Vet. Pathol.

20: 687-697 ( I 983)

Erythrocyte Macrocytosis in Feline Leukemia Virus Associated

Anemia

M. G. WEISERand G. J. KOCIBA

Department of Veterinary Pathobiology, Ohio State University, Columbus, Ohio

Abstract. Using erythrocyte volume distribution histograms (erythrograms), erythrocyte

macrocytosis and anisocytosis were quantitated in 139 cats tested for feline leukemia virus
group-specific antigen. Feline leukemia virus-negative cats with non-regenerative anemia or
normal packed cell volumes had normal mean corpuscular volume values. Uninfected cats
with regenerative anemia had prominent, significantly increased macrocytosis and anisocy-
tosis ( p < 0.01). Ninety percent of 62 feline leukemia virus-positive cats had altered
erythrograms. Thirty-three feline leukemia virus-positive cats with non-regenerative anemia
*
had marked macrocytosis. Their mean corpuscular volume values (mean 60 fl 2 fl standard
error, reference range of 37-49 fl) were significantly greater than those of feline leukemia
virus-negative cats except for those with regenerative anemia. Feline leukemia virus-positive,
non-anemic cats had significantly increased mean corpuscular volume values of intermediate
magnitude. Nine adult cats experimentally infected with feline leukemia virus developed
non-regenerative anemia with significant increases in mean corpuscular volume and aniso-
cytosis. However, the macrocytosis observed in these cats was considerably less than in
naturally occurring feline leukemia virus-positive cats with non-regenerative anemia. These
observations indicate there are events in the pathogenesis of feline leukemia virus-associated
anemia other than simple erythroid hypoplasia. We suggest that hemolysis and erythrocyte
regeneration occur before erythroid hypoplasia and may partially account for macrocytosis
observed in the face of non-regenerative anemia.

Feline leukemia virus may cause anemia in both natural [4, 5, 161 and experi-
mental [8, 141 infection. In most studies the anemia has been characterized as
normocytic, normochromic, and non-regenerative on the basis of low reticulocyte
counts and bone marrow erythroid hypoplasia [4,5,8, 14, 161. Although occasional
macrocytic erythrocytes were observed in some of the cats, no mean corpuscular
volume values were determined [4, 5 , 161. The anemia was characterized as initially
regenerative and subsequently non-regenerative in a few instances [4, 5 , 161. In
contrast, one group of experimentally infected kittens had transient regenerative,
macrocytic, hemolytic anemia prior to developing lymphosarcoma [ 141. Non-
regenerative anemia with macrocytosis and an increased mean corpuscular volume
value was observed in cats with myeloproliferative disorders [7, 15, 23, 271 and
687
688 Weiser and Kociba

lymphosarcoma [ 141. These features were attributed to abnormal erythrogenesis

associated with virus infection and neoplastic marrow disease [27].
By direct measurement of mean corpuscular volume [29], we have observed a
high frequency of increased mean corpuscular volume values in feline leukemia
virus-infected cats. Erythrograms are even more sensitive than mean corpuscular
volume values for detecting disturbances of erythrocyte size [ 1, 2, 201. The purpose
of this study was to use erythrograms to characterize erythrocyte changes in feline
leukemia virus-infected cats.

