Research Directory

Report on a European Science Foundation Workshop on Invasive Aspergillosis

21st - 22nd October, 1998

Introduction

Dr Denning welcomed the participants. He described the size of the problem of invasive aspergillosis (IA) in terms of its incidence in different patient populations:

He noted the high mortality of invasive aspergillosis (50 to100 %) even with therapy and discussed the report from Frankfurt (Groll et al., J Infection 1996; 33: 23-32) which showed a 14-fold rise (over 12 years) in the incidence of invasive aspergillosis at autopsy. He presented antifungal sales figures from 1997 from Europe, USA and Japan of which $370 M from a total of $1230 M was spent on the only two licensed drugs with activity against Aspergillus - amphotericin B and itraconazole. He emphasised that invasive aspergillosis is now the leading cause of early death in many transplant centres and has a major impact on the management of leukaemia.

Dr Denning then presented an overview of strategies to address this problem (Figure I.). Subsequent contributions to the workshop were specifically targeted to one or more of these issues.

Finally, he informed the workshop of progress towards getting the genome of Aspergillus fumigatus sequenced. A pilot project to create and map a bacterial artificial library (BAC) of a pathogenic isolate of A. fumigatus is under consideration by Beowulf, the pathogen sequencing arm of the Wellcome Trust. This project is a collaboration initially between the Sanger centre in Cambridge and the University of Manchester. The project proposes to sequence ten clones - about 1 Mb of the estimated 30+ Mb genome. The region to be sequenced corresponds to a well characterised area of A. nidulans chromosome VIII (containing the nitrate assimilation gene cluster; crnA, niiA and niaD). The workshop unanimously agreed that a genome sequencing project was important to improve the identification of novel diagnostic antigens/targets, the development of new antifungal targets, the understanding of antifungal resistance and the understanding of the biology and pathogenicity of A.fumigatus. The sequence data could also generate new commercial enzyme or metabolic products because of the remarkable thermotolerance of A. fumigatus.

IA in neutropenia and bone marrow transplant recipients risk factors and outcome
Dr Martino gave an overview of the size of the problem of filamentous fungal infection in leukaemia and bone marrow transplantation. He noted that in the group he represented (GIMEMA group in Italy - 60 centres), 48 % of these infections occurred at the onset of leukaemia, for example during the first course of induction remission therapy. These infections were often fatal; either directly because antifungal treatment failed or because chemotherapy for the leukaemia was delayed, compromising its treatment. Filamentous fungal infections therefore present a serious problem in the management of leukaemia today. Dr Martino also discussed the control of Aspergillus in the environment and focussed on the recent descriptions of Aspergillus in water supplies and showerheads, the significance of which is currently unclear.

IA in solid organ transplant recipients risk factors & outcome
Dr Pahissa presented an overview of the problems of invasive aspergillosis in solid organ transplant recipients. He described the group set up in Spain to study infection in these patients (GESITRA - SEIMC). This group, comprising transplant surgeons, transplant physicians and infectious disease physicians collectively cares for about 1400 renal, 1700 liver, 900 heart and 60 lung transplant recipients annually. He noted that in their experience the mortality of invasive aspergillosis varied from 27 to 77 %. He also observed that the median time to diagnosis was six days, and that 78 % of diagnoses were made antemortem. In their multivariate analysis the following were all independent risk factors for death: renal insufficiency, the need for additional surgery because of surgical complications during the transplantation procedure, a lack of cotrimoxazole prophylaxis, chronic rejection and retransplantation. Only the use of AmBisome was an independent predictor of survival.

Clinical and radiological diagnosis in non neutropenic patients
Dr Grossi described diagnostic approaches in solid organ transplant patients with a particular emphasis on lung transplant patients. These approaches were based on the experiences of the 12 centres in the Italian Study Group for Fungal Infections in Thoracic Organ Transplantation. He noted that most cases occurred two to three months after transplantation. In lung transplant recipients their experience of bronchoalveolar lavage for diagnosing invasive aspergillosis is shown below:
 

Sensitivity = 100 %, specificity = 92 %
Positive predictive value = 16 %
Negative predictive value = 100 %

Some patients had multiple positive cultures and did not develop IA although almost all were given oral itraconazole. He observed that among the lung transplant recipients, 35 % of those with IA had dissemination to other organs and overall, the mortality in their hands was about 35 %.

