Allogeneic Th1 Cells Home to Host Bone Marrow and Spleen and Mediate IFNγ-Dependent Aplasia

Abstract

Bone marrow graft failure and poor graft function are frequent complications following hematopoietic stem cell transplantation and result in significant morbidity and mortality. Both conditions are associated with graft versus host disease (GVHD), although the mechanism remains undefined. Here we show in two distinct murine models of GVHD (complete MHC- and class II-disparate) that mimic human peripheral blood stem cell transplantation that Th1 CD4⁺ cells induce bone marrow failure in allogeneic recipients. Bone marrow failure following transplant of allogeneic naïve CD4⁺ T cells was associated with increased CD4⁺ Th1 cell development within bone marrow and lymphoid tissues. Using IFNγ-reporter mice, we found that Th1 cells generated during GVHD induced bone marrow failure following transfers into secondary recipients. Homing studies demonstrated that transferred Th1 cells express CXCR4, which was associated with accumulation within bone marrow and spleen. Allogeneic Th1 cells were activated by radiation-resistant host bone marrow cells and induced bone marrow failure through an IFNγ-dependent mechanism. Thus, allogeneic Th1 CD4⁺ cells generated during GVHD traffic to hematopoietic sites and induce bone marrow failure via IFNγ-mediated toxicity. These results have important implications for prevention and treatment of bone marrow graft failure following hematopoietic stem cell transplantation.

Introduction

Hematopoietic stem cell transplantation (HSCT) is an increasingly utilized therapy for treatment of malignant and non-malignant disorders. Although outcomes continue to improve, significant morbidity and mortality continues to limit this treatment for many patients. Bone marrow graft failure (GF) and poor graft function (PGF) occur in up to 25% of patients undergoing HSCT and both are associated with an increased risk of infection and death (1, 2). Risk factors for development of GF and PGF include infection, medication side effects, and graft versus host disease (GVHD) (1). The mechanistic basis for the relationship between GVHD and bone marrow (BM) failure remains poorly defined.

Previous adoptive transfer studies have demonstrated that allogeneic Th17 cells, produced in vitro, induce an atypical form of GVHD manifested primarily by skin and lung disease (3). Similar studies with Th2 cell transfers indicated a decreased pathogenicity of these cells in HSCT mouse models (4, 5). Transfer studies using in vitro generated Th1 cells have been limited by previous isolation methods, and no studies have conclusively determined the role of committed Th1 cells in GVHD using adoptive transfer methodology (6, 7). Here, using a previously reported IFNγ-reporter mouse model (8), we describe GVHD mediated by purified, committed Th1 cells in clinically relevant murine models.

Th1 development is under control of the transcription factor, Tbet, which can be upregulated by interleukin (IL)-12 and other signals (9). Th1 cells produce the signature cytokine, IFNγ, which acts to further promote Th1 development and suppress the development of other lineages. T-bet is elevated in T cells from aplastic anemia patients with bone marrow failure (10). Previous studies have also demonstrated an important role for IFNγ in bone marrow suppression and failure (11–16). In addition, a direct negative effect of IFNγ on CD34⁺ cord blood hematopoietic stem cells has been demonstrated (17). Elegant studies using IFNγ-receptor-deficient recipients revealed increased levels of IFNγ present in recipient blood and tissues, which was associated with hematopoietic failure and lymphoid aplasia. Disease in these mice was dependent on both IFNγ and Fas-FasL (18). IFNγ is a ubiquitous cytokine produced by multiple cell lineages within the immune system, including Th1 cells. CD8⁺ cells in particular are important source of IFNγ, and several studies have indicated that CD8⁺ cells are critical for inducing bone marrow disease (11, 16). Previous work using polyclonal, allogeneic CD4⁺ cells indicated that IFNγ was important for bone marrow disease in the setting of sublethal conditioning, but not in lethal conditioning (13). Other studies exploring CD4⁺ mediated bone marrow suppression have implicated IFNγ-independent mechanisms. Fas-FasL interactions, in particular, seemed to be important in mediating the bone marrow manifestations in these studies (19, 20). It remains uncertain, therefore, whether allogeneic Th1 cells directly mediate suppression of recipient bone marrow function, and, if so, the mechanism(s) of this suppression.

This study significantly extends previous work by definitely demonstrating that allogeneic Th1 cells directly mediate host hematopoietic failure. In addition, we have performed novel studies, through the use of transgenic reporter mouse systems, determining allogeneic Th1 cell homing, and detailed analyses, including mechanism, of Th1-mediated suppression of host hematopoiesis.

Material and Methods

Mice

Mice were purchased from Jackson Laboratory and/or bred at our facility: BALB/cJ (BALB/c), B6.C-H-2bm12 (bm12), C57BL/6J (B6), C57BL/6.Ly5.2 (CD45.1-homozygous), B6.MRL-Fas^lpr/J (Fas deficient), and B6.129S7-Ifngr1^tm1Agt/J (IFNγ receptor deficient). The Ifng-reporter BAC-In transgenic mice were previously described (8). The 3BBM74 transgenic TCR mice were a kind gift from E. Palmer (University Hospital-Basel, Switzerland). All animals were bred and maintained in accordance with the University of Alabama at Birmingham Institutional Animal Care and Use Committee (IACUC) regulations.

Cell Preparation

CD4⁺ cells were purified from pooled spleen and/or lymph nodes by magnetic bead positive selection (Dynal beads, Invitrogen). CD25-depletion was performed using phycoerythrin (PE)-labeled anti-CD25 antibody (eBiosciences) followed by anti-PE magnetic bead depletion (Miltenyi Biotech). Post-sort purity was confirmed by flow cytometry.

Th1 cell purification was performed by first isolating lymphocytes from pooled spleens. Splenocytes were then treated with PE-labeled anti-Thy1.1 (eBiosciences) followed by anti-PE magnetic bead selection (Miltenyi Biotech).

Donor Bone Marrow cells were prepared by pooling cells from both femurs and tibias of donor mice. In all experiments, bone marrow was obtained from CD45.1- or CD45.2-homozygous donor mice to differentiate from co-administered donor CD4⁺ cells. T cell depletion was performed using PE-labeled Thy1.2 antibody (BD Biosciences) followed by anti-PE magnetic bead depletion (Miltenyi Biotech).

