Clinical and Immunological Effects of Granulocyte-Macrophage Colony-Stimulating Factor Coadministered with Interleukin 2: a Phase IB Study 1

Vol. 2, 319-330, February 1996 Clinical Cancer Research 319

Clinical and Immunological Effects of Granulocyte-Macrophage Colony-stimulating Factor Coadministered with Interleukin 2: A Phase IB Study 1

Joan H. Schiller, 2 Jacquelyn A. Hank, tivity and CD16 expression on monocytes and lymphocytes Masoud Khorsand, Barry Storer, and CD56 expression on lymphocytes were significantly lower in patients receiving GM-CSF simultaneously with Agnes Borchert, Karen Huseby-Moore, IL-2 than in patients receiving the sequential treatment. Dan Burns, Osvaldo Wesly, Antitumor activity was observed in the lungs of four of Mark R. Albertini, George Wilding, eight renal cell carcinoma patients with pulmonary metas- and Paul M. Sondel tases treated with concurrent GM-CSF and IL-2. Although University of Wisconsin Comprehensive Cancer Center, Departments no or minimal shrinkage was observed in the patients' large of Human Oncology and Medicine, Madison, Wisconsin 53792 primary tumors, these results warrant further study. The recommended initial Phase II dose and schedule is 1.25 p~g/kg/day GM-CSF, given concurrently with 1.5 million ABSTRACT Roche units/m2/day (4.5 x 106 international units/m2/day) Interleukin 2 (IL-2) and granulocytes-macrophage col- IL-2, with subsequent escalation of GM-CSF to 2.5 Ixg/kg/ ony-stimulating factor (GM-CSF) are activators of the lym- day after careful observation for toxicities. phocyte and granulocyte/macrophage series, respectively. We conducted a phase IB trial to identify the maximally INTRODUCTION tolerated dose and to assess immunological effects of the The failure of cytotoxic chemotherapy to induce long- combination. Thirty-four patients with incurable cancers lasting remissions in the majority of solid tumors has prompted received 2.5, 5, or 10 Ixg/kg GM-CSF s.c. either before or investigation of new treatment modalities, including antitumor concurrently with 1.5 or 3.0 million units/m2/day IL-2. immunotherapies. The ability of various biological response The most common laboratory and clinical side effects modifiers, such as IL-23 and IFNs, to enhance cellular immunity included an elevation of the total WBC or eosinophil count suggests a potential therapeutic role for these agents in certain due to GM-CSF, and constitutional symptoms due to IL-2. neoplastic diseases, such as melanoma and renal cell carcinoma Grade 3 or 4 toxicities included hypotension, thrombocyto- (1-5). IL-2, a M r 15,000 protein produced by helper T cells penia, elevations in aspartate aminotransferase or bilirubin, following activation by antigens or mitogens, can activate ef- renal toxicity, gastrointestinal hemorrhage, arrhythmia, and fector T-cell populations and NK cells to lyse tumor cells in constitutional symptoms. Two patients receiving 5.0 Ixg/kg vitro and in vivo and can augment ADCC in vitro when used in GM-CSF plus concurrent 3.0 million units IL-2 experienced combination with MoAbs (6-10). As a single agent, it can dose-limiting grade 3 or 4 neurological toxicity, which re- induce antitumor responses in 10-15% of patients with renal versed almost completely. cell carcinoma or metastatic melanoma (1-5, 11). Toxicities, An increase in the serum-soluble IL-2 c~ chain receptor however, are significant, thus promoting continued efforts at was observed with administration of GM-CSF, IL-2, or the identifying methods to enhance antitumor effects by increasing combination. IL-2 therapy enhanced lymphokine-activated the specificity of the immune-mediated tumor destruction while killer activity, antibody-dependent cellular cytotoxicity, and modifying treatment programs to reduce toxicity (2, 11). lymphocyte activation, with increased CD16 and CD56 ex- One possible method for enhancing antitumor effects in- pression. GM-CSF increased expression of human leukocyte cludes combining IL-2 with other immunotherapies. Although antigen DR on peripheral blood monocytes and decreased IL-2 has been combined with IFNs and certain MoAbs, such as surface expression of CD16 on circulating monocytes and the anti-CD3 MoAb, with limited success (12-16), IL-2 has not polymorphonuclear cells. Lymphokine-activated killer ac- been studied extensively in combination with CSFs. CSFs are a family of regulatory glycoproteins that were first identified because of their ability to stimulate precursor hematopoietic

Received 7/26/95; revised 10/6/95; accepted 10/18/95. 1 This research was supported by NIH Grants and Contracts CA53441, CA05436, CA32685, CM87290, CA14520, CA13539, CM47669, 3 The abbreviations used are: IL, interleukin; NK, natural killer; ADCC, HL02143, and RR03186, and American Cancer Society Grant CH-237. antibody-dependentcellular cytotoxicity; MoAb, monoclonal antibody; J. H. S. and G. W. are also supported in part by the Veterans Admin- GM-CSF, granulocyte-macrophage colony-stimulating factor; M-CSF, istration. monocyte CSF; BRMP, Biological Response Modifiers Program; PBM, 2 To whom requests for reprints should be addressed, at Department of peripheral blood mononuclear; PMN, polymorphonuclear;E:T, effector: Medicine, University of Wisconsin Comprehensive Cancer Center, K4/ target; LAK, lymphokine-activatedkiller; HLA, human leukocyte anti- 666 Clinical Science Center, 600 Highland Avenue, Madison, WI gen; IL-2R, IL-2 receptor; MRI, magnetic resonance imaging; TNF, 53792. Phone: (608) 263-8600; Fax: (608) 263-8613. tumor necrosis factor.

cells to form colonies of progenitor cells. However, it is now Table 1 Schema of treatment clear that they are not simply proliferative stimuli but can Dose of regulate the functional activity of mature hematopoietic cells No. of GM-CSF Dose of IL-2 (17). GM-CSF, for example, is necessary for cell survival in Sequence patients (lxg/kg/day) (million units/mZ/day) vitro of both progenitor and mature granulocytes and macro- Sequential, 6 5.0 3.0 phages and also can affect the activation, cell mobility, phago- GM-CSF ~ IL-2 6 10.0 3.0 cytic activity, and ADCC of these cells (17). The effects of Concurrent, 6 2.5 1.5 GM-CSF + IL-2 4 2.5 3.0 GM-CSF and M-CSF on neutrophils and monocytes may trans- 12 5.0 1.5 late into antitumor effects in vitro or in preclinical models (18-25). Clinically, treatment with GM-CSF or M-CSF can enhance monocyte activation and induce monocytosis (26-28). Patients with significant central nervous system disease, central Combining IL-2, a lymphocyte activator, with GM-CSF, nervous system metastases, or serious recent infections, or those an activator of the monocyte/macrophage and granulocyte lin- who required continued therapy with corticosteroids, aspirin, or eage, may be a reasonable approach to the treatment strategy of nonsteroidal antiinflammatory agents were ineligible. Patients activation of multiple effector cell populations. Therefore, we with histories of chronic obstructive pulmonary disease were conducted a Phase IB trial of GM-CSF and IL-2 to determine required to have adequate ventilatory function (forced expira- the maximally tolerated dose, to identify toxicities, and to assess tory volume, ->60% of predicted). All patients underwent place- the immunological effects of the combination. ment of in-dwelling venous catheters.

