Stem cell mobilisation

Transplantation of haematopoietic stem cells
Autologous stem cell transplantation (ASCT) is widely used for the curative treatment of hematological malignancies and of haemoglobinopathies. The vast majority of ASCTs are performed with the support of peripheral blood stem cells (PBSCs), thus making their mobilisation and collection an important part of ASCT. In fact, the rapid and sustained recovery of the hematopoietic function after ASCT correlates with the number ofCD34+ hematopoietic stem cells infused. CD34+ cells reside mainly in the bone marrow (BM) niche(s) but they can be effectively mobilised to peripheral blood (PB) by the administration of growth factors such as granulocyte colony-stimulating factor (G-CSF) (filgrastim, lenograstim, pegfilgrastim) or granulocyte-macrophage colony-stimulating factor (GM-CSF) (sargramostim) alone or combined with disease-specific chemotherapy (chemomobilisation). The minimum dose of CD34+ cells to provide a high likelihood of successful engraftment is generally considered to be ≥ 2 × 106 cells/kg, whereas the ‘optimal’ number of PBSCs for transplantation is 4–6 × 106 CD34+ cells/kg in both adult and pediatric patients. The finding that higher numbers of re-infused CD34+ cells have been correlated, at least in some studies, with earlier engraftment after transplantation and with better disease-free and overall survival than lower cell doses, has led many transplant centres to attempt the collection of the optimal PBSC number (‘target cell dose’) rather than the minimum dose. CD34+ stem/progenitor cell collection correlates with the absolute number of circulating CD34+ cells prior to the apheresis. Peak mobilisation after G-CSF alone usually occurs 4–5 days after the initiation of G-CSF, whereas peak mobilisation following chemotherapy-based regimens is more variable and may occur 10–20 days from the start of chemotherapy. A significant proportion of cancer patients eligible for ASCT fails to mobilise a sufficient number of CD34+ hematopoietic stem/progenitor cells due to various pre-mobilisation (predictive) factors such as prior treatment with stem cell toxic drugs, underlying disease, age, prior radiotherapy and BM involvement. The failure rate with current strategies in adults is estimated to range from 5% to 40%,  leading to repeated apheresis sessions, suboptimal grafts associated with delayed hematopoietic recovery, need for re-mobilisation and, sometimes, to treatments other than ASCT. The percentage of “poor mobilisers” across different studies is variable depending on definitions, disease categories and lack of standard mobilisation and collection practices, so that there are no commonly accepted criteria to define the success/failure rates. Thus, there is a medical need of more effective mobilisation strategies for patients with advanced or relapsed lymphomas or patients with MM who may be successfully treated with high-dose chemotherapy followed by ASCT.

Strategies for PBSC mobilisation based on growth factors: risks and benefits
G-CSF (e.g., filgrastim, lenograstim) are the only approved mobilisation agents in Europe for both adult and pediatric patients. Recent data demonstrate that over 80% to 90% of all ASCT worldwide are performed using either cytokine- or chemotherapy - followed - by - cytokine - mobilised PBSCs.

G-CSF Alone
The approved dosing for non-pegylated G-CSF for stem cell mobilisation is 10 μg/kg s.c., although some investigators use it at higher doses (i.e., up to 32 μg/kg s.c. daily) to rescue poor mobilisers. G-CSF is initiated 4 days prior to the first apheresis session and its administration is continued until the last day of apheresis. CD34+ cell levels in the blood usually peak on the fifth day of G-CSF. The reported total yield of collected CD34+ cells across a number of controlled studies ranged from 2.5 to 5.8 × 106/kg (median values) during a median of two to five apheresis sessions. The addition of chemotherapy to G-CSF increases yields at the expense of more side-effects, although the reported failure rates (defined as CD34+</SUP> cell yields of <2.0 × 106/kg) are not different between the two treatments, with failures rates of up to 23%. After transplantation, the median time to granulocyte engraftment with G-CSF alone has been reported to be 11 days, and for platelet engraftment approximately 11–14 days. G-CSF is generally well tolerated. Common side effects include bone pain, headache, anaemia and decreased platelet counts. Rare but potentially fatal splenic rupture has also been reported. In addition, screening for thrombophilia is recommended in normal donors who report a familiar or personal history of previous thrombosis due to some suggestions of thrombotic events during PBSC mobilisation with G-CSF in healthy donors.

