AS Lau, D Lehman, J Liu

Abstract

Cytokines are soluble signaling proteins responsible for cell-to-cell communication and play critical roles in many biological processes including growth and development, immunity, and hematopoiesis. In response to pathogen invasion, these pleiotropic molecules are produced by immune cells and other cells including fibroblasts and endothelia. Interaction among cytokines results in propagation of inflammation and host defense against pathogens. The biological functions of prototype cytokines including interferons, tumor necrosis factor-alpha, interleukins, and granulocyte colony stimulating factors are reviewed. With special references to children, therapeutic uses of interferon-α in viral diseases and interferon-γ in chronic granulomatous disease are discussed. The clinical applications of granulocyte/macrophage colony stimulating factors in patients with neutropenia due to cytotoxic chemotherapy, bone marrow transplantation, and congenital agranulocytosis are presented. We also examine the role of tumor necrosis factor-alpha and interleukins in the pathogenesis of septic shock and discuss the various possible strategies to suppress the undesirable effects of over-expression of proinflammatory cytokines. In summary, advances in biotechnology enable the use of interferons and colony stimulating factors to provide novel approaches to treat children and adults with viral infections, cancer, or autoimmune diseases. In addition, manipulating the cytokine systems by inhibiting the activity of proinflammatory cytokines may prove useful in the treatment of patients with sepsis.

Introduction

Host defense mechanisms are activated, following infection with pathogens, with the release of signaling molecules, cytokines, to mobilize different branches of the immune system. Cytokines comprise a rapidly growing collection of potent, pleiotropic proteins that act as local and/or systemic intercellular regulatory factors. They play crucial roles in many biological processes including microbial infections, inflammation, immunity, and hematopoiesis. They are produced by cells including macrophages/monocytes, lymphocytes, fibroblasts, and endothelial cells. To date, a large number of cytokines including interferons (IFN), interleukins (IL), tumor necrosis factors (TNF), growth factors (e.g. epidermal growth factors), and differentiating factors (e.g. colony stimulating factors, CSF) have been identified.^1-7 It has been demonstrated that specific growth factors are involved in embryo development, and growth and differentiation factors (e.g. granulocyte/macrophage-CSF and granulocyte-CSF) are required for maintenance of hemostasis and hematopoiesis in adults. Other cytokines, including IFN, TNF-α, IL-1 and IL-6, are involved in immunological responses to microbial infection or injury.^5-7 Table I is a partial list of cytokines that have been characterized. Here, the roles of IFN-α, IFN-γ, TNF-α, IL-1, IL-6, GM-CSF and G-CSF in the pathogenesis of infectious diseases and their potential clinical applications will be discussed.

While important in immune defense, cytokines have been described as double-edged swords in immune responses. It is well known that over-production of cytokines can cause tissue injury leading to systemic pathology. For instance, IL-1 is known to be a key mediator of the inflammation and tissue damage found in rheumatoid arthritis.⁶ This has been substantiated by previous reports that administration of recombinant IL-1 into rabbit knee joints would reproduce many of the pathological features characteristic of rheumatoid arthritis. In addition, it is well documented that TNF-α, IL-1, and IL-6 are the major mediators of acute phase response in sepsis.^5-7 Thus, inhibition of the synthesis of specific cytokines or blockade of their activity may provide therapeutic approaches to modify the progression of inflammation in autoimmune or infectious diseases. This has led to many clinical trials targeted at mitigating the effects of inflammatory cytokines including TNF-α and IL-1. Recent advances in recombinant DNA technology have provided the opportunity for cloning and production of large quantities of purified recombinant cytokine products and their antagonists for therapeutic trials. Clinically, trials of these cytokines and growth factors in infections and cancer have reached the stage of licensing for specific uses in human diseases. In the pediatric setting, therapeutic application of cytokines and IFN is still in its infancy. Ongoing research would address many of the issues specific to the use of cytokines and immunomodulators in children. Where data are available on specific disease entities in children, special emphasis will be placed and the literature will be analyzed. In most cases, the following is a general discussion that will cover disease spectrum in patients of all ages.

Interferons

In response to viral infection, IFNs are produced by most, if not all, vertebrates (reviewed in references^8-10). These glycoproteins possess antitumor and immunomodulatory activities in addition to their antiviral effects. IFNs can be classified into three major groups, namely, alpha, beta, and gamma (Table II). While IFN-α and -β are similar in protein structure and bind to the same receptor, IFN-γ is quite distinct and has its own receptor system. The IFN-α family comprises at least 14 subtypes inducible by viruses, bacteria, and tumor cells. They are produced primarily by macrophages/monocytes, dendritic cells and B lymphocytes. Human IFN-β is induced by viral and other foreign nucleic acids in fibroblasts, macrophages, and epithelial cells. Along with other lymphokines, IFN-γ is made primarily by T-cells in response to foreign antigens and mitogens. IFN-γ appears to be mainly an immunomodulator instead of being an antiviral protein. Compared to other IFNs, IFN-γ is more effective in inhibiting intracellular microorganisms other than viruses, e.g., some rickettsia, bacteria (Listeria), and protozoa.

Biological activities of interferon

To elicit its antiviral and biological effects, IFN binds to its cognate receptor to induce transmembrane signaling including the activation of protein kinases including tyrosine kinase 2 (tyk 2) and Janus kinase (Jak), and possibly phospholipase A2 (reviewed in references^11,12). These events result in activation of specific STAT factors (signal transducers/activators of transcriptions) to form protein complexes,¹¹ which are consequently translocated into the nucleus to initiate transcription of specific IFN-stimulated genes. Following translation, the IFN-stimulated gene products mediate antiviral and other biological activities of IFNs.

The antiviral actions of IFN are thought to be mediated by at least two pathways^8,9,12: (i) degradation of RNA via the activation of a latent, specific ribonuclease, and (ii) inhibition of protein synthesis due to activation of an IFN-regulated kinase (double-stranded RNA-dependent kinase, PKR) which renders translation initiation factor 2α inactive (Fig. 1). The first pathway involves increased activities of 2'-5'-linked oligoadenylate (2-5A) synthetase. This enzyme generates 2-5A which, in turn, reversibly activate another enzyme, a latent ribonuclease. This 2-5A dependent ribonuclease can cleave viral and cellular mRNAs, resulting in cessation of viral and cellular mRNA synthesis. In the latter pathway, PKR is active only in the presence of double-stranded RNA, which is a common stage utilized by viruses during its replication cycle. The activated kinase inactivates the translation initiation factor 2α, resulting in inhibition of translation including viral as well as host cellular proteins. Besides these enzymes, other polypeptides have recently been found to be induced by IFN-γ. The IFN-γ regulated gene products may be involved in additional pathways regulating cellular metabolism.

Fig. 1 Activation of antiviral pathways in cells treated with interferons. Refer to text for description.

With respect to immune effects, IFNs regulate the expression of immunoglobulin, MHC antigens, and cytokines (reviewed in reference⁹). In addition, IFN-γ enhances the differentiation and function of macrophages. IFNs may play a role in inducing production of certain growth factors such as IL-2 and other cellular products (e.g. cytokines) by lymphocytes and macrophages. These changes all affect cell recognition and immune cell functions. The antitumor effects of IFNs may be exerted through several mechanisms as follows: i) direct antiproliferative effects on tumor cells, ii) autocrine growth inhibitor, iii) activation of cytotoxic effectors of the host immune system including natural killer cells, activated macrophages, and sensitized T-lymphocytes, iv) enhanced production of cytokines including TNF-α, and v) down-regulation of oncogenes. The induction of cytokines results in a cascade of cellular processes that may further perpetuate the antitumor activity of the host immune system.

