Hematological Profiles in Dairy Cows Treated with Recombinant Bovine Somatotropin1r2 J. L. Burton*, B. W. McBride*#3, B. W. Kennedy*, J. H. Burton*, T. H. Elsassel.4, and B. Woodward+ Department of *Animal and Poultry Science and +Nutritional Sciences, University of Guelph, Guelph, Ontario, Canada N1G 2Wl
Key Words: Somatotropin, Hematology, Leukocytes, Insulin-Like Growth Factor I, Dairy Cows J. Anim. Sci. 1992. 70:1488-1495
Introduction Recombinant bovine somatotropin (bST) is galactopoietic when injected into dairy cows (reviewed by McBride et al., 1988). Kelley (1989)has reviewed further roles for somatotropin (ST)in hematopoiesis and maintenance of thymic cellularity and function in ST-deficient animals. Blood concentrations of ST influence liver production of insulin-like growth factor I (IGF-I), which is
'The authors express their appreciation to the staff of the Clinical Pathology Laboratory (Ontario Veterinary College, University of Guelph) for performing hematology and leukocyte differentials and acknowledge the Ontario Ministry of Agriculture and Food and the Natural Sciences and Engineering Council of Canada for support of this work. %ecombinant bovine somatotropin was donated by American Cyanamid (Princeton, NJ). 3To whom correspondence should be addressed. 4USDA, ARS, Ruminant Nutr. Lab., Beltsville, MD 20705. Received August 16, 1991. Accepted December 20, 1991.
elevated in the blood of ST-treated animals (reviewed by McBride et al., 1988).There is evidence that specific IGF-I receptors are present on cells derived from the hematopoietic system (reviewed by Burton, 1991); therefore, IGF-I may be partially responsible for effects of injected ST on these cell types (Gruler et al., 1988; Kelley, 1989). Effects of bST on certain hematological variables have been reported (Eppard et al., 1987; Solderholm et al., 1988; Phipps, 1989; Annexstad et al., 1990; Burton et al., 1990; McGuffey et al., 1990; Vicini et al., 1990). Consistent across trials was the observation of decreased hematocrits in cows treated with 2 20.6 mg of bST/d. Relatively few studies have been reported, however, that determine the effects of long-term administration of bST in dairy cows in previous and current lactations, and(or1 of blood IGF-I concentrations, on detailed hematological profiles obtained in the current lactation. These were the objectives of the present study.
1488
Downloaded from https://academic.oup.com/jas/article-abstract/70/5/1488/4705328 by 04860000 user on 12 January 2019
tered hematocrits (P = .lo), mean corpuscular volume (P = .03),mean corpuscular hemoglobin (P = .009), and fraction of blood lymphocytes (P = .09). A week x bST treatment interaction also contributed to variation in hematocrit (P = .09), mean corpuscular hemoglobin (P = .05),fraction of neutrophils (P = .O2), and fraction of lymphocytes (P = .04). Blood IGF-I concentration influenced fractions and counts of neutrophils (P = .06, .09), lymphocytes (P = .04, .02), and monocytes (P = .08, .05). Further in vivo and in vitro studies are warranted because this preliminary evidence s u g gests that bST and(or1 IGF-I contribute to regulation of hematopoiesis in mature dairy cows.
ABSTRACT: Recombinant bovine somatotropin (bST) was administered at 0, 10.3, or 20.6 mg per cow per day to 32 Holsteins for 38 wk. Fifteen currently treated cows had been treated in the previous lactation. Eighteen hematological variables and blood concentration of insulin-like growth factor I (IGF-I) were measured at five sample periods. The objectives of the study were to test the effects of bST treatments on hematological profiles and to relate blood IGF-I concentration to these variables. Results indicated little influence (P > .lo1 of previous bST treatments on hematological profiles measured in the current lactation. Current bST treatments, however, al-
BST AND HEMATOLOGICAL PROFILES
Materials and Methods
the fraction of a particular leukocyte type by the total number of leukocytes previously determined by the Coulter counter method. Insulin-Like Growth Factor I Concentration. The concentrations of IGF-I in blood sera were assayed by RIA, according to the procedure of Elsasser et al. (1989). These results were reported previously (Burton et al., 1991al. Statistical Analyses. Before statistical analysis, all data were log,-transformed for normality of distributions. The GLM procedure of SAS (SAS, 1982) was used for all ANOVA. A preliminary ANOVA was performed to test effects of current and previous bST treatment groupings of cows on hematological variables measured pretreatment (wk 3). Neither effect was significant ( P > .lo). Previous treatments similarly did not affect (P > .lo) variables from subsequent sample weeks, except for whole blood hemoglobin and hematocrit at wk 10 IP < .lo). However, previous treatment was not included in the model used to assess variation in lactational hematological profiles because it was confounded with a cow(treatment1 effect. The model used was as follows: Yijklm = p + weeki + seasonyrj + b l pretrmtkl + trmtk + cow(trmt)kl + week x trmtik + eijklm, where Yijklm = observed hematological/leukocyte differential measurement for cow ijklm; p = population mean; weeki = fixed sample week effect (i = 10, 26, 35, 48); seasonyrj = fixed seasonyear effect [j = winter 1987, 1988; spring 1987, 1988; summer 1987, 1988; and fall 1987, where winter = December, January, February; spring = March, April, May; summer = June, July, August; fall = September, October, November); bl = regression coefficient of Ykl on pretreatment values for cow kl; trmtk = fixed treatment effect (k = 0, 10.3, 20.6 of mg bST/d); cow(trmtIk1 = random cow-nestedwithin-bST treatment effect; wk x trmtik = fixed week x bST treatment interaction effect; and eijkh = random error term. A contrast comparing controls with all treated cows was also performed. The bST treatment and contrast effects were tested using the Type 111 mean square for cow(treatment1 as the denominator in the F-test. Pretreatment covariates were tested against the Type I mean square for cow(treatment1. All other effects were tested against the residual mean square. Possible statistical relationships between blood IGF-I concentrations and hematological variables were tested by two additional ANOVA. In one, a n IGF-I concentration covariate was added to the model previously described. If the significance of the bST treatment effect for any hematological variable was altered by this covariate, then IGF-I might have been considered part of the bST treatment effect. In the other ANOVA an IGF-I
Downloaded from https://academic.oup.com/jas/article-abstract/70/5/1488/4705328 by 04860000 user on 12 January 2019
Animals and Treatments. Thirty-two Holsteins in second or higher parity were used. The bST treatments were given daily S.C. starting between wk 4 and 5 of lactation and continued for 38 wk. Cows were assigned to bST treatment groups according to several criteria. If cows were returning to trial from a previous lactation with bST, they were put in the same treatment regimen as before. If the cows were new to bST trials, they were evenly distributed across treatment groups according to age and previous milk yield. Twelve cows were injected with sterile saline (controls), and 10 cows in each of two groups received either 10.3 or 20.6 mg of bST/d. Of the cows receiving bST, 15 had received the same treatment in the previous lactation and five (two in the 10.3 mg/d group and three in the 20.6 mg/d group) were new to the trial. Controls did not receive bST in the previous lactation. Blood Sampling Schedule. Peripheral blood was collected by tail venipuncture at 0900 of each sample period. Blood for hematological profiling was collected into 10-mL evacuated tubes (Vacu tainer, Fisher Scientific, Don Mills, ON) containing EDTA and that for IGF-I quantification was collected into 10-mL tubes lacking anticoagulant. Blood sample periods were chosen based on lactation curves from the previous bST trial (McBride et al., 1990). Three samples per cow were collected over three consecutive days in wk 3 of lactation to determine pretreatment values of each blood variable and to test potential carryover effects of bST between one lactation and the next. Additional single samples were collected at wk 10, 26, and 35 of lactation to give hematological profiles from peak milk yield through mid- and late lactation, respectively. A sample was also taken a t wk 46 of lactation to determine the effect of bST treatment termination. Hematological Profiles. Variables measured by means of a Coulter Counter (Coulter Plus 4, Coulter Electronics, Hialeah, FL) included leuko, counts (cells cyte counts (cells x r o Q / ~ )erythrocyte x 10l2/L), whole blood hemoglobin (g/L), hematocrit (%I, mean corpuscular volume (E),mean corpuscular hemoglobin (pgl, mean corpuscular hemoglobin concentration (g/L1, red cell distribution width (O/O), platelet counts (cells x 10Q/L),and mean platelet volume (E). Leukocytes were enumerated in 100-cell differential counts of blood smears stained with Wright’s stain and results were expressed as the fraction (percentage) of neutrophils, lymphocytes, monocytes, or eosinophils in the smear. Absolute counts of these leukocyte types in whole blood (cells x 10g/L) were then estimated by multiplying
1489
1490
BURTON ET AL.
