SEARCH

SEARCH BY CITATION

Keywords:

  • CD4 T cells;
  • homeostasis;
  • memory;
  • thymoglobulin;
  • transplantation

Abstract

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Acknowledgments
  8. References

T-cell depletion reportedly leads to alterations in the T-cell compartment with predominant survival of memory phenotype CD4 T cells. Here, we asked whether the prevalence of memory T cells postdepletion results from their inherent resistance to depletion and/or to the homeostatic expansion of naive T cells and their phenotypic conversion to memory, which is known to occur in lymphopenic conditions. Using a ‘mosaic memory’ mouse model with trackable populations of alloreactive memory T cells, we found that treatment with murine antithymocyte globulin (mATG) or antilymphocyte serum (ALS) effectively depleted alloreactive memory CD4 T cells, followed by rapid homeostatic proliferation of endogenous CD4 T cells peaking at 4 days postdepletion, with no homeostatic advantage to the antigen-specific memory population. Interestingly, naive (CD44lo) CD4 T cells exhibited the greatest increase in homeostatic proliferation following mATG treatment, divided more extensively compared to memory (CD44hi) CD4 T cells and converted to a memory phenotype. Our results provide novel evidence that memory CD4 T cells are susceptible to lymphodepletion and that the postdepletional T-cell compartment is repopulated to a significant extent by homeostatically expanded naive T cells in a mouse model, with important important implications for immune alterations triggered by induction therapy.


Introduction

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Acknowledgments
  8. References

T-cell depletion through antibody-mediated induction therapy is routinely used to promote long-term graft survival in solid organ transplantation. Although a number of T-cell depleting and nondepleting antibody preparations are used as induction therapy (1), their use in kidney transplantation is dominated by antithymocyte globulin (ATG: 42.3%) (2), resulting in a dramatic decrease in circulating T cells lasting weeks to months (3–6). However, the immune system has the capacity to recover T-cell numbers through the homeostatic proliferation of peripheral T cells (7), and by newly generated naive T cells emerging from the thymus. Understanding the origin and composition of the postdepletional T-cell compartment is critical for optimizing immunotherapies to prolong graft survival.

Irradiation-induced lymphopenia is known to trigger rapid homeostatic proliferation of the remaining peripheral T cells (8). In mouse models, transfer of CD4 or CD8 T cells into irradiated hosts or into congenitally lymphocyte-deficient mice results in extensive T-cell proliferation, resulting in naive T cells adopting memory phenotypes and function (9,10). T cells that acquire a memory phenotype solely due to homeostatic expansion and in the absence of primary antigen stimulation can mediate secondary responses to pathogens (8,11), indicating that a lymphopenic environment can promote polyclonal T-cell differentiation. Whether ATG therapy can trigger homeostatic conversion of endogenous naive T cells to memory T cells has not been demonstrated.

In transplantation, the presence of memory T cells specific for alloantigens can be detrimental to graft survival due to their enhanced activation, functional and migration properties compared to naive counterparts (12). Alloreactive memory T cells can be generated from a previous transplant, pregnancy, transfusion or via heterologous cross-reactivity (13). It is well known that presensitized patients experience increased rates of acute rejection leading to decreased graft survival (14,15). Because of their distinct activation and functional properties, memory T cells can be differentially susceptible and/or resistant to immunosuppressive therapies (16) and interfere with tolerance induction regimens (17). Therefore, evaluating the presence of memory T cells has recently emerged as an important consideration for clinical optimization of immunosuppression.

Studies in mice and in patients indicate that memory-phenotype T cells predominate in peripheral blood following T-cell depletion (10,16), suggesting that memory T cells may be either resistant to depletional therapy and/or may be more rapidly replenished during homeostasis. In a clinical study, patients treated exclusively with T-cell depletional agent exhibited a predominant effector-memory CD4 T-cell profile in the remaining lymphocytes that was associated with acute rejection (16). Whether this memory predominance was due to depletion-resistant memory T cells and/or phenotypic conversion of homeostatically expanded naive T cells was not determined.

