ORIGINAL ARTICLE

T-lymphocyte reconstitution following rigorously T-cell-depleted versus

unmodified autologous stem cell transplants

PAW te Boekhorst1, CHJ Lamers2, MR Schipperus3, RQ Hintzen4, B van der Holt5, JJ Cornelissen1,B Lowenberg1 and JW Gratama2

1Department of Hematology, Erasmus Medical Center, Rotterdam, The Netherlands; 2Department of Medical Oncology, ErasmusMedical Center, Rotterdam, The Netherlands; 3Department of Hematology, HAGA Hospitals, The Hague, The Netherlands;4Department of Neurology, Erasmus Medical Center, Rotterdam, The Netherlands and 5Department of Trials & Statistics, ErasmusMedical Center, Rotterdam, The Netherlands

We compared the kinetics of T-cell recovery afterextensive ex vivo and in vivo T-cell depleted autologousstem cell transplantation (SCT) for multiple sclerosis(MS; n¼ 8) with unmodified SCT for hematologicalmalignancies (HM; n¼ 39). Both patient group showeda very protracted recovery of ‘naive’ CD4þ , 45R0�

(ECD45RAþ ) T-cells. Within the ‘primed’ CD4þ ,45R0þ T-cells, the ‘central memory’ cells expressing theCD62L and CD27 markers were the slowest to recover.The repopulating T-cells were highly activated, as shownby increased expression of HLA-DR and the apoptosismarker CD95. The capability of CD4þ and CD8þ

T-cells to produce IFN-c, IL-2 and TNF-a had reachednormal ranges from 2 months post SCT onwards.Unexpectedly, the kinetics of T-cell recovery between 3and 12 months post transplant was similar in T-depletedand unmodified SCT. Before SCT, the HM patientsshowed lymphopenia of all T-cell subsets, upregulatedHLA-DR and CD95 expression and increased cytokineresponses. We suggest that the similar kinetics of T-cellrecovery in the two patient groups may be explained bythe susceptibility to apoptosis of the activated CD4þ

T-cells in the autografts of the HM patients. Thissusceptibility to apoptosis would interfere with a swiftand sustained CD4þ T-cell regeneration post SCT.Bone Marrow Transplantation (2006) 37, 763–772.doi:10.1038/sj.bmt.1705333; published online 6 March2006Keywords: autologous stem cell transplantation; T-celldepletion; T lymphocytes; immune reconstitution

Introduction

High-dose immunosuppressive therapy followed by auto-logous or allogeneic stem cell transplantation (SCT) mightreduce autoimmune disease activity. For example, experi-ments using this approach in murine multiple sclerosis(MS) showed control of the disease after SCT.1,2 Earlierclinical studies indicated that long-term remissions ofseveral autoimmune diseases could be achieved afterSCT.3–5 Here, we applied extensive ex vivo and in vivoT-cell depletion in combination with autologous SCT assalvage treatment for MS. Ex vivo T-cell depletion wasdone by preparing autografts enriched for CD34þ hema-topoietic stem cells, while in vivo T-cell depletion wasaccomplished by administration of antithymocyte globuline(ATG) and total body irradiation (TBI).6 Althoughimmune reconstitution has been extensively studied afterunmodified and CD34-enriched autologous SCT,7–14 onlyrecently data have become available on autologous SCTrecipients pretreated with intensive ex vivo and in vivo T-celldepletion in the setting of autoimmune disease.15,16 Here,we compared the kinetics of T-cell recovery in MS patientstreated with this approach with that in patients withhematological malignancies (HM) receiving unmodifiedautografts. We studied absolute T-cell subset counts,expression of T-cell activation markers, and simultaneousanalyses of CD4þ and CD8þ T-cell function as measuredby the induction of various cytokine responses after in vitrolymphocyte stimulation during the first year post SCT. Aswe anticipated that the T-cell recovery in the intensivelyT-depleted MS patient group would be delayed and couldpredispose these patients for infectious problems, wecompared the occurrence of clinically significant infectiousepisodes with the rate of immune reconstitution in bothpatient groups.

Patients and methods

PatientsAfter informed consent, eight patients with rapidlyprogressive secondary MS (median age, 41 (range 30–50)

Received 30 August 2005; revised 2 February 2006; accepted 3 February2006; published online 6 March 2006

Correspondence: Dr PAW te Boekhorst, Department of Hematology,Erasmus Medical Center Rotterdam, PO Box 2040, 3000 CA Rotterdam,The Netherlands.E-mail: p.teboekhorst@erasmusmc.nl

Bone Marrow Transplantation (2006) 37, 763–772& 2006 Nature Publishing Group All rights reserved 0268-3369/06 $30.00

www.nature.com/bmt

years; three were male) and 39 patients with hematologicalmalignancies (median age, 51 (range 20–64) years; 21 weremale) were included in the study. The difference in agebetween both patient groups was not significantly different.The MS patients had not received any prior cytoreductivetherapy. Approximately 6–8 weeks before conditioning andstem-cell reinfusion, autologous bone marrow cells wereharvested under general anesthesia from these patients.After CD34 positive selection, the CD34 positive cellfraction of the graft was cryopreserved in liquid nitrogenuntil further use. The conditioning regimen consisted of horseantithymocyte globulin (ATG; Merieux (Marcy-l’Etoile,France)) at a dose of 15mg/kg intravenously (i.v.) fromdays �7 to �3. On days �4 and �3, cyclophosphamide(Cy; 60mg/kg i.v.) was given combined with mesnum(15mg/kg, 4� daily). Total body irradiation (TBI) wasgiven in two fractions of 5Gy each at days –2 and –1.On day 0, the autografts were thawed and reinfused. Thediagnostic indications for autologous SCT of the HMpatients were non-Hodgkin’s lymphoma (n¼ 21), multiplemyeloma (n¼ 11), acute lymphoblastic leukemia (n¼ 3),acute myelogenous leukemia (n¼ 3) and Hodgkin’s disease(n¼ 1). These patients had been pretreated with chemo-therapy for remission induction and maintenance beforemyeloablative conditioning and SCT. Their peripheralblood stem cells were harvested after stimulation withgranulocyte colony-stimulating factor (G-CSF, 5 mg/kg s.c.twice daily), and cryopreserved without further modifica-tion. Conditioning regimens consisted of procarbazine,etoposide, cytarabine and Cy (n¼ 21); Cy and TBI (n¼ 12);Cy and procarbazine (n¼ 3); procarbazine and TBI (n¼ 2);and Cy, procarbazine and etoposide (n¼ 1). Followingthese conditioning regimens, the autografts were thawedand reinfused.

