Progenitor Cell Mobilization Effect of Prior Chemotherapy on Hematopoietic Stem Cell Mobilization

Progenitor cell mobilization Effect of prior chemotherapy on hematopoietic stem cell mobilization

CD Ford, W Green, S Warenski and FB Petersen

Utah Blood and Bone Marrow Transplantation Program, LDS Hospital and University of Utah, Salt Lake City, UT USA

Summary: The identification of factors that correlate with a patient’s ability to obtain high concentrations of circulating CD34 þ A number of studies have suggested that prior chemo- cells during the mobilization process has been the objective therapy correlates negatively with the efficiency of of a number of prior studies. These have disagreed over the hematopoietic stem cell mobilization. However,little data influence of age,1–4 gender,4–6 interval from last chemother- exist with regard to the relative effects of the specific apy,1,3,4,7,8 bone marrow involvement,1,3,4,8,9 tumor pri- chemotherapeutic drug classes. We retrospectively re- mary,5,9,10 disease status,3–5 and prior radiation,3,4,9,11,12 viewed the records of 201 consecutive patients with suggesting that these are, at best, weak predictive factors. nonmyeloid malignancies undergoing CD34 þ cell mobi- Overall, however, one of the most consistent negative lization with chemotherapy þ granulocyte colony-stimu- influences on mobilization efficiency has been the amount lating factor (G-CSF). The number of prior chemotherapy of prior chemotherapy. courses correlated negatively with the peripheral CD34 þ Although the number of courses of prior chemotherapy cell concentration (pCD34) on the first day of collection usually correlates negatively with mobilization efficiency, (Po0.001). No significant correlation was found for age, rather minimal and conflicting data exist with regard to the gender,tumor primary,mobilization chemotherapy regi- relative effects of the specific chemotherapeutic drug classes. men,disease status,marrow involvement,prior radiation For example, while some studies have suggested that therapy,or dose and timing of G-CSF administration. alkylating agents may be more damaging,12,13 other analyses When the number of courses of individual classes of have yielded no differences.3 In this study, we have retro- chemotherapeutic agents was correlated with pCD34,only spectively analyzed the effect of prior chemotherapy with the exposures to platinum compounds (P ¼ 0.001) and alky- various drug classes on the peripheral blood CD34 þ cell lating agents (P ¼ 0.01) were found to be independent concentration (pCD34) at the initiation of harvest. In order negative predictive factors for pCD34. Within classes, to minimize the effect of the mobilization regimen, we DNA crosslinking agents and etoposide appeared possibly restricted the analysis to patients receiving chemotherapy and more damaging than DNA methylating agents and granulocyte colony-stimulating factor (G-CSF). doxorubicin,respectively. None of the drug classes showed evidence of recovery. We conclude that exposure to Methods chemotherapy,especially platinum compounds and alkylating agents,should be minimized prior to mobilization. Patients Bone Marrow Transplantation (2004) 33, 901–905. doi:10.1038/sj.bmt.1704455 Data from 201 consecutive adult patients with nonmyeloid Published online 8March 2004 malignancies and available peripheral CD34 þ cell con- Keywords: hematopoietic stem cell; mobilization; che- centrations on the day of first collection were included. All motherapy; CD34 patients had undergone mobilization with chemotherapy and recombinant G-CSF at LDS Hospital or the University of Utah. Most patients had received G-CSF 10 (n ¼ 161) or High-dose therapy with peripheral blood stem cell (PBSC) 6(n ¼ 33) mg/kg/day subcutaneously beginning on day 5 support has been successfully used in the management of a (n ¼ 100) or 6 (n ¼ 79) following the initiation of chemo- number of malignancies. PBSCs are usually mobilized from therapy. Complete blood counts were followed daily and, the marrow by the action of hematopoietic growth factors when the WBC had risen to greater than 800 cells/ml, daily (with or without chemotherapy), harvested by a cell pCD34 was determined. PBSC collections were begun at separator, and quantified by CD34 þ cell enumeration. the discretion of the attending physician, usually following a recovery of the peripheral white blood count (pWBC) to 45.0 Â 109 cells/l (median: 10 Â 109 cells/l) and a pCD34 410 Â 106 cells/l (median: 60 Â 106 cells/l). Correspondence: Dr CD Ford, Blood and Bone Marrow Transplant Program, LDS Hospital, Eighth Avenue and C Street, Salt Lake City, UT 84143, USA; Prior chemotherapeutic agents E-mail: [email protected] Received 2 October 2003; accepted 5 December 2003 Data on prior chemotherapy were obtained by chart Published online 8March 2004 review. Chemotherapeutic agents received by the patients Stem cell mobilization CD Ford et al 902 were divided into seven classes: platinum analogs, alkylat- Table 1 Characteristics of the 201 patients ing agents, topoisomerase interactive agents, antimicro- Age (years) 46 (12)a tubule agents, antimetabolites, corticosteroids, and Gender (F/M) 122/79 bleomycin. The individual drugs were usually administered in combination and a course of the combination therapy Diagnosis was considered a course for the individual agent. For the Breast cancer 83 Lymphoma 67 small number of instances of oral continuous administra- Myeloma 31 tion, 1 month of therapy was considered a course. Ovary 11 Other 9 #Prior chemotherapy courses 6.1 (3.3)a CD34 þ cell determination #Prior radiation courses 0.27 (0.51)a