In this study, we used a mouse model with a trackable population of alloreactive memory CD4 T cells to determine their in vivo susceptibility to two T-cell depleting agents: antilymphocyte serum (ALS) and ATG, and evaluate whether peripheral naive, memory and/or regulatory T cells (Tregs) could reconstitute the postdepletion compartment. We demonstrate here that alloreactive memory T cells are readily depleted by ALS or mATG, and that both memory and naive CD4 T cells undergo postdepletional homeostatic proliferation. Notably, naive CD4 T cells exhibit the highest level of homeostatic expansion postdepletion compared to memory T cells or Tregs, which results in their conversion to a memory phenotype. Our findings demonstrate that induction therapy can deplete preexisting memory T cells and trigger previously quiescent naive T cells to expand and adopt memory phenotypes, with potential implications for immune alterations in the postdepletion T-cell compartment.

Materials and Methods

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Acknowledgments
  8. References

Mice

C57BL/6 and BALB/c mice were purchased from The National Cancer Institute Biological Testing Branch (Frederick, MD) and used at 8–12 weeks of age. BALB/c(Thy1.1) mice were bred as homozygotes and maintained in the Animal Facility at the University of Maryland (Baltimore, MD) under specific pathogen-free conditions. All studies involving animal subjects were approved by the Institutional Animal Care and Use Committee at the University of Maryland.

Antibodies and flow cytometry

The following antibodies were purchased from Bio-X-Cell (West Lebanon, NH): anti-CD8 (TIB105), anti-CD4 (GK1.5), anti-I-Ad (212.A1) and antiThy-1 (TIB238). PE-conjugated anti-CD90 Thy 1.1 and anti-CD62L44; FITC-conjugated anti-CD25; PerCP-conjugated anti-CD4 and anti-CD4; APC-conjugated anti-CD44; FITC-conjugated anti-CD3; PE-conjugated anti-FoxP3 were purchased from BD Pharmingen (San Diego, CA). For analysis of surface phenotype, cells were incubated with fluorescently coupled antibodies in stain buffer as described (18) and analyzed by flow cytometry using the LSRII flow cytometer (BD-Biosciences, San Jose, CA) and FACSDiva (BD-Biosciences) or FlowJo software (Tree Star Inc., Ashland, OR).

Treatment of mice with T-cell depleting agents

Rabbit antimurine lymphocyte serum (ALS) was purchased from Accurate Chemical & Scientific Corporation (Westbury, NY). Anti-mouse thymocyte globulin (mATG), generated by immunizing rabbits with a mixture of thymocytes from eight different strains of mice (C57BL/6, BALB/c, DBA/2, 129, C3H, SJL, Swiss Webster, ICR) as described (19), and control rabbit IgG were provided by Genzyme Corporation (Cambridge, MA), where tests for quality control and functional activities were performed (19). In some cases, ATG and control rabbit IgG were purchased from Accurate Chemical and Scientific Corporation. Mouse hosts were administered 0.2 mL ALS or PBS control intraperitoneally (i.p.) daily for 3 days, or mATG or control IgG in two doses of 500 μg/0.2 mL i.p. 48 h apart as described (20). Peripheral blood, spleen and lymph nodes were recovered at indicated times posttreatment.

Generation of mosaic-memory mice

‘Mosaic-memory’ mice containing a marked population of alloantigen-specific memory CD4 T cells were generated as previously described (18). Briefly, CD4 T cells purified from spleens of BALB/c mice (21) were cultured (2 × 106/mL) with mitomycin C-treated APC (6 × 106/mL) prepared from C57BL/6 splenocytes in supplemented Clicks media (Irvine Scientific, Irvine, CA) for 3–4 days at 37°C and 5% CO2. The alloantigen-primed fraction was magnetically sorted for upregulation of CD25 expression (Miltenyi Biotec, Auburn, CA) (18), and intravenously transferred (2 × 106/mouse) into congenic BALB/c(Thy1.1) hosts, who had previously undergone either a thymectomy (THX) or sham THX (Sham) procedure as described (22), and 6–8 weeks later, the resultant mosaic-memory mice (18) were administered mATG, IgG, ALS or PBS as above. In some cases mice were administered BrdU (100 μL of 10 mg/mL i.p.; BD-Biosciences) for 4 consecutive days as described (22), beginning 1 day after the last treatment.

