- Email: [email protected]
Analysis of feline dual lymphocyte populations observed by flow cytometry
Veterinary
Veterinary
ELSEVIER
hmE=zl”g” immunopathology
Immunology and Immunopathology 48 (1995) 11-25
Analysis of feline dual lymphocyte populations observed by flow cytometry C. Walker
*, S. Bao, P.J. Canfield
Department of Veterinary Pathology, University ofSydney, Sydney, N.S. W. 2006, Australia Accepted
21 December
1994
Abstract Two discrete
were observed commonly on flow cytometric analysis subsets. The identity of these populations as small and large lymphocytes was established by correlating data from FCM with that from peripheral blood films. Dual lymphocyte populations were more likely to be seen in feline immunodeficiency virus-positive (FIV- + ve) cats but their occurrence was not affected by health status, age, gender or breed. FIV- + ve cats had a significantly higher proportion of large lymphocytes than FIV-negative (FIV- - ve) cats. However, FIV- + ve cats had significantly fewer small lymphocytes than FIV- - ve cats but similar numbers of large lymphocytes. Lymphocyte subset analysis revealed that small lymphocytes had a greater proportion of CD4 + cells than large lymphocytes, regardless of the FIV or health status of the cat. In FIV- - ve cats, small lymphocytes had a greater proportion of Pan T + lymphocytes than large lymphocytes, but the converse was seen in FIV- + ve cats. The proportion of CD8 + cells was higher in small lymphocytes than large lymphocytes in well FIV- - ve cats but this distinction was not seen in sick FIV- - ve cats or FIV- + ve cats of any health status. Regardless of health status, FIV- + ve cats had a lower absolute count of small lymphocytes which were T cells (due to lower numbers of both CD4 + and CD8 + cells) than FIV- - ve cats. The numbers of small B cells were similar for both FIV- + ve and FIV- - ve cats. However, there were no differences between FIV- + ve and FIV- - ve cats in the absolute values of any subset of the large lymphocytes, which suggested that FIV may affect only small lymphocytes. Statistically, the inclusion or exclusion of the large lymphocyte population for routine lymphocyte subset analysis did not affect the overall results. However, because there were significant differences in subset distribution between small and large lymphocytes, analysis of both groups should be included in studies examining the role of lymphocytes in disease.
(FCM)
of
feline
lymphocyte
Keywords: Lymphocytes;
* Corresponding
populations
lymphocyte
Subsets; Feline; Flow cytometry;
author at: School of Biological
SSDI 0165-2427(95)05421-9
Sciences,
FIV
Macquarie
University,
NSW 2109, Australia.
12
C. Walker et al. / Veterinary Immunology and Immunopathology
48 (1995) 11-25
1. Abbreviations FCM, flow isothiocyanate; FIV-positive; munoglobulin; scatter; WCC,
cytometric analysis; FeLV, feline leukaemia virus; FITC, fluorescein FIV, feline immunodeficiency virus; FIV- - ve, FIV-negative; FIV- + ve, FSC, forward scatter; HIV, human immunodeficiency virus; Ig, immAb, monoclonal antibody; PBS, phosphate buffered saline; SSC, side white cell count.
2. Introduction Lymphocytes of varying size are routinely observed on feline peripheral blood films and have been variously categorised as small, medium and large or just small and large (Gilmore et al., 1964; Penny et al., 1970; Anderson et al., 1971; Jain, 1993). Values reported in the literature are scant and describe marked individual variation: 57-860/o small and 14-41% large lymphocytes (Penny et al., 1970; Anderson et al., 1971), with differentiation being based on cytoplasmic characteristics rather than defined measurements. To date, the functions of morphologically different lymphocytes have not been well studied. Tompkins et al. (1989) cultured large granular lymphocytes, which had abundant cytoplasm and large azurophilic granules, and tentatively characterised them functionally as natural killer cells or lymphokine-activated killer cells. No flow cytometric (FCM) studies were done to characterise their immunophenotype. In FCM, forward angle light scatter (FSC) measures light refraction and thus identifies cells according to their size, whereas the right angle light scatter, or side scatter @SC), assesses nuclear and cytoplasmic characteristics and thus cell complexity (Rabinowitz et al., 1992). With the development of monoclonal antibodies (mAbs) specific to feline lymphocytes, these techniques are now being utilised to increase our understanding of the feline immune system. In this study, we wish to report the detection of two distinct lymphocyte populations observed by FCM. As they varied in their FSC, these populations were presumed to correspond to small and large lymphocytes present in peripheral blood films (Fig. 1). They were further characterised by the application of T and B lymphocyte markers. In addition, the properties of these distinct populations were compared between healthy and sick cats with and without FIV infection. Finally, the significance of including both lymphocyte populations in the routine analysis of feline lymphocyte subsets was examined.
3. Materials and methods 3.1. Animals Sixty-eight purebred and domestic cats, all feline leukaemia virus (FeLV)-negative (FeLV-antigen ELISA (enzyme-linked immunosorbent assay), Virachek/ FeLV; Synbiotics San Diego, CA, USA), were accessed through veterinary hospitals or commercial
C. Walker et al. / Veterinary Immunology and Immunopathology 48 (I 995) I I-25
13
catteries. The health status of each cat was determined by clinical examination and history as either ‘well’ or ‘sick’. Well cats were those which presented as clinically normal with no history of either chronic or intermittent disease. Sick cats either presented with one or more clinical signs or had a history of clinical signs occurring intermittently. All cats were tested both with a commercially available FIV-antibody ELISA (CITE; Agritech Systems, Portland, ME, USA) and by Western blot analysis to determine FIV status (Walker et al., 1994). 3.2. Sample collection Blood was collected by jugular venipuncture from cats that were non-tranquillised/ non-anaesthetised, tranquillised with ketamine/ diazapam, or anaesthetised with ketamine or ketamine/ acetylpromazine. Generally, samples were collected in the morning into EDTA (ethylenediaminetetraacetic acid), stored at room temperature and processed on the same day. 3.3. Haematology The total white cell counts (WCC) were determined with a Coulter Counter DN (Coulter Electronics, Luton, UK). Blood films were stained with Giemsa and 100 leucocytes were differentiated for every 10 X lo9 WCC 1-l. The absolute lymphocyte count was calculated by multiplying the WCC by the observed percentage of lymphocytes from the differential count. 3.4. Flow cytometry Samples were prepared as described previously (Walker et al., 1994). Briefly, whole blood was combined in separate tubes with mAbs which recognise feline Pan T + cells (Ackley and Cooper, 19921, CD4 + cells (Ackley et al., 1990a) and CD8 + cells (Klotz and Cooper, 1986) (Southern Biotechnology, Birmingham, AL, USA). Secondary staining was performed by incubating with fluorescein isothiocyanate (FITC)-conjugated goat anti-mouse immunoglobulin (Ig) G (ICN Biomedicals, Costa Mesa, CA, USA). B lymphocytes were enumerated by prewashing whole blood five times at room temperature in phosphate buffered saline (PBS)/ azide and labelling with FITC-conjugated goat anti-cat IgG (H + L) antibody (Southern Biotechnology) to detect surface Ig. Analyses were performed on a FACScan flow cytometer (Becton Dickinson, Lane Cove, N.S.W.) on the same day as cell preparation. A dotplot of forward (FSC) and side scatter (SSC) was examined and used to construct a live lymphocyte gate of 5000 cells for each antigenic marker, including the negative control. Both lymphocyte populations (small and large) were included in this collection gate (Fig. 1). Fluorescence histograms of cell numbers versus fluorescence intensity were analysed for each lymphocyte population and gates delineating the positive regions set against the negative control. 3.5. Examination of smears Blood films were made immediately after venipuncture and stained with Giemsa. For those samples which demonstrated two lymphocyte populations on FCM, 50 lympho-
14
C. Walker et al. / Veterinary Immunology and Immunopathologv 48 (1995) I I-25
cytes were counted using a modification of the battlement technique (Jain, 1986), viewing both horizontal edges and central fields by traversing the slide. Each smear was scanned to appraise the variation in lymphocyte sizes present and the lymphocytes were divided into categories based on relative size alone: small, medium and large. The actual measurements of the cells varied between samples and, depending on the area of the smear being counted, with cells towards the tail of the smear being larger than those in the body. On average, small lymphocytes were less than 9 pm, medium lymphocytes were between 9 pm and 11 pm, and large lymphocytes were greater than 11 pm in diameter. Cytoplasmic (degree of basophilia, presence of granules) or nuclear differences (shape, chromatin pattern) were not considered. 3.6. Calculations and statistics To examine the proportions of the two lymphocyte populations observed on the dotplot, the negative control was used to record cell numbers from each population. Using a commercial statistical software package (Minitab Inc., State College, PA, USA), the numbers were expressed as a percentage of the total lymphocyte population. The Pearson product moment correlation coefficient was used to compare these flow cytometric values with those obtained from examining the smears. The absolute count of each lymphocyte population was calculated by multiplying the absolute lymphocyte count by the percentage observed by FCM. The effects of health status and FIV infection on the occurrence and distribution of the two lymphocyte populations as observed by FCM were examined for statistical significance (P < 0.05) using the chi-square statistic and a Student’s C-test. The differences in lymphocyte subset distribution between the two lymphocyte populations were also analysed using a Student’s t-test. The effect of including the larger lymphocyte population in the analysis of lymphocyte subsets was investigated by comparing the percentage of cells positive for an antigenic marker in the smaller lymphocyte population alone with those positive for both the smaller and larger populations together.
