Identification of stem cell transcriptional programs normally expressed in embryonic and neural stem cells in alloreactive CD8+ T cells mediating graft-versus-host disease

Abstract

A hallmark of graft-versus-host-disease (GVHD), a life-threatening complication after allogeneic hematopoietic stem cell transplantation, is the cytopathic injury of host tissues mediated by persistent alloreactive effector T cells (T(E)). However, the mechanisms that regulate the persistence of alloreactive T(E) during GVHD remain largely unknown. Using mouse GVHD models, we demonstrate that alloreactive CD8(+) T(E) rapidly diminished in vivo when adoptively transferred into irradiated secondary congenic recipient mice. In contrast, although alloreactive CD8(+) T(E) underwent massive apoptosis upon chronic exposure to alloantigens, they proliferated in vivo in secondary allogeneic recipients, persisted, and caused severe GVHD. Thus, the continuous proliferation of alloreactive CD8(+) T(E), which is mediated by alloantigenic stimuli rather than homeostatic factors, is critical to maintaining their persistence. Gene expression profile analysis revealed that although alloreactive CD8(+) T(E) increased the expression of genes associated with cell death, they activated a group of stem cell genes normally expressed in embryonic and neural stem cells. Most of these stem cell genes are associated with cell cycle regulation, DNA replication, chromatin modification, and transcription. One of these genes, Ezh2, which encodes a chromatin modifying enzyme, was abundantly expressed in CD8(+) T(E). Silencing Ezh2 significantly reduced the proliferation of alloantigen-activated CD8(+) T cells. Thus, these findings identify that a group of stem cell genes could play important roles in sustaining terminally differentiated alloreactive CD8(+) T(E) and may be therapeutic targets for controlling GVHD.

Fig.1

Alloreactive CD8⁺ T_E are highly replicating cells causing GVHD. (A) CFSE-labeled donor CD8⁺ T_N (2×10⁶) derived from normal C3H.SW mice were transplanted with B6/SJL TCD BM (0.5×10⁶) into lethally irradiated B6/SJL mice. Donor CD8⁺ T cells were recovered at day 14 after transplantation from the spleen and lymph node of 3 recipients, numerated and stained for flow cytometric analysis. Histograms show the cell division based on CFSE dilution and production of IFN-γ by CD8⁺ T_MSC, T_CM and T_E. Data are representative of three independent experiments. (B) Real-time RT-PCR analysis shows the relative mRNA expression of selected genes in donor day-14 CD8⁺ T_E, T_CM and T_MSC. Data are representative of three independent experiments. (C and D) Day-14 CD8⁺ T_E were re-labeled with CFSE and adoptively transferred into lethally irradiated secondary B6/SJL mice. CFSE-labeled C3H.SW CD8⁺ T_N were transferred as controls. Donor T cells were recovered at day 7 after adoptive transfer, numerated and analyzed for flow cytometry. Histogram shows the cell division of indicated cells. Non-stimulated CFSE-labeled CD8⁺ T_N were used as non-dividing cell control. Contour plots show the expression of Ki67 (C). For BrdU labeling, secondary recipient mice were given drinking water containing BrdU for 3 days prior to the end of the experiments. At day 7, donor cells were recovered, stained with BrdU, and analyzed by flow cytometry. Dot plots show the labeling of BrdU in gated CD8⁺ T_E population (Mean ± SD) (D). (E and F) As described above in (C), the number of donor T cells recovered from these secondary recipients (n=3 for each group) was calculated (E) and the percentage of apoptotic cell was shown (F). Data are representative of two independent experiments. (G) Donor C3H.SW TCD BM (5×10⁶) were transplanted alone, or with donor day-14 CD8⁺ T_E (0.5×10⁶) and donor CD8⁺ T_N (1.0×10⁶) into lethally irradiated secondary B6/SJL mice. Survival of animals was analyzed with the Kaplan-Meier method. *, P<0.05, significant difference.

Fig.2

Proliferation and persistence of alloantigen-specific CD8⁺ T_E depend on the presence of chronic alloantigen. (A) CFSE-labeled donor T cells (CD4⁺ and CD8⁺ T cells, 3×10⁶ for each) from normal B6/SJL mice (CD45.1) were transplanted, together with B6 TCD BM (CD45.2, 5×10⁶), into lethally irradiated (10Gys) primary BALB/b recipients (CD45.2). Dot plots show percent of host miHA H60-specific CD8⁺ T_E that were recovered at day 12 after transplantation from spleens and livers of primary GVHD BALB/b recipients (n=4) of B6/SJLTCD BM and B6/SJL T cells. Data are representative of three independent experiments. (B, C and D) Donor alloreactive B6/SJL H60⁺CD8⁺ T_E (5×10⁴) and CD4⁺ T cells (3×10⁵) were isolated 12 days after transplantation, respectively, from these primary GVHD BALB/b recipients (CD45.2) of B6/JL T cells, re-labeled with CFSE, and adoptively transferred together with B6 TCD BM (5×10⁶) into lethally irradiated secondary BALB/b recipients and congenic B6 recipients (CD45.2), respectively. The percentage of donor alloreactive H60⁺CD8⁺ T_E in the peripheral blood of these secondary recipients at day 14, 28 and 43 after adoptive transfer (B). Dot plots show the fraction of H60⁺CD8⁺ T_E in the peripheral blood (C). Donor alloreactive H60⁺CD8⁺ T_E recovered at day 43 from secondary recipients were calculated (D). Data are presented as mean ± SD. * P<0.01. (E) Histograms show the surface markers of donor H60⁺CD8⁺ T_E isolated from the secondary allogeneic BALB/b mice (left), and dot plots show the production of IFN-γ and GZMB (right). Data shown in B, C, D and E are representative of two independent experiments. (F) Survival rate of animals was analyzed by the Kaplan-Meier Method. *, P<0.05, significant difference.

