The James Ewing Laboratory of Developmental Hematopoicsis, Memorial Sloan-Kettering Cancer Center, New York, NY 10021, USA * Supported by grants CA 20194, CA 32156, and CA 31780 from the National Cancer Institute; American Cancer Society Grant CH-3K, and the Gar Reichman Foundation.

A. Introduction

The Frederick Stohlman Memorial Lectures have, over the years, ref1ectcd the progression of leukemia research in areas initially thought to be as diverse as retrovirology /oncogenes, chemotherapy /immunotherapy, and hematopoietic growth factors. At the Seventh Wilsede Meeting we heard how convergent these areas were. The polypeptide growth factors which affect Iymphohematopoietic ce1l proliferation and differentiation have expanded to at least 11 distinct gene products with pleiotropic and overlapping functions (G-, GM-, M-CSF, erythropoietin, and interleukins 1- 7 [1 ]. The early characterization of hematopoietic growth factors was based on their ability to stimulate the clonal proliferation of human bone marrow progenitor ce1ls. In addition, these factors were shown to activate a variety of functions of mature lymphoid and hematopoietic ce1ls. The purification, characterization, and gene cloning of human G-CSF fo1lowed shortly after the equivalent characterization of human GM-CSf [2, 3]. With the availability of recombinant G-CSF in quantities sufficient for in vivo evaluation, its action on hematopoietic function was assessed in normal murine and primate systems, and fo1lowing chemotherapy or radiation-induced myelosuppression [1, 4, 5]. Efficacy was rapidly demonstrated and opened the way to clinical trials in human bone marrow suppression or failure systems [6 8]. Without reviewing this area in detail, 1 wish to touch on various aspects that i1lustrate the utility of G-CSF in various pathophysiological situations.

B. Hematopoietic Role of G-CSF

I. G-CSF Treatment of Mice Receiving Cyclophosphamide, Myleran, or 5-rluorouracil
In order to investigate the potential of G-CSF in preventing episodes of neutropenia fo1lowing high-dose chemotherapy with cyclophosphamide (CY), C3H/ HeJ mice were suhject to weekly injections of 200 mg/kg CY intraperitonea1ly followed by G-CSF therapy (1.75 g x 2 daily) beginning 2 h after CY treatment and finished 48 h before the second cycle of CY. As shown in Fig. 1, this protocol prevented the subscquent nadirs of neutrophil counts between 4 and 5 days after CY treatment, and in a1l ten cycles of CY therapy G-CSF ahrogated the neutrophil nadirs. The mice receiving CY alone had a substantial mortality evident by the eighth cycle of treatment. In eight cyc1es of CY treatment over the course of 70 days the G-CSF-treated animals were neutropenic (1000 ANC/mm3) for only 3 days, whereas thc untreatcd animals were neutropenic for 24 days. The time to recovery of absolute neutrophil counts (ANCs) to control levels was also substantially affected, with the non-G-CSF-treated groups requiring an additional16 days for recovery. Neutropenia at levels

Fig.2. Absolute neutrophil counts in the peripheral blood of C3H/HcJ mice administered 100 mg/kg myleran orally by intubation with or without twice daily i.p. injections of rh-G-CSF at 2 µg/mouse for 21 days. Control animals received corn oil alone with G-CSF. Four mice per group. Note the log scale