Cell Culture

Purified wild-type B6 CD4⁺ cells were cultured in vitro with irradiated B6 splenocytes in Th1 conditions with 1ng/mL rmIL12 (R&D Systems) and 10μg/mL anti-IL4 antibody (clone 11B11), along with 2.5μg/mL anti-CD3 (clone 145-11) stimulation. Cells were cultured for three days and purified for transfer.

Transplant Procedure

Donor and recipient mice were 4–8 weeks of age at time of transplant. Transplants were performed according to UAB IACUC approved protocols. Recipient mice received 900 cGy of total body irradiation in 2 split fractions three hours apart from an x-irradiator (X-RAD 320, Precision X-ray Inc.). At least one hour later, T-cell depleted bone marrow, with or without purified CD4⁺ cells, was administered to anesthetized mice via intravenous injection. Cell doses and experimental groups specified in text and legends. All animals were given water supplemented with trimethoprim/sulfamethoxazole for 4 weeks after transplant.

Mice were weighed at least twice weekly and mice exhibiting severe disease, evidenced by lethargy, severe skin disease, hunching, or weight loss >20% original weight were euthanized and scored as dead.

In order to simplify the manuscript nomenclature, the authors refer to suppression of bone marrow in the mouse recipients of allogeneic cells as “suppression of recipient bone marrow.” However, these recipient mice are lethally irradiated and reconstituted with bone marrow from donor mice.

Flow Cytometric Analysis

Intracellular staining was performed as previously described(21). Live cells were identified using LIVE/DEAD Fixable Far Red Dead Cell Stain (Invitrogen). All fluorescent antibodies for analysis purchased from eBiosciences, unless otherwise specified. Data were acquired on an LSRII flow cytometer (BD Biosciences) and analyzed using FlowJo software (Tree Star).

Bone Marrow Progenitor Analysis

Red blood cells were lysed using ammonium chloride, and bone marrow cells were stained for 30 minutes in HBSS, 2% FBS containing the following antibody combinations. Both stains included anti-c-Kit (2B8; BD Pharmingen) and Sca-1 (E13-161.7; Biolegend) antibodies conjugated to APC and Pacific Blue, respectively, and, for the lineage stains, PE-Cy7 conjugated antibodies against CD3 (145-2C11), CD4 (RM4-5), CD8 (53-6.7), B220 (RA3-6B2), CD19 (1D3), Mac-1 (M1/70), Gr-1 (RB6-8C5), and TER119 (TER119). PE-conjugated anti-IL7-Ra (SB/199: BD Pharmingen) was included in the HSC/lymphoid progenitor stain. Anti-CD34 (RAM34) and CD16/32 (2.4G2; BD Pharmingen) antibodies conjugated to FITC and PE, respectively, were included in the myeloid progenitor stain.

Histology

Tissue samples were fixed in 10% phosphate-buffered formalin solution and labeled without experimental details. Slides were embedded in paraffin, sectioned, and stained by the UAB Animal Resources Program Comparative Pathology Laboratory, and then read by a pathologist blinded to experimental details (T. Schoeb). Tissues were scored according to the extent of disease and the severity of disease, and the product of these scores was reported.

Laboratory Analysis

Complete blood counts were performed on heparinized whole blood samples obtained by tail bleeding, and analysis was performed by the UAB Animal Resources Program Comparative Pathology Laboratory using Abaxis VetScan HMII7.

Statistical Analysis

Statistical significance was determined using either unpaired or paired Student’s t test, as indicated in text. P values are reported in text. Survival analysis was performed using Mantel Cox Test.

Results

GVHD induced by naïve CD4⁺ cells is associated with bone marrow and spleen hypocellularity and Th1 cell predominance

GVHD occurs in the setting of allo-disparity between hematopoietic stem cell donor and recipient, and can be mediated by both CD4⁺ and CD8⁺ T cells. To investigate effector CD4⁺ T cell development in the setting of GVHD and bone marrow (BM) failure, we used the MHC class II-disparate C57BL/6 (B6) to B6.C-H-2bm12 (bm12) model, in which GVHD is mediated solely by CD4⁺ T cells (22, 23).

Bm12 recipients of allogeneic naïve B6 CD4⁺ T cells succumb to GVHD within three weeks of transfer of 10⁵ or 10⁶ cells (Figure 1A). Similarly, BALB/c recipients of naïve B6 CD4⁺ T cells (complete MHC mismatch) develop GVHD with a similar time course (Figure 1B). In the absence of GVHD, irradiated murine recipients of BM alone are usually well engrafted within 3 weeks of transplantation (not shown). Irradiated CD45.2 bm12 recipients of 10⁴ allogeneic CD4⁺CD25⁻CD45.1 T cells were monitored for at least 6 weeks post-transplantation. Death from GVHD generally occurred between weeks 2 to 4 post-transplantation for those mice experiencing severe disease (Figure 1A). To ensure adequate time for engraftment of donor BM cells, mice dying after 3 weeks were selected for histological evaluation. GVHD target organs (lung, skin, liver, gastrointestinal tract) were evaluated histologically, as were hematopoietic sites (BM and spleen). In addition to disease in typical GVHD target organs, BM and spleen demonstrated loss of normal architecture and moderate to severe aplasia, whereas those tissues from the negative control groups had absent or mild disease (Figure 1C).

Figure 1. Naïve CD4⁺ T cell transfer induces lethal graft versus host disease, including severe bone marrow aplasia.

Wild-type C57BL/6 (B6) naïve CD4⁺ cells were purified and transferred to lethally irradiated allogeneic B6.C-H-2bm12 (bm12) at varying doses, along with T cell-depleted bone marrow (TCD BM). Negative controls were syngeneic B6 recipients (syngeneic control) or bm12 recipients of TCD BM. Kaplan-Meier survival curves are shown in (A) for these experiments, with the number of mice in each group indicated in legend. In (B) are Kaplan-Meier curves for similar experiments using allogeneic BALB/c (H-2^d) and syngeneic B6 recipients of 1×10⁶ naïve B6 (H-2^b) CD4⁺ cells, with TCD BM. Moribund mice from (A) were sacrificed (along with negative control mice) and tissues sent for histology scoring (extent x severity) by an independent pathologist blinded to experimental details. Shown in (C) are the scoring results from these tissues. Error bar indicates standard error of the mean (SEM). Results are from five independent experiments.

