To The Editor:
We have read with interest the article by Juan et al1 in the July 1, 1997 issue of Blood. The study follows a previously published paper in Blood2 from the same laboratory. Both papers focus on the biology of flt3 ligand (FL). In the Juan study, a retrovirus was constructed that expressed the full-length FL molecule and was used to infect stem cells that repopulated the hematopoietic system of lethally irradiated mice. Both reports devote a significant amount of their allotted page space to predictions of FL toxicity and suggest that the studies performed by Juan et al will be useful to predict clinical adverse events in patients receiving FL.
We believe that the model used by Juan et al is an artificial system of sustained, local, high-level production of a cytokine that inadequately mimics the effects of short-term systemic administration of growth factors. One key issue with a study such as this is the level of FL made in reconstituted animals. Figure 2 shows that FL mRNA is clearly detectable from the retroviral vector, but the endogenous FL message is virtually undetectable. This suggests that the level of vector-encoded FL is extremely high, because endogenous FL transcripts are abundant in hematopoietic organs such as the spleen.3 Unfortunately, there was no quantitation, either by enzyme-linked immunosorbent assay, histochemistry, or bioassay of how high the expression levels of FL are in these studies. Furthermore, it has been elegantly shown that the membrane-bound and soluble forms of stem cell factor elicit qualitatively different hematologic responses.4 Because FL and stem cell factor are structurally related cytokines, it stands to reason that such a difference in biologic response may exist with FL as well. There are no data in Juan et al to help understand how much of the membrane-bound and soluble forms of biologically active FL were found in experimental animals reconstituted with the FL-expressing retrovirus.
Several studies have shown that soluble, recombinant FL can be safely administered to rodents (including Molineux et al) and primates5-8 and, therefore, do not support the prediction in the discussion by Juan et al of significant risk with this growth factor, even with short courses of administration. We have completed a formal primate toxicology program with FL at doses of up to 2 mg/kg daily for 30 days and fail to see the types of tissue changes reported with retroviral expression. Furthermore, we have recently reported on the preliminary results of human clinical testing of FL at doses up to 100 μg/kg/d for 14 days in normal human subjects and have found the molecule to be very well tolerated.9 Obviously, these studies are preliminary in nature and will require further clinical testing to confirm, but the data from preclinical studies and early clinical testing are encouraging and do not support the presumption of toxicity by Juan et al.
