Abstract
PI3K/AKT/mTOR pathway hyperactivation is frequent in T-cell acute lymphoblastic leukemia/lymphoma (T-ALL/LBL). To model inhibition of mTOR, pre–T-cell lymphoblastic leukemia/lymphoma (pre-T LBL) tumor development was monitored in mice with T lymphocyte–specific, constitutively active AKT (Lck-MyrAkt2) that were either crossed to mTOR knockdown (KD) mice or treated with the mTOR inhibitor everolimus. Lck-MyrAkt2;mTOR KD mice lived significantly longer than Lck-MyrAkt2;mTOR wild-type (WT) mice, although both groups ultimately developed thymic pre-T LBL. An increase in survival was also observed when Lck-MyrAkt2;mTOR WT mice were treated for 8 weeks with everolimus. The transcriptional profiles of WT and KD thymic lymphomas were compared, and Ingenuity Pathway Upstream Regulator Analysis of differentially expressed genes in tumors from mTOR WT versus KD mice identified let-7 and miR-21 as potential regulatory genes. mTOR KD mice had higher levels of let-7a and miR-21 than mTOR WT mice, and rapamycin induced their expression in mTOR WT cells. CDK6 was one of the most downregulated targets of both let-7 and miR21 in mTOR KD tumors. CDK6 overexpression and decreased expression of let-7 in mTOR KD cells rescued a G1 arrest phenotype. Combined mTOR (rapamycin) and CDK4/6 (palbociclib) inhibition decreased tumor size and proliferation in tumor flank transplants, increased survival in an intravenous transplant model of disseminated leukemia compared with single agent treatment, and cooperatively decreased cell viability in human T-ALL/LBL cell lines. Thus, mTOR KD mice provide a model to explore drug combinations synergizing with mTOR inhibitors and can be used to identify downstream targets of inhibition.
Introduction
T-cell acute lymphoblastic leukemia/lymphoma (T-ALL/LBL) accounts for 15% of childhood and 25% of adult lymphoblastic leukemias and is often characterized by hyperactivation of the PI3K/AKT/mTOR pathway (1). In one study, deletion or mutational inactivation of PTEN or activating PI3K or AKT mutations were found in approximately one-half of all primary T-ALL samples (2). Frequent activation of this pathway and its role in tumor cell proliferation, survival, and chemotherapeutic resistance has suggested mTOR as a target for pharmacologic inhibition, leading to clinical trials of mTOR inhibitors in ALL [https://clinicaltrials.gov (3, 4)]. In vitro pharmacologic inhibition decreases the viability of human T-ALL/LBL cell lines (5–7). However, mTOR inhibitors can be associated with feedback loops and resistance (8), thus combining them with drugs that target genes in either the same or different pathways may be more effective or serve to suppress resistance mechanisms.
To mimic drug treatment, mice with constitutive reduction in mTOR, mTOR knockdown (KD; refs. 9–11) were crossed to Lck-MyrAkt2 (Lck-Akt2; lymphocyte cell–specific protein-tyrosine kinase, myristoylated) mice (12–14), which express constitutively active AKT kinase in immature thymocytes and spontaneously develop thymic pre-T-cell lymphoblastic leukemia/lymphoma (pre-T LBL) after 75–200 days of age. Comparative analysis of gene expression profiles coupled with Ingenuity Pathway Analysis (IPA) of the differentially expressed genes in tumors from Lck-Akt2;mTOR wild-type (WT) and Lck-Akt2;mTOR KD mice identified let-7 (lethal 7) and miR21 (miRNA 21) as prominent upstream regulators. CDK6 (cyclin-dependent kinase 6), previously identified as a target of both let-7 (15) and miR21 (16, 17), was the most downregulated gene in tumors of mice with constitutive reduction in mTOR. CDK6 overexpression in mTOR KD cells rescued the cell-cycle arrest phenotype seen in these cells. Because CDK6 is also often highly expressed in patient samples (18, 19), we evaluated the effect of combination therapy involving mTOR and CDK4/CDK6 inhibitors to limit pre-T LBL and T-ALL tumors.
Materials and Methods
Mice
Animal studies were performed in compliance with NIH Animal Care and Use Committee Protocol LG009. The Lck-Akt model used was the Lck-MyrAkt2 (founder line 72; C57BL/6 background) transgenic mouse in which the T lymphocyte–specific proximal Lck promoter drives a constitutively active, myristoylated Akt2 protein (12–14, 20, 21). These mice were crossed with heterozygous Mtortm1. Lgm mice (mTOR KD mice with approximately 70% reduction in mTOR expression: C57BL/6;129 background; Fig. 1A; ref. 11) for two generations; offspring positive for Lck-Akt2 and homozygous for mTOR WT or KD were selected. Mice were genotyped for the Lck-Akt2 transgene (primers 5′-AGG CAC TGC CCT CTT GAA GC-3′ and 5′-TTT GGG GTT CTG AAT GTG AG-3′), and for the disrupted mTOR gene as described previously (11). Mouse survival was assessed by Kaplan–Meier survival curves and log-rank tests (GraphPad Prism). Tumors were collected at humane endpoint; a portion was fixed in Telly fixative and the remainder was frozen. In a separate experiment to control for the allelic variant (628C) of mTOR in the KD mice, Mtortm1.1Lgm mice with the allelic variant and normal mTOR levels (11) were crossed with Lck-Akt2 mice; no differences in survival were observed between Lck-Akt2 mice with or without the allelic variant (data not shown). Lck-Akt2 transgenic mice from founder line 55 (C57BL/6 background; similar tumor incidence to founder line 72 but with more rapid tumor formation (13)) were given everolimus (RAD001; 20 mg/kg) or vehicle (50 μL) by gavage twice weekly, beginning at 8 weeks of age and continuing for 8 weeks. Survival was assessed at humane endpoint.