Results and Discussion Pretreatment Means Pretreatment means for the 18 hematological variables are given in Table 1. Preliminary ANOVA detected no effect of previous bST treatments on these data (not shown). Similarly, current bST treatments did not affect these values (data not shown). Data in Table 1 were also within published reference ranges (Lumsden et al., 1980). Therefore, treatment groups of cows started the trial with similar, clinically normal hematological profiles.
Hematology Non-Bovine Somatotropin Effects. Fixed effects of sample week and season-year contributed to variation in erythrocyte count ( P = .09, .lo), mean corpuscular volume ( P = .01, .07),and mean corpuscular hemoglobin ( P = .0001, .007).Red cell distribution width was affected by season-year ( P = .O2) but not by sample week. The pretreatment covariate was significant for the above red cell variables ( P e .001), as well as for hematocrit ( P = .08), mean platelet volume ( P = .00051, and leukocyte count ( P = .01). Cow(treatment1, however, was the main source of variation for leukocyte count ( P = .0001), erythrocyte count ( P = .003),hematocrit ( P = .lo), mean corpuscular volume ( P = .00011, mean corpuscular hemaglobin ( P = .00011, red cell distribution width ( P = .0006), and platelet count ( P = .05). The model was relatively inefficient in explaining variation in whole-blood hemoglobin and mean corpuscular hemoglobin concentration (model R2 values were .47 and .41, respectively). Bovine Somatotropin Effects. The bST treatment influenced hematocrit ( P = .lo), mean corpuscular volume ( P = .03), and mean corpuscular hemoglobin (P = .0091. Pooled (wk 10 to 461 treatment least squares means showed 5.8% (P = .03)and 4.9% ( P = .06)decreases in hematocrit for the 20.6 mg/d group vs controls and vs the 10.3 mg/d group,
Table 1. Pretreatment mean and bST treatment least squares mean hematological values Hematological variable Hematology Leukocyte count, cells x 10s/L Erythrocyte count, cells x 1Ol2/L Whole blood hemoglobin, g/L Hematocrit,O/O Mean corpuscular volume, fL Mean corpuscular hemoglobin, pg Mean corpuscular hemoglobin concentration, g/L Red cell distribution width, o h Platelet count, cells x iOg/L Mean platelet volume, fL Leukocyte differential Fraction, O/O, of Neutrophils Lymphocytes Monocytes Eosinophils Counts, cells x I O ~ / L ,of Neutrophils Lymphocytes Monocytes Eosinophils
Pretreatment
Pooled bST Treatment LS Meansb
Meana
SD
7.43 5.67 99.48 28.00 49.83 17.64 354.19 18.88 548.07 5.94
2.40 .63 8.40 2.35 3.32 1.20 6.67 1.12 103.57 .49
40.00 51.00 5.33 5.00
9.00 9.32 4.79 3.87
2.94 3.77 .33 .39
1.07 1.44 .23 .38
0 mg/d 7.39 6.05 102.51 26.94' 44.70' 16.28' 357.81 17.46 387.61 5.99
&Means of three replicates per cow and pooled over bST treatment groups. bLeast squares means pooled over wk 10 to 46. 'VdBeLeast squares means within a row with different superscripts differ ( P c .05). f,gLeast squares means within a row with different superscripts differ (P < .io).
4 1.90 49.16' 1.87 4.6QCf 2.97 3.90 1 .oo .35
10.3 mg/d
20.6 mg/d
8.00 6.17 102.51 28.65c*dsf 5 1 .42d 17.46d 354.25 17.81 454.86 6.11
7.85 6.05 96.54 27.25& 44.26' 15.80e 354.25 17.99 415.72 5.93
42.32 47.24' 2.26 3.47cpd.g
49.16 38.67d 3.05 3.14d
3.53 3.32
3.49 3.78
.68
SO
.27
.24
Downloaded from https://academic.oup.com/jas/article-abstract/70/5/1488/4705328 by 04860000 user on 12 January 2019
covariate was included and the bST treatment effect was deleted from the previously described model. This analysis allowed determination of the influence of blood IGF-I concentration on hematological variables without adjustment for bST treatment effects. Unless otherwise stated, a significance level of P e .10 was assumed. Least squares means in tables and figures were converted from loge values to original units of measurement, so standard errors of means were not reported. Indicated significance levels of paired t-tests between estimated means, however, were based on analysis of loge-transformed data.
1491
BST AND HEMATOLOGICAL PROFILES
respectively (Table 1). Preliminary ANOVA indicated that previously treated cows had lower ( P e .lo) hematocrits at wk 10 than previously untreated cows; therefore, possible bias in the estimation of the current bST treatment effect was acknowledged. Such bias may have been negligible, however, because current bST treatments did not affect hematocrit until wk 26 (Figure 1). The week x bST treatment interaction affected (P = .09) hematocrit. Interaction least squares means demonstrated 1 1.3%( P = .009) and 1 1.7% (P = .01) lower hematocrits at wk 26 and 35, respectively, for the 20.6 mg/d group vs controls, with no differences detected between controls and the 10.3 mg/d group (Figure 1). Values for the 20.6 mg/d group increased to control values after treatments were terminated (Figure 1). Results were consistent with numerous other observations of decreased hematocrits in cows treated with high doses (5 20.6 rng/dl of bST (Eppard et al., 1987; Solderholm et al., 1988; Phipps et al., 1989; Annexstad et al., 1990; Burton et al., 1990; McGuffey et al., 1990; Vicini et al., 19901. The period of maximum effects of bST on milk yield in the current lactation coincided with detectable effects on hematocrit. During mid- to late lactation milk yield was increased by approximately 25% for the 20.6 mg/d group vs controls (Gibson, McBride, and Burton; unpublished observations). It is possible, therefore, that decreased hematocrits resulted from increased plasma volume in response to increased milk output, as was suggested by Phipps (19891 and by Annexstad et al. (19901. Plasma volume, however, has not been estimated directly in the reported studies. Because plasma volume may influence concentrations of all cell types in whole blood, investigation of effects of bST on plasma volume in treated cows is indicated.
34i 32 ..
-
1
0 Controls 10.3 rng
30 1
bST/d
28 0
0
20.6 rng bST/d
26
E 24
20
10
26
35
Week of Lactation
46
Figure 1. Week x bST treatment least squares means for hematocrit, with treatment differences shown (bars within a week with different letters differ [P < .05]]. A difference also occurred between wk 35 and 46 (P = .08) for the 20.6 mg/d group.