Pegylated Granulocyte Colony-Stimulating Factor (Pegylated G-CSF)
The potential of the pegylated form of G-CSF (pegfilgrastim), a longer-lasting variant of G-CSF, to mobilise PBSCs has been investigated in clinical trials. Its long plasma half-life of 33 hours makes a single dose sufficient to induce stem cell mobilisation, whereas G-CSF with a plasma half-life of 3 to 4 hours must be administered daily. The safety profile of pegfilgrastim is similar to that of G-CSF and like all current mobilisation methods there is a significant failure rate of around 25%. Interestingly, both filgrastim and pegfilgrastim are widely recognized as regulators of the immune system by mainly inducing modulatory cells.

Granulocyte-Macrophage Colony-Stimulating Factor (GM-CSF)
GM-CSF is used less often than G-CSF, and only in USA, for PBSC mobilisation because it is less efficient (both when given alone and in combination with chemotherapy) and has a more unfavourable safety and tolerability profile than G-CSF. GM-CSF is sometimes used in combination with G-CSF in patients who failed an initial mobilisation attempt.

Chemomobilisation
Most mobilisation regimens combine treatment with G-CSF (and rarely GM-CSF) after administration of a disease-specific chemotherapy regimen to achieve higher CD34+</SUP> cell yields than treatment with G-CSF alone, both in patients with MM and NHL (although failure rates with G-CSF plus chemotherapy seem to be as high as with G-CSF alone). For instance, Moskowitz et al. reported that mobilisation with G-CSF alone (10 μg/kg daily) yielded 1.5 × 106/Kg CD34+</SUP> cells compared with 6.7 × 106/Kg CD34+</SUP> cells when chemotherapy plus G-CSF were used. Additional benefits of chemomobilisation include fewer number of required apheresis sessions compared to G-CSF alone. More importantly, there is indication that chemomobilisation, particularly in lymphoma, reduces, in vivo, the tumour load and tumour cell contamination in the apheresis product. In fact, PBSC mobilisation is often part of a cycle of induction or salvage treatment for lymphoma patients thus avoiding additional costs and risks associated with the use of unnecessary chemotherapy for mobilisation. Chemomobilisation is also commonly used in MM using a single dose of cyclophosphamide. In this case, the benefit of higher cell yields (than with G-CSF alone) may be offset by less predictability of timing and an increased risk for the patient (i.e., increased morbidity, greater risk of infection and febrile neutropenia, more hospital admissions, transfusions, antibiotic therapy, and drug-specific toxicities) without any well documented anti-tumor effect. One potential problem related to the use of chemotherapy is that PBSC mobilisation is less predictable and may vary substantially between patients. Thus, is it necessary to monitor leukocytes and CD34+</SUP> cell counts over several days to determine when to begin apheresis.3 Overall, the addition of a myelosuppressive regimen to a cytokine may result in a higher cell yield than cytokine alone, but this result needs to be balanced against the increased risks for the patient and the greater resource utilization unless chemotherapy is part of the treatment strategy.

