The innate immune system represents the initial arm of host defense against pathogenic bacteria, fungi, and parasites. Neutrophils, monocytes, and tissue-based macrophages are major cellular components of this system. The potential ability to augment activity of the innate immune system has increased dramatically during the past 2 decades, with the discovery and development of cytokines. Four cytokines, namely granulocyte colony-stimulating factor (G-CSF), granulocyte-macrophage colony-stimulating factor (GM-CSF), macrophage colony-stimulating factor (M-CSF), and interferon (IFN)-γ, have received increasing attention as potential adjunctive agents for the treatment of infectious diseases. In various animal models of infection, therapeutic administration of each of the 4 cytokines has been shown to enhance pathogen eradication and to decrease morbidity and/or mortality. However, variable therapeutic efficacy has been reported in clinical trials conducted to date. This review summarizes the current status of the use of G-CSF, GM-CSF, M-CSF, and IFN-γ in the treatment of infectious diseases.
Two different genes encoding the amino acid sequence of GCSF initially were isolated independently from different tissue sources. Nagata et al. [1] isolated cDNA for G-CSF, which encoded a predicted amino acid sequence of 177 aa, from a squamous cell line. Because of alternativemRNAsplicing sites, Souza et al. [1] isolated a different cDNA for G-CSF, which coded for a polypeptide of 174 aa from a bladder carcinoma cell line. The G-CSF gene is located on chromosome 17 q21–22 near other genes involved in the development of PMNL, and the G-CSF receptor is encoded by a single gene on chromosome 1 p35–p34.3 [3].
G-CSF is produced primarily by monocytes/macrophages, fibroblasts, and endothelial cells [4]. The major target cells of G-CSF are PMNL precursors and mature PMNL [4]. G-CSF plays an essential role in normal regulation of PMNL development, as demonstrated in the G-CSF knockout mouse. Mice renderedG- CSF-deficient by targeted disruption of the G-CSF gene in embryonic stemcells developed chronic neutropenia associated with a 50% reduction of granulocyte precursor cells in bonemarrow. G-CSF knockout mice exhibited a markedly impaired ability to control infection by Listeria monocytogenes and failed to develop sepsis-related neutrophilia [5].
In healthy persons, mean (±SD) serum levels of G-CSF are 25 ± 19:7 pg/mL. During the acute stage of an infection, serum levels of G-CSF increase by ∼30-fold; in endotoxin-induced shock, levels increase to ∼200 ng/mL [6–8]. This increase is mediated by bacterial products and inflammatory mediators, such as tumor necrosis factor (TNF)-α [9]. On the basis of a model of sepsis in rats, it appears that G-CSF may also function as a negative feedback signal to inhibit synthesis and release of TNF-α [10]. Similar to erythropoietin in anemic patients, serum levels of G-CSF can be elevated in individuals with a variety of both congenital and acquired forms of neutropenia, including idiopathic aplastic anemia,myelodysplastic syndrome, severe congenital neutropenia, and cyclic neutropenia [11–14]. However, similar to the reported behavior of thrombopoietin in immune thrombocytopenic purpura, G-CSF production in autoimmune neutropenia of infancy does not appear to increase, despite the associated low absolute neutrophil count (ANC) [15].
G-CSF in animal models of infection. Prophylactic treatment withG-CSF has been reported to decrease sepsis-relatedmortality in mice [10]. In various other neutropenic and nonneutropenic animal studies involving severe burns, intramuscular abscesses, streptococcal infections, pneumonias, viral infections, peritonitis, and sepsis, the administration of G-CSF has been associated with reduced mortality [16, 17]. For example, G-CSF reduced mortality from 90% to 28% in nonneutropenic mice with experimental acute disseminated Candida albicans infection [18]. Furthermore, outgrowth of C. albicans from the kidneys, spleens, and livers of G-CSF-treated mice was significantly reduced. Both PMNL count and PMNL killing activity were significantly increased in rodents who received prophylactic G-CSF prior to experimental Escherichia coli peritonitis [19]. Survival was also favorably affected by G-CSF pretreatment, with mortality reduced from 78% in control animals to 38% in G-CSF-treated animals.
