Despite significant progress in understanding the genetic landscape of T-cell acute lymphoblastic leukemia (T-ALL), the discovery of novel therapeutic targets has been difficult. Our results demonstrate that the levels of PIM1 protein kinase is elevated in early T-cell precursor ALL (ETP-ALL) but not in mature T-ALL primary samples. Small-molecule PIM inhibitor (PIMi) treatment decreases leukemia burden in ETP-ALL. However, treatment of animals carrying ETP-ALL with PIMi was not curative. To model other pathways that could be targeted to complement PIMi activity, HSB-2 cells, previously characterized as a PIMi-sensitive T-ALL cell line, were grown in increasing doses of PIMi. Gene set enrichment analysis of RNA sequencing data and functional enrichment of network modules demonstrated that the HOXA9, mTOR, MYC, NFκB, and PI3K-AKT pathways were activated in HSB-2 cells after long-term PIM inhibition. Reverse phase protein array–based pathway activation mapping demonstrated alterations in the mTOR, PI3K-AKT, and NFκB pathways, as well. PIMi-tolerant HSB-2 cells contained phosphorylated RelA-S536 consistent with activation of the NFκB pathway. The combination of NFκB and PIMis markedly reduced the proliferation in PIMi-resistant leukemic cells showing that this pathway plays an important role in driving the growth of T-ALL. Together these results demonstrate key pathways that are activated when HSB-2 cell line develop resistance to PIMi and suggest pathways that can be rationally targeted in combination with PIM kinases to inhibit T-ALL growth.

This article is featured in Highlights of This Issue, p. 1761

T-cell acute lymphoblastic leukemia (T-ALL) is an aggressive hematopoietic malignancy that is treated by intensive chemotherapy (1). To improve risk stratification, patients with T-ALL are classified into distinct subtypes via immunophenotyping and gene expression analysis (2, 3). Early T-cell precursor ALL (ETP-ALL) is a high-risk subtype of T-ALL with a unique immunophenotype including myeloid and early progenitor in addition to T-cell lineage markers (4). Whole-genome and -transcriptome sequencing of ETP-ALL reveals several activating gene fusions, alterations, and mutations within IL7R and JAK-STAT pathways that could result in constitutively activated tyrosine kinases (5–7), whereas non-ETP or mature ALL often display mutations in PI3K/AKT and NOTCH signaling pathways (8). Despite the advances that have been made in decoding the genetic background of T-ALL, the translation of novel targeted therapies toward clinical practice has remained elusive (8).

Our group and others have shown that the PIM (Provirus Integration sites for Moloney murine leukemia virus) kinase is a potential target in T-ALL (9–13). Recently, two studies identified a similar TCRβ-PIM1 translocation in patients with T-ALL that is essential for aberrant activation of PIM1 kinase (12, 13). The PIM kinases are a highly conserved family of three serine/threonine kinases involved in cell cycle, apoptosis, transcription, translation, cell metabolism, and drug resistance through phosphorylation of downstream targets (14–16). PIM kinase levels are increased by cytokines that stimulate JAK-STAT signaling (17). Previously, we reported that PIM1 is overexpressed in ETP-ALL and a small percentage non-ETP-ALL patient-derived xenografts (PDX; ref. 10), suggesting that this kinase may be an ideal therapeutic target. Using ETP-ALL PDXs, we demonstrate the PIMi activity in this disease. To learn more about the signal transduction pathways that are essential for T-ALL cells to survive and potentially complement PIMi, we have modeled the pathways that are activated using both proteomic and transcriptomic approaches when cells develop resistance to these inhibitors. A key finding from these analyses is that PIMi-resistant cells contain an activated NFκB pathway. The combination of NFκB and PIM inhibitors (PIMi) therapy suppresses T-ALL growth, suggesting an approach that can be investigated to improve therapy for a subset of patients with T-ALL with increased PIM kinase activity.

T-ALL PDXs engraftment in NCG (NOD CRISPR Prkdc Il2rγ) mice

All in vivo experiments were conducted on protocols approved by the Institutional Animal Care and Use Committee of the University of Arizona (Tucson, AZ). Deidentified T-ALL PDXs were supplied by Children's Cancer Institute, Children's Oncology Group, and St. Jude Children's Research Hospital. These PDX samples had previously been characterized by gene expression profiling, whole-exome or -genome sequencing, or targeted sequencing of a gene panel (7, 18). NCG mice were irradiated and inoculated by tail-vein injection with 5 million cells per 100-μL PBS. Mice were monitored for leukemia engraftment by flowcytometric analysis of peripheral blood (PB) using antibodies against human and mouse CD45. At necropsy, spleens were minced and mononuclear cells were cryopreserved for subsequent experiments.

Human T-ALL cell lines and cell culture

T-ALL cell lines including HSB-2, SUP-T1, DU.528, CUTLL1, KOPT-K1, and HPB-ALL were a kind gift of Dr. Jon C. Aster (Department of Pathology, Brigham and Women's Hospital, Boston, Massachusetts, USA) and cultured in RPMI supplemented with 2 mmol/L glutamax and 10% FBS. All cell lines were maintained at 37°C in 5% CO2 and were authenticated by short tandem repeat DNA profiling performed by the University of Arizona Genetics Core Facility (Tucson, AZ). The cell lines were routinely tested for Mycoplasma and used for fewer than 50 passages.

Derivation of PIMi-resistant HSB-2 cells

PIMi resistance in HSB-2 cells was induced by growing cells in the presence of increasing concentrations of AZD1208 (AZDR1 and AZDR2) or LGB321 (LGBR2) over a 4-month period until resulting cell populations were able to maintain their growth in 1 μmol/L AZD1208 or LGB321. While the HSB-2 naïve cells had an IC50 < 1 μmol/L, all resistant populations had IC50 > 10 μmol/L when incubated with AZD1208 or LGB321 over the 3-day XTT growth study.

Establishment of xenograft models

Xenograft models were established as described previously (7). All experiments were conducted on protocols approved by the Institutional Animal Care and Use Committee of the University of Arizona (Tucson, AZ). For efficacy studies, sublethally irradiated (2.0 Gy) NCG mice (Charles River) were randomized to treatment or vehicle (5 mice/arm) once xenografts had engrafted with sufficient disease burden to detect 1% peripheral human CD45+ cells. PIMi (AZD1208; 30 mg/kg) or vehicle was administered for the entire 3-week treatment period by oral gavage. AZD1208 was formulated using a Cremophore/Ethanol/PBS—24/6/70 ratio. Studies were terminated, and mice were euthanized after 3 weeks of treatment. PB, spleen, and bone marrow (BM) were harvested at the time of necropsy. Disease burden was assessed every 4–5 days by flow cytometric measurement of human CD45+ cells in PB, and at necropsy by measuring leukemia burden by analyzing the percent of human CD45+ cells in PB, spleen, and BM by flow cytometry.

Cell viability assay

Cryopreserved PDX cells were thawed at 37°C and washed in RPMI1640 media supplemented with 10% FBS, 100 mg/mL streptomycin, 100 U/mL penicillin, and 2 mmol/L l-glutamine. After centrifugation (500 × g for 5 minutes) and a second wash, cell viability was determined using Trypan blue exclusion assay. Cells were resuspended at the appropriate cell density in QBSF-60/F [containing QBSF-60 media (Quality Biological) supplemented with 20 ng/mL FMS-like tyrosine kinase 3 ligand (Flt-3L; ProSpec), 100 U/mL penicillin, 100 mg/mL streptomycin, and 2 mmol/L l-glutamine]. T-ALL PDX cells (5 × 104 cells/well) were seeded into 96-well plates and treated with PIMi (AZD1208 and LGB321) as indicated. DMSO was added to keep concentrations of DMSO (<0.1%) equal in all wells. After 72 hours, cell viability was measured by using an adenosine triphosphate–based luminescence assay (ATPlite 1-step; PerkinElmer) according to the manufacturer's protocol. Because PDX cells do not proliferate in these conditions, this assay reflects cell survival only. The survival of DMSO control cells was considered as 100% and percent of cell death after individual treatments is reported relative to the DMSO control.

qRT-PCR

To assess the differential expression of genes related to development of PIMi resistance in HSB-2 cells, total RNA was isolated the RNeasy Kit (Qiagen, catalog no. 74104). cDNA was synthesized from 1 μg RNA using the iScript cDNA Synthesis Kit (Bio-Rad, catalog no. 1708890) according to the manufacturer's instructions. qRT-PCR amplification and analysis were conducted using the SsoAdvanced Universal SYBR Green Supermix (Bio-Rad, catalog no. 1725271) and CFX96 Real-Time System. Samples were assayed in triplicate and RNA levels were normalized to 18S expression levels. Primer sequences for qRT-PCR are provided in the Supplementary Table S1.