Indirect evidence from the present study suggests that decreased hematocrits in cows treated with 20.6 mg bST/d may indeed be more meaningful than simple increases in blood volume in response to increased milk output. For example, hematocrit and erythrocyte count were positively correlated at all sample periods for the 20.6 mg/d group (r = .92 [P < .011, .63, .67, and .64 [P e .lo1 a t wk 10, 26, 35, and 46, respectively). In addition, there was a 13% increase (P = .00011 in mean corpuscular volume in the 10.3 mg/d vs the average of the other groups (Table 2). Furthermore, a 6.8% increase (P = .0009) in the 10.3 mg/d group and a 3.0% decrease (P = .0006) in the 20.6 mg/d group relative to controls was observed for mean corpuscular hemoglobin, and the 10.3 mg/d
Table 2. Analysis of variance of leukocyte differential variables Tvoe 111 s u m s of sauares F-test P-value for Independent variable
Fraction df
Neut'
Lymph
(oh)
of
MonoC
Eosind
Counts (cells
x 1oB/L)
Lymp
Mono
.13 .16 .59
SO
.18 .49 .26
.oo1
.08
.05
.19 .27 .18 .64
.14 .40 .19 .60
31 .71 .24 .55
Neut
of Eosin
~
Week Season-year Pretreatment covariate Cow(treatment1 Week x treatment bST treatment Contraste Model R2
3 6 1 29 6 2 1
.03
.10
.09 .22 .01 .02 .12 .24
.OB .39 ,008 .04 .09
.65
.58
.05
'Neutrophils. bLymphocytes. 'Monocytes. dEosinophils. eContrast compared controls with all bST-treated cows.
.12 .15 .47 .o 1 .6 1 .79 .12 .55
.14
.03 .05 .06
.63 .11 .05 .64
.63 ,001
.09 .07 .15 .03 .22 .22 .ll .67
Downloaded from https://academic.oup.com/jas/article-abstract/70/5/1488/4705328 by 04860000 user on 12 January 2019
22
BURTON ET AL.
1492
65T
b
65i
_ .60
b
v
0 Controls Is1 10.3 rng bST/d
55 50 0 LL 45 40 35 L 30 3 25 _20_
a
c
0 Controls
0 'E
2
::
46
10
26 35 46 Week of Lactation
Figure 2. Week x bST treatment least squares means for fraction of neutrophils with treatment differences shown (bars within a week with different letters differ [P < .OS]). Differences also occurred between wk 35 and 46 (P = .04) for the 10.3 mg/d group and between wk 10 and 26 (P = .lo), 10 and 35 (P = .08), 26 and 46 (P = .09), and 35 and 46 ( P = .03) for the 20.6 mgld group.
Figure 3. Week x bST treatment least squares means for fraction of lymphocytes with treatment differences shown (bars within a week with different letters differ [ab P < .05; ef P c .lo]). Differences also occurred between wk 35 and 46 (P = .01) for the 10.3 mgld group and between wk 26 and 46 (P = .07), and 35 and 46 (P = .06) for the 20.6 mg/d group.
group had 9.5% higher (P = .0001) values than the 20.6 mg/d group (Table 2).Vicini et al. (19901 also found decreased hematocrits and erythrocyte counts after administering two massive doses of bST into pregnant cows. Some contribution of altered erythropoiesis to hematocrit in bST treated cows, therefore, cannot be ruled out.
similar effect of bST on eosinophils after administration of two massive doses of bST, but others have detected no average effects of bST on lymphocytes or eosinophils (Annexstad et al., 1990; McGuffey et al., 1990). It may be argued that the absolute counts of leukocyte types, rather than their fractions, are important during physiological response to infection. There were, however, no differences (P > .lo) in incidence of naturally occurring infectious diseases among the present groups of cows, and cows from each treatment group remained relatively healthy throughout the trial (Burton, 1991). Therefore, physiological response to infection under long-term bST administration could not be determined from results of the present study. For several reasons, however, effects of bST on blood leukocytes per se may have been more appropriately determined from fraction as opposed to count data in this study. First, absolute counts of leukocyte types were estimated by multiplying fraction data by total leukocyte counts on each sample week. Large variation existed among cows for total leukocyte count (7.9 +_ 2.2 [pooled mean f SDI), which contributed to large variation in counts of the various leukocyte types. Second, effects of bST on blood plasma volume have not been determined conclusively. Alterations in plasma volume may affect leukocyte counts of whole blood but would have no effect on differential counts from blood smears. Third, because leukocyte differential data were collected
Leukocyte Differentials Non-Bovine Somatotropin Effects. Compared with the hematological data discussed previously, sample week, season-year, and the pretreatment covariate contributed relatively little to variation in blood leukocyte profiles (Table 2). However, cowttreatmentl was a major source of variation for these variables (Table 2). Bovine Somatotropin Effects. The main effect of bST treatment did not influence absolute counts of any leukocyte type in whole blood, although there was a tendency (P = .091 for the lymphocyte fraction to be affected (Table 2). Furthermore, the contrast effect was significant for both lymphocyte (P = ,051and eosinophil fractions (P = .05;Table 2). Pooled treatment least squares means showed decreases of 21.3% (P = .002) and 33% (P= .02) for lymphocyte and eosinophil fractions, respectively, in the 20.6 mg/d group vs controls (Table 1). There was also an overall decrease of 26% (P = .09)for eosinophil fraction in the 10.3 mg/d group vs controls (Table 1). Vicini et al. (1990) found a
Downloaded from https://academic.oup.com/jas/article-abstract/70/5/1488/4705328 by 04860000 user on 12 January 2019
35
1493
BST AND HEMATOLOGICAL PROFILES
significant biological effects of bST on granulopoiesis and cellular activation are possible, and further studies to test this are warranted. Lymphocyte fraction was also influenced by the week x bST treatment interaction (P = .04; Table 2). Relative to controls, the 20.6 mg/d group had 32.7% (P = .0008)and 33.3% (P = .002)decreases in lymphocyte fraction at wk 26 and 35,respectively (Figure 3). By wk 35, the 10.3 mg/d group had a 16.8% (P = .061 decrease in this variable compared to controls. When treatments were terminated, values increased ( P e .lo) in both treated groups such that they were not different from control values by wk 46 (Figure 3). There are few published data with which to compare present lymphocyte results. This, combined with the fact that health status of the present cows was not different among bST treatment groups and was generally good, made interpretation of present results difficult. It may be speculated, however, that the decreased fraction of lymphocytes did not reflect their biological activity, because peripheral blood lymphocytes from all treated cows displayed elevated ( P < .01) proliferative responsiveness to mitogen in culture (Burton et al., 1991a1, and because there were increased (P e .05)concentrations of immunoglob-
Table 3. Comparison of Type I11 sums of squares P-values for the bST treatment effect and an IGF-I concentration covariate from three ANOVA of hematological variables Blood variable Hematology Leukocyte count, cells x IOO/L Erythrocyte count, cells x 10l2/L Whole blood hemoglobin, g/L Hematocrit, YO Mean corpuscular volume, fL Mean corpuscular hemoglobin, pg Mean corpuscular hemoglobin concentration, g/L Red cell distribution width, YO Platelet count, cells x IOO/L Mean platelet volume, fL Leukocyte differential Fraction, oh, of Neutrophils Lymphocytes Monocytes Eosinophils Counts, cells x I O ~ / L ,of Neutrophils Lymphocytes Monocytes Eosinophils
Model l a
Model
bST effect
bST effect
.81 .81 .17 .10 .03 .01 .55 .95 .58 .65
.73 .84 .23 .15
.12 .09
ab
Model 3c
IGF-I covar.
IGF-I
COVW.
9 1 .70 .54 .41 .20 .20 .86 .76 .11 .74
.Q1 .70 .54 .42 .24 .26 .86 .75 .ll .74
.OB
,7Q .ll
.52 .41 .46 .17
.OB .04 .08 .72
.27 .40 .71 .22
.74 .85 .50 .14
.10 .03 .06 .84
.04
.01 .48 .e7
.40 .63
.05
.OQ .71
aThe original model, including the fixed bST treatment effect. bAn IGF-I concentration covariate was added to the original model. ' A n IGF-I concentration covariate was added and the bST treatment effect ignored from the original model
.OQ .02 .05 .85
Downloaded from https://academic.oup.com/jas/article-abstract/70/5/1488/4705328 by 04860000 user on 12 January 2019
before, during, and after the bST treatment period, more detailed assessment of bST effects on fraction data could be made by relating such effects to the initiation and termination of bST treatments. Information of this type was gained by including a week x bST treatment interaction in the statistical model previously described. Indeed, ANOVA showed that the week x bST treatment interaction influenced fraction of neutrophils (P = .021 and lymphocytes (P = .04; Table 2). Interaction least squares means for the control group were relatively invariant across sample week, whereas those for treated cows were reactively more variable (Figures 2 and 3). Fraction of neutrophils in the 20.6 mg/d group rose by 27.4% ( P = .002)and 29.8% (P = .002) at wk 26 and 35, respectively, relative to control values (Figure 2). However, the dose of 10.3 mg of bST/d did not affect neutrophil fraction. When bST treatments were terminated, neutrophil fraction for the 20.6 mg/d group dropped dramatically and reached control values by at least wk 46 (Figure 2). These results are consistent with the notion that ST is granulopoietic (reviewed by Kelley, 1989).Furthermore, there has been some indication that ST is capable of augmenting superoxide anion production in neutrophils (Fu et al., 1991). Therefore,
BURTON ET AL.
