FUNDAMENTAL
AND
APPLIED
Zidovudine
TOXICOLOGY
14,764-775
(
1990)
Toxicity to Cats Infected with Feline Leukemia Virus
WANDA M. HASCHEK, RONALD M. WEIGEL, GAIL SCHERBA, MA CRISTINA DEVERA, RHONDA FEINMEHL, PHILIP SOLTER, MARY B. TOMPKINS,’ AND WAYNE A. F. TOMPKINS’ Department yf’Puthobiology, College of L’eterinary Medicine, University qflllinois,
Urbana, Illinois 61801
Received July 27, 1989; accepted December 18, 1989
Zidovudine
W. M.. M. B., AND TOMPKINS. W. A. F. (1990). Fundam. Appl. To.xicoi. 14, 764-715. Feline leukemia virus (FeLV) infection of cats is a model for the acquired immunodeficiency syndrome in humans. The toxicity of zidovudine was evaluated in SPF cats experimentally infected with FeLV. At initiation of the zidovudine study, all cats were antibody positive for FeLV antigens but clinically asymptomatic. Four cats were also viremic. Thirteen, 6- to IO-month-old cats were divided into five dosage groups and given zidovudine po at 0. 7.5. 15, 30. or 60 mg/kg daily in three equally divided doses for 32 to 34 days. Titers of circulating virus antigen remained constant: however, three of six cats receiving the higher doses of zidovudine (a30 mg/kg) showed an increase in antibody titers to FeLV. Administration of zidovudine resulted in a progressive anemia, dependent upon dose and time. Macrocytes were observed prior to the development of anemia and were also found in several nonanemic cats. Repeated measures regression analyses indicated that an increased dose of zidovudine was associated with decreased packed cell volume, red blood cell count, and hemoglobin. As determined from the packed cell volume, the analyses indicate that anemia is induced only by the two highest doses ofzidovudine. The regression model indicates that daily doses of 60 and 30 mg/kg are expected to induce anemia by Day 4 and Day 13. respectively. Progressive absolute neutropenia was observed in the a30 mg/kg groups. Histopathologic lesions consisted of marked bone marrow hypercellularity in cats given >30 mg/kg zidovudine and splenic extramedullary hematopoiesis in cats given > 15 mg/kg. Thus, oral toxicity of zidovudine in the cat is manifested by a dose-related anemia and neutropenia as observed in humans. 01 1990 Society ofToxicology. WEIGEL,
R. M.,
Toxicity to Cats Infected with Feline Leukemia Virus. SCHERBA,
G.. DEVERA.
M. C., FEINMEHL,
Zidovudine, also known as azidothymidine (3’-azido-3’-deoxythymidine) or Retrovir, a thymidine analog, is widely used as a treatment for AIDS or AIDS-related complex in humans. Zidovudine is phosphorylated to AZT triphosphate which competes with cellular nucleotides for incorporation into viral DNA by the viral reverse transcriptase (Mitsuya et al., 1985). In this manner, zidovudine inhibits the replication of human immunodeficiency virus (HIV) in vitro (Mitsuya et al., ’ Present address: North Carolina State University, College of Veterinary Medicine, 4700 Hillsborough Street, Raleigh, NC 27606. 0272-0590/90 $3.00 Copyright 0 1990 by the Society of Toxicology. All rights of reproduction in any form reserved.
764
R.. SOLTER,
HASCHEK, P., TOMPKINS,
1985) and has been shown to decrease viremia, lengthen survival, and decrease frequency of opportunistic infections in AIDS patients (Fischl et al., 1987; Yarchoan et al., 1986; Yarchoan and Broder, 1987). Although significant clinical benefits of zidovudine therapy have been documented, hematologic toxicity consisting of macrocytosis, anemia, and neutropenia has also been observed (Dournon et al., 1988; Fischl et al., 1987; Gill et al., 1987; Richman et al., 1987). Anemia appears to be due to red cell hypoplasia or aplasia based on bone marrow examination of anemic patients that showed red cell aplasia, erythroid hypoplasia, or meg-
ZIDOVUDINE
TOXICITY
aloblastic erythropoiesis (Walker et al., 1988). Marked zidovudine toxicity for bone marrow hematopoietic precursor cells is also evidenced by inhibition of granulocyte-macrophage CFU and erythroid burst-forming cells in vitro (Sommadossi and Carlisle, 1987). Although zidovudine has been used to treat retrovirus infection in other species used as models for human AIDS (Ruprecht et al., 1986; Tavares et al., 1987), there is little information on toxicity of the drug in species other than man. Recent studies indicate that cynomolgus monkeys are quite susceptible to zidovudine-induced hematologic toxicity, while rats are resistant (Ayers, 1988). Cats infected with feline leukemia virus (FeLV) have been studied as a possible model for HIV infection, since they develop similar immunodeficiency and neoplastic syndromes (Hoover et al., 1989; Tavares et al., 1989). Among the immunopathological similarities to HIV infection, FeLV-infected cats show a deficiency in T helper (TH, ) cell production of IL-2 (Tompkins et al., 1989a) and reduced lymphokine-activated killer (LAK) cell function (Tompkins et al., 1989b). Although feline immunodeficiency virus (FIV) is much more closely related to HIV than is FeLV, it has a long and variable latent period, and does not produce consistent clinical signs and lesions in infected cats (Pedersen et al., 1989). Therefore, at the present time, FIV is not a suitable model for evaluation of potential therapeutic agents. Data obtained from this study however will also be applicable to FIV-infected cats as the model is further developed. Zidovudine has been evaluated in FeLVinfected cats and shown to have therapeutic and preventive potential (Tavares et al., 1987); however, its hematologic toxicity has not been evaluated. As part of a large program to evaluate promising biological response modifiers and antiviral chemicals for efficacy and safety, the subacute oral toxicity of zidovudine was examined in cats experimentally infected with the immunosuppressive Rickard strain of FeLV. The study was designed to determine the dose at which tox-
IN
765
CATS
