Conceived and designed the experiments: SXL LWK A-MC-A RA AMH CT VMH JAR CL GM OA MRMvdB. Performed the experiments: SXL LWK A-MC-A RA AMH CT VMH JAR CL GM MS DS CK UKR NY JLB RRJ OP I-KN. Analyzed the data: SXL JAR CL GM OA MRMvdB. Contributed reagents/materials/analysis tools: RSB FM KVH NB. Wrote the paper: SXL OA MRMvdB.
¶ These authors also contributed equally to this work.
The authors have declared that no competing interests exist.
Allogeneic bone marrow transplantation (allo-BMT) is a potentially curative therapy for a variety of hematologic diseases, but benefits, including graft-versus-tumor (GVT) activity are limited by graft-versus-host-disease (GVHD). Carcinoembryonic antigen related cell adhesion molecule 1 (Ceacam1) is a transmembrane glycoprotein found on epithelium, T cells, and many tumors. It regulates a variety of physiologic and pathological processes such as tumor biology, leukocyte activation, and energy homeostasis. Previous studies suggest that Ceacam1 negatively regulates inflammation in inflammatory bowel disease models.
We studied Ceacam1 as a regulator of GVHD and GVT after allogeneic bone marrow transplantation (allo-BMT) in mouse models.
We conclude that Ceacam1 regulates T cell activation, GVHD target organ damage, and numbers of donor T cells in lymphoid organs and GVHD target tissues. In recipients of allo-BMT, Ceacam1 may also regulate tissue radiosensitivity. Because of its expression on both the donor graft and host tissues, this suggests that targeting Ceacam1 may represent a potent strategy for the regulation of GVHD and GVT after allogeneic transplantation.
Ceacam1 is a member of a large family of carcinoembryonic antigen proteins
Some Ceacam1 isoforms contain intracellular ITIM motifs, and activation of Ceacam1 results in the recruitment of the SHP-1 and SHP-2 phosphatases
In addition to immune regulation, Ceacam1 exerts a wide variety of other biological functions. It is a cell-cell adhesion molecule
Allo-BMT is an established therapy with curative intent for a variety of hematologic malignancies and non-malignant conditions
Donor-recipient antigenic disparity, donor T cells, and tissue injury resulting in inflammation due to the conditioning regimen all contribute to GVHD, which primarily affects intestines, liver, skin and thymus
Ceacam1 is expressed both on leukocytes (especially T cells), as well as on epithelial and endothelial cells, which are prominent components of the parenchyma of the above-mentioned GVHD target organs. In addition, Ceacam1 is upregulated on many tumors. In this report, we assess the impact of Ceacam1 on alloreactive T cells in the donor allograft, as well as the effects of Ceacam1 deficiency on recipients of allo-BMT with respect to GVHD and GVT activity.
We assessed Ceacam1 regulation of GHVD on donor T cells or recipients in two well-described major histocompatibility complex (MHC) class I/II-disparate models C57BL/6 (B6, H-2b)→BALB/c (H-2d) and BALB/c→B10.BR (H-2k). We used Ceacam1−/− B6 mice
We first transplanted irradiated BALB/c mice with B6 TCD-BM with WT or Ceacam1−/− T cells, and observed that recipients of Ceacam1−/− T cells had significantly increased mortality compared to recipients of WT T cells (
A) LEFT: 5×106 TCD-BM ± 1×106 T→BALB/c. N = 10/group. Representative data from one of four experiments. RIGHT: 5×106 TCD-BM ± 2×106 T→B10.BR. N = 20/group combined from two experiments. Square: BM only. Triangle: WT T. Circle: Ceacam1−/− T. B) LEFT: 5×106 TCD-BM ± 0.5×106 T→BALB/c. N = 10/group. Representative data from one of three experiments. RIGHT: 5×106 TCD-BM ± 1×106 T→BALB/c. N = 10/group. Representative data from one of two experiments. Square: BM only. Triangle: WT T. Circle: Ceacam1-Tg T. C) 5×106 TCD-BM ± 1×106 T→BALB/c. Combined N = 55 in groups receiving T cells from six experiments. Diamond: Ceacam1−/− recipients, BM only. Square: WT recipients, BM only. Triangle: WT recipients, BM+T. Circle: Ceacam1−/− recipients, BM+T.
