in the diagnosis of IgE-mediated penicillin allergy but little information of RAST with minor antigenic determinants was available.

The mechanisms involved in allergic reactions to penicillins can be antibody dependent or T-cell mediated, although antibody production is a phenomenon controlled by T cells. Classically, antibody-mediated reactions are induced by IgE antibodies and are classified as immediate reactions. Serum IgE levels are a quantitative trait associated with atopic disease including allergic rhinitis, bronchial asthma, and atopic dermatitis. IgE-dependent activation of mast cells and basophils through the high-affinity IgE receptor (FcɛRI) is involved in the pathogenesis of allergen-induced immune responses in atopic diseases. The FcɛRIβ gene is located on chromosome 11q13, which is a region that has shown linkage to atopy and asthma (4, 5). Recently, animal models have provided unequivocal genetic evidence that FcɛRIβ functions as an amplifier of immune responses, especially those induced by mast cells and basophils (6). Mutations in the FcɛRIβ gene could alter IL-4 production and thus modify IgE levels. An amino acid substitution (glutamic acid → glycine) at position 237 in the FcɛRIβ gene has been associated with total and specific IgE levels and with atopic asthma (7, 8). The 237 glycine allele (*237G) is present in approximately 5% of Japanese and White populations (9).

To elucidate the underlying mechanism of allergic reactions to penicillins and improve the diagnostic methods, we explore the relationships between penicillins allergy, specific IgE antibodies and FcɛRIβ genetic polymorphism. We assay eight kinds of specific IgE antibodies to different penicillins antigenic determinants in sera by using RAST. There has been no report to date studying the association between penicillins allergy and the coding variant of FcɛRIβ E237G. Therefore we analyze whether this coding variant of FcɛRIβ shows association with penicillins allergy.

Material and methods

Subjects

The study included 448 patients with penicillins allergy (225 male and 223 female; age: 5–80 years) and 101 control subjects (75 male and 26 female; age: 16–60 years). All subjects were mainly recruited from the first affiliated hospital of Zhengzhou University. Among them, patients with penicillin G (PG), ampicillin (AMP), and amoxicillin (AX) allergy were 414, 21, and 13, respectively. To obtain information about their illnesses, the time, and nature of previous apparently allergic reactions, a detailed form was developed. Patient selection was based on a positive skin test or clinical symptoms after penicillins therapy. Their sera of 101 control subjects with no allergic reactions (including drug, food, pollen, and so on) and 30 fetal cord blood were selected to establish the RAST cutoff point. All of them had no history of allergic diseases, cancer, cardiovascular disease or other atopic diseases. Blood samples were drawn from each subjects at the time of the clinical manifestations and/or at the time of positive skin test. After clotting for 2 h at room temperature the sera were separated and stored at −30°C until the in vitro tests were performed.

RAST

Major and minor antigenic determinant paper discs made by ourselves included benzylpenicilloyl-polylysine (BPO-PLL), phenoxomethylpenicilloyl-polylysine (PVO-PLL), ampicilloyl-polylysine (APO-PLL), amoxicilloyl-polylysine (AXO-PLL), benzylpenicillanyl-polylysine (BPA-PLL), phenoxomethylpenicillanyl-polylysine (PVA-PLL), ampicillanyl-polylysine (APA-PLL), and amoxicillanyl-polylysine (AXA-PLL).

Cyanogen–bromide-activated cellulose discs (Filter paper 542, Whatman International, Maidstone, UK) were coupled to PLL and then treated with PG, PV (Phenoxymethylpenicillin), AMP, and AX. Antibodies were detected as previously described (10). The RAST results reflected the radiolabeled substance uptake, expressed as a percentage of the maximum counts per minute (cpm) added. Values over the mean + 2.33 SD of the results in the control group for BPO, PVO, APO, AXO, BPA, PVA, APA, and AXA were considered to be positive. Meanwhile, the fetal cord sera was used as negative control.