4. Results 4.1. Demonstration
of dual lymphocyte populations
by FCM
Using a dotplot of FSC and SSC, 84% (57/68) of samples examined showed dual lymphocyte populations on FCM (Fig. 1). To display these pdpulations more fully, using the events gated as lymphocytes in the negative control tube, a three-dimensional plot of FSC versus SSC versus cell numbers was examined (Fig. 2). This demonstrated that the lymphocytes comprised two discrete populations of different size (based on differing FSC) but which had comparable shape and complexity (based on similar SCC). 4.2. Experimental group The cats demonstrating dual populations (57/68) included 40 castrated males with the remainder being ovariohysterectomised females. Nine were purebred cats. Their
C. Walker et af. / Veterinary Immunology and Immunopathoiogy 48 (1995) I1 -25
15
I
Fig. 1. Dotplot of forward scatter (FSC) versus side scatter populations. a, small lymphocytes; b, large lymphocytes.
health status, Clinical signs ve5; FIV- + disease (FIV-
(SSC)
demonstrating
the two lymphocyte
FIV status and mean ages for each subgroup are presented in Table 1. observed within the ‘sick’ group included chronic oral disease (FIV-ve:ll), miliary dermatitis (FIV- - ve:3; FIV- + ve:9), respiratory tract - ve:2; FIV- + ve:7), cryptococcosis (FIV- - ve:l; FIV- + ve:2), feline
Fig. 2. Characterisation of the dual lymphocyte (SSC) versus cell numbers. The two populations shape and complexity (SCC).
population using forward scatter (FSC) versus side scatter comprised cells of two different sixes (FSC) but similar cell
16
C. Walker et al. / Veterinary Immunology and Immunopathology 48 (1995) 11-25
Table 1 Data for those cats which demonstrated
dual lymphocyte
populations
Health status
FIV status
Number in group
Age (years, mean f SD)
Well
FIV-negative FIV-positive
16 4
3.5 + 3.2 7.1 L-4.1
Sick
FIV-negative FIV-positive
10 27
7.8k4.2 8.9k3.6
SD, standard
deviation.
lower urinary tract disease (FIV- - ve:l; FIV- + ve:4), renal failure (FIV- - ve:2; FIV+ ve:l) and weight loss (FIV- + ve:4). Those cats which did not demonstrate dual populations (11/68) included eight castrated males, two ovariohysterectomised and one entire female. Two were purebred. The data for these cats are presented in Table 2. The sick cats within this group displayed chronic oral disease (FIV- - ve:2; FIV- + ve:l), miliary dermatitis (FIV- + ve:l), respiratory tract disease (FIV- - ve:2; FIV- + ve:4), cryptococcosis (FIV- - ve:2), renal failure (FIV- - ve:l) and weight loss (FIV- - ve:l; FIV- + ve:2). FIV- + ve cats (31/33, 94%) were more likely to demonstrate dual lymphocyte populations than FIV- - ve cats (26/35, 75%) ( x2 = 4.839; d.f. = 1; 0.02 < P < 0.05). However, health status ( x2 = 0.426; d.f. = 1; 0.5 < P < 0.71, age ( x2 = 18.495; d.f. = 15; 0.2 < P < 0.3) gender ( x2 = 0.029; d.f. = 1; 0.8
The percentages of cells in the smaller and larger lymphocyte population seen on FCM and the percentages of cells in the various groups (small, medium and large) as counted on the corresponding blood films are presented in Table 3. When the smaller lymphocyte population, i.e. those lymphocytes with lower FSC, was compared with the combined values of the small and medium lymphocytes, as
Table 2 Data for those cats which did not demonstrate
dual lymphocyte
populations
Health status
FIV status
Number in group
Well
FIV-negative FIV-positive
4 0
4.5 f 2.7 _
Sick
FIV-negative FIV-positive
5 2
9.6k4.7 10.5 + 2.1
SD, standard
deviation.
Age (years, mean + SD)
C. Walker et al. / Veterinary Immunology and Immunopathology Table 3 Percentages
of lymphocytes
of various sizes observed
by flow cytometry
48 (1995) 11-25
and blwd
films (n = 57)
Size of lymphocyte
Mean + SD (o/o)
Flow cytometry
Low FSC (smaller) High FSC (larger)
72.7 f 14.1 27.3 f 14.1
Blood film
Small Medium Small + medium Large
18.5 56.3 72.9 25.5
SD, standard
deviation;
FSC, forward
17
k 16.8 zk 16.1 + 15.4 zk 14.2
scatter.
counted on the blood films, the correlation coefficient was 0.891. When the larger lymphocyte population, i.e. those lymphocytes with higher FSC, was compared with the large lymphocytes as counted on the blood films, the correlation coefficient was 0.803. Thus the smaller population of lymphocytes observed by FCM correlated well with small and medium lymphocytes, and the larger population with large lymphocytes observed on peripheral blood films. For the remainder of this report, ‘small’ lymphocytes will refer to the smaller group of lymphocytes as observed by FCM and the combination of small and medium lymphocytes as observed on peripheral blood films. Similarly, ‘large’ lymphocytes will refer to the larger group of lymphocytes seen on FCM and large lymphocytes seen on peripheral blood films.
4.4. The effect of health status and FIV status on the distribution lymphocytes
of small and large
Using FCM, the percentages of small and large lymphocytes were compared between well and sick cats. No significant differences in the proportion of small and large lymphocytes were found between sick and well cats (Fig. 3(a)). Similarly, the absolute numbers of neither small nor large lymphocytes differed between well and sick cats (Fig. 3(b)). This was true for both FIV- + ve and FIV- - ve cats. However, when the percentages of small and large lymphocytes were compared between FIV- + ve and FIV- - ve cats regardless of health status, FIV- + ve cats had a lower percentage (P = 0.016) of small lymphocytes and a higher percentage (P = 0.016) of large lymphocytes than FIV- - ve cats. This was due to FIV- + ve cats having significantly fewer (P = 0.0044) small lymphocytes than FIV- - ve cats, yet similar numbers of large lymphocytes. The FIV effect on lymphocyte proportions was only observed in well cats (P = 0.019), as there was no significant difference in the percentages of small and large lymphocytes between sick FIV- + ve and sick FIV- - ve cats. The possibility that this difference in lymphocyte numbers may be age-related was examined. There was statistically no difference between the mean ages of the well FIV- - ve cats and the well FIV- + ve cats (P = 0.2). Because of marked intercat variability, there were no significant differences in absolute values of either small or large lymphocytes between FIV- - ve and FIV- + ve cats after they had been grouped as ‘well’ and ‘sick’.
18
C. Walker et al. / Veterinary Immunology and Immunopathology 48 (1995) 11-25
4.5. The effect of health status alone on the distribution of antigenic markers between small and large lymphocytes Most reports of feline lymphocyte subset values have analysed the effects of immunosuppressive viral infections, e.g. FIV (Ackley et al., 1990b; Walker et al., 1994) or FeLV (Quackenbush et al., 1990; Tompkins et al., 1991), yet there have been no accounts of how lymphocyte subsets vary between health and ill-health generally. To
m
(a)
m
(bl
Small lymphocytes
We11
Sick
Largelymphocytes
Well
Sick
Small lymphocytes
Well
I
L_I
Sick
Large lymphocytes
Well
Sick
-
Fig. 3. Compatison of (a) proportions (%) and (b) absolute values (X lo9 l- ‘1 of small and large lymphocytes between well and sick FIV-negative (FIV( - 1) and FIV-positive (FIV( + )) cats (mean f standard error).
C. Walker et al. / Veterinary Immunology and Immunopathology
48 (1995) 11-25
m
Small lymphocytes -well cats
m
Small lymphocytes - sick cats
a
Large lymphocytes - well cats
B
Large lymphocytes - sick cats
19
~~
so
3
PanT+
CD4+
Fig. 4. Comparison of the subset proportions FIV-negative cats (mean +_standard error).
