Bcl11b is a haploinsufficient tumor suppressor, mutations or deletion of which has been found in 10–16% of T-cell acute lymphoblastic leukemias. Bcl11bKO/+ heterozygous mice are susceptible to thymic lymphomas, a model of T-cell acute lymphoblastic leukemia, when γ-irradiated, and irradiated Bcl11bKO/+ mice generate clonally expanding or premalignant thymocytes before thymic lymphoma development. Cells with radiation-induced DNA damages are assumed to be the cells of origin in tumors; however, which thymocyte is the tumor cell origin remains obscure. In this study we generated Bcl11bflox/+;Lck-Cre and Bcl11bflox/+;CD4-Cre mice; in the former, loss of one Bcl11b allele occurs in thymocytes at the immature CD4−CD8− stage, whereas in the latter the loss occurs in the more differentiated CD4+CD8+ double-positive stage. We examined clonal expansion and differentiation of thymocytes in mice 60 days after 3 Gy γ-irradiation. Half (9/18) of the thymuses in the Bcl11bflox/+;Lck-Cre group showed limited rearrangement sites at the T-cell receptor-β (TCRβ) locus, indicating clonal cell expansion, but none in the Bcl11bflox/+;CD4-Cre group did. This indicates that the origin of the premalignant thymocytes is not in double-positive cells but immature thymocytes. Interestingly, those premalignant thymocytes underwent rearrangement at various different sites of the TCRα locus and the majority showed a higher expression of TCRβ and CD8, and more differentiated phenotypes. This suggests the existence of a subpopulation of immature cells within the premalignant cells that is capable of proliferating and continuously producing differentiated thymocytes.
Mouse thymic lymphoma, a model of human T-cell acute lymphoblastic leukemias (T-ALL), is induced by fractionated whole-body γ-irradiation with a high incidence.[1-4] An aggressive malignancy of thymocytes, T-ALL accounts for approximately 15% of all ALL cases.[5, 6] Radiation can damage DNA within the cell, and DNA-damaged cells are assumed to be the cells of origin in radiation-induced tumors. However, as in most other cancers, which cell within the thymus is the tumor cell origin remains obscure. Furthermore, classic studies have indicated that the tissue of radiation target may not be the thymus but other tissues such as bone marrow.
The Bcl11b gene encodes zinc-finger transcription proteins[7-9] and is a haploinsufficient tumor suppressor gene in mouse thymic lymphomas and T-ALL.[10, 11] Recently, Bcl11b mutations or deletion at only one allele was found in 10–16% of T-ALL.[11, 12] Bcl11bKO/+ heterozygous mice rarely develop thymic lymphomas spontaneously until 600 days after birth but are susceptible to them when γ-irradiated at a single dose of 3 Gy. Therefore, efficient induction of thymic lymphomas in mice requires the loss of one Bcl11b allele and γ-irradiation. Analysis of thymocytes in irradiated Bcl11bKO/+ mice at early stages before thymic lymphoma development revealed the generation of clonally expanding or premalignant thymocytes, some of which probably give rise to thymic lymphomas after acquiring additional mutations. Similar premalignant thymocytes are found in another T-ALL model that overexpresses the Lmo2 oncogene. Interestingly, Lmo2-induced premalignant immature thymocytes produce various T-cell subsets with normal properties for more than 1 year, comprising cells with the stem cell-like self-renewal property that functions for bone marrow stem cells.
In this study, we aimed to investigate which thymocyte is the origin of clonally expanding cells in irradiated mice by introducing conditional deletion of one Bcl11b allele at different differentiation stages. Thus, we developed Bcl11bflox/+;Lck-Cre and Bcl11bflox/+;CD4-Cre mice. The former mice lose one allele at the immature CD4− CD8− double-negative (DN) stage of differentiation without expression of CD4 or CD8 cell surface markers, whereas the latter lose that at the more differentiated CD4+ CD8+ double-positive (DP) stage. Using these mouse models, we examined development and properties of the clonally expanding or premalignant thymocytes after γ-irradiation. We found the clonal expansion in the former model, but not in the latter model, suggesting that the origin of premalignant thymocytes leading to thymic lymphomas is in immature thymocytes but not in DP thymocytes, the cell subset found in a majority of thymic lymphomas.[16-19] Of interest, a majority of the clonally expanded thymocytes possessed phenotypes of differentiated thymocytes. This suggests the presence of a cell subpopulation within the clonally expanded thymocytes that is capable of continuously producing differentiated thymocytes.
