Low Density Lipoproteins Amplify Cytokine-signaling in Chronic Lymphocytic Leukemia Cells

Abstract

Recent studies suggest there is a high incidence of elevated low-density lipoprotein (LDL) levels in Chronic Lymphocytic Leukemia (CLL) patients and a survival benefit from cholesterol-lowering statin drugs. The mechanisms of these observations and the kinds of patients they apply to are unclear. Using an in vitro model of the pseudofollicles where CLL cells originate, LDLs were found to increase plasma membrane cholesterol, signaling molecules such as tyrosine-phosphorylated STAT3, and activated CLL cell numbers. The signaling effects of LDLs were not seen in normal lymphocytes or glycolytic lymphoma cell-lines but were restored by transduction with the nuclear receptor PPARδ, which mediates metabolic activity in CLL cells. Breakdown of LDLs in lysosomes was required for the amplification effect, which correlated with down-regulation of HMGCR expression and long lymphocyte doubling times (LDTs) of 53.6±10.4months. Cholesterol content of circulating CLL cells correlated directly with blood LDL levels in a subgroup of patients. These observations suggest LDLs may enhance proliferative responses of CLL cells to inflammatory signals. Prospective clinical trials are needed to confirm the therapeutic potential of lowering LDL concentrations in CLL, particularly in patients with indolent disease in the "watch-and-wait" phase of management.

Keywords: Cholesterol; Chronic lymphocytic leukemia; HMGCR; Janus kinases; Lipoproteins; Lysosomal lipase; Nuclear receptors; Ruxolitinib; STAT3.

Graphical abstract

Fig. 1

Effect of LDLs on growth, cellular lipids, and phosphorylated STAT3 levels in activated CLL cells in vitro. A. Purified CLL cells (1 × 10⁶ cells/ml) were stimulated with IL-2 and Resiquimod (2S) in lipid-poor media in the indicated concentrations of LDLs. Viable cells were counted manually by Tryphan Blue exclusion after 4 days. The average and standard deviation of the results from 6 different patient samples are shown. B,C. Mean fluorescence intensities (MFIs) of Nile Red staining (n = 19) to indicate lipid droplets (B) and PFO staining (n = 4) to indicate plasma membrane cholesterol levels (C) were measured by flow cytometry after 18 h and 24 h, respectively, in the presence or absence of 0.5 mM LDLs. Averages and standard deviations are shown. D. Protein extracts were made after 18 h and phospho-STAT3 expression measured by immunoblotting and quantified by densitometry with β-actin (right panel) or total STAT3 (left panel) used as loading controls. Averages and standard deviations are indicated and representative examples of immunoblots are shown below the graphs. *p < 0.05; **p < 0.01.

Fig. 2

Effect of LDLs on normal lymphocytes and leukemic cell-lines. A–C. Purified normal PBMCs (1 × 10⁶ cells/ml) were stimulated with IL-2 and Resiquimod (2S) and cultured for 18 h with or without LDL (0.5 mM). Nile Red- (A) (n = 4) and PFO (B)-staining (n = 5) were measured by flow cytometry. C. Changes in pSTAT3 levels were determined by immunoblotting with total STAT3 protein levels used as a loading control (n = 5). D. Oncomine analysis of the Haferlach data base comparing 74 normal PBMC and 448 CLL samples indicates significantly lower SORL1 expression in CLL cells compared to PBMCs and other hematologic malignancies. E–F. Vector control and PPARD^hi Daudi were cultured in RPMI-1640 ± LDL (0.5 mM). Cells were counted after 4 days (E) and pSTAT3 levels measured after 18 h (F). G. Cholesterol levels in control and PPARD^hi Daudi cells in exponential growth in RPMI + 5% FCS were measured by Amplex red staining. *p < 0.05; **p < 0.01; ns = not significant.

Fig. 3

Effect of LDLs on cytokine-signaling in activated CLL cells in vitro. Purified CLL cells were cultured in lipid-poor conditions with or without IL-2 and Resiquimod and with or without addition of LDLs (0.5 mM) in the presence or absence of IL-6 antibodies (10 ng/ml), IL-10 antibodies (10 ng/ml) or Ruxolitinib (Rux) (500 nm). A. After 18 h, phospho-STAT3 levels were determined in 4 patient samples by immunoblotting and densitometry and normalized to the results for β-actin. The averages and standard deviations of the relative densitometric values are indicated in the graph and a representative immunoblot is shown. B, C, D. After 48 h, IL10 levels in the culture supernatants were measured by ELISAs (n = 5) (B), IL10 mRNA was measured by quantitative RT-PCR for Pt.117 (C), and mean fluorescence intensities of IL10-receptor (IL10R)-staining were determined by flow cytometry for 5 patient samples (D). Averages and standard deviations are shown. *p < 0.05; **p < 0.01; ns, non significant.

