Search

Purpose of Review: To summarize the data of both preclinical studies and initial clinical trials of a novel allergoid for allergen specific immunotherapy. This allergoid consists of allergen covalently coupled to immunostimulatory oligodeoxynucleotide DNA sequences., Recent Findings: Recently, immunostimulatory oligodeoxynucleotide sequences, also called unmethylated cytosin-guanine dinucleotide motifs, have been discovered that act as strong T helper 1 response inducing adjuvants in mice. Although mixing allergens with immunostimulatory DNA sequences induces T helper 1 responses in T helper 2 biased mice, the allergens in such mixes could still cause anaphylactic reactions when used in humans which is one of the reasons why immunotherapy has gradually been falling out of favor. Therefore, we made allergen-immunostimulatory oligodeoxynucleotide conjugates and investigated their immunogenicity and allergenicity in animal models of allergy. These conjugates were highly immunogenic for inducing T helper 1-like antiallergen responses and reversed T helper 2 responses and symptoms of asthma in mouse models. They were also less allergenic, as shown by the reaction with human immunoglobulin E antibodies and by histamine release from basophils of allergic patients. Preliminary phase I and II trials in ragweed allergic patients showed that allergen-immunostimulatory oligodeoxynucleotide conjugates are well tolerated, less allergenic and induce immunoglobulin G antiallergen antibodies more rapidly than allergen extracts without significantly increasing the immunoglobulin E titer., Summary: Allergen-immunostimulatory DNA conjugates induce T helper 1 and down regulate preexisting T helper 2 anti-allergen responses in mice. Initial phase I and II trials in ragweed allergic patients showed that ragweed allergen-DNA conjugates are well tolerated and induce a rapid immunoglobulin G but not E response. The data show that allergen-DNA conjugates are a novel type of allergoid that have great potential for a safe and potent form of allergen specific immunotherapy.

Background: Immunotherapy has gradually fallen out of favor for the treatment of many allergic diseases because of the overall convenience, safety, and efficacy of medications. However, investigations suggest that allergen/immunostimulatory sequence oligodeoxynucleotide (ISS-ODN) conjugates (AICs) might have improved safety and efficacy compared with allergen extracts., Objective: We determined whether changes in the ISS-ODN conjugation ratio would effect the immunogenicity and allergenicity of AIC., Methods: Immunogenicity was determined by means of AIC vaccination of mice, followed by analysis of antigen-specific antibody and cytokine responses. The allergenicity of AIC was determined in mast cell release studies and in murine models of anaphylaxis and the Arthus reaction., Results: AIC induced a stronger immune response than allergen alone or allergen mixed with ISS-ODN, but higher-level ISS-ODN conjugation reduced its immunogenicity modestly. In mast cell degranulation studies AIC was approximately 100-fold less allergenic than native allergen, with stepwise increases in the ODN conjugation ratio leading to stepwise decreases in allergenicity. In anaphylaxis studies death rates were reduced from 100% with native allergen challenge to as low as 0% with high-ratio ISS-ODN AIC challenge. Similar results were obtained in an Arthus reaction model., Conclusion: These investigations establish that AIC is both significantly more immunogenic and less allergenic than native allergens and the techniques used might have further utility for the standardization and optimization of AIC formulations for use in allergic patients.

Allergen-specific immunotherapy, although efficacious, is now less frequently used because of potential adverse reactions. Recently, two new types of allergen immunotherapy have been developed that appear to overcome this problem, namely allergen gene vaccination and vaccination with allergen-immunstimulatory DNA conjugates. In animal models of allergy, both have been shown to induce nonallergic T-helper cell type 1 immune responses to allergens and downregulate pre-existing T-helper cell type 2 responses. In initial clinical trials with allergic patients, allergen-immunostimulatory DNA conjugates were well-tolerated, induced immunoglobulin-G but not immunoglobulin-E antibodies and appeared to have great potential as a novel, safe and efficacious type of allergen specific immunotherapy.

Although excellent pharmacological treatments for allergies exist, they do not change the underlying pathogenesis of allergic diseases and do not cure the disease. Only allergen-specific immunotherapy, the injection of small but increasing amounts of allergen, has been shown to change a pre-existing allergic Th2 immune response to a non-allergic Th1 response. However, since injection of allergen is associated with the risk of allergic and sometimes even life-threatening anaphylactic reactions, immunotherapy is no longer used as extensively as in the past. In the search for a novel immunotherapy having a low risk-to-benefit ratio, immunostimulatory CpG motif DNA sequences have recently been shown to provide an excellent tool for designing safer and more efficient forms of allergen immunotherapy. These DNA-based immunotherapeutics include allergen gene vaccines, immunization with allergen-DNA conjugates and immunomodulation with immunostimulatory oligodeoxynucleotides. All three DNA-based immunotherapeutics have been shown to be very effective in animal models of allergic diseases and, at present, allergen-DNA conjugates are being tested for their safety and efficacy in allergic patients. This review describes the preclinical findings and the data of the first clinical trials in allergic patients of DNA-based immunotherapeutics for allergic disorders.

Background: Immunostimulatory sequence oligodeoxynucleotide (ISS-ODN) is a potent antiallergic immunomodulating agent in mice. However, few studies have addressed its antiallergic potential in human subjects., Objective: We sought to determine whether a phosphoro-thioate ISS-ODN could inhibit IL-4-dependent IgE synthesis by human B cells., Methods: Initially, nonatopic- and atopic-donor PBMCs were incubated with ISS-ODN or mutated oligodeoxynucleotide, and cytokine production and B-cell expression of IFN-gamma receptor and IL-4 receptor were measured by using ELISA and flow cytometry, respectively. In subsequent studies atopic-donor PBMCs were incubated with IL-4 alone or with ISS-ODN or mutated oligodeoxynucleotide. After 14 days, IgE production and IgM, IgG, and IgA production were determined by using ELISA. In select IgE studies cytokines were neutralized with mAbs., Results: ISS-ODN induced IL-12, IFN-alpha, IFN-gamma, IL-10, and IL-6 production from both nonatopic- and atopic-donor PBMCs. ISS-ODN also increased IFN-gamma receptor and inhibited IL-4 receptor expression on B cells from both donor populations. Furthermore, ISS-ODN inhibited IL-4-dependent IgE production by atopic-donor PBMCs. Neutralization of IL-12, IFN-alpha, IFN-gamma, and IL-10, but not IL-6, attenuated the inhibitory activity of ISS-ODN on IgE production. In contrast to its inhibition of IgE synthesis, ISS-ODN stimulated the production of IgM, IgG, and IgA., Conclusion: These in vitro studies demonstrate that phos-phorothioate ISS-ODN elicits an innate immune response by PBMCs, which inhibits IL-4-dependent IgE synthesis. In addition, these results provide further support for consideration of ISS-ODN therapy for the treatment of allergic disease in clinical practice.