Assessment of downstream targets of mTOR and T-cell development in pretumor thymi of Lck-Akt2;mTOR WT and KD mice. A, Schematic summarizing the characteristics of mouse strains bred to produce the Lck-Akt2;mTOR WT and Lck-Akt/mTOR KD mice. Lck-MyrAkt2 mice have constitutively active Akt2 due to myristoylation exclusively in T lymphocytes, which is driven by a proximal Lck promoter; these mice spontaneously develop thymic lymphomas. mTOR KD mice have decreased mTOR expression due to a neomycin cassette inserted in the mTOR gene in cells throughout the body (11). The resulting offspring are referred to throughout as WT and KD based on mTOR status, and both are positive for the Lck-Akt2 transgene. B, Protein levels of downstream targets of mTOR, phospho-4E-BP1Thr37/46 and phospho-AKTS473, were evaluated in the thymocytes of 4-week-old WT and KD mice, as shown on digital representation of a capillary-based protein electrophoresis from representative mice (see Supplementary Fig. S1A for electrophoresis chromatogram). C and D, Flow cytometric evaluation of developing T-cell populations in the thymi of 4-week-old WT and KD mice. Absolute numbers and percentages of CD4+ and CD8+ single populations (C) were assessed, as were double negative populations (DN1-4) labeled for CD25 and CD44 (D). Three WT and three KD thymi were assessed, and error bars represent SEM. Absolute numbers of cells for each population were compared between WT and KD by Student t test (*, P < 0.05).
For in vivo drug combination studies, tumor fragments from primary tumors arising in Lck-Akt2;mTOR (WT) mice were expanded in vivo to yield sufficient starting material by subcutaneous passage in nude mice. For the preliminary pharmacodynamic study, 250 μL of freshly isolated tumor homogenate was injected bilaterally into the flanks of 16 nude mice. When tumors reached palpable size, mice were randomized to four treatment groups (4 mice/group), and received vehicle (gavage and intraperitoneal), 150 mg/kg palbociclib (PD-0332991, Ibrance; oral gavage), 2.5 mg/kg rapamycin (intraperitoneal injection), or a combination of palbociclib and rapamycin once daily for 5 days. Tumor weights were taken at necropsy, and tissues were collected for subsequent analysis.
To assess the individual and combination drug efficacy in a disseminated leukemia setting, nude mice were injected intravenously with a single cell suspension of 5.0 × 106 Lck-Akt2;WT tumor cells in PBS. Mice were randomized into four treatment groups of 8 mice each, and then received once daily treatments (Monday–Friday) beginning 4 days posttransplantation of either vehicle (oral and intraperitoneal), palbociclib (LC Labs, gavage at 50 mg/kg in 1% Tween-80 in sterile water), rapamycin (LC labs, intraperitoneally at 2.5 mg/kg in 5% Tween-80;5%PEG in sterile saline), or a combination of palbociclib and rapamycin. Treatment continued for two cycles of 3 weeks on treatment/1 week off. Mice were weighed twice weekly and monitored for signs of disease progression and duress. Animals were euthanized in accordance with NCI IACUC humane euthanasia criteria. At necropsy, spleen and liver weights were recorded, and tissues (femurs, visibly enlarged lymph nodes, and other tissues that appeared on gross examination to contain tumor infiltrates) were collected for histologic confirmation of disseminated disease. Statistical differences in survival were determined by Mantel–Cox log-rank testing (with Bonferroni corrections) of the Kaplan–Meier survival curves.
Analysis of the dose response to CDK4/6 inhibition by palbociclib in low passage tumor cell lines from WT and KD mice was done using nonlinear regression modeling implemented in GraphPad Prism 7.05 in the “log(inhibitor) vs. response–variable slope” module. Comparisons of nested models and the IC50 best-fit parameter was performed using extra sum-of-square F test. Significance level of 0.05 was applied for rejecting the null hypotheses.
Flow cytometry, transfections, protein analysis, and IHC
Single cell suspensions from thymi of 4-week-old Lck-Akt2;mTOR WT, and KD mice and from thymic pre-T LBL were stained for CD4 and CD8 to determine double (CD4+,CD8+) and single (CD4+ or CD8+) positive T-cell populations by flow cytometry. Double negative (DN; CD4−,CD8−-) T cells were identified as described previously (22). Tumor (pre-T LBL) cells treated in vitro for 48 hours with 1 nmol/L rapamycin or 625 nmol/L palbociclib, and flank tumor transplants treated as described above were evaluated for apoptosis and cell cycle using PE/Annexin Apoptosis Detection Kit 1 and PI/RNase staining buffer (Becton Dickson), respectively. Cells were analyzed using FACSCalibur (Becton Dickinson), FlowJo 8.7 and ModFit LT software.
Protein was isolated with RIPA lysis buffer plus protease inhibitors (Santa Cruz Biotechnology). Twelve micrograms of lysates were electrophoresed on 4%–12% Bis-Tris gels (Novex, Life Technologies) and then transferred to nitrocellulose membranes via the iBLOT system (Life Technologies). Antibodies were purchased from Abcam (c-MYC: ab32072), Santa Cruz Biotechnology (p16:sc-1207), or Cell Signaling Technology (mTOR:2972, CDK6:3136, phospho-4E-BP1Thr37/46:2855, 4E-BP1:9452, phospho-AKTS473:9271, AKT:9272, phospho-RBS780:8180, RB:9313, phopsho-S6S240/244:2215, and S6:2217;β-actin:3700). A size-based, automated, capillary immunoassay system (Simple Western, Protein Simple; ref. 23) was utilized by the CCR Collaborative Protein Technology Resource to measure phospho/total 4E-BP1 and AKT protein expression.
