SEARCH

SEARCH BY CITATION

Keywords:

  • humanized mice;
  • IL-2R ‘common’ gamma chain;
  • SCID;
  • T cell

Summary

  1. Top of page
  2. Summary
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. Disclosure
  9. References
  10. Supporting Information

Immunodeficient mice bearing targeted mutations in the IL2rg gene and engrafted with human immune systems are effective tools for the study of human haematopoiesis, immunity, infectious disease and transplantation biology. The most robust human immune model is generated by implantation of human fetal thymic and liver tissues in irradiated recipients followed by intravenous injection of autologous fetal liver haematopoietic stem cells [often referred to as the BLT (bone marrow, liver, thymus) model]. To evaluate the non-obese diabetic (NOD)-scid IL2rγnull (NSG)–BLT model, we have assessed various engraftment parameters and how these parameters influence the longevity of NSG–BLT mice. We observed that irradiation and subrenal capsule implantation of thymus/liver fragments was optimal for generating human immune systems. However, after 4 months, a high number of NSG–BLT mice develop a fatal graft-versus-host disease (GVHD)-like syndrome, which correlates with the activation of human T cells and increased levels of human immunoglobulin (Ig). Onset of GVHD was not delayed in NSG mice lacking murine major histocompatibility complex (MHC) classes I or II and was not associated with a loss of human regulatory T cells or absence of intrathymic cells of mouse origin (mouse CD45+). Our findings demonstrate that NSG–BLT mice develop robust human immune systems, but that the experimental window for these mice may be limited by the development of GVHD-like pathological changes.


Introduction

  1. Top of page
  2. Summary
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. Disclosure
  9. References
  10. Supporting Information

Immunodeficient mice engrafted with human immune systems represent a promising alternative for the in-vivo study of human immune systems without placing patients at risk [1-4]. These ‘humanized’ mice are created by the engraftment of immunodeficient mice with mature human immune cell populations, human haematopoietic stem cells (HSC) or human fetal tissues [5-7]. Early humanized models using immunodeficient mice bearing the Prkdcscid (scid) recombination activating gene 1 (Rag1null) or 2 (Rag2null) mutations were limited by low levels of systemic engraftment of human immune cells, variability in the overall levels of human cell survival and limited functionality of the human immune system [8]. The limitations of these initial immunodeficient mouse models were largely overcome by the introduction of targeted mutations in the interleukin (IL)-2 receptor common gamma chain (IL2rg) gene [8]. The IL-2rγ-chain is required for high-affinity ligand binding and signalling through multiple cytokine receptors, including those for IL-2, IL-4, IL-7, IL-9, IL-15 and IL-21 [9]. Immunodeficient mice bearing a targeted mutation within the IL2rg gene support higher levels of human haematolymphoid engraftment than all previous immunodeficient stocks and permit the engraftment of functional human immune systems [10-19].

Although a number of engraftment strategies are currently being used to produce humanized mice [8], the implantation of human fetal thymic and liver tissues accompanied by intravenous (i.v.) injection of human fetal liver HSC offers a number of advantages, including robust levels of human cell chimerism, development of functional human T cells and education of T cell progenitors on autologous human thymic epithelium [20, 21]. This fetal thymus/liver model is often referred to as the BLT (bone marrow, liver, thymus) model [2, 6, 22, 23]. The standard protocol to generate BLT mice involves the implantation of human fetal thymic and liver tissues into irradiated mice and then injection of HSC derived from the autologous fetal liver tissues [23-25]. Alternatively, human HSC derived from allogeneic sources will also allow human T cell development [6, 26]. BLT mice have been used to study a number of aspects of human biology, including human haematopoiesis [27-36], immune responses to Epstein–Barr virus (EBV), dengue virus, HIV, West Nile virus and xenogeneic tissues [23, 24, 37-42], EBV pathogenesis [43], HIV pathogenesis and anti-HIV therapies [17, 39, 44-53]. However, BLT mice have been shown to develop a graft-versus-host disease (GVHD)-like syndrome at later points post-engraftment and disease onset has been associated with T cell activation [26, 54].

In this study we evaluate various parameters for establishing the non-obese diabetic (NOD)-scid IL2rγnull (NSG)–BLT model, and potential mechanisms underlying their ultimate development of the GVHD-like syndrome. Variation of the engraftment parameters has a significant effect on the levels of chimerism achieved and the development of T cells. Development of the GVHD-like syndrome correlated with the activation of human T cells and increased levels of human immunoglobulin (Ig), suggesting a spontaneous activation and loss of ‘self-tolerance’ of the human immune system. The onset of GVHD was not delayed in NSG mice lacking murine major histocompatibility complex (MHC) classes I or II and was not associated with a loss of human regulatory T cells (Treg) or absence of intrathymic mouse antigen-presenting cells (APCs) in the developing human thymus. Together these observations define the ideal conditions for generating human immune system-engrafted NSG–BLT mice and the optimal time-frame for their experimental use.

Materials and methods

  1. Top of page
  2. Summary
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. Disclosure
  9. References
  10. Supporting Information

Mice

NOD.Cg-PrkdcscidIl2rgtm1Wjl/SzJ (NOD-scid IL2rγnull, NSG) mice, NOD.Cg-PrkdcscidIl2rgtm1WjlH2-Ab1tm1Gru/Sz (NOD-scid IL2rγnull Ab°, NSG-Abo) mice, which do not express murine MHC class II molecules on the cell surface [55, 56], and NOD.Cg-PrkdcscidIl2rgtm1Wjl H2-K1tm1Bpe H2-D1tm1Bpe/Sz [NSG-(KbDb)null] mice, which do not express murine MHC class I molecules, were obtained from colonies developed and maintained by LDS at The Jackson Laboratory (Bar Harbor, ME, USA). The [NSG-(KbDb)null] mice were developed by first crossing STOCK-H2-(KbDb)null mice [57] with NOD-scid/scid mice and back-crossing the (KbDb)null double knock-out for 12 generations onto the NOD-scid strain. After fixing both scid and (KbDb)null to homozygosity, NOD-scid/scid (KbDb)null mice were crossed with NSG mice and additional genetic crosses were carried out to fix the scid, IL2rgnull and (KbDb)null mutations to homozygosity. The stock is maintained by matings of [NSG-(KbDb)null] sibs. All animals were housed in a specific pathogen-free facility in microisolator cages, given autoclaved food and maintained on sulphamethoxazole–trimethoprim medicated water (Goldline Laboratories, Ft Lauderdale, FL, USA) and acidified autoclaved water on alternating weeks. All animal use was in accordance with the guidelines of the Animal Care and Use Committee of the University of Massachusetts Medical School and The Jackson Laboratory and conformed to the recommendations in the Guide for the Care and Use of Laboratory Animals (Institute of Laboratory Animal Resources, National Research Council, National Academy of Sciences, 1996).

Human peripheral blood mononuclear cells (PBMC) collection

Human PBMC were collected in heparin from healthy volunteers under signed informed consent in accordance with the Declaration of Helsinki and approval from the Institutional Review Board of the University of Massachusetts Medical School.