of 500 ANC/mm3 indicated an even more significant difference, with such low levels observed on only 1 day out of 70 in the G-CSF + CY -treated mice, in contrast to 15 days in the CY-treated group. This cyclophosphamide model allowed us to investigate the timing of initiation of G-CSF therapy following a single injection of a high dose of CY. Administration of G-CSF coincident with CY treatment resulted in an improvernent in recovery of neutrophil counts, but this was substantially less than observed when GCSF was administered 2 h after CY therapy. Administration of G-CSF 24 and 48 h after CY was less effective than after 2 h, indicating that treatment early in the course of chemotherapy is likely to be more effective. Myleran (MY), administered orally (100 mg/kg), was also associated with a significant reduction in neutrophil counts and in bone marrow progenitor populations. G-CSF therapy initiated within 6 h of MY treatment led to an accelerated recovery of neutrophils so that by 1 week the neutrophil counts had returned to normal values, whereas the MY -treated controls rernained profoundly neutropenic (ANC 200/rnm3) (Fig. 2). By 14 days MY + G-CSF-treated animals exhibited a neutrophilleukocytosis of between 104 and 2 x 104/rnm3 at times when the MY control animals had 100 neutrophils/rnm3. Treatment with 5-fluorouracil (5-FU 150 mg/kg) led to profound myelosuppression in three different strains of mice tested. In C3H/HeJ mice, recovery of neutrophil counts to control values was delayed for 12-14 days. Administration of G-CSF accelerated recovery of neutrophil counts to normal values within 8 days [5].

II. Clinical Trials of G-CSF Following Chemotherapy

The use of rh-G-CSF in clinical trials has demonstrated efficacy in promoting regeneration of functioning neutrophil granulocytes. Our first clinical trial involved the administration of G-CSF in a phase 1/1I study of patients receiving MVAC (methotrcxate, vinblastinc, doxarubicin, and cisplatin) in 22 patients with transitional ceIl carcinoma of the uroepitheliurn [6]. In all patients a specific dose-dependent increase in ANC was seen to levels of up to 12-fold above normal at the higher levels of G-CSF administration. At these high doses (10-60 g/ kg) a tenfold increase in monocytes, but not in other hernatopoietic lineages, was evident. Treatment after chernotherapy significantly reduced the number of days on which antibiotics were used. This resulted in a significant increase in the number of patients qualified to receive planned chernotherapy on day 14 of the treatment cycle (100% vs. 29% ). In addition, the incidence of mucositis significantly decreased. Investigators at other centers have confirmed these observations. In patients with advanced small cell lung cancer receiving high-dose chemotherapy, repeated every 3 weeks, G-CSF was given to each patient for 14 days on alternate cycles of chemotherapy. This resulted in the reduction of the period of absolute neutropenia with return to normal or supranormallevels of ANC within 2 weeks [8]. This therapy also drastically reduced episodes of severe infections observed during these cycles of chemotherapy .In advanced malignancies treated with melphalan, neutropenia was significantly reduced even at dose levels of 1 or 3µg/kg G-CSF. Preliminary clinical studies have also demonstrated that G-CSF hastens granulocyte recovery in Hodgkin's disease after high-dose chernotherapy and autologous bone marrow transplantation [9]. In all of these studies no significant dose-Iirniting toxicities have been observed to date.

