endothelia (Mealey et al. 2001 ; Roulet et al. 2003 ). Only the amino-terminus of the
protein is generated (7.1% of the fully functional protein) and the ATP-binding,
substrate-binding, and phosphorylation sites required for normal function are
absent. For sensitive dogs, the phenotypic consequences are considerable;P-gp
exports a wide variety of structurally-unrelated hydrophobic therapeutic drugs
including loperamide, cyclosporin, digoxin and ivermectin. Atypical hypersensitiv-
ity to all such agents has been observed inP-gp-deficient Collies (Mealey et al.
2001 , 2003 ). Consequently, a genetic test for the homozygous mutation is now
commercially available from several international laboratories for determining the
MDR1 genotype of herding-breed dogs prior to the administration of potentialP-gp
substrates.
Selamectin exhibits a wider safety margin in phenotypically-sensitive dogs than
ivermectin following topical or oral administration of high doses (Bishop et al.
2000 ; Novotny et al. 2000 ). Although selamectin is equipotent to ivermectin as a
P-gp substrate and inhibitor in vitro (Griffin et al. 2005 ), recent in vivo data from
Abcb1a(/) mice showed that selamectin accumulates in the brain to a far lesser
extent than ivermectin, even though it was confirmed as aP-gp substrate (Geyer
et al. 2009 ). Even when present in high concentrations in the CNS, it did not
provoke neurotoxicity in mice. Neurotoxicity is normally detected in sensitive
Collies after oral consumption of either a concentrated ivermectin pour-on designed
for farm livestock or from off-label high oral dose ivermectin treatment of canine
demodectic mange, whereas selamectin is administered only topically to dogs in
a spot-on formulation (Sarasola et al. 2002 ). Mealey ( 2008 ) argues, as a general
principle, that administration of selamectin and the related compounds, moxidectin
and milbemycin, in doses exceeding those recommended for heartworm prevention
may have the potential to induce neurotoxic symptoms in ABCB1-1Ddogs.
In support, doramectin and milbemycin oxime caused neurotoxicosis following
administration at high dose rates to ivermectin-sensitive dogs (Yas-Natan et al.
2003 ; Tranquilli et al. 2001 ), data recently confirmed for milbemycin oxime in a
separate study of the treatment of demodicosis in dogs carrying theABCB1-1D
genotype (Barbet et al 2009 ). The argument is however unlikely to apply to
selamectin, given the PK data reported by Geyer et al. ( 2009 )inAbcb1a(/)
mice. Although Paul et al. ( 2000 ) observed a wider safety margin of moxidectin in
avermectin-sensitive Collies compared to ivermectin and milbemycin, cases of
moxidectin poisoning have still been documented, resulting from accidental inges-
tion of high doses inABCB1-1DCollies (Beal et al. 1999 ) and other breeds
(Snowden et al. 2006 ).
The discovery of theP-gp mutation in Collies and herding breeds has implica-
tions for potential dose-dependent drug interactions between co-administeredP-gp
substrates and inhibitors in both wild-type and mutant dogs. There are numerous
examples of drugs, excipients, and nutraceutical components used in veterinary
medicine that areP-gp substrates or inhibitors or both (Mealey 2004 ; Martinez et al.
2008b). Selected relevant examples of substrates include doxorubucin, dexametha-
sone, ketoconazole and digoxin, while inhibitors include fluoxetine, verapamil, and
cyclosporine. Formulation excipients that inhibitP-gp range from PEGs to pluronic
Drug Delivery Systems in Domestic Animal Species 101