STAT3 and NF-κB cooperatively control in vitro spontaneous apoptosis and poor chemo-responsiveness in patients with chronic lymphocytic leukemia

Abstract

Chronic lymphocytic leukemia (CLL) is an adult disease characterized by in vivo accumulation of mature CD5/CD19/CD23 triple positive B cells and is currently incurable. CLL cells undergo spontaneous apoptosis in response to in vitro cell culture condition but the underlying mechanism is unclear. We hypothesize that the sensitivity of CLL cells to spontaneous apoptosis may be associated with the constitutive activities of transcription factors STAT3 and/or NF-κB. We now show that the sensitivity of fresh CLL cells to spontaneous apoptosis is highly variable among different patients during 48 hours' cell culture and inversely correlated with in vivo constitutively activated STAT3 and NF-κB (p < 0.001). Both activated STAT3 and NF-κB maintain the levels of anti-apoptotic protein Mcl-1/Bcl-xL and autocrine IL-6 production. CLL cells with higher susceptibility to in vitro spontaneous apoptosis show the greatest chemosensitivity (p < 0.001), which is reflected clinically as achieving a complete response (CR) (p < 0.001), longer lymphocyte doubling times (p < 0.01), time to first treatment (p < 0.01), and progression free survival (p < 0.05). Our data suggest that the sensitivity of CLL cells to in vitro spontaneous apoptosis is co-regulated by constitutively activated STAT3 and NF-κB and reflects the in vivo chemo-responsiveness and clinical outcomes.

Keywords: NF-kB; STAT3; chronic lymphocytic leukemia; prognosis; spontaneous apoptosis.

Figure 1. In vitro spontaneous apoptosis of fresh CLL cells

Freshly isolated CLL cells were incubated in complete culture medium for 48 hours. A. Apoptotic cell death was measured by flow cytometry at 0 hour (without incubation) and after in vitro incubation for 48 hours in 2 CLL cases. B. Apoptotic cell death in 51 cases was determined by flow cytometry (mean ± SD). C. The level of spontaneous apoptosis in different clinical prognostic biomarker groups: Binet stage at presentation (A vs B/C, n = 51), treated vs untreated (n = 51), CD38 (negative or positive; cut-off 20%, n = 47), cytogenetic (absence or presence of 11q and 17p deletions, n = 48), ZAP-70 (negative or positive; cut-off 20%, n = 26) and IgHV mutated vs unmutated (n = 16) cases. D. CLL cells were incubated with or without 20μM Z-VAD.fmk for 24, 48 and 72 hours, then spontaneous apoptosis was analyzed by flow cytometry (n = 6). ** p < 0.01, ***p < 0.001.

Figure 2. In vivo and in vitro cytokine production by CLL cells

5×10⁵/ml of fresh CLL cells (CD19+/CD5+ more than 95%) were cultured in RPMI-1640 complete medium for 24 hours. The conditioned medium and corresponding plasma samples were incubated with cytokine beads and the levels of cytokine production were measured by flow cytometry: A. IL-2, B. IL-4, C. IL-6, D. IL-10, E. TNF-α and F. VGEF concentrations (n = 16). G. CLL cells were incubated with 10ng/ml of IL-2, 10ng/ml of IL-4, 10ng/ml of IL-6, 10ng/ml of IL-10, 10ng/ml of TNF-α or 10ng/ml of VGEF for 48 hours. Decreased percentages of apoptotic cell death mediated by cytokines were assessed by flow cytometry and compared with the untreated controls (n = 5). * p < 0.05, ** p < 0.01, ***p < 0.001.