Infectious episodesThe occurrence of infections was evaluated during the firstyear after SCT. Clinically significant infections were definedas grade X2 according to the NCI common toxicitycriteria. We analyzed infections during the first 100 dayspost SCT separately from those occurring after day 100, aswe considered the first group of infections to be related tothe direct effects of cytoreductive therapy such as mucosaltoxicity and neutropenia. Infections were considered as‘proven’ in case of clinical symptoms and microbiologicaland/or pathological proof. Alternatively, ‘clinical’ infec-tions were defined as those with clinical symptoms, butwithout microbiological or pathological proof. Dermalherpes zoster infections were considered as ‘proven’irrespective of the outcome of viral cultures. Infectionswith a protracted course were each scored as a singleinfectious event. Infections occurring after progression ofMS or relapse of HM were excluded from this analysis.

Study design for immunological monitoringT-cell reconstitution studies, as detailed below, wereperformed before SCT, that is, immediately before con-ditioning, and at 2, 3, 6, 9 and 12 months post SCT.In three HM patients these studies were discontinuedprematurely because of relapse of disease. In addition,

neutrophil counts were assessed three times per weekduring hospitalization, and at larger intervals thereafter, inorder to determine the time to neutrophil recovery (i.e.,0.5� 109 cells/l). For the T-cell reconstitution studies,reference ranges (defined by 5th and 95th percentiles) weredetermined in 160 healthy volunteer donors. The number ofviable T-cells in the grafts (per kg body weight of therecipient) administered to the MS and HM patients wasassessed by flow cytometry17 and were compared to those ina control group of 19 allogeneic grafts obtained fromhealthy sibling donors.

Enumeration of lymphocyte subsetsBlood samples (3ml each) were anticoagulated withethylene diamine tetra-acid (EDTA). Absolute lymphocytecounts were assessed using a 3-color stain-lyse-no washsingle platform method.18 A volume of 100ml blood waspipetted in duplicate in TRUCOUNT tubes (Table 1, tube 1;BD Biosciences, San Jose, CA, USA) and incubated for30min with a mixture of monoclonal antibodies (mAb;10 ml each) at room temperature (RT). Thereafter, ery-throcytes were lysed by adding 900 ml FACS LysingSolution, followed by incubation for 15 to 120min at 41Cin the dark until flowcytometric data acquisition using aFACScalibur instrument equipped with CellQuest software(all from BD Biosciences). Ungated list mode datacontaining a minimum of 5000 lymphocytes (defined asCD45bright events with low to intermediate forward (FSC)and sideward scatter signals (SSC)) were acquired fromeach tube. In this staining, T-lymphocytes were defined asCD45bright SSClow and CD3þ .

The T-cells were further characterized with six additionalstainings (Table 1). Here, CD4þ helper T-cells were definedas CD45bright SSClow and CD4þ ;19 CD8þ cytotoxic T-cellsas CD45bright SSClow and CD8bright TCRabþ T-cells asCD45bright SSClow and TCRabþ , and TCRgdþ T-cells asCD45bright SSClow and TCRgdþ . As markers for T-cellactivation, the percentage coexpression of HLA-DR wasassessed on the major TCRabþ T-cell subset, and that ofCD95 (Fas) on the CD8bright and CD8� (ECD4þ ) T-cellsubsets. Finally, CD4þ T-cells were subdivided intoCD45R0þ (‘primed’) and CD45R0� (‘naive’) cells; withinthe CD4þ , 45R0þ subset, cells were further divided into‘central memory’ (CD27þ or CD62Lþ ) cells and ‘effector

Table 1 Staining panel for lymphocyte subset enumeration

Tube no. Fluorochrome

FITC PE PE-Cy5

1 CD3 CD16+CD56 CD452 TCRab CD8 CD453 TCRgd CD4 CD454 CD57 HLA-DR TCRab5 CD8 CD95 TCRab6 CD45R0 CD27 CD47 CD45R0 CD62L CD4

Abbreviations: CD¼ cluster of differentiation; FITC¼ fluorescein isothio-cyanate; PE¼phycoerythrin; PE-Cy5¼PE-Cy5 tandem conjugate;TCR¼T-cell receptor.

Immune reconstitution in autologous stem cell transplantationPAW te Boekhorst et al

764

Bone Marrow Transplantation

memory’ cells (CD27� or CD62L�).20,21 These six stainingswere performed using a lyse-stain-wash assay.22 Here,Falcon 2052 tubes not containing counting beads wereused; following erythrocyte lysis using 2ml FACS LysingSolution, the cells were spun down, the supernatantremoved by pipetting and the cell pellet resuspended in250ml phosphate-buffered saline (PBS) and kept in the darkfor 2 h maximum until flow cytometric data acquisition.

All mAb used were purchased from BD Biosciences withthe exception of CD45mAb conjugated to PE-Cy5,TCRab/PE-Cy5 and CD4/PE-Cy5 which were purchasedfrom Beckman Coulter (Miami, FL, USA); CD45RO/FITC and CD8/PE were purchased from Dako (Glostrup,Denmark). Absolute numbers of lymphocyte subsets werecalculated by multiplying the absolute lymphocyte count(tube 1) by the percentages of each subset (expressed asproportion of lymphocytes). For the absolute counts oflymphocytes and CD3þ T-cells, the mean result ofduplicate assessments (tube 1) were used.