Cytokine assay

The frequency of alloantigen-specific CD4 T cells producing IL-2 and IFN-γ was determined using ELISPOT as described (18). Briefly, CD4 T cells isolated from IgG- or mATG-treated mice were cultured (105 cells/well) in triplicate with media, BALB/c APC or allogeneic C57BL/6 APC (3 × 105/well) in 96-well MultiScreen-IP plates (Millipore, Billerica, MA). ELISPOTs were quantitated using the Immunospot reader (Cellular Technology, Cleveland, OH) as described (23).

In vivo CFSE proliferation assays

Naive (CD44lo) and memory (CD44hi) CD4 T cells were sorted from BALB/c(Thy1.1) spleens by MACS separation using APC-anti-CD44 antibodies and anti-APC-conjugated microbeads (Miltenyi Biotec) as described (24). Purified naive and memory subsets were labeled with 5 μM CFSE (Molecular Probes, Eugene, OR) and transferred intravenously (2 × 106/mouse) into mATG- or control-IgG-treated BALB/c mice. Spleens were recovered from recipient mice 7 days posttransfer and CFSE content determined by flow cytometry. The percentage of cells entering 1–6 cell divisions was calculated as follows (% cells divided = total number of cells divided/(total number of cells divided + number of undivided cells) (25).

Statistical methods

All data are reported as means ± standard deviation. A nonpaired Student's t-test was used to evaluate differences between groups. Significance was determined at the 95% confidence interval.

Results

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Acknowledgments
  8. References

Antilymphocyte serum treatment efficiently depletes alloreactive memory T cells

We evaluated the susceptibility of alloreactive memory T cells to depletion using a ‘mosaic-memory’ model in which in vitro-primed alloreactive effector cells (Thy 1.2+) are adoptively transferred into BALB/c(Thy1.1+) congenic mouse hosts, resulting in the development and persistence of a trackable population of alloreactive memory CD4 T cells (Thy1.2+) among a full complement of endogenous T cells (Thy1.1+) (18). In addition, we generated both euthymic (Sham) and thymectomized (THX) mosaic-memory mouse hosts to monitor effects of thymic output and model pediatric and adult transplant recipients, respectively. We treated Sham and THX groups of mosaic-memory mice with ALS or control PBS (see section ‘Materials and Methods’) and recovered spleen and lymph nodes at 4, 7 and 14 days after injection. We found that alloreactive memory CD4 T cells were efficiently depleted (85%) and remained at low to undetectable frequencies for up to 3 weeks after ALS administration compared to their persistence in control-treated mice (p < 0.05) (Figure 1A and data not shown), demonstrating that memory CD4 T cells are susceptible to antibody-mediated depletion in vivo. In addition, we found that ALS administration led to a persistent and significant depletion in the frequency of naive CD4 T cells in both Sham (51%) and THX mice (85%) compared to controls (Figure 1B).

image

Figure 1. Alloreactive memory CD4 T cells are depleted and do not rapidly reconstitute following ALS treatment. (A) Mosaic memory mice containing an alloreactive anti-H-2b population of memory CD4 T cells were generated in BALB/c mice (see section ‘Materials and Methods’) and groups underwent either thymectomy (THX) or sham operation (Sham) before being treated with antilymphocyte serum (ALS) or PBS. Flow cytometry plot shows the frequency of alloreactive memory Thy 1.2+ CD4 T cells at 4, 7 and 14 days following ALS or PBS treatment. (B) Frequency of naive CD4 T cells before and 4–14 days after ALS administration in euthymic (Sham) or thymectomized (THX) animals as in (A). (C) Homeostatic proliferation of memory and endogenous CD4 T cells after ALS depletional therapy. Graph shows percentage of BrdU+ CD4 T cells among Thy 1.2+ alloreactive memory or endogenous CD4 T-cell population over time after the administration of ALS or PBS in euthymic (Sham) or thymectomized (THX) hosts from five mice per group.