MATERIALS AND METHODS Treatment Program Reagents Patients were scheduled to receive 2.5, 5, or 10 ~g/kg GM-CSF and either 1.5 or 3.0 × 10 6 units/m2/day IL-2 on a IL-2 and GM-CSF sequential or concurrent schedule (Table 1). Patients were en- Human recombinant IL-2 and GM-CSF for clinical treat- tered in groups of six to ensure adequate numbers for statistical ment and in vitro studies were provided by Hoffmann-LaRoche evaluation of the immunological studies. Patients without dose- Inc. (Nutley, NJ) and Immunex Corp. (Seattle, WA), respec- limiting toxicity or progressive disease in the first course were tively, through the Cancer Therapy Evaluation Program of the eligible to receive a second course of treatment following a National Cancer Institute. IL-2 and GM-CSF are highly purified 2-week rest period. products produced in Escherichia coli and yeast expression Patients were registered sequentially to receive one of two vectors, respectively, using recombinant DNA techniques. IL-2 different treatment schedules (Table 1 and Fig. 1): GM-CSF for units are expressed in BRMP or Hoffmann-LaRoche units, and this 12 days followed by IL-2 (sequential treatment) or concurrently compares to the international standard (in which 1 BRMP unit = with IL-2 starting on day 8. -3 international units; Ref. 29). Thus, the 1.5 × 106-unit/m2/day Sequential GM-CSF Followed by IL-2. To assess the regimen used here with Hoffman-LaRoche IL-2 would correspond effects of GM-CSF and IL-2 individually, the first 12 patients to 4.5 × 106 international units/m2/day commercially available received GM-CSF followed by IL-2 (Fig. 1). GM-CSF was Aldesleukin (IL-2; Chiron, Inc., Emeryville, CA). given s.c. on days 1-12 or until hematological toxicity occurred, as defined below. Following a 3-day rest period, three consec- Clinical Protocol utive weekly 4-day continuous infusions of IL-2 commenced (days 15-18, 22-25, and 29-32). Patients GM-CSF Administered Concurrently with IL-2. To Thirty-four patients were enrolled in this Phase I trial determine the clinical and biological effects of a combination of (National Cancer Institute BRMP protocol B92-0001). All pa- IL-2 and GM-CSF, patients 13-34 received IL-2 as a 4-day tients had refractory cancers, for which other proven effective continuous infusion for 3 weeks (days 1-4, 8-11, and 15-18) treatments were not available, and all signed approved informed with GM-CSF starting the second week of IL-2 (day 8). Again, consent forms. Patients had an Eastern Cooperative Oncology GM-CSF was administered for 12 days (days 8-19) or until Group performance status of 0-1 (Karnofsky, 80-100) and a treatment-limiting toxicity occurred (Fig. 1). life expectancy of at least 12 weeks. Eligibility criteria included normal hematological parameters (hemoglobin, -> 10 g/dl; gran- Toxicity Grading and Dose Modifications ulocyte count, >2000/mm3; and platelet count, ->100,000/mm3), The Common Toxicity Criteria were used for grading of adequate liver function (total serum bilirubin, <2.0 mg/dl; and toxicities. The University of Wisconsin Comprehensive Cancer serum glutamic oxaloacetic transaminase, <3 times the upper Center clinical toxicity grading scale was used for weight gain, limit of normal), and adequate renal function (serum creatinine, systolic blood pressure, temperature, and decline in performance <2.0 mg/dl; and/or creatinine clearance, >50 ml/min). Criteria status (8, 12). Mild (grade 1) toxicity corresponded to a for exclusion included major surgery within the last 3 weeks or -<20-mm Hg decrease in systolic blood pressure, a 5-10% treatment with cytotoxic chemotherapy, radiation therapy, or weight gain, temperature _<38°C, and a decline in performance other immunosuppressive therapy within 4 weeks of entry into status of two grades. Moderate toxicity corresponded to grade 2 this clinical trial. Patients with clinically significant cardiac toxicity of a 21-39-mm Hg decrease in systolic blood pressure, abnormalities or significant pleural effusions were ineligible. 11-14% weight gain, temperature of 38.1-39.9°C, and a decline of

Sequential GM-CSF ~ IL-2: Courses 1 and 2 Week: ...... Week 1 Week 2 Week 3 Week 4 Week 5

Day: On-study 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 ..... GM-CSF IL-2

Biologic X X X X X monitoring

Concurrent GM-CSF + IL-2: Courses 1 and 2 Week: Week 1 Week 2 Week 3

Day: On-study 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 GM-CSF

IL-2

Biologic X X X X X X X monitoring

Fig. 1 Schematicdiagram of sequential and concurrent administration of IL-2 and GM-CSF, and timing of biological monitoring. In both schedules, there was a 3-week rest period between courses. *, serum samples were obtained 24 h after the last dose of GM-CSF.

three grades in performance status. Severe toxicity corresponded to PBM and PMN Cell Isolation grade 3 with a ->40-mm Hg decrease in systolic blood pressure, PBM cells from patients and healthy donors were separated ->15% weight gain, temperature ->40°C, and a decrease of four from whole blood by centrifugation on a Ficoll-hypaque gradi- grades in performance status. For grade 3 toxicity, treatment was ent and washed twice with PBS. PMN cells (granulocytes) were withheld until the toxic reaction(s) improved to grade 1 or symp- obtained from the dense RBC layer of the gradient by adding 4 toms returned to baseline. Treatment was then resumed at a 50% ml of the patient's plasma (saved from the gradient) per 10 ml dose. whole blood and 1.6 ml dextran (4.5% in 0.85% NaC1 with 0.2% Initially, GM-CSF was administered for 12 days or until Tris, adjusted to pH 7.4, and sterile filtered) per 10 ml whole the absolute neutrophil count reached 50,000/mm 3. Because of blood at room temperature. After allowing the preparation to a grade 4 central nervous system toxicity occurring in patient 17, settle for 40-60 min, the upper dextran layer containing PMN associated with an eosinophil count of 73,000/mm 3, the protocol cells was obtained, and the contaminating RBCs were lysed with was amended so that dose modifications were to occur for an a lysing buffer. Neutrophils were not separated from this pop- eosinophil count of >20,000/mm 3 and/or a total WBC count of ulation. After centrifuging, the PMN pellet was washed with >50,000/mm 3. Patients 18-20 were considered to have dose- cold PBS. limiting toxicity if this occurred. Starting with patient 21, the protocol was amended so that doses of IL-2 and/or GM-CSF ADCC would be reduced or held for these parameters; however, this All ADCC assays were performed in RPMI-HS (RPMI was not considered dose-limiting toxicity. 1640 supplemented with 10% human serum, 25 rnM HEPES, 100 units/ml penicillin, and 100 ~g/ml streptomycin sulfate) and Response Criteria in RPMI-HS supplemented with IL-2 or GM-CSF. Effector cells A partial response was defined as a 50% or greater shrink- in a total volume of 50 ~1 were plated in quadruplicate into age in the sum of the areas of all known disease, lasting at least 96-well, U-bottom microliter plates at E:T ratios of 50:1, 16.7:1, 1 month. A complete response was defined as a complete and 5.6:1 for PBM cells and 60:1, 20:1, and 6.7:1 or 180:1, 60:1, disappearance of all known disease for at least 1 month. and 20:1 for PMN cells. Fifty microliters of RPMI-HS, RPMI-HS supplemented with IL-2 at a final concentration of Immunological Monitoring 100 units/ml (to stimulate PBM cells), or RPMI-HS supplemented with GM-CSF at a final concentration of 1 Ixg/ml (to Sample Times stimulate PMN cells) were added to effector cells and incubated Two baseline blood samples were obtained before treat- for 20-30 min at 37°C. Just prior to the addition of target cells ment from all patients. Blood samples to analyze the in vivo (LA-N-5 neuroblastoma cell line), 50 ILl chimeric antibody effects of GM-CSF followed by IL-2 were obtained on days 2, 14.18 were added to quadruplicate wells at a final concentration 6, 9, 13, 20, and 34 of treatment. To assess the effects of of 0.25 ~g/ml. Effector cells in media and in IL-2 or GM-CSF GM-CSF combined with IL-2, blood samples were obtained on alone were also plated to determine the effectors' ability to days 6, 9, 13, 16, and 20 (Fig. 1). mediate lysis of target cells in the absence of antibody. Target