In vitro induction of cytokine responsesFor this assay, 3ml blood anti-coagulated with Na-heparinwas drawn on the same time points as outlined above. Onemilliliter of blood was incubated for 4 h at 371C in a 50-mltube with 25 ng/ml phorbol-12-myristate 13-acetate (PMA;Sigma, St Louis, MO, USA), 1mg/ml ionomycin (Calbio-chem, La Jolla, CA, USA) in RPMI-1640 medium (Gibco,Grand Island, NY, USA) in a final volume of 2ml. To thismixture, 10 mg/ml brefeldin A (Epicentre Technologies,Madison, WI, USA) had been added from the start toinhibit the secretion of the produced cytokines. As negativecontrol, a blood sample that had not been stimulated withPMA and ionomycin was included in each assay. There-after, 4ml PBS containing 1% bovine serum albumin (PBS-BSA) was added, the cells spun down and the supernatantremoved by pipetting. Erythrocytes were then lysed byadding 20ml FACS Lysing Solution for 10min at roomtemperature. Following centrifugation, the supernatant wasremoved and the leukocytes resuspended and permeabilizedusing 3ml FACS Permeabilizing Solution (BD Biosciences)for 10min at room temperature. The cells were then spundown, the supernatant removed and the cell pellet stainedwith a mixture of anti-TCRab PE-Cy5þCD8 FITC mAb.Following 30min incubation at RT in the dark, the cellswere washed once using PBS-BSA, resuspended in 300 mlPBS-BSA and aliquoted over six Falcon 2052 tubes forlabeling with anti-IFNg PE, anti-IL2 PE, anti-TNFa PE,anti-IL4 PE, anti-IL13 PE and mouse IgG1 PE (isotypecontrol) mAb (all from BD Biosciences), respectively.Incubation with mAb was again for 30min at RT, followedby washing and resuspending in 250 ml PBS until flowcytometry within 2 h. Ungated list mode data containing aminimum of 5000 TCRabþ lymphocytes were acquiredfrom each tube.

For data analysis, the induction of cytokine responseswas analyzed within the TCRabþ , CD8bright lymphocytes(representing the CD8þ T-cells) and the TCRabþ , CD8�

lymphocytes (representing the CD4þ T-cells). This proce-dure, rather than identifying CD4þ T-cells directly using aCD4mAb, was followed because stimulation with PMA

strongly downregulates CD4 expression.23,24During dataanalysis, a threshold was set for each cell sample on thecorresponding isotype control staining to distinguishbetween positive and negative fluorescence signals, andthe proportions of cytokine-positive cells within theTCRabþ , CD8� and TCRabþ , CD8þ T-cell subsets weredetermined. Finally, any background signal in the corre-sponding unstimulated controls (i.e., the correspondingblood samples that had not been activated) was subtractedfrom that of the stimulated samples.

Statistical analysesPatients experiencing a relapse of their original diseasefollowing SCT were censored at the time of relapse.Wilcoxon’s rank sum test was used to compare, perparameter, data from two patient groups. To comparethe incidence of infections between the MS and HM patientgroups, the incidence rates were calculated and comparedas described elsewhere.25 All P-values were two-sided, andP-values p0.05 were considered statistically significantunless as detailed below. For the comparison of data fromMS or HM patients with those of healthy controls, or thecomparison of data fromMS with those of HM patients foreach time point relative to SCT (Figures 2–6), Wilcoxon’srank sum test was used with adjustment for multipletesting, that is, 25 parameters ((15 T-cell markersþfivecytokine assessments on CD4þT cells)þ (five cytokineassessments on CD8þT cells))� six time points yieldsa factor of 150, so that only P-values less than0.05C150¼ 0.0003 were considered significant. Thematched-pair t-test was used to compare, per patient andimmune parameter, of post transplant data with thecorresponding pre-transplant data, again with adjustmentfor multiple testing (see legends to Figures 2–6).

Results

T-cell content of the stem cell graftsThe CD34þ cell-selected autografts of the MS patientscontained three to four logs less T-cells than the non-T-cell-depleted autografts of the HM patients (mediano0.01� 106 versus 109� 106 T-cells per kg body weightof the recipient; P¼ 0.0001; Figure 1). However, thenon-depleted autografts of the HM patients contained stillfewer T-cells than the non-depleted grafts from allogeneic,healthy stem cell donors (median 109� 106 versus261� 106 T-cells per kg body weight of the recipient;P¼ 0.01), reflecting the T-lymphopenia of the HM patientsat the time of hematopoietic stem cell mobilization.

Kinetics of neutrophil recoveryThe time for neutrophils to reach the 0.5� 109/l thresholdwas significantly longer in the MS patients (i.e., median 27(range 12–38) days) than in the HM patients (i.e., median15 (range 11–73) days; Po0.01).

Recovery of circulating T-cell subsetsBefore transplant, the median group values of the absolutecounts of the T-cells as defined by their receptor type

Immune reconstitution in autologous stem cell transplantationPAW te Boekhorst et al

765

Bone Marrow Transplantation

(i.e., TCRab and TCRgd), as well as the CD4þ and CD8þ T-cell subsets of the MS patients were all within their normalranges (Figure 2). This was also the case with both CD4þ

T-cell subsets as defined by their CD45R0 expression(Figure 3), as well as the subsets within the CD4þ , 45R0þ

T cells defined by their expression of CD62L and CD27(Figure 4). In contrast, the median group values of all pre-transplant T-cell subset counts of the HM patients werebelow or around the lower limit of their normal ranges.The low pre-transplant T-cell counts of the HM patientswere particularly apparent for the CD4þ T-cells (Figure 2,lower left panel), irrespective of their CD45R0 expres-sion (Figure 3), and were observed for the CD4þ45R0þ

T-cells irrespective of their CD62L and CD27 expression(Figure 4).

Only the CD8þ T-cell subset showed a rapid recoveryafter SCT in both MS and HM patients. Its median groupvalue had attained the normal range by 2 months (Figure 2,lower right panel). In contrast, the reappearance of CD4þ

T-cells was remarkably slow in both patient groups andtheir median values for both patient groups had not yetreached the lower limit of the normal range at 12 months.Of note, the rate of CD4þ T-cell reconstitution during thefirst 6 months tended to be slower in the MS patients thanin the HM patients (Figure 2, lower left panel). Similarly,the reconstitution of the TCRabþ and TCRgdþ T-cells inthe MS patients lagged slightly behind that in the HMpatients (Figure 2, upper panels).

1000

100

10

1

0.1

<.01

ALLO HM

Type of SCT donor

MS

Nu

mb

er o

f C

D3+

T c

ells

× 1

06

per

kg

bo

dy

wei

gh

t

Figure 1 T-cell (CD3þ ) content of stem cell grafts obtained fromallogeneic, healthy stem cell donors (ALLO; n¼ 19), patients withhematological malignancies (HM; n¼ 39) and patients with multiplesclerosis (MS; n¼ 8). The horizontal dashes indicate the median of eachgroup. Abbreviation: SCT, stem cell transplant.