Download figure to PowerPoint

To determine whether the endogenous or memory CD4 T cells remaining after ALS treatment exhibited proliferative expansion, we administered BrdU for 4 days prior to recovery and analyzed BrdU incorporation from endogenous and memory CD4 T-cell populations. Both endogenous and alloreactive memory CD4 T cells showed extensive proliferation early after ALS treatment, peaking 4 days after depletional therapy and declining by 14 days (Figure 1C). The highest percentage of proliferating CD4 T cells was observed in ALS-treated thymectomized mice (Figure 1C, solid lines), compared to the ALS-Sham group (Figure 1C, dashed lines), with negligible proliferation observed in control-treated animals, indicating that rapid homeostatic expansion occurs as a result of ALS-mediated depletion, and that the presence of thymic output dampens the extent of lymphopenia-induced proliferation. Notably, proliferation of endogenous CD4 T cells following ALS treatment greatly exceeded that of alloreactive memory CD4 T cells in both thymectomized and euthymic mice (Figure 1C), demonstrating that a specific population of memory T cells does not exhibit a homeostatic advantage in repopulating the postdepletional T-cell compartment compared to endogenous (mostly naive) lymphocytes. These results provided the rationale to pursue the dynamics of naive and memory T-cell depletion and reconstitution using a more clinically relevant depletional agent.

mATG treatment rapidly alters naive: memory CD4 T-cell ratio

We used a murine form of thymoglobulin (mATG) as a clinically relevant agent to assess how naive and memory T-cell subsets were depleted and reconstituted, and the effect of thymic output on this process. We treated THX and Sham mice with mATG or IgG, and assessed the kinetics of depletion and CD4 T-cell reconstitution by analyzing absolute numbers and phenotype of peripheral blood T cells in mice at 1–3 weeks following treatment. We found that mATG led to a significant decline in CD4+CD3+ counts in both THX (p < 0.01) and Sham hosts (p = 0.029) compared to control IgG that persisted for up to 2 weeks (Figure 2A). By the third week after mATG injection, Sham hosts had recovered their CD3+CD4+ counts to the level of IgG-treated hosts (Figure 2A, p = 0.20), whereas T-cell numbers were still reduced in mATG-treated THX hosts. These results demonstrate that lymphopenia is transient in the presence of thymic output and prolonged in thymectomized hosts.

image

Figure 2. Administration of murine thymoglobulin (mATG) results in an altered ratio of circulating naive and memory CD4 T cells. Thymectomized (THX) or euthymic (Sham) BALB/c mice were treated with mATG or IgG control and peripheral blood was analyzed by flow cytometry at the indicated time points. (A) Graph shows the frequency of circulating CD4+CD3+ T cells prior to (Pre) and at 1–3 weeks after treatment in Sham-ATG (black bars or circles), THX-ATG (grey bars or circles) or control rabbit IgG (white bars or circles). (B) Representative flow cytometry plots showing naive and memory phenotypes of peripheral blood CD4 T cells pre- and postdepletion. (C) The frequency of memory CD44hi (top) and naive CD44lo (B) phenotype CD4 T cells is shown prior to and at 1–3 weeks post-mATG treatment. Data are expressed as mean ± SD with *p < 0.05 vs. Pre and **p < 0.05 vs. IgG from four mice per group.

Download figure to PowerPoint

Treatment with mATG treatment also resulted in a significant increase in the proportion of CD44hi memory-phenotype cells at 1–2 weeks postinduction (p = 0.001) in both Sham and THX hosts that returned to baseline levels by week 3 in Sham but not in THX mice (Figure 2B,C) Conversely, we observed a significant decline in the CD44lo naive phenotype CD4 T cells at 1 week following mATG administration in both THX (p < 0.001) and Sham (p = 0.001) animals (Figure 2B,C), which persisted for 2 weeks in Sham (p = 0.03) and 3 weeks in THX mice (p = 0.0015). Administration of control IgG did not alter the percentage of circulating naive (CD44lo: 81%) or memory (CD44hi: 19%) cells during the 3 weeks after injection (Figure 2C). These results demonstrate that mATG treatment results in a dramatic shift in CD4 T-cell subset distribution from predominantly naive to a predominantly memory phenotype, and that the duration of this shift depends on thymic output.