1494
ulins G and G2 in the blood of the 10.3 mg/d group (Burton et al., l991b). The practical implications of these combined observations will only be clarified with further investigation of bST-lymphocyte interactions and with specific infectious disease challenges of bST-treated cows.
Relationship Between Insulin-Like Growth Factor I and Hematological Variables
Implications Blood cells are responsible for maintaining health in animals. In the present study recombi-
Literature Cited Annexstad, R. J., D. E. Otterby, J. G. Linn, W. P. Hansen, C. G. Solderholm, and J. E. Wheaton. 1990. Somatotropin treatment for a second consecutive lactation. J. Dairy Sci. 73: 2423. Burton, J. H., G. K. MacLeod, B. W. McBride, J. L. Burton, K. Bateman, I. McMillan, and R. G. Eggert. 1990. Overall efficacy of chronically administered recombinant bovine somatotropin to lactating dairy cows. J. Dairy Sci. 73:2157. Burton, J. L.1991. Effect of recombinant bovine somatotropin on immune responses and associated parameters in lactating dairy cows. Ph.D. Thesis. Dept. of Anim. and Poult. Sci., Univ. of Guelph, Guelph, ON, Canada. Burton, J. L., B. W. McBride, B. W. Kennedy, J. H. Burton, T. H. Elsasser, and B. Woodward. 1991a. The influence of exogenous bovine somatotropin on the responsiveness of peripheral blood lymphocytes to mitogen. J. Dairy Sci. 74:916. Burton, J. L., B. W. McBride, B. W. Kennedy, J. H. Burton, T. H. Elsasser, and B. Woodward. 1991b. Evaluation of blood serum immunoglobulin profiles of dairy cows chronically treated with recombinant bovine somatotropin. J. Dairy sci. 74:1589. Elsasser, T. H., T. S. Rumsey, and A. C. Hammond. 1989. Influence of diet on basal and growth hormone-stimulated plasma concentrations of IGF-I in beef cattle. J. Anim. Sci. 87:128. Eppard, P. J., D. E. Bauman, C. R. Curtis, H. N. Erb, G. M. Lanza, and M. J. DeGeeter. 1987. Effect of 188-day treatment with somatotropin on health and reproductive performance of lactating dairy cows. J. Dairy Sci. 70:582. Fu, Y.-K., S.Arkins. B. S. Wang, and K. W. Kelley. 1991. A novel role of growth hormone and insulin-like growth factor I: priming of neutrophils for superoxide anion secretion. J. Immunol. 146:1602. Gruler, H.P.,J. Zapf, E. Scheibiller, and E. R. Froesch. 1988. Recombinant insulin-like growth factor I stimulates growth and has distinct effects on organ size in hypophysectomized rats. Proc. Natl. Acad. Sci. USA 85:4889. Kelley. K. W. 1989. Commentary: Growth hormone, lymphocytes and macrophages. Biochem. Pharmacol. 38:705. K. Mullen, and R. %we. 1980. Hematology and Lumsden, J. H., biochemistry reference values for female Holstein cattle. Can. J. Comp. Med. 44:24. McBride, B. W., J. L. Burton, and J. H.Burton. 1988. Review The influence of bovine growth hormone (somatotropin) on animals and their products. Res. Dev. Agric. 5:l. McBride, B. W., J. L. Burton, J. P. Gibson, J. H. Burton, and R. Eggert. 1980. Use of recombinant bovine somatotropin for up to two consecutive lactations on dairy production traits. J. Dairy Sci. 73:3248. McGuffey, R. K., H. B. Green, R. P. Basson, and T. H. Ferguson.
Downloaded from https://academic.oup.com/jas/article-abstract/70/5/1488/4705328 by 04860000 user on 12 January 2019
Elevated (P < .051 concentrations of IGF-I in blood of the present group of bST-treated vs control cows has recently been reported (Burton et al., 1991a). It was of interest, therefore, to test possible statistical relationships between IGF-I concentration and the hematological variables of the present study. Two ANOVA were performed for which small alterations were made to the model previously described. For one ANOVA, an IGF-I concentration covariate was added to the original model, and for the other the IGF-I covariate was kept but the bST treatment effect was deleted. There were no alterations in P-values for the bST treatment effect on total leukocyte, red cell, hemoglobin, or platelet variables when IGF-I was added to the original model, but this analysis removed the significant bST effect for lymphocyte fraction (compare Columns 1 and 2 in Table 3). Elevated blood IGF-I concentration may have accounted for the decreased lymphocyte fractions in bST-treated cows (Figure 3). The IGF-I covariate further contributed to variation in fractions and absolute counts of neutrophils, lymphocytes, and monocytes, regardless of whether ANOVA included (Column 3; Table 31 or ignored (Column 4; Table 3) the bST treatment effect. Blood IGF-I concentration, therefore, was at least partially responsible for variation in neutrophil fractions between bST treatment groups (Figure 21, and IGF-I concentration, but not bST treatment, caused variation in monocyte variables. This was consistent with positive IGF-I regression estimates for neutrophil fraction ( P c .01) and count (P < .051 and negative estimates for lymphocyte IP < .01) and monocyte ( P < .05) fractions and counts. Present observations in combination with evidence that neutrophils and lymphocytes possess cellular receptors for IGF-I, and that IGF-I can be produced by these leukocytes (reviewed by Burton, 19911, indicate that further research on the mode of action of IGF-I on leukocytes is warranted. Such research may be especially informative if it is performed in combination with infectious disease challenges.
nant bovine somatotropin caused significant alterations in some red cell and leukocyte variables, particularly at the dose of 20.6 mg per cow per day. Somatotropin treatment effects were generally detected after many weeks of injections, were lost after treatments were terminated, and were not carried over between lactations. The effects on leukocytes were partially attributed to insulin-like growth factor I, which was elevated in the blood of treated cows. The reported effects of somatotropin and(or1 insulin-like growth factor I were for healthy cows, so the practical significance of such effects with respect to infectious disease resistance requires further clarification.
BST AND HEMATOLOGICAL PROFILES 1990. Lactation response of dairy cows receiving bovine somatotropin via daily injections or in a sustained-release vehicle. J. Dairy Sci. 73:763. Phipps, R. H. 1989. A review of the influence of somatotropin on health, reproduction and welfare in lactating dairy cows. In: K. Sejrsen, M. Vestergaard, and A. Neimann-Sorensen (Ed.) Use of Somatotropin in Livestock Production. pp 88-1 19. Elsevier Applied Science, London. SAS. 1982. SAS User’s Guide: Statistics. SAS Inst. Inc., Cary, NC.
1495
Solderholm, C. G., D. E. Otterby, J. G. Linn, and F. R. Ehle, J. W. Wheaton. W. P. Hansen, and R. J. Annexstad. 1988. Effects of recombinant bovine somatotropin on milk production, body composition and physiological parameters. J. Dairy sci. 71:355. Vicini, J. L., S . Hudson, W. J. Cole, M. A. Miller, P. J. Eppard, T. C. White, and R. J. Collier. 1990. Effect of acute challenge with a n extreme dose of somatotropin in a prolonged release formulation on milk production and health of dairy cows. J. Dairy Sci. 73:2093.