Donor T cells recovered from spleen and lymph nodes of bm12 recipients were distinguished by the CD45.1/45.2 allotype and phenotyped by flow cytometry. In agreement with previous studies (24–26), the majority of donor CD4⁺ T cells produced IFNγ, consistent with a Th1 allo-response (Figure 2A, B). TNFα levels were slightly decreased in donor CD4⁺ T cells (Figure 2C). There was no significant increase in IL-17, IL-4, or IL-13 in donor CD4⁺ T cells (Figure 2A; not shown) suggesting that the IFNγ-positive T cells in the bone marrow developed via the Th1 pathway. Studies performed using BALB/c recipients gave similar results (Figure 1B and 2A).

Figure 2. Th1 effector cells are present within allogeneic lymphoid tissues and bone marrow in the setting of graft versus host disease.

Naïve CD45.1⁺CD4⁺ T cells (1×10⁶) from C57BL/6 (B6) donors were transferred, along with T cell depleted CD45.2⁺ B6 BM, to lethally irradiated B6.C-H-2bm12 (bm12), BALB/c, or B6 recipients. Six days later, recipient mice were sacrificed and intracellular cytokine or FoxP3 staining and flow cytometric analysis was performed on lymphocytes collected from peripheral lymph nodes (LN), mesenteric lymph nodes (mLN), and spleen. Shown in (A) is representative data for transferred and bone marrow-derived CD4⁺ cells recovered from peripheral lymph nodes of syngeneic B6 control (left column) and allogeneic bm12 (middle column), and BALB/c (right column) recipient mice. Cells were gated by lymphocyte gate, followed by live cells (live/dead staining), and CD4 and CD45.1-positive cells. In (B) are pooled results for IFNγ staining for donor CD4⁺ cells (CD45.1⁺) isolated from LN, mLN, and spleen of B6 and bm12 recipients. Columns indicate mean value with individual percentages displayed by circles. Error bars indicate standard error of the mean (SEM). Percentage of donor CD4⁺ TNF-α positivity is displayed in (C) from the same tissues of B6 and bm12 recipients. Columns indicate mean value and error bars indicate SEM. Bone marrow was harvested from allogeneic and syngeneic control recipients and flow cytometric analysis performed. Cells were gated by lymphocyte gate, followed by live cells (live/dead staining) and CD4-positive cells. Donor T cells were identified by CD45.1-staining. Representative IFNγ staining from a bm12 (allogeneic) recipient is shown in (D). Pooled IFNγ staining results from recipient bone marrow are shown in (E). IFNγ-positive staining was significantly increased in donor T cells from allogeneic recipients compared to syngeneic recipients (p=0.001). Error bars indicate standard error of the mean (SEM). Results are derived from 5 separate experiments.

Due to the presence of severe bone marrow disease in allogeneic recipients of naïve CD4⁺ T cells (Fig. 1C), we isolated donor CD4⁺ T cells from recipient bone marrow. Compared to syngeneic control recipients, donor CD45.1 T cells were present at 2–3-fold increased numbers in allogeneic bone marrow (not shown). Intracellular cytokine staining revealed markedly increased IFNγ in donor CD4⁺ cells from allogeneic compared to syngeneic recipients (Figure 2D, E), indicating a predominance of donor Th1 cells within the affected bone marrow.

In vivo-generated IFNγ-reporter-positive Th1 cells demonstrate marked alloreactivity and expansion in the allogeneic host

To assess the stability and pathogenic potential of allogeneic Th1 cells, we performed transfer experiments using purified Th1 cells. To avoid the possibility that Th1 cells derived ex vivo might not accurately recapitulate all functional features of Th1 cells normally developing in GVHD, we designed a unique model likely to more closely replicate human clinical HSCT. To this end, we utilized transgenic IFNγ-reporter (Ifng/Thy1.1 BAC-In) mice, described previously (8). In this model, the thy1.1 reporter is under control of an Ifng BAC transgene such that Ifng expression is reported without perturbing endogenous Ifng alleles. Use of these reporter mice enables isolation of T cells that have undergone Ifng transcription by stably marking them with surface Thy1.1 expression. This provides a means for isolating naturally arising Th1 cells directly ex vivo without requirement for stimulation. Because Thy1.1 is not expressed in either B6 or bm12 mice, IFNγ-competent (Thy1.1⁺) T cells could be isolated by magnetic sorting.

Alloreactive IFNγ-positive Th1 cells were generated in vivo by transplanting naïve Ifng/Thy1.1 BAC-In CD45.2 CD4⁺ T cells into irradiated bm12 mice (termed “donor” bm12). Six days following transfer, spleens were harvested from these mice and Thy1.1-positive CD4⁺ T cells isolated (Supplemental Figure 1). Thy1.1 purity was ≥95 in all experiments, and Th1 commitment was confirmed in the Thy1.1 population by demonstrating Tbet positivity by flow cytometry (Supplemental Figure 1). Intracellular cytokine staining of Thy1.1-positive cells demonstrated low levels of the lineage-associated cytokines for Th2 (IL-4), Th9 (IL-9), Th17 (IL-17), and Th22 (IL-22) T cells (Supplemental Figure 1). To test the functional capacity of isolated Th1 cells, 1×10⁶ Thy1.1⁺ cells were transferred into lethally irradiated bm12 (“recipient” bm12) or syngeneic B6 mice, along with CD45.1, T-cell depleted BM (2×10⁶ cells per mouse). Ten days following transplantation, spleens from recipient mice were harvested and analyzed by flow cytometry (Figure 3 A, B, C).

Figure 3. Th1 cells expand and retain Th1 phenotype in allogeneic but not syngeneic hosts.

Purified Ifng/Thy1.1 BAC-In CD45.2⁺ Thy1.1-positive cells (1 × 10⁶) were transferred to lethally irradiated allogeneic bm12 or syngeneic B6 control recipients, along with CD45.1⁺ TCD BM (2 × 10⁶). Ten days later, spleens were harvested from recipient mice and analyzed by flow cytometry. Shown in (A) are representative FACS analyses of Thy1.1 expression on live-gated CD45.2⁺ cells recovered from spleens of bm12 (allogeneic) and B6 (syngeneic) recipients. Shown in (B) and (C) are the cumulative results for Thy1.1 and IFNγ staining, respectively, from donor CD4⁺ cells recovered from recipient splenic tissue. Recovered Thy1.1⁺ cell numbers are shown in (D). Results are representative of three independent experiments.

