Stem cell mobilization in HIV seropositive patients with lymphoma

Abstract

High-dose chemotherapy with autologous peripheral blood stem cell rescue has been reported as feasible and effective in HIV-associated lymphoma. Although a sufficient number of stem cells seems achievable in most patients, there are cases of stem cell harvest failure. The aim of this study was to describe the mobilization policies used in HIV-associated lymphoma, evaluate the failure rate and identify factors influencing mobilization results. We analyzed 155 patients who underwent attempted stem cell mobilization at 10 European centers from 2000–2012. One hundred and twenty patients had non-Hodgkin lymphoma and 35 Hodgkin lymphoma; 31% had complete remission, 57% chemosensitive disease, 10% refractory disease, 2% untested relapse. Patients were mobilized with chemotherapy + G-CSF (86%) or G-CSF alone (14%); 73% of patients collected >2 and 48% >5 × 10⁶ CD34⁺ cells/kg. Low CD4+ count and refractory disease were associated with mobilization failure. Low CD4⁺ count, low platelet count and mobilization with G-CSF correlated with lower probability to achieve >5 × 10⁶ CD34⁺ cells/kg, whereas cyclophosphamide ≥3 g/m² + G-CSF predicted higher collections. Circulating CD34⁺ cells and CD34/WBC ratio were strongly associated with collection result. HIV infection alone should not preclude an attempt to obtain stem cells in candidates for autologous transplant as the results are comparable to the HIV-negative population.

Introduction

High-dose chemotherapy (HDT) with autologous stem cell transplantation (ASCT) is a potentially curative treatment for several hematologic malignancies, including Hodgkin’s lymphoma (HL) and non-Hodgkin’s lymphoma (NHL), with peripheral blood as the preferred hematopoietic stem cell (SC) source.1,2 The lowest SC dose to safely support HDT conditioning regimens in patients with lymphoma is considered to be 2 × 10 CD34 cells/kg3–7 and although this is achievable in most patients, there are cases of stem cell harvest failure. In the HIV-negative population, failure rates are estimated to be between 5% and 30%, with different mobilization regimens and patient populations, and up to 60% in high-risk patients such as those exposed to fludarabine.8–10 Indeed, there is much interest in novel agents and strategies to minimize mobilization failure.9,11 The chance of cure for HIV-infected patients with lymphoma has greatly increased after the advent of combination antiretroviral therapy (cART) in 1996,12,13 and, more recently, HDT with ASCT in HIV-positive patients with lymphoma has been reported to be as feasible and effective as in HIV-negative counterparts.14–18 However, although the mechanism is not completely understood, depletion of hematopoietic progenitor cells has been described in HIV-infected subjects, as measured by reduction in long-term colony-initiating cell (LTCIC) numbers and increased rate of hematopoietic SC apoptosis.19,20 Moreover, reduced CD34 cell mobilization using G-CSF has been reported in patients with severe immunodeficiency.21 Several groups reported successful SC mobilization and ASCT in HIV-positive patients receiving cART as either rescue or consolidation of treatment for NHL or HL, usually in small series of selected patients. Effective antiretroviral therapy could help to correct the defective hematopoiesis and finally protect from mobilization failure.22 In the HIV-negative patients, several parameters have been identified predicting poor SC collections (including older age, type and status of underlying hematologic disease, number and type of prior treatments, prior radiotherapy, marrow involvement and thrombocytopenia at mobilization).23–26 Proper analyses in an HIV setting are missing. The purpose of the present study was to describe the mobilization policies used in HIV-associated lymphoma, to evaluate the failure rate, and identify factors influencing mobilization results. Moreover, the role of ‘ongoing’ parameters (circulating pre-apheresis peripheral blood CD34 cells and the ratio between CD34 count/WBC count evaluated the same day) in predicting the collection outcome was assessed as potential early markers of failure.

Table 1.Characteristics of 155 patients at first mobilization and distribution among successful mobilization and failure.