Tissue transplantation

Human fetal thymus and fetal liver (gestational age between 16 and 20 weeks) specimens were provided by Advanced Bioscience Resources (Alameda, CA, USA) or StemExpress (Placerville, CA, USA). Upon receipt, tissues were washed with RPMI supplemented with penicillin G (100 U/ml), streptomycin (100 mg/ml), fungizone (0·25 μg/ml) and gentamycin (5 μg/ml) and then 1 mm3 fragments were prepared from the thymus and liver for transplantation. When indicated 1 mm3 fragments of fetal NSG mouse liver were co-implanted with the human tissues. The remaining human fetal liver was processed to recover human HSC as described below. Indicated groups of recipient mice were irradiated with 200 cGy and then implanted with a fetal thymus and fetal liver fragment together in the renal subcapsular space or subcutaneously in the ventral area. Following surgery, recipient mice received a subcutaneous injection of gentamycin (0·2 mg) and cefazolin (0·83 mg).

Enrichment of CD34+ HSC from fetal liver tissue

To recover human HSC, fetal liver was minced and digested at 37°c for 20 min with a collagenase-dispase buffer (liver digest medium; Gibco, Carlsbad, CA, USA). The recovered cell suspension was then washed with RPMI supplemented with 10% fetal bovine serum (FBS) and filtered through a metal sieve. Red blood cells were removed by Ficoll-Hypaque density centrifugation. The fetal liver cells were then depleted of CD3+ cells using a magnetic bead separation technique (Miltenyi Biotec, Inc., Auburn, CA, USA) and the percentage of CD34+ cells determined by flow cytometry. At a minimum of 4 h after irradiation of recipient mice, CD3-depleted fetal liver cells were injected i.v. with 1 to 5 × 105 CD34+ HSC per mouse.

Antibodies and flow cytometry

For analysis of human haematopoietic engraftment, monoclonal antibodies specific for mouse CD45 (30-F11), human CD45 (2D1), CD3 (UCHT1), CD4 (RPA-T4), CD8 (RPA-T8), CD10 (HI10A), CD11c (B-ly6), CD14 (HCD14), CD20 (2H7), CD27 (M-T271), CD33 (WM53), CD34 (581), CD38 (HIT2), CD45RA (HI100), CD123 (AC145) and IgD (IAG-2) were purchased from either BD Biosciences, Inc. (San Jose, CA, USA), eBiosciences (San Diego, CA, USA) or BioLegend (San Diego, CA, USA). Single-cell suspensions of bone marrow and spleen were prepared from engrafted mice, and whole blood was collected in heparin. Single-cell suspensions of 1 × 106 cells in a 50 μl or 100 μl of whole blood were washed with fluorescence activated cell sorter (FACS) buffer [phosphate-buffered saline (PBS) supplemented with 2% FBS and 0·02% sodium azide] and then preincubated with rat anti-mouse CD16/CD32 (clone 2.4G2) to block Fc binding. Specific antibodies were then added to the samples and incubated for 30 min at 4°C. Stained samples were then washed and fixed with 2% paraformaldehyde for cell suspensions or treated with BD FACS lysing solution for whole blood. At least 50 000 events were acquired on LSRII or FACSCalibur instruments (BD Biosciences). Data analysis was performed with FlowJo (Tree Star, Inc., Ashland, OR, USA) software.

In-vitro cytokine production assay

Cytokine production by human CD4 and CD8 T cells was quantified using the BD Cytofix/Cytoperm Kit Plus GolgiStop (BD Biosciences), according to the manufacturer's instructions. Splenocytes were recovered from the indicated mice at 12 weeks after implant of fetal tissues. Red blood cells were lysed and 1 × 106 cells were then left unstimulated or stimulated with phorbol myristate acetate (PMA) (0·5 μg/ml) and ionomycin (0·5 μg/ml) in the presence of GolgiStopTM (0·1 μg/ml) for 4 h at 37°C in 5% CO2. Cells were then fixed and permeabilized using Cytofix/Cytoperm solution and stained with monoclonal antibodies (mAb) to interferon (IFN)-γ (clone 4S.B3; eBioscience), IL-2 (clone MQ1-17H12; eBioscience), IL-17A (clone eBio64DEC17; eBioscience) and IL-22 (clone IL22JOP; eBioscience). Stained samples were analysed as described above.

Detection of human Treg

CD4+ human Treg were identified in the blood of NSG–BLT mice by staining with antibodies specific for human CD25 (clones MA-251 and 2A3), CD127 (clone A019D5) and forkhead box protein 3 (FoxP3) (clone 236A/E7). For staining, 100 μl of whole blood were washed with FACS buffer and then preincubated with rat anti-mouse FcR11b. Antibodies specific for human cell surface markers (CD45, CD3, CD4, CD25 and CD127) were added to the samples and incubated for 30 min at 4°C. Stained blood samples were then treated with BD FACS lysing solution for whole blood. Cells were incubated in eBioscience fixation/permeabilization buffer for 60 min and stained with antibodies specific for human FoxP3 in eBioscience permeabilization buffer for 60 min. Stained samples were analysed as described above.

Quantitation of human IgM and IgG levels

Heparinized blood samples from engrafted mice were centrifuged and the plasma was stored at −80°C. Human IgM and IgG levels were measured using an enzyme-linked immunosorbent assay (ELISA) kit (Bethyl Laboratories, Inc., Montgomery, TX, USA) according to the manufacturer's instructions and an EMax Endpoint ELISA microplate reader (Molecular Devices, Sunnyvale, CA, USA).

Haematological analyses

Blood samples were collected in 1·5 ml Microvette haematological tubes coated with ethylenediamine tetraacetic acid (EDTA) (Sarstedt, Newton, NC, USA), and haematological analyses including haematocrit (HCT), red blood cell (RBC), platelet (PLT) and haemoglobin (HGB) values were recorded on a Heska CBC-Diff Veterinary Hematology analyser (Loveland, CO, USA).

Histological analyses

Thymic implants were recovered, fixed in 10% neutral buffered formalin and processed as described previously for histology and histochemistry [18]. Briefly, fixed tissues were embedded in paraffin and 5-μM sections were prepared from the blocks. Sections were stained for haematoxylin and eosin (H&E) and immunostained with a monoclonal antibody specific for human CD45 [either clones 2B11 and PD7/26 from Dako (Glostrup, Denmark) or clone HI30 from BD] or mouse CD45 (clone 30-F11, BD), as described previously [18, 58]. Sections were maintained without any medium. Digital light microscopic images were recorded at room temperature (RT) with either a Nikon EclipseE600 microscope (with ×10 and ×20 Nikon objective lenses), a Diagnostic Instruments Spot RT colour camera and Spot version 5.0 Basic Software or with a Hamamatsu Nanozoomer 2.0HT equipped with an Olympus UPlanSApo 20x/0.75NA objective and NDP.serve software.

Statistical analyses

To compare individual pairwise groupings, we used one-way analysis of variance (anova) with Bonferroni post-tests and Kruskal–Wallis test with Dunn's post-test for parametric and non-parametric data, respectively. Significant differences were assumed for P-values < 0·05. Statistical analyses were performed using GraphPad Prism software (version 4.0c; GraphPad, San Diego, CA, USA).

