Endotoxin Mastitis in Cows Milked Four Times Daily 1 D. E. SHUSTER,2 R. J. HARMON, J. A. JACKSON, and R. W. HEMKEN Department of Animal Sciences University of Kentucky lexington 4Q546.{)215 ABSTRACT
synthesis of the different milk components. (Key words: endotoxin mastitis, milk production, somatic cell count)
As part of a project to identify the pathophysiological cause or causes of mastitic hypogalactia, midlactation cows were infused in two homolateral quarters with 10 /lg of endotoxin while being milked four times daily to resolve better the temporal changes in mammary synthetic activity during endotoxin mastitis. Milk fat was decreased by the first milking (5 h) postinfusion and then recovered rapidly. In contrast, milk yield and the yields of protein and lactose were not significantly inhibited until the second milking, and these yields recovered slowly thereafter. The decline in milk yield by infused quarters was only 20% greater than the decline by uninfused quarters in this experiment. Mammary inflammation developed rapidly in infused quarters as milk serum albumin concentration was maximal at the first milking. Milk see and NAGase were also elevated at this time, and maximal levels occurred at miIkings 2 to 4. Increased temperature, increased cortisol, and a mild anorexia were apparent at the first milking only. Endotoxin treatment had no effect on serum prolactin or glucose. These data suggest that the delayed hypogalactia is consequent to the mammary inflammation and systemic responses following endotoxin infusion. The results indicate that different pathophysiological events may inhibit
Abbreviation key: BSA = bovine serum albumin, TNF = tumor necrosis factor. INTRODUCTION
Received September 17, 1990. Accepted January IS, 1991. l1bis manuscript (90-5-134) is published with the approval of the director of the Kentucky Agricu1lura1 Experiment Station. This study was supported, in part, by a grant from The Upjohn Company, Kalamazoo, MI. 2current address: National Animal Disease Center, PO Box 70, Ames, IA 50010. 1991 J Dairy Sci 74:1527-1538
The negative relationship of elevated see with milk production has been well established (3,4,9, 15). However, the fundamental causes for the reduced synthetic activity of mammary secretory cells during mastitis remain unknown. Common suggestions as to the cause or causes of the hypogalactia include anorexia, bacterial toxin or inflammatory damage to mammary tissue, hormonal changes, and effects of inflammatory mediators. These effects can alter the synthetic activity of the mammary epithelium in one of three general ways: physical damage to the epithelial cells, interference with substrate availability for milk synthesis, or alteration of the metabolic activity of milk-producing cells either by a reduced concentration of a galactopoietic hormone or an increased concentration of an inhibitory hormone or inflammatory mediator. During a series of experiments to investigate pathophysiological events that may mediate the suppression of milk production during endotoxin mastitis, it became clear that many of the productive, inflammatory, and systemic responses were so rapid that a definitive study of the temporal pattern of these responses was not possible when cows were milked twice daily [(8); Shuster et al., unpublished data]. In this experiment, endotoxin mastitis was induced in cows that were milked four times daily so that temporal resolution could be improved. The objectives were to determine which inflammatory and systemic responses occur prior to the decline in milk synthesis so that possible cause and effect relationships could be identified, and so that changes in synthesis of
1527
SHUSTER ET AL.
1528
different milk components could be compared to determine whether these components may be regulated differently. MATERIALS AND METHODS
Cows
Healthy, midlactation Holstein or Jersey cows, free of intramammary infection and producing 15 to 30 kg milk/d, were used for this experiment. Four cows were assigned to the endotoxin treatment, and two cows served as untreated controls. Cows received a total mixed ration consisting of alfalfa and corn silages plus concentrate and mineral supplement. Just prior to each milking, orts were weighed and 1.5 kg of a pelleted concentrate were offered. Cows were milked every 6 h (four times per day) with a quarter milking machine. To ensure more complete and consistent milk removal at each milking with the intensive milking frequency, cows were given 10 ill of oxytocin via the coccygeal artery or vein immediately after the milking machine was removed. The machine was reattached within 1 min to harvest the residual milk. Cows were allowed a 3-d adjustment period to become accustomed to the milking frequency and experimental conditions before any data were collected. The following four milkings served as a preliminary period to determine baseline values for all cows and quarters prior to treatment. Treatments were administered approximately 1 h following the fourth preliminary period milking. Responses were then followed during the next 3.5 d. sampling Procedures
Coccygeal blood samples and rectal temperatures were taken as the rust step in the routine at each milking. The blood was refrigerated (4°C) until the serum was harvested within 15 h of collection. Two milk samples were taken from each quarter reservoir of the milking machine. One sample was preserved with potassium dichromate and refrigerated for subsequent milk composition and SCC analysis. The other milk sample was stored (4"C) for NAGase analysis and whey sample preparation. Whey samples were prepared within 15 h of milk collection. Milk was centrifuged (9000 x Journal of Dairy Science Vol. 74, No.5, 1991
g, 10 min, 4"C) after which the skim was warmed in a water bath (37°C) before acidification to pH 4.5 with glacial acetic acid. The acidified whey was centrifuged (9000 x g, 10 min, 4"C), and a portion of the supernatant was stored frozen (-20°C).