(%) of small
cD8+ and large
B Cell lymphocytes
for well and sick
investigate the effect of health status on the differences in distribution of antigenic markers between small and large lymphocytes, the subset percentages in FIV- - ve well cats and FIV- - ve sick cats were compared, which excluded the effect of FIV (Fig. 4). Small lymphocytes expressed a significantly greater percentage of CD4 + cells for both well cats (P = 0.037) and for sick cats (P = 0.0011) and a significantly lesser percentage of B cells for both well cats (P = 0.0026) and sick cats (P = 0.0028) than large lymphocytes. Only the distribution of CDS + cells varied with health status. In well cats, small lymphocytes demonstrated a higher percentage of CD8 + cells than large lymphocytes (P = 0.0084) but there was no difference in the distribution of CD8 + cells between small and large lymphocytes in sick cats. As with the absolute numbers of small and large lymphocytes, there were no differences between well and sick FIV- - ve cats in the absolute values of any subset of either small or large lymphocytes (data not shown). 4.6. The effect of FIV infection and large lymphocytes
on the distribution
of antigenic
markers between small
The differences in percentages and absolute values of each antigenic marker between small and large lymphocytes were analysed for all FIV- + ve cats and compared with the differences of each antigenic marker between small and large lymphocytes for FIV- - ve cats (Figs. 5(a) and 5(b)). In FIV- - ve cats, small lymphocytes had a greater proportion of Pan T + cells than large lymphocytes (P = 0.0211, but the converse was seen in FIV- + ve cats (P = 0.011) (Fig. 5(a)). The small lymphocytes of FIV- - ve cats had a significantly higher proportion of Pan T + cells than the small lymphocytes of FIV- + ve
20
C. Walker et al. /Veterinary
LI
Immunology and Immunopathology
48 (1995) 11-25
/--
I
PanT+
cD8+
CD4+
B Cell
la)
m
Small lymphocytes - FIV-negative
cats
m
Small lymphocytes - FIV-positive
cats
m
Large lymphocytes - FIV-negative
cats
B
Large lymphocytes - FlV-pcsitive
cats
(b)
PanT+
cD4+
CD8+
B Cell
Fig. 5. Comparison of (a) the subset proportions (%I and (b) the absolute values (X lo9 I-‘) of subsets of small and large lymphocytes between well and sick FIV-negative cats and well and sick FIV-positive cats (mean f standard error).
cats (P = 0.019) and the large lymphocytes of FIV- - ve cats had a significantly lower proportion of Pan T + cells than FIV- + ve cats (P = 0.016) (Fig. 5(a)). FIV- + ve cats had significantly fewer small lymphocytes (P = 0.002) which were Pan T + than FIV- - ve cats, but had similar numbers (P = 0.44) of large Pan T + lymphocytes (Fig. 5(b)).
C. Walker et al. / Veterinary Immunology
and Immunopathology
48 (1995) 11-25
21
In both FIV- - ve and FIV- + ve cats, small lymphocytes had proportionately more CD4 + cells than large lymphocytes (P = 0.0002 and P = 0.0017 respectively). Both small and large lymphocytes of FIV- - ve cats displayed proportionately more (P = 0.001 and P = 0.0021 respectively) cells positive for the CD4 marker than the corresponding lymphocytes of FIV- + ve cats (Fig. 5a). However, when absolute numbers of lymphocytes were examined, it was only for small lymphocytes that FIV- + ve cats had fewer CD4 + cells (P = 0.0001) as there was no difference in the numbers of large CD4 + lymphocytes between FIV- + ve and FIV- - ve cats (P = 0.17) (Fig. 5(b)). Unlike FIV- - ve cats, FIV- + ve cats displayed no difference in the percentage of CD8 + cells between small and large lymphocytes, regardless of health status. There were no differences in the proportions of cells which were CD8 + between FIV- - ve and FIV- + ve cats for either small or large lymphocytes (Fig. 5(a)). Again, FIV- - ve cats had significantly more small CD8 + lymphocytes than FIV- + ve cats (P = 0.0024) but there was no difference in the absolute number of large CD8 + lymphocytes (P = 0.36) (Fig. 5(b)). The FIV- + ve, but not the FIV- - ve cats had a higher CD4:CD8 ratio for small lymphocytes compared with large lymphocytes (P = 0.0054). The finding that small lymphocytes had a lower percentage of B cells than large lymphocytes (P = 0.0001) in FIV- - ve cats was not observed in FIV- + ve cats. The small lymphocytes of FIV- + ve cats had a significantly higher percentage of B cells (P = 0.014) and the large lymphocytes a significantly lower percentage (P = 0.012) than those of FIV- - ve cats (Fig. 5(a)). There were no differences between FIV- + ve and FIV- - ve cats in the absolute numbers of B cells for either small or large lymphocytes (Fig. 5(b)). 4.7. The effect of health status on the distribution and large lymphocytes in FIV-positive cats
of antigenic
markers between small
In the FIV- + ve cats which were well, there were no differences in subset distribution between small and large lymphocytes (Fig. 6). However, in the FIV-+ ve cats which were sick, the small lymphocytes had a significantly lower percentage (P = 0.011) of PanT + cells, a higher percentage of CD4 + lymphocytes (P = 0.0017) and higher CD4:CD8 ratios (P = 0.0002) but no difference in the distribution of CD8 + or B cells when compared with the large lymphocytes. As with the FIV- - ve cats, there were no differences between well and sick FIV- + ve cats in the absolute values of any subset of either the small or large lymphocyte (data not shown). 4.8. Significance
of including both lymphocyte populations
in lymphocyte
subset analysis
Conventionally in subset analysis of human lymphocytes, there is only one population of lymphocytes, represented usually by a discrete tight population of cells with low FSC and SSC. However, as this study has shown, in feline samples there are commonly two discrete lymphocyte populations. Reports of feline subset analysis in the literature do not discuss these dual populations, or in fact which lymphocyte population provided the data. The data from all cats in this study were analysed to examine the effect on the total subset values of including the group of large lymphocytes. The values obtained for
C. Walker et al. / Veterinary Immunology and Immunopathology 48 (1995) 11-25
22
m
Small lymphocytes - well cats
m
Small lymphocytes
- sick cats
L_i
Large lymphocytes -well cats
B
Large lymphocytes
- sick cats
.90 r
3 d
4s
i
32
2 .o B
16.
2 4 0
PanT+
cD4+
Fig. 6. Comparison of the subset proportions FN-positive cats (mean f standard error).
(o/o) of small
CD8+ and large
B Cell lymphocytes
for well
and sick
the small lymphocyte population, which comprised the majority of lymphocytes, were compared with those obtained from analysis of the small and large lymphocyte populations together. There was no significant difference for any lymphocyte subset between the two groups regardless of FIV infection or health status (data not shown).
5. Discussion In this study, the flow cytometric observation of two discrete lymphocyte populations, differing in size, was a common finding. Large lymphocytes comprise only 2% of total lymphocytes in humans (Douglas, 1983) which may explain why there are no flow cytometric reports of dual populations of lymphocytes based on variations in their FSC. It must remain only an assumption that both the populations are in fact lymphocytes, and not, for instance, monocytes. However, the good correlation between FCM results and examination of blood films observed in this study would support both populations being lymphocytes. Moreover, it has been reported in the literature that none of the feline mAbs used in our study reacted with monocytes (or granulocytes, erythrocytes or platelets) when isolated cell populations were examined for antigen expression (Klotz and Cooper, 1986; Ackley et al., 1990a; Ackley and Cooper, 1992). FIV infection appeared to cause a significantly higher percentage of large lymphocytes due to a reduction in small lymphocyte numbers but no alteration in large lymphocyte numbers (Fig. 3(b)). This suggested that FIV affects small lymphocytes rather than large lymphocytes. It is not known what the functional differences between small and large lymphocytes are in cats. In humans, larger lymphocytes often represent
C. Walker et al. / Veterinary Immunology and Immunopathology
48 (1995) 11-25
23
activated cells (Weiss, 1993), so perhaps it is resting cells which are depleted by FIV infection, resulting in proportionately more activated lymphocytes. This FIV effect appeared to be negated by the onset of illness. In humans, the proportion of large lymphocytes (of unreported phenotype) is seen to increase with acute viral illness (Douglas, 1983). Th e sick cats in this study had chronic disease and perhaps this explains why there was no significant difference between healthy and sick cats in their proportions of small and large lymphocytes. In humans, variations in SSC (reflecting cell structure) have been observed and utilised to analyse differences in subset distribution (Terstappen et al., 1986; Rabinowitz et al., 1992). Using similar methodology, this study has demonstrated differences in distribution of CD4 + , CD8 + and B cells between small and large lymphocytes in cats. The study was able to demonstrate the effect of ill-health alone, although numbers precluded subset analysis among specific disease states. The significance of CD8 + cells alone varying proportionally with health status between the two lymphocyte groups is not known. It is possible that the percentage of large lymphocytes expressing the CD8 + marker tends to increase in response to illness, whilst the distribution of CD4 + and B cells remains similar to that found in health. These differences may reflect alterations in subpopulations of CD8 + cells. There were significant differences between FIV- + ve and FIV- - ve cats in the distribution of lymphocyte subsets between small and large lymphocytes. The differences between FIV- - ve and FIV- + ve cats occurring in this study in subsets of the small lymphocytes in this study concur with those of other workers (Ackley et al., 1990b; Novotney et al., 1990; Barlough et al., 1991; Tompkins et al., 1991; Torten et al., 1991; Hoffmann-Fezer et al., 1992). However, the finding that the subsets of large lymphocytes appear unaffected by FIV infection is novel. The fact that there were no differences in the distribution of CD8 + cells between small and large lymphocytes, in either well or sick FIV- + ve cats, suggests that, even when well, FIV- + ve cats have more CD8 + cells which are larger (and thus possibly activated). Perhaps these CD8 + cells have a cytotoxic function. It is well accepted that human immunodeficiency virus (HIV) causes B cell activation in all stages of infection (Mizuma et al., 1988) and the B cells of HIV-positive individuals are larger than those of uninfected controls (Martinez-Maza et al., 1987). The significance of the variation in subset distribution between large and small lymphocytes and the involvement of FIV infection has yet to be clarified. It is not known whether activated B cells in the cat are larger than resting B cells as they are in other species (Kincade and Gimble, 1993). If activated B cells are larger than resting B cells and FIV causes B cell activation, it may have been expected that FIV- + ve cats would have had more larger B cells (either relatively or absolutely) than FIV- - ve cats. This was not found. Thus it would seem from this study that there is no evidence, based on variation in cell size between FIV- - ve and FIV- + ve cats, of B cell activation in FIV- + ve cats. It may be possible to explore this further by double staining lymphocytes for T or B cell markers and an FIV-specific marker (Steinman et al., 1990) to investigate the proportion of infected cells in the two lymphocyte populations. mAbs which act as activation markers, as are available for human T and B lymphocytes (Formenti et al.,
24
C. Walker et al. / Veterinary Immunology and Immunopathology 48 (I 995) 11-25
19891, would be useful to determine whether the various sizes of feline lymphocytes represent differing stages of activation. Also, the development of mAbs specific for immature and mature cell types would help clarify the significance of small and large lymphocytes. It is unknown whether or not small lymphocytes are more mature than large lymphocytes in cats as they are in other species (Beagley et al., 1989). Similarly, size differences and differences in lymphocyte subset distribution between the small and large lymphocytes give no indication of the functional differences either in vitro or in vivo which may exist between the two populations. Given that large lymphocytes comprise a considerable proportion of feline circulating lymphocytes, and have been shown to have significant differences in lymphocyte subset distribution, it is important to consider their inclusion or exclusion in routine flow cytometric analysis. This study demonstrated that, because the absolute numbers of large lymphocytes in feline samples are often much lower than those of small lymphocytes, their effect overall on analysis of lymphocyte subsets is statistically insignificant. Thus, for routine subset analysis, inclusion or exclusion of the large lymphocyte population in the analysis gate will not affect the overall results. However, further investigation of the feline immune system, especially for particular disease states, e.g. FIV infection, and health status should involve examination of both lymphocyte populations by FCM.
