L-Selectin Is Dispensable for T Regulatory Cell Function Postallogeneic Bone Marrow Transplantation

Abstract

In murine models, the adoptive transfer of CD4⁺/CD25⁺ regulatory T cells (T_regs) inhibited graft-versus-host disease (GvHD). Previous work has indicated a critical role for the adhesion molecule L-selectin (CD62L) in the function of T_regs in preventing GvHD. Here we examined the capacity of naive wild-type (WT), CD62L^−/− and ex vivo expanded CD62L^Lo T_regs to inhibit acute GvHD. Surprisingly, we found that CD62L^−/− T_regs were potent suppressors of GvHD, whereas CD62L^Lo T_regs were unable to inhibit disease despite being functionally competent to suppress allo T cell responses in vitro. Concomitant with improved outcomes, WT and CD62L^−/− T_regs significantly reduced liver pathology and systemic pro-inflammatory cytokine production, although CD62L^−/− T_regs were less effective in reducing lung pathology. While accumulation of CD62L^−/− T_regs in GvHD target organs was equivalent to WT T_regs, CD62L^−/− T_regs did not migrate as well as WT T_regs to peripheral lymph nodes (PLNs) over the first 2 weeks posttransplantation. This work demonstrated that CD62L was dispensable for T_reg-mediated protection from GvHD.

Introduction

Allogeneic hematopoeitic stem cell transplantation (allo-HSCT) is a potentially curative therapy for numerous blood born-malignant and nonmalignant disorders (1,2). Although allo-HSCT holds much promise, the prevalence of graft-versus-host disease (GvHD) limits its widespread use (3). CD4⁺/CD25⁺ naturally occurring T regulatory (T_reg) cells offer a potential solution to the prevention of GvHD. Importantly, T_regs can suppress allo-reactive T cell responses, including those involved in solid organ and skin allograft rejection (4). Multiple groups, including our own, have demonstrated that T_regs are capable of inhibiting GvHD without impacting the GvL response (5,6).

L-selectin (CD62L) is a member of the selectin family that is involved in leukocyte homing (7). CD62L is constitutively expressed by myeloid cells, nave lymphocytes and central memory T cells (7,8). We, and others, have shown the importance of T_reg homing molecule expression in preventing GvHD (5,9,10). These studies provided evidence that the phenotype of the CD62L^Hi T_reg population was responsible for inhibition of GvHD; however, they did not directly assess the role of CD62L in this process. Decreased expression of CD62L may be indicative of a T_reg subset that is biologically distinct from CD62L^hi T_regs independent of the function of CD62L.

Using a clinically relevant model of allo-HSCT we now show that CD62L expression by T_regs was not required for the inhibition of GvHD as CD62L^−/− T_regs provided similar protection from lethal acute GvHD compared to WT T_regs. In addition, CD62L expression was not critical for T_reg migration to GvHD target organs. However, the expression of CD62L was important for the prompt migration of T_regs to PLNs.

Methods

Mice

Donor mice consisted of male C57BL/6J (B6), (H-2^b; The Jackson Laboratory, Bar Harbor, ME), Thy1.1⁺ mice (H-2^b; The Jackson Laboratory) and CD62L^−/− mice, which have been described previously (11). CD62L^−/− mice were crossed with B6 mice expressing the enhanced GFP (eGFP) protein to generate eGFP-expressing CD62L^−/− mice. The generation of B6-eGFP mice has been described (12). In some experiments T_reg cells were isolated from FIR mice (expressing red fluorescent protein (RFP) under the FoxP3 promoter) as described (13). Recipient mice were male (C57BL/6JXDBA/2) FI mice, (B6D2) (H-2^bxd; The Jackson Laboratory).Within each experiment, all recipient and donor mice were male mice ranging from 9 to 14 weeks. All animal experiments were performed in accordance with protocols approved by the University of North Carolina Institutional Animal Care and Use Committee.

Antibodies and flow cytometry

Antibodies with the following specificities were purchased from eBiosciences (San Diego, CA): anti-CD4 (RM4.5), CD62L (Mel-14), CD25 (PC61), CD8 (53-6-7), Thy1.1 (HIS51) and FoxP3 (FJK-16s). Acquisition was performed on a FACSCalibur using CellQuest software (BD Biosciences; San Jose, CA). Analysis was performed using FlowJo (Treestar Inc., Ashland, OR) software.

