Cell-free miRNAs may indicate diagnosis and docetaxel sensitivity of tumor cells in malignant effusions

Our study was approved by the Ethics Review Board of Nanjing University, Nanjing, China. Informed consent was obtained from all patients. All effusion samples were recruited from the Affiliated Drum Tower Hospital, Medical School of Nanjing University, from Jan 2006 to Dec 2007. Malignant effusions were confirmed by pathological diagnosis. Samples were collected before chemotherapy. Effusions were diagnosed as benign based on the clinical context and the absence of malignant cells in 3 separate samples from the same patients. No cancers occurred in patients with benign effusions over the next 6 months. All samples were transported to the laboratory within 30 min after their collection. The information for patients is presented in Table 1.

Primary tumor cells were prepared by Ficoll-Hypaque (specific gravity 1.077, Pharmacia) density centrifugation. The number and viability of cells were determined by cytological examination with trypan blue dye. The test drug concentration of docetaxel (DOC, donated by Jiangsu Hengrui Medicine Company, purity > 99.9%) was 12.4 μM, derived from pharmacokinetic data and adjusted for protein-binding to approximate the concentration achievable in the patient plasma. Cells were seeded in U-shape-96-well plates at 5× 10⁴ cells/well and treated with drugs in triplicate for 72 h. WST-8 reagent (10 μl each) from Cell Counting Kit-8 (Dojindo, Kumamoto, Japan) was added and absorbance was determined with a multiwell spectrophotometer (BioTek, VT, USA) at 460 nm. Inhibition rate = (1-absorbance of treated cells/control cells) ×100%. Only the specimens containing at least 50% cancer cells were enrolled for chemosensitivity test. And we have proved the increase of tumor cells number by the culture in our preliminary test. Serum free medium and polypropylene U-shape-96-well microplates can enrich tumor cells up to 80-90% [11].

Samples for miRNA quantification were centrifuged at 8000 g for 20 min and the supernatants were collected. After adding 750 μl Trizol LS (Invitrogen, CA, USA) into 250 μl effusion samples, another 25 μl ath-miR156a (20 nM, synthesized by Takara, Japan) was supplemented into each sample tube. Total RNA was extracted using Trizol LS reagent (Invitrogen, CA, USA) according to the manufacture's instruction. The miRNA expression levels were quantified by real-time PCR using miRNA detection kit (Applied Biosystems, Foster city, CA, USA). Briefly, miRNAs were reverse-transcribed to cDNA with the AMV reverse transcriptase (Takara, Japan). Subsequently, real-time PCR was performed on Mx3000P cycler (Stratagene, USA). All reactions were run in triplicate. Ct data were determined using default threshold settings. The relative expression levels of miRNAs were calculated as 2^-ΔΔCt [12], using ath-miR156a as reference and a mixture of RNA extracted from human lung cancer A549 cells and gastric cancer BGC-823 cells (1:1) as the calibrator.

We firstly tested the stability of cell-free miRNAs in the effusion samples kept at room temperature for different time periods, or treated with multiple freezing and thawing, and digestion with RNase or DNase. β-actin mRNA transcripts and 18S RNA representing large molecular RNAs were analyzed in this study. qRT-PCR was carried out to quantify the RNAs of three samples with various treatments. Multiple freeze-thaw cycles had hardly any effect on cell-free RNAs in effusion samples (Figure 1A). When effusion samples were left at room temperature for 3 h or 24 h, cell-free RNAs began to degrade at 3 h. β-actin mRNA degraded quickly and significantly, while 18S RNA and the two miRNAs were more stable, with more than 20% remaining after 24 h (Figure 1B). When treated with RNase, miRNAs were more stable than large RNA molecules although significant deposition also occurred in miRNAs. All cell-free RNAs were resistant to DNase digestion (Figure 1C).

Bild vergrößern

In order to evaluate the miRNAs expression in the supernatant of effusion samples, we selected 22 miRNAs from the list of deregulated cell-free miRNAs in the serum of lung cancer patients published previously [8]. qRT-PCR was used to examine the possible expression levels in two samples, which were pooled from 18 malignant and 12 benign effusions (equivalent amounts of RNA were added to each sample). Ct values of each miRNAs were obtained and assessed. miRNAs with Ct >35 were discarded, resulting in 13 miRNAs (Table 2).

Next, we measured the expression levels of 13 miRNAs in 29 samples (benign n = 11; malignant n = 18). The expression levels of the 13 miRNAs were compared between malignant and benign groups using a Mann-Whitney U test. miR-24, miR-26a and miR-30d were expressed significantly differently between groups (P = 0.006, 0.021 and 0.011, respectively). All three miRNAs were significantly over-expressed in the supernatant of malignant effusions (Figure 2). However, no significant difference was identified after Bonferroni correction for the three miRNAs differently expressed between malignant and benign groups.

Bild vergrößern

The WST-8 test has been confirmed to be a reliable chemosensitivity test for primary tumor cells of malignant effusions [13]. In the current study, the chemosensitivity of primary tumor cells isolated from fifteen malignant effusion samples were examined. The mean inhibition rate was 0.49 (range: 0.2 to 0.86, n = 15). Malignant effusion samples with an inhibition rate < mean were considered docetaxel resistant, and those with an inhibition rate ≥ mean were considered as docetaxel sensitive. The docetaxel sensitive group has a higher level of cell-free miR-152 in the effusion supernatant (P = 0.005, Figure 3A). Spearman correlation coefficient was also used to compare the miRNAs expression levels and sensitivity. The expression levels of miR-152 correlated with the inhibition rates of docetaxel (r = 0.60, P = 0.016, Figure 3B). miR-320 expression levels were also associated with docetaxel inhibition rates (r = 0.52, P = 0.042), but no significant difference on miR-320 expression levels was observed between docetaxel resistant and sensitive groups (P = 0.087).

Bild vergrößern