Results

  1. Top of page
  2. Summary
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. Disclosure
  9. References
  10. Supporting Information

Evaluating the role of host mouse irradiation and human thymic tissue implantation for human cell chimerism in the NSG–BLT models

The BLT mouse model allows for the development of a complete human immune system including the efficient generation of peripheral human T cells [59]. The standard protocol for generating BLT mice includes the irradiation of recipient immunodeficient mice prior to tissues implant [59]. However, whether or not irradiation of the murine host in establishing haematopoietic chimerism in the BLT model is required for optimal engraftment of the human tissues and subsequent T cell development has not been reported. We first evaluated the importance of irradiation for human cell chimerism in adult NSG mice injected with fetal liver-derived human HSC only (no thymic implant) and compared levels of chimerism in mice implanted with human thymic and liver tissues and injected with human HSC (thymic implant). Levels of human CD45+ cells were examined in the blood at 12 weeks (Fig. 1a) after implant and in the blood (Fig. 1b), spleen (Fig. 1c,d) and bone marrow (Fig. 1e) at 16 weeks after implant. Significantly higher levels of human CD45+ cells were detected at 12 (Fig. 1a) and 16 (Fig. 1b) weeks in the blood of NSG mice that were irradiated and implanted with fetal thymic and liver tissues compared to non-irradiated groups and irradiated NSG mice injected with human HSC only. In the spleen, the percentage of human CD45+ cells (Fig. 1c) was similar between the groups, with the exception of non-irradiated mice injected with human HSC only. However, the total number of human CD45+ cells in the spleen (Fig. 1d) was significantly higher in NSG mice that were irradiated and implanted with fetal thymic and liver tissues. In the bone marrow (Fig. 1e), irradiated groups had higher percentages of human CD45+ cells compared to non-irradiated groups, although the difference in CD45 percentages for the non-irradiated recipients with or without thymic implants was not significant. CD34+/CD38-positive human HSC (Fig. 1f, expressed as a percentage of human CD45+ cells) were detectable in all groups of mice, with a slightly higher percentage in non-irradiated mice injected with HSC only. The increased percentage of CD34+ HSC in the bone marrow of non-irradiated mice injected with HSC only was attributed to the overall low levels of human CD45+ cells in the bone marrow. As described in Materials and methods, NSG recipient mice were injected with a range in number of CD34+ HSC (1 × 105–5 × 105), depending on cell recovery and number of mice implanted. To determine if this fivefold range influenced the levels of human cell engraftment, NSG mice that were either non-irradiated or irradiated and then implanted with human fetal thymic and liver tissues and HSC were evaluated for human CD45+ chimerism in the peripheral blood at 12 weeks (Supporting information, Fig. S1). Surprisingly, there was no correlation between the number of HSC-injected and levels of CD45+ cells in the peripheral blood, suggesting that the inherent variability in human cell chimerism between individual donor tissues is not overcome by a fivefold increase in HSC number for the BLT model. Together these results suggest that optimal human cell chimerism after implant of human HSC mice requires irradiation, but that a significant level of chimerism can be achieved by co-implantation of human thymic tissues in the absence of irradiation.

figure

Figure 1. Effects of irradiation and implantation of human thymic tissues in the establishment of human CD45+ cell chimerism in non-obese diabetic (NOD)-scid IL2rγnull (NSG) mice. NSG mice were irradiated with 200 cGy or not irradiated (0 cGy) and mice from each group were then implanted with 1 mm3 fragments of human fetal thymus and liver in the renal subcapsular space (thymic implant) or left unmanipulated (no thymic implant). All mice were then injected intravenously with 1 × 105 to 5 × 105 CD34+ haematopoietic stem cells (HSC) derived from the autologous human CD3-depleted fetal liver. The peripheral blood of recipient NSG mice was screened for human CD45+ cell chimerism at 12 weeks after implant (a). At 16 weeks after implant human cell chimerism was determined in the blood (b), spleen (c,d) and bone marrow (e). The bone marrow was also evaluated for the presence of human HSC (CD34+/CD38) and the values are shown as a percentage of total human CD45+ cells (f). *P < 0·05; **P < 0·01; ***P < 0·001; ****P < 0·0001.

Download figure to PowerPoint

In addition, we have compared the levels of human CD45+ cells at 12 weeks in the peripheral blood of female or male NSG mice that were irradiated and implanted with fetal thymus and liver tissues and HSC (standard BLT mice) from either male or female donors (Supporting information, Fig. S2). The data show that tissues from both male and female donors engraft NSG mice effectively. Moreover, for five of eight sets of tissues, female NSG recipients engrafted at slightly higher levels with human CD45+ cells compared to NSG male mice, as described previously for human umbilical cord blood-derived HSC [60]. This preferential engraftment of female mice was evident for tissues from both female and male donors.

T cell development in NSG–BLT mice does not require irradiation

The presence of human thymic tissue within the BLT model allows for high-level development of human T cells following injection of HSC [21, 59]. We next evaluated the importance of host mouse irradiation on T cell development in either NSG mice injected with human HSC only or in NSG mice implanted with human thymic and liver tissues and injected with autologous HSC. In irradiated NSG mice that were injected with HSC only, human CD3+ T cells were detectable in the blood at 12 weeks (Fig. 2a) and in the blood (Fig. 2b) and spleen (Fig. 2c,d) at 16 weeks. Irradiation was required for T cell development in NSG mice injected with HSC, with only very low levels of human CD3+ cells detected in non-irradiated mice in the absence of a thymus implant. In contrast, human T cell development was not significantly different between non-irradiated and irradiated HSC-engrafted NSG mice that were implanted with human thymic tissue. Moreover, human thymic tissues recovered from non-irradiated and irradiated NSG mice showed no structural differences by H&E (Fig. 2e,f) or human CD45 staining (Fig. 2g,h). Slightly higher numbers of human CD45+ cells were recovered from thymic tissues of irradiated NSG mice at 12 weeks compared to non-irradiated mice (Supporting information, Fig. S3a), but the proportions of CD4 and CD8 single-positive thymocytes and double-positive thymocytes were similar (Supporting information, Fig. S3b). In all groups of mice that developed detectable levels of human CD3+ T cells, CD4 T cells were present at higher levels compared to CD8 T cells (Fig. 2i,j). We also evaluated if the number of CD34+ HSC injected influenced the levels of human T cells developing in the periphery. For this, NSG mice that were either non-irradiated or irradiated and then implanted with human fetal thymic and liver tissues and HSC were evaluated for human CD3+ T cells in the peripheral blood at 12 weeks (Supporting information, Fig. S1b,d). As seen with human CD45+ levels, there was no correlation between the number of HSC-injected and levels of human T cells in peripheral blood.

figure

Figure 2. Irradiation is not required for human T cell development in haematopoietic stem cells (HSC)-engrafted non-obese diabetic (NOD)-scid IL2rγnull (NSG) mice implanted with human thymic tissues. NSG mice were irradiated with 200 cGy or not irradiated (0 cGy) and mice from each group were then implanted with 1 mm3 fragments of human fetal thymus and liver in the renal subcapsular space (thymic implant) or left unmanipulated (no thymic implant). All mice were then injected intravenously with 1 × 105 to 5 × 105 CD34+ HSC derived from the autologous human CD3 depleted fetal liver. The peripheral blood of recipient NSG mice was screened for human CD3+ T cell levels (shown as the percentage of human T cells among CD45+ cells) at 12 weeks after implant (a). At 16 weeks after implant the level of human CD3+ T cells was determined in the blood (b) and spleen (c,d). Thymus/liver grafts in non-irradiated NSG mice (e,g) or NSG mice irradiated with 200 cGy (f,h) were evaluated by haematoxylin and eosin (H&E) staining (e,f) and by immunostaining for human CD45 (g,h). The original magnification was ×100. The ratio of CD4+ T cells to CD8+ T cells is shown for the blood (i) and the spleen (j) at 16 weeks. *P < 0·05; **P < 0·01; ***P < 0·001; ****P < 0·0001.