Treatment of mice with G-CSF initiated prior to cecal ligation and puncture and subsequent treatment with antibiotic therapy improved survival, decreased splenic bacterial colony-forming units and serum TNF-α levels, and increased serum interleukin (IL)-10 levels, compared with treatment with antibiotics alone or with saline [20]. G-CSF treatment of rats with experimental bacterial pneumonia resulted in divergent clinical responses, depending on the bacterial species causing infection [21]. In Staphylococcus aureus pneumonia, G-CSF increased PMNL count and bacterial clearance, resulting in decreased rates of pulmonary injury and death. However, in E. coli pneumonia, which is associated with high TNF-α levels, bacterial clearance was impaired in G-CSF-treated animals, resulting in increased pulmonary injury and mortality. It also has been reported that pretreatment with G-CSF promotes replication of bacteria in liver and spleen in a murine Klebsiella pneumoniae pneumonia model, suggesting that G-CSF may impair, rather than enhance, antibacterial activity in certain situations in vivo [22].
Clinical use ofG-CSF. G-CSF has been employed therapeutically to increase the PMNL count in congenital neutropenia, idiopathic neutropenia, leukemic neutropenia, and aplastic anemia [23–33]. The use of G-CSF in congenital and acquired severe chronic neutropenia, including cyclic neutropenia (hematopoiesis), has had a major positive impact on morbidity, mortality, and quality of life for affected individuals. The clinical benefits observed have led to the widespread acceptance of prolonged G-CSF therapy as the standard of care for these relatively rare hematologic disorders [23–33]. In cases of severe congenital neutropenia that are unresponsive to standardG-CSF treatment, combination therapy with corticosteroids and G-CSF may be beneficial [34, 35].
G-CSF is widely used clinically to prevent febrile neutropenia or to shorten the duration of neutropenia associated with chemotherapy, radiotherapy, or myelosuppressive drugs. Opportunistic bacterial and fungal infections continue to be a major cause of morbidity and mortality in patients with cancer receiving myelosuppressive chemotherapy. Both the degree and duration of neutropenia have been shown to affect the risk for and clinical outcome of opportunistic infection [36–38]. On the basis of randomized trials of patients with chemotherapy-induced neutropenia, G-CSF has been approved for acceleration of myeloid recovery in patients after standard-dose chemotherapy for solid tumors and hematologic malignancies [37, 38]. When administered prophylactically to individuals undergoing chemotherapy in clinical trials, G-CSF has significantly reduced the resulting period of neutropenia and decreased the incidence of episodes of febrile neutropenia in high-risk patients [37–40]. However, the proper role of G-CSF in the treatment of a patient who develops febrile neutropenia has remained controversial, despite several previous randomized trials designed to address this issue [41–43].
Patients with febrile neutropenia are not a homogeneous group and vary in terms of clinical prognosis. Prospectively validated risk factors for complications and poor clinical outcome include prior inpatient status, serious independent comorbidity, and uncontrolled malignant disease [44, 45]. Subset analysis in one clinical trial suggested that G-CSF treatment provides the greatest benefit to patients with an ANC of <100 cells/μL and/ or documented infection [41].
The efficacy of adding G-CSF to empiric broad-spectrumantibiotic treatment was examined recently in a prospective, multicenter, randomized clinical trial, involving 210 patientswith solid tumors and high-risk febrile neutropenia. All patients enrolled in this study had an axillary temperature of>38°C with an ANC of <500 cells/mL and met at least one of the following high-risk criteria: (1) profound neutropenia (ANC of <100 cells/mL); (2) short latency from previous chemotherapy cycle (<10 days); (3) sepsis or clinically documented infection; (4) severe comorbidity (respiratory failure, congestive heart failure, uncontrolled cardiac arrhythmia, renal failure, hepatic dysfunction, severe emesis, severe mucositis, severe diarrhea, symptomatic hypercalcemia, or uncontrolled bleeding); (4) performance status >3 on the Eastern Cooperative Oncology Group scale [46]; (5) prior inpatient status; and (6) failure of out-patient management of lowrisk neutropenia. In this patient population with high-risk neutropenia, G-CSF treatment was associated with shortened duration of neutropenia, decreased duration of antibiotic therapy and hospitalization, and reduced overall hospital costs [47].
Much interest has emerged recently regarding the use of GCSF to mobilize CD34+ hematopoietic stem cells from marrow to blood (peripheral blood stem cells [PBSC]) for use in lieu of bone marrow cells in hematopoietic transplantation [48]. The discovery of G-CSF as a means to elevate blood PMNL counts also has rekindled interest in granulocyte transfusion therapy for treatment of serious bacterial and fungal infections in neutropenic patients [49].