Genetyping

We genotyped the FcɛRIβ polymorphism in 158 patients and 87 healthy subjects using polymerase chain reaction (PCR)-restricted endonuclease fragment length polymorphism analysis. Genomic DNA was isolated from peripheral blood by standard phenol-chloroform extraction followed by ethanol precipitation (11). DNA was amplified by the PCR with primers 5′-CAG GTT CCA GAG GAT CGT GTT CTG-3′ (upstream) and 5′-GAT TCT TAT AAA TCA ATG GGA GGA AAC-3′ (downstream) to incorporate the E237G polymorphic site into an XmnI recognition site (9). PCR was performed with reagents composed of 0.25 μmol/l primers, 200 μmol/l dNTPs, 1.5 mmol/l MgCl2, 2.5 μl 10×Buffer, 100–200 ng DNA and 1 U TaqDNA polymerase in a 25 μl reaction solution. The program was accomplished by 38 cycles at 95°C for 30 s, 55°C for 30 s, 72°C for 50 s, and a final 8 min extention at 72°C. The PCR product was precipitation with 135 μl iced ethanol and 40 μl 10 mol/l NH4AC, dissolved in 15 μl distilled water, and 10 μl of this was digested with 5 U Xmnl (Promega, Madison, WI, USA) at 37°C overnight. Digested fragments were separated on a 3% agrose gel. The results of PCR products amplified from three genotypes were EE(AA), EG(GA), GG(GG) (Fig. 1).

Figure 1.

Polymerase chain reaction (PCR)-restricted endonuclease fragment length polymorphism analysis with XmnI for the E237G polymorphism of the FcɛRIβ gene. The size of the PCR product was 107 bp, and digested fragments derived from the E237 allele were 80 bp and 27 bp. PCR products from the 237G allele were not cut with Xmnl. Lanes 1–3: digested PCR products from a E237 homozygote (lane 1), a E/G237 heterozygote (lane 2), and a 237G homozygote (lane 3). Lane 4: an undigested control.

Statistical analysis

IgE values among groups were compared with One-way anova. The association of the E237G variant between the patient groups and control group was analyzed by chi-square test (spss v. 10.0, SPSS Inc., Chicago, IL, USA). Only P values of <0.05 were considered significant.

Results

The positive rate of specific IgE antibodies

The positive rate of specific IgE antibodies in 448 patients was 58.26% (261), which comprised 53.07% (95/179) patients with allergic history and 62.03% (165/266) patients with positive skin test as well as 33.33% (1/3) patients with family history. It was necessary to note the positive rate of IgE antibodies to the male was significantly higher than that to the female (P = 0.001) (data not shown).

Allergic reaction and kinds of specific IgE antibodies

Among eight kinds of antigenic determinants, the positive rates of AXA- and PVA-IgE were all the highest (24.78%), followed by BPA-IgE (24.11%), and that of APA-IgE was the lowest (10.49%). According to allergic symptoms, the positive rates of eight kinds of antigenic determinants were presented in Table 1. From the values given in the table, we can see that the lowest positive rate was APA-IgE in three groups and the highest belonged to minor determinants except for other symptoms group. The positive rates of antibodies to major and minor antigenic determinants were 37.28% (167) and 47.09% (211), respectively. Clearly, it showed significant differences (P = 0.003). The positive rates of antibodies to major and minor antigenic determinants showed differences between patients with anaphylactic shock, those with urticaria and those with other symptoms, but they were not quite significant. It was evident that the positive rate of specific IgE had increased significantly from 37.28 to 58.26% because of examining minor antigenic determinants. The results signified the more kinds of antigenic determinants, the higher the positive rate of specific IgE antibodies. The minor antigenic determinants probably play an important role in the process of allergic reaction.

Table 1.  The relationship between specific IgE antibodies and allergic symptoms in 179 patients with allergic history
GroupCasesSpecific IgE antibodies n (%)
BPOPVOAPOAXOBPAPVAAPAAXA
  1. BPO, benzylpenicilloyl; PVO, phenoxomethylpenicilloyl; APO, ampicilloyl; AXO, amoxicilloyl; BPA, benzylpenicillanyl; PVA, phenoxomethylpenicillanyl; APA, ampicillanyl; AXA, amoxicillanyl.

Anaphylactic shock479 (19.15)5 (10.64)11 (23.40)4 (8.51)10 (21.28)13 (27.66)3 (6.38)5 (10.64)
Urticaria8815 (17.05)16 (18.18)12 (13.64)12 (13.64)24 (27.27)18 (20.45)7 (7.95)17 (19.32)
Other symptoms4410 (22.73)7 (15.91)8 (18.18)5 (11.36)6 (13.64)6 (13.64)2 (4.55)7 (15.91)
Total17934 (18.99)28 (15.64)31 (17.32)21 (11.73)40 (22.35)37 (20.67)12 (6.70)29 (16.20)

The time interval and specific IgE antibodies

Of the 179 patients with allergic history, 70.83% (17/24) patients had positive antibodies within 30 days, while 45.28% (24/53) had positive antibodies after 5 years. These were significantly different (P = 0.037). In addition, we observed 7069 patient with serious urticaria within 8 months and assayed the specific IgE antibodies in serum once every month. The values of positive antibodies to major antigenic determinants decreased dramatically with increasing time, especially BPO- and PVO-IgE antibodies. There was not a significant decline in the values of positive antibodies to minor antigenic determinants (Fig. 2). The results showed specific IgE antibodies in serum decreased significantly with increasing time, and there was a close relationship between major antigenic determinants and urticaria.