Acknowledgements This work was supported in part by a Faculty of Veterinary Science grant. The authors thank the clinical pathology laboratory for use of their facilities. Professsor Wayne Robinson kindly provided the FIV antigen preparation for the Western blots. Thanks also to Christine Bligh and Becton Dickinson for advice and materials for flow cytometry, and Dr. John Quinn, from Westmead Hospital, for advice on B cell staining. The FeLV tests were generously supplied by Commonwealth Serum Laboratories. Thanks also to the many veterinary practitioners who referred cases to us for this study.
References Ackley, C.D. and Cooper, M.D., 1992. Characterization of a feline T-cell-specific monoclonal antibody reactive with a CDS-like molecule. Am. J. Vet. Res., 53: 466-471. Ackley, C.D., Hoover, E.A. and Cooper, M.D., 1990a. Identification of a CD4 homologue in the cat. Tissue Antigens, 35: 92-98. Ackley, C.D., Yamamoto, J.K., Levy, N., Pedersen, N.C. and Cooper, M.D., 1990b. Immunologic abnormalities in pathogen-free cats experimentally infected with feline immunodeficiency virus. J. Virol., 64: 5652-5655. Anderson, L., Wilson, R. and Hay, D., 1971. Haematological values in normal cats from four weeks to one year of age. Res. Vet. Sci., 12: 579-583. Barlough, J.E., Ackley, C.D., George, J.W., Levy, N., Acevedo, R., Moore, P.F., Rideout, B.A., Cooper, M.D. and Pedersen, N.C., 1991. Acquired immune dysfunction in cats with experimentally induced feline immunodeficiency virus infection: comparison of short-term and long-term infections. J. Acq. Immun. Defic. Syndr., 4: 219-227. Beagley, K.W., Eldridge, J.H., Lee, F., Kiyono, H., Everson, M.P., Koopman, W.J., Hirano, T., Kishimoto, T.
C. Walker et al. / Veterinary Immunology and Immunopathology
48 (1995) 11-25
25
and McGhee, J.R., 1989. Interleukins and IgA synthesis. Human and murine interleukin 6 induce high rate IgA secretion in &A-committed B cells. J. Exp. Med., 169: 2133-2148. Douglas, SD., 1983. Morphology of lymphocytes and plasma cells. In: W.J. Williams, E. Beutler, A.J. Erslev and M.A. Lichtman (Editors), Haematology. McGraw-Hill, New York, pp. 883-902. Formenti, SC., Turner, R.R., de Martini, R.M., Boone, D.C., Bishop, P.C., Levine, A.M. and Parker, J.W., 1989. Immunophenotypic analysis of peripheral blood leukocytes at different stages of HIV infection. Am. J. Clin. Pathol., 92: 300-307. Gilmore, C.E., Gilmore, V.H. and Jones, T.C., 1964. Bone marrow and peripheral blood of cats: technique and normal values. Pathol. Vet., 1: 18-40. Hoffmann-Fezer, G., Thum, J., AckIey, C., Herbold, M., Mysliwietz, J., Thefeld, S., Hartmann, K. and Kraft, W., 1992. Decline in CD4-positive cell numbers in cats with naturally acquired feline immunodeficiency virus infection. J. Virol., 66: 1484-1488. Jain, N.C., 1986. Schalm’s Veterinary Haematology. Lea&Febiger, Philadelphia, 1221 pp. Jain, N.C., 1993. Essentials of Veterinary Haematology. LeaBtFebiger, Philadelphia, 417 pp. Kincade, P.W. and Gimble, J.M., 1993. B lymphocytes. In: W.E. Paul (Editor), Fundamental Immunology. Raven Press, New York, pp. 43-73. Klotz. F.W. and Cooper, M.D., 1986. A feline thymocyte antigen defined by a monoclonal antibody (FI’2) identifies a subpopulation of non-helper cells capable of specific cytotoxicity. J. Immunol., 136: 2510-2514. Martinez-Maza, O., Crabb, E., Mitsuyasu, R.T., Fahey, J.L. and Giorgi, J.V., 1987. Infection with the human immunodeficiency virus (HIV) is associated with an in vivo increase in B lymphocyte activation and immaturity. J. Immunol., 138: 3720-3724. Mizuma, H., Litwin, S. and Zoller-Pazner, S., 1988. B-cell activation in HIV infection: relationship of spontaneous immunoglobulin secretion to various immunological parameters. Clin. Exp. Immunol., 71: 410-416. Novotney. C.. English, R.V., Housman, J., Davidson, M.G.. Nasisse, M.P., Jeng, C.R., Davis, W.C. and Tompkins. M.B., 1990. Lymphocyte population changes in cats naturally infected with feline immunodeficiency virus. AIDS, 4: 1213-1218. Penny. R.H.C., Carlisle, C.H. and Davidson, H.A., 1970. The blood and marrow picture of the cat. Br. Vet. J.. 126: 459-454. Quackenbush, S.L., Donahue, P.R., Dean, G.A., Myles, M.H.. Ackley, C.D., Cooper, M.D., Mullins, J.I. and Hoover, EA., 1990. Lymphocyte subset alterations and viral determinants of immunodeficiency disease induction by the feline leukaemia virus FeLV-FAIDS. 3. Virol., 64: 5465-5474. Rabinowitz, R.. Granot, E., Deckelbaum, R. and Schlesinger, M., 1992. Antigenic differences between subsets of peripheral blood lymphocytes differing in their right angle light scatter in flow cytometric analysis. Int. Arch. Allergy Immunol., 97: 200-204. Steinman. R.. Dombrowski, J., O’Connor, T., Monteolaro, R.C., Tonelli, Q., Lawrence, K., Seymour. C.. Goodness. J., Pedersen, N.C. and Andersen, P.R., 1990. Biochemical and immunological characterization of the major structural proteins of feline immunodeficiency virus. J. Gen. Virol.. 71: 701-706. Terstappen, L.W.M.M., de Grooth. B.G., Nolten, G.M.J., ten Napel. C.H.H.. van Berkel, W. and Greve, J., 1986. Physical discrimination between human T-lymphocyte subpopulations by means of light-scattering, revealing IWO populations of TB-positive cells. Cytometry, 7: 178-183. Tompkins, M.B., Pang, V.F., Michaely, P.A.. Feinmehl, R.I., Basgall, E.J., Baszler, T.V., Zachary, J.F. and Tompkins. W.A.A., 1989. Feline cytotoxic large granular lymphocytes induced by recombinant human IL-2. J. Immunol., 143: 749-754. Tompkins, M.B., Nelson, P.D., English, R.V. and Novotney, C., 1991. Early events in the immunopdthogenesis of feline retrovirus infections. J. Am. Vet. Med. Assoc.. 199: 1311-1315. Torten, M., Franchini, M., Barlough, J.E., George, J.W., Mozes, E., Lutz, H. and Pedersen, N.C., 1991. Progressive immune dysfunction in cats experimentally infected with feline immunodeficiency virus. J. Virol., 65: 2225-2230. Walker. C.. Canfield, P.J. and Love, D.N., 1994. Analysis of leucocytes and lymphocyte subsets for different clinical stages of naturally-acquired feline immunodeficiency virus infection. Vet. Immunol. Immunopathol., 44: I-12. Weiss. A., 1993. T lymphocyte activation. In: W.E. Paul (Editor). Fundamental Immunology. Raven Press, New York, pp. 467-504.
