Case Presentations: Adverse Drug Reactions Involving P-Glycoprotein
ACVIM 2008
Katrina L. Mealey, DVM, PhD, DACVIM, DACVCP
Pullman, WA, USA


P-glycoprotein (P-gp) is a drug transport pump encoded by the ABCB1 gene, (MDR1 gene by earlier nomenclature). P-gp is expressed in cancer cells where it was originally discovered to play a key role in chemotherapeutic drug resistance by pumping chemotherapeutic drugs out of cancer cells. Several years later P-gp was found to also be expressed in various mammalian tissues including the apical border of intestinal epithelial cells, brain capillary endothelial cells, biliary canalicular cells, and renal proximal tubular epithelial cells. At these strategic locations P-gp appears to function in a protective capacity for the organism by decreasing exposure to potentially toxic xenobiotics found in the environment. P-gp expressed on enterocytes transports drugs into the lumen of the intestine, preventing their absorption. On brain capillary endothelial cells, P-gp contributes to the blood brain barrier by transporting substrate drugs back into the capillary lumen, limiting access of drugs to sensitive brain tissue. P-gp expressed on biliary canalicular cells and renal tubular cells actively effluxes substrate drugs that have entered these cells to the lumen of the bile canaliculus or renal tubule, respectively, excreting the drug from the body.

Mammalian P-gp has wide substrate specificity, transporting a number of drugs with diverse chemical structures, including anticancer agents (Vinca alkaloids, doxorubicin), immunosuppressants (cyclosporine, tacrolimus), antiparasitic agents (ivermectin, selamectin, milbemycin, moxidectin), corticosteroids (cortisol, dexamethasone), and others1. The mechanism by which P-gp can recognize and transport such a structurally diverse range of compounds is not known. Interestingly, most P-gp substrates are natural compounds, or synthetic derivatives of natural compounds, reinforcing the notion that P-gp's principal physiological role is to protect the organism from potentially toxic environmental xenobiotics. Lists of P-gp substrates are available in several review articles8,9. Many drugs have not been specifically evaluated with regard to their status as P-gp substrates, so it is likely that other drugs used in veterinary medicine will be identified as such and added to these lists. Thus far it appears that P-gp substrate specificity crosses species lines. That is, drugs determined to be substrates for human P-gp are also substrates for canine P-gp2, although studies specifically addressing this question have not been performed.

P-Glycoprotein's Role in Drug Disposition

The discovery in 20013 that some herding breed dogs have a mutation in the ABCB1 gene (ABCB1-1Δ, a mutation that leads to complete loss of P-gp function) explained a number of clinical observations, the most well-known of these being ivermectin sensitivity in Collies. Dogs with the ABCB1-1Δ mutation are susceptible to other adverse drug reactions as well. Some of the most important adverse drug reactions documented to date will be described below. Additionally, because P-gp's function can be inhibited by certain drugs, adverse drug reactions involving P-gp can occur in ABCB1 wildtype dogs. An example of this type of drug interaction will also be described.

P-gp and Intestinal Drug Absorption--Digoxin Toxicity

Digoxin toxicity was recently documented in a collie homozygous for the ABCB1-1Δ mutation4. A 9 year old female spayed Collie presented to the University of Wisconsin-Madison VTH for evaluation of tachycardia. Pertinent past history included surgery for a patent ductus arteriousus at 9 months of age. Although the ductus remained patent after surgery, the dog displayed no clinical signs until the current presentation. ECG findings were consistent with paroxysmal ventricular tachycardia. Treatment consisted of lidocaine, furosemide, digoxin, and enalapril. Once the arrhythmia was controlled the lidocaine was discontinued, mexiletine was initiated and the dog was discharged from the hospital. Seven days after the initial presentation the dog re-presented for anorexia and abnormal mentation. A serum digoxin concentration obtained 8 hours after dosing was 4.5 ng/mL (reference range 0.8-1.2 ng/mL). This patient developed an unusually high serum digoxin concentration resulting in digoxin toxicity despite the administration of 60% of the calculated daily digoxin dosage (i.e., 0.002 mg/kg in the AM and 0.004 mg/kg in the PM; recommended dosage, 0.005 mg/kg BID). Because other factors that precipitate digoxin toxicity, such as obesity, hypokalemia, or azotemia, were not present, it seems likely that a functional lack of P-gp led to increased oral bioavailability and decreased intestinal and/or renal excretion of digoxin in this patient. Studies in abcb1 knockout mice have shown increased oral bioavailability and decreased urinary clearance of digoxin1.