CD34 þ cells were enumerated in the peripheral blood by Mobilization chemotherapyb flow cytometry utilizing a Beckman Coulter Epics XL- C30 MCL according to the recommendations of the Interna- CT 4 tional Society of Hematotherapy and Graft Engineering14 CE 38 modified for three-color fluorescence. CTP 49 ESHAP/DHAP 15 IcbE 11 Statistical analysis Standard-dose chemotheapy 14 Statistical analyses were performed with SPSS Version 10.1 Mobilization G-CSF G-CSF dose(mg/kg/Day) 9.3 (1.8)a or higher. Pearson’s correlation (continuous variables) or Day G-CSF started 5.4 (1.1)a one-way analyses of variance (ANOVA) (categorical variables) were utilized to determine the significant aMean(s.d.). associations with pCD34. Pairwise comparisons were bC ¼ cyclophosphamide; T ¼ paclitaxel; E ¼ etoposide; P ¼ cisplatin; performed if ANOVA or ANCOVA were significant at Cb ¼ carboplatin; I ¼ ifosphamide. the 0.05 level. Univariate regression models and a parsimonious multivariable regression model were created to investigate the effect of the chemotherapy classes on exposure to platinum compounds and alkylating agents pCD34. Results were considered significant if Po0.05 had the strongest negative effect on pCD34. Only the (two-sided). number of platinum courses was more predictive than the total number of chemotherapy courses. Within drug classes, DNA crosslinking agents appeared more damaging Results than methylating agents and DNA-intercalating agents (doxorubicin) than nonintercalating topoisomerase inhibi- The characteristics of the 201 patients are summarized in tors (etoposide). On a multivariate analysis, which included Table 1. The patients had received a mean of 6.173.3 the total number of chemotherapy courses, only the premobilization chemotherapy courses utilizing various number of prior courses of platinum compounds agents. In all, 40 patients were mobilized following (P ¼ 0.001) and alkylating agents (P ¼ 0.01) were found to standard-dose salvage chemotherapy used in the treatment be independent negative predictive factors for pCD34. of their specific primary tumors, while the remainder Regression analysis suggested that each course of platinum received various intermediate-dose chemotherapy regimens decreased the pCD34 on the first day of collection by 18 designed for mobilization. The patients were relatively cells/ml, and each alkylating agent course by 6.6 cells/ml. uniform in the dosage and timing of G-CSF. The potential for recovery from the effects of the The specific chemotherapeutic agents received by the 201 different drug classes was investigated by correlating the patients prior to mobilization are summarized in Table 2. time from the last drug administration with the pCD34 on Cyclophosphamide and doxorubicin had been used most the first day of harvest (Table 5). Although some of the often. Some agents such as fludarabine and venorelbine patients had relatively long intervals, we could find were received too infrequently for a significant evaluation evidence of recovery for none of the drug classes. of their effects on pCD34. The correlation of several possible predictive factors for pCD34 in our patient population is shown in Table 3. The Discussion mobilization chemotherapy regimen, G-CSF dose, and G-CSF schedule did not correlate with pCD34. A statistical In this study, we have examined a number of factors trend was noted for prior radiation. However, only the potentially related to the ability to obtain higher levels of number of prior chemotherapy courses was significant pCD34 at the initiation of harvest. Of several possible end (P ¼ 0.001). Regression modeling suggested that each points for mobilization efficiency, we chose pCD34 on the course of chemotherapy resulted in a decrease of 9.4 cells/ml first day of harvest as the most direct measurement. For for the pCD34 on the first day of harvest. practical reasons, many other studies have utilized the The relation between pCD34 and number of prior harvest yield. We and others have shown a high correlation courses of drugs from the various chemotherapeutic agent between these two end points.15 However, we have also classes is shown in Table 4. On univariate analysis, prior suggested previously that the end point of harvest yield

