Abstract
Transdifferentiation (TD) is a direct conversion of one cell type into another without passage through induced pluripotent stem cells. Direct TD of normal somatic cells of one lineage to malignant cells of another lineage is not well established. p53 is a “genome guardian” and its deletions/mutations are observed in multiple cancers. It was suggested that p53 may control reprogramming and TD to cancer without in vivo experimental evidence for this. LMO2 is implicated in reprogramming and T-cell acute lymphoblastic leukemia (T-ALL). We showed that Lmo2 expression in B-cells (Lmo2+Mb1-Cre mice) caused TD to T-ALL. We evaluated the in-vivo role of p53 in TD to T-ALL using Lmo2+Mb1-Cre (n=86) p53+/-(n=95), p53-/-(n=88), p53+/-;Lmo2+Mb1-Cre(n=100) and p53-/-;Lmo2+Mb1-Cre(n=77) mice strains. The animals showed normal hematopoietic and B-cell lineage development at birth, with uniform Lmo2 expression (GFP+) only in B cells in the bone marrow, spleen and peripheral blood and no GFP expression in myeloid or T cells (CD4, CD8). The mice strains exhibited statistically significant different overall survival (OS) (log rank p<0.0001) with shortest OS observed in p53-/-;Lmo2+Mb1-Cre mice and longest in Lmo2+Mb1-Cre mice. The OS of the p53+/-;Lmo2+Mb1-Cre animals was shorter than of the p53+/- animals (p=0.017). The OS of the p53-/-:Lmo2+Mb1-Cre animals was shorter than of the p53-/- animals (p=0.007). The leading cause of death was T-ALL in p53+/-(17%), p53-/-(67%), p53+/-;Lmo2+Mb1-Cre (33%) and p53-/-;Lmo2+Mb1-Cre(72%) but not in the Lmo2+Mb1-Cre (16%)animals. The T-ALL cumulative incidence over time was significantly higher in the p53+/-;Lmo2+Mb1-Cre versus p53+/-(p=0.007) and p53-/-;Lmo2+Mb1-Creversus p53-/- (p=0.015) mice. T-ALL developed sooner in the p53-/- than p53+/-animal strains. The median latency time to development of T-ALL was similar in the p53+/-and p53+/-;Lmo2+Mb1-Cre animals but was shorter in the p53-/-;Lmo2+Mb1-Creversus p53-/- animals (t test p=0.045). In all strains, the lymphoblasts were positive for CD3, TdT, CD34, Lmo2 and had high Ki67 expression. The cells were most commonly double CD4+CD8+ positive, or single positive for CD8 or CD4. Flow cytometric analyses of thymic T cell populations in pre-T-ALL animals showed absence of Lmo2 expression as reflected by GFP, however GFP+ (Lmo2 expressing) cells appeared and were detected in T-ALL blasts infiltrating thymuses in all animal strains. T-ALL cells that originated from pro-B cells harbored clonal rearrangements of T-cell receptor and VDJ region genes encoding immunoglobulin heavy chain locus that might originate from transdifferentiated B-cells. The whole exome sequencing of CD4+CD8+ lymphoblasts in 4 animals from each strain of mice detected 450 somatic mutations in 187 genes, including mutations in 39 genes listed in the COSMIC, including in Notch1, Pten and Fat3 genes that are also mutated in human T-ALL. None of the animals exhibited mutations in p53. Notch1 mutations, observed in 50% of human T-ALL, were most observed (in 9 (45%) of analyzed mice), including all 4 analyzed Lmo2+Mb1-Cre mice, 2 of 4 p53+/-mice, 2 of 4 p53+/-;Lmo2+Mb1-Cre mice and 1 of 4 p53-/- mice. RNA expression analysis of T-ALL cells demonstrated loss of B cell and gain of T cell expression programs. including expression of Gata3 and Bcl11b. Expression of direct p53 target genes (Cdkn2a encoding p19ARF, Dsp, EpCAM, Pank1, Sulf2, Me1 and Nlrc4) was significantly higher in the p53-/-;Lmo2+Mb1-Cre and p53+/-;Lmo2+Mb1-Cre animals versus animals with wild type p53. DNA methylation profiles from sorted pro B cells isolated from 15-week-old mice and corresponding T-ALL cells from all the strains revealed genotype-specific epigenetic differences. Noteworthy, several T-cell–associated genes (e.g., Gata3, Lck, Runx3) exhibited partial demethylation in LMO2-expressing pro-B cells, particularly in those with concurrent p53 loss and the same genes were also demethylated in the resulting T-ALL cells. These findings suggest epigenetically primed lineage reprogramming at the pro-B cell stage with further stabilization during TD, offering in vivo proof that p53 plays an essential role in preventing cell reprogramming/TD and show that p53 inactivation in tumors may be permissive for TD of B-cells into T-ALL. As LMO2 is frequently expressed in human T-ALL and p53 deletions and mutations are observed in both newly diagnosed and relapsed T-ALL, TD may be more common than we were assuming.
