MNT suppresses T cell apoptosis via BIM and is critical for T lymphomagenesis

MNT loss in lymphoid cells reduces competitive fitness

To avoid embryonic lethality conferred by MNT loss [15], we deployed the Rag1Cre transgene, which is expressed only in early lymphoid progenitors [22]. This strategy enabled us to compare MNT-deficient vs normal lymphopoiesis in adult mice, using competitive bone marrow reconstitution. Lethally irradiated C57BL/6-Ly5.1⁺ mice were injected with a 50:50 mixture of Ly5.1⁺ WT cells and test Ly5.2⁺ Mnt^fl/flRag1Cre or Mnt^+/+Rag1Cre bone marrow cells (Fig. 1A). Analysis by flow cytometry after 12 weeks (Supplementary Fig. S1) revealed that the bone marrow cells from Mnt^fl/flRag1Cre mice had competed poorly against WT cells in regenerating lymphoid populations compared to those from Mnt^+/+Rag1Cre mice. Thymi displayed a significantly lower proportion of Ly5.2⁺ Mnt^fl/flRag1Cre cells (gold bars) than Ly5.2⁺ Mnt^+/+Rag1Cre cells (brown bars) in all major thymic sub-populations (Fig. 1B). Similarly, the spleen of reconstituted mice contained significantly fewer Mnt^fl/flRag1Cre than Mnt^+/+Rag1Cre CD4⁺ or CD8⁺ T cells (Fig. 1C). Comparable outcomes were noted for B lineage cells (CD19⁺) in the spleen and bone marrow (Fig. 1C, D). In contrast, as anticipated, Ly5.2⁺ Mnt^+/+Rag1Cre and Ly5.2⁺ Mnt^fl/flRag1Cre myeloid cells (Mac1⁺), were present in comparable numbers in the spleen and bone marrow in competitively reconstituted mice. We conclude that MNT loss puts both T and B lymphopoiesis at a significant competitive disadvantage.

A Protocol for competitive reconstitution. C57BL/6 mice (Ly5.1⁺) (3–4 per test) were lethally irradiated (2 × 5.5 Gy) and injected with a 1:1 mixture of bone marrow cells (10⁶ cells per genotype) from Ly5.1⁺ C57BL/6 mice (competitor cells) and either Mnt^+/+Rag1Cre or Mnt^fl/flRag1Cre mice (Ly5.2⁺ test cells). B–D Relative proportions of reconstituted test cells (Ly5.2⁺) and competitor cells (Ly5.1⁺) of the indicated cell types in the thymus (B), spleen (C) and bone marrow (D), 12 weeks post-transplantation. Analysis was performed by flow cytometry (see Supplementary Fig. S1). Data shown are from four independent reconstitution experiments; genotypes are WT (Ly5.1⁺, black), Mnt^+/+Rag1Cre (Ly5.2⁺, brown) and Mnt^fl/flRag1Cre (Ly5.2⁺, gold). Dots indicate individual reconstituted mice; bars show mean ± SEM; statistical significance is shown only for Mnt^+/+Rag1Cre versus Mnt^fl/flRag1Cre cells; ****P ≤ 0.0001, ns not significant.

Impact of MNT loss on T lymphopoiesis

We have previously investigated the impact of MNT loss in B lymphopoiesis [15]. To investigate how MNT loss affects T lymphopoiesis, we analysed the thymus and spleen in 6–7 wk-old Mnt^fl/flRag1Cre, Mnt^+/+Rag1Cre and WT mice. Mnt deletion mediated via the Rag1Cre transgene was very efficient, as shown by PCR and Western blot analysis (Fig. 2A).