Bone Marrow Transplantation Stem cell mobilization CD Ford et al 903 Table 2 Numbers of patients exposed to individual chemother- Table 3 Signiﬁcance of patient characteristics in prediction of apeutic classes and agents pCD34

Drug class Drug #Patients #Courses of P-value exposed drug class/ patient exposed Age (years) 0.15 (mean(range)) Gender (F/M) 0.23 Patient weight (kg) 0.31 Platinum analogs 63 2.9 (1–14) Primary tumor 0.46 Cisplatinum 382.3 (1–8) Patient disease status (7NED) 0.31 Carboplatinum 24 2.9 (1–8) Active marrow involvement 0.47 #Prior chemotherapy courses 0.001 Alkylating agents 144 6.1 (1–23) # Prior radiation courses 0.075 Crosslinking 137 5.7 (1–23) Mobilization chemotherapy regimen 0.34 Cyclophoshamide 116 5.4 (1–16) Dose of G-CSF 0.21 Ifosphamide 23 2.3 (1–9) Day G-CSF started 0.29 Melphalan 9 7.0 (1–17) Nitrogen mustard 3 2.0 (1–3) Chlorambucil 1 3.0 Carmustine 1 2.0 Methylating 19 5.4 (2–8) doses of G-CSF received by our patients is not sufficient to DTIC 184.9 (2–8) show a dose effect in a study of our sample size. If so, our Procarbazine 5 3.0 (1–6) results would be consistent with those of Andre et al,20 who found no significant difference in CD34 þ cell harvest Topoisomerase 174 5.3 (2–14) inhibitors yields between 131 patients randomized to receive 5 or Doxorubicin 1684.8(1–11) 10 mg/kg/day of G-CSF following mobilization chemother- Etoposide 39 2.2 (1–6) apy. That doses greater than 10 mg/kg/day may be superior Mitoxantrone 6 2.7 (1–6) to lesser doses, or that wider disparity in the doses administered compared to the 5 mg/kg/day difference in Microtubular 1185.1 (1–11) Paclitaxel 20 3.9 (1–8) the study arms in the trial of Andre et al, are needed to Docetaxel 3 3.0 (2–4) show an effect was suggested by the small study of Demirer Vincristine 87 4.9 (1–11) et al,19 which found superior collections following 16 as Vinblastine 16 4.9 (2–8) compared to 8 mg/kg/day following mobilization chemo- Venorelbine 3 2.0 (2.0) therapy. Antimetabolites 69 5.4 (1–18) It has also been suggested that patient case mix may Methotrexate 17 5.4 (1–9) affect the sensitivity of detecting significant differences 5-fluorouracil 385.8(2–12) among groups receiving varying doses of G-CSF. Thus, Cytosine 282.1 (1–6) using G-CSF alone, Weaver et al18 noted that the arabinoside Fludarabine 1 6.0 differential effect of an increased dose (30 vs 10 mg/kg/ day) on CD34 þ cell collections seemed to be greater in Corticosteroids 91 6.1 (1–23) patients who had received lesser doses of premobilization chemotherapy. However, this result was not seen in the Bleomycin 22 4.7 (2–8) study of Andre et al. These findings may suggest that differences between harvest results from good and poor risk patients are minimized with chemotherapy þ G-CSF does add additional confounding variables not related to mobilization regimens and/or use of G-CSF doses mobilization such as cell separator collection efficiency.16 o10 mg/kg/day such as were received by our patient We also attempted to design our study in a manner that population and that of Andre et al. would minimize the effect of mobilization regimen vari- In agreement with previous studies,3,4,6–8 the amount of ables as confounding factors. As randomized trials have prior chemotherapy is a significant predictor for mobiliza- demonstrated the superiority of chemotherapy þ growth tion efficiency in our study. Prior radiation therapy appears factors over growth factors alone,17 we included only to have less effect, a result that conflicts with some3,4,9 but patients receiving both chemotherapy and G-CSF. Rando- not other11,12 studies. However, our ability to detect an mized trials have also suggested that the dose of G-CSF influence was likely related to the use of radiation in only correlates with the mobilization efficiency when G-CSF is 24% of our patients, and only one course of radiation was used alone18 or following chemotherapy,19 although the received by 88% of these. While such limited radiation latter finding has not been consistent.20 appears to have little impact, our results do not rule out a In our patient population, the G-CSF dose did not negative effect from more extensive treatments. correlate with pCD34. There are several possible reasons More data are needed with regard to the effects of why such an effect, if it exists, may have been more difficult individual classes of drugs received prior to stem cell to detect. The first is that our patient population is mobilization. Drake et al21 proposed a scoring system relatively homogeneous in terms of the G-CSF dose based largely on murine studies that evaluated the received. Thus, 80% of our patients received a dose of repopulating ability of bone marrow cells that had been 10 mg/kg/day and all but seven patients received doses of 6 exposed to various chemotherapeutic agents. In general, or 10 mg/kg/day. A second reason may be that the range of Drake’s system classified DNA alkylating agents as most,