Ki67, a marker of cell proliferation, was evaluated immunohistochemically in the untreated and treated Lck-Akt2;mTOR (WT) flank transplants and in the untreated KD flank transplants. Slides from three tumors per treatment group were deparaffinized, blocked in 3% hydrogen peroxide and Background Buster (Innovex Biosciences), and exposed to steam antigen retrieval in pH 9.0 antigen retrieval solution (Dako). Slides were incubated with primary antibody (Ki67, Abcam 66155; 1:3,000) and secondary antibody (Dako; 1:400). The signal was amplified using AB complex (Vector Laboratories) and slides were incubated with DAB (diaminobensidine, Dako) and counterstained with hematoxylin. The Ki67 proliferative index was scored by counting positively labeled nuclei/5,000 tumor cells.
Tumor clonality
Analysis for T-cell clonality was performed on genomic DNA extracted from eight paraffin-embedded tumor sections using QIAamp DNA FFPE Tissue Kit (Qiagen). PCR amplification was performed using a two-step nested PCR that detects intralocus T-cell receptor gamma (TCRG) rearrangements (24) utilizing previously described primers (23). PCRs were prepared in 25 μL total volume, with 3 μL DNA, 5 pmol of each primer, 0.5 U of Taq polymerase (Invitrogen), and final concentrations of 80 μmol/L DNTP, 2 mmol/L magnesium chloride, 20 mmol/L Tris-hydrogen chloride, and 50 μL KCl. Cycling conditions were: 94°C for 4 minutes; 40 cycles at 94°C for 1 minute; 55°C for 1 minute; and 72°C for 5 minutes. All PCRs were run in duplicate. Amplicons were visualized by high-resolution capillary electrophoresis using a QIAxcel Advanced Instrument (Qiagen) and QIAxcel Screengel Software. Two samples from normal mouse spleens and lymph nodes were amplified as controls.
In vitro tumor cell culture, retroviral delivery of CDK6 and let-7, and FISH
Thymic pre-T LBL single cell suspensions from Lck-Akt2;mTOR WT and KD mice were cultured in Iscove modified Dulbecco medium (Life Technologies) supplemented with 10% FBS, 1% penicillin-streptomycin, l-glutamine, nonessential amino acids, sodium pyruvate, and 0.1% β-mercaptoethanol). CDK6 was overexpressed in low passage Lck-Akt2;mTOR KD cells by retroviral delivery. Mouse cDNA of CDK6 (BC119360) was produced by RT-PCR using primers containing the cloning linkers (forward primer: mCDK6-EcoR1f2: 5′-TTGGGAATTCATGGAGAAGGACAGCCTGA-3′; reverse primer: mCDK6-XhoIr2: 5′-TTGGCTCGAGATGTGTTCAGATGACGAGC-3′); the full-length cDNA was cloned into with EcoR1 and XhoI restriction sites of the MIGR1 retroviral plasmid and CDK6 sequences were verified in the CCR Genomics Core. Retrovirus production and infection: Gryphon retroviral packaging cell lines (Ampho, ABP-RVC-10001, Allele Biotechnology) were used to produce retrovirus. Control (MIGR1, GFP-only control vector retroviral plasmid, a gift from Warren Pear (Addgene plasmid, catalog no. 27490; http://www.addgene.org/27490/), or CDK6 (MIGR1-CDK6) vectors were transfected into Ampho cells using lipofectamine 2000. Retrovirus was collected at 48–72 hours after transfection and 4–5 mL of medium (collection medium: RPMI1640 plus 10% FBS without antibiotics) containing retrovirus was used to infect Lck-Akt2; mTOR KD cells in a 6-well plate (2 million cells/well). Cells were spun at 2,500 rpm for 90 minutes, and then the media were replaced with 4–5 mL of T-cell media (RPMI1640, 10% FBS, penicillin-streptomycin, l-glutamine). Lck-Akt2; mTOR KD cells were transfected with either empty vector (MIGR1) or MIGR1 vector containing CDK6; after 48 hours, the cells were stained with propidium iodide (PI)/RNase Staining Buffer (BD Pharmingen Cell cycle detection kit) using the manufacturer's suggested protocols and analyzed on a FACSCalibur, as above. Retroviruses expressing GFP or let-7 sponge were prepared by adapting previously described methods (25) in the following manner: PlatE cells were transfected with pMSCV PIG (Puro IRES GFP empty vector, a gift from David Bartel, Addgene, catalog no. 21654) or MSCV puro let-7 sponge (a gift from Phil Sharp, Addgene, catalog no. 29766) plasmid by Lipofectamine 3000 (Invitrogen), and recombinant retrovirus were collected 48–72 hours after transfection. The pan T cells were isolated from primary tumors arising in Lck-Akt2;mTOR (KD) mice by using Pan T-cell isolation kit II (Miltenyi Biotec, catalog no. 130-095-130), and transduced twice with retroviruses by “spin inoculation” (650 g for 90 minutes). The cells were cultured for an additional 3 days before being analyzed for cell-cycle analysis by FACS.
For FISH, thymic tumor cells from Lck-Akt2;mTOR WT and KD mice were cultured with 0.2% N-desacetyl-N-methylocolchicine (colcemid) for 2 hours to arrest cells in metaphase, lysed with 0.075 mol/L KCl, and fixed with 3:1 methanol/glacial acetic acid for cytogenetic analysis. Probes for Tcra (BAC probe provided by Thomas Reid, NCI) and Myc were labeled and hybridized overnight to metaphase spreads from fixed cells. Slides were incubated with FITC-labeled anti-DIG and avidin-Alexa 568, stained with DAPI, and imaged at 100×. At least 50 metaphase spreads were assessed for the presence of t(14;15) translocations/sample (13). For in vitro drug studies, low passage WT thymic tumor cells were treated in quadruplicate wells with 1 nmol/L rapamycin and/or 5 nmol/L-5 μmol/L palbociclib (SelleckChem) dissolved in DMSO. Cell viability was assessed after 24 or 48 hours using CellTiter96 Aqueous One Solution (MTS Assay, Promega). All experiments were performed at least twice, and data from a representative experiment are shown unless otherwise indicated in the text.
