Immunocompetent T-Cells with a Memory-Like Phenotype are the Dominant Cell Type Following Antibody-Mediated T-Cell Depletion

Authors

  • Jonathan P. Pearl,

    1. Transplantation Branch, National Institute of Diabetes, Digestive and Kidney Diseases, National Institutes of Health, Department of Health and Human Services, Bethesda, Maryland, USA
    2. Department of Surgery, National Naval Medical Center, Bethesda, Maryland, USA
    Search for more papers by this author
  • Jeremy Parris,

    1. Transplantation Branch, National Institute of Diabetes, Digestive and Kidney Diseases, National Institutes of Health, Department of Health and Human Services, Bethesda, Maryland, USA
    Search for more papers by this author
  • Douglas A. Hale,

    1. Transplantation Branch, National Institute of Diabetes, Digestive and Kidney Diseases, National Institutes of Health, Department of Health and Human Services, Bethesda, Maryland, USA
    2. Organ Transplant Service, Walter Reed Army Medical Center, Washington, District of Columbia, USA
    Search for more papers by this author
  • Steven C. Hoffmann,

    1. Transplantation Branch, National Institute of Diabetes, Digestive and Kidney Diseases, National Institutes of Health, Department of Health and Human Services, Bethesda, Maryland, USA
    Search for more papers by this author
  • Wendy B. Bernstein,

    1. Division of Retrovirology, Walter Reed Army Institute of Research, Rockville, Maryland, USA
    Search for more papers by this author
  • Kelly L. McCoy,

    1. Transplantation Branch, National Institute of Diabetes, Digestive and Kidney Diseases, National Institutes of Health, Department of Health and Human Services, Bethesda, Maryland, USA
    2. Department of Surgery, National Naval Medical Center, Bethesda, Maryland, USA
    Search for more papers by this author
  • S. John Swanson,

    1. Transplantation Branch, National Institute of Diabetes, Digestive and Kidney Diseases, National Institutes of Health, Department of Health and Human Services, Bethesda, Maryland, USA
    2. Organ Transplant Service, Walter Reed Army Medical Center, Washington, District of Columbia, USA
    Search for more papers by this author
  • Roslyn B. Mannon,

    1. Transplantation Branch, National Institute of Diabetes, Digestive and Kidney Diseases, National Institutes of Health, Department of Health and Human Services, Bethesda, Maryland, USA
    2. Department of Surgery, National Naval Medical Center, Bethesda, Maryland, USA
    Search for more papers by this author
  • Mario Roederer,

    1. Immunotechnology Section, Vaccine Research Center, National Institute of Allergy and Infectious Diseases, National Institutes of Health, Department of Health and Human Services, Bethesda, Maryland, USA
    Search for more papers by this author
  • Allan D. Kirk

    Corresponding author
    1. Transplantation Branch, National Institute of Diabetes, Digestive and Kidney Diseases, National Institutes of Health, Department of Health and Human Services, Bethesda, Maryland, USA
    2. Department of Surgery, National Naval Medical Center, Bethesda, Maryland, USA
    3. Organ Transplant Service, Walter Reed Army Medical Center, Washington, District of Columbia, USA
    Search for more papers by this author

*Corresponding author: Allan D. Kirk, AllanK@intra.niddk.nih.gov

Abstract

T-cell depletion facilitates reduced immunosuppression following organ transplantation and has been suggested to be pro-tolerant. However, the characteristics of post-depletional T cells have not been evaluated as they relate to tolerance induction. We therefore studied patients undergoing profound T-cell depletion with alemtuzumab or rabbit anti-thymocyte globulin following renal transplantation, evaluating the phenotype and functional characteristics of their residual cells. Naïve T cells and T cells with potential regulatory function (CD4+CD25+) were not prevalent following aggressive depletion. Rather, post-depletion T cells were of a single phenotype (CD3+CD4+CD45RA-CD62L-CCR7-) consistent with depletion-resistant effector memory T cells that expanded in the first month and were uniquely prevalent at the time of rejection. These cells were resistant to steroids, deoxyspergualin or sirolimus in vitro, but were calcineurin-inhibitor sensitive. These data demonstrate that therapeutic depletion begets a limited population of functional memory-like T cells that are easily suppressed with certain immunosuppressants, but cannot be considered uniquely pro-tolerant.