Bone Marrow Transplantation Stem cell mobilization CD Ford et al 904 Table 4 Univariate analysis of the correlation of pCD34 with CD34 þ cells is not entirely clear. Studies in rodents have exposure to different classes of chemotherapy agents suggested that alkylating agents and platinum analogs can Drug class #Courses Pearson’s P-value produce a sustained reduction in the number of effective correlation marrow stem cells as evidenced by decreased repopulating ability.23 While this may result, at least in part, from a Platinum 0.9(1.8)a À0.249 o0.001 direct effect on stem cells, cyclophosphamide has also been Total #chemo courses 6.1(3.3) À0.233 0.001 shown to negatively affect the ability of the marrow stroma Alkylating agents 4.4(3.9) À0.194 0.006 24,25 DNA crosslinking agents 3.9(3.7) À0.156 0.027 to support hematopoietic progenitors. DNA methylating agents 0.5(1.8) À0.091 0.20 We conclude that prior chemotherapy is the most Topoisomerase inhibitors 4.6(2.8) À0.150 0.034 important potentially controllable variable for stem cell Etoposide 0.4(1.1) À0.185 0.008 mobilization. Minimization of exposure to chemotheraputic Adriamycin 4.1(2.4) À0.072 0.31 Corticosteroids 2.8(3.8) À0.142 0.044 agents, especially platinum analogs and DNA crosslinking Microtubular agents 3.0(3.1) À0.116 0.10 agents, should increase CD34 þ cell harvest yields. Taxanes 0.4(1.5) À0.125 0.077 VCR/VBL 2.5(3.0) À0.054 0.44 Antimetabolites 1.8(3.7) 0.009 0.90 Bleomycin 0.5(1.6) À0.091 0.20 References

