Page 138 - Haematologica June

1040
S. Grgurevic et al.
MEC-1 and MEC-2, respectively. MEF cell lines were grown in Dulbecco modified Eagle medium with glutamax supplemented with 10% fetal bovine serum and 1% penicillin/streptomycin at a density of 1.5 x 105 cells/10 cm dish. Primary peripheral blood mononuclear cells from CLL patients were cultured in Roswell Park Memorial Institute medium supplemented with 1% gluta- mine, 10% fetal bovine serum and 1% penicillin/streptomycin at a density of 1 x 107 cells/mL allowing long-term viability.
DNA synthesis
DNA synthesis was monitored by a Click-iT EdU imaging assay (Invitrogen, Saint-Aubin, France). 5-ethynyl-2’-deoxyuridine (EdU) was added at the concentration of 25 mM for 30 minutes (min) as indicated. During DNA staining, cells were incubated with propidium iodide at a concentration of 50 mg/mL, 0.1% RNase A and 0.1% Triton X-100 in a phosphate-buffered saline solution for 10 min away from light.
DNA combing in nanochannels
Cells were incorporated into agarose plugs (200,000 cells/plug) and each experiment was carried out with half of an agarose plug deposited in a 1.5 mL tube filled with 800 mL of 0.5X TBE buffer supplemented with 2% poly-vinylpyrrolidone (40 kDa, below the overlapping concentration of 7%) and 5% dithiothreitol. Agarose was melted by heating the tube at 70°C for 15 min. The tempera- ture was set at 42°C, and 2 mL of β-agarase (New England Biolabs) were added for overnight agarose digestion. Finally, DNA was stained with 2.5 mM fluorophore SYTOX-orange (Molecular Probes). Nanofluidic chips were fabricated using a two-step pho- tolithography process to generate the array of nanochannels by projection lithography (ECI 1.2 mm photoresist, Stepper Canon 3000i4); these were then transferred into silicon by Reactive Ion Etching over a depth of 250 nm. Conventional photolithography was then performed to etch lateral microchannels of 20 mm in width and 7 mm in depth. For optical mapping, we used hydrody- namics to force the uptake of genomic DNA in nanochannels. Imaging was performed with a wide-field inverted Zeiss micro- scope equipped with a 40X lens (NA=1.4). The light source was a LED engine (Lumencor) with 542/33 nm emission with the filter sets Cy3-4040C (Semrock) for SYTOX-orange visualization. Images were collected with an Andor Zyla camera operating with a 2x2 binning (pixel size = 325 nm). The velocity of molecules was set at 200 ± 20 mm/s using a pressure source operating at 80-100 mbar. All images presented in the manuscript were filtered using the FFT bandpass filter implemented in ImageJ using minimum and maximum cut-offs of 3 and 40 pixels, then subtracting the background.
Statistical analysis
The patients’ clinical data are summarized by frequency and percentage for categorical variables and by median and range for continuous variables. Time to progression was defined as the peri- od from the first-line therapy to progression or last follow up (cen- sored data) and estimated by the Kaplan-Meier method with 95% confidence intervals. The minimum P-value approach was used to dichotomize POLN expression,17 which selects the threshold that best discriminates patients’ outcomes. Selected values of the prog- nostic factor are examined as candidates for the threshold, after eliminating the top and bottom 10% of the extreme values. The value that best separates patients’ outcomes according to a mini- mum P-value obtained by the log-rank test is chosen. The P-value is adjusted using the Altman correction to account for the problem of multiple testing. Stability of the threshold was assessed using bootstrap internal validation. Backward selection was performed with a Cox proportional hazards model to identify clinical factors
associated with time to progression. Multivariate analysis was also performed using a Cox proportional hazards model to study the influence of POLN on time to progression after adjusting for the clinical factors previously identified. Two-sided P-values of less than 0.05 were considered statistically significant. All statisti- cal analyses were performed using STATA 12.0 software. Viability data obtained by the MTS assay were analyzed in GraphPad Prism using an ANOVA multiple test. The length of replicated domains assessed in the DNA combing in nanochannels experi- ments was analyzed in GraphPad Prism using the Mann-Whitney rank-sum t-test. FACS data were analyzed in GraphPad Prism using a paired or unpaired Student t-test depending on the experi- mental conditions.
Results
POLN determines time to progression in chronic lymphocytic leukemia
In order to investigate the implication of 3R genes in the clinical course of CLL, we first performed a large, high- throughput quantitative reverse transcriptase polymerase chain reaction-based gene expression analysis on 99 pri- mary samples obtained from treatment-naïve patients included in the CLL 2007 FMP clinical trial (see above). The patient’s baseline characteristics are summarized in Table 1. Unsupervised clustering on 3R gene expression values confirmed, as previously published,18 clear underly- ing differences between samples from healthy donors and CLL lymphocytes (Figure 1A) and showed a significantly higher expression in CLL of POLN, the gene encoding for the specialized DNA polymerase ν (Figure 1B), as com- pared to most of the 3R genes, including other specialized DNA polymerases or genes involved in double-stranded break repair (Online Supplementary Figure S1). Next, we performed a cross-analysis on the gene expression infor- mation obtained and the patients’ clinical data. The results of this analysis reveal that among all 3R genes, POLN was the only one linked to the therapeutic outcome of CLL. In univariate analysis, a high level of POLN expression was associated with a shorter time to relapse after first-line therapy, i.e. time to progression (threshold=11.9x10-2, sta- bility=47.42%; P=0.0009, adjusted P=0.0227), as shown in (Figure 1C). Interestingly, we found that POLN expression maintained a significant correlation with time to progres- sion in multivariate analysis after adjustments using previ- ously identified prognostic factors (Table 2). The adjusted Hazard Ratio (HR) was 4.14 with a 95% Confidence Interval of 1.60-10.72 and a P-value of 0.003, defining POLN expression as the strongest prognostic marker of time to progression after fludarabine-based treatment, independently of Binet stage or IGHV mutational status.
Proliferating chronic lymphocytic leukemia lymphocytes over-express POLN
Leukemic lymphocytes circulating in the CLL patients’ peripheral blood are arrested in the G0 and G1 phases of the cell cycle.5 To proliferate, CLL lymphocytes need to receive proliferative and pro-survival stimuli released by the accessory cells residing in the lymph node pseudofol- licules.4 In order to investigate CLL replication parameters, we mimicked the proliferative lymph node microenviron- ment by stimulating primary CLL lymphocytes with inter- leukin-2 and DSP30. After obtaining proliferating leukemic cells, as confirmed by the CFSE dilution assay
haematologica | 2018; 103(6)