Biophysical and pharmacokinetic properties of five analogs of ISIS 3082, a 20-mer phosphorothioate oligodeoxynucleotide that inhibits the expression of mouse intercellular adhesion molecule 1, were evaluated. Compared to the parent compound, ISIS 3082, the 2'-propoxy modified phosphodiester, ISIS 9044 and the 2'-propoxy phosphorothioate, ISIS 9045, had greater affinity for complementary RNA and were more lipophilic. A chimeric oligonucleotide comprised of 2'-propoxy diester wings and a phosphorothioate deoxy center (ISIS 9046) had equal affinity. It was also more lipophilic than ISIS 3082, but less so than the other 2'-propoxy modified analogs. The two analogs with 5'-lipophilic conjugates, ISIS 9047 (5'-octadecylamine) and ISIS 8005 (5'-(2'-O-hexylamino-carbonyl-oxycholesterol) were more lipophilic than ISIS 3082 (3- and 7-fold, respectively) but had similar affinity for complementary RNA. Binding of ISIS 3082 to bovine serum albumin was salt-dependent and, at physiological concentration (320 mOsmol), the dissociation constant (Kd) was 140 microM. Similarly, the 2'-propoxy phosphodiester, ISIS 9044, displayed salt-dependent bovine serum albumin binding, but not binding was measurable at physiological salt conditions. In contrast, the more lipophilic phosphorothioate analogs displayed much higher affinity to bovine serum albumin at 320 mOsmol than ISIS 3082. After bolus injection to mice, the initial volumes of distribution of the more lipophilic phosphorothioate analogs, ISIS 9045, ISIS 9047 and ISIS 8005, were less and the initial clearance from plasma was slower than ISIS 3082. The pharmacokinetics of the other analogs was similar to ISIS 3082. Distribution of ISIS 3082 into peripheral tissues was similar to that reported for other phosphorothioates with liver and kidney accumulating the highest fraction of the dose. The only modification to markedly influence distribution was the very lipophilic cholesterol conjugate (ISIS 8005), which increased substantially the fraction of the dose accumulated by the liver. Little intact drug was found in urine or feces for any analog, and the patterns of metabolites suggested that for all analogs the principal metabolic pathway was due to 3'-exonuclease activity. The metabolism of ISIS 3082 was similar to that reported for other phosphorothioates. After 2 hr, most of the radioactivity in plasma represented metabolites but, in tissues, intact ISIS 3082 was present for much longer periods of time and metabolites accumulated more slowly. The 24-hr exposure to ISIS 3082 of liver and kidney was 20.7 and 67.9 microM/hr, respectively. The rates of metabolism in plasma, liver and kidney of the two 5'-conjugates, ISIS 9047 and ISIS 8005, were similar to ISIS 3082, as was the pattern of metabolism. The rate of metabolism of ISIS 9044 (2'-propoxy phosphodiester oligonucleotide) was much more rapid in liver and plasma, but surprisingly much slower in the kidney. ISIS 9045 (full 2-propoxy phosphorothioate) was much more stable than ISIS in all tissues, the enhanced stability of ISIS 9045 resulted in increased exposure of liver and kidney to the drug, whereas the exposure of the liver to the two more lipophilic analogs, ISIS 9047 and ISIS 8005, was greater because a higher fraction of the dose was distributed to the liver. The exposure of the kidney to ISIS 9044 was also greater than that to ISIS 3082 due to the surprising stability of the drug in the kidney.