The safety and survival data from animal models of infection, combined with the favorable toxicity profile in humans, have led to several clinical trials of G-CSF as adjunctive therapy in a variety of conditions. For example, the administration of GCSF has reversed or prevented neutropenia in human immunodeficiency virus (HIV)-infected patients [50–59]. Neutropenia is a frequent complication of HIV infection, occurring in 35%–75% of patients with AIDS [51, 52]. The etiology of neutropenia in this setting is multifactorial and includes bone marrow infiltration or infection, depressed production of endogenous myeloid growth factors, myelosuppressive drug therapies, the presence of HIV antibodies and autoimmune mechanisms, and accelerated neutrophil apoptosis [51, 52, 60–67]. Neutropenia represents a significant risk factor for bacterial infection in HIV-infected patients [50–52, 68–71]. Furthermore, a variety of functional activities, including chemotaxis, phagocytosis, expression of cellular adhesion molecules, and production of reactive oxygen intermediates, reportedly are impaired in the PMNL of HIV-infected individuals [50–52, 72–77]. These functional deficits can be partially corrected by G-CSF, and clinical trials have shown that G-CSF administration can reduce the incidence of bacterial infections and the number of consequent days of hospitalization required for HIV-infected individuals [50–52, 54–59, 72, 78].
Rates of opportunistic fungal infections have increased substantially in both Europe and North America [79]. Because PMNL constitute the main mechanism of host defense against opportunistic fungi, including Candida and Aspergillus species, these infections occur predominantly in patients with neutropenia or impaired PMNL function [79, 80]. The ability of PMNL to induce damage to fungal pathogens is augmented by G-CSF [81]. The candidacidal activity of human PMNL was significantly increased following pretreatment with G-CSF [82]. Moreover, G-CSF may prevent, at least partially, corticosteroid-induced suppression of PMNL-mediated activity against opportunistic fungi [83]. However, controlled clinical trials are necessary to establish a definitive role for G-CSF as an adjunctive immunomodulatory agent in fungal infections.
Therapeutic administration of G-CSF has been investigated in several studies of severe community-acquired pneumonia. A double-blind, controlled, multicenter trial enrolled 756 patients with community-acquired bacterial pneumonia to receive intravenous antibiotics plus either G-CSF (300 μg/day for up to 10 days; n = 380) or placebo (n = 376) [84]. Outcome measures included time to resolution of morbidity (TRM), 28-day mortality, length of hospital stay, and adverse events.A microbial cause for pneumonia was identified in 56% of patients, and antimicrobial use was judged to be appropriate for 98% of the patients by an independent review group. Administration of G-CSF increased the peripheral blood PMNL count 3-fold but failed to affect TRM, mortality, or length of hospitalization. However, radiographic infiltrates resolved more rapidly in the G-CSF-treated group and only 1 patient developed empyema, compared with 6 patients in the placebo group (P = .068). Not only was the administration of G-CSF safe and well tolerated but the development of sepsis-related organ failure, acute (adult) respiratory distress syndrome, and disseminated intravascular coagulopathy was significantly reduced in G-CSF recipients (P < .017 and P < .007, respectively). The clinical benefits of G-CSF therapy appeared to be more pronounced in the subgroup of patients with multilobar pneumonia.
To further evaluate this latter observation, 480 patients with multilobar community-acquired pneumonia were subsequently randomized to receive, in addition to standard therapy, G-CSF (300 μg/day; n = 237) or placebo (n = 243) for a maximum of 10 days [85]. No significant toxicity was associated with G-CSF administration in this patient population. However, although there was a trend toward reduced mortality in the subpopulation of G-CSF-treated patients with pneumococcal bacteremia, GCSF failed to improve clinical outcome in the overall study population [85].
Other pilot studies have yielded promising results when GCSF has been used to treat a variety of infectious conditions in nonneutropenic patients. In a randomized, double-blind, placebocontrolled trial performed at a single center, 40 insulin-dependent diabetic patients with foot infections were randomly assigned to receiveG-CSF (n = 20) or placebo (n = 20) for 7 days [86]. Both groups were treated with similar antibiotic and insulin regimens. G-CSF treatment was associated with significantly earlier eradication of pathogens (median, 4 vs. 8 days), quicker resolution of cellulitis (median, 7 vs. 12 days), a shorter hospital stay (median, 10 vs. 17.5 days), and a shorter duration of antibiotic treatment (median, 8.5 vs. 14.4 days). No G-CSF-treated patient required surgery, whereas 2 placebo recipients underwent amputations and 2 required extensive debridement under anesthesia. Corroboration of these observations in a randomized multicenter trial will be required before consideration of G-CSF treatment of diabetic foot infection as standard clinical practice.