Antibodies and reagents

The following antibodies were purchased from Cell Signaling Technology: anti-phospho- IRS1 (S1101, catalog no. 2385), anti-phospho-S6 (S240/244, catalog no. 5364), anti-phospho-TSC2 (S939, catalog no. 3615), anti-TSC2 (catlog no. 3990), anti-PTEN (catalog no. 138G6), anti-Phospho-NFκB p65 (S536, catalog no. 93H1), anti-NFκB p65 (catalog no. D14E12), anti-phospho-4EBP1 (S65, catalog no. 13443), anti-4EBP1(catalog no. 9644), and anti-phospho-GSK3β (S9, catalog no. 5558). Anti-IRS1 (catalog no. 06-248) antibody was purchased from Merck Millipore. Anti-DDIT4 (catalog no.10638-1-AP) and anti-SESN2 (catalog no. 10795-1-AP) antibodies were purchased from ProteinTech Group. Horseradish peroxidase (HRP)-conjugated anti-β-actin (catalog no. A3854) was purchased from Sigma. HRP-linked mouse IgG (catalog no. NA931V) and rabbit IgG (catalog no. NAV934V) were purchased from GE Healthcare Life Sciences. IgG (Isotype control catalog no. 400107), FITC-conjugated anti-human CD45 (catalog no. 304006), and APC-conjugated anti-mouse CD45 (catalog no. 103112) were purchased from BioLegend. Anti-S6 (catalog no. SC74459) antibody was purchased from Santa Cruz laboratories. LGB-321 (catalog no. A14420), AZD1208 (catalog no. A13203), and IKK16 (catalog no. A12836) were purchased from AdooQ Bioscience. TNFα (catalog no. PHC3015) was purchased from Thermo Fisher Scientific.

Supplementary Methods include a description of the Western blot analysis, RNA-seq data analysis, gene set enrichment analysis (GSEA), RPPA, active module analysis, lentiviral production, the NFκB reporter assay, and statistics.

RNA-seq data is available in the Gene Expression Omnibus under accession number GSE135545.

PIM kinase inhibition decreases leukemia burden in ETP-ALL PDXs

We have examined the level of three PIM isoforms in six ETP-ALL and four mature T-ALL PDXs. Importantly, qRT-PCR of these 10 PDXs demonstrated that PIM1 expression, but not PIM2 or PIM3, was elevated specifically in ETP-ALL and not in the mature T-ALL PDXs (Fig. 1A). PIM1 protein expression was high in ETP-ALL as compared with mature T-ALL PDXs (Supplementary Fig. S1A) and PIM1 expression was correlated with STAT5 activation. This result suggested that the PIM1 could be targeted to inhibit ETP-ALL growth. To examine whether these PDXs are sensitive to PIMi, the cells were incubated with two pan-PIMis AZD1208 (19) or LGB321 (20). LGB321 is a tool compound related to Pim447, which is being used to treat patients (21). PIMi suppressed ETP-ALL growth by 25% to 50% while the effects on more mature T-ALL PDXs was significantly less (Fig. 1B).

Figure 1.

Pan-PIM kinase inhibitors reduces leukemia burden in ETP-ALL PDX models. A, Relative mRNA expression of PIM1, PIM2, and PIM3 in T-ALL PDX cells. B, T-ALL PDX cells were incubated with the AZD1208 or LGB321 (AZD 3 μmol/L or LGB 1 μmol/L) for 72 hours and then the percentage of viable cells was quantified by the ATPlite assay. The growth of DMSO control cells was considered 100% and percent cell growth after individual treatment is reported relative to the DMSO. The data shown are the average ± SD of three independent experiments. C–E, The PB, spleen, and BM were harvested and analyzed hCD45+ cells by flow cytometry. Percent hCD45+ data shown are the average ± SD from five samples per treatment. F, Western blot (WB) analysis of lysates collected from BM cells using the specified antibodies. Each lane on the WB represents an individual tumor.

Figure 1.

Pan-PIM kinase inhibitors reduces leukemia burden in ETP-ALL PDX models. A, Relative mRNA expression of PIM1, PIM2, and PIM3 in T-ALL PDX cells. B, T-ALL PDX cells were incubated with the AZD1208 or LGB321 (AZD 3 μmol/L or LGB 1 μmol/L) for 72 hours and then the percentage of viable cells was quantified by the ATPlite assay. The growth of DMSO control cells was considered 100% and percent cell growth after individual treatment is reported relative to the DMSO. The data shown are the average ± SD of three independent experiments. C–E, The PB, spleen, and BM were harvested and analyzed hCD45+ cells by flow cytometry. Percent hCD45+ data shown are the average ± SD from five samples per treatment. F, Western blot (WB) analysis of lysates collected from BM cells using the specified antibodies. Each lane on the WB represents an individual tumor.

Close modal

To evaluate the PIMi ability to block leukemic growth in vivo, ETP13 PDXs were injected intravenously into NCG mice. ETP13 PDX was selected for testing because of the ability of this PDX to rapidly engraft in the mice (Supplementary Fig. S1B) and its intermediate PIM1 expression. Mice were treated with AZD1208 (10) or vehicle once daily for a duration of 3 weeks (Supplementary Fig. S1C) without significant change in body weight (Supplementary Fig. S1D). Our results demonstrate markedly decreased hCD45+ leukemic cells in the PB (∼70% decrease; Fig. 1C; Supplementary Fig. S1E), spleen (∼50% decrease; Fig. 1D), and BM (30% decrease; Fig. 1E) from the AZD1208-treated mice when compared with the vehicle. The modest response of PIMi in BM suggests a possible nutrient rich microenvironment in the BM that might activate compensatory signaling pathways in these leukemic cells (22) and protect from PIMi. In addition, AZD1208-treated mice showed reduced spleen size (Supplementary Fig. S1F) and a red colored BM (Supplementary Fig. S1G) indicating reestablishment of normal hematopoiesis and decreased leukemic burden. Western blot analysis of total BM cells collected after 3 weeks of drug treatment (Fig. 1F) showed a significant decrease in IRS1 phosphorylation, a PIM kinase substrate, demonstrating PIMi activity in the BM cells of these mice (23). As expected from our previous results (10), PIMi treatment also decreased mTORC1 signaling as shown by decreased S6 and 4EBP1 phosphorylation levels. These results supported the potential use of PIMi treatment to decrease tumor load in patients with ETP-ALL.

