To the Editor:
Based on experimental evidence from studies in the murine model, Gyger et al1 have stated that this is not the case; on the contrary, granulocyte colony-stimulating factor (G-CSF) most probably leads to significant early depletion of stem cells from their marrow niches.
However, in the human system (ie, patients), there are several observations2-5 indicating that the effect of stimulation of hematopoiesis may differ in mouse and human! First, after 5 days of treatment with recombinant human G-CSF (rhG-CSF) the bone marrow cellularity increases 10- to 100-fold without any change in CD34+, colony-forming units (CFU), long-term culture-initiating cells (LTCIC), and pre-CFU frequencies, corresponding to a 1 to 2 log increase in total numbers per volume.2-4 Second, the expansion was confirmed by magnetic resonance imaging, documenting a moderate increase on days 5 to 7 that surprisingly continued to expand on days 12 to 14.5 Based on these findings, it is very unlikely that the marrow was in a negative stem/progenitor cell balance on day 4 postcytokine therapy, which is the time when marrow harvest was performed in most studies.6-12 Third, several clinical studies have, until now, documented cytokine-primed marrow as a sufficient autograft without any reported graft failures in more than 100 patients.2,4,6-12
The reason for the putative discrepancy seen in humans compared with mice may be differences in the hematopoietic system. One important difference is that spleen hematopoiesis in mice13,14differs from humans as blood circulating progenitors/stem cells in mice lodge in the spleen and grow as colonies that can be readily counted 7 to 12 days later. It may be that spleen homing progenitors are the reason for the early depletion of stem cells from their marrow niches after G-CSF priming.1,13-16
The hypothetical interpretation from Gyger et al1 on peripheral blood contamination at the time of marrow harvest to explain engraftment capability cannot be supported by data from our center. We have performed analysis of stem/progenitor cell numbers, CD4/CD8, lactate dehydrogenase (LDH), and growth factor concentration without documenting any significant blood contamination in marrow aspirates up to a size of 10 to 20 mL. On the contrary, our impression is an altered and enhanced microcirculation in the marrow in parallel with G-CSF–mediated expansion of hematopoiesis.17This is reflected by a change of bone marrow punctures into easy and fast aspirations of a sufficient number of stem cells.18
In conclusion, there seems to be sound evidence from studies in humans suggesting that G-CSF can indirectly expand human stem/progenitor cell numbers after short-term priming.2-5 Consequently, it may be feasible in some poor mobilizers to perform stem cell harvest after rhG-CSF priming and expansion with the aim to obtain a safe autograft with three-lineage engraftment. It is our experience that such a harvest can be performed with success 5 to 7 days after stopping rhG-CSF administration to allow maximal expansion and harvest of greater than 1 × 106 CD34+ marrow cells/kg during one or two procedures.
In consequence of the results published by Damiami et al,9we now need to set the clock back and prospectively evaluate the outcome of high-dose therapy and autografting of quality-assessed primed marrow or mobilized peripheral blood stem/progenitor cells, taking into consideration not only time to three-lineage recovery, but also the risk of severe infections and relapse. It is our hypothesis that such primary endpoints may validate the use of rhG-CSF–primed bone marrow quality assessed in accordance with actual practice.
We thank Dr Johnsen for his reply to our recent letter to the editor. The purpose of our reply to Damiani et al1-1 was to address concerns about the capacity of G-CSF alone, as given for peripheral blood stem cell mobilization, to prime the most primitive stem cell compartment of the bone marrow (CD34+, CD38−, Lin− cells).
We have raised these concerns for two reasons.