Definition of “poor mobiliser” and risk factors
As mentioned, the definition of “poor mobiliser” varies according to different parameters analyzed to evaluate PBSC mobilisation: peak of CD34+</SUP> cells in PB, fold-increase of circulating CD34+</SUP> cells, CD34+</SUP> cells collected, number of candidate patients undergoing ASCT. As a consequence, different criteria have been proposed to define a successful PBSC mobilisation and the adequate apheresis yield, but these data are difficult to analyze and compare to each other. The extensive review of predictive factors for poor mobilisation is beyond the scope of this article (see    However, it should be kept in mind that in addition to baseline parameters, during- and post-mobilisation factors have been poorly exploited due to the lack, so far, of rescue strategies. For instance, febrile neutropenia is one major complication after administration of mobilising chemotherapy. The release of pro-inflammatory cytokines may negatively affect stem cell proliferation and mobilisation. Furthermore, genetic factors as well as polymorphisms in cytokine gene receptors are believed to be responsible for the great variability in mobilisation responses in allogeneic donors. In patients receiving chemomobilisation, slow leukocyte and platelet recovery as well as anemia indicate poor marrow function. However, type and dose of chemotherapy may influence the risk of mobilisation failure as severe thrombocytopenia induced by alkylating agents administered during mobilisation can be a risk factor for mobilisation failure while high-dose cytarabine mobilisation regimen often induces severe thrombocytopenia and neutropenia without negatively affecting stem cell mobilisation. Other factors predicting mobilisation failure are: delayed or anticipated timing of apheresis (due to insufficient circulating stem cells monitoring) and/or small volume of processed blood which may affect PBSC collection even in patients showing a satisfactory peak of CD34+</SUP> cells in the PB. For these reasons, a working group promoted by GITMO (Italian Group for Stem Cell transplantation) proposed the definition of “poor mobiliser” identifying “proven poor mobiliser” and “predicted poor mobiliser”. In order to develop criteria for the definition of “poor mobiliser”, the working group used the analytic hierarchy process (AHP) which had been developed to establish priorities and to make the best decision when both the quantitative and qualitative aspects of a decision need to be considered and a poor information base is available. AHP is a multistep process that includes four major phases: 1) defining the goal; 2) decomposing the problem and identifying critical issues; 3) categorizing/framing the main criteria; 4) defining a hierarchy of the criteria. GITMO panel selected two conceptual criteria to identify the “proven poor mobiliser”: the peak of circulating CD34+</SUP> cells during mobilisation and the absolute number of harvested CD34+</SUP> cells. All participants agreed that pre-apheresis CD34+</SUP> count in PB is the best predictor of CD34+</SUP> cells in the aphaeresis products     and, operationally, considered a peak of CD34+</SUP> cells >20 μl in PB, as a reliable indicator of a satisfactory mobilisation ability. Moreover, the GITMO panel identified 2.0×106 CD34+</SUP> cells/kg as the minimum safe dose for ensuring rapid neutrophil and platelet recovery both in lymphoma and in MM patients to be achieved with a maximum number of 3 aphereses. These parameters and indicators applied to both chemomobilisation and G-CSF alone strategies although the timing of CD34+</SUP> cells peak and doses of G-CSF are different and should be considered. Furthermore, GITMO panel selected 3 major and 5 minor criteria to identify the “predicted poor mobiliser”. The most important criteria were felt to be: previous cytotoxic chemotherapy, irradiation on BM bearing bones and failure of previous mobilisation attempt. Among the other factors associated with unsuccessful mobilisation, GITMO panel selected advanced phase disease (i.e. at least 2 prior cytotoxic lines), refractory disease, extensive BM involvement at mobilisation, BM cellularity <30% at mobilisation and age >65 years as minor criteria. The proposed definitions should be validated in prospective clinical studies. In conclusion, poor mobilisation of PBSCs is a major limitation for patients eligible for ASCT. The availability of new drugs, aimed at optimizing PBSC mobilisation, requires a stringent definition of “poor mobilisation”. In this view, GITMO panel recommended that patients previously failing at least one mobilisation attempt should be candidate for new mobilising strategies. In addition, the use of standard criteria for identifying both the “proven and the predicted poor mobiliser” before planning the use of new mobilising agents was recommended. To this end, the GITMO working group tried to define simple, but stringent operational criteria for the identification/prediction of “poor mobiliser” in the setting of lymphoproliferative diseases.

