Anti-cytokine autoantibodies: epiphenomenon or critical modulators of cytokine action

Anti-cytokine autoantibodies: mechanistic insights and disease associations

Anti-cytokine autoantibodies (ACAAs) are increasingly recognized as modulating disease severity in infection, inflammation and autoimmunity. By reducing or augmenting cytokine signalling pathways or by altering the half-life of cytokines in the circulation, ACAAs can be either pathogenic or disease ameliorating. The origins of ACAAs remain unclear. Here, we focus on the most common ACAAs in the context of disease groups with similar characteristics. We review the emerging genetic and environmental factors that are thought to drive their production. We also describe how the profiling of ACAAs should be considered for the early diagnosis, active monitoring, treatment or sub-phenotyping of diseases. Finally, we discuss how understanding the biology of naturally occurring ACAAs can guide therapeutic strategies.

The chemokine landscape: one system multiple shades

Leukocyte trafficking is mainly governed by chemokines, chemotactic cytokines, which can be concomitantly produced in tissues during homeostatic conditions or inflammation. After the discovery and characterization of the individual chemokines, we and others have shown that they present additional properties. The first discoveries demonstrated that some chemokines act as natural antagonists on chemokine receptors, and prevent infiltration of leukocyte subsets in tissues. Later on it was shown that they can exert a repulsive effect on selective cell types, or synergize with other chemokines and inflammatory mediators to enhance chemokine receptors activities. The relevance of the fine-tuning modulation has been demonstrated in vivo in a multitude of processes, spanning from chronic inflammation to tissue regeneration, while its role in the tumor microenvironment needs further investigation. Moreover, naturally occurring autoantibodies targeting chemokines were found in tumors and autoimmune diseases. More recently in SARS-CoV-2 infection, the presence of several autoantibodies neutralizing chemokine activities distinguished disease severity, and they were shown to be beneficial, protecting from long-term sequelae. Here, we review the additional properties of chemokines that influence cell recruitment and activities. We believe these features need to be taken into account when designing novel therapeutic strategies targeting immunological disorders.

[Anti-cytokine autoantibodies: Review of the literature]

Anti-cytokine antibodies (ACA) are an emerging cause of acquired immunodeficiency, especially in previously healthy adults. The most frequently reported are anti-IFN-γ responsible for disseminated non-tuberculous mycobacteria infections, and anti-GM-CSF mainly in mycobacteria, cryptococcosis and nocardiosis infections. The presence of anti-IFN-α in severe COVID-19 infections has recently been described. The search for and detection of these ACAs in an unusual infection situation makes it possible to set up specific therapies in addition to the anti-infective treatment. ACAs are also frequent in various autoimmune pathologies where, in addition to being indicators of the breakdown of immune tolerance, they can modulate the activity of the disease according to their cytokine target. In this review of the literature, we will focus on the epidemiology and the clinical impact of these ACAs in healthy subjects and in infectious or dysimmune diseases.

Anti-Cytokine Autoantibodies in Systemic Lupus Erythematosus

Cytokine dysregulation is characteristic of systemic lupus erythematosus (SLE), a systemic autoimmune disease of considerable heterogeneity. Insights gained about the cytokine dysregulation in SLE have the potential for identifying patient subsets before the onset of clinical disease and during established disease. Clustering patients by cytokine and disease activity subsets is more informative than isolated cytokine studies, as both pro inflammatory and immunoregulatory cytokines contribute to the cytokine dysregulated state in SLE. Endogenous anti-cytokine autoantibodies (ACAAs) may be involved in the regulation of cytokine biology by reducing excessive production or by prolonging their half-life in the circulation through the formation of cytokine-antibody immune complexes. Although endogenous ACAAs may have deleterious effects such as contributing to immunodeficiency states, their role in the pathophysiology of autoimmune conditions such as SLE has yet to be clearly elucidated. The aim of the present article is to provide a focused review of the current knowledge of ACAAs in SLE.

