PHYSIOLOGY AND MANAGEMENT Effect of High Doses of a Sustained-Release Bovine Somatotropin on Antibody Formation In Dairy Cows1 P. J. EPPARD, G. J. ROGAN, B. G. BOYSEN,2 M. A. MILLER,3 R. L HINTZ, B. G. HAMMOND, A. R. TORKELSON, R. J. COLUER, and G. M. LANZA4 Monsanto Company Animal Sciences Division St. Louis, MO 63198 ABSTRACT
x lOS Umol. Cows considered to be clinically positive had performance similar to those of their herdmates having binding 25% (positives) during the 1st yr. At the .6-g dose level, no binding was detected after wk 15. Seven of the 13 positive cows were among the group randomly selected to continue on study during yr 2. In the 2nd yr, only 2 out of 24 bST-treated cows were positive. Binding activity was associated with the IgG fraction in serum. Binding capacities of antibodies ranged from .625 to 3.04 mg of bSTlL. and affinities ranged from 1.14 x lOS to 3.14
Abbreviation somatotropin.
hST
=
human
INTRODUCTION
Received January 29, 1992. Accepted June 2, 1992. 'No official support or endorsement by the FDA is intended nor should be inferred. 2Present address: Hazleton Wisconsin, Inc., 3301 Kins· man Boulevard, PO Box 7545, Madison, WI 53707. 3Present address: Food and Drug Administration, Center for Veterinary Medicine, Division of Toxicology, HFV 150, 7500 Standish Place, Rockville, MD 20855. 4Present address: Northwestern University Medical School, Department of Medicine, Section of Cardiology, 250 East Superior Street, Chicago, IL 60611. 1992 J Dairy Sci 75:2959-2967
key:
The safety of bST in the dairy cow has only recently been addressed. Most previous bST studies have used relatively lower doses (Q5 mgld) or have been longer term clinical studies primarily designed to study efficacy rather than cow health. However, two acute bST toxicity studies recently demonstrated that massive. short-term dosages of bST were well tolerated with no indication of toxicity (17, 30). Multiple lactation, bST toxicology data (6, 9) have demonstrated minimal clinical health changes. However, more subtle health changes, such as immune response to bST, have been postulated but not demonstrated (15). Exogenous hormones elicit mild immunological responses in human subjects. Therapeutic administration of human pituitary or recombinantly derived somatotropin (hST) to humans resulted in the production of circulating anti-hST antibodies (14, 24, 25, 31). The titer and binding capacity of anti-hST antibodies were low, and the growth response was not attenuated (1, 14,24,25). Approximately 20 to 40% of patients treated with recombinant
2959
2960
EPPARD ET AL.
methionyl hST developed measurable antibodies (25, 31). Immune-mediated tissue damage was not observed, e.g., glomerular lesions secondary to activation of complement or deposition of immune complexes (20). Growth attenuation in humans was documented for early pituitary hST preparations (29) and for recombinant hST in humans with idiopathic somatotropin deficiency (22). Early work with pituitary hST indicated that use of preparations with low concentrations of aggregated hST caused a transient presence of antibodies (20, 31). Jordan (12) also reported that, after minor improvements in the production process, the hST was considerably less antigenic. The highest purity methionyl hST cited in that report (12) did not produce a significant antigenic response in the rhesus monkey. In contrast, methionyl hST elicited a slightly higher antibody response in humans or monkeys than a nonmethionyl hST (25, 32). Recent reports (4, 5, 33) have examined the effect of bST on immune function of the dairy cow. Most cows treated with 960, 2880, or 4800 mg/28 d of bST (somidobove) developed low concentrations of antibody against bST (33), but no adverse health or performance effects were observed. Burton et al. (5) also reported no detrimental effect of daily injection of up to 20.6 mg of bST (266 d) on blood concentrations of Ig. Modest increases in IgG and IgG2 were observed, but all mean concentrations of Ig isotypes were within normal ranges (5). Thus, the humoral immune response to recombinant bST preparations has not been associated with adverse health effects or attenuated milk production. Published data indicate that recombinant bST preparations can be chemically similar to pituitary bST (16). However, antigenic differences may be expected between specific molecular variants that differ structurally depending on the source of recombinant bST. Available molecular variants of bST vary considerably at the amino terminus: three additional amino acids (5, 13), nine additional amino acids (4, 13, 33), and methionine substituted for alanine (9). Therefore, the possibility of attenuation of lactation by antibody production was examined in dairy cows treated with zinc methionyl bST (sometribove).