cells were labeled with 250 i~Ci 51Cr in 0.2 ml RPMI-HS. Table 2 Patient characteristics Target cells were mixed every 30 min during labeling to keep Characteristics No. of patients the cells in suspension. After washing twice, 5 × 103 target cells Sex were added to wells containing effector cells and centrifuged at Male 20 200 × g for 5 rain. The plates were incubated at 37°C for 4 h, Female 14 and the supernatants were harvested using the Skatron Harvest- Age (yr) Mean 51.4 ing System (Skatron, McLean, VA). Percent cytotoxicity values Range 29-66 were calculated for each E:T ratio as: %cytotoxicity = 100 × Diagnosis [(experimental release - spontaneous release)/(maximum re- Renal 29 lease - spontaneous release)]. Melanoma 2 The maximum 51Cr release was measured by lysing target Breast 1 Transitional cell 1 cells with 2% centrimide solution (Sigma Chemical Co., St. Hepatoma 1 Louis, MO), and spontaneous release was measured with Prior treatment targets incubated in RPMI-HS alone. The results are ex- Surgery 14 Radiation therapy 7 pressed as lytic units in which 1 lytic unit is the number of Chemotherapy 4 effector cells necessary to achieve 20% lysis of 5 × 103 Hormonal therapy 1 targets with maximum lysis of 100% by centrimide treatment. Immunotherapy 5 Performance status 0 14 LAK Cell Functional Assays 1 20 To test PBM cells for LAK cell function, PBM cells were assayed for their ability to kill the NK-cell-resistant Daudi target. PBM cells in 50 pA RPMI-HS were plated in quadruplicate in 96-well, U-bottom microtiter plates at E:T ratios of 50:1, 16.7:1, and 5.6:1. RPMI-HS or RPMI-HS supplemented with HB8784 at a concentration of 2 ~xg/ml in carbonate buffer (pH IL-2 at a final concentration of 100 units/ml was added to 9.6) overnight at 4°C, then washed and blocked with 5% nonfat quadruplicate wells. Target and effector cells were handled as dry milk in PBS for 1 h at room temperature. After washing with described above. PBS/0.05% Tween 20, 100 ixl/well standards, controls, and test samples were incubated overnight at 4°C in a wet chamber. Surface Marker Analysis Following washing, biotinylated secondary antibody GL439 All fluorescence labeling was performed at 4°C in the dark was added, and the plate was incubated for 3 h at room tem- for 30 min. Patient PBM and PMN cell populations were char- perature and washed as above; alkaline phosphatase-conjugated acterized by incubating PBM and PMN cells in 100 lxl HBSS ExtrAvidin (Sigma) was added (1:5000 dilution in Tris:Tween), and 1% fetal bovine serum with FITC-conjugated or phyco- and plates were incubated for 1 h at room temperature. After erythrin-conjugated antibodies (Becton Dickinson, Mountain washing with PBS/Tween, 1 mg/ml substrate (p-nitrophenyl View, CA) at the recommended concentrations. The cells were phosphate; Sigma) in diethanolamine buffer (pH 9.8) was washed of unbound antibody by adding 2 ml cold PBS and 1% added. After a 30-min incubation at room temperature in the fetal bovine serum. One hundred microliters of propidium io- dark, the enzymatic reaction was stopped, and color intensity dide were added just prior to analysis on FACS to exclude dead was measured at 405 versus 490 nm using an ELISA reader cell populations from the analysis. The percentages of lympho- (EAR 400 AT; SLT Labinstruments, Salzburg, Austria), and the cytes, monocytes, and granulocytes making up the PBM cell concentration of soluble IL-R was determined using SOFT-2000 population were characterized by CD45 and CD14. The Fc software (SLT). Positive samples assayed previously were receptor (CD16) and NK marker (CD56) were examined on the population gated for lymphocytes. The lymphocyte and mono- included in each assay to provide internal controls, allowing cyte populations were examined for expression of MHC class II for comparison of soluble IL-2 results between different (HLA-DR) molecules and Fc receptors. In addition, the mono- assays. cyte population was examined for the Fc receptor (32.2). PMN cells were examined for Fc receptors I and III using antibodies Statistical Analysis (32.2 and Leu l la) and for the expression of MHC class II Assessments of immunological effects within treatment molecules (HLA-DR). groups were made using paired t tests. A comparison of effects between treatment groups were made using a two- Soluble IL-2R-a ELISA sample t test. A sandwich assay using two monoclonal anti-Tac antibodies, American Type Culture Collection clones HB8784 and GL439 from R. Robb (DuPont, Boston, MA), was developed. RESULTS This assay was standardized according to the commercially Thirty-four patients were enrolled on the protocol. Their available kit from T-Cell Diagnostics (Cambridge, MA). characteristics are shown in Table 2. Twenty-seven patients Briefly, 96-well, C-bottom polystyrene plates (MaxiSorp; Nunc, were evaluable for biological response assessment. Nine pa- Roskilde, Denmark) were coated with 150 pMwell antibody tients went off study prior to completion of one cycle of therapy,

Table 3 Number of courses with grade 3 toxicities in patients Table 4 Number of courses with grade 3 toxicities in patients receiving sequential GM-CSF followed by IL-2 receiving concurrent GM-CSF plus IL-2 GM-CSF dose (l~g/kg) 5 10 GM-CSF dose (l~g/kg) 2.5 2.5 5.0 IL-2 dose (million units/m2/day) 3 3 IL-2 dose (million units/me/day) 1.5 3.0 1.5 No. of courses administered 13 12 No. of courses administered 13 7 23 GM-CSF Hepatic Pain/reaction at injection site 0 1 Aspartate aminotransferase 0 1 1 Bone pain/myalgia 0 1 Bilirubin 1 1 0 Hepatic Cardiac Aspartate aminotransferase 0 0 Hypotension 0 1 1~ B ilirubi n 0 0 Arrhythmias 0 0 1a Cardiac Hematological Hypotension 0 3 WBC >60,000 0 1 1 Arrhythmias 0 0 WBC > 40,000 0 0 2 Hematological Eosinophils WBC >60,000 0 0 20,000-50,000 5 2 1 Eosinophils >50,000 0 0 1~ >10,000 with dose roods 0 1 Thrombocytopenia 1 0 2 and I a 20,000--50,000 0 1 Neurological 0 0 1 and 1~ >50,000 0 0 Other Thrombocytopenia 0 1 Renal 0 1 1 Neurological 0 0 Performance status 0 1 0 Other Fever >40°C 3 0 1 Performance status 0 0 Hypocalcemia Fever >40°C 3 1 Total 0 2 1 Hypocalcemia Corrected 0 0 0 Total 11 4 Diarrhea 0 0 1 Corrected 0 0 ~Grade 4 toxicities. Diarrhea 0 1 Renal 0 0 Gastrointestinal hemorrhage 0 0