We next asked whether T cells overexpressing Ceacam1 would cause less disease, and transplanted BALB/c recipients with 0.5×106 or 1×106 donor WT or Ceacam 1-Tg T cells. At both doses, recipients of Ceacam1-Tg T cells showed attenuated mortality (
Finally, we assessed the role of Ceacam1 on tissues of allo-BMT recipients, and transferred TCD-BM+T cells into WT vs. Ceacam1−/− BALB/c recipients. This revealed that Ceacam1−/− recipients had increased early (but not overall) mortality, with nearly 50% of mice succumbing within the first week (
We next asked whether Ceacam1 regulated GVHD target organ damage, and again assessed effects of Ceacam1 deficiency or overexpression on donor T cells, and Ceacam1−/− allo-BMT recipients.
We observed that recipients of Ceacam1−/− T cells had more severe large intestinal GVHD (
A) 5×106 TCD-BM ± 1×106 T→BALB/c. Day 21. Histopathology score. N = 5–12/group. SB, small bowel. LB, large bowel. B) Thymocyte count and apoptotic skin cells from (A). N≥7/group. Day 21. One of two representative experiments. C) Apoptotic skin cells from (A). N≥7/group. Day 21. D) 5×106 TCD-BM ± 1×106 T→BALB/c. Histopathology score. Day 21. N≥7/group. SB, small bowel. LB, large bowel. E) Thymocyte counts from (E): N≥7/group. One of two independent experiments. DP: CD4+CD8+ double-positive thymocytes. Day 21. F) Apoptotic skin cells from (E). N≥7/group. Day 21. G) 5×106 TCD-BM ± 1×106 T→WT or Ceacam1−/− BALB/c. Histopathology score. Day 14. N≥12/group. SB, small bowel. LB, large bowel. H) Thymocyte count from (I). N≥8/group. One of two independent experiments. DP: CD4+CD8+ double-positive thymocytes. Day 14.
In experiments comparing recipients of WT and Ceacam1-Tg T cells on day 21 post-transplant, we found that recipients of Ceacam1-Tg T cells demonstrated significantly less GVHD of the liver, intestines, and thymus compared to recipients of WT T cells, but similar skin GVHD (
Finally, we assessed Ceacam1−/− allo-BMT recipients on day 14 post-transplant. In correspondence with increased early GVHD mortality, Ceacam1−/− allo-BMT recipients showed increased large bowel damage and thymic GVHD (
We next assessed the numbers of donor CD4 and CD8 effector T cells after transfer of Ceacam1−/− or Ceacam1-Tg T cell-containing allografts, or in Ceacam1−/− allo-BMT recipients.
Comparing recipients of WT T cells with those receiving Ceacam1−/− T cells, we observed increased numbers of Ceacam1−/− donor alloactivated effector T cells in the spleen, MLN, and IEL of allo-BMT recipients (
A) 5×106 TCD-BM ± 1×106 T→BALB/c. Day 14. Donor CD44+CD62L− effector cells were enumerated. Donor T cells are shown. N = 5–18/group. Combined data from two to three experiments. B) 5×106 TCD-BM ± 1×106 T→BALB/c. Day 14. Donor CD44+CD62L− effector cells were enumerated. N = 7 group. Combined data from two experiments. C) 5×106 TCD-BM ± 1×106 T→BALB/c. Day 14. Donor CD44+CD62L− effector cells were enumerated. N = 7–18 group. Combined data from two to four experiments.
When we analyzed organs of recipients of allografts containing WT vs. Ceacam1-Tg T cells, we noted decreased numbers of donor effector T cells in the MLN, PLN, and liver (
The accelerated early mortality of Ceacam1−/− allo-BMT recipients, together with increased accumulation of donor T cells in GI tract and mesenteric lymph nodes, but decreased numbers peripheral lymph nodes (
A) WT and Ceacam1−/− BALB/c mice were irradiated as a single-dose. N = 5/group. One of two independent experiments. B) Experiment as in A. Quantification of surviving and regenerating crypts from the terminal ileum. N = 3/group, 9 sections per terminal ileum taken. C) WT and Ceacam1−/− B6 mice were irradiated as a single-dose. N = 5/group. One of two independent experiments. D) Experiment as in C. Quantification of surviving and regenerating crypts from the terminal ileum. N = 3/group, 9 sections per terminal ileum taken.