Microarray and RT-PCR
Total RNA (1 μg) was isolated from flash-frozen WT and KD thymic tumors using TRIzol reagent (Invitrogen, Life Technologies) and purified with RNeasy MiniElute Cleanup Kit (Qiagen). 1 μg RNA was labeled and amplified with Message Amp-11-Biotin Enhanced Kit (Ambion, Life Technologies) and hybridized overnight to Affymetrix Mouse Genome 430 2.0 array cartridges. Chips were washed in a GeneChip Fluidic wash station (Affymetrix) and scanned. Microarray analyses (ArrayExpress accession E-MTAB-3242, www.ebi.ac.uk/arrayexpress) were performed using Affymetrix GCOS, Partek Genomics Suite, and Ingenuity Pathway Analysis (IPA). The raw data were normalized with MAS5 algorithm. Prior to the analyses, genes that had low expression across all samples were removed (sum of the means in KD and WT groups less than 100) and MAS5 expression values were transformed to log base 2 scale. ANOVA was run to identify genes differentially expressed between KD and WT mice and subsequently, genes with P < 0.05 and at least two-fold change in expression were selected for the IPA Upstream Regulator analysis. Gene set enrichment analysis (GSEA; ref. 26) was run using Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway database (MSigDB version 6; ref. 27) and IPA mapped mouse-human orthologues in the whole filtered dataset (∼10K unique gene symbols). The preranked GSEA was performed with signal-to-noise estimate using R Bioconductor fgsea library (28) and the false discovery adjusted P-value less than 0.05 was used to identify enriched pathways.
To assess levels of let-7a and miR-21 in cells, RNA was first isolated from cultured cells using TRIzol reagent (Invitrogen, Life Technologies). cDNA was then prepared using the TaqMan MicroRNA Reverse Transcription Kit (Applied Biosystems). qRT-PCR was performed using the TaqMan Universal PCR Master Mix (Applied Biosystems) on a 7500 Fast RT-PCR system (Applied Biosystems) with the following primers for miR-21 (Thermo Fisher Scientific, TM397)3′- CAACAGCAGUCGAUGGGCUGUC-5′) and for let-7a-5p (Thermo Fisher Scientific, TM000377) 3′-UGAGGUAGUAGGUUGUAUAGUU-5′). Values were normalized to snoRNA234 (3′ CTTTTGGAACTGAATCTAAGTGATTTAACAAAAATTCGTCACTACCACTGAGA-5′).
To assess levels of Cdk6, cDNA was prepared using TaqMan RT Reagents(Life Technologies) and RT PCR (qRT-PCR) was performed using SYBR-Green master mix (ABI) on a 7500 Fast RT-PCR system (Applied Biosystems) with the following primers for Cdk6:3′-CCAAATCTGCTCAACCCATCG-5′ and 3′-CAGGTTGTCCTTGTATCTCTCCAG-5′ and for Cdkn2a/p16Ink4a [cyclin-dependent kinase inhibitor INK4a (p16)] as described previously (29). Values were normalized to β-actin or 18S RNA.
Matrix dose-response screen in human T-ALL/LBL cell lines
Human T-ALL/LBL cell lines (CEM, CUTLL1, HALL, P12-ICHIKAWA, Jurkat, TIB-153, J.CaM1.6, and MOLT-3) were grown in RPMI (0.2% normocin,1% penicillin-streptomycin and l-glutamine, 10% FBS). To assess the activity of the drugs in combination, cell lines were treated with a range of five doses of rapamycin (mTORi, 0–100 nmol/L) and palbociclib(CDK4/6i, 0–1 μmol/L) individually and in combination [matrix dose response screen (30)]. Viability at 48 hours posttreatment was assessed by MTS assay. Each screen was repeated twice, with duplicate matrices plated in each experiment. In addition, each cell line was treated with 500 nmol/L palbociclib, 1 nmol/L rapamycin, or the combination for 48 hours and collected for protein isolation. Drug activity of the combination was evaluated using excess over highest single agent (EOHSA; ref. 31), and synergy scores were calculated using combination index (CI) methods [CompuSyn software http://www.combosyn.com/;CI<1 is synergistic(32–34)]. Heatmaps and CI plots for the dose matrices were generated with R v2.15.1 (35).
Availability of data and materials
Publicly archived microarray data generated in this study are available in ArrayExpress (accession number E-MTAB-3242) at www.ebi.ac.uk/arrayexpress.
Ethics approval and consent to participate
Animal studies were performed in compliance with NIH Animal Care and Use Committee Protocols.
Results
Pretumor thymocytes from Lck-Akt2/mTOR KD mice have decreased mTOR activity
Lck-MyrAkt2 mice have histologically normal thymi at 4 weeks of age, but spontaneously develop thymic tumors characterized as pre-T-LBL by 10–20 weeks of age (12, 13). mTOR signaling and T-cell development were evaluated in pretumor thymocytes from 4-week-old Lck-Akt2;mTOR WT (WT) and Lck-Akt2;mTOR KD (KD) mice prior to tumor formation (Fig. 1A and B). The downstream targets of mTOR, phospho-4E-BP1Thr37/46 (TORC1) and phospho-AKTS473 (TORC2), were decreased in thymocytes from KD mice compared with those from WT mice (Fig. 1B; Supplementary Fig. S1A). Total thymocyte number and size were similar in WT (n = 3) and KD (n = 3) pretumor thymi based on flow cytometry, and there was no difference in the absolute number of double positive (DP) CD4+,CD8+ and single positive CD4+ or CD8+ thymocytes (Fig. 1C). Four stages of DN thymocyte maturation are recognized, DN1–4. A slight T-cell developmental delay was observed in KD thymocytes at the DN1 stage (P = 0.031), which resolved as thymocyte populations were similar between KD and WT through the remaining developmental stages, indicating the mTOR KD did not drastically disrupt thymic T-cell development (Fig. 1D).