Download figure to PowerPoint

To determine if irradiation influences the activation status of human T cells developing in HSC-engrafted mice, the expression of CD45RA was examined on human CD4+ and CD8+ cells in the blood at 12 and 16 weeks and in the spleen at 16 weeks (Supporting information, Fig. S4). CD45RA expression levels are not shown for mice injected with human HSC in the absence of irradiation due to the extremely low levels of T cell development. For NSG mice implanted with human thymic tissues and injected with HSC, irradiation did not change the CD45RA expression levels significantly on human CD4 and CD8 T cells in the peripheral blood (Supporting information, Fig. S4a,b,d,e) and spleen (Supporting information, Fig. S4c,f) compared to mice that did not receive irradiation. Interestingly, T cells from NSG mice that were irradiated and injected with HSC only were consistently lower in the expression of CD45RA compared to mice also implanted with thymic tissues, consistent with a recently published study [21], suggesting that the development of human T cells on human thymic tissue helps to maintain a naive phenotype of human T cells. Representative flow plots displaying CD45RA and CD62L staining of human CD4 (Supporting information, Fig. S4g,h) and CD8 (Supporting information, Fig. S4i,j) are shown from mice that were implanted with fetal thymus and liver tissues and injected with autologous HSC in the absence or presence of irradiation. In addition, we examined the ability of human CD4 and CD8 T cells from NSG mice implanted with human thymic and liver tissues and injected with autologous HSC to produce cytokines following an in-vitro polyclonal stimulation with PMA and ionomycin (Supporting information, Fig. S5). CD4 T cells from mice that received no irradiation or 200 cGy were able to produce IFN-γ, IL-2, IL-17A and IL-22, with slightly higher levels of IL-2-producing CD4 T cells detected in mice that were not irradiated. IFN-γ and IL-2-producing CD8 T cells were detectable from both groups of mice. Higher levels of CD8 T cell-producing IFN-γ were detectable in the 200 cGy group, and higher levels of IL-2-producing CD8 T cells were detected in the 0 cGy group. Together, these data indicate that the implantation of human thymic tissue into NSG mice supports high levels of T cell development in the absence of irradiation following injection of autologous HSC.

Human B cells develop in NSG–BLT mice in the absence of irradiation but their ability to produce human Ig is reduced

Human B cells develop in the standard BLT model, and these cells are functional, producing antigen-specific Ig following viral infections [24, 38]. We therefore evaluated the importance of irradiation for B cell development and function in either NSG mice injected with human HSC only or NSG mice implanted with human thymic and liver tissues and injected with autologous HSC. CD20+ B cells accounted for a large proportion of the human CD45+ cells in the blood at 12 weeks (Fig. 3a) and in the blood (Fig. 3b) and spleen (Fig. 3c) at 16 weeks in NSG mice that were injected with HSC only. In HSC-engrafted NSG mice that were implanted with human thymic tissues, the percentages of human B cells in the blood and spleen were not significantly different between mice that were non-irradiated versus irradiated. However, there was a significant decrease in the total number of human B cells in spleen of mice that did not receive irradiation (Fig. 3d). To assess the overall functionality of the human B cells, the levels of human IgM and IgG present in the serum of engrafted mice were determined at 12 weeks. NSG mice that received irradiation had significantly higher levels of human IgM compared to mice that were not irradiated (Fig. 3e). Human IgG levels were detected at very low levels in all groups of mice (Fig. 3f), and this is consistent with other studies using BLT mice [37, 38]. To determine if irradiation influences the maturation of human B cell subsets, we used lineage-specific markers to define immature/transitional (CD10+/CD27/CD38+/IgD), transitional (CD10/CD27/CD38/IgDdim), naive (CD10/CD27/CD38/IgD+) and memory (CD10/CD27+) CD20+ B cells in the blood and spleen of NSG mice that have been implanted with fetal thymic and liver tissues and injected with HSC (Supporting information, Fig. S6). The gating strategy used to define the human B cell subsets is shown in Supporting information, Fig. S6a. Irradiation of the recipient did not alter the maturation of human B cell populations, with similar proportions of each subset detected in blood and spleen of NSG mice implanted with fetal thymic and liver tissues and injected with autologous HSC (Supporting information, Fig. S6b–e). In addition, B cell subsets developing in the NSG–BLT mice were compared to the populations in human blood. As described previously, there are higher levels of immature and transitional B cells in the blood of NSG–BLT mice compared to humans [37]. Together, these results suggest that irradiation is not necessary for B cell development but is required to obtain optimal number of B cells and for Ig production.

figure

Figure 3. Irradiation is not required for human B cell development in haematopoietic stem cells (HSC)-engrafted non-obese diabetic (NOD)-scid IL2rγnull (NSG) mice implanted with human thymic tissues but enhances immunoglobulin (Ig) production. NSG mice were either irradiated with 200 cGy or not irradiated (0 cGy) and mice from each group were then implanted with 1 mm3 fragments of human fetal thymus and liver in the renal subcapsular space (thymic implant) or left unmanipulated (no thymic implant). All mice were then injected intravenously with 1 × 105 to 5 × 105 CD34+ HSC derived from the autologous human fetal liver. The peripheral blood of recipient NSG mice was screened for human CD20+ B cell levels (percentage of human CD45+ cells) at 12 weeks after implant (a). At 16 weeks after implant the level of human CD20+ B cells was determined in the blood (b) and spleen (c,d). Levels of human IgM (e) and human IgG (f) were quantified by enzyme-linked immunosorbent assay (ELISA) in plasma of engrafted NSG mice at 12 weeks. *P < 0·05; **P < 0·01; ***P < 0·001; ****P < 0·0001.

Download figure to PowerPoint

Human innate immune cells develop in NSG–BLT mice in the absence of host mouse irradiation

We next evaluated the development of human innate immune cells in the BLT model established with or without irradiation conditioning (Supporting information, Fig. S7). The gating strategy used to define the human innate immune subsets is shown in Supporting information, Fig. S7a. At 16 weeks post-implant the development of human monocyte/macrophage (CD14+/CD33+), myeloid dendritic cells (mDC, CD11c+/CD33+) and plasmacytoid DC (pDC, CD123+/CD33+) was assessed in the blood, spleen and bone marrow (Supporting information, Fig. S7b–d). Significantly higher percentages of human monocyte/macrophage were detected in the blood of NSG–BLT mice that had received irradiation compared to non-irradiated NSG–BLT mice, and there was a trend towards increased levels in the spleen and bone marrow, although these differences were not significant (Supporting information, Fig. S7b). The levels of mDC (Supporting information, Fig. S7c) and pDC (Supporting information, Fig. S7d) were similar in irradiated and non-irradiated NSG–BLT mice. In addition, innate cell subsets developing in the NSG–BLT mice were comparable to the populations in human blood. Together, these results suggest that irradiation conditioning of the recipient slightly enhances human macrophage development in NSG–BLT mice but is not necessary for mDC or pDC development.