Long-term PIM inhibition in HSB-2 cells exhibit altered gene and protein expression

The observation that after 3 weeks of treatment, leukemic cells remained in the BM and spleen suggested that there were additional pathways being stimulated, which continued to support T-ALL growth even when the PIM kinases were inhibited. To model the biochemical nature of these pathways, previously characterized PIMi-sensitive T-ALL cell line HSB-2 (10) was grown in the presence of increasing concentrations of AZD1208 or LGB321 until the cell populations grew continuously in 1 μmol/L AZD1208 or LGB321. Previously, we have shown this cell line has gene expression signature that is consistent with other ETP-ALL cells (10). The approach used to develop PIMi resistance cells was based on published study examining resistance to a NOTCH inhibitor in T-ALL (24). Elevated levels of LCK expression secondary to t(1;7)(p34;q34) chromosomal translocation is found in the HSB-2 cells (25). However, using six T-ALL cell lines, we showed that the PIMi sensitivity was independent of LCK expression (Supplementary Fig. S2A and S2B). Sensitivity to PIMi in HSB-2 cells is comparable with that found in DU.528 and KOPT-K1 cell lines that do not have LCK translocation. Our group and others have shown (10, 26, 27) in various cancers that PIMi sensitivity is correlated with the level of PIM expression. We did not observe any correlation between PIMi sensitivity (Fig. 1B) and LCK expression in T-ALL PDXs as shown in Supplementary Fig. S2C. While the HSB-2 naïve cells had an IC50 of less than 1 μmol/L for these drugs, the drug-tolerant cell lines had IC50 values greater than 10 μmol/L with either resistance to AZD1208 (AZDR1 and AZDR2) or LGB321 (LGBR2; Fig. 2A and B). The drug-tolerant cell lines behaved similarly to the mature T-ALL cell line, SUP-T1 which is insensitive to PIMi (10), suggesting that they likely had alternate pathways activated that maintained their growth even when incubated with the inhibitor. Like in HSB-2 naïve and SUPT1 cells incubated short term with PIM kinase inhibitor (24 hours), PIMi-resistant cell lines demonstrated a complete blockade of IRS1 phosphorylation. Importantly, this indicated that incubation with AZD1208 or LGB321 either short- or long-term blocked PIM kinase activity (23), thus demonstrating that the drug was entering into these resistant cells. However, PIMi treatment reduced mTORC1 activity only in HSB-2 naïve cells where growth is inhibited, but not in PIMi-resistant cell lines or the naturally insensitive SUPT1 cells (Fig. 2C; Supplementary Fig. S2D). These results suggest that mTORC1 plays a role in resistance to the PIMi while IRS1, although a good PIM kinase substrate did not regulate cell growth when the PIM pathway was blocked. This result demonstrates that, when PIM is inhibited the mTORC1 is critical to maintain cell growth and thus could be targeted as part of a long-term combination treatment strategy.

Figure 2.

Characterization of HSB-2 naïve and PIMi-resistant cells. A and B, Using an XTT assay, the percentage of viable cells was measured after each cell line was exposed to increasing doses of AZD or LGB over 72 hours and compared with vehicle. Data are represented as average ± SD of three independent experiments. C, WB comparing the mTOR signaling activity in HSB-2 naïve (AZD1 μmol/L; 24 hours) and PIMi-resistant cell populations. D, Hierarchical clustering on the differentially expressed genes with high-significance (FDR < 0.01) from AZDR RNA-seq data reveals significantly enriched biological pathways. The heatmap represents the row z-score of log2-transformed RNA-seq count values. Optimally clustered (by silhouette width) membership is shown with colors on the left side of the heatmap while clusters cut at height h = 2 are shown with colors on the right side of the heat map. Selected GO term annotations of each cluster are represented on the right side of the heatmap. Total of 2,225 genes with n = 3 in each group were analyzed for the heatmap clustering analysis.

Figure 2.

Characterization of HSB-2 naïve and PIMi-resistant cells. A and B, Using an XTT assay, the percentage of viable cells was measured after each cell line was exposed to increasing doses of AZD or LGB over 72 hours and compared with vehicle. Data are represented as average ± SD of three independent experiments. C, WB comparing the mTOR signaling activity in HSB-2 naïve (AZD1 μmol/L; 24 hours) and PIMi-resistant cell populations. D, Hierarchical clustering on the differentially expressed genes with high-significance (FDR < 0.01) from AZDR RNA-seq data reveals significantly enriched biological pathways. The heatmap represents the row z-score of log2-transformed RNA-seq count values. Optimally clustered (by silhouette width) membership is shown with colors on the left side of the heatmap while clusters cut at height h = 2 are shown with colors on the right side of the heat map. Selected GO term annotations of each cluster are represented on the right side of the heatmap. Total of 2,225 genes with n = 3 in each group were analyzed for the heatmap clustering analysis.

Close modal

To further understand the differences between naïve and resistant cell lines, we performed RNA sequencing (RNA-seq) on HSB-2 naïve and AZDR1 cells finding 3,346 genes differentially expressed (FDR < 0.05). To evaluate the function of these expression changes, we used the 2,225 genes with highest significance (FDR < 0.01) as input for hierarchical clustering (Fig. 2D). Silhouette width analysis on the resulting dendrogram indicated that the optimal number of clusters is 2 (Supplementary Fig. S3A–S3C). These clusters were enriched for genes in RNA processing, immune response, inflammatory response, and cell migration (Supplementary Table S2). Further cutting the tree down (at height of 2) led to 14 clusters, and Gene Ontology (GO) analysis showed that upregulated genes in AZDR1 cells were enriched for mitochondrial DNA repair, leukocyte differentiation, T follicular helper cell differentiation, and RNA processing genes. In contrast, genes involved in ephrin receptor signaling pathway, immune response, lymphocyte activation, NFκB signaling, and glycolysis were downregulated (Fig. 2D; Supplementary Table S3). We note that GO gene sets include both activators and inhibitors of the annotated process, so upregulation or downregulation of the mRNA expression values does not necessarily correspond to activation or repression of the corresponding signaling pathway but suggests these pathways are modified in cells after long-term PIM inhibition. We validated the RNA-seq results by performing qPCR on genes with the most significant (FDR < 1 × 10−17) change in mRNA expression and demonstrated that BMPR1B, WNK2, SCEL, and DEFA6 genes were upregulated, whereas HKDC1 and DDIT4 were downregulated in AZDR1 compared with naïve HSB-2 cells (Supplementary Fig. S3D).

Activation of mTOR, MYC, and disruption of NFκB and HOXA9 signaling in long-term PIMi-treated HSB-2 cells

In addition to clustering the differentially expressed genes, we used GSEA to compare global gene expression in HSB-2 naïve and AZDR1 cells. Genes were ranked by differential expression P-value and by upregulation or downregulation and assessed for enrichment in curated canonical pathways and oncogenic signatures (Supplementary Table S4). In agreement with published studies (13, 28), in AZDR1 cells, glycolysis was significantly changed while cell cycle–related processes were upregulated (Supplementary Figs. S4 and S5A; Supplementary Table S4). Similarly, expression of genes in pathways such as JAK-STAT, mTOR, MAPK, and cytokine signaling were significantly altered in AZDR1 cells suggesting that continued PIMi treatment is associated with major changes in the signal transduction pathways that regulate cell survival and growth. GSEA identifies these pathways as key to the survival of HSB-2 cells.

To discover cancer drivers that may not themselves be altered at the mRNA level, but might cause the observed changes in gene expression, we carried out GSEA using the “oncogenic perturbation” gene sets in MSigDB (29). We found evidence that several oncogenic pathways were activated in AZDR1 cells, including mTOR, HOXA9, RelA, and MYC. (Fig. 3; Supplementary Fig. S5B–S5E; Supplementary Table S5). A recent study showed increased HOXA9 expression in T-ALL cases with JAK-STAT activation, and in these cases PIM kinase was identified as a possible drug target (11). RelA (NFκB P65) is a major subunit of NFκB and its activation has been shown to promote survival in leukemic cells (30). MYC and PIM kinase activities are known to be regulated similarly. MYC overexpression is a common driver of many hematologic cancers (31, 32). Together, these results identify multiple compensatory pathways that are activated in cells surviving after prolonged PIMi treatment.

Figure 3.

Oncogenic signatures enriched in transcriptome of long-term PIMi-treated cells compared with HSB-2 naïve cells. Genes were ranked by differential expression and analyzed using GSEA testing against curated oncogenic signatures. The size of the bubbles in the bubble plots represents the normalized enrichment score (NES), which summarizes the number and differential expression intensity of enriched genes. The absolute value of the NES generally varies between 0 (no enrichment) and 5 (extremely high enrichment). The significance of the pathway enrichment is measured by the FDR (x-axis). Open bubbles indicate downregulated pathways and closed indicate upregulated pathways. All oncogenic signatures shown have FDR < 0.25 and GSEA significance is defined as FDR < 0.25.