We recovered increased numbers of transferred in vivo-generated Th1 cells from allogeneic recipients compared to syngeneic controls (Figure 3D). These studies demonstrated that naturally arising Th1 cells retain this phenotype while undergoing marked expansion in the allogeneic host. Studies of allogeneic Th1 cells isolated from BALB/c mice transplanted with naïve CD4⁺ T cells from Ifng/Thy1.1 BAC-In mice and then transferred into a second cohort of BALB/c recipients showed similar results (not shown). Similar results were also obtained using Th1 cells generated in vitro from polyclonal wild-type B6 CD4⁺ cells or antigen-specific CD4⁺ T cells from 3BBM74 transgenic TCR mice (Supplemental Figure 2). The 3BBM74 TCR T cells are specific for the MHC class I-A bm12 mutation and are known to induce GVHD in bm12 recipients (27, 28). These studies demonstrate that Th1 cells given at high doses are capable of lethality, even with a polyclonal TCR population.

Allogeneic Th1 cells induce bone marrow and spleen aplasia in a dose-dependent manner

To assess GVHD manifestations induced by naturally arising allogeneic Th1 cells, we transplanted purified Thy1.1⁺ cells at de-escalating doses into lethally irradiated “recipient” bm12 mice and syngeneic B6 controls, along with T cell-depleted BM (Figure 4A). Survival curves revealed that a low dose of alloreactive Thy1.1⁺ T cells (1 × 10³) were capable of inducing lethal disease in half of the bm12 recipients. Lethal disease in low dose recipients of Thy1.1⁺ cells was overcome, however, by increasing the cell dose of concomitantly administered TCD BM (Supplemental Figure 3).

Figure 4. Allogeneic Th1 cells mediate spleen and bone marrow aplasia.

Purified Ifng/Thy1.1 BAC-In CD4⁺Thy1.1⁺ cells were harvested from “donor” bm12 mice and transferred to lethally irradiated syngeneic B6 and allogeneic “recipient” bm12 mice. Kaplan-Meier survival curves for different doses of transferred cells are displayed in (A). The number of mice included in each curve is shown in the legend. Allogeneic mice succumbing to disease were sacrificed and histologic scoring performed by an independent, blinded pathologist. Scores for indicated tissues are shown in (B). Results are cumulative from 5 independent experiments. Shown in (C) are representative images of splenic tissue from syngeneic (left column) and allogeneic (right column) recipients. Images are at 10x magnification. Cumulative histologic scoring results for white pulp (WP) and red pulp (RP) from syngeneic and allogeneic recipients are shown in (D). P value indicates paired Student’s t test analysis for allogeneic red pulp compared to allogeneic white pulp. Representative images (10x magnification) and histologic scoring results for bone marrow tissues are shown in (E) and (F), respectively. Cumulative tissue scores were obtained by multiplying extent and severity (see Methods). Error bars indicate SEM.

Tissues of recipient mice were assessed to identify the mechanism of lethality. Histopathology revealed the absence of disease in typical GVHD target organs (skin, liver, lung, and small and large intestine) (Figure 4B). Stomach, pancreas, thymus, and the central nervous system were also unaffected (not shown). The spleen and bone marrow, however, revealed severe disease (Figure 4B). Specifically, recipients of allogeneic CD4⁺Thy1.1⁺ T cells demonstrated absence of normal splenic architecture compared to negative controls (Figure 4C, D). Both red and white pulp areas were severely disrupted, with the white pulp demonstrating more severe cellular depletion (Figure 4D). Peripheral blood CBC data revealed increased suppression of white blood cells, compared with other hematopoietic lineages (Supplemental Figure 4A). The bone marrow from recipients of allogeneic CD4⁺Thy1.1⁺ T cells demonstrated severe aplasia (Figure 4E, F). Notably, flow cytometric analysis of BM revealed a marked decrease in donor hematopoietic chimerism in recipients of Thy1.1⁺ cells (Supplemental Figure 4B, C).

Allogeneic Th1 cells demonstrate increased toxicity to the lymphoid compartment, including bone marrow lymphoid progenitors

Histology and peripheral blood counts indicated that allogeneic Th1 cells induce increased toxicity to the lymphoid compartment in allogeneic recipients (Figure 4D and Supplemental Figure 4A). Flow cytometric analyses of recipient bone marrow and spleen were performed to further assess the presence of a lineage-specific suppression by allogeneic Th1 cells. Total cell counts were reduced by 50–90% in recipients of allogeneic CD4⁺Thy1.1⁺ T cells compared to syngeneic CD4⁺Thy1.1⁺ control recipients and allogeneic recipients of BM only. Myeloid and lymphoid total cell numbers were both reduced, but lymphoid cells were more severely affected (Supplemental Figure 5). There were decreased percentages of B cells (by CD19 and B220 staining), CD8⁺ cells, and BTLA-expressing B and T lymphocytes. NK cell percentages between allogeneic and syngeneic recipients were not significantly different. Although the absolute number of myeloid cells was also reduced, there was an increased percentage of CD11c-expressing cells and CD11b⁺Gr-1⁺ neutrophils in allogeneic recipients (Supplemental Figure 5).

Additional analyses were performed to examine the effect of allogeneic CD4⁺Thy1.1⁺ cells on recipient hematopoietic stem cell progenitor populations. Bone marrow harvested from recipients of allogeneic CD4⁺Thy1.1⁺ T cells and negative controls four weeks following transplantation was analyzed by flow cytometry. Recipients of allogeneic CD4⁺Thy1.1⁺ cells demonstrated reduced percentages and absolute numbers of c-Kit⁺Sca-1⁺lineage marker⁻ (KSL), IL7Ra⁺ common lymphoid progenitor cells (CLPs) (29) compared to control recipients (Figure 5A). No differences were seen in percentages of common myeloid progenitor cells (CMP) (c-Kit⁺Sca-1⁻ CD34⁺CD16/32⁻), megakaryocyte-erythroid progenitor cells (MEP) (c-Kit⁺Sca-1⁻ CD34⁻CD16/32⁻), and granulocyte-macrophage progenitor cells (GMP) (c-Kit⁺Sca-1⁻ CD34⁺CD16/32⁺); however, total numbers of all populations were significantly decreased in recipients of allogeneic CD4⁺Thy1.1⁺ cells compared to controls (Figure 5B, D). Finally, an increased percentage of KSL hematopoietic stem cells was noted in allogeneic Thy1.1⁺ cell recipients compared to controls, but total numbers of these cells were similar in both groups (Figure 5C, D). Collectively, these results indicate global suppression of recipient hematopoietic lineages with a greater effect on committed lymphoid populations.