Downloaded from https://academic.oup.com/jas/article-abstract/70/5/1488/4705328 by 04860000 user on 12 January 2019
Key Words: Somatotropin, Hematology, Leukocytes, Insulin-Like Growth Factor I, Dairy Cows J. Anim. Sci. 1992. 70:1488-1495
Introduction Recombinant bovine somatotropin (bST) is galactopoietic when injected into dairy cows (reviewed by McBride et al., 1988). Kelley (1989)has reviewed further roles for somatotropin (ST)in hematopoiesis and maintenance of thymic cellularity and function in ST-deficient animals. Blood concentrations of ST influence liver production of insulin-like growth factor I (IGF-I), which is
'The authors express their appreciation to the staff of the Clinical Pathology Laboratory (Ontario Veterinary College, University of Guelph) for performing hematology and leukocyte differentials and acknowledge the Ontario Ministry of Agriculture and Food and the Natural Sciences and Engineering Council of Canada for support of this work. %ecombinant bovine somatotropin was donated by American Cyanamid (Princeton, NJ). 3To whom correspondence should be addressed. 4USDA, ARS, Ruminant Nutr. Lab., Beltsville, MD 20705. Received August 16, 1991. Accepted December 20, 1991.
elevated in the blood of ST-treated animals (reviewed by McBride et al., 1988).There is evidence that specific IGF-I receptors are present on cells derived from the hematopoietic system (reviewed by Burton, 1991); therefore, IGF-I may be partially responsible for effects of injected ST on these cell types (Gruler et al., 1988; Kelley, 1989). Effects of bST on certain hematological variables have been reported (Eppard et al., 1987; Solderholm et al., 1988; Phipps, 1989; Annexstad et al., 1990; Burton et al., 1990; McGuffey et al., 1990; Vicini et al., 1990). Consistent across trials was the observation of decreased hematocrits in cows treated with 2 20.6 mg of bST/d. Relatively few studies have been reported, however, that determine the effects of long-term administration of bST in dairy cows in previous and current lactations, and(or1 of blood IGF-I concentrations, on detailed hematological profiles obtained in the current lactation. These were the objectives of the present study.
1488
Downloaded from https://academic.oup.com/jas/article-abstract/70/5/1488/4705328 by 04860000 user on 12 January 2019
tered hematocrits (P = .lo), mean corpuscular volume (P = .03),mean corpuscular hemoglobin (P = .009), and fraction of blood lymphocytes (P = .09). A week x bST treatment interaction also contributed to variation in hematocrit (P = .09), mean corpuscular hemoglobin (P = .05),fraction of neutrophils (P = .O2), and fraction of lymphocytes (P = .04). Blood IGF-I concentration influenced fractions and counts of neutrophils (P = .06, .09), lymphocytes (P = .04, .02), and monocytes (P = .08, .05). Further in vivo and in vitro studies are warranted because this preliminary evidence s u g gests that bST and(or1 IGF-I contribute to regulation of hematopoiesis in mature dairy cows.
ABSTRACT: Recombinant bovine somatotropin (bST) was administered at 0, 10.3, or 20.6 mg per cow per day to 32 Holsteins for 38 wk. Fifteen currently treated cows had been treated in the previous lactation. Eighteen hematological variables and blood concentration of insulin-like growth factor I (IGF-I) were measured at five sample periods. The objectives of the study were to test the effects of bST treatments on hematological profiles and to relate blood IGF-I concentration to these variables. Results indicated little influence (P > .lo1 of previous bST treatments on hematological profiles measured in the current lactation. Current bST treatments, however, al-
BST AND HEMATOLOGICAL PROFILES
Materials and Methods
the fraction of a particular leukocyte type by the total number of leukocytes previously determined by the Coulter counter method. Insulin-Like Growth Factor I Concentration. The concentrations of IGF-I in blood sera were assayed by RIA, according to the procedure of Elsasser et al. (1989). These results were reported previously (Burton et al., 1991al. Statistical Analyses. Before statistical analysis, all data were log,-transformed for normality of distributions. The GLM procedure of SAS (SAS, 1982) was used for all ANOVA. A preliminary ANOVA was performed to test effects of current and previous bST treatment groupings of cows on hematological variables measured pretreatment (wk 3). Neither effect was significant ( P > .lo). Previous treatments similarly did not affect (P > .lo) variables from subsequent sample weeks, except for whole blood hemoglobin and hematocrit at wk 10 IP < .lo). However, previous treatment was not included in the model used to assess variation in lactational hematological profiles because it was confounded with a cow(treatment1 effect. The model used was as follows: Yijklm = p + weeki + seasonyrj + b l pretrmtkl + trmtk + cow(trmt)kl + week x trmtik + eijklm, where Yijklm = observed hematological/leukocyte differential measurement for cow ijklm; p = population mean; weeki = fixed sample week effect (i = 10, 26, 35, 48); seasonyrj = fixed seasonyear effect [j = winter 1987, 1988; spring 1987, 1988; summer 1987, 1988; and fall 1987, where winter = December, January, February; spring = March, April, May; summer = June, July, August; fall = September, October, November); bl = regression coefficient of Ykl on pretreatment values for cow kl; trmtk = fixed treatment effect (k = 0, 10.3, 20.6 of mg bST/d); cow(trmtIk1 = random cow-nestedwithin-bST treatment effect; wk x trmtik = fixed week x bST treatment interaction effect; and eijkh = random error term. A contrast comparing controls with all treated cows was also performed. The bST treatment and contrast effects were tested using the Type 111 mean square for cow(treatment1 as the denominator in the F-test. Pretreatment covariates were tested against the Type I mean square for cow(treatment1. All other effects were tested against the residual mean square. Possible statistical relationships between blood IGF-I concentrations and hematological variables were tested by two additional ANOVA. In one, a n IGF-I concentration covariate was added to the model previously described. If the significance of the bST treatment effect for any hematological variable was altered by this covariate, then IGF-I might have been considered part of the bST treatment effect. In the other ANOVA an IGF-I
Downloaded from https://academic.oup.com/jas/article-abstract/70/5/1488/4705328 by 04860000 user on 12 January 2019
Animals and Treatments. Thirty-two Holsteins in second or higher parity were used. The bST treatments were given daily S.C. starting between wk 4 and 5 of lactation and continued for 38 wk. Cows were assigned to bST treatment groups according to several criteria. If cows were returning to trial from a previous lactation with bST, they were put in the same treatment regimen as before. If the cows were new to bST trials, they were evenly distributed across treatment groups according to age and previous milk yield. Twelve cows were injected with sterile saline (controls), and 10 cows in each of two groups received either 10.3 or 20.6 mg of bST/d. Of the cows receiving bST, 15 had received the same treatment in the previous lactation and five (two in the 10.3 mg/d group and three in the 20.6 mg/d group) were new to the trial. Controls did not receive bST in the previous lactation. Blood Sampling Schedule. Peripheral blood was collected by tail venipuncture at 0900 of each sample period. Blood for hematological profiling was collected into 10-mL evacuated tubes (Vacu tainer, Fisher Scientific, Don Mills, ON) containing EDTA and that for IGF-I quantification was collected into 10-mL tubes lacking anticoagulant. Blood sample periods were chosen based on lactation curves from the previous bST trial (McBride et al., 1990). Three samples per cow were collected over three consecutive days in wk 3 of lactation to determine pretreatment values of each blood variable and to test potential carryover effects of bST between one lactation and the next. Additional single samples were collected at wk 10, 26, and 35 of lactation to give hematological profiles from peak milk yield through mid- and late lactation, respectively. A sample was also taken a t wk 46 of lactation to determine the effect of bST treatment termination. Hematological Profiles. Variables measured by means of a Coulter Counter (Coulter Plus 4, Coulter Electronics, Hialeah, FL) included leuko, counts (cells cyte counts (cells x r o Q / ~ )erythrocyte x 10l2/L), whole blood hemoglobin (g/L), hematocrit (%I, mean corpuscular volume (E),mean corpuscular hemoglobin (pgl, mean corpuscular hemoglobin concentration (g/L1, red cell distribution width (O/O), platelet counts (cells x 10Q/L),and mean platelet volume (E). Leukocytes were enumerated in 100-cell differential counts of blood smears stained with Wright’s stain and results were expressed as the fraction (percentage) of neutrophils, lymphocytes, monocytes, or eosinophils in the smear. Absolute counts of these leukocyte types in whole blood (cells x 10g/L) were then estimated by multiplying
1489
1490
BURTON ET AL.
