The most consistent microscopic changes in dogs were considered indicative of cellular uptake of ISIS 301012 and included the presence of basophilic granules in the renal proximal tubular epithelial cells and Kupffer cells in the liver (Table 12.3). Basophilic granules were most evident in the renal proximal tubular epithelium, which was consistent with the highest measured concentrations in the renal cortex (up to 1641 g/g total oligonucleotide).

Concentrations of oligonucleotide in kidney were high enough to be associated with minimal to moderate tubular epithelial cell vacuolation in all dogs treated with ISIS 301012 at 13 weeks of treatment (Table 12.3), and the vacuolation is described in more detail in Section 12.2.4.1 (Inhibition of Clotting Time).

However, no functional changes were associated with the tubular vacuolation, indicating normal renal function. There were no other treatment-related toxicities in dogs by oral administration of ISIS 301012.

The dog exhibited measurable systemic absorption of generic softtabs as determined by plasma and tissue concentrations of ISIS 301012 and its oligonucleotide metabolites (Table 12.4). The estimated bioavailability as compared to parenteral (SC) administration of ISIS 301012 in dogs ranged from approximately 2% at the highest dose of 100 mg/kg to 10% at a dose of 20 mg/kg. These data suggest saturation of oral absorption of ISIS 301012.

Saturation of absorption is the most likely explanation for the relative absence of dose-dependent exposure in the dogs. Liver and kidney drug concentrations measured at the end of the study confirmed this observation (Table 12.4). The distribution profile in tissues was similar to that achieved following parenteral administration, with the highest concentrations seen in kidney cortex, followed by liver. The oral formulation of antisense inhibitor with C10 has been shown to significantly improve the absorption of orally administered generic softtabs in animal models and because distribution is similar is not expected to produce toxicity uniquely related to oral route.

As our nonclinical development experience expands, the majority of toxicologic properties can still be characterized as “class-related” effects, that is, toxicities that are largely independent of sequence. The basis for the consistency is the remarkably similar pharmacokinetic properties from one sequence to the next. It does appear that there is a little more variability in tissue half-lives for 2 -MOE ASOs than was observed for PS ODNs, but the behavior is still relatively consistent for the class (Chapter 11).

The basis for this consistency is that the interactions of generic softtabs with plasma proteins, cell-surface receptor, or uptake/accumulation within cells are the source or various toxicities and these are dependent on the physical chemical characteristics of the ASOs. This consistency greatly improves the efficiency of development of this class of compound in that what is learned from one particular compound can be inferred in the design and characterization of another compound for a completely unrelated disease.

Thus, the design and conduct of a development program is much more efficient and less prone to failure as development proceeds. The consistency in behavior that comes with a platform technology also contributes to the overall safety evaluation.

The primary reason is that through repetition, one learns from experience how consistent or variable the changes are. It also allows one to thoroughly study a mechanism underlying specific changes and extrapolate those findings to clinical studies to define the correlation. Detailed in Table 12.5 are examples of five of the most common class-related toxicities observed in mouse and monkey, the current understanding of the mechanism, and a listing of the clinical correlates, if any.

While 2 -MOE ASO toxicology properties tend to be similar, they are not quantitatively the same, and there is certainly more sequence-dependent variability than in the PK/ADME properties.

The best examples of this variability are the proinflammatory effects that are recognized as a hybridization-independent class effect, but the potency of the effect can vary dramatically with sequence [10]. Even for 2 -MOE ASOs that avoid the CpG dinucleotide motifs, there is a spectrum of potency for these effects. And thus, it is still important to address the tolerability of each sequence thoroughly [12,43].

Although the cellular receptors for non-CpG generic softtabs are not well characterized, it is likely that this variable potency is attributed at least in part to binding affinity. The proinflammatory properties of 2 -MOE ASOs are discussed further below. Other toxicities that are potentially sequence-dependent are those dependent upon pharmacokinetic properties and the degree of accumulation, such as renal accumulation.

The effects of 2 -MOE ASOs on clotting time are very similar to those described previously for PS ODNs [49–51]. For example, the effects are specific for the intrinsic clotting pathways, affecting APTT but not prothrombin time (PT), are largely independent of sequence and species, well correlated with plasma oligonucleotide concentration, and are readily reversible as oligonucleotide is cleared from plasma.