Immunogenicity screening assay development for a novel human-mouse chimeric anti-CD147 monoclonal antibody (Metuzumab)

The clinical effect of patient immune responses to therapeutic antibodies affect product safety and efficacy, which makes the development of valid, sensitive immune assays a key aspect of antibody drug development. In this paper, we reported the generations of mouse monoclonal and Cynomolgus monkey polyclonal antibodies against the anti-CD147 antibody (Metuzumab) as the internal standards and the positive controls. Seven mouse monoclonal antibodies were shown to recognize both (Fab)2 and full length of Metuzumab, but not the control normal human IgGs, and monoclonal anti-Metuzumab, Clone 2D9 was chosen to be used as the internal standard for anti-Metuzumab study. A Bridging ELISA assay was developed by coating the wells with the antibody drug, and the anti-drug antibody (ADA) in the animal sera were detected by enzyme-labeled antibody. Its limit of detection (LOD) was determined to be 0.39ng/ml of anti-Metuzumab antibody (ADA) with linear range between 0.39-50ng/ml and R(2)=0.994. For normal monkey sera, a minimal dilution was determined to be 1:80. However, very different from peptide or other protein drugs, strong interferences from the residual antibody drugs were observed from most of the testing monkey sera in the preclinical study. It was experimentally determined that the concentration of the residual antibody drug in the assay have to be lower than 1μg/ml, so the assays were carried out at 1:100 dilution of the monkey sera. In the pre-clinical study, 32 monkeys were treated with escalating doses of Metuzumab between 0, 10, 50, 200mg/kg for 13 times over 13weeks of time period. 16 of them were terminated right after the last injection, while the other 16 were rested for additional 4weeks before termination. Afraid to miss any positive response to antibody drug, sera samples were collected at six time points, including 2-, 6- and 10-weeks post 1st dose, prior to last dose, and 2-, 4-weeks into recovery. The highest positive rates were seen with the Medium- and High-dose group 2-weeks post the first injection, 6 out 8 monkeys in the High-dose were positive for free ADA. However, no significant pathologic and clinic adversary effect was observed in those monkeys.

Neutralizing antibodies and fatigue as predictors of low response to interferon-beta treatment in patients with multiple sclerosis

Background

The clinical impact of neutralizing antibodies against interferon-beta (NAb) is controversial. Their presence can lead to a decrease in interferon-beta (IFNβ) efficacy. Fatigue reported in patients with multiple sclerosis (MS) may be associated with an unfavorable clinical course. We conducted a prospective multicentre study to assess the association between response to IFNβ, NAb and fatigue.

Methods

Patients with relapsing-remitting MS on IFNβ treatment were included. During the second year of treatment, the patients were analyzed for NAb status and non-response criteria to IFNβ (number of relapses ≥1 during the follow-up period, increase in the Expanded Disability Status Scale ≥0.5). The score on the Modified Fatigue Impact Scale (MFIS pathological if score ≥35) was noted for each patient.

Results

Of the 176 patients included: 22.3% were NAb positive, 54.5% presented non-response criteria to IFNβ, and 57.4% had a pathological MFIS score. Fatigue was increased in NAb + patients (p = 0.0014) and they were more likely to present non-response criteria to IFNβ (p = 0.041) than NAb- patients. Multivariate logistic regression analysis showed that the presence of NAb was related to fatigue (p = 0.0032) and denoted disease activity in these patients (p = 0.026).

Conclusions

This study demonstrates the impact of NAb on the non-clinical response to IFNβ. Fatigue assessment is an indicator of IFNβ responsiveness and a predictive biomarker of deterioration on patient’s neurological status.

Protein microarray analysis reveals BAFF-binding autoantibodies in systemic lupus erythematosus

Autoantibodies against cytokines, chemokines, and growth factors inhibit normal immunity and are implicated in inflammatory autoimmune disease and diseases of immune deficiency. In an effort to evaluate serum from autoimmune and immunodeficient patients for Abs against cytokines, chemokines, and growth factors in a high-throughput and unbiased manner, we constructed a multiplex protein microarray for detection of serum factor-binding Abs and used the microarray to detect autoantibody targets in SLE. We designed a nitrocellulose-surface microarray containing human cytokines, chemokines, and other circulating proteins and demonstrated that the array permitted specific detection of serum factor-binding probes. We used the arrays to detect previously described autoantibodies against cytokines in samples from individuals with autoimmune polyendocrine syndrome type 1 and chronic mycobacterial infection. Serum profiling from individuals with SLE revealed that among several targets, elevated IgG autoantibody reactivity to B cell-activating factor (BAFF) was associated with SLE compared with control samples. BAFF reactivity correlated with the severity of disease-associated features, including IFN-α-driven SLE pathology. Our results showed that serum factor protein microarrays facilitate detection of autoantibody reactivity to serum factors in human samples and that BAFF-reactive autoantibodies may be associated with an elevated inflammatory disease state within the spectrum of SLE.