Genetic and pharmacologic mTOR inhibition prolongs survival of Lck-Akt2 mice
Over a period of approximately 400 days, a similar proportion of WT (75%) and KD (71%) mice developed thymic pre-T LBL; these tumors were histologically similar (Supplementary Fig. S1B). However, KD mice had a notably increased survival time compared with WT mice (log-rank test, P = 0.002; Fig. 2A); KD mice survived an average of 168 days (median survival: 151 days), whereas WT mice survived an average of 99 days (median: 92 days). As in the pretumor thymocytes, phospho-AKTS473 levels were decreased in tumors from KD mice, as were phospho-4E-BP1Thr37/46 levels, albeit to a lesser extent (Fig. 2B; Supplementary Fig. S1C). WT tumors (n = 2) were a mixture of CD8+ and DP neoplastic T cells, whereas cells from KD tumors (n = 3) were predominately CD8+ (representative tumors shown in Fig. 2C). PCR for T-cell antigen receptor rearrangement revealed that 6 of 8 tumors tested were clonal (3/4 WT, 3/4 KD); pseudoclones were observed in the remaining two samples, likely a result of PCR failure in these samples (Fig. 2D).
Survival associated with thymic pre–T-cell lymphoblastic leukemia/lymphoma (pre-T LBL) in mice with genetic and pharmacologic mTOR inhibition. A, Survival associated with thymic pre-T LBL development/progression was evaluated in KD (n = 28) and WT (n = 57) mice (housed at NCI) by Kaplan–Meier survival analysis and a log-rank test. B, Protein levels of downstream phospho-targets of mTOR, phospho-4E-BP1Thr37/46 and phospho-AKTS473, were evaluated in WT and KD tumor cells, as shown on a digital representation of a capillary-based protein electrophoresis (representative tumors; see Supplementary Fig. S1C for electrophoresis chromatogram). C, The CD4+, CD8+ status of WT (n = 2) and KD (n = 3) tumor cells was assessed by flow cytometry (representative scatterplots shown for a WT and KD tumor). D, To determine clonality, WT (n = 3) and KD (n = 3) tumors were assessed by nested PCR for T-cell antigen receptor rearrangement (PCR reactions were run in duplicate). The normal spleen samples had the expected polymorphic (nonclonal) pattern. Negative control was water, and positive control was known clonal T-cell lymphoma. E, Representative metaphase spreads from WT and KD tumor cells with t(14;15) at 100× magnification. Myc signals were green, and Tcra signals were red. Juxtaposed green and red signals (very close signals appear yellow) are indicative of the t(14;15) translocated chromosomes. Note that both metaphases showed two copies of the rearranged chromosome 15, der(15). F, Percentages of WT (n = 11) and KD (n = 8) tumors with or without Tcra/Myc translocation by FISH. G, MYC protein levels in representative WT and KD pretumor thymi (from 4-week-old mice) and thymic Pre-T LBL from WT and KD mice. H, Lck-Akt2 transgenic mice (housed at Fox Chase) treated with everolimus (n = 12) or vehicle (n = 10) starting at 8 weeks of age and continuing for 8 weeks or until mice showed signs of illness. Arrows indicate start and completion dates of the drug treatment. Survival was evaluated by Kaplan–Meier survival analysis and a log-rank test.
As Lck-Akt2 mice frequently develop a t(14;15) translocation involving a Tcra;Myc rearrangement (13), we hypothesized that the increased survival in KD mice could be associated with decreased translocation frequency. However, FISH screening for t(14;15) in tumor cells from 11 WT and 8 KD mice showed similar rates of translocations in KD cells (50% vs. 45% in WT; Fig. 2E and F). Four of each WT and KD tumors did not have the translocation, and two WT tumors had three copies of Myc indicative of trisomy 15 (13). MYC expression levels were similar in WT and KD tumor cells without the translocation and were lower in pretumor thymocytes (Fig. 2G). The similar proportion of t(14:15) observed in the WT and KD tumors indicates that the increased survival in the KD mice is not associated with decreased frequency of Myc translocations.
In parallel, Lck-Akt2; mTOR WT mice from an independent founder line (maintained at Fox Chase) were treated biweekly with the mTOR inhibitor everolimus which targets TORC1 (20 mg/kg; n = 12) or vehicle (n = 10), beginning at 8 weeks of age and continuing for 8 weeks or until humane endpoint. Mice treated with everolimus had significantly improved survival (median survival: 155 days vs. 73.5 days in vehicle-treated mice; P < 0.001; Fig. 2H), echoing results seen using a genetic approach with our Lck-MyrAkt2;mTOR KD (TORC1 and 2) mice.
let-7a is upregulated and CDK6 is downregulated in mTOR KD and rapamycin-treated WT tumors
We next performed gene expression profiling of thymic pre-T LBL tumors from WT and KD mice. A total of 342 genes were differentially expressed between WT and KD tumors (P < 0.05, fold change >2 or ≤−2; Supplementary Table S1).