Veterinary
ELSEVIER
hmE=zl”g” immunopathology
Immunology and Immunopathology 48 (1995) 11-25
Analysis of feline dual lymphocyte populations observed by flow cytometry C. Walker
*, S. Bao, P.J. Canfield
Department of Veterinary Pathology, University ofSydney, Sydney, N.S. W. 2006, Australia Accepted
21 December
1994
Abstract Two discrete
were observed commonly on flow cytometric analysis subsets. The identity of these populations as small and large lymphocytes was established by correlating data from FCM with that from peripheral blood films. Dual lymphocyte populations were more likely to be seen in feline immunodeficiency virus-positive (FIV- + ve) cats but their occurrence was not affected by health status, age, gender or breed. FIV- + ve cats had a significantly higher proportion of large lymphocytes than FIV-negative (FIV- - ve) cats. However, FIV- + ve cats had significantly fewer small lymphocytes than FIV- - ve cats but similar numbers of large lymphocytes. Lymphocyte subset analysis revealed that small lymphocytes had a greater proportion of CD4 + cells than large lymphocytes, regardless of the FIV or health status of the cat. In FIV- - ve cats, small lymphocytes had a greater proportion of Pan T + lymphocytes than large lymphocytes, but the converse was seen in FIV- + ve cats. The proportion of CD8 + cells was higher in small lymphocytes than large lymphocytes in well FIV- - ve cats but this distinction was not seen in sick FIV- - ve cats or FIV- + ve cats of any health status. Regardless of health status, FIV- + ve cats had a lower absolute count of small lymphocytes which were T cells (due to lower numbers of both CD4 + and CD8 + cells) than FIV- - ve cats. The numbers of small B cells were similar for both FIV- + ve and FIV- - ve cats. However, there were no differences between FIV- + ve and FIV- - ve cats in the absolute values of any subset of the large lymphocytes, which suggested that FIV may affect only small lymphocytes. Statistically, the inclusion or exclusion of the large lymphocyte population for routine lymphocyte subset analysis did not affect the overall results. However, because there were significant differences in subset distribution between small and large lymphocytes, analysis of both groups should be included in studies examining the role of lymphocytes in disease.
(FCM)
of
feline
lymphocyte
Keywords: Lymphocytes;
* Corresponding
populations
lymphocyte
Subsets; Feline; Flow cytometry;
author at: School of Biological
SSDI 0165-2427(95)05421-9
Sciences,
FIV
Macquarie
University,
NSW 2109, Australia.
12
C. Walker et al. / Veterinary Immunology and Immunopathology
48 (1995) 11-25
1. Abbreviations FCM, flow isothiocyanate; FIV-positive; munoglobulin; scatter; WCC,
cytometric analysis; FeLV, feline leukaemia virus; FITC, fluorescein FIV, feline immunodeficiency virus; FIV- - ve, FIV-negative; FIV- + ve, FSC, forward scatter; HIV, human immunodeficiency virus; Ig, immAb, monoclonal antibody; PBS, phosphate buffered saline; SSC, side white cell count.
2. Introduction Lymphocytes of varying size are routinely observed on feline peripheral blood films and have been variously categorised as small, medium and large or just small and large (Gilmore et al., 1964; Penny et al., 1970; Anderson et al., 1971; Jain, 1993). Values reported in the literature are scant and describe marked individual variation: 57-860/o small and 14-41% large lymphocytes (Penny et al., 1970; Anderson et al., 1971), with differentiation being based on cytoplasmic characteristics rather than defined measurements. To date, the functions of morphologically different lymphocytes have not been well studied. Tompkins et al. (1989) cultured large granular lymphocytes, which had abundant cytoplasm and large azurophilic granules, and tentatively characterised them functionally as natural killer cells or lymphokine-activated killer cells. No flow cytometric (FCM) studies were done to characterise their immunophenotype. In FCM, forward angle light scatter (FSC) measures light refraction and thus identifies cells according to their size, whereas the right angle light scatter, or side scatter @SC), assesses nuclear and cytoplasmic characteristics and thus cell complexity (Rabinowitz et al., 1992). With the development of monoclonal antibodies (mAbs) specific to feline lymphocytes, these techniques are now being utilised to increase our understanding of the feline immune system. In this study, we wish to report the detection of two distinct lymphocyte populations observed by FCM. As they varied in their FSC, these populations were presumed to correspond to small and large lymphocytes present in peripheral blood films (Fig. 1). They were further characterised by the application of T and B lymphocyte markers. In addition, the properties of these distinct populations were compared between healthy and sick cats with and without FIV infection. Finally, the significance of including both lymphocyte populations in the routine analysis of feline lymphocyte subsets was examined.
3. Materials and methods 3.1. Animals Sixty-eight purebred and domestic cats, all feline leukaemia virus (FeLV)-negative (FeLV-antigen ELISA (enzyme-linked immunosorbent assay), Virachek/ FeLV; Synbiotics San Diego, CA, USA), were accessed through veterinary hospitals or commercial
C. Walker et al. / Veterinary Immunology and Immunopathology 48 (I 995) I I-25
13
catteries. The health status of each cat was determined by clinical examination and history as either ‘well’ or ‘sick’. Well cats were those which presented as clinically normal with no history of either chronic or intermittent disease. Sick cats either presented with one or more clinical signs or had a history of clinical signs occurring intermittently. All cats were tested both with a commercially available FIV-antibody ELISA (CITE; Agritech Systems, Portland, ME, USA) and by Western blot analysis to determine FIV status (Walker et al., 1994). 3.2. Sample collection Blood was collected by jugular venipuncture from cats that were non-tranquillised/ non-anaesthetised, tranquillised with ketamine/ diazapam, or anaesthetised with ketamine or ketamine/ acetylpromazine. Generally, samples were collected in the morning into EDTA (ethylenediaminetetraacetic acid), stored at room temperature and processed on the same day. 3.3. Haematology The total white cell counts (WCC) were determined with a Coulter Counter DN (Coulter Electronics, Luton, UK). Blood films were stained with Giemsa and 100 leucocytes were differentiated for every 10 X lo9 WCC 1-l. The absolute lymphocyte count was calculated by multiplying the WCC by the observed percentage of lymphocytes from the differential count. 3.4. Flow cytometry Samples were prepared as described previously (Walker et al., 1994). Briefly, whole blood was combined in separate tubes with mAbs which recognise feline Pan T + cells (Ackley and Cooper, 19921, CD4 + cells (Ackley et al., 1990a) and CD8 + cells (Klotz and Cooper, 1986) (Southern Biotechnology, Birmingham, AL, USA). Secondary staining was performed by incubating with fluorescein isothiocyanate (FITC)-conjugated goat anti-mouse immunoglobulin (Ig) G (ICN Biomedicals, Costa Mesa, CA, USA). B lymphocytes were enumerated by prewashing whole blood five times at room temperature in phosphate buffered saline (PBS)/ azide and labelling with FITC-conjugated goat anti-cat IgG (H + L) antibody (Southern Biotechnology) to detect surface Ig. Analyses were performed on a FACScan flow cytometer (Becton Dickinson, Lane Cove, N.S.W.) on the same day as cell preparation. A dotplot of forward (FSC) and side scatter (SSC) was examined and used to construct a live lymphocyte gate of 5000 cells for each antigenic marker, including the negative control. Both lymphocyte populations (small and large) were included in this collection gate (Fig. 1). Fluorescence histograms of cell numbers versus fluorescence intensity were analysed for each lymphocyte population and gates delineating the positive regions set against the negative control. 3.5. Examination of smears Blood films were made immediately after venipuncture and stained with Giemsa. For those samples which demonstrated two lymphocyte populations on FCM, 50 lympho-
14
C. Walker et al. / Veterinary Immunology and Immunopathologv 48 (1995) I I-25
cytes were counted using a modification of the battlement technique (Jain, 1986), viewing both horizontal edges and central fields by traversing the slide. Each smear was scanned to appraise the variation in lymphocyte sizes present and the lymphocytes were divided into categories based on relative size alone: small, medium and large. The actual measurements of the cells varied between samples and, depending on the area of the smear being counted, with cells towards the tail of the smear being larger than those in the body. On average, small lymphocytes were less than 9 pm, medium lymphocytes were between 9 pm and 11 pm, and large lymphocytes were greater than 11 pm in diameter. Cytoplasmic (degree of basophilia, presence of granules) or nuclear differences (shape, chromatin pattern) were not considered. 3.6. Calculations and statistics To examine the proportions of the two lymphocyte populations observed on the dotplot, the negative control was used to record cell numbers from each population. Using a commercial statistical software package (Minitab Inc., State College, PA, USA), the numbers were expressed as a percentage of the total lymphocyte population. The Pearson product moment correlation coefficient was used to compare these flow cytometric values with those obtained from examining the smears. The absolute count of each lymphocyte population was calculated by multiplying the absolute lymphocyte count by the percentage observed by FCM. The effects of health status and FIV infection on the occurrence and distribution of the two lymphocyte populations as observed by FCM were examined for statistical significance (P < 0.05) using the chi-square statistic and a Student’s C-test. The differences in lymphocyte subset distribution between the two lymphocyte populations were also analysed using a Student’s t-test. The effect of including the larger lymphocyte population in the analysis of lymphocyte subsets was investigated by comparing the percentage of cells positive for an antigenic marker in the smaller lymphocyte population alone with those positive for both the smaller and larger populations together.