Treatments Endotoxin, a trichloroacetic acid extract from Escherichia coli 055:B5 (Sigma Chemical Co., St. Louis, MO) was dissolved in pyrogenfree Hanks balanced salt solution and filter (.2 ~m) sterilized. Treated cows were infused in each of two homolateral quarters with 10 ~g of endotoxin in 10 ml of Hanks balanced salt solution. Control cows remained untreated. Assays
The fat, protein, and lactose composition of milk were determined by infrared spectrophotometry. Milk sec was determined with a Coulter counter (Coulter Electronics, Hialeah, FL). The NAGase assay was a modification of the fluorometric assay described by Kitchen et al. (10). Lactoferrin and bovine serum albumin (BSA) concentrations of whey samples were measured with a sandwich enzyme immunoassay (18). The intraassay and interassay coefficients of variation were 8 and 14% for the BSA assay and 5 and 10% for the lactoferrin assay. The lactoferrin assay yielded values slightly lower than those given by an electroimmunodiffusion assay (18). Serum glucose was measured using a commercial assay kit that utilizes glucose oxidase (510-A, Sigma Chemical Co., St. Louis, MO). Semm cortisol was measured with a solidphase radioimmunoassay kit (Coat-A-Count Cortisol, Diagnostic Products Corporation, Los Angeles, CA) as adapted for bovine serum by Schillo et al. (17). The assay used antibody coated tubes and 12SI-Iabeled cortisol. The intraassay and interassay coefficients of variation were 4 and 12%. Any samples that read below the standard curve, i.e., below 2.5 ng/mI, were assigned a value of 2.0 ng/ml. All samples were assayed in duplicate. Serum prolactin was measured by double antibody radioimmunoassay essentially as described by Forrest et al. (5). Prolactin for iodination (bPrl-Il) and standards (USDA-bPrI-Bl) and antisera (DJB-7-0330)
ENDOTOXIN MASTITIS WITII FREQUENT MILKING
were graciously provided by D. J. Bolt (Beltsville, MD). The intraassay and interassay coefficients of variation were less than 15%. All samples for each trial were analyzed in duplicate in the same assay to negate effects of interassay variation. Statistical Analysis
This experiment was conducted in two separate trials with a control and two endotoxintreated cows in each trial. The preliminary period, the 1st d (four milkings) following the 3-d adjustment period, was used to detennine baseline values for all parameters. To correct the data for initial differences among cows and quarters, all data were converted to percentages by dividing all values by the arithmetic preliminary period means for the corresponding cow and quarter. Data for sec, NAGase, BSA, lactoferrin, cortisol, and prolactin were then logarithmically transfonned. All treatment comparisons were made by t test (16). An F test of variance equality was made at the a-level of .05 and the appropriate t test was used for treatment comparisons at each milking. Milk parameter data for all quarters receiving the same treatment within each cow were averaged after conversion to percentages and logarithmic transformation. These data were then used for between cow treatment comparisons, i.e., control versus infused or uninfused quarters. For within-cow comparisons (i.e., infused versus uninfused quarters), a paired t test was made using the data from each cow as one data pair. For all tests, an a level of .05 was used to detennine significance. This analysis correctly accounted for the much larger variance in the treatment group relative to the control group that resulted from the heterogeneity in responsiveness of cows to endotoxin. RESULTS
Inflammation, as indicated by udder edema, was very marked at the first milking (5 h) after infusion. Despite this inflammation in treated cows, milk yields ranged from 90 to 94% of the preliminary period level for control cows compared with 88 to 100% for the infused and uninfused quarters of endotoxin cows (Figure 1). Milk production was not reduced signifi-
1529
cantly in any quarters of the four treated cows by milking 1 (P > .05). By the second milking, milk production had dropped precipitously in endotoxin treated cows. Milk production then recovered during the next several milkings. Only a slight difference (in magnitude) in milk production existed between infused and uninfused quarters of treated cows. In contrast to the delayed effect on milk yield, the fat content of milk: was reduced (P < .05) by the first milking (Figure 2). The fat response was biphasic with a nonsignificant elevation at the third milking. The milk protein response was delayed with peak milk protein composition at third milking. The lactose content of milk from endotoxin-infused quarters declined rapidly and remained low for many milkings; little change occurred in milk from uninfused quarters (P > .05). Because component yields, and not percentage composition, reflect actual synthetic rates, fat, protein, and lactose yields were calculated. Changes in component yields from uninfused and infused quarters were similar (Figure 3). In both, a more rapid decline and recovery of fat yield occurred than the yields of protein and lactose. Fat yield was reduced at the first and second milking and had recovered rapidly by the third milking. In contrast, lactose and protein yields remained normal at the first milking, declined markedly by the second milking, and recovered slowly during the next several milkings. Mammary inflammation developed rapidly and was generally limited to infused quarters. Milk sec and NAGase were elevated (P < .05) by the first milking following endotoxin infusion with peak levels occurring during milkings 2 to 4 (Figure 4). The BSA response appeared more rapid in that the BSA concentration of milk was highest at the first milking and gradually declined to normal levels thereafter (Figure 5). Lactoferrin showed the characteristic slow response with peak concentrations occurring roughly 2 d after endotoxin infusion. The systemic responses developed rapidly after endotoxin infusion. A nonsignificant increase in temperature was observed at milking 1 (P> .05) (Figure 6). Much more striking was the 5- to lo-fold increase in serum cortisol at this milking. Both temperature and cortisol had returned to normal by the second milking. Prolactin concentrations, which were measured at milkings -1 to 4, were unaffected by endotoxin treatment (P > .05). No effect of endotoxin mastitis on senun glucose was evident in this 10umaI of Dairy Science Vol. 74, No.5, 1991
1530
SHUSTER BT AL. CONTROL UNINFUSED ...... INFUSED
.....Ol
-oz X
i=
()
1.5
::J
o
o a: a..
:x: ...J
~
.5
T -3
,t o
4
MILKING
8
12
(4X/day)
Figure 1. Milk production by infused and uninfused quarters of cows given endotoxin 1 h after milking 0 (arrow) and by quarters of control cows. Treatments at the same milking with the same or no letter do not differ significantly (P > .05).
experiment (P > .05) (Figure 7). Endotoxin cows offered pelleted concentrate at first milking appeared moderately anorexic. This observation was supported by the data, which showed a nonsignificant reduction in feed consumption during the period immediately following endotoxin infusion (P > .05). During this experiment, which was conducted in July and August, the cows suffered moderate heat stress, as evidenced by basal temperatures of some cows that exceeded 4O"C. This stress may have accounted for some of the seemingly large fluctuations in rectal temperature and feed intake, which partially explains the lack of statistical significance for some responses in these parameters. DISCUSSION
The results of this experiment generally confirmed results obtained in cows milked twice Journal of Dairy Science Vol. 74. No.5. 1991
daily during endotoxin mastitis (l, Shuster et al., unpublished data). In these earlier experiments, endotoxin infusion reduced milk yield and lactose composition of milk, increased protein composition, and induced an inflammatory response that was primarily limited to infused quarters. These results show that endotoxin infusion in two quarters reduces milk yield in all quarters despite the absence of mammary inflammation in uninfused quarters. These data indicate that part of the suppression in milk synthesis is mediated systemically. The observation that milk and lactose yields changed after the initiation of mammary inflammation is consistent with the theory that these changes are a consequence of local and systemic pathophysiological changes associated with mammary inflammation. Further experiments are required to establish a relationship between cause and effect. The early 10% decline in milk lactose composition did not
ENDOTOXIN MAS11TIS -
CONlROl
• _.
UNINFUSED
wrm
1531
FREQUENT MILKING
•••••• INFUSED
5
le(
u..
3
'i
3.6
--iiiz I-
oa:: Q.
3.0
5
""' w
~
rn
0
I-
0
e(
..J
4 T
-3
,t
o
4
MILK ING
8
12
(4X/day)
Figure 2. Fat, protein, and lactose composition of milk from infused and uninfusod quarters of cows given endotoxin 1 h after milking 0 (arrow) and from qUMtcrs of control cows. Trealmc:nts at the same milking with the same or no letter do not differ significantly (P > .OS).
contribute to a large decrease in lactose yield and may have resulted from the escape of lactose into the blood. The blood-milk barrier was very penneable at this time as shown by high BSA levels. Furthennore, the excretion of lactose into urine is increased shortly after in-
tramammary endotoxin infusion (Shuster et al., unpublished data). The earlier suppression of fat yield relative to changes in milk volume and yields of protein and lactose indicate that different pathophysiological changes mediate the suppression of fat secretion, or, alternatively, JoumaI of Dairy Science Vol. 74, No.5, 1991
1532
SHUSTER ET AL.