Because esophagectomy for patients with esophageal carcinoma is associated with substantial infectious complications, the use of G-CSF to reduce the number of postoperative infections in this patient population was investigated at a single center [87]. Nineteen patients were treated perioperatively by daily administration of G-CSF for 10 days beginning 2 days before surgery. Their outcome was compared historically with 77 patients with esophageal cancer who did not receive G-CSF.Within the first 10 days after surgery, 23 untreated patients but none of the G-CSF-treated patients developed infections (29.9%vs. 0%; P = .005). After cessation of G-CSF therapy, 2 cases of infection developed in the G-CSF-treated group, compared with 6 additional cases of infection in the untreated group (10.5% vs. 37.7%; P < .05). Although the rate of fatal hospital infections was reduced in the G-CSF group, this difference did not reach statistical significance (G-CSF-treated patients, 9%; historical control patients, 0%).
G-CSF is approved by the Food and Drug Administration (FDA) for treatment of patients with cancer receiving myelosuppressive chemotherapy, after induction or consolidation chemotherapy for acute myelogenous chemotherapy, for myeloid reconstitution after hematopoietic stem cell transplantation, for mobilization and collection of PBSC for transplantation, and for severe chronic neutropenia. Results of clinical studies conducted to date suggest potential use of G-CSF as a therapeutic agent to reduce complications of infections in not only neutropenic but also nonneutropenic patient populations. However, only large, randomized, controlled clinical trials will establish the rational role of G-CSF for the prevention and treatment of specific infections, particularly in patients with compromised PMNL reserves or function.
The principal cellular sources of GM-CSF are T lymphocytes, monocytes/macrophages, fibroblasts, and endothelial cells [4]. GM-CSF stimulates a variety of functional activities in PMNL, eosinophils, and monocytes/macrophages but does not appear to play an essential role in normal PMNL development. Mice lacking GM-CSF as a result of targeted genetic disruption develop impaired pulmonary homeostasis, manifested as alveolar proteinosis, and increased splenic hematopoietic progenitors; however, steady-state hematopoiesis is not impaired [88, 89].
GM-CSF is produced as a polypeptide of 144 amino acids, from which a 17-aa segment of the amino terminus is cleaved to yield the mature 127-aa protein [90]. The gene encoding human GM-CSF is located on chromosome 5, region 5q21–5q32 [91]. Other genes in this area include those for M-CSF, IL-3, IL-4, and IL-5. In knockout mice lacking functional IL-3, GMCSF, and IL-5, no hematologic defect other than a reduced number of eosinophils has been detected [92].
To elucidate and characterize the effects of GM-CSF on cells of the granulocyte and monocyte lineages, recombinant human GM-CSF was administered to 7 healthy volunteers daily for 14 days [93]. All volunteers developed eosinophilia, and peripheral blood PMNL counts rose gradually to peak at a level 3.5-fold greater than baseline. Marrow aspirates on day 5 of GM-CSF treatment revealed a statistically significant increase in the proportion of promyelocytes and myelocytes, which was accompanied by a significant decrease in bands and segmented PMNL. Similar effects were observed when HIV-infected patients were treated with GM-CSF [94].
GM-CSF in animal models of infection. Variable responses have been reported when GM-CSF has been studied as a therapeutic immunomodulatory agent in experimental animal models of infection. A critical role for GM-CSF in pulmonary host defense was demonstrated recently in mice rendered deficient in GM-CSF by targeted genetic disruption (GM-CSF knockout mice). Susceptibility to group B streptococcal pneumonia was increased in the mice [95]. In a Pneumocystis carinii pneumonia model induced by intratracheal inoculation, CD4+ lymphocyte- depletedmicewere treated with GM-CSF [96]. This study demonstrated that GM-CSF plays a critical role in the inflammatory response to P. carinii under conditions of impaired cell-mediated immunity. When administered prophylactically, GM-CSF also significantly improved survival in a neonatal rat model of staphylococcal sepsis [97]. In contrast, GM-CSF failed to improve survival in rats with peritonitis induced by cecal ligation and puncture [98].