III. In Vivo Studies of G-CSF in Congenital and Idiopathic Neutropenia

Cyclic neutropenia is an inherited disease of man and gray collie dogs, characterized by regular oscillations of the number of peripheral blood cells and of bone marrow progenitor cells. Cycling of se rum or urine CSF levels has also been reported. However, induction or cure of cyclic neutropcnia by bone marrow transplantation has suggested that this disorder is a disease of pluripotential hematopoietic stern cells. It should be noted that the hematopoietic stem cell itself can generate cells (macrophages, T cells ) capable of hematopoietic growth [actor production and thus it is not possible to exclude intrinsic growth factor production defects in the pathophysiology of cyclic neutropenia. In the dog model we have demonstrated that daily administration 0f rh-G-CSF (5 µg/kg x 2 daily [or 30 days) caused an immediate (within 12 h) and persistent leukocytosis ( > 40 000 W HC) in both cyclic and normal dogs due to a ten [old increase in the numbers 0[ circulating neutrophils and monocytes [10]. This therapy eliminated two predicted neutropenic episodes and suppressed the cycling of CFU-GM in the bone marrow. Comparable therapy with rh-GM-CSF induces a monocytosis and neutrophilia in normal dogs but did not eliminate the recurrent neutropenia in cyclic dogs. Analysis 0[ the serum levels of CSF indicated a striking periodicity with peak levels of G-CSF coinciding with the peaks of monocytes and the nadirs of neutrophils. The continuing cycling of CSF levels was also observed even when animals showed a major neutrophil leukocytosis under exogenous GCSF therapy. A disorder of neutrophil regulation is also seen in chronic idiopathic neutropenia in which the peripheral neutrophil count is reduced to 2000 cells/mm3 for prolonged periods, with maturation arrest of neutrophil granulocyte precursors in the bone marrow. Other hematopoietic cell counts are usually normal and antineutrophil antibodies are absent. Clinically, these patients experience recurrent episodes of life-threatening infections, ulcers of the mucus membrane, and periodontal disease. We have had the opportunity of studying a number of patients with this disorder utilizing primary clonogcnic assay and in vitro suspension cultures of patient bone marrow cells [7]. The incidence of CFU-GM observed in primary bone marrow culture was always in the high range of normal with respect to each species of CSF (G-, GM-, M-, and IL-3), indicating that the neutropenia was not attributed to a shortage of appropriate myeloid progenitors. furthermore, morphological studies showed normal neutrophil maturation within the colonies developing in the presence of G-, GM-CSF, and IL-3, with eosinophil maturation and macrophage development a significant feature with GM-CSF stimulation. In suspension cultures GCSF was particularly effective in generating mature segmented neutrophils which were absent in input bone marrow, and did not develop spontaneously in over 3 weeks of culture. With the addition of G-CSF, myelocytes expanded in the 1st week 0[ culture. By the 2nd and 3rd week high levels of production of segmented, functionally normal neutrophils were found. Based upon these in vitro studies, patients were treated with subcutaneous G-CSF on a daily continuous basis. In the first patient studied, the neutrophil count rose rapidly, and by 20 days the patient's ANC was > 1000 cells/mm³ and plateaued in the range of 2000-3000/mm³ with evidence of a 40-day cycle (observed even when the dose 0[ G-CSF was increased). The neutrophils were functionally normal and the patient has remained on this continuous G-CSF treatment without any notable toxic side effects [7]. A more severe form of neutropenia is found in patients with Kostmann's syndrome (congenital agranulocytosis). In this disorder, marked by sevcre neutropenia and maturatjon arrest at the myelocyte level, five patients received GCSF therapy, and the neutrophils increased from 0% -1% to 10%-72% with clinical resolution of preexisting infections [11].

C. Hematopoietie Role of Interleukin-1

Interleukin-1 ( I L-1 ) was first identified as an endogenous pyrogen, produced by macrophages following activation by endotoxin. Subsequent studies implicated IL-1 as a mediator of a variety of inflammatory phenomena involving production of cascades of cytokines and cyclooxagenase products [12]. Various lymphokines are released by the interaction of I L-1 with T cells, and IL-1 induces expression of G-, GM-, and MCSF by a direct action on stroma] cells such as fibroblasts and endothelial cells, both in vitro and in vivo [13 -15]. Recent studies have revealed a more direct role for IL-1 in hematopoiesis, involving an interaction with the early hematopoietic stern cell [5,16-18].

I. Interleukin-1 Identity with Hemopoietin 1/Synergistic Activity

Early studies involving mice treated with 5-fluorouracil (5-FU) suggested the necessity of a synergistic interaction between two factors, one being a direct hematopoietic colony stimulus such as MCSF or IL-3, and the other, which lacked direct colony-stimulating activity, was shown to synergize with the preceding species of CSFs [19]. Syncrgistic activity was identified in conditioned media or extracts of various human tissues and was also identified and purified to homogeneity from the human bladder cancer cell line, 5637 [20]. This latter cellline had also been used to identify, purify, and clone G- and GM-CSF [2, 3, 21]. The constitutive expression of hemopoietin 1 (H-1), which was shown to be a 17000kd protein [20], led us to attempt to purify, sequence, and clone this molecule, using similar strategies to those that we had employed to clone the G-CSF gene from 5637 cells. The identity of [L-1 and H-11 synergistic activity was established on the basis of the following factors :