Upstream regulator analysis was performed using the 342 genes differentially expressed in WT versus KD tumors (Fig. 3A, IPA). The top network from this analysis identified the miRNAs, let-7a and miR-21, as potential master regulators, along with Cdkn2a, Gata6 (GATA-binding protein 6), Hic1 (hypermethylated in cancer 1 protein), and Mitf (microphthalmia-associated transcription factor). Because no differences were observed in the microarray analysis between WT and KD tumors with respect to the genes Gata6, Hic1, Mitf, and Cdkn2a [validated for Cdkn2a: mRNA (qRT-PCR; Supplementary Fig. S2A], we evaluated expression levels of the miRNAs, miR-21, and let-7a, in WT versus KD tumors (Fig. 3B) as well as in response to rapamycin treatment in WT cells (Fig. 3C). In both cases, reductions in mTOR levels resulted in increase in miR21 (1.6–2.2-fold) and let-7a (3–3.3-fold) expression levels (Fig. 3B and C). Targets of let-7a and miR21 (TargetScan, Release 7.2) were among the most affected by constitutive reductions in mTOR (mTOR KD; Fig. 3A). KEGG analysis of genes differentially expressed (Supplementary Table S1) demonstrate the pathways significantly enriched (adjusted P value of ≤ 0.05); one of the most upregulated pathways was T-cell receptor signaling, and the several disease-associated pathways (e.g., Systemic Lupus Erythematosus, Parkinsons, and Huntingtons) were downregulated (Supplementary Fig. S2C; Supplementary Table S2).
Transcriptional profiling of WT and KD tumors. A, Upstream regulator analysis (Ingenuity) of genes differentially expressed (Supplementary Table S1) between WT (n = 4) and KD (n = 3) thymic lymphomas (green designates downregulated in KD mice and pink designates upregulated; numbers indicate fold change in KD mice relative to WT). Upstream regulators, miR-21 and let-7a, were increased in KD tumors versus WT (B) and in rapamycin-treated tumors versus untreated (C). Values for the target genes were normalized to endogenous control (SnoRNA-234) with the WT or WT-untreated values standardized to 1. D, Levels of Cdk6 transcripts were evaluated in WT and KD tumor cells by RT-PCR (P = 0.05). E, CDK6 protein levels were compared between pretumor thymocytes and tumors from representative WT and KD mice (representative mice). F, CDK6 and pS6 levels in untreated versus rapamycin-treated mTOR WT tumor cell lines (3 independent lines; see Supplementary Fig. S2B for CDK6 and CDK4 levels in additional untreated tumors). G, Cell viability of WT and KD tumor cells (n = 3/group) treated in vitro for 48 hours with 1 nmol/L rapamycin (mTORC1 inhibitor) relative to untreated cells. Error bars indicate the SE. H, Cell-cycle analysis [propidium idodide (PI) staining of untreated and treated (1 nmol/L rapamycin, 48 hours)] mTOR WT cells. I, Cell-cycle analysis (PI staining) of WT, KD, and KD cells transfected with CDK6 (from a representative experiment). J, Cell-cycle analysis (PI staining) of KD tumor cells and KD tumor cells transfected with a let-7 sponge (representative experiment). K, Low passage WT and KD tumor cell lines (n = 3 each) treated for 24 hours with a CDK4/6 inhibitor palbociclib over a range of doses. IC50 values between WT and KD tumor cell lines were significantly different (P <0.03). L, Low passage WT cells treated with single agents, palbociclib (P; 500 nmol/L) and rapamycin (R; 1 nmol/L), or the combination of the two drugs (R+P).
Cdk6, a target of both let-7a and miR-21, was one of the most downregulated genes in lymphomas from KD mice (P = 0.001, fold change = −10.7), and Cdk6 mRNA (Fig. 3D) and corresponding protein (Fig. 3E; Supplementary Fig. S2B) were strongly downregulated in lymphomas from KD mice. Although pretumor thymi from WT and KD mice had similar, low levels of CDK6 protein, WT tumors had increased levels of CDK6, whereas levels in KD tumors were undetectable (Fig. 3E; Supplementary Fig. S2B). CDK4 protein levels were similar between mTOR WT and KD thymic tumors (Supplementary Fig. S2B). In addition, when tumor cell lines generated from WT tumors (low passage) were treated with 1 nmol/L of rapamycin, an mTORC1 inhibitor, they did not have detectable levels of CDK6 (Fig. 3F). In fact, both low passage WT and KD tumor cells were sensitive to rapamycin (Fig. 3G), suggesting that chronically low mTOR levels did not lead to rapamycin resistance in the KD tumors. Treatment of low passage WT cells with rapamycin for 48 hours led to an increased percentage of cells in G1 phase (∼75% compared with ∼66% in untreated cells; Fig. 3H), similar to that seen in low passage KD cells (∼73% in G1; Fig. 3I). In three separate experiments, when Cdk6 was overexpressed in lymphoma cells from KD mice (Supplementary Fig. S2D), the cells were rescued from G1 arrest (∼60% in G1; Fig. 3I). Similarly, when let-7 activity was decreased using a let-7 sponge, CDK6 levels were increased (Supplementary Fig. S2E) and the number of cells in G1 decreased (Fig. 3J).
At this point, we explored the consequences of inhibiting CDK6 activity with palbociclib alone (Fig. 3K) and also coupled with mTOR inhibition by rapamycin (Fig. 3L) in low passage cells from WT tumors. KD tumor cells were not as sensitive (IC50 = 1.04 μmol/L) to palbociclib treatment as WT cells (IC50 = 0.02 μmol/L) 24 hours after treatment (P < 0.03; Fig. 3K). WT tumor cell viability was decreased to a greater extent by a combination of rapamycin (mTORi, 1 nmol/L) and palbociclib (CDK4/6i, IC50 = 625 nmol/L) than by either single agent alone (Fig. 3L). From these data, we have developed a hypothetical signaling pathway (Fig. 4) for mTOR KD or rapamycin treatment leading either directly or indirectly to upregulation of let-7 and miR-21 and downregulation of Cdk6.
Hypothetical signaling model for upregulation of let-7 and miR21 leading to the downregulation of Cdk6 after genetic or pharmacologic downregulation of mTOR in AKT-induced tumors. Genetic KD of mTOR expression in transgenic mice via expression of Mtortm1.Lgm (mTOR KD) instead of WT Mtor or by treatment with the mTOR inhibitor rapamycin (Rapa) leads to upregulation of let-7 and miR-21 and downregulation of Cdk6. Genetic or pharmacologic downregulation of mTOR signaling combined with inhibition of CDK6 activity by palbociclib (Palb) results in decreased tumor cell proliferation compared with treatment with Rapa or Palb alone.