Interferon and cytokines in viral infections

In view of the broad spectrum antiviral effects of IFNs on viral replication observed in vitro and in animal models, there has been enormous interest in using IFNs therapeutically for the last three decades (see next section). However, it has long been recognized that while IFNs have significant antiviral activity against many viruses, not all viruses are equally susceptible to IFN-α in vivo. Certain viruses can evade the actions of IFNs, some being more successful than others, by inhibiting the activity of IFN regulated enzymes or down regulating the expression of IFN genes. DNA-containing viruses are in general less sensitive to IFN-α than RNA viruses. This may be in part due to the fact that little, if any, double-stranded RNA is synthesized during DNA virus infection. In addition, some DNA viruses have evolved specific strategies for evading the action of IFNs (reviewed in reference⁸). For instance, herpes simplex virus induces the synthesis of 2-5A analogs that would compete with the natural 2-5A for binding with the 2-SA dependent ribonuclease, thereby rendering the nuclease inactive against viral RNA. In contrast, RNA viruses inhibit the 2-5A system by a different mechanism which appears to be independent of the 2-5A analog competitors. With regard to the PKR kinase system, it is known that adenovirus type 5 encodes small RNA molecules called VA1 which block autophosphorylation of PKR. In contrast, hepatitis B virus (HBV) utilizes a different mechanism: HBV core antigen acts as a viral trans-acting factor that can suppress transcription of the IFN-β gene.¹³

An additional important example is the effect of HIV on the IFN antiviral system. We have recently reviewed a number of possible mechanisms for HIV to evade the IFN-α system and to inhibit the synthesis of IFN in HIV infected cells, thereby contributing to the pathogenesis of AIDS.¹⁴ To discuss a few examples, we have previously demonstrated a reduction in IFN-α receptor expression on primary blood monocytes and lymphocytes during the progression of HIV infection from an asymptomatic state to fulminant AIDS.¹⁵ In addition, it has been shown that HIV tat protein is capable of blocking the activation of 2-5A synthetase^16,17 and down-regulating PKR activity during active HIV replication.¹⁸ Similarly, it has been reported that the tat-responsive element in the 5' untranslated region of HIV mRNA (termed TAR region), because of its double-stranded nature, is capable of binding and inhibiting PKR activation in vitro and in vivo.^16,19 Recently, we have an original observation which provided direct evidence that PKR activity is absolutely required in the signaling pathway controlling transcription of the IFN genes.²⁰ Taken together, these results indicate that HIV can accomplish the goal of suppressing the intrinsic cellular antiviral system and IFN induction by inhibiting PKR activity. Although IFN-α has been reported to induce the regression of HIV viremia²¹ and Kaposi's,^22,23 the findings discussed above suggest that IFN-α may not be useful in late stages of AIDS when the HIV-1 infection is fully established.

In addition to evading the IFN system, HIV appears to take advantage of immune response in enhancing its replication. For example, it has been shown that proinflammatory cytokines including TNF-α induce the replication of HIV-1.^24,25 The molecular mechanisms underlying this process have been partially elucidated. This is due to the fact that HIV incorporates a binding site for transcription factor NF-kB in its HIV promoter.²⁵ In the human cellular system, this transcription factor complex is one of master switches for controlling immune cell activation as well as synthesis of immunoglobulin, MHC antigens and acute phase proteins. Since TNF-α is produced in response to bacterial sepsis, it is possible that frequent bacterial infections, commonly found in pediatric AIDS patients, provide a milieu for further activation of HIV and henceforth rapid progression of AIDS in children. Thus uncontrolled or chronic induction of TNF-α may play a role in the pathogenesis of HIV infection.

Clinical applications of interferons

By 1996, the Food and Drug Administration has approved IFNs for a number of clinical indications in the United States (Table III). IFN-α is approved for treatment of viral infections including hepatitis B and C virus and human papilloma virus (condyloma acuminata), and malignancies including hairy cell leukemia, chronic myelogenous leukemia, malignant melanoma as well as Kaposi's sarcoma in HIV-infected patients. IFN-γ is approved for prophylactic use in patients with chronic granulomatous disease to prevent recurrence of bacterial infections. IFN-β was approved in the United States two years ago for the treatment of patients with multiple sclerosis. In addition, IFNs have demonstrable efficacy in laryngeal papillomatosis, early stages of HIV infection, multiple myeloma, and basal cell and cutaneous squamous cell carcinoma. In some European countries, IFNs are approved for a number of these diseases. While there has been significant work done on children, most of the cytokine clinical trials to date are on adults. Throughout this review, in areas where there are data available, the use of cytokines in children will be emphasized. Otherwise, results on adult patients will be discussed.

Recombinant and natural IFN-α (derived from leukocyte cell lines) as well as recombinant IFN-γ are available for clinical use. Dosages of IFNs range from one million units per treatment in hairy cell leukemia to more than 30 million units per dose in Kaposi's sarcoma.²⁶ The duration of therapy is also highly variable, ranging from a few weeks for condyloma acuminata to more than 6 months for hepatitis C. The adverse effects of IFNs are dose-dependent and found mostly in patients receiving more than a few million units of IFNs. The most common side effect consists of a flu-like syndrome characterized by fever, chills, fatigue, headache, myalgia, and malaise.²⁶ These undesirable side effects can be ameliorated by administering acetaminophen. The more serious side effects including relative leukopenia, thrombocytopenia, and hepatotoxicity are observed in patients receiving more than twenty million units (approximately 0.1 mg) of IFN-α or -γ per dose. In general, even these changes are transient and are rapidly reversible following discontinuation of IFN therapy. To date, IFNs have been used as monotherapy for viral and neoplastic diseases. However, laboratory and clinical investigations indicate that IFNs can be used in combination with conventional treatment modalities including surgery, chemotherapy, or radiotherapy to provide additive or synergistic therapeutic effects. Thus, IFNs serve as prototypes of cytokines to be used as a therapeutic agent in immunomodulation of the immune system.

Bacterial versus mammalian cell-produced cytokines

Clinical trials of IFNs were aided by recent advances in biotechnology which employ genetic engineering as well as modem pharmaceutical techniques to produce mass quantities of purified IFNs for therapeutic uses. Prior to the advent of biotechnology, a single culture of about one million human cells was needed to produce only one picogram of human IFN-α following painstaking purification processes. Consequently, billions of cells were required to obtain one microgram of IFN-α (0.2 million units) for clinical studies. At present, there are three major approaches to large scale production of IFNs for therapeutic uses: i) recombinant IFNs produced in bacterial cells (e.g., IFN-α2 or IFN-β in E. coli), ii) recombinant IFN-β produced in mammalian cells (e.g., Chinese hamster ovarian cells); or iii) natural IFNs from leukocytic cells freshly isolated from human blood or from lymphoblastoid cell lines following stimulation with viruses. All three types of IFNs are now being produced in large amounts for therapeutic uses.

In the first two groups, both recombinant IFN-α2 and IFN-β are cloned products; and each is homogeneously made up of only one IFN subtype. Due to the sources for IFN production, i.e., bacteria versus mammalian cells, there are differences in the biochemical and immunological properties of the IFN preparations. In the first group, recombinant IFN-α2 or IFN-β is cheaper and more efficient to make in E. coli which expresses the cloned IFN gene. However, these recombinant proteins from E. coli are not glycosylated because bacteria do not have the proper machinery for glycosylation. In contrast, in the second group, recombinant IFN-β produced in mammalian cells has the proper structure and glycosylation, but they are more expensive to produce. To add another level of complexity, there is the third group of natural leukocytic IFNs which consist of the entire complement of IFN-α subtypes and IFN-β; and they are produced in human leukocytic or lymphoblastoid cells after stimulation with viruses as inducers. Because they are produced in human cells, their composition and glycosylation are closest or identical to the IFNs induced in vivo in humans. However, they are expensive, time-consuming and cumbersome to produce. In this case of IFNs from fresh leukocytes, extensive purification processes are used to minimize the potential risk of human pathogen contamination from a large pool of blood donors. As for IFNs from lymphoblastoid cells, there have been concerns regarding potential contamination with animal virues.^28a These products require constant monitoring to minimize the risks. For better patient care, further research to develop an economical way to produce leukocytic IFNs without the risk of human pathogen contamination or without the use of live viruses as inducers is warranted.