Figure 3.

Oncogenic signatures enriched in transcriptome of long-term PIMi-treated cells compared with HSB-2 naïve cells. Genes were ranked by differential expression and analyzed using GSEA testing against curated oncogenic signatures. The size of the bubbles in the bubble plots represents the normalized enrichment score (NES), which summarizes the number and differential expression intensity of enriched genes. The absolute value of the NES generally varies between 0 (no enrichment) and 5 (extremely high enrichment). The significance of the pathway enrichment is measured by the FDR (x-axis). Open bubbles indicate downregulated pathways and closed indicate upregulated pathways. All oncogenic signatures shown have FDR < 0.25 and GSEA significance is defined as FDR < 0.25.

Close modal

Active network module analysis reveals PIMi-resistant cells have an active protein complex enriched with cancer-associated signaling pathways

We identified “modules” of interacting proteins that are transcriptionally dysregulated in AZDR cells. Using the active network modules (ANM) approach (28), we overlaid adjusted P-values from the RNA-seq differential expression results comparing drug-resistant and vehicle-treated HSB-2 cells over the STRING protein–protein interaction network (33). We then applied the jActiveModules algorithm to find differentially expressed protein complexes (34). Figure 4A shows the statistically significant ANM that was identified.

Figure 4.

ANM analysis on long-term PIMi-treated cells finds changes in protein complex enriched for signaling and growth pathways. A, ANM analysis is shown with nodes representing genes and edge length representing the strength of the protein–protein interaction, with shorter edge lengths indicating higher weight interactions according to STRING database. The ellipse shaped nodes represent genes in either apoptosis, EGFR tyrosine kinase inhibitor resistance (EGFR-TKIR), JAK/STAT, mTOR, NFκB, or PI3K-AKT signaling pathways. The color of the nodes (fill color) indicates the log2 (fold change) of the RNA-seq expression in AZDR1 cells compared with HSB-2 naïve cells. The red fill indicates significantly increased (FDR < 0.05) proteins and blue fill indicates significantly decreased (FDR < 0.05) proteins while grey fill indicates not significant (FDR ≥ 0.05). The border color of each node represents pathway membership (e.g., yellow bordered genes are part of the PI3K-AKT signaling pathway). B, Enrichment of KEGG pathways in active module. The size of the bubble represents gene ratio calculated by taking the number of pathway-related genes within the active modules divided by the total number of genes in the pathway. The significance of the pathway enrichment is measured by the FDR (y-axis). Significance is defined as FDR < 0.05.

Figure 4.

ANM analysis on long-term PIMi-treated cells finds changes in protein complex enriched for signaling and growth pathways. A, ANM analysis is shown with nodes representing genes and edge length representing the strength of the protein–protein interaction, with shorter edge lengths indicating higher weight interactions according to STRING database. The ellipse shaped nodes represent genes in either apoptosis, EGFR tyrosine kinase inhibitor resistance (EGFR-TKIR), JAK/STAT, mTOR, NFκB, or PI3K-AKT signaling pathways. The color of the nodes (fill color) indicates the log2 (fold change) of the RNA-seq expression in AZDR1 cells compared with HSB-2 naïve cells. The red fill indicates significantly increased (FDR < 0.05) proteins and blue fill indicates significantly decreased (FDR < 0.05) proteins while grey fill indicates not significant (FDR ≥ 0.05). The border color of each node represents pathway membership (e.g., yellow bordered genes are part of the PI3K-AKT signaling pathway). B, Enrichment of KEGG pathways in active module. The size of the bubble represents gene ratio calculated by taking the number of pathway-related genes within the active modules divided by the total number of genes in the pathway. The significance of the pathway enrichment is measured by the FDR (y-axis). Significance is defined as FDR < 0.05.

Close modal

We next tested the genes in the ANM for functional enrichment in Kyoto Encyclopedia of Genes and Genomes (KEGG) pathways (33). Using this approach many pathways that are found activated in tumor cells, including MAPK, PI3K-AKT, NFκB, p53, glycolysis, and apoptosis were among the most enriched pathways (Fig. 4B; Supplementary Table S6). This set of pathway enrichments in the ANM is consistent with results in different tumor types (13, 28, 35, 36). It is possible that the protein–protein network from any literature-based database, including STRING, could be biased toward known cancer pathways. To check for such a bias, we simulated data by randomly assigning P-values to sets of genes in the RNA-seq data and performed jActiveModules search to show that the results above could not have arisen through intrinsic bias in the PPI network (Supplementary Fig. S6A–S6D; Supplementary Tables S7–S10). The ANM approach allows the discovery of not only enriched pathways, but also reveals new pathway members that are not altered at the mRNA level but interact strongly with genes that are differentially expressed. These proteins identified in HSB-2 cells could be important candidates for further study of essential signaling in T-ALL with high PIM1 expression (Fig. 4A).

Long-term PIM inhibition in cells exhibit protein level alterations in mTOR, PI3K-AKT, and NFκB signaling

To gain a deeper insight into the functional relationship between PIMi-dependent tumor suppression and the development of drug tolerance, we measured the relative expression/activation levels of 200 proteins and phosphoproteins using reverse phase protein arrays (RPPA). A pairwise differential abundance analysis between five treatment groups compared HSB-2 naïve or short-term (AZD1 μmol/L/LGB1 μmol/L) with long-term (AZDR1/LGBR2) PIMi-treated cells (Fig. 5AD). When compared with naïve cells, the long-term PIMi-treated cells exhibited differential abundance of 79 proteins in AZDR1 and 95 proteins in LGBR2 (Fig. 5A and C). When compared with the short-term PIMi-treated cells, only 17 proteins in AZDR1 and 9 proteins in LGBR2 were differentially abundant (FDR<0.05; Fig. 5B and D). This pattern was visualized by volcano plots (Fig. 5AD; Supplementary Tables S11–S14).

Figure 5.

AZD1208 and LGB321 resistant HSB-2 cells exhibit significant changes in protein expression. A–D, Volcano plot [log2(fold change) as x-axis and −log10(P) as y-axis] of RPPA differential expression comparing HSB-2 naïve versus AZDR1 (A), AZD 1 μmol/L (short-term) versus AZDR1 (long-term; B), Naïve versus LGBR2 (C), and LGB 1 μmol/L (short-term) versus LGBR2 (long-term; D) indicate that many proteins have significant changes in expression. Dots indicate Benjamin and Hochberg FDR < 0.05 and x-mark indicates FDR ≥ 0.05. E, Validation of RPPA data generated from comparing HSB-2 naïve and AZDR1 or LGBR2 cells by WB. F and G, RelA-S536 abundance is shown on the y-axis and time of treatment with PIMi on the x-axis. Total of 200 proteins were analyzed with n = 6 for each sample. Results are expressed as mean±SEM; ***P <0.001.

Figure 5.

AZD1208 and LGB321 resistant HSB-2 cells exhibit significant changes in protein expression. A–D, Volcano plot [log2(fold change) as x-axis and −log10(P) as y-axis] of RPPA differential expression comparing HSB-2 naïve versus AZDR1 (A), AZD 1 μmol/L (short-term) versus AZDR1 (long-term; B), Naïve versus LGBR2 (C), and LGB 1 μmol/L (short-term) versus LGBR2 (long-term; D) indicate that many proteins have significant changes in expression. Dots indicate Benjamin and Hochberg FDR < 0.05 and x-mark indicates FDR ≥ 0.05. E, Validation of RPPA data generated from comparing HSB-2 naïve and AZDR1 or LGBR2 cells by WB. F and G, RelA-S536 abundance is shown on the y-axis and time of treatment with PIMi on the x-axis. Total of 200 proteins were analyzed with n = 6 for each sample. Results are expressed as mean±SEM; ***P <0.001.