aMean(s.d.) for the 201 patients. 1 Kotasek D, Shepherd KM, Sage RE et al. Factors affecting blood stem cell collections following high-dose cyclophosphamide mobilization in lymphoma, myeloma and solid Table 5 Correlation between time from last chemotherapy class tumors. Bone Marrow Transplant 1992; 9: 11–17. and pCD34 2 Chabannon C, Le Coroller AG, Faucher C et al. Patient Drug class Mean time Pearson’s P-value condition affects the collection of peripheral blood progenitors (days) correlation after priming with recombinant granulocyte colony-stimulating factor. J Hematother 1995; 4: 171–179. Last platinum 78(134)a À0.143 0.26 3 Koumakis G, Vassilomanolakis M, Hatzichristou H et al. Last alkylating agent 303(642) À0.036 0.67 Predictive factors affecting mobilization and peripheral blood Last topoisomerase 163(455) À0.071 0.35 stem cell (PBSC) collection using single apheresis (SA) for Last tubular 288(635) À0.057 0.54 rescuing patients after high-dose chemotherapy (HD, CHE) Last antimetabolite 318(821) À0.073 0.55 in various malignancies. Bone Marrow Transplant 1996; 18: Last steroids 178(560) À0.052 0.63 Last bleomycin 599(803) 0.016 0.95 1065–1072. 4 Haas R, Mo¨ hle R, Fru¨ hauf S et al. Patient characteristics aMean(s.d.). associated with successful mobilizing and autografting of No chemotherapy class shows evidence of recovery. peripheral blood progenitor cells in malignant lymphoma. Blood 1994; 83: 3787–3794. 5 Keung YK, Cobos E, Dunn D et al. Determining factors for platinum analogs and topoisomerase inhibitors as inter- the outcome of peripheral blood progenitor cells harvests. J Clin Apheresis 1996; 11: 23–26. mediate, and antitubular agents as least damaging. Both 21 22 6 Humpe A, Riggert J, Vehmeyer K et al. Comparison of Drake et al and Clark and Brammer found a correlation CD34+ cell numbers and colony growth before and after between the scoring system and CD34 þ cell collections. cryopreservation of peripheral blood progenitor and stem cell Some problems of the Drake’s system, for current use, harvests: influence of prior chemotherapy. Transfusion 1997; include a lack of initial grounding in human patient data, 37: 1050–1057. exclusion of antimetabolites (especially when 5-FU, a 7 Olivieri A, Offidani M, Ciniero L et al. Optimization of the commonly used drug, had shown adverse effects in the yield of PBSC for autotransplantation mobilized by high-dose murine studies that formed the basis of their system), and a chemotherapy and G-CSF: proposal for a mathematical lack of inclusion of newer and frequently used agents such model. Bone Marrow Transplant 1994; 14: 273–278. as the taxanes. 8Shepherd KM, Charles P, Sage RE et al. Mobilization of haemopoietic stem cells by cyclophosphamide into the In order to address some of these issues, we performed a peripheral blood of patients with haematological malignancies. retrospective study correlating the exposure to seven Clin Lab Haematol 1991; 13: 25–32. chemotherapeutic drug classes in common use with 9 Bensinger W, Appelbaum F, Rowley S et al. Factors that mobilization efficiency. Our results revealed some simila- influence collection and engraftment of autologous peripheral- rities and some differences with those of Drake et al. In our blood stem cells. J Clin Oncol 1995; 13: 2547–2555. patients, prior exposure to DNA crosslinking agents and 10 Imrie K, Stewart K, Crump RM et al. Blood stem cell platinum analogs appeared to carry the greatest risk of collection: factors influencing the recovery of granulocyte– impairment, while antimetabolites seemed to have little, if macrophage colony forming cells. Cell Transplant 1996; 5: any, effect. In addition, we could find no evidence of 379–383. recovery over time for any of the drug classes. This result is 11 McQuaker IG, Haynes AP, Stainer C et al. Stem cell mobilization in resistant or relapsed lymphoma. Superior yield consistent with some3,4 but not other1,7,8 prior work that of progenitor cells following a salvage regimen comprising has examined the relation between time since last chemo- ifosphamide, etoposide and epirubicin compared to intermedi- therapy and mobilization efficiency. ate-dose cyclophosphamide. Br J Haematol 1997; 98: 228–233. The mechanism(s) by which prior chemotherapy de- 12 Goldschmidt H, Hegenbart U, Wallmeier M et al. Factors creases the concentration of mobilized peripheral blood influencing collection of peripheral blood progenitor cells