Preparation of cells for transplant and bone marrow transplants

T cell-depleted bone marrow (TCD BM) cells, T_eff cells and T_regs were isolated and infused as described (14).

T_reg cell expansion

CD4⁺/CD25⁺/RFP⁺/CD62L^high cells were sorted on a MoFlo cell sorter (Dako A/S, Glostrup, Denmark) from the spleens of RFP-FoxP3 mice. Sorted cells were expanded with plate-bound anti-CD3 (145-2C11, 15µg/mL; eBioscience) and CD28 (37.51, 10µg/mL; eBioscience) supplemented with IL-2 (500 units/mL; Peprotech; Rocky Hill, NJ) for 12 days. After 12 days, cells were recovered, stained for CD62L and sorted. Sorted cells were always >95% CD4⁺/mRFP⁺/CD62L^Lo.

In vitro suppression assay

In vitro suppression assays were performed as previously described (10).

GvHD grading

Mice were observed twice weekly for signs of GvHD using the previously described clinical scoring system (15).

Fluorescence microscopy

Animals were anesthetized with avertin and organs were imaged with a Zeiss SteREO Lumar V12 microscope with eGFP bandpass filter (Carl Zeiss MicroImaging, Inc., Thornwood, NY) as described (14).

Competitive T_reg migration assay

Competitive migration of WT versus CD62L −/− T_reg cells was done as described (14).

Histopathology

The sections were scored by one of us (A.P.-M.) who was blinded to the treatment given using a previously described method (14).

Quantitation of chemokine receptor transcripts

RNA was isolated from sort-purified T_regs using the Qiagen RNeasy Kit (Qiagen; Valencia, CA). Quantitative RT-PCR for chemokine receptor transcripts was performed using primers and probes to CCR1, CCR2, CCR4, CCR5, CCR7, CCR8, CCR9, CCR10, CXCR3 and CXCR4 (Applied Biosystems; Corvalis, OR). The ΔCt method was used to normalize transcripts to 18S RNA and to calculate fold induction.

Measurement of serum IFN-γ

Serum samples were obtained from mice receiving whole naïve T cells, with or without WT T_regs, CD62L^−/− T_regs, CD62L^Lo T_regs or BM only. The samples were recovered when animals reached a clinical GvHD score of 4. IFN-γ concentrations were determined according to the manufacturer’s instructions using ELISA (Biolegend San Diego CA).

Statistical analysis

For GvHD scoring, we used Student’s t-test; for overall survival we used Fisher’s exact test, and for median survival we used the Mann–Whitney log rank test. p values ≤ 0.05 were considered significant.

Results

CD62L^−/− T_regs mediate protection against lethal GvHD

To determine the precise requirement for CD62L expression in T_reg-mediated protection during GvHD, we isolated fresh CD4⁺/CD25⁺ cells from WT or CD62L deficient animals (CD62L^−/−). Unexpectedly, we did not observe a significant difference in the overall survival (p = 1.0) or median survival time (p = 0.86) in recipient mice given WT compared to CD62L^−/− T_regs (Figure 1A). Both WT and CD62L^−/− T_regs recipients had significantly improved overall survival (p < 0.001) compared to recipients of WT T cells alone. Our previous work has demonstrated that in vitro expanded CD62L^Lo T_regs were unable to ameliorate GvHD pathology (5); however, our subsequent analysis of expanded CD4⁺/CD25⁺ cells has revealed considerable contamination by FoxP3⁻ cells in the CD62L^Lo fraction (M. Carlson, J. Serody; unpublished observation). We therefore isolated cells from FIR mice in which the red fluorescent protein is expressed under control of the FoxP3 promoter (13), and thus, T_regs can be identified from the CD4⁺/CD25⁺ fraction by their expression of mRFP. Recipients of ex vivo expanded mRFP⁺/CD62L^Lo T_regs displayed only a very modest improved overall (p = 0.09) and median survival time (p = 0.12) relative to animals receiving T cells alone (Figure 1A). These results demonstrated that CD62L^−/− T_regs were capable of providing protection from lethal acute and GvHD. These data also demonstrate that contamination of CD62^Lo T_reg cells with effector cells was not an explanation for the lack of activity of CD62L^Lo T_regs in the current study. The paucity of CD62L^Lo T_regs present in FIR mice precluded the evaluation of this population of cells without ex vivo expansion.