probably due to the high WBC count with the eosinophilia, the patient was given 2 g hydroxyurea p.o. on day 19, with a six due to toxicity and three due to progressive disease. All resultant decrease in the WBC count to 46,000/mm 3 and 23,000/ patients were considered evaluable for toxicity and response. mm 3 over the next 2 days. Head MRI revealed multiple small lesions in the white matter in the front lobe bilaterally, with Clinical Results smaller lesions in the medial aspect of the right parietal lobe and right parietal temporal junction. Toxicities The patient's neurological findings resolved slowly. Ap- Toxicities secondary to treatment are shown in Tables 3 proximately 1 month later, he was seen in the clinic, where he and 4. The majority of toxicities were elevations in total WBC was found to have some residual deficits in his attention span or eosinophil count secondary to GM-CSF. Other nonhemato- and ability to perform math calculations, clumsiness of his fight logical toxicities included fevers, hypotension, diarrhea, in- hand, and some visual complaints. A computed tomographic creases in hepatic enzymes, and arrhythmias. scan of his chest showed an improvement in his pulmonary No dose-limiting (grade 4) toxicities of sequential GM- nodules. Two months following IL-2 and GM-CSF, neurologi- CSF and IL-2 (Table 3) were observed. Dose-limiting toxicities cal symptoms and complaints had completely resolved, with the of concurrent GM-CSF and IL-2 were primarily neurological exception of a visual-field defect, which was thought to be and cardiovascular (Table 4). related to an occipital lesion, and subtle fine motor deficiencies Grade 3 neurological toxicity was observed in patient 17, in his fight hand. Head MRI revealed no change in the ischemic who received concurrent GM-CSF (5 i~g/kg) and IL-2 (1.5 lesions. The patient refused all but palliative treatment subse- million units/me/day). The patient was a 48-year-old male who quent to GM-CSF and IL-2, and he developed a spinal cord began experiencing forgetfulness, confusion, and lethargy on compression secondary to bone metastases 11 months later. day 18 of his treatment. At that time, his WBC count was noted Another patient (patient 33) on the concurrent arm of 5 to be 75,800/mm 3 with an absolute eosinophil count of 57,000/ txg/kg GM-CSF and 1.5 million units/m 2 IL-2 developed grade mm 3 and a platelet count of 67,000/mm 3. The GM-CSF was 4 neurological toxicity. This 53-year-old male who had renal held, and IL-2 was stopped. Nonetheless, the patient's neuro- cell carcinoma with nodal and adrenal metastases received six logical symptoms continued to worsen, and on day 19, he was doses of GM-CSF during his second week of IL-2 therapy (two experiencing dizziness, lethargy, confusion, headaches, halluci- doses at 5 ~g/kg, two at 2.5 p~g/kg, and two at 1.25 txg/kg, with nations, short-term memory loss, right-hand motor difficulties, dose reductions due to grade 3 diarrhea and eosinophilia). On and intermittent mild, expressive aphasia. The maximum total day 18 of his therapy, the patient became confused and lethar- WBC count was 99,400/mm 3, and the maximum total eosino- gic, with weakness in all extremities, especially the right arm, phil count was 73,000/mm 3 on day 19, 31 h after the last dose and he developed total vision loss. Treatment with IL-2 and of GM-CSF. As it was thought that these symptoms were GM-CSF was stopped. This patient's maximum total WBC and

Table 5 Antitumor effects of GM-CSF and IL-2 by groups and dose level Sequential, Concurrent, GM-CSF ~ IL-2 GM-CSF + IL-2 Total GM-CSF dose (p,g/kg) 5 10 2.5 2.5 5.0 IL-2 dose (million units/m2/day) 3 3 1.5 3.0 1.5 No. of patients per dose level 6 6 6 4 12 34 No. of patients with metastatic renal cell 2 5 3 1 4 15 carcinoma and pulmonary metastases No. with >--50% reduction in pulmonary 0 1 1 0 3 5 metastases only No. of responses 0 0 0 0 0 0

A B 30 30 .E 25 25

i= 2o 20 Fig. 2 Fold increase from base- 9 line in soluble IL-2Rs in patients ~ 15 15 receiving sequential (A) or concurrent (B) GM-CSF and IL-2. The abscissa shows the day samples ~ 10 10 were obtained in relation to IL-2 o e- and GM-CSF administration.

"O 5 O 0 0 6 13 20 34 day 6 13 20 day IL-2 IL-2 == == GM-CSF t i GM-CSF [ I

eosinophil counts were 43,800/mm 3 and 24,000/mm 3, respec- remained normal. He remained in the intensive care unit tively, on day 18. Head MRI showed multiple new foci of overnight and did not experience any more problems; he was abnormalities in the cerebrum consistent with thrombosis or not retreated. multiple small vessel emboli; a lumbar puncture and transesoph- ageal echocardiogram were negative. The patient's neurological Antitumor Effects signs began improving on day 19, and vision was totally resolved None of the 34 patients had a >50% shrinkage of the sum within 24 h. The majority of his symptoms resolved gradually over of all disease. Fifteen patients had stable disease at the end of the next 3 months, and he was left with a residual mild weakness two courses of therapy, and 19 patients developed progressive in the fight arm. Neither of the two patients experiencing neuro- disease over the same time interval. logical toxicity had any predisposing risk factors or any evidence Although no patient had a partial response, as defined by a on head MRI of predisposing atherosclerotic disease. >50% shrinkage of the sum of all disease, the number of patients Grade 4 cardiac toxicity was observed in patient 15 while with metastatic renal cell carcinoma who had reductions in their receiving concurrent GM-CSF (5 p,g/kg) and IL-2 (1.5 million 2 pulmonary disease is of note (Table 5). Overall, 15 patients had units/m/day). The patient was a 65-year-old male with renal cell carcinoma who had a prior history of atrial flutter and was metastatic renal cell carcinoma with pulmonary metastases; of taking digoxin; however, he had passed an exercise thallium test these, 11 patients had their primary renal tumors intact. Of the 15 prior to going on study. On day 10, the patient began experi- patients with metastatic renal cell carcinoma, 5 (33%) had >50% encing hypotension and tachycardia. He was found to have an reduction in the size of their pulmonary metastases, although none elevated creatinine level (2.9 mg/dl) and hyperkalemia (6.0). In of them qualified as a partial responder because of a lack of addition, an electrocardiogram showed 2:1 atrial flutter. The patient measurable response in the kidneys or other sites of metastases. was transferred to the cardiac intensive care unit for monitoring, Also noteworthy was that 4 of 8 patients with renal cell carcinoma where he received i.v. fluids and Kayexalet. The patient and pulmonary metastases on the GM-CSF and concurrent IL-2 spontaneously converted back to normal sinus rhythm, and regimen had >50% reduction in the size of their pulmonary his serum potassium returned to normal; his cardiac enzymes metastases.

Table 6 Significant differences in immunological parameters between elevated through administration of GM-CSF. Although LAK patients in the sequential and concurrent groups following cycle 1 activity seemed to be declining from the peak level noted at day Change in % of cells positive from 6, these differences were not statistically significant (P = 0.14). baseline to completion of IL-2 The LA-N-5 neuroblastoma target was used to measure Sequential, Concurrent, ADCC activity with the chimeric 14.18 human and mouse Assay day 34 day 20 P antibody, which recognizes the ganglioside GD2 on the neuro- Surface markers blastoma target. Treatment for 12 days with GM-CSF in the se- Lymphocyte CDI6 34 22 0.02 quential group demonstrated that there was a slight down-modula- Lymphocyte CD56 50 28 0.0009 tion of ADCC activity from baseline to days 6 and 13. Initiation of Monocyte CD16 28 - 1 <0.001 IL-2 induced a significant increase in ADCC noted at days 20 and Change in lyric units from 34. In the concurrent group, ADCC increased after an initial week baseline to completion of IL-2 of continuous infusion of IL-2 and remained elevated following 2 weeks of IL-2 combined With GM-CSF. Sequential, Concurrent, Assay day 34 day 20 P When comparing LAK activity in the two groups following 3 weeks of IL-2, LAK activity in patients receiving GM-CSF LAK activity simultaneously with IL-2 was significantly lower than that seen Daudi, medium 14 2 0.03 in patients receiving the sequential treatment (Table 6). This was Daudi, IL-2 233 62 0.03 LA-N-5, medium 78 5 0.08 noted against the Daudi target with PBM cells tested in medium LA-N-5, IE-2 295 28 0.012 (P = 0.03) and PBM cells tested on both Daudi and LA-N-5 ADCC activity targets in IL-2 (P = .03 and 0.012, respectively) (Table 6). LA-N-5, antibody 480 306 0.5 There was no significant difference in the augmentation of LA-N-5, + IL-2 735 552 0.56 ADCC activity induced with the two treatment regimens.