TCR��+ T lymphocytes TCR��+ T lymphocytes

CD4+ T lymphocytes CD8+ T lymphocytes

100002439596

100

10

1

10000

1000

176

12

1

10000

1704

399100

10

1

10000

1031

191100

10

1

0 2 3 6 9 12 0 2 3 6 9 12

0 2 3 6 9 12 0 2 3 6 9 12

Months after autologous SCT

Ab

solu

te c

ell c

ou

nt

Figure 2 Absolute numbers of TCRabþ T-cells (upper left panel), TCRgdþ T-cells (upper right panel), CD4þ T-cells (lower left panel) and CD8þ T-cells(lower right panel) prior to SCT (indicated with 0) and at 2, 3, 6, 9 and 12 months post SCT. Each symbol (HM patients, o; MS patients, þ ) indicates asingle data point. For each time point and patient group the median is indicated by a horizontal dash. In each panel, the upper horizontal line indicates the95th percentile, and the lower horizontal line the 5th percentile of 160 apparently healthy control persons. Logarithmic scales were used for the y-axes inorder to compress the figure. After correction for the number of comparisons made (see Patients and methods), the MS patient data differed significantlyfrom controls as follows: TCRabþ T-cells: 3 and 6 months; TCRgdþ T-cells: 3, 6 and 12 months; CD4þ T cells: 2, 3, 6 and 12 months. Post transplant, theMS patient data differed significantly from the corresponding pre-transplant data as follows: TCRabþ T-cells: 3 months; CD4þ T-cells: 3 months. The HMpatient data differed significantly from controls as follows: TCRabþ T-cells: 0–9 months; TCRgdþ T-cells: 0–9 months; CD4þ T-cells: 0–12 months; andCD8þ T cells: 0 months. Post transplant, the HM patient data differed significantly from the corresponding pre-transplant data at none of the studied timepoints. The MS and HM patient data differed significantly from each other: CD4þ T-cells, 0 months.

Immune reconstitution in autologous stem cell transplantationPAW te Boekhorst et al

766

Bone Marrow Transplantation

The delayed reconstitution of CD4þ T-cells was mainlydue to a slow recovery of ‘naıve’ CD4þCD45RO� T-cellsin both patient groups (Figure 3, left panel). In contrast, the‘primed’ CD4þCD45ROþ T-cells recovered at a fasterrate, that is, their median group value reached the lowerlimit of the normal range at 2 months in the HM patients,and at 12 months in the MS patients (Figure 3, right panel).The median group values of the ‘effector memory’ CD4þ

T-cells (CD4þ , 45R0þ , 27� and CD4þ , 45R0þ , 62L�)were by and large within their normal ranges at all timeintervals post SCT in both patient groups (Figure 4, right

panels). In contrast, the median values of the ‘centralmemory’ CD4þ T-cells (CD4þ , 45ROþ , 27þ and CD4þ ,45R0þ , 62Lþ ) remained below the lower limits of theirnormal ranges during the 12 months of follow-up in bothpatient groups (Figure 4, left panels).

Expression of HLA-DR and CD95 markersBefore SCT, HLA-DR expression by the major TCRabþ

T-cell subset was normal in MS patients but increased inHM patients. After SCT, T-cell activation as indicated by

10000

868

118

10

1

10000

848

198

10

10 2 3 6 9 12 0 2 3 6 9 12

Months after autologous SCT

Ab

solu

te c

ell c

ou

nt

CD4+, 45RO– T lymphocytes CD4+, 45RO+ T lymphocytes

Figure 3 Absolute numbers of CD4þ , 45R0� ‘naıve’ T-cells (left panel) and CD4þ , 45R0þ ‘primed’ T-cells (right panel) before and at 2, 3, 6, 9 and 12months post SCT. After correction for the total number of comparisons made, the MS patient data differed significantly from controls as follows: CD4þ ,45R0þ T-cells: 2, 3, 6 and 12 months; CD4þ , 45R0� T-cells: 2, 3, 6 and 12 months. Post transplant, the MS patient data differed significantly from thecorresponding pre-transplant data as follows: CD4þ , 45R0� T-cells: 2 months. The HM patient data differed significantly from controls as follows: CD4þ ,45R0þ T-cells: 0–12 months; CD4þ , 45R0� T-cells: 0–12 months. Post transplant, the HM patient data did not differ significantly from the correspondingpre-transplant data at any of the studied time points. The MS and HM patient data differed significantly from each other: CD4þ , 45R0� T-cells, 0 months.See further the legend to Figure 2.

10000

14451000

324100

10

1

10000

14561000

331100

10

1

10000

1000362100

10

50

1

10000

1000

226100

1023

1

0 2 3 6 9 12 0 2 3 6 9 12

0 2 3 6 9 120 2 3 6 9 12

CD4+, 45RO+, 62L+ T lymphocytes

CD4+, 45RO+, 27+ T lymphocytes CD4+, 45RO+, 27–, T lymphocytes

CD4+, 45RO+, 62L–, T lymphocytes

Months after autologous SCT

Ab

solu

te c

ell c

ou

nt

per

�l

Figure 4 Absolute numbers of CD4þ , 45R0þ , 62Lþ (upper left panel) and CD4þ , 45R0þ , 27þ (lower left panel) ‘memory-effector’ T-cells, and CD4þ ,45R0þ , 62L� (upper right panel) and CD4þ , 45R0þ , 27� (lower right panel) ‘terminally differentiated effector’ T-cells before and at 2, 3, 6, 9 and 12 monthspost SCT. After correction for the total number of comparisons made, the MS patient data differed significantly from controls as follows: CD4þ , 45R0þ ,62Lþ T-cells: 2, 3, 6 and 12 months; CD4þ , 45R0þ , 27þ T-cells: 2, 3, 6 and 12 months; CD4þ , 45R0þ , 62L� T-cells: 3 and 6 months; CD4þ , 45R0þ , 27�

T-cells: 3 months. Post transplant, the MS patient data did not differ significantly from the corresponding pre-transplant data at any of the studied timepoints. The HM patient data differed significantly from controls as follows: CD4þ , 45R0þ , 62Lþ T-cells: 0–12 months; CD4þ , 45R0þ , 27þ T-cells: 0–12months; CD4þ , 45R0þ , 62L� T-cells: 0, 2 and 6 months; CD4þ , 45R0þ , 27� T-cells: 0 months. Post transplant, the HM patient data did not differsignificantly from the corresponding pre-transplant data at any of the studied time points. The MS and HM patient data differed significantly from eachother: CD4þ , 45R0þ , 27þ T-cells, 0 months. See further the legend to Figure 2.

