Cellular distribution of phosphorothioate oligodeoxynucleotides in normal rodent tissues.

The distribution of intravenously injected phosphorothioate oligodeoxynucleotides (P = S ODN) was studied in vivo in rodent tissues using three histologic methods: immunohistochemistry with a monoclonal antibody that recognizes P = S ODN ISIS 2105; direct fluorescence microscopy of P = S ODN ISIS 2105 conjugated to rhodamine; and autoradiography of 14C-labeled P = S ODN ISIS 2302. All three methods gave the same pattern of oligonucleotide distribution, and the intensity of the histologic signal agreed with previously published pharmacokinetic data on the relative concentration of P = S ODN in different organs. Proximal tubule cells in the kidney and Kupffer and endothelial cells in the liver were among the most heavily labeled with P = S ODN at all doses and time-points. Connective tissues proper, such as the lamina propria and submucosa of the intestine and the dermis and subcutaneous layer of the skin, were also labeled, whereas the P = S ODN signal was weak or negative in epithelial and muscle cells in the skin and intestine. At 2 hours postinjection, P = S ODN were clearly detectable in the extracellular matrix in loose and dense connective tissues, although by 24 hours, the label was predominantly intracellular. Large, nucleated cells in red marrow, and the connective tissues around bone and skeletal muscle cells and lining the knee joint, were positive for oligonucleotide, whereas P = S ODN were not detected in erythrocytes, cartilage, compact bone, and skeletal muscle. In spleen, white pulp was negative for P = S ODN, whereas cells surrounding the sinusoids and nucleated cells in the red pulp were strongly positive for P = S ODN. Our results provide specific information on the tissue and cellular localization of P = S ODN within organs in vivo. The data presented will be used as a reference for studies of P = S ODN distribution in diseased tissues and the distribution of modified oligonucleotides. Furthermore, because our results indicate which cell types are likely to be affected by antisense oligonucleotides, they can be used to guide future in vivo applications of the technology.