Results and Discussion Pretreatment Means Pretreatment means for the 18 hematological variables are given in Table 1. Preliminary ANOVA detected no effect of previous bST treatments on these data (not shown). Similarly, current bST treatments did not affect these values (data not shown). Data in Table 1 were also within published reference ranges (Lumsden et al., 1980). Therefore, treatment groups of cows started the trial with similar, clinically normal hematological profiles.
Hematology Non-Bovine Somatotropin Effects. Fixed effects of sample week and season-year contributed to variation in erythrocyte count ( P = .09, .lo), mean corpuscular volume ( P = .01, .07),and mean corpuscular hemoglobin ( P = .0001, .007).Red cell distribution width was affected by season-year ( P = .O2) but not by sample week. The pretreatment covariate was significant for the above red cell variables ( P e .001), as well as for hematocrit ( P = .08), mean platelet volume ( P = .00051, and leukocyte count ( P = .01). Cow(treatment1, however, was the main source of variation for leukocyte count ( P = .0001), erythrocyte count ( P = .003),hematocrit ( P = .lo), mean corpuscular volume ( P = .00011, mean corpuscular hemaglobin ( P = .00011, red cell distribution width ( P = .0006), and platelet count ( P = .05). The model was relatively inefficient in explaining variation in whole-blood hemoglobin and mean corpuscular hemoglobin concentration (model R2 values were .47 and .41, respectively). Bovine Somatotropin Effects. The bST treatment influenced hematocrit ( P = .lo), mean corpuscular volume ( P = .03), and mean corpuscular hemoglobin (P = .0091. Pooled (wk 10 to 461 treatment least squares means showed 5.8% (P = .03)and 4.9% ( P = .06)decreases in hematocrit for the 20.6 mg/d group vs controls and vs the 10.3 mg/d group,
Table 1. Pretreatment mean and bST treatment least squares mean hematological values Hematological variable Hematology Leukocyte count, cells x 10s/L Erythrocyte count, cells x 1Ol2/L Whole blood hemoglobin, g/L Hematocrit,O/O Mean corpuscular volume, fL Mean corpuscular hemoglobin, pg Mean corpuscular hemoglobin concentration, g/L Red cell distribution width, o h Platelet count, cells x iOg/L Mean platelet volume, fL Leukocyte differential Fraction, O/O, of Neutrophils Lymphocytes Monocytes Eosinophils Counts, cells x I O ~ / L ,of Neutrophils Lymphocytes Monocytes Eosinophils
Pretreatment
Pooled bST Treatment LS Meansb
Meana
SD
7.43 5.67 99.48 28.00 49.83 17.64 354.19 18.88 548.07 5.94
2.40 .63 8.40 2.35 3.32 1.20 6.67 1.12 103.57 .49
40.00 51.00 5.33 5.00
9.00 9.32 4.79 3.87
2.94 3.77 .33 .39
1.07 1.44 .23 .38
0 mg/d 7.39 6.05 102.51 26.94' 44.70' 16.28' 357.81 17.46 387.61 5.99
&Means of three replicates per cow and pooled over bST treatment groups. bLeast squares means pooled over wk 10 to 46. 'VdBeLeast squares means within a row with different superscripts differ ( P c .05). f,gLeast squares means within a row with different superscripts differ (P < .io).
4 1.90 49.16' 1.87 4.6QCf 2.97 3.90 1 .oo .35
10.3 mg/d
20.6 mg/d
8.00 6.17 102.51 28.65c*dsf 5 1 .42d 17.46d 354.25 17.81 454.86 6.11
7.85 6.05 96.54 27.25& 44.26' 15.80e 354.25 17.99 415.72 5.93
42.32 47.24' 2.26 3.47cpd.g
49.16 38.67d 3.05 3.14d
3.53 3.32
3.49 3.78
.68
SO
.27
.24
Downloaded from https://academic.oup.com/jas/article-abstract/70/5/1488/4705328 by 04860000 user on 12 January 2019
covariate was included and the bST treatment effect was deleted from the previously described model. This analysis allowed determination of the influence of blood IGF-I concentration on hematological variables without adjustment for bST treatment effects. Unless otherwise stated, a significance level of P e .10 was assumed. Least squares means in tables and figures were converted from loge values to original units of measurement, so standard errors of means were not reported. Indicated significance levels of paired t-tests between estimated means, however, were based on analysis of loge-transformed data.
1491
BST AND HEMATOLOGICAL PROFILES
respectively (Table 1). Preliminary ANOVA indicated that previously treated cows had lower ( P e .lo) hematocrits at wk 10 than previously untreated cows; therefore, possible bias in the estimation of the current bST treatment effect was acknowledged. Such bias may have been negligible, however, because current bST treatments did not affect hematocrit until wk 26 (Figure 1). The week x bST treatment interaction affected (P = .09) hematocrit. Interaction least squares means demonstrated 1 1.3%( P = .009) and 1 1.7% (P = .01) lower hematocrits at wk 26 and 35, respectively, for the 20.6 mg/d group vs controls, with no differences detected between controls and the 10.3 mg/d group (Figure 1). Values for the 20.6 mg/d group increased to control values after treatments were terminated (Figure 1). Results were consistent with numerous other observations of decreased hematocrits in cows treated with high doses (5 20.6 rng/dl of bST (Eppard et al., 1987; Solderholm et al., 1988; Phipps et al., 1989; Annexstad et al., 1990; Burton et al., 1990; McGuffey et al., 1990; Vicini et al., 19901. The period of maximum effects of bST on milk yield in the current lactation coincided with detectable effects on hematocrit. During mid- to late lactation milk yield was increased by approximately 25% for the 20.6 mg/d group vs controls (Gibson, McBride, and Burton; unpublished observations). It is possible, therefore, that decreased hematocrits resulted from increased plasma volume in response to increased milk output, as was suggested by Phipps (19891 and by Annexstad et al. (19901. Plasma volume, however, has not been estimated directly in the reported studies. Because plasma volume may influence concentrations of all cell types in whole blood, investigation of effects of bST on plasma volume in treated cows is indicated.
34i 32 ..
-
1
0 Controls 10.3 rng
30 1
bST/d
28 0
0
20.6 rng bST/d
26
E 24
20
10
26
35
Week of Lactation
46
Figure 1. Week x bST treatment least squares means for hematocrit, with treatment differences shown (bars within a week with different letters differ [P < .05]]. A difference also occurred between wk 35 and 46 (P = .08) for the 20.6 mg/d group.
Indirect evidence from the present study suggests that decreased hematocrits in cows treated with 20.6 mg bST/d may indeed be more meaningful than simple increases in blood volume in response to increased milk output. For example, hematocrit and erythrocyte count were positively correlated at all sample periods for the 20.6 mg/d group (r = .92 [P < .011, .63, .67, and .64 [P e .lo1 a t wk 10, 26, 35, and 46, respectively). In addition, there was a 13% increase (P = .00011 in mean corpuscular volume in the 10.3 mg/d vs the average of the other groups (Table 2). Furthermore, a 6.8% increase (P = .0009) in the 10.3 mg/d group and a 3.0% decrease (P = .0006) in the 20.6 mg/d group relative to controls was observed for mean corpuscular hemoglobin, and the 10.3 mg/d
Table 2. Analysis of variance of leukocyte differential variables Tvoe 111 s u m s of sauares F-test P-value for Independent variable
Fraction df
Neut'
Lymph
(oh)
of
MonoC
Eosind
Counts (cells
x 1oB/L)
Lymp
Mono
.13 .16 .59
SO
.18 .49 .26
.oo1
.08
.05
.19 .27 .18 .64
.14 .40 .19 .60
31 .71 .24 .55
Neut
of Eosin
~
Week Season-year Pretreatment covariate Cow(treatment1 Week x treatment bST treatment Contraste Model R2
3 6 1 29 6 2 1
.03
.10
.09 .22 .01 .02 .12 .24
.OB .39 ,008 .04 .09
.65
.58
.05
'Neutrophils. bLymphocytes. 'Monocytes. dEosinophils. eContrast compared controls with all bST-treated cows.
