Skip to main page content
Access keys NCBI Homepage MyNCBI Homepage Main Content Main Navigation
. 2014 Jun 19;510(7505):402-6.
doi: 10.1038/nature13239. Epub 2014 May 4.

PTEN Action in Leukaemia Dictated by the Tissue Microenvironment

Free PMC article

PTEN Action in Leukaemia Dictated by the Tissue Microenvironment

Cornelius Miething et al. Nature. .
Free PMC article


PTEN encodes a lipid phosphatase that is underexpressed in many cancers owing to deletions, mutations or gene silencing. PTEN dephosphorylates phosphatidylinositol (3,4,5)-triphosphate, thereby opposing the activity of class I phosphatidylinositol 3-kinases that mediate growth- and survival-factor signalling through phosphatidylinositol 3-kinase effectors such as AKT and mTOR. To determine whether continued PTEN inactivation is required to maintain malignancy, here we generate an RNA interference-based transgenic mouse model that allows tetracycline-dependent regulation of PTEN in a time- and tissue-specific manner. Postnatal Pten knockdown in the haematopoietic compartment produced highly disseminated T-cell acute lymphoblastic leukaemia. Notably, reactivation of PTEN mainly reduced T-cell leukaemia dissemination but had little effect on tumour load in haematopoietic organs. Leukaemia infiltration into the intestine was dependent on CCR9 G-protein-coupled receptor signalling, which was amplified by PTEN loss. Our results suggest that in the absence of PTEN, G-protein-coupled receptors may have an unanticipated role in driving tumour growth and invasion in an unsupportive environment. They further reveal that the role of PTEN loss in tumour maintenance is not invariant and can be influenced by the tissue microenvironment, thereby producing a form of intratumoral heterogeneity that is independent of cancer genotype.

Conflict of interest statement

The authors declare no competing financial interest.