Abbreviations: 
CsA

cyclosporine A;

DSG

deoxyspergualin;

FITC

fluorescein isothiocyanate;

Mab

monoclonal antibody;

PBMC

peripheral blood mononuclear cells;

RATG

rabbit anti-thymocyte globulin;

SEB

Staphylococcal enterotoxin B;

TCR

T-cell receptor.

Introduction

Allograft recipients typically receive induction immunosuppression, to counter the non-specific immune-activating effects of implantation (e.g. allograft reperfusion), and more tolerable maintenance immunosuppression chronically (1–4). Recently, it has been shown that induction drugs mediating aggressive but transient T-cell depletion acutely reduce the subsequent need for maintenance immunosuppression (5–11). Indeed, T-cell depletion in experimental animals occasionally induces tolerance (12–16). Thus, aggressive induction in humans has been suggested to facilitate a near-tolerant state (5,7).

We have previously demonstrated that induction with the humanized CD52-specific monoclonal antibody (Mab) alemtuzumab or the polyclonal antibody preparation rabbit anti-thymocyte globulin (RATG) at the time of transplantation results in rapid and sustained reduction of T cells peripherally and in secondary lymphoid tissues (9,10,17). However, without maintenance immunosuppression, patients uniformly experience rejection despite exceptionally low T-cell counts, suggesting that their residual T cells are well suited to mediate alloimmunity. Thus, it is not clear whether aggressive depletional induction is pro-tolerant and could be expected to facilitate other tolerance-inducing therapies (e.g. costimulation blockade or T-cell regulation), or if it simply allows for more manageable immunosuppression. For example, it is not clear if depletion homogeneously affects all T-cell subsets. This distinction is critical since T-cell subsets differ in their activation requirements and immune modulating functions. Indeed, previously activated cells are costimulation independent (18,19) and other T cells exert pro-tolerant regulation such that their depletion could hinder tolerance (reviewed in 20).

We therefore studied patients undergoing therapeutic T-cell depletion, and human T cells in vitro, specifically asking whether T cells of various phenotypes and activation requirements (naïve, memory or regulatory) were equivalently depleted or repopulated with similar kinetics. We also looked at the sensitivity of depletion-refractory cells to immunosuppressants. We demonstrate that CD4+ T cells with a surface phenotype consistent with effector memory cells predominate following depletion, and that calcineurin inhibitors, especially tacrolimus, most efficiently restrict the activity of these cells.

Methods

Patients, clinical therapies and blood samples

Patients (n = 13) were enrolled in clinical protocols approved by the Institutional Review Board at the National Institutes of Health. Experimental therapies were rendered under Food and Drug Administration Investigational New Drug Applications. The patients, aged 14–56 years (6 males, 7 females), were recipients of HLA-non-identical live donor renal allografts. Five patients received T-cell depletion with alemtuzumab (Campath-1H, Millennium Pharmaceuticals, Cambridge, MA) 0.3 mg/kg × 4 doses (days 0, 1, 3 and 5 post-transplant) in combination with deoxyspergualin (DSG, Nippon, Japan) intravenously at an initial dose of 4 mg/kg followed by 2.5 mg/kg daily for 13 days (17). These patients received no maintenance immunosuppression. Five patients received high-dose RATG (Thymoglobulin, Sangstat Medical Corporation, Fremont, CA) given as eight 2.5 mg/kg doses and maintenance with sirolimus monotherapy, as previously described (9). Three patients received daclizumab (2-mg induction and 1 mg on days 3, 14 and 28) as well as maintenance with tacrolimus (Fujisawa, Deerfield, IL) and mycophenolate mofetil (Roche, Nutley, NJ). These patients served as non-depleting controls. There was no graft loss or mortality. All patients in the alemtuzumab group (those without maintenance therapy) experienced rejection during the 1-month study period. No other patients experienced rejection during the study period.

Blood samples for analysis were drawn prior to depletion and weekly for 4 weeks. Lymphocyte counts were determined from clinical laboratory reports. Individual T-cell phenotype counts were calculated by multiplying the absolute lymphocyte count by the respective cell population percentages as determined by simultaneously drawn flow cytometry samples (below). Peripheral blood mononuclear cells (PBMC) for study were isolated by density centrifugation using Ficoll Hypaque (Amersham Biosciences AB, Uppsala, Sweden).