Next, we studied a variety of possible mechanisms by which Ceacam1 may regulate donor T cell function. We analyzed donor WT and Ceacam1−/− alloactivated splenic T cells on day 14 after allo-BMT for trafficking molecules, and found that Ceacam1−/− CD8+ CD44+CD62L− effector T cells expressed higher levels of integrin β7 subunit and the gut homing integrin α4β7 (
A) 5×106 TCD-BM ± 1×106 WT or Ceacam1−/− T→BALB/c. Day 14. N = 6 to 10/group. One of two independent experiments. B) 5×106 TCD-BM ± 1×106 WT or Ceacam1-Tg T→BALB/c. Day 14. N = 6 to 10/group. One of two independent experiments. C) 5×106 TCD-BM ± 1×106 T→WT or Ceacam1−/− BALB/c. Day 14. N = 5 to 8/group. One of two independent experiments.
When we assessed the expression of trafficking molecules in recipients of WT vs. Ceacam1-Tg T cell allografts, we found no significant differences in levels of β7 subunit, integrin α4β7 (
Ceacam1 can be found on activated T cells
A) 107 CFSE-labeled B6 or B6 CD45.1 splenic T cells→BALB/c (8.5 Gy). Spleens analyzed day 2. Donor CD4 and CD8 T cells were analyzed by number of divisions viaCFSE dilution. Ceacam1 median fluorescence intensity (MFI) is shown. N = 6, combined from three experiments. * p<0.05 for MFI difference vs. non-dividing cells. B) 107 WT or Ceacam1−/− CFSE-labeled T→BALB/c. Day 3, splenic T cells. Percentage of donor CD25 positive CD4 CFSElo cells are shown. N = 5–10/group. Combined data from three experiments. C) Percentage of CD62L positive CD4 CFSElo cells from experiment as in (A). Combined from four experiments, total N = 10/group. D) Percentage of donor CFSElo, fast-proliferating allo-activated T cells from experiment as in (A). N = 5/group, one of three experiments shown. E) 107 CFSE+ B6 or CEACAM1-Tg T→BALB/c. Day 3 spleen. CFSE dilution on donor T cells are shown. Shaded: Ceacam1-Tg; open: WT. N = 2/group. Representative data from one of two experiments. F) 107 CFSE+ B6 T→WT or Ceacam1−/− BALB/c. Percentage of donor CD25 positive CD4 CFSElo cells are shown. N = 10/group. Combined data from four experiments.
Because the expression of Ceacam1 on alloreactive T cells after adoptive transfer occurred
We observed that relative to isotype control staining, an increased percentage of alloactivated CFSElo CD4 Ceacam1−/− T cells were positive for the alloactivation marker CD25, and that a greater percentage of these cells downregulated CD62L than WT T cells (
We repeated these experiments with alloreactive Ceacam1-Tg T cells and as expected, observed a decrease in numbers of CFSElo T cells as assessed by CFSE dilution (
Lastly, we assessed the role of Ceacam1 expression on radio-resistant cells in allo-BMT recipients for donor T cell alloactivation. We transferred CFSE-labeled B6 T cells into irradiated WT vs. Ceacam1−/− BALB/c mice, and analyzed donor T cells in spleens on day 3. Here, we did not observe differences in proliferation (data not shown), but donor CD4 T cells in Ceacam1−/− allogeneic recipients did exhibit an increase in alloactivation as measured by CD25 (
We measured serum cytokines in recipients of WT, Ceacam1-Tg and Ceacam1−/− T cells on days 7 and 14 post-transplant, and observed that levels of IFNγ, TNF, IL-2, IL-4, IL-6, IL-10, and IL-12p70 were similar (data not shown). Percentages of FoxP3+ donor regulatory T cells and expression of T-bet were also similar between recipients of WT, Ceacam1-Tg and Ceacam1−/− T cells (data not shown and
|
|
|
|
Survival | ↓ | ↑ | ↓ |
Complete blood count | = | = | = |
Engraftment/chimerism | = | = | = |
Serum TNF, IFNγ, IL-2, IL-12 | = | = | = |
Donor T cells | |||
Activation/Proliferation | ↑ | ↓ | ↑ |
T-bet, FoxP3, IFNγ | = | = | = |
Trafficking molecules | ↑ | = | ↑ |
Cytotoxicity | = | n/a | n/a |
Numbers in | |||
Spleen | ↑ | = | = |
MLN/intestine | ↑ | ↓ | ↑ |
PLN | ↓ | ↓ | ↓ |
Splenic donor DCs | |||
percentage, apoptosis | = | = | = |
Maturation (CD80/86) | = | = | = |
Target organ damage | |||
Liver | = | ↓ | = |
Small intestine | = | ↓ | = |
Large intestine | ↑ | ↓ | ↑ |
Thymus | ↓ | ↓ | ↑ |
As Ceacam1 can regulate the cytolytic responses of lymphocytes
Finally, we assessed the GVT activity of Ceacam1−/− donor alloreactive T cells against A20 lymphoma and RENCA renal cell carcinoma. Recipients of Ceacam1−/− donor T cells had improved survival in the A20 lymphoma model (
A) 5×106 B6 BM+0.5×106 A20±0.5×106 B6 WT or Ceacam1−/− T cells→BALB/c (8.5 Gy). N = 20/group for groups receiving T cells, combined from two experiments. B) 5×106 B6 BM±0.1×106 RENCA±0.8×106 B6 WT or Ceacam1−/− T cells→BALB/c (8.5 Gy). N = 9 for groups receiving T cells. C) A20 lymphoma cells and RENCA tumor cells in culture were stained for Ceacam1 expression.