140
120
~
•
~
Milkfot
Protein Lactose
".......
~
0
.t: Q)
a..
.g
100
...a..
Q)
~ ........
80
9 w
>=
60
120t
........ "0 0 .t:
100
Q)
a..
.g Q)
...c..
80
~ ........
9 w
60
>= -1
o
2
3
4
5
6
7
MILKING (4X/day) Figure 3. Fat, protein, and lactose yields, as a percentage of the average yield during the preliminary period, from infused (top) and uninfused (bottom) quarters of cows given endotoxin 1 h after milking O. Asterisks denote values that differ significantly from the corresponding yield of control cows (P < .05).
that the synthesis of milk fat has a different temporal responsiveness to the same effector(s). The elevated fat content of milk at milking 3 in this experiment and probably the higher fat content of milk during endotoxin mastitis in cows milked twice daily (Shuster et al., unpublished data) do not represent hypersecretion of milk fat, as fat yields did not greatly exceed Journal of Dairy Science Vol. 74, No.5, 1991
basal levels. This high fat composition of milk results from a more rapid recovery of fat secretion than milk yield so that the same amount of fat is secreted in a smaller volume of milk. Thus, milk fat synthesis underwent a rapid, short-duration suppression, and at no time was milk fat synthesis enhanced. The elevated protein content of milk that occurred in infused
1533
HNDOTOXIN MASn11S WI1H FREQUENT MILKING -
CONT~OL.
• _. UNINFUSED .••••. INFUSED
3
,... E ..... III
Qi
to>
oJ'
() () (/)
,...
b b .! . ,..---
6
j
.....E c:
::J
:
~
I
ell
:
3
co
.~ _..!- -.:. -
/:/
o
c(
>
c
z
•
-3
o
. .
""'~' ,~_.,~, ,~ •
".
~...~..'...~'.='.~~
•
...
• •
4
MILKING
~
.!'••...... ~
"
•
8
..
..... _ _ w
12
(4X/day)
Figure 4. Somatic cell cOUDt and NAGase activity of milk from infused and uninfused quancrs of cows given endotoxin 1 h after milking 0 (arrow) and from quancrs of control cows. Treatments at the same milking with the same or no letter do not differ significantly (P > .05).
and uninfused quarters did not result primarily from the influx of serum proteins because the milk BSA response, and thus the permeability changes, clearly occurred earlier than the peak in milk protein composition. The high protein composition more likely reflects a relatively greater suppression of milk volume than milk protein synthesis. Like fat secretion, milk protein secretion never exceeded basal levels, as indicated by the protein yields, which were suppressed by endotoxin infusion. Similarly, only the early decline in lactose composition
can be explained by permeability changes and the low milk lactose concentration later in the response probably reflects relatively greater suppression of lactose synthesis than milk volume. Studies in other species have established that tumor necrosis factor a (TNF) inhibits lipoprotein lipase activity when these animals are given endotoxin (2). This inhibition causes an increase in the lipid content of blood because this enzyme, which hydrolyzes fatty acids from triacylglycerol, is required for the uptake Joarnal of Dairy Science Vol. 74, No.5, 1991
SHUS1l!R ET AL.
1534 _
CONTROL
.-
UNINFUSEO
.•••.. INFLJSEO
.6
,r······ E
:
......
.4
CIl
\a
E
< en m
.
b
: f
.2
b/-
/-'~"""~~
. -......
.....•..
•
..."'--_.-
~ ....... ~~..::.:------::>-~~~ '. b b b :;:b~
.2
.....E CIl
E
z a: a: w
.1
LL.
o
~
(J
< ~ '3
o
4
MILKING
8
12
(4X/day)
Figure 5. Bovine serum albumin (BSA) and Iactoferrin concentration of milk from infused and uninfused quarters of cows given endotoxin 1 h after milking 0 (arrow) and from quarteR of control cows. Treatments at the same milking with the same or no letter do DOt differ significantly (P > .05).
of fatty acids by many tissues of the body. Because the mammary gland is also dependent upon lipoprotein lipase for the uptake of longchain fatty acids for milk synthesis (11), TNFinduced inhibition of lipoprotein lipase could presumably reduce the synthesis of milk fat by the mammary gland Following endotoxin treatment of rodents, TNF is rapidly produced and circulates with a short half-life (2). Intravenous Journal of Dairy Science Vol. 74, No.5, 1991
or intramammary administration of endotoxin to cattle is thought to result in circulating TNF (13). Thus, the rapid and short-duration decline in fat synthesis is consistent with possible mediation by TNF. Part of the mastitic hypogalactia could ~ tentially be mediated by incomplete milk ejection. Obvious milk ejection failure is occasionally observed during natural cases of mastitis.
1535
ENDOTOXIN MAS1lTIS WI1H FREQUENT MILKING
42
-
CONTROL.
......
INFUSED
,....
0°..... w
a:
:::>
I-
< a:
40
w
Q. ~
W I-
38
50
,....
E ....
01
.....c: -J
30
0
(/)
I-
a: 0
°
10
~
80
01
.....c: Z I-
°~
~"""""""""',,::-.---? ................
40
o
a:
Q.
'3
o
4
MILKING
8
12
(4X/day)
Figure 6. Rectal temperature and serum cOitisol and prolactin concentrations of cows infused with endotoxin 1 h after milking 0 (arrow) and of control cows. Treatments at the same milking with the same or no letter do not differ significantly (P > .05).
In this experiment, the cows were injected with exogenous oxytocin and milk ejection appeared to be complete, but a decline in milk yield still occurred. It is possible that congestion of the udder during mastitis prevented oxytocin from
reaching the myoepithelial cells, and milk removal could still have been incomplete. Because udder congestion and mastitic algesia seemed most intense at the first milking postinfusion, when milk yields were normal, udder Journal of Dairy Science Vol. 74, No. S, 1991
1536
SHUSTER. ET AL.
-
CONTROL
..•••.