Bone Marrow Transplantation Stem cell mobilization CD Ford et al 905 following high-dose cyclophosphamide and granulocyte colony- granulocyte colony-stimulating factor (rhG-CSF): a rando- stimulating factor in patients with multiple myeloma. Br J mized evaluation of different doses of rhG-CSF. Br J Haematol 1997; 98: 736–744. Haematol 2002; 116: 468–474. 13 Morton J, Morton A, Bird R et al. Predictors for optimal 20 Andre M, Baudoux E, Bron D et al. Phase III randomized mobilization and subsequent engraftment of peripheral blood study comparing 5 or 10 g per kg per day of filgrastim for progenitor cells following intermediate dose cyclophosphamide mobilization of peripheral blood progenitor cells with chemo- and G-CSF. Leukemia Res 1997; 21: 21–27. therapy, followed by intensification and autologous transplan- 14 Sutherland R, Anderson L, Keeney M et al. The ISHAGE tation in patients with nonmyeloid malignancies. Transfusion guidelines for CD34+ cell determination by flow cytometry. 2003; 43: 50–57. J Hematother 1996; 5: 213–226. 21 Drake M, Ranaghan L, Morris TCM et al. Analysis of the 15 Ford CD, Chan K, Reilly WF, Petersen FB. An evaluation of effect of prior therapy on progenitor cell yield: use of a predictive factors for CD34+ cell harvest yields from patients chemotherapy scoring system. Br J Haematol 1997; 98: mobilized with chemotherapy and growth factors. Transfusion 745–749. 2003; 43: 622–625. 22 Clark RE, Brammer CG. Previous treatment predicts the 16 Ford CD, Greenwood J, Strupp A, Lehman CM. Change in efficiency of blood progenitor cell mobilization: validation of a CD34+ cell concentration during peripheral blood stem cell chemotherapy scoring system. Bone Marrow Transplant 1998; collection: effects on collection efficiency and efficacy. Trans- 22: 859–863. fusion 2002; 42: 904–911. 23 Neben S, Hemman S, Montgomery M et al. Hematopoietic 17 Narayanasami U, Kanteti R, Morelli J et al. Randomized trial stem cell deficit of transplanted bone marow previously of filgrastim versus chemotherapy and filgrastim mobilization exposed to cytotoxic agents. Exp Hematol 1993; 21: 156–162. of hematopoietic progenitor cells for rescue in autologous 24 Molineux G, Xu C, Hendry J, Testa NG. A cellular analysis of transplantation. Blood 2001; 98: 2059–2064. long-term haematopoietic damage in mice after repeated 18Weaver CH, Birch R, Greco L et al. Mobilization and treatment with cyclophosphamide. Cancer Chemother Phar- harvesting of peripheral blood stem cells: randomized evalua- macol 1986; 18: 11–16. tions of different doses of filgrastim. Br J Haematol 1998; 100: 25 Adler SS, Kuznetshy RD. Effects of chemotherapeutic agents 338–347. on hemopoietic stromal tissue: I. Effects of cyclophosphamide 19 Demirer T, Ayli M, Ozcan M et al. Mobilization of peripheral and bleomycin on hemopoiesis in subcutaneous femur blood stem cells with chemotherapy and recombinant human implants. Exp Hematol 1984; 12: 153–161.

Bone Marrow Transplantation