For short term or low dose therapy, the differences in glycosylation may not be discernible to cause significant side effects. On the contrary, in long term or high dose therapy, the lack of glycosylation in bacterially produced cytokines could be significant from a clinical standpoint. Indeed, previous studies have shown much higher rates of rejection for the bacterially produced proteins, particularly after long term or high dose usage. For instance, two previous reports demonstrated that the incidence of rejection as reflected by antibody formation can be as high as 20 to 38% for bacterially produced IFN-α compared with only 1.2% for leukocytic IFN-α.^27,28 Similar findings were reported in a recent study with regard to absence of anti-IFN antibodies development during leukocytic IFN therapy.^28b In addition, IFN-binding or -neutralizing antibodies were documented in hairy cell leukemia patients treated with recombinant IFN-α2a.²⁹ In this report, patients who developed anti-IFN antibodies following recombinant IFN-α2a therapy can be successfully treated with a subsequent course of leukocytic IFN. Similar results were obtained in patients with idiopathic mixed cryoglobulinemia or polycythemia vera, who developed IFN-antibodies and -resistance after prolonged recombinant IFN-α treatment. They can be successfully re-induced into remission by leukocytic IFN-α.^30,31 Most recently, a preliminary report on a multicenter, randomized open study demonstrated that leukocytic IFN-α induced normalization of liver functions in some patients with hepatitis C virus (HCV) infection who previously failed to respond to recombinant IFN-α therapy.^31a

This issue of improper structure and glycosylation appears to be a more common problem for the clinically useful cytokines. A further example is the use of GMCSF in immunocompromised patients or bone marrow transplant recipients (see Table IV). In a review of 32 clinical studies, the median incidence of side effects for bacterially manufactured GM-CSF is almost twice that of the ones produced in eukaryotic cells including mammalian and yeast cells.³² Increased incidence of side effects may effectively compromise the long term use of any cytokines. Currently, most clinically approved cytokines, including IFN-α2, IFN-β, GM-CSF and G-CSF, are produced in E. coli. To date, there are no sufficient data to absolutely prove that lack of glycosylation and proper structures similar to the natural products would give major disadvantages in using recombinant proteins clinically. Further studies will be needed to unequivocally determine whether the bacterially produced cytokines are more prone to be associated with IFN resistance, side effects, and treatment failure. While it is reasonable to continue the use of these recombinant bacterially produced cytokines with demonstrable efficacy, it would be most desirable to use mammalian cell derived proteins in future trials of new cytokines for clinical use.

Antiproliferative activities of IFNs

IFNs in general have greater therapeutic effectiveness for hematological malignancy than for solid tumors. In chronic leukemias, myeloid and hairy cell, the use of IFN-α has resulted in a therapeutic response in more than 70% of patients with a reduction in leukemic cell mass and an improvement in the patients' quality of life.^33,34 It is very interesting to note that IFN-induced therapeutic responses in chronic myelogenous leukemia patients are associated with a gradual reduction in the frequency of cells with 9-22 chromosomal translocation defect, which is the hallmark of the leukemia. With regard to solid tumors, IFN-α elicits a response rate of 15% in metastatic melanoma, a level comparable to results with cytotoxic chemotherapeutic agents. There are clinical trials for other malignancies including multiple myeloma and B- and T-cell lymphoma as well as renal carcinoma.^9,35 In children, an IFN-α trial on patients with chemotherapy-refractory T-cell lymphoblastic leukemia showed some initially encouraging but inconclusive results.³⁶ In HIV-related Kaposi's sarcoma, the overall response rate for recombinant IFN-α2 is around 40%.^22,23 Further studies on combination therapy with IFN-α2 and antiviral agents, e.g., zidovudine or dideoxycytosine, may further improve the response rates in Kaposi's sarcoma.

Hemangiomas are benign tumors of the blood vessels occurring in 1% of newborns. After birth, some hemangiomas may continue to grow progressively to massive lesions, causing severe illness associated with bleeding or obstruction of vital organs. In view of the antiproliferative effects of IFNs, the effects of IFN on angiogenesis were examined. It has been shown that IFN inhibits the motility of capillary endothelial cells and formation of new capillary blood vessels.³⁷ This has led to several clinical trials of using IFN-α for the treatment of hemangiomas with remarkable results.^38,39 In the latest study, it has been reported that 18 out of 20 children showed a dramatic response to daily subcutaneous injections of IFN-α.³⁹

Antiviral activities of IFNs

With respect to human papilloma virus, IFN-α has been used for condyloma acuminata, plantar warts, and juvenile laryngeal papillomatosis. In patients with refractory condyloma acuminata, intralesional administration (one million units per lesion three times weekly for three weeks) appears to be most effective with complete response in 36% to 62% of patients, while topical and parenteral routes have also been used.^40-42 However, the clinical efficacy of IFN-α on condyloma acuminata is significantly less in AIDS patients. Considering the cost, IFN-α should be reserved for patients with recalcitrant disease not responding to topical podophyllin or laser therapy. Plantar warts seem to be less responsive than are genital warts.

Following systemic IFN-α therapy of juvenile laryngeal papillomatosis, most of the children show clinical responses in terms of decrease in the number of lesions. However, long term response to parenteral IFN-α therapy is quite variable, probably due to persistence of the papilloma viral DNA in the lesions.⁴³ For instance, one previous study, using natural leukocyte derived IFN-α (two million units per m² three times per week for 12 months), showed significant reductions in papilloma growth rates during the first six months. Initially, this resulted in obviating the need for surgical ablation of the benign tumor. Nevertheless, no long term benefits was observed when comparing parenteral IFN-α therapy with laser therapy alone.⁴⁴ In addition, epidermodysplasia verruciformis, a rare human papilloma virus-related condition, may respond to intralesional IFN-α therapy.

IFN-α and hepatitis virus infections

HBV infects a significant percentage of the world population, in fact at epidemic proportions in some areas of the world. Up to 200 million people worldwide have chronic hepatitis B and they are carriers of HBV surface antigens. The carrier state is often associated with chronic liver dysfunction and long term risk of development of hepatocellular carcinoma.⁴⁵ For these patients with chronic HBV infection, approximately one-third respond to IFN-α (dosage, five or ten million units daily) therapy for four to six months with significant clinical improvement.^46-48 Elevated liver function is normalized in nearly all patients with a sustained loss of viral replication markers including HBV e antigen and DNA polymerase from the patients' serum. Recurrence is not common once a response to IFN is established. In addition, in the initial non-responders, at least another 15% showed delayed response, with disappearance of HBV surface antigens, months or a year after cessation of IFN therapy leading to termination of the carrier state. It therefore has been suggested that IFN-α should be considered for those patients with chronic hepatitis as evidenced by the presence of hepatitis B e antigen, serum aminotransferase levels more than double the upper limit of the normal range, and low to moderate levels of HBV DNA in the serum. In 1992, the U.S. Food and Drug Administration approved IFN-α2b for the treatment of chronic hepatitis B.

In children, the response to IFN therapy depends on the study population, the mode of infection, and the histological feature of the disease presentation. These variations in patient selection and demographics have resulted in conflicting results. For instance, the incidence of HBV e antigen disappearance in sera from patients on IFN therapy have been as low as 18% (same as controls without IFN treatment) in Chinese children to as high as 46% (with controls at 17%) in European children.^49-51 In these studies, the differences in response may have been related to the mode of infection, vertical transmission in Chinese versus horizontal in European children. IFN-α appeared to be well tolerated by children in dosage range of less than 10 million units per m² as used in these studies. Similar to adults, clinical response, including normalization of liver function and improvement in liver histology, was associated with the clearance of HBV e antigen and HBV DNA. To boost response to IFN therapy, corticosteroid was used to prime patients prior to therapeutic trials. The results from two studies, one on Chinese and the other European children, again showed discrepancy with more definitely more remarkable response in the latter group.^52,53 Nevertheless, steroid priming appeared to be useful in both groups. At this time, the advantages of steroid priming before IFN therapy remain to be established and should not used as standard therapy.

HCV infection represents the common type of transfusion-related hepatitis, prevalent in chronically debilitated patients with renal disease or malignancy. With conventional screening technique, it is estimated that 5% of transfusions may involve exposure to HCV.⁵⁴ Half of these acute hepatitis C patients may develop chronic hepatitis, with 20% of whom develop cirrhosis. In chronic hepatitis C, 40 to 50% of patients may respond to IFN-α with resolution of abnormal liver function tests.⁵⁵ In this study, the dosage of IFN-α was one or three million units per dose, three times per week for 24 weeks. The response to IFN occurred rapidly, with about 85% of the responses documented within the first three months of therapy. The clinical response is associated with improvement in liver function as well as histological changes on liver biopsy. However, a significant portion of the responders will relapse following cessation of IFN-α therapy. For those who relapse, re-introduction of IFN-α therapy is necessary. Overall, up to 20% of patients may achieve a sustained remission following initial IFN-α therapy. A recent study attempted to address this issue of incomplete response or relapse.⁵⁶ They showed that one third of these patients achieved complete response with a significant decrease in HCV RNA levels. However, it appears that the criteria for selection of patient subgroups may affect the outcome of repeat IFN therapy in these studies.