Background: Allergen immunotherapy is inconvenient and associated with the risk of anaphylaxis. Efforts to improve the safety of immunotherapy by means of chemical modification of allergens have not been successful because it greatly reduced their antigenicity. Recently, immunostimulatory DNA sequences (ISS or CpG motifs) have been shown to act as strong T(H)1 response-inducing adjuvants., Objective: We sought to determine whether conjugation of ISS to the major short ragweed allergen Amb a 1 results in enhanced immunotherapeutic potential in mice and decreased allergenicity in human subjects., Methods: A 22-mer ISS oligodeoxynucleotide (ISS-ODN) was coupled to Amb a 1 and used for immunization of mice, rabbits, and monkeys., Results: In mice the Amb a 1-ISS conjugate induced a T(H)1 response (IFN-gamma secretion), whereas Amb a 1 induced a T(H)2 response (IL-5 secretion). The T(H)1 response was not observed with an Amb a 1-non-ISS conjugate. Coinjection of Amb a 1 with ISS-ODN was much less effective in inducing a T(H)1 response. In mice primed for a T(H)2 response, injection with Amb a 1-ISS conjugate induced a de novo T(H)1 response and suppressed IgE antibody formation after challenge with Amb a 1. Amb a 1-ISS conjugate induced high-titer anti-Amb a 1 IgG antibodies in rabbits and cynomolgus monkeys, whereas Amb a 1 alone or Amb a 1 coinjected with ISS-ODN did not induce a detectable response. Amb a 1-ISS conjugate was less allergenic than Amb a 1 alone, as shown by a 30-fold lower histamine release from human basophils of patients with ragweed allergy, whereas mixing ISS-ODN with Amb a 1 did not reduce histamine release., Conclusion: Amb a 1-ISS conjugate has an enhanced T(H)1-biased immunogenicity and reduced allergenicity. It may offer a more effective and safer approach for allergen immunotherapy than currently available methods.

Immunostimulatory DNA sequences (ISS) are a potent Th1 adjuvant. We hypothesized that conjugation of ISS to protein antigens would strongly enhance their immunogenicity because both antigen and adjuvant (ISS) would be delivered to the same locale/antigen-presenting cell. To test this hypothesis, we conjugated a 22-mer immunostimulatory oligodeoxynucleotide (ISS-ODN) to two test antigens of differing intrinsic immunogenicity, namely Escherichia coli beta-galactosidase and the HIV-1 envelope glycoprotein gp120. We show that the antigen-ISS conjugates rapidly induce Th1 cells secreting high levels of IFN-gamma, strong CTL activity, and high titer IgG2a and HIV-neutralizing antibodies, exceeding gene and protein vaccination alone or immunization with mixtures of antigen and ISS-ODN. The data suggest that this procedure generates a novel and unique vaccine that rapidly triggers strong humoral and cell-mediated immunity.

Allergic disorders are characterized by the prevalence of immunoglobulin (Ig) isotype E antibodies, and are considered to result from enhanced T helper type-2 (Th2) responses to allergens. A Th2 response is characterized by enhanced humoral responses and the production of IL-4 and IL-5 by CD4(+) T cells (Th2) in response to antigen (1,2). These "Th2 cytokines" enhance the allergic response by inducing B cell isotype switching to IgE (3-5), by inducing undifferentiated Th0 cells to further differentiate into Th2 cells, and by inducing eosinophil growth, and differentiation (5). In addition, the Th2 cytokines inhibit Th0 differentiation into Th1 cells, thereby reducing the recruitment of interferon gamma (IFN-γ) producing Th1 cells that could then down-regulate or modulate the Th2 responses (3,5). Consequently, the ability of a Th2 response to allergens to exert a positive feedback effect leads to a vicious cycle and may explain the exacerbation of allergic responses that follows continued exposure to allergens in atopic humans (2,6).

The nature of the immune response (Th1/Th2) in mice to protein antigens or allergens was compared to that of immunization with pDNA encoding the same antigens. pDNA immunization induced a Th1 response and no IgE antibodies whereas the proteins induced a Th2 response and IgE antibodies. Furthermore, the pDNA induced Th1 response dominated over the protein elicited Th2 response in a secondary immune response. In particular, a preexisting Th2 response (as is the case in allergic patients) did not prevent a new Th1 response to an allergen-pDNA booster injection. The major reason why pDNA immunization induced a Th1 response to allergens was the presence of immunostimulatory non-coding DNA sequences (ISS) in the plasmid constructs having a CpG motif. These ISS caused antigen presenting cells to secrete INF-alpha, INF-beta and IL-12, all cytokines that induce naive T cells to differentiate into CD4+ Th1 cells and CD8+ Tc1 cells. Passive transfer of both Th1 and Tc1 cells from pDNA immunized mice into naive mice inhibited a Th2 response and IgE antibody formation to a subsequent injection of allergen in alum. pDNA immunization or ISS-oligonucleotide injection prior to allergen challenge reduced both immediate type airway sensitivity and late phase allergen induced eosinophil filtration of the lung. Allergen-pDNA immunization may provide a novel type of immunotherapy for the treatment of allergic diseases in man. Since only small amounts of allergen are secreted by the allergen-pDNA transformed cells, allergen-pDNA immunotherapy will unlikely carry the risk of the anaphylactic reactions that are associated with classical allergen injection immunotherapy.