4. Results 4.1. Demonstration
of dual lymphocyte populations
by FCM
Using a dotplot of FSC and SSC, 84% (57/68) of samples examined showed dual lymphocyte populations on FCM (Fig. 1). To display these pdpulations more fully, using the events gated as lymphocytes in the negative control tube, a three-dimensional plot of FSC versus SSC versus cell numbers was examined (Fig. 2). This demonstrated that the lymphocytes comprised two discrete populations of different size (based on differing FSC) but which had comparable shape and complexity (based on similar SCC). 4.2. Experimental group The cats demonstrating dual populations (57/68) included 40 castrated males with the remainder being ovariohysterectomised females. Nine were purebred cats. Their
C. Walker et af. / Veterinary Immunology and Immunopathoiogy 48 (1995) I1 -25
15
I
Fig. 1. Dotplot of forward scatter (FSC) versus side scatter populations. a, small lymphocytes; b, large lymphocytes.
health status, Clinical signs ve5; FIV- + disease (FIV-
(SSC)
demonstrating
the two lymphocyte
FIV status and mean ages for each subgroup are presented in Table 1. observed within the ‘sick’ group included chronic oral disease (FIV-ve:ll), miliary dermatitis (FIV- - ve:3; FIV- + ve:9), respiratory tract - ve:2; FIV- + ve:7), cryptococcosis (FIV- - ve:l; FIV- + ve:2), feline
Fig. 2. Characterisation of the dual lymphocyte (SSC) versus cell numbers. The two populations shape and complexity (SCC).
population using forward scatter (FSC) versus side scatter comprised cells of two different sixes (FSC) but similar cell
16
C. Walker et al. / Veterinary Immunology and Immunopathology 48 (1995) 11-25
Table 1 Data for those cats which demonstrated
dual lymphocyte
populations
Health status
FIV status
Number in group
Age (years, mean f SD)
Well
FIV-negative FIV-positive
16 4
3.5 + 3.2 7.1 L-4.1
Sick
FIV-negative FIV-positive
10 27
7.8k4.2 8.9k3.6
SD, standard
deviation.
lower urinary tract disease (FIV- - ve:l; FIV- + ve:4), renal failure (FIV- - ve:2; FIV+ ve:l) and weight loss (FIV- + ve:4). Those cats which did not demonstrate dual populations (11/68) included eight castrated males, two ovariohysterectomised and one entire female. Two were purebred. The data for these cats are presented in Table 2. The sick cats within this group displayed chronic oral disease (FIV- - ve:2; FIV- + ve:l), miliary dermatitis (FIV- + ve:l), respiratory tract disease (FIV- - ve:2; FIV- + ve:4), cryptococcosis (FIV- - ve:2), renal failure (FIV- - ve:l) and weight loss (FIV- - ve:l; FIV- + ve:2). FIV- + ve cats (31/33, 94%) were more likely to demonstrate dual lymphocyte populations than FIV- - ve cats (26/35, 75%) ( x2 = 4.839; d.f. = 1; 0.02 < P < 0.05). However, health status ( x2 = 0.426; d.f. = 1; 0.5 < P < 0.71, age ( x2 = 18.495; d.f. = 15; 0.2 < P < 0.3) gender ( x2 = 0.029; d.f. = 1; 0.8
The percentages of cells in the smaller and larger lymphocyte population seen on FCM and the percentages of cells in the various groups (small, medium and large) as counted on the corresponding blood films are presented in Table 3. When the smaller lymphocyte population, i.e. those lymphocytes with lower FSC, was compared with the combined values of the small and medium lymphocytes, as
Table 2 Data for those cats which did not demonstrate
dual lymphocyte
populations
Health status
FIV status
Number in group
Well
FIV-negative FIV-positive
4 0
4.5 f 2.7 _
Sick
FIV-negative FIV-positive
5 2
9.6k4.7 10.5 + 2.1
SD, standard
deviation.
Age (years, mean + SD)
C. Walker et al. / Veterinary Immunology and Immunopathology Table 3 Percentages
of lymphocytes
of various sizes observed
by flow cytometry
48 (1995) 11-25
and blwd
films (n = 57)
Size of lymphocyte
Mean + SD (o/o)
Flow cytometry
Low FSC (smaller) High FSC (larger)
72.7 f 14.1 27.3 f 14.1
Blood film
Small Medium Small + medium Large
18.5 56.3 72.9 25.5
SD, standard
deviation;
FSC, forward
17
k 16.8 zk 16.1 + 15.4 zk 14.2
scatter.
counted on the blood films, the correlation coefficient was 0.891. When the larger lymphocyte population, i.e. those lymphocytes with higher FSC, was compared with the large lymphocytes as counted on the blood films, the correlation coefficient was 0.803. Thus the smaller population of lymphocytes observed by FCM correlated well with small and medium lymphocytes, and the larger population with large lymphocytes observed on peripheral blood films. For the remainder of this report, ‘small’ lymphocytes will refer to the smaller group of lymphocytes as observed by FCM and the combination of small and medium lymphocytes as observed on peripheral blood films. Similarly, ‘large’ lymphocytes will refer to the larger group of lymphocytes seen on FCM and large lymphocytes seen on peripheral blood films.
4.4. The effect of health status and FIV status on the distribution lymphocytes
of small and large
Using FCM, the percentages of small and large lymphocytes were compared between well and sick cats. No significant differences in the proportion of small and large lymphocytes were found between sick and well cats (Fig. 3(a)). Similarly, the absolute numbers of neither small nor large lymphocytes differed between well and sick cats (Fig. 3(b)). This was true for both FIV- + ve and FIV- - ve cats. However, when the percentages of small and large lymphocytes were compared between FIV- + ve and FIV- - ve cats regardless of health status, FIV- + ve cats had a lower percentage (P = 0.016) of small lymphocytes and a higher percentage (P = 0.016) of large lymphocytes than FIV- - ve cats. This was due to FIV- + ve cats having significantly fewer (P = 0.0044) small lymphocytes than FIV- - ve cats, yet similar numbers of large lymphocytes. The FIV effect on lymphocyte proportions was only observed in well cats (P = 0.019), as there was no significant difference in the percentages of small and large lymphocytes between sick FIV- + ve and sick FIV- - ve cats. The possibility that this difference in lymphocyte numbers may be age-related was examined. There was statistically no difference between the mean ages of the well FIV- - ve cats and the well FIV- + ve cats (P = 0.2). Because of marked intercat variability, there were no significant differences in absolute values of either small or large lymphocytes between FIV- - ve and FIV- + ve cats after they had been grouped as ‘well’ and ‘sick’.
18
C. Walker et al. / Veterinary Immunology and Immunopathology 48 (1995) 11-25
4.5. The effect of health status alone on the distribution of antigenic markers between small and large lymphocytes Most reports of feline lymphocyte subset values have analysed the effects of immunosuppressive viral infections, e.g. FIV (Ackley et al., 1990b; Walker et al., 1994) or FeLV (Quackenbush et al., 1990; Tompkins et al., 1991), yet there have been no accounts of how lymphocyte subsets vary between health and ill-health generally. To
m
(a)
m
(bl
Small lymphocytes
We11
Sick
Largelymphocytes
Well
Sick
Small lymphocytes
Well
I
L_I
Sick
Large lymphocytes
Well
Sick
-
Fig. 3. Compatison of (a) proportions (%) and (b) absolute values (X lo9 l- ‘1 of small and large lymphocytes between well and sick FIV-negative (FIV( - 1) and FIV-positive (FIV( + )) cats (mean f standard error).
C. Walker et al. / Veterinary Immunology and Immunopathology
48 (1995) 11-25
m
Small lymphocytes -well cats
m
Small lymphocytes - sick cats
a
Large lymphocytes - well cats
B
Large lymphocytes - sick cats
19
~~
so
3
PanT+
CD4+
Fig. 4. Comparison of the subset proportions FIV-negative cats (mean +_standard error).