Subcutaneous implant of human thymic tissue does not promote T cell development

The standard implantation site for thymic and liver fragments in the BLT model is within the subcapsular space of the kidney. However, this procedure is considered survival surgery for the mice and is labour-intensive. As an alternative to the renal capsule, we tested whether implantation of thymic and liver fragments subcutaneously would support high levels of T cell development. NSG mice were irradiated with 200 cGy, implanted with 1 mm3 fragments of human fetal thymus and liver either in the renal subcapsular space or subcutaneously, and then injected i.v. with human HSC derived from the fetal liver. At 18 weeks post-implant the mice were evaluated for total human cell chimerism (CD45+ cells), for human T cell development (CD3+ cells) and for human B cell development (CD20+) in the blood and spleen (Fig. 4a–c). No significant differences were detected for the percentage of CD45+ cells in the blood and spleen (Fig. 4a) and in the total number in the spleen (47·9 ± 18 × 106 for subcapsular and 37·9 ± 27 × 106, P = 0·467) between the groups. However, significantly higher levels of T cells were detected in NSG mice that were implanted in the renal subcapsular space of the kidneys compared to the subcutaneous site (Fig. 4b). No structural differences were observed between thymus tissues recovered from either site (Fig. 4d–k), although the size of the tissue recovered from the subcutaneous site was consistently smaller. Moreover, well-formed Hassall's corpuscles, a structure characteristic of human thymus, were detected readily within the thymic medulla of tissues recovered from either renal subcapsular or subcutaneous sites (Fig. 4e,i,g,k) [61]. Significantly higher levels of B cells were detected in NSG mice implanted in the subcutaneous site (Fig. 4c), although no significant differences were detected in human IgM and IgG in the plasma of mice from either group (Fig. 4l,m). These findings indicate that subcutaneous implantation of human fetal thymic tissues is less efficient than subrenal implantation for generation of human T cells in the BLT model.

figure

Figure 4. Subcutaneous implantation of human thymic tissue leads to reduced human T cell generation. Non-obese diabetic (NOD)-scid IL2rγnull (NSG) mice were irradiated with 200 cGy, implanted with 1 mm3 fragments of human fetal thymus and liver in the renal subcapsular space or subcutaneously. All implanted mice were then injected intravenously with 1 × 105 to 5 × 105 CD34+ haematopoietic stem cells (HSC) derived from the autologous human CD3-depleted fetal liver. The percentages of human CD45+ cells (a) CD3+ cells (b) and CD20+ cells in the blood and spleen is shown from mice at 18 weeks post-implant. Thymus/liver grafts in NSG mice implanted in the subcapsular space (d,e,h,i) or subcutaneously in mice irradiated with 200 cGy (f,g,j,k) were evaluated by haematoxylin and eosin (H&E) staining (d,e,f,g) and by immunostaining for human CD45 (h,i,j,k) at 18 weeks post-implant. The original magnifications were either ×15 (d,h,f,j) or ×630 (e,i,g,k). Levels of human IgM (l) and human IgG (m) were quantified by enzyme-linked immunosorbent assay (ELISA) in plasma of engrafted NSG mice at 12 weeks. **P < 0·01.

Download figure to PowerPoint

BLT mice maintain high levels of human cell chimerism for 3–4 months but the human immune system undergoes spontaneous activation

To evaluate the long-term maintenance of human cell chimerism in BLT mice, NSG mice were irradiated (200 cGy), implanted with human thymic and liver tissues and injected with human HSC as described in Materials and methods. Between 26 and 28 weeks post-implant, NSG–BLT mice were screened for total human cell chimerism (CD45+ cells) for human T cell (CD3+ cells) and B cell (CD20+ cells) development in the blood and spleen (Fig. 5a–c). Human leucocyte levels were very high in mice that had been engrafted for greater than 25 weeks. However, both T and B cells were transitioning to an activated phenotype at these later time-points. For example, there was a significant decrease in the percentage of CD45RA+ CD4 and CD8 T cells in the blood at 26 weeks compared to 12 weeks (Fig. 5d). CD45RA is not expressed exclusively by naive T cells, but still provides a reliable estimation of the activation status [62]. In the spleen of BLT mice, the average percentage of CD45RA+ CD4 and CD8 T cells was less than 60% between 26 and 28 weeks after implant (Fig. 5e). Moreover, there was a significant increase in human IgM and IgG levels in plasma of BLT mice at 26 to 28 weeks after implant compared to 12 and 19 weeks (Fig. 5f,g). The activation of the human immune system also correlated with a decrease in platelet (PLT), red blood cell (RBC) and haemoglobin (HGB) values (Fig. 5h–j, respectively). Together these data suggest that human cell chimerism is maintained long term in BLT mice, but the human immune system becomes activated spontaneously.

figure

Figure 5. The engrafted human immune system in bone marrow, liver, thymus (BLT) mice undergoes spontaneous activation at late points after implant. Non-obese diabetic (NOD)-scid IL2rγnull (NSG) mice were irradiated with 200 cGy, implanted with 1 mm3 fragments of human fetal thymus and liver in the renal subcapsular space and then injected intravenously with 1 × 105 to 5 × 105 CD34+ haematopoietic stem cells (HSC) derived from the autologous human fetal liver. The percentages of human CD45+ cells (a), human CD3+ cells (b) and human CD20+ cells (c) in the blood and spleen is shown from mice at 26–28 weeks post-implant. At 12 and 26 weeks (d) human CD4+ and CD8+ T cells in the peripheral blood were examined for the expression of CD45RA. The values shown represent the percentage of human CD4+ or CD8+ T cells co-expressing CD45RA. The percentage of human CD4+ or CD8+ T cells expressing CD45RA in the spleen (e) is shown at 26–28 weeks post-implant. Levels (ng/ml) of human immunoglobulin (Ig)M (f) and human IgG (g) were quantified by enzyme-linked immunosorbent assay (ELISA) in plasma of engrafted NSG mice at the indicated times after implant. Haematological analyses (h, platelets; i, red blood cells; j, haemoglobin) of blood from NSG–BLT mice were performed over time as described in Materials and methods. *P < 0·05; **P < 0·01; ***P < 0·001; ****P < 0·0001.

Download figure to PowerPoint

Development of xeno-GVHD in NSG–BLT mice and the potential role of immune tolerance

NSG–BLT mice support the human immune system engraftment for an extended time-frame; however, these animals have been reported to develop a xeno-GVHD late after implant [54]. At approximately week 20 post-implant, NSG–BLT mice generated in our laboratory displayed a significantly increased rate of mortality compared to NSG mice that were only irradiated (P = 0·026, Fig. 6a). Survival levels of NSG–BLT mice were 51·1% (24 of 47 mice surviving) by 28 weeks post-implant compared to 86·7% (14 of 16 mice surviving) survival of irradiated-only control NSG mice that did not receive human tissues. We next evaluated if the number of CD34+ HSC injected influenced the incidence of xeno-GVHD in NSG–BLT mice, as indicated by the time of death. NSG mice that were irradiated and then implanted with human fetal thymic and liver tissues and injected with the indicated number of CD34+ HSC were monitored for survival over 200 days (Supporting information, Fig. S8a). The data show that there is no correlation between the number of CD34+ HSC injected and the incidence of xeno-GVHD. In addition, we found no correlation between the percentages of CD3+ T cells in the peripheral blood of NSG–BLT mice at 12 weeks and incidence of xeno-GVHD (Supporting information, Fig. S8b). We also found no differences in the incidence of xeno-GVHD between NSG–BLT mice implanted with female and male tissues (Supporting information, Fig. S8c). The decrease in naive phenotype human CD4 and CD8 T cells in older NSG–BLT mice (Fig. 5) suggests that these T cells are being activated and mediating a xenogeneic GVHD. We hypothesized that the development of xeno-GVHD in NSG–BLT mice might result from a lack of negative selection against murine antigens in the human thymus or by a lack of peripheral regulation. Our previous studies showed that the xenogeneic GVHD occurring after the injection of human PBMC into NSG mice is mediated by T cell recognition of murine MHC (H2) classes I and II [55, 56]. To test if H2-reactive human T cells escape negative selection and contribute to the mortality of older NSG–BLT mice, NSG mice lacking the expression of murine MHC class I [NSG-(KbDb)null] or class II (NSG-Abo), were used to engraft fetal thymic and liver tissues. NSG-(KbDb)null and NSG-Abo BLT mice did not have increased overall survival compared to standard NSG–BLT mice (Fig. 6a). Unexpectedly, the survival of engrafted NSG-(KbDb)null mice was reduced significantly compared to NSG–BLT mice (P < 0·001, Fig. 6a). Human cell chimerism (huCD45+ cells) was compared in the blood at 12, 16 and 20 weeks in NSG mice, NSG-(KbDb)null and NSG-Abo mice (Fig. 6b). Human CD45+ cell chimerism was comparable in the three NSG strains. Together, these data suggest that elimination of either murine class I or murine class II is not sufficient to overcome the mortality of older NSG–BLT mice.