Figure 2.

Changes of four kinds of IgE antibodies to major and minor antigenic determinants of serum 7069 in different time.

RAST and skin test results

Among 122 patients with instant positive skin test, the more the positive reaction degree of skin test, the higher the positive rates of specific IgE antibodies. If the positive reaction of skin test was over ‘+’ as positive criterion, the coincidence between RAST and skin test would be 61.48%. The coincidence between RAST and skin test was 71.05, 88.24, and 100% when the positive reaction degree of skin test was ‘++’, ‘+++’, and ‘++++’, respectively. The positive reaction degree of skin test was absolutely correlated with specific IgE antibodies (P = 0.047).

Cross-reactivity

Among 261 patients with positive antibodies, the positive rates of different antibodies and drugs were shown in Table 2. We concluded that there was a distance between 37.55% (98) patients with only one kind of positive IgE and 261 patients with including any kind of positive IgE. Of the four drugs, the specific positive rates of IgE antibodies to PG, PV, AMP and AX were 12.26, 14.18, 3.49, and 8.81%, respectively, while the total positive rate of IgE antibodies to them were 59.77, 63.60, 36.40, and 51.34%, respectively. The results showed that partial cross-reactivity existed in different penicillin derivatives.

Table 2.  The positive rate of different antibodies and drugs in the allergic patients with positive IgE antibodies (n = 261)
GroupPG [n (%)]PV [n (%)]AMP [n (%)]AX [n (%)]Total [n (%)]
BPOBPAPVOPVAAPOAPAAXOAXA
  1. A: positive cases only to corresponding antibody and drug; B: positive cases with including corresponding antibody and drug; C: cross-reaction positive cases.

  2. BPO, benzylpenicilloyl; PVO, phenoxomethylpenicilloyl; APO, ampicilloyl; AXO, amoxicilloyl; BPA, benzylpenicillanyl; PVA, phenoxomethylpenicillanyl; APA, ampicillanyl; AXA, amoxicillanyl.

A
 Antibody14 (5.36)16 (6.13)7 (2.68)30 (11.49)8 (3.07)1 (0.38)5 (1.92)17 (6.51)98 (37.55)
 Drug32 (12.26) 37 (14.18) 9 (3.49) 23 (8.81) 101 (38.70)
B
 Antibody104 (39.85)106 (40.61)91 (34.87)111 (42.53)67 (25.67)47 (18.01)57 (21.84)111 (42.53)261 (100)
 Drug156 (59.77) 166 (63.60) 95 (36.40) 134 (51.34) 261 (100)
C
 Antibody90 (34.49)90 (34.48)84 (32.18)81 (31.03)59 (22.61)46 (17.62)52 (19.92)94 (36.02)163 (62.45)
 Drug124 (47.51) 129 (49.42) 86 (32.91) 111 (42.53) 160 (61.30)

FcɛRIβ genetic polymorphism

The frequencies of the E237G alleles and the genotypes were presented in Table 3. We found no significant differences in the prevalence of E237G polymorphism between any group studied and control group.

Table 3.  Genotype and allele distribution of FcɛRIβ E237G polymorphism in control and allergic subjects
 nAA (EE) (%)AG (EG) (%)GG (GG) (%)A (E) (%)G (G) (%)
  1. IHR, immediate hypersensitivity reactions; DHR, delayed hypersensitivity reactions.

  2. A, Allergic patients with positive IgE in RAST test; B, Allergic patients with negative IgE in RAST test; A+B, All allergic patients; C, Allergic patients with positive IgE to major antigenic determinant in RAST; D, Allergic patients with positive IgE to minor antigenic determinant in RAST.

Control8763 (72)23 (26)1 (1)149 (86)25 (14)
 A10065 (65)33 (33)2 (2)163 (82)37 (19)
 B5841 (71)14 (24)3 (5)96 (83)20 (17)
 A+B158106 (67)47 (30)5 (3)259 (82)57 (18)
Allergy types
 IHR7051 (73)15 (21)4 (6)117 (84)23 (16)
 DHR5536 (65)19 (35)0 (0)91 (83)19 (17)
Allergy symptoms
 Shock3527 (77)7 (20)1 (3)61 (87)9 (13)
 Urticaria6842 (62)24 (35)2 (3)108 (79)28 (21)
 Others2115 (71)5 (24)1 (5)35 (83)7 (17)
Drugs
 Penicillin G12787 (69)36 (28)4 (3)210 (83)44 (17)
 Amoxycillin149 (64)4 (29)1 (7)22 (79)6 (21)
 Ampicillin83 (38)5 (63)0 (0)11 (69)5 (31)
Positive antibodies
 C6239 (63)21 (34)2 (3)99 (80)25 (20)
 D7850 (64)28 (36)0 (0)128 (82)28 (18)