Figure 1. CD62L^−/− T_regs protect from lethal acute GvHD and are potent suppressers in vitro.

1.25 × 10⁶ WT (of which approximately 80% expressed CD62L and 67% had high levels of expression of CD62L), CD62L^Lo T_regs, or CD62L^−/− T_regs were transferred with 3 × 10⁶ TCD BM cells into lethally irradiated B6D2 recipients on day 0. 4 × 10⁶ whole splenic T cells from WT mice were then transferred on day +2 (n = 9 WT T_regs, n = 11 CD62L^Lo T_regs, n = 12 CD62L^−/− T_regs, n = 10 Effectors alone, n = 4 BM only). Animals were monitored for (A) survival and (B) signs of GvHD. Data represent mean score ± SEM at each time point. (C) Suppression of WT responder cell (CD4⁺/CD25⁻) proliferation in response to B6D2 alloantigen by WT (■) or CD62L^−/− T_regs () was determined as described in the Methods section.

Next, we determined disease severity using a defined clinical scoring system (15). Although the survival outcomes were not significantly different, WT T_regs did afford reduced clinical GvHD scores compared to CD62L^−/− T_regs during the first 21 days posttransplant (p < 0.04 for days 7–21) (Figure 1B). Starting on day 24, and for the duration of the experiment, GvHD scores were not significantly different (p > 0.05) in recipients given either WT or CD62L^−/− T_regs. Consistent with no improvement in overall or median survival, CD62L^Lo T_regs did not reduce clinical manifestations of GvHD as compared to T cells alone (Figure 1B). Collectively, these data demonstrated that CD62L^−/− T_regs were able to protect animals from lethal GvHD, albeit they did not suppress clinical GvHD manifestations as well as WT T_reg cells in the first 3 weeks posttransplant. In addition, CD62L^−/− T_regs functioned more efficiently to prevent GvHD than mRFP⁺/CD62L^Lo T_regs posttransplantation.

CD62L^−/− T_regs function normally to suppress T cell responses to allo-antigen in vitro

Because we observed significant differences early posttransplant in the clinical appearance of GvHD between recipients of WT and CD62L^−/− T_regs, we sought to determine the ability of CD62L^−/− T_regs to inhibit effector T cell responses to allo-antigen. To address this question, freshly isolated CD4⁺/CD25⁺ cells from WT and CD62L^−/− mice were co-cultured with WT CD4⁺/CD25⁻ responder cells stimulated with irradiated B6D2 splenocytes. CD62L^−/− and WT T_regs displayed equivalent suppression of WT effector T cells up to a 1:8 T_reg:Effector cell ratio (Figure 1C). Therefore, the early elevated GvHD scores of animals given CD62L^−/− T_regs was not due to an intrinsic defect in their suppressive function. As described, CD62L^Lo T_regs were potent suppressors of allo-reactive T cells in vitro up to a ratio of 1:32 T_regs: Effector cells (Figure S1).

GvHD target organ histopathology

Given the differences observed in clinical GvHD scores, we were interested in determining the impact that phenotypically different T_regs had on individual organ pathology. Histopathology scores in the colon were not statistically different between any of the groups (Figure 2A). Recipients of WT T_regs demonstrated less pathological damage in the lung as compared to recipients of CD62L^−/− T_regs (p = 0.05) (Figure 2B). Examination of the liver demonstrated that both WT and CD62L^−/− T_regs significantly inhibited GvHD pathology (p < 0.03) compared to recipients of effector T cells alone (Figure 2C). Interestingly, despite the modest difference in tissue pathology, there were significant differences in serum IFN-γ levels in mice given effector T cells alone compared to WT or CD62L^−/− T_regs (p < 0.01) (Figure 2D). Overall, these results demonstrated that with the exception of worsened lung pathology, CD62L^−/− T_regs functioned as well as WT T_regs to prevent GvHD, while both were potent in their ability to inhibit systemic IFN-γ production.

