Knockdown of Cbfb::MYH11 (CM) decreases leukemic burden in the spleen and peripheral blood, but not the bone marrow
To test the role of Cbfb::MYH11 (CM) in leukemia maintenance in vitro, we used leukemia cells from Cbfb+/56M, Mx1-Cre+ (CM+) mice which conditionally express CM from the endogenous Cbfb promoter and develop leukemia that faithfully recapitulates human inv(16) AML [18]. We isolated leukemia cells from terminal mice and transduced them with a doxycycline (Dox) inducible shRNA against MYH11 (shMYH11) which also constitutively expresses GFP [19]. We sorted GFP+ cells and cultured them with or without Dox for 48 h (hrs) (Fig. 1A). As expected, we observed that Dox induced CM KD resulted in a significant decrease of CM expression at both the mRNA and protein levels (Fig. 1B, C). We also found that CM KD induced increased apoptosis and decreased colony formation (Fig. 1D, E). These findings indicate that CM is required for the growth and survival of frank leukemia cells in vitro.
Fig. 1: Inducible knockdown of CM causes apoptosis and impairs colony formation in vitro.
A Schematic showing the experimental design in which Cbfb+/56M (CM+) leukemia cells are transduced with a vector expressing an inducible shRNA against MYH11 (shMYH11) and GFP under a constitutive promoter, sorted, and either used for experiments in vitro or transplanted into mice for in vivo testing. This figure was created with BioRender.com. B Bar graph showing quantitative reverse transcription PCR (qRT-PCR) of Cbfb::MYH11 in shMYH11 transduced CM+ leukemia cells treated in vitro with either control (CTRL) or doxycycline (Dox) for 48 h. C Representative western blot (left) and bar graph of relative CBFβ::SMMHC protein levels (right) in shMYH11 transduced leukemia cells treated as in (B). D Bar graph showing the percentage (%) of shMYH11 transduced leukemia cells treated with control or Dox that are positive for Annexin V. E Bar graph showing relative colony numbers generated by cells treated as in (A). N = 4 biological replicates, with each biological replicate in B and E performed in triplicate., *=p ≤ 0.05 compared to CTRL.
To determine the effect of decreased CM expression in vivo, we transplanted 500,000 sorted, shMYH11 transduced cells into sub-lethally irradiated mice (4.25 Gy) and allowed them to engraft. The peripheral blood was analyzed for the presence of GFP+ cells to confirm establishment of disease. Once mice showed at least 30% GFP+ cells in the peripheral blood (SFig. S1A, B), mice were given Dox in their water to induce CM KD or control water, sacrificed at days 4, 7, 14, 21, and 28 days, and the percentage of GFP+ cells in the blood, spleen, and bone marrow determined. We observed that CM KD induced a significant decrease in GFP+ leukemia cells in the peripheral blood and spleen starting at day 7, with near complete elimination by day 14 (Fig. 2A, B). In addition, we observed a notable decrease in spleen size with CM KD (SFig. S1C, D). In the bone marrow, there was a significant decrease in leukemia cells starting at day 7 of Dox treatment. However, leukemia cells continued to make up 20-30% of the bone marrow through day 28 (Fig. 2C). To determine if this persistence was due to a failure of Dox to penetrate the bone marrow, we sorted GFP+ leukemia cells from the bone marrow of control and Dox treated mice at days 7 and 21. Quantitative real time PCR (qRT-PCR) and western blot analysis of these cells showed significant decrease in CM RNA and protein at both time points indicating efficient KD in the bone marrow (Fig. 2D–F).
Fig. 2: CM knockdown decreases leukemic burden in vivo.
A Line graph showing the percentage (%) of GFP+ shMYH11 transduced leukemia cells (left) and representative flow cytometry plots (right) in the peripheral blood (PB), B spleen (SP) and C bone marrow (BM) of mice treated with doxycycline (Dox) or control (CTRL) for 4, 7, 14, 21, or 28 days. D Bar graph showing relative expression of CM in sorted GFP+ leukemia cells from the BM of mice treated with Dox or CTRL for 7 and E 21 days. F Representative western blot (left) and bar graph (right) of relative CBFβ::SMMHC protein levels in sorted GFP+ cells from the BM of mice treated with either CTRL or Dox for 7 days. N = 4 biological replicates per group at each time point (panels A–C) and N = 3 (D–F) biological replicates, with each biological replicate in (D and E) performed in triplicate, ns not significant, *=p ≤ 0.05.