Immune reconstitution in autologous stem cell transplantationPAW te Boekhorst et al

767

Bone Marrow Transplantation

increased HLA-DR expression by the TCRabþ T-cells wasapparent in both patient groups; this expression washighest at 2 months and returned to its normal range at12 months post SCT (Figure 5, left panel).

The expression of the apoptosis receptor CD95 by CD4þ

and CD8þ T-cells showed a large variation in normalsubjects (Figure 5, middle and right panels, respectively).Before SCT, the HM patients had significantly higher levelsof CD95 expression by their CD4þ and CD8þ T-cells thanthe MS patients. After SCT, the CD4þ T-cells of the MSpatients showed a strong increase of their CD95 expressionby their CD4þ T-cells with median group values exceedingthe normal ranges between 3 and 9 months. The expressionof CD95 by the CD4þ T-cells of the HM patients remainedstrongly elevated during the entire period of follow-up(Figure 5, middle panel). After SCT, the CD95 expressionby the CD8þ T-cells of the MS patients increased to similarlevels as seen on those of HM patients. However, themedian group values of CD95 expression by the CD8þ

T-cells in both MS and HM groups remained withinthe normal range, although a large variation betweenpatients was observed (Figure 5, right panel).

Cytokine responses by CD4þ and CD8þ T-lymphocytesThe intracellular expression of IFN-g (Figure 6, left panels)and TNF-a (data not shown) by CD8� (ECD4þ ) T-cellsafter stimulation with PMA and ionomycin showed aconsiderable variation between apparently healthy donors.This variation was smaller for IL-2 (Figure 6, right panels),while hardly any induction of IL-4 and IL-13 expressionwas observed (data not shown).

Before SCT, the group median values of all cytokineresponses of the CD4þ and CD8þ T-cells of the MSpatients were within the normal range. In contrast, allcytokine responses except IL-2 of both CD4þ and CD8þ

T-cells of the HM patients were increased as compared tothe healthy controls, indicating their activated state (Figure6 and data not shown).

Following SCT, the group median values of the IFN-gand TNF-a responses of the CD8þ T-cells of the MSpatients fluctuated around the upper limit of the normalrange, while the corresponding responses of the CD4þ

T-cells were not increased (Figure 6, left panels, and datanot shown). In contrast, the IL-2 responses of the CD4þ

and CD8þ T-cells between 2 and 6 months were relativelylow (Figure 6, right panels), while hardly any inductionof IL-4 and IL-13 was seen (data not shown). The cyto-kine responses of the CD4þ and CD8þ T-cells of theHM patients followed, by and large, similar patterns postSCT as described for the MS patients, except for higherIFN-g but not TNF-a responses by the CD4þ T-cells(Figure 6, upper left panel, and data not shown).

Infectious complications post SCTDuring the first 100 days post SCT no statisticallysignificant differences in the occurrence of infectiouscomplications between both patient groups were observed(Table 2). Infectious complications between 3 and 12months post SCT were rare among the MS patients andmore frequent among the HM patients, but this differencedid not reach significance.

Discussion

We set out to compare the rate of T-cell reconstitution andthe occurrence of infectious complications after autologousSCT in two settings: one with extensive ex vivo and in vivoT-cell depletion administered to patients with MS, and oneof unmodified autografts given to patients as consolidationtherapy for HM. We expected that the rate of T-cellrecovery in the intensively T-depleted MS patients wouldbe delayed and that, therefore, these patients wouldbe predisposed for significant infectious complications.Reportedly, both viral and bacterial infections reportedlyoccur more frequently after autologous SCT that had been

100

80

60

40

23

2

100

807260

40

17

2

100

807260

40

2050

0 2 3 6 9 12 0 2 3 6 9 12 0 2 3 6 9 12

% DR+ cells (fraction of TCRab+ T ly.) % CD95+ cells (fraction of CD4+ T ly.) % CD95+ cells (fraction of CD8+ T ly.)

Months after autologous SCT

Perc

ent

po

siti

ve c

ells

Figure 5 Percentages of TCRabþ T cells expressing HLA-DR (left panel), percentages of CD4þ T-cells expressing CD95 (middle panel) and percentagesof CD8þ T-cells expressing CD95 (right panel) before and at 2, 3, 6, 9 and 12 months post SCT. Linear scales were used for the y-axes. After correction forthe number of comparisons made, the MS patient data differed significantly from controls as follows: % HLA-DRþ cells (fraction of TCRabþ T-cells): 2, 3and 6 months; % CD95þ cells (fraction of CD4þ T-cells): 3 and 6 months. Post transplant, the MS patient data differed significantly from thecorresponding pre-transplant data as follows: % HLA-DRþ cells (fraction of TCRabþ T-cells): 3 and 6 months; % CD95þ cells (fraction of CD4þ T-cells):3 and 6 months. The HM patient data differed significantly from controls as follows: % HLA-DRþ cells (fraction of TCRabþ T cells): 0–12 months; %CD95þ cells (fraction of CD4þ T-cells): 0–12 months; % CD95þ cells (fraction of CD8þ T-cells): 0, 2, 3 and 9 months. Post transplant, the HM patientdata did not differ significantly from the corresponding pre-transplant data at any of the studied time points. The MS and HM patient data differedsignificantly from each other: % HLA-DRþ cells (fraction of TCRabþ T-cells): 0 months. See further the legend to Figure 2.

Immune reconstitution in autologous stem cell transplantationPAW te Boekhorst et al

768

Bone Marrow Transplantation

enriched for CD34þ hematopoietic stem cells (and thushad been depleted of T-cells) than after unmodifiedautografting.26 However, the rates of infections in ourMS and HM patient cohorts were not significantlydifferent, neither during the first 100 days nor between100 and 365 days post SCT. During the latter period, weobserved only two clinically significant infectious episodesin eight MS patients.