The Th1/Th2 type immune response to E. coli beta-galactosidase (beta-gal) was compared to that to gene vaccination with plasmid (p) DNA encoding beta-gal. BALB/c mice were immunized with beta-gal in alum or a pDNA construct consisting of a CMV-based promoter and the beta-gal gene (pCMV-LacZ). Beta-gal in alum induced IgG1 and IgE antibodies and the CD4+ T cells from these mice secreted interleukin 4 (IL-4) and IL-5 but no interferon-gamma (IFN-gamma) after in vitro antigen stimulation. In contrast, mice immunized with pCMV-LacZ formed predominantly IgG2a antibodies and their CD4+ T cells secreted IFN-gamma but no IL-4 and IL-5. These data indicate that beta-gal induced a Th2 and the pCMV-LacZ a Th1 response to beta-gal. The pDNA induced Th1 response dominated over the Th2 response. Mice primed with pCMV-LacZ failed to produce IgE antibodies after a booster injection of beta-gal in alum. Boosting of mice primed with beta-gal in alum with pCMV-LacZ resulted in a 75% decrease in the IgE antibody titer within 6 weeks and IgG2a antibody formation and CD4+ T cells that secreted IFN-gamma in amounts similar to T cells from pDNA primed mice. As shown by adoptive cell transfer, both CD4+ and CD8+ T cells from pDNA immunized mice inhibited an IgE response to beta-gal in alum in the recipient mice. pDNA immunization also inhibited the eosinophilic infiltration of the lung of ovalbumin (OVA) immunized mice after OVA inhalation challenge in an animal model of the late phase reaction. The mechanism of the pDNA induced Th1 immune response was shown to be the result of stimulation by distinct non-coding immunostimulatory DNA sequences (ISS) in the backbone of the pDNA. The ISS induced antigen presenting cells to secrete cytokines that cause naive T cells to differentiate into Th1 cells (e.g. IFN-alpha, IL-12). The data indicate that gene vaccination induces a Th1 immune response that is capable of down-regulating a preexisting Th2 response and IgE antibody formation. Thus, immunization with pDNA encoding for allergens may provide a novel type of immunotherapy for allergic diseases.

Background: We previously showed that immunization of mice with plasmid DNA (pDNA) encoding the Escherichia coli beta-galactosidase gene (pCMV-LacZ) induces a Th1 response, whereas beta-galactosidase (beta-gal) in saline or alum induces a Th2 response. Furthermore, the Th1 response dominates over the Th2 response and downregulates preexisiting IgE antibody formation. Here, we determined by passive transfer of CD4+ or CD8+ lymphocytes and by immunizing beta2-microglobulin knockout (beta2-M KO) mice whether CD4+ and/or CD8+ cells from pDNA-immunized mice suppress IgE antibody production., Methods: BALB/c mice were injected with either CD4+ or CD8+ lymphocytes from naive beta-gal-in-alum or pCMV-LacZ-immunized mice, then immunized with beta-gal in alum, and the IgE antibody formation was determined. Second, C57BL/6 wild-type (WT) or beta2-M KO mice were immunized with beta-gal orpCMV-LacZ, and the IgE antibody production was assessed., Results: Passive transfer of both CD4+ and CD8+ lymphocytes from pDNA-immunized mice suppressed the IgE antibody response by 90% compared to transfer of CD4+ T cells from naive or beta-galin-alum immunized mice. beta2-M KO mice produced 3 times more IgE than the WT control mice both in the primary and secondary response., Conclusion: Both CD4+ and CD8+ subsets of T cells from pDNA-immunized mice can suppress IgE antibody production by affecting the primary response and/or by propagating the Th1 memory response in a passive cell transfer system. Immunization with pDNA-encoding allergens may be an effective new form of immunotherapy for atopic diseases.

Dendritic cells (DC) have been reported to be the most potent antigen-presenting cells (APC) for the activation of naive T cells and to be 10-100-fold more potent APC than monocytes (M phi) in the mixed lymphocyte reaction. In this study the authors compared human blood DC with M phi and B cells for their ability to activate cloned rye grass allergen Lol p I specific CD4+ Th1 and Th2 cells. In the presence of Lol p I, all three types of APC activated Th1 and Th2 cells to a similar extent, as shown by T-cell proliferation and interferon-gamma, interleukin-2 (IL-2) or IL-4 secretion. However, at low APC:T cell ratios, M phi were the most potent APC for both Th1 and Th2 cells followed in decreasing order by DC and B cells. This hierarchy was observed with APC preparations isolated by negative selection or highly purified by positive selection using fluorescent cell sorting for HLA-DR(high)-DC, CD14(pos)-M phi and CD19(pos)-B cells. The data demonstrate that, in contrast to what has been reported for naive T cells, human blood DC activate cloned memory Th1 and Th2 cells to a similar extent as M phi and B cells presumably because the requirements for activation of memory type T cells are less stringent than those for naive T cells.

We compared the antigen-specific antibody isotypes and lymphokine secretion by CD4+ T cells in BALB/c mice immunized intradermally with either Escherichia coli beta-galactosidase (beta-gal) or plasmid DNA (pDNA) encoding beta-gal in a cytomegalovirus-based expression vector (pCMV-LacZ). pCMV-LacZ induced mainly IgG2a, whereas beta-gal in saline or alum induced IgG1 and IgE beta-gal-specific antibodies. In addition, splenic CD4+ T helper (Th) cells isolated from pDNA-immunized mice secreted interferon-gamma but not interleukin (IL)-4 and IL-5, whereas Th cells from beta-gal-injected mice secreted IL-4 and IL-5 but not interferon-gamma after in vitro stimulation with antigen. Together these data demonstrate that pDNA immunization induced a T helper type 1 (Th1) response, whereas protein immunization induced a T helper type 2 (Th2) response to the same antigen. Interestingly, priming of mice with pCMV-LacZ prevented IgE antibody formation to a subsequent i.p. beta-gal in alum injection. This effect was antigen-specific, because priming with pCMV-LacZ did not inhibit IgE anti-ovalbumin antibody formation. Most importantly, intradermal immunization with pCMV-LacZ (but not pCMV-OVA) of beta-gal in alum-primed mice caused a 66-75% reduction of the IgE anti-beta-gal titer in 6 weeks. Also, pCMV-LacZ induced specific IgG2a antibody titers and interferon-gamma secretion by Th cells in the beta-gal in alum-primed mice. The data demonstrate that gene immunization induces a Th1 response that dominates over an ongoing protein-induced Th2 response in an antigen-specific manner. This suggests that immunization with pDNA encoding for allergens may provide a novel type of immunotherapy for allergic diseases.