Figure 2. CD62L^−/− and WT T_regs suppress liver pathology.

1.25 × 10⁶ WT or CD62L^−/− T_regs were transferred with 3 × 10⁶ TCD BM cells into lethally irradiated B6D2 recipients on day 0. 4 × 10⁶ whole splenic T cells from WT mice were then transferred on day +2 (n = 6 WT T_regs (■), n = 7 CD62L^−/− T_regs, (), n = 8 Effectors alone (), n = 3 BM only (□). Animals were recovered when clinical scores reached a total of >4. Animals that did not reach a score of 4 were recovered on days 25–27 posttransplant. Histopathological assessment of the (A) colon, (B) lung, (C) liver. (D) Serum was recovered from animals at the time of histopathology assessment and analyzed by ELISA for levels of IFN-γ.

CD62L^−/− T_regs traffic to secondary lymphoid tissues and GvHD target organs

Because we observed differences early on in the clinical manifestation of GvHD between recipients of WT and CD62L^−/− T_regs, we were interested in determining the trafficking pattern of these T_regs. To evaluate in vivo T_reg trafficking, we used a competitive lymphocyte migration assay (16). As illustrated in Figure 3A, 6 days after transfer, CD62L^−/− T_regs were found at a similar frequency as WT T_regs in the liver, lung, spleen, bone marrow and mesenteric lymph node (MLN), although as expected, there were substantially fewer CD62L^−/− T_regs in the PLNs of recipient animals. Further analysis 16 days post-T_reg transfer showed no difference between WT and CD62L^−/− T_reg migration to liver, spleen, bone marrow or MLN (Figure 3B). However, although not statistically different, there were fewer CD62L^−/− T_regs detected in the lung and PLN on day 16 compared to WT T_regs, which correlated with the enhanced GvHD in the lung of recipient animals receiving CD62L^−/− T_regs (Figure 3B).

Figure 3. CD62L^−/− and WT T_reg trafficking.

5 × 10⁵ WT-GFP (■) or CD62L^−/−-GFP () T_regs were transferred with 5 × 10⁵ Thy1.1⁺ T_regs along with 3 × 10⁶ TCD BM cells into lethally irradiated B6D2 recipients on day 0. 2 × 10⁶ whole splenic T cells from WT mice were then transferred on day +2. On days 6 (A) and 16 (B) post-T_reg transfer, lymphocytes were isolated from GvHD target organs and secondary lymphoid tissues as described in the Methods section. The ratio of eGFP⁺/FoxP3⁺:Thy1.1⁺/FoxP3⁺ cells are shown (n = 4 for each group). N = 4 animals/time point (MLNs were pooled for each group, and PLNs were pooled for each group). 1.0 × 10⁶ WT-GFP (top) or CD62L^−/−-GFP (bottom) T_regs were transferred along with 3 × 10⁶ TCD BM cells into lethally irradiated B6D2 recipients on day 0. 2 × 10⁶ whole splenic T cells from WT mice were then transferred on day +2. 16 days post-T_reg transfer animals were anesthetized with avertin and organs were imaged with a Zeiss SteREO Lumar.V12 microscope with eGFP bandpass filter. Brightfield images (left), and GFP images (middle) were taken for each organ. GFP intensities (right) were determined by software analysis. (C) MLN (D) spleen (E) lung (F) peripheral lymph node. Original magnification: lung = 25X, spleen = 40X, MLN = 45X, ILN = 45X. Data represent mean score ± SEM for each organ.

To confirm our findings regarding the function of CD62L in the migration of T_regs in a lymphopenic environment, we performed in vivo imaging using fluorescence stereomicroscopy. In the MLN(Figure 3C) and spleen (Figure 3D)we found similar distribution and GFP signal intensity by WT and CD62L^−/− T_regs, indicating that the migration and accumulation of CD62L^−/− T_regs was indistinguishable from WT T_regs in these organs 16 days posttransplantation. Interestingly, we observed fewer GFP⁺ CD62L^−/− T_regs in the lung (Figure 3E) and PLN (Figure 3F) at this time point. Taken together, these observations illustrated that CD62L^−/− T_regs home to GvHD target organs, similar to WT T_regs, with the exception of a modest impairment in migration to the lung. Differences in the migration of WT compared to CD62L^−/− T_regs to PLNs were found in the first week posttransplantation demonstrating the importance of CD62L in the initial migration of T_regs to PLNs. However, at day 16 these differences were minimized indicating that CD62L was not absolutely required for the eventual migration of T_regs to PLNs.