New approaches to optimize HSC mobilisation
In adult life, the chemokine receptor CXCR4 and its ligand stromal cell-derived factor-1 (SDF-1/CXCL12) are critically regulating the retention of hematopoietic stem cells in the BM. Under physiological conditions (i.e. in absence of “danger signals”) the release of hematopoietic stem cells from the BM occurs infrequently and follows a circadian loop. Tissue damage, infections or flogosis induce the exit of stem cells from the BM to contribute to tissue repair. Disruption of the CXCR4/CXCL12 axis in the BM, which can be directly achieved by CXCR4 antagonists or indirectly by G-CSF through the development of a proteolytic enviroment, increases the motility of hematopoietic stem cells and their egress from the BM. Plerixafor (formerly AMD3100) is a CXCR4 chemokine antagonist that has been shown to increase the number of circulating CD34<SUP>+</SUP> cells in healthy volunteers and cancer patients alone or with G-CSF. The key feature of this chemokine receptor/ligand interaction is the rapidity of the mobilisation process and stands in clear contrast with G-CSF-based mobilisation where up to four days of treatment are required before the significant increase of circulating CD34<SUP>+</SUP> cells is observed. Consistent with its antagonistic activity on the CXCR4 receptor, plerixafor also increases the number of circulating leukocytes. There are two distinct phases of stem cell mobilisation according to different routes of administration: the peak occurs approximately four hours after intravenous injection while 10–12 hours are required for stem cell release from the BM after subcutaneous administration. By 24 hours the mobilising effect of plerixafor is returned to baseline or close to baseline. Therefore, the rapid biological activity of plerixafor allows its administration “on demand” without planning the timing of administration in advance. Early studies in patients with NHL and MM suggested the superiority of G-CSF plus plerixafor over G-CSF alone in regard to mobilisation efficiency. More recently, a compassionate use study including 115 patients who had failed at least one previous mobilisation attempt showed a success rate for re-mobilisation with G-CSF plus plerixafor of 60% for NHL, 71% for MM and 76% for HD. Similar results have been shown in an European compassionate use study including 56 patients with lymphoma or MM, where the success rate was 75%. In two phase III randomized placebo-controlled studies in MM and NHL patients, the combination of G-CSF plus plerixafor was found to be safe and superior in terms of mobilisation efficacy as compared to G-CSF plus placebo. In MM patients randomized to G-CSF plus plerixafor, 71.6% of the patients achieved the primary study endpoint (collection of at least 6 × 106/kg CD34<SUP>+</SUP> cells with less or equal to two aphereses) compared to only 34.4% of patients receiving G-CSF and placebo. Similarly, 59% of NHL patients achieved the primary study endpoint (collection of at least 5 × 106/kg CD34<SUP>+</SUP> cells with less or equal to four aphereses) compared to only 19.6% of patients mobilised with G-CSF plus placebo. plerixafor-mobilised PBSCs did show rapid and sustained engraftment after high-dose therapy in both studies. Noteworthy, Maziarz and co-workers performed a post-hoc analysis based on data from the randomized trial of plerixafor + G-CSF vs. placebo + G-CSF in NHL patients. The investigators evaluated the efficacy of the addition of plerixafor to G-CSF on the evening of day 4 in patients with pre-plerixafor circulating CD34<SUP>+</SUP> cell count < 10 × 106/l, to achieve the collection of the minimum (≥ 2 × 106/kg) or the target (≥ 5 × 106/kg) cell dose. The results demonstrated that patients who had been randomized to receive plerixafor in addition to G-CSF showed a 6 fold-increase of PB CD34<SUP>+</SUP> cells on day 5 compared to only 1.6 fold-increase in patients receiving G-CSF and placebo. These data resulted in a significantly higher cumulative number of CD34<SUP>+</SUP> cells after 2 apheresis days in plerixafor-treated patients as compared to placebo patients (2.92 vs. 0.94 × 106/kg). Overall, 78% of patients in the plerixafor + G-CSF group achieved the primary end point compared to only 34.2% in the control group. Thus, the addition of plerixafor to G-CSF enabled the collection of the minimal transplantable dose in the majority of patients with a PB CD34<SUP>+</SUP> cell count < 10 × 106/L on day 4. A statistically significant increase in PBSC collections was also obtained in patients mobilised with G-CSF plus plerixafor and with PB CD34<SUP>+</SUP> count < 20 × 106/L on day 4. Taken together, these results provide a clear example of the potential of ‘early intervention’ with novel strategies to rescue cancer patients who can be considered ‘proven poor mobilisers’ as they have < 10–20 × 106/L PB CD34<SUP>+</SUP> cells at the peak time of mobilisation after G-CSF mobilisation. However, clinical studies involving the use of plerixafor in children are needed to confirm its potential for PBSC mobilisation in this patient population.