1. The bladder cancer cells evidently produced high levels of IL-1 active in the thymocyte comitogenesis assay.
2. The synergistic activity and I L-1 could be copurified using various protein purification procedures.
3. Abundant levels of IL-1-alpha and -beta mRNA could be detected in the 5637 cells -comparable to levels in activated macrophages.
4. F'ractions of in RNA from 5637 expressed in oocytcs resulted in intracellular production of hematopoietic synergistic activity and thymocyte comitogenesis factor production.
5. The synergistic activity and thymocyte comitogenesis activity were completely neutralized by monoclonal and polyclonal antibodies to IL-1. 6. rhlL-1 alpha and beta at 0.110 U/ml synergized with G-, GM- M-CSF, and IL-3 in stimulating high proliferative potential (HPP)-CFU in clonogenic assays of 5-FU-treated murine bone marrow [1, 5, 16-18].

IL-1 has a direct effect upon hematopoietic stem cells in addition to its ability to elicit production of various CSF species by accessory cell populations within hematopoietic tissues. In our original studies it was not possible to conclude that I L-1 was acting directly on early stern cells since accessory cell populations were not depleted from the target bone marrow cell population. In more recent studies we have established a linear doseresponse relationship between the numbers of HPP-CFU, and the number of bone marrow cells plated, with a highly significant correlation (r = 0.97) indicative of a single-hit phenomenon.

II. Action ofIL-1 in Short-term Marrow Suspension Culture (Delta Assay)

The rationale behind the delta assay is to demonstrate the ability of hematopoictic growth factors to promote the survival, recruitment, or expansion of stern cells and/or progenitor cells in relatively short term suspension culture systems. As originally developed, we utilized bone marrow from mice that had been treated with 5-ru for 24 hand then subjected to a 4- to 7-day suspension culture in the

Table I. Interleukin-1 and CSF-induced amplification of CFU-GM in 7-day
suspension cultures of 25 x 100.000 murine bone marrow cells obtained 24 h post 5-fluorouracil treatment

Femoral bone marrow cells taken from B6D2F1 mice 24 h after a single i.v. injectiol1 of 5-nuorouracil (150 mg/kg) were incubated at 25 x 100.000 cells/ml in Iscove's modified Dulbecco's medium with 20% fetal calf serum il1 24-well cluster plates containing 100 units rh-1 L-I alpha. or 2000 u/ml rh-G-CSF; 1000 u/ml m-GM-CSF (purified from muril1e post-endotoxil1 lung CM) 200 u m-IL-3 (purified from WEHI-3 cell line CM), or 1000 u CSF-1 (purified from L-cell CM). After 7 days of incubation, cells were recovered and assayed for CFU-GM in agarose cultures stimulated by the various CSF species alone or in combination with IL-1. The delta valuc (CFU-GM output/input) was calculated on a recovery from triplicate clonal assays from triplicate suspension cultures

presence of I L-1 alone, CSFs alone, or combinations of IL-1 with various CSF species [5, 16]. At the end of the suspension culture phase, total cellularity and morphology was determined, and cells were cloned in semisolid culture, again in the presence of IL-1 alone, CSFs alone, or combinations of I L-l and CSFs. Table 1 shows that IL-l caused an expansion of the numbers of CFU-GM recovered after 7 days of culture. Neither M-CSF nor G-CSF alone supported survival or expansion of CFU-GM; however, the combination of IL-l and CSFs demonstrated additivc or synergistic effects on the expansion of thcsc progenitors. GM-CSF or IL-3 alone caused some expansion of progenitor populations, but again the combination with lL-1 evidenced synergtsm. Human systems based on in vitro purging with 4-hydroperoxcyclophosphamide (4HC) and positive selection by "panning" with MY10 monoclonal anti body for CD34 + cells followed by 7-day suspension culture also demonstrated synergistic interactions between 1 L-1 and various CSF species. The most dramatic effect was observed with the combination of IL-1 and IL-3, where an up to 85-fold increase in progenitor cells was noted following 7-day suspension culture (Table 2). In contrast to the murine system, synergism between IL-1 and CSF-1 was not evident.

