INTRODUCTION

TOP ABSTRACT INTRODUCTION CASE REPORT DISCUSSION REFERENCES

The role of cytokines and interferons in the pathogenesis of and potential therapy for human disease is increasingly being recognized. Two of the more commonly used biologic response modifiers in clinical trials are interleukin-2 (IL-2) and interferon- (IFN-). Endogenous IL-2 is produced by activated T-cells and serves to activate several types of immune cells (1). Exogenous IL-2 has been used in human clinical trials since the mid-1980s as adjuvant immunotherapy for advanced malignancies such as melanoma, renal cell carcinoma, and lymphoma. Immunotherapy with IL-2 alone and IL-2 combined with lymphokine-activated killer (LAK) cells has been shown to induce tumor regression in phase II trials (2, 3). Endogenous IFN- is produced by T-cells and natural killer cells. IFN- upregulates the expression of HLA-DR antigen, which is thought to render cancer cells more sensitive to cell-mediated cytotoxicity (4, 5). IL-2 and IFN-, along with cyclosporine, are used together in stem cell transplantation in order to induce a graft-versus-tumor effect (6).

The clinical use of IL-2 has been limited by the development of a vascular leak syndrome characterized by weight gain, edema, oliguria, and the development of pulmonary infiltrates and dyspnea (3, 7). Radiographically, the therapeutic use of IL-2 has been associated with the demonstration of pulmonary edema, pleural effusions, and pericardial effusions (8). The specific etiology of this vascular leak syndrome is unknown and has been attributed to both direct and indirect effects of IL-2 (9). Here we report a case of pulmonary infiltrates, hypoxemia, and dyspnea after immunotherapy with IL-2 and IFN- that was associated with massive deposition of the toxic eosinophil granule, MBP, throughout the pulmonary interstitium.

	CASE REPORT

TOP ABSTRACT INTRODUCTION CASE REPORT DISCUSSION REFERENCES

A 49-yr-old Caucasian woman with a history of hypothyroidism underwent a modified left radical mastectomy for breast carcinoma followed by adjuvant chemotherapy with cyclophosphamide/methotrexate/5-fluorouracil. Recurrent carcinoma was noted at the left pectoral muscle 4 yr later and was surgically resected. Four lymph nodes were positive for malignancy at the time of recurrence, but the surgical margins were negative. The patient subsequently received four cycles of adriamycin/cyclophosphamide. Because of the high risk for further recurrence, the patient then underwent autologous stem cell transplantation with a preparative regimen consisting of high dose cyclophosphamide and thiotepa. On the day of the transplant, cyclosporine was begun per protocol at 2.5 mg/kg of ideal body weight.

Five days after the transplant the patient developed a neutropenic fever and was begun on antibacterial and subsequently antifungal therapy. An extensive fever workup did not reveal a source of infection. One week after the stem cell transplant the patient was begun on IL-2 and IFN- per protocol. IFN- was administered subcutaneously at a dose of 0.025 mg/m² for 10 d and IL-2 was administered subcutaneously at a dose of 0.935 million units/d for 11 d. Because of the appearance of peripheral eosinophilia (counts as great as 9,259/mm³), a negative fever workup, and a lack of response to antimicrobial therapy, drug fever was suspected and antibiotics were discontinued. The fever abated and the patient was discharged 25 d after receiving infusion of her stem cell transplant.

Two days after discharge, the patient was readmitted with a nonproductive cough, worsening dyspnea with mild exertion, worsening fatigue, and recrudescence of fever to 39.2° C. The physical examination was notable for bilateral crackles at the lung bases and 2+ pedal edema extending to the knees. Room air blood gas revealed a pH of 7.46, PCO₂ of 26 mm Hg, PO₂ of 60 mm Hg, and a calculated HCO₃ of 18. A radiograph of the chest revealed bilateral perihilar infiltrates and a small right pleural effusion. CT of the chest was consistent with either a pneumonic infiltrate or pulmonary vascular congestion (see Figure 1A). These perihilar infiltrates were not present on a chest CT performed 1 mo earlier. The differential diagnosis of the infiltrates and dyspnea included cardiogenic pulmonary edema, drug reaction, and infectious pneumonia. The patient was begun on diuretic therapy, antibiotics, and subsequently ganciclovir. A transthoracic echocardiogram revealed mild left ventricular hypertrophy and normal left ventricular function. The patient underwent bronchoscopy with bronchoalveolar lavage that was nondiagnostic for an infectious etiology. Despite a diuresis of 3 kg and the development of prerenal azotemia, the infiltrates persisted. The IL-2 and IFN- were suspected as a cause of the pyrexia, but fevers continued despite the completion of the cytokine protocol.