Extended Data Figure 1
Extended Data Figure 1. Pten shRNA-transgenic mice enable conditional expression of PTEN and develop tumors after prolonged PTEN knockdown
a, Western blot (WB) analysis of PTEN protein knockdown in NIH 3T3 cells infected with different Pten shRNAs at low multiplicity of infection. (s.e.: short exposure, l.e.: long exposure). b, PTEN protein knockdown assessed by WB in ES cell clones targeted with two different Pten shRNAs, either treated with doxycycline (Dox) or left untreated. c, MEFs from Rosa26-rtTA;shPten.1522 transgenic mice, wild type control mice, or Pten+/− mice were treated with Dox for the indicated times and analyzed for PTEN, GFP, and ACTB expression by WB. d, Overall survival of mice receiving bone marrow cells from tTA-transgenic mice infected with an inducible TRE-GFP-miR30 (TGM) retroviral vector expressing shPten.1522, shPten.2049 or control after irradiation with 600 rad. e, Fluorescence image of a CAGGS-rtTA;shPten.1522 mouse on Dox for 5 days and a CAGGS-rtTA only control mouse. f, Flow cytometric analysis of the peripheral blood of a CAGGS-rtTA;shPten.1522 mouse on Dox and an off Dox control mouse for myeloid (CD11b) and GFP marker expression 10 days after initiating Dox food. g, Overall survival curve of CMV-rtTA;shPten.1522 double-transgenic and control mice (single transgenic shPten.1522 or CMV-rtTA). Dox treatment for shRNA induction was started after weaning (at ~4 weeks of age). h, Situs of a tumor-bearing CAGGS-rtTA;shPten.1522 double-transgenic mouse. A large thymic tumor (full arrow), as well as enlarged lymph nodes (dashed arrows) and spleen (arrowhead), are visible. i, Immunohistochemical staining of kidney sections from a CAGGS-rtTA;shPten.1522 mouse for the indicated antigens. Arrows highlight a tumor infiltrate around a kidney venule. Scale bars show 100 μm.
Extended Data Figure 2
Extended Data Figure 2. Vav-tTA;shPten transgenic mice with targeted shPten expression in the hematopoietic lineage display thymic hyperplasia by 6 weeks, and a subset develops thymic tumors infiltrating multiple peripheral organs
a, Brightfield (left) and fluorescence (right) images of spleen and thymus from Vav-tTA;shLuc (control) and Vav-tTA;shPten double-transgenic mice at 5 weeks. b, FACS analysis of spleen and thymus single cell suspensions from Vav-tTA;shLuc/shPten mice for CD4/CD8 expression. c, Immunohistochemical (IHC) analysis of thymic tissue from 6 week old Vav-tTA;shLuc and Vav-tTA;shPten mice. Sections were stained with haematoxylin/eosin (H/E), anti-GFP or anti-PTEN antibodies, showing heterogeneous GFP staining and correspondingly variable PTEN knockdown. Scale bars are 200 μm for H/E and GFP, and 100 μm for PTEN. The insets are 2× magnifications. d, Thymus weight of 6-week old Vav-tTA;shLuc and Vav-tTA;shPten mice (n=5 for both groups, P<0.006 by t-test). e, Brightfield and GFP-fluorescence images of a Vav-tTA;shPten mouse with tumors. f, IHC staining of spleen, liver, and kidney tissues from a Vav-tTA;shPten mouse with primary T-cell disease. Sections were stained for H/E, GFP, PTEN and phospho-AKT(S473) as indicated, showing heterogeneous staining due to variable tumor infiltration. Scale bars represent 100 μm (insets 25 μm).
Extended Data Figure 3
Extended Data Figure 3. Immunophenotype, chromosomal aberrations and Notch1 mutations observed in murine shPten tumors
a, Flow cytometric analysis of organ infiltration by primary tumors in Vav-tTA;shPten.1522 transgenic mice. Single cell suspensions of indicated tissues were analyzed for EGFP, Thy1.2, CD4 and CD8 expression. LN: lymph node, BM: bone marrow, PB: peripheral blood. b, Spectral karyotyping analysis of a T-cell tumor arising in Vav-tTA;shPten.1522 mice, showing a t(14;15) translocation. c, Comparative genomic hybridization (CGH) analysis of a Vav-tTA;shPten.1522 leukemia. Genomic tumor DNA was analyzed on Affymetrix CGH SNP arrays and compared to normal skin tissue. X-axis indicates genomic coordinate and y-axis represents log2(tumor/germline). d, Schematic of the murine NOTCH1 protein generated using protein paint (, highlighting the different NOTCH1 protein domains and the mutations detected in the murine shPten T-ALL tumors.
Extended Data Figure 4
Extended Data Figure 4. Summary of karyotyping and Notch1 sequencing of shPten TALL tumors