Methods

This is a retrospective multicenter analysis of mobilization (and remobilization) attempts in HIV-positive patients with lymphoma, performed consecutively and registered in the ASCT database of 10 European centers from April 2000 to May 2012. All HIV-positive patients diagnosed with HL or NHL who were potential candidates for ASCT and who had started SC mobilizing procedures were eligible; at least one CD34 cell measurement on peripheral blood should have been performed on the predicted day of collection.

This study is a collaborative effort within the Cooperative European Group on AIDS and Tumors (GECAT). All patients had given written informed consent to PBSC mobilization and collection either within Ethical Committee approved clinical trials or in the context of standard clinical practice.

Data regarding SC collection attempts were analyzed on clinical records to evaluate mobilization and remobilization success/failure rates and to identify predicting factors. “Mobilization failure” was defined as a collection of <2 × 10 CD34 cells/kg and “optimal mobilization” when >5 × 10 CD34 cells/kg were collected. The variables evaluated at the time of mobilization and considered for correlation with mobilization results (including demographic, lymphoma and HIV infection-related factors and mobilization-related parameters) and the parameters registered during mobilization attempt (“ongoing parameters”) as predictors of SC yield are reported in the Online Supplementary Appendix. The impact of all these variables was evaluated both for the probability of mobilization failure and for the capability to achieve an optimal SC collection. A mobilization attempt was defined as “first mobilization” when patients had never undergone a mobilization attempt before or “remobilization” when patients had previously experienced a mobilization failure. Stem cell harvest by apheresis was performed according to the policies of the center concerned. Collection procedures started after a peripheral blood CD34 cell count of <20/mcL was reached.

Table 2.Univariate and multivariate statistical analysis of factors influencing mobilization failure (CD34⁺ < 2×10⁶/kg) in 155 first mobilization attempts.

Statistical analysis

Standard descriptive statistics, such as median, range and proportions were used to summarize the patient sample. The χ test was used to compare differences in percentage, and the Mann-Whitney U test was used to compare continuous values. Logistical regression model was used for univariate and multivariate analysis of predictor variables for minimal (>2 × 10 CD34 cells/kg) and optimal (>5 × 10 CD34 cells/kg) SC collection. Continuous variables were categorized as follows: each variable was first divided into 4 categories at approximately the 25, 50, and 75 percentile. If the Odds Ratios (OR) in 2 or more adjacent categories were not substantially different, these categories were grouped together. If no clear pattern was observed, the median was taken as the cutoff point. Variables found to be significant (P<0.1) in univariate analysis were tested in multivariate analysis, which was performed using a stepwise logistical regression model. P<0.05 was considered statistically significant.

Table 3.Univariate and multivariate statistical analysis of factors influencing optimal mobilization (CD34⁺ >5×10⁶/kg) in 155 first mobilization attempts.

Results

Patient population and CD34+ cells yield at first mobilization attempt

We analyzed 155 HIV-positive patients who underwent initial SC mobilization. One hundred and twenty patients had NHL and 35 HL. Thirty-one percent of patients were in complete remission, 57% had chemosensitive disease, 10% refractory disease and 2% untested relapse. The majority of patients were mobilized with CT + G-CSF (86%) and the remainder with G-CSF alone (14%). All patients but 4 were on cART (including zidovudine, which is known for hematopoietic toxicity, in 10) 27 and 2 of them started cART at mobilization. Overall, in 73% of patients (113 of 155) a collection of >2 × 10 CD34 cells/kg was achieved and in 48% (74 of 155) an optimal yield (>5 × 10 CD34 cells/kg) was obtained. The median number of CD34 cells collected was 6.12 × 10/kg (range 2–33 × 10 6/kg), after a median of 2 aphereses (range 1–4). Interestingly, a higher number of CD34 cells was collected after CT + G-CSF compared to G-CSF alone: 6.55 × 10 CD34 cells/kg (range 2.0–33.0) and 3.85 × 10 CD34 cells/kg (range 2.3–31.0), respectively (P=0.01). Mobilization failure occurred in 27% of patients (42 of 155). Patients’ characteristics and their distribution among successful (>2 × 10 CD34 cells/kg) and failed (<2 × 10 CD34 cells/kg) mobilization are reported in Table 1. Interestingly, 6 of 10 (60%) patients receiving zidovudine collected >2 × 10 CD34 cells/kg, with optimal collection in 4 (40%).