Therapeutic uses of IFN-γ

Chronic granulomatous disease is a rare genetic syndrome characterized by recurrent life-threatening infections including hepatic and pulmonary abscesses due to Staphylococcus aureus and Aspergillus.^57,58 A large-scale placebo-controlled trial was conducted in 128 evaluable pediatric patients with this disease by subcutaneous injection of IFN-γ1b (5.0 ug/m² body surface area), three times weekly for up to 12 months.⁵⁹ This study showed that long term IFN-γ therapy produces a two-third reduction in the incidence of serious infections requiring hospitalization. While all patients' clinical course improved regardless of age, sex, use of prophylactic antibiotics, or genetic patterns of inheritance, the greatest therapeutic benefit was found in patients less than 10 years of age. The drug was well tolerated with common side effects (e.g. fever, headache, chills, injection site erythema) usually being mild and can be alleviated by symptomatic treatment.⁵⁹ Thus, IFN-γ has been approved as a prophylactic agent to reduce the risk of life threatening infections in patients with chronic granulomatous disease.

Cytokines in Sepsis and Their Clinical Implications

Biology of proinflammatory cytokines

Many of the features of septic shock result from bacterial endotoxin triggering a complex cascade of cytokine synthesis including TNF-α, IL-1, and IL-6. TNF-α and TNF-β are cytotoxic polypeptides released from activated monocytes/macrophages and lymphocytes, respectively, in response to inflammation, bacterial or viral infection.^5,60,61 At the cellular level, TNF-α causes degranulation, production of superoxide radicals, induction of procoagulant activity and suppression of thrombomodulant resulting in thrombosis. While crucial in combating infection, high levels of TNF-α in vivo can be detrimental be cause they induce severe metabolic acidosis, hypotension, and hemorrhagic necrosis of vital organs,^60,61 thereby mediating the pathophysiological process of septic shock. Recent reports indicate that high serum levels of TNF-α correlate with the morbidity and mortality of patients with meningococcemia and fulminant hepatic failure.^62,63 In experimental models of meningitis, it has been shown that TNF-α is a mediator of inflammation and its presence in the cerebrospinal fluid appears to be predictive of neurological damage.^64,65 Similarly, we and others have also shown that TNF-α synthesis is enhanced in patients with Kawasaki disease.⁶⁶ Therefore, deregulation of its synthesis in vivo may contribute to the pathogenesis of inflammatory diseases.

IL-6 is a pleiotropic cytokine produced by various cell types for the regulation of immune responses including acute phase reactions, and hematopoiesis.⁷ Some examples of IL-6 functions include: i) differentiation of B-cells into antibody producing plasma cells, ii) induction of C-reactive protein, complement C3, and fibrinogen, iii) differentiation of T cells, and iv) regulation of mesangial cell growth.^7,67 The abnormal production of IL-6 has been reported in disease states including autoimmune diseases, plasma cell neoplasia, glomerulonephritis, sepsis and meningitis. Moreover, IL-6 has been implicated in the pathogenesis of HIV infection and the development of B-cell lymphoma and Kaposi's sarcoma found in AIDS patients.⁶⁸

Similar to IL-6 and TNF-α, IL-1 is one of the prototypes of pro-inflammatory cytokines produced by immune cells and fibroblasts. The major immunological properties of IL-1 include activation of T- and B-cells, enhancement of natural killer cell activity, stimulation of endothelial cells, and induction of cytokine gene expression such as G-CSF and GM-CSF. In experimental models, it is interesting to note that IL-1 induces non-specific resistance to infection and injury if given at low dose at least 24h prior to bacterial infection, lethal radiation, or malaria infection.^69,70 However, it has been shown that massive doses of IL-1 administration (>1 mg/kg) into animals results in hypotension, depressed myocardial function, pulmonary edema, renal toxicity, and bone resorption.⁷⁰

Clinical implications for interleukins and TNF

For decades, endotoxin is thought to be the bacterial product responsible for the mediation of septic shock syndrome. It has recently become clear that many of the clinical features of septic shock result from a complex cascade of endogenous cytokine synthesis and intertwined interactions, triggered by endotoxin.⁶¹ In studying serum samples from patients with meningococcal disease or septicemia, detectable levels of TNF-α have been found in 10 out of the 11 patients who died, compared to the presence of TNF-α in sera from only 8 of the 68 survivors.^71,72 In these studies, all patients with serum TNF-α levels over 0.1 ng/ml succumbed. Another report on children with septicemia and purpura fulminans also demonstrated a correlation between the morbidity and mortality with high serum levels of TNF-α, IL-1 and IFN-γ.⁶² Many other subsequent reports confirmed these findings. It is interesting to note that while TNF-α may appear transiently and repetitively during burn injury, higher levels of TNF-α were detected in burn patients with sepsis.⁷³ Elevated levels of TNF-α have also been reported in patients with parasitic infections including Plasmodium falciparum and Leshmania.^60,61,74

Similar to TNF-α, both IL-1 and IL-6 have therefore been implicated in the pathogenesis of bacterial meningitis and sepsis. Both cytokines have been found in the serum, co-existing with TNF-α in the systemic circulation during meningococcal shock.⁷¹ In addition, elevated levels of IL-6 have been found in the cerebrospinal fluid of pediatric patients with bacterial meningitis.⁷⁵ Similarly, higher plasma levels of cytokines (including TNF-α, IL-1, and IL-6) were found to be correlated with mortality rate in children with both Gram-positive and Gram-negative sepsis.^62,76

Abnormal levels of cytokines in the body fluids may be an useful indicator of microbial infections. While TNF-α, IL-1, and IL-6 are present in the cerebrospinal fluid of patients with bacterial meningitis,^64,65,75 the predominant cytokines present in the cerebrospinal fluid of patients with viral meningitis are IFN-α and IL-6. For example, it was shown that IL-6 is present in the cerebrospinal fluid of 89% of patients with aseptic meningitis. In addition, the cerebrospinal fluid IL-6 level of these patients correlates with the degree of acute inflammation as measured by cerebrospinal fluid polymorphonuclear cell counts.⁷⁷ In contrast, IFN-α is usually not found in patients with bacterial meningitis. Thus, measurement of specific cytokines, namely IFN-α and TNF-α, may be useful in differentiating bacterial from viral etiology in patients with clinical presentations of meningitis. Further studies are needed to evaluate the prognostic value of the levels of inflammatory cytokines in patients with bacterial meningitis or sepsis.

Modulation of Cytokine Systems in Therapy

Many attempts have been made to alter the cytokine systems in order to ameliorate the severity of immunological injury and hemodynamic changes in sepsis. While most of these are at early experimental stages (Fig. 2), a number of possible strategies for manipulating the cytokine systems include: (i) Inhibition of cytokine gene transcription by cyclosporin A, FK506, or corticosteroids. (ii) Blocking antibodies: the use of anti-endotoxin antibody to inhibit the binding of endotoxin to its target cells. Some preliminary clinical trials suggested that human antibody to endotoxin may be useful in the treatment of Gram-negative septic shock with reduction in mortality. (iii) Inhibition of mRNA translation: Tenidap or SK&F86002 inhibits IL-1 synthesis. (iv) Inhibition of cytokine release: compound IX207887 or pentamidine prevents the release of IL-1 from activated monocytes without affecting synthesis. (v) Soluble cytokines as targets for neutralization by anti-cytokine antibodies or soluble receptors, blockade of cytokine binding to its specific cognate receptor by receptor antagonists, or inhibition of cytokine action on susceptible cells. (vi) Blockade of intracellular actions by interfering with transmembrane signaling (reviewed in reference³).

Fig. 2 Potential strategies for modulating the cytokine systems: (i) blocking antibodies for endotoxin; (ii) inhibition of cytokine gene transcription; (iii) suppression of protein synthesis; (iv) inhibition of cytokine release; (v) soluble cytokines as targets for neutralization; (vi) blockade of intracellular action. Ab, antibodies.