Figure 5. Allogeneic Thy1.1 cells demonstrate toxicity to common lymphoid progenitors.

Purified Ifng/Thy1.1 BAC-In CD4⁺Thy1.1⁺ cells (CD45.2⁺) were transferred to lethally irradiated B6.C-H-2bm12 (bm12) mice at a dose of 1 × 10⁵ cells/mouse, along with 3 × 10⁶ TCD BM cells (BM + Th1). Control bm12 mice were given only TCD BM cells (BM only). Four weeks later, all mice were sacrificed, BM removed from both tibias and femurs, and cell counts performed. Cells were then analyzed by flow cytometry to analyze BM progenitors. The flow cytometric gating procedure is detailed in Materials and Methods. Frequency of common lymphoid progenitors is shown in (A). Shown in (B) are the frequencies of myeloid progenitor populations: CMP, GMP, and MEP. The frequencies of hematopoietic stem cells (HSCs) in the experimental and control groups are shown in (C). Total BM cell numbers obtained from recipients of BM only and BM + Th1 cells are shown in (D). Analysis performed using an unpaired Student’s t test demonstrated a statistically significant difference (p=0.002). Data represents cumulative results from 5 mice per group. Error bars indicate SEM.

Th1 cells express the chemokine receptor CXCR4 and demonstrate preferential homing to bone marrow and spleen in allogeneic recipients

The induction of BM and spleen aplasia by transferred Th1 cells in the absence of disease in typical GVHD target tissues suggested tissue-specific homing of these cells. To assess this, the surface expression pattern of homing receptors on CD4⁺Thy1.1⁺ T cells isolated from spleen and peripheral lymph nodes of recipients of allogeneic Th1 cells was analyzed, using splenic CD4⁺ cells from non-irradiated bm12 mice as controls. Most spleen-derived CD4⁺ cells had a CD44^hiCD62L^lo phenotype, consistent with previous activation (Figure 6A). Specific analysis of Thy1.1⁺ Th1 cells revealed a similar pattern, with a slightly increased frequency of CD62L^hi expression in lymph nodes. CCR7 expression was low to absent on Thy1.1⁺ cells (30). Intracellular staining for T-bet and FoxP3 showed that nearly all Thy1.1⁺ T cells were T-bet⁺Foxp3⁻, consistent with a stable Th1 effector-memory phenotype (Figure 6B) (31).

Figure 6. Allogeneic Thy1.1 cells home to bone marrow and spleen.

Purified Thy1.1-positive Ifng/Thy1.1 BAC-In CD45.2⁺CD4⁺ cells were harvested from allogeneic “donor” bm12 mice. Staining was performed for CD62L and CD44, and cumulative flow cytometry results are show in (A). Cells were gated first by lymphocyte gate, followed by live cells (live/dead staining) and CD4- and CD45.2-positive cells. Control cells were obtained from spleens of non-irradiated bm12 mice. Analysis performed on CD4-positive cells within spleen and lymph node is labeled allogeneic spleen and allogeneic LN, respectively. Analysis performed on Thy1.1-positive cells is labeled spleen Thy1.1 positive and LN Thy1.1 positive, respectively. Error bars indicate SEM. Intracellular FACS staining for T-bet and FoxP3 was performed on Thy1.1 positive cells harvested from spleen and lymph node, and cumulative results are shown in (B). Cumulative staining results for selected chemokine receptors are shown in (C). These results are from two independent experiments using 10 mice. Purified Ifng/Thy1.1 BAC-In CD4⁺Thy1.1⁺ cells (CD45.2⁺) were harvested from “donor” bm12 mice and transferred to lethally irradiated syngeneic B6 and allogeneic “recipient” bm12 mice, along with CD45.1⁺ TCD BM. Two weeks later, mice were sacrificed and tissues analyzed for the presence of donor Thy1.1⁺ cells. Shown in (D) are representative CD4⁺ histograms from various labeled tissue sites. Cells were gated by lymphocyte gate, followed by live cells (live/dead staining) and CD45.2-positive cells. The cumulative percentage of Thy1.1-expressing CD4⁺ cells within these tissues is shown in (E). Results are from two independent experiments.

Additional staining was performed to assess expression of chemokine receptors. Notably, Thy1.1⁺ cells obtained from the spleen expressed very high levels of CXCR4 (Figure 6C), a receptor known to be responsible for CD4⁺ cell homing to BM and spleen (32, 33). Additionally, these cells lacked receptors known to facilitate homing to the GI tract, specifically CCR6 and CCR9 (Figure 6C) (34, 35). Thy1.1⁺ cells also expressed high levels of CCR5 and CXCR3, receptors associated with Th1 cells (36, 37).

Homing experiments were performed to determine whether Thy1.1⁺ cells isolated from spleen preferentially migrated to the BM and spleen after transfer to “recipient” bm12 mice. Thy1.1⁺ T cells were purified and transferred to lethally irradiated bm12 and syngeneic B6 mice. Two weeks later, tissues were harvested and analyzed for presence of donor T cells. As shown in Figure 6D, donor T cells were identified within recipient BM and spleen. Insufficient cells were obtained from GI tissues for analysis, but analysis of mesenteric LNs revealed no donor CD4⁺Thy1.1⁺ cells (Figure 6D). Donor CD4⁺ cells were also identified within lung tissue of recipient mice (Figure 6D), consistent with known CXCL12 expression (38). Thy1.1 expression analysis revealed that these donor CD4⁺ cells maintained a Th1 phenotype within these tissues (Figure 6E).

Allogeneic Th1 cells are activated by radiation-resistant recipient bone marrow and spleen cells and induce aplasia through an IFNγ-dependent mechanism

Similar to myeloablative HSCT in humans, in these studies donor bone marrow and Th1 cells are from the same MHC background. In the setting of lethal irradiation, however, radiation-resistant hematopoietic and non-hematopoietic recipient cells persist within BM and other tissues (39, 40), suggesting that donor Th1 cells might generate an allo-response against these remaining recipient BM and spleen cells.