CD62L^Hi, CD62L^−/− and CD62L^Lo T_regs display differential expression of chemokine receptors

The finding of CD62L^−/− T_regs in the PLNs of irradiated recipients was somewhat surprising given the role that CD62L plays in T cell rolling and homing to lymph nodes. This observation suggests that in the absence of CD62L other proteins important for T cell migration may serve a similar function. To this end we examined the phenotypic profile of T_regs based on CD62L expression (Figures 4A–C). As shown in Figure 4D, CD62L^Hi and CD62L^−/− T_regs have increased expression of CCR7 mRNA relative to CD62L^Lo T_regs. We then compared the three T_reg types to naïve CD4⁺/mRFP⁻ cells in their mRNA expression of other chemokine receptors. CD62L^−/− T_regs resembled CD62L^Lo T_regs in the expression of CCR2, CCR4 and CXCR3, and resembled CD62L^Hi T_regs in the expression of CCR9. CD62L^−/− T_regs had intermediate levels of CCR5 and CCR8, with distinctive expression of CCR1 and CCR10. Collectively, these data demonstrated that the migratory profile of CD62L^−/− T_regs was that of an intermediate activated phenotype with higher expression of pro-inflammatory chemokine receptors compared to CD62L^Hi T_regs and much greater expression of CCR7 compared to CD62L^Lo T_regs.

Figure 4. T_reg chemokine receptor expression based on CD62L expression.

Sort purification of cells for quantitative real-time PCR analysis. (A) CD4⁺/CD25⁺/mRFP⁺/CD62L^Hi T_regs (bottom) and CD4⁺/CD25⁻/mRFP⁻/CD62L^Hi naïve T cells (top) were sort purified from mRFP-FoxP3 mice. (B) CD4⁺/CD25⁺ T_regs were sort purified from CD62L^−/− mice. (C) CD4⁺/mRFP⁺/CD62L^Lo T_regs were sort purified following in vitro expansion. RNA was extracted and real-time PCR performed as described in the Methods section. (D) CCR7 expression on mRFP⁺/CD62L^Hi T_regs (), CD4⁺/CD25⁺/CD62L^−/− T_regs (■), naïve CD4⁺/CD25⁻ T cells () and mRFP⁺/CD62L^Lo Tregs (). Data are shown as relative change in expression (logarithmic scale) compared to CD4⁺/mRFP⁺/CD62L^Lo T_regs. (E) Chemokine receptor expression on mRFP⁺/CD62L^Hi T_regs (), CD4⁺/CD25⁺/CD62L^−/− T_regs (■), mRFP⁺/CD62L^Lo T_regs () and naïve CD4⁺/CD25⁻ T cells (). Data are shown as relative change in expression (logarithmic scale) compared to naïve CD4⁺/CD25⁻ T cells. Data are representative of three independent experiments.

Discussion

In the current work, we were interested in determining whether CD62L itself was critical for T_reg function and migration into lymphoid tissue. We demonstrated that CD62L was not critical for T_reg function to prevent GvHD lethality as CD62L^−/− T_regs afforded substantial protection from lethal acute GvHD in the clinically relevant model employed. WT T_regs yielded reduced clinical scores compared to CD62L^−/− T_regs during the first 3 weeks posttransplant, which correlated with delayed migration of CD62L^−/− T_regs to PLNs. Histopathological analysis of GvHD target organs correlated with the clinical scores, as recipients of WT T_regs showed improved pathology in the lung and similar pathology in the colon and liver compared to CD62L^−/− T_regs. Lastly, we demonstrated differential chemokine receptor expression of T_regs based on CD62L expression, where the CD62L^−/− T_regs displayed a phenotype that appeared to be an intermediate between the naïve CD62L^Hi and activated CD62L^Lo.