III. In Vivo Interaction Between IL-1 and G-CSF in mice Treated with 5-FU

In mice trated with 5-FU, G-CSF administration restores neutrophil counts to normal values some 5-6 days earlier than in mice not receiving the factor, but aperiod of profound neutropenia is still observed [5]. Administration of IL-l alone, giving postchemotherapy for 410 days twice daily at doses of 0.2 µg/ mouse per day, reduced the severity of

Table 2. Interleukin-1 and CSF-induced amplification of CFU-GM in 7-day
suspension cultures of 25 x 100.000 4-HC-treated, CD34 + human bone marrow cells

Bone marrow cells, obtained with informed consent from normal volunteers, were separated over Ficoll-Hypaque, subjected to plastic adherence, and incubated for 30 min with 100 µM 4-hydropcroxycyclophosphamide (4-HC). Cells were then treated with antimyl monoclonal antibody (anti-HPCA-1 Beckon Dickinson, Mountain View Ca), on ice for 45 min, washed. al1d incubated for 1 h at 4 Celsius on bacteriological-grade plastic petri dishes previously coated with goat anti-mouse IgCJ. Adherent CD34 + cells were harvested by vigorous pipcttil1g. 25 x 100.000 harvested cells/well were il1cubated in lscove's moditied Dulbecco's medium (IMDM) plus 20%., fetal calf serum in 24-well cluster plates containing test stimuli. In suspension phase, stimuli were 10 ng/ml rh-G-CSF, rh-GM-CSF, rh-IL-3 (Amgen), rh CSF-1 (Cetus), and rh-IL-1 alpha (Roche). 4-HC purged, CD34 + marrow cells were plated at 2 x 10.000 cells/ml in semisolid agarose culture in the presence or 10 ng/ml GM-CSF, G-CSF, CSF-1, or IL-3 alone or in combination with IL-I both pre- and postsuspension culture. Colony formation was assessed after 12 days. The delta value (CFU-GM output/input) was calculated on recovery from triplicate clonal assays from duplicate suspension cultures

the neutrophil nadir, and accelerated the recovery of the neutrophil count to an extent greater than observed with G-CSF alone. The combination of G-CSF and IL-1 administered after 5-FU therapy also resulted in accelerated hematopoictic reconstitution, although the results were additive rather than synergistic [5]. Analysis of total hematopoietic cell reconstitution in %-FU-treated mice also showed the efficacy of I L-1 therapy, or IL-l plus G-CSF therapy, in accelerating total recovery of erythroid as weil as granulocytic elements in the marrow, spleen, and blood . The potential of combination biotherapy as an effective means or accelerating hemopoietic cell differentiation raises the issue of whether premature exhaustion of the stern cell and progenitor cell population may occur. Measurement of colony forming units (cru-s, CFU-c, BFU-c, CFU-GEMM, CFU-Meg) on an incidence basis proved to be misleading because of the redistribution or hematopoietic precursors in regenerating murine tissues. By quantitation of total numbers of these cell populations in marrow, spleen, blood, and other tissues, we demonstrated that there was an absolute increase in the recovery of all of these populations in 5-FU-treated mice exposed to both I L-1 alone and I L-1 plus G-CSF. The results indicate that the mechanism of action of these cytokines, both alone or in combination, is not simply mediated by accelerated differentiation, but involves an absolute expansion of the most primitive stern cell populations that can be measured, e. g., by the day 12 murine spleen colony (CFU-s) assay, or the high proliferative potential in vitro colony assay requiring IL-1 plus CSF as the read-out stimuli [1, 5, 16]. This is an important observation for extrapolation of these studies to clinical situations because it lays to rest some of the concerns that combination biotherapy may accelerate exhaustion of the stem cell compartment in patients receiving myeloablative therapy.

