Recombinant Hematopoietic Cytokines Of Therapeutic Importance

The discovery of hematopoietic cytokines, predominantly in the 1970s and 1980s, followed the development of assays to detect their activity like the in vitro colony-forming cell assays introduced above. However, the larger challenge at that time was purifying proteins with separate activities from the complex biological fluids used as the starting material. Macrophage (M)-CSF (also known as CSF-1) was the first hematopoietic growth factor to be purified, initially from human urine and later from medium conditioned by a murine fibroblast cell line (7). This was followed in the same year by the discovery of granulocyte-macrophage (GM)-CSF in medium conditioned by tissues from the lungs of mice previously treated with bacterial lipopolysaccharide (8). A few years later, a third myeloid growth factor was identified: granulocyte (G)-CSF (9). It was after some years that the genes that encoded these proteins were cloned—cloning was a relatively nascent technology at that time; thus, 1985 saw the cloning of human M-CSF (10), EPO (11,12), and GM-CSF (13,14), and 1986 saw the cloning of G-CSF (15,16), IL-3 (17), and IL-5 (18).

The natural versions of most hematopoietic cytokines are glycosylated, for example, IL-3 (17), IL-5 (19), IL-6 (20), IL-7 (21), GM-CSF (13), G-CSF (22), M-CSF (23), SCF (24), and EPO (25). In several cases, however, the carbohydrate has been shown not to be required to maintain activity, for example, the O-linked carbohydrate at threonine 133 on natural G-CSF. In one celebrated case however, that of EPO, the carbohydrate component was found to be not only obligatory for in vivo action but also amenable to manipulation to therapeutic advantage (26). Endogenous cytokines are frequently heterogeneous at some level, often because of posttranslational modifications such as glycosylation, sulfation, proteolytic cleavage, etc. Recombinant forms may not therefore be identical to the natural prototype and will vary markedly depending on the host cell in which they are produced, method of purification, and a number of other factors. Overall, the precise biochemical nature and activity of endogenous cytokines remain largely unknown as does their comparability with recombinant preparations. Comparisons can be made to define relative potency, but other aspects of product performance, for example, pharmacokinetics, safety, etc., must be studied carefully in animals or humans and often in large numbers of subjects and over extended periods before their safety and efficacy can be definitively established.

With respect to the clinical development and subsequent consideration of therapeutic proteins by regulatory agencies, it has been suggested that the protein product is in essence the process used to manufacture it. This perspective presents a considerable hurdle in comparing related products like, for example, follow-on biologics (FOBs), subsequent entry biologics, or biosimilars intended to offer alternative products after innovator patent expiry. Thus, the term "generic" is difficult to apply given the likely nonidentity of proteins produced in different host cell systems that are purified and formulated using different methods—presenting an interesting challenge for regulatory authorities for which differing solutions are being developed in different countries.

From a drug development perspective, the general observation that has emerged from the medical exploitation of hematopoietic cytokines is that plei-otropy is an undesirable property for such agents. More lineage-restricted cytokines have, in general, proven more useful (27), as exemplified by the clinical utility of EPO (28), G-CSF (29), and GM-CSF (30) and the promise of a thrombopoietin (TPO) mimetic. In the following sections, the discovery and development of these hematopoietic growth factors with demonstrated clinical utility, and their pharmacokinetic (PK) and pharmacodynamic (PD) properties, will be discussed.


Also known as mast cell growth factor (MGF), kit ligand (KL), and steel factor, SCF is the ligand for the cognate tyrosine kinase receptor c-kit. It is approved for clinical use in limited countries as a coadministration with G-CSF for hema-topoietic stem and progenitor cell mobilization based on phase 3 clinical trial data in breast cancer patients (31). Despite its use in stem cell mobilization, all patients require prophylactic administration of H1 and H2 antihistamines and a bronchodilator to ameliorate the collateral effects of SCF in stimulating mast cells.