figure

Figure 6. Evaluating the role of central and peripheral tolerance in the development of xeno-graft-versus-host disease (GVHD) in non-obese diabetic (NOD)-scid IL2rγnull-bone marrow, liver, thymus (NSG–BLT) mice. (a,b) NSG, NSG-Abo or NSG-(KbDb)null mice were irradiated with 200 cGy, implanted with 1 mm3 fragments of human fetal thymus and liver in the renal subcapsular space and then injected intravenously with 1 × 105 to 5 × 105 CD34+ haematopoietic stem cells (HSC) derived from the autologous human fetal liver. An additional group of NSG received irradiation only (200 cGy). The survival of engrafted NSG (n = 47), NSG-Abo (n = 12) and NSG-(KbDb)null mice (n = 12) and survival of NSG mice that received irradiation only (n = 15) mice was monitored over the next 28 weeks (a). The graph shown in (a) represents the percentage survival over time. The percentages of human CD45+ cells in the blood are shown from mice at 12, 16 and 20 weeks post-implant (b). (c,d,e) Alternatively, a cohort of NSG–BLT mice was generated by the implant of 1 mm3 fragments of fetal NSG mouse liver (fml) in addition to the human fetal thymus and liver tissues. The thymic organoids from NSG–BLT mice and NSG–BLT mice implanted with fml were evaluated by immunohistochemistry for the presence of mouse cells by staining with monoclonal antibodies specific for mouse CD45 (c). The original magnification was ×100. The percentages of human CD45+ cells (d) in the blood are shown from mice at 12 weeks post-implant. The survival of NSG–BLT mice (n = 14) and NSG–BLT mice implanted with fetal mouse liver (n = 8) was monitored over 36 weeks (e). (f) The percentage of CD4+ human regulatory T cells (Treg) [CD25+/CD127dim/forkhead box protein 3 (FoxP3+)] levels was monitored in the blood of NSG–BLT mice over time. *P = 0·026 NSG–BLT compared to irradiated NSG; **P < 0·001 NSG–BLT compared to irradiated NSG. (g) Survival of NSG–BLT mice that were either non-irradiated (n = 10) or irradiated with 200 cGy (n = 10) was monitored over 30 weeks (g). ****P < 0·0001.

Download figure to PowerPoint

We next compared the engraftment and survival of NSG–BLT mice to BLT mice that were co-implanted under the renal capsule with 1 mm3 fragment of fetal mouse liver (fml) and human thymic tissue, in an attempt to enhance negative selection against murine antigens. Co-implant of fml did not increase the proportion of mouse cells (murine CD45+ staining) detected within human thymic organoid (Fig. 6c). Overall engraftment in the blood of both sets of mice was similar at 12 weeks after implant (Fig. 6d), but overall survival was not enhanced by co-implant of the fml (Fig. 6e). To determine if xeno-GVHD resulted from a loss of peripheral tolerance, we evaluated the levels of human Treg detectable in the blood of standard NSG–BLT mice (with irradiation) over time (Fig. 6f). The percentage of CD25+/CD127dim/FoxP3+ cells in the blood of NSG–BLT mice did not decrease over time. To determine the contribution of irradiation in the development of xeno-GVHD in BLT mice, we compared the survival of NSG–BLT mice that were either irradiated or non-irradiated (Fig. 6g). Overall, there was an increased survival of non-irradiated NSG–BLT mice; however, these animals ultimately developed GVHD-like symptoms.

Discussion

  1. Top of page
  2. Summary
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. Disclosure
  9. References
  10. Supporting Information

The BLT mouse, also referred to as the Thy/Liv mouse, is an ideal model to study human immune and T cell functions, as the implant of human thymic tissues and autologous human HSC enable the efficient development of HLA-restricted human CD4 and CD8 T cells [63]. Following implantation into the subcapsular renal space, the human fetal thymus grows significantly, is populated with a normal distribution of human thymocyte subsets and allows high levels of human T cells to repopulate the peripheral lymphoid tissues [21-23]. The BLT model is based on the severe compromised immunodeficient-humanized (SCID-hu) mouse described by McCune and colleagues [6]. The original SCID-hu model was created using CB17-scid mice and involved the transplant of human fetal thymic tissues in the renal subcapsular space and i.v. injection of autologous or allogeneic HSC derived from the fetal liver. The SCID-hu mouse enabled the development of human T cells, which required both the implant of thymic tissues and injection of HSC. However, in CB17-scid mice the persistence of human T cells in the peripheral tissues was transient, as CD3+ cells were not detectable in the peripheral blood at 12 weeks post-implant and the ability of these cells to mediate an immune response was limited [64]. The persistence and functionality of human T cells was improved significantly by the use of NOD-scid mice as recipients of human thymic and liver tissues [22, 23]. However, engraftment of fetal thymic and liver tissues into NSG mice enhances human cell chimerism significantly, including reconstitution of a mucosal immune system, compared to other mouse strains [17, 65]. Continued improvement of the NSG mouse by the transgenic expression of human-specific cytokines and growth factors and expression of HLA that will allow matching with the donor tissues will further augment the development of human immune systems in BLT mice [3, 66]. In an effort to provide an analysis of optimal parameters for establishing the NSG–BLT model, we have assessed the requirement for irradiation to attain high-level human cell chimerism, the optimal implantation sites for thymic tissues, the stability of human cell chimerism and the longevity of engrafted mice.

Low-dose irradiation is a commonly used pre-conditioning regimen to promote engraftment of human HSC in immunodeficient mice and has been incorporated into the standard protocol for generating BLT mice [22, 23, 67]. One mechanism by which irradiation is thought to enhance HSC engraftment is by stimulating the release of factors that improve the homing and survival of stem cells such as stem cell factor (SCF) [63] and SDF-1 [68]. However, total body irradiation has a number of negative consequences, including stunting growth and impairing neuronal function [19, 69]. Recent work from our laboratory and others have demonstrated that both adult and newborn NSG mice will support human HSC engraftment in the absence of irradiation [69, 70]. Moreover, the transgenic expression of human SCF improves human HSC engraftment significantly in non-irradiated NSG mice [69]. In this study we show that irradiation is not essential for the human immune system development in NSG–BLT mice, although irradiation increases levels of human chimerism. One significant difference for non-irradiated NSG–BLT mice was the lower level of human IgM detected in the serum compared to NSG–BLT mice that were preconditioned with irradiation. The reduced levels of IgM may be attributed to the slightly reduced levels of human B cells in the spleens of non-irradiated NSG–BLT mice. To allow for complete analysis of the engraftment data, we have also presented the human cell chimerism levels shown in Figs 1-3 (human CD45+, human CD3+ T cells and human CD20+ B cells) for each unique set of human fetal tissues (Supporting information, Fig. S9).