P-gp and Biliary Excretion--Vincristine Toxicity

A 6-year-old Collie diagnosed with lymphoma (Stage IIIa) was treated with vincristine (0.6 mg/M2). Two days later the dog became lethargic, anorexic and began vomiting. The dog's condition deteriorated over the next few days. The dog did not receive chemotherapy 1 week later because of neutropenia (1,100/µl). Two weeks after the initial treatment, the dog received cyclophosphamide (200 mg/M2) and did not develop neutropenia. A 2nd treatment with vincristine (0.5 mg/M2) resulted in neutropenia and apparent peripheral neuropathy. Nine days after the 2nd vincristine treatment the dog became febrile (104oF) and died (presumed cause of death was sepsis). This dog was homozygous for the ABCB1-1Δ mutation.

P-gp mediated excretion of vincristine through the biliary excretion is an important route of elimination for vincristine. In a rodent study, pharmacologic inhibition of P-gp resulted in a 6-fold reduction in biliary excretion of vincristine5. In nuclear scintigraphy studies involving normal dogs and dogs with the ABCB1-1Δ mutation, the role of P-gp in biliary excretion was evident. Radiolabeled P-gp substrate drug clearly accumulates in the gall bladders of ABCB1 wildtype dogs, but is absent from gall bladders of dogs with the ABCB1-1Δ mutation. Altered biliary excretion may play a role in the apparent increased sensitivity of herding breeds to chemotherapeutic drugs that are P-gp substrates (vincristine, doxorubicin). Dogs with the ABCB1-1Δ mutation (both heterozygotes and homozygotes) are extremely susceptible to myelosuppression and GI toxicity induced by the chemotherapeutic agents vincristine and doxorubicin (P-gp substrates) even at low doses, but appear to tolerate cyclophosphamide (not a P-gp substrate) at the full dose.

P-gp and CNS Drug Distribution--Loperamide Toxicity

P-gp is an important component of the blood-brain barrier minimizing the distribution of substrate drugs to brain tissue. Dogs that lack P-gp (those homozygous for the ABCB1-1Δ mutation) experience profound neurological effects when given "normal" doses of ivermectin. While "ivermectin sensitivity" is the classic adverse drug reaction associated with the ABCB1-1Δ mutation, ivermectin is not the only drug that can cause neurological toxicity in these dogs. It is important to point out that ivermectin causes neurological toxicity only at doses much higher than the label dose for heartworm prevention. Doses used for treating mange (> 300 µg/kg) will cause toxicity in dogs that are homozygous for the ABCB1-1Δ mutation. At high doses other macrocyclic lactones (milbemycin, moxidectin, selamectin) will also cause neurological toxicity in dogs that are homozygous for the ABCB1-1Δ mutation. An unrelated drug, the over-the-counter antidiarrheal agent loperamide (Imodium®), also causes severe neurological toxicity in dogs with the ABCB1-1Δ mutation.

A 3-year-old spayed female Collie was referred to the University of Wisconsin Veterinary Teaching Hospital (UWVTH)6 with a 2-day history of decreased appetite, vomiting, and diarrhea. The dog had been adopted from a Collie rescue agency 2 weeks before presentation, and previous historic information was not available. The referring veterinarian reported no abnormalities on physical examination and treated the dog with loperamide (0.14 mg/kg PO q12h). The dog presented to the referring veterinarian again the next morning. She had vomited once during the night and had developed profuse ptyalism. The veterinarian performed a diagnostic workup consisting of a CBC, serum biochemistry, venous blood gas analysis, and abdominal radiography. The only abnormalities detected were mild hypokalemia (3.4 mEq/L; reference range, 3.7-5.8 mEq/L) and hyperglycemia (131 mg/dL; reference range, 60-110 mg/dL). Treatment consisted of subcutaneous fluid administration and loperamide at the dosage previously specified. Over the next several hours, ptyalism continued and the dog developed progressively more severe neurologic signs, including rear limb weakness, apparent difficulty holding up the head, vocalization, disorientation, and ataxia. That evening, the dog was taken to an emergency veterinary hospital, where she was treated with additional subcutaneous fluid and metoclopramide (0.3 mg/kg SC) and discharged. The next day, the dog's owners contacted the Collie rescue organization regarding their dog's medical condition and were advised to discontinue loperamide because of possible breed sensitivity to the drug. The dog was presented to the UWVTH that afternoon for evaluation of neurologic and gastrointestinal signs, but by the time of presentation the dog's neurologic status had begun to improve. Neurologic examination identified disorientation, hyperexcitability, and hyperresponsiveness to noise. The dog was recumbent but able to walk when encouraged. She appeared ataxic without head tilt or intention tremor. Results of the cranial nerve examination were normal. Conscious proprioceptive deficits were present in all 4 limbs. Diagnostic procedures performed on admission included a CBC and serum biochemistry. All results were within reference ranges. ABCB1 genotyping was performed and the patient was found to be homozygous for the ABCB1-1Δ mutation. In subsequent studies the author has demonstrated that the recommended dose of loperamide (0.2 mg/kg) cause severe CNS depression in dogs with the ABCB1-1Δ mutation yet will produce no adverse effects in dogs that express P-gp normally (ABCB1 wildtype dogs). Loperamide is an opioid that is generally devoid of CNS activity because it is normally excluded from the brain by P-gp7.