Table 4.Circulating CD34⁺ cells, WBC and CD34/WBC ratio in 155 patients at first mobilization and their distribution among successful mobilization and failure.

Univariate and multivariate analysis for mobilization failure and for optimal mobilization

Logistical regression analysis of predictive factors for mobilization failure is reported in Table 2.

In univariate analysis, refractory disease, platelet count less than 160 × 10/L, CD4 count less than 237/mcL, cyclophosphamide 1.5 g/m as mobilizing treatment, two or more lines of prior therapies and bone marrow disease at mobilization were significantly associated with the risk of mobilization failure. In multivariate analysis, only low CD4 cell count and refractory disease remained significant. Indeed, 44% of patients with CD4 count below 100/mcL failed mobilization, as did 64% of patients with refractory disease. Univariate and multivarite analyses for optimal mobilization are reported in Table 3. In univariate analysis, refractory disease, platelet count <160 × 10/L, CD4 count <237/mcL, G-CSF alone as mobilizing treatment compared to G-CSF + CT and cyclophosphamide 1.5 g/m were associated with failure to achieve a collection of >5 × 10 CD34 cells/kg, while cyclophosphamide >3 g/m correlated with optimal collection. In multivariate analysis, only low CD4 cell count, low platelet count and mobilization with G-CSF alone remained significantly associated with lower probability of optimal collection, whereas cyclophosphamide >3 g/m + G-CSF remained significantly correlated with optimal collection.

Peripheral CD34+ cell counts and CD34+ cells/WBC counts ratio

Table 4 shows the peripheral blood WBC counts, circulating CD34 cells and their ratio (CD34/WBC ratio) measured at predicted day of collection in the whole series of 155 first mobilization cases and their distribution in the 2 groups of mobilization success and failure. A clear difference was seen in absolute CD34 count and in CD34/WBC ratio between the 2 groups. To verify if pre-apheresis CD34 count and CD34/WBC ratio could overcome the predictive significance of ‘base-line parameters’ on collection results, we performed the logistical regression multivariate analysis for collection failure and for optimal collection including the ‘ongoing parameters’. Both CD34 absolute count and CD34/WBC ratio were independently associated with the probability of mobilization failure (P<0.001 for both) and CD34/WBC ratio alone (P<0.001) with the probability of optimal collection. However, low CD4 count and refractory disease maintained their independent impact on mobilization failure, and low platelet count maintained its adverse impact in obtaining an optimal collection (data not shown). To further examine the correlation between ‘ongoing parameters’ and the probability of SC collection, we segregated the CD34/WBC ratio into quartiles. The lowest quartile was for ratio <0.7 and the higher for >5.4. The percentage of patients who failed SC harvest correlated with the quartile, with 68%, 24%, 9% and 0% of failure rate, respectively, for lower to higher quartile.

Table 5.Stem cell mobilization and remobilization attempts in 155 HIV positive patients.

Remobilization

A total of 42 of 155 patients failed initial mobilization and 23 of these 42 patients were remobilized, in 15 cases with CT + G-CSF, in 7 cases with G-CSF alone and in one case with plerixafor + G-CSF. Following remobilization, only 7 of 23 (30%) patients achieved at least 2 × 10 CD34 cells/kg. Two more patients achieved >2 × 10 CD34 cells/kg after pooling first and second attempts. Few patients proceeded to further attempts (Table 5). Among remobilization strategies, G-CSF alone and cyclophosphamide 1.5 g/m had the highest failure rates, respectively 86% (6 of 7 patients) and 83% (5 of 6 patients), compared to 44% (4 of 9 patients) for other CT regimens plus G-CSF. The only patient who received plerixafor + G-CSF failed collection. Multivariate analysis of factors influencing failure at remobilization was limited by the small number of cases, but showed a low CD4 count as the only independent significant factor correlated to failure (OR 2.9; P=0.02); however, adding the ‘ongoing parameters’ to the analysis, only CD34/WBC ratio resulted significantly correlated with collection (OR 0.17; P=0.03) (data not shown).