A Mnt deletion mediated by Rag1Cre gene in lymphoid progenitors is highly efficacious. Typical PCR (left) and western blot (right) analysis of the indicated T cell sub-populations sorted from thymi of 6–7 wk-old WT and Mnt^fl/fl Rag1Cre (KO) mice. In left panel, *indicates primer dimer. B Mnt^fl/fl Rag1Cre mice exhibit modest T cell deficit. (L to R) Weight, cellularity and flow cytometric quantification of DN, DP, SP CD4 + and SP CD8⁺ T cells in thymi of 6–7 wk-old mice of indicated genotypes. C DN sub-population analysis (see Supplementary Fig. 1) reveals significant reduction of DN4 (CD4^–CD8^–CD25^–CD44^–) cells in Mnt^fl/fl Rag1Cre mice. D Splenic T and B cells, but not myeloid cells, are reduced in Mnt^fl/fl Rag1Cre mice. Cell preparations from thymi and spleens of 6–7 wk-old mice were analysed by flow cytometry (n = 4 to 21); genotypes shown are WT (black), Mnt^+/+Rag1Cre (brown) and Mnt^fl/flRag1Cre (gold); dots indicate individual mice; bars show mean ± SEM; *P ≤ 0.05, **P ≤ 0.01; ***P ≤ 0.001; ****P ≤ 0.0001, ns not significant.

Thymic weight and cellularity were reduced to ~65% of normal in Mnt^fl/flRag1Cre mice, primarily due to fewer DP T cells (P ≤ 0.001), although DN and SP CD4⁺ populations were also significantly reduced (Fig. 2B). DN4 (CD25^–CD44^–) cells were more affected than DN2 (CD25⁺CD44⁺) or DN3 (CD25⁺CD44^–) cells (Fig. 2C and Supplementary Fig. S2A). This deficit was not due to a failure of Τcrβ gene rearrangement because intracellular TCRβ protein was readily detectable in Mnt^fl/flRag1Cre DN4 cells (Supplementary Fig. S2B).

Spleen cellularity was also reduced in young Mnt^fl/flRag1Cre mice (Fig. 2D), primarily due to decreased B lymphoid cells (~37%; p < 0.0001), as reported previously [15]. In addition, T cells were reduced, particularly CD4⁺ T cells (~60%; p < 0.001), but the relative proportions of naïve, memory and effector T cells were equivalent between WT and Mnt-deficient T cells (Supplementary Fig. S2C, D). Myeloid (Mac1⁺) cellularity was unaffected (Fig. 2D), as expected from the lack of Rag1Cre expression in myeloid cells [22].

MNT loss increases T cell apoptosis

The T cell deficit in Mnt^fl/flRag1Cre mice seemed likely to reflect increased apoptosis and/or reduced MYC levels. All four major thymic sub-populations in Mnt^fl/flRag1Cre mice displayed a significantly increased proportion of annexin V-positive cells compared to their WT or Mnt^+/+Rag1Cre control counterparts (Fig. 3A, Supplementary Fig. S3) and there was a similar trend for CD4⁺ and CD8⁺ T cells in the spleen (Fig. 3E). However, MNT loss did not alter endogenous MYC protein levels in any of these cell populations, as shown by flow cytometric and immunoblot analysis (Fig. 3B–D).

A MNT loss increases apoptosis in all major thymic T cell sub-populations. B–D MNT loss does not alter MYC levels in thymic T cell sub-populations. B Typical intracellular MYC staining. Thymocytes were stained with antibodies against CD4 and CD8 and then fixed, permeabilised and stained with an antibody against MYC (pink) or Ig isotype-matched control antibody (blue). C Mean fluorescence intensity (MFI) for intracellular MYC ± SEM for the indicated thymocyte sub-populations. D Typical western blot showing MNT, MYC and ACTIN (protein loading control) protein levels in the indicated thymocyte sub-populations. Note the lower MYC levels in DP cells, a well-documented feature of this predominantly quiescent T lymphoid cell population [57]. E MNT loss has minor impact on apoptosis of splenic CD4⁺ and CD8⁺ T cells. Apoptosis was quantified by flow cytometry after staining cells with AnnexinV-FITC. Genotypes analysed were WT (black), Mnt^+/+Rag1Cre (brown) and Mnt^fl/flRag1Cre (gold). Dots indicate individual mice; bars show mean ± SEM; **P ≤ 0.01; ***P ≤ 0.001; ****P ≤ 0.0001; ns not significant.