Close modal

In comparison with naïve cells, long-term PIMi-treated cells demonstrate significant changes in the abundance/activation state of various phospho-proteins, for example, GSK3, NFκB p65 (RelA), Tyk2, Rb, and ErbB2, and changes in total proteins including LAMP2, TGFβ, PTEN, and TNFα. As predicted from our previous studies and those of others, ERK phosphorylation at the T202/Y204 sites (9, 10) was significantly upregulated in HSB-2 cells after long-term PIM inhibition. Both TNFα and TGFβ protein levels were increased in AZDR1 and LGBR2 cells (Fig. 5A and C; Supplementary Tables S11 and S13), and TGFβ levels were also induced in HSB-2 cells after short-term PIMi treatment (Fig. 5B and D; Supplementary Tables S12 and S14), suggesting a possible compensatory mechanism that might maintain PIM kinase levels.

We then validated RPPA data comparing HSB-2 naïve and PIMi short- and long-term treated cells by Western blot analysis (Fig. 5E). We found an increased expression of NFκB and activation of the AKT pathway possibly induced by decreased PTEN levels. Activated AKT after long-term PIM inhibition caused an increase in GSK3β phosphorylation in these resistant HSB-2 cells. Activation of the AKT pathway after long-term PIM inhibition has been noted previously in other cell types (35). The protein levels of DDIT4 and SESN2 (both negative regulators of mTORC1) were downregulated and TSC2 phosphorylation was increased in long-term PIMi-treated cells. The observation that the mTORC1 pathway was activated as part of PIMi resistance was consistent with our finding in cell culture and on RNA-seq. In comparison, short-term PIMi treatment of naïve HSB-2 cells showed significant suppression of mTORC1 signaling (Fig. 2C) while in AZDR1, AZDR2, and LGBR2 cells, PIMi treatment showed no effect on this pathway.

We observed that the differentially abundant proteins in long-term PIMi-treated cells found by RPPA were significantly enriched in the ANM constructed from transcriptomic data (P = 0.0211 by hypergeometric test), with the overlapping proteins including NOS3, IGF1R, PDK1, PRKCD, STAT4, and CASP3. Therefore, both the gene and protein expression changes in long-term PIMi-treated cells converge on common biological functions.

Targeting NFκB signaling overcomes resistance to PIMi treatment

Computational analysis of both RNA-seq and RPPA data indicated activation of signaling pathways including EGFR, HOXA9, mTOR, MYC, NFκB, and PI3K-AKT after long-term PIM inhibition in HSB-2 cells. In other cell types, the EGFR, mTOR, and PI3K-AKT pathway activity have been closely associated with PIMi resistance (35–37). However, the NFκB activation has not previously been identified. After 48-hour PIMi treatment, HSB-2 cells exhibited a significant increase in the phosphorylated form of RelA (NFκB P65) on S536. This modification denotes transcription factor activation (Fig. 5F and G). As seen on the RPPA, the heightened level of NFκB-S536 phosphorylation was maintained during long-term PIM inhibition with either AZD1208 (F2,15 = 24.57, P = 1.85 × 10−5) or LGB321 (F2,15 = 17.43, P = 0.00012; Fig. 5F and G). This result is consistent with the RNA-seq analysis which also identified increases in RelA in HSB-2-AZDR cells.

A possible mechanism that could maintain NFκB activity in PIMi-treated HSB-2 cells is the autocrine secretion of growth factors that activate signaling pathways that stimulate NFκB. To test this possibility, conditioned media (CM) from HSB-2 naïve or AZDR1 or LGBR2 cells was incubated with HEK-293T cells stably expressing lenti-NFκB-luc-GFP (38). The CM from long-term PIMi-treated cells showed a two-fold to three-fold increase in NFκB activation compared with CM from naïve cells (Fig. 6A; Supplementary Fig. S7A). Addition of a selective IκB kinase (IKK) inhibitor, IKK16 (39, 40) to the CM, inhibited activation of this reporter. TNFα addition to these 293T cells increased NFκB-driven luciferase activity while the addition of IKK16 inhibits TNFα, AZDR1, and LGBR2 cells CM induced NFκB reporter activity (Fig. 6B). Importantly, NFκB reporter activity was highly correlated with NFκB-S536 phosphorylation levels in these experiments (Fig. 6C). This result was consistent with our observation that a 60-minute treatment with a PIMi did not induce RelA-S536 phosphorylation, while overnight treatment with inhibitors did, suggesting that sufficient secretion of growth factors occurred after 18 hours to activate this pathway (Supplementary Fig. S7B and S7C). In a PIMi-insensitive mature T-ALL cell line, SUP-T1 (10), AZD1208 treatment had no effect on S536 phosphorylation and did not activate the NFκB pathway. Because of the LCK translocation in HSB-2 cells, we also validated the activation of NFκB using PIM kinase–dependent DU.528 and KOPT-K1 cell lines with no LCK translocation (Supplementary Fig. S7D), further establishing that the NFκB activation after PIM kinase inhibition is independent of LCK translocation or expression. These results suggest that PIMi treatment in HSB-2 cells appears to be associated with the increased secretion of growth factors that in an autocrine fashion activate NFκB signaling.

Figure 6.

NFκB activation in PIMi-treated HSB-2 cells. A, HEK-293T-NFκB-luc cells were incubated with CM from HSB-2 naïve or AZDR1 and LGBR2 cells with or without IKK16 (2 μmol/L) for 6 hours. B and C, HEK-293T-NFκB-luc cells were incubated with media containing TNFα (20 ng/mL) alone or in combination with IKK16 (2 μmol/L) or CM from HSB-2 LGBR2 or AZDR1 with or without IKK16 (2 μmol/L) for 6 hours. NFκB-mediated luciferase activity was measured using One-Glo Luciferase Assay System. Cell lysates were Western blotted with the specified antibodies. LE, long exposure. D, Cell viability of HSB-2 naïve and AZDR1 cells treated with indicated concentrations of AZD alone and/or in combination with IKK16 for 72 hours. E, T-ALL PDX cells were incubated with the indicated concentrations of LGB321 (LGB) alone or in combination with IKK16 for 72 hours and then the percentage of viable cells was quantified by the ATPlite assay. D and E, the growth of DMSO control cells was considered 100% and percent cell growth after individual treatment is reported relative to the DMSO. The data shown are the average ± SD of three independent experiments.

Figure 6.

NFκB activation in PIMi-treated HSB-2 cells. A, HEK-293T-NFκB-luc cells were incubated with CM from HSB-2 naïve or AZDR1 and LGBR2 cells with or without IKK16 (2 μmol/L) for 6 hours. B and C, HEK-293T-NFκB-luc cells were incubated with media containing TNFα (20 ng/mL) alone or in combination with IKK16 (2 μmol/L) or CM from HSB-2 LGBR2 or AZDR1 with or without IKK16 (2 μmol/L) for 6 hours. NFκB-mediated luciferase activity was measured using One-Glo Luciferase Assay System. Cell lysates were Western blotted with the specified antibodies. LE, long exposure. D, Cell viability of HSB-2 naïve and AZDR1 cells treated with indicated concentrations of AZD alone and/or in combination with IKK16 for 72 hours. E, T-ALL PDX cells were incubated with the indicated concentrations of LGB321 (LGB) alone or in combination with IKK16 for 72 hours and then the percentage of viable cells was quantified by the ATPlite assay. D and E, the growth of DMSO control cells was considered 100% and percent cell growth after individual treatment is reported relative to the DMSO. The data shown are the average ± SD of three independent experiments.