Alloreactive memory CD4 T cells are sensitive to antithymocyte globulin and show diminished recall cytokine response to foreign antigens

We asked whether alloreactive memory CD4 T cells exhibited a similar susceptibility to depletion by mATG as we saw with ALS (Figure 1). We set up similar groups of THX and Sham mosaic-memory mice, and treated them with either mATG or control IgG. Similar to our results with ALS, mATG administration effectively depleted the alloreactive memory cell population by 86.5 ± 2.4% in THX mice and by 87.9 ± 3.5% in Sham (euthymic) mice (Figure 3A,B). These results demonstrate that memory T cells are not intrinsically resistant to mATG depletion nor uniquely sensitive to ALS. To determine whether ATG likewise induced a functional loss of alloreactive memory CD4 T cells, we assessed the presence of allo-specific memory CD4 T cells before and after mATG treatment using ELISPOT. Mosaic-memory mice have a high frequency of alloreactive memory CD4 T cells producing IFN-γ and IL-2 (Figure 3C and (18)), compared to the baseline alloreactive memory population in naive BALB/c mice (dashed line). Following treatment with mATG, there was a marked reduction of alloantigen-specific memory CD4 T cells in both euthymic and THX hosts, compared to IgG control-treated hosts (Figure 3C), establishing that mATG functionally depletes alloreactive memory CD4 T cells to the baseline level found in naive animals.

image

Figure 3. Alloreactive memory T cells are sensitive to mATG-induced lymphodepletion and maintain a diminished cytokine response to foreign antigen. Mosaic memory mice were generated as in Figure 1, and groups underwent either thymectomy (THX) or sham operation (Sham) before being treated with mATG or IgG as in Figure 2. (A) Frequency (A) and absolute numbers (B) of CD4+ Thy 1.1+ alloreactive memory CD4 T cells following the administration of mATG or control IgG to euthymic (Sham) and thymectomized (THX) hosts, respectively. Data are expressed as mean ± SD with *p < 0.05 vs. IgG from eight mice per experimental group. (C) Functional analysis of alloreactivity before and after mATG administration. CD4 T cells were isolated from mosaic memory mice prior to and 4 days after mATG or IgG administration as in (A), stimulated with allogeneic APC from C57BL/6 mice and the frequency of IFN-γ and IL-2 producers was determined by ELISPOT. The alloreactive precursor frequency in naive mice is indicated by a dotted line for both graphs. Results are shown as an average of four mice per group (mean ± SD).

Download figure to PowerPoint

Relative homeostasis of regulatory, naive and memory T cells following mATG treatment

We analyzed polyclonal naive and memory phenotype CD4 T cells in control and mATG-treated BALB/c mice (Figure 4) to determine the contributions of naive and memory subsets to postdepletion T-cell reconstitution and homeostasis. We also analyzed in vivo homeostasis of the regulatory T cells subset (Tregs) marked by nuclear expression of FoxP3 as Tregs undergo homeostasis in steady-state and lymphopenic conditions (26,27). Following the administration of mATG to THX or Sham mice, both FoxP3+ and FoxP3 CD4 T cells exhibited increased proliferation compared to IgG-treated animals (p = 0.001), with a more striking increase in FoxP3 CD4 T-cell proliferation compared to FoxP3+ Tregs (Figure 4A, p = 0.003), demonstrating that the non-Treg (FoxP3) population comprising naive and memory subsets is the predominant group mediating homeostatic proliferation following mATG.

image

Figure 4. Relative changes in homeostatic proliferation of regulatory, naive and memory CD4 T cells following mATG administration. BALB/c mice were thymectomized or sham operated, treated with mATG or IgG as in Figure 2 and subsequently administered 10 μg/mL BrdU daily for 4 successive days, after which the spleens were recovered and analyzed. Representative flow cytometry dot plots (left) and corresponding histograms (right) show percent BrdU incorporation by FoxP3+ regulatory and FoxP3 nonregulatory CD4 T cells (A) and by FoxP3 CD44lo (naive) and CD44hi (memory) CD4 T cells (B) following mATG or IgG administration in Sham and THX hosts. (C) Compiled analysis of BrdU incorporation from naive, memory and regulatory subsets in mATG and control-treated THX and Sham hosts. Data are presented as mean ± SD with *p < 0.05 vs. CD44hi or Foxp3+, respectively; **p < 0.05 as indicated from five mice per group, representative of four experiments.