icity was manifested, the target organs of toxicity, and the highest nontoxic dose for future therapeutic studies. Because these cats were infected with FeLV, virological and immunological parameters were also measured. Hematologic toxicity was observed and similarities to the human condition are discussed. METHODS Experimental design. Thirteen, specific pathogen-free (SPF), Mill Hill strain cats (Liberty Lab, Liberty, NJ) were experimentally infected intravenously with approximately 5 X lo4 focus forming units of the Rickard strain of FeLV (obtained from Dr. L. Mathes, Ohio State University, Columbus) at 6 to 8 weeks of age. All cats developed subclinical infections with circulating antibodies to FeLV antigen and four were viremic when the zidovudine study was initiated. At 6 to 10 months of age, the cats were divided into five dosage groups (Table 1): 0 (n = 2). 7.5 (n = 3) I5 (n = 2), 30 (n = 3) and 60 (n = 3) mg/kg/ day zidovudine. Zidovudine was administered in three equally divided doses in moist cat food for 32 to 34 days, with doses adjusted weekly according to body weight. Cats were weighed and evaluated daily by physical examination. Blood samples (5 ml) were drawn weekly from the jugular vein, beginning 1 week prior to the start of zidovudine treatment. for hematologic and clinical chemistry evaluations and for viral and immunological measurements. At the end of the study all cats were euthanized with T-6 I euthanasia solution (Taylor Pharmacal Co., Decatur, IL) and complete necropsies performed. Zidovudine. Zidovudine (Retrovir) or 3’-azido-3’deoxythymidine (AZT) was generously provided by Burroughs Welcome Co. (Research Triangle, Park, NC). Hematology. The peripheral blood compartment was evaluated as follows. Percentage packed cell volume (PCV) was determined with an Autocrit II centrifuge (Fisher Scientific Co., Fairlawn, NJ). Total erythrocyte (RBC) and leukocyte (WBC) counts were obtained according to standard techniques (Jain, 1986). Hemoglobin (Hb) concentration was assayed by the cyanmethemoglobin method (Sigma Chemical Co., St. Louis, MO). Mean corpuscular volume (MCV) and mean corpuscular hemoglobin concentration (MCHC) were calculated as MCV = IO X (PCV/RBC) and MCHC = 100 X (Hb/ PCV). Leukocyte differentials were performed to determine absolute numbers of neutrophils, lymphocytes, eosinophils, and monocytes. Bone marrow smears made at necropsy were examined to interpret more precisely the peripheral blood data. Serum biochemistry. The concentrations of serum total protein. albumin, globulin, creatinine, blood urea nitrogen, glucose, cholesterol, phosphorus, calcium, sodium, chloride, potassium. and total bilirubin as well as the activities of alkaline phosphatase, aspartate amino-
766
HASCHEK
transferase, alanine aminotransferase, and lactate dehydrogenase were determined with an autoanalyzer (Hycel Super Seventeen Autoanalyzer, Houston, TX). Patho1og.v. At necropsy, the following organs were weighed: brain, adrenals, heart, kidneys, liver, spleen, thymus, testes, and epididymus. Sections from the above. as well as from thyroid and parathyroid. lymph nodes, bone marrow, lung, pancreas, ovary, stomach. and intestines, were fixed in 10% buffered neutral formalin; except for testes and ovaries which were fixed in Bouin’s solution. Tissue sections were embedded in paraffin, sectioned at 4 to 6 pm. and stained with hematoxylin and eosin. All tissues were examined from control and 60 mg/kg zidovudine groups while only bone marrow, spleen, thymus, and lymph nodes were examined in the remaining groups. Impression smears were made from femoral bone marrow, fixed in 70% alcohol, and stained with Wright-Giemsa. Bone marrow was evaluated cytologically and the ratio of myeloid to erythroid cells was determined. Assayfir FeLVantigen. FeLV viremia was monitored by testing the plasma fraction of blood for FeLV group specific (gs) antigen by an ELISA system (TechAmerica). Assay for FOCMA antibodies. Serums were assayed for antibodies to the FeLV-induced feline oncorna virus cell membrane-associated antigen (FOCMA) by indirect immunofluorescence using a FOCMA-positive FL74 lymphoma cell line (obtained from Dr. W. Hardy, Sloan Kettering Institute, New York) as the target. Briefly, I X lo6 viable FL74 cells were incubated with 100 ~1 of serial dilutions (1:4- 1:64) of plasma from FeLV-infected cats for 20 min at 4°C. After incubation, the cells were washed in ice-cold PBS plus 2% FBS. fixed for 5 min with 1% paraformaldehyde. washed, incubated in 100 pg of fluorescein-conjugated goat anti-feline IgG (heavy and light chain) for 20 min, and then washed three times in PBS/FBS. The cell pellets were suspended in 50 ~1 PBS: glycerol (1:9) and mounted under a glass coverslip for viewing with a fluorescence microscope. Positive reactions are revealed as a beaded fluorescence on the surface of the cell. Statisticul methods. The effect of zidovudine on hematologic measures over the course of therapy was evaluated using a repeated measures multiple regression model (Cohen and Cohen. 1983). For each hematologic measure analyzed, two analyses were conducted. In the between subjects variance analysis, the dependent variable was the mean of the two hematologic measures taken during the 5th (i.e., final) week oftherapy, and the independent variable was zidovudine dose with baseline level of the hematologic variable (which was the mean of the two pretreatment measurements) as a covariate. The average of two measures was utilized to increase reliability (Ederer, 1972). In the within subjects variance analysis, the dependent variable was the value of the hematologic measurement at each of the five post-treatment times, and the independent variables were pretreatment level of the outcome variable, days since onset of zido-
ET AL. vudine therapy, and a dose by time interaction term. An u level of 0.05 was chosen for all statistical analyses. In both the between subjects and within subjects regression analyses, presence or absence of viremia was also added to the model as an independent variable in supplementary analyses. However, no effect of viremia was apparent, so this variable was excluded from the regression models. The relationship between zidovudine dose and several dichotomous outcomes were investigated post hoc in 2 X 2 contingency table analyses. In each case, zidovudine dose was dichotomized as “~30” (0, 7.5, and 15 mg/kg/ day doses; total n = 7) versus “>30” (30 and 60 mg/kg/ day; total n = 6). The departure from random association was examined statistically using the Fisher Exact Test.