Antibodies and polychromatic flow cytometry

We used polychromatic flow cytometry to ask whether post-depletional T cells (CD3+) were predominantly CD4+ or CD8+, naïve (CD45RA+CD62L+); one of three phenotypes associated with functional memory cells (21): CD45RA-CD62L- (effector), CD45RA-CD62L+ (central) or CD45RA+CD62L-(RA+) memory; or memory cells with potential regulatory function (CD45RA-CD4+CD25+) (20). These phenotypes have profoundly different activation and homing requirements that significantly influence rejection in some experimental models (18,19,21–23). We also evaluated the CCR7 expression on cells from 3 patients by conventional flow cytometry, since CCR7 defines the homing characteristics of some memory cells (21).

To determine these phenotypes, clinical samples and whole blood incubation samples (see below) were stained with the following Mab conjugates: CD3 Cascade Blue, CD4 Cy5.5 Phycoerythrin (PE), CD8 Texas Red PE, CD11a Cy7 allophycocyanin (APC), CD14 Cy7PE, CD25 PE, CD45RA Cy5PE, CD45RO Cy5.5APC, CD62L fluorescein isothiocyanate (FITC), CD56 APC, CCR7 PE. Purified antibodies to the cell-surface markers above were obtained from BD/Pharmingen (Franklin Lakes, NJ) and conjugated to the indicated fluorochromes using standard protocols (http://drmr.com/abcon). Cascade Blue, FITC and Texas Red PE were obtained from Molecular Probes (Eugene, OR). Phycoerythrin and APC were obtained from ProZyme (San Leandro, CA). Cy5, Cy5.5 and Cy7 were obtained from Amersham Life Sciences (Pittsburgh, PA). Data were collected on a modified FACSDiVa (Becton Dickinson, Franklin Lakes, NJ) or a modified LSRII (Beckton Dickinson) and analyzed with FlowJo software (Tree Star, San Carlos, CA) (24–26).

Variable ß T-cell receptor (TCR) repertoire analysis

Since depletional induction therapies are administered at the time of transplantation, the initial post-depletional TCR repertoire is not initially influenced by alloantigen. If based on selective survival of memory cells, cells that exist as a result of prior environmental antigenic (or in the case of patients with autoimmune nephritis or diabetes, autoantigenic) exposure, the repertoire should vary widely based on the immune history brought to transplantation. Occasional losses in diversity manifested by clonal deletions and/or expansions within Vβ families should be expected. Conversely, homeostatic repopulation of naïve cells would yield a more diverse repertoire. To address this, we evaluated the peripheral TCR Vβ repertoire in depleted individuals during the first month after depletion and subsequently during the sixth month.

Peripheral blood mononuclear cells were stored in Trizol solution, then processed into RNA and cDNA using the manufacturer's recommended method. DNA was amplified using Vβ region primers (Sigma-Genosys, Woodlands, TX) for the 24 Vβ families and an unlabeled primer from the Cβ region (27–29). The PCR and runoff reactions were carried out as previously described (30). Jβ analysis was performed using 13 6-FAM-labeled Jβ primers in a 5 cycle runoff reaction of the amplified Vβ product as previously described (30). Electrophoresis was performed on a 3100 Applied Biosystems Genetic Analyzer (Applied Biosystems, Foster City, CA). Data were collected with 3100 Data Collection software version 1.1 (Applied Biosystems), and analyzed with GeneScan Analysis software, version 3.7 (Applied Biosystems).

Whole blood in vitro depletion

To model depletion in vitro, in the absence of homeostatic regulation, heparinized (1 u/mL) whole blood from healthy donors was incubated in 25-mL breathable cell culture flasks (Costar, Corning Incorporated, Corning, NY) at 37°C with 5% CO2. Alemtuzumab (0.4, 4 or 40 μg/mL) or RATG (5, 50 and 500 μg/mL) was added, and incubated for 8 h. The intermediate concentrations were representative of the serum concentrations achieved clinically during induction therapy. Live cell counts were performed by trypan blue exclusion and lymphocyte counts were established using an automated particle size analyzer (Coulter Z2 AccuComp, Beckman Coulter, Inc, Miami, FL). Polychromatic flow cytometry was performed as described above. All in vitro experiments were repeated five times.