Materials and Methods

Blood samples obtained were from 139 cats examined for a wide variety of illnesses at the
Ohio State University Veterinary Teaching Hospital. An immunofluorescence assay for
feline leukemia virus group-specific antigen [ I01 was done as part of the diagnostic evaluation
and this was the criterion used for selecting the cats. At the time this test was done, one
hemogram and erythrogram were done on blood from each cat. A few cats had multiple
hematologic evaluations before and/or after this time. Hemograms were done on an
automated, multi-channel cell counting system (Coulter Counter S-Senior, Coulter Electron-
ics, Hialeah, FL). This instrument is modified with an increased red cell bath aperture
current to provide a lower threshold of about ten femtoliters (fl). This system provides feline
mean corpuscular volume values comparable to those obtained on a single channel electronic
counter having a threshold range of 7 to 115 fl [29]. Erythrograms were done as described
previously [29]. An index of anisocytosis value, defined as the volume difference in fl
between cells occupying the two 50% maximum frequency channels on the erythrogram
[29], was determined for each cat. Absolute aggregate reticulocyte counts were done on most
cats with anemia [28]. Anemia was defined by a packed cell volume value of less than 25
and regenerative anemia was defined by an aggregate reticulocyte count of greater than
60,000//*1.
Based on the hematologic findings at the time the immunofluorescence assay was done,
the above cats were subdivided into six groups for purposes of analysis. Feline leukemia
virus-positive cats were divided as follows: 1) non-anemic, 2) non-regenerative anemia, and
3) regenerative anemia. Feline leukemia virus-negative cats also were divided: 4) non-
anemic, 5) non-regenerative anemia, and 6) regenerative anemia.
A group of 75 clinically normal adult cats [29] served as a reference group. Using analysis
of variance and Hochberg’s GT-2 test for unplanned comparisons among multiple means
[25], these groups were evaluated for significant differences in mean corpuscular volume
and index of anisocytosis values.
Using methyl-prednisolone pretreatment, nine adult, specific pathogen free cats [2 I ] were
infected with the Kawakami-Theilen strain of feline leukemia virus as described previously
[ 121. Four cats similarly treated only with methyl-prednisolone served as controls. Immu-
nofluorescence tests for feline leukemia virus group-specific antigen were done on blood
films from all cats-before and each two weeks after inoculation. Hemograms, reticulocyte
counts, and erythrograms were done at day 0 and once between days 30 and 33 after
inoculation. After developing severe anemia, each infected cat was killed between days 50
and 80. The hematologic determinations were done again at this time. Control cats were
last sampled at 80 days after the start of the experiment. Using initial and final values for
each cat, a paired t-test was used to evaluate control and inoculated cats for significant
changes in packed cell volume, mean corpuscular volume, and index of anisocytosis. Using
analysis of variance and Hochberg’s GT-2 test [25], naturally occurring feline leukemia
virus-positive cats with non-regenerative anemia were compared with inoculated cats and
 Erythrocyte Macrocytosis 689

70-

60-
h

40J

GROUP: NORMAL FeLV FeLV FeLV FeLV FeLV FeLV

- - + + - +
NO NR NO NR REGEN REGEN
ANEMIA ANEMIA ANEMIA ANEMIA ANEMIA ANEMIA
NUMBER
75 61 5 17 33 11 12

control cats at days 50 to 80 for significant differences in packed cell volume, mean
corpuscular volume, and index of anisocytosis values.

Results
The mean corpuscular volume and index of anisocytosis values for the seven
groups of cats are included in figures 1 and 2, respectively. Feline leukemia virus-
positive cats with regenerative anemia or non-regenerative anemia and feline
leukemia virus-negative cats with regenerative anemia had marked macrocytosis
and anisocytosis. These groups had significantly greater mean corpuscular volume
and index of anisocytosis values than all other groups (p < 0.01). Of 33 feline
690 Weiser and Kociba

i i

GROUP: NORMAL FeLV FeLV FeLV FeLV FeLV FeLV

- - + + + -
NO NR NO NR REGEN REGEN
ANEMIA ANEMIA ANEMIA ANEMIA ANEMIA ANEMIA
NUMBER
75 61 5 17 33 12 11

leukemia virus-positive cats with non-regenerative anemia, 23 had mean corpus-

cular volume values greater than the upper reference limit of 49 fl [29] and 30 had
index of anisocytosis values greater than the upper reference limit of 15 fl [29]. In
this group, hemobartonellosis was diagnosed in four cats, Heinz bodies were
observed in four others, and no defects associated with hemolysis were observed in
the remaining 25 cats. Of five cats in this group having multiple complete blood
counts up to several weeks earlier, all had reticulocytosis at least once.
Regenerative anemia in 12 feline leukemia virus-positive cats was attributed to
hemobartonellosis in four cats, Heinz body hemolysis in two cats, hemorrhage in
one cat, and undefined hemolysis in five cats. Six of these cats had subsequent
hematologic evaluations and all developed non-regenerative anemia within three
 Erythrocyte Macrocytosis 69 1

- 0 CELL VOLUME (11)