RESULTS All cats remained alert and clinically healthy throughout the study, with exception of pale mucous membranes and a slow capillary refill time that developed in cats given the two highest doses of zidovudine. Table 1 summarizes the observed hematological changes. Tables 2a and 2b present descriptive statistics on mean values of body weight and hematologic measures as a function of dose and time. It is apparent that body weight was not affected by zidovudine (Table 2a). Administration of zidovudine did result in a progressive anemia, using the criterion of a decrease in PCV to less than 30% (Table 2a). All six cats in the highest dose groups (a30 mg/kg/day) developed anemia, in contrast to the seven cats in the lower dose groups (~30 mg/kg/day) (one-tailed p = 0.0047, Fisher Exact Test). Direct microscopic examination of the peripheral blood smears characterized the anemia as normochromic with anisocytosis distinguished by many erythrocytes measuring up to 8 pm in diameter (macrocytes). Normal erythrocyte diameter in the cat is 5.5 to 6.3 pm. All six cats in the highest dose groups (230 mg/kg/day), compared to three of the seven cats in the lower dose groups (~30 mg/kg/day), developed macrocytes (one-tailed p = 0.049, Fisher Exact Test). Macrocytes were observed in all six cats of the highest dose groups prior to the development of anemia. Reticulocyte counts were not performed; however, the macro-
ZIDOVUDINE
TOXICITY
767
IN CATS
TABLE I AGE, SEX, VIREMIC
Daily zidovudine dose’ (m&kg) 0 1.5 15 30 60
STATUS, AND HEMATOLOGIC CHANGES
OF CATS TREATED WITH ZILIOVUDINE
ID6
Cat age (months)
Sex
Macrocytes observed
TWI NT3 TY2 oc4 NT5 TV2 UB2 TX2 NZl NT4 UB3 NU4 NZ6
6 6 10 10 10 10 10 6 6 6 6 6 6
M M M M F M M M M F M M F
+ + + fC +’ +’ +c +< +c
Anemia + + + + + +
Neutropenia + + + + -
a Zidovudine was given orally in three equally divided doses for 32 to 34 days. b Identification number. Cats with same first letter belong to same litter. ’ Macrocytes observed prior to development of anemia.
cytes did not appear to be immature cells as they were not basophilic. Regression analyses confirmed that 230 mg/kg doses of zidovudine caused significant decreases in three hematologic variables: PCV, RBC, Hb, and increased MCHC (Table 3). Although MCV appeared to increase over time in the higher dose groups and macrocytes were noted on peripheral blood smears, in the regression analysis, zidovudine did not significantly effect the MCV. Data from repeated measures regression analysis (Table 3) show that increasing dose of zidovudine during the 5th week of treatment was associated with decreased RBC, decreased PCV, decreased Hb, and increased MCHC and MCH. The increase in MCHC and MCH with increasing zidovudine dose reflects a more rapid decline in PCV relative to Hb with increasing dose. In the absence of lipemia, or large numbers of Heinz bodies on peripheral blood examination, this is suggestive of free hemoglobin in the plasma. There were also significant dose by time interaction effects for RBC, PCV, and Hb, as is apparent from Figs. 1, 2, and 3, plotting the
regression lines for each dose group over time. In addition, the decrease in these red blood cell measures from baseline levels occurs more rapidly with increasing dose. Using a criteria of PCV < 30% as a definition of anemia, the regression analyses suggest an induction of anemia by Day 4 with a daily zidovudine dose of 60 mg/kg and by Day 13 with 30 w/k. No significant association of zidovudine dose with WBC hematologic measures was apparent from the regression analyses. However, a decrease in absolute numbers of neutrophils without a left shift was observed at high doses (Table 2b). Four of the six cats in the two highest dosage groups (230 mg/kg/ day) developed an absolute neutropenia, using the criterion of a decrease in neutrophil numbers to less than 3000. None of the seven cats in the three lower dose groups developed neutropenia (one-tailedp = 0.02 1; Fisher Exact Test). The thrombocytic series in all cats appeared normal. Although evaluation of bone marrow smears was complicated by technical problems, the numbers of immature myeloid and
768
HASCHEK TABLE DESCRIPTIVE
STATISTICS:
BODY WEIGHT
ET AL. 2a AND ERYTHROCYTE Zidovudine
(Sample
size):
0 (fl = 2)
7.5 (n = 3)
PARAMETERS
dose (mg/kg/day) 15(n=2)
30 (n = 3)
60 (n = 3)
Week
Mean
SEM
Mean
SEM
Mean
SEM
Mean
SEM
Mean
SEM
Pre 2 3 4-5
3.95 3.90 3.90 4.00 3.90
0.95 0.70 0.90 1.oo 0.90
3.20 3.13 3.20 3.27 3.40
0.42 0.52 0.53 0.64 0.58
3.75 3.60 4.00 3.90 4.10
0.15 0.20 0.20 0.10 0.10
3.13 3.27 3.20 3.27 3.13
0.43 0.29 0.35 0.35 0.33
2.88 3.00 2.93 3.13 3.00
0.32 0.31 0.37 0.41 0.31
Red blood cell (X lob/d) [normal range: 5-101
Pre 1 2 3 4-5
7.98 8.00 7.87 8.39 7.53
0.70 0.69 1.14 0.17 0.39
6.98 6.85 5.96 5.07 6.67
0.21 0.14 0.18 0.45 0.26
7.54 7.52 7.48 6.89 6.64
0.18 0.03 0.05 0.44 0.08
6.42 6.77 4.63 5.09 4.58
0.72 0.76 0.47 0.88 0.42
6.50 5.91 4.12 5.54 3.85
0.12 1.12 0.17 0.60 0.13
Hemoglobin [normal
Pre
1 1.45 11.75 13.10 12.60 1 1.63
1.45 1.15 0.80 0.40 0.03
11.43 11.13 10.10 9.93 11.83
0.77 0.52 0.64 1.42 0.94
11.60 11.55 12.05 11.95 II.20
1.45 1.55 1.45 1.95 1.39
10.20 10.73 9.40 9.20 8.68
0.44 1.19 0.25 0.70 0.65
10.08 10.50 8.03 8.37 6.70
0.32 0.85 0.20 0.87 0.40
34.15 36.00 38.30 38.70 35.00
4.50 2.80 6.80 0.50 0.60
33.80 33.80 29.87 29.97 36.07
2.53 2.08 1.54 3.39 2.18
34.98 35.35 36.40 34.45 34.10
3.07 3.25 3.20 5.25 3.70
31.23 33.90 23.23 26.16 25.70
1.19 4.31 0.58 2.78 1.99
30.05 28.20 18.80 24.67 18.37
0.84 5.31 0.86 2.26 0.87
43.09 45.00 48.40 46.15 46.63
2.03 0.40 1.70 1.55 1.63
48.27 49.27 50.17 52.77 53.83
2.36 2.23 2.43 2.19 2.94
46.53 47.00 48.70 52.50 51.35
5.18 4.50 4.60 3.99 4.65
49.58 50.37 5 1.00 52.50 57.13
4.14 4.40 3.91 3.98 7.18
46.23 47.70 45.60 44.63 47.87
0.68 0.31 0.32 0.72 0.75
33.58 32.55 34.90 32.53 33.25
0.22 0.65 4.10 0.63 0.50
33.88 33.00 33.80 36.63 32.78
0.34 0.76 0.42 1.11 0.64
33.05 32.50 33.00 34.60 32.68
1.25 1.40 1.10 0.40 1.43
32.75 31.77 40.47 35.37 33.88
0.59 0.52 0.56 1.05 0.16
33.77 39.47 42.80 33.90 37.80
0.58 6.14 1.14 1.47 0.23
Variable Weight
(kg)
(g/dl) range: 8- 151
2 3 4-5 Packed cell volume [normal range: 30-451
(%)
Mean corpuscular volume (p’) [normal range: 39-551
Mean corpuscular hemoglobin concentration (%) [normal range: 30-361
Pre 2 3 4-5 Pre 2 3 4-5 Pre 2 3 4-5
erythroid cells in the highest dose groups appeared disproportionately high, suggesting an abnormality in the maturation of the myeloid and erythoid cell series. Cats that did not develop anemia but in which macrocytes were observed showed an increase in the MCV but less evidence of anisocytosis in later peripheral blood smears. The bone marrow smears from these cats showed normal myeloid to erythroid ratios, and all cell lines appeared to be maturing normally.