Then, of the whole series of 155 patients who underwent SC mobilization, 113 collected enough CD34 cells at first mobilization attempt, plus 9 after remobilization and pooling harvests (total 122 of 155, 79%), while 33 of 155 (21%) failed to collect enough cells to perform ASCT even after repeated attempts. Finally, 104 patients actually received HDT with SC rescue. Neutrophil engraftment, defined as an absolute neutrophil count >0.5 × 10/L, occurred in all patients at a median of 11 days (range 8–33 days). Platelet engraftment, defined as self-supporting platelet count >20 × 10/L, occurred in all evaluable patients (4 patients died early before platelet engraftment), except in 2 who were lost to follow up five and nine months after transplant, at a median of 14 days (range 7–455 days).

Discussion

This is the first study to address SC mobilization and collection in HIV-positive patients with lymphoma. Despite prior concerns that HIV might inhibit hematopoiesis and reduce SC mobilization,19–21 in a series of 155 consecutive HIV-associated lymphoma, the majority of patients (73%) achieved enough CD34 cells to proceed to transplant at the first mobilization attempt and almost half (48%) reached an optimal SC collection. Moreover, engraftment kinetics in patients who actually received ASCT appear reassuring and comparable with the general HIV-negative population.2 The 27% rate of mobilization failure, with no difference between HL and NHL, is still troublesome, but it does not appear much different from the HIV-negative counterpart, where a wide range of failure rate is reported, and might depend on different mobilization strategies and patient populations.8–10

The widespread use of cART, which seems capable of correcting the defective hematopoiesis of HIV-infected subjects,22 might play a pivotal role in this. The mechanism by which antiretroviral therapy favorably impacts on hematopoiesis in HIV-infected subjects is not completely understood, since HIV does not directly infect the CD34 cells.28In vitro studies showed that Ritonavir, a protease inhibitor frequently used in antiretroviral combination treatments, markedly decreased apoptosis and increased colony formation if added to CD34 cells cultures from HIV-infected patients and even from HIV-negative controls. Thus, it has been hypothesized that cART could overcome inhibition of hematopoiesis with a mechanism unrelated to its antiviral activity.29

In our series, only two factors were independently associated with mobilization failure: chemo-refractory disease and a low CD4 count, which can be regarded as the expression of patients’ immune status and of responsiveness to antiretroviral therapy. Numerous previous studies have demonstrated the influence of patients’ immune status (expressed by the number of CD4 cells) on clinical aggressiveness of HIV-associated lymphoma, response to treatment and risk of relapse.30–32 In addition, a recent study from the GICAT (Gruppo Italiano Cooperativo AIDS e Tumori) showed a correlation between CD4 count and the chance for HIV-positive patients with lymphoma to receive HDT and SC transplantation in the salvage setting.33 The present study further shows that the CD4 count correlates even with the possibility of mobilizing and harvesting hematopoietic SC.

Our data also emphasize that different methods of hematopoietic SC mobilization are in use without an established standard. The majority of patients were mobilized after CT + G-CSF (with a wide variety of CT regimens) and only a small proportion received G-CSF alone. Mobilization after CT + G-CSF, a strategy widely used in Europe, has, in this study, allowed higher CD34 cell yields compared to G-CSF alone, as previously reported in the general HIV-negative population.34 While the use of G-CSF alone has been advocated to reduce toxicity and costs,10 adding chemotherapy to G-CSF as mobilizing agent allows further tumor cytoreduction and in vivo purging of mobilized tumor cells, besides allowing higher CD34 yield.9,35 Even if a target of 2 × 10/Kg CD34 cells is considered appropriate to ensure engraftment after myeloablative treatment, in the HIV-negative population, a higher amount of CD34 cells reinfused has a favorable impact on patient’s outcome, mainly in terms of prompt engraftment.36,37 Such a correlation between CD34 cell dose and quality of engraftment has been suggested for both neutrophil and platelet delayed engraftment also in a small series of HIV-associated lymphoma.17 Thus, achieving an optimal CD34 collection, defined as a CD34 cell number >5 × 10/kg, might be preferred, particularly in a subset of patients, such as HIV-positive patients with lymphoma, potentially at higher risk of delayed engraftment and infectious complication.