View larger version (124K): [in this window] [in a new window]   Figure 1.   (A) Chest CT at the level of the hilus showing perihilar infiltrates and bilateral pleural effusions. Some motion artifact is present. (B) Transbronchial biopsy (H&E stain; original magnification: ×100): low-magnification photomicrograph showing absence of alveolar interstitial thickening, and a few eosinophils within the interstitium (arrows). This section contained the greatest number of eosinophils of any section examined. (C ) Transbronchial biopsy (H&E stain; original magnification: ×250): on higher magnification, punctate eosinophilic structures consistent with intact eosinophil granules were evident within the submucosa (arrows). (D) Transbronchial biopsy (indirect immunofluorescence [IF]) using control antisera (original magnification: ×160): no background immunofluorescence was detected. (E ) Transbronchial biopsy (indirect IF with anti-MBP, original magnification: ×160): a few intact eosinophils (brightly fluorescent ovals) are observed infiltrating the tissue with extensive diffuse and granular extracellular MBP deposition. Intense perivascular extracellular MBP staining pattern is indicated by the arrow. (F  ) Transbronchial biopsy (indirect IF with anti-MBP, original magnification: ×160): several intact eosinophils infiltrate this section with extensive diffuse and granular extracellular MBP deposition.

A second bronchoscopy was performed and transbronchial lung biopsies were performed. The pathologic examination showed occasional eosinophils infiltrating the bronchial wall and alveolar interstitium, but it was otherwise unremarkable (see Figures 1B and 1C). Special stains for fungi and mycobacteria were negative (not shown). The patient was begun on prednisone 1 mg/kg/d with a rapid improvement in exercise tolerance, hypoxemia, and the perihilar infiltrates. By the time of her discharge 16 d after readmission, the peripheral edema, eosinophilia, and pulmonary infiltrates had completely resolved. Steroid therapy was discontinued as an outpatient over several weeks and there has not been a recrudescence of dyspnea or the pulmonary infiltrates at 1-yr follow-up.

Subsequently, serial sections of the transbronchial biopsies were examined after indirect immunofluorescence staining using affinity-chromatography purified antibodies specific for eosinophil major basic protein as previously described (10). Massive deposition of this toxic granule protein was observed throughout the bronchial submucosa and alveolar interstitium (see Figures 1D-1F).

	DISCUSSION

TOP ABSTRACT INTRODUCTION CASE REPORT DISCUSSION REFERENCES

Stem cells are undifferentiated, primitive bone marrow cells with the ability both to multiply and to differentiate into specific hematopoietic lineages. In autologous stem cell transplantation, stem cells are initially collected using leukapheresis after G-CSF or GM-CSF administration and are then reintroduced or transplanted after high dose chemotherapy. Autologous stem cell transplantation is being used with increasing frequency and now is more often practiced than bone marrow transplantation for breast cancer. Breast cancer is currently the most common indication for stem cell transplantation (11). This procedure is used both for locally advanced primary tumors and for metastatic disease.

Cyclosporine, IL-2, and IFN- were used together with stem cell transplantation in this case in order to facilitate a graft-versus-tumor effect and potentially a more durable remission. The rationale behind the induction of autologous graft-versus-host disease involves the generation of self-reactive T-cell clones during recovery from intensive chemotherapy. Prolonged treatment with cyclosporine is thought to prevent the deletion of these clones (5, 6, 12). IL-2 has antitumor effects that are thought to be attributable to its: (1) stimulation of the development of lymphokine-activated killer cells, which act on tumor cells in a nonmajor histocompatibility complex (MHC) restricted manner (13); (2) stimulation of tumor-specific cytotoxic T-cells in a MHC-restricted manner (2, 16, 17); and (3) potential activation of IL-5-mediated eosinophil cytotoxicity (discussed further below) (18). IFN- serves to increase the expression of MHC Class II antigen on target cells (4, 19), which allows recognition by the CD4+ and CD8+ T-cells elaborated by the induced graft-versus tumor response.

Infection is of particular concern in IL-2-treated patients because this cytokine has been shown to induce a profound but reversible defect in neutrophil chemotaxis that may explain the increased incidence of bacteremia that occurs after IL-2 administration (20). Among the other potential side effects of IL-2 therapy are eosinophilia, malaise, headache, fever, chills, hyperbilirubinemia, hypotension, diffuse erythroderma, anemia, thrombocytopenia, and a capillary leak syndrome (7, 21). The capillary leak syndrome is characterized by weight gain, edema, and oliguria. Such fluid retention is common and, late in the course of IL-2 therapy, pulmonary edema and interstitial infiltrates associated with dyspnea may develop (3). In a review of 423 treatment courses of IL-2 therapy with or without LAK cells or cyclophosphamide, Lee and colleagues (22) found that 9.2% of patients developed respiratory distress requiring supplemental oxygen and an additional 6.4% required temporary intubation. In this same report, decreased systemic vascular resistance, mean arterial blood pressure, stroke index, and left ventricular stroke work was identified in a study group of 10 subjects receiving IL-2. Notably, the pulmonary capillary wedge pressure was not significantly elevated in these subjects. These and other observations suggest that IL-2 may induce both cardiogenic and noncardiogenic pulmonary edema (22, 23).