INFUSEO
70
30
w ~
.OS).
congestion apparently did not interfere with milk ejection. Furthermore, no compensatory increase in milk yield was seen at a later milking as would be expected if milk ejection had been inhibited. The findings that systemic responses such as pyrexia, enhanced serum cortisol concentration, and anorexia are short-lived and occur soon after intramammary endotoxin infusion corroborate the results of others (8, 14). However, serum glucose concentrations remained unaffected, in contrast to results in cows milked twice daily (Shuster et al., unpublished data) and with the results of Paape et al. (14) in which serum glucose concentrations were Journal of Dairy Science Vol. 74. No.5, 1991
elevated during endotoxin mastitis. The reason for the discrepancy is unknown. The large cortisol response is consistent with the response being a possible mediator of mastitic hypogalactia as other studies have demonstrated that elevated glucocorticoid concentrations may impair lactational perfonnance (7, 19). The lack of a detectable prolactin response agrees with earlier findings (8) that intramammary endotoxin administration produces little change in serum prolactin so prolactin probably does not mediate the mastitic hypogalactia. The brief anorexia observed in this experiment was of such low magnitude that the anorexia probably was not a major contributor to the
ENDOTOXIN MASTI11S WI1H FREQUENT MD..KING
hypogalactia This conclusion is supported by the absence of a hypoglycemic period associated with the anorexia In cows milked twice daily, a much greater inhibition of milk production occurred in endotoxin-infused quarters relative to uninfused quarters (Shuster et al., unpublished data). This greater depression in infused quarters was referred to as the local suppression of milk production because this suppression was attributed to the local effects of inflammation in these quarters. The finding that milk production is inhibited more in infused quarters is a consistent finding in studies of endotoxin or infectious mastitis in cows milked twice daily (1, 6, 12; Shuster et al., unpublished data). Therefore, the small difference in milk yield between infused and uninfused quarters in this experiment was surprising. The local suppression of milk: production was diminished. A number of differences are apparent between this experiment and other endotoxin experiments including heat stress, oxytocin injections, and four times daily milking frequency. Only the higher milking frequency reasonably explains the small local suppression of milk production. Although this experiment was not designed to test the effects of milking four times daily, the results indicate that higher milking frequency may diminish local effects of mammary inflammation on milk production. The local suppression is probably caused by penneability changes in the bloodmilk barrier, mammary edema associated congestion and pressure, the leukocyte influx, or locally produced inhibitors. More frequent milking would remove locally produced inhibitors and reduce intramammary pressure, which presumably would decrease the escape of milk components and minimize detrimental effects of edematous pressure on epithelial secretory activity. More frequent milking probably would not be able to modify adverse effects of the leukocyte influx as a prominent leukocytosis still occurred. CONCLUSIONS
Mammary inflammation and systemic responses develop before any alteration in mammary synthetic activity except a reduction in milk fat synthesis. These data support the hypothesis that changes in lactational perfonnance are consequent to mammary inflammation
1537
and systemic responses. Milk fat synthesis declines before, and recovers sooner, than other aspects of mammary function, which indicates that pathophysiological mechanisms affecting milk fat synthesis may be different and may involve TNF. Incomplete milk letdown and anorexia probably play only a minor role in mastitic hypogalactia. Frequent milking appears to abrogate the local suppression of milk production. ACKNOWLEDGMENTS
The assistance and hospitality offered by Frank Button and the Coldstream dairy farm personnel during this experiment are gratefully acknowledged. REFERENCES 1 Anderson, K. L., A. R Smith, R. D. Shanks, L. E. Davis, and B. K. Gustafsson. 1986. Efficacy of flunixin meglumiDe for the treatment of endotoxin-induced b0vine mastitis. Am. I. VeL Res. 47:1366. 2 Beutlez, B., and A. Cerami. 1986. Cachectiu and tumour necrosis factor as two sides of the same biological coin. Nature (Lond.) 320:584. 3 Eberhart, R I., L. I. Hutchinson, and S. B. Spencer. 1982. Relationships of bulk tank somatic cell counts to prevalence of intramammary infection and to indices of herd production. I. Food Prot. 45:1125. 4 Fetrow, I .• K. Anderson, S. Sexton, and K. Butcher. 1988. Herd composite somatic cell counts: average linear score and weighted average somatic cell count score and milk production. I. Dairy Sci. 71:257. 5 Forrest, D. W., I. L. Fleeger, C. R. Long. A. M Sorenson, Ir., and P. G. Harms. 1980. Effect of exogenous prolactin on peripheral luteinizing hormone levels in ovariectomized cows. BioI. Reprod. 22: 197. 6 Harmon, R. I., F. L. Schanbacher, L. C. Ferguson, and K. L. Smith. 1976. Changes in lactoferrin, immunoglobulin G, bovine serum albumin, and a-lactalbumin during acute experimental and natural coliform mastitis in cows. lnfecL Immun. 13:533. 7 Hartmann, P. E., and D. S. Kronfeld. 1973. Mammary blood flow and glucose uptake in lactating cows given dexamethasone. I. Dairy Sci. 56:896. 8 Iackson, I. A., D. E. Shuster, W. I. Silvia, and R J. Harmon. 1990. Physiological responses to intramammary or intravenous treatment with endotoxin in lactating dairy cows. I. Dairy Sci. 73:627. 9 lones, G. M., R. E. Pearson, G. A. Clabaugh, and C. W. Heald. 1984. Relationships between somatic cell counts and milk production. I. Dairy Sci. 67:1823. 10 Kitchen, B. I., G. Middleton, and M Salmon. 1978. Bovine milk N-acetyl-~D-glucosaminidaseand its significance in the detection of abnormal udder secretions. J. Dairy Res. 45:15. 11 Larson, B. L., ed. 1985. Biosynthesis and cellular secretion of milk. Page 129 in Lactation. Iowa State Univ. Press, Ames. 12 Lefcourt, A. M, I. BilInall, D. L. Wood, H. Tao, B. Iournal of Dairy Science Vol. 74, No.5, 1991
1538
SHUSTER ET AL.
Stroud, and W. D. Schultze. 1983. Radiotelemetry temperature respouses of mammary glaDd and body to intnunammary infusion of E. coli endotoxin and S. agalactiae in lactating dairy cows. 1. Dairy Sci. 66(Suppl. 1):214.(Abslr.) 13 Lohuis, l.A.C.M., J.H.M. Verlieijden, C. Burvenich, and A.SJP.A.M. van Mien. 1988. Pathophysiological effects of endotoxins in rumiDants. 1. Changes in body temperature and reticulo-rumen motility, and the effect of repeated administration. Vet. Q. 10:109. 14 Paape, M. J., W. D. Schultze, C. Desjardins, and R. H. Miller. 1974. Plasma corticosteroid, ciJculating leukocyte and milk somatic cell respouses to Escheric/UQ coli endotoxin-induced mastitis. Proc. Soc. Exp. BioI. Med. 145:553. 15 Raubertas, R. F., and G. B. Shook. 1982. Relationship
10urnal of Dairy Science Vol. 74, No.5, 1991
between lactation measures of somatic cell concentration and milk yield. J. Dairy Sci. 65:419. 16 SAS~ User's Guide: Statistics. 1982. SAS Inst., Inc., Cary, NC. 17 Schillo, K.. K., M. A. Green, and S. H. Hayes. 1988. Effects of adrenalectomy on photoperiod-induced changes in release of luteinizing hormone and prolactin in ovariectomized ewes. J. Reprod. Fertil. 83:431. 18 Shuster, D. B., and R. J. Harmon. 1990. Enzyme immunoassays of bovine Iactoferrin and serum albumin in acid precipitated and ultraeentrifugal wheys. J. Dairy Sci. 73:3104. 19 Varner, M. A., and B. H. Johnson. 1983. Influence of adrenocorticotropin upon milk production, milk constituents, and endocrine measures of dairy cows. J. Dairy Sci. 66:458.