a, Results from a multiplex FISH analysis of three different primary shPten induced T-ALL tumors. At least 10 cells were analyzed for each sample, and chromosomal gains, deletions or translocations are highlighted. b, Summary of Notch1 mutations identified in shPten and Pten KO tumors. c, Sequence of all shRNAs targeting murine Pten that were tested in the study. Sense and guide strand are highlighted in red.
Extended Data Figure 5
Extended Data Figure 5. Gene Set Enrichment Analysis shows similar gene expression patterns in human and mouse T-ALL lacking Pten
a, Gene Set Enrichment analysis (GSEA) of a mouse shPten signature in PTEN-altered human T-ALL was tested after establishing the shPten-dependent signature using the 100 most upregulated genes in shPten T-ALL samples (untreated, n=3) against PTEN-restored samples (Dox-treated, n=4) as determined by RNA-seq analysis (data not shown). Publicly available human T-ALL gene expression profiles (GSE28703, n=47) were processed using RMA (quantile normalization) and supervised for PTEN status (PTEN altered including PTEN deletion, mutation or both, n=10; PTEN wild-type (WT), n=37) according to the published sample annotation. Statistical significance of GSEA results was assessed using 1000 samples permutations. b, For enrichment of human PTEN TALL signature in mouse shPten knockdown (kd) T-ALL (Dox-off) profiles against Pten-restored (Dox-on) profiles, a human PTEN-disrupted signature was generated by including the 100 most upregulated genes in PTEN-disrupted vs PTEN-wild type T-ALL samples. Mouse genes were ranked by supervising untreated to Dox-treated shPten T-ALLs. Statistical significance of human PTEN-disrupted signature enrichment was assessed using 1000 gene set permutations.
Extended Data Figure 6
Extended Data Figure 6. Secondary recipients of shPten T-ALL cells display extensive intestinal tumor infiltration similar to a subset of human patients characterized by peripheral T-cell lymphoma and low PTEN expression
a, Overall survival of sublethally irradiated Rag1−/− mice transplanted with 1×105 T-ALL cells from primary Vav-tTA;shPten.1522 mice compared to untransplanted mice (n=5 for both groups, P<0.003). b, IHC staining for EGFP expression in the indicated tissues from secondary T-ALL transplant recipients. Scale bars represent 400μm and 100 μm for insets. c, Overall survival of PTEN normal (WT) vs. PTEN altered patients with T-ALL analyzed from published data on patients with T-ALL, P=0.02). PTEN altered (n=20) include patients with PTEN deletion, mutation, underexpression (<0.8 sigma after z scoring) and any combination of such alterations, PTEN normal (n=62) include all other patients with available data. d, PTEN IHC staining of tissue micro-arrays of tumor sections from MSKCC patients with peripheral T-cell lymphomas (PTCL). Examples of low (upper panel) and high (lower panel) PTEN expression samples are shown. e, Contingency table showing a significant association (p<0.003; Fisher Exact Test) between low expression of PTEN and intestinal infiltration in PTCL patients. f, Overall survival of Rag1−/− mice transplanted with T-ALL cells from Ptenfl/fl; Lck-Cre mice ±Dox (n=5 for each group). g, Weight of spleen (n=4) and lymph nodes (n=8) in Rag1−/− mice transplanted with Vav-tTA;shPten leukemic cells untreated or treated with Dox for 5 days. h, MRI of Rag1−/− mice transplanted with Vav-tTA;shPten leukemic cells untreated or treated with Dox for 5 days, 14 days after transplant. Arrows highlight lymph nodes (LN) and increased signals in the liver. Representative images for one out of 3 analyzed mice per condition are shown. i, MRI imaging of the intestine and liver of the same mice as in h are shown. Dashed arrows highlight the liver, full arrows the intestine.
Extended Data Figure 7
Extended Data Figure 7. Pten reactivation affects multiple pathways and increases apoptosis in tumor cells infiltrating the intestine, but not in the spleen