Among patients with positive specific IgE, significant differences of E237G genotype were observed between patients with positive BPA-IgE and control group (P = 0.015). Similar results were observed in E237G genotype between patients with positive PVO- or APO-IgE and control group (P = 0.015 and 0.008, respectively). We also found significant differences in E237G genotype between positive and negative BPA-, PVO- as well as APO-IgE patients (P = 0.014, 0.02, and 0.011, respectively) (Table 4). Therefore, there were associations between genotype and allergic reactions to BPA, PVO, and APO antigens. The results showed that FcɛRIβ E237G gene was associated with the increase of BPA-, PVO-, and APO-IgE, respectively.

Table 4.  Genotype and allele distribution of FcɛRIβ E237G polymorphism in patients with positive and negative specific IgE antibodies
GroupnGenotype frequency (%)Allele frequency (%)
AA (EE) (%)AG (EG) (%)GG (GG) (%)A (E) (%)G (G) (%)
  1. BPO, benzylpenicilloyl; PVO, phenoxomethylpenicilloyl; APO, ampicilloyl; AXO, amoxicilloyl; BPA, benzylpenicillanyl; PVA, phenoxomethylpenicillanyl; APA, ampicillanyl; AXA, amoxicillanyl.

  2. *P < 0.05, **P < 0.01: (+) group vs control.

  3. P < 0.05: (−) vs (+) group.

Control8763 (72)23 (26)1 (1)149 (86)25 (14)
BPO(+)3622 (61)12 (33)2 (6)56 (78)16 (22)
PVO(+)3115 (48)16* (52)0* (0)46 (74)16 (26)
APO(+)2511 (44)14** (56)0** (0)36 (72)14 (28)
AXO(+)1912 (63)7 (37)0 (0)31 (82)7 (18)
BPA(+)3819 (50)19* (50)0* (0)57 (75)19 (25)
PVA(+)3624 (67)12 (33)0 (0)60 (83)12 (17)
APA(+)149 (64)5 (36)0 (0)23 (82)5 (18)
AXA(+)3624 (67)12 (33)0 (0)60 (83)12 (17)
BPO(−)6443 (67)21 (33)0 (0)107 (84)21 (16)
PVO(−)6950 (72)17† (25)2† (3)117 (85)21 (15)
APO(−)7554 (72)19† (25)2† (3)127 (85)23 (15)
AXO(−)8153 (65)26 (32)2 (2)132 (81)30 (19)
BPA(−)6246 (74)14† (23)2† (3)106 (85)18 (15)
PVA(−)6441 (64)21 (33)2 (3)103 (80)25 (20)
APA(−)8656 (65)28 (33)2 (2)140 (81)32 (19)
AXA(−)6441 (64)21 (33)2 (3)103 (80)25 (20)

Discussion

Penicillins allergy continues to be a serious problem; it is unpredictable and diverse in nature, ranging from relatively mild skin rashes or fever, through to life-threatening severe skin reaction or anaphylactic reactions. A major difficulty in prediction and avoidance of these reactions is that the immunological processes underlying penicillin allergy are poorly understood at a mechanistic level (12).

In our investigation RAST performed with eight kinds of major and minor determinants was found to be positive in 261 (58.26%) of 448 patients with penicillins allergy. Our results showed the positive rate of specific IgE had increased significantly from 37.28 to 58.26% because of examining minor antigenic determinants. Two main clinical manifestations exist in immediate allergic reactions to penicillins: anaphylactic shock and urticaria. Blanca et al. (13) believed minor determinants mixture (MDM) not only improved the capacity of diagnostic tests, but also was usually associated with a great risk of developing anaphylactic shock. Torres et al. (14) thought positivity to minor determinants of penicillin was associated more with anaphylactic shock than urticaria. This paper also showed that MDM played an important role in allergic reaction to penicillins. In addition, we observed one patient with serious urticaria within 8 months and assayed the specific IgE antibodies once every month. The results showed there was a close relationship between major antigenic determinants and urticaria.