Previous reports examining the role of CD62L in T_reg-mediated inhibition of GvHD suggested either that (1) CD62L itself was critically important in the function of T_regs or that (2) the CD62L^Hi phenotype functioned differently than CD62L^Lo T_regs but that CD62L itself was not critical (5,9). Our data demonstrated that CD62L itself was not critically required for the prevention of GvHD lethality or for the ability to migrate into LN post transplantation. While there was no difference in either overall or median survival time, our data indicated that CD62L did serve as an accessory molecule given the statistical difference in clinical scores between WT and CD62L^−/− T_regs during the first 3 weeks posttransplantation. It is of interest that we also observed no statistical difference in clinical scores between CD62L^−/− and CD62L^Lo T_regs for the first 2 weeks posttransplantation suggesting CD62L serves early on to promote T_reg inhibition of GvHD most likely by enhancing the migration of T_regs into lymphoid tissue.

One concern in our previous studies in which we expanded CD4⁺/CD25⁺ T cells to obtain a CD62L^Lo population was the difficulty in eliminating CD25⁺ effector cells from the T_reg infusion (5). Here, we have circumvented this concern by using T_regs from FIR mice in which mRFP is under control of the FoxP3 promoter and thus cells expressing FoxP3 can be detected using flow cytometry (13). Our data confirm previous observations that CD62L^Lo T_regs were not sufficient to prevent GvHD in the overwhelming majority of transplanted recipients. The possibility of impaired suppressive function of these cells was ruled out by in vitro analysis in which CD62L^Lo T_regs were more proficient suppressors of T cell responses to allo-antigen, consistent with the previously published data (17). Therefore, the inability of CD62L^Lo T_regs to provide protection against GvHD could not be explained by impaired function but may be due to impaired homing to lymphoid tissue or diminished survival after infusion.

Examination of the pathology in individual organs revealed that WT and CD62L^−/− T_regs ameliorated disease in the liver, whereas WT T_reg recipients displayed reduced pathology in the lung as compared to CD62L^−/− T_reg recipients. The increased lung pathology correlated with modestly impaired CD62L^−/− T_reg migration to the lung. The accumulation of IFN-γ in the serum has been shown to be a predictor of GvHD mortality (18). We also documented a substantial reduction in the level of IFN-γ in the serum of animals receiving either WT or CD62L^−/− T_regs an effect not seen in recipients of mRFP⁺/CD62L^Lo T_regs (data not shown). Here again, a functional distinction was made between CD62L^−/− and CD62L^Lo T_regs.

While it is clear that T_regs do inhibit effector T cell expansion and suppress effector functions, it is less clear as to whether the inhibition is in lymphoid tissues or GvHD target organs. In the current report, we demonstrated that CD62L^−/− T_regs migrate to GvHD target organs with similar efficiency as WT T_regs; however, their accumulation within the PLNs was delayed. It is interesting to note that this delay corresponded with increased clinical GvHD scores, thus supporting the hypothesis that entry into lymph nodes by T_regs was important in inhibiting the initial expansion of donor T cells. The inability of CD62L^Lo T_regs to inhibit GvHD has been attributed to ineffective trafficking to secondary lymphoid tissues (9). Normal trafficking seen in CD62L^−/− T_regs provides another distinction between CD62L^Lo and CD62L^−/− phenotypes.

Other studies have examined chemokine receptor expression on T_reg subsets, including the CD62L^Hi and CD62L^Lo populations (19,20). Our data are in agreement that the CD62L^Hi fraction expressed high levels of the lymph node homing chemokine receptor CCR7. Of interest, the CD62L^−/− population also expressed high levels of CCR7, providing a plausible mechanism for their migration to secondary lymphoid tissues. In keeping with an activated status the CD62L^Lo T_regs expressed high levels of CCR5 and CCR8 while the CD62L^−/− T_regs displayed intermediate expression.

In summary, our data demonstrate that posttransplant, CD62L was dispensable for T_reg inhibition of GvHD lethality.

Abbreviations

Allo-HSCT

allogeneic hematopoietic stem cell transplantation

APC

antigen presenting cell

eGFP

enhanced green fluorescent protein

GvHD

graft-versus-host disease

GvL

graft-versus-leukemia

PLN

peripheral lymph nodes

MLN

mesenteric lymph node

TCD BM

T cell depleted bone marrow

T_regs

regulatory T cells

WT

wild type.