Close modal

To evaluate whether the NFκB pathway inhibition could enhance T-ALL growth suppression by PIMi, increasing doses of IKK16 were added to HSB-2 naïve or AZDR cells with and without AZD1208. IKK16 alone suppressed the growth of both naïve and AZDR cells in a dose-dependent fashion (Fig. 6D) but was significantly more effective in the AZDR cells. This result is consistent with the NFκB pathway activation in long-term PIMi-treated cells. In addition, when combined with AZD1208, IKK16 was able to inhibit the growth of AZDR cells to a greater extent than the naïve cells, suggesting that the role of NFκB in driving cell growth had been increased by PIMi treatment. Notably, Combosyn analysis demonstrated that the AZD1208 and IKK16 combination was highly synergistic in suppressing the growth of AZDR as compared with naïve HSB-2 cells (Supplementary Fig. S7E). Finally, the combination of PIM and NFκB inhibitors treatment lead to a moderate but significant decrease in the viability of ETP-ALL PDXs when compared with mature T-ALL PDXs (Fig. 6E). Thus, inhibition of NFκB by small molecules can be used to enhance the growth inhibitory effect of PIMi in ETP-ALL cells.

In this study, we demonstrate that PIM1 expression is markedly increased in ETP-ALL but not in mature T-ALL PDXs, suggesting that this kinase could be a therapeutic target to inhibit the growth of ETP-ALL. The crucial role of PIM kinases in hematopoietic tumors has fueled the development of a number of pan-PIMis which are in phase IB/II trials for multiple myeloma, myelodysplastic syndrome, myelofibrosis, and lymphoma (21). Independent studies (11–13) demonstrate that PIM1 function is necessary to support T-ALL growth. In both pediatric and adult T-ALL, activation of PIM1 is driven by a rare but recurrent TCRβ-PIM1 translocation (12, 13), and plays a clear role in tumor growth. In vitro and in vivo results demonstrate that PIMi reduces T-ALL growth. These results suggest that PIMi could be used to treat patients with chemotherapy-resistant ETP-ALL with high PIM1 expression. However, PIMi alone is not completely effective in ETP-ALL PDX models, suggesting that other agents are needed to hit complementary targets. To help define those signaling pathways, we derived PIMi-resistant cells using HSB-2 cell line and analyzed the pathways that were enhanced. In addition, our previous work demonstrated that the genes expressed in the PIMi-sensitive cells, including HSB-2 were highly correlated with the gene expression profile of ETP-ALL PDXs rather than the mature non-ETP ALL PDXs (10), suggesting that this cell line could serve as a model for PIMi-resistant ETP-ALL.

To investigate these pathways after long-term PIMi treatment, we first analyzed the transcriptomes of PIMi-resistant HSB-2 cells. RNA-seq oncogenic enrichment analysis demonstrated activation of mTOR, MYC, NFκB, and HOXA9 pathways in long-term PIMi-treated cells. PIM kinases have been implicated in the control of mTOR pathway (9, 10). mTOR plays a key role in early T-lymphopoiesis and eradicating mTORC1 blocks the development of T-ALL (41). Short-term treatment of HSB-2 cells with PIMi blocks mTORC1 activation as demonstrated by S6 and 4EBP1 phosphorylation, while mTOR is continuously activated in long-term PIMi-treated cells. In AML, AZD1208 resistance occurs through feedback activation of mTOR signaling, which is mediated by reactive oxygen species, p38, and AKT (36). One possible mechanism accounting for mTOR activation in PIMi-resistant HSB-2 cells is the compensatory decrease in DDIT4 (REDD1) and Sestrin2 (SESN2) expression. Decreases in these proteins lead to activation of the two GTP-binding proteins, Rheb and Rag, which regulate mTOR and control the activation of protein synthesis. MYC plays an important role in driving T-ALL growth and is increased by NOTCH activation. PIM kinase regulates MYC levels through phosphorylation-dependent activation and PIMis are known to decrease MYC levels (42). MYC levels are also regulated by mTOR, and the constitutive activation of this enzyme in long-term PIMi-treated cells could maintain the activation of this signaling pathway. NFκB and HOXA9, positive regulators of PIM kinase expression, are often abnormally activated in AML (43, 44). The activation of NFκB and HOXA9 signaling pathways appear to be a compensatory response to the prolonged inhibition of PIM kinases.

Although many pathways are involved in the development of PIMi resistance in HSB-2 cells, this is the first demonstration of the contribution of the NFκB pathway after long-term PIMi treatment. Through RNA-seq and RPPA profiling, we found compelling evidence that PIM inhibition activates RelA (NFκB P65) in AZDR cells compared with naïve HSB-2. PIM kinase facilitates the activation of the NFκB pathway through phosphorylation-dependent stabilization of RelA (45). The sustained elevation of phosphorylated NFκB S536 in HSB-2 cells during PIM kinase inhibition indicates a dysfunctional regulation of the NFκB pathway where PIM kinase inhibition fails to deactivate NFκB. Because mTOR is an important regulator of NFκB (46), it is possible that this kinase plays a role in the aberrant regulation. Alternatively, we demonstrate using CM that long-term PIMi-treated cells produce growth factors that activate NFκB. It is possible that one or more ILs known to activate NFκB, for example, IL6, could be secreted in increased quantities (47). The observation that concurrent inhibition of PIM kinase and NFκB leads to reduced viability in ETP-ALL PDXs suggests that activation of this pathway is important to the survival of ETP-ALL cells. Considering that the effect of PIM and NFκB inhibitors when administered together is modest, targeting additional pathways like PI3K/AKT and mTOR activated along with long-term PIM inhibition will be needed to completely kill T-ALL. Future in vivo PDX studies using a triple kinase inhibitor IBL-302, a compound that hits PIM/PI3K/mTOR (48) in combination with NFκB inhibitor could potentially show a broader response.

Furthermore, our network-based analysis identified NFκB pathway members such as ATM (ataxia telangiectasia mutated), a tyrosine kinase whose protein-level alteration has been linked extensively to leukemia (49). Likewise, the ANM suggests that LYN, PIDD1, and VCAM1 are additional NFκB-associated protein markers that differentiate between long-term and short-term PIM kinase inhibition provides further evidence that NFκB signaling plays a significant role in PIMi resistance. Future studies focused on proteins discovered in the ANM could provide additional insight into the role of the NFκB signaling in modulating ETP-ALL growth.

Our unbiased RNA-seq analysis and targeted RPPA assay of long-term PIMi-treated HSB-2 cells demonstrated that key survival pathways are turned on with prolonged incubation with this inhibitor. These results obtained in HSB-2 cells suggest the potential of combining NFκB and PIMis to treat a subset of T-ALL with high PIM1 expression.

E.F.Petricoin reports personal fees from Perthera, Inc., Avant Diagnostics, Inc., and Ceres Nanosciences, Inc. outside the submitted work. D.T. Teachey reports other from Amgen (advisory board), La Roche (advisory board), Sobi (advisory board), and Janssen (advisory board) outside the submitted work. No potential conflicts of interest were disclosed by the other authors.

J.T. Lim: Conceptualization, data curation, software, formal analysis, validation, investigation, visualization, methodology, writing-original draft, writing-review and editing. N. Singh: Data curation, formal analysis, investigation, visualization, methodology, writing-review and editing. L.A. Luevano: Data curation, formal analysis, investigation, visualization, methodology. V.S. Calvert: Data curation, formal analysis, visualization, methodology. E.F. Petricoin: Resources, data curation, software, formal analysis, visualization, methodology. D.T. Teachey: Resources, methodology. R.B. Lock: Resources, methodology, development of key experimental models. M. Padi: Conceptualization, data curation, software, formal analysis, validation, investigation, visualization, methodology, writing-original draft, writing-review and editing. A.S. Kraft: Conceptualization, resources, supervision, funding acquisition, methodology, writing-original draft, project administration, writing-review and editing. S.K.R. Padi: Conceptualization, data curation, formal analysis, supervision, validation, investigation, visualization, methodology, writing-original draft, project administration, writing-review and editing.