The cats were normal at necropsy except for the pallor noted on physical examination in the two higher dose groups. Organ weights were within normal limits (data not shown). Microscopic lesions were confined to the spleen, lymph nodes, and bone marrow. All cats had reactive lymphoid tissue in spleen and lymph nodes. Bone marrow hypercellularity was prominent in the cats receiving 230 mg/kg zidovudine and was also present in one control cat. This was characterized by
ZIDOVUDINE
TOXICITY TABLE
769
IN CATS
2b
DESCRIPTIVE STATISTICS: LEUKOCYTE MEASURES Zidovudine dose (mg/kg/day) (Sample size): Variable
Week
White blood cell (X 103) [normal range: 5.5-19.51
Neutrophils (X 103) [normal range: 2.5-12.51
Lymphocytes (X 103) [normal range: 1.5-7.01
Eosinophils (X 103) [normal range: O-1.51
Monocytes (X 103) [normal range: O-8.51
0 (?I = 2)
7.5 (n = 3)
Mean
SEM
Pre
19.18
2.58
15.58
1 2 3 4-5
25.20 16.15 16.80 17.90
6.80 2.15 2.00 I .45
23.20 14.13 I 1.43 17.42
Pre
9.73 12.91 7.09 7.14 9.09
0.93 6.29 0.23 0.19 1.94
5.48 10.16 6.55 5.92 8.07
7.73 10.13 7.63 7.49 7.02
I .84 0.75 1.33 1.72 1.13
8.56 10.38 6.32 4.75 7.40
1 2 3 4-5
Pre I 2 3 4-5
Mean
SEM
15(n=2)
30 (n = 3)
Mean
SEM
2.10
28.10
10.25
5.87 2.52 4.37 2.74
30.05 21.00 14.75 19.33
4.45 5.60 2.05 4.52
0.92 2.91 1.56 3.11 1.79
8.78 8.66 6.14 4.36 7.33
1.38 1.07 0.02 0.68 1.26
1.34 1.65 1.30
16.80 19.29 12.89 8.98 10.04
1.11 0.84
Mean
60 (n = 3)
SEM
Mean
SEM
10.78
3.13
13.13 11.33 12.60 12.22
2.05 3.52 3.36 2.30
16.67 12.03 7.60 11.63 21.65
0.93 1.83 0.26 0.52 4.85
3.98 3.51 4.31 4.59 4.57
0.41 0.92 1.07 1.54 1.17
8.07 4.70 3.56 4.12 7.10
0.65 1.52 0.35 0.65 1.82
8.32 5.21 5.19 1.10 3.08
5.51 8.30 5.13 6.44 6.36
2.32 1.81 1.74 1.12 1.38
6.87 5.83 3.14 6.46 12.28
0.74 1.62 0.13 0.99 4.31
Pre
1.29
0.13
1.16
0.32
1 2 3 4-5
1.74 1.06 1.79 1.10
0.46 0.22 0.46 0.10
2.56 1.09 0.28 1.24
1.23 0.57 0.14 0.09
I .92 1.68 1.57 0.99 1.77
0.76 0.39 0.03 0.35 0.69
0.93 1.03 1.32 1.29 0.87
0.40 0.33 0.70 0.64 0.30
1.20 1.04 0.66 0.82 1.44
0.53 0.29 0.28 0.20 0.46
Pre
0.18 0.41 0.37 0.30 0.30
0.07 0.23 0.37 0.30 0.11
0.37 0.73 0.09 0.42 0.60
0.13 0.32 0.06 0.29 0.27
0.36 0.43 0.13 0.25 0.39
0.36 0.08 0.13 0.25 0.27
0.29 0.28 0.44 0.22 0.34
0.15 0.06 0.39 0.11 0.12
0.40 0.46 0.24 0.19 0.69
0.12 0.35 0.13 0.04 0.34
I 2 3 4-5
increased numbers of immature cells of both the myeloid and the erythroid series. Splenic
extramedullary hematopoiesis was observed in all cats receiving > 15 mg/kg zidovudine. This change was most marked in cats receiving a30 mg/kg of zidovudine. Two of the cats with the greatest changes were also anemic and neutropenic. All clinical chemistry values were within normal range. Plasma from zidovudine-treated cats was measured weekly for virus antigen (gs antigen) by ELISA and for antibodies to FOCMA by indirect immunofluorescence on FL74 cells which are productively infected with FeLV. Four of 13 cats were positive and 9 were negative for FeLV antigen at the begin-
ning of the study; this pattern did not change for the duration of treatment (Table 4). Although data are not presented for the measurements between the sample taken prior to zidovudine treatment and the necropsy sample, the ELISA titers for gs antigen did not change appreciably in treated or control cats at any time point. All of the cats were positive for FOCMA antibodies at the beginning of the zidovudine study (Table 4). Interestingly, 4 of 6 cats treated with high doses of zidovudine (330 mg/kg/day) had an increased FOCMA antibody titer when pretreatment samples were compared with necropsy samples. In these cats, increased FOCMA antibody titers were evident by 2 to 3 weeks after
770
HASCHEK
ET AL.