Fig.3

Array based mRNA arrays of alloreactive CD8⁺ T cells. (A) Donor CD44^loCD62L^hi CD8⁺ T_N (2×10⁶) from C3H.SW mice (CD45.2) were transplanted with B6/SJL TCD BM(0.5×10⁶) into lethally irradiated B6/SJL mice (CD45.1). Donor CD8⁺ T cells were recovered at day 14 after transplantation from the spleens and lymph nodes of these recipients, magnetically purified and stained with indicated Abs for flow cytometry cell sorting. Affymetrix Mouse Genome 430A 2.0 Array was used to broadly compare the transcription profiles of these CD8⁺ T_MSC and T_E to that of T_N cells. (B) Real-time RT-PCT analysis shows the relative expression of mRNA in each T cell subset. Data are representative of three independent preparations of alloreactive T cells.

Fig.4

Acquisition of stem cell transcriptional programs in alloreactive CD8⁺ T cells. (A) Genes distinctively expressed by CD8⁺ T_MSC and T_E relative to T_N were analyzed for functional set enrichment analysis using curated gene lists from the Ramalho-Santos's array data. (B) Acquisition of stem cell transcriptional programs in P14 LCMV-gp33-specific CD8⁺ T cells. Genes differentially expressed by CD8⁺ T_E-KLRG1^Hi, CD8⁺ T_E-KLRG1^Hint and CD8⁺ memory T cells relative to T_N were analyzed for enrichment for the same stem cell gene lists. Sizes of stem cell gene lists differ slightly from that in Fig.4A since the Sarkar's study used Affymetrix Mouse_430_2, which also had slightly different probe-set annotation than our arrays.

Fig.5

Characterization of stem cell genes in alloreactive CD8⁺ T cells. (A) The relative expression of stem cell genes in alloreactive CD8⁺ T_E and CD8⁺ T_MSC are shown with genes identified by functional set enrichment analysis using curated gene lists from MSigDBv2 (Fig.4). (B) Real-time RT-PCR analysis shows the relative mRNA expression of selected genes in each T cell subset. Data are representative of three independent preparations of alloreactive T cells. (C) Western blot analysis shows the expression of EZH2 protein in purified day-14 CD8⁺ T_E, T_MSC and control T_N. (D) donor CD8⁺ T_N and day-14 CD8⁺ T cells were stained with anti-Ezh2 Ab using intracellular staining and analyzed by flow cytometry. Dot plots show the expression of Ezh2 in cell subset of donor T cells. Mean fluorescent intensity shows the amount of testing antigen.

Fig.6

Identification of stem cell gene Ezh2 in alloreactive CD8⁺ T cells. (A) The relative expression of 23 Ezh2 target or partner genes that were also selected as increased in alloreactive CD8⁺ T_E relative to T_N. (B) Western blot shows Ezh2 protein in CD8⁺ T_N expressing with Ezh2-shRNA or Control-shRNA. These Ezh2-shRNA GFP⁺CD8⁺ T_N were derived from B6 mice reconstituted with HSCs infected by Ezh2-shRNA/GFP-pLVPT^off. GFP⁺CD8⁺ T_N from B6 mice reconstituted with Con-shRNA/GFP-pLVPT^off were isolated as controls. (C) Sorted Ezh2-shRNA GFP⁺CD8⁺ T_N and Control-shRNA GFP⁺CD8⁺ T_N (1×10⁵ / well, 96 well plate) were stimulated with anti-CD3Ab and anti-CD28 (2.5 μg/ml for each). Five days later, cells were recovered and analyzed with flow cyotmetry for measuring GFP⁺CD8⁺ T cells. (D) Unfractionated Ezh2-shRNA GFP⁺CD8⁺ T_N were stimulated with BABL/c mouse-derived DCs, or with IL-7 alone. Five days later, cells were recovered and analyzed using flow cytometry for measuring GFP⁺CD8⁺ T cells. Fold change of GFP⁺CD8⁺ T cells was calculated based on the output number of GFP⁺CD8⁺ T cells after culture divided by the input number of GFP⁺CD8⁺ T cells before culture. Data shown in (B, C and D) are representative of two independent experiments.