IV. Interleukin-1 and CSFs as Radioprotective Agents

Interleukin-1 has been shown to mediate a radioprotective action when given to mice 20 h before what would otherwise be a lethal dose of irradiation [12, 15]. This radioprotective effect cannot be duplicated by preadministration of various CSF species. Using BALB/c mice which are particularly susceptible to irradiation, we observed that 0.2 ~g IL-1 given 20 h before 850 rads total body irradiation completely prevented what would otherwise be 100% mortality by 14 days, confirming earlier reports. However, we were able to show that administration of IL-l or G-CSF after irradiation was also signil1cantly radioprotective. While the mechanism of IL-l radio protection remains controversial, our down observations suggest that one mechanism may involve accelerated recruitment of primitive Go stem cells into the cell cycle with expression of CSF receptors. The administration of 750 rads total body irradiation to BALH/c mice, which resulted in 80% mortality within 14 days, provided a model in which we could test the interactions of IL-l and G-CSf in reconstitution of hematopoiesis and in radioprotection. Recovery of neutrophils to normallevels in irradiated mice was delayed for up to 3 weeks; G-CSF administration begun immediately after irradiation, and carried out for 3 weeks, accelerated recovery of neutrophils and total hcmatopoietic cells, so that normal levels were observed by 14 days [22]. IL-1 administration as a single dose prior to irradiation resulted in accelerated recovery of hematopoiesis. When administered daily postirradiation for 4 days, IL-1 reduced the nadir of blood neutrophils and produced a degree of recovery or hematopoictic function comparable to that seen with G-CSF administration (Fig. 3). A significant synergistic interaction was noted when G-CSF was combined with IL-1 postirradiation, with up to a tenfold greater increase in blood neutrophils, hematopoietic cells, and both stem and progenitor cells populations in both spleen and bone marrow.

Fig.4. Absolute neutrophil counts in the peripheral blood of B6D2F1 mice subject to 950 rads total irradiation and injected with varying numbers of syngeneic bone marrow cells. with or without subsequent twice daily administration of 2 µg rhG-CSF i. p. Five mice were used per group



V. Interactions Between IL-1 and G-CSF in Autologous Bone Marrow Transplantantion

Transplantation of autologous bone marrow into lethally irradiated (950 rads) C3H/HeJ and B6D2F1 mice was undertaken using three different dose levels of donor marrow cells, 100.000, 1million, and 10million cells/mouse. The LD5o in such transplanted animals is observed at dose levels of 1.0-2.5 x 1.million marrow cells/ mouse. It is generally accepted that the delayed recovery in hematopoietic cel1s in the bone marrow and spleen and in peripheral neutrophil counts was attributed to limitations in the number of donor stem cel1s as measured by the CFU-s assay. As can be seen in Fig. 4, the peripheral neutrophil counts ref1ected a marrow cell dose-dependent difference in a return to normal values. While this observation (which was also ref1ected in bone marrow and spleen hematopoietic cell reconstitution) might lead to the assumption that stem cell availability is the sole critcrion for hematopoietic reconstitution, the administration of exogenous G-CSF suggested additional variables in the transplant equation. In all instances fol1owing autologous bone marrow transplantation at the different dose levels, exogenous G-CSF administration enhanced the recovery of neutrophils in the peripheral blood by three- to tenfold. This in turn was accompanied by accelerated regeneration of bone marrow and spleen hematopoietic progenitor ccl1s and of pluripotential stem cel1s as measured by the CFU-s assay. Studies in autologous bone marrow transplantation in cynomolgus monkeys confirmed the ability of G-CSF to accelerate bone marrow regeneration and recovery of peripheral blood neutrophil counts. These studies indicated that higher doses of G-CSF were required to enhance the recovery of hematopoietic parameters to normal, suggesting that the role of G-CSF involved recruitment of earlier stem cel1s by some indirect mechanism involving cytokine cascades or, alternatively, reflecting the relative paucity of G-CSF receptors on primitive hematopoietic cells. Based upon our in vitro studies indicating that IL-1 upregulated the receptors for G-CSF and other CSF species on primitive hematopoietic stern cells that were resistant to irradiation and chemotherapy, we cultured bone marrow from mice treated with 5FU for 24 h with IL-1 prior to autologous transplantation in lethal1y irradiated mice. Fol1owing transplantation, recipients received 1 L-1 alone for 4 days, G-CSF alone for 21 days, or IL-1 plus G-CSF, and hematopoietic reconstitution was assessed. Preliminary studies indicated that in vitro pretreatment of bone marrow enhanced bone marrow reconstitution and accelerated recovery of peripheral neutrophil counts in mice that subsequently received IL-1, G-CSF, or a combination of IL-1 plus G-CSF in the post-transplant period. As measured by reconstitution or total stern cel1s and progenitor cells, hematopoietic recovery, and peripheral neutrophil counts, the most efficacious combination involved pretreatment of bone marrow in vitro with IL-1, followed by post-transplant treatment with a combination of IL-1 and G-CSF.