Depleting antibody surface receptor density

To determine the ligand density for CD52 or RATG ligands, CD4+cells were isolated from PBMC using goat anti-mouse IgG-labeled beads (Dynabeads M-450, Dynal Biotech, Oslo, Norway) and a Mab sorting preparation (mouse anti-human CD14, CD20, CD11b, CD16, HLA-DR, CD8). Samples were stained with the following fluorescence-labeled Mabs: CD4FITC, CD8PE, CD14FITC, CD20PE, CD45RAFITC, CD62LPE (Becton Dickinson) and FITC-labeled alemtuzumab or RATG. Data were collected on a FACScan flow cytometer and analyzed with Cellquest software (Becton Dickinson).

Immunosuppressive incubations and intra-cellular cytokine staining

Given the variable rates of rejection seen in recent maintenance minimization protocols (5–11), and the markedly differing influence of certain classes of drugs in tolerance regimens; to wit, calcineurin inhibitors are thought to be antagonistic with tolerance induction regimens relying on T-cell regulation or costimulation blockade (31,32), we sought to determine whether clinically used immunosuppressants differed in their ability to control CD4+ effector memory phenotype T cells, specifically their IL-2 and interferon (IFN)-γ secretion upon TCR stimulation, in vitro.

Human PBMC were either evaluated in bulk culture or sorted by flow cytometry based on the phenotype CD45+, CD3+, CD4+, CD45RA- and CD62L- and suspended in RPMI (Invitrogen Corporation, Carlsbad, CA) supplemented with 10% FCS (Invitrogen) at 106 cells/mL. Immunosuppressive agents were added: sirolimus (Wyeth Pharmaceuticals, Madison, NJ) 100 μg/mL, DSG 100 μg/mL, methylprednisolone (Pfizer, New York, NY) 100 μg/mL and 1 mg/mL (consistent with levels achieved during treatment of rejection), cyclosporine A (CsA; Novartis International AG, Basel, Switzerland) 100 ng/mL, tacrolimus 10 ng/mL. After treatment, the preparation was stimulated with 100-μg/mL SEB (Sigma Chemical Co, St Louis, MO) for 8 h. Cells were washed in PBS (BioWhittaker, Walkersville, MD) and surface stained with the following Mab conjugates: CD3 Biotin (Becton Dickinson)-Quantum Dot 605 streptavidin (Quantum Dot Corporation, Hayward, CA) CD4-Cy5.5APC, CD8-TRPE, CD11a-Cy7APC, CD14-Cy7PE, CD20-Alexa430, CD45RA-Cy5.5PE, CD62L-Cy5PE. Cells were fixed and permeabilized (Cytofix/Cytoperm Kit, BD/Pharmingen). Intra-cellular cytokine staining was performed with IL-2PE and IFN-γ APC (Becton Dickinson).

To test the effect of immunosuppression on the proliferative capabilities of T cells, human PBMC were suspended in PBS at 106 cells/mL. CFSE (0.25 μM, Molecular Probes, Eugene, OR) was added and cells were incubated at 37°C for 7 min, washed with PBS supplemented with 10% FCS, and then resuspended (RPMI with 10% FCS). Staphylococcal enterotoxin B (SEB; 100 μg/mL) was added and the cells were incubated at 37°C for 96 h. Cells were washed and subsequently stained with CD3 Biotin-Quantum Dot 605 streptavidin, CD4 Cascade Blue, CD8-TRPE, CD11a-Cy7APC, CD14-Cy7PE, CD45RA-Cy5.5PE.

Statistics

Statistical differences between groups were determined by 2-tailed Student's t-test with significance defined as p ≤ 0.05.

Results

The predominant T-cell type present following antibody-mediated T-cell depletion is an activated memory-like T cell

We first evaluated the distinguishing characteristics of residual and repopulating T cells in transplant recipients undergoing depletional induction with alemtuzumab/deoxyspergualin (n = 5) or induction with RATG and maintenance with sirolimus (n = 5) compared to non-depleted patients (n = 3). The pre-transplant phenotypic distribution in all patients was similar to that seen in normal volunteers indicating that the peripheral T cells of adult humans, unlike many experimental animal models, is largely composed of cells with a memory phenotype. Following treatment with either alemtuzumab or RATG, patients had profoundly reduced lymphocyte counts from 1574 ± 159 cells/μL to 55 ± 10 cells/μL, and all T-cell phenotypes were, in absolute terms, greatly reduced (Figure 1A). However, differences in the relative depletion of various phenotypes were readily apparent in all depleted patients at all time points.