To confirm that the loss of leukemia cells in Dox treated mice is due to decreased CM expression and not the effect of Dox, we used CM expressing leukemia cells transduced with a non-targeted shRNA construct. We observed no difference in the percentage of GFP+ leukemia cells between Dox and control treated mice (SFig. S1E, F). These findings indicate that cells with reduced CM can survive in bone marrow, but not the peripheral blood or spleen.
Deletion of Cbfb::MYH11 impairs the growth and survival but does not eliminate colony formation of leukemia cells in vitro
To validate our finding that a population of leukemia cells can survive without the fusion gene, we generated a new knock-in mouse model that allows deletion of the Cbfb::MYH11 fusion gene by Cre Recombinase (Cre). We used homologous recombination to knock-in the Cbfb exon5::MYH11 fusion flanked by LoxP sites into the endogenous Cbfb locus (Cbfb+/flMYH11) (Fig. 3A). We identified embryonic stem cell (ES) clones with the knock-in allele by Southern blot and confirmed CBFβ::SMMHC expression by western blot (SFig. S2A, B). Three independent Cbfb+/flMYH11 ES clones were used to generate chimeric mice. To test that the CbfbflMYH11 allele can be fully excised, chimeric mice were crossed with mice expressing Cre from the Actin B promoter (ActB-Cre+), which should allow for excision of the fusion gene in all cells of an embryo. We harvested embryos at day 12.5 of development. In Cbfb+/flMYH11 embryos lacking the ActB-Cre transgene, we observed central nervous system hemorrhaging, indicating expression of Cbfb::MYH11, as in previous knock-in models (Fig. 3B) [23]. Importantly, this phenotype was not observed in Cbfb+/flMYH11; ActB-Cre+, indicating excision of the fusion gene in the presence of Cre. PCR using primers specific for the unexcised and excised alleles confirmed that CbfbflMYH11 was efficiently excised in Cbfb+/flMYH11; ActB-Cre+ mice. To generate leukemia cells expressing the CbfbflMYH11 allele, chimeric mice were treated with N-ethyl-N-nitrosourea and monitored for disease onset. Once leukemia was detected, cells were collected and transplanted into congenic recipient mice and monitored for survival (Fig. 3C). Recipient mice transplanted with leukemia cells from 3 different ES cell clones developed disease with a similar immunophenotype and histological appearance as previous CM knock-in mouse models (Fig. 3D, SFig. 2C, D) [18, 31]. These results indicate that Cbfb+/flMYH11 mice develop a frank leukemia with a similar phenotype to previous models and that the fusion gene can be deleted by Cre.
Fig. 3: Generation of flCbfb::MYH11 mice.
A Schematic representation of the embryonic stem (ES) cell targeting strategy. The probe (0.2 C) and restriction enzyme sites (NcoI) used for screening are indicated. The black arrows represent PCR primers for the unexcised allele, and the orange arrows represent primers for the excised allele. B Pictures of representative embryos harvested at embryonic day 12.5 generated by crossing Cbfb+/flMYH11 chimeras with transgenic mice expressing Cre recombinase (Cre) from the β–Actin promoter (ActB-Cre+). The right panel shows a representative PCR for the excised and unexcised Cbfb+/flMYH11 allele in DNA from embryos of the indicated genotype. C Kaplan-Meyer survival curve of recipient mice transplanted with Cbfb+/flMYH11 leukemic cells from three different chimeric donor mice. D Representative Wright-Giemsa-stained peripheral blood smear and hematoxylin & eosin-stained spleen and liver section from a leukemic Cbfb+/flMYH11 recipient mouse. N = 3 biological replicates.
To test the effect of CM deletion in vitro, we transduced Cbfb+/flMYH11 leukemia cells with a lentivirus expressing Cre and GFP from an IRES (Fig. 4A). Transduced cells were sorted, and CM excision was confirmed by qPCR and western blot (Fig. 4B, C). We found that CM deletion caused a statistically significant increase in Annexin V, indicating induction of apoptosis (Fig. 4D). To test the effect of CM deletion on colony forming ability, we plated equal numbers of control and Cre transduced Cbfb+/flMYH11 leukemia cells in methylcellulose. After 14 days, colonies were counted, resuspended as single cells, and equal numbers of cells replated in methylcellulose. At both the 1st and 2nd plating, we found that Cbfb+/flMYH11 cells transduced with Cre generated significantly fewer colonies (Fig. 4E). To determine if the colonies that were able to grow from Cre-infected cells had successfully deleted CM, individual colonies were picked and qPCR for the CbfbflMYH11 allele was performed. We found that most colonies retained the unexcised CM allele. However, 5-20% of the colonies showed successful deletion of CM (SFig. S3A, B). These findings are similar to what we observed with in vivo knockdown of the fusion protein in which the majority of leukemia cells require continued expression of the fusion gene after leukemia initiation, but that there is a population of cells that are able to survive without CM.