Figure 3. Co-regulation of STAT3 and NF-κB on autocrine IL-6 production

A. and B. Prediction of the interaction between STAT3 and NF-κB by STRING v10 database. A. Interaction network of STAT3 and NF-κB. The interaction network is generated by manual input of multiple human transcriptional factors. The thickness of the connection line reflects the binding score of two proteins. B. Combined association scores, including physical and functional interaction. C. D. Fresh CLL cells were incubated with 10μM of Stattic, 10μM of 5,15-DDP, 10μM of CAPE or 10μM of JSH-23 for 24 hours in complete PRMI-1640 medium respectively. C. Autocrine IL-6 production (n = 9); D. IL-2, IL-4, IL-10, TNFα or VEGF (n = 4) production in the conditioned medium was determined by human CBA Flex kit and analyzed by flow cytometry.

Figure 4. Regulation of STAT3 and RelA on Bcl-xL/Mcl-1 expression and spontaneous apoptosis

Fresh CLL cells were incubated with 10μM of Stattic, 10μM of 5,15-DDP, 10μM of CAPE or 10μM of JSH-23 for 24 hours respectively. (A and B) Expression of Bcl-xL, Mcl-1, Bcl-2, and Bax was determined by Western blotting. C. Spontaneous apoptotic cell death for 48 hours incubation was detected by flow cytometry (n = 5). D. Down-regulation of STAT3 and RelA by siRNA. Cells were transfected with STAT3-siRNA, RelA-siRNA or control siRNA. Expression of Mcl-1, Bcl-xL, STAT3, and RelA were determined by Western blotting. E. Cells were transfected with STAT3-siRNA, RelA-siRNA, Mcl-1-siRNA, Bcl-xL-siRNA or control siRNA and spontaneous apoptosis was determined by flow cytometry after transfection for 24 hours(n = 5).

Figure 5. Correlation of constitutive activities of STAT3 and RelA with in vitro IL-6 production and spontaneous apoptosis

A. Constitutive activities of STAT3 and RelA. Expression of p-STAT3 and p-RelA was determined on 9 fresh CLL samples by Western Blotting. B. The association of expression of p-STAT3/p-RelA with IL-6 production and spontaneous cell death from 10 CLL cases (CLL ID) was expressed by a heat map. The levels of p-STAT3 and p-RelA expression ranged from 15-140 arbitrary units (AU) (analyzed by densitometry using GelScan v 5.1 software); autocrine IL-6 ranged from 0-40pg/ml and cell death ranged from 0-50%. (C, D, E) Correlation between p-STAT3 and p-RelA C. the correlation of IL-6 production to p-RelA D. or p-STAT3 E. was analyzed in 20 CLL cases. Correlation between spontaneous cell death and p-STAT3 F. or p-RelA G. was analyzed in 14 CLL cases. H. Correlation between IL-6 production and spontaneous cell death was analyzed in 38 CLL cases.

Figure 6. Correlation between constitutively activated STAT3 and RelA with chemosensitivity

5×10⁶/ml of fresh CLL cells were treated with 20μg/ml of chlorambucil, cell death was determined by flow cytometry after 24 hours. A. and B. Correlation between chlorambucil-induced cell death and constitutive p-STAT3 (n = 15) or p-RelA (n = 15) expression. C. Correlation between spontaneous apoptosis and chlorambucil-induced cell death was analyzed in 38 CLL cases. D. Schematic illustration of the mechanisms by which STAT3/RelA co-regulate Bcl-xL and Mcl-1 expression, IL-6 production, and the chemoresistance.

Figure 7. Spontaneous apoptosis and clinical prognosis

Differential sensitivities to spontaneous apoptosis were analyzed. A. lymphocyte doubling time (LDT) more than 12 months and less than 12 months CLL cases; B. CLL cells from patients achieving a complete remission (CR, p < 0.01, n = 10) or non-CR (i.e., a partial response, stable disease and progressive disease; n = 19, p < 0.001). C. The significant difference between the CLL cells that were sensitive (i.e., more than average, n = 11) and those were resistant (less than average, n = 18) to spontaneous cell death in the time to first treatment (TTFT). D. The significant difference between the CLL cells that were sensitive (i.e., more than average, n = 23) and those that were resistant (less than average, n = 28) to spontaneous cell death in the progression free survival (PFS).