Other than the use of G-CSF alone, a lower likelihood of achieving an optimal SC collection was predicted in our series by a low platelet count at the time of mobilization, a potential indicator of a reduced bone marrow reserve, which is a well known predictor of poor collection also in HIV-negative populations.23–26 Moreover, the use of cyclophosphamide at the dosage of 1.5 g/m as mobilizing regimen predicted a lower yield of CD34 cells as already reported in lymphoma series in the general population.38 On the other hand, high-dose cyclophosphamide (at least 3 g/m) correlated with higher CD34 cell collection compared to the other CT regimens.

Cyclophosphamide monotherapy was the most used regimen in this series, and its mobilization potential is well known, since it was the traditional regimen employed in seminal studies on PBSC collection.39 Its retrospective comparison with other CT mobilizing programs in our study is hampered by the wide variety of regimens used, which leads to an excessive fragmentation of data.

On the whole, our study shows that, as in the HIV-negative population, also in HIV-positive subjects the identified prognostic factors, available before starting mobilization, were not strong enough to accurately predict the collection result. The impact of pre-apheresis peripheral blood CD34 cell count and CD34/WBC ratio was also evaluated as potentially useful ongoing indicators of collection success or failure. Indeed, both CD34 cell count and CD34/WBC ratio resulted strongly associated to subsequent SC yield. Given the recent availability of mobilizing agents like plerixafor, whose use ‘on demand’ might boost ‘ongoing’ CD34 cell mobilization, in cases at risk of failure such parameters may be shown to be clinically useful. Accordingly, within the HIV-negative population, standard definition of “predicted poor mobilizers” has been recently proposed to identify patients who may potentially benefit from early intervention with new mobilizing agents.40 While the results of plerixafor in HIV-negative populations seem encouraging,9,10,11,41 no experiences have been reported so far on its use as mobilizing agent in HIV-positive patients, although it was originally introduced as a potentially useful anti-HIV agent.42 Our results would support a trial on plerixafor in HIV-positive subjects predicted to be at high risk of mobilization failure. The rarity of the use of plerixafor in the present series might lie in the small number of patients who received a second mobilization attempt and on the limited availability of this agent in Europe during the study period. Only 55% of our patients who failed first mobilization went forward for a second attempt, and overall, only 21% reached a sufficient number of CD34 cells after the second attempt. That means that HIV-positive patients failing a first mobilization are actually very unlikely to proceed to autologous transplant. There could be several reasons for this, including early disease progression, but it is also due to the difficulty of harvesting SC after a first collection failure.

Taken together, our data suggest that HIV should not preclude an attempt to obtain SC in candidates for autologous transplant as the results are comparable to the HIV-negative population. CT + plus G-CSF seems to mobilize better than G-CSF alone and, if using cyclophosphamide, at least 3 g/m seems recommendable. However, we need to optimize current mobilization protocols, and more effective mobilization agents are welcome in HIV-positive patients, as in the general population. Improving antiretroviral therapy is highly advisable even to improve SC collection, and a better understanding of variables associated with mobilization success may further optimize SC collection. The results of this study might help to decide optimal mobilizing strategy in HIV-related lymphoma and could provide the framework for a rational investigation into the use of new mobilizing agents.

Footnotes

• Authorship and Disclosures Information on authorship, contributions, and financial & other disclosures was provided by the authors and is available with the online version of this article at www.haematologica.org.

• Received April 11, 2013.
• Accepted August 20, 2013.