Journal of Dairy Science Vol. 75, No. 11, 1992
MATERIALS AND METHODS General Methods
Eighty-two first, second, and third parity Holstein cows were assigned to treatments in a randomized block design. The cows had received no previous bST treatment. Treatments were 0, .6, 1.8, and 3.0 g/14 d of zinc methionyl bST in an oil-based formulation. The biopotency of the bST was approximately 1.1 IU/mg of protein (14% coefficient of variation) as determined by a hypophysectomized rat bioassay (18). Biweekly intramuscular injections were administered in the left or right gluteus, semitendinosus, or semimembranosus muscles (five alternating sites). At 3 and 10 d after bST administration, injection sites were scored for swelling (0 none, 1 minimal, 2 moderate, 3 severe). Treatment began at 60 ± 3 d postpartum and continued until the end of lactation or necropsy. A subset of 38 cows continued through a second lactation with the same treatment regimen. Cows were removed from study only for 1) scheduled necropsies following the first or second treatment periods or at the start of the third lactation or 2) unscheduled necropsies when a cow died or was declared moribund (6). The 38 cows selected to continue into the second lactation were chosen randomly before the end of the first treatment period without consideration of health or production response. At necropsy, sections from at least 30 tissues per cow were obtained for microscopic evaluation. Additional details of the study procedures were presented elsewhere (6, 9). The study was conducted according to FDA Good Laboratory Practice Regulations (10). The National Institutes of Health Guide for the Care and Use of Laboratory Animals (21) was followed.
=
=
=
=
Blood Analyses
Jugular blood samples were drawn at least 2 h postrnilking during wk -2, -1, 3, and 7 and then every 8 wk (yr 1) or every 4 wk (yr 2). Although all samples were analyzed, for consistency, only data for the 8-wk sampling pattern for each year are presented. No blood samples were drawn during the cows' dry peri-
SOMATOTROPIN ANTIBODIES IN COWS
ods. Blood creatinine concentrations were analyzed using an ACA IV Clinical Analyzer (DuPont, Wilmington, DE). Serum anti-bST binding was measured as previously described (7, 19) with the following modifications. Cow serum (25 lJ,1) was incubated overnight with 75 lJ,1 of horse serum (K. C. Biologicals, Lenexa, KS) and 100 lJ,1 of 125I-Iabeled recombinant methionyl bST (.35 ng per tube; 10 to 15 lJ,Ci/ lJ,g; Monsanto Co., St. Louis, MO) in buffer (1 % bovine serum albumin, .5% polyoxyethylenesorbitan [Sigma Chemical Co., St. Louis, MO] and 50 mM barbital [Fisher Scientific, S1. Louis, MOJ). A positive control, consisting of 25 III of 1; 100 (vol/vol) rabbit polyclonal anti-bST antibody (lot R609, Monsanto Co.), 75 j..l1 of equine serum, and 100 lJ,1 of 125I-labeled bST was included in each assay. The negative control was pretreatment (lst yr) serum from each cow. Bound bST was precipitated by addition of 200 lJ,1 of ice-cold 25% polyethylene glycol and centrifugation at 2156 x g for 30 min at 4°C. The supernatant was aspirated, and the pellet was washed with .4 ml of cold 12.5% polyethylene glycol barbital, followed by recentrifugation and aspiration of the supernatant. Counts were measured by a Micromedic gamma counter (Huntsville, AL) and expressed as the percentage of relative binding (corrected for each cow's individual negative control, the 1st yr pretreatment value) compared with the rabbit antibody positive control and were not corrected for endogenous bST concentration. Intraassay and interassay coefficients of variation averaged 1.3 and 10.5%, respectively. Binding greater than 25% was considered to be clinically positive (l9). All serum samples with binding over 25% were chromatographed on a PD-lO buffer exchange column (Pharrnacia, Uppsala, Sweden) to equilibrate the serum in 30 roM NaKP04 buffer (pH 6.5) and to remove lipids. The samples were then rechromatographed on a Perkin-Elmer HPLC (Richfield, Cf) using a mono-Q HR5/5 I-ml anion-exchange column (Pharrnacia). The column was equilibrated with 30 ml of the NaKP04 buffer after which I ml of dilute (1:4, vol/vol) serum was loaded. The column was then washed with 12 ml of the NaKP04 buffer. A step gradient was employed for separation of IgG from IgM. The
2961