In vivo treatment with the mTORi/CDK6i combination results in reduced tumor size and causes G1 arrest
Nude mice (N = 4/group) bearing palpable tumors (from transplanted Lck-MyrAkt2;mTOR WT tumor cells) were randomized into one of four treatment groups (vehicle, rapamycin, palbociclib, or a combination of rapamycin and palbociclib) and treated for 5 days (photomicrographs of flank tumors in Supplementary Fig. S3A). Tumors were significantly smaller in all treatment groups compared with the vehicle control cohort, but tumors were smallest with the rapamycin/palbociclib combination (Fig. 5A). The average percentage of cells positive for Ki67, a proliferation rate marker, was reduced in all treatment groups compared with vehicle (P < 0.01), and to the greatest extent in the palbociclib alone and combination groups (Fig. 5B). Individual drugs and the drug combination induced greater G1 arrest of cells isolated from the flank tumors than vehicle treatment (Fig. 5C; P < 0.01). Apoptosis, as measured by Annexin V/PI staining of tumor cells collected at treatment day 5, was not significantly altered by any treatment (Supplementary Fig. S2F). These flank tumor data are consistent with the decreased cell viability noted in the low passage murine tumor cells treated with the drug combination. Flank tumors treated with the combination therapy for only 5 days had lower levels of CDK6, p-RB (indicating an increase in RB activity), p-AKT, and p-S6 (Fig. 5D).
In vivo treatment of tumor transplants, an in vivo model of disseminated leukemia, and human T-ALL cell lines with rapamycin, palbociclib, and the combination. A, The average weight of tumors generated from Lck-Akt2;mTOR WT tumor cells injected into the flanks of nude mice and treated individually or with the combination of mTOR (rapamycin, R) and CDK4/6 (palbociclib, P) inhibitors for 5 days. Error bars represent SEM for the tumors from the 4 mice/group. B, Tumors from 3 mice/treatment group were labeled for Ki67 (proliferation marker), and percent positive Ki67 nuclei in each tumor group is depicted. C, Percent cells in G1 in tumor transplants treated with single agents or the combination. D, Immunoblots from representative WT flank tumor transplants after 5-day treatment with vehicle, palbociclib, rapamycin, and the combination. Proteins designated with a “p-” are the phosphorylated forms of the protein. E, Survival curves of mice (N = 8/group) tail vein injected with 5 × 106 WT pre-T LBL tumor cells and treated with vehicle, either single agent, or the combination. Red arrows indicate the cycle of drug treatment. Mice were off drug treatment for 1 week between cycles 1 and 2. Statistical analysis was done using Mantel–Cox log-rank testing with Bonferroni correction. F, Graphs show the EOHSA for T-ALL/LBL cell lines treated for 48 hours with increasing doses of palbociclib (CDK6i, 0 nm-1 μmol/L, starting at 62.5 nmol/L and increasing 2-fold for each dose) and rapamycin (mTORi, 0 nmol/L to 100 nmol/L, increasing 10-fold for each dose), representing the difference in viability between cells treated by the single agent and the combination at each dose. Values are averaged results from four replicate experiments. See Supplementary Fig. S4A and S4B for dose matrix and synergy scores.
To better model disseminated leukemic disease, 5 × 106 WT tumor cells were injected intravenously into 32 nude mice. Mice were randomized into four treatment groups (8 mice/group) 4 days after transplant, and were treated with vehicle, rapamycin, palbociclib, or both drugs once daily, 5 days per week with treatment cycles of 3 weeks on treatment/1 week off. Mice treated with the drug combination had a significantly prolonged average survival of 45 days (log-rank test with Bonferroni correction, P = <0.0006), whereas mice treated with rapamycin or palbociclib survived an average of 20 days (Fig. 5E). Vehicle-treated mice succumbed to the tumor burden by an average of 14 days, and tumor cells were widely distributed in examined tissues, including kidney, bone marrow, spleen, and lymph nodes (Supplementary Fig. S3B).
The combination of mTOR and CDK4/6 inhibition is cooperative in human T-ALL/LBL cell lines
To extend the translational implications of our experimental mouse model, we examined the combined effect of mTOR and CDK4/6 inhibition in human T-ALL/LBL cell lines. Human T-ALL cell lines were selected for study as both T-ALL and T-LBL fall under the same WHO classification, and have only minor variations based on bone marrow involvement and differential gene expression; these tumors share morphologic, immunophenotypic, and genetic similarities (36).
Treatment of T-ALL/LBL cell lines with palbociclib and rapamycin in a combination dose matrix showed dose-responsive combined activity on cell viability in all of the cell lines with cooperative action (positive values) seen in EOHSA (Fig. 5F; Supplementary Fig. S4A). Notably, synergy CI scores indicated a high degree of synergy (CI < 1; ref. 32) in most cell lines (Supplementary Fig. S4B).
Discussion
AKT is considered a pivotal player in tumorigenesis, as its hyperactivation results in the phosphorylation of a wide variety of downstream target proteins involved in regulating cell survival and other aspects of tumor formation (37, 38). One of its targets, mTOR kinase, stimulates cell proliferation through the regulation of ribosomal biogenesis and mRNA translation of several growth regulation genes (39, 40) and also functions as a nutrient sensor to promote changes in cell size and cell cycling (41). The mTOR pathway, through AKT signaling, also regulates cell-cycle progression by phosphorylating and inhibiting GSK3β, a kinase that phosphorylates and mediates the degradation of cyclin D1, as well as by phosphorylating the cell-cycle inhibitors p21WAF1 and p27Kip1, causing their cytoplasmic retention and preventing their binding/inhibition of cyclin/CDK complexes (37). Because mTOR represents only one of many prooncogenic downstream effectors of AKT, and mTOR KD results in only partial loss of mTOR kinase activity, it is not surprising that Lck-MyrAkt2;mTOR KD mice eventually develop thymic tumors. The delayed tumor onset is likely due to the cytostatic effect of mTOR KD and its consequent inhibition of cell cycling.