Enhanced apoptosis probably also explains the reduced DN4 population in Mnt^fl/flRag1Cre mice (Fig. 2C). When sorted DN3 and DN4 cells were cultured on OP9-DL1 stromal cells with IL-7, conditions which are permissive for T lymphoid cell proliferation and differentiation (Fig. 4A), MYC levels and proliferation were unaffected (Supplementary Fig. S4A, B). However, the Mnt KO DN4 cells produced considerably fewer viable cells, of all differentiation stages, than Mnt WT DN4 cells (Fig. 4B, C), while the DN3 cell cultures showed no major differences. Thus, MNT loss apparently renders DN4 cells, but not DN3 cells, more vulnerable to apoptosis during culture.

A–C MNT loss enhances predisposition of DN4 but not DN3 progenitor T cells to undergo apoptosis in vitro. DN3 and DN4 cells sorted by FACS from thymi of Mnt^fl/flRag1Cre (gold) and control Mnt^+/+Rag1Cre (brown) mice were cultured in IL-7 on OP9-DL1 stromal cells and (A) analysed on d4 by flow cytometry for (B) viability and (C) differentiation. D–F MNT loss enhances apoptosis of stimulated CD4⁺ but not CD8⁺ splenic T cells. CD4⁺ and CD8⁺ T cells sorted from WT and Mnt^fl/flRag1Cre spleens were cultured in 2 mL medium containing 20 ng/mL PMA and 1 μg/mL ionomycin for 72 h and were then analysed by flow cytometry and immunoblot analysis. D Annexin V-staining (left panels) and viable cell number (right panels). E Typical western blot of MNT, MYC, BCL-2, MCL-1 and BIM protein before and after stimulation, with ACTIN serving as a loading control. F Quantification of BIM, MCL-1 and MYC protein in 3 western blots, relative to ACTIN in the same blot, in T lymphoid cells after stimulation. Values for Mnt^fl/flRag1Cre cells (gold) were normalised to those for WT cells (black). Dots indicate individual mice; bars show mean ± SEM; *P ≤ 0.05, **P ≤ 0.01, ****P ≤ 0.0001; ns not significant.

MNT loss increased apoptosis of splenic CD4⁺ T cells activated in vitro by PMA and ionomycin. The proportion of annexin-V-positive cells was ~2-fold higher in the Mnt KO than the Mnt^+/+ CD4⁺ T cell population and there were fewer viable cells (Fig. 4D, Supplementary Fig. S4C). In contrast, MNT loss had little consequence for CD8⁺ T cells under these conditions.

Taken together, these observations suggest that increased apoptosis is the major determinant of the T cell deficit in Mnt^fl/flRag1Cre mice.

BIM is a critical mediator of apoptosis in MNT-null T cells

Cellular stress causes cell death via the mitochondrial apoptosis pathway, which is regulated by opposing factions of the BCL-2 family [27, 28] and extensive genetic studies have identified BIM (BCL2L11), a pro-apoptotic BH3-only protein, as a key trigger of lymphocyte death [29,30,31,32]. We therefore hypothesised that BIM contributed to the enhanced apoptosis of MNT-deficient T cells.

Notably, western blot analysis and intracellular flow cytometry revealed increased BIM protein in Mnt KO DP thymocytes compared to WT DP thymocytes but no significant change in anti-apoptotic MCL-1, an important regulator of T cell survival [33] (Fig. 5A, B). A modest increase in Bim transcription in MNT-deficient T cells (Fig. 5C) may partly account for the increased BIM protein.