Fig. 4

Effect of LDL components on STAT3-phosphorylation in activated CLL cells. A. CLL cells were activated with IL-2 and Resiqimod (labeled “2S”) in the presence and absence of LDLs (0.5 mM) and/or Lalistat (LAL) (1 μM). Densitometric values of p-STAT3 normalized to β-actin expression were determined after 18 h (left panel) (n = 5), cell counts were measured after 4 days by trypan blue exclusion (middle panel) (n = 3), and mean fluorescence intensities of PFO staining to indicate plasma membrane cholesterol levels were measured by flow cytometry after 24 h (right panel) (n = 4). B. 2S-activated CLL cells from 3 different patients were cultured with or without LDL (0.5 mM), oleic acid (OA) (5 μm), linoleic acid (LA) (5 μm), heptanoic acid (HA) (5 μm), propanoic acid (PrA) (5 μm), or octanoic acid (OcA) (5 μm) and normalized phospho-STAT3 levels determined after 18 h. C. 2S-activated CLL cells from 20 different patients were cultured with or without cholesterol (15 μm) and/or methyl-β-cyclodextrin (Cycd) (0.5 mM). Expression of pSTAT3 normalized to β-actin (left panel) was determined for each sample after 18 h and Nile Red-staining measured by flow cytometry to confirm lipid-loading and stripping (right panel). Averages and standard deviations are shown in each graph. *p < 0.05; **p < 0.01; ns, non significant.

Fig. 5

Effect of LDL components on oxidative stress and signaling in CLL cells. A. CLL cells were activated with IL2 and resiquimod (2S) and with or without LDL (0.5 mM) or Vitamin E (5 μM). Mean fluorescence intensities (MFIs) of DCFH staining were determined by flow cytometry after 12 h (left and middle panels) and densitometric values of p-STAT3 expression relative to β-actin were determined after 18 h (right panel). Shown are the average results and standard errors from the number of patient samples indicated in each graph. An example of an immunoblot is shown in the insert. B. Purified CLL cells (1 × 10⁶ cells/ml) were cultured over-night in fatty acid free media (FAF) or 1% CD Lipid extract, consisting of a mixture of free cholesterol, fatty acids, and Vitamin E. The cells were then treated with IL10 (10 ng/ml) but not otherwise activated with IL2 and resiquimod. Levels of pSTAT3 were measured by immunoblotting after 15 min, 30 min, and 1 h using STAT3 as the loading control. Averages and standard deviations of relative pSTAT3 densitometry values for 6 different patient samples are plotted with a representative immunoblot shown below the graph. C. Relative pSTAT3 values were measured in CLL cells from 6 different patients that had been activated with IL2 and resiquimod for 18 h in fatty acid free media (FAF) or 1% CD Lipid extract with or without Lalistat (LAL) (1 μM). Averages and standard deviations are shown. D. CLL cells were cultured with or without LDL (0.5 mM) for 18 h and stimulated with anti-IgM antibodies (10 ng/ml) (n = 3; left panel) or phorbol dibutyrate (5 ng/ml) (n = 3; right panel). Phospho-AKT (pAKT) levels were determined after 30 min by immunoblotting with β-actin as a loading control. *p < 0.05; **p < 0.01; ns, non significant.

Fig. 6

Effect of LDLs on STAT3-phosphorylation and HMGCR mRNA expression in vitro and cholesterol content of CLL cells in vivo. A. CLL cells from individual patients were activated in lipid-poor media with IL2 and resiquimod with or without 0.5 mM LDLs. After 18 h, pSTAT3 levels were quantified by densitometry and normalized to the results for β-actin. HMGCR transcripts were measured by quantitative RT-PCR. The differences between the LDL-treated cells compared to the untreated cells (ΔpSTAT3 and ΔHMGCR) were calculated and their relationship explored with a scatter plot. Note that the negative value of ΔHMGCR is shown on the y-axis to reflect LDL-uptake. B. A cutoff of ΔpSTAT3 = 0.6 was used to empirically classify patients into group 1 (ΔpSTAT3 ≥ 0.6; n = 18) and group 2 (ΔpSTAT3 < 0.6; n = 11). Highly significant differences were seen in ΔHMGCR, white blood cell (WBC) numbers, and lymphocyte doubling times (LDTs) between the two groups. C, D. Cholesterol content of CLL cells from 30 other patients and normal B cells from 8 healthy donors were measured by Amplex-red fluorescence (C). LDL concentrations at the time of cell collection were taken from the patients' electronic medical records (D). The results are shown as 4 groups: normals, patients on statins, and CLL cholesterol concentrations higher or lower than 20 fluorescence units. Average results for each group were used for statistical analysis. E. Protein extracts were frozen immediately from CLL cells purified from the blood of a patient before and 3.5 months after starting Simavastin (20 mg daily). LDL concentrations pre- and post-statin therapy were 2.82 and 1.76 mM. Levels of pSTAT3 in the thawed extracts were determined at the same time using STAT3 as the loading control. The immunoblot is shown on the right with densitometry readings on the left. **p < 0.01; *p < 0.05.