Natural antibodies against bone morphogenic proteins and interferons in healthy donors and in patients with infections linked to type-1 cytokine responses

In patients receiving recombinant therapeutic proteins, the production of antibodies against the therapeutics is a rising problem. The antibodies can neutralize and interfere with the efficacy and safety of drugs and even cause severe side effects if they cross-react against the natural, endogenous protein. Various factors have been identified to influence the immunogenic potential of recombinant human therapeutics, including several patients' characteristics. In recent years, so-called naturally occurring antibodies against cytokines and growth factors have been detected in naive patients before start of treatment with recombinant human therapeutics. The role of naturally occurring antibodies is not well understood and their influence on production of anti-drug antibodies is not known. One might speculate that the presence of naturally occurring antibodies increases the likelihood of eliciting anti-drug antibodies once treatment with the corresponding recombinant therapeutic protein is started. We screened serum samples from 410 healthy controls and patients for auto-antibodies against bone morphogenetic proteins (BMPs) 2 and 7 and interferon (IFN)-α, -β, and -γ in a new 3-step approach: rough initial screening, followed by competition and protein A/G depletion. Naturally occurring antibodies against these proteins were detected in 2% to 4% of the tested sera. Individuals who are 65 years or older had a slightly higher occurrence of naturally occurring antibodies. Auto-antibodies against BMP-7 and IFN-α were mainly comprised of IgM isotypes, and natural antibodies against BMP-2, IFN-β, and -γ were mainly IgG. To ensure assay specificity, assays were also used to detect antibodies against BMP-7 in patients being treated with rhBMP-7 before and after surgical procedure. Fifty percent of the treated patients had persistent anti-BMP-7 antibodies over time. The 3-step approach provides an attractive tool to identify naturally occurring antibodies in naive patients.

Cytokine-deficient mice as a model for generation of autologous anti-cytokine monoclonal antibodies

The possibility of inducing immune responses to murine interleukin-4 (IL-4) and IL-13 and generating anti-cytokine monoclonal antibodies (mAbs) was studied in IL-4- and IL-13-knockout mice. The minimal doses of IL-4 or IL-13 that could induce significant anti-cytokine responses with titers of 5000-10,000 were 20 microg per injection with the total doses of 100 microg. The highest titers in a range of 20,480-40,960 were achieved by triple immunization of IL-13-knockout mice with 30 microg of IL-13 per injection. Anti-IL-4 mAbs were generated at an antibody serum titer about 400; however, only 0.5% of primary hybridoma clones were anti-IL-4-positive. Anti-IL-13 cell fusions were successful at titers of 20,480 and 40,960 with 50% of the primary clones positive. Both hybridomas secreted low-affinity IgMkappa mAbs and were IL-6-dependent. These data demonstrate that IL-4- and IL-13-deficient mice may develop high polyclonal immune responses to "syngeneic" murine cytokines but fail to generate high-affinity mAbs at the tested conditions.

High levels of neutralizing IL-6 autoantibodies in 0.1% of apparently healthy blood donors

IL-6-specific autoantibodies (aAb-IL-6) have been reported in diseased and healthy individuals. We recently established a model for aAb-IL-6 in different mouse strains, based on vaccination with immunogenic IL-6 analogues, in which titers of aAb-IL-6 above 1,000 resulted in an in vivo IL-6 deficiency. Here, we examined aAb-IL-6 in 4,230 blood donors. Stable low titers of aAb-IL-6 were found in 9% of the donors, while 1% had titers ranging from 64 to greater than 10,000. Such aAb-IL-6-positive donors appeared normal with no overt signs of pathology. Natural and recombinant forms of IL-6 bound avidly to their IgG, and their plasma strongly neutralized IL-6 in vitro. Slightly elevated concentrations of IL-6 exclusively in the form of IL-6-IgG complexes were present in their circulation. The complexes did not contain soluble IL-6 receptors. Titers of 0.1% of the blood donors were as positive as the vaccination-induced IL-6-deficient mice. Such donors might be IL-6 deficient, and if so, IL-6 seems be dispensable for several months in otherwise healthy individuals. Such highly positive donors also explain why normal human IgG for pharmaceutical use may contain high anti-IL-6 activity. Finally, transfusion of plasma with a high titer of aAb-IL-6 might, temporally, render a recipient IL-6 deficient.