The authors acknowledge the Experimental Mouse Shared Resource and Cytometry Core facility at the University of Arizona Cancer Center (UACC) for helping with in vivo experiments and flow cytometry, respectively. The authors acknowledge Sammed Mandape in the UACC Bioinformatics Shared Resource for providing preliminary data analysis. The authors wish to thank the University of Arizona Genetics Core (UAGC) for providing RNA sequencing services. This research was supported by University of Arizona Cancer Center support grant P30CA023074, NIH award R01CA173200, and DOD award W81XWH-12-1-0560 (to A.S. Kraft).

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

1.
Hunger
SP
,
Lu
X
,
Devidas
M
,
Camitta
BM
,
Gaynon
PS
,
Winick
NJ
, et al
Improved survival for children and adolescents with acute lymphoblastic leukemia between 1990 and 2005: a report from the children's oncology group
.
J Clin Oncol
2012
;
30
:
1663
9
.
2.
Yeoh
EJ
,
Ross
ME
,
Shurtleff
SA
,
Williams
WK
,
Patel
D
,
Mahfouz
R
, et al
Classification, subtype discovery, and prediction of outcome in pediatric acute lymphoblastic leukemia by gene expression profiling
.
Cancer Cell
2002
;
1
:
133
43
.
3.
Cleaver
AL
,
Beesley
AH
,
Firth
MJ
,
Sturges
NC
,
O'Leary
RA
,
Hunger
SP
, et al
Gene-based outcome prediction in multiple cohorts of pediatric T-cell acute lymphoblastic leukemia: a Children's Oncology Group study
.
Mol Cancer
2010
;
9
:
105
.
4.
Coustan-Smith
E
,
Mullighan
CG
,
Onciu
M
,
Behm
FG
,
Raimondi
SC
,
Pei
D
, et al
Early T-cell precursor leukaemia: a subtype of very high-risk acute lymphoblastic leukaemia
.
Lancet Oncol
2009
;
10
:
147
56
.
5.
Zhang
J
,
Ding
L
,
Holmfeldt
L
,
Wu
G
,
Heatley
SL
,
Payne-Turner
D
, et al
The genetic basis of early T-cell precursor acute lymphoblastic leukaemia
.
Nature
2012
;
481
:
157
63
.
6.
Neumann
M
,
Heesch
S
,
Schlee
C
,
Schwartz
S
,
Gokbuget
N
,
Hoelzer
D
, et al
Whole-exome sequencing in adult ETP-ALL reveals a high rate of DNMT3A mutations
.
Blood
2013
;
121
:
4749
52
.
7.
Maude
SL
,
Dolai
S
,
Delgado-Martin
C
,
Vincent
T
,
Robbins
A
,
Selvanathan
A
, et al
Efficacy of JAK/STAT pathway inhibition in murine xenograft models of early T-cell precursor (ETP) acute lymphoblastic leukemia
.
Blood
2015
;
125
:
1759
67
.
8.
Liu
Y
,
Easton
J
,
Shao
Y
,
Maciaszek
J
,
Wang
Z
,
Wilkinson
MR
, et al
The genomic landscape of pediatric and young adult T-lineage acute lymphoblastic leukemia
.
Nat Genet
2017
;
49
:
1211
8
.
9.
Lin
YW
,
Beharry
ZM
,
Hill
EG
,
Song
JH
,
Wang
W
,
Xia
Z
, et al
A small molecule inhibitor of Pim protein kinases blocks the growth of precursor T-cell lymphoblastic leukemia/lymphoma
.
Blood
2010
;
115
:
824
33
.
10.
Padi
SKR
,
Luevano
LA
,
An
N
,
Pandey
R
,
Singh
N
,
Song
JH
, et al
Targeting the PIM protein kinases for the treatment of a T-cell acute lymphoblastic leukemia subset
.
Oncotarget
2017
;
8
:
30199
216
.
11.
de Bock
CE
,
Demeyer
S
,
Degryse
S
,
Verbeke
D
,
Sweron
B
,
Gielen
O
, et al
HOXA9 cooperates with activated JAK/STAT signaling to drive leukemia development
.
Cancer Discov
2018
;
8
:
616
31
.
12.
La Starza
R
,
Messina
M
,
Gianfelici
V
,
Pierini
V
,
Matteucci
C
,
Pierini
T
, et al
High PIM1 expression is a biomarker of T-cell acute lymphoblastic leukemia with JAK/STAT activation or t(6;7)(p21;q34)/TRB@-PIM1 rearrangement
.
Leukemia
2018
;
32
:
1807
10
.
13.
De Smedt
R
,
Peirs
S
,
Morscio
J
,
Matthijssens
F
,
Roels
J
,
Reunes
L
, et al
Pre-clinical evaluation of second generation PIM inhibitors for the treatment of T-cell acute lymphoblastic leukemia and lymphoma
.
Haematologica
2019
;
104
:
e17
20
.
14.
Yan
B
,
Zemskova
M
,
Holder
S
,
Chin
V
,
Kraft
A
,
Koskinen
PJ
, et al
The PIM-2 kinase phosphorylates BAD on serine 112 and reverses BAD-induced cell death
.
J Biol Chem
2003
;
278
:
45358
67
.
15.
Zhang
F
,
Beharry
ZM
,
Harris
TE
,
Lilly
MB
,
Smith
CD
,
Mahajan
S
, et al
PIM1 protein kinase regulates PRAS40 phosphorylation and mTOR activity in FDCP1 cells
.
Cancer Biol Ther
2009
;
8
:
846
53
.
16.
Warfel
NA
,
Kraft
AS
. 
PIM kinase (and Akt) biology and signaling in tumors
.
Pharmacol Ther
2015
;
151
:
41
9
.
17.
Chen
XP
,
Losman
JA
,
Cowan
S
,
Donahue
E
,
Fay
S
,
Vuong
BQ
, et al
Pim serine/threonine kinases regulate the stability of Socs-1 protein
.
Proc Natl Acad Sci U S A
2002
;
99
:
2175
80
.
18.
Dolai
S
,
Sia
KC
,
Robbins
AK
,
Zhong
L
,
Heatley
SL
,
Vincent
TL
, et al
Quantitative phosphotyrosine profiling of patient-derived xenografts identifies therapeutic targets in pediatric leukemia
.
Cancer Res
2016
;
76
:
2766
77
.
19.
Keeton EK
MK
,
Dillman
KS
,
Palakurthi
S
,
Cao
Y
,
Grondine
MR
,
Kaur
S
, et al
AZD1208, a potent and selective pan-Pim kinase inhibitor, demonstrates efficacy in preclinical models of acute myeloid leukemia
.
Blood
2014
;
123
:
905
13
.
20.
Garcia
PD
,
Langowski
JL
,
Wang
Y
,
Chen
M
,
Castillo
J
,
Fanton
C
, et al
Pan-PIM kinase inhibition provides a novel therapy for treating hematologic cancers
.
Clin Cancer Res
2014
;
20
:
1834
45
.
21.
Zhang
X
,
Song
M
,
Kundu
JK
,
Lee
MH
,
Liu
ZZ
. 
PIM kinase as an executional target in cancer
.
J Cancer Prev
2018
;
23
:
109
16
.
22.
Chiarini
F
,
Lonetti
A
,
Evangelisti
C
,
Buontempo
F
,
Orsini
E
,
Evangelisti
C
, et al
Advances in understanding the acute lymphoblastic leukemia bone marrow microenvironment: from biology to therapeutic targeting
.
Biochim Biophys Acta
2016
;
1863
:
449
63
.
23.
Song
JH
,
Padi
SKR
,
Luevano
LA
,
Minden
MD
,
DeAngelo
DJ
,
Hardiman
G
, et al
Insulin receptor substrate 1 is a substrate of the Pim protein kinases
.
Oncotarget
2016
;
7
:
20152
65
.
24.