Fig. 4: Deletion of Cbfb::MYH11 induces apoptosis and reduces colony formation in vitro.
A Schematic representation of the experimental design to test the effect of Cbfb::MYH11 deletion in vitro. This figure was created with BioRender.com. B Bar graph of qRT-PCR of the unexcised or excised Cbfb+/flMYH11 alleles in leukemia cells transduced with a lentivirus expressing Cre Recombinase (Cre) or Control (CTRL) and sorted for GFP. C Representative western blot of CBFβ::SMMHC in Cbfb+/flMYH11 leukemia cells infected with CTRL or Cre virus. D Bar graph of % Annexin V in CTRL or Cre transduced Cbfb+/flMYH11 leukemia cells. E Bar graph showing the relative number of colonies from equal numbers of CTRL or Cre transduced Cbfb+/flMYH11 leukemia cells. Colonies were scored on day 14. N = 3 biological replicates, with each biological replicate in B and E performed in triplicate, *=p ≤ 0.05, *** = p ≤ 0.001 as compared to CTRL.
KD of CM alters apoptosis and proliferation
To investigate potential mechanisms for the persistence of CM KD cells in the bone marrow, we analyzed apoptosis and proliferation in the leukemic cells of Dox and control treated mice. Mice were transplanted and treated as described in Fig. 2 and sacrificed at Day 7 and Day 21 of treatment. At Day 7, we found that Dox, but not control, treated mice had a significant increase in apoptosis in the spleen, but not the bone marrow (Fig. 5A left). However, at day 21, a time point in which control mice have high leukemic burden, we found high levels of apoptotic leukemia cells in the bone marrow of control treated mice, but not Dox treated mice (Fig. 5A right). To determine if the reduced apoptosis in the bone marrow with CM KD is related to the levels of anti-apoptotic proteins, we sorted leukemia cells from the spleen and bone marrow of control and Dox treated mice on day 7 and performed western blots. We found that KD of CM induced decreased BCL-2 in the spleen but caused no change in MCL-1. In leukemia cells from the bone marrow of both control and Dox treated mice, BCL-2 was barely detectable. In contrast, MCL-1 was higher in leukemia cells from the bone marrow of both control and Dox treated mice, as compared to cells from the spleen (Fig. 5B, C). These findings imply that MCL-1, but not BCL-2, may contribute to the survival of CM KD cells in the bone marrow.
Fig. 5: CM Knockdown induces apoptosis in leukemia cells in the spleen, but not the bone marrow.
A Bar graph showing the percentage (%) of Annexin V+, GFP+ cells from the spleen (SP) and bone marrow (BM) of mice treated with control (CTRL) or doxycycline (Dox) for 7 (left, N = 5 biological replicates) or 21 days (right, N = 4 biological replicates). B Representative western blot and (C). Bar graph of quantification for the indicated proteins from leukemia cells sorted for GFP from the SP and BM after 7 days of treatment with CTRL or Dox. (N = 3 biological replicates). D Bar graph of the % of BrdU+, GFP+ leukemia cells harvested from the SP and BM of mice after 21 days of treatment with CTRL or Dox (N = 4 biological replicates). E Bar graph showing the % of GFP+ cells expressing the mature myeloid markers Mac-1 and Gr-1 from the SP and BM of mice treated as in (D) (N = 3 biological replicates). F Bar graph showing the % of GFP+ cells expressing the indicated leukemia stem cell markers in the BM from of mice treated as in (D) (N = 4 biological replicates). ns = not significant, * = p ≤ 0.05, ** = p ≤ 0.01, *** = p ≤ 0.001 as compared to CTRL.