RETICULOCYTES O---o

20 -.
o. _ _ _ -o---
-- --__
--
CELL VOLUME ( f l l TIME, days

Fig. 3: Serial erythrograms, packed cell volume values, and reticulocyte counts from a
feline leukemia virus-positive cat. The initial regenerative anemia became non-regenerative
after partial recovery.

weeks. One cat was evaluated with multiple hematologic and erythrogram deter-
minations over a 60-day period (fig. 3 ) . This cat initially had two erythrocyte
populations associated with marked reticulocytosis. As the packed cell volume
returned toward normal, there was complete repopulation by macrocytes. The
macrocyte population persisted while the cat developed non-regenerative anemia.
Regenerative anemia in 1 1 feline leukemia virus-negative cats was attributed to
experimentally induced Heinz body hemolysis in five cats [30], hemorrhage in one
cat, hemobartonellosis in one cat, and undefined hemolysis in the remaining four
cats. All cats with regenerative anemia had mean corpuscular volume and index of
anisocytosis values greater than the upper reference limit for normal cats.
The feline leukemia virus-negative cats with non-regenerative anemia were
uncommon. All five of the cats had mean corpuscular volume values in the normal
692 Weiser and Kociba

range and two had index of anisocytosis values slightly greater than the upper
reference limit.
The non-anemic feline leukemia virus-positive cats had significantly increased
mean corpuscular volume values (p < 0.05) compared to feline leukemia virus-
negative, non-anemic cats and normal cats and their index of anisocytosis values
were significantly greater than those of normal cats (p < 0.05) (figs. 1, 2). Of 17
cats in this group, nine had a mean corpuscular volume greater than the upper
reference value and 12 had an index of anisocytosis value above the upper reference
limit.
Four types of erythrograms were observed in feline leukemia virus-positive cats
and feline leukemia virus-negative cats with regenerative anemia (fig. 4). The first
pattern (fig. 4, Curve A) reflects a heterogeneous population which includes the
normal size range. This was observed in 30% of feline leukemia virus-positive cats.
Two other curve types (fig. 4, Curves B and C) reflect various degrees of blood
repopulation with a distinct subpopulation of cells having volume about twice
normal. These curves were observed in 60% of feline leukemia virus-positive cats.
Normal erythrograms occurred in 10% of virus-positive cats.
Cats experimentally infected with feline leukemia virus became positive for the
group-specific antigen two weeks after inoculation and remained positive. Control
cats remained negative for group-specific antigen of feline leukemia virus. The
sequential changes in packed cell volume, mean corpuscular volume, and index of
anisocytosis values for these cats are presented in figure 5. By days 50 to 80,
inoculated cats had significant increases in mean corpuscular volume (p < 0.001)
and index of anisocytosis values (< 0.0 1 ) and a significant decrease in packed cell
volume (p < 0.001). Control cats had no significant changes in these values. Only
two of nine inoculated cats developed a mean corpuscular volume greater than the
upper reference limit, however, all developed index of anisocytosis values above
the upper reference limit. Naturally occurring feline leukemia virus-positive cats
with non-regenerative anemia had significantly greater mean corpuscular volume
values than inoculated cats (p < 0.01) (compare figs. 1 and 5). These two groups
had no significant differences in index of anisocytosis or packed cell volume values.
No reticulocytosis was recognized in the experimentally inoculated cats.

Discussion
Macrocytosis and increased anisocytosis occurred in most feline leukemia virus-
infected cats and these findings were most prominent in those with anemia. These
findings are observed commonly in man with preleukemia [13, IS]. Of greatest
interest was the observation that feline leukemia virus-positive cats with non-
regenerative anemia had macrocytosis and anisocytosis values comparable to those
of other groups with regenerative anemia and significantly greater than those of
feline leukemia virus-negative cats with non-regenerative anemia. In our experience
and from data presented here, cats which are not anemic or have non-regenerative
anemia and have a mean corpuscular volume of greater than 50 fl have a high
 Erythrocyte Macrocytosis 693

Fig. 4: Representative types of erythrograms observed in feline leukemia virus-positive

cats and feline leukemia virus-negative cats with regenerative anemia (solid lines). Broken
lines are reference feline erythrograms. Curve A represents a heterogeneous erythrocyte
population overlapping the normal size range. Curve B and C show a distinct macrocyte
subpopulation with various degrees of repopulation.
694 Weiser and Kociba