a, Heatmap of top 30 upregulated and b, downregulated genes after Pten reactivation as determined by RNAseq on CD4-sorted leukemic samples isolated from the spleen. Three mice with Pten knocked down and four mice with reactivated Pten were analyzed. Pten is one of the top 50 upregulated genes after reactivation, but not included on the list. c, Bubblegraph visualization of the most significantly affected pathways as determined by DAVID pathway analysis. Y-axis represents relative pathway enrichment in Pten reactivated vs. Pten knockdown leukemic cells, and size of the bubble graph is inversely proportional to p-value. d, IHC analysis for expression of GFP and PTEN in spleen, lymph node (LN) and liver from shPten-tumor transplanted mice ±Dox treatment (5 days after start of Dox treatment; n=3 per group). Representative sections are shown. Scale bars are 100 μm for full images and 20 μm for insets. e, In vivo BrdU uptake in leukemic cells isolated from the lymph nodes of mice transplanted with Vav-tTA;shPten primary T-ALL tumors ±Dox. n=3 for each group. f, TUNEL staining of spleen and intestinal sections of Rag1−/− mice serially transplanted with Vav-tTA;shPten leukemia cells and either left untreated or treated with Dox 24h before sectioning. Scale bars are 200μm (2.5×) and 50μm (10×). g, Quantification of TUNEL stained sections from the intestinal sections in f. TUNEL positive cells from three representative areas of 1 mm2 from two different intestine sections were counted for each condition (P<0.01).
Extended Data Figure 8
Extended Data Figure 8. Akt and S6 protein phosphorylation is affected by PTEN reactivation in the intestine
a, Immunohistochemical (IHC) staining for phospho-S6 (pS235/236-S6) and phospho-AKT (pS473AKT) of spleen sections from Rag1−/− mice transplanted with Vav-tTA;shPten tumor cells from primary mice and either treated with Dox or left untreated two days after treatment begin (n=3 per group). Scale bars are 100 μm, 20 μm for insets. Representative images are shown. b, IHC staining for pS473-AKT (bottom) in the intestine, showing very low pAKT signal in the intestinal epithelial cells independent of Dox treatment status (arrows; bottom left and right panels), conversely strong staining for pAKT was detected in some of the infiltrating tumor cells (arrow heads). The signal was reduced concomitantly with the overall reduction of the Pten-shRNA linked GFP signal (top) after 36 h of Dox treatment (+ Dox; right panels). c, Representative histogram of flow cytometric analysis for intracellular pS6 signal in CD4+ cells isolated from spleen and intestine of Rag1−/− mice transplanted with shPten tumor cells and either treated with Dox for 5 days or left untreated. d, Flow cytometric quantification of pS6 signal in CD4+ cells isolated from the intestine and e, spleens of Rag1−/− mice transplanted with primary shPten tumors and treated ±Dox for 5 days (n=4 for each condition, P<0.04 for the intestine and n.s. for the spleen by paired t-test). MFI: mean fluorescent intensity.
Extended Data Figure 9
Extended Data Figure 9. NCR mice display a reduced intestinal tumor infiltration, which is not dependent on the absence of the thymus
a, Brightfield pictures of the intestinal situs of Rag1−/− and NCR nude mice serially transplanted with shPten tumors (upper four panels) and fluorescence images (FI) of cells infiltrating the small intestine in these mice (lower four panels). Scale bars are 800 μm (upper FI panels) and 100 μm (lower FI panels). Pictures were taken on a Nikon SMZ 1000 stereomicroscope. b, Quantification of the intestinal infiltration in transplanted Rag1−/− or NCR mice by flow cytometry (P<0.03). c, Weight of lymph nodes (P<0.01) and d, spleens (P=n.s.) in transplanted Rag1−/− and NCR mice. e, CCL25 expression in the small intestine of Rag1−/− and NCR mice measured by ELISA. f, Western blot analysis of CCL25 expression in the small intestine of Rag1−/− and NCR mice. g, Overall survival of Rag1−/− and thymectomized Rag1−/− mice after transplant with shPten T-ALL cells (n=5 per group). h, H/E and immunohistochemical analysis of CD3 expression of spleen, liver and intestine from Rag1−/−and thymectomized Rag1−/− mice transplanted with shPten T-ALL cells. Scale bars represent 200 μm for spleen and liver and 100 μm for intestinal samples. i, Flow cytometric measurement of CCR9 expression on shPten leukemia cells either in the absence of Dox (Pten knocked down) or Dox treated (Pten reactivated). One representative analysis out of four analyzed on/off Dox pairs is shown. A CCR9 negative B-cell line was used as control. j, Immunoblot analysis of PTEN, phospho-AKT(S473) and ACTB expression in human HBP-ALL T-ALL cells infected with either a control shRNA (shRenilla) or a shRNA targeting PTEN, and either starved or stimulated for 15 min with 500 ng/ml CCL25. k, shPten tumor cell migration across a Boyden chamber in the presence or absence of 1 μg/ml Dox and 500 ng/ml CCL25. One representative experiment of two is shown, samples were run in triplicate; ** p<0.01, *** p<0.001 by t-test.