To determine if a difference in proliferation could also contribute to the persistence of CM KD cells in the bone marrow, we used BrdU incorporation. Leukemia cells from both the spleen and bone marrow of day 21 Dox treated mice had significantly lower proliferation compared to controls (Fig. 5D). To test if this decrease was due to a change in differentiation of the leukemia cells, we examined markers of mature myeloid cells, Mac-1 and Gr-1. In the spleen of Dox treated mice, we observed an increase in Mac1+, GR-1+ leukemia cells, consistent with differentiation. In contrast, there was no difference in the presence of Mac1 or GR-1 on the bone marrow leukemia cells (Fig. 5E). To test whether bone marrow CM KD cells were instead enriched for leukemia stem cell (LSC) markers, we examined expression of KIT, CD59a, IL1RL1, and CD123. Work by us and others previously showed that KIT, CD59a, and IL1RL1 are enriched on the LSC population driven by CM expression, and CD123 is an established LSC marker in human AML [20, 32, 33]. We observed a small yet statistically significant increase in CD123 and decrease in IL1RL1 in the leukemia cells in the bone marrow of Dox treated mice (Fig. 5F). These results indicate that the decreased proliferation observed in the bone marrow CM KD cells is not caused by differentiation to a post-mitotic mature myeloid cell or enrichment for LSCs.
To test if differences in the immune microenvironment could contribute to the survival of CM KD cells in the bone marrow, we examined the presence of the non-leukemic GFP- myeloid and lymphoid cells in the spleen and bone marrow of control and Dox treated mice at day 14 of KD. We found no differences in the percentage of granulocytes (GFP-, Mac-1+, Gr-1+) or monocytes (GFP-, Mac-1+, Gr-1-) with CM KD in either the spleen or the bone marrow, but a decrease in B-cells (GFP-, B220+) in the bone marrow of Dox treated mice (SFig. S4A, B). In both the spleen and the bone marrow, we observed no statistical difference in the total T-cell (GFP-, CD3+) or cytotoxic T-cell (GFP-, CD8+) populations with the KD of CM (SFig. S4C, E). Interestingly, both the GFP-, CD3+, and CD8+ populations showed increased expression of PD-1 and decreased Tim-1, which is suggestive of an exhausted phenotype (SFig. S4D, F) [34].
CM KD is associated with distinct changes in gene expression in spleen and bone marrow
To investigate the transcriptomic differences associated with the survival of CM KD leukemic cells in the bone marrow, but not the spleen, we performed bulk RNA sequencing (RNA-Seq) on GFP+ leukemia cells from both populations on day 7 of control or Dox treated mice. We identified a total of 348 differentially expressed genes (DEGs); (221 upregulated, 127 downregulated) in leukemia cells from the spleen of Dox treated animals, as compared to control. In leukemia cells from the bone marrow, we identified a total of 4650 DEGs (2328 upregulated, 2322 downregulated) in cells from Dox treated mice, as compared with control (Supplementary Table 2). Hierarchical clustering shows the bone marrow replicates from control and Dox treated mice have similar DEG patterns (SFig. S5A). While spleen control replicates cluster together, one Dox treated sample clustered more closely with the control spleen samples, suggesting variability in response to CM KD (SFig. S5B, C). In the bone marrow, several known targets of the CM fusion protein, including Cebpe, Cebpd, Runx1, and Il1rl1, are downregulated with CM KD (SFig. S5D). In addition, genes associated with an abnormal myeloid progenitor (AMP) phenotype, including Rhag, Cldn13, Add2, Gata2 were found to be upregulated (SFig. S5E) [32]. Overall, these changes in genes expression are consistent with loss of CM activity but imply that CM KD cells retain transcriptional changes associated with leukemic blasts (Fig. 6A).
Fig. 6: Increased expression of autophagy related genes in CM knockdown leukemia cells from the bone marrow.
A Volcano plot showing distribution of the statistically significant differentially expressed genes (DEGs) in the GFP+ sorted leukemia cells from the bone marrow (BM) of mice treated for 7 days with control (CTRL) or doxycycline (Dox). A few of the key genes with a log2 fold change ≥0.5 and -log10 (adjusted p value) are highlighted. B Venn diagram showing overlap of significantly upregulated DEGs in CM KD leukemia cells from the spleen (SP) and BM, as compared to CTRL. C Graph of top KEGG associated with upregulated DEGs in Dox treated cells from the BM. D. Box plot depicting relative expression of indicated autophagy associated genes in CTRL and Dox leukemia cells from the BM. N = 3 biological replicates, * = padj ≤0.05 as compared to CTRL.