Also unexpected was the similar kinetics of T-cellreconstitution from 3 to 12 months post SCT in the MSand HM groups, both in terms of absolute counts ofcirculating T-cell subsets as in functional terms (i.e.,expression of activation markers and intracellular IFN-g,IL-2 and TNF-a following universal lymphocyte stimula-tion). The reconstitution of the CD4þ T-cells, in particularthat of the ‘primed’ CD4þ , 45R0þ subset, tended to bedelayed in the MS patients, but this difference with the HMpatients did not reach significance. A limitation of ourstudy design was that T-cell monitoring was only started by2 months post SCT, and that any differences between thetwo groups before that time have not been recorded.Several factors would have contributed to an initiallyslower rate of (T-cell) engraftment in the MS patients: (i)the 3–4 log T-cell depletion of their grafts; (ii) the depletionof residual (and any incoming) T-cells by ATG treatment;and (iii) the source of stem cells (bone marrow derivedversus peripheral blood derived in the HM patients).27 Weconsider it unlikely that the somewhat younger age of theMS patients would have offset their disadvantage with

respect to T-cell reconstitution rate as compared to theHM patients,14,28 as the ages of both groups were notsignificantly different. However, both groups did differwith respect to the rate of neutrophil engraftment, whichwas delayed in the MS patients. Here, the source of thestem cells (bone marrow) and the T-cell depletion of theautografts may have played a role.

Our immunophenotypic studies confirm and extendearlier observations on T-cell reconstitution after auto-logous SCT: (i) a very protracted recovery of CD4þ T-cells,9,15,29,30 in particular that of the ‘naive’ CD4þ , 45R0�

(ECD45RAþ ) subset;16,31–33 within the ‘primed’ CD4þ ,45R0þ T-cells, the ‘central memory’ cells expressingthe CD62L and CD27 markers were the slowest torecover;15,16 (ii) a relatively slow recovery of TCR-gdþ

T-cells; (iii) a rapid normalization of CD8þ T-cellcounts;8,15,29–31,33 (iv) initially, a strong upregulation ofHLA-DR expression by TCRabþ T-cells, graduallydeclining to normal at 12 months;12,30 and (v) upregulationof the apoptosis marker CD95, in particular by the CD4þ

T-cells.16,34 With respect to cytokine responses, ourobservation that the capability of CD4þ and CD8þ

T-cells to produce IFN-g, IL-2 and TNF-a hadreached their normal ranges from 2 months post SCTonwards is consistent with that of others.30 However, thepost transplant levels of IFN-g and TNF-a production bythe CD8þ T-cells in our two patient groups weresupranormal, which may fit with their activated state (seealso below).

100 100

80

60

40

20

3

%

80

64

40

20

080

100 100

80

60

402720

1

807160

40

20

06

INF-� (fraction of CD8– Tly) % IL-2 (fraction of CD8– Tly)

% IL-2 (fraction of CD8+ Tly)INF-� (fraction of CD8+ Tly)

0 2 3 6 9 12

0 2 3 6 9 12 0 2 3 6 9 12

0 2 3 6 9 12

Months after autologous SCT

Perc

ent

po

siti

ve c

ells

%

Figure 6 Percentages of CD8� (ECD4þ ) T-cells expressing IFN-g (upper left panel) or IL-2 (upper right panel), and percentages of CD8þ T-cellsexpressing IFN-g (lower left panel) or IL-2 (lower right panel) following in vitro stimulation with PMA and ionomycin before and at 2, 3, 6, 9 and 12 monthspost SCT. Linear scales were used for the y-axes. After correction for the number of comparisons made, the MS patient data differed significantly fromcontrols as follows: % IFN-gþ cells (fraction of CD8þ T-cells): 3 and 6 months. Post transplant, the MS patient data did not differ significantly from thecorresponding pre-transplant data at any of the studied time points. The HM patient data differed significantly from controls as follows: % IFN-gþ cells(fraction of CD4þ T-cells): 2, 3 and 9 months; % IFN-gþ cells (fraction of CD8þ T-cells): 0–12 months. Post transplant, the HM patient data did not differsignificantly from the corresponding pre-transplant data at any of the studied time points. The MS and HM patient data did not differ significantly fromeach other at any of the studied time points.

Immune reconstitution in autologous stem cell transplantationPAW te Boekhorst et al

769

Bone Marrow Transplantation

Before conditioning and SCT, we observed importantdifferences between the circulating T-cells in the MS andHM patients. While the MS patients had normal T-cellsubset counts, normal levels of the activation markersHLA-DR and CD95 and normal cytokine responses bytheir CD4þ and CD8þ T-cells, the HM patients showedlymphopenia of all T-cell subsets, upregulated HLA-DRand CD95 expression and increased cytokine responsesexcept for IL-2. Similar observations have been made inother studies of adult patients treated with high-dosechemotherapy35–37 with, as a result, an impact on thephenotype of the T-cells in their autografts.38 In thiscontext it is important that the autografts of our HMpatients contained significantly less T-cells than allograftsobtained from otherwise healthy donors (Figure 1).Following high-dose chemotherapy, the ‘naive’ CD4þ

and CD8þ T-cells in the HM patients were especially slowto regenerate,35,36 while the main wave of regeneration from‘primed’ CD4þ T-cells was only temporary, decliningbetween 9 and 12 months after chemotherapy.35 Thehigh levels of CD95 expression by post chemotherapy

CD4þ T-cells explain their increased susceptibility toapoptosis following in vitro stimulation with mitogens,34,35

which may reflect a regulatory response to their increasedlevel of activation. Of note, triggering of CD27 – expressedon ‘central memory’ CD4þ , 45R0þ T-cells – has beenshown to sensitize these cells for CD95-induced apopto-sis.39 Therefore, we suggest that the progeny of theactivated CD4þ T-cells in the autografts of our HMpatients are prone to apoptosis and, as a consequence,cannot achieve a swift and sustained T-cell regenerationpost SCT. The net result would be a delayed CD4þ T-cellregeneration similar to as observed in the MS patients aftertheir ex vivo and in vivo T-cell depletion.