By crossing a genetically engineered mouse model that spontaneously develops pre-T LBL tumors with a model of constitutively decreased mTOR expression (9–11), we were able to evaluate the molecular sequelae of mTOR inhibition during T-cell tumorigenesis in vivo. Both genetic and pharmacologic mTOR inhibition was found to significantly impair tumor development and progression, with a substantial increase in overall survival for both mTOR-inhibited groups. The miRNAs let-7a and miR-21 were found to be upstream regulators differentially expressed between tumors from Lck-Akt2;mTOR WT and Lck-Akt2;mTOR KD mice. CDK6, a critical cell-cycle regulator involved in T-cell development (42) was determined to be one of the most repressed targets in mTOR KD tumors and is known to be a direct target of both let-7a and miR21, as validated by luciferase reporter assays (15, 16). Indeed, reduction of let-7a levels in the mTOR KD tumor cells, increased CDK6 levels. As we (30, 43) and others (44) have found that mTOR inhibition alone delays rather than blocks tumor progression, we evaluated combined inhibition of mTOR and CDK4/6 activity in vivo in the setting of disseminated disease in mice and in vitro in human T-ALL cell lines. Although the antitumor activity of single agent mTOR or CDK4/6 inhibition was found to be limited in this aggressive transplanted tumor model, the combination of the two agents showed cooperativity in extending survival. It is notable that in the disseminated leukemic model, all mice given rapamycin or palbociclib alone died in between treatment cycles, while mice given the combination lived twice as long.
There remains a critical need for additional therapies to treat both children and adults with T-ALL, especially in cases of early relapse after current chemotherapy regimens, where prognosis is particularly poor. The addition of targeted agent combinations to the clinical armamentarium may offer treatment strategies that may avoid some of the side effects seen with many current chemotherapeutic regimens. Drug combinations that include CDK4/6 inhibitors have shown preclinical efficacy in a number of cancers, including T-ALL (19, 45–48). Our data suggest that cotargeting mTOR and CDK6 may be the equivalent of targeting drugs in the same pathway.
Our mouse model that mimics inhibition of all mTOR complexes provides an efficient way to evaluate drug combinations that synergize with mTOR inhibition. Our studies involving the cotargeting of mTOR and CDK6 suggest that this particular combination may be limited in that its effects are predominantly cytostatic resulting in a large proportion of cells arrested in G1, but not undergoing apoptosis. These studies point to using our mTOR mouse model to focus on targets that when inhibited will result in apoptotic cell death, and to explore interrupting pathways distinct from mTOR.
Disclosure of Potential Conflicts of Interest
No potential conflicts of interest were disclosed by the authors.
Authors' Contributions
J.M. Gary: Conceptualization, data curation, formal analysis, validation, investigation, methodology, writing-original draft, writing-review and editing. J.K. Simmons: Conceptualization, data curation, formal analysis, validation, investigation, visualization, methodology, writing-original draft. J. Xu: Data curation. S. Zhang: Data curation, validation, methodology, writing-review and editing. T.J. Peat: Data curation, formal analysis, validation, methodology, writing-review and editing. N. Watson: Data curation, visualization. B.J. Gamache: Data curation, formal analysis, writing-review and editing. K. Zhang: Data curation, formal analysis, methodology. A.L. Kovalchuk: Data curation, methodology, writing-review and editing. A.M. Michalowski: Data curation, software, formal analysis. J.-Q. Chen: Data curation, formal analysis. T. Thaiwong: Data curation, formal analysis. M. Kiupel: Data curation, formal analysis, investigation. S. Gaikwad: Data curation, formal analysis, investigation, writing-review and editing. M. Etienne: Data curation. R.M. Simpson: Data curation, formal analysis. W. Dubois: Data curation, formal analysis, methodology. J.R. Testa: Conceptualization, resources, data curation, formal analysis, writing-original draft, writing-review and editing. B.A. Mock: Conceptualization, resources, data curation, software, formal analysis, supervision, funding acquisition, validation, investigation, visualization, methodology, writing-original draft, project administration, writing-review and editing.
Acknowledgments
We thank Doug Lowy, Peter Aplan, Remy Bosselut, Hussain Perwez, and Ashish Lal for their suggestions during experimental design and manuscript preparation. We also thank Shelley Hoover, Jennifer Dwyer, Adrian Eiden, Dena Tran, Zaw Phyo, Solomon Lynch, and Kenneth Felsenstein for insights and assistance with the studies. J.M. Gary was a fellow in the NIH Comparative Biomedical Scientist Training Program supported by the NCI in a partnership with Michigan State University.
Work performed by J. Xu and J.R. Testa was supported by grants R01 CA77429 and P30 CA06927 from the NCI. The remainder of the work was supported by the Intramural Research Program of the NIH, NCI, Center for Cancer Research ZIA BC011065. The content of this publication does not necessarily reflect the views or policies of the Department of Health and Human Services, nor does mention of trade names, commercial products, or organizations imply endorsement by the U.S. government.
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.
Footnotes
Note: Supplementary data for this article are available at Molecular Cancer Therapeutics Online (http://mct.aacrjournals.org/).
Mol Cancer Ther 2020;19:2221–32
- Received July 5, 2019.
- Revision received January 14, 2020.
- Accepted July 13, 2020.
- Published first August 3, 2020.
- ©2020 American Association for Cancer Research.
References
Volume 19, Issue 10