D. Conclusion

Numerous clinical trials are now underway using G-CSF and GM-CSF in iatrogenic myelosuppressive situations associated with cancer chemotherapy and irradiation therapy, and in conjunction with autologous and al1ogeneic bone marrow transplantation. Efficacy is also under study in congenital disorders of neutrophil production, and in myelodysplastic syndromes and myeloid leukemias. The initial promise of thc CSFs and interleukins suggests that they may provide a major new therapeutic modality but only if we can develop a deeper insight into their modes of action and interaction. The choice of the type or factor to be administered will be influenced by the particulars of the pathology being treat ed, making it meaningless to generalize as to whether one factor is "better" than another. It is already clear that the timing, dose, and route of administration of hematopoietic growth factors arc important variables and efficacy will have to be balanced against potential adverse side effects (which fortunately have not emerged as serious limitations in the case of G-CSF and GM-CSf trials). In addition, preclinical studies point to the value of combination biotherapy using two or more factors administered at the same time, or sequentially. Finally, there is a need to recognize that proliferative processes arc self-limiting and physiological mechanisms probably exist to counteract the action of hematopoietic growth factors. Unraveling these issues wil1 undoubtedly occupy the agenda of many future Wilsede Meetings.

References

1. Moore MAS ( 1988) The use or hematopoictic growth and differentation factors for bone marrow stimulation. Ill: De Vita VT, Hellman S, Rosenberg SA (cds) Important advances ill oncology 1988. Lippincott, Philadelphia, pp 31-54
2. Welte K, Platzer E, Lu L, Gabrilovc JL, Levi E, Mertelsmann R, Moore MAS (1985) Puriricatioll and biochemical characterization or human pluripotent hematopoictic colony-stimulating factor. Proc Natl Acad Sci USA 82.1526-1530
3. Souza L, Boolle TC, Gabrilove JL, Lai PH, Zsebo KM, Murdock DC, Chazill VR, Bruszcwski J, Lu H, Chell KK, Barelldt J, Platzer E, Moore MAS, Mertelsmann R, Welte K (1986) Recombinant human granulocyte colony-stimulating factor; effects On normal and leukemic cells. Science 232.61-65
4. Welte K, Bonilla MA, Ginio AP, Bone TC, Potter GK, Gabrilove JL, Moore MAS, O'Reilly L Souza LM (1987) Rccombinant human granulocyte colony stimulating factor. effects On hematopoiesis in normal and cyclophosphamidetreated primates. J Exp Med 1645: 941-948
5. Moore MAS, Warrell W ( 1987) Inter Ieukill-1 and G-CSF synergism. ill vivo slimulaliol1 of stem cell recovery al1d hematopoictic rcgcl1cratiol1 followil1g 5-fluorouracil treatmel1t il1 mice" Proc Natl Acad Sci USA 84" 7134- 7138
6. Gabrilovc JL, Jakobowski A, Scher H, Slerl1bcrg C, Wong G, Grous J, Yagoda A, Fain K, Moore MAS, Clarkson B, Oettgen HF, Alton K, Welte K, Souza L (1988) Effect of granulocytc colony-stimulating factor on neutropenia and associated morbidily due to chemotherapy for transiliol1al-cell carcinoma of the urothelium" N Engl J Med 318:1414-1422