In this report, we show that Ceacam1, which is found on both donor alloreactive T cells as well as non-hematopoietic tissues such as gastrointestinal and hepatic epithelium, can regulate both donor T cell function and the sensitivity of allo-BMT recipients to radiation-containing preparative regimens. In addition, Ceacam1 on donor T cells and tumors may modulate GVT activity. Ceacam1 on both the donor allograft and recipient tissues thus appears to represent an important regulator of GVHD and GVT morbidity and mortality via both T cell dependent and independent mechanisms, suggesting that therapeutic approaches which modulate Ceacam1 may need to assess and balance GVHD vs. GVT.
Ceacam1 on T cells has previously been shown to restrain CD4 T cell polarization, cytokine secretion and cytotoxicity. In our GVHD model systems however, we found similar T cell polarization and cytokine secretion when we analyzed donor alloreactive T cells
We also assessed the role of Ceacam1 in allo-BMT recipients. In our model systems, WT T cells in a Ceacam1-deficient environment showed a phenotype similar to that of Ceacam1−/−alloactivated T cells: both showed increased activation, selective damage to the large intestines, and preferential accumulation in the MLN and intestinal parenchyma of mice with GVHD, and correspondingly decreased infiltration of the liver and PLN, ultimately leading to exacerbation of disease, with accelerated mortality in the first two weeks post-transplant. This suggests that Ceacam1 on donor T cells interacts with recipient tissues, and that Ceacam1 “fraternal” interactions between cells of the donor graft, were not sufficient to restrain GVHD in Ceacam1−/− recipients. However, the increased early mortality of Ceacam1−/− allo-BMT recipients with GVHD also led us to ask whether Ceacam1−/− mice were sensitive to radiation injury. While Ceacam1−/− mice were not significantly defective for hematopoiesis after sublethal irradiation at 3.5 and 4.5 Gy (data not shown), they did exhibit significantly increased damage to the small intestines after lethal irradiation (
Ceacam1 also directly regulates intestinal epithelia. Due to enhanced Wnt/β-catenin signaling, Ceacam1−/− jejunal and ileal enterocytes exhibit higher levels of the positive cell cycle regulators c-Myc and cyclin D1
It is difficult to directly assess the relative importance of gastrointestinal radiation sensitivity versus increased GVHD in Ceacam1−/− allo-BMT recipients, as radiation-induced gut damage may both be directly manifested in intestinal pathology, yet transmural migration of bacterial superantigens is an important first step for the initiation of GVHD
In experiments with Ceacam1−/− donor T cells, we also observed a trend for splenic donor CD8 alloactivated T cells to express higher levels of α4β7. Although integrin α4β7 is important for GVHD pathogenesis, and we have previously shown that β7−/− T cells cause a sustained decrease in acute systemic and intestinal GVHD
Moreover, recipients of Ceacam1-Tg T cells also had reduced intestinal infiltrates despite similar integrin α4β7 expression, suggesting that Ceacam1 regulates the accumulation of donor T cells in target tissues via multiple mechanisms. Thus, our results on donor lymphocyte infiltrates into GVHD target tissues and secondary lymphoid tissues must be interpreted cautiously, as they must be influenced by T cell proliferation, retention and apoptosis, in addition to trafficking.