A, B BIM protein is elevated in Mnt KO DP thymocytes. DP T cells were sorted from thymi of 3 WT and 5 Mnt^fl/fl Rag1Cre mice and BIM levels were quantified by immunoblot and intracellular FACS analysis. A Left panel shows one of two western blots and right panel quantifies BIM and MCL-1 levels for both blots relative to ACTIN (protein loading control) and normalised to that in WT cells. B Left panel shows a typical FACS profile, comparing BIM level in DP thymocytes from Mnt^fl/flRag1Cre (gold), WT (black) and, as a negative control, Bim^−/− mice (grey). Right panels show % BIM-positive cells and Mean Fluorescence Intensity (MFI) for BIM in three independent experiments ± SEM. C Bim transcription is elevated in Mnt KO DP thymocytes. Quantitative RT-PCR of Bim transcripts in Mnt^fl/fl Rag1Cre (gold) compared to WT (black) DP thymocytes. Values were calculated relative to Gapdh control. D Bim heterozygosity reduces apoptosis of the major thymic T cell sub-populations in Mnt^fl/flRag1Cre mice. E Bim heterozygosity restores cellularity of major thymic T cell sub-populations in Mnt^fl/flRag1Cre mice. F Bim heterozygosity increases the cellularity of DN2, DN3 and DN4 progenitor cell populations in the thymus of Mnt^fl/flRag1Cre mice. G Bim heterozygosity increases CD4⁺ and CD8⁺ T cells in the spleen of Mnt^fl/fl Rag1Cre mice. H Bim heterozygosity prevents increased apoptosis in activated CD4⁺ and CD8⁺ splenic T cells caused by MNT loss. Splenic CD4⁺ and CD8⁺ T cells from mice of the indicated genotypes were cultured in PMA + ionomycin for 72 h. Apoptosis was quantified by flow cytometry after staining cells with Annexin-V-FITC. Genotypes analysed include WT (black), Mnt^fl/flRag1Cre (gold), Bim^+/−Mnt^fl/fl Rag1Cre (rust) and Bim^+/− (grey). Dots indicate individual mice; bars show mean ± SEM; *P ≤ 0.05, **P ≤ 0.01; ***P ≤ 0.001; ****P ≤ 0.0001; ns not significant. Data include mice in Fig. 2 plus additional mice.

BIM protein was also notably higher in mitogen-activated Mnt KO CD4⁺ splenic T cells, but not in activated CD8⁺ T cells (see Fig. 4E, F), paralleling their apoptosis susceptibility under these conditions (Fig. 4D). These results suggest that MNT suppresses Bim expression in T lymphoid cells, as we previously proposed for B lymphoid cells [15].

To directly test the importance of BIM in the apoptosis of MNT-deficient T cells, we bred Bim^+/− Mnt^fl/flRag1Cre mice (Bim is functionally haplo-insufficient [34]). Indeed, apoptosis in thymocyte populations from Bim^+/− Mnt^fl/flRag1Cre mice (rust bars) was significantly less than in those from Mnt^fl/fl Rag1Cre mice (gold bars), and comparable to that in WT mice (black bars) (Fig. 5D). Notably, the cellularity of the major thymic sub-populations was restored (Fig. 5E), as was that of the DN4 sub-population (Fig. 5F). Splenic T cell cellularity was also restored to normal in the Bim^+/− Mnt^fl/flRag1Cre mice (Fig. 5G). Furthermore, loss of one Bim allele prevented the enhanced apoptosis of Mnt^fl/flRag1Cre CD4⁺ splenic T cells stimulated in vitro by PMA + ionomycin (Fig. 5H). In summary, MNT loss upregulates BIM, thereby enhancing the vulnerability of T cells to apoptosis during normal T lymphopoiesis.

MNT also constrains BIM levels in other cell types. Thus, using CRISPR/Cas9, we found that human MNT KO HEK 293 T and HeLa cells, and Mnt KO Bax^−/−Bak^−/− mouse embryonic fibroblasts (MEFs) express more BIM protein than their parental cells (Supplementary Fig. S5). Since apoptosis is blocked in Bax^−/−Bak^−/− MEFs, MNT regulation of BIM levels must occur upstream of mitochondrial permeability changes. Of note, the mechanism is reversible, because when the MEFs (Mnt^+/+ or Mnt^−/−) cells were infected with MntERT2 retrovirus and treated with 4-OHT to activate the exogenous MNTERT2 protein [35], BIM levels were again reduced (Supplementary Fig. S5I).

Mnt deletion prevents T lymphoma development in MYC10
^hom transgenic mice

To examine the impact of MNT loss on MYC-driven T lymphomagenesis, we utilised our MYC10^hom mice [24, 36], which are homozygous for a transgene expressing human MYC cDNA via the pan-haemopoietic VavP transgenic vector [37]. In these mice, expression of transgenic MYC protein in T lymphoid cells is significantly higher than in B lymphoid and myeloid cells, and thymic T lymphoma is the major cause of morbidity, although these mice can also develop disseminated histiocytic myeloid (monocyte/macrophage) (Mac1⁺F4/80⁺Gr1^–) tumours affecting the spleen and other organs [36].