Download figure to PowerPoint

We found extensive proliferation of both naive (CD44lo) and memory (CD44hi) CD4 T cells early after mATG treatment (20%–28% proliferating), with two important differences in homeostasis between these subsets (Figure 4B). First, naive CD4 T cells exhibited the greatest increase compared to memory and regulatory T cells in homeostatic proliferation following mATG treatment, with a 15-fold increase in BrdU incorporation in mATG-treated vs. IgG-treated THX mice and a 10-fold increase in mATG-treated THX mice vs. IgG-treated animals (p = 0.001). By contrast, memory and Treg proliferation in mATG compared to IgG-treated hosts increased less than two fold, as these subsets already exhibit a significant level of steady-state proliferation in replete hosts. Second, only naive T cells in thymectomized mice showed a marked increase in cellular proliferation compared to the sham-operated group (p = 0.007), suggesting that de novo thymic output of naive T cells during the period of homeostatic expansion may specifically compete with peripheral naive CD44lo cells. These results indicate that naive CD4 T cells undergo the greatest increase in homeostatic expansion following lymphodepletion and exhibit the highest level of division in thymectomized hosts (Figure 4C).

Quantitative analysis of postdepletion homeostasis

To quantitatively analyze the extent of homeostatic proliferation by naive and memory CD4 T-cell subsets following mATG treatment, we employed a sensitive in vivo CFSE dilution assay. We purified naive CD44lo and memory CD44hi phenotype CD4 T cells from BALB/c(Thy1.1) mouse spleens, CFSE-labeled them, transferred them into BALB/c(Thy1.2+) congeneic hosts treated 4 days previously with mATG or IgG (the time point for peak depletion/homeostasis shown in Figures 1 and 2) and assessed in vivo proliferation after 7 days. We observed two distinct populations undergoing cellular division in mATG-treated hosts—a fast-dividing population that completely diluted out CFSE, and a slower-dividing population that exhibited only two to three divisions (Figure 5A), similar to the pattern observed when CD4 T cells are transferred into irradiated hosts (9). We utilized a previously validated mathematical formula (25) to calculate the overall percentage of dividing cells in each group following mATG treatment as well as the percentage in each division cycle (Figure 5B,C). The results show that naive CD4 T cells divide at a significantly greater rate than those with a memory phenotype in ATG-treated mice (p = 0.014). Consistent with BrdU incorporation results, we observed more extensive cell division in the naive CD44lo compared to memory CD44hi fraction in thymectomized animals (p = 0.001). We also observe a notable increase in CD44lo cell division in mATG-treated THX mice compared to their euthymic cohorts, whereas the extent of CD44hi cell division was comparable in ATG-treated THX and euthymic hosts (Figure 5B). These results demonstrate that naive T cells experience more homeostatic expansion as a result of mATG treatment compared to memory counterparts, particularly in THX hosts.

image

Figure 5. Quantitative analysis of naive and memory CD4 T-cell proliferation in mATG-treated hosts reveals extensive naive CD4 T-cell homeostasis. CD44hi memory or CD44lo naive T cells were purified from BALB/c(Thy1.1) mice, CFSE-labeled and transferred (2 × 106/mouse) into thymectomized or sham treated BALB/c mice treated with mATG or IgG control as in Figure 2 at 5 days posttreatment. Spleen cells were recovered 7 days later. (A) Representative histogram showing the percent of dividing cells from the total CD44lo (upper) or CD44hi (lower) precursor population in IgG- or mATG-treated thymectomized (THX) mice. (B) Effect of thymectomy on homeostatic proliferation of naive and memory CD4 T cells. Percentage of naive CD44lo (black bars) and memory CD44hi (white bars) T cells undergoing division, as determined by CFSE dilution as in (A), in mATG-treated euthymic (Sham) or thymectomized (THX) hosts. Results are shown as mean ± SD from seven mice per group. (C) Extent of division by naive and memory CD4 T cells in mATG-treated hosts. Graphs show relative percentage of naive (CD44lo, black bars) and memory (CD44hi, white bars) T cells undergoing mATG-induced homeostatic proliferation at each cell cycle compared to their respective starting population in Sham-operated (left) and thymectomized (THX) hosts. Data are presented as mean ± SD from seven mice per group with *p < 0.05 vs. CD44lo; **p < 0.05 vs. Sham.