In a murine model, IL-2 has been shown to induce fluid extravasation into several organs including the lung, kidney, liver, thymus, and spleen (9). Interestingly, fluid accumulation in the lung, liver, and spleen was markedly attenuated with concurrent cortisone acetate treatment or pretreatment with cyclophosphamide or irradiation. Additionally, IL-2 did not produce extravasation of fluid in the nude mature T-cell-deficient mouse (9). These data suggest that it is not IL-2 that directly causes the increased vascular permeability and the capillary leak syndrome, but rather some other factor dependent on the presence of T-cells. A recent study found that IL-2 can directly increase albumin permeability of cultured endothelial cells in vitro (24). It is currently not clear why similar effects were not observed in vivo (9).

The eosinophilia associated with IL-2 treatment is now known to be mediated by the stimulation of interleukin-5 (IL-5) production from lymphocytes in vivo (25, 26). IL-5 promotes the differentiation, proliferation, and activation of eosinophils (27, 28). Interestingly, van Haelst Pisani and colleagues (29) reported elevations of plasma concentrations of eosinophil granule major basic protein (MBP) that began even before eosinophilia developed, suggesting early eosinophil degranulation in response to systemic IL-2 therapy (29). Skin biopsies in that study obtained after initiation of IL-2 therapy showed extracellular MBP deposition in perivascular areas of the dermis along with infiltrating eosinophils. Because MBP is a known endothelial toxin in vitro (30, 31), it has been suggested that the intravascular release of toxic eosinophil products might contribute to the development of the capillary leak syndrome that accompanies IL-2 immunotherapy (29). Further evidence for the injurious role of MBP is suggested by its presence in the airways of asthmatics (10) and in the skin of patients with atopic dermatitis (32). The side effects of IFN- can include fevers, chills, malaise, nausea, vomiting, leukopenia, mild anorexia, headache, and abnormal liver function tests (33). Interestingly, IFN- can directly activate eosinophils in vitro (34). Because the clinical use of IFN- by itself has not been associated with the development of a capillary leak syndrome, the pulmonary signs and symptoms in this case are most likely attributable to the use of IL-2. The stem cell transplantation protocol used in this patient was complex, however, and it is difficult to determine the precise cause of eosinophil activation in one case study.

The findings presented in this report suggest that activation of eosinophils within the lung parenchyma and subsequent deposition of MBP contributed to a localized pulmonary vascular leak syndrome in this patient. Several observations support this contention. First, none of the blood or BAL cultures showed evidence of an infectious etiology. Second, there was neither a history of nor evidence for cardiogenic dysfunction by echocardiography, and the patient's symptoms and the pulmonary infiltrates did not improve with a vigorous diuresis, which resulted in a prerenal azotemia. Third, the patient improved rapidly while receiving glucocorticoids, which are known to effectively suppress eosinophilopoiesis and induce eosinophil apoptosis (35). Fourth, the observation that the pulmonary infiltrates and dyspnea occurred late in the course of immunotherapy and coincident with the development of eosinophilia argues against a direct role for IL-2-induced capillary permeability in this case. Although we cannot rule out the possibility that the MBP deposition in the interstitium was secondary to the extravasation of plasma known to have elevated levels of MBP (29), the presence of intact eosinophils and eosinophil granules on hematoxylin-eosin (H&E) staining argues against this possibility (see Figures 1B and 1C). It is currently not apparent why this phenomenon was limited to the perihilar region of the lung and did not involve more distal areas of the pulmonary parenchyma.

The radiographic perihilar appearance in this case was difficult to distinguish from cardiogenic pulmonary edema. It was, however, quite distinct from the peripheral infiltrates characteristic of chronic eosinophilic pneumonia, an entity in which the toxicity of eosinophilic granules and MBP has also been implicated (36). It is noteworthy that there was a marked discrepancy between the presence of intact eosinophils in the transbronchial biopsies detected by routine histologic analysis (Figures 1B and 1C) and the massive deposition of MBP observed in adjacent sections by immunostaining (Figures 1E and 1F). This observation suggests that the role of eosinophils in this syndrome is underestimated by routine histology. A discrepancy between tissue eosinophilia seen by H&E and tissue MBP deposition has been observed in the airways of asthmatics (10) and the skin of patients with atopic dermatitis (32) and likely results from the rapid activation and degranulation of extravascular eosinophils in these diseases. It is worth noting that activated eosinophils apart from MBP may also play a direct role in lung damage (31).

We conclude that eosinophil activation and the deposition of eosinophilic products in the lung may result in a localized vascular leak syndrome. More clinical experience with this treatment regimen is required in order to better characterize this syndrome and to determine its incidence. This syndrome should be considered in the differential diagnosis of pulmonary infiltrates that occur acutely after bone marrow transplantation with cytokine augmentation. The optimal treatment regimen for this condition is not known. Although our patient appeared to improve rapidly while receiving glucocorticoids, the decision to use these agents must be made after carefully considering their many immunomodulatory effects. The intrapulmonary deposition of eosinophil granule MBP with an associated capillary leak syndrome after cytokine therapy represents a distinct and iatrogenic form of eosinophilic lung disease.