In conclusion, a remarkable finding in this autologousSCT study was that the CD4þ T-cell recovery followingchemotherapy and unmodified SCT was delayed to asimilar rate as observed after chemoradiotherapy andextensive ex vivo and in vivo T-cell depletion. In thiscontext it is interesting to recall the observations by Porrataand co-workers describing the beneficial effect of earlylymphocyte recovery on the overall survival time afterautologous SCT (reviewed in Porrata and Markovic40).Although the mechanism of this effect is not certain, it isfair to speculate that lymphocytes in the grafts contributeto this effect. Extrapolation of our current results to thoseof Porrata et al. would imply that autograft collectionshould be timed when the patients are immunologically stillnormal, for example, have normal CD4þ T-cell countswithout an activated immunophenotype.

Acknowledgements

We are grateful to J Kraan, R van der Linden, N de Leeuw,K van Rooyen and J Doekaharan–van der Sluis for theirtechnical assistance.

References

1 Van Gelder M, Kinwel-Bohre EP, Van Bekkum DW.Treatment of experimental allergic encephalomyelitis in ratswith total body irradiation and syngeneic BMT. Bone MarrowTransplant 1993; 11: 233–241.

2 Karussis DM, Vourka-Karussis U, Mizrachi-Koll R, Abrams-ky O. Acute/relapsing experimental autoimmune encephalo-myelitis: introduction of long lasting, antigen-specific toleranceby syngeneic bone marrow transplantation. Multiple Sclerosis1999; 5: 17–21.

3 Traynor AE, Schroeder J, Rosa RM, Cheng D, Stefka J,Mujais S et al. Treatment of severe systemic lupus erythe-matosus with high-dose chemotherapy and haemopoieticstem-cell transplantation: a phase I study. Lancet 2000; 356:701–707.

4 Openshaw H, Nash RA, McSweeney PA. High-dose immuno-suppression and hemopoietic stem cell transplantation inautoimmune diseases: clinical review. Biol Blood MarrowTransplant 2002; 8: 223–248.

5 Nash RA, Bowen JD, McSweeney PA, Pavletic SZ, MaravillaKR, Park MS et al. High-dose immunosuppressive therapyand autologous stem cell transplantation for severe multiplesclerosis. Blood 2003; 102: 2364–2372.

Table 2 Infectious complications before and after day 100 post

SCTa

Description of infectiousepisodes

Number of infectious episodes

Multiple sclerosis(8 patients)

Hematologicalmalignancies(39 patients)

Prior to day 100 post SCTTotal number of episodes 9 21

BacterialStaphylococcus epidermidis 2 3Streptococcus pneumoniae 0 1Clostridium difficile 1 0Bacillus cereus 0 1

ViralHSV-1 3 3CMV 0 2RSV 0 1

Fungal/parasitic 0 0Clinical 3 10

Between days 100 and 365 post SCTTotal number of episodes 2 25

BacterialGram-positive rod 0 1

ViralHSV-1 1 2VZV 0 7

Fungal/parasiticToxoplasma gondii 0 2Candida species 0 1

Clinical 1 12

aGrade X2 according to NCI common toxicity criteria.Abbreviations: CMV¼ cytomegalovirus; HSV¼ herpes simplex virus;RSV¼ respiratory syncytial virus; SCT¼ stem cell transplantation;VZV¼ varicella zoster virus.

Immune reconstitution in autologous stem cell transplantationPAW te Boekhorst et al

770

Bone Marrow Transplantation

6 Samijn JPA, Te Boekhorst PAW, Mondria T, Van Doorn PA,Flach HZ, van der Meche FG et al. Intense T-cell depletionfollowed by autologous bone marrow transplantation forsevere multiple sclerosis. J Neurol Neurosurg Psychiatry 2006;77: 46–50.

7 Guillaume T, Rubinstein DB, Symann M. Immune reconstitu-tion and immunotherapy after autologous hematopoietic stemcell transplantation. Blood 1998; 92: 1471–1490.

8 Bomberger C, Sing-Jairam M, Rodey G, Guerriero A, YeagerAM, Flemming WH et al. Lymphoid reconstitution afterautologous PBSC transplantation with FACS-sorted CD34+

hematopoietic progenitors. Blood 1998; 91: 2588–2600.9 Vescio R, Schiller G, Stewart AK, Ballester O, Noga S, RugoH et al.Multicenter phase III trial to evaluate CD34+ selectedversus unselected autologous peripheral blood progenitorcell transplantation in multiple myeloma. Blood 1999; 93:1858–1868.

10 Rutella S, Rumi C, Laurenti L, Pierelli L, Sora’ F, Sica S et al.Immune reconstitution after transplantation of autologousperipheral CD34+ cells: analysis of predictive factors andcomparison with unselected progenitor transplants. Br JHaematol 2000; 108: 105–115.

11 Douek DC, Vescio RA, Betts MR, Brenchley JM, Hill BJ,Zhang L et al. Assessment of thymic output in adults afterhaematopoietic stem-cell transplantation and prediction ofT-cell reconstitution. Lancet 2000; 355: 1875–1881.

12 Nachbaur D, Kropshofer G, Heitger A, Latzer K, Glassl H,Ludescher C et al. Phenotypic and functional lymphocyterecovery after CD34+-enriched versus non-T cell-depletedautologous peripheral blood stem cell transplantation.J Hematother Stem Cell Res 2000; 9: 727–736.

13 Damiani D, Stocchi R, Masolini P, Michelutti A, Geromin A,Sperotto A et al. CD34+-selected versus unmanipulatedautologous stem cell transplantation in multiple myeloma:impact on dendritic and immune recovery and on complica-tions due to infections. Ann Oncol 2003; 14: 475–480.

14 Malphettes M, Carcelain G, Saint-Mezard P, Leblond V, AltesHK, Marolleau JP et al. Evidence for naive T-cell repopulationdespite thymus irradiation after autologous transplantation inadults with multiple myeloma: role of ex vivo CD34+ selectionand age. Blood 2003; 101: 1891–1897.

15 Storek J, Zhao Z, Lin E, Berger T, McSweeney PA, Nash RAet al. Recovery from and consequences of severe iatrogeniclymphopenia (induced to treat autoimmune diseases). ClinImmunol 2004; 113: 285–298.

16 Muraro PA, Douek DC, Packer A, Chung K, Guenaga FJ,Cassiani-Ingoni R et al. Thymic output generates a new anddiverse TCR repertoire after autologous stem cell transplant-ation in multiple sclerosis patients. J Exp Med 2005; 201:805–816.

17 Sutherland DR, Keeney M, Gratama JW. Enumeration ofCD34+ hematopoietic stem and progenitor cells. Curr ProtCytometry 2003, 6.4.1–6.4.25.