Initial in vitro studies determined whether radio-resistant cells within recipient bone marrow were capable of stimulating Thy1.1⁺ donor T cells. B6 (control) and bm12 mice were administered lethal doses of irradiation and six days later (prior to death) bone marrow cells were removed and placed into culture with purified CD4⁺Thy1.1⁺ T cells. Flow cytometric analysis performed 48 hours later demonstrated that the CD4⁺Thy1.1⁺ cells retained IFNγ expression (as measured by Thy1.1 staining) and secreted increased levels of IFNγ (by ELISA) when stimulated with allogeneic radio-resistant BM cells to levels comparable to the positive control (Figure 7A and not shown). Furthermore, Thy1.1⁺ cells also underwent proliferation after co-culture with allogeneic bm12 BM cells, compared to control BM cells, as measured by carboxyfluorescein succinimidyl ester (CFSE)-staining (not shown).

Figure 7. Allogeneic Th1 cells are activated by radiation-resistant recipient bone marrow cells and mediate bone marrow failure through an IFNγ-dependent mechanism.

B6 (syngeneic) and bm12 (allogeneic) mice were administered lethal irradiation and bone marrow harvested from tibias and femurs six days later, prior to death. RBCs were lysed and recovered bone marrow cells were added to culture. Purified CD4⁺Thy1.1⁺ cells from “donor” bm12 mice were labeled with CFSE and added to bm12 (bm12 BM) or B6 BM cells. Positive control (labeled “Control”) included purified Thy1.1-positive cells co-cultured with B6 (syngeneic) BM cells along with anti-CD3 antibody stimulation. Cells were cultured for 48 hours and analyzed for Thy1.1 expression (A). Cells were gated by lymphocyte gate, followed by live cells (live/dead staining) and CD45.2-positive cells.

Purified CD4⁺Thy1.1⁺ cells were harvested from “donor” BALB/c mice (H-2^d) and transferred to lethally irradiated “recipient” BALB/c mice at a dose of 0.5–1×10⁵ cells, along with 1×10⁶ T cell-depleted B6 (H-2^b) BM cells. BALB/c recipients of BM only and bm12 recipients (H-2^b) of B6 BM were used as negative controls. Mice were sacrificed four weeks later, and bone marrow isolated from femurs and tibias. FACS staining was performed, and representative histograms are shown in (B). Dead cells were removed from analysis using live/dead staining. Results are shown for BALB/c recipient of BM only (thick black line) and BALB/c recipient of Thy1.1 cells and BM (thin black line). The gray line is for bm12 recipient of BM only. Cumulative H-2^d staining is shown in (C). Error bars indicate SEM. Results are from two independent experiments using 24 mice.

Purified CD4⁺Thy1.1⁺ cells (0.5 × 10⁵ cells/mouse) were harvested from “donor” bm12 mice and transferred to lethally irradiated bm12 “recipient” mice, along with 2×10⁶ BM cells from either wild-type (n=15), IFNγ receptor knock-out (IFNγR KO) (n=8), or Fas knockout B6 mice (n=8). Syngeneic B6 recipients of Thy1.1-purified cells and BM were used as negative control. Four weeks later, surviving mice were sacrificed and bone marrow sent for histologic scoring by independent pathologist blinded to experiment. Cumulative histologic scoring results are shown in (D). Error bars indicate SEM. Shown in (E) are Kaplan-Meier survival curves. Statistical analysis revealed a significant difference in survival in recipients of wild type BM (p=0.01). Analysis of IFNγR KO BM and FasKO BM recipients revealed a significant trend by log-rank trend analysis, but Mantel-Cox testing was not significant (p=0.07). Statistical results are shown on figure. Results are from 3 independent experiments. Statistical analysis performed using unpaired Student’s t test with p value shown.

In order to confirm these findings in vivo, the same transplant schema described previously was used except that BALB/c rather than bm12 mice were used as recipients to enable discrimination of donor and recipient cell populations. Using this experimental design, BALB/c cells remaining in the bone marrow after irradiation were identified based on their expression of H-2^d. Purified B6 CD4⁺Thy1.1⁺ cells (0.5–1×10⁵ cells/mouse) were transplanted into lethally irradiated BALB/c mice along with T cell-depleted B6 BM cells. BALB/c recipients of BM only and bm12 recipients (H-2^b) of B6 BM were used as controls. All mice were sacrificed after four weeks, and BM isolated. BM cellularity was markedly reduced in recipients of allogeneic Thy1.1⁺ Th1 cells compared to BM only recipients (not shown). Furthermore, analysis of H-2^d expression revealed reduced numbers of BALB/c-derived H-2^d expressing cells within the marrow of Th1 cell recipients compared to recipients of BM only (Figure 7B, C). Recipients of Th1 cells demonstrated reduced BM cell numbers and loss of cells of recipient origin (BALB/c) within the BM, indicating that the transferred Thy1.1⁺ cells likely recognize remaining radio-resistant recipient cells.

Th1 cells can mediate cytotoxic allogeneic effects through several mechanisms, including IFNγ secretion and Fas-FasL interactions. To determine whether either mechanism might be responsible for the BM and spleen aplasia induced by alloreactive Th1 cells, B6 CD4⁺Thy1.1⁺ cells (0.5 × 10⁵ cells/mouse) were transplanted into bm12 recipients along with TCD BM cells from IFNγ receptor (IFNγR)-deficient, Fas-deficient, or WT mice (all B6 background). Recipients of IFNγR-deficient BM demonstrated significantly reduced histologic scoring compared to WT controls (Figure 7D). Recipients of Fas-deficient BM had reduced disease that did not achieve statistical significance, despite increased expression of CD178 (Fas ligand) by the Thy1.1⁺ Th1 cells (not shown). Recipients of IFNγR-deficient and Fas-deficient demonstrated improved survival compared to recipients of WT BM (p=ns by Mantel-Cox Test) (Figure 7E). Collectively, these data indicate that allograft Th1 cells recognize radiation-resistant cells in the recipient BM that induce the release of IFNγ, which is directly toxic to donor hematopoietic cells.

Discussion

The studies herein utilize a novel reporter mouse system in order to determine the alloresponse of highly purified Th1 cells in several distinct mouse models of graft versus host disease. Our studies have extended previous work by definitively demonstrating that Th1 cells directly mediate bone marrow failure in allogeneic host mice. Furthermore, we have significantly advanced these findings to demonstrate that allogeneic Th1 cells home preferentially to spleen and marrow, and induce a marked suppression of the lymphoid compartment, including common lymphoid progenitors. Our results provide preliminary evidence indicating that allogeneic Th1 cells recognize radiation-resistant host cells remaining within host bone marrow, inducing activation and IFNγ release. Finally, we have shown that Th1-mediated bone marrow suppression occurs through an IFNγ-dependent mechanism.