Corticosteroid in sepsis and meningitis

Because of its anti-inflammatory as well as cardiovascular effects, there has been long-standing interests in using corticosteroid to treat sepsis and meningitis. While corticosteroids have been shown to protect animals against mortality from experimental sepsis, such protection has not been observed in patients with septic shock receiving very high doses of methylprednisolone.^78,79 This discrepancy could be explained by the fact that in experimental sepsis, steroids were given prophylactically prior to the onset of bacterial infection. Indeed, there has been concern about excessive immunosuppression and secondary infections in patients who received high dose steroids.

In experimental Hemophilus influenza meningitis, McCracken et al showed that intravenous dexamethasone administration results in concomitant reduction in meningeal inflammation and TNF-α levels in the cerebrospinal fluid.⁶⁴ In children with bacterial meningitis, simultaneous use of dexamethasone and antibiotics resulted in lower levels of proinflammatory cytokines, prostaglandin I₂, and lactates in the cerebrospinal fluid.^65,80 In these studies, approximately 75% of the patients were infected with H. influenza. Early intravenous administration of dexamethasone, 0.15 mg/kg/dose every 6 hours for 4 days, in addition to appropriate antibiotics, resulted in shorter duration of fever and less neurological sequelae including hearing loss.⁸¹ However, subsequent studies did not confirm these findings.⁸² In these studies,^80,81 dexamethasone therapy was not associated with delayed sterilization of cerebrospinal fluid cultures. Approximately one fifth to two third of patients developed transient, low grade fever one to two days after discontinuation of dexamethasone. While some unusual cases of gastrointestinal bleeding were observed necessitating blood transfusion, there were few serious side effects attributable to dexamethasone.

The discrepancy in clinical response of patients in different studies may be due to many reasons including selection of patient population, severity of diseases, or time lapsed between the initiation of antibiotics and administration o steroids. From the biological standpoint, in order to block the initiation of cytokine cascade, it is important that dexamethasone should be initiated the earlier the better, preferably prior to the infusion of antibiotics. Corticosteroid suppresses the transcription of cytokine genes and has relatively little effects on translation. Once the cytokine mRNAs are synthesized following activation of monocytes by bacterial endotoxin, further use of steroids will not be very effective.

In view of these studies, the American Academy of Pediatrics Committee on Infectious Diseases recommends the following in the 1994 Red Book: Dexamethasone therapy should be considered when bacterial meningitis in infants and children 6 weeks and older is diagnosed or strongly suspected on the basis of the cerebrospinal fluid tests, including Gram-stained smears, after the physician has weighed the benefits and possible risks and before the etiology has been established.⁸³ Dexamethasone is recommended for treatment of infants and children with H. influenzae meningitis while its efficacy for pneumococcal or meningococcal meningitis is not proven. The recommended dose is 0.6 mg/kg/d in four divided doses, given intravenously, for the first 4 days of antibiotics treatment. The drug should be given as early as possible, preferably at the time of, or shortly after, the first dose of antibacterial therapy. Dexamethasone should not be used for suspected or proven nonbacterial meningitis. If dexamethasone had been started before the diagnosis of nonbacterial meningitis was made, it should be discontinued.

Anti-endotoxin antibody in sepsis

Most monoclonal antibodies are produced from hybrid cells of mouse splenic lymphocytes, following immunization of the mouse with a specific antigen of interest, with mouse myeloma cells. While they are relatively easy to generate in mass quantities, they have significant drawbacks. They can trigger immune responses, including anaphylactic reaction as manifested by hypotension and tachycardia, in as much as half of the recipients following repetitive injections. These adverse effects therefore have prompted the development of a human monoclonal antibody against lipid A moiety of endotoxin (HA-1A, commercially marketed as Centoxin). An initial study showed that HA-1A was not immunogenic and did not induce antibodies to neutralize itself.⁸⁴ This was followed by two randomized, double blind, placebo controlled clinical trials of HA-1A conducted in as many as 24 centers in the United States.^85-87 In the controlled trial,⁸⁵ it was reported that HA-1A significantly reduced mortality by 39% in adults with sepsis and Gram-negative bacteremia when compared to treatment with placebo. However, the second trial, designated as Centocor HA-1A Efficacy in Septic Shock trial, was terminated prematurely after an interim analysis of the initial 500 enrolled patients. In those patients with sepsis not caused by Gram-negative bacteria who were treated with HA-1A, they had a higher mortality rate than untreated controls.^86,87 The reasons underlying this phenomenon remain to be determined. Studies of new agents that neutralize or antagonize the cellular effects of endotoxin are in development. In future trials, in order to allow for immunomodulation at multiple steps of the cytokine cascade, it would be important to consider combining the use of anti-endotoxin antibodies with anti-cytokine antibodies.

Anti-cytokine antibodies in sepsis

Since the importance of TNF-α as a major mediator of inflammation has been confirmed by a number of studies, anti-TNF-α antibodies have been investigated as an adjunctive agent to treat sepsis. At present, there are a number of completed studies on the use of anti-TNF-α antibodies using animal models of sepsis. In brief, these reports showed that polyclonal or monoclonal antibodies to TNF-α protect animals including mice, rats, and baboons from the lethal effects or multi-organ failure induced by endotoxin or Gram negative bacteria.^88-91

With these encouraging findings on the efficacy of anti-TNF-α antibodies in animal models, a number of clinical trials were initiated to examine the use of these agents in human sepsis, sponsored by pharmaceutical companies including Celltech, Bayer-Miles, and Centocor. To date, there are several preliminary reports on a phase 1 and phase 2 trials using murine monoclonal anti-TNF-α antibodies on human subjects with severe septic shock. The antibodies were found to be safe without acute adverse effects,^92,93 and possibly beneficial since it was associated with some improvement in hypotension.⁹³ In another study, treatment of ten patients with septic shock was associated with transient improvement in ventricular function and arterial oxygenation.⁹⁴ However, human anti-mouse antibodies were detected in the patients' serum which may limit the use of murine monoclonal antibodies in repeat treatment of these patients.⁹²

Natural inhibitors of cytokines: soluble receptors and receptor antagonists

In addition to anti-cytokine antibodies, another effective means to neutralize a cytokine would be through the use of its soluble receptors. Shedding of cytokine receptor is a common physiological phenomenon and may form part of a normal homeostatic regulatory system. The presence of specific soluble receptors in the microenvironment surrounding the immune cells would neutralize cytokines and remove them from acting on other cells, thereby limiting the propagation of inflammation. Soluble forms of receptors for TNF-α, IFN-γ, IL-1, IL-2, IL-4, and IL-6 have been reported. There has been much interest in the therapeutic use of soluble TNF receptors in the treatment of septic shock.^3,95 There are also attempts to develop designer molecules of TNF soluble receptors with enhanced affinity and circulating longevity.96

Receptor antagonists refer to molecules that are capable of binding to cytokine receptors, thereby blocking the binding of its cognate and biologically active ligand, without eliciting any physiological response. While the development of receptor antagonists theoretically appears to be ideal for cytokine inhibition, the relatively high molecular weights of cytokine antagonists and the extremely high affinity of cytokine-receptor interaction render the design and practical use of this class of molecules very difficult. The best known example is the clinical use of IL-1 receptor antagonist (IL-1ra) which is produced by adherent monocytes.⁶ Recombinant IL-1ra blocks the activity of IL-1α and IL-1β both in vitro and in vivo by binding to IL-1 receptors without inducing biological activity. The antagonist, when administered at very high concentrations, has been shown to inhibit a wide range of diseases, including septic shock and cerebral malaria, in animal models. Clinical trials using recombinant IL-1ra for the treatment of sepsis was performed⁹⁷ and the therapeutic efficacy was marginal, with no significant differences (p=.22) between the mortality rates of patients treated with placebo (34%) and high dose of IL-1ra at 2 mg/kg/hr (29%). The disappointing results can be analyzed as follows: First, in the blockade of receptors, it would require a large amounts of blocking agents (e.g. IL-1ra) to block every single receptor sites on the immune cells throughout the body. This is reflected by the enormous high doses of IL-1ra used versus the amount of antibodies used in anti-TNF or anti-endotoxin studies cited above. Second, the avidity of interaction (as reflected by K_D, dissociation constant) between a cytokine and its cognate receptors is of several hundred to thousand fold lower than that of the antibody and antigen binding reactions, thus rendering the soluble antagonists approach not effective. In all, from a biological and biochemical standpoint, it is better to use anti-endotoxin or anti-cytokine antibodies for therapeutic trials instead of modulating the downstream events such as receptor antagonists to block cytokine receptors.