Download figure to PowerPoint

We further quantitated the extent of cell division within each division cycle. We found that the majority of naive and memory CD4 T cells underwent one division in mATG-treated hosts, with significant numbers undergoing two divisions, particularly in thymectomized hosts (Figure 5C). Interestingly, although a qualitative spike in maximally divided cells is observed for both naive and memory subsets in mATG-treated hosts, the actual frequency of cells undergoing maximal division is minimal, emphasizing the importance of quantitative analysis of CFSE dilution studies.

Phenotypic shift in naive T cells during homeostasis

We asked whether mATG-triggered homeostatic proliferation of naive T cells would result in conversion to a memory phenotype, similar to that shown in irradiated or congenitally lymphopenic mouse hosts (9,10). We found that naive CD44lo CD4 T cells express higher levels of CD44 with successive cell divisions, such that maximally divided cells exhibit a CD44hi phenotype (Figure 6). Memory phenotype CD44hi CD4 T cells maintained their CD44hi phenotype at all division cycles. This phenotypic conversion of naive to memory CD4 T cells was comparable in mATG-treated thymectomized and euthymic hosts. These results demonstrate that lymphopenia induced by mATG treatment can trigger robust homeostatic expansion of naive T cells, resulting in their phenotypic conversion to memory T cells.

image

Figure 6. Naive CD44lo T cells convert to a memory phenotype during mATG-induced homeostatic proliferation. Animals were treated in a similar manner as described in Figure 5 and analyzed. (Top) Flow cytometry plots showing the phenotypic conversion of naive CD44lo T cells to a CD44hi phenotype with cell division, as determined by CFSE dilution, following the administration of mATG compared to controls (IgG). Cellular division without phenotypic conversion in the memory T-cell population exposed to mATG is shown below. Results are representative of seven mice per group.

Download figure to PowerPoint

Discussion

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Acknowledgments
  8. References

T-cell depletional strategies such as ATG now constitute a routine part of induction therapy in solid organ transplantation. In this study, we provide direct evidence in a mouse model that memory T cells are susceptible to T-cell depletional strategies and that preexisting alloreactive memory can be functionally depleted by ATG. We further demonstrate rapid homeostatic expansion in the early postdepletional period by naive and memory CD4 T cells, with naive CD4 T cells exhibiting the greatest increase in homeostatic turnover, particularly in thymectomized hosts. Homeostatically expanding naive T cells can readily expand and convert to memory phenotype cells following depletional therapy, suggesting the potential for polyclonal T-cell priming in T-cell-depleted hosts.

Previous studies have identified memory T-cell predominance in the postdepletional T-cell compartment (10,16), leading to the prevailing view that memory T cells may be relatively resistant to depletional therapies compared to their naive counterparts. However, these studies did not address early homeostatic expansion and may have missed the dynamic changes and early phenotypic transitions we observed in this study. Depletional therapy leads to a significant decline in peripheral T-lymphocytes through multiple processes ranging from antibody-mediated/complement-dependent cytotoxicity, opsonization of lymphocytes and Fas/Fas-ligand mediated apoptosis (3–5). The depleting antibodies within human ATG are known to recognize pan T-cell markers such as CD2, CD3, CD4, CD8, CD25 and CD45, and murine ATG recognizes a comparable array of surface molecules (19). Moreover, human memory T cells are known to be susceptible to in vivo depletion using anti-CD2 antibodies (Alefacept) (28), indicating that memory T cells are not inherently resistant to antibody-mediated depletion regimes. We also show here that regulatory T-lymphocytes constitute only a small proportion of the homeostatically dividing cell lines following mATG, similar to findings in other systems (10,19,29).