Serological diagnosis
Dr Latgé described approaches to serological diagnosis that have been used. He mentioned interest in mannitol and certainly this metabolite of Aspergillus has been detected in patients and in experimental infections. However mannitol has not been systematically used in the clinical setting. Antibody detection is of great use in the diagnosis of aspergilloma and allergic bronchopulmonary aspergillosis, but has been disappointingly negative in patients with invasive aspergillosis. All controls have low levels of detectable antibody and so the standard cut-off for a positive test is a one in ten dilution. Study of lower levels of positivity or a fall off in levels as antibody is 'consumed' were presented as possible ways forward. Another suggestion was to improve the actual antigen(s) and standardise these with recombinant technology.

Dr Latgé also described attempts to localise infection using directly injected optical brighteners. This needs more work but is conceptually attractive. Similarly the concept of monoclonal antibody imaging has not been developed, but is attractive.

The only marketed antigen test in Europe is Sanofi Pasteur's Platelia sandwich ELISA. Dr Latgé remarked on its utility, noting that the galactomannan that is detected is produced by the fungus during infection irrespective of the host response. However it is chemically proven that many foods, especially noodles, contain galactomannan and this could account for the relative frequency of false positive tests. Little or no data exist on the usefulness of the test in non-haematology patients. He contrasted this antigen test with the so-called G-test that is marketed in Japan and detects 1-3, ß-D-glucan. The glucan test is likely to be positive (it is very sensitive) in patients with some degree of immune response as the release of 1-3, ß-D-glucan only occurs when the fungal cell wall is damaged. Comparative tests between ß-1-3 glucan and galactomannan have been started.
 

Polymerase Chain Reaction detection of Aspergillus DNA in samples
Dr Bretagne described the PCR tests that have been developed in his laboratory and their relative merits and problems. He outlined the various contamination problems that can beset PCR. These include extraneous contamination of the original sample or of the PCR mix with Aspergillus conidia and more importantly, contamination of the PCR mix with amplicons from prior reactions. Several stringent quality control measures were advocated to prevent false positive results, including the systematic use of uracil-DNA-glycosylase (UNG) which cleaves any contaminating amplicons prior to amplification. Dr Bretagne also advised against the use of nested PCR which is more liable to false positives. He also noted that inhibitors of the amplification reaction could yield false negative results. This inhibition can be detected by amplifying an internal control reaction in each sample tested. Finally Dr Bretagne argued that only high quality samples should be analysed to avoid interpretation problems.

PCR and T cells responses to Aspergillus
Dr Hebart described his group's results with a PCR test that firstly amplifies DNA from all pathogenic fungi and then utilises a hybridisation step to differentiate the underlying fungal pathogen. He presented the results of a prospective screening program on blood samples from recipients of an allogeneic stem cell transplant. In patients with proven and probable invasive aspergillosis, PCR positivity preceded clinical signs of IA by a mean of 21 days. PCR positivity in the later post transplant period was found to be associated with a graft versus host disease (GvHD) and corticosteroid treatment, major risk factors for IA. Based on these results, the group led by Drs Hebart and Einsele has started a multicenter trial comparing PCR-based versus empirical antifungal therapy after allogeneic stem cell transplantation.

Dr Hebart also described preliminary results on Aspergillus-specific lymphoproliferation in healthy volunteers and patients with IA. Lymphoproliferative T-cell responses were highest in patients surviving IA, whereas these responses were not detectable in recipients of an allogeneic stem cell transplant in the early post-transplant phase (up to day + 100). Further details are warranted.

How does immunosuppression, radio- and chemotherapy affect epithelium and immune cells?
Dr Poynton noted that little research had addressed the issue of local and systemic damage wrought by chemotherapy and radiotherapy relative to a) the number of patients treated and b) the size of the problem. In the context of acute induction therapy for leukaemia, he described experiments using mannitol and lactulose to quantify damage to the gut. Mannitol is actively reabsorbed in the normal gut, but lactulose absorption is a measure of damage to intercellular boundaries. A ratio of the two allows direct quantification of gut damage, which is considerable in the few patients they have studied so far. Further work has shown that the level of endotoxin in blood is directly related to gut damage. Thus fever in neutropenia could reflect gut damage and risk of infection as much as infection itself. Glutamine is somewhat protective of epithelial damage and could be important in protecting patients from aspergillosis or may be a marker of damaged epithelial surfaces.