GM-CSF and G-CSF

Both GM-CSF and G-CSF are growth factors categorized on the basis of their biological activity on the hematopoietic system.^98-106 G-CSF is a late-acting hematopoietin, mainly restricting its effects to the neutrophil lineage.⁹⁸ G-CSF has major activity on the differentiation and development of neutrophils including phagocytosis, superoxide release, antibody-dependent cellular cytotoxicity, and migration. In contrast, GM-CSF is a multilineage hematopoietin and acts upon multiple steps of myeloid differentiation from the early stem cell to end cell functions. While relatively specific for the development of both granulocytes and monocytes/ macrophages from myeloid progenitors, GM-CSF has growth and differentiation effects on other immune cell precursors.^98-100 In addition, GM-CSF has effects on end cell functions including phagocytosis, neutrophil migration, and metabolic activity. Due to the multi-lineage activities of GM-CSF and its effects on rendering monocytes hypersensitive to endotoxin stimulation of TNF synthesis, there is a possible, at least theoretical, risk involved in using GM-CSF in patients who are prone to develop bacterial infections. Nevertheless, areas of therapeutic potential for these two cytokines overlap because of their common biological activities. For an overview here, both agents will be discussed simultaneously for their therapeutic uses.

Recombinant human GM-CSF and G-CSF are used as adjunctive therapy, in many centers as standard therapy, in recipients of bone marrow transplantation and cancer chemotherapy, and in patients with dysfunctional hematopoiesis (Table IV). To date, clinical trials on human subjects have obtained impressive results validating the in vitro effects of these CSF on myelopoiesis in vivo.⁹⁸ Both GM-CSF and G-CSF have clearly reduced neutropenia and infection rate when given to patients after conventional cytotoxic chemotherapy and high-dose chemotherapy preceding autologous bone marrow transplantation. Similar results have been described in other leukemias, including acute myelogenous and hairy cell, following chemotherapy, as well as other diseases including aplastic anemia, myelodysplasia, drug-induced agranulocytosis, and chronic neutropenia.^101,103-105 In general, the recipients showed increases in peripheral blood granulocyte counts, less opportunistic infection, and shorter periods of hospitalization. Treatment appears to be well tolerated, despite common occurrence of mild to moderate flu-like symptoms and occasional fluid retention problems. However, there is one potential concern whether GM-CSF or G-CSF may stimulate residual abnormal clones in patients with myelodysplasia or acute myelogenous leukemia.¹⁰² While this undesirable effect has been observed in vitro on myelogenous leukemic cells, current data suggest that with appropriate monitoring and exclusion of high-risk patients, this has not been a significant problem clinically.

In all, recombinant CSFs constitute a new class of therapeutic modality. They have demonstrated major potential in the management of myelosuppression following cytotoxic chemotherapy or bone marrow transplantation. They can also be used as primary treatment for other causes of neutropenia. Application of these growth factors to the bedside represents a major contribution of biotechnology to a difficult area of therapeutics in febrile, neutropenic patients.

Summary

Recent advances in basic research and clinical trials have demonstrated that IFNs and proinflammatory cytokines play crucial roles in the pathogenesis of infectious diseases. They are important mediators of immune responses to pathogen invasion. Interaction among cytokines leads to propagation of inflammation and host defense against pathogens. In this paper, we have reviewed the biological activity of IFNs, TNF-α, interleukins, and GM-CSF and G-CSF. Since IFN-α plays crucial roles in antiviral responses, its therapeutic applications in hepatitis B and C, human papilloma virus, HIV infection, and malignancy were discussed. The use of IFN-γ in bacterial prophylaxis in patients with chronic granulomatous disease was also presented. In patients with neutropenia due to cytotoxic chemotherapy, bone marrow transplantation, and congenital agranulocytosis, the use of GM-CSF and G-CSF was reviewed. Since over-expression of certain cytokines may cause undesirable effects and contributes to tissue pathology, we have also discussed recent advances in using immunomodulators and cytokine antagonists to suppress cytokines in order to mitigate the undesirable effects of excessive inflammatory cytokines in disease processes.

Acknowledgments

The authors wish to thank Drs. M YEUNG and F. GEERTSMA for their helpful discussion, Professor CY YEUNG for reviewing the manuscript, and Ms. M LAU for assistance in literature research and organization of the materials.

References

1. Cohen S. The role of cell-mediated immunity in the induction of inflammatory responses. Am J Pathol 1977;88:502-8.

2. Kohase M, May LT, Vilcek J. A cytokine network in human diploid fibroblasts: interaction of β-IFN, TNF, PDGF and IL-1. Mol Cell Biology 1987;7:273-8.

3. Henderson B, Blake S. Therapeutic potential of cytokine manipulation. Immunol Today 1992;13:145-52.

4. Gutterman JU. Cytokine therapeutics: lessons from interferon-alpha. Proc Natl Acad Sci USA. 1994;91:1198-205.

5. Aggarwal BB, Vilcek J (editors). Tumor necrosis factor: structure, function, and mechanism of action. Marcel Dekker Inc, New York, 1992.

6. Dinarello C. Interleukin-1 and interleukin-1 antagonism. Blood 1991;77:1627-52.

7. Akira 5, Kishimoto T. Interleukin-6 and NF-IL-6 in acute phase response and viral infection. Immunol Review 1992;127:25-50.

8. Taylor JL, Grossberg SE. Recent advances in interferon research: molecular mechanisms of regulation, action, and virus circumvention. Virus Research 1990;15:1-26.

9. Baron S (editor) Interferon: basic science and medical application. Univ Texas Medical Branch Press, Galveston, Texas, 1992.

10. Lau AS. Cytokines in the pathogenesis and treatment of infectious diseases. Adv Pediat Infect Dis 1994;9:211-36.

11. Helm MH, Kerr IM, Stark GR, Darnell JE. Contribution of STAT SH2 groups to specific IFN signaling by the Jak-STAT pathway. Science 1995;267:1347-9.

12. Williams BRG. Transcriptional regulation of interferon stimulated genes. Eur J. Biochem 1991;200:1-11.

13. Whitten TM, Quets AT, Schloemer RH. Identification of the hepatitis B virus factor that inhibits expression of the beta interferon gene. J Virol 1991;65:4699-704.

14. Lau AS, Der SD, Menill JD, Yeung MC. Evasion of the interferon-a system by HIV. Pasteur Res Immunol 1994;145:668-73.

15. Lau AS, Read SE, Williams BRG. Down regulation of interferon alpha but not gamma receptor expression in vivo in AIDS. J Clin Invest 1988;82:1415-21.

16. SenGupta DN, Silverman RH (1989) Activation of interferon-regulated dsRNA-dependent enzymes by HIV-1 leader RNA. Nucleic Acids Res 17:969-78.

17. Schroder HC, Ugarkovic D, Wenger R, et al. Binding of tat protein to TAR region of HIV-1 blocks TAR-mediated activated of 2-5A synthetase. AIDS Res Human Retroviruses 1990;6: 659-72.

18. Roy S, Katze MG, Pakin NT, et al. Control of the interferon-induced 68-kilodalton protein kinase by HIV-1 tat gene product. Science 1990;247:1216-9.

19. Roy S, Agy MB, Hovanessian AG, et al. The integrity of the stem structure of HIV-1 Tat-responsive sequence RNA is required for interaction with the IFN-induced 68,000-Mr protein kinase. J Virol 1991;65:632-40.

20. Der SD, Lau AS. Involvement of the double-stranded-RNA-dependent kinase PKR in the interferon expression and interferon-mediated antiviral activity. Proc Natl Acad Sci USA. 1995;92:8841-5.

21. Lane HC, Davey V, Kovacs JA. Interferon-alpha in patients with asymptomatic HIV infection. Ann Intern Med 1990;112:805-11.

22. Krown SE, Real FX, Cunningham-Rundles T. Preliminary observation on the effect of recombinant leukocyte A interferon in homosexual men with Kaposi's sarcoma. N Engl J Med 1983;308:1071-6.