Mnt^fl/flMYC10^hom/Rag1Cre mice survived significantly longer than the control Mnt^+/+MYC10^hom and Mnt^+/+MYC10^hom/Rag1Cre mice (median of 158 d compared to 136 d and 148 d; p ≤ 0.001, p ≤ 0.01, respectively) (Fig. 6A) and autopsy of euthanised sick mice revealed a major difference in pathology. Whereas the control mice presented with massively enlarged thymi and/or splenomegaly, Mnt^fl/flMYC10^hom/Rag1Cre mice presented with splenomegaly but not thymic enlargement (Fig. 6B).

A Kaplan–Meier survival curves showing significantly delayed morbidity for Mnt^fl/flMYC10^hom/Rag1Cre mice (green curve; median survival 165 d; n = 27: 16 F, 11 M) compared to control Mnt^+/+MYC10^hom (blue curve; median survival 134 d; n = 26: 17 F, 9 M) and Mnt^+/+MYC10^hom/Rag1Cre (red curve; median survival 148 d; n = 24: 12 F, 12 M) mice. B MNT loss prevents thymus but not spleen enlargement. Sick mice were euthanised and the weights of spleens and thymi measured. Bars show mean thymus and spleen weights ± SEM, and dots indicate values for individual euthanised mice. C MNT loss prevents T lymphoma development. Proportion of T (black) and B (grey) thymic lymphomas and myeloid tumours (white) in euthanised Mnt^+/+MYC10^hom, Mnt^+/+MYC10^hom/Rag1Cre and Mnt^fl/flMYC10^hom/Rag1Cre mice. (See text and Tables 1–3). D Polyclonality of thymic T cells in Mnt^fl/flMYC10^hom/Rag1Cre mice. PCR analysis of Tcrβ gene rearrangements was performed on DNA isolated from DP thymic cells from Mnt^fl/flMYC10^hom/Rag1Cre mice # 1313, #1316 (green) and, as a control, a clonal T lymphoma Mnt^+/+MYC10^hom/Rag1Cre mouse #1466 (red). Lower 4 panels show PCR genotype analysis. E Clonality of thymic T lymphomas from Mnt^+/+MYC10^hom mice. DNA was isolated from individual T lymphomas (blue) and genomic PCR analysis performed for indicated Tcrβ gene rearrangements, MYC10 transgene and, as a loading control, Thy1. F Enlarged spleens from both Mnt^+/+ and Mnt^fl/fl MYC10^hom/Rag1Cre mice have elevated proportion of Mac1⁺ myeloid cells. Mouse genotypes analysed included WT (black), Mnt^+/+MYC10^hom/Rag1Cre (red) and Mnt^fl/flMYC10^hom/Rag1Cre mice (green). Dots indicate individual mice; bars show mean % cellularity ± SEM; *P ≤ 0.05, **P ≤ 0.01; ***P ≤ 0.001; ****P ≤ 0.0001; ns not significant.

Importantly, Rag1Cre-mediated Mnt deletion specifically prevented T lymphoma development in MYC10^hom mice (Fig. 6C). None of the 26 mice in the Mnt^fl/fl MYC10^hom/Rag1Cre cohort developed thymic T lymphomas (Supplementary Table S3) and, where analysed, their thymic T cells had polyclonal rather than monoclonal TCR gene rearrangement, consistent with not being transformed (eg # 1313 and #1316 in Fig. 6D). In contrast, 12/26 Mnt^+/+MYC10^hom and 7/24 Mnt^+/+MYC10^hom/Rag1Cre control mice developed massive thymi (up to 1440 mg) and 14/15 of those immunophenotyped were CD4⁺CD8⁺ T lymphomas (the other being a CD19⁺ B lymphoma) (Supplementary Fig. S6A–C and Tables S1, S2). Seven of these thymic T lymphomas analysed by PCR showed 1 or 2 dominant Tcrβ gene rearrangements, indicative of mono- or bi- clonality (Fig. 6D, E). Curiously, the T lymphomas were also Mac-1-positive (Supplementary Fig. S6A, B), which may be due to high MYC expression, because activated CD8⁺ T cells express Mac1 [38].