The PK parameters of SCF in humans have not been extensively studied but appear relatively unremarkable. A phase 1 trial in cancer patients indicated a predose serum SCF level of around 1 mg/mL, with a Cmax 12 to 17 hours after first administration, reducing with subsequent injections (32). Clearance was linear, with a half-life of approximately 35 hours. More intriguing were the data obtained for recombinant SCF administered to mice. Following intravenous administration, radiolabeled material distributed very quickly to the lungs of treated mice and was then eliminated via the kidney and liver with a half-life of around two hours. Sl/Sld mice, which lack mast cells because of a genetic lesion in the SCF gene, also accumulated SCF in the lungs but did not suffer the effects of mast cell degranulation seen in their wild-type littermates (33).

The link between the PK and PD of SCF is not particularly clear. The major PD endpoint measured in phase 3 trials was the mobilization of CD34+ cells. However, mobilization is an indirect result of neutrophil-derived proteases cleaving adhesion molecules that tether stem and progenitor cells to the bone marrow stroma (34). Thus, mobilization is mechanistically related to the granulocyte response rather than a direct effect of SCF. Since SCF has been shown to interact with intracellular G-CSF signaling (35), the phenomenon observed and exploited in patients is understandable. This outcome may not be directly linked to SCF, and so it may be causally distinct from the PK. In contrast, the side effects (or at least unintended effects) on mast cells are better understood and more satisfactorily linked to drug exposure.


GM-CSF is one of two myeloid cytokines approved for clinical use in cancer patients in the European Union and the United States, the other being G-CSF. GM-CSF does not have the breadth of application that G-CSF has, with its approved clinical uses being confined to acute leukemia and in transplant settings. As the name implies, GM-CSF is more pleiotropic than G-CSF. Among the documented effects of GM-CSF are stimulation of progenitor cell proliferation (36), neutrophil function (37,38), monocyte activation (39), and dendritic cell function (40), especially, as a vaccine adjuvant.

In a recent study (41), GM-CSF was administered daily for 10 days to cancer patients; PK analysis showed a dose-dependent increase in drug level several hours after the first administration when none had been detectable beforehand. By the time of the next daily dose, about half the patients still had low but detectable GM-CSF in their blood. In common with many cytokines, SC administration prolonged the half-life of GM-CSF, possibly via delayed absorption, with nonlinear clearance for escalating doses (42). With repeated administration, clearance of GM-CSF gradually increases (41,43,44). Though the mechanism for this effect is not well defined, it may include target cell-mediated clearance as will be discussed later for M-CSF, G-CSF, and TPO. Intravenous administration of GM-CSF illustrates two distinct phases of disposition: the first, presumably representing initial distribution, is quick (Ti/2 less than five minutes); the second phase is slower, with T1/2 of two to three hours (45) representing clearance.

Hematological (PD) responses to administration of GM-CSF include increases in circulating lymphocytes, monocytes, neutrophils, and eosinophils, with small or no changes in erythrocytes and platelets (41). Though these effects, especially on neutrophils, may be used to define the PK/PD relationship in, for example, neutropenia after bone marrow transplantation, the desired PD in other settings may not be so clear. For instance, in the deployment of GM-CSF for immunotherapy applications, the increased leukocyte count, which relates to both dose and duration of GM-CSF treatment, correlated positively with the absolute number of putative immune effector (GM-CSFRa+/CD14+, GM-CSFRa+/CD66b+) cells. In contrast, high doses of GM-CSF impaired antibody-dependent cellular cytotoxicity (ADCC) in in vitro assays of harvested cells. This suggests that dose and schedule need to be optimized for this application, but the predictable PK of GM-CSF should make this relatively straightforward as long as the nature of the desired biological effect is well defined. In practice, the cell types required to elicit optimal immune function are not fully understood and will require further study to define the desired PD of GM-CSF in what would appear to be its most useful application.

0 0

Post a comment