Human rye grass allergen Lol p I-specific T helper cell clones of Thp, Th0, Th1 and Th2 subtype were activated with Lol p I and monocytes or B cells as antigen-presenting cells, and cell proliferation, interleukin (IL)-2, interferon-gamma, and IL-4 secretion were measured. Monocytes induced activation of T cell clones of all four T helper cell subsets and were usually more potent antigen-presenting cells than B cells. B cells and monocytes similarly induced proliferation and IL-4 secretion by Th2 clones, whereas B cells, in contrast to monocytes, only weakly activated Th1 clones. However, exceptions to this rule existed within each T helper cell subset suggesting that individual T cell clones, regardless of the subset to which they belong, may have quantitatively and/or qualitatively different requirements for secondary activation signals which are provided by the antigen-presenting cells. The data demonstrate that, in general, monocytes are more effective than B cells in activating human T cell clones of all subtypes and that B cells were efficient antigen-presenting cells only for Th2 cells. However, individual T cell clones of any given T helper cell subset vary with respect to their activation by monocytes or B cells.

In the mouse, activation of T cells by T cell receptor (TCR) crosslinking with anti-CD3 antibodies in the absence of a costimulatory signal induces Th1 but not Th2 cell anergy. Furthermore, anti-CD3 induces anergy of Th1- but not Th2-type lymphokine secretion in Th0 cells. This study was designed to determine whether this is also the case in man. Human rye grass allergen Lol p I-specific cloned CD4+ T helper cells of subtypes Th0, Th1, and Th2 were treated with immobilized anti-CD3. The cells were rested for 4 days and then activated under optimal conditions with antigen and antigen-presenting cells (APCs). Cell proliferation and IL-2, IFN-gamma, and IL-4 secretion was determined to test for the anergic state. The initial anti-CD3 treatment induced cell proliferation, IL-2, IFN-gamma, and/or IL-4 secretion by T cells of all three subsets which was followed by an anergic state in Th0, Th1, and Th2 cells as shown by a 51 to > 94% decrease in cell proliferation and IFN-gamma and/or IL-4 secretion after subsequent APC and Lol p I activation. Addition of IL-2 or IL-4 during anti-CD3 treatment of the cells did not prevent unresponsiveness. However, the addition of IL-2 but not IL-4 during APC and Lol p I stimulation partially reversed the anergic state. These data demonstrate that, contrary to the mouse, cloned T cells of all three human T helper cell subtypes are anergized by anti-CD3 TCR activation in the absence of costimulatory signals. The fact that human Th2 cells can be anergized may be important for the development of new treatments in Th2-mediated allergic disorders.

1,25-Dihydroxyvitamin D3 [1,25-(OH)2-D3] is known to be an immunosuppressive hormone. This review primarily deals with in vitro and in vivo effects of 1,25-(OH)2-D3 and analogue, 1,25-dihydroxy-16ene-vitamin D3 [1,25-(OH)2-16ene-D3], on T helper subsets type 1 (Th1) or type 2 (Th2) that have distinctive functional characteristics in humans. Th1 secrete interferon (IFN-gamma), interleukin (IL-2) and induce B cells to produce immunoglobulin IgG2a while Th2 secrete IL-4, IL-10 and induce the production of IgG1 and IgE by B cells. The sterol inhibits the secretion of IL-12, a cytokine produced by monocytes and B cells, which leads to the activation and differentiation of Th1. In addition, 1,25-(OH)2-D3 directly inhibits IFN-gamma secretion by Th1 clones while it has little effect on IL-4 secretion by Th2 clones. The analogue, 1,25-(OH)2-16ene-D3, is 100-fold more potent than 1,25-(OH)2-D3 in inhibiting IFN-gamma secretion but also has little effect on IL-4 secretion. In mice, when given in vivo, the sterol prevents the induction of spontaneous and induced autoimmune diseases and inhibits Th1 induce IgG2a responses. These actions of the vitamin D3 compounds suggest that it may have potential therapeutic applications in Th1-mediated clinical situations such as autoimmunity and transplantation.

Three hundred fifty-nine serum samples from patients with immunoglobulin M (IgM) or IgG monoclonal gammopathies were tested for binding to the capsular polysaccharide (PS) of Neisseria meningitidis group B (MenB PS, poly-alpha[2-->8]-N-acetylneuraminic acid). Of 159 IgM paraproteins, 7 (4.4%) were positive, compared with 0 of 200 IgG paraproteins (P < 0.05). Since MenB PS reactivity was limited to the IgM paraproteins, the 159 IgM paraproteins were tested by enzyme-linked immunosorbent assay (ELISA) for reactivity with seven other bacterial PSs. None reacted with meningococcal A or C, Haemophilus influenzae type b, or Streptococcus pneumoniae type 3, 6, 14, or 23 PS. The specificity of the MenB PS-reactive antibodies was confirmed by demonstration of binding to N. meningitidis group B cells but not to a capsular PS-deficient mutant and by specific inhibition of binding to solid-phase MenB PS by soluble MenB PS in an ELISA. Five of five antibodies tested protected infant rats from bacteremia caused by Escherichia coli K1, an organism with a PS capsule that also is composed of poly-alpha[2-->8]-N-acetylneuraminic acid. Each of the seven MenB PS-reactive paraproteins had autoantibody activity as defined by binding to homogenates of calf brain in a radioimmunoassay. For six of the seven antibodies, binding to calf brain was inhibited by the addition of soluble MenB PS. Thus, approximately 4% of human IgM paraproteins have autoantibody activity to poly-alpha[2-->8]-N-acetylneuraminic acid, an antigen expressed in fetal brain and cross-reactive with the MenB capsular PS. The reason for this skewing of the IgM paraprotein repertoire toward reactivity with poly-alpha[2-->8]-N-acetylneuraminic acid antigenic determinants is unknown.

T cell lines (TCL) and CD4+ T cell clones (TCC) with specificity for the rye grass allergen Lolium perenne (Lol p) I were isolated from the blood of nine donors, six having active atopic disease, two being in remission, and one having IgE anti-Lol pI Abs but not atopic disease. The T cell epitopes of Lol pI were determined by TCLs and TCCs reactivity with 23 overlapping, 20 amino acid-long peptides spanning the entire length of the 230 amino acid-long allergen. In addition, the Th subsets (Th1, Th2, Th0, Thp) were determined by measuring IL-2, IFN-gamma, and IL-4 in the supernatants of TCC activated with Lol pI and irradiated APC. TCC from individuals from which a large panel of clones were obtained from 10(5) PBMC initial cultures recognized multiple peptides (5-9) and 23 overlapping peptides a total of 16 were recognized by at least one TCC from one of the patients. These 16 peptides were derived from all areas of the Lol pI molecule, indicating the ability of human Th cells to recognize many peptide epitopes on Lol pI. Although no clear cut immunodominant peptides were detected, T cell clones of 50% of the patients reacted with peptide 191-210. There was no correlation between peptide epitope reactivity and lymphokine secretion pattern of the TCC. Of 12 TCC obtained from six patients with active atopic disease, four (33%) were of Th1, five (42%) of Th2, one (8%) of Thp, and two (17%) of Th0 type. Of 14 TCCs isolated from three atopic donors in remission, five (36%) were of Th1, three (21%) of Th2, four (29%) of Thp, and two (14%) of Th0 type. The data demonstrate that T cells from rye grass pollen allergic patients can recognize multiple peptide epitopes on Lol pI scattered over the entire molecule. No correlation existed between epitope reactivity and lymphokine secretion pattern of the TCC.