To identify biological functions that promote survival of bone marrow CM KD cells, we compared DEGs in spleen and bone marrow leukemia cells and performed pathway analysis on only those genes altered in the bone marrow (Fig. 6B). Of the 2162 unique DEGs upregulated in bone marrow leukemia cells we identified autophagy, a known promoter of AML survival, as well as mTOR signaling, AMPK signaling, and mitophagy which have been implicated in promoting autophagy, survival, and stress adaptation in AML (Fig. 6C) (Supplementary Table 3) [35,36,37,38,39]. Analysis of individual autophagy related genes showed significantly increased expression of Atf6, Atg14, Atg2A, Becn1, and mTor (Fig. 6D). These results imply a role for autophagy in the survival of CM KD cells in the bone marrow. Pathway analysis of leukemia cells from the spleen indicated that the top deregulated pathway with KD of CM is apoptosis, consistent with the Annexin V staining and western blot analysis in Fig. 5 (SFig. S5F).
To determine if CM KD induces distinct gene expression changes in different subpopulations of leukemia cells, we performed single-cell RNA sequencing (scRNA-Seq). We harvested cells from the bone marrow and spleen of control and Dox treated mice at day 7 of treatment. Single cell suspensions were lineage depleted and analyzed by flow cytometry to confirm enrichment of leukemia cells. All samples showed >60% leukemia cells, with the bone marrow approaching 80% in both control and Dox treated samples after lineage depletion (SFig. S6A). Single-cell transcript libraries were prepared and sequenced.
Unbiased clustering analysis revealed multiple subpopulations of leukemia cells in both the spleen and bone marrow. This analysis shows that there is dramatic cellular heterogeneity among the leukemia cells in both bone marrow and spleen, with different subpopulations differentially dependent on CM expression (Fig. 7A, SFig. S6B). Interestingly, CM KD affected the relative proportion of each cluster, with the number of cells increased in some clusters while others decreased in samples from Dox treated mice (Fig. 7A, B). In the spleen, Clusters 2 and 4 appeared to be the most sensitive to CM KD (SFig. S6C). In the bone marrow, Clusters 3, 4, 5, 6, 9, and 10 showed marked reduction with CM KD while the number of cells in Clusters 0, 1, 2, 7 and 8 increased, indicating these clusters do not require CM expression for survival (Fig. 7B). To understand more about these clusters, we used the Mouse RNA-seq data (GSE14833 and GSE6506) in the Bloodspot Database to assign cell identifies [40]. We found many of the clusters in the spleen and bone marrow expressed markers of progenitor cells and mature myeloid cells, although we were not able to assign identities to all clusters (Fig. 7B, SFig. S6C, Supplementary Table 4).
Fig. 7: Single cell transcriptomics reveals distinct CM-dependent and CM-independent subpopulations in the bone marrow.
A Umap projection of leukemia cells from the bone marrow (BM) of control (CTRL) or doxycycline (Dox) treated mice. Each dot represents a single cell and is colored according to the assigned transcriptional cluster. B Bar graph of the number of cells in each cluster in BM samples from CTRL or Dox treated mice. C Heat map showing whether the of the 456 genes common among the five CM-independent clusters (0, 1, 2, 7, and 8) are expressed (blue) or undetected (yellow) across all clusters in the BM from CTRL mice. D Top 5 molecular functions associated with the 456 shared genes from (C).
To determine if the expansion of CM-independent clusters is related to altered expression of specific genes, we compared the genes differentially expressed between control and CM KD cells in each cluster. Surprisingly, we did not identify any genes or pathways commonly deregulated in CM-independent clusters but not CM-dependent clusters (Supplementary Table 4). This finding implies that the CM-independent clusters may have inherent transcriptional programs promoting their survival in the bone marrow. To test this, we compared the transcripts expressed in the clusters that do not require CM. We identified 456 genes expressed in all CM-independent clusters (SFig. S36D). Among these, several genes including Cxcl9, Ccr10, Elane, Ebf3, Col4a1, CD72, Cebpe, Ppbp, Prdm1, and Trbc are all associated with leukemia differentiation, survival and proliferation (Fig. 7C) [41,42,43,44,45,46,47]. Consistent with this finding, pathway analysis indicates that the genes commonly expressed in the CM-independent clusters are significantly enriched for molecular functions involving cytokine and chemokine activity (Fig. 7D) (Supplementary Table 5). In contrast, the common genes expressed in CM-dependent clusters were associated with non-immune molecular functions (SFig. S6E). Collectively, these results imply that cytokine and chemokine signaling within the bone marrow microenvironment may be a critical mechanism promoting survival of CM-independent leukemia cells.
