Identification Of Diffuse Large b-Cell Lymphoma (DLBCL) Subtypes Defined By The Expression Of Stromal Proteins and Their Association With Proangiomirs and Antiangiomirs

1) to classify DLBCL cases according to the signature stromal-1 and stromal-2 using tissue microarray (TMA) and immunohistochemistry; 2) to evaluate the expression of pro- and antiangiomiRs in paraffin embedded tissues affected by DLBCL and correlate them with the signatures of the tumor stroma.

111 DLBCL (NOS and variants, all HIV negative) admitted to the Hospital São Paulo between 2000 and 2010 were included in this study. Expression of CD68 (stromal-1 marker) was classified into four (0-3) levels. To analyze the expression of CD34 (MVD, stromal-2), we conducted a manual count of microvessels in the entire field of TMA. We performed miRNA extraction from paraffin samples using RecoverAll ™ Total Nucleic Acid Isolation Kit for FFPE Tissues (Applied Biosystems). Real-time quantitative PCR was performed using TaqMan® Small RNA kit Assays and normalized with U18 and RNU44. Pro-angiomirs Let-7f, miR-92a, miR-130a, miR-210, miR-296 and miR-378 and anti-angiomiRs miR-16, miR-20b, miR-221 and miR-328 were selected for analyses and were considered differentially expressed when tumor samples levels were 1.2 times higher or lower than normal samples.

Stromal-1 cases had high expression of CD68 (scores 2-3) and low expression of CD34 (quartiles I-II) (32.6% of cases), and as stromal-2 cases had low expression of CD68 (scores 0-1) and high expression of CD34 (quartiles III-IV) (11.6% of cases). The 40 cases that had high expression of both CD68 and CD34 (42.1%) were additionally considered as stromal-2, due to high MVD scores. Cases of low expression of CD68 and CD34 (13.7% of cases) were not scored. We observed a statistically significant relationship between the median expression of CD34 and Ann Arbor stage (I-II versus III-IV), with a predominance of high MVD expression in patients with advanced stage disease (III-IV). The median of vessels in stages I-II was 82.5 (range 9 to 240) and in stages III-IV was 111 (range 12-297) (p = 0.0276, Mann-Whitney). For the marker CD68 (stromal-1) and for the other variables, there were no statistically significant associations between groups. 101 of 111 initial cases were considered suitable for analysis of microRNAs. We observed overexpression of angiomiRs miR-92a, miR-296, miR-210, miR-378, Let-7f and miR-130a in 100%, 56.5%, 55,5%, 48.5%, 14.9% and 13.9% of cases, respectively. Among antiangiomiRs we found underexpression of miR-20b, miR-221, miR-16 and miR-328 in respectively 65.7%, 45.5%, 27.7% and 12.9%. We also observed a correlation between overexpression of miR-20b and low MVD (p = 0.0225, Mann-Whitney) and overexpression of miR-221 and low MVD (p = 0.0339, Mann-Whitney), suggesting a correlation between increased antiangiomiRs expression and low angiogenesis profile. For other variables, there were no statistically significant differences between groups.

Stromal-2 signature was found in 53.7% of cases. The higher MVD in cases of DLBCL with advanced stage (Ann Arbor III-IV) shows a possible association between angiogenesis and more aggressive/disseminated disease, being a possible target for antiangiogenic therapy in cases that do not achieve a complete response with R-CHOP. Two angiomiRs emerge as potential new targets: proangiomiR miR-92a, overexpressed in 100% of DLBCL, and antiangiomiR miR-20b, underexpressed in 65.7% of cases. Our study provides new information about the importance of the role of microRNAs in the development of DLBCL, which can be exploited as a therapeutic target for patients with this malignant disease of high prevalence. (Supported by FAPESP 2010/17668-6).

No relevant conflicts of interest to declare.