We studied the role of IL-4 in human IgE formation in severe combined immunodeficient mice engrafted with peripheral blood mononuclear leukocytes (hu-PBL-SCID). PBL from four nonatopic donors produced only small (< 20 ng/ml) or undetectable amounts of IgE in SCID mice whereas engrafted PBL from seven atopic donors secreted IgE with IgE serum levels reaching a mean +/- SE of 184 +/- 37 ng/ml (n = 20). Serum IgE levels peaked 2-3 wk after PBL transfer and declined thereafter with a half-life of 1-2 wk. In contrast, IgG of all subclasses reached maximum serum levels 5-7 wk after PBL transfer and declined little thereafter. Injection of a neutralizing monoclonal antibody to the human IL-4 receptor (IL-4R) on day 0 inhibited completely the IgE formation and caused an approximate twofold reduction of IgG production of all subclasses. The anti-IL-4 R antibody had no effect on IgE secretion when administered 4 wk after PBL engraftment. Incubation of PBL with IL-4 before engraftment resulted in a 10-fold increase in IgE production and could be further enhanced by 100 fold if, in addition to preincubation with IL-4, IL-4 was injected daily for 5 d after PBL transfer. This treatment with IL-4 also induced two- to threefold increase in IgG levels. IFN-gamma had no effect on either IgE or IgG subclass production. In approximately 50% of the mice, one or more IgG subclasses increased disproportionally 5 wk after PBL injection as a result of monoclonal IgG formation. These data demonstrate that PBL from atopic donors secrete IgE in SCID mice in an IL-4-dependent manner, and that IgE production can be enhanced 10- to 100-fold with exogenous human IL-4 in these mice. This mouse model is amenable for the in vivo study of immunomodulators on human IgE formation.

Controversial data have been reported regarding the ability of peripheral blood T cells to secrete interferon-gamma (IFN-gamma) and interleukin-4 (IL-4) from atopic patients as compared to nonatopic healthy controls. In most of these studies, T cells in peripheral blood mononuclear cell preparations (PBMC) were stimulated with polyclonal T cell activators. Some of these activators are able to activate cells other than T cells in the PBMC preparations which may influence the lymphokine levels in supernatants of PBMC. To evaluate this, we compared the IFN-gamma and IL-4 levels in PBMC and isolated T cell preparations after activation with phytohemagglutinin (PHA), Concanavalin A (ConA), anti-CD3 plus phorbol myristate acetate (PMA), or ionomycin plus PMA. The IFN-gamma and IL-4 levels in the supernatants were calculated based on the percent T cells in the preparations. Whereas all activators induced significant IFN-gamma secretion, only ionomycin plus PMA stimulation induced large IL-4 secretion. In virtually all cases, the IFN-gamma levels calculated on a per T cell basis differed for PBMC versus isolated T cells. Whereas in some donors the IFN-gamma levels were higher in PBMC preparations than in T cells, in others it was the opposite. Similarly, in about one half of both normal and atopic donors tested, the IL-4 levels of activated PBMC were 2- to 7-fold lower than levels in isolated T cells. The data suggest that non-T cells have a significant effect on the IFN-gamma and IL-4 levels in supernatants of polyclonally activated PBMC.(ABSTRACT TRUNCATED AT 250 WORDS)

The effect of IL-4 and transforming growth factor-beta (TGF-beta) on immunoglobulin secretion in vitro by peripheral blood mononuclear cells (PBMC) or purified B cells activated with murine EL4 thymoma cells and phorbol myristate acetate (PMA) was investigated. As previously reported, IL-4 induced IgE and IgG4 secretion by B cells in PBMC preparations and B cells activated with EL4 cells and PMA. However, when B cells, either in PBMC preparations or purified and activated with EL4 cells and PMA, spontaneously secreted large quantities of immunoglobulin, IL-4 suppressed the immunoglobulin secretion of all isotypes. IL-4 also suppressed the IgE secretion by B cells from an atopic dermatitis patient. This suppressive effect was not reversed by adding IL-2 or interferon-gamma (IFN-gamma) to the cultures. We also showed that TGF-beta suppressed the immunoglobulin secretion by purified B cells activated by EL4 cells and PMA. To investigate whether IL-4 or TGF-beta suppressed immunoglobulin secretion by in vivo 'switched' and isotype-committed B cells, sIgD- B cells were isolated, activated with EL4 cells and PMA and cultured with IL-4 or TGF-beta. Such activated B cells secreted large quantities of IgG1, IgG2, IgG3, IgA1, IgA2 and IgM, and IL-4 and TGF-beta suppressed all these isotypes by greater than 80%. The data demonstrated that IL-4 and TGF-beta suppress immunoglobulin secretion in vitro by in vivo isotype-committed sIgD- B cells, suggesting that these lymphokines may play a down-regulatory role on differentiated isotype-committed B cells in an isotype-unrestricted manner. The data also showed that IL-4 and TGF-beta acted directly on isolated B cells.