Considerable research effort has been directed toward the potential use of GM-CSF as adjunctive treatment for fungal infection. In one study of experimental candidemia in mice, GMCSF treatment significantly enhanced survival through day 15 of infection and improved clearance of C. albicans from liver and spleen but not kidney [99]. In a neutropenicmousemodel, overall survival was increased when GM-CSF was administered prophylactically prior to experimental C. albicans, Pseudomonas aeruginosa, or S. aureus infection [100].
Clinical use of GM-CSF. The predominant clinical use of GM-CSF is to acceleratemarrow recovery after cancer chemotherapy. Its efficacy has been evaluated in patients with cancer chemotherapy-induced myelosuppression and in patients undergoing bone marrow or peripheral hematopoietic stem cell transplantation [101–105]. Moreover, considerable interest has focused recently on the use ofGM-CSF for ex vivo expansion of hematopoietic stem and progenitor cells for a variety of applications, including in vitro tumor cell purging and reduction in the volume of blood required for processing during leukapheresis procedures for collection of hematopoietic stem cells [105].
During the past decade, GM-CSF therapy has been investigated in a variety of clinical settings. Candida and Aspergillus species are consistently noted as the most important fungal pathogens in patientswith cancer [106]. In a small, uncontrolled trial, neutropenic patients with documented fungal infections received amphotericin B plus GM-CSF [107]. Six of 8 evaluable patients responded, including 4 who experienced complete clearance of infection. In the same study, favorable clinical responses alsowere reported for patients with bacterial infections who received GMCSF treatment. In another study, the fungistatic activity of human PMNL and monocytes against C. albicans was significantly increased by in vitro culture with GM-CSF for 24 h [108].
Contrasting results have been reportedwhenGM-CSF has been studied as an immunomodulatory agent to augment host defense against Mycobacterium avium. In a randomized study, 30 HIV infected patients without M. avium infection were randomized to receive GM-CSF, azithromycin, or both agents [109]. After GM-CSF therapy, neither PMNL nor monocytes could significantly reduce M. avium growth. Furthermore, at the dose used in this study (250 μg/m2/day for 5 days), GM-CSF slightly increased HIV load and frequently caused adverse reactions such as bone pain andmyalgia. In another study, GM-CSF significantly improved HIV-1-mediated impairment in the ability of monocyte- derived macrophages to phagocytose M. avium complex in vitro, suggesting a possible role for GM-CSF as a therapeutic agent to restore monocyte/macrophage-mediated antimycobacterial activity in HIV-infected individuals [110]. In addition, it was shown recently that GM-CSF inhibits HIV-1 replication in monocyte-derived macrophages and, therefore, may exert a favorable immunomodulating effect against HIV infection itself [111].
GM-CSF is approved by the FDA for patients after induction chemotherapy for acute myelogenous leukemia or after transplantation of autologous PBSC, for mobilization and collection of autologous PBSC for transplantation, for myeloid reconstitution after autologous or allogeneic hematopoietic stemcell transplantation, and for individuals experiencing hematopoietic stem cell transplantation failure or engraftment delay. A number of studies have suggested a potential role for clinical use of GMCSF in addition to these approved indications. Most of these studies, however, have examined a relatively small number of patients. Therefore, the use of GM-CSF in clinical practice remains relatively limited, compared with the use of G-CSF.
M-CSF is a heavily glycosylated, disulfide-linked homodimer that was first cloned and produced as a recombinant protein in 1985 [112]. M-CSF, which is produced endogenously by monocytes/ macrophages, fibroblasts, and endothelial cells, acts specifically on cells of the monocyte/macrophage lineage to enhance cytotoxicity, superoxide production, phagocytosis, chemotaxis, and secondary cytokine production [113, 114]. PMNL function does not appear to be affected by M-CSF administered either in vitro or in vivo. In vitro, M-CSF has been shown to augment the antifungal activity of monocytes/macrophages against both conidia and hyphae of Aspergillus fumigatus, partly via enhancement of oxidation-dependent mechanisms [115].
M-CSF M-CSF has been reported to render human monocyte-derived macrophages more susceptible to HIV-1 infection in vitro [116]. M-CSF increased the frequency with which monocyte-derived macrophages became infected, the amount of HIV mRNA expressed per infected cell, and the level of proviralDNAexpressed per infected culture. Thus, M-CSF may function in an autocrine/ paracrine manner to sustain HIV replication, and possible therapeutic strategies for decreasing virus load after HIV infection could be the administration ofM-CSF monoclonal or polyclonal antibodies or soluble M-CSF receptors to HIV-1-infected patients [117].