IgG was eluted with 165 mM Na2P04 (pH 6.7); IgM was eluted with 300 mM Na2P04. After validation of the chromatography procedure, fractions routinely were collected beginning after 7 min (14 ml) of elution. The fractions were lyophilized on a speed vacuum concentrator (Savant Instruments, Inc., Farmingdale, NY) and resolubilized with .2 m1 of water for identification of IgG, IgM, and antibST binding activity. Serum fractions were analyzed for IgG or IgM using double diffusion in agar (23). Resolubilized serum fractions (30 lJ,1) were placed in the exterior wells of immunodiffusion disks (Miles, Inc., Kankakee, IL), and 30 lJ,1 of either polyclonal goat anti-bovine IgG (Kirkegaard and Perry Labs, Inc., Gaithersburg, MD) or of rabbit anti-bovine IgM (Cappel, Durham, NC) were then added to the center well. The presence of precipitation bands identified IgG or IgM in the serum fractions. Fractions with IgG or IgM activity were then reassayed using the antibody screening procedure. Scatchard analysis (27) was used to characterize the binding capacity and affinity of the anti-bST antibodies. Statistics
The effect of treatment on serum anti-bST binding was analyzed separately for each sample by chi-square analysis using the frequency procedure on SAS software (26). For each sampling period, the relative binding data were categorized into five percentage unit levels (binding ~5%, binding >5 but ~1O%, binding >10 but ~15%, etc.) and analyzed for significance (P < .10). Blood creatinine and blood urea N were analyzed by ANOV A as previously described (9). Data from 1 cow were excluded from all analyses because of bovine paratuberculosis (6). Two cows also were excluded from the 1st yr analyses because of early removal from study (6). RESULTS AND DISCUSSION
Sometribove increased (P < .10) the frequency of anti-bST binding because only bSTtreated cows developed antibodies (Table 1). By wk 3 of study, most bST-treated cows Journal of Dairy Science Vol. 75, No. II, 1992
2962
EPPARD ET AL.
TABLE 1. Frequencyl of anti-bST antibody responses to somebibove administered during the first of two consecutive lactations. Week of study
Relative binding
Overall
bST
1}
Control
.6 g
1.8 g
3.0 g
~
20120
20120
18/18
21121
~5
20120 0120 0120
9120 8120 3120
4118 11118 3/18
3121 13/21 5/21
25 7
~
5-25 >25 15
~5
5-25 >25 23
~5
5-25 >25 31
~
5-25 >25
6120
lFrequency of relative anti-bST binding reported as the proportion of cows in each group with~, 5 to 25, or >25% (positive) binding. Chi-square analyses conducted for frequency across 5% intervals, e.g., ~5%, >5 but ~1O%. >10 but ~15%, etc. Binding levels during treatment are adjusted for background titers (wk -2, -1).
experienced an increase in relative anti-bST binding compared with their individual pretreatment binding percentages (Figure 1). Relative binding averaged -1 ± 1.7, 7 ± 11.7. 10 ± 16.1. and 15 ± 23.0% (mean ± SE) for the control and groups receiving 6. 1.8. and 3.0 g. respectively. However. only 13 out of 59 cows receiving sometribove developed anti-bSTbinding activity >25% (positives): 4 cows receiving .6 g. 3 cows receiving 1.8 g, and 6 cows receiving 3.0 g (Table 1). In all cases. the increase in binding was detected within the first 7 wk of treatment. and. over the course of lactation. the number of cows with anti-bST binding generally declined. A similar initial increase in binding followed by a gradual decline was reported previously for hSTtreated humans (25, 31). Five cows in the groups receiving 1.8 and 3.0 g sustained circulating anti-bST binding greater than 25% through the entire first lactation. In the .6-g group, no positive cows (binding >25%) were detected after wk 15 of the fIrst treatment period. Within bST treatment groups. the FCM production of positive cows was similar to that of herdmates (Figure 2). Journal of Dairy Science Vol. 75. No. II, 1992
During the 2nd yr. relative binding averaged -1 ± 3.4. 2 ± 3.3, 3 ± 4.2, and 16 ± 25.1 % for the control. .6-. 1.8-. and 3.O-g groups. respectively. Of the cows with binding greater than 25% during the 1st yr. no cows were declared moribund or died. but 6 cows were selected randomly for necropsy at the end that lactation. Of the remaining 7 positive responders from the 1st yr. 1 cow receiving .6 g, 1 cow receiving 1.8 g, and 3 cows receiving 3.0 g had no increase in binding during the 2nd yr (relative to their pretreatment values for yr 1). Only 2 cows. both in the 3.0-g group, exhibited anti-bST binding >25% (Table 2). The relative binding in these 2 cows' sera was >75% after wk 3 of treatment. It is not apparent why these cows responded differently than their herdmates. However. the binding apparently did not attenuate because milk production of these cows (37.7 and 48.0 kg/d) was among the highest for the study (9). A low level of anti-bST binding during each year of treatment is consistent with previous results (4. 5, 33). Mean IgG concentrations (not specifIc for bST) were within normal
2963
SOMATOTROPIN ANTIBODIES IN COWS 80
70 •
'0
60
Ii!! E3
1.8 g of bST
o
3.DgofbST
.6gOfbST
•
•
COntrol
lso '" 'g 40
•
ST BIndIng ('\0)
3.5% FCM (kg/d)
Figure I. Percentage of anti-bST-binding activity in bST-treated cows at wk 3 of the 1st yr of study. Binding is expressed relative to pretreatment and positive control values.