synthesis of the different milk components. (Key words: endotoxin mastitis, milk production, somatic cell count)
As part of a project to identify the pathophysiological cause or causes of mastitic hypogalactia, midlactation cows were infused in two homolateral quarters with 10 /lg of endotoxin while being milked four times daily to resolve better the temporal changes in mammary synthetic activity during endotoxin mastitis. Milk fat was decreased by the first milking (5 h) postinfusion and then recovered rapidly. In contrast, milk yield and the yields of protein and lactose were not significantly inhibited until the second milking, and these yields recovered slowly thereafter. The decline in milk yield by infused quarters was only 20% greater than the decline by uninfused quarters in this experiment. Mammary inflammation developed rapidly in infused quarters as milk serum albumin concentration was maximal at the first milking. Milk see and NAGase were also elevated at this time, and maximal levels occurred at miIkings 2 to 4. Increased temperature, increased cortisol, and a mild anorexia were apparent at the first milking only. Endotoxin treatment had no effect on serum prolactin or glucose. These data suggest that the delayed hypogalactia is consequent to the mammary inflammation and systemic responses following endotoxin infusion. The results indicate that different pathophysiological events may inhibit
Abbreviation key: BSA = bovine serum albumin, TNF = tumor necrosis factor. INTRODUCTION
Received September 17, 1990. Accepted January IS, 1991. l1bis manuscript (90-5-134) is published with the approval of the director of the Kentucky Agricu1lura1 Experiment Station. This study was supported, in part, by a grant from The Upjohn Company, Kalamazoo, MI. 2current address: National Animal Disease Center, PO Box 70, Ames, IA 50010. 1991 J Dairy Sci 74:1527-1538
The negative relationship of elevated see with milk production has been well established (3,4,9, 15). However, the fundamental causes for the reduced synthetic activity of mammary secretory cells during mastitis remain unknown. Common suggestions as to the cause or causes of the hypogalactia include anorexia, bacterial toxin or inflammatory damage to mammary tissue, hormonal changes, and effects of inflammatory mediators. These effects can alter the synthetic activity of the mammary epithelium in one of three general ways: physical damage to the epithelial cells, interference with substrate availability for milk synthesis, or alteration of the metabolic activity of milk-producing cells either by a reduced concentration of a galactopoietic hormone or an increased concentration of an inhibitory hormone or inflammatory mediator. During a series of experiments to investigate pathophysiological events that may mediate the suppression of milk production during endotoxin mastitis, it became clear that many of the productive, inflammatory, and systemic responses were so rapid that a definitive study of the temporal pattern of these responses was not possible when cows were milked twice daily [(8); Shuster et al., unpublished data]. In this experiment, endotoxin mastitis was induced in cows that were milked four times daily so that temporal resolution could be improved. The objectives were to determine which inflammatory and systemic responses occur prior to the decline in milk synthesis so that possible cause and effect relationships could be identified, and so that changes in synthesis of
1527
SHUSTER ET AL.
1528
different milk components could be compared to determine whether these components may be regulated differently. MATERIALS AND METHODS
Cows
Healthy, midlactation Holstein or Jersey cows, free of intramammary infection and producing 15 to 30 kg milk/d, were used for this experiment. Four cows were assigned to the endotoxin treatment, and two cows served as untreated controls. Cows received a total mixed ration consisting of alfalfa and corn silages plus concentrate and mineral supplement. Just prior to each milking, orts were weighed and 1.5 kg of a pelleted concentrate were offered. Cows were milked every 6 h (four times per day) with a quarter milking machine. To ensure more complete and consistent milk removal at each milking with the intensive milking frequency, cows were given 10 ill of oxytocin via the coccygeal artery or vein immediately after the milking machine was removed. The machine was reattached within 1 min to harvest the residual milk. Cows were allowed a 3-d adjustment period to become accustomed to the milking frequency and experimental conditions before any data were collected. The following four milkings served as a preliminary period to determine baseline values for all cows and quarters prior to treatment. Treatments were administered approximately 1 h following the fourth preliminary period milking. Responses were then followed during the next 3.5 d. sampling Procedures
Coccygeal blood samples and rectal temperatures were taken as the rust step in the routine at each milking. The blood was refrigerated (4°C) until the serum was harvested within 15 h of collection. Two milk samples were taken from each quarter reservoir of the milking machine. One sample was preserved with potassium dichromate and refrigerated for subsequent milk composition and SCC analysis. The other milk sample was stored (4"C) for NAGase analysis and whey sample preparation. Whey samples were prepared within 15 h of milk collection. Milk was centrifuged (9000 x Journal of Dairy Science Vol. 74, No.5, 1991
g, 10 min, 4"C) after which the skim was warmed in a water bath (37°C) before acidification to pH 4.5 with glacial acetic acid. The acidified whey was centrifuged (9000 x g, 10 min, 4"C), and a portion of the supernatant was stored frozen (-20°C).