CM KD leukemia cells can re-establish disease
The persistence of a CM KD leukemic population in the bone marrow raises the possibility that these cells may be able to re-establish disease. To test this possibility, we maintained mice transplanted with shMYH11 transduced leukemia cells on Dox or control drinking water until they showed signs of lethal disease, typically between day 28 and 42 of treatment. Notably, all control treated mice developed lethal disease by day 21, implying that KD of CM increases survival. All dox treated mice showed initial loss of GFP+ cells from the blood, but upon sacrifice, 10 out of 18 Dox treated mice showed re-emergence of GFP+ cells in the blood and/or spleen of at least 4% (Fig. 8A). To determine if these cells expressed CM, we sorted for GFP+ cells and performed qRT-PCR and western blot (Fig. 8B–D). We found that re-emergent GFP+ cells had reduced CM at both the mRNA and protein levels, as compared to parental leukemia cells from control treated mice. In addition, we found that the re-emergent CM KD leukemia cells showed similar expression levels of cKit, Mac-1, and Gr-1 to parental leukemia cells from control mice, implying they are arrested at a similar stage of differentiation (SFig. S7A, B). To understand the molecular mechanism allowing CM KD cells to survive in the blood, we performed whole transcriptome sequencing on the re-emergent and parental leukemia cells. The re-emergent GFP+ cells from mice with secondary (2°) disease were harvested from the spleen and compared with either spleen or bone marrow from control mice. Principal component analysis (PCA) showed that samples did not cluster with parental leukemia cells harvested from either the bone marrow or spleen indicating that 2° disease is accompanied by widespread alterations in transcription (Fig. 8E). In addition, only 3 of the 4 relapsed samples clustered together, implying that there may be multiple transcriptional programs that promote re-emergence of CM KD cells from the bone marrow. Comparing the transcriptome of relapsed samples to control, we observed a total of 930 DEGs (359 upregulated, 571 downregulated) in the 2° disease samples compared to parental leukemia cells (Supplementary Table 6). Further, pathway analysis indicates that the DEGs are associated with multiple pathways implicated in leukemia survival and relapse, including interferon regulator factor (IRF) signaling (Fig. 8F) (Supplementary Table 7). Specifically, Irf1, Irf7, and Irf9 are decreased in the 2° disease samples compared to controls (Fig. 8G). Notably, loss of these genes is associated with relapse in AML patients [46,47,48]. To identify potential driver mutations that cause 2° disease, we compared our RNA-Seq data to the mouse reference genome and identified nucleotide variants. We selected variants with a high likelihood to have deleterious consequences on the encoded protein in each 2° disease sample but not found in any of the control samples. The majority of the genes with potentially deleterious variants were unique to each 2° disease sample (Supplementary Table 8). None of the identified variants were in established oncogenes or tumor suppressor genes, but pathway analysis showed they were associated with pathways affecting gene and protein expression, as well as kinase activity (SFig. S7C).
Fig. 8: The surviving CM knockdown cells in the bone marrow can give rise to secondary disease.
A Line graph showing the percentage (%) of GFP+, shMYH11 transduced cells in PB of mice treated for up to 42 days with either CTRL or Dox. Each line represents individual mice monitored longitudinally. Day -5 refers to the GFP burden 5 days prior to starting treatment at Day 0. N = 5 control biological replicates and N = 12 biological replicates. B Bar graph of relative CM expression in GFP+ sorted leukemia cells from Dox treated mice with secondary (2°) disease, defined as the re-emergence of >4% GFP+ cells in the blood. Data shown are from mice in A and a second cohort exhibiting similar trends of re-emergent disease. N = 3 control and no 2˚ disease biological replicates, N = 5 2˚ disease biological replicates, will all biological replicates assayed in triplicate. C Representative western blot and D. Bar graph of relative CBFβ::SMMHC levels from sorted GFP+ cells harvested from CTRL or Dox treated mice N = 4 biological replicates. E Principal Component Analysis (PCA) plots of whole transcriptome sequencing data from GFP sorted 2° disease (N = 4) or CTRL (N = 3) leukemic cells harvested from the SP or BM. F Pathway analysis of the differentially expressed genes in 2° disease samples compared to the combined SP and BM CTRL leukemia samples. G Bar graph depicting the Irf genes deregulated in Dox treated mice after development of 2° disease. * = p ≤ 0.05 as compared to CTRL.
Collectively, this study indicates that while CM is required for the survival of most frank leukemia cells, a population of leukemia cells in the bone marrow do not strictly require CM activity and can re-establish fatal disease.