Figure 2. Effect of serum anti-bST-binding activity on each cow's average 3.5% FCM production during yr I.
ranges following daily injection of 10.3 or 20.6 mg of bST/d (5), and anti-bST binding was unaffected by daily injection of up to 40.5 mg of bST (2). In contrast, Zwickl et al. (33) reported that nearly all bST-treated cows deve-
loped anti-bST IgG antibodies with no clinical health effects during each of two lactations of treatment. The molecular variants of bST utilized in each of these studies probably varied in structure at the amino terminus: three addi-
TABLE 2. Frequency! of anti-bST antibody responses to sometribove administered during the second of two consecutive lactations. Week of study
Relative binding
Overall
bST Control
.6 g
14/14 0114 0114 12/14 2/14 0/14 14114 0114 0114 14/14 0114 0114 14114 0114 0114
8/8 0/8 0!8 8/8 0/8 0/8 8/8 0/8
;(2
1.8 g
3.0 g
P value
m
719 1/9 1/9
.362
(%) -2, -I
3
7
15
23
$5 5-25 >2.5 $5 5-25 >25 $5 5-25 >25 $5 5-25 >25 $5
5-25 >25
017 017
5n 2n 017 6n
In
018 618 2/8
017 4n
0/8
017
5/6 1/6 0/6
5n
3n
2n 017
5/9 219 S/9 419 3/9 2/9
.172
619 1/9 2/9
.062
5/9 219 2/9
.296
.133
!Frequency of relative anti-bST binding reported as the proportion of cows in each group with $5. 5 to 25. or >25% (positive) binding. Chi-square analyses conducted for frequency across 5% intervals. e.g., $5%, >5 but $10%, >10 but $15%, etc. Binding levels during treatment are adjusted for background titers (wk -2, -1 of yr I). Most cows were dried off prior to the wk-31 sampling. Journal of Dairy Science Vol. 75. No. II, 1992
2964
EPPARD ET AL.
tional amino acids (5, 13), nine additional amino acids (4, 13, 33), and methionine substituted for alanine in the bST used in the present study. Although molecular variants of bST probably affect antibody response, to date no positive or negative clinical health effects have been noted for any bST variant (4, 5, 33). Moreover, only 2 out of 42 cows injected for 21 d with 25 mgld of one of four recombinant bST variants or pituitary-derived bST (8) developed anti-bST binding >25% (data not presented). Interstudy differences in antibody response may be due to dissimilar antibody assays, molecular aggregation of bST in vivo, bST dosages, or frequency of treatment. As previously described (I, 5, 33), a principle Ig that is produced in response to exogenous somatotropin is IgG. The anti-bSTbinding activity in sera from cows with binding >25% in the present study was further analyzed by HPLC to identify IgG and IgM. An example chromatogram of the serum fractions with associated IgG- and IgM-binding activities is presented in Figure 3. Immunoglobulin G was found in fractions 22 to 27, primarily eluting in fractions 22 and 23, and IgM eluted in fractions 35 to 37. The fractions were then reassayed for anti-bST-
80 .30
binding activity. In all cases, anti-bST binding was associated with the IgG peak (fractions 22 and 23). Thus, the binding activity probably is not due to an endogenous bST-binding protein as has been described for humans (3). Scatchard analysis indicated that binding capacities were highly variable, ranging from .625 to 3.04 mg of bSTlL, and the affinities ranged from 1.14 x 108 to 3.14 X 108 LImo\. These antibodies did not appear to affect the galactopoietic response to bST (Figure 2) but may have resulted in slightly underestimated serum bST concentrations (9). At 3 and 10 dafter bST administration, injection sites were scored for swelling (0 = none, 1 = minimal, 2 = moderate, 3 = severe). Injection site response did not appear to be related to anti-bST binding (Figure 4). Of the 13 cows with >25% relative binding, only 2 cows (both receiving 3.0 g of bST) had repeated incidences of adverse injection site swelling. Moreover, immune-mediated tissue damage, such as glomerular lesions secondary to activation of complement and the resultant compromise in renal function, was not observed in cows (6), regardless of serum bST binding. At necropsy (6), interstitial fibrosis or regeneration of tubular epithelium was noted in kidneys from 3 of 13 cows with elevated anti-bST binding (>25%) at some time during the study. However, similar lesions were also noted in controls, and all but one of the lesions were of only minimal or slight severity. In
60
=> 00
.20 40
..