Material and Methods
The Bcl11bflox/+ C57BL/6 mouse harboring a Bcl11b-flox allele was generated by Katsuragi et al. (unpublished report). Bcl11bflox/+ mice were mated with Lck-Cre or CD4-Cre mice of C57BL/6J background. Efficient deletion of Bcl11b floxed allele was observed in the thymus of Bcl11bflox/+;Lck-Cre and Bcl11bflox/+;Cd4-Cre mice (Fig. S1). Their progeny were 3 Gy γ-irradiated at 8 weeks of age (1 Gy/min). Left and right lobes of the thymus were separately isolated 60 days after irradiation and subjected to analyses. Mice used in this study were maintained under specific pathogen-free conditions in the animal colony of Niigata University (Niigata, Japan). All animal experiments complied with the guidelines set by the ethics committee for animal experimentation of the university.
Single cell suspensions of thymocytes were prepared from thymus and 2–4 × 106 cells were incubated with antibodies in PBS containing 2% FCS and 0.2% NaN3 for 15 min at 4°C. The mAbs used were anti-CD4-PerCP-Cy5.5 or -APC (clone, RM4-5), anti-CD8-PE (53–6.7), and anti-TCRβ-FITC (H57-597) (BioLegend, San Diego, CA, USA). To prevent non-specific binding of mAbs, we added CD16/32 (93; eBioscience, San Diego, CA, USA) before staining with labeled mAbs. Dead cells and debris were excluded from the analysis by appropriate gating of forward scatter (FSC) and side scatter. After the treatment, cells were analyzed by a FACScalibur (Becton-Dickinson, Franklin Lakes, NJ, USA) flow cytometer, and data were analyzed using FlowJo software (TreeStar, Ashland, OR, USA).
Incorporation of BrdU
Sixty days after irradiation, mice were injected i.p. with 100 μL BrdU solution (10 mg/mL; Sigma, St Louis, MO, USA) and the thymus was isolated 1 h later. Thymocytes were analyzed with the use of a BrdU Flow Kit (BD Pharmingen, San Diego, CA, USA) according to the manufacturer's instructions. In brief, cells were suspended at a concentration of 1–2 × 106 cells/mL, fixed, permeabilized, treated with DNase to expose incorporated BrdU, and incubated with a murine anti-BrdU antibody for 20 min at room temperature.
For cell cycle analysis of thymocytes in Bcl11bflox/+;Lck-Cre and Bcl11bflox/+ mice, we injected 100 μL BrdU solution (10 mg/mL) i.p. Thymuses were isolated 5 h after BrdU injection and analyzed as described above. In indicated cases, 1 Gy irradiation was carried out 1 h after BrdU injection to examine its effect.
Detection of D-J rearrangement at TCRβ locus
To determine D-J rearrangement patterns at the T-cell receptor-β (TCRβ) locus, PCR was carried out as described previously.[14, 22] The PCR products were analyzed by 1.5% agarose gel electrophoresis and visualized by ethidium bromide staining.
Detection of Vα-Cα rearrangement at TCRα locus
Total RNA was prepared from thymocytes by the RNeasy Mini kit (Qiagen, Valencia, CA, USA) according to the protocol recommended by the manufacturer. cDNA was synthesized from 5 μg total RNA with a oligo (dT) primer using SuperScript II reverse transcriptase (Invitrogen, Carlsbad, CA, USA). The cDNA was subjected to PCR using primer sets that were specific for Vα2, Vα3, Vα6, Vα8, Vα14, Vα17, and Vα19 regions, and for Cα in the constant region as common reverse primers, as described in Hu et al. As a control, specific primers for GAPDH were used. The cycling conditions were as follows: denaturation of 1 min at 94°C; 38 cycles of 30 s at 94°C, 30 s at 56°C, and 30 s at 72°C; and a final extension of 5 min at 72°C. The products were separated by 1.5% agarose gel electrophoresis and visualized by ethidium bromide staining.