The grossly enlarged spleens arising in either Mnt^+/+ or Mnt KO MYC10^hom mice contained a high proportion of transplantable Mac-1⁺ myeloid cells (Fig. 6F, Supplementary Table S4). Histological review revealed invasion of many other tissues by these tumour cells, as described previously [36]. Although the splenic CD4⁺ T cells in these mice were clearly activated (CD44⁺CD62L^–) (Supplementary Fig. S6D), they were not transplantable (Supplementary Table S4).

In summary, lymphoid cell-specific Mnt deletion prevented the development of MYC-driven thymic T lymphomas in MYC10^hom transgenic mice but not their myelomonocytic tumours. Whether the MYC-driven myeloid tumorigenesis requires MNT is not addressed by these studies as Rag1Cre is expressed only in lymphoid progenitors [39].

MNT loss impairs T cell development in MYC10
^hom transgenic mice

To clarify why T lymphomagenesis was abrogated in Mnt^fl/flMYC10^hom/Rag1Cre mice, we analysed healthy young ie premalignant (8 wk-old) mice. PCR and western blot analysis of DP thymocytes confirmed efficient Mnt deletion (not shown). Of note, thymic cellularity was reduced ~50% in Mnt^fl/flMYC10^hom/Rag1Cre mice (green) compared to control Mnt^+/+MYC10^hom mice (blue) (p ≤ 0.001), and all thymocyte sub-populations were reduced around two-fold (Fig. 7A).

A–E Thymus: A Reduction in cellularity of all major thymic T cell sub-populations in 8 wk-old premalignant Mnt^fl/flMYC10^hom/Rag1Cre mice (green; n = 10: 3 F, 7 M) vs premalignant Mnt^+/+MYC10^hom (blue; n = 13: 8 F, 5 M) transgenic mice and control WT (black; n = 13: 7 F, 6 M) mice. Cell number was determined by flow cytometry after immunostaining. B, C Elevated apoptosis of thymic T cells in premalignant MYC10^hom transgenic mice was further enhanced by MNT loss. B Quantification of AnnexinV-positive cells in the four major thymic T cell sub-populations from WT (black; n = 8: 3 F, 5 M), Mnt^+/+MYC10^hom (blue; n = 9: 6 F, 3 M) and Mnt^fl/flMYC10^hom/Rag1Cre (green; n = 9: 3 F, 6 M) mice. C MNT loss enhances death of premalignant MYC10^hom DP thymic T cells in vitro. Sorted DP T cells were cultured at 37 ^oC in OptiMEM medium + 10% FCS and viability determined by PI staining at 0, 20, 32 and 48 h. Values are expressed relative to viability at t = 0 h. D–F MNT loss is associated with elevated BIM expression. D Left panel. Typical Western blot showing elevated BIM levels in sorted DP thymic T cells from 2 independent Mnt^fl/flMYC10^hom/Rag1Cre mice vs control Mnt^+/+MYC10^hom or WT mice. Right panel shows quantification of BIM and MCL-1 protein in 2 independent western blots, normalised to ACTIN (protein loading control) first and then normalised to expression in WT cells (see Supplementary Fig. S7A for MNT, MYC and BCL-X_L quantification). E Elevated intracellular BIM staining in DP thymic T cells from Mnt^fl/flMYC10^hom/Rag1Cre vs control Mnt^+/+MYC10^hom mice. Upper panel shows a typical FACS analysis of intracellular BIM expression and lower panel presents BIM Mean Fluorescence Intensity (MFI) ± SEM for DP thymic T cells from 2 independent WT and 3 independent Mnt^+/+MYC10^hom and Mnt^fl/flMYC10^hom/Rag1Cre mice. DP thymic T cells from Bim^−/− mice served as a negative control. F Elevated Bim transcripts in DP thymic T cells from Mnt^fl/flMYC10^hom/Rag1Cre vs control Mnt^+/+MYC10^hom mice revealed by quantitative RT-PCR. Values were calculated relative to Gapdh mRNA used as a loading control. G, H Spleen: G Marked reduction of T and B lymphoid but not myeloid cells in spleens of 8 wk-old premalignant MNT-null MYC10^hom transgenic mice. H Elevated apoptosis of splenic CD4⁺ and CD8⁺ T cells in MNT-null MYC10^hom transgenic mice, quantified by Annexin-V staining. Dots indicate individual mice. Mice used for G, H were those analysed in A and B above. Bars show mean ± SEM; *P ≤ 0.05, **P ≤ 0.01; ***P ≤ 0.001, ****P ≤ 0.0001, ns not significant.