C :
x lOS Umol. Cows considered to be clinically positive had performance similar to those of their herdmates having binding 25% (positives) during the 1st yr. At the .6-g dose level, no binding was detected after wk 15. Seven of the 13 positive cows were among the group randomly selected to continue on study during yr 2. In the 2nd yr, only 2 out of 24 bST-treated cows were positive. Binding activity was associated with the IgG fraction in serum. Binding capacities of antibodies ranged from .625 to 3.04 mg of bSTlL. and affinities ranged from 1.14 x lOS to 3.14
Abbreviation somatotropin.
hST
=
human
INTRODUCTION
Received January 29, 1992. Accepted June 2, 1992. 'No official support or endorsement by the FDA is intended nor should be inferred. 2Present address: Hazleton Wisconsin, Inc., 3301 Kins· man Boulevard, PO Box 7545, Madison, WI 53707. 3Present address: Food and Drug Administration, Center for Veterinary Medicine, Division of Toxicology, HFV 150, 7500 Standish Place, Rockville, MD 20855. 4Present address: Northwestern University Medical School, Department of Medicine, Section of Cardiology, 250 East Superior Street, Chicago, IL 60611. 1992 J Dairy Sci 75:2959-2967
key:
The safety of bST in the dairy cow has only recently been addressed. Most previous bST studies have used relatively lower doses (Q5 mgld) or have been longer term clinical studies primarily designed to study efficacy rather than cow health. However, two acute bST toxicity studies recently demonstrated that massive. short-term dosages of bST were well tolerated with no indication of toxicity (17, 30). Multiple lactation, bST toxicology data (6, 9) have demonstrated minimal clinical health changes. However, more subtle health changes, such as immune response to bST, have been postulated but not demonstrated (15). Exogenous hormones elicit mild immunological responses in human subjects. Therapeutic administration of human pituitary or recombinantly derived somatotropin (hST) to humans resulted in the production of circulating anti-hST antibodies (14, 24, 25, 31). The titer and binding capacity of anti-hST antibodies were low, and the growth response was not attenuated (1, 14,24,25). Approximately 20 to 40% of patients treated with recombinant
2959
2960
EPPARD ET AL.
methionyl hST developed measurable antibodies (25, 31). Immune-mediated tissue damage was not observed, e.g., glomerular lesions secondary to activation of complement or deposition of immune complexes (20). Growth attenuation in humans was documented for early pituitary hST preparations (29) and for recombinant hST in humans with idiopathic somatotropin deficiency (22). Early work with pituitary hST indicated that use of preparations with low concentrations of aggregated hST caused a transient presence of antibodies (20, 31). Jordan (12) also reported that, after minor improvements in the production process, the hST was considerably less antigenic. The highest purity methionyl hST cited in that report (12) did not produce a significant antigenic response in the rhesus monkey. In contrast, methionyl hST elicited a slightly higher antibody response in humans or monkeys than a nonmethionyl hST (25, 32). Recent reports (4, 5, 33) have examined the effect of bST on immune function of the dairy cow. Most cows treated with 960, 2880, or 4800 mg/28 d of bST (somidobove) developed low concentrations of antibody against bST (33), but no adverse health or performance effects were observed. Burton et al. (5) also reported no detrimental effect of daily injection of up to 20.6 mg of bST (266 d) on blood concentrations of Ig. Modest increases in IgG and IgG2 were observed, but all mean concentrations of Ig isotypes were within normal ranges (5). Thus, the humoral immune response to recombinant bST preparations has not been associated with adverse health effects or attenuated milk production. Published data indicate that recombinant bST preparations can be chemically similar to pituitary bST (16). However, antigenic differences may be expected between specific molecular variants that differ structurally depending on the source of recombinant bST. Available molecular variants of bST vary considerably at the amino terminus: three additional amino acids (5, 13), nine additional amino acids (4, 13, 33), and methionine substituted for alanine (9). Therefore, the possibility of attenuation of lactation by antibody production was examined in dairy cows treated with zinc methionyl bST (sometribove).