40
T T

30
>
a
0

20

0 30-33 50-80
TIME- DAY OF EXPERl MENT
Fig. 5: Packed cell volume, mean corpuscular volume, and index of anisocytosis values
for nine cats experimentally infected with feline leukemia virus (clear bars) and four control
cats (shaded bars). Vertical bars represent k 1 standard error of the mean. Infected cats had
significant increases in mean corpuscular volume (p < 0.001) and index of anisocytosis (p
< 0.01) and decrease in packed cell volume (p < 0.001). No significant changes occurred in
the control cats. PCV = packed cell volume; MCV = mean corpuscular volume; IA = index
of anisocytosis.

probability of being feline leukemia virus-positive. An alternative would be post-

regenerative macrocytosis [30]. This is in contrast to previous reports that the non-
regenerative anemia of feline leukemia virus-positive cats is normocytic [4, 5 , 8,
14, 161. Since mean corpuscular volume was not measured in these studies, it
appears the anemia was described as normocytic based on absence of reticulocy-
tosis. However, the description of some macrocytic erythrocytes on blood films of
some cats with non-regenerative anemia suggests that macrocytosis may have been
 Erythrocyte Macrocytosis 695

present [4, 5 , 161. Cats with myeloproliferative disorders, in which mean corpus-
cular volume was measured, had macrocytosis comparable to our observation [7,
15. 23, 271.
The development of macrocytosis in feline leukemia virus-positive cats suggests
that a regenerative response occurs before erythroid hypo-proliferation. Compatible
with this view is the observation in this study and others [4, 5 , 161 of feline leukemia
virus-positive cats with regenerative anemia converting to non-regenerative anemia.
It recently has been demonstrated in cats with hemolysis that macrocytosis persisted
for up to 40 days after packed cell volume and reticulocyte counts returned to
normal [30]. The sequence of events in feline leukemia virus-related anemia may
be similar to those we observed in one cat available for multiple hematologic
observations (fig. 3). Apparent hemolytic disease accompanied by reticulocytosis
may result in various degrees of blood repopulation by macrocytes. Within a short
period of time the erythroid bone marrow becomes hypoproliferative. Non-regen-
erative anemia then is observed when the cat is presented for veterinary care.
Shortened erythrocyte survival has been observed in feline leukemia virus-
positive cats [9] and a cat with erythroleukemia [22]. Erythrocyte survival studies
in mice infected with Friend leukemia virus, Rauscher virus, and murine erythro-
blastosis virus have demonstrated shortened survival which is likely immune
mediated [3, 6, 11, 311. While some feline leukemia virus-infected cats have been
reported to have a positive Coomb’s test [15, 241, the frequency and role of
antibodies attached to erythrocytes are uncertain in development of the anemia.
Circumstantial evidence supportive of shortened erythrocyte survival includes
observing some feline leukemia virus-infected cats with reticulocytosis before the
anemia is non-regenerative. Also, prominent erythrophagocytosis was observed in
cats with myeloproliferative disorders [22, 23, 271 and in experimentally infected
kittens with erythroid hypoplasia [8]. No reticulocytosis was observed in the
experimentally infected cats in this study. However, some degree of macrocyte
output must have occurred to account for the increase in mean corpuscular volume
and index of anisocytosis values. It is possible that the duration of erythroid
regeneration before erythroid hypoplasia is short and that no reticulocytosis was
detected because of infrequent sampling.
Hemobartonellosis was observed frequently in feline luekemia virus-infected cats
in this and other studies [4, 5 , 16, 191. It is possible that this organism occurs more
frequently in these cats than is recognized. Heinz bodies, frequently observed in
cats, also could contribute to hemolysis. It is interesting to note that the experi-
mentally infected cats developed considerably less macrocytosis than naturally
infected cats. The reason for this difference is not clear, but it is possible that a
hemolytic event such as hemobartonellosis during virus infection potentiates the
development of macrocytosis. Also, the possibility that development of macrocy-
tosis depends on the viral subgroup can not be excluded [ 14, 171.
Alternative explanations for macrocytosis must be considered. Macrocytic ane-
mia occurs in humans with vitamin Bl2 or folate deficiency. Cats placed on folic
696 Weiser and Kociba