(1) The presently proposed molecular mechanisms for stem cell mobilization do not favor a clear potential for G-CSF–mediated expansion of the uncommitted marrow stem cell pool. Indeed, some studies demonstrate that the rapid stem cell mobilization after G-CSF administration is potentially mediated by modulation of adhesion molecules on CD34+ cells, mostly very late antigen-4 (VLA-4), leukocyte function-associated molecule-1 (LFA-1), and L-selectin (CD62L).1-2,1-3 This could lead to a modification of the adhesion capacity to the stroma. Furthermore, Papayannopoulou et al1-4 recently suggested that growth factors such as G-CSF could initiate mobilization in neutrophils. In support of this theory, the metalloproteinase gelatinase-B (MMP-9) secreted by neutrophils has been shown to be involved in the cleaving of the extracellular matrix molecules to which stem cells are attached.1-5 G-CSF induced injury to the extracellular matrix and/or downregulation of adhesion molecules could lead to a breakdown in the stem-stromal cell homeostasis and promote egress and shift of the primitive stem cell compartment to the peripheral blood. In concordance with these observations, we have summarized data from the murine model that argue that the true stem cell compartment (uncommitted stem cells) was indeed in a negative balance after a short course of G-CSF.1-6
(2) Marrow harvest is invariably associated with hemodilution, and the timing of G-CSF–primed bone marrow harvest coincides with the peak incidence of peripheral blood progenitors.1-1 We have thus raised the hypothesis that this hemodilution, rather than G-CSF–induced stem cell expansion, could at least be partly responsible for the enhanced engraftment reported in Damiani et al.1-1 Is G-CSF bone marrow priming a misconception?
In their reply, the investigators state that human and murine stem cell physiology are quite different and conclude that there seems to be sound evidence in humans suggesting that G-CSF can indirectly expand human/stem progenitor cell numbers after short-term administration, based on the following evidence that they report. (1) There is an increase in bone marrow cellularity after 5 days of G-CSF treatment of candidates for autologous transplantation. (2) This can be confirmed by magnetic resonance. (3) Several clinical studies have documented cytokine primed marrow as a sufficient autograft without any reported graft failures in more than 100 patients. Lastly, the investigators rule out any significant peripheral blood stem cell contamination that could have partly contributed to enhanced engraftment.
We fully agree that a short-course administration of G-CSF will result in increased marrow cellularity by triggering differentiation along the myeloid lineage of committed progenitors such as colony-forming unit–granulocyte-macrophage (CFU-GM). However, what is at stake here is the primitive CD34+, CD38−, Lin− stem cell compartment that accounts for approximately 1 in 1 × 105 marrow cells. This compartment is obviously difficult to assess through magnetic resonance. At least two studies show that, in humans, a short course of G-CSF administration can lead to a reduction of the primitive marrow stem cell compartment. Mandelam et al1-7 documented a lower incidence of the primitive CD34+, Lin− subset after 3 days of G-CSF administration in the bone marrow of normal volunteers. Likewise, Martinez et al1-8 noted that, although after 4 days of G-CSF administration to normal donors, there was an increase in the total number of nucleated BM cells from a median of 41.5 × 109/L before G-CSF to 56 × 109/L, the proportion of primitive BM hematopoietic progenitors (CD34+ CD38−, CD34+HLDR−, and CD34+ CD117−) decreased during G-CSF administration, as well as the total number of CD34+ cells. Interestingly, Dicke et al1-9 have shown that the administration of G-CSF for 2 days to stage IV breast cancer patients shifted the stem cell balance in favor of differentiation over self-renewal. The primitive HLA-DR+component, although slightly increased after the initial G-CSF injections, returned to baseline values after the second day of cytokine therapy. Concern about further decreases and possible exhaustion of the CD34+ HLA-DR− cells with further G-CSF therapy was even raised by the investigators.
We question how the present investigators can harvest 1 L of marrow during a 1.0- to 2.5-hour procedure of 200 to 300 aspirations without documenting any significant contamination of peripheral blood with aspirates up to a size of 10 to 20 mL.1-10 Contamination of peripheral blood during marrow harvest has been a well-documented phenomenon since the early days of BM transplantation and has been the object of an excellent study by Batinic et al1-11demonstrating significant contamination with peripheral blood. This is a crucial point in this discussion, because it constitutes a potential mechanism for G-CSF primed marrow enhanced recovery. The study of lymphoid subsets in the marrow is useful for evaluating the degree of blood contamination.1-12 Moreover, mobilized peripheral blood stem cells have a markedly distinct immunophenotype characterized by a reduced expression of c-kit, CD11a, CD18, CD49d, and CD62L.1-2,1-3 It would be interesting to assess the incidence of these immunophenotypically distinct stem cell subsets in a normal versus G-CSF–primed marrow.
In summary, most of the presently proposed molecular mechanisms of G-CSF peripheral blood stem cell mobilization and the results of some murine/human data suggest that a short course of G-CSF administration (4 days) has the potential to decrease the primitive marrow stem cell pool. We would like to see more studies focusing on the biology and immunology of G-CSF primed marrow to evaluate its clinical utility and advantages, if any, over mobilized peripheral blood stem cells.