TABLE 3 RESULTS OF REPEATED-MEASURES REGRESSIONANALYSIS
Model
Partial regression coefficient
Unique contribution toR*
t
P
Baseline RBC Dose Days Dose X days
B/S B/S W/S W/S
0.498 -0.050 -0.008 -0.00 I
0.042 0.395 0.003 0.094
2.541 -7.741 -0.43 1 -3.562
ns
AND
APPLIED
Zidovudine
TOXICOLOGY
14,764-775
(
1990)
Toxicity to Cats Infected with Feline Leukemia Virus
WANDA M. HASCHEK, RONALD M. WEIGEL, GAIL SCHERBA, MA CRISTINA DEVERA, RHONDA FEINMEHL, PHILIP SOLTER, MARY B. TOMPKINS,’ AND WAYNE A. F. TOMPKINS’ Department yf’Puthobiology, College of L’eterinary Medicine, University qflllinois,
Urbana, Illinois 61801
Received July 27, 1989; accepted December 18, 1989
Zidovudine
W. M.. M. B., AND TOMPKINS. W. A. F. (1990). Fundam. Appl. To.xicoi. 14, 764-715. Feline leukemia virus (FeLV) infection of cats is a model for the acquired immunodeficiency syndrome in humans. The toxicity of zidovudine was evaluated in SPF cats experimentally infected with FeLV. At initiation of the zidovudine study, all cats were antibody positive for FeLV antigens but clinically asymptomatic. Four cats were also viremic. Thirteen, 6- to IO-month-old cats were divided into five dosage groups and given zidovudine po at 0. 7.5. 15, 30. or 60 mg/kg daily in three equally divided doses for 32 to 34 days. Titers of circulating virus antigen remained constant: however, three of six cats receiving the higher doses of zidovudine (a30 mg/kg) showed an increase in antibody titers to FeLV. Administration of zidovudine resulted in a progressive anemia, dependent upon dose and time. Macrocytes were observed prior to the development of anemia and were also found in several nonanemic cats. Repeated measures regression analyses indicated that an increased dose of zidovudine was associated with decreased packed cell volume, red blood cell count, and hemoglobin. As determined from the packed cell volume, the analyses indicate that anemia is induced only by the two highest doses ofzidovudine. The regression model indicates that daily doses of 60 and 30 mg/kg are expected to induce anemia by Day 4 and Day 13. respectively. Progressive absolute neutropenia was observed in the a30 mg/kg groups. Histopathologic lesions consisted of marked bone marrow hypercellularity in cats given >30 mg/kg zidovudine and splenic extramedullary hematopoiesis in cats given > 15 mg/kg. Thus, oral toxicity of zidovudine in the cat is manifested by a dose-related anemia and neutropenia as observed in humans. 01 1990 Society ofToxicology. WEIGEL,
R. M.,
Toxicity to Cats Infected with Feline Leukemia Virus. SCHERBA,
G.. DEVERA.
M. C., FEINMEHL,
Zidovudine, also known as azidothymidine (3’-azido-3’-deoxythymidine) or Retrovir, a thymidine analog, is widely used as a treatment for AIDS or AIDS-related complex in humans. Zidovudine is phosphorylated to AZT triphosphate which competes with cellular nucleotides for incorporation into viral DNA by the viral reverse transcriptase (Mitsuya et al., 1985). In this manner, zidovudine inhibits the replication of human immunodeficiency virus (HIV) in vitro (Mitsuya et al., ’ Present address: North Carolina State University, College of Veterinary Medicine, 4700 Hillsborough Street, Raleigh, NC 27606. 0272-0590/90 $3.00 Copyright 0 1990 by the Society of Toxicology. All rights of reproduction in any form reserved.
764
R.. SOLTER,
HASCHEK, P., TOMPKINS,
1985) and has been shown to decrease viremia, lengthen survival, and decrease frequency of opportunistic infections in AIDS patients (Fischl et al., 1987; Yarchoan et al., 1986; Yarchoan and Broder, 1987). Although significant clinical benefits of zidovudine therapy have been documented, hematologic toxicity consisting of macrocytosis, anemia, and neutropenia has also been observed (Dournon et al., 1988; Fischl et al., 1987; Gill et al., 1987; Richman et al., 1987). Anemia appears to be due to red cell hypoplasia or aplasia based on bone marrow examination of anemic patients that showed red cell aplasia, erythroid hypoplasia, or meg-
ZIDOVUDINE
TOXICITY
aloblastic erythropoiesis (Walker et al., 1988). Marked zidovudine toxicity for bone marrow hematopoietic precursor cells is also evidenced by inhibition of granulocyte-macrophage CFU and erythroid burst-forming cells in vitro (Sommadossi and Carlisle, 1987). Although zidovudine has been used to treat retrovirus infection in other species used as models for human AIDS (Ruprecht et al., 1986; Tavares et al., 1987), there is little information on toxicity of the drug in species other than man. Recent studies indicate that cynomolgus monkeys are quite susceptible to zidovudine-induced hematologic toxicity, while rats are resistant (Ayers, 1988). Cats infected with feline leukemia virus (FeLV) have been studied as a possible model for HIV infection, since they develop similar immunodeficiency and neoplastic syndromes (Hoover et al., 1989; Tavares et al., 1989). Among the immunopathological similarities to HIV infection, FeLV-infected cats show a deficiency in T helper (TH, ) cell production of IL-2 (Tompkins et al., 1989a) and reduced lymphokine-activated killer (LAK) cell function (Tompkins et al., 1989b). Although feline immunodeficiency virus (FIV) is much more closely related to HIV than is FeLV, it has a long and variable latent period, and does not produce consistent clinical signs and lesions in infected cats (Pedersen et al., 1989). Therefore, at the present time, FIV is not a suitable model for evaluation of potential therapeutic agents. Data obtained from this study however will also be applicable to FIV-infected cats as the model is further developed. Zidovudine has been evaluated in FeLVinfected cats and shown to have therapeutic and preventive potential (Tavares et al., 1987); however, its hematologic toxicity has not been evaluated. As part of a large program to evaluate promising biological response modifiers and antiviral chemicals for efficacy and safety, the subacute oral toxicity of zidovudine was examined in cats experimentally infected with the immunosuppressive Rickard strain of FeLV. The study was designed to determine the dose at which tox-
IN
765
CATS