Surface Marker Analysis Immunological Results Lymphocyte, monocyte, and PMN cell (granulocyte) populations were assessed for phenotypic changes and expression of Soluble IL-2R activation antigens (Fig. 3). Increases in the soluble IL-2R a chain of 3.5- and 6.3-fold Effects of GM-CSF Alone. As seen in Fig. 4, GM-CSF were observed following 6 and 12 days, respectively, of treat- alone did not affect the expression of CD16, CD56, or HLA-DR ment with GM-CSF in the 12 patients in the sequential arm (P on lymphocytes. However, CD 16 was significantly down-mod- < 0.0001; Fig. 2). These data demonstrate that GM-CSF alone ulated in both monocytes and PMN cells. In addition, GM-CSF induces a significant rise in this receptor level. The soluble induced an increase in the monocyte population expressing the IL-2R continued to rise with subsequent continuous infusion of HLA class II (DR) antigen. IL-2 such that at day 34, there was a 28.6-fold increase from Effects of IL-2 with and without GM-CSF. Lympho- baseline (P < 0.0001). cytes. As noted in previous studies, the percentage of lympho- Soluble IL-2R increased 16.2-fold in patients in the se- cytes expressing CD56 and CD16 increased following IL-2 quential group when measured 24 h after completion of the first administration in patients in both the concurrent and sequential 96-h continuous infusion of IL-2. This change was significantly treatment groups. In addition, the percentage of lymphocytes greater than the 6.6-fold increase following 1 week of IL-2 alone expressing HLA-DR was also increased in both groups during in patients in the concurrent arm. These results indicate that the time of IL-2 administration. GM-CSF given prior to one 96-h continuous infusion of IL-2 Monocytes. IL-2 alone induced a significant increase in results in a higher overall release of soluble IL-2R than does CD16 expression on monocytes in patients in the sequential group. treatment with IL-2 alone. Note, however, that the patients No increase in the percentage of monocytes positive for CD16 was receiving the sequential treatment were treated with IL-2 at 3 × observed when IL-2 was given concurrently with GM-CSF. 106 units/m2, whereas only 4 of the patients in the concurrent Although treatment with GM-CSF alone induced increased group received this dose. The overall change following 3 weeks expression of HLA-DR on monocytes, this returned to baseline of IL-2 for the two groups was not significantly different, with during the IL-2 infusion in patients receiving sequential treat- a 28.6-fold increase on day 34 for the sequential arm and a ment. There was no increase in monocyte DR expression in the 28.5-fold increase at day 20 for the concurrent arm (P = 0.11). concurrent patients following 1 week of IL-2. Increased DR expression was noted following the second week of IL-2; how- LAK and ADCC Activity ever, at the end of the third week of IL-2, the level was not Twelve days of GM-CSF alone did not induce LAK activ- significantly above the pretreatment level. ity in patients in the sequential arm. Following 12 days of PMN Cells. IL-2 alone induced a significant decrease in GM-CSF, treatment for 96 h with IL-2 alone induced LAK CD 16 expression on PMN cells in patients in both the sequential activity when measured in the presence of IL-2 during the 4-h and concurrent groups. This level continued to fall during con- 51Cr release assay. LAK activity was also detected in patients in current IL-2 and GM-CSF treatment. No significant changes the concurrent arm following a 96-h continuous infusion of IL-2 were noted in cells expressing the Fc I receptor. prior to initiation of GM-CSF (day 6). This activity remained The expression of HLA-DR on PMN cells was examined

A. Lymphocyte CD56 100 --~- CD16 80 DR 60

~,sO 40

I I I I I I I I I B. Monocyte 100 -

m 80 ---{3 m m 60 Fig. 3 Percentage of lymphocytes, (A) monocytes (B), or PMN cells (C) express- o 40 ing CD56, CD16, or HLA-DR in patients

I receiving sequential or concurrent GM- 20 0..""0'"'-..0.__<3 CSF and IL-2. The abscissa shows the day PBM specimens were obtained in relation "<3- - - "~ to IL-2 and GM-CSF administration. 0 I I I I I I I I t C. PMN 100 q (3,% % % 80 - ~ % % % 60 % ~%%% 40 b % "0 20 "O

0 I I I I I 0 6 13 20 34 0 6 13 20 day Ik-2 ~ ~ IL-2 ~ ~ ~ GM-CSF I i GM-CSF I I Sequential Concurrent

only in patients treated in the concurrent arm. Resting PMN in this group. This increased expression of CDI6 and CD56 and cells did not express HLA-DR. Following 1 week of IL-2 alone, LAK cytotoxicity in the sequential group also may be due to the there was a significant increase, with 10% of the PMN cells higher dose of IL-2 received by these patients, making any expressing HLA-DR. With the combination of IL-2 and GM- direct comparison difficult. However, there was no significant CSF, this increased to 34% at day 13 and was at 23% on day 20. difference in the ADCC activity when comparing the sequential and concurrent arms. Because only a fraction of cells mediating Comparative Analyses between Groups LAK activity can mediate ADCC (30), this result may suggest Comparisons of changes between immunological parame- differential activation of the LAK and ADCC effectors. ters at the end of treatment in the concurrent group compared Significant differences in immunological parameters between with the sequential group are shown in Table 6. The increase in the 18 patients in the concurrent treatment arm who showed no CD16 and CD56 expression on lymphocytes and CD16 expres- antitumor effects from treatment compared with the 4 patients who sion on monocytes were significantly greater in the patients had at least 50% reductions in their pulmonary nodules (responding receiving the sequential regimen. In addition, LAK cytotoxicity patients) are shown in Fig. 4 (P < 0.05). Although both responding against the Daudi and LA-N-5 targets was significantly greater and nonresponding patients had significant increases in lymphocyte