Plerixaflor combined with chemomobilisation
At present chemomobilisation is considered the mobilisation standard in many transplant centres especially in lymphoma patients. However, published results indicate that the addition of chemotherapy to G-CSF does not prevent poor mobilisation. Limited data is available on the effects of the administration of plerixafor added to chemomobilisation to enhance the mobilisation of PBSCs. Dugan and co-workers evaluated prospectively the safety and efficacy of plerixafor combined with chemotherapy and G-CSF in an open-label, multicenter trial. In this study, 40 patients (26 MM, 14 NHL) received various chemotherapy regimens followed by G-CSF plus plerixafor. The mean fold-increase of PB CD34<SUP>+</SUP> cells was 1.7 fold after plerixafor. The combination was well-tolerated. However, based on the results published on the peak number of circulating CD34<SUP>+</SUP> cells and apheresis yields, most of the patients could not be considered as hard-to-mobilise as pre-plerixafor median PB CD34<SUP>+</SUP> counts were 33 × 106/L in NHL patients and 150 × 106/L in MM patients, respectively. Recently, the addition of plerixafor to chemotherapy plus G-CSF mobilisation was tested in patients who mobilise poorly (i.e. re-mobilisation or first mobilisation with low blood CD34<SUP>+</SUP> counts or poor collection yields). Based on the mechanism of action, plerixafor causes a rapid release (5–11 hours) of CD34<SUP>+</SUP> cells from the BM to circulation, which makes the drug suitable for pre-emptive or ‘on demand’ use in patients who are hard-to-mobilise. Patients were classified as ‘poor mobilisers’ based on daily monitoring of PB CD34<SUP>+</SUP> cell counts during the recovery phase after chemotherapy and G-CSF and/or the collection of PBSCs was felt to be inadequate to proceed to ASCT. By considering only the studies with more detailed information available,   28 out of 34 patients (85%) collected ≥ 2 × 106/kg CD34<SUP>+</SUP> cells after the first mobilisation attempt with a median of two plerixafor injections. The analysis of published data suggests that in poor mobilisers plerixafor may not be effective in inducing CD34<SUP>+</SUP> cell mobilisation when the leukocyte count is very low. Therefore, the critical issue of the optimal timing for plerixafor addition cannot be addressed conclusively. Too early addition of plerixafor may not be cost-effective as many patients may be successfully collected by waiting 1–2 days especially if PB CD34<SUP>+</SUP> cell and leukocyte counts are rising. On the other hand, waiting too long may be deleterious as the mobilisation induced by chemotherapy plus G-CSF may diminish and hence late addition of plerixafor might be less effective. Thus, future studies should test prospectively well-defined algorithms, perhaps based on leukocyte and CD34<SUP>+</SUP> cell counts and/or the results of first day collection, to optimize the use of plerixafor after chemomobilisation. Altogether, few patients reported in four small series  mobilised with chemotherapy/G-CSF plus plerixafor have been transplanted so far. Only two patients, who received grafts containing 1.8 and 2.1 × 106 CD34<SUP>+</SUP> cells/kg, respectively, were reported to have slow platelet engraftment. In the German series, all 24 patients mobilised with chemomobilisation plus plerixafor engrafted. Thus, based on the study by Dugan et al. and five patient series on add-on use of plerixafor after chemomobilisation, this combination appears to be safe and no major side-effects attributable to plerixafor have been reported. In addition, plerixafor-mobilised PBSCs did show rapid and sustained engraftment. This finding supports the data from randomized phase III studies showing that patients mobilised with G-CSF plus plerixafor have stable and sustained engraftment after high-dose therapy. Beside stem cell mobilisation, few additional topics related to plerixafor administration should be mentioned. It is known that mobilisation with G-CSF plus plerixafor results in different graft composition when compared to G-CSF alone mobilisation, including more CD34<SUP>+</SUP>CD38− cells as well as more NK-cells and T cells. There are no data on the graft content, other than CD34<SUP>+</SUP> cell dose in patients mobilised with a combination of chemotherapy, G-CSF and plerixafor. As graft content may be of importance also for immune reconstitution and long-term patient outcomes, this issue deserves further studies. These differences in graft composition might also contribute to the combination of plerixafor with G-CSF performing better in thalassaemic patients than other combinations tested. G-CSF alone has proven inefficient and potentially hazardous for collection of haematopoietic stem and progenitor cells from patients with haemoglobinopathies. The combination of G-CSF with plerixafor, however, resulted in long-term multilineage engraftment in immudeficient mice and in improved short-term engraftment compared to other mobilisation treatments. For gene-therapy applications using lentiviral vectors, the combination moreover resulted in long-term persistence of high vector copy numbers in virus-treated cells and in higher levels of vector-derived gene expression for a given VCN, indicating mobilisation of CD34+ by plerixafor combined with G-CSF as a superior graft for stem-cell based gene therapy of β-thalassaemia. Plerixafor is not recommended in patients with acute myeloid leukemia due to the mobilisation of leukemic cells into circulation, while available data indicate that the use of plerixafor is not associated with increased mobilisation of tumour cells in myeloma or lymphoma patients. The activity of plerixafor on BM microenviroment and the subsequent release of leukemic cells has been recently exploited therapeutically to enhance the antileukemic effect of conventional chemotherapy in resistant/relapsed patients (chemosensitization).