icity was manifested, the target organs of toxicity, and the highest nontoxic dose for future therapeutic studies. Because these cats were infected with FeLV, virological and immunological parameters were also measured. Hematologic toxicity was observed and similarities to the human condition are discussed. METHODS Experimental design. Thirteen, specific pathogen-free (SPF), Mill Hill strain cats (Liberty Lab, Liberty, NJ) were experimentally infected intravenously with approximately 5 X lo4 focus forming units of the Rickard strain of FeLV (obtained from Dr. L. Mathes, Ohio State University, Columbus) at 6 to 8 weeks of age. All cats developed subclinical infections with circulating antibodies to FeLV antigen and four were viremic when the zidovudine study was initiated. At 6 to 10 months of age, the cats were divided into five dosage groups (Table 1): 0 (n = 2). 7.5 (n = 3) I5 (n = 2), 30 (n = 3) and 60 (n = 3) mg/kg/ day zidovudine. Zidovudine was administered in three equally divided doses in moist cat food for 32 to 34 days, with doses adjusted weekly according to body weight. Cats were weighed and evaluated daily by physical examination. Blood samples (5 ml) were drawn weekly from the jugular vein, beginning 1 week prior to the start of zidovudine treatment. for hematologic and clinical chemistry evaluations and for viral and immunological measurements. At the end of the study all cats were euthanized with T-6 I euthanasia solution (Taylor Pharmacal Co., Decatur, IL) and complete necropsies performed. Zidovudine. Zidovudine (Retrovir) or 3’-azido-3’deoxythymidine (AZT) was generously provided by Burroughs Welcome Co. (Research Triangle, Park, NC). Hematology. The peripheral blood compartment was evaluated as follows. Percentage packed cell volume (PCV) was determined with an Autocrit II centrifuge (Fisher Scientific Co., Fairlawn, NJ). Total erythrocyte (RBC) and leukocyte (WBC) counts were obtained according to standard techniques (Jain, 1986). Hemoglobin (Hb) concentration was assayed by the cyanmethemoglobin method (Sigma Chemical Co., St. Louis, MO). Mean corpuscular volume (MCV) and mean corpuscular hemoglobin concentration (MCHC) were calculated as MCV = IO X (PCV/RBC) and MCHC = 100 X (Hb/ PCV). Leukocyte differentials were performed to determine absolute numbers of neutrophils, lymphocytes, eosinophils, and monocytes. Bone marrow smears made at necropsy were examined to interpret more precisely the peripheral blood data. Serum biochemistry. The concentrations of serum total protein. albumin, globulin, creatinine, blood urea nitrogen, glucose, cholesterol, phosphorus, calcium, sodium, chloride, potassium. and total bilirubin as well as the activities of alkaline phosphatase, aspartate amino-
766
HASCHEK
transferase, alanine aminotransferase, and lactate dehydrogenase were determined with an autoanalyzer (Hycel Super Seventeen Autoanalyzer, Houston, TX). Patho1og.v. At necropsy, the following organs were weighed: brain, adrenals, heart, kidneys, liver, spleen, thymus, testes, and epididymus. Sections from the above. as well as from thyroid and parathyroid. lymph nodes, bone marrow, lung, pancreas, ovary, stomach. and intestines, were fixed in 10% buffered neutral formalin; except for testes and ovaries which were fixed in Bouin’s solution. Tissue sections were embedded in paraffin, sectioned at 4 to 6 pm. and stained with hematoxylin and eosin. All tissues were examined from control and 60 mg/kg zidovudine groups while only bone marrow, spleen, thymus, and lymph nodes were examined in the remaining groups. Impression smears were made from femoral bone marrow, fixed in 70% alcohol, and stained with Wright-Giemsa. Bone marrow was evaluated cytologically and the ratio of myeloid to erythroid cells was determined. Assayfir FeLVantigen. FeLV viremia was monitored by testing the plasma fraction of blood for FeLV group specific (gs) antigen by an ELISA system (TechAmerica). Assay for FOCMA antibodies. Serums were assayed for antibodies to the FeLV-induced feline oncorna virus cell membrane-associated antigen (FOCMA) by indirect immunofluorescence using a FOCMA-positive FL74 lymphoma cell line (obtained from Dr. W. Hardy, Sloan Kettering Institute, New York) as the target. Briefly, I X lo6 viable FL74 cells were incubated with 100 ~1 of serial dilutions (1:4- 1:64) of plasma from FeLV-infected cats for 20 min at 4°C. After incubation, the cells were washed in ice-cold PBS plus 2% FBS. fixed for 5 min with 1% paraformaldehyde. washed, incubated in 100 pg of fluorescein-conjugated goat anti-feline IgG (heavy and light chain) for 20 min, and then washed three times in PBS/FBS. The cell pellets were suspended in 50 ~1 PBS: glycerol (1:9) and mounted under a glass coverslip for viewing with a fluorescence microscope. Positive reactions are revealed as a beaded fluorescence on the surface of the cell. Statisticul methods. The effect of zidovudine on hematologic measures over the course of therapy was evaluated using a repeated measures multiple regression model (Cohen and Cohen. 1983). For each hematologic measure analyzed, two analyses were conducted. In the between subjects variance analysis, the dependent variable was the mean of the two hematologic measures taken during the 5th (i.e., final) week oftherapy, and the independent variable was zidovudine dose with baseline level of the hematologic variable (which was the mean of the two pretreatment measurements) as a covariate. The average of two measures was utilized to increase reliability (Ederer, 1972). In the within subjects variance analysis, the dependent variable was the value of the hematologic measurement at each of the five post-treatment times, and the independent variables were pretreatment level of the outcome variable, days since onset of zido-
ET AL. vudine therapy, and a dose by time interaction term. An u level of 0.05 was chosen for all statistical analyses. In both the between subjects and within subjects regression analyses, presence or absence of viremia was also added to the model as an independent variable in supplementary analyses. However, no effect of viremia was apparent, so this variable was excluded from the regression models. The relationship between zidovudine dose and several dichotomous outcomes were investigated post hoc in 2 X 2 contingency table analyses. In each case, zidovudine dose was dichotomized as “~30” (0, 7.5, and 15 mg/kg/ day doses; total n = 7) versus “>30” (30 and 60 mg/kg/ day; total n = 6). The departure from random association was examined statistically using the Fisher Exact Test.