Figure 1.

T-cells with an effector memory phenotype are the predominant T cells present following antibody-mediated T-cell depletion. Results from polychromatic flow cytometry studies on (A) patients treated with alemtuzumab or RATG compared to (B) non-depleted transplant recipients, with particular emphasis on memory, naïve and regulatory T-cell phenotypes. The absolute number of all T-cell phenotypes is profoundly reduced by depletional therapy (panel A). However, differences in the relative depletion of CD4+ T-cell phenotypes were apparent. CD8+ T cells (only bulk population shown), CD4+ naïve T-cells (TN) and both CD45RA+ memory (TRA) and central memory (TCM) cells were depleted by greater than 99%. However, both CD4+CD25+ T-cells (TReg) and CD4+ effector memory cells (TEM) were comparatively spared, being depleted by 90%. CD4+ T-cells with an effector memory phenotype accounted for the vast majority of cells following depletion.

Subsets of CD8+ cells were homogenously depleted by >98% (Figure 1A). Similarly, CD4+ naïve T cells, and both RA+ and central memory cells were depleted by >99%. However, both CD4+CD25+ T cells and CD4+ CD45RA-CD62L- T cells were relatively spared, depleted by 90% (p = 0.001, Figure 1A). Lymphocytes were uniformly CCR7- throughout the first 3 weeks post-depletion (data not shown). This is consistent with the CD45RA-CD62L- cells being effector memory T cells. CD4+CD25+ T cells represented only 2.1 ± 0.3% of residual T cells, not significantly different than in un-manipulated individuals. In contrast, CD4+CD45RA-CD62L- effector memory cells accounted for 88 ± 3% of depletion-resistant T-cells; markedly different than their prevalence of 10–20% in normal volunteers or renal failure patients (unpublished observations). Both the predominance of CD4+ effector memory-like cells and the relative paucity of other populations were sustained over the first 3 weeks post-transplantation and markedly contrasted with a comparator group of non-depleted transplant recipients (n = 3) (Figure 1B).

Activated CD4+ effector memory-like T cells predominate peripherally and in the allograft during rejection

We then asked whether effector memory-like T cells were associated with allograft rejection. We limited our study to patients induced with alemtuzumab and DSG without maintenance immunosuppression as they uniformly experienced rejection within the study period (17). In these patients, the lymphocyte counts declined from 1591 ± 546 cells/μL to 56 ± 27 cells/μL. Despite this, all 5 patients experienced an episode of acute rejection within 1 month, at which time immunosuppression was initiated. All patients regained normal allograft function and were maintained rejection-free thereafter on monotherapy immunosuppression (sirolimus or tacrolimus).

Homeostatic repopulation as measured by lymphocyte counts was not evident in the weeks preceding rejection. The numbers of B-cells, NK-cells (data not shown), and all CD8+ subsets (Figure 2) remained extremely low, less than 3 cells/μL, throughout the pre-rejection period as were the numbers of CD4+ naïve, RA+ memory and central memory cells (Figure 2). In contrast, CD4+ effector memory phenotype cells represented the majority of the non-depleted cells and their numbers increased significantly (p = 0.03) preceding rejection, rising from a nadir of 18 ± 5 cells/μL to 41 ± 10 cells/μL. At rejection, they accounted for 64 ± 11% of lymphocytes, 83 ± 6% of T cells and 95 ± 2% of CD4+ cells (Figure 2). These cells remained >80% CCR7- in the 3 patients studied for this marker (data not shown).

Figure 2.

CD4+ memory cells predominate peripherally during lymphopenic rejection episodes. Polychromatic flow cytometry studies in 5 patients who received depletion as their lone immunotherapy are shown. CD8+ cells (only bulk population shown), CD4+ naïve cells (TN), CD4+CD45RA+ memory cells (TRA) and CD4+ central memory cell (TCM) numbers were extremely low and stable at less than 3 cells/μL in the weeks preceding rejection. In contrast, CD4+ T-cells with an effector memory phenotype (TEM) represented the majority of the non-depleted cells and their numbers increased significantly (p = 0.03) in the days preceding the time of rejection. At rejection, CD4+ effector memory phenotype cells were the predominant cell type.