.12 .15 .47 .o 1 .6 1 .79 .12 .55
.14
.03 .05 .06
.63 .11 .05 .64
.63 ,001
.09 .07 .15 .03 .22 .22 .ll .67
Downloaded from https://academic.oup.com/jas/article-abstract/70/5/1488/4705328 by 04860000 user on 12 January 2019
22
BURTON ET AL.
1492
65T
b
65i
_ .60
b
v
0 Controls Is1 10.3 rng bST/d
55 50 0 LL 45 40 35 L 30 3 25 _20_
a
c
0 Controls
0 'E
2
::
46
10
26 35 46 Week of Lactation
Figure 2. Week x bST treatment least squares means for fraction of neutrophils with treatment differences shown (bars within a week with different letters differ [P < .OS]). Differences also occurred between wk 35 and 46 (P = .04) for the 10.3 mg/d group and between wk 10 and 26 (P = .lo), 10 and 35 (P = .08), 26 and 46 (P = .09), and 35 and 46 ( P = .03) for the 20.6 mgld group.
Figure 3. Week x bST treatment least squares means for fraction of lymphocytes with treatment differences shown (bars within a week with different letters differ [ab P < .05; ef P c .lo]). Differences also occurred between wk 35 and 46 (P = .01) for the 10.3 mgld group and between wk 26 and 46 (P = .07), and 35 and 46 (P = .06) for the 20.6 mg/d group.
group had 9.5% higher (P = .0001) values than the 20.6 mg/d group (Table 2).Vicini et al. (19901 also found decreased hematocrits and erythrocyte counts after administering two massive doses of bST into pregnant cows. Some contribution of altered erythropoiesis to hematocrit in bST treated cows, therefore, cannot be ruled out.
similar effect of bST on eosinophils after administration of two massive doses of bST, but others have detected no average effects of bST on lymphocytes or eosinophils (Annexstad et al., 1990; McGuffey et al., 1990). It may be argued that the absolute counts of leukocyte types, rather than their fractions, are important during physiological response to infection. There were, however, no differences (P > .lo) in incidence of naturally occurring infectious diseases among the present groups of cows, and cows from each treatment group remained relatively healthy throughout the trial (Burton, 1991). Therefore, physiological response to infection under long-term bST administration could not be determined from results of the present study. For several reasons, however, effects of bST on blood leukocytes per se may have been more appropriately determined from fraction as opposed to count data in this study. First, absolute counts of leukocyte types were estimated by multiplying fraction data by total leukocyte counts on each sample week. Large variation existed among cows for total leukocyte count (7.9 +_ 2.2 [pooled mean f SDI), which contributed to large variation in counts of the various leukocyte types. Second, effects of bST on blood plasma volume have not been determined conclusively. Alterations in plasma volume may affect leukocyte counts of whole blood but would have no effect on differential counts from blood smears. Third, because leukocyte differential data were collected
Leukocyte Differentials Non-Bovine Somatotropin Effects. Compared with the hematological data discussed previously, sample week, season-year, and the pretreatment covariate contributed relatively little to variation in blood leukocyte profiles (Table 2). However, cowttreatmentl was a major source of variation for these variables (Table 2). Bovine Somatotropin Effects. The main effect of bST treatment did not influence absolute counts of any leukocyte type in whole blood, although there was a tendency (P = .091 for the lymphocyte fraction to be affected (Table 2). Furthermore, the contrast effect was significant for both lymphocyte (P = ,051and eosinophil fractions (P = .05;Table 2). Pooled treatment least squares means showed decreases of 21.3% (P = .002) and 33% (P= .02) for lymphocyte and eosinophil fractions, respectively, in the 20.6 mg/d group vs controls (Table 1). There was also an overall decrease of 26% (P = .09)for eosinophil fraction in the 10.3 mg/d group vs controls (Table 1). Vicini et al. (1990) found a
Downloaded from https://academic.oup.com/jas/article-abstract/70/5/1488/4705328 by 04860000 user on 12 January 2019
35
1493
BST AND HEMATOLOGICAL PROFILES
significant biological effects of bST on granulopoiesis and cellular activation are possible, and further studies to test this are warranted. Lymphocyte fraction was also influenced by the week x bST treatment interaction (P = .04; Table 2). Relative to controls, the 20.6 mg/d group had 32.7% (P = .0008)and 33.3% (P = .002)decreases in lymphocyte fraction at wk 26 and 35,respectively (Figure 3). By wk 35, the 10.3 mg/d group had a 16.8% (P = .061 decrease in this variable compared to controls. When treatments were terminated, values increased ( P e .lo) in both treated groups such that they were not different from control values by wk 46 (Figure 3). There are few published data with which to compare present lymphocyte results. This, combined with the fact that health status of the present cows was not different among bST treatment groups and was generally good, made interpretation of present results difficult. It may be speculated, however, that the decreased fraction of lymphocytes did not reflect their biological activity, because peripheral blood lymphocytes from all treated cows displayed elevated ( P < .01) proliferative responsiveness to mitogen in culture (Burton et al., 1991a1, and because there were increased (P e .05)concentrations of immunoglob-
Table 3. Comparison of Type I11 sums of squares P-values for the bST treatment effect and an IGF-I concentration covariate from three ANOVA of hematological variables Blood variable Hematology Leukocyte count, cells x IOO/L Erythrocyte count, cells x 10l2/L Whole blood hemoglobin, g/L Hematocrit, YO Mean corpuscular volume, fL Mean corpuscular hemoglobin, pg Mean corpuscular hemoglobin concentration, g/L Red cell distribution width, YO Platelet count, cells x IOO/L Mean platelet volume, fL Leukocyte differential Fraction, oh, of Neutrophils Lymphocytes Monocytes Eosinophils Counts, cells x I O ~ / L ,of Neutrophils Lymphocytes Monocytes Eosinophils
Model l a
Model
bST effect
bST effect
.81 .81 .17 .10 .03 .01 .55 .95 .58 .65
.73 .84 .23 .15
.12 .09
ab
Model 3c
IGF-I covar.
IGF-I
COVW.
9 1 .70 .54 .41 .20 .20 .86 .76 .11 .74
.Q1 .70 .54 .42 .24 .26 .86 .75 .ll .74
.OB
,7Q .ll
.52 .41 .46 .17
.OB .04 .08 .72
.27 .40 .71 .22
.74 .85 .50 .14
.10 .03 .06 .84
.04
.01 .48 .e7
.40 .63
.05
.OQ .71
aThe original model, including the fixed bST treatment effect. bAn IGF-I concentration covariate was added to the original model. ' A n IGF-I concentration covariate was added and the bST treatment effect ignored from the original model
.OQ .02 .05 .85
Downloaded from https://academic.oup.com/jas/article-abstract/70/5/1488/4705328 by 04860000 user on 12 January 2019
before, during, and after the bST treatment period, more detailed assessment of bST effects on fraction data could be made by relating such effects to the initiation and termination of bST treatments. Information of this type was gained by including a week x bST treatment interaction in the statistical model previously described. Indeed, ANOVA showed that the week x bST treatment interaction influenced fraction of neutrophils (P = .021 and lymphocytes (P = .04; Table 2). Interaction least squares means for the control group were relatively invariant across sample week, whereas those for treated cows were reactively more variable (Figures 2 and 3). Fraction of neutrophils in the 20.6 mg/d group rose by 27.4% ( P = .002)and 29.8% (P = .002) at wk 26 and 35, respectively, relative to control values (Figure 2). However, the dose of 10.3 mg of bST/d did not affect neutrophil fraction. When bST treatments were terminated, neutrophil fraction for the 20.6 mg/d group dropped dramatically and reached control values by at least wk 46 (Figure 2). These results are consistent with the notion that ST is granulopoietic (reviewed by Kelley, 1989).Furthermore, there has been some indication that ST is capable of augmenting superoxide anion production in neutrophils (Fu et al., 1991). Therefore,
BURTON ET AL.