M-CSFin animal models of infection. Expression of M-CSF was increased in the serum, brain, kidney, liver, and lung after infection of mice with a low-virulence strain of C. albicans [118]. M-CSF remained elevated during the 2-week period preceding the onset of specific T cell-dependent immunity. The number of monocytic precursor cells increased in the bone marrow of infected mice. Mice with systemic C. albicans infection also benefited by combination therapy with amphotericin B and M-CSF, which was superior to amphotericin B alone [119].
The therapeutic efficacy of recombinant human M-CSF alone and in combination with the antifungal agent fluconazole was evaluated in a large study of 344 rats with acute and chronic candidiasis [120]. Daily subcutaneous bolus injections of M-CSF for 10 days plus a single bolus dose of fluconazole improved the median survival time from 5 days (32% survival)with fluconazole alone to >30 days (88% survival) in the M-CSF/fluconazole-treated rats. In the chronic model, a 10-day course of M-CSF plus a single bolus dose of fluconazole decreased the median titer of C. albicans by 10-fold at 15 and 30 days after onset of infection.
Clinical trials with M-CSF. In contrast to G-CSF and GMCSF, clinical experience with M-CSF is rather limited. Because of its specificity for cells of the monocyte lineage, M-CSF has not been used for stem cell harvesting [121].
M-CSF has been used for treatment of infectious complications during neutropenia. In a phase 1 trial, 24 patients with invasive fungal infections after bone marrow transplantation receivedMCSF in addition to standard antifungal treatment [122]. Infection resolved in only 6 of the treated patients, and no M-CSF-mediated effects on PMNL, monocyte, or lymphocyte counts were observed. Long-termfollow-up has also been reported on a series of bone marrow transplant patients with severe fungal infections who received M-CSF as adjunctive therapy [123]. The overall survival of M-CSF-treated patients was reportedly greater than that of an historical control group (27% vs. 5%). Multivariate analysis of the patients who received M-CSF revealed 2 factors that correlated with poor survival—Karnofsky score >20% and invasive aspergillosis. To our knowledge, no prospective, randomized, controlled clinical trials ofM-CSF for adjunctive therapy of fungal infections have been published.
IFN-γ is produced endogenously by 3 major types of lymphocytes: CD4+ and CD8+ T cells and NKcells [4, 124]. Monocytes/ macrophages andPMNLare the major target cells of this cytokine. However, the spectrum of activity of IFN-γ extends beyond the conventional cellular components of immunity to nonprofessional host defense cells, including platelets, endothelial and epithelial cells, fibroblasts, hepatocytes, astrocytes, and microglia cells [124]. Of interest, granuloma formation in response to intracellular pathogens has been shown to be an IFN-γ-dependent process [124].
In vitro, IFN-γ has been reported to stimulate the ability of humanor animal macrophages to kill or inhibit a wide variety of bacteria, fungi, and protozoa [124–126]. IFN-g is recognized to play a critical regulatory role in macrophage-mediated killing of important intracellular pathogens, including Mycobacterium, Leishmania, Rickettsia,Legionella,and Chlamydia species [124–126].
IFN-γ in animal models of infection. The pivotal role played by IFN-γ in the host immune response has been demonstrated in mice with targeted genetic disruptions of either IFN-γ or the IFN-γ receptor. These knockout mice have impaired ability to control infections by intracellular pathogens, such as viruses, Mycobacterium tuberculosis, bacillus Calmette-Guérin, and L. monocytogenes [124, 125, 127, 128].
In vivo, beneficial effects have been reported for IFN-g treatment in a wide variety of experimental infections in animal models. For example, systemic administration of IFN-g enhanced resistance against acute disseminated C. albicans infection and invasive aspergillosis in mice [129, 130]. In corticosteroid-treated rats, local administration of intratracheal IFN-g significantly augmented pulmonary defenses against experimental legionellosis [131]. In mice infected with Cryptococcus neoformans, mice treated with IFN-γ, compared with untreated mice, had significantly increased survival and decreased lung colony-forming unit counts [132]. To date, administration of IFN-γ has been reported to be therapeutically beneficial in experimental models of infection by at least 22 bacterial, fungal, protozoan, and helminthic pathogens [125, 126].