Naturally occurring anti-IFN-gamma autoantibody and severe infections with Mycobacterium cheloneae and Burkholderia cocovenenans

Recently various genetic defects in immunity mediated by interferon gamma (IFN-gamma) have been described, including mutations in the IFN-gamma receptor 1 (IFN-gammaR1) and receptor 2 (IFN-gammaR2), signal transducer and activator of transcription 1 (STAT 1), and interleukin 12 receptor beta 1 (IL-12Rbeta1), and IL-12 p40 genes. These mutations are associated with the occurrence of severe infections with intracellular pathogens especially nontuberculous mycobacteria and vaccine-associated bacilli Calmette-Guérin (BCG). Here we report data on a previously healthy adult patient primarily presenting with severe infections with Burkholderia cocovenenans and subsequently Mycobacterium cheloneae. We found a strong inhibitory anti-IFN-gamma activity in the patient's plasma and identified a high-affinity neutralizing anti-IFN-gamma autoantibody. Unfortunately, the patient died due to severe sepsis before we knew the nature of the inhibitory activity. The application of alternative therapeutic approaches such as intravenous immunoglobulin or immunoadsorption may have been beneficial in this case. Screening for neutralizing anti-IFN-gamma autoantibodies should supplement testing for IFN-gamma and IL-12 pathway defects in patients with recurrent infections with intracellular pathogens, especially with nontuberculous mycobacteria.

Identification of natural antibodies to interleukin-18 in the sera of normal humans and three nonhuman primate species

Natural antibodies to cytokines can be found in the sera of normal healthy individuals in the absence of specific immunostimulation. However, the function, impact, and purpose of natural antibody development have yet to be fully elucidated. Interleukin (IL)-18 is a cytokine that exerts proinflammatory activities and induces natural killer (NK) cell activity. Recombinant human IL-18 (rHuIL-18) is currently in development as a cancer immunotherapy. In this study, the presence of natural antibodies to IL-18 in the sera of normal humans and three nonhuman primate species was evaluated by electrochemiluminescence immunoassay (ECLIA). Of the human sera tested, 6 of 47 samples were positive for natural antibodies to IL-18. Of the nonhuman primate sera tested, 22 of 80 cynomolgus monkey samples, 4 of 31 rhesus monkey samples, and 2 of 20 chimpanzee samples were positive for natural antibodies to IL-18. Natural anti-IL-18 antibodies were neutralizing in 5 of 22 cynomolgus and 2 of 4 rhesus sera. None of the chimpanzee or human sera were able to neutralize IL-18 induction of interferon (IFN)-gamma in vitro. In vivo activity of rHuIL-18 was compared in IL-18 natural antibody-positive and -negative cynomolgus monkeys. The presence of natural antibodies to IL-18 did not alter rHuIL-18 systemic exposure levels, induction of neopterin, or induction of treatment-induced antibodies following intravenous administration of rHuIL-18. In conclusion, our data indicate that, as has been found with other cytokines, natural anti-IL-18 antibodies are relatively common. Moreover, natural anti-IL-18 antibodies do not appear to influence rHuIL-18 activity in vivo and are not predictive of a heightened immune response, suggesting that natural anti-IL-18 antibodies do not impact IL-18 therapy. Finally, our data suggest that the ability to detect natural anti-cytokine antibodies may be a useful measure of the adequacy of an assay for deployment in clinical trials.

Anti-cytokine autoantibodies in autoimmunity: preponderance of neutralizing autoantibodies against interferon-alpha, interferon-omega and interleukin-12 in patients with thymoma and/or myasthenia gravis