Our findings demonstrate that alloreactive Th1 cells that develop during GVHD are capable of inducing lethal bone marrow failure in fully myeloablated MHC-disparate recipients, and that graft failure is mediated through an IFNγ-dependent mechanism. This is the first study to directly demonstrate a Th1 cell-mediated allo-response in bone marrow and was enabled by implementation of IFNγ-reporter mice that allows isolation of differentiated Th1 cells without requirement for T cell stimulation, which can induce Th1 cell apoptosis. This model has the further advantage that it utilizes effector T cells that develop in the context of a natural allo-response in vivo, thereby recapitulating the normal expression of chemokine and other receptors present in transplanted cellular grafts and more closely mimicking the human transplant setting.

A role for IFNγ in mediating BM and spleen aplasia in sublethal models of HSCT was previously reported (12, 13); however, it appeared to play a protective role in the setting of ablative conditioning (13). Our studies have focused on the myeloablative HSCT setting, where recipient hematopoiesis is eliminated. Studies using lethally irradiated recipients demonstrated a role for CD8⁺ and CD4⁺ cells in mediating BM aplasia (18, 20). In these studies, CD4⁺ T-cell mediated BM aplasia occurred primarily through Fas-FasL interactions and B cells were primarily affected (20). Fas-FasL interactions have also been shown to be important in mouse models of aplastic anemia (41). Our studies differ from the aforementioned studies because we have utilized adoptive transfer methodologies to determine a direct role for purified, committed Th1 cells in mediating BM failure. These experiments revealed that CD4⁺ cells are capable of inducing severe BM failure that is primarily mediated through IFNγ.

Our studies indicate that allogeneic Th1 cells are activated by radiation-resistant hematopoietic or non-hematopoietic bone marrow cells. Recent studies have shown that GVHD can be mediated by recipient non-hematopoietic APCs, and that these cells appear even more potent in generating GVHD than donor-derived hematopoietic APCs (42, 43). Our studies have demonstrated that these radiation-resistant APCs were sufficient for activation of allogeneic Th1 cells in the allograft, resulting in IFNγ-induced toxicity that targeted bone marrow hematopoietic cells through a mechanism yet to be defined (Figure 7). Although a minor role for Fas-dependent cytotoxicity could not be excluded based on survival curves (Figure 7E), histologic analysis demonstrated that BM toxicity mediated by IFNγ was clearly dominant (Figure 7D). It is possible, however, that IFNγ and Fas-FasL interactions are not mutually exclusive, and future studies will address possible mechanisms of IFNγ-mediated aplasia.

A profound suppressive effect was noted on lymphoid, myeloid and erythroid lineages within the peripheral blood and BM by allogeneic Th1 cells (Supplemental Figure 4A and 5). More severe damage to splenic white pulp was also consistent with peripheral lymphocyte destruction (Figure 4D). These findings are in agreement with previous studies (11, 16). Flow cytometry data revealed a lesser effect on myeloid populations, with increased percentages of CD11c- and Gr-1-positive cells within recipient BM likely indicating the preferential targeting of lymphoid lineages (Supplemental Figure 5). Our studies also revealed decreased hematopoietic progenitors from all lineages within the allogeneic host BM, with a more pronounced effect seen on common lymphoid progenitors (Figure 5). Collectively, these studies indicate that allogeneic Th1 cells induce BM failure, with a greater suppressive effect on the lymphoid lineage (both committed lymphocytes and CLPs), perhaps secondary to differential IFNγ receptor expression (44, 45). Future studies will explore the mechanism of this lymphoid “bias” by allogeneic Th1 cells.

Thy1.1⁺ cells within the BM and spleen of allogeneic recipients expressed high levels of the chemokine receptor, CXCR4. This receptor has previously been shown to be critical for CD4⁺ T cell trafficking between BM and peripheral blood (32). In addition to homing, CXCR4 also has a role in T cell signaling and upregulation of anti-apoptotic proteins (46, 47). We have demonstrated that murine naïve CD4⁺ T cells grown in vitro in Th1 conditions also upregulate CXCR4 (unpublished data). These studies demonstrate that transferred Thy1.1⁺ effector Th1 cells accumulate within allogeneic BM associated with increased CXCR4 expression. Further studies are required to determine if CXCR4 expression is required for Th1-induced BM failure.

IFNγ has also been shown to be critical in mediating gastrointestinal GVHD (48, 49). We found Th1 cells present in GI tissues following allogeneic naïve CD4⁺ transplant, consistent with previous studies (25, 48, 49). However, we did not find disease in these organs following purified Thy1.1⁺ cell transfer, nor was there any evidence of homing to the GI tract by the Th1 cells used herein. This is likely due to the absence of important gut-homing chemokine receptors (CCR6 and CCR9) on the transferred effector Th1 cells (34, 35), and likely reflects the fact that the Thy1.1⁺ T cells used were harvested from donor spleens. Despite expressing the chemokine receptors CXCR3 and CCR5, we were unable to detect pathology within skin or liver of recipient mice. It is currently not known whether this is secondary to absence of trafficking to these tissues or whether committed Th1 cells do not induce disease in these tissues (25). Interestingly, transferred allogeneic Th1 cells were recovered from the lungs of recipient mice (Figure 6D), but there was no significant pathology in this tissue. This is consistent with previous reports that IFNγ is not an important mediator of lung GVHD (48, 50). It is possible that Th1-mediated disease occurs within the bone marrow because the marrow is simply more a sensitive organ to allogeneic T cell effects compared to other end-organs. However, we feel it is less likely as the disease pattern in recipient mice tissues was similar, even at higher doses of allogeneic Th1 cells (Figure 4).

The current studies were designed to mimic human stem cell transplantation. For this reason, Thy1.1⁺ cells were generated in vivo and acquired from only the spleens (and not lymph nodes) of donor mice because these cells reflect peripheral blood composition which is consistent with the clinical scenario in human HSCT using peripheral blood or bone marrow harvested cells (51–53). In contrast to the clinical setting, however, we did not utilize cytokine mobilized/conditioned bone marrow for these studies. These studies also used relatively lower doses of transferred CD4⁺ T cells and bone marrow compared to other models (13, 24). Future studies will address whether cytokine mobilization of donor bone marrow can alter the suppressive effect of Th1 cells and/or IFNγ.