The NSG–BLT mouse has sustained high levels of human cell chimerism and T cells in the peripheral lymphoid tissues. However, many NSG–BLT mice succumb ultimately to a GVHD-like syndrome [54] which has also been reported for BLT mice generated on the NOD-scid background [26]. The development of the delayed GVHD-like syndrome in NSG–BLT mice correlated with the transition of human T cells to an activated phenotype and increased levels of human IgM and IgG in the serum. This late, spontaneous activation of the human immune systems suggests that a peripheral tolerance mechanism is abrogated as NSG–BLT mice age, and this loss of tolerance allows the human immune system to respond to the murine host. T cells are a primary effector population mediating tissue damage during classic GVHD [71], and the high levels of human T cell chimerism in the NSG–BLT mice suggest that these cells are key mediators of the disease pathology. Our data show that the development of GVHD in NSG–BLT mice does not require the expression of murine MHC classes I or II, indicating that either human CD4 or CD8 T cells or both probably mediate GVHD, or that murine MHC classes I or II are not necessary for disease development. We are initiating studies to evaluate further the mechanism mediating GVHD in NSG–BLT mice by generating NSG mice that lack both murine classes I and II and by the depletion of human T cell subsets at precise time-points. Unexpectedly, survival of engrafted NSG-(KbDb)null mice was reduced significantly compared to NSG–BLT mice and studies are ongoing to characterize the disease process in these mice. Interestingly, a recent study has proposed that GVHD developing in immunodeficient mice implanted with thymic tissues and human HSC is a result of mature thymocyte populations residing within the thymic tissues that are not tolerant to the murine host and expand following emigration to the periphery [26]. In this study, the development of GVHD in NSG recipient mice was minimized with depletion of thymocyte populations by using thymic tissues that were initially cryopreserved and then thawed prior to implant and by the treatment of mice with a monoclonal antibody to human CD2. However, implanted NSG mice were followed only for 20 weeks post-implant for the development of disease, and it remains to be determined whether this treatment approach will reduce the late-onset GVHD that our results show develops after 20 weeks.

The onset of xeno-GVHD in NSG–BLT mice may be a direct result of a breakdown in tolerance mechanisms [72]. It is possible that the levels of mouse cells within the human thymic organoid are not sufficient to enable the negative selection of human T cells that are reactive with mouse MHC (H2). This would result in the development of mature human T cells that recognize mouse MHC as a xeno-antigen and ultimately mediate a GVHD. Our data show that co-implantation of mouse fetal liver with the human thymic tissues was insufficient to prevent or delay the onset of GVHD in NSG–BLT mice. Interestingly Hassall's corpuscles were readily detectable within the BLT thymic organoid. Hassall's corpuscles are typical of human thymic tissue, and the presence of these structures in the medulla suggests that the BLT thymus is developing a normal architecture [73]. Moreover, Hassall's corpuscles have been proposed to be critical for supporting the development of thymic dendritic cells, which induce the differentiation of human Treg [61]. CD4+/CD25+/FoxP3+/CD127low human Treg are detectable in the periphery of BLT mice [31], and our data show that development of GVHD in NSG–BLT mice was not associated with a decline in peripheral human Treg numbers. We are currently comparing the functionality of human Treg from younger and older NSG–BLT mice to determine if the onset of GVHD can be correlated with a loss in Treg function. An additional parameter that may influence the development of GVHD in NSG mice implanted with fetal thymic and liver tissues may be the use of antibiotics in the drinking water, which may change the microbiota of the mice and alter immune regulation [74].

In summary, we have shown that: (i) irradiation is not required for human immune system development in NSG–BLT mice, (ii) subcutaneous implant of thymic tissues is suboptimal for T cell development in NSG–BLT mice and (iii) human cell chimerism is sustained long term in NSG–BLT mice but that many of these mice will develop a GVHD-like syndrome at later time-points. The BLT mouse has become widely used to study human immunobiology, and the findings presented here highlight important parameters for the generation of this model and its use. Overall, our data indicate that optimal human cell engraftment of BLT mice requires subrenal implant of thymic tissues and low-dose irradiation. However, reasonable engraftment levels can be achieved in the absence of irradiation, and these BLT mice have an extended life span. Importantly, our study underscores the importance for considering the duration of experiments when using NSG–BLT mice, as these animals develop an activated human T cell population after 20 or more weeks post-implant in most cohorts.

Acknowledgements

  1. Top of page
  2. Summary
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. Disclosure
  9. References
  10. Supporting Information

We thank Jamie Kady, Meghan Dolan, Pamela St Louis, Linda Paquin, Michael Bates, Bruce Gott, Allison Ingalls, Michelle Farley and Rebecca Riding for excellent technical assistance. This work was supported by National Institutes of Health research grants AI046629 and DK032520, an institutional Diabetes Endocrinology Research Center (DERC) grant DK32520, a grant from the University of Massachusetts Center for AIDS Research, P30 AI042845 and grants from the Juvenile Diabetes Research Foundation, International and the Helmsley Charitable Trust. The contents of this publication are solely the responsibility of the authors and do not necessarily represent the official views of the National Institutes of Health.

Disclosure

  1. Top of page
  2. Summary
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. Disclosure
  9. References
  10. Supporting Information

Michael A. Brehm is a consultant for The Jackson Laboratory. No other authors have conflicts of interest to declare.

References

  1. Top of page
  2. Summary
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. Disclosure
  9. References
  10. Supporting Information

Supporting Information

  1. Top of page
  2. Summary
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. Disclosure
  9. References
  10. Supporting Information
FilenameFormatSizeDescription
cei12180-sup-0001-si.docx2001K

Fig. S1. Influence of the number of injected human CD34+ haematopoietic stem cells (HSC) on human cell chimerism in non-obese diabetic (NOD)-scid IL2rγnull- bone marrow, liver, thymus (NSG–BLT) mice. NSG mice were irradiated with 200 cGy (a,b) or non-irradiated (c,d) were implanted with 1 mm3 fragments of human fetal thymus and liver in the renal subcapsular space and then injected intravenously with the indicated number of CD34+ HSC derived from the autologous human CD3-depleted fetal liver. The peripheral blood of recipient NSG mice was screened for human CD45+ cell chimerism (a,c) and development of human CD3+ T cells (b,d) at 12 weeks after implant. Each point shown represents an individual mouse.

Fig. S2. Engraftment levels of human CD45+ cells in female or male non-obese diabetic (NOD)-scid IL2rγnull (NSG) mice implanted with tissues from either male or female donors. Male or female NSG mice were irradiated with 200 cGy, implanted with 1 mm3 fragments of human fetal thymus and liver in the renal subcapsular space and then injected intravenously with 1 × 105 to 5 × 105 CD34+ haematopoietic stem cells derived from the autologous human CD3-depleted fetal liver cells. Tissues both male (a) and female donors (b) were used. The peripheral blood of recipient NSG mice was screened for human CD45+ cell chimerism at 12 weeks after implant. Each letter on the x-axis represents a unique set of donor tissues and each point represents an individual mouse. *P < 0·05; **P < 0·01; ***P < 0·001.