Although Ceacam1−/− and Ceacam1-Tg T cells displayed overall symmetric and opposite phenotypes, we also noted differences. Ceacam1-Tg T cells primarily showed decreased proliferation, whereas Ceacam1−/− T cells showed changes in proliferation, but also trafficking and activation. Some of these differences may be due to our models: on WT T cells, Ceacam1 is only briefly and transiently upregulated during activation. Consequently, Ceacam1−/− T cells are “missing” Ceacam1 only transiently, while Ceacam1-Tg T cells constitutively over-express Ceacam1. Furthermore, while Ceacam1−/− T cells are effectively insensitive to all Ceacam1 ligands and interactions, Ceacam1-Tg T cells which over-express the protein may have increased fraternal Ceacam1 interactions with other donor T cells, but may not necessarily experience increased Ceacam1 interactions with donor BM or host hematopoietic and non-hematopoietic components. These differences may explain why their activation and trafficking phenotypes are not directly opposed.
We were interested to note that in our GVT experiments, recipients of Ceacam1−/− T cells had significantly improved survival when challenged with A20 lymphoma but not renal cell carcinoma. Although both A20 lymphoma and renal cell carcinoma express Ceacam1, A20 cells uniformly expressed Ceacam1 at high levels, while only a subset of RENCA cells showed (somewhat lower) expression. Indeed, a number of hematologic tumors, including EL4 leukemia, P815 mastocytoma, and C1498 myeloid leukemia all express substantial levels of Ceacam1 (data not shown), whereas some solid tumors, such as mouse 4T1 breast epithelial cancer and CT51 colon tumor normally express only lower or even minimal levels of Ceacam1, similar to the lower level of expression we found with RENCA (not shown).
Therefore, one possibility is that the GVT activity of T cells can be negatively regulated by tumors expressing high levels of Ceacam1, but is less important for tumors that express low levels or only on a subset of cells in the first place. However, RENCA in our GVT model systems is found primarily in the liver, and to a lesser extent, the lungs. Since donor allografts with Ceacam1−/− T cells showed decreased numbers of donor alloreactive T cells in the liver as compared with wildype in GVHD experiments (
In conclusion, our results show that Ceacam1 on both donor T cells and allo-BMT recipients controls the proliferation, activation, and trafficking of donor alloreactive T cells, and the sensitivity of gastrointestinal tissues to irradiation. Consequently, Ceacam1 may represent a viable target for reducing radiation-associated gastrointestinal toxicity, for the control of GVHD and GVT activity after allo-BMT.
All animal protocols were approved by the Memorial Sloan-Kettering Cancer Center (MSKCC) Institutional Animal Care and Use Committee (protocol #99-07-025).
C57BL/6 (B6, H-2b), BALB/c (H-2d), and B10.BR (H-2k) mice were obtained from The Jackson Laboratory (Bar Harbor, ME). B6 and BALB/c Ceacam1−/− mice, and B6 Ceacam1-Tg mice were generated at McGill University (B6 and BALB/c from Harlan (Montreal, Quebec, Canada)), and maintained at Memorial Sloan-Kettering Cancer Center. Mice used were between 8 and 12 weeks old.
BM cells removed from femurs and tibias were T cell-depleted (TCD) with anti-Thy-1.2 and low-TOX-M rabbit complement (Cedarlane Laboratories, Hornby, ON, Canada). Enriched splenic T cells were obtained by nylon wool column passage. Cells were resuspended in DMEM and injected into lethally irradiated recipients on day 0 after total body irradiation (137Cs source) as a split dose 3 hours apart.
Purified splenic T cells were incubated with CFSE (Invitrogen, Carlsbad, CA) at a concentration of 2.5–5 µM in PBS (5×107 cells/mL) at 37°C for 20 minutes, washed twice with PBS, resuspended in DMEM and infused intravenously into lethally irradiated allogeneic recipients. Splenocytes from recipients were harvested at varying time points and analyzed by FACS as described.
Survival was monitored daily, and mice were scored weekly for 5 clinical parameters (weight, posture, activity level, fur ruffling, and skin lesions) on a scale from 0 to 2. A clinical GVHD score was generated by summation of the 5 criteria scores; mice scoring 5 or greater were considered moribund and euthanized.
Small and large bowel, liver, and skin were assessed by experts in a blinded fashion. Organs were preserved in formalin, transferred to 70% ethanol, and then embedded in paraffin, sectioned, stained with hematoxylin and eosin, and scored with a semi-quantitative scoring system. Bowel and liver were scored for 19 to 22 different parameters associated with GVHD (detailed in
Histopathologic scores, median fluorescence intensities and cell counts were compared between groups using the nonparametric unpaired Mann-Whitney U test; the Mantel-Cox log-rank test was used for survival data.
Additional methods are described in
(TIFF)
(DOC)
(DOC)
The authors thank the staff of the Memorial Sloan-Kettering Cancer Center Research Animal Resources Center for excellent animal care.