Another key issue, because of its clinical relevance to diagnosis and selection of appropriate alternative therapy, is the immunological cross-reactivity of betalactam antibiotics. In addition to the antigenic determinants that are formed from the β-lactam ring structure, the side-chain group that distinguishes the different penicillins also may elicit the production of IgE antibodies that are clinically significant. The importance of side-chain-specific antibodies was recently demonstrated in a study by Baldo (15) in which the IgE-binding specificity was evaluated in patients who had reacted to flucloxacillin. Quantitative hapten inhibition studies demonstrated that only dicloxacillin, cloxacillin, and oxacillin (penicillins that have an R group similar to that in flucloxacillin) were able to strongly inhibit IgE binding. Penicillins that did not posses a methyl-phenyl-isoxazolyl side-chain determinant were poor inhibitors. Our previous studies revealed that the combing site of the specific IgE antibody was likely to be the side-chain of drug or the overwhelming drug structure (10). These findings suggested that different penicillins may be cross-reactive, not only by virtue of their shared β-lactam and thiazolidine rings but also by virtue of shared or similar side-chain determinants. This paper also demonstrated partial cross-reactivity existed in different penicillins.

The RAST and the skin test are frequently used to diagnose allergy. Our results showed the positive reaction degree of skin test was significantly correlated with specific IgE antibodies. This fact suggests that RAST is a safe, effective approach to identifying patients at risk or those not at risk for allergic reaction to penicillin. Moreover, detection of IgE antibodies in the skin or serum depends on the time interval between the reaction and the evaluation. Different studies have shown that the longer the interval, the lesser the possibility of detecting IgE antibodies (16). Subjects with a history of penicillin allergy may have negative skin test responses for several reasons, the most common reason is loss of skin test sensitivity, particularly in cases evaluated some time after reaction. The reason for the variability in the skin test response or RAST may be genetic or environmental because uncontrolled or hidden contact with β-lactam can occur, which could maintain sensitization (17). The results of our study showed that 70.83% of the patients with allergic history had positive antibodies within 30 days, and 58% had positive antibodies after 2 years, whereas 45.28% had positive antibodies after 5 years. In a recent prospective study carried out by Patriarca et al. (16), approximately 50% of patients had positive responses after 3 years, and around 30% had positive responses after 5 years. These were probably related to patients selected.

To our knowledge, this is the first report studying the association between a candidate gene for atopy and penicillins allergy. The results showed that there were associations between genotype and allergic reactions to BPA, PVO, and APO antigens. Moreover, E237G gene was associated with the increase of BPA, PVO, and APO-IgE antibodies, respectively. These results suggest that the E237G variant of the FcɛRIβ gene is involved in the development of penicillin allergy through the process for the production of specific IgE antibodies. One of the possible mechanisms for this is that the 237G allele elevates the signaling activity and expression of FcɛRI on mast cells and let them release more proallergic cytokines including IL-4, which is essential for synthesis of IgE (18). Indeed, the β-chain of FcɛRI was reported to act as an amplifier of FcɛRI-mediated cell activation signals and of receptor expression (19), and the E237G site was located in close proximity to the immunoreceptor tyrosine activation motif in the β-chain (20). However, Donnadien et al. (21) recently showed that transfection of a variant β-chain with 237G to human monocytic and basophilic cell lines had no direct effect on the amplifier functions of the β-chain. Moreover, Furumoto et al. (22) reported that introduction of E228G, which was a mouse homologue for the E237G variant in human, to murine mast cells did not elevate IgE-mediate mast cell activation. Therefore, the basic mechanisms for the elevation of IgE antibodies and the development of penicillin allergy by the E237G polymorphism remain to be clarified. As suggested by Donnadieu et al. (21), the variant β-chain may be in linkage disequilibrium with other more relevant polymorphisms or affecting unknown β-chain functions.

The value of RAST has been shown in this paper but it remains to be discovered what causes the drug side-effects in the remaining patients. Further information is necessary about specific antibodies of all the immunoglobulin classes that are present at the time of the allergic reaction. However, the role of IgE in the clinical picture must not be forgotten, as could be seen in one of these patients who gave evidence of shock when skin tests were performed. In addition, because the mechanisms for the development of atopic disease should vary somewhat, more genetic studies on penicillin allergy would help us in understanding the significance of FcɛRIβ E237G polymorphism in the development of atopic diseases.

Acknowledgments

We thank Professor B.A. Baldo and Dr Zhen-Jun Zhao (Molecular Immunology Unit, Kolling Institute of Medical Research, Royal North Shore Hospital of Sydney, Sydney, New South Wales, Australia) for their excellent technical assistance. This project was supported by the Science Foundation for Distinguished Young Scholars of Henan Province (No. 0312002100) and the Nature Science Foundation of Henan Province (No. 0211040100).

Ancillary