To determine if the peripheral phenotype was also represented in the allograft, we performed immunohistochemical analysis. As previously described, the infiltrate was predominantly monocytic (10,33). However, essentially all T cells in the allograft were consistent with the peripheral phenotype in that they were CD45RO+ and predominantly CD4+ (data not shown). Given that T cells assume a memory phenotype upon activation, this finding was not surprising.

T cells with a CD4+ effector memory phenotype are relatively resistant to antibody-mediated depletion

Several processes might account for an in vivo predominance of CD4+ effector memory cells during repopulation including antigen-induced activation, and homeostatic proliferation (pseudo-memory phenotype) (23,34,35). Furthermore, antigen-experienced cells might have properties promoting their survival that could render them relatively resistant to complement-mediated lysis, apoptosis or other mechanisms of depletion. RATG, but not alemtuzumab, also contains antibodies with the potential for activating T cells (e.g. anti-CD3), and inducing a memory phenotype (36,37). To determine whether relative resistance to depletion was a mechanism contributing to selective cell sparing independent of homeostatic responses in vivo, we studied CD4+ T-cell depletion in vitro.

When treated with alemtuzumab or RATG, most lymphocytes were depleted as indicated by a >90% decline in viable cell counts (data not shown). However, CD4+ effector memory cells represented the vast majority of both alemtuzumab and RATG resistant cells (Figure 3). Following alemtuzumab treatment, CD4+ naïve cells, RA+ memory and central memory cells were all markedly depleted to 3 ± 2%, 5 ± 2% and 7 ± 2% of the remaining CD4+ cells, respectively. However, CD4+ effector memory cells increased from 28 ± 7 to 84 ± 6% of CD4+ cells. Following RATG treatment, a similar decrement in cell counts for CD4+ naïve cells, RA+ memory cells and central memory cells to 2 ± 1%, 5 ± 2% and 12 ± 4%, respectively. However, CD4+ effector memory cells again increased from 26 ± 8 to 82 ± 6% of CD4+ T cells. The differential effects were evident at all concentrations of both agents. Therefore, CD4+ effector memory cells are relatively resistant to antibody-mediated depletion in vitro, and this, in addition to the potential effects of homeostatic repopulation, can contribute to the observed predominance of these cells. The resistance was not due to differential expression of the ligands for RATG or alemtuzumab. Binding occurred equally highly amongst cell phenotypes without detectable ligand-negative cells (Figure 4).

Figure 3.

CD4+ effector memory phenotype cells are relatively resistant to antibody-mediated depletion in vitro. Blood from healthy volunteers were treated with increasing concentrations of either alemtuzumab (0.4, 4 and 40 μg/mL), or RATG (5, 50, 500 μg/mL), and then evaluated by polychromatic flow cytometry. The y-axis of each panel shows the percentage of CD4+ cells remaining after the indicated treatment. CD4+ effector memory cells (lower panel) represented the vast majority of both alemtuzumab- and RATG-resistant cells following the 8-h study period.

Figure 4.

Alemtuzumab and RATG bind equally to all CD4+ memory phenotypes. Shown are flow cytometry histograms from (A) alemtuzumab-FITC-labeled cells and (B) RATG-FITC-labeled cells. There are no detectible ligand-negative cells, and all cells are tightly grouped indicating equal expression of alemtuzumab and RATG ligands, respectively.

The post-depletional T-cell repertoire is highly skewed, oligoclonal and becomes increasingly diverse over time

Patients evaluated in the first month had highly variable repertoires, sometimes expressing as few as 2 TCR Vβ families (Figure 5). Jβ analysis of selected Vβ families confirmed oligoclonality with limited expression of the 13 Jβ families and markedly skewed oligoclonal spectratypes among those expressed (data not shown). By 6 months, spectratypes had become more Gaussian-like with polyclonal distribution of CDR3 lengths among both Vβ and Jβ families, and was as, if not more, diverse than the pre-transplant repertoire (data not shown). Thus, the post-depletional repertoire is consistent with an initial selective oligoclonal survival advantage that over time gives way to a more diverse repertoire having benefited from the results of homeostatic repopulation.