Data were analyzed for statistical significance by Student's t-test.
Clonal expansion of thymocytes in irradiated mice of different Bcl11b genotypes
CD4− CD8− DN immature thymocytes, which differentiate into CD4+ CD8+ DP thymocytes and CD4+ or CD8+ single-positive (SP) mature thymocytes, are divided into four DN1–DN4 subpopulations based on the surface expression of CD44 and CD25. Their developmental progression is CD44+ CD25− (DN1) to CD44+ CD25+ (DN2) to CD44− CD25+ (DN3), and then to CD44− CD25− (DN4) cells. DNA rearrangement at the TCRβ and TCRα loci takes place at the DN3 and DP stages, respectively. TCRβ is highly expressed in a small fraction of DP cells and mature CD4SP and CD8SP cells, whereas much lower expression is seen in immature CD8+ SP (ISP) cells. The expression of Lck is developmentally regulated and takes place in DN2 thymocytes.
We generated Bcl11bflox/+;Lck-Cre mice of C57BL/6 (B6) background by crossing Bcl11bflox/flox mice with Lck-Cre mice. The mice lose one Bcl11b allele in thymocytes at the DN2/3 stages. Mice at 8 weeks of age were subjected to 3 Gy γ-irradiation, and 60 days after irradiation left and right lobes of the thymus were separately isolated. These thymic lobes were atrophic and contracted, harboring a reduced number of thymocytes as shown in previous published reports.[1, 3, 25] Clonality of the thymocytes was determined by assaying specific V(D)J rearrangements with three PCR primer sets designed for the TCRβ locus. Figure 1(a) shows electrophoretic patterns of PCR products. Unirradiated thymus (lane Th) gave six different bands corresponding to possible recombination sites between D and J regions by Dβ1-Jβ1, Dβ2-Jβ2, and Dβ1-Jβ2 probe sets, and one band for germline DNA by the former two probe sets. In contrast, the thymic lymphoma DNA (Ly) gave one band only by the Dβ2-Jβ2 probe set used, indicating an identical rearrangement, and brain DNA (Br) gave the germline DNA band by Dβ1-Jβ1 and Dβ2-Jβ2 probe sets.
Half (9/18) of the thymuses showed only a few bands or limited numbers of bands different from the normal thymus pattern, indicating the existence of clonally expanded thymocytes that consisted of cells of the same origins. Those thymocytes were designated as C-type thymocytes (C stands for clonal expansion) and the thymus containing C-type thymocytes was named C-type thymus, for the purposes of this article. The other thymuses showed rearrangement patterns identical or similar to the control thymus (designated as T-type thymus; T stands for thymus). All 16 control Bcl11bflox/+ mouse thymuses were T-type. These results suggest that the Lck-Cre-induced loss of one Bcl11b allele in thymocytes at the DN2/DN3 stages promotes the development of C-type or premalignant thymocytes after γ-irradiation. Figure 1(b) shows PCR patterns of spleen DNA. No change in band patterns was detected, whereas a prominent band was detected in spleen of the mice with overt thymic lymphomas (Fig. S2). This indicated that descendant T cells from the C-type thymocytes did not dominate in spleen.
Figure 1(c) shows a Kaplan–Meier analysis of thymic lymphoma development in Bcl11bflox/+;Lck-Cre and Bcl11bflox/+ mice. Four of the five Bcl11bflox/+;Lck-Cre mice developed thymic lymphomas, whereas none of the 10 Bcl11bflox/+ mice did. Those thymic lymphomas retained the wild-type Bcl11b allele (Fig. 1d), as found in Bcl11bKO/+ mouse thymic lymphomas. The result suggests that Lck-Cre-induced loss of a Bcl11b allele contributes to thymic lymphoma development, probably by generating C-type thymocytes in Bcl11bflox/+;Lck-Cre mice.