As reported previously [36], the level of MYC protein in thymocytes of MYC10^hom transgenic mice greatly exceeds endogenous MYC levels (compare first 2 tracks in Fig. 7D). Concomitantly, MNT levels are also elevated 3-fold (Fig. 7D, Supplementary Fig. S7A).

MNT loss did not affect MYC protein level or cell size in pre-malignant MYC10^hom T cells (Supplementary Fig. S7B–D). However, the proportion of annexin-V-positive cells was significantly higher in Mnt KO MYC10^hom than Mnt^+/+ MYC10^hom thymocyte sub-populations (compare green to blue bars), which in turn tended to be higher than comparable WT sub-populations (compare blue to black bars; Fig. 7B). Furthermore, when cultured in vitro, Mnt KO MYC10^hom DP thymocytes died faster than their Mnt^+/+MYC10^hom or WT counterparts (Fig. 7C). Thus, an overt consequence of MNT loss was increased apoptosis.

The increased apoptosis paralleled elevated BIM protein levels (Fig. 7D, E) and increased Bim transcription (Fig. 7F) in thymic DP T cells. Anti-apoptotic MCL-1 protein levels were higher in MYC10^hom than WT DP thymocytes, but not affected by MNT loss (Fig. 7D). BCL-X_L levels were comparable in cells from all three genotypes (Supplementary Fig. S7A). The tumour suppressor p53, which can activate apoptosis via transcriptional induction of pro-apoptotic BH3-only proteins PUMA and NOXA [40], was not detectable, by either western blot or qRT-PCR analyses (not shown).

MNT loss also resulted in a deficit of CD4⁺ and CD8⁺ T cells in the spleen of pre-malignant MYC10^hom mice (Fig. 7G, green bars), and annexin V staining (Fig. 7H) indicated greater predisposition to apoptosis. Pertinently, the high MYC levels did not further increase in the absence of MNT (Supplementary Fig. S8A, B). Staining for CD44 and CD62L indicated that, as expected, the MYC transgenic splenic T cells were enlarged and highly activated (Supplementary Fig. S8C, D).

MNT loss also greatly reduced CD19⁺ B lymphoid cells in the spleen and bone marrow of the young MYC10^hom mice, as reported previously for Mnt^fl/fl Eμ-Myc/Rag1Cre mice [15], but myeloid cells (Mac1⁺, Gr1⁺, Mac1⁺Gr1⁺) were unaffected (Fig. 7G, Supplementary Fig. S9C). Indeed, myeloid cell numbers were still normal at this age in both MYC10^hom genotypes, despite the disseminated myeloid disease that inevitably develops as these mice age.

We conclude that abrogation of thymic T lymphoma development in Mnt KO MYC10^hom mice is largely due to the increased apoptosis of the highly proliferative pre-malignant thymic T cells, driven by elevated BIM levels.

MNT loss prevents γ-radiation-induced T lymphoma development

To investigate the MNT-dependency of T lymphomagenesis in the absence of a Myc transgene, we performed serial total body γ-irradiation of C57BL/6 mice. In this well-studied model [26, 41], γ-irradiation decimates leukocytes and the thymus is repeatedly regenerated from bone marrow-derived haemopoietic stem/progenitor cells, some of which have sustained γ-irradiation-induced oncogenic mutation.