7. Jakubowski AA, Sou7a L, Kelly F, Fain K, Budman D, Clarkson B, Bonilla MA, Moore MAS, Gabrilove JL (1989) Effects of human granulocyte colony stimulating factor on a patient with idiopathic neutropenia New Engl J Med 320 38 -42
8. Hernandez Bronchud MH, Scargge JH, Thatcher N, Crowther D, Souza LM, AIton NK, Tesla NG, Dexter TM (1987) Phase 1/11 study of recombinant human granulocyte colony-stimulating factor in patients receiving intensive chemotherapy for small cell lung cancer" Br J Cancer 56:809 -813
9. Taylor K, Spitzcr G, Jagannath S, Dicke K, Vincent M, Souza L (1988) rhG-CSF hastens granulocyte recovery in Hodgkin's disease after high-dose chemotherapy and aulologous bone marrow transplant" Exp Hcmaloll6"413 (abstr)
10. Lothrop Cl) Jr, Warren DJ, Souza LM, Jones JB, Moore MAS (1988) Correction of canine cyclic hematopoiesis with recombinant human granulocyte colony stimulating factor" Blood 72" 1324-1328
11.Bonilla MA, Gillio AP, Ruggiero M, Kernan NA, Brochstein JA, f'umagalli L, Bordignon C, Vincent M, WeIte K, Souza LM, O'Reilly RJ (1988) Corrcclion of neutropenia in patients with congenital agranulocytosis with recombinant human granulocyte colony stimulating factor in vivo" Exp Hemalol t 6 520 (abstr)
12. Oppenheim J, Kovacs E, Malsushima K, Durum SK (1986) There is more than one interlcukin-1 Immunol Today 7:45-56
13. Lovhaug D, Pelus LM, Nordie EM, Boyum A, Moore MAS (1986) Monocyteconditioned medium and inlcrleukin I in duce granulocylc macrophage colony stimulating jaclor production in the adherent cell layer of murine bone marrow cultures Exp Hematol 14:1037-1042
14. Bagby GC, Dinarello CA, Wallace P, Wagner C, Hefeneider S, McCall E (1986) Interleukin I stimulates granulocyle macrophage colony stimulating activity release by vascular endothelial cells J Clin Invest 78:1316-1320
15. Neta R, Oppenhcim JJ (1988) Why should internists be interested in il1lerleukin 1 ') Ann Intern Med 109.1-3
16. Moore MAS, Warren DJ, Souza L (1987) Synergistic interaction between inlerIcukin-1 and CSF's in hemalopoiesis" In" Gale RP, Golde l)W (eds) Recent advances in Ieukemia and lymphoma" UCLA symposium on molecular and cellular biology Liss, New York, pp 445-458
17. Mochizuki l)Y, Eisenman J R, Conlon PJ, Larscn AD, Tushinski RJ (1987) Interleukin I regulaled hemalopoielic activity, a role previously ascribed to hemopoietin 1 Proc Natl Acad Sci USA 84: 5627
18. Warren DJ, Moore M AS (1988) Synergism among interleukin I, interleukin 3, and inlcrleukil1 5 in lhc produclion of cosinophils from primilivc hcmopoictic slem cclls J ImmunoI 140:94-99
19. Stanlcy ER, Bartocci A, Patinkin D, Rosendaal M, Bradley TR (1986) Regulation of very primitive multipotcnt hemopoietic cells by hemopoietin-1 Cell 45:667-674
20. Jubil1sky PI, Stal1lcy ER (1985) Purificalion of hcmopoietin-l" a mullilineagc hcmopoictic growth factor" Proc Natl Acad Sci USA 82: 2764-2767
21. Gabrilove J, Welte K, Harris P, Platzer E, Lu L, Levi E, Mertelsmann R, Moore MAS (1986) Pluripoielin alpha" a second human hcmatopoietic colony-stimulating factor produced by the human bladder carcinoma cell line 5637 Proc Natl Acad Sci USA 83: 2478-2482
22. Moore MAS, Welte K, Gabrilove JL, Souza LM ( 1986) In vivo action of recombinant human G-CSF on chemotherapy or radiation myclosuppressed mice" Blood 68 [Suppl l]:173a (abstr)