Treatments Endotoxin, a trichloroacetic acid extract from Escherichia coli 055:B5 (Sigma Chemical Co., St. Louis, MO) was dissolved in pyrogenfree Hanks balanced salt solution and filter (.2 ~m) sterilized. Treated cows were infused in each of two homolateral quarters with 10 ~g of endotoxin in 10 ml of Hanks balanced salt solution. Control cows remained untreated. Assays
The fat, protein, and lactose composition of milk were determined by infrared spectrophotometry. Milk sec was determined with a Coulter counter (Coulter Electronics, Hialeah, FL). The NAGase assay was a modification of the fluorometric assay described by Kitchen et al. (10). Lactoferrin and bovine serum albumin (BSA) concentrations of whey samples were measured with a sandwich enzyme immunoassay (18). The intraassay and interassay coefficients of variation were 8 and 14% for the BSA assay and 5 and 10% for the lactoferrin assay. The lactoferrin assay yielded values slightly lower than those given by an electroimmunodiffusion assay (18). Serum glucose was measured using a commercial assay kit that utilizes glucose oxidase (510-A, Sigma Chemical Co., St. Louis, MO). Semm cortisol was measured with a solidphase radioimmunoassay kit (Coat-A-Count Cortisol, Diagnostic Products Corporation, Los Angeles, CA) as adapted for bovine serum by Schillo et al. (17). The assay used antibody coated tubes and 12SI-Iabeled cortisol. The intraassay and interassay coefficients of variation were 4 and 12%. Any samples that read below the standard curve, i.e., below 2.5 ng/mI, were assigned a value of 2.0 ng/ml. All samples were assayed in duplicate. Serum prolactin was measured by double antibody radioimmunoassay essentially as described by Forrest et al. (5). Prolactin for iodination (bPrl-Il) and standards (USDA-bPrI-Bl) and antisera (DJB-7-0330)
ENDOTOXIN MASTITIS WITII FREQUENT MILKING
were graciously provided by D. J. Bolt (Beltsville, MD). The intraassay and interassay coefficients of variation were less than 15%. All samples for each trial were analyzed in duplicate in the same assay to negate effects of interassay variation. Statistical Analysis
This experiment was conducted in two separate trials with a control and two endotoxintreated cows in each trial. The preliminary period, the 1st d (four milkings) following the 3-d adjustment period, was used to detennine baseline values for all parameters. To correct the data for initial differences among cows and quarters, all data were converted to percentages by dividing all values by the arithmetic preliminary period means for the corresponding cow and quarter. Data for sec, NAGase, BSA, lactoferrin, cortisol, and prolactin were then logarithmically transfonned. All treatment comparisons were made by t test (16). An F test of variance equality was made at the a-level of .05 and the appropriate t test was used for treatment comparisons at each milking. Milk parameter data for all quarters receiving the same treatment within each cow were averaged after conversion to percentages and logarithmic transformation. These data were then used for between cow treatment comparisons, i.e., control versus infused or uninfused quarters. For within-cow comparisons (i.e., infused versus uninfused quarters), a paired t test was made using the data from each cow as one data pair. For all tests, an a level of .05 was used to detennine significance. This analysis correctly accounted for the much larger variance in the treatment group relative to the control group that resulted from the heterogeneity in responsiveness of cows to endotoxin. RESULTS
Inflammation, as indicated by udder edema, was very marked at the first milking (5 h) after infusion. Despite this inflammation in treated cows, milk yields ranged from 90 to 94% of the preliminary period level for control cows compared with 88 to 100% for the infused and uninfused quarters of endotoxin cows (Figure 1). Milk production was not reduced signifi-
1529
cantly in any quarters of the four treated cows by milking 1 (P > .05). By the second milking, milk production had dropped precipitously in endotoxin treated cows. Milk production then recovered during the next several milkings. Only a slight difference (in magnitude) in milk production existed between infused and uninfused quarters of treated cows. In contrast to the delayed effect on milk yield, the fat content of milk: was reduced (P < .05) by the first milking (Figure 2). The fat response was biphasic with a nonsignificant elevation at the third milking. The milk protein response was delayed with peak milk protein composition at third milking. The lactose content of milk from endotoxin-infused quarters declined rapidly and remained low for many milkings; little change occurred in milk from uninfused quarters (P > .05). Because component yields, and not percentage composition, reflect actual synthetic rates, fat, protein, and lactose yields were calculated. Changes in component yields from uninfused and infused quarters were similar (Figure 3). In both, a more rapid decline and recovery of fat yield occurred than the yields of protein and lactose. Fat yield was reduced at the first and second milking and had recovered rapidly by the third milking. In contrast, lactose and protein yields remained normal at the first milking, declined markedly by the second milking, and recovered slowly during the next several milkings. Mammary inflammation developed rapidly and was generally limited to infused quarters. Milk sec and NAGase were elevated (P < .05) by the first milking following endotoxin infusion with peak levels occurring during milkings 2 to 4 (Figure 4). The BSA response appeared more rapid in that the BSA concentration of milk was highest at the first milking and gradually declined to normal levels thereafter (Figure 5). Lactoferrin showed the characteristic slow response with peak concentrations occurring roughly 2 d after endotoxin infusion. The systemic responses developed rapidly after endotoxin infusion. A nonsignificant increase in temperature was observed at milking 1 (P> .05) (Figure 6). Much more striking was the 5- to lo-fold increase in serum cortisol at this milking. Both temperature and cortisol had returned to normal by the second milking. Prolactin concentrations, which were measured at milkings -1 to 4, were unaffected by endotoxin treatment (P > .05). No effect of endotoxin mastitis on senun glucose was evident in this 10umaI of Dairy Science Vol. 74, No.5, 1991
1530
SHUSTER BT AL. CONTROL UNINFUSED ...... INFUSED
.....Ol
-oz X
i=
()
1.5
::J
o
o a: a..
:x: ...J
~
.5
T -3
,t o
4
MILKING
8
12
(4X/day)
Figure 1. Milk production by infused and uninfused quarters of cows given endotoxin 1 h after milking 0 (arrow) and by quarters of control cows. Treatments at the same milking with the same or no letter do not differ significantly (P > .05).
experiment (P > .05) (Figure 7). Endotoxin cows offered pelleted concentrate at first milking appeared moderately anorexic. This observation was supported by the data, which showed a nonsignificant reduction in feed consumption during the period immediately following endotoxin infusion (P > .05). During this experiment, which was conducted in July and August, the cows suffered moderate heat stress, as evidenced by basal temperatures of some cows that exceeded 4O"C. This stress may have accounted for some of the seemingly large fluctuations in rectal temperature and feed intake, which partially explains the lack of statistical significance for some responses in these parameters. DISCUSSION
The results of this experiment generally confirmed results obtained in cows milked twice Journal of Dairy Science Vol. 74. No.5. 1991
daily during endotoxin mastitis (l, Shuster et al., unpublished data). In these earlier experiments, endotoxin infusion reduced milk yield and lactose composition of milk, increased protein composition, and induced an inflammatory response that was primarily limited to infused quarters. These results show that endotoxin infusion in two quarters reduces milk yield in all quarters despite the absence of mammary inflammation in uninfused quarters. These data indicate that part of the suppression in milk synthesis is mediated systemically. The observation that milk and lactose yields changed after the initiation of mammary inflammation is consistent with the theory that these changes are a consequence of local and systemic pathophysiological changes associated with mammary inflammation. Further experiments are required to establish a relationship between cause and effect. The early 10% decline in milk lactose composition did not
ENDOTOXIN MAS11TIS -
CONlROl
• _.
UNINFUSED
wrm
1531
FREQUENT MILKING
•••••• INFUSED
5
le(
u..
3
'i
3.6
--iiiz I-
oa:: Q.
3.0
5
""' w
~
rn
0
I-
0
e(
..J
4 T
-3
,t
o
4
MILK ING
8
12
(4X/day)
Figure 2. Fat, protein, and lactose composition of milk from infused and uninfusod quarters of cows given endotoxin 1 h after milking 0 (arrow) and from qUMtcrs of control cows. Trealmc:nts at the same milking with the same or no letter do not differ significantly (P > .OS).
contribute to a large decrease in lactose yield and may have resulted from the escape of lactose into the blood. The blood-milk barrier was very penneable at this time as shown by high BSA levels. Furthennore, the excretion of lactose into urine is increased shortly after in-
tramammary endotoxin infusion (Shuster et al., unpublished data). The earlier suppression of fat yield relative to changes in milk volume and yields of protein and lactose indicate that different pathophysiological changes mediate the suppression of fat secretion, or, alternatively, JoumaI of Dairy Science Vol. 74, No.5, 1991
1532
SHUSTER ET AL.
140
120
~
•
~
Milkfot
Protein Lactose
".......