The effect of interleukin-2 (IL-2) on IL-4-induced IgE and IgG4 secretion by B cells in peripheral blood mononuclear cell (PBMC) preparations from non-atopic healthy humans and atopic dermatitis patients was investigated. PBMC were cultured at an optimal concentration of recombinant IL-4 with or without addition of IL-2 for 10 days. Native and recombinant IL-2 inhibited the IL-4-induced IgE and IgG4 secretion in a dose-dependent manner by cells from both normal and atopic donors. Rabbit antibodies to IL-2 or to the monoclonal anti-IL-2 receptor antibody anti-TAC reversed the IL-2 effect. Culturing cells with IL-4 and IL-2 for 1 or 2 days only slightly suppressed the IgE and IgG4 secretion whereas addition of IL-2 to IL-4 containing cultures on day 4 or 5 inhibited the IgE and IgG4 secretion more effectively. This is in contrast to interferon-gamma (IFN-gamma) which inhibited the IL-4 induced IgE and IgG4 secretion when added for the first 24 or 48 h but had no effect when added on days 4 or 5. The data demonstrate that both IL-2 and IFN-gamma act as antagonists in the IL-4-induced IgE and IgG4 secretion by human B cells; while IL-2 appears to inhibit relatively late in culture, IFN-gamma has an early inhibitory effect, suggesting that the two lymphokines inhibit the IL-4 effect by different mechanisms.

Peripheral blood mononuclear cells from 8 normals and 8 patients with atopic dermatitis (AD) were cultured with recombinant interleukin-4 (IL-4) and the IgE and IgG subclass levels in the culture supernatants measured by radioimmunoassays. IL-4 induced IgE and IgG4 secretion by B cells from both normals and AD patients whereas it has no consistent effect on IgG1, IgG2 and IgG3 secretion. The IL-4 dose response was similar for IgE and IgG4 secretion by cells from both normals and AD patients. On the average, the patients' cells secreted more IgE and less IgG4 than the cells from normals, but because of a large variation, the differences were not significant. However, the ratio of IgG4:IgE secretion was significantly greater for normals than AD patients (mean +/- SEM 7.1 +/- 1.6:1 vs. 1.5 +/- 0.4:1; p less than 0.01). The data demonstrate that IL-4 induces IgE and IgG4 secretion by B cells from both normals and AD patients and suggest that the IL-4 induced switch from IgM to IgG4 or IgE secretion may proceed preferentially to IgE in AD patients as compared to normals.

Immunoglobulin production, particularly IgE, is known to be dysregulated in graft-vs-host disease (GVHD). We examined serum levels of the highly T-dependent Ig isotypes, IgE, IgG1, and IgG2a, in two different mouse models of GVHD. GVHD across minor histocompatibility barriers is produced by injection of B10.D2 spleen cells into 600 rad irradiated BALB/c hosts. Both strains are H2d and mls b, but differ at the minor histocompatibility antigens. As GVHD progresses there is a rapid rise in serum IgE (300-fold) and IgG1 (2.5-fold) with a peak at Day 14. Concomitantly, IgG2a falls. Serum immunoglobulin levels return to normal by 11 weeks. The rise in IgE is abolished by increased (900 rad) recipient irradiation, suggesting that host-derived factors are important. GVHD across major histocompatibility barriers is produced by injection of DBA/2 spleen cells into unirradiated or 600 rad irradiated (B6 x DBA/2)F1 hosts. Only in the irradiated recipients is there severe Ig dysregulation. In this situation there is a 100-fold rise in IgE, and 5- to 10-fold rises in IgG1 and IgG2a. While the results in GVHD across minor barriers suggest stimulation of T helper cells secreting IL-4, the increase in IgE, IgG1, and IgG2a levels in GVHD across major barriers suggests activation of IL-4 and IFN-gamma-secreting T cells. These results indicate that different mechanisms may be operating in these two models of GVH. Murine GVHD can serve as a model for studying dysgammaglobulinemias in general and for hyper-IgE formation in particular.

IgE binds to two types of Fc receptors, called Fc epsilon R1 (or high-affinity Fc epsilon R) and Fc epsilon R2 (or low-affinity Fc epsilon R). The Fc epsilon R1 is composed of four polypeptide chains, one alpha, one beta, and two gamma chains. The alpha chain contains the IgE binding site and is a member of the immunoglobulin supergene family. The Fc epsilon R2, also called CD23, consists of one polypeptide chain which shows homology to animal lectin receptors. Fc epsilon R1 are expressed on mast cells and basophils. Crosslinking of the Fc epsilon R1 induces immediate release of mediators of inflammation such as histamine and leukotrienes and delayed secretion of interleukins 4, 5, and 6. Fc epsilon R2 are expressed on resting mu delta + B cells, monocytes/macrophages (M phi), eosinophils, and platelets but rarely on T cells. Interleukin-4 upregulates Fc epsilon R2 expression on B cells and M phi. The functions of Fc epsilon R2 on the different cell types are not fully established and are controversial. Fc epsilon R2 on M phi, eosinophils, and platelets mediate cytotoxicity to schistosomules, enhance phagocytosis, and induce the release of granule enzymes. However, M phi from patients with atopic dermatitis expressing significantly more Fc epsilon R2 than M phi from normals do not release more leukotriene C4, prostaglandin E2, or beta-glucuronidase after incubation with aggregated IgE than normal monocytes. Furthermore, aggregated IgG1 is much more efficient than IgE in inducing mediator release from M phi and IgG1 antibodies are not known to induce immediate-type hypersensitivity reactions. Therefore, definitive proof that Fc epsilon R2 are involved in the pathogenesis of allergic disorders is still lacking. IL-4 appears to play a central role in immediate-type hypersensitivity. It induces human B cells to secrete IgE and IgG4, Ig isotypes typical for antibodies to helminthic parasites and allergens. IL-4 stimulates mast cell growth and upregulates Fc epsilon R2 expression. Interferon-gamma and IL-2 inhibit the IL-4-induced IgG4 and IgE secretion. Whether the abnormally high IgE antibody production in atopic patients is the result of overproduction of IL-4 or deficient IFN-gamma/IL-2 production is presently unknown.