Next, we generated Bcl11bflox/+;CD4-Cre mice by crossing Bcl11bflox/flox mice with CD4-Cre mice. The mice lose one Bcl11b allele in thymocytes at the DP stage. At 8 weeks of age, the mice were subjected to 3 Gy γ-irradiation. Figure 1(e) shows PCR patterns of clonality assay. None of the 10 Bcl11bflox/+;CD4-Cre thymuses showed changes in band patterns, revealing no C-type thymuses developed in those animals. This suggests that the CD4-Cre-induced loss of one Bcl11b allele at the DP stage does not contribute to clonal growth of thymocytes.
Differentiation capability of C-type thymocytes
We examined differentiation of the C-type thymuses. Figure 2(a) shows representative results of flow cytometric analyses for control, T-type, and C-type thymocytes using differentiation markers. The left panels show CD4 and CD8 expressions. The percentage of CD8SP cells was increased in C-type thymuses, suggesting a higher production and/or a prolonged retention of CD8SP cells within the thymus. The middle and right panels show TCRβ expression in total thymocytes and in the CD8+ fraction, respectively. Figure 2(b) summarizes the percentage of TCRβhigh cells in total thymocytes. This percentage in the C-type thymuses in irradiated Bcl11bflox/+;Lck-Cre mice was approximately 30% on average, higher than that in irradiated Bcl11bflox/+ mouse thymuses and T-type thymuses in irradiated Bcl11bflox/+;Lck-Cre mice. The result indicated that C-type thymocytes consisted of TCRβhigh differentiated cells more than control and T-type thymocytes. Figure S3 shows flow cytometry analyses of thymocytes and the percentage of TCRβhigh cells in irradiated Bcl11bflox/+;CD4-Cre mice. No significant changes were observed.
We examined DNA rearrangement at the TCRα locus using cDNA prepared from thymocytes. Seven sets of PCR primers on different Vα regions were used (Fig. 2c), which covered major recombination sites. We compared the proportion of rearrangement sites between unirradiated thymuses, irradiated Bcl11bflox/+ thymuses, and C-type thymuses using three different concentrations of the cDNA template (Fig. 2d). No prominent band was detected in C-type thymuses with any primer pairs used, and comparison between C-type and control thymuses did not show any marked difference. This indicated that rearrangements occurred at various sites on the TCRα locus in C-type thymuses to produce cells with various TCRα chains. This suggests no clonal expansion of the mature TCRβhigh thymocytes that rearranged DNA at the TCRα locus.
Cell proliferation of C-type thymocytes
We examined cell proliferation in C-type and other thymocytes. Figure 3(a) shows flow cytometry of various thymocyte subsets in mice 1 h after BrdU injection. The horizontal axis shows BrdU incorporation and the vertical axis indicates cell number. Figure 3(b) summarizes the percentage of BrdU-incorporated cells in total thymocytes, in thymocytes with TCRβ expression, and in cells of ISP, DP, CD4SP, and CD8SP fractions. In total and DP thymocytes, the percentage of BrdU+ cells was increased in C-type thymuses in irradiated Bcl11bflox/+;Lck-Cre mice relative to irradiated Bcl11bflox/+ mouse thymuses and control unirradiated thymuses. This increase was also observed in T-type thymuses. Interestingly, a marked increase was observed in thymocytes with TCRβ expression and in CD8SP cells, although the normal TCRβhigh and CD8SP cells did not proliferate. These results indicated elevation of cell proliferation in the DP and CD8SP cells in C-type thymuses.