(%) of small
cD8+ and large
B Cell lymphocytes
for well and sick
investigate the effect of health status on the differences in distribution of antigenic markers between small and large lymphocytes, the subset percentages in FIV- - ve well cats and FIV- - ve sick cats were compared, which excluded the effect of FIV (Fig. 4). Small lymphocytes expressed a significantly greater percentage of CD4 + cells for both well cats (P = 0.037) and for sick cats (P = 0.0011) and a significantly lesser percentage of B cells for both well cats (P = 0.0026) and sick cats (P = 0.0028) than large lymphocytes. Only the distribution of CDS + cells varied with health status. In well cats, small lymphocytes demonstrated a higher percentage of CD8 + cells than large lymphocytes (P = 0.0084) but there was no difference in the distribution of CD8 + cells between small and large lymphocytes in sick cats. As with the absolute numbers of small and large lymphocytes, there were no differences between well and sick FIV- - ve cats in the absolute values of any subset of either small or large lymphocytes (data not shown). 4.6. The effect of FIV infection and large lymphocytes
on the distribution
of antigenic
markers between small
The differences in percentages and absolute values of each antigenic marker between small and large lymphocytes were analysed for all FIV- + ve cats and compared with the differences of each antigenic marker between small and large lymphocytes for FIV- - ve cats (Figs. 5(a) and 5(b)). In FIV- - ve cats, small lymphocytes had a greater proportion of Pan T + cells than large lymphocytes (P = 0.0211, but the converse was seen in FIV- + ve cats (P = 0.011) (Fig. 5(a)). The small lymphocytes of FIV- - ve cats had a significantly higher proportion of Pan T + cells than the small lymphocytes of FIV- + ve
20
C. Walker et al. /Veterinary
LI
Immunology and Immunopathology
48 (1995) 11-25
/--
I
PanT+
cD8+
CD4+
B Cell
la)
m
Small lymphocytes - FIV-negative
cats
m
Small lymphocytes - FIV-positive
cats
m
Large lymphocytes - FIV-negative
cats
B
Large lymphocytes - FlV-pcsitive
cats
(b)
PanT+
cD4+
CD8+
B Cell
Fig. 5. Comparison of (a) the subset proportions (%I and (b) the absolute values (X lo9 I-‘) of subsets of small and large lymphocytes between well and sick FIV-negative cats and well and sick FIV-positive cats (mean f standard error).
cats (P = 0.019) and the large lymphocytes of FIV- - ve cats had a significantly lower proportion of Pan T + cells than FIV- + ve cats (P = 0.016) (Fig. 5(a)). FIV- + ve cats had significantly fewer small lymphocytes (P = 0.002) which were Pan T + than FIV- - ve cats, but had similar numbers (P = 0.44) of large Pan T + lymphocytes (Fig. 5(b)).
C. Walker et al. / Veterinary Immunology
and Immunopathology
48 (1995) 11-25
21
In both FIV- - ve and FIV- + ve cats, small lymphocytes had proportionately more CD4 + cells than large lymphocytes (P = 0.0002 and P = 0.0017 respectively). Both small and large lymphocytes of FIV- - ve cats displayed proportionately more (P = 0.001 and P = 0.0021 respectively) cells positive for the CD4 marker than the corresponding lymphocytes of FIV- + ve cats (Fig. 5a). However, when absolute numbers of lymphocytes were examined, it was only for small lymphocytes that FIV- + ve cats had fewer CD4 + cells (P = 0.0001) as there was no difference in the numbers of large CD4 + lymphocytes between FIV- + ve and FIV- - ve cats (P = 0.17) (Fig. 5(b)). Unlike FIV- - ve cats, FIV- + ve cats displayed no difference in the percentage of CD8 + cells between small and large lymphocytes, regardless of health status. There were no differences in the proportions of cells which were CD8 + between FIV- - ve and FIV- + ve cats for either small or large lymphocytes (Fig. 5(a)). Again, FIV- - ve cats had significantly more small CD8 + lymphocytes than FIV- + ve cats (P = 0.0024) but there was no difference in the absolute number of large CD8 + lymphocytes (P = 0.36) (Fig. 5(b)). The FIV- + ve, but not the FIV- - ve cats had a higher CD4:CD8 ratio for small lymphocytes compared with large lymphocytes (P = 0.0054). The finding that small lymphocytes had a lower percentage of B cells than large lymphocytes (P = 0.0001) in FIV- - ve cats was not observed in FIV- + ve cats. The small lymphocytes of FIV- + ve cats had a significantly higher percentage of B cells (P = 0.014) and the large lymphocytes a significantly lower percentage (P = 0.012) than those of FIV- - ve cats (Fig. 5(a)). There were no differences between FIV- + ve and FIV- - ve cats in the absolute numbers of B cells for either small or large lymphocytes (Fig. 5(b)). 4.7. The effect of health status on the distribution and large lymphocytes in FIV-positive cats
of antigenic
markers between small
In the FIV- + ve cats which were well, there were no differences in subset distribution between small and large lymphocytes (Fig. 6). However, in the FIV-+ ve cats which were sick, the small lymphocytes had a significantly lower percentage (P = 0.011) of PanT + cells, a higher percentage of CD4 + lymphocytes (P = 0.0017) and higher CD4:CD8 ratios (P = 0.0002) but no difference in the distribution of CD8 + or B cells when compared with the large lymphocytes. As with the FIV- - ve cats, there were no differences between well and sick FIV- + ve cats in the absolute values of any subset of either the small or large lymphocyte (data not shown). 4.8. Significance
of including both lymphocyte populations
in lymphocyte
subset analysis
Conventionally in subset analysis of human lymphocytes, there is only one population of lymphocytes, represented usually by a discrete tight population of cells with low FSC and SSC. However, as this study has shown, in feline samples there are commonly two discrete lymphocyte populations. Reports of feline subset analysis in the literature do not discuss these dual populations, or in fact which lymphocyte population provided the data. The data from all cats in this study were analysed to examine the effect on the total subset values of including the group of large lymphocytes. The values obtained for
C. Walker et al. / Veterinary Immunology and Immunopathology 48 (1995) 11-25
22
m
Small lymphocytes - well cats
m
Small lymphocytes
- sick cats
L_i
Large lymphocytes -well cats
B
Large lymphocytes
- sick cats
.90 r
3 d
4s
i
32
2 .o B
16.
2 4 0
PanT+
cD4+
Fig. 6. Comparison of the subset proportions FN-positive cats (mean f standard error).
(o/o) of small
CD8+ and large
B Cell lymphocytes
for well
and sick
the small lymphocyte population, which comprised the majority of lymphocytes, were compared with those obtained from analysis of the small and large lymphocyte populations together. There was no significant difference for any lymphocyte subset between the two groups regardless of FIV infection or health status (data not shown).
5. Discussion In this study, the flow cytometric observation of two discrete lymphocyte populations, differing in size, was a common finding. Large lymphocytes comprise only 2% of total lymphocytes in humans (Douglas, 1983) which may explain why there are no flow cytometric reports of dual populations of lymphocytes based on variations in their FSC. It must remain only an assumption that both the populations are in fact lymphocytes, and not, for instance, monocytes. However, the good correlation between FCM results and examination of blood films observed in this study would support both populations being lymphocytes. Moreover, it has been reported in the literature that none of the feline mAbs used in our study reacted with monocytes (or granulocytes, erythrocytes or platelets) when isolated cell populations were examined for antigen expression (Klotz and Cooper, 1986; Ackley et al., 1990a; Ackley and Cooper, 1992). FIV infection appeared to cause a significantly higher percentage of large lymphocytes due to a reduction in small lymphocyte numbers but no alteration in large lymphocyte numbers (Fig. 3(b)). This suggested that FIV affects small lymphocytes rather than large lymphocytes. It is not known what the functional differences between small and large lymphocytes are in cats. In humans, larger lymphocytes often represent
C. Walker et al. / Veterinary Immunology and Immunopathology
48 (1995) 11-25
23
activated cells (Weiss, 1993), so perhaps it is resting cells which are depleted by FIV infection, resulting in proportionately more activated lymphocytes. This FIV effect appeared to be negated by the onset of illness. In humans, the proportion of large lymphocytes (of unreported phenotype) is seen to increase with acute viral illness (Douglas, 1983). Th e sick cats in this study had chronic disease and perhaps this explains why there was no significant difference between healthy and sick cats in their proportions of small and large lymphocytes. In humans, variations in SSC (reflecting cell structure) have been observed and utilised to analyse differences in subset distribution (Terstappen et al., 1986; Rabinowitz et al., 1992). Using similar methodology, this study has demonstrated differences in distribution of CD4 + , CD8 + and B cells between small and large lymphocytes in cats. The study was able to demonstrate the effect of ill-health alone, although numbers precluded subset analysis among specific disease states. The significance of CD8 + cells alone varying proportionally with health status between the two lymphocyte groups is not known. It is possible that the percentage of large lymphocytes expressing the CD8 + marker tends to increase in response to illness, whilst the distribution of CD4 + and B cells remains similar to that found in health. These differences may reflect alterations in subpopulations of CD8 + cells. There were significant differences between FIV- + ve and FIV- - ve cats in the distribution of lymphocyte subsets between small and large lymphocytes. The differences between FIV- - ve and FIV- + ve cats occurring in this study in subsets of the small lymphocytes in this study concur with those of other workers (Ackley et al., 1990b; Novotney et al., 1990; Barlough et al., 1991; Tompkins et al., 1991; Torten et al., 1991; Hoffmann-Fezer et al., 1992). However, the finding that the subsets of large lymphocytes appear unaffected by FIV infection is novel. The fact that there were no differences in the distribution of CD8 + cells between small and large lymphocytes, in either well or sick FIV- + ve cats, suggests that, even when well, FIV- + ve cats have more CD8 + cells which are larger (and thus possibly activated). Perhaps these CD8 + cells have a cytotoxic function. It is well accepted that human immunodeficiency virus (HIV) causes B cell activation in all stages of infection (Mizuma et al., 1988) and the B cells of HIV-positive individuals are larger than those of uninfected controls (Martinez-Maza et al., 1987). The significance of the variation in subset distribution between large and small lymphocytes and the involvement of FIV infection has yet to be clarified. It is not known whether activated B cells in the cat are larger than resting B cells as they are in other species (Kincade and Gimble, 1993). If activated B cells are larger than resting B cells and FIV causes B cell activation, it may have been expected that FIV- + ve cats would have had more larger B cells (either relatively or absolutely) than FIV- - ve cats. This was not found. Thus it would seem from this study that there is no evidence, based on variation in cell size between FIV- - ve and FIV- + ve cats, of B cell activation in FIV- + ve cats. It may be possible to explore this further by double staining lymphocytes for T or B cell markers and an FIV-specific marker (Steinman et al., 1990) to investigate the proportion of infected cells in the two lymphocyte populations. mAbs which act as activation markers, as are available for human T and B lymphocytes (Formenti et al.,
24
C. Walker et al. / Veterinary Immunology and Immunopathology 48 (I 995) 11-25
19891, would be useful to determine whether the various sizes of feline lymphocytes represent differing stages of activation. Also, the development of mAbs specific for immature and mature cell types would help clarify the significance of small and large lymphocytes. It is unknown whether or not small lymphocytes are more mature than large lymphocytes in cats as they are in other species (Beagley et al., 1989). Similarly, size differences and differences in lymphocyte subset distribution between the small and large lymphocytes give no indication of the functional differences either in vitro or in vivo which may exist between the two populations. Given that large lymphocytes comprise a considerable proportion of feline circulating lymphocytes, and have been shown to have significant differences in lymphocyte subset distribution, it is important to consider their inclusion or exclusion in routine flow cytometric analysis. This study demonstrated that, because the absolute numbers of large lymphocytes in feline samples are often much lower than those of small lymphocytes, their effect overall on analysis of lymphocyte subsets is statistically insignificant. Thus, for routine subset analysis, inclusion or exclusion of the large lymphocyte population in the analysis gate will not affect the overall results. However, further investigation of the feline immune system, especially for particular disease states, e.g. FIV infection, and health status should involve examination of both lymphocyte populations by FCM.