GM-CSF and M-CSF as monocyte activators when the gene is 60 "A. Lymphocyte CD16 introduced directly into cells are ongoing (32-34). 50 Clinical studies with M-CSF indicate that this molecule has 40 biological and immunological activity. A Phase I study of M- 30 CSF demonstrated that M-CSF increased the percentage of peripheral blood monocytes expressing the CD16 low-affinity 20 Fc-',/ receptor for aggregated immunoglobulin and the CD14 10 receptor (28). Systemic M-CSF induces peripheral blood mono- 0 cytosis and thrombocytopenia (28, 35-37). The data regarding the effectiveness of GM-CSF as a mono- 60 B. Lymphocyte CD56 cyte activator in humans are somewhat limited. A Phase I study demonstrated that retreatment with GM-CSF led to a statisti- 40 cally significant enhancement in direct monocyte cytotoxicity against HT29 cells, but without increasing serum TNF-oL or 30 IL-l[3 levels and without any consistent in vitro induction of 20 TNF-oL or IL-l[3 (26). One Phase I study of GM-CSF demon- 10 strated activation of monocyte-mediated tumorlytic properties in only one of seven patients and failed to demonstrate stimulation 0 of IL-1 or TNF by peripheral monocytes (27). However, patients received variable amounts of GM-CSF, and GM-CSF was ad- 20 C. Absolute Eosinophil Count ministered as a continuous infusion. In another Phase I study using the 17-1A MoAb, ADCC and the number of Fc receptor- bearing monocytes increased during therapy with GM-CSF (38). GM-CSF currently has practical advantages over M-CSF in that it is commercially available, and its toxicities and schedule 5 --m- Responding Patients of administration for hematopoietic stimulation are well defined. ÷ Nonresponding Patients Therefore, the objective of this protocol was to determine the biological and immunological effects of this molecule on circu- 0 0 16 20 clay lating blood lymphocytes and peripheral neutrophils and mono- IL-2 1 1 I cytes alone and in combination with IL-2. Simultaneous activa- GM-CSF I I tion of multiple immune system effector cells has the theoretical advantage of enhanced direct antitumor cytotoxicity as well as Fig. 4 Significant differences in percentage of lymphocytes expressing CD16 (A) or CD56 (B), and absolute eosinophil counts (C) in four renal activation of effectors capable of mediating ADCC for eventual cell carcinoma patients with a >50% reduction in their pulmonary use with tumor-specific MoAbs. metastases (D) and in 18 patients treated on the concurrent arm with In our study, we observed an increase in soluble IL-2Rs with GM-CSF and IL-2 (0; P < 0.05). The abscissa shows the day samples administration of GM-CSF, IL-2, or the combination. Expected were obtained in relation to IL-2 and GM-CSF administration. effects of IL-2 on LAK activity, ADCC, and lymphocyte activation were also observed and did not seem to be qualitatively different following pretreatment with GM-CSF when compared with results CD16 and CD56 expression, the responding patients did not have obtained in our laboratory from other clinical IL-2 trials (7-9, 12, increases as great at day 20. The eosinophil counts in the respond- 15). Interestingly, these stimulatory effects on circulating lympho- ing patients were significantly greater at day 16. cytes were decreased with concurrent GM-CSF treatment, with a significantly lower level of LAK activity in patients receiving DISCUSSION simultaneous treatment compared with sequential GM-CSF. GM- Preclinical studies indicate that the CSFs, which activate CSF increased expression of HLA-DR on monocytes; this was not monocytes, macrophages, and granulocytes, have immunologi- maintained with concurrent or subsequent IL-2 therapy. cal effects that may result in antitumor activity. M-CSF en- Intriguingly, surface expression of CD16 on monocytes hances ADCC in cynomolgus monkey, human, and murine and PMN cells decreased during GM-CSF treatment, whereas monocytes (18, 19, 25, 31). GM-CSF induces macrophage tu- the morphology of the circulating PMN cells showed signs of moricidal activity more effectively than IFN-'? (21). GM-CSF activation (i.e., toxic granulations) clearly while on therapy. enhances expression of TNF, IL-1, and IL-6 on both monocytes Although GM-CSF may cause down-regulation of CD16 in and alveolar macrophages (22). Peritoneal macrophages from neutrophils (39), GM-CSF is also known to cause more rapid mice inoculated with GM-CSF produced greater levels of mem- production and release of PMN precursors from the bone mar- brane-bound and secreted IL-2, exhibited greater levels of Fc row, and Fc receptors are expressed late in PMN cell maturation receptor, and had an increased ability to phagocytose opsonized (40). Thus, we postulate that the increase in circulating PMN sheep RBC in vitro (23). Human monocytes cultured in human cell and monocyte numbers with the decreased expression of Fc M-CSF ingested up to 100% of fluorochrome-labeled melanoma receptors on these cells represents two concurrent changes from and neuroblastoma target cells in the presence of a MoAb (19). the baseline physiological status: first, the appearance in the Current preclinical and clinical studies exploring the use of blood of immature PMN cells and monocytes that have not yet

expressed their Fc receptors; and second, the more rapid egress form of thrombomodulin potently. Endothelial and endocardial from the circulation of these activated PMN cells and mono- cells maintain an anticoagulant surface, in part through the expres- cytes due to their stimulation by GM-CSF and due to endothelial sion of thrombomodulin on their surfaces (48-50); therefore, inhi- activation by the associated cytokines released in response to bition of thrombomodulin by hypereosinophilia may lead to throm- IL-2 and GM-CSF. This should result in a greater number of boembolic cerebrovascular accidents. It remains unclear whether activated cells that are maturing and expressing CD 16 within the this mechanism or other possible mechanisms are responsible for tissues. This could not be tested directly in this study and empha- the observed neurological toxicity in this study. Because the neu- sizes difficulties in evaluating changes in cell behavior in tissue rological toxicity was not related clearly to the dose of GM-CSF or compartments based solely on measurement of serum or peripheral the total WBC or eosinophil count, we recommend the Phase II blood cells from patients undergoing biological therapy. starting dose of 1.25 ixg/kg/day GM-CSF with 1.5 million Roche Interestingly, several studies indicate that GM-CSF may ac- units/mZ/day IL-2, with subsequent escalation of GM-CSF to 2.5 tivate lymphocytes either directly or indirectly in vivo. An increase txg/kg/day after careful observation for toxicities. in the absolute lymphocyte count, serum IL-2R, serum CD8 levels, We observed an increase in soluble IL-2Rs with adminis- and CD25 ÷ and CD4 ÷ lymphocytes was observed in patients with tration of GM-CSF, IL-2, or the combination. IL-2 enhanced malignant lymphoma receiving GM-CSF (41). GM-CSF potenti- LAK activity, ADCC, and lymphocyte activation with and with- ated the long-term growth of nonactivated T cells in the presence of out pretreatment with GM-CSF; LAK activity was lower in IL-2 and short-term proliferation of lectin-stimulated T cells (42) patients receiving concurrent GM-CSF and IL-2 treatment. GM- and augmented low-dose IL-2 induction from murine splenocytes CSF increased the expression of HLA-DR on monocytes; this was (43). In our study, we did not observe any effects of GM-CSF on not maintained with concurrent or subsequent IL-2 therapy. GM- the induction of LAK or ADCC activity or on lymphocyte activa- CSF decreased the surface expression of CD16 on monocytes and tion as assessed by increases in CD56 +, CD16 +, or HLA-DR- PMN cells. The antitumor responses observed in four of eight renal positive lymphocytes in the peripheral blood. An increase in solu- cell carcinoma patients with pulmonary metastases treated with ble IL-2Rs was observed, indicating that GM-CSF may have an concurrent GM-CSF and IL-2 are provocative and warrant further effect on lymphocyte activation, although monocytes also may study. Although the relationship between the neurological toxicity have been the source of the soluble IL-2R. and the dose and schedule of the combination, if any, is unclear, the Four of 5 patients who had 50% or greater reduction in recommended initial Phase II dose and schedule is 1.25 p~g/kg/day their pulmonary metastases were receiving concurrent treatment GM-CSF, given concurrently with 1.5 million Roche units/m2/day with GM-CSF and IL-2. The percentage of lymphocytes posi- IL-2, with subsequent escalation of GM-CSF to 2.5 Ixg/kg/day after tive for CD16 and CD56 in these patients was not as great as the careful observation for toxicities. increase noted in the 18 nonresponding patients. Soluble IL-2Rs and ADCC and LAK activity were not decreased in this sub- group of patients, indicating that immunological activation still ACKNOWLEDGMENTS occurred. Although the numbers of patients are small, it is We thank Tammy Rivest and Marilyn Larson for their help with tempting to speculate that CD16 + and CD56 + lymphocytes data management and protocol development; Jean Surfus, Kathy Schell, were less prevalent in the circulation because of greater egress and Jacek Gan for conducting the laboratory studies; and Kathy Edge for into tissues, possibly including sites of pulmonary disease. preparation of the manuscript. Clinically, the treatment was well tolerated, with side effects typical of those commonly seen with IL-2 or GM-CSF. REFERENCES The etiology of the two grade 3 or 4 neurological toxicities is 1. Borden, E., and Sondel, P. Lymphokines and cytokines as cancer treatment: immunotherapy realized. Cancer (Phila.), 65: 800-814, unclear. Although we have not seen similar toxicity in more than 1990. 150 patients treated with IL-2 at our institution, 6 other patients 2. Rosenberg, S. New approaches to the immunotherapy of cancer receiving IL-2 at other institutions, none of whom were also using interleukin-2. Ann. Intern. Med., 108: 853-864, 1988. receiving GM-CSF, have been described as having similar 3. Whittington, R., and Faulds, D. Interleukin-2, a review of its phar- toxicities; patients had confusion, delirium, coma, ataxia, macologic properties and therapeutic use in patients with cancer. Drugs, hemiparesis, and cortical blindness, with multiple cerebral 46: 446-514, 1993. lesions on MRI (44). In this study, the neurotoxicity was not 4. Anderson, P. M., and Sorenson, M. A. Effects of route and formu- uniformly the result of the elevated WBC count and eosino- lation on clinical pharmacokinetics of interleukin-2. Clin. Pharmacoki- philia, because these were markedly elevated in only one of net., 27: 19-31, 1994. the two patients with neurotoxicity. Similarly, they were not 5. Rosenberg, S. A., Yang, J. C., Topalian, S. L., Schwartzentruber, D. J., Weber, J. S., Parkinson, D. R., Seipp, C. A., Einhorn, J. H., and clearly related to the dose of GM-CSF, because the second White, D. E. Treatment of 283 consecutive patients with metastatic patient had received two doses of 5 i~g/kg, two doses of 2.5 melanoma or renal cell cancer using high dose bolus interleukin-2. Ixg/kg, and two doses of 1.25 lxg/kg GM-CSF. JAMA, 271: 907-913, 1994. Although the etiology of this neurotoxicity is not clear, 6. Cheever, M., Greenberg, P., Fefer, A., and Gillis, S. Augmentation of thromboembolic cerebrovascular accidents have been described the antitumor therapeutic efficacy of long-term cultured T lymphocytes in patients with hypereosinophilia (45, 46). Most of the de- by in vitro administration of purified interleukin-2. J. Exp. Med., 155: 968-980, 1982. scribed patients had endocardial disease associated with idio- 7. Hank, J., Kohler, P., Hillman, G., Rosenthal, N., Moore, K., Storer, pathic hypereosinophilic syndrome. Slungaard et al. (47) B., Minkoff, D., Bradshaw, J., Bechhofer, R., and Sondel, P. In vivo showed that eosinophilic cationic proteins, in particular major induction of the lymphokine-activated killer phenomenon: interleukin basic protein, inhibit anticoagulant activities of the glycosylated 2-dependent human non-major histocompatibility complex-restricted