1494
ulins G and G2 in the blood of the 10.3 mg/d group (Burton et al., l991b). The practical implications of these combined observations will only be clarified with further investigation of bST-lymphocyte interactions and with specific infectious disease challenges of bST-treated cows.
Relationship Between Insulin-Like Growth Factor I and Hematological Variables
Implications Blood cells are responsible for maintaining health in animals. In the present study recombi-
Literature Cited Annexstad, R. J., D. E. Otterby, J. G. Linn, W. P. Hansen, C. G. Solderholm, and J. E. Wheaton. 1990. Somatotropin treatment for a second consecutive lactation. J. Dairy Sci. 73: 2423. Burton, J. H., G. K. MacLeod, B. W. McBride, J. L. Burton, K. Bateman, I. McMillan, and R. G. Eggert. 1990. Overall efficacy of chronically administered recombinant bovine somatotropin to lactating dairy cows. J. Dairy Sci. 73:2157. Burton, J. L.1991. Effect of recombinant bovine somatotropin on immune responses and associated parameters in lactating dairy cows. Ph.D. Thesis. Dept. of Anim. and Poult. Sci., Univ. of Guelph, Guelph, ON, Canada. Burton, J. L., B. W. McBride, B. W. Kennedy, J. H. Burton, T. H. Elsasser, and B. Woodward. 1991a. The influence of exogenous bovine somatotropin on the responsiveness of peripheral blood lymphocytes to mitogen. J. Dairy Sci. 74:916. Burton, J. L., B. W. McBride, B. W. Kennedy, J. H. Burton, T. H. Elsasser, and B. Woodward. 1991b. Evaluation of blood serum immunoglobulin profiles of dairy cows chronically treated with recombinant bovine somatotropin. J. Dairy sci. 74:1589. Elsasser, T. H., T. S. Rumsey, and A. C. Hammond. 1989. Influence of diet on basal and growth hormone-stimulated plasma concentrations of IGF-I in beef cattle. J. Anim. Sci. 87:128. Eppard, P. J., D. E. Bauman, C. R. Curtis, H. N. Erb, G. M. Lanza, and M. J. DeGeeter. 1987. Effect of 188-day treatment with somatotropin on health and reproductive performance of lactating dairy cows. J. Dairy Sci. 70:582. Fu, Y.-K., S.Arkins. B. S. Wang, and K. W. Kelley. 1991. A novel role of growth hormone and insulin-like growth factor I: priming of neutrophils for superoxide anion secretion. J. Immunol. 146:1602. Gruler, H.P.,J. Zapf, E. Scheibiller, and E. R. Froesch. 1988. Recombinant insulin-like growth factor I stimulates growth and has distinct effects on organ size in hypophysectomized rats. Proc. Natl. Acad. Sci. USA 85:4889. Kelley. K. W. 1989. Commentary: Growth hormone, lymphocytes and macrophages. Biochem. Pharmacol. 38:705. K. Mullen, and R. %we. 1980. Hematology and Lumsden, J. H., biochemistry reference values for female Holstein cattle. Can. J. Comp. Med. 44:24. McBride, B. W., J. L. Burton, and J. H.Burton. 1988. Review The influence of bovine growth hormone (somatotropin) on animals and their products. Res. Dev. Agric. 5:l. McBride, B. W., J. L. Burton, J. P. Gibson, J. H. Burton, and R. Eggert. 1980. Use of recombinant bovine somatotropin for up to two consecutive lactations on dairy production traits. J. Dairy Sci. 73:3248. McGuffey, R. K., H. B. Green, R. P. Basson, and T. H. Ferguson.
Downloaded from https://academic.oup.com/jas/article-abstract/70/5/1488/4705328 by 04860000 user on 12 January 2019
Elevated (P < .051 concentrations of IGF-I in blood of the present group of bST-treated vs control cows has recently been reported (Burton et al., 1991a). It was of interest, therefore, to test possible statistical relationships between IGF-I concentration and the hematological variables of the present study. Two ANOVA were performed for which small alterations were made to the model previously described. For one ANOVA, an IGF-I concentration covariate was added to the original model, and for the other the IGF-I covariate was kept but the bST treatment effect was deleted. There were no alterations in P-values for the bST treatment effect on total leukocyte, red cell, hemoglobin, or platelet variables when IGF-I was added to the original model, but this analysis removed the significant bST effect for lymphocyte fraction (compare Columns 1 and 2 in Table 3). Elevated blood IGF-I concentration may have accounted for the decreased lymphocyte fractions in bST-treated cows (Figure 3). The IGF-I covariate further contributed to variation in fractions and absolute counts of neutrophils, lymphocytes, and monocytes, regardless of whether ANOVA included (Column 3; Table 31 or ignored (Column 4; Table 3) the bST treatment effect. Blood IGF-I concentration, therefore, was at least partially responsible for variation in neutrophil fractions between bST treatment groups (Figure 21, and IGF-I concentration, but not bST treatment, caused variation in monocyte variables. This was consistent with positive IGF-I regression estimates for neutrophil fraction ( P c .01) and count (P < .051 and negative estimates for lymphocyte IP < .01) and monocyte ( P < .05) fractions and counts. Present observations in combination with evidence that neutrophils and lymphocytes possess cellular receptors for IGF-I, and that IGF-I can be produced by these leukocytes (reviewed by Burton, 19911, indicate that further research on the mode of action of IGF-I on leukocytes is warranted. Such research may be especially informative if it is performed in combination with infectious disease challenges.
nant bovine somatotropin caused significant alterations in some red cell and leukocyte variables, particularly at the dose of 20.6 mg per cow per day. Somatotropin treatment effects were generally detected after many weeks of injections, were lost after treatments were terminated, and were not carried over between lactations. The effects on leukocytes were partially attributed to insulin-like growth factor I, which was elevated in the blood of treated cows. The reported effects of somatotropin and(or1 insulin-like growth factor I were for healthy cows, so the practical significance of such effects with respect to infectious disease resistance requires further clarification.
BST AND HEMATOLOGICAL PROFILES 1990. Lactation response of dairy cows receiving bovine somatotropin via daily injections or in a sustained-release vehicle. J. Dairy Sci. 73:763. Phipps, R. H. 1989. A review of the influence of somatotropin on health, reproduction and welfare in lactating dairy cows. In: K. Sejrsen, M. Vestergaard, and A. Neimann-Sorensen (Ed.) Use of Somatotropin in Livestock Production. pp 88-1 19. Elsevier Applied Science, London. SAS. 1982. SAS User’s Guide: Statistics. SAS Inst. Inc., Cary, NC.
1495
Solderholm, C. G., D. E. Otterby, J. G. Linn, and F. R. Ehle, J. W. Wheaton. W. P. Hansen, and R. J. Annexstad. 1988. Effects of recombinant bovine somatotropin on milk production, body composition and physiological parameters. J. Dairy sci. 71:355. Vicini, J. L., S . Hudson, W. J. Cole, M. A. Miller, P. J. Eppard, T. C. White, and R. J. Collier. 1990. Effect of acute challenge with a n extreme dose of somatotropin in a prolonged release formulation on milk production and health of dairy cows. J. Dairy Sci. 73:2093.
Downloaded from https://academic.oup.com/jas/article-abstract/70/5/1488/4705328 by 04860000 user on 12 January 2019