In summary, we have here conclusively demonstrated that allogeneic Th1 cells induce secondary bone marrow failure through IFNγ-mediated and perhaps other minor mechanisms. This study illuminates at least one possible etiology for GF and PGF associated with human GVHD and should impact the design of future trials for improved hematopoietic stem cell graft manipulation in humans.

Supplementary Material

01. Supplemental Figure 1. Thy1.1 generation and purification.

Purified naïve CD4⁺ cells from C57BL/6 (B6) IFNγ Bac-In transgenic reporter mice (Ifng/Thy1.1 BAC-In) were transferred to lethally irradiated allogeneic B6.C-H-2bm12 (bm12) mice (“donor” bm12). Six days later, these mice were sacrificed and spleens removed. Lymphocytes were isolated using Ficoll separation and then analyzed for FACS phenotype. Thy1.1-expressing cells were purified using magnetic bead sorting and transferred to lethally irradiated allogeneic bm12 mice (“recipient” bm12) or C57BL/6 (B6) syngeneic mice. Schema for in vivo generation of purified Thy1.1-positive Th1 cells is shown in (A). Shown in (B) are representative FACS for unstimulated (left) and PMA and ionomycin 4-hour-stimulated CD4⁺ cells isolated on day 6 from “donor” bm12 mice. Cells were gated by lymphocyte gate, followed by live cells (live/dead staining) and CD4-positive cells. IFNγ is shown on the x-axis and the y-axis is Thy1.1 staining. Cells were purified for Thy1.1 expression using magnetic bead separation. Shown in (C) is post-sorting FACS histogram for Thy1.1 expression. A representative FACS histogram for Tbet staining in Thy1.1-positive cells is shown in (D). The filled histogram depicts the control staining. Cells in (C) and (D) were gated by lymphocyte gate, followed by live cells (live/dead staining) and CD4-positive cells. Results are representative of seven independent experiments. Shown in (E) are intracellular FACS staining results for lineage-associated cytokines in recovered Thy1.1-positive CD4⁺ cells. Staining results are shown for Th2- (IL-4), Th9- (IL-9), Th17- (IL-17), and Th22-associated (IL-22) cytokines. Results are representative of two separate experiments.

02. Supplemental Figure 2. In vitro-derived Th1 cells induce a lethal allo-response.

Naïve CD4⁺ cells from wild-type C57BL/6 (B6) or 3BBM74 transgenic TCR mice (3BBM) were purified and cultured in vitro with anti-CD3 and irradiated B6 antigen-presenting cells (APCs) in Th1 conditions (1ng/mL rmIL-12 and 10μg/mL anti-IL4 antibody) for three days. Live cells were isolated by Ficoll separation and 1 × 10⁶ cells were transferred to lethally irradiated allogeneic B6.C-H-2bm12 (bm12) or syngeneic B6 recipients, along with T cell depleted, CD45.1⁺ BM. Shown in (A) are the pre-transplant FACS phenotypes. Cells were first gated on lymphocyte gate, followed by live cells (by live/dead stain) and CD45.2 expression. Kaplan-Meier survival curve is shown in (B) for allogeneic bm12 recipients of polyclonal wild-type B6 or 3BBM transgenic Th1 cells, as well as B6 recipients of polyclonal B6 Th1 cells (syngeneic control). Mice were followed for 12 days and surviving mice were sacrificed (indicated by arrow). Lymph tissues were harvested and lymphocytes were analyzed by flow cytometry. Cumulative percentage of donor CD4⁺ IFNγ-positivity is shown in (C). Error bars indicate SEM.

03. Supplemental Figure 3. Th1 (Thy1.1-positive) lethality can be overcome with increased bone marrow dose.

Purified Thy1.1-positive Ifng/Thy1.1 BAC-In CD4⁺ cells were harvested from allogeneic bm12 mice (“donor” bm12) and 1×10⁴ cells were transferred to lethally irradiated syngeneic B6 and allogeneic bm12 (“recipient”) mice, along with increasing doses of T cell depleted bone marrow (TCD BM). Kaplan-Meier dose-response survival curves for each bone marrow dose are shown.

04. Supplemental Figure 4. Thy1.1 cells mediate severe toxicity to the bone marrow in allogeneic recipients and induce a pronounced peripheral blood lymphopenia.

Purified Ifng/Thy1.1 BAC-In CD45.2⁺ Thy1.1-positive cells were transferred to lethally irradiated allogeneic B6.C-H-2bm12 (bm12) or syngeneic B6 control recipients. Peripheral blood was obtained from syngeneic and allogeneic mice three weeks following transplant. Shown in (A) are CBC results from samples obtained three weeks after transplant, with hemoglobin results contained in left panel and white blood cell counts shown in right panel. In (B), Thy1.1⁺ CD45.2⁺ cells were transferred along with 5 × 10⁶ CD45.1⁺ T-cell depleted B6 bone marrow (Thy1.1 & BM) cells to lethally irradiated allogeneic bm12 recipients. Control bm12 mice were given bone marrow only (BM only). Four weeks later, bone marrow was harvested from recipients from both groups. Shown are representative FACS plots of bone marrow donor chimerism. Cells were gated by lymphocyte gate, followed by live cells (live/dead staining). CD45 chimerism is shown as histogram of CD45.1. Shown in (C) are the cumulative results for CD45.1 chimerism for both Thy1.1 and BM and BM only recipients. Total nucleated cell counts were 2.5–3 times lower in the recipients of Thy1.1-positive cells compared to BM only control.

05. Supplemental Figure 5. Allogeneic Thy1.1 cells demonstrate greater toxicity to the lymphoid compartment of recipient spleen and bone marrow.

Purified Ifng/Thy1.1 BAC-In Thy1.1-positive cells (CD45.2-homozygous) were transferred to lethally irradiated B6.C-H-2bm12 (bm12) or syngeneic B6 control recipients. Moribund mice were sacrificed and cells removed from spleen and bone marrow. FACS analysis was only performed on allogeneic mice with sufficient cells, eliminating approximately 25% of allogeneic mice. FACS results for syngeneic and allogeneic spleen cells are shown in two columns on left, and results for bone marrow cells shown to right. Cells were gated on live cells (by live/dead staining) and CD45.2-negative cells. Histograms are labeled appropriately. Results are representative of two separate experiments with ten mice per group.

Abbreviations used in this article

GF

graft failure

PGF

poor graft function

HSCT

hematopoietic stem cell transplantation

BM

bone marrow

GVHD

graft versus host disease

Lymph node

LN