Clinical use of IFN-γ. IFN-γ is approved by the FDA for the treatment of individuals with chronic granulomatous disease (CGD). On the basis of the in vitro observation, using CGD cells, that IFN-γ can stimulate the human monocyte oxidative burst- independent antimicrobial mechanisms and subsequent confirmation of enhanced CGD monocyte and PMNL activity after administration of IFN-γ, 128 patients with CGD were treated twice a week with IFN-γ or placebo. The IFN-γ-treated group developed significantly fewer infections (both total and serious) and required fewer days of hospitalization [133, 134]. Subsequent long-term follow-up of these patients who received IFN-γ (50 μg/m2 subcutaneously 3 times weekly) revealed no unexpected toxicity, developmental delay, or growth impairment [124, 135]
IFN-γ may also have therapeutic potential as an immunomodulatory agent for other specific immunodeficiency syndromes. The hyperimmunoglobulinemia E (hyper-IgE) syndrome, also called Job syndrome, is a rare, distinct primary immunodeficiency syndrome characterized by recurrent skin and pulmonary abscesses, pneumonia, eczema, eosinophilia, and highly elevated levels of serum IgE [136, 137]. The principal pathogen responsible for infections in affected individuals is S. aureus [136, 137]. Several investigators have reported that the in vitro chemotaxis response is depressed in PMNL of patients with hyper-IgE syndrome [138, 139]. Furthermore, lymphocytes of patients with hyper-IgE syndrome have an impaired response to IL-12, resulting in decreased IFN-γ production in response to bacterial antigens [140, 141].On the basis of these observations, a small, uncontrolled trial has been performed, suggesting that systemic therapy with recombinant human IFN-γ might improve the clinical course in patients with hyper-IgE syndrome [142, 143].
During the past 2 decades, treatment trials using both locally and systemically administered IFN-γ have been performed in patients with several types of infections. IFN-γ appears to be an effective adjunctive therapeutic agent when used in conjunction with conventional antimicrobial chemotherapy for patients with cutaneous and visceral leishmaniasis, disseminated atypical mycobacterial infection, or lepromatous leprosy [125, 144, 145]
Because IFN-γ activates alveolar macrophages, which represent important effector cells in host immunity against M. tuberculosis, aerosolized IFN-γ therapy has been used for treatment of a limited number of patients with multidrug-resistant tuberculosis. In an open-label study, 5 patients with smear-positive, pulmonary tuberculosis received aerosol IFN-γ 3 times per week for 1 month [146]. Sputum acid-fast bacillus smears converted from positive to negative in all 5 patients. Treatment with IFN-γ for 2 months led to a reduction in the size of cavitary lesions in each patient. Unfortunately, sputum smears reverted to positive in 4 of the patients within 1 month of discontinuation of IFN-γ therapy. However, these preliminary data suggest that aerosolized IFN-g is a promising, adjunctive treatment for patients with multidrug-resistant tuberculosis who are otherwise not responding well to therapy. A limited number of patients with antimicrobial refractory respiratory tract infections caused by atypical (nontuberculous) mycobacteria have also been reportedly treated successfully with both systemic and aerosolized IFN-γ [144, 147–149].
Table 1 provides an overview of cellular sources and biologic activities of G-CSF, GM-CSF, M-CSF, and IFN-γ. Table 2 summarizes approved clinical indications for each of these 4 cytokines.
Major cellular sources and biologic actions of granulocyte colony-stimulating factor (G-CSF), granulocyte-macrophage colony-stimulating factor (GM-CSF), macrophage colony-stimulating factor (M-CSF), and interferon (IFN)-γ.
G-CSF. The most commonly reported side effect of G-CSF is mild-to-moderate bone and/or musculoskeletal pain, which occurs in 20%–30% of individuals [150–160]. Aside from these effects, there are relatively few adverse effects, even after years of administration in patients with severe chronic neutropenia [161]. Other adverse effects include headache, anemia, splenomegaly, thrombocytopenia, and, rarely, allergic and injectionsite reactions [150–160]. Uncommon adverse events include cutaneous neutrophilic vasculitis (Sweet's syndrome), G-CSF-induced lung toxicity after cancer chemotherapy, and splenic enlargement in healthy donors during G-CSF-induced mobilization of allogeneic PBSC [162–165]. Two episodes of splenic rupture have been reported in apparently healthy individuals after G-CSF-based regimens for harvest of PBSC [166, 167]. Both individuals received serial administration of G-CSF at a dose of 5–16 μg/kg of body weight on a daily basis for 3–5 consecutive days [166, 167]. Adverse effects related to G-CSF administration generally resolve rapidly after discontinuation of treatment. However, it is clear that the long-term safety of G-CSF administration to healthy donors for mobilization of PBSC or granulocytes requires careful monitoring [168–171].