Figure 5.

The post-depletional T-cell repertoire is skewed. Shown are 22 Vβ families (pseudogenes Vβ10 and Vβ19 omitted) of 7 patients 1 month following T-cell depletion with alemtuzumab. There is marked loss of diversity with loss of expression of many Vβ families and oligoclonal distribution of CDR3 lengths in those expressed.

Calcineurin inhibitors most efficiently inhibit the function of CD4+ memory phenotype T cells

As expected (18), isolated CD4+ CD45RA-CD62L- T cells rapidly produced IL-2 and IFN-γ when treated with SEB (Figure 6A). Treatment with methylprednisolone, sirolimus or DSG had little effect on the sorted cells' cytokine production. However, both calcineurin inhibitors, CsA and tacrolimus, reduced the production of IL-2 by >95% and IFN-γ by >60%. Similarly, when PBMC were stained with carboxyfluorescein succinimidyl ester, and treated as described above, CD4+CD45RA- memory cells treated with sirolimus, DSG or methylprednisolone, underwent robust proliferation with up to four rounds of division (Figure 6B). Cells treated with CsA also divided, although the percentage of cells was reduced substantially. Contrariwise, there was no evident cell division in tacrolimus-treated memory cells. In memory cells treated with sirolimus, DSG and methylprednisolone, 20 ± 13% remained undivided, compared to 21 ± 8% in the control (p > 0.9). However, 55 ± 11% of CsA-treated cells and 91 ± 1% of tacrolimus-treated memory cells do not proliferate (p < 0.001). Calcineurin inhibitors, particularly tacrolimus, at low therapeutic concentrations, inhibited the proliferation and cytokine production of isolated or bulk culture memory phenotype cells.

Figure 6.

(A) Calcineurin inhibitors uniquely inhibit the function of CD4+ memory cells. Shown are the results of intra-cellular cytokine staining for IL-2 and IFN-γ of human CD4+CD11abrightCD45RA- memory cells sorted and then stimulated with SEB alone or in the presence of the indicated concentrations of sirolimus, DSG, methylprednisolone, CsA or tacrolimus and assessed their production of IL-2 and IFN-γ. Cytokine production is inhibited by CsA and tacrolimus, but not by the other agents. (B) Tacrolimus uniquely inhibits the proliferation of CD4+ memory cells. Shown are flow cytometry histograms depicting the proliferation of SEB-treated CD4+CD11abrightCD45RA- memory cells stained with carboxyfluorescein succinimidyl ester alone or in the presence of the indicated concentrations of sirolimus, DSG, methylprednisolone, CsA or tacrolimus. The cells treated with sirolimus, DSG or methylprednisolone, underwent robust proliferation with up to four rounds of division. The percentage of dividing cells was reduced substantially when treated with CsA. There was no evident cell division in tacrolimus-treated memory cells.

Discussion

We have shown that peripheral T cells from transplant recipients undergoing aggressive T-cell depletion are almost exclusively of the effector memory surface phenotype. These cells are the only ones with detectable proliferation immediately post-depletion, and are prevalent prior to and during rejection, peripherally and in the allograft. Furthermore, we have shown that T cells with this phenotype have selective resistance to therapeutic depletion, are resistant to several immunosuppressants, but are effectively inhibited by low-dose calcineurin pathway blockade. These data contravene the simplistic view of T-cell depletion as a pro-tolerant maneuver, and underscore the importance of considering the relative make-up of resurgent T-cell populations in designing anti-rejection therapies.

Although depletion greatly decreases cell number, our data suggest that the remaining cells are efficient mediators of alloimmunity. In a lymphopenic environment, both naïve and memory T cells undergo homeostatic proliferation (34,35). Naïve cells expand and express a pseudo-memory phenotype, maintaining some memory cell properties, but being suboptimal mediators of rejection (23). Conversely, homeostatic proliferation of established memory cells begets fully functional cells capable of vigorous rejection without the need for reexposure to antigen in the secondary lymphoid environment. Our data suggest that true memory cells are more likely the initial progenitors of cells instigating rejection, although the effects of homeostatic repopulation likely contribute and may dominate over time (38). As demonstrated by the rejection seen in lymphopenic patients, the residual cells, regardless of their mode of activation, are quite functional. These data are consistent with experiments suggesting a beneficial effect of T-cell repopulation on tumor immunity (39). Although concerning with regard to rejection, this may explain why few infectious complications have been reported following depletion combined with minimal maintenance immunosuppression (5–11).