Extended Data Figure 10
Extended Data Figure 10. CCR9 inactivation by shRNA knockdown or by pharmacologic inhibition attenuates intestinal tumor infiltration
a, CCR9 expression on the surface of shPten tumor cells either infected with a control shRNA against Renilla luciferase (left) or with a shRNA targeting Ccr9 (right) as measured by flow cytometry, compared to uninfected cells respectively. b, Flow cytometry-based quantification of CCR9 suppression in shCcr9 infected shPten T-ALL cells compared to shRenilla-infected cells, n=5 for each cohort. c, Raw percentage of shRenilla/shCcr9 expressing shPten T-ALL cells in different tissue compartments of mice 12 days after transplantation, determined by flow cytometry, n=5 for each cohort. p<0.0005 (intestine). d, Immunohistochemical (IHC) analysis for mCherry (left: shRenilla-mCherry; right:shCcr9-mCherry) expressing cells in tissue sections of mice from c. Spleen, liver and intestinal sections of mice transplanted with shRenilla- or shCcr9-infected T-ALL cells were analyzed for mCherry expression. Representative stains from one mouse out of three analyzed mice are shown. Scale bars represent 100 μm (insets 20 μm). e, IHC staining for EGFP expression in representative sections of small intestine, liver and spleen of Vav-tTA;shPten tumor bearing mice treated with vehicle or the CCR9 inhibitor CCX8037 (n=3). Scale bars are 400 μm for the 2.5× and 100 μm for the 10× images. f, Flow cytometric quantification of intestinal tumor infiltration in Rag1−/− mice transplanted with Vav-tTA;shPten leukemia cells and treated with vehicle (n=4) or a small molecule inhibitor of CCR9 (n=5). *P<0.05 by t-test. g, Immunoblot analysis of phospho-AKT expression 15 min after stimulation of shPten leukemia cells with CCL25 in the absence or presence of indicated concentrations of CCX8037. h, IHC analysis of EGFP and phospho-AKT signal in representative sections of small intestine from Vav-tTA;shPten tumor bearing mice treated with vehicle or the CCR9 inhibitor CCX8037. Scale bars are 100 μm (25 μm for insets).
Figure 1
Figure 1
Pten shRNA transgenic mice develop disseminated CD4/CD8 double-positive (DP) T-cell leukemia. (A) Outline of the targeting construct and the ES cell targeting strategy. SA –splice acceptor site. pA – polyadenylation site. TRE – tetracycline responsive element promoter. EGFP – enhanced green fluorescent protein. PGK – phosphoglycerate kinase promoter. ATG*-truncated ATG sequence. FRT – FLP recognition target. *Hygromycin – ATG-less hygromycin cDNA. (B) Immunoblot (WB) analysis of murine embryonic fibroblasts from shPten.1522;Rosa26-rtTA2 transgenic mice ±doxycycline (Dox) for 5 days at indicated timepoints after stimulation with 100 nM insulin. (C) Overall survival of Vav-tTA;shPten mice (n=49) and controls (n=98, P<0.001 by log-rank). (D) Flow cytometric analysis of a representative primary Vav-tTA;shPten tumor for EGFP, Thy1.2, CD4 and CD8 (n=10). (E) WB analysis of T-cell tumors from Trp53−/−, Ptenfl/fl;Lck-Cre and Vav-tTA;shPten mice for the indicated proteins. (F) PTEN immunohistochemistry (IHC) of bone marrow samples of 31 human patients with T-ALL categorized as positive (upper left panel) or low/negative (lower left panel). Association of PTEN expression with status for disseminated disease was calculated using a contingency table (Fisher’s Exact Test).
Figure 2
Figure 2