As expected, almost all γ-irradiated WT mice and Mnt^+/+/Rag1Cre controls developed thymic T lymphomas (median survival 172 and 204 d respectively). These usually presented in the thymus and often also in the spleen, and had either a CD4⁺CD8⁺, CD8⁺ or mixed surface marker expression profile (Fig. 8A–C, Supplementary Table S5). Western blot analysis (Fig. 8D) showed that, with one exception, MYC protein was lower in γ-irradiation-induced T lymphomas than in MYC10^hom transgenic T lymphomas, but nevertheless still far higher than in a normal thymus. NOTCH1 and p53/p19Arf pathway mutations were frequent, as reported previously [42], and all lymphomas expressed MNT and BIM.

A Kaplan–Meier survival curves for γ-irradiated WT (n = 14: 11 F, 3 M), Mnt^+/+ Rag1Cre (n = 10: 6 F, 4 M) and Mnt^fl/fl Rag1Cre (n = 15: 9 F, 6 M) C57BL/6 mice. Starting at 33 days of age, mice of the indicated genotypes received 4 doses of 1.5 Gy γ-irradiation at weekly intervals. In contrast to the control genotypes, all γ-irradiated Mnt^fl/fl Rag1Cre remained healthy for at least 250 days of age. B The γ-irradiated WT and Mnt^+/+ Rag1Cre mice developed an enlarged thymus and/or spleen. By comparison, the mean weight of the thymus and spleen from healthy unirradiated 8 wk-old C57BL/6 mice is 68 ± 3.4 mg and 87 ± 4.1 mg respectively. C Tumours were dominated by large CD4⁺CD8⁺ and/or CD8⁺ T lymphoid cells (see Supplementary Table 5). Flow cytometric analysis comparing enlarged thymus of a γ-irradiated WT mouse to that of a non-irradiated WT mouse. First two panels display cell surface markers in all major thymic sub-populations; the third panel compares size of DP T cells in tumour (blue) with those in a non-irradiated control mouse (pink). D Western blot analysis of T lymphomas that developed in control γ-irradiated Mnt^+/+ and Mnt^+/+ Rag1Cre mice. Mouse #2319 had a mixed DP (CD4⁺CD8⁺) (designated 2319) and SP (CD8⁺) (designated 2319*) thymic T lymphoma. Mutations that inactivate p53 result in elevated p19Arf protein, due to the absence of the p53-mediated negative feedback loop [58]. ACTIN levels serve as a protein loading control. E–G Thymic T cell analysis following a single dose of γ-irradiation at 33 days of age. E, F Cell number (E) and apoptosis (F) in major thymic T cell sub-populations, analysed by flow cytometry on d2 (WT n = 2 (2 M); Mnt^fl/fl Rag1Cre n = 2 (1 F, 1 M) and d21 (WT n = 5 (3 F, 2 M); Mnt^+/+ Rag1Cre n = 5 (2 F, 3 M); Mnt^fl/fl Rag1Cre n = 8 (4 F, 4 M) after irradiation. Unirradiated control mice (WT n = 9 (3 F, 6 M); Mnt^+/+ Rag1Cre n = 3 (2 F, 1 M); Mnt^fl/fl Rag1Cre n = 4 (2 F, 2 M) were analysed on d21. Apoptosis was determined by staining with Annexin-V. G MNT loss increases BIM protein levels in γ-irradiated (and non-irradiated) thymic T cells. BIM levels in CD4⁺CD8⁺ thymic T cells obtained from mice analysed in E, F above were quantified by intracellular staining followed by FACS. Each dot represents data from an individual mouse.

Remarkably, however, none of the γ-irradiated Mnt^fl/fl/Rag1Cre mice developed lymphomas (Fig. 8A). To understand why, we analysed thymi recovering from the first γ-irradiation dose (Fig. 8E–G). All major T cell populations were greatly reduced on d2 compared to unirradiated controls but had largely recovered by d21, irrespective of genotype (Fig. 8E). However, the Mnt KO T cells exhibited greater levels of apoptosis compared to controls (Fig. 8F) and intracellular staining revealed significantly elevated BIM (Fig. 8G). We infer that MNT loss ‘enhances BIM-induced apoptosis of cells repopulating the thymus, including any clones expanding from stem cells carrying irradiation-induced oncogenic mutations. Consequently, lymphoma development is prevented.