We have screened for spontaneous anticytokine autoantibodies in patients with infections, neoplasms and autoimmune diseases, because of their increasingly reported co-occurrence. We tested for both binding and neutralizing autoantibodies to a range of human cytokines, including interleukin-1alpha (IL-1alpha), IL-1beta, IL-2, IL-4, IL-6, IL-8, IL-10, IL-12, IL-18, interferon-alpha2 (IFN-alpha2), IFN-omega, IFN-beta, IFN-gamma, tumour necrosis factor alpha (TNF-alpha), transforming growth factor beta-1 (TGF-beta1) and granulocyte-macrophage colony stimulating factor (GM-CSF), in plasmas or sera. With two notable exceptions described below, we found only occasional, mostly low-titre, non-neutralizing antibodies, mainly to GM-CSF; also to IL-10 in pemphigoid. Strikingly, however, high-titre, mainly IgG, autoantibodies to IFN-alpha2, IFN-omega and IL-12 were common at diagnosis in patients with late-onset myasthenia gravis (LOMG+), thymoma (T) but no MG (TMG-) and especially with both thymoma and MG together (TMG+). The antibodies recognized other closely related type I IFN-alpha subtypes, but rarely the distantly related type I IFN-beta, and never (detectably) the unrelated type II IFN-gamma. Antibodies to IL-12 showed a similar distribution to those against IFN-alpha2, although prevalences were slightly lower; correlations between individual titres against each were so modest that they appear to be entirely different specificities. Neither showed any obvious correlations with clinical parameters including thymoma histology and HLA type, but they did increase sharply if the tumours recurred. These antibodies neutralized their respective cytokine in bioassays in vitro; although they persisted for years severe infections were surprisingly uncommon, despite the immunosuppressive therapy also used in most cases. These findings must hold valuable clues to autoimmunizing mechanisms in paraneoplastic autoimmunity.

Feline leukemia virus DNA vaccine efficacy is enhanced by coadministration with interleukin-12 (IL-12) and IL-18 expression vectors

The expectation that cell-mediated immunity is important in the control of feline leukemia virus (FeLV) infection led us to test a DNA vaccine administered alone or with cytokines that favored the development of a Th1 immune response. The vaccine consisted of two plasmids, one expressing the gag/pol genes and the other expressing the env gene of FeLV-A/Glasgow-1. The genetic adjuvants were plasmids encoding the feline cytokines interleukin-12 (IL-12), IL-18, or gamma interferon (IFN-gamma). Kittens were immunized by three intramuscular inoculations of the FeLV DNA vaccine alone or in combination with plasmids expressing IFN-gamma, IL-12, or both IL-12 and IL-18. Control kittens were inoculated with empty plasmid. Following immunization, anti-FeLV antibodies were not detected in any kitten. Three weeks after the final immunization, the kittens were challenged by the intraperitoneal inoculation of FeLV-A/Glasgow-1 and were then monitored for a further 15 weeks for the presence of virus in plasma and, at the end of the trial, for latent virus in bone marrow. The vaccine consisting of FeLV DNA with the IL-12 and IL-18 genes conferred significant immunity, protecting completely against transient and persistent viremia, and in five of six kittens protecting against latent infection. None of the other vaccines provided significant protection.

Anti-interleukin 8 autoantibody: interleukin 8 complexes in the acute respiratory distress syndrome. Relationship between the complexes and clinical disease activity

Increased levels of interleukin 8 (IL-8) are found in bronchoalveolar lavage (BAL) fluids from patients with the acute respiratory distress syndrome (ARDS). However, IL-8 is not an efficient predictor of the course of ARDS. Our prior studies demonstrated that IL-8 present in lung fluids from patients with ARDS is associated with anti-IL-8 autoantibodies (anti-IL-8:IL-8 complexes). These data led us to hypothesize that the complexes might better predict the development of acute lung injury. Accordingly, we measured concentrations of free and complexed IL-8 in BAL fluids from 19 patients at risk and 45 with established ARDS on Days 1, 3, 7, 14, and 21 after the onset of ARDS. The concentrations of anti-IL-8:IL-8 complexes in patients with ARDS on Day 1 were significantly higher than in patients at risk (p < 0.05). There was a significant association between anti-IL-8:IL-8 complex concentrations and the onset of ARDS (p = 0.03). Similarly, anti-IL-8:IL-8 complex concentrations were significantly higher in patients on Day 1 of ARDS who later died (p < 0.05), and the association between high anti-IL-8: IL-8 complex concentrations and the probability of dying was significant (p = 0.03). The presence of anti-IL-8:IL-8 complexes in BAL fluids of patients with ARDS is an important prognostic indicator for the development and outcome of ARDS.