acid-deficient diets had megaloblastic changes in the marrow but did not have
increases in mean corpuscular volume [26]. Additionally, several cats with myelo-
proliferative disease and macrocytosis did not show changes in mean corpuscular
volume following treatment with these vitamins [27]. Macrocytosis in human
preleukemia is unresponsive to vitamin replacement [ 13, IS]. These findings suggest
that such a deficiency is not responsible for the macrocytosis. Another explanation
for macrocytosis in cats with myeloproliferative disorders is that the presence of
virus and/or the proliferative disease intrinsically alters erythrocyte production
[27]. This possibility can not be excluded in cats of this study. However, altered
erythropoiesis would have to occur without severe erythroid hypoplasia for a
considerable period of time to account for the degree of repopulation observed.
It is important to interpret mean corpuscular volume values with respect to
values for the method used for measurement. The most widely accepted values,
determined by the microhematocrit method (39-55 fl) [22], have a considerably
higher upper limit than values determined electronically in our laboratory (37-49
fl) [29]. Macrocytosis frequently may be overlooked when interpreting electroni-
cally determined values with respect to reference values established by the micro-
hematocrit method. Anisocytosis, as quantitated on erythrograms, appears more
sensitive than mean corpuscular volume values in detecting the presence of
macrocytic erythrocytes. For example, the experimentally infected cats had a
significant increase in mean corpuscular volume and all cats developed index of
anisocytosis values greater than the upper reference limit. While this indicated
development of considerable macrocytosis, only two of nine cats had a mean
corpuscular volume value greater than the upper reference limit.

References
1 BESSMAN, D.: Red cell volume heterogeneity. Johns Hopkins Med J 146:226-230, 1980
2 BESSMAN, J.D.; JOHNSON, R,K.: Erythrocyte volume distribution in normal and abnormal
subjects. Blood 46:369-379, 1975
3 BRODSKY, I.; DENNIS,L.H.; KAHN,S.B.: Erythropoiesis in Friend leukemia: Red blood
cell survival and ferrokinetics. Cancer Res 26: 1887-1892, 1966
4 COTTER,S.M.: Anemia associated with feline leukemia virus infection. J Am Vet Med
ASSOC175: 1191-1 194, 1975
5 COTTER,S.M.; HARD,W.D.; ESSEX,M.: Association of feline leukemia virus with
lymphosarcoma and other disorders in the cat. J Am Vet Med Assoc 166:449-454, 1975
6 Cox, K.O.; KEAST,D.: Rauscher virus infection, erythrocyte clearance studies, and
autoimmune phenomena. JNCI 50:94 1-946, 1973
7 HENNESS, A.M.; CROW,S.E.; ANDERSON,B.C.: Monocytic leukemia in three cats. J Am
Vet Med Assoc 170: 1325-1328, 1977
8 HOOVER, E.A.; KOCIBA, G.J.; HARDY,W.D.; YOHN,D.S.: Erythroid hypoplasia in cats
inoculated with feline leukemia virus. JNCI 53: 1271-1276, 1974
9 HOOVER,E.A.; KOCIBA,G.J.; OLSEN,R.G.: Experimental erythroid aplasia in cats
induced by feline leukemia virus. Proceedings of the Conference on Asplastic Anemia:
A Stem Cell Disease, p. 306. National Institutes of Health Publication No. 81-1008,
June, 1981
10 HOOVER, E.A.; MATHES, L.E.; ROJKO,J.L.; SCHALLER, J.P.; OLSEN,R.G.: Modifications
 Erythrocyte Macrocytosis 691

of the immunofluorescence assay for feline leukemia virus group-specific antigens. Am