Fig. S3. Thymocyte populations from non-obese diabetic (NOD)-scid IL2rγnull- bone marrow, liver, thymus (NSG–BLT) not irradiated and mice from each group were then implanted with 1 mm3 fragments of human fetal thymus and liver in the renal subcapsular space. All mice were then injected intravenously with 1 × 105 to 5 × 105 CD34+ haematopoietic stem cells derived from the autologous human CD3-depleted fetal liver. At 12 weeks post-implant, thymic tissues were recovered and the total number of CD45+ cells (a) and the proportion of CD4 and CD8 single-positive and double-positive cells (b) were determined using flow cytometry. **P < 0·001.

Fig. S4. Irradiation does not alter the activation status of human T cells in haematopoietic stem cells-engrafted non-obese diabetic (NOD)-scid IL2rγnull (NSG) mice implanted with human thymic tissues. NSG mice were irradiated with 200 cGy or not irradiated (0 cGy) and mice from each group were then implanted with 1 mm3 fragments of human fetal thymus and liver in the renal subcapsular space (thymic implant) or left unmanipulated (no thymic implant). All mice were then injected intravenously with 1 × 105 to 5 × 105 CD34+ haematopoietic stem cells derived from the autologous human CD3-depleted fetal liver. Human CD4+ T cells (a,b,c) and CD8+ T cells (d,e,f) were examined for the expression of CD45RA in the peripheral blood at 12 (a,d) and 16 (b,e) weeks and in the spleen at 16 weeks (c,f). The values shown represent the percentages of human CD4+ or CD8+ T cells expressing CD45RA. Data from NSG mice injected with human HSC in the absence of irradiation is not shown due to the very low levels of T cell development. Representative flow cytometry histograms for expression of CD45RA and CD62L on CD4+ (g,h) and CD8+ (i,j) T cells is shown for mice implanted with human fetal thymus and liver tissues. *P < 0·05; **P < 0·01; ***P < 0·001; ****P < 0·0001.

Fig. S5. Human CD4 and CD8 T cells from non-obese diabetic (NOD)-scid IL2rγnull-bone marrow, liver, thymus (NSG–BLT) mice produce cytokines following in-vitro stimulation.

NSG mice were either irradiated with 200 cGy or not irradiated and mice from each group were then implanted with 1 mm3 fragments of human fetal thymus and liver in the renal subcapsular space. All mice were then injected intravenously with 1 × 105 to 5 × 105 CD34+ haematopoietic stem cells derived from the autologous human CD3-depleted fetal liver. The ability of human CD4 T cells (a,c,e,g) and human CD8 T cells (b,d,f,h) from the spleens of mice from each group to produce interferon (IFN)-γ (a,b), interleukin (IL)-2 (c,d), IL-17A (e,f) and IL-22 (g,h) was determined at 12 weeks after tissue implant. Splenocytes were stimulated ex vivo with phorbol myristate acetate (PMA) and ionomycin for 5 h in a standard intracellular cytokine assay, as described in Materials and methods. *P < 0·05; ***P < 0·001.

Fig. S6. Irradiation does not alter human B cell maturation in non-obese diabetic (NOD)-scid IL2rγnull-bone marrow, liver, thymus (NSG–BLT) mice. NSG mice were either irradiated with 200 cGy or not irradiated (0 cGy) and mice from each group were then implanted with 1 mm3 fragments of human fetal thymus and liver in the renal subcapsular space. All mice were then injected intravenously with 1 × 105 to 5 × 105 CD34+ haematopoietic stem cells derived from the autologous human CD3-depleted fetal liver. Human B cell subsets were defined as follows: immature/transitional (CD10+/CD27/CD38+/IgD), transitional [CD10/CD27/CD38+/immunoglobulin (Ig)Ddim], naive (CD10/CD27/CD38/IgD+) and memory (CD10/CD27+) CD20+ B cells. The gating strategy used to identify the human B cell subsets is shown in (a). The proportion of immature/transitional (b), transitional (c), naive (d) and memory (e) CD20+ B cells is shown for the blood and spleen at 16 weeks post-implant and for human blood. *P < 0·05; **P < 0·01; ****P < 0·0001.

Fig. S7. Irradiation does not alter human innate immune cell development in non-obese diabetic (NOD)-scid IL2rγnull-bone marrow, liver, thymus (NSG–BLT) mice. NSG mice were irradiated with 200 cGy or not irradiated (0 cGy) and mice from each group were then implanted with 1 mm3 fragments of human fetal thymus and liver in the renal subcapsular space. All mice were then injected intravenously with 1 × 105 to 5 × 105 CD34+ haematopoietic stem cells derived from the autologous human CD3-depleted fetal liver. Human innate immune cell subsets were defined as follows: macrophage (CD14+/CD33+), myeloid dendritic cells (mDC, CD11c+/CD33+) and plasmacytoid dendritic cells (DC) (pDC, CD123+/CD33+). The gating strategy used to identify the human innate subsets is shown in (a). The proportion of monocyte/macrophage (b), mDC (c) and pDC (d) is shown for the blood, spleen and bone marrow at 16 weeks post-implant and for human blood. **P < 0·01; ***P < 0·001.

Fig. S8. Influence of the number of injected human CD34+ haematopoietic stem cells (HSC) and T cell levels on the incidence of xeno-graft-versus-host disease (GVHD) in non-obese diabetic (NOD)-scid IL2rγnull-bone marrow, liver, thymus (NSG–BLT) mice. NSG mice were irradiated with 200 cGy and implanted with 1 mm3 fragments of human fetal thymus and liver in the renal subcapsular space and then injected intravenously with the indicated number of CD34+ HSC derived from the autologous human CD3-depleted fetal liver. (a) NSG–BLT mice were monitored for survival and the day of death compared to the number of injected HSC is shown. (b) The peripheral blood of recipient NSG mice was screened for development of human CD3+ T cells at 12 weeks after implant and compared to the day of death. (c) The incidence of GVHD was also compared for male NSG mice engrafted with either female or male donor tissues. Each point shown represents an individual mouse. Survival was monitored over 200 days after implant.

Fig. S9. Comparison of human chimerism levels in individual cohorts of non-obese diabetic (NOD)-scid IL2rγnull(NSG) mice implanted with fetal tissues from the same donors.

NSG mice were irradiated with 200 cGy or not irradiated (0 cGy) and mice from each group were then implanted with 1 mm3 fragments of human fetal thymus and liver in the renal subcapsular space (thymic implant) or left unmanipulated (no thymic implant). All mice were then injected intravenously with 1 × 105 to 5 × 105 CD34+ haematopoietic stem cells derived from the autologous human CD3-depleted fetal liver. At 12 weeks (a,b,c) and 16 weeks (d,e,f) after implant, the peripheral blood of recipient NSG mice was screened for human CD45+ cell chimerism (a,d), T cell development (b,e) and B cell development (c,f). Each colour represents a unique set of donor tissues, and each symbol type indicates the specific implant protocol used to generate the mice. Each point represents an individual mouse.

Please note: Wiley Blackwell is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.