Physiological levels of the PTEN-PI3K-AKT axis activity are required for maintenance of Burkitt lymphoma

Abstract

In addition to oncogenic MYC translocations, Burkitt lymphoma (BL) depends on the germinal centre (GC) dark zone (DZ) B cell survival and proliferation programme, which is characterized by relatively low PI3K-AKT activity. Paradoxically, PI3K-AKT activation facilitates MYC-driven lymphomagenesis in mice, and it has been proposed that PI3K-AKT activation is essential for BL. Here we show that the PI3K-AKT activity in primary BLs and BL cell lines does not exceed that of human non-neoplastic tonsillar GC DZ B cells. BLs were not sensitive to AKT1 knockdown, which induced massive cell death in pAKT^high DLBCL cell lines. Likewise, BL cell lines show low sensitivity to pan-AKT inhibitors. Moreover, hyper-activation of the PI3K-AKT pathway by overexpression of a constitutively active version of AKT (myrAKT) or knockdown of PTEN repressed the growth of BL cell lines. This was associated with increased AKT phosphorylation, NF-κB activation, and downregulation of DZ genes including the proto-oncogene MYB and the DZ marker CXCR4. In contrast to GCB-DLBCL, PTEN overexpression was tolerated by BL cell lines. We conclude that the molecular mechanisms instrumental to guarantee the survival of normal DZ B cells, including the tight regulation of the PTEN-PI3K-AKT axis, also operate in the survival/proliferation of BL.

Fig. 1

Low PI3K-AKT activity in BL in comparison with GCB-DLBCL cell lines and tonsillar GC DZ B cells. a, b Expression and phosphorylation status AKT in BL and GCB-DLBCL cell lines was analysed by immunoblot. TUBB served as loading control. A representative of two independent experiments is shown. a Short exposure. b Long exposure. c Immunoblots of pAKT^T308 and total AKT in three FACS-sorted tonsillar GC DZ B cells as well as two BL cell lines with lowest and highest pAKT^T308 expression levels. d Expression and phosphorylation status of AKT in FACS-sorted tonsillar GC DZ B cells (sample #2 as shown in c) compared with BL and GCB-DLBCL cell lines with lowest and highest pAKT^T308 levels was analysed by immunoblot

Fig. 2

Physiological PI3K-AKT activation in BL. a Immunohistochemical detection of pAKT^T308. Normal tonsillar GC B cells (GC) are weakly positive (+) while follicular mantle B cells (MZ) are strongly positive (+++). In BL pAKT^T308 was weakly positive in most cases. The majority of GCB-DLBCL was clearly positive (++). Left column: Overview. Scale bar represents 200 µm. Right column: Details. Scale bar represents 100 µm. b Table gives details of all cases examined

Fig. 3

BL cell lines are less sensitive to AKT1 knockdown than pAKT^high GCB-DLBCL cell lines. BL and GCB-DLBCL cell lines were transduced with vectors expressing AKT1 shRNA (A1sh) vs. scrambled (scr) control. a Knockdown efficiencies of AKT1sh vs. scr control. Transduced cells were FACS-sorted 5 days post transduction and total AKT expression was analysed. AKT expression was quantified using ImageJ software. A representative of two independent experiments is shown. b Cells were FACS-sorted 4 days post transduction, followed by cell counting using a cell viability analyser. Initial number of cells was set as 1. Data are shown as mean ± SD (N = 3). c Cell death analysis of cells expressing A1sh or scr. Transduced cells were sorted 4 days post transduction, and incubated in complete medium for 6 days followed by Annexin V-FITC/PI staining. Specific Apoptosis (SA) was calculated as SA (%) = 100 × (AE − AC)/(100 − AC), where AE equals the percentage of apoptotic cells in the experimental group and AC equals the percentage of apoptotic cells in the control group. Data are shown as mean ± SD (N = 3). The data were analysed by two-sided T-test. ****P < 0.0001. d Representative dot-plot images as shown in c

Fig. 4

Low sensitivity of BL cell lines to AKT inhibitors. a Sensitivity of BL and GCB-DLBCL cell lines to AZD5363 (AZD). Cells were exposed to various concentrations of AZD for 5 days. Cell viability was assessed using MTT assay and IC50 values were calculated using GraphPad Prism software. Data are shown as mean ± SD (N = 3). b Percentage of viable BL or GCB-DLBCL cells 5 days post treatment with AZD5363. Cells were counted using a cell viability analyser. Initial number of cells was set as 100%. Data are shown as mean ± SD (N = 3). c Sensitivity of BLs with reported t(8;14) MYC translocations, not otherwise specified GCB-DLBCL, and AKT^high GCB-DLBCL cell lines to MK-2206 or GSK690693. IC50 values were obtained from cancerrxgene.org and analysed using Mann–Whitney U test with help of Social Science Statistics Calculator (socscistatistics.com). MK-2206: BL vs. total GCB-DLBCL **p = 0.004. BL vs. pAKT^high GCB-DLBCL **p = 0.001. GSK690693: BL vs. total GCB-DLBCL ns, p = 0.060. BL vs. pAKT^high GCB-DLBCL **p = 0.009. Maximum screening concentration: MK-2206 = 4 μM; GSK690693 = 10.2 μM