The impact of PTEN reactivation on leukemia viability is influenced by anatomical site. (A) Overall survival of Rag1−/− mice transplanted with 1×105 cells from Vav-tTA;shPten (shPten) tumors and treated with Dox (Pten on; n=14), or untreated controls (Pten off; n=15), P<0.0001 by log-rank test. (B) WB analysis of splenic tumor cells from control, untreated, and mice treated with Dox for 5d. (C) Brightfield and EGFP images of lymph nodes and spleen from an untreated mouse (Pten off) and mouse treated with Dox for 5d (Pten on) (n=10). (D) Flow cytometric analysis of CD4, EGFP and CD8 expression in tumor cells from the peripheral blood of mice ±Dox for 5d (n=10). (E) IHC analysis for CD3 expression in the spleen, liver and small intestine from shPten T-ALL transplanted mice ±Dox for 5d (n=3 per group). Scale bars:100 μm (20 μm in insets). (F) Relative tumor infiltration in the indicated organs of transplanted Rag1−/−mice off (n=7) and on (n=7) Dox, quantified by flow cytometric analysis of CD4+ cells; * P<0.05, ** P<0.01 by t-test. (G) IHC staining for cleaved caspase 3 in the spleen and intestine from mice 10d after transplant with shPten T-ALL either treated with Dox for 36h or left untreated. Representative sections from one of three mice per cohort are shown. Scale bars: 100 μm (20 μm in insets).
Figure 3
Figure 3
Tissue-dependent effects of PTEN reactivation on PI3K signaling. (A) Small intestinal sections from shPten T-ALL transplanted mice ±Dox were stained with hematoxylin/eosin (HE) or by IHC for the indicated molecules. Representative sections from one of three mice per cohort are shown. Scale bars represent 100 μm (20 μm in insets). (B) Serial 18F-FDG PET analysis of shPten T-ALL transplanted mice before and 2d after beginning of Dox treatment. White arrows: BM, red arrow: spleen, green arrow: liver/intestine. Representative images from two out of 12 analyzed mice are shown. (C) CD3 IHC staining of shPten tumor infiltrations in the liver of mice ±Dox 4d after treatment initiation (n=3 per group). Arrows highlight CD3+ tumor infiltrates. Scale bars represent 500 μm. (D) 18F-FDG PET/CT analysis of shPten T-ALL transplanted mice ±Dox 4d after beginning of Dox treatment. Full arrow: spleen, dashed arrow: liver/intestine. Representative images from two out of six analyzed mice are shown.
Figure 4
Figure 4
CCL25-CCR9 chemokine signaling contributes to leukemia dissemination. (A) Overall survival of Rag1−/− mice (n=9; −Dox, n=7; +Dox) and NCR mice (n=7; −Dox, n=7; +Dox) transplanted with 1×105 shPten leukemia cells. Survival of Rag1−/−(−Dox) vs. NCR(−Dox) mice; P<0.0001 by log-rank test. (B) IHC staining for EGFP in intestinal sections from Rag1−/− and NCR mice transplanted with shPten leukemia (n=3 per cohort). Scale bars: 100 μm. Numbers show mean fraction (±SD) of infiltrating EGFP+ tumor cells of total viable cells as determined by flow cytometry (P<0.03 by t-test). (C) Representative images of lymph nodes and spleens from Rag1−/− and NCR mice (n=7). (D) CCR9 receptor expression on shPten leukemia cells, normal CD4/CD8 double-positive and CD4 or CD8 single-positive thymic T-cells measured by flow cytometry (n=3 per group). (E) WB analysis of indicated proteins in shPten leukemic cells ±Dox after stimulation with 500 ng/ml CCL25 for the indicated time. (F) WB of indicated proteins in human T-ALL1 cells infected with either shRenilla control or shPTEN.1522 and ±stimulation with CCL25 for 15 min. (G) Outline of competition experiment of untransduced vs. shCcr9-mCherry- or control shRenilla-mCherry transduced EGFP+ shPten tumor cells. (H) Normalized ratio of mCherry+;EGFP+ cells over all EGFP+ cells isolated from spleen, liver and intestine of 5 mice per cohort from 2 independent transplantations. Cells were analyzed by flow cytometry and normalized to the mCherry/EGFP ratio in the spleen to account for differences in transduction rate.

Similar articles

See all similar articles

Cited by 23 articles

See all "Cited by" articles


    1. Li J, et al. PTEN, a putative protein tyrosine phosphatase gene mutated in human brain, breast, and prostate cancer. Science. 1997;275:1943–1947. - PubMed
    1. Engelman JA, Luo J, Cantley LC. The evolution of phosphatidylinositol 3-kinases as regulators of growth and metabolism. Nat Rev Genet. 2006;7:606–619. doi: 10.1038/nrg1879. - DOI - PubMed
    1. Salmena L, Carracedo A, Pandolfi PP. Tenets of PTEN tumor suppression. Cell. 2008;133:403–414. doi: 10.1016/j.cell.2008.04.013. - DOI - PubMed
    1. Dickins RA, et al. Probing tumor phenotypes using stable and regulated synthetic microRNA precursors. Nature genetics. 2005;37:1289–1295. doi: 10.1038/ng1651. - DOI - PubMed
    1. Dickins RA, et al. Tissue-specific and reversible RNA interference in transgenic mice. Nat Genet. 2007;39:914–921. doi: 10.1038/ng2045. - DOI - PMC - PubMed

Publication types

MeSH terms