18 Schnizlein-Bick CT, Spritzler J, Wilkening CL, Nicholson JK,O’Gorman MR. Evaluation of TruCount absolute-count tubesfor determining CD4 and CD8 cell numbers in humanimmunodeficiency virus-positive adults. Site Investigators andThe NIAID DAIDS New Technologies Evaluation Group.Clin Diag Lab Immunol 2000; 7: 336–343.

19 Glencross D, Scott LE, Jani IV, Barnett D, Janossy G. CD45-assisted PanLeucogating for accurate, cost-effectivedual-platform CD4+ T-cell enumeration. Cytometry 2002;15: 69–77.

20 Sallusto F, Lenig D, Forster R, Lipp M, Lanzavecchia A.Two subsets of memory T lymphocytes with distincthoming potentials and effector functions. Nature 1999; 401:708–712.

21 Baars PA, Maurice MM, Rep M, Hooibrink B, Van LierRAW. Heterogeneity of the circulating human CD4+ T cellpopulation. J Immunol 1995; 154: 17–25.

22 Gratama JW, Van Esser JWJ, Lamers CHJ, Tournay C,Loweberg B, Bolhuis RL et al. Tetramer-based quantificationof cytomegalovirus (CMV)-specific CD8+ T lymphocytes inT-cell-depleted stem cell grafts and after transplantation mayidentify patients at risk for progressive CMV infection. Blood2001; 98: 1358–1364.

23 Bigby M, Wang P, Fierro JF, Sy MS. Phorbol myristateacetate-induced down-modulation of CD4 is dependent oncalmodulin and intracellular calcium. J Immunol 1990; 144:3111–3116.

24 Luo T, Downing JR, Garcia JV. Induction of phosphorylationof human immunodeficiency virus type 1 Nef and enhancementof CD4 downregulation by phorbol myristate acetate. J Virol1997; 71: 2535–2539.

25 Rothman KJ. Modern epidemiology. Little, Brown andCompany: Boston, 1986.

26 Crippa F, Holmberg L, Carter RA, Hooper H, Marr KA,Bensinger W et al. Infectious complications after autologousCD34-selected peripheral blood stem cell transplantation. BiolBlood Marrow Transplant 2002; 8: 281–289.

27 Talmadge JE, Reed E, Ino K, Kessinger A, Kuszynski C,Heimann D et al. Rapid immunologic reconstitution followingtransplantation with mobilized peripheral blood stem cells ascompared to bone marrow. Bone Marrow Transplant 1997; 19:161–172.

28 Kalwak K, Gorczynska E, Toporski J, Turkiewicz D, SlociakM, Ussowicz M et al. Immune reconstitution after haemato-poietic cell transplantation in children: immunophenotypeanalysis with regard to factors affecting the speed of recovery.Br J Haematol 2002; 118: 74–89.

29 Hoepfner S, Haut PR, O’Gorman M, Kletzel M. Rapidimmune reconstitution following autologous hematopoieticstem cell transplantation in children: a single institutionexperience. Bone Marrow Transplant 2003; 31: 285–290.

30 Schlenke P, Sheikhzadeh S, Weber K, Wagner T, Kirchner H.Immune reconstitution and production of intracellular cyto-kines in T lymphocyte populations following autologousperipheral blood stem cell transplantation. Bone MarrowTransplant 2001; 28: 251–257.

31 Koehne G, Zeller W, Stockschlaeder M, Zander AR.Phenotype of lymphocyte subsets after autologous peripheralblood stem cell transplantation. Bone Marrow Transplant 1997;19: 149–156.

32 Divine M, Boutolleau D, Delfau-Larue MH, Beaujeau F,Jovault H, Reyes F et al. Poor lymphocyte recovery followingCD34-selected autologous peripheral blood stem cell trans-plantation for non-Hodgkin’s lymphoma. Br J Haematol 1999;105: 349–360.

33 Galy A, Rudraraju S, Baynes R, Klein J. Recovery oflymphocytic and dendritic cell subsets after autologousCD34+ cell transplantation. Bone Marrow Transplant 2000;25: 1249–1255.

34 Singh RK, Varney ML, Buyukberber S, Ino K, Ageitos AG,Reed E et al. Fas-FasL-mediated CD4+ T-cell apoptosisfollowing stem cell transplantation. Cancer Res 1999; 59:3107–3111.

35 Hakim FT, Cepeda R, Kaimei S, Mackall CL, McAtee N,Zujewski J et al. Constraints on CD4 recovery postchemo-therapy in adults: thymic insufficiency and apoptoticdecline of expanded peripheral CD4 cells. Blood 1997; 90:3789–3798.

36 Mackall CL, Fleisher TA, Brown MR, Andrich MP, Chen CC,Feuerstein IM et al. Distinctions between CD8+ and CD4+T-cell regenerative pathways result in prolonged T-cell

Immune reconstitution in autologous stem cell transplantationPAW te Boekhorst et al

771

Bone Marrow Transplantation

subset imbalance after intensive chemotherapy. Blood 1997; 89:3700–3707.

37 Singh RK, Varney ML, Ino K, Vose JM, Bierman PJ,Talmadge JE. Immune dysfunction despite high levels ofimmunoregulatory cytokine gene expression in autologousperipheral blood stem cell transplanted non-Hodgkin’s lym-phoma patients. Exp Hematol 2000; 28: 499–507.

38 Fagnoni FF, Lozza L, Zibera C, Zambelli A, Ponchio L, JibelliN et al. Cytotoxic chemotherapy preceding apheresis ofperipheral blood progenitor cells can affect the early recon-

stitution phase of naıve T cells after autologous transplant-ation. Bone Marrow Transplant 2003; 31: 31–38.

39 Arens R, Baars PA, Jak M, Tesselaar K, Van Der Valk M,Van Oers MH et al. Cutting edge: CD95 maintains effectorT cell homeostasis in chronic immune activation. J Immunol2005; 174: 5915–5920.

40 Porrata LF, Markovic SN. Timely reconstitution ofimmune competence affects clinical outcome followingautologous stem cell transplantation. Clin Exp Med 2004;4: 78–85.

Immune reconstitution in autologous stem cell transplantationPAW te Boekhorst et al

772

Bone Marrow Transplantation