Journal of Dairy Science Vol. 75, No. 11, 1992
MATERIALS AND METHODS General Methods
Eighty-two first, second, and third parity Holstein cows were assigned to treatments in a randomized block design. The cows had received no previous bST treatment. Treatments were 0, .6, 1.8, and 3.0 g/14 d of zinc methionyl bST in an oil-based formulation. The biopotency of the bST was approximately 1.1 IU/mg of protein (14% coefficient of variation) as determined by a hypophysectomized rat bioassay (18). Biweekly intramuscular injections were administered in the left or right gluteus, semitendinosus, or semimembranosus muscles (five alternating sites). At 3 and 10 d after bST administration, injection sites were scored for swelling (0 none, 1 minimal, 2 moderate, 3 severe). Treatment began at 60 ± 3 d postpartum and continued until the end of lactation or necropsy. A subset of 38 cows continued through a second lactation with the same treatment regimen. Cows were removed from study only for 1) scheduled necropsies following the first or second treatment periods or at the start of the third lactation or 2) unscheduled necropsies when a cow died or was declared moribund (6). The 38 cows selected to continue into the second lactation were chosen randomly before the end of the first treatment period without consideration of health or production response. At necropsy, sections from at least 30 tissues per cow were obtained for microscopic evaluation. Additional details of the study procedures were presented elsewhere (6, 9). The study was conducted according to FDA Good Laboratory Practice Regulations (10). The National Institutes of Health Guide for the Care and Use of Laboratory Animals (21) was followed.
=
=
=
=
Blood Analyses
Jugular blood samples were drawn at least 2 h postrnilking during wk -2, -1, 3, and 7 and then every 8 wk (yr 1) or every 4 wk (yr 2). Although all samples were analyzed, for consistency, only data for the 8-wk sampling pattern for each year are presented. No blood samples were drawn during the cows' dry peri-
SOMATOTROPIN ANTIBODIES IN COWS
ods. Blood creatinine concentrations were analyzed using an ACA IV Clinical Analyzer (DuPont, Wilmington, DE). Serum anti-bST binding was measured as previously described (7, 19) with the following modifications. Cow serum (25 lJ,1) was incubated overnight with 75 lJ,1 of horse serum (K. C. Biologicals, Lenexa, KS) and 100 lJ,1 of 125I-Iabeled recombinant methionyl bST (.35 ng per tube; 10 to 15 lJ,Ci/ lJ,g; Monsanto Co., St. Louis, MO) in buffer (1 % bovine serum albumin, .5% polyoxyethylenesorbitan [Sigma Chemical Co., St. Louis, MO] and 50 mM barbital [Fisher Scientific, S1. Louis, MOJ). A positive control, consisting of 25 III of 1; 100 (vol/vol) rabbit polyclonal anti-bST antibody (lot R609, Monsanto Co.), 75 j..l1 of equine serum, and 100 lJ,1 of 125I-labeled bST was included in each assay. The negative control was pretreatment (lst yr) serum from each cow. Bound bST was precipitated by addition of 200 lJ,1 of ice-cold 25% polyethylene glycol and centrifugation at 2156 x g for 30 min at 4°C. The supernatant was aspirated, and the pellet was washed with .4 ml of cold 12.5% polyethylene glycol barbital, followed by recentrifugation and aspiration of the supernatant. Counts were measured by a Micromedic gamma counter (Huntsville, AL) and expressed as the percentage of relative binding (corrected for each cow's individual negative control, the 1st yr pretreatment value) compared with the rabbit antibody positive control and were not corrected for endogenous bST concentration. Intraassay and interassay coefficients of variation averaged 1.3 and 10.5%, respectively. Binding greater than 25% was considered to be clinically positive (l9). All serum samples with binding over 25% were chromatographed on a PD-lO buffer exchange column (Pharrnacia, Uppsala, Sweden) to equilibrate the serum in 30 roM NaKP04 buffer (pH 6.5) and to remove lipids. The samples were then rechromatographed on a Perkin-Elmer HPLC (Richfield, Cf) using a mono-Q HR5/5 I-ml anion-exchange column (Pharrnacia). The column was equilibrated with 30 ml of the NaKP04 buffer after which I ml of dilute (1:4, vol/vol) serum was loaded. The column was then washed with 12 ml of the NaKP04 buffer. A step gradient was employed for separation of IgG from IgM. The
2961
IgG was eluted with 165 mM Na2P04 (pH 6.7); IgM was eluted with 300 mM Na2P04. After validation of the chromatography procedure, fractions routinely were collected beginning after 7 min (14 ml) of elution. The fractions were lyophilized on a speed vacuum concentrator (Savant Instruments, Inc., Farmingdale, NY) and resolubilized with .2 m1 of water for identification of IgG, IgM, and antibST binding activity. Serum fractions were analyzed for IgG or IgM using double diffusion in agar (23). Resolubilized serum fractions (30 lJ,1) were placed in the exterior wells of immunodiffusion disks (Miles, Inc., Kankakee, IL), and 30 lJ,1 of either polyclonal goat anti-bovine IgG (Kirkegaard and Perry Labs, Inc., Gaithersburg, MD) or of rabbit anti-bovine IgM (Cappel, Durham, NC) were then added to the center well. The presence of precipitation bands identified IgG or IgM in the serum fractions. Fractions with IgG or IgM activity were then reassayed using the antibody screening procedure. Scatchard analysis (27) was used to characterize the binding capacity and affinity of the anti-bST antibodies. Statistics
The effect of treatment on serum anti-bST binding was analyzed separately for each sample by chi-square analysis using the frequency procedure on SAS software (26). For each sampling period, the relative binding data were categorized into five percentage unit levels (binding ~5%, binding >5 but ~1O%, binding >10 but ~15%, etc.) and analyzed for significance (P < .10). Blood creatinine and blood urea N were analyzed by ANOV A as previously described (9). Data from 1 cow were excluded from all analyses because of bovine paratuberculosis (6). Two cows also were excluded from the 1st yr analyses because of early removal from study (6). RESULTS AND DISCUSSION
Sometribove increased (P < .10) the frequency of anti-bST binding because only bSTtreated cows developed antibodies (Table 1). By wk 3 of study, most bST-treated cows Journal of Dairy Science Vol. 75, No. II, 1992
2962
EPPARD ET AL.