Peripheral blood mononuclear cells (PBMCs) from healthy nonallergic donors were cultured with recombinant interleukin-4 (rIL-4), and the Ig of different isotypes was quantitated in the culture supernatants by radioimmunoassays. Recombinant IL-4 induced IgG4 and IgE secretion in a dose-dependent manner, whereas it had no consistent effect on the secretion of the other isotypes. In the absence of rIL-4, B cells in the PBMC preparations secreted less than 1 ng IgE/ml and a mean of 5 ng IgG4/ml. In the presence of the optimal dose of 100 U rIL-4/ml, PBMCs from five donors secreted a mean +/- SEM of 37 +/- 8 ng IgE/ml and 66 +/- 25 ng IgG4/ml. In kinetic studies, no IgG4 or IgE secretion was detected during the first 5 days of culture, and approximately 50% of the IgG4 and IgE secreted by day 15 was detected in supernatants on day 7. Cycloheximide, actinomycin-D, and mytomycin-C completely inhibited the rIL-4-induced IgG4 and IgE secretion, indicating that de novo protein, RNA, and DNA synthesis was required. As shown by Percoll buoyant density centrifugation, rIL-4 induced B cells in the high-density fraction to secrete IgG4 and IgE, whereas it inhibited spontaneous IgG4 secretion by low-density B cells. Interferon-gamma inhibited IL-4-induced IgG4 and IgE secretion. The data demonstrate that IL-4 induces small, dense, peripheral blood B cells to secrete not only IgE but also IgG4, which parallells the IL-4-induced IgE and IgG1 secretion by murine B cells.(ABSTRACT TRUNCATED AT 250 WORDS)

The ability of human peripheral blood mononuclear cells to synthesize IgE in vitro in response to pokeweed mitogen (PWM) is controversial. To determine whether the conflicting results obtained by different laboratories could be due to inherent qualitative differences in the anti-IgE antibodies used to measure low concentrations of IgE in culture supernatants, we compared the specificities of anti-IgE reagents prepared by various methods. Immunoadsorbent-purified antibodies were isolated from a goat antiserum to the lambda, IgE myeloma protein PS and a rabbit antiserum to the kappa, IgE protein Bed in three ways: 1) antibodies to IgE PS (anti-PS) were isolated from the goat antiserum by affinity chromatography with PS coupled to Sepharose 4B; these antibodies consisted of anti-epsilon chain-specific and anti-idiotypic antibodies to protein PS; 2) antibodies specific for the epsilon-chain (anti-epsilon) were purified by affinity chromatography with IgE myeloma proteins that were not used for immunization; and 3) antibodies to idiotypic determinants of proteins PS (anti-id PS) and Bed (anti-id Bed) were isolated on affinity columns with the respective myeloma proteins after absorption of the epsilon-chain-specific antibodies. These three types of antibodies were then used in a solid phase radioimmunosorbent test to quantitate the amount of "IgE" synthesized by peripheral blood mononuclear cells from nonatopic and atopic donors cultured for 7 days in the presence and absence of PWM. ANTI-PS antibodies detected a PWM-induced "IgE formation" in cell culture supernatants of both non-atopic and atopic donors. (ABSTRACT TRUNCATED AT 250 WORDS)

Peripheral blood T lymphocytes from nonatopic control donors, asymptomatic atopic donors, and patients with moderate to severe atopic dermatitis were analyzed for Fc receptors for IgE (T epsilon cells) and IgG (T gamma cells) by rosette assays and were characterized with monoclonal antibodies. The T cells were reacted first with monoclonal antibodies, followed by fluoresceinated F(ab')2 goat antimouse Ig; they were then rosetted, and subsequently the rosetting cells were examined for immunofluorescence. Seven nonatopic control donors had less than 0.1% T epsilon cells and a mean +/- SD of 10.5% +/- 4.1 T gamma cells. Seven asymptomatic atopic donors with low IgE levels (2 to 233 IU/ml) varied from less than 0.1 to 1.3% T epsilon cells and 7.2% +/- 3.7 T gamma cells. Six of seven patients with moderate to severe atopic dermatitis and IgE levels of 1339 to 24,261 IU/ml had less than 0.1% T epsilon cells and significantly fewer T gamma cells (3.1% +/- 2.7, p less than 0.01) than the nonatopic control donors and the atopic donors in remission. Both T epsilon and T gamma cells reacted with the pan-T cell antibody Lyt-3 (anti-sheep red cell receptor) but not with antibodies OKT3, OKT4, or OKT6. Subpopulations of both T epsilon and T gamma cells reacted with antibodies OKT8 and the antimonocyte antibody OKM1. The OKM1+ cells did not appear to be monocytes, however, because the T cells did not react with another antimonocyte antibody, BRL.2, and were negative for nonspecific esterase activity. (ABSTRACT TRUNCATED AT 250 WORDS)

BALB/c mice were injected ip with 1 microgram ovalbumin (OVA) in alum or complete Freund's adjuvant (cFA) and the changes of the IgE and IgG subclass serum levels and isotypes of the anti-OVA specific antibodies determined by radioimmunoassays. By Day 10, OVA in alum had induced a 5- to 10-fold increase of the IgE serum level and an initial decrease of the IgG subclass levels which subsequently increased to two to threefold over the preinjection level. OVA in cFA induced a gradual twofold increase of the IgE serum level, a rapid fourfold increase of the IgG2a level occurring by Day 7, and a gradual two to threefold increase of the other IgG subclasses. Over 90% of the anti-OVA antibodies were of the IgGl isotype with both adjuvants; OVA in alum induced slightly more IgGl anti-OVA antibodies than cFA. In contrast, the OVA in alum injected mice formed significantly more (5- to 10-fold) IgE anti-OVA antibodies than the cFA-injected mice. OVA in alum also induced a large nonspecific increase of the IgE serum level because only approximately 40% of the increase observed on Day 14 was absorbable with OVA, whereas approximately 90% the IgE increase in cFA injected mice was absorbable with OVA. The data demonstrate that mice form mainly IgGl and IgE antibodies to OVA irrespective of the adjuvant. The low specific and lack of nonspecific IgE formation by mice injected with OVA in cFA may be the result of cFA-induced interferon-gamma (IFN-gamma) production because IFN-gamma has been shown to stimulate IgG2a and inhibit IgE secretion in vitro.

Release of immunoglobulin-binding factors was induced from cultured human lymphoblastoid cells in a serum-free environment by incubation at 37 degrees. RPMI-8866 cells expressing Fc receptors for IgE (Fc epsilon R) and Daudi cells expressing Fc receptors for IgG (Fc gamma R) released IgE- and IgG-binding factors, respectively. The factors induced were shown to have: (i) specificity of rosette inhibition; (ii) specificity of Ig-binding; and (iii) a glycoprotein nature. The IgE-binding factor remained stable, whereas the IgG-binding factor showed a time-dependent degradation and a concurrent change in specificity from IgG- to IgE-binding. The occurrence of this change was paralleled by a disappearance of a 125,000 mol. wt. protein on autoradiograms of SDS-PAGE of the IgG-binding factor preparation. These results raise the possibility that the released IgG-binding factor is found in a polymerized form which may dissociate into subunits in aqueous solution, and that common subunits may exist with regard to IgE- and IgG-binding factors, some of which may either lose their activity, or bind to both IgE and IgG with different affinities.