Dr Poynton went on to discuss the relationship between cytokine expression and cytokine levels using a small dataset he had collected. It was already apparent that there was no relationship in the cells they examined (IL4 &IL12 positive CD4+ cells). Finally, he discussed the possible role of nitrous oxide (NO) and speculated that the anergy noted just before death could be attributable to 'disregulation of NO'.
 

How do phagocytes kill Aspergillus, an extracellular pathogen?
Dr Roilides gave an overview of the killing of Aspergillus conidia and hyphae by the oxidative and non-oxidative mechanisms of phagocytes (neturophils, monocytes, macrophages). He noted that the subcellular and biochemical mechanisms of phagocytes still remain unknown. He also emphasised the enhancing effects of hemopoietic and Th1 cytokines on certain antifungal activities of phagocytes against Aspergillus. He went on to discuss the unknown importance of non-oxidative enzymes and other substances including NO and its possible role in the killing of conidia by macrophages. Antimicrobial peptides of leukocytes (e.g. defensins, lactoferrin lysozyme, bactericidal permeability-inducing protein, calprotectin, protegrins, cecropins, etc) have not been investigated to any degree, but are probably important. Xoma, a California-based company, has developed a series of such compounds one of which has substantial in vivo activity in an animal model of aspergillosis. In other fungi substantial work has been done on the interactions of phagocytes, hemopoietic cytokines (e.g. granulocyte colony stimulating factor) and antifungal drugs (e.g. azoles), a synergism which may be very important in patients with fungal infections. However, similar work has not been performed with Aspergillus or other antifungal drugs (e.g. lipid formulations of amphotericin B, echinocandins, pneumocandins, etc).

Cytokine responses to Aspergillus infection
Dr Romani spoke about the substantial variation in susceptibility to invasive aspergillosis in different strains of mice. In a murine model of invasive aspergillosis, production of Th1 and Th2 cytokines were different in mice resistant or susceptible to the infection. Development of protective acquired immunity required IL-12 and was associated with the production of IFN-g from Th1 cells upon in vitro antigen-specific stimulation. In contrast, production of IL-4 and IL-10 by antigen-activated Th2 cells was associated with disease progression and down-regulation of protective Th1 immunity. IL-4 neutralisation or IFN-g administration had a curative effect in invasive aspergillosis, indicating the critical role of these cytokines in the pathogenesis of this infection. Similar to observations in mice with invasive aspergillosis, innate immunity and the generation of Th1 cytokines play an essential role in host defenses against invasive pulmonary aspergillosis (IPA). Resistance to infection was associated with the unimpaired innate antifungal activity of pulmonary phagocytic cells. This was concomitant with high-level production of TNF-a  and IL-12 and the presence of interstitial lymphocytes producing IFN-g. Conversely, production of TNF-a and IL-12 was down-regulated in highly susceptible mice showing defective innate antifungal immunity and high-level production of IL-4 and IL-10 by lung lymphocytes. Resistance was increased in susceptible mice upon local IL-4 or IL-10 neutralisation or IL-12 administration

Circumstantial evidence in humans indicates that dysregulated production of Th cytokines may contribute to the pathogenesis of IPA. These include the pathogenetic role IL-10 may have in aspergillosis in patients with HIV infection, the occurrence of pulmonary aspergillosis in patients with the hyperimmunoglobulin E syndrome and the recently reported ability of human lymphocytes to respond to A.fumigatus conidia in vitro with activation, proliferation, and production of Th1 cytokines. Should this last finding be extended to patients, it may be possible to identify persons at increased risk of aspergillosis by the quality of the Th cell response to in vitro Aspergillus antigens.

Genetics of cytokine responses to micro-organisms
Dr Ollier gave a broad introduction to genetic factors determining immune response. He spoke of polymorphisms in the promoter regions of TNF-a  and the correlation between IL-10 polymorphisms and circulating levels of this cytokine. He summarised the data on mannose binding lectin (MBL) which binds to Aspergillus, is important in the clearance of pathogens and immune complexes and is an activator of the alternate complement cascade. Five mutations have been described, two in the promoter region (H/L and Y/X) and three in the open reading frame (at codons 52, 54 and 57). These lead to more or less functional circulating MBL. The codon 54 mutation is present in 16 % of caucasians and is homozygous in 2 %. Combinations of promoter mutations lead to higher or lower serum MBL levels so that the HY genotype has high levels and the LX genotype is essentially deficient. MBL deficiency leads to an increased frequency of bacterial infections particularly otitis media and pneumonia. No one has investigated its possible role in any of the forms of aspergillosis.