Decreased numbers of thymocytes in irradiated Bcl11bflox/+;Lck-Cre mice
Our previous study showed a decrease in the number of thymocytes in Bcl11bKO/+, but not Bcl11b+/+, thymuses at 60 days after γ-irradiation. We thus examined cell numbers in irradiated Bcl11bflox/+;Lck-Cre mice (Fig. 4a). The number was significantly decreased in C-type thymuses and also T-type thymuses, but it was retained in irradiated Bcl11bflox/+ mice at a similar level to that of non-irradiated mice. The result suggests impairment in the maintenance of thymocytes in Bcl11bflox/+;Lck-Cre mice after γ-irradiation, as found in irradiated Bcl11bKO/+ mice. Figure 4(b) shows the number of thymocytes in irradiated Bcl11bflox/+CD4-Cre mice. No decrease was observed.
Effect of loss of one Bcl11b allele on differentiation and cell cycle regulation
We examined differentiation of thymocytes in non-irradiated Bcl11bflox/+ and Bcl11bflox/+;Lck-Cre mice. Figure 5(a) shows representative flow cytometric analyses using CD4, CD8, and TCRβ, and Figure 5(b) summarizes the percentage of DN, ISP, DP, CD4SP, and CD8SP cells. The percentage of DN and ISP cells was increased in Bcl11bflox/+;Lck-Cre mice relative to that of Bcl11bflox/+ mice, whereas the percentage of DP, CD4SP, and CD8SP cells was decreased. This indicated that the Lck-Cre-induced loss of one Bcl11b allele provided two- or three-fold elevation in the percentage of immature thymocytes. Figure S4 shows flow cytometry analyses and the summary of the percentages of those cells in Bcl11bflox/+;CD4-Cre mice. No significant change was observed.
We previously showed impairment of cell cycle arrest at the S phase in response to γ-irradiation in ISP cells of Bcl11bKO/+ mice. We thus examined the arrest in Bcl11bflox/+ and Bcl11bflox/+;Lck-Cre mice. To monitor the cell cycle, we injected BrdU 1 h before 1 Gy γ-irradiation, and analyzed thymocytes 4 h after irradiation. Figure 6(a) shows results of flow cytometric analysis. The BrdU incorporation is shown on the vertical axis and FSC values on the horizontal axis, an indication of cell size. The BrdU+ FSClarge fraction represents BrdU-incorporated cells in S and G2/M phases of the cell cycle, whereas the BrdU+ FSCsmall fraction represents G1 cells that have passed S phase after BrdU incorporation. Figure 6(b) summarizes the percentage of cells in the BrdU+ FSClarge and BrdU+ FSCsmall fractions in ISP cells. Both fractions showed higher percentages in Bcl11bflox/+;Lck-Cre mice than in Bcl11bflox/+ mice, indicating more cells in the S phase and in the G1 cells that have passed S phase. The result suggests that the Lck-Cre-induced loss of one Bcl11b allele attenuates radiation-induced cell cycle arrest at the S phase in ISP cells.
In this paper, we examined clonal expansion of thymocytes in Bcl11bflox/+;Lck-Cre and Bcl11bflox/+CD4-Cre mice after γ-irradiation at an early stage prior to the time of thymic lymphoma development. In those mice, loss of one Bcl11b allele occurs in thymocytes at the DN2/DN3 and DP developmental stages, respectively. We found clonally expanding or C-type thymocytes in a half of Bcl11bflox/+;Lck-Cre thymuses but not in Bcl11bflox/+CD4-Cre thymuses, suggesting the origin of premalignant thymocytes in cells before the DP stage. The result is of interest because thymic lymphomas in T-ALL mouse models having mutation in the Ku70, FBXW7, or Pten genes consisted of DP cells.[16-19] Furthermore, DP cells are assumed to be the origin of leukemia stem cells in Pten-null mice.