TABLE 1. Frequencyl of anti-bST antibody responses to somebibove administered during the first of two consecutive lactations. Week of study
Relative binding
Overall
bST
1}
Control
.6 g
1.8 g
3.0 g
~
20120
20120
18/18
21121
~5
20120 0120 0120
9120 8120 3120
4118 11118 3/18
3121 13/21 5/21
25 7
~
5-25 >25 15
~5
5-25 >25 23
~5
5-25 >25 31
~
5-25 >25
6120
lFrequency of relative anti-bST binding reported as the proportion of cows in each group with~, 5 to 25, or >25% (positive) binding. Chi-square analyses conducted for frequency across 5% intervals, e.g., ~5%, >5 but ~1O%. >10 but ~15%, etc. Binding levels during treatment are adjusted for background titers (wk -2, -1).
experienced an increase in relative anti-bST binding compared with their individual pretreatment binding percentages (Figure 1). Relative binding averaged -1 ± 1.7, 7 ± 11.7. 10 ± 16.1. and 15 ± 23.0% (mean ± SE) for the control and groups receiving 6. 1.8. and 3.0 g. respectively. However. only 13 out of 59 cows receiving sometribove developed anti-bSTbinding activity >25% (positives): 4 cows receiving .6 g. 3 cows receiving 1.8 g, and 6 cows receiving 3.0 g (Table 1). In all cases. the increase in binding was detected within the first 7 wk of treatment. and. over the course of lactation. the number of cows with anti-bST binding generally declined. A similar initial increase in binding followed by a gradual decline was reported previously for hSTtreated humans (25, 31). Five cows in the groups receiving 1.8 and 3.0 g sustained circulating anti-bST binding greater than 25% through the entire first lactation. In the .6-g group, no positive cows (binding >25%) were detected after wk 15 of the fIrst treatment period. Within bST treatment groups. the FCM production of positive cows was similar to that of herdmates (Figure 2). Journal of Dairy Science Vol. 75. No. II, 1992
During the 2nd yr. relative binding averaged -1 ± 3.4. 2 ± 3.3, 3 ± 4.2, and 16 ± 25.1 % for the control. .6-. 1.8-. and 3.O-g groups. respectively. Of the cows with binding greater than 25% during the 1st yr. no cows were declared moribund or died. but 6 cows were selected randomly for necropsy at the end that lactation. Of the remaining 7 positive responders from the 1st yr. 1 cow receiving .6 g, 1 cow receiving 1.8 g, and 3 cows receiving 3.0 g had no increase in binding during the 2nd yr (relative to their pretreatment values for yr 1). Only 2 cows. both in the 3.0-g group, exhibited anti-bST binding >25% (Table 2). The relative binding in these 2 cows' sera was >75% after wk 3 of treatment. It is not apparent why these cows responded differently than their herdmates. However. the binding apparently did not attenuate because milk production of these cows (37.7 and 48.0 kg/d) was among the highest for the study (9). A low level of anti-bST binding during each year of treatment is consistent with previous results (4. 5, 33). Mean IgG concentrations (not specifIc for bST) were within normal
2963
SOMATOTROPIN ANTIBODIES IN COWS 80
70 •
'0
60
Ii!! E3
1.8 g of bST
o
3.DgofbST
.6gOfbST
•
•
COntrol
lso '" 'g 40
•
ST BIndIng ('\0)
3.5% FCM (kg/d)
Figure I. Percentage of anti-bST-binding activity in bST-treated cows at wk 3 of the 1st yr of study. Binding is expressed relative to pretreatment and positive control values.