23. Krown SE. Approaches to interferon combination therapy in the treatment of AIDS. Semin Oncology 1990;17:11-5.

24. Folks TM, Clouse KA, Justement J, et al. Tumor necrosis factor-a induces expression of HIV in a chronically infected T-cell clone. Proc Natl Acad Sci USA 1989;86:2365-8.

25. Osborn L, Kunkel S, Nabel GJ. Tumor necrosis factor-a and interleukin-1 stimulate the HIV enhancer by activation of the nuclear factor kB. Proc Natl Acad Sci USA 1989;86:2336-40.

26. Don RT. Interferon-alpha in malignant and viral diseases. Drugs 1993;45:177-211.

27. Antonelli G, Cunenti M, Tutniziani 0, Dianzani F. Neutralizing antibodies to interferon-alpha: relative frequency in patients treated with different interferon preparations. J Infect Dis 1991;163:882-5.

28. Quesada JR, Rios A, Swanson D, Trown P, Gutterman JU. Antitumor activity of recombinant-derived interferon alpha in metastatic renal cell carcinoma. J Clin Oncology 1985;3:1522-8.

28a. Mahapatra BD, Schwartz J, Nagabhushan TL. Detection of squinel monkey retrovirus (SMRV) sequences in samples of commercially available interferon products by polymerase chain reaction. International Society for IFN and Cytokine Research Meeting Abstract. Baltimore, November 1995.

28b. Zhao X, Hua J, Kuchler M, et al. Absence of neutralizing antibody development in patients treated with interferon alfaN3. International Society for IFN and Cytokine Research Meeting Abstract. Baltimore, November 1995:S141.

29. von Wussow P, Pralle H, Hochkeppel HK, et al. Effective natural IFN-alpha therapy in recombinant IFN-alpha-resistant patients with hairy cell leukemioa. Blood 1991;78:38-43.

30. Casato M, Antonelli G, Maggi F, et al. Resistance to recombinant alpha IFN therapy in idiopathic mixed cryoglobulinemia: reinduction of remission by natural alpha IFN both in antibody-positive and -negative patients. J Biol Regulators Homeostatic Agents 1994;8:56-9.

31. Brand CM, Leadbeater L, Budiman R, et al. Successful retreatment of anti-IFN resistant polycythemia vera patient with lymphoblastoid IFN-alpha N1 and in vitro studies on the specificity of the antibodies. Brit J Hematology 1994;86:216-8.

31a. Gordon SC, Everson G, McHutchison JG, et al. Natural alfa interferon (Alferon) treatment of chronic hepatitis C in patients refractory to recombinant interferon alfa 2B. Hepatology 1995;22:No. 4, Pt. 2, Abstract 41.

32. Don RT. Clinical properties of yeast-derived versus E. coli-derived GM-CSF. Chin Therap 1993;15:19-29.

33. Talpaz M, Kantarjian HM, McCredi K, et al. Hematologic remission and cytogenetic improvement induced by recombinant human IFN aA in chronic myelogenous leukemia. N Engl J Med 1986;314(17):1065-9.

34. Ratain MJ, Golomb HM, Bardawil RG, Vardiman JW, et al. Durability of response to IFN α2b in advanced hairy cell leukemia. Blood 1987;69:872-9.

35. Nelson BE, Borden EC. Interferons: biological and clinical effects. Semin Surg Oncol 1989;5:391-401.

36. Lauer SJ, Ochs J, Pollock BH, et al. Recombinant alpha-2b interferon treatment for childhood T-lymphocytic disease in relapse. Cancer 1994;74:197-202.

37. Brouty-Boye D, Zetter BR. Inhibition of endothelial cell proliferation by gamma-interferon. Science 1980;208:516-8.

38. White CW, Sondheimer HM, Crouch EC, et al. Treatment of pulmonary hemangiomatosis with recombinant IFNa2a. N Engl J Med 1989;320:1197-200.

39. Ezekowitz RA, Mulliken J, Folkman J. Alpha interferon 2a therapy for life threatening hemangioma of infancy. N Engl J Med 1992;326:1456-63.

40. Eron EJ, Judson F, Tucker S, et al. Interferon therapy for condyloma acuminata. N Engl J Med 1986;315:1059-1064.

41. Friedman-Kien AE, Eron U, Conant M, et al. Natural interferon alpha for treatment of condylomata acuminata. JAMA 1988;259:533-8.

42. Reichman RC, Oakes D, Bonnez W, et al. Treatment of condyloma acuminata with three different interferons administered intralesionally. Ann Intern Med 1988;108:675-9.

43. Lusk RP, MaCabe BF, Mixon JH. Three-year experience of treating recurrent respiratory papilloma with interferon. Ann Otol Rhinol Laryngol 1987;19:158-62.

44. Healy GB, Gelber RD. Trowbridge AL, et al. Treatment of recurrent respiratory papillomatosis with human leukocyte interferons. N Engl J Med 1988;319:401-7.

45. Beasley RP, Hwang LY. Epidemiology of hepatocarcinoma. In: Vyas, et al. (Eds) Viral hepatitis and liver disease. Grune & Stratton, Orlando, Florida 1984:209-24.

46. Hoofnagle JH, Peters M, Jones DB, et al. Randomized, controlled trial of recombinant human IFN-α in patients with chronic hepatitis B. Gastroenterlogy 1988;95;1318-25.

47. Perrillo RP. Treatment of chronic hepatitis B with interferon: experience in Western countries. Semin Liver Diseases 1989;9:240-8.

48. Korenman J, Baker B, Waggoner J, et al. Long term remission of chronic hepatitis B after alpha interferon therapy. Ann Int Med 1991;114:629-34.

49. Lai CL, Lin HJ, Yeung CY, et al. Placebo-controlled trial of recombinant α2 interferon in Chinese HBsAg-carrier children. Lancet 1987;2:877-80.

50. Ruiz-Moreno M, Rua MJ, Molina J, et al. Prospective randomized trial of interferon-alpha in children with chronic hepatitis B. Hepatology; 1991;13:1035-9.

51. Utili R, Sagnelli E, Galanti B, et al. Prolonged treatment of children with chronic hepatitis B with recombinant alpha 2a interferon: a controlled, randomized study. Am J Gastroenterology 1991;86:327-30.

52. Lai CL, Lin HJ, Lau JN, et al. Effect of recombinant alpha2 IFN with or without prednisone in Chinese HBsAg carrier children. Quart I Med 1991;78:155-63.

53. Utili R, Sagnelli E, Gaeta GB, et al. Treatment of chronic hepatitis B in children with prednisone followed by alpha interferon: a controlled, randomized study. J Hepatology 1994;20:163-7.

54. Kuo G, Choo QL, Alter HI, et al. An assay for circulating antibodies to a major etiologic virus of non-A, non-B hepatitis. Science 1989;244:362-4.

55. Davis GL, Balart LA, Schiff ER, et al. Treatment of chronic hepatitis C with recombinant IFN-alpha. A multicenter randomized controlled trial. N Engl J Med 1989;321:1501-6.

56. Toyoda H, Nakano S, Takedo I, et al. Retreatment of chronic hepatitis C with IFN. Am J Gastroenterology 1994;89:1453-7.

57. Sechler JMG, Malech HL, White CJ, Gallin JI. Recombinant human interferon-gamma reconstitutes defective phagocyte function in patients with chronic granulomatous disease of childhood. Proc Natl Acad Sci (USA) 1988;4874-8.

58. Ezekowitz RAB, Dinauer MC, Jaffe HS, et al. Partial correction of phagocyte defects in patients with X-linked chronic granulomatous disease by cutaneous interferon-a. N Engl J Med 1988;319:146-51.

59. International Chronic Granulomatous Disease Cooperative Study Group. A controlled trial of interferon gamma to prevent infection in chronic granulomatous disease. N Engl J Med 1991;324:509-16.

60. Cer A, Beuffer B. Cachectin: more than a tumor necrosis factor. N Engl J Med 1987;316:379-85.

61. Cerami A, Beutler B. The role of cachectin/TNF in endotoxin shock. Immunol Today 1988;9:28-31.

62. Girardin E, Grau G, Dayer JM, Roux-Lombard P, J5 Study Group, et al. Tumor necrosis factor and interleukin-1 in the serum of children with severe infectious purpura. N Engl J Med 1988;319:397-400.