cytotoxicity generated in vivo during administration of human recom- 24. Charak, B., Agah, R., and Mazumder, A. Granulocyte-macrophage binant interleukin 2. Cancer Res., 48: 1965-1971, 1988. colony-stimulating factor-induced antibody-dependent cellular cytotox- 8. Sosman, J. A., Kohler, P. C., Hank, J., Moore, K. H., Bechhofer, R., icity in bone marrow macrophages: application in bone marrow trans- Storer, B., and Sondel, P. M. Repetitive weekly cycles of interleukin-2. plantation. Blood, 81: 3474-3479, 1993. II. Clinical and immunological effects of dose, schedule, and addition of 25. Sanda, M., Bolton, E., MuM, J. J., and Rosenberg, S. A. In vivo indomethacin. J. Natl. Cancer Inst., 80: 1451-1461, 1988. administration of recombinant macrophage colony-stimulating factor 9. Hank, J., Robinson, R., Surfus, J., Mueller, B., Reisfeld, R., Cheung, induces macrophage-mediated antibody-dependent cytotoxicity of tu- N., and Sondel, P. Augmentation of antibody dependent cell mediated mor cells. J. Immunother., 12: 132-137, 1992. cytotoxicity following in vivo therapy with recombinant interleukin 2. 26. Chachoua, A., Oratz, R., Hoogmoed, R., Caron, D., Peace, D., Cancer Res., 50: 5234-5239, 1990. Liebes, L., Blum, R., and Vilcek, J. Monocyte activation following 10. Smith, K., Lachman, L., Oppenheim, J., and Favata, M. The func- systemic administration of granulocyte-macrophage colony-stimulating tional relationship of the interleukins. J. Exp. Med., 154: 1551-1556, factor. J. Imunother., 15: 217-224, 1994. 1980. 27. Kleinerman, E., Knowles, R., Lachman, L., and Gutterman, J. 11. Rosenberg, S. A. The immunotherapy and gene therapy of cancer. Effect of recombinant granulocyte/macrophage colony-stimulating fac- J. Clin. Oncol., 10: 180-199, 1992. tor on human monocyte activity in vitro and following intravenous 12. Schiller, J., Hank, J., Storer, B., Borchert, A., Moore, K., Albertini, M., administration. Cancer Res., 48: 2604-2609, 1988. Bechhofer, R., Wesley, O., Brown, R., Bastin, A., and Sondel, P. A direct 28. Weiner, L., Li, W., Holmes, M., Catalono, R., Dovnarsky, M., comparison of immunological and clinical effects of interleukin 2 with and Padavic, K., and Alpaugh, R. Phase I trial of recombinant macrophage without interferon-a in humans. Cancer Res., 53: 1286-1292, 1993. colony-stimulating factor and recombinant ~/interferon: toxicity, mono- 13. Sznol, M., Clark, J., Smith, J., Steis, R., Urba, W., Rubinstein, L., cytosis, and clinical effects. Cancer Res., 54: 4084-4090, 1994. VanderMolen, L., Janik, J., Sharfman, W., Fenton, R., Creekmore, S., 29. Gearing, A. J. H., and Thorpe, R. The international standard for Kremers, R., Conlon, K., Hursey, J., Beveridge, J., and Longo, D. Pilot human interleukin-2: calibration by international collaborative study. J. study of interleukin-2 and lymphokine-activated killer cells combined Immunol. Methods, 114: 3-9, 1988. with immunomodulatory doses of chemotherapy and sequenced with 30. Weil-Hillman, G., Fisch, P., Prieve, A. F., Sosman, J. A., Hank, interferon a-2a in patients with metastatic melanoma and renal cell J. A., and Sondel, P. M. Lymphokine-activated killer activity induced by carcinoma. J. Natl. Cancer Inst., 84: 929-937, 1992. in vivo interleukin 2 therapy: predominant role for lymphocytes with 14. Budd, G. T., Osgood, B., Barna, B., and Bukowski, R. M. Phase I increased expression of CD2 and Leul9 antigens but negative expres- clinical trial of interleukin 2 and a-interferon: toxicity and immunologic sion of CD16 antigens. Cancer Res., 49: 3680-3688, 1989. effects. Cancer Res., 49: 6432-6436, 1989. 31. Bock, S., Cameron, R., Kragel, P., Mule, J., and Rosenberg, S. Biological and antitumor effects of recombinant human macrophage 15. Hank, J. A., Albertini, M., Wesly, O. H., Schiller, J. H., Borchert, colony-stimulating factor in mice. Cancer Res., 51: 2649-2654, 1991. A., Moore, K., Bechhofer, R., Storer, B., Gan, J., Gambacorti, C., Sosman, J., and Sondel, P. M. Clinical and immunological effects of 32. Walsh, P., Dorner, A., Duke, R. C., Su, L. J., and Glode, L. M. treatment with murine anti-CD3 monoclonal antibody along with inter- Macrophage colony-stimulating factor complementary DNA: a candi- leukin 2 in patients with cancer. Clin. Cancer Res., 1: 481-491, 1995. date for gene therapy in metastatic melanoma. J. Natl. Cancer Inst., 87: 809-816, 1995. 16. Sosman, J. A., Weiss, G. R., Margolin, K. A., Aronson, F. R., Sznol, M., Atkins, M. B., O'Boyle, K., Fisher, R. I., Boldt, D. H., Doroshow, 33. Lee, S. S., Eisenlohr, L. C., McCue, P. A., Mastrangelo, M. J., Fink, E., and Lattime, E. C. In vivo gene therapy of murine tumors using J., Ernest, M. L., Fisher, S. G., Mier, J., Vachino, G., and Caliendo, G. recombinant vaccinia virus encoding GM-CSF. Proc. Am. Assoc. Can- Phase IB clinical trial of anti-CD3 followed by high-dose bolus inter- cer Res., 36: 248, 1995. leukin-2 in patients with metastatic melanoma and advanced renal cell carcinoma: clinical and immunologic effects. J. Clin. Oncol., 11: 1496- 34. Dranoff, G., Jaffee, E., Lazenby, A., Golumbek, P., Levitsky, H., 1505, 1993. Brose, K., Jackson, V., Hamada, H., Pardoll, D., and Mulligan, R. C. Vaccination with irradiated tumor cells engineered to secrete murine 17. Metcalf, D. The molecular biology and functions of the granulo- granulocyte-macrophage colony stimulating factor stimulates potent, cyte-macrophage colony stimulating factors. Blood 67: 35-43, 1987. specific, and long-lasting antitumor immunity. Proc. Natl. Acad. Sci. 18. Munn, D., Garnick, M., and Cheung, N. Effects of parenteral USA, 90: 3539-3543, 1993. recombinant human macrophage colony-stimulating factor on monocyte 35. Cole, D., Sanda, M., Yang, J., Schwartzentruber, D., Weber, J., number, phenotype, and antitumor cytotoxicity in nonhuman primates. Ettinghausen, S., Pockaj, B., Kim, H., Levin, R., Pogrebniak, H., Blood, 75: 2042-2048, 1990. Balkissoon, J., Fenton, R., DeBarge, L., Kaye, J., Rosenberg, S., and 19. Munn, D., and Cheung, N-K. Phagocytosis of tumor cells by human Parkinson, D. Phase I trial of recombinant human macrophage colony- monocytes cultured in recombinant macrophage colony stimulating stimulating factor administered by continuous intravenous infusion in factor. J. Exp. Med., 172: 231-237, 1990. patients with metastatic cancer. J. Natl. Cancer Inst., 86: 39-45, 1994. 20. Munn, D., and Cheung, N. Antibody-dependent antitumor cytotox- 36. Sanda, M., Yang, J., Topalian, S., Groves, E., Childs, A., Belfort, icity by human monocytes cultured with recombinant macrophage col- R., Jr., de Smet, M., Schwartzentruber, D., White, D., Lotze, M., and ony-stimulating factor. J. Exp. Med., 170: 511-526, 1989. Rosenberg, S. Intravenous administration of recombinant human mac- 21. Grabstein, K., Urdal, D., Tushinski, R., Mochizuki, D., Price, V., rophage colony-stimulating factor to patients with metastatic cancer: a Cantrell, M., Gillis, S., and Conlon, P. Induction of macrophage tumori- phase I study. J. Clin. Oncol., 10: 1643-1649, 1992. cidal activity by granulocyte-macrophage colony stimulating factor. 37. Zamkoff, K., Hudson, J., Groves, E., Childs, A., Konrad, M., and Science (Washington DC), 232: 506-508, 1986. Rudolph, A. A phase I trial of recombinant human macrophage colony- 22. Thomassen, M., Ahmad, M., Barna, B., Antal, J., Wiedmann, H., stimulating factor by rapid intravenous infusion in patients with refrac- Meeker, D., Klein, J., Bauer, L., Gibson, V., Andresen, S., and tory malignancy. J. Immunother., 11: 103-110, 1992. Bukowski, R. Induction of cytokine messenger RNA and secretion in 38. Ragnhammar, P., Magnusson, I., Masucci, G., and Mellstedt, H. alveolar macrophages and blood monocytes from patients with lung The therapeutic use of the unconjugated monoclonal antibodies (MAb) cancer receiving granulocyte-macrophage colony-stimulating factor. 17-1A in combination with GM-CSF in the treatment of colorectal Cancer Res., 51: 857-862, 1991. carcinoma (CRC). Med. Oncol. Tumor Pharmacother., 10: 61-70, 1993. 23. Morrissey, P., Bressler, L., Charrier, K., and Alpert, A. Response of 39. Kerst, J. M., de Haas, M., van der Schoot, C. E., Slaper-Cortenbach, resident murine peritoneal macrophages to in vivo administration of I. C., Kleijer, M., von dem Borne, A. E., and van Oers, W. H. Recom- granulocyte-macrophage colony-stimulating factor. J. Immunol., 140: binant granulocyte colony-stimulating factor administration to healthy 1910-1915, 1988. volunteers: induction of immunophenotypically and functionally altered