NIH funding
Dr Dennis Dixon, Chief of the Bacteriology and Mycology Branch of the NIH/NIAID/DMID, outlined the basic approaches to NIH funding and the internal structures of the NIH. He indicated that funding for fungal projects could come from several sources within the NIH, although most is handled, at least in part, by the Division of Microbiology and Infectious Diseases. Over the last three years (1995-1997) funding for the mycology program within that Division had risen to around $15 M annually. The program goals for mycology were:
 

1) Diagnosis, treatment and prevention of serious mycoses
2) Mechanisms of pathogenesis
3) Improvements in treatment through basic and clinical studies, including comparative clinical trials.

He identified the gap areas in mycology as:

1) Training in contemporary mycology
2) Research on fungal vaccines
3) Research on Aspergillus and other moulds
4) Fungal genome research on human pathogens

He noted that the NIH would be very supportive of any EU effort to sequence the A.fumigatus genome. The NIH website carries much information - including reports on mycology workshops
 

In vitro and animal models for the study of pathogenesis
Dr Latgé described the animal models that have been used to study virulence. He summarised the findings to date as essentially negative, despite the considerable investment of time and resources in generating and evaluating isogenic single and double mutants. Many questions remain: for example, the fate of conidia, the different steps in the infection process and why there are so many different manifestations of aspergillosis.

Dr Latgé also summarised the issues regarding adhesion. Conidia adhere to several separate components of the external milieu including collagen, complement, fibrinogen, laminin, fibronectin, surfactant, and a phagocyte receptor. The conidial cell wall has three identified binding components, rodlet proteins, pigment and lectin-like coating. Binding to fibrinogen and laminin is for example, sialic acid dependant. Given this multiplicity of possible binding combinations, it may be unrealistic to consider blocking it as a therapeutic avenue.

Dr Latgé then addressed the question of the mechanisms of conidial killing by alveolar macrophages and concluded that we do not know how they kill conidia. The same is almost true with neutrophils.

He raised a critical point for future work, namely a definition of virulence. Most workers have been using increased or reduced mortality in a mouse model as the definition, but this is a crude approach and subject to much experimental error. He recommended that such a definition be agreed in the near future. Identification of genes differentially expressed during infection would avoid any wrong a priori selection of putative virulence factors. A program to approach this strategy is presently being undertaken at the Pasteur Institute.

Single gene disruption in Aspergillus
Dr Monod presented the different strategies available to obtain single disruption events in A. fumigatus. Two drug resistance markers (hygromycin and phleomycin) have been used to successfully transform A. fumigatus. Dr. Monod gave a list of the different genes which have been disrupted to date including various proteases, catalases, ribonuclease and hydrophobin. Some experimental data supported a putative role for these proteins during infection. Unfortunately, in a mouse model of IA, no differences were seen between the mutants and parental strains. These results called into question the animal model used to test the virulence of A. fumigatus strains and raised again the question of a polygenic nature of virulence in this opportunistic fungus.

Dr. Monod also introduced us to Invitrogen's Pichia pastoris system used successfully for the production of recombinant proteins of A. fumigatus. Pichia pastoris respects the glycosylation of the protein and allows the production of high concentrations of protein (0.1 mg/ml) which can be secreted in the culture medium.

New molecular approaches
Dr Jeremy Brown presented an overview of the restriction enzyme mediated integration (REMI) approach to virulence gene identification and described his work over three years. The signature tagged mutagenesis - restriction enzyme mediated integration (STM-REMI) approach allows the production of a bank of mutants by unique insertion of a resistance marker at a random site in the A. fumigatus genome. Every mutant is tagged at a different locus which facilitates simultaneous testing of 96 different mutants in the same mouse. Any mutant(s) which have not grown in the mouse must have a gene essential for virulence disrupted and these genes can be cloned owing to their unique flanking sequence tag. He described how 4648 mutants were tested (56 pools of mutants on three to five mice per pool). From this large screen, two mutants of interest were found. The first was less virulent and had a disruption event in a small ORF; this ORF revealed no homologues. The second mutant was a para-amino benzoic acid (paba) auxotroph and was fully avirulent (not germinating in the mouse lung). These results also questioned the importance of single mutations in the study of A. fumigatus virulence and would suggest that virulence is multifactorial.