The mechanisms involved in memory cell sparing remain undefined. Antigen-experienced cells have lower activation requirements and might up-regulate cell survival factors more readily than naïve cells. However, baseline expression of apoptotic-related genes (Bcl-2, Bcl-xl, Bax) or selective hyper-expression of complement regulatory genes (CD46, CD55, CD59) do not appear distinguishing between naïve and memory cells (unpublished observations). Alternatively, trafficking of memory cells could provide some protection from therapeutic antibodies since the serum concentration of antibodies is typically lower than tissue concentrations. Regardless, the effect is not absolute as most cells regardless of phenotype are depleted. Modest resistance confers a suitable cell survival advantage.

This is not to say that there are no benefits to depletion. Aggressively depleted patients have markedly delayed rejection and can be maintained rejection-free with limited maintenance immunosuppression (5–10). There are many plausible reasons for this. In the absence of naïve cells, a depleted individual is likely dependent on heterologous immunity (40,41). By limiting the available T-cell repertoire, as indicated by the markedly skewed repertoires in our patients, the number of cells with direct alloreactivity is reduced, thereby attenuating a predominant mechanism for acute rejection (42). A sparse population is also less prone to by-stander activation, and one can reasonably surmise fewer effectors available to mediate cytotoxic damage.

Practically, our data also suggest that calcineurin inhibitors are better suited for controlling post-depletional cells. Memory cells are exquisitely sensitive to low concentrations of tacrolimus both in terms of proliferation and cytokine production. These data may be germane to the design of immunosuppressive minimization strategies and are consistent with recent studies by Calne, Knechtle and Leventhal showing that depletional regimens combined with calcineurin inhibitors have lower acute rejection rates than those relying on sirolimus (5,6,8,11).

This study highlights potential limitations of depletion in some tolerance protocols. For example, rodent studies have shown that environmental antigen exposure prior to transplantation evokes resistance to tolerance induction (41,43–46). It has also been established that memory cells predominantly carry out murine allograft rejection in the face of aggressive induction therapy. A similar phenomenon should be expected in humans who have extensive exposure to viral antigens and resultant cross-reactive memory clones, which, as shown in this study, are less easily depleted.

The studies reported herein speak to a more complex view of depletion than has been appreciated clinically. Following depletion, patients do not simply have fewer cells, but rather have a cadre of cells with unique capabilities compared to non-depleted individuals. Thus, adding depletion to other modes of immunosuppression cannot be assumed to be synergistic. For example, cells with a memory phenotype are less dependent on costimulatory signals (18,19,45–47). Although depletion decreases the allo-specific precursor frequency, the remaining cells are less susceptible to therapies that interrupt naïve cell activation. It is also inappropriate to suggest that depletion will necessarily favor reconstitution of T cells in an environment dominated by regulatory cells. For depletion to take advantage of a pro-regulatory environment, some additional maneuver will likely be required.

Conceptually, depletional induction is increasingly being described as toleragenic rather than immunosuppressive, with the argument being that a reduced requirement for maintenance immunosuppression indicates that a person is more tolerant than one who is on high-dose immunosuppression (5,7). However, there is no biological basis for defining relative degrees of tolerance. Our data suggest that the observed benefits of depletion are less indicative of tolerance than demonstrated limitations of post-depletional immunity. Post-depletional T cells are easily controlled by calcineurin inhibitors, oligoclonal, and severely limited with regard to CD8+ cytotoxic effectors. They are not, however, inherently incapable of responding to the allograft. Thus, while depleted patients may be more easily immunosuppressed, they are not necessarily more tolerant.

Acknowledgments

The authors gratefully acknowledge the nurses and transplant coordinators of the Organ and Tissue Transplant Research Center of the Warren G. Magnuson Clinical Center for their excellent patient care skills; Christine Chamberlain for her clinical pharmacology support, Millennium Pharmaceutical Corporation, for their generous gift of alemtuzumab, Sangstat Medical Corporation, for their generous gift of Thymoglobulin, Steve Perfetto for his assistance with the cell sorts and Dr. Ronald Germain for helpful hallway conversations.

Ancillary