~
0
.t: Q)
a..
.g
100
...a..
Q)
~ ........
80
9 w
>=
60
120t
........ "0 0 .t:
100
Q)
a..
.g Q)
...c..
80
~ ........
9 w
60
>= -1
o
2
3
4
5
6
7
MILKING (4X/day) Figure 3. Fat, protein, and lactose yields, as a percentage of the average yield during the preliminary period, from infused (top) and uninfused (bottom) quarters of cows given endotoxin 1 h after milking O. Asterisks denote values that differ significantly from the corresponding yield of control cows (P < .05).
that the synthesis of milk fat has a different temporal responsiveness to the same effector(s). The elevated fat content of milk at milking 3 in this experiment and probably the higher fat content of milk during endotoxin mastitis in cows milked twice daily (Shuster et al., unpublished data) do not represent hypersecretion of milk fat, as fat yields did not greatly exceed Journal of Dairy Science Vol. 74, No.5, 1991
basal levels. This high fat composition of milk results from a more rapid recovery of fat secretion than milk yield so that the same amount of fat is secreted in a smaller volume of milk. Thus, milk fat synthesis underwent a rapid, short-duration suppression, and at no time was milk fat synthesis enhanced. The elevated protein content of milk that occurred in infused
1533
HNDOTOXIN MASn11S WI1H FREQUENT MILKING -
CONT~OL.
• _. UNINFUSED .••••. INFUSED
3
,... E ..... III
Qi
to>
oJ'
() () (/)
,...
b b .! . ,..---
6
j
.....E c:
::J
:
~
I
ell
:
3
co
.~ _..!- -.:. -
/:/
o
c(
>
c
z
•
-3
o
. .
""'~' ,~_.,~, ,~ •
".
~...~..'...~'.='.~~
•
...
• •
4
MILKING
~
.!'••...... ~
"
•
8
..
..... _ _ w
12
(4X/day)
Figure 4. Somatic cell cOUDt and NAGase activity of milk from infused and uninfused quancrs of cows given endotoxin 1 h after milking 0 (arrow) and from quancrs of control cows. Treatments at the same milking with the same or no letter do not differ significantly (P > .05).
and uninfused quarters did not result primarily from the influx of serum proteins because the milk BSA response, and thus the permeability changes, clearly occurred earlier than the peak in milk protein composition. The high protein composition more likely reflects a relatively greater suppression of milk volume than milk protein synthesis. Like fat secretion, milk protein secretion never exceeded basal levels, as indicated by the protein yields, which were suppressed by endotoxin infusion. Similarly, only the early decline in lactose composition
can be explained by permeability changes and the low milk lactose concentration later in the response probably reflects relatively greater suppression of lactose synthesis than milk volume. Studies in other species have established that tumor necrosis factor a (TNF) inhibits lipoprotein lipase activity when these animals are given endotoxin (2). This inhibition causes an increase in the lipid content of blood because this enzyme, which hydrolyzes fatty acids from triacylglycerol, is required for the uptake Joarnal of Dairy Science Vol. 74, No.5, 1991
SHUS1l!R ET AL.
1534 _
CONTROL
.-
UNINFUSEO
.•••.. INFLJSEO
.6
,r······ E
:
......
.4
CIl
\a
E
< en m
.
b
: f
.2
b/-
/-'~"""~~
. -......
.....•..
•
..."'--_.-
~ ....... ~~..::.:------::>-~~~ '. b b b :;:b~
.2
.....E CIl
E
z a: a: w
.1
LL.
o
~
(J
< ~ '3
o
4
MILKING
8
12
(4X/day)
Figure 5. Bovine serum albumin (BSA) and Iactoferrin concentration of milk from infused and uninfused quarters of cows given endotoxin 1 h after milking 0 (arrow) and from quarteR of control cows. Treatments at the same milking with the same or no letter do DOt differ significantly (P > .05).
of fatty acids by many tissues of the body. Because the mammary gland is also dependent upon lipoprotein lipase for the uptake of longchain fatty acids for milk synthesis (11), TNFinduced inhibition of lipoprotein lipase could presumably reduce the synthesis of milk fat by the mammary gland Following endotoxin treatment of rodents, TNF is rapidly produced and circulates with a short half-life (2). Intravenous Journal of Dairy Science Vol. 74, No.5, 1991
or intramammary administration of endotoxin to cattle is thought to result in circulating TNF (13). Thus, the rapid and short-duration decline in fat synthesis is consistent with possible mediation by TNF. Part of the mastitic hypogalactia could ~ tentially be mediated by incomplete milk ejection. Obvious milk ejection failure is occasionally observed during natural cases of mastitis.
1535
ENDOTOXIN MAS1lTIS WI1H FREQUENT MILKING
42
-
CONTROL.
......
INFUSED
,....
0°..... w
a:
:::>
I-
< a:
40
w
Q. ~
W I-
38
50
,....
E ....
01
.....c: -J
30
0
(/)
I-
a: 0
°
10
~
80
01
.....c: Z I-
°~
~"""""""""',,::-.---? ................
40
o
a:
Q.
'3
o
4
MILKING
8
12
(4X/day)
Figure 6. Rectal temperature and serum cOitisol and prolactin concentrations of cows infused with endotoxin 1 h after milking 0 (arrow) and of control cows. Treatments at the same milking with the same or no letter do not differ significantly (P > .05).
In this experiment, the cows were injected with exogenous oxytocin and milk ejection appeared to be complete, but a decline in milk yield still occurred. It is possible that congestion of the udder during mastitis prevented oxytocin from
reaching the myoepithelial cells, and milk removal could still have been incomplete. Because udder congestion and mastitic algesia seemed most intense at the first milking postinfusion, when milk yields were normal, udder Journal of Dairy Science Vol. 74, No. S, 1991
1536
SHUSTER. ET AL.
-
CONTROL
..•••.
INFUSEO
70
30
w ~
.OS).