Individuals with severe congenital neutropenia are at increased risk for development of myelodysplastic syndrome (MDS) and progression to acute myeloid leukemia (AML). Recent evidence indicates that mutations in the gene for the G-CSF receptor are associated with this disease progression [172]. Of 352 patients with severe congenital neutropenia receiving long-term G-CSF therapy while enrolled in the Severe Chronic Neutropenia International Registry (SCNIR; Seattle), 31 developed transformation to MDS/AML. In contrast, none of the 344 patients with idiopathic or cyclic neutropenia followed by the SCNIR developed MDS/ AML while receiving long-term G-CSF therapy [173]. Thus, progression to MDS/AML appears to be intrinsic and unique to the severe congenital neutropenia syndrome. Currently, G-CSF therapy is not believed to be the cause of malignant transformation in severe congenital neutropenia. However, G-CSF may accelerate manifestation of the MDS/AML clinical phenotype in affected individuals. Careful monitoring for MDS/AML, including periodic cytogenetic analysis, is advised for patients with severe congenital neutropenia receiving long-term G-CSF therapy.
GM-CSF. Compared with G-CSF, GM-CSF treatment has been associated with a greater frequency of side effects, possibly as a result of stimulation of proinflammatory responses in monocytes/ macrophages. The most frequently reported side effect is fever, which occurs in >20% of patients and can serve as a confounding factor in the evaluation of treatment responses during infection [174]. Fever is often accompanied by myalgias and an influenza-like syndrome. First-dose reactions, consisting of flushing, tachycardia, hypotension, musculoskeletal pain, dyspnea, nausea, vomiting, and arterial oxygen desaturation, have been reported in 5% of recipients [175]. High doses ofGM-CSF have been reported to cause a generalized capillary leak syndrome [176].
M-CSF. Most phase 1/2 trials using M-CSF reported no severe side effects; however, a transient, dose-related thrombocytopenia, requiring dose reduction in some cases, has been observed by most investigators [122, 123]. The etiology of thrombocytopenia is unknown, although experimental models suggest that enhanced function of splenic phagocytesmay relate to the thrombocytopenia [177].Although it can promote proinflammatory cytokine production by monocytes and macrophages, M-CSF does not appear to strongly induce or exacerbate graft-versus-host disease when used in patients after allogeneic hematopoietic stem cell transplantation [122].
IFN-γ. Adverse reactions to systemic IFN-γ include fever, chills, fatigue,myalgias, and headache [124, 125]. Although IFN-γ-associated influenza-like symptoms are typically mild, decrease over time, and can usually be managed with prophylactic antipyretics, these side effects can be significantly reduced when IFN-γ is administered via inhalation for the treatment of pulmonary infections [146]. Isolated laboratory evidence of toxicity, including reversible neutropenia, may occur. However, data from large clinical trials indicate that clinically significant hematologic abnormalities are infrequent even when patients are treated for years [125]. Although IFN-γ treatment has been reported to produce beneficial effects in patients with rheumatoid arthritis and systemic sclerosis, it can exacerbate multiple sclerosis [125]. Therefore, IFN-γ should be used cautiously in patients with inflammatory diseases.
Table 3 provides an overview of common side effects associated with systemic administration of G-CSF, GM-CSF, M-CSF, and IFN-γ.
G-CSF, GM-CSF, M-CSF, and IFN-γ play critical roles in the host defense response during infection. In vitro, all 4 cytokines have been shown to up-regulate functional activity in monocytes/ macrophages and/or PMNL. Studies conducted in animal models have demonstrated potential utility for each of these cytokines in the treatment of infections caused by a variety of bacterial, fungal, and parasitic pathogens. Unfortunately, positive treatment responses for these cytokines in preclinical animal models often have failed to be replicated in clinical trials. The reason for this limited success in clinical trials is most likely multifactorial. Therefore, therapeutic potential must be substantiated in controlled clinical trials, to define the proper role of G-CSF, GM-CSF, M-CSF, and IFN-γ therapy in the clinical management of specific infections.
Financial support: Deutsche Krebshilfe (to K.H.); National Institutes of Health (HL-62995 to W.C.L.).
IDSA Members: For your free access to this journal, log in via the IDSA members area.
Open access options for authors visit Oxford Open
This journal enables compliance with the NIH Public Access Policy