neutrophils via an effect on myeloid progenitor cells. Blood, 82: 3265- 45. Moore, P. M., Harley, J. B., and Fauci, A. S. Neurologic dysfunc- 3272, 1993. tion in the idiopathic hypereosinophilic syndrome. Ann. Intern. Med., 40. Valerius, T., Repp, R., deWit, T. P., Berthold, S., Platzer, E., 102: 109-114, 1985. Kalden, J. R., Gramatzki, M., and van de Winkel, J. G. Involvement of 46. Gardner-Thorpe, G., Harriman, D. G., Parsons, M., and Rudge, P. the high-affinity receptor for IgG (Fc gamma RI; CD64) in enhanced Loffier's eosinophilic endocarditis with Balint's syndrome. Q. J. Med. tumor cell cytotoxicity of neutrophils during granulocyte colony-stim- 40: 249-260, 1971. ulating factor therapy. Blood, 82: 931-939, 1993. 47. Slungaard, A., Vercellotti, G. M., Tran, T., Gleich, G. J., and 41. Ho, A. D., Haas, R., Wulf, G., Knauf, W., Ehrhardt, R., Heilig, B., Key, N. S. Eosinophil cationic granule proteins impair thrombo- K6rbling, M., Schulz, G., and Hunstein, W. Activation of lymphocytes modulin function. A potential mechanism for thromboembolism in induced by recombinant human granulocyte-macrophage colony-stimulat- hypereosinophilic heart disease. J. Clin. Invest., 91: 1721-1730, ing factor in patients with malignant lymphoma. Blood, 75:203-212, 1990. 1993. 42. Santoli, D., Clark, S. C., Kreider, B. L., Maslin, P. A., and Rovera G. Amplification of IL-2-driven T cell proliferation by recombinant 48. Dittman, W. A., and Majerns, P. W. Structure and function of human IL-3 and granulocyte-macrophage colony-stimulating factor. J. thrombomodulin: a natural anticoagulant. Blood 75: 329-336, 1990. Immunol., 141: 519-526, 1988. 49. Esmon, C. T. The roles of protein C and thrombomodulin in the 43. Stewart-Akers, A. M., Cairns, J. S., Tweardy, D. J., and McCarthy, regulation of blood coagulation. J. Biol. Chem., 264: 4743-4746, S. A. Effect of granulocyte-macrophage colony-stimulating factor on 1989. lymphokine-activated killer cell induction. Blood, 81:2671-2678, 1993. 50. Maruyama, I., Bell, C. E., and Majerus, P. W. Thrombomodulin is 44. Karp, B., Yang, J., and Khorsand, M. Multiple cerebral lesions found on endothelium of arteries, veins, capillaries, and lymphatics and complicating therapy with interleukin-2 (Abstract). Neurology, 45 syncytiotrophoblast of human placenta. J. Cell Biol. 101: 363-371, (Suppl. 4): A400, 1995. 1985.

Downloaded from clincancerres.aacrjournals.org on September 23, 2021. © 1996 American Association for Cancer Research. Clinical and immunological effects of granulocyte-macrophage colony-stimulating factor coadministered with interleukin 2: a phase IB study.

J H Schiller, J A Hank, M Khorsand, et al.

Clin Cancer Res 1996;2:319-330.

Updated version Access the most recent version of this article at: http://clincancerres.aacrjournals.org/content/2/2/319

E-mail alerts Sign up to receive free email-alerts related to this article or journal.

Reprints and To order reprints of this article or to subscribe to the journal, contact the AACR Publications Subscriptions Department at [email protected].

Permissions To request permission to re-use all or part of this article, use this link http://clincancerres.aacrjournals.org/content/2/2/319. Click on "Request Permissions" which will take you to the Copyright Clearance Center's (CCC) Rightslink site.