Neutralizing antibodies to granulocyte-macrophage colony-stimulating factor, interleukin-1alpha and interferon-alpha but not other cytokines in human immunoglobulin preparations

Human immunoglobulin preparations are used therapeutically for various disorders. Such therapy is generally safe but adverse effects occasionally occur in recipients. It has been suggested that antibodies to cytokines present in clinical immunoglobulin products may contribute to undesirable effects in recipients. Therefore, we investigated intravenous and intramuscular immunoglobulin products for the presence of cytokine-specific neutralizing antibodies. Using validated bioassays, we detected neutralizing activity against human granulocyte-macrophage colony-stimulating factor (GM-CSF), interferon-alpha2a (IFN-alpha2a) and interleukin-1alpha (IL-1alpha) in immunoglobulin products. We found no neutralization of granulocyte colony-stimulating factor, macrophage colony-stimulating factor, stem cell factor, IL-1beta, IL-2, IL-3, IL-4, IL-6, IL-9, IL-10, IL-12, tumour necrosis factor-alpha, oncostatin M (OSM) and IFN-gamma. Most batches which neutralized IFN-alpha2a activity also neutralized other IFN-alpha subtypes, IFN-omega and IFN-beta. Most products (94%) neutralized the biological activity of GM-CSF. No correlation between batches and their ability to neutralize bioactivities of GM-CSF, IFN-alpha2a and IL-1alpha was found. This neutralizing activity could be traced to plasma pools used for manufacture of immunoglobulins. The neutralization was mediated by specific cytokine antibodies contained within immunoglobulin products as it was present in specific immunoglobulin G (IgG) fractions eluted from cytokine affinity chromatography columns. Specific binding of such IgG fractions to cytokines in immunoblots and in enzyme-linked immunosorbent assays (ELISAs) was observed. This contrasts with the broad non-specific recognition of cytokine proteins observed using unfractionated immunoglobulins in ELISAs. This is the first comprehensive study showing the presence of neutralizing antibodies against GM-CSF, IL-1alpha, or IFN-alpha2a in immunoglobulin products.

Evolution of a cytokine using DNA family shuffling

DNA shuffling of a family of over 20 human interferon-alpha (Hu-IFN-alpha) genes was used to derive variants with increased antiviral and antiproliferation activities in murine cells. A clone with 135,000-fold improved specific activity over Hu-IFN-alpha2a was obtained in the first cycle of shuffling. After a second cycle of selective shuffling, the most active clone was improved 285,000-fold relative to Hu-IFN-alpha2a and 185-fold relative to Hu-IFN-alpha1. Remarkably, the three most active clones were more active than the native murine IFN-alphas. These chimeras are derived from up to five parental genes but contained no random point mutations. These results demonstrate that diverse cytokine gene families can be used as starting material to rapidly evolve cytokines that are more active, or have superior selectivity profiles, than native cytokine genes.

Spontaneously occurring neutralizing antibodies against granulocyte-macrophage colony-stimulating factor in patients with autoimmune disease

There is increasing evidence that spontaneous anticytokine autoantibodies are associated with chronic infections and autoimmune diseases. We report the sporadic occurrence in autoimmune diseases of such autoantibodies to granulocyte-macrophage colony-stimulating factor (GM-CSF), a cytokine involved in inflammation and the regulation of proliferation, differentiation and function of granulocytic and monocytic cell lineages. In 41 of 425 patients tested, we found low to moderate levels of autoantibodies binding to GM-CSF in serum or plasma. These were most prevalent in patients with myasthenia gravis (MG). However, neutralizing autoantibodies against GM-CSF were very rare, being found in only three patients. Two had autoimmune MG, one with thymoma (Patient A) and the other (Patient B) with 'seronegative' MG, i.e. without the antiacetylcholine receptor autoantibodies characteristic of most MG patients, and a third (Patient D) had multiple sclerosis. Only very limited amounts of Patient A and Patient D serum/plasma were available for analysis and therefore further studies were carried out on the more plentiful samples from Patient B. The anti-GM-CSF autoantibodies of Patient B were predominantly polyclonal immunoglobulin G and strongly neutralized recombinant human (rh) GM-CSF derived from different expression systems. They had similar immunological and immunochemical characteristics to anti-GM-CSF antibodies that developed in immunocompetent colorectal carcinoma patients following (rh)GM-CSF therapy. In serial samples from Patient B, the anti-GM-CSF autoantibodies were undetectable from diagnosis at age 8 years until at least age 13, but then developed spontaneously during (temporary) withdrawal of immunosuppressive treatment. Their neutralizing activity has persisted since their first detection at age 15 years 1 month, and was at its highest level recently at age 17 years 7 months. There was no obvious association with other autoimmune phenomena, nor were any haematological deficiencies overtly manifested, suggesting that any loss of GM-CSF function may have been compensated for by other cytokines.