RESULTS All cats remained alert and clinically healthy throughout the study, with exception of pale mucous membranes and a slow capillary refill time that developed in cats given the two highest doses of zidovudine. Table 1 summarizes the observed hematological changes. Tables 2a and 2b present descriptive statistics on mean values of body weight and hematologic measures as a function of dose and time. It is apparent that body weight was not affected by zidovudine (Table 2a). Administration of zidovudine did result in a progressive anemia, using the criterion of a decrease in PCV to less than 30% (Table 2a). All six cats in the highest dose groups (a30 mg/kg/day) developed anemia, in contrast to the seven cats in the lower dose groups (~30 mg/kg/day) (one-tailed p = 0.0047, Fisher Exact Test). Direct microscopic examination of the peripheral blood smears characterized the anemia as normochromic with anisocytosis distinguished by many erythrocytes measuring up to 8 pm in diameter (macrocytes). Normal erythrocyte diameter in the cat is 5.5 to 6.3 pm. All six cats in the highest dose groups (230 mg/kg/day), compared to three of the seven cats in the lower dose groups (~30 mg/kg/day), developed macrocytes (one-tailed p = 0.049, Fisher Exact Test). Macrocytes were observed in all six cats of the highest dose groups prior to the development of anemia. Reticulocyte counts were not performed; however, the macro-
ZIDOVUDINE
TOXICITY
767
IN CATS
TABLE I AGE, SEX, VIREMIC
Daily zidovudine dose’ (m&kg) 0 1.5 15 30 60
STATUS, AND HEMATOLOGIC CHANGES
OF CATS TREATED WITH ZILIOVUDINE
ID6
Cat age (months)
Sex
Macrocytes observed
TWI NT3 TY2 oc4 NT5 TV2 UB2 TX2 NZl NT4 UB3 NU4 NZ6
6 6 10 10 10 10 10 6 6 6 6 6 6
M M M M F M M M M F M M F
+ + + fC +’ +’ +c +< +c
Anemia + + + + + +
Neutropenia + + + + -
a Zidovudine was given orally in three equally divided doses for 32 to 34 days. b Identification number. Cats with same first letter belong to same litter. ’ Macrocytes observed prior to development of anemia.
cytes did not appear to be immature cells as they were not basophilic. Regression analyses confirmed that 230 mg/kg doses of zidovudine caused significant decreases in three hematologic variables: PCV, RBC, Hb, and increased MCHC (Table 3). Although MCV appeared to increase over time in the higher dose groups and macrocytes were noted on peripheral blood smears, in the regression analysis, zidovudine did not significantly effect the MCV. Data from repeated measures regression analysis (Table 3) show that increasing dose of zidovudine during the 5th week of treatment was associated with decreased RBC, decreased PCV, decreased Hb, and increased MCHC and MCH. The increase in MCHC and MCH with increasing zidovudine dose reflects a more rapid decline in PCV relative to Hb with increasing dose. In the absence of lipemia, or large numbers of Heinz bodies on peripheral blood examination, this is suggestive of free hemoglobin in the plasma. There were also significant dose by time interaction effects for RBC, PCV, and Hb, as is apparent from Figs. 1, 2, and 3, plotting the
regression lines for each dose group over time. In addition, the decrease in these red blood cell measures from baseline levels occurs more rapidly with increasing dose. Using a criteria of PCV < 30% as a definition of anemia, the regression analyses suggest an induction of anemia by Day 4 with a daily zidovudine dose of 60 mg/kg and by Day 13 with 30 w/k. No significant association of zidovudine dose with WBC hematologic measures was apparent from the regression analyses. However, a decrease in absolute numbers of neutrophils without a left shift was observed at high doses (Table 2b). Four of the six cats in the two highest dosage groups (230 mg/kg/ day) developed an absolute neutropenia, using the criterion of a decrease in neutrophil numbers to less than 3000. None of the seven cats in the three lower dose groups developed neutropenia (one-tailedp = 0.02 1; Fisher Exact Test). The thrombocytic series in all cats appeared normal. Although evaluation of bone marrow smears was complicated by technical problems, the numbers of immature myeloid and
768
HASCHEK TABLE DESCRIPTIVE
STATISTICS:
BODY WEIGHT
ET AL. 2a AND ERYTHROCYTE Zidovudine
(Sample
size):
0 (fl = 2)
7.5 (n = 3)
PARAMETERS
dose (mg/kg/day) 15(n=2)
30 (n = 3)
60 (n = 3)
Week
Mean
SEM
Mean
SEM
Mean
SEM
Mean
SEM
Mean
SEM
Pre 2 3 4-5
3.95 3.90 3.90 4.00 3.90
0.95 0.70 0.90 1.oo 0.90
3.20 3.13 3.20 3.27 3.40
0.42 0.52 0.53 0.64 0.58
3.75 3.60 4.00 3.90 4.10
0.15 0.20 0.20 0.10 0.10
3.13 3.27 3.20 3.27 3.13
0.43 0.29 0.35 0.35 0.33
2.88 3.00 2.93 3.13 3.00
0.32 0.31 0.37 0.41 0.31
Red blood cell (X lob/d) [normal range: 5-101
Pre 1 2 3 4-5
7.98 8.00 7.87 8.39 7.53
0.70 0.69 1.14 0.17 0.39
6.98 6.85 5.96 5.07 6.67
0.21 0.14 0.18 0.45 0.26
7.54 7.52 7.48 6.89 6.64
0.18 0.03 0.05 0.44 0.08
6.42 6.77 4.63 5.09 4.58
0.72 0.76 0.47 0.88 0.42
6.50 5.91 4.12 5.54 3.85
0.12 1.12 0.17 0.60 0.13
Hemoglobin [normal
Pre
1 1.45 11.75 13.10 12.60 1 1.63
1.45 1.15 0.80 0.40 0.03
11.43 11.13 10.10 9.93 11.83
0.77 0.52 0.64 1.42 0.94
11.60 11.55 12.05 11.95 II.20
1.45 1.55 1.45 1.95 1.39
10.20 10.73 9.40 9.20 8.68
0.44 1.19 0.25 0.70 0.65
10.08 10.50 8.03 8.37 6.70
0.32 0.85 0.20 0.87 0.40
34.15 36.00 38.30 38.70 35.00
4.50 2.80 6.80 0.50 0.60
33.80 33.80 29.87 29.97 36.07
2.53 2.08 1.54 3.39 2.18
34.98 35.35 36.40 34.45 34.10
3.07 3.25 3.20 5.25 3.70
31.23 33.90 23.23 26.16 25.70
1.19 4.31 0.58 2.78 1.99
30.05 28.20 18.80 24.67 18.37
0.84 5.31 0.86 2.26 0.87
43.09 45.00 48.40 46.15 46.63
2.03 0.40 1.70 1.55 1.63
48.27 49.27 50.17 52.77 53.83
2.36 2.23 2.43 2.19 2.94
46.53 47.00 48.70 52.50 51.35
5.18 4.50 4.60 3.99 4.65
49.58 50.37 5 1.00 52.50 57.13
4.14 4.40 3.91 3.98 7.18
46.23 47.70 45.60 44.63 47.87
0.68 0.31 0.32 0.72 0.75
33.58 32.55 34.90 32.53 33.25
0.22 0.65 4.10 0.63 0.50
33.88 33.00 33.80 36.63 32.78
0.34 0.76 0.42 1.11 0.64
33.05 32.50 33.00 34.60 32.68
1.25 1.40 1.10 0.40 1.43
32.75 31.77 40.47 35.37 33.88
0.59 0.52 0.56 1.05 0.16
33.77 39.47 42.80 33.90 37.80
0.58 6.14 1.14 1.47 0.23
Variable Weight
(kg)
(g/dl) range: 8- 151
2 3 4-5 Packed cell volume [normal range: 30-451
(%)
Mean corpuscular volume (p’) [normal range: 39-551
Mean corpuscular hemoglobin concentration (%) [normal range: 30-361
Pre 2 3 4-5 Pre 2 3 4-5 Pre 2 3 4-5
erythroid cells in the highest dose groups appeared disproportionately high, suggesting an abnormality in the maturation of the myeloid and erythoid cell series. Cats that did not develop anemia but in which macrocytes were observed showed an increase in the MCV but less evidence of anisocytosis in later peripheral blood smears. The bone marrow smears from these cats showed normal myeloid to erythroid ratios, and all cell lines appeared to be maturing normally.