Knoechel
B
,
Roderick
JE
,
Williamson
KE
,
Zhu
J
,
Lohr
JG
,
Cotton
MJ
, et al
An epigenetic mechanism of resistance to targeted therapy in T cell acute lymphoblastic leukemia
.
Nat Genet
2014
;
46
:
364
70
.
25.
Wright
DD
,
Sefton
BM
,
Kamps
MP
. 
Oncogenic activation of the Lck protein accompanies translocation of the LCK gene in the human HSB2 T-cell leukemia
.
Mol Cell Biol
1994
;
14
:
2429
37
.
26.
Horiuchi
D
,
Camarda
R
,
Zhou
AY
,
Yau
C
,
Momcilovic
O
,
Balakrishnan
S
, et al
PIM1 kinase inhibition as a targeted therapy against triple-negative breast tumors with elevated MYC expression
.
Nat Med
2016
;
22
:
1321
9
.
27.
Braso-Maristany
F
,
Filosto
S
,
Catchpole
S
,
Marlow
R
,
Quist
J
,
Francesch-Domenech
E
, et al
PIM1 kinase regulates cell death, tumor growth and chemotherapy response in triple-negative breast cancer
.
Nat Med
2016
;
22
:
1303
13
.
28.
Chatterjee
S
,
Chakraborty
P
,
Daenthanasanmak
A
,
Iamsawat
S
,
Andrejeva
G
,
Luevano
LA
, et al
Targeting PIM kinase with PD1 inhibition improves immunotherapeutic antitumor T-cell response
.
Clin Cancer Res
2019
;
25
:
1036
49
.
29.
Subramanian
A
,
Tamayo
P
,
Mootha
VK
,
Mukherjee
S
,
Ebert
BL
,
Gillette
MA
, et al
Gene set enrichment analysis: a knowledge-based approach for interpreting genome-wide expression profiles
.
Proc Natl Acad Sci U S A
2005
;
102
:
15545
50
.
30.
Bueso-Ramos
CE
,
Rocha
FC
,
Shishodia
S
,
Medeiros
LJ
,
Kantarjian
HM
,
Vadhan-Raj
S
, et al
Expression of constitutively active nuclear-kappa B RelA transcription factor in blasts of acute myeloid leukemia
.
Hum Pathol
2004
;
35
:
246
53
.
31.
Zippo
A
,
De Robertis
A
,
Serafini
R
,
Oliviero
S
. 
PIM1-dependent phosphorylation of histone H3 at serine 10 is required for MYC-dependent transcriptional activation and oncogenic transformation
.
Nat Cell Biol
2007
;
9
:
932
44
.
32.
Schaub
FX
,
Dhankani
V
,
Berger
AC
,
Trivedi
M
,
Richardson
AB
,
Shaw
R
, et al
Pan-cancer alterations of the MYC oncogene and its proximal network across the Cancer Genome Atlas
.
Cell Syst
2018
;
6
:
282
300.e2
.
33.
Szklarczyk
D
,
Gable
AL
,
Lyon
D
,
Junge
A
,
Wyder
S
,
Huerta-Cepas
J
, et al
STRING v11: protein-protein association networks with increased coverage, supporting functional discovery in genome-wide experimental datasets
.
Nucleic Acids Res
2019
;
47
:
D607
13
.
34.
Ideker
T
,
Ozier
O
,
Schwikowski
B
,
Siegel
AF
. 
Discovering regulatory and signalling circuits in molecular interaction networks
.
Bioinformatics
2002
;
18
:
S233
40
.
35.
Silverbush
D
,
Grosskurth
S
,
Wang
D
,
Powell
F
,
Gottgens
B
,
Dry
J
, et al
Cell-specific computational modeling of the PIM pathway in acute myeloid leukemia
.
Cancer Res
2017
;
77
:
827
38
.
36.
Brunen
D
,
Garcia-Barchino
MJ
,
Malani
D
,
Jagalur Basheer
N
,
Lieftink
C
,
Beijersbergen
RL
, et al
Intrinsic resistance to PIM kinase inhibition in AML through p38alpha-mediated feedback activation of mTOR signaling
.
Oncotarget
2016
;
7
:
37407
19
.
37.
Aziz
AUR
,
Farid
S
,
Qin
K
,
Wang
H
,
Liu
B
. 
PIM kinases and their relevance to the PI3K/AKT/mTOR pathway in the regulation of ovarian cancer
.
Biomolecules
2018
;
8
:
7
.
38.
Wilson
AA
,
Kwok
LW
,
Porter
EL
,
Payne
JG
,
McElroy
GS
,
Ohle
SJ
, et al
Lentiviral delivery of RNAi for in vivo lineage-specific modulation of gene expression in mouse lung macrophages
.
Mol Ther
2013
;
21
:
825
33
.
39.
Duckworth
C
,
Zhang
L
,
Carroll
SL
,
Ethier
SP
,
Cheung
HW
. 
Overexpression of GAB2 in ovarian cancer cells promotes tumor growth and angiogenesis by upregulating chemokine expression
.
Oncogene
2016
;
35
:
4036
47
.
40.
Waelchli
R
,
Bollbuck
B
,
Bruns
C
,
Buhl
T
,
Eder
J
,
Feifel
R
, et al
Design and preparation of 2-benzamido-pyrimidines as inhibitors of IKK
.
Bioorg Med Chem Lett
2006
;
16
:
108
12
.
41.
Hoshii
T
,
Kasada
A
,
Hatakeyama
T
,
Ohtani
M
,
Tadokoro
Y
,
Naka
K
, et al
Loss of mTOR complex 1 induces developmental blockage in early T-lymphopoiesis and eradicates T-cell acute lymphoblastic leukemia cells
.
Proc Natl Acad Sci U S A
2014
;
111
:
3805
10
.
42.
Beharry
Z
,
Mahajan
S
,
Zemskova
M
,
Lin
YW
,
Tholanikunnel
BG
,
Xia
Z
, et al
The Pim protein kinases regulate energy metabolism and cell growth
.
Proc Natl Acad Sci U S A
2011
;
108
:
528
33
.
43.
Zhu
N
,
Ramirez
LM
,
Lee
RL
,
Magnuson
NS
,
Bishop
GA
,
Gold
MR
. 
CD40 signaling in B cells regulates the expression of the Pim-1 kinase via the NF-kappa B pathway
.
J Immunol
2002
;
168
:
744
54
.
44.
Hu
YL
,
Passegue
E
,
Fong
S
,
Largman
C
,
Lawrence
HJ
. 
Evidence that the Pim1 kinase gene is a direct target of HOXA9
.
Blood
2007
;
109
:
4732
8
.
45.
Nihira
K
,
Ando
Y
,
Yamaguchi
T
,
Kagami
Y
,
Miki
Y
,
Yoshida
K
. 
Pim-1 controls NF-kappaB signalling by stabilizing RelA/p65
.
Cell Death Differ
2010
;
17
:
689
98
.
46.
Dan
HC
,
Cooper
MJ
,
Cogswell
PC
,
Duncan
JA
,
Ting
JP
,
Baldwin
AS
. 
Akt-dependent regulation of NF-{kappa}B is controlled by mTOR and Raptor in association with IKK
.
Genes Dev
2008
;
22
:
1490
500
.
47.
Guzman
ML
,
Neering
SJ
,
Upchurch
D
,
Grimes
B
,
Howard
DS
,
Rizzieri
DA
, et al
Nuclear factor-kappaB is constitutively activated in primitive human acute myelogenous leukemia cells
.
Blood
2001
;
98
:
2301
7
.
48.
Mohlin
S
,
Hansson
K
,
Radke
K
,
Martinez
S
,
Blanco-Aparicio
C
,
Garcia-Ruiz
C
, et al
Anti-tumor effects of PIM/PI3K/mTOR triple kinase inhibitor IBL-302 in neuroblastoma
.
EMBO Mol Med
2020
;
12
:
e11749
.
49.
Wu
ZH
,
Shi
Y
,
Tibbetts
RS
,
Miyamoto
S
. 
Molecular linkage between the kinase ATM and NF-kappaB signaling in response to genotoxic stimuli
.
Science
2006
;
311
:
1141
6
.

Supplementary data