Figure 2. Effect of serum anti-bST-binding activity on each cow's average 3.5% FCM production during yr I.
ranges following daily injection of 10.3 or 20.6 mg of bST/d (5), and anti-bST binding was unaffected by daily injection of up to 40.5 mg of bST (2). In contrast, Zwickl et al. (33) reported that nearly all bST-treated cows deve-
loped anti-bST IgG antibodies with no clinical health effects during each of two lactations of treatment. The molecular variants of bST utilized in each of these studies probably varied in structure at the amino terminus: three addi-
TABLE 2. Frequency! of anti-bST antibody responses to sometribove administered during the second of two consecutive lactations. Week of study
Relative binding
Overall
bST Control
.6 g
14/14 0114 0114 12/14 2/14 0/14 14114 0114 0114 14/14 0114 0114 14114 0114 0114
8/8 0/8 0!8 8/8 0/8 0/8 8/8 0/8
;(2
1.8 g
3.0 g
P value
m
719 1/9 1/9
.362
(%) -2, -I
3
7
15
23
$5 5-25 >2.5 $5 5-25 >25 $5 5-25 >25 $5 5-25 >25 $5
5-25 >25
017 017
5n 2n 017 6n
In
018 618 2/8
017 4n
0/8
017
5/6 1/6 0/6
5n
3n
2n 017
5/9 219 S/9 419 3/9 2/9
.172
619 1/9 2/9
.062
5/9 219 2/9
.296
.133
!Frequency of relative anti-bST binding reported as the proportion of cows in each group with $5. 5 to 25. or >25% (positive) binding. Chi-square analyses conducted for frequency across 5% intervals. e.g., $5%, >5 but $10%, >10 but $15%, etc. Binding levels during treatment are adjusted for background titers (wk -2, -1 of yr I). Most cows were dried off prior to the wk-31 sampling. Journal of Dairy Science Vol. 75. No. II, 1992
2964
EPPARD ET AL.
tional amino acids (5, 13), nine additional amino acids (4, 13, 33), and methionine substituted for alanine in the bST used in the present study. Although molecular variants of bST probably affect antibody response, to date no positive or negative clinical health effects have been noted for any bST variant (4, 5, 33). Moreover, only 2 out of 42 cows injected for 21 d with 25 mgld of one of four recombinant bST variants or pituitary-derived bST (8) developed anti-bST binding >25% (data not presented). Interstudy differences in antibody response may be due to dissimilar antibody assays, molecular aggregation of bST in vivo, bST dosages, or frequency of treatment. As previously described (I, 5, 33), a principle Ig that is produced in response to exogenous somatotropin is IgG. The anti-bSTbinding activity in sera from cows with binding >25% in the present study was further analyzed by HPLC to identify IgG and IgM. An example chromatogram of the serum fractions with associated IgG- and IgM-binding activities is presented in Figure 3. Immunoglobulin G was found in fractions 22 to 27, primarily eluting in fractions 22 and 23, and IgM eluted in fractions 35 to 37. The fractions were then reassayed for anti-bST-
80 .30
binding activity. In all cases, anti-bST binding was associated with the IgG peak (fractions 22 and 23). Thus, the binding activity probably is not due to an endogenous bST-binding protein as has been described for humans (3). Scatchard analysis indicated that binding capacities were highly variable, ranging from .625 to 3.04 mg of bSTlL, and the affinities ranged from 1.14 x 108 to 3.14 X 108 LImo\. These antibodies did not appear to affect the galactopoietic response to bST (Figure 2) but may have resulted in slightly underestimated serum bST concentrations (9). At 3 and 10 dafter bST administration, injection sites were scored for swelling (0 = none, 1 = minimal, 2 = moderate, 3 = severe). Injection site response did not appear to be related to anti-bST binding (Figure 4). Of the 13 cows with >25% relative binding, only 2 cows (both receiving 3.0 g of bST) had repeated incidences of adverse injection site swelling. Moreover, immune-mediated tissue damage, such as glomerular lesions secondary to activation of complement and the resultant compromise in renal function, was not observed in cows (6), regardless of serum bST binding. At necropsy (6), interstitial fibrosis or regeneration of tubular epithelium was noted in kidneys from 3 of 13 cows with elevated anti-bST binding (>25%) at some time during the study. However, similar lesions were also noted in controls, and all but one of the lesions were of only minimal or slight severity. In
60
=> 00
.20 40
..
C :