T-cell reconstitution in the lymphopenic environment is attributed to a combination of de novo thymic replenishment and peripheral turnover of residual T-cell populations. Thymic replenishment occurring through positive and negative selection is highly dependent upon T-cell receptor (TCR) engagement (30). By contrast, the rapid peripheral turnover of T cells in lymphopenic conditions is highly dependent upon cytokines as well as TCR interactions (31). We found that naive T-cell proliferation due to mATG-induced lymphopenia was even greater in the absence of thymic output in thymectomized mice, whereas proliferation of memory T cells in mATG-treated hosts was not affected by thymic output and was similar in thymectomized and euthymic hosts. These results suggest that naive CD4 T cells specifically compete for niche space with other naive T cells (but not with memory or Tregs), perhaps due to competition for homeostatic cytokines and/or MHC class II/peptide ligands on antigen-presenting cells (9, 32). The lack of increased homeostatic expansion by memory T cells in thymectomized hosts was not due to dispersion to extranodal sites, as we did not find significant numbers of CD44hi CD4 T cells in nonlymphoid tissue such as lung and liver (data not shown).

We show that mATG-induced homeostatic proliferation can induce the phenotypic conversion of naive CD44lo T cells into a memory phenotype without any foreign antigenic stimulation. A similar phenomenon has been reported to occur in nude mice (33) as well as in mice lacking secondary lymphoid organs (34). Therefore, the increase in the proportion of CD44hi phenotype CD4 T cells following mATG can be attributed to homeostatically expanding naive T cells that adopt memory phenotypes. We investigated whether there was biased expansion of certain naive T cells based on TCR repertoire and found similar distribution of 14 Vβ family members in nondepleted and the postdepletional T cells (data not shown), indicating polyclonal T-cell expansion in the ATG-induced lymphopenic environment. In addition, ELISPOT analysis of alloantigen-specific CD4 T cells prior to and 4–14 days after ATG treatment of naive mice did not reveal an increase in the frequency of alloreactive CD4 T cells in the postdepletional T-cell compartment (data not shown). Together, these results suggest an intrinsic, unbiased homeostatic expansion of naive CD4 T cells, at least in the absence of an allograft.

In Figure 7, we depict three distinct homeostatic states identified in this study. In steady-state or immune replete condition, memory T cells exhibit the highest level of turnover followed by Tregs, with naive T cells exhibiting minimal turnover. Following ATG-mediated depletion, T-cell homeostasis is increased, with naive T cells now experiencing significant proliferative turnover, comparable or higher than that of memory T cells. The duration of lymphopenia and the magnitude of naive T-cell homeostasis also depends on thymic output, with increased homeostasis and lymphopenia in the absence of a thymus, suggesting differences in T-cell reconstitution in adult vs. pediatric transplant recipients.

image

Figure 7. Proposed model for T-cell reconstitution and homeostasis during the early period following T-cell depletion. Diagram showing relative homeostasis of naive (N) memory (M) and regulatory (T) CD4 T cells in steady-state conditions (replete, top), T-cell depleted and euthymic host or pediatric transplant patient (middle) and T-cell depleted and thymectomized host or adult transplant patient (bottom).

Download figure to PowerPoint

Our results suggest potential clinical consequences of induction therapy. Depletion of preexisting memory T cells by ATG can diminish preexisting alloreactive or autoreactive memory T-cell populations in presensitized or autoimmune individuals, respectively. However, the concomitant and rapid homeostatic expansion of residual naive T cells following ATG therapy and their conversion to memory, could results in memory generation of T-cell specificities that are normally quiescent in healthy conditions. The physiological consequences of this altered naive T cells compartment will require an understanding of the mechanisms that control homeostatic expansion in thymoglobulin-induced lymphopenia. Further study of T-cell depletion and reconstitution in the presence of a transplanted organ and/or maintenance immunosuppression will provide valuable insight for identifying and treating immune alterations brought about by induction therapies.

Acknowledgments

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Acknowledgments
  8. References

The authors thank Dr. Stephen Bartlett for critical review of this paper. This work was supported by a Roche Organ Transplantation Research Foundation (ROTRF) grant awarded to D.L.F. A.S. is supported by the American Society of Transplant Surgeons-Novartis Fellowship in Transplantation Award.

References

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Acknowledgments
  8. References