A human IgG1 myeloma protein that has a delection in the third constant domain of the heavy chain (Cgamma3) and forms two-chain half-molecules was studied for its in vivo turnover and its ability to fix C1q and hemolytic complement, to bind to human lymphocytes, neutrophils, and monocytes, and to induce a passive cutaneous reaction in guinea pigs. In both man and monkeys, the half-molecule was rapidly catabolized and in part excreted into the urine. The half-life in man was 4.3 days and the fractional turnover 165% per day; 7.6% of the intravascular pool was excreted into the urine per day. Although the 7S four-chain myeloma protein could not be obtained in a pure form, the elimination from the serum of a partially purified preparation suggested that it was also rapidly catabolized. The unaggregated half-molecule neither formed complexes with C1q, cound to human lymphocytes, neutrophils, and monocytes, nor elicited a reverse passive cutaneous reaction in guinea pigs. In contrast, the aggregated half-molecule fixed hemolytic complement and bound to the human white cells similarly to an intact IgG1 myeloma protein. In order to explain the biological activities of this half-moleculr, it is postulated that IgG1 may have several (at least two) submolecular sites for a given biological activity that are localized on both the Cgamma2 and Cgamma3 domains. Proteins having both sites would be capable of binding to C1q and Fc cell receptors in unaggregated in order to obtain half-molecule, must be aggregated in order to obtain this binding.

Pneumococcus-induced serotonin release from human platelets is greatly facilitated by a factor present in normal human plasma and serum. We have identified this factor as immunoglobulin by: (a) removing if from plasma and serum with solid phase antiFab antibody; (b) demonstrating its absence from the serum of an individual with severe immunoglobulin deficiency; and, (c) showing its presence in IgG preparations isolated from normal individuals. Evidence suggesting that the release reaction is triggered by pneumococcal antigen-antibody complexes rather than by nonimmune interaction of immunoglobulin with pneumococcus includes: (a) the failure of isolated IgG myeloma proteins to support release; (b) a lack of correlation between IgG concentration and "releasing factor activity" in normal human sera; (c) the identification of a normal serum that supports release by types II and III pneumococci but not type VII; and, (d) the ability of most normal sera to support release by pneumococca serotypes II and VII, though these types have not shown nonimmune reactivity with the Fc portion of the IgG molecule. The ability of antibodies present in normal serum to support pneumococcus-induced serotonin release suggests that the thrombocytopenia seen in pneumococcal infection may at least in part be caused by pneumococcal antigen-antibody complexes.

Peripheral blood lymphocytes from 10 nonallergic donors and 7 patients suffering from seasonal allergic rhinitis and receiving desensitization therapy were analyzed by rosette assays for Fc receptors for IgE (Fc epsilon R) and IgG (Fc gamma R) before, during and after the grass pollen season. Six of seven patients had moderately elevated IgE levels (330 +/- 268 IU/ml), all had high titers of skin sensitizing antibodies to grass pollens and serum IgE antibodies as measured by radio-allergosorbent tests (RAST). Seven of the nonallergic donors had 2-30 IU/ml IgE and negative RAST, whereas three had 91-267 IU/ml IgE and two were RAST positive to the grass pollens. In March, when the patients were asymptomatic, the mean +/-SD of the Fc epsilon R+ lymphocytes did not significantly differ from the nonallergic control group: nonallergic Fc epsilon R+ 1.2 +/- 0.9% (29 +/- 20/,mm3), allergic Fc epsilon R+ 2.0 +/- 3.1% (48 +/- 52/mm3). In contrast, during the grass pollen season in May and June, when the patients developed symptoms of allergic rhinitis, they had significantly (P less than 0.01) more Fc epsilon R+ lymphocytes than the controls: nonallergic Fc epsilon R+ 1.7 +/- 1.9% (40 +/- 46/mm3), allergic Fc epsilon R+ 4.7 +/- 1.2% (134 +/- 69/mm3). In the postpollen period, August-October, most of the patients again had low numbers of Fc epsilon R+ lymphocytes: nonallergic Fc epsilon R+ 1.4 +/- 0.9% (26 +/- 13/mm3), allergic Fc epsilon R+ 2.1 +/- 1.9% (62 +/- 82/mm3). The nonallergic control donors with elevated IgE levels and positive RAST always had low numbers of Fc epsilon R+ lymphocytes. In contrast, two other nonallergic donors, who had a 2-7 IU/ml IgE and negative RAST, showed significant increases of Fc epsilon R+ lymphocytes over several weeks during the grass pollen season. No statistically significant changes in Fc gamma R+ lymphocytes occurred in both nonallergic and allergic donors. The total and specific IgE serum levels did not vary much in the nonallergic donors and patients during the period of study and any changes that did occur did not correlate with the changes in Fc epsilon R+ lymphocytes. The data demonstrate that Fc epsilon R+ peripheral blood lymphocytes increase in allergic patients during natural antigen exposure and active disease in the absence of measurable increases of total and specific serum IgE. Because two nonallergic control donors also had temporary increases of Fc epsilon R+ lymphocytes, an increase of peripheral blood Fc epsilon R+ lymphocytes may be a sensitive indicator of an ongoing IgE immune response.

Murine B cells were activated for 24 h with either lipopolysaccharide (LPS) or polyribosyl-ribitol-phosphate (PRP), followed by addition of interleukin 4 (IL-4) and the immunoglobulin isotypes secreted into the supernatant quantitated. Without IL-4, both LPS and PRP induced mainly IgM and IgG3 and no IgE secretion. Addition of IL-4 to both LPS- and PRP-activated cells decreased the IgM and IgG3 secretion and induced a large IgG1 production. In contrast to IgG1, only LPS-activated cells secreted large amounts of IgE, demonstrating that the nature of the polyclonal B cell activator also plays an important role in the IL-4 induced IgE formation. The effect of LPS and IL-4 on high- and low-density sIgM+/sIgD+ cells was also investigated. LPS and IL-4 induced IgG1 and IgE secretion by both populations. Low-density B cells from mice infected with the parasite Nippostrongylus brasiliensis formed more IgE than low-density B cells from normal mice, presumably because these mice have more in vivo preactivated B cells committed to IgE formation. The data show that IL-4 can act on both small high-density resting B cells as well as on in vivo preactivated low-density B cells to induce IgG1 and IgE secretion.