Individual C-type thymocytes possessed common rearrangement sites at the TCRβ locus but underwent rearrangement at various different sites at the TCRα locus. This suggests the existence of a subpopulation of precursors within the C-type thymocytes that retains the capability of DNA rearrangement at the DP stage to produce cells with different TCRα chains. The C-type precursors are cells after the rearrangement at the TCRβ locus and before the rearrangement at the TCRα locus. Hence, they may be the immature cells present during the DN3 to DP stages, including ISP cells. However, the majority of C-type thymocytes showed committed or more mature phenotypes after the DP TCRβlow stage, and a higher expression of TCRβ and CD8. This suggests that the C-type precursors continuously produce C-type thymocytes with differentiated phenotypes. In terms of proliferation and producing differentiated cells, the C-type precursors resemble, although are not parallel to, the premalignant DN3 cells that develop in Lmo2-transgenic mice, as briefly described above. The Lmo2-DN3 cells give rise to thymic lymphomas after a long latency, and when transplanted, they develop DN3 cells within the recipient thymus that function for the bone marrow stem cells, continuously producing differentiated thymocytes. In contrast, transplanted Bcl11bKO/+ premalignant immature cells were able to form thymic lymphomas, as shown in other premalignant immature thymocytes,[2, 3] but were not able to produce committed DP thymocytes (Fig. S5). This suggests that acquisition of an Lmo2-transgene, but not loss of one Bcl11b allele, confers the stem cell-like property to DN cells.
The Lck-Cre-induced loss of one Bcl11b allele in thymocytes provided two- or three-fold elevation in the percentage of immature thymocytes, indicating the increase in cells of origin of the premalignant thymocytes described above. Therefore, this increase may be a contribution of the loss of one Bcl11b allele to thymic lymphoma development. Another contribution may be deregulation of the cell cycle. The ISP cells showed attenuation of arrest at S phase in Bcl11bflox/+;Lck-Cre mice after γ-irradiation. This propensity of attenuation may influence the response to DNA damage and DNA replication stress, leading to genomic instability in thymocytes. Increased genomic instability is a key risk factor for leukemic transformation.[27, 28] Consistent with this, a similar attenuated response was detected in Bcl11bKO/+ mice. One of the targets of Bcl11b transcription repressor is the MDM2 gene, which functions as a negative regulator of p53, a key tumor suppressor. Thus, loss of one Bcl11b allele may affect the p53–MDM2 feedback loop to reduce the p53 expression level.
Clonally expanding thymocytes were detected in Bcl11bKO/+ mice of (MSM/Ms × BALB/c) F1 genetic background at 60 and 80 days after 3 Gy γ-irradiation. They also showed a decrease in the number of thymocytes, suggesting a relationship between C-type cell generation and the cell number decrease. This is consistent with the interpretation by classic histological studies that the tumors arose in small, lymphocyte-depleted thymuses. However, the C-type thymocytes differ in phenotype from those in irradiated Bcl11bflox/+;Lck-Cre mice. One difference is that most C-type thymocytes consisted of immature cells but not mature thymocytes highly expressing TCRβ, indicating differentiation arrest of thymocytes before the DP stage. Another difference is the decrease in cell proliferation. In Bcl11bKO/+ mice, loss of one Bcl11b allele exists in almost all cells of T-cell lineage in the thymus, whereas in Bcl11bflox/+;Lck-Cre mice the loss occurs only in thymocytes after the DN2/DN3 stages. Although some of these differences may have been attributable to the different genetic background of the mice used, the results suggest that loss of one Bcl11b allele in different thymocyte subsets leads to premalignant cells with different properties, and that the origin of premalignant thymocytes can be cells at different developmental stages depending on the stage of the driver mutation.
These findings may parallel, in part, results published in a recent report that tests the effect of cell of origin on thymic lymphoma development by using three different Cre-transgenic mouse strains carrying Vav-1-Cre, Lck-Cre, and CD4-Cre. The strains can induce Sleeping Beauty transposase expression at three different developmental stages, leading to insertional mutagenesis in different types of thymocytes or their progenitors. All three strains developed thymic lymphomas, and the thymic lymphomas showed different latencies and different distributions of driver mutations in the three strains. Although phenotypes of the thymic lymphomas were not examined, the study shows that the cell of origin is a key determinant of the effectiveness and genetic selection in thymic lymphoma development.
This work was supported by grants-in-aid of the Third Term Comprehensive Control Research for Cancer from the Ministry of Health, Labor and Welfare of Japan, and for Cancer Research from the Ministry of Education, Science, Technology, Sports, and Culture of Japan.