Acknowledgements This work was supported in part by a Faculty of Veterinary Science grant. The authors thank the clinical pathology laboratory for use of their facilities. Professsor Wayne Robinson kindly provided the FIV antigen preparation for the Western blots. Thanks also to Christine Bligh and Becton Dickinson for advice and materials for flow cytometry, and Dr. John Quinn, from Westmead Hospital, for advice on B cell staining. The FeLV tests were generously supplied by Commonwealth Serum Laboratories. Thanks also to the many veterinary practitioners who referred cases to us for this study.
References Ackley, C.D. and Cooper, M.D., 1992. Characterization of a feline T-cell-specific monoclonal antibody reactive with a CDS-like molecule. Am. J. Vet. Res., 53: 466-471. Ackley, C.D., Hoover, E.A. and Cooper, M.D., 1990a. Identification of a CD4 homologue in the cat. Tissue Antigens, 35: 92-98. Ackley, C.D., Yamamoto, J.K., Levy, N., Pedersen, N.C. and Cooper, M.D., 1990b. Immunologic abnormalities in pathogen-free cats experimentally infected with feline immunodeficiency virus. J. Virol., 64: 5652-5655. Anderson, L., Wilson, R. and Hay, D., 1971. Haematological values in normal cats from four weeks to one year of age. Res. Vet. Sci., 12: 579-583. Barlough, J.E., Ackley, C.D., George, J.W., Levy, N., Acevedo, R., Moore, P.F., Rideout, B.A., Cooper, M.D. and Pedersen, N.C., 1991. Acquired immune dysfunction in cats with experimentally induced feline immunodeficiency virus infection: comparison of short-term and long-term infections. J. Acq. Immun. Defic. Syndr., 4: 219-227. Beagley, K.W., Eldridge, J.H., Lee, F., Kiyono, H., Everson, M.P., Koopman, W.J., Hirano, T., Kishimoto, T.
C. Walker et al. / Veterinary Immunology and Immunopathology
48 (1995) 11-25
25
and McGhee, J.R., 1989. Interleukins and IgA synthesis. Human and murine interleukin 6 induce high rate IgA secretion in &A-committed B cells. J. Exp. Med., 169: 2133-2148. Douglas, SD., 1983. Morphology of lymphocytes and plasma cells. In: W.J. Williams, E. Beutler, A.J. Erslev and M.A. Lichtman (Editors), Haematology. McGraw-Hill, New York, pp. 883-902. Formenti, SC., Turner, R.R., de Martini, R.M., Boone, D.C., Bishop, P.C., Levine, A.M. and Parker, J.W., 1989. Immunophenotypic analysis of peripheral blood leukocytes at different stages of HIV infection. Am. J. Clin. Pathol., 92: 300-307. Gilmore, C.E., Gilmore, V.H. and Jones, T.C., 1964. Bone marrow and peripheral blood of cats: technique and normal values. Pathol. Vet., 1: 18-40. Hoffmann-Fezer, G., Thum, J., AckIey, C., Herbold, M., Mysliwietz, J., Thefeld, S., Hartmann, K. and Kraft, W., 1992. Decline in CD4-positive cell numbers in cats with naturally acquired feline immunodeficiency virus infection. J. Virol., 66: 1484-1488. Jain, N.C., 1986. Schalm’s Veterinary Haematology. Lea&Febiger, Philadelphia, 1221 pp. Jain, N.C., 1993. Essentials of Veterinary Haematology. LeaBtFebiger, Philadelphia, 417 pp. Kincade, P.W. and Gimble, J.M., 1993. B lymphocytes. In: W.E. Paul (Editor), Fundamental Immunology. Raven Press, New York, pp. 43-73. Klotz. F.W. and Cooper, M.D., 1986. A feline thymocyte antigen defined by a monoclonal antibody (FI’2) identifies a subpopulation of non-helper cells capable of specific cytotoxicity. J. Immunol., 136: 2510-2514. Martinez-Maza, O., Crabb, E., Mitsuyasu, R.T., Fahey, J.L. and Giorgi, J.V., 1987. Infection with the human immunodeficiency virus (HIV) is associated with an in vivo increase in B lymphocyte activation and immaturity. J. Immunol., 138: 3720-3724. Mizuma, H., Litwin, S. and Zoller-Pazner, S., 1988. B-cell activation in HIV infection: relationship of spontaneous immunoglobulin secretion to various immunological parameters. Clin. Exp. Immunol., 71: 410-416. Novotney. C.. English, R.V., Housman, J., Davidson, M.G.. Nasisse, M.P., Jeng, C.R., Davis, W.C. and Tompkins. M.B., 1990. Lymphocyte population changes in cats naturally infected with feline immunodeficiency virus. AIDS, 4: 1213-1218. Penny. R.H.C., Carlisle, C.H. and Davidson, H.A., 1970. The blood and marrow picture of the cat. Br. Vet. J.. 126: 459-454. Quackenbush, S.L., Donahue, P.R., Dean, G.A., Myles, M.H.. Ackley, C.D., Cooper, M.D., Mullins, J.I. and Hoover, EA., 1990. Lymphocyte subset alterations and viral determinants of immunodeficiency disease induction by the feline leukaemia virus FeLV-FAIDS. 3. Virol., 64: 5465-5474. Rabinowitz, R.. Granot, E., Deckelbaum, R. and Schlesinger, M., 1992. Antigenic differences between subsets of peripheral blood lymphocytes differing in their right angle light scatter in flow cytometric analysis. Int. Arch. Allergy Immunol., 97: 200-204. Steinman. R.. Dombrowski, J., O’Connor, T., Monteolaro, R.C., Tonelli, Q., Lawrence, K., Seymour. C.. Goodness. J., Pedersen, N.C. and Andersen, P.R., 1990. Biochemical and immunological characterization of the major structural proteins of feline immunodeficiency virus. J. Gen. Virol.. 71: 701-706. Terstappen, L.W.M.M., de Grooth. B.G., Nolten, G.M.J., ten Napel. C.H.H.. van Berkel, W. and Greve, J., 1986. Physical discrimination between human T-lymphocyte subpopulations by means of light-scattering, revealing IWO populations of TB-positive cells. Cytometry, 7: 178-183. Tompkins, M.B., Pang, V.F., Michaely, P.A.. Feinmehl, R.I., Basgall, E.J., Baszler, T.V., Zachary, J.F. and Tompkins. W.A.A., 1989. Feline cytotoxic large granular lymphocytes induced by recombinant human IL-2. J. Immunol., 143: 749-754. Tompkins, M.B., Nelson, P.D., English, R.V. and Novotney, C., 1991. Early events in the immunopdthogenesis of feline retrovirus infections. J. Am. Vet. Med. Assoc.. 199: 1311-1315. Torten, M., Franchini, M., Barlough, J.E., George, J.W., Mozes, E., Lutz, H. and Pedersen, N.C., 1991. Progressive immune dysfunction in cats experimentally infected with feline immunodeficiency virus. J. Virol., 65: 2225-2230. Walker. C.. Canfield, P.J. and Love, D.N., 1994. Analysis of leucocytes and lymphocyte subsets for different clinical stages of naturally-acquired feline immunodeficiency virus infection. Vet. Immunol. Immunopathol., 44: I-12. Weiss. A., 1993. T lymphocyte activation. In: W.E. Paul (Editor). Fundamental Immunology. Raven Press, New York, pp. 467-504.