Drug Interactions Involving P-Glycoprotein--Ivermectin Toxicity in an ABCB1 Wildtype Dog

The WSU Veterinary Clinical Pharmacology Laboratory was contacted regarding apparent ivermectin toxicity in a 2-year-old female spayed Shiba Inu that was being treated for demodectic mange. The dose of ivermectin was within the recommended range (300-600 µg/kg/day), but the clinical signs were consistent with ivermectin toxicity. Although the ABCB1-1Δ mutation has not been reported in this breed, samples were submitted for determination of ABCB1 genotype. The owner was queried about exposure to other potential toxins or other drugs. The owner reported that the dog had been receiving ketoconazole (5.8 mg/kg q 12 H) for concurrent Malassezia dermatitis. The owner was instructed to discontinue administration of ivermectin until the ABCB1 genotype results were available. Neurological signs subsided after discontinuation of the ivermectin. Because the dog's genotype was ABCB1 wildtype/wildtype, we speculated that ketoconazole could be inhibiting P-gp resulting in an 'ivermectin sensitive' phenotype. When ivermectin was re-administered (after the ketoconazole was discontinued for 5 days), the dog was able to tolerate ivermectin.

A wide variety of drugs are known to be substrates of P-gp, but in addition there are a wide variety of drugs that have been shown to inhibit P-gp function. Ketoconazole is a highly effective inhibitor of P-gp in human cells and was recently shown to inhibit canine P-gp as well**. Other drugs that may inhibit P-gp in canine patients at clinically used doses include cyclosporin, tamoxifen, and verapamil.

Pesticides and P-gp--Abamectin toxicity

A five-year-old neutered male Australian Shepherd, Kahlua, presented to his veterinarian for acute onset of neurological signs which included ataxia, hypersalivation, and weakness. The owner suspected that the dog had developed toxicity as a result of exposure to a pesticide that had been applied to the yard. Kahlua received supportive therapy (parenteral fluids). The exterminator was contacted and reported that the pesticide used was Advance®, which contains abamectin, a mixture of avermectin 1a and avermectin 1b. Ivermectin is actually a combination of 22,23-dihydroavermectin B1a + 22,23-dihydroavermectin B1b, therefore it is reasonable to assume that abamectin would not only cause clinical signs similar to ivermectin, but that abamectin would also be a substrate for P-gp. Kahlua's owner contacted the WSU Veterinary Clinical Pharmacology Laboratory for additional diagnostic and therapeutic advice. Kahlua tested homozygous for the ABCB1-1Δ mutation. The author is aware of several other dogs that have experienced abamectin toxicity as a result of exposure to pesticides. Until recently, these pesticides were not available to anyone but a licensed applicator. However, over-the-counter products are now available that contain abamectin.

Abcb1-1Δ Testing in Clinical Patients

Thus far, the ABCB1-1Δ mutation has been identified in 10 different dog breeds. Nine of those breeds and the frequency of the mutation are indicated in (Table 1)--the 10th breed is the English Shepherd. The above examples demonstrate that the presence of the ABCB1-1Δ mutation in a particular patient has broad clinical implications with respect to veterinary medicine. A test is commercially available so that veterinarians can screen dogs for the ABCB1-1Δ mutation prior to administering drugs that are potentially problematic (Veterinary Clinical Pharmacology Laboratory, College of Veterinary Medicine, Washington State University, Pullman, Washington;

Table 1. ABCB1-1Δ frequency in a North American population of dogs.


Genotype (%)



ABCB1 wt/wt

ABCB1 mut/wt

ABCB1 mut/mut

Number of dogs

Australian Shepherd





Border Collie










German Shepherd





Herding Breed Cross





Long-haired Whippet





Min. Australian Shepherd





Old English Sheepdog





Shetland Sheepdog





Silken Windhound










Other purebred dogs





*Owner or DVM either did not list breed or reported breed as unknown


1.  Schinkel A, et al. J Clin Invest 1995;96:1698.

2.  Page RL, et al. Anticancer Res 2000;20:3533.

3.  Mealey et al. Pharmacogenetics 2001;11:727.

4.  Henik RA, et al. J Vet Int Med 2006;20:415.

5.  Song S, et al. Drug Metab Disp 1999;27:689.

6.  Sartor LL. J Vet Int Med 2004:18:117.

7.  Wandel C, et al. Anesthesiology 2002;96:913.

8.  Ford, et al. Cytotechnology 1993;12:171.

9.  Mealey J Vet Pharmacol Ther. 2004; 27:257.

Speaker Information
(click the speaker's name to view other papers and abstracts submitted by this speaker)

Katrina Mealey, DVM, PhD, DACVIM, DACVCP
Washington State University
Pullman, WA

MAIN : ACVCP : P-glycoprotein
Powered By VIN