The cats were normal at necropsy except for the pallor noted on physical examination in the two higher dose groups. Organ weights were within normal limits (data not shown). Microscopic lesions were confined to the spleen, lymph nodes, and bone marrow. All cats had reactive lymphoid tissue in spleen and lymph nodes. Bone marrow hypercellularity was prominent in the cats receiving 230 mg/kg zidovudine and was also present in one control cat. This was characterized by
ZIDOVUDINE
TOXICITY TABLE
769
IN CATS
2b
DESCRIPTIVE STATISTICS: LEUKOCYTE MEASURES Zidovudine dose (mg/kg/day) (Sample size): Variable
Week
White blood cell (X 103) [normal range: 5.5-19.51
Neutrophils (X 103) [normal range: 2.5-12.51
Lymphocytes (X 103) [normal range: 1.5-7.01
Eosinophils (X 103) [normal range: O-1.51
Monocytes (X 103) [normal range: O-8.51
0 (?I = 2)
7.5 (n = 3)
Mean
SEM
Pre
19.18
2.58
15.58
1 2 3 4-5
25.20 16.15 16.80 17.90
6.80 2.15 2.00 I .45
23.20 14.13 I 1.43 17.42
Pre
9.73 12.91 7.09 7.14 9.09
0.93 6.29 0.23 0.19 1.94
5.48 10.16 6.55 5.92 8.07
7.73 10.13 7.63 7.49 7.02
I .84 0.75 1.33 1.72 1.13
8.56 10.38 6.32 4.75 7.40
1 2 3 4-5
Pre I 2 3 4-5
Mean
SEM
15(n=2)
30 (n = 3)
Mean
SEM
2.10
28.10
10.25
5.87 2.52 4.37 2.74
30.05 21.00 14.75 19.33
4.45 5.60 2.05 4.52
0.92 2.91 1.56 3.11 1.79
8.78 8.66 6.14 4.36 7.33
1.38 1.07 0.02 0.68 1.26
1.34 1.65 1.30
16.80 19.29 12.89 8.98 10.04
1.11 0.84
Mean
60 (n = 3)
SEM
Mean
SEM
10.78
3.13
13.13 11.33 12.60 12.22
2.05 3.52 3.36 2.30
16.67 12.03 7.60 11.63 21.65
0.93 1.83 0.26 0.52 4.85
3.98 3.51 4.31 4.59 4.57
0.41 0.92 1.07 1.54 1.17
8.07 4.70 3.56 4.12 7.10
0.65 1.52 0.35 0.65 1.82
8.32 5.21 5.19 1.10 3.08
5.51 8.30 5.13 6.44 6.36
2.32 1.81 1.74 1.12 1.38
6.87 5.83 3.14 6.46 12.28
0.74 1.62 0.13 0.99 4.31
Pre
1.29
0.13
1.16
0.32
1 2 3 4-5
1.74 1.06 1.79 1.10
0.46 0.22 0.46 0.10
2.56 1.09 0.28 1.24
1.23 0.57 0.14 0.09
I .92 1.68 1.57 0.99 1.77
0.76 0.39 0.03 0.35 0.69
0.93 1.03 1.32 1.29 0.87
0.40 0.33 0.70 0.64 0.30
1.20 1.04 0.66 0.82 1.44
0.53 0.29 0.28 0.20 0.46
Pre
0.18 0.41 0.37 0.30 0.30
0.07 0.23 0.37 0.30 0.11
0.37 0.73 0.09 0.42 0.60
0.13 0.32 0.06 0.29 0.27
0.36 0.43 0.13 0.25 0.39
0.36 0.08 0.13 0.25 0.27
0.29 0.28 0.44 0.22 0.34
0.15 0.06 0.39 0.11 0.12
0.40 0.46 0.24 0.19 0.69
0.12 0.35 0.13 0.04 0.34
I 2 3 4-5
increased numbers of immature cells of both the myeloid and the erythroid series. Splenic
extramedullary hematopoiesis was observed in all cats receiving > 15 mg/kg zidovudine. This change was most marked in cats receiving a30 mg/kg of zidovudine. Two of the cats with the greatest changes were also anemic and neutropenic. All clinical chemistry values were within normal range. Plasma from zidovudine-treated cats was measured weekly for virus antigen (gs antigen) by ELISA and for antibodies to FOCMA by indirect immunofluorescence on FL74 cells which are productively infected with FeLV. Four of 13 cats were positive and 9 were negative for FeLV antigen at the begin-
ning of the study; this pattern did not change for the duration of treatment (Table 4). Although data are not presented for the measurements between the sample taken prior to zidovudine treatment and the necropsy sample, the ELISA titers for gs antigen did not change appreciably in treated or control cats at any time point. All of the cats were positive for FOCMA antibodies at the beginning of the zidovudine study (Table 4). Interestingly, 4 of 6 cats treated with high doses of zidovudine (330 mg/kg/day) had an increased FOCMA antibody titer when pretreatment samples were compared with necropsy samples. In these cats, increased FOCMA antibody titers were evident by 2 to 3 weeks after
770
HASCHEK
ET AL.
TABLE 3 RESULTS OF REPEATED-MEASURES REGRESSIONANALYSIS
Model
Partial regression coefficient
Unique contribution toR*
t
P
Baseline RBC Dose Days Dose X days
B/S B/S W/S W/S
0.498 -0.050 -0.008 -0.00 I
0.042 0.395 0.003 0.094
2.541 -7.741 -0.43 1 -3.562
ns