Fig. 5

myrAKT overexpression is toxic for BL cell lines. a–f BL cell lines were transduced with lentiviral plasmids expressing constitutively active myrAKT or empty vector (EV). a Transduced cells were FACS-sorted 4 days post transduction and analysed for expression of total AKT and FOXO1 and pAKT^T308 and pFOXO1^T24 levels. TUBB served as loading control. A representative of two independent experiments is shown. b Percentage of transduced cells was measured every 3 days using flow cytometry. First measurement was performed 4 days post transduction and the percentage of GFP⁺ cells was set as 100. Data are shown as mean ± SD (N ≥ 3). c myrAKT downregulates CXCR4. BL cell lines expressing myrAKT or EV were stained with CXCR4-APC or isotype control 5–7 days post transduction. Dot-plots show percentages of CXCR4⁺/GFP⁺ and CXCR4⁻/GFP⁺ cells. For histograms, only transduced GFP⁺ cells were included. Data are shown as mean ± SD (N = 3). d Downregulation of critical GC genes after myrAKT overexpression. Transduced cells were sorted 3 days post transduction and RNA expression levels were analysed by qRT-PCR. qRT-PCR data were quantified by the 2^−ΔΔCT method. Data are shown as mean ± SD (N = 3). The data were analysed by two-sided T-test. For all genes and cell lines p < 0.05 with exception of BL-41 RAD51AP1 p = 0.135. e MYB expression in sorted GFP⁺ cells was analysed by immunoblot 3 days post transduction with myrAKT or EV. TUBB served as loading control. A representative of two independent experiments is shown. f myrAKT activates NF-κB. BL cell lines expressing myrAKT or EV were FACS-sorted 3 days post transduction and pRELA^S536 and total RELA were analysed by immunoblot. TUBB served as loading control. A representative of two independent experiments is shown. g Luciferase reporter assay. Namalwa cells stably expressing a NF-κB-dependent luciferase reporter (3 × κB.luc) [60] were transduced with a vector expressing myrAKT. GFP⁺ cells were sorted 4 days post transduction. Luminescence was measured as described in Supplementary Methods. Data are shown as mean ± SD (N = 3)

Fig. 6

PTEN expression is essential for BL cell lines. a Expression of PTEN in BL was analysed by immunoblot. GCB-DLBCL cell lines were included as controls. A representative of two independent experiments is shown. b–h BL cell lines and/or GCB-DLBCL cell lines were transduced with vectors expressing PTEN shRNAs (PTsh#1, PTsh#2) or scr control. b Knockdown efficiencies of PTsh vs. scr control in BL cell lines and PTEN-positive GCB-DLBCL cell line OCI-Ly19. Transduced cells were selected using 4 µg/mL puromycin or FACS-sorted and PTEN expression was analysed. A representative of two or three independent experiments is shown. c Growth dynamics of transduced BL and GCB-DLBCL cell lines. The percentage of RFP⁺ cells was measured every 3 days using flow cytometry. First measurement was performed 4 days post transduction and the percentage of RFP⁺ cells was set as 100. Data are shown as mean ± SD (N ≥ 3). d Transduced cells were FACS-sorted 4–5 days post transduction and PTEN and pAKT^T308 levels were analysed using immunoblot. A representative of two independent experiments is shown. e MYB expression levels in transduced and FACS-sorted BL cell lines expressing PTsh#2 or scr control were analysed by immunoblot 3 days post transduction. A representative of two independent experiments is shown. f BL cell lines expressing PTsh#2 or scr control were FACS-sorted 4 days post transduction and pRELA^S536 and total RELA were analysed by immunoblot. TUBB served as loading control. A representative of two independent experiments is shown. g PTsh#2 downregulates CXCR4. BL cell lines expressing PTsh#2 or scr control were stained with CXCR4-APC or isotype control 4–7 days post transduction. Dot-plots show percentages of CXCR4⁺/RFP⁺ and CXCR4⁻/RFP⁺ cells. For histograms, only transduced RFP⁺ cells were included. Data are shown as mean ± SD (N = 3). h AZD5363 (AZD) treatment partially rescues BL cell lines from the growth inhibitory effect of PTEN knockdown. BL cell lines were treated with 1 µM AZD 4–5 days post transduction with PTsh#2 and the percentage of RFP⁺ cells was measured every 3 days using flow cytometry. The percentage of RFP⁺ cells at first measurement was set as 100. Data are shown as mean ± SD (N = 3)

Fig. 7

PTEN overexpression induces growth inhibition in GCB-DLBCL but not BL cell lines. a–c BL and GCB-DLBCL cell lines were transduced with vectors expressing PTEN vs. empty vector (EV) control. a Transduction efficiencies were ranging from 11% to 100%. Transduced cells were lysed 4–5 days post transduction and PTEN overexpression was confirmed by immunoblot. GAPDH served as loading control. A representative of two independent experiments is shown. b Transduced cells were FACS sorted 4 days post transduction and analysed for expression of PTEN, FOXO1, pAKT^T308, and pFOXO^T24. TUBB served as loading control. A representative of two independent experiments is shown. c Growth dynamics of transduced BL and GCB-DLBCL cell lines. The percentage of GFP⁺ cells was measured every 3 days using flow cytometry. First measurement was performed 4–5 days post transduction and the percentage of GFP⁺ cells was set as 100. Data are shown as mean ± SD (N = 3). d Illustration of the maintenance of physiological mechanisms of PI3K-AKT regulation in BL. BLs maintain the original PI3K-AKT and IKK-NF-κB activation status to prevent extinguishing of the DZ programme, including CXCR4 and MYB. Part of the image is adapted from Motifolio Drawing Toolkits (www.motifolio.com)