J Vet Res 39:1877-1880, 1978
11 KIRSTEN, W.H.; MAYER,L.A.; WOLLMAN, R.L.; PIERCE,M.I.: Studies on a murine
erythroblastosis virus. JNCI 38: 1 1 7- 139, 1967
I2 KOCIRA, G.J.; HOOVER, E.A.; SCIULLI, V.M.; OLSEN,R.G.: Feline leukemia virus-induced
erythroid aplasia: Corticosteroid enhancement in vivo and FeLV suppression of erythroid
colony development in vitro. In: Advances in Comparative Leukemia Research 198 1,
ed. Yohn and Blakeslee, pp. 21 1-212. Elsevier North Holland, New York, 1982
13 LINMAN, J.W.; BAGBY,G.C.: The preleukemic syndrome (hematopoietic dysplasia).
Cancer 42: 854-864, 1978
14 MACKEY,L.; JARRETT, W.; JARRETT, 0.; LAIRD,H.: Anemia associated with feline
leukemia virus infection in cats. JNCI 54:209-3 17, 1975
15 MADEWELL, B.R.; JAIN,N.C.; WELLER,R.E.: Hematologic abnormalities preceding
myeloid leukemia in three cats. Vet Pathol 16:510-5 19, 1978
16 MAGGIO,L.; HOFFMAN, R.; COTTER,S.M.; DAINIAK, N.; MOONEY,S.; MAFFEI,L A :
Feline preleukemia: An animal model of human disease. Yale J Biol Med 51:469-476,
1978
17 ONIONS, D.; JARRETT, 0.;TESTA, N.; FRASSONI, F.; TOTH,S.: Selective effect of feline
leukemia virus on early erythroid precursors. Nature 296: 156- 158, 1982
18 PIERRE,R.V.: Preleukemic states. Semin Hematol 11:73-92, 1974
19 PRIESTER, W.A.; HAYES,H.M.: Feline leukemia after feline infectious anemia. JNCI
51~289-291, 1973
20 PROCTOR,S.: COX, J.; SHERIDAN, T.: Anisocytosis and the C-1000 channelyzer in
macrocytic anemia. J Clin Pathol 29:7 19-723, 1976
21 ROHOVSY, M.; GRIESEMER, R.; WOLFE,L.: The germ-free cat. Lab Anim Care 16:52-59,
1966
22 SCHALM, O.W.; J A I N N.C.:
, CARROLL, E.J.: Veterinary Hematology, pp. 573-589, 3rd
ed. Lea and Febiger, Philadelphia, 1975
23 SCHALM, O.W.; THEILEN, G.H.: Myeloproliferative disease in the cat, associated with c-
type leukovirus particles in bone marrow. J Am Vet Med Assoc 157: 1686-1696, 1970
24 SCOTT,D.W.: SCHULTZ, R.D.; POST,J.E.; BOLTON,G.R.; BALDWIN, C.A.: Autoimmune
hemolytic anemia in the cat. J Am Anim Hosp Assoc 9:530-539, 1973
25 SOKAL,R.R.: ROIKF,F.J.: Biometry, pp. 242-262, 2nd ed. W.H. Freeman, San Fran-
cisco, 1981
26 THENEN, S.W.; RASMUSSEN, K.M.: Megaloblastic erythropoiesis and tissue depletion of
folk acid in the cat. Am J Vet Res 39: 1205- 1207, 1978
27 WARD,J.M.; SODIKOFF, C.H.; SCHALM, 0.W.: Myeloproliferative disease and abnormal
erythrogenesis in the cat. J Am Vet Med Assoc 155:879-888, 1969
28 WEISER,M.G.: Hematologic techniques. I n ; Veterinary Clinics of North America, ed.
Jain and Zinkl, vol. 1 1 , pp. 277-288. W.B. Saunders, Philadelphia, 1981
29 WEISEK,M.G.: Erythrocyte volume distribution analysis in healthy dogs, cats, horses,
and dairy cows. Am J Vet Res 43: 163-166, 1982
30 WEISER,M.G.; KOCIBA, G.J.: Persistent macrocytosis assessed by erythrocyte subpopu-
lation analysis following erythrocyte regeneration in cats. Blood 60: 295-303, I982
31 WOLLMAN, R.L.; PANG,E.J.; EVANS,A.E.; KIRSTER, W.H.: Virus-induced hemolytic
anemia in mice. Cancer Res 30: 1003-1010, 1970

Request reprints from G. J. Kociba, Department of Veterinary Pathobiology, 1925 Coffey

Rd., Columbus, Ohio 43210 (USA).