Identifying new candidates as virulence determinants biochemically
Dr Ruchel presented what little was known about secreted enzymes in Aspergillus. Various secreted or cellular proteins of the opportunistic pathogen A.fumigatus have been suspected of being pathogenicity factors. However, targeted deletion of individual corresponding genes has met with discouraging results from subsequent experimental infections. The normal habitat of A.fumigatus is not the warm blooded host, but rather moist organic matter in the environment. Therefore, the existence of such "virulence factors", expressed by the fungus for survival in the human host, is questionable. However, we assume that fungal proteins may still warrant attention because proteins critical in essential biochemical cellular pathways or for fungal growth in the host could be targets for future pharmacotherapy. These proteins should be enzymes, and thus targets of inhibitors and must differ from human proteins. Candidates are enzymes of the cell wall, of the vacuole, or of other cellular compartments which are specific to fungi. Dr Ruchel favoured vacuole-associated enzymes.

Mycotoxins and Aspergillus fumigatus
Dr Galtier described what was known about the toxins of Aspergillus. He noted that the tryptostatins (A, B) inhibit cell cycle progression at M phase, but the relative toxicity of each is not known. He noted that there were four tryptoquivalines all with low toxicity potential. In contrast, verruculogen was very toxic causing tremors and fits. Its mechanism of action is not fully characterised, but it may interact with the neurotransmitter - amino butyric acid (GABA). Similar but in decreasing order of toxicity are the fumitremorgens A, B, C and TR-2 and C.

Dr Galtier then focussed on gliotoxin which is produced by A.fumigatus, A.flavus and C.albicans. A six day culture of each organism in Eagle's or YES medium with 5 % fetal calf serum will yield relatively large quantities of gliotoxin. It has been found in necrotic bovine udders, peritoneal washings in infected mice and in lung tissue of turkeys in very variable concentrations. It is rapidly taken up by cells and causes reduced phagocytosis and apotosis probably via multiple mechanisms including RNA polymerase inhibition, inactivation of intracellular proteins and possibly the generation of intracellular hydrogen peroxide. Gliotoxin may also cause damage to epithelial cells. Dr Galtier outlined the probable synthetic pathway for gliotoxin and noted that the pertinent enzymes were not known or cloned.

Epidemiology of resistance in Aspergillus
Dr Rodriguez-Tudela gave a short summary of progress in antifungal susceptibility testing over the last decade. He outlined the impressive progress towards standardisation in yeast susceptibility testing by the NCCLS subcommittee in the US. He outlined problems with the methodology and breakpoint interpretation including the testing of amphotericin B and the poor growth of some yeasts (especially Cryptococcus) in RPMI without glucose supplementation. He then spoke about 'problem species' and the need to standardise methods and conduct contemporary surveys of resistance in Aspergillus. He mentioned the formation of a new European Committee for Susceptiblity Testing (EUCAST) subcommittee on antifungal susceptibility testing that would address many of these issues.

Molecular basis for resistance in Aspergillus
Dr Denning described what little was known about itraconazole resistance in Aspergillus fumigatus. Two mechanisms of resistance were postulated a) overexpression of an efflux pump leading to reduced intracellular concentrations of itraconazole and b) mutation of the itraconazole target, 14a -demethylase. Work proceeds on cloning the relevant genes. Dr Denning then discussed the issue of cross resistance as animal models and in vitro studies have indicated a degree of cross resistance with SCH 56592 (an azole in phase II development), but no cross resistance to voriconazole (an azole in phase III development). The structure SCH 56592 is similar to that of itraconazole, whereas that of voriconazole is not.

He then showed some animal model data consistent with amphotericin B resistance in one isolate of Aspergillus. He mentioned a recent paper by Dr Lass Florl in the Journal of Antimicrobial Chemotherapy which purported to be able to separate, with a breakpoint of >2 mg/ml, amphotericin B resistant isolates from susceptible ones. Those patients infected with the 'resistant' isolates, all died. There was some missing data from the paper, but overall it seems possible that amphotericin B susceptibility testing will soon be possible for Aspergillus. The mechanism of action of amphotericin B in Aspergillus is unknown.

Dr Denning then bid farewell to all the participants and thanked them for their contributions and attendance.

Prepared by D.W. Denning and J.P. Latge, March 1999

Participants