63. Muto Y, Noun-Aria KT, Alexander GJ, et at. Enhanced tumor necrosis factor and interleukin-1 in fulminant hepatic failure. Lancet 1988;2(8602):72-4.

64. Mustafa MM, Ramilo O, Olsen KD, McCracken GH, et al. Tumor necrosis factor in mediating experimental H. influenzae type b meningitis. J Clin Invest 1989;84:1253-9.

65. Mustafa MM, Lebel MH, McCracken GH, et al. Correlation of interleukin-1b and cachectin concentrations in CSF and outcome from meningitis. J Pediat 1989;115:208-13.

66. Lange B, Silverman E, Laxer R, Lao AS. Spontaneous production of tumor necrosis factor in Kawasaki disease. J Pediatrics 1989;115:939-43.

67. Hirano T. Interleukin-6 and its relationship to inflammation and disease. Clin Immunol and Immunopathol 1992;62(1):S60-S65.

68. Miles SA, Rezai AR, Salazar-Gonzalez JF, et al. AIDS Kaposi sarcoma-derived cells produce and respond to interleukin-6. Proc Nail Acad Sci (USA) 1990;87:4068-72.

69. Cheers C, Zhan YF, Egan P1. In vivo interleukin-1 potentiates both specific and non-specific arms of immune response to infection. Immunology 1990;70:411-3.

70. Dinarello CA. Role of interleukin-1 in infectious diseases. Immunol Review 1992;127;119-46.

71. Waage A, Brandtzaeg P, Halstensen A, et al. The complex pattern of cytokines in serum from patients with meningococcal septic shock. J Exp Med 1989;169:333-8.

72. Waage A, Halstensen A, Espevik T. Association between tumor necrosis factor in serum and fatal outcome in patients with meningococcal disease. Lancet 1987;ii:355-7.

73. Marano MA, Fang Y, Moldawer LL, et at. Serum cachectin/ tumor necrosis factor in critically ill patients with burn correlates with infection and mortality. Surg Gynecol Obstet 1990;107:32-8.

74. Kern P, Hemmer CJ, Van Damme J, et al. Elevated tumor necrosis factor alpha and interleukin-6 serum levels as markers for complicated Plasmodium falciparum malaria. Am J Med 1989;87:139-43.

75. Ruscone F, Parizzi F, Garlaschi L. Interleukin-6 activity in infants and children with bacterial meningitis. Pediatr Infect Dis J 1991;10:117-21.

76. Sullivan IS, Kilpatrick L, Costarino AT, et al. Correlation of plasma cytokine elevations with mortality rate in children with sepsis. J Pediat 1992;120:510-5.

77. Chavanet P, Bonnotte B, Guiguet M, et al. High concentrations of intrathecal IL-6 in human bacterial and non-bacterial meningitis. J Infect Dis 1992;166:428-31.

78. Hinshaw L, Peduzzi P, Young E, et al. Effect of high-dose glucocorticoid therapy on mortality in patients with clinical signs of systemic sepsis. N Engl J Med 1987;317:659-65.

79. Bone RC, Fisher CJ, Clemmer TP, et al. A controlled clinical trial of high-dose methylprednisolone in the treatment of severe sepsis and septic shock. N Engl J Med 1987;317:653-8.

80. Mustafa MM, Ramilo O, Saez-Llorens X, McCraken GH, et at. CSF prostaglandin, IL-1b and TNF in bacterial meningitis: clinical and laboratory correlation in placebo-treated and dexamethasone treated patients. Am J Dis Child 1990;144:883-7.

81. Odio CM, Faingezicht I, McCracken GH, et al. The beneficial effects of early dexamethasone administration in infants and children with bacterial meningitis. N Engl J Med 1991;324:1525-31.

82. King SM, Law B, Wang E, Gold R, et al. A randomized controlled trial of dexamethazone versus placebo in children with bacterial meningitis. Internatl Conf Antimicrob Agent Chemotherapy Abstract 1992;pp122.

83. 1994 Red Book, Report of the Committee on Infectious Diseases, American Academy of Pediatrics 1994:558-9.

84. Fisher CJ, Zimmerman J, Khazeli MB, et al. Initial evaluation of human monoclonal anti-lipid A antibody (HA-1A) in patients with sepsis syndome. Crit Care Med 1990;18:1311-5.

85. Ziegler El, Fisher CJ, Sprung CL, et al. Treatment of gram-negative bacteremia and septic shock with HA-1A human monoclonal antibody against endotoxin. N Engl J Med 1991;324:429-36.

86. Luce JM. Introduction of new technology in critical care practice: A history of HA-1A human monoclonal antibody against endotoxin. Crit Care Med 1993;21:1233-41.

87. Gibaldi M. Anatomy of an antibody, and related misadventures in developing an effective treatment for septic shock. Pharmacotherapy 1993;13:302-8.

88. Beutler B, Milsak IW, Cerami AC. Passive immunization against cachectin/TNF protects mice from lethal effects of endotoxin. Science 1985;229:869-71.

89. Tracey KJ, Fong Y, Hesse DG, et al. Anti-cachectin /TNF monoclonal antibodies prevent septic shock during lethal bacteremia. Nature 1987;330:662-4.

90. Opal SM, Cross AS, Kelly NM, et al. Efficacy of a monoclonal antibody directed against TNF in protecting neutropenic rats from lethal infection with Pseudomonas aeruginosa. J Infect Dis 1990;161:1148-52.

91. Hinshaw LB, Tekamp-Olson P, Chang ACK, et at. Survival of primates in LD100 septic shock following therapy with antibody to tumor necrosis factor. Circ Shock 1990;30:279-92.

92. Exley AR, Cohem J, Buurman W, et al. Monoclonal antibody to TNF in severe septic shock. Lancet 1990;335:1275-6.

93. Exley AR, Cohen J, Burman W, et al. Monoclonal antibodies to TNF-a in severe septic shock. Lancet 1990;335:1275-7.

94. Vincent JL, Bakker J, Marecaux G, et al. Administration of anti-TNF antibody improves left ventricular function in septic sphoock patients: Results of a pilot study. Chest 1992;101 :810-5.

95. Christmann J, Jlolden EP, Blackwell TS. Critical Care Med 1995;23:955-63.

96. Aderka D, Engelman H, Maor Y, et at. Stabilization of the bioreactivity of TNF by its soluble receptors. J Exp Med 1992;175:323-29.

97. Fisher CJ, Dhainault JF, Pribble JP, et at. Recombinant human IL-1 receptor antagonists in the treatment of apteients with sepsis syndrome. A randomized, open-label, placebo-controlled multicenter trial. Phase III rhIL-1ra Sepsis Study Group. JAMA 1994;27:1836-43.

98. Grosh WW, Quesenberry PJ. Recombinant human hematopoietic growth factors in the treatment of cytopenias. Clin Immunol Immunopathol 1992;62:525-538.

99. Robinson BE, Quesenberry PJ. Hematopoietic growth factors. Am J Med Sci 1990;300:163-70,237-44,311-21.

100. Metcalfe D. The colony stimulating factors: discovery, development, and clinical applications. Cancer 1990;65:2185-95.

101. Delannoy A. GM-CSF therapy drug-induced agranulocytosis. J Intern Med 1992;231:269-71.

102. Vallenga E, Young DC, Wagner K, et al. The effects of GMCSF and G-CSF in promoting growth of clonogenic cells in acute myeloblastic leukemia. Blood 1987;69:1771-76.

103. Furman, WL, Crist WM. Biology an5 clinical applications of hemopoietins in pediatric practice. Pediatr 1992;90:716-28.

104. Nemunaitis J, Rabinowe SN, Singer JW, et al. Recombinant granulocyte- macrophage colony-stimulating factor after autologous bone marrow transplantation for lymphoid cancer. N Eng J Med 1991;324:1773-78.

105. Riikonen P, Saarinen UM, Makripemaa A, et al. Recombinant human granulocyte- macrophage conlony-stimulating factor in the treatment of febrile neutropenia: a double blind placebo-controlled study in children. Pediatr Infect Dis J 1994;13:197-202.

106. Wright DG, Kenny RF, Oette DH, et al. Contrasting effects of recombinant human granulocyte-macrophage colony-stimulating factor (CSF) and granulocyte CSF treatment of cycling of blood elements in childhood-onset cyclic neutropenia. Blood 1994;84:1257-67.