congestion apparently did not interfere with milk ejection. Furthermore, no compensatory increase in milk yield was seen at a later milking as would be expected if milk ejection had been inhibited. The findings that systemic responses such as pyrexia, enhanced serum cortisol concentration, and anorexia are short-lived and occur soon after intramammary endotoxin infusion corroborate the results of others (8, 14). However, serum glucose concentrations remained unaffected, in contrast to results in cows milked twice daily (Shuster et al., unpublished data) and with the results of Paape et al. (14) in which serum glucose concentrations were Journal of Dairy Science Vol. 74. No.5, 1991
elevated during endotoxin mastitis. The reason for the discrepancy is unknown. The large cortisol response is consistent with the response being a possible mediator of mastitic hypogalactia as other studies have demonstrated that elevated glucocorticoid concentrations may impair lactational perfonnance (7, 19). The lack of a detectable prolactin response agrees with earlier findings (8) that intramammary endotoxin administration produces little change in serum prolactin so prolactin probably does not mediate the mastitic hypogalactia. The brief anorexia observed in this experiment was of such low magnitude that the anorexia probably was not a major contributor to the
ENDOTOXIN MASTI11S WI1H FREQUENT MD..KING
hypogalactia This conclusion is supported by the absence of a hypoglycemic period associated with the anorexia In cows milked twice daily, a much greater inhibition of milk production occurred in endotoxin-infused quarters relative to uninfused quarters (Shuster et al., unpublished data). This greater depression in infused quarters was referred to as the local suppression of milk production because this suppression was attributed to the local effects of inflammation in these quarters. The finding that milk production is inhibited more in infused quarters is a consistent finding in studies of endotoxin or infectious mastitis in cows milked twice daily (1, 6, 12; Shuster et al., unpublished data). Therefore, the small difference in milk yield between infused and uninfused quarters in this experiment was surprising. The local suppression of milk: production was diminished. A number of differences are apparent between this experiment and other endotoxin experiments including heat stress, oxytocin injections, and four times daily milking frequency. Only the higher milking frequency reasonably explains the small local suppression of milk production. Although this experiment was not designed to test the effects of milking four times daily, the results indicate that higher milking frequency may diminish local effects of mammary inflammation on milk production. The local suppression is probably caused by penneability changes in the bloodmilk barrier, mammary edema associated congestion and pressure, the leukocyte influx, or locally produced inhibitors. More frequent milking would remove locally produced inhibitors and reduce intramammary pressure, which presumably would decrease the escape of milk components and minimize detrimental effects of edematous pressure on epithelial secretory activity. More frequent milking probably would not be able to modify adverse effects of the leukocyte influx as a prominent leukocytosis still occurred. CONCLUSIONS
Mammary inflammation and systemic responses develop before any alteration in mammary synthetic activity except a reduction in milk fat synthesis. These data support the hypothesis that changes in lactational perfonnance are consequent to mammary inflammation
1537
and systemic responses. Milk fat synthesis declines before, and recovers sooner, than other aspects of mammary function, which indicates that pathophysiological mechanisms affecting milk fat synthesis may be different and may involve TNF. Incomplete milk letdown and anorexia probably play only a minor role in mastitic hypogalactia. Frequent milking appears to abrogate the local suppression of milk production. ACKNOWLEDGMENTS
The assistance and hospitality offered by Frank Button and the Coldstream dairy farm personnel during this experiment are gratefully acknowledged. REFERENCES 1 Anderson, K. L., A. R Smith, R. D. Shanks, L. E. Davis, and B. K. Gustafsson. 1986. Efficacy of flunixin meglumiDe for the treatment of endotoxin-induced b0vine mastitis. Am. I. VeL Res. 47:1366. 2 Beutlez, B., and A. Cerami. 1986. Cachectiu and tumour necrosis factor as two sides of the same biological coin. Nature (Lond.) 320:584. 3 Eberhart, R I., L. I. Hutchinson, and S. B. Spencer. 1982. Relationships of bulk tank somatic cell counts to prevalence of intramammary infection and to indices of herd production. I. Food Prot. 45:1125. 4 Fetrow, I .• K. Anderson, S. Sexton, and K. Butcher. 1988. Herd composite somatic cell counts: average linear score and weighted average somatic cell count score and milk production. I. Dairy Sci. 71:257. 5 Forrest, D. W., I. L. Fleeger, C. R. Long. A. M Sorenson, Ir., and P. G. Harms. 1980. Effect of exogenous prolactin on peripheral luteinizing hormone levels in ovariectomized cows. BioI. Reprod. 22: 197. 6 Harmon, R. I., F. L. Schanbacher, L. C. Ferguson, and K. L. Smith. 1976. Changes in lactoferrin, immunoglobulin G, bovine serum albumin, and a-lactalbumin during acute experimental and natural coliform mastitis in cows. lnfecL Immun. 13:533. 7 Hartmann, P. E., and D. S. Kronfeld. 1973. Mammary blood flow and glucose uptake in lactating cows given dexamethasone. I. Dairy Sci. 56:896. 8 Iackson, I. A., D. E. Shuster, W. I. Silvia, and R J. Harmon. 1990. Physiological responses to intramammary or intravenous treatment with endotoxin in lactating dairy cows. I. Dairy Sci. 73:627. 9 lones, G. M., R. E. Pearson, G. A. Clabaugh, and C. W. Heald. 1984. Relationships between somatic cell counts and milk production. I. Dairy Sci. 67:1823. 10 Kitchen, B. I., G. Middleton, and M Salmon. 1978. Bovine milk N-acetyl-~D-glucosaminidaseand its significance in the detection of abnormal udder secretions. J. Dairy Res. 45:15. 11 Larson, B. L., ed. 1985. Biosynthesis and cellular secretion of milk. Page 129 in Lactation. Iowa State Univ. Press, Ames. 12 Lefcourt, A. M, I. BilInall, D. L. Wood, H. Tao, B. Iournal of Dairy Science Vol. 74, No.5, 1991
1538
SHUSTER ET AL.
Stroud, and W. D. Schultze. 1983. Radiotelemetry temperature respouses of mammary glaDd and body to intnunammary infusion of E. coli endotoxin and S. agalactiae in lactating dairy cows. 1. Dairy Sci. 66(Suppl. 1):214.(Abslr.) 13 Lohuis, l.A.C.M., J.H.M. Verlieijden, C. Burvenich, and A.SJP.A.M. van Mien. 1988. Pathophysiological effects of endotoxins in rumiDants. 1. Changes in body temperature and reticulo-rumen motility, and the effect of repeated administration. Vet. Q. 10:109. 14 Paape, M. J., W. D. Schultze, C. Desjardins, and R. H. Miller. 1974. Plasma corticosteroid, ciJculating leukocyte and milk somatic cell respouses to Escheric/UQ coli endotoxin-induced mastitis. Proc. Soc. Exp. BioI. Med. 145:553. 15 Raubertas, R. F., and G. B. Shook. 1982. Relationship
10urnal of Dairy Science Vol. 74, No.5, 1991
between lactation measures of somatic cell concentration and milk yield. J. Dairy Sci. 65:419. 16 SAS~ User's Guide: Statistics. 1982. SAS Inst., Inc., Cary, NC. 17 Schillo, K.. K., M. A. Green, and S. H. Hayes. 1988. Effects of adrenalectomy on photoperiod-induced changes in release of luteinizing hormone and prolactin in ovariectomized ewes. J. Reprod. Fertil. 83:431. 18 Shuster, D. B., and R. J. Harmon. 1990. Enzyme immunoassays of bovine Iactoferrin and serum albumin in acid precipitated and ultraeentrifugal wheys. J. Dairy Sci. 73:3104. 19 Varner, M. A., and B. H. Johnson. 1983. Influence of adrenocorticotropin upon milk production, milk constituents, and endocrine measures of dairy cows. J. Dairy Sci. 66:458.
