milk progesterone enzyme immunoassay: modifications and a field trial for pregnancy detection in...

Milk Progesterone Enzyme Immunoassay: Modifications and a Field Trial for Pregnancy Detection in Dairy Cows 1'2

T. H. WIMPY, C. F. CHANG, V. L. ESTERGREEN, and J. K. HILLERS

Department of Animal Sciences Washington State University

College of Agriculture and Home Economics Pullman 99164-6332

ABSTRACT

A milk progesterone enzyme immunoassay was modified to shorten the antibody:antigen incubation time and tested in a field trial. Coupling the antibody to paper fibers at pH 7 increased the binding activity of the paper-antibody conjugate to allow incubation for 3 h at room temperature with no significant loss of competitive binding of progesterone. The relationship between progesterone concentrations measured by the modified and original enzyme immunoassays was r=.94 (n=80 pairs). Milk samples (n=67) collected 21 d after artificial insemination were classified by the original and modified methods as being from pregnant or nonpregnant cows using both spectrophotometric and visual evaluations. Com- parison to subsequent rectal palpation or return to estrus showed that the two methods were comparable. A field trial was conducted involving 622 cows and 40 producers using the modified enzyme immunoassay method. The overall field trial accuracy of the enzyme immunoassay in diagnosing the cow's reproductive status for a single sample on d 21 after breeding was 71% for correct diagnosis of pregnancy and 81% for correct diagnosis of nonpregnancy. It was concluded that the modified method is a valuable technique for rapid monitoring of milk progesterone concentrations.

Received May 21, 1985. Scientific Paper Number 7147. College of Agricul-

ture and Home Economics Research Center, Washing- ton State University, Projects 0353 and 7353.

Supported in part by the National Association of Animal Breeders, Columbia, MO.

INTRODUCTION

Numerous workers have demonstrated the usefulness of measuring milk progesterone for monitoring the cow's reproductive status (2, 3, 6, 10, 11, 12, 13, 16)s ince progesterone was first identified in milk in 1972 (7). As a test for pregnancy, milk progesterone at d 21 to 24 after breeding is 75 to 85% accurate in diagnosing animals classified pregnant and is 92 to 100% accurate for cows classified nonpregnant (3, 5, 8, 12, 14, 18).

Determining progesterone concentrations with the radioimmunoassay (RIA) requires a central laboratory and special equipment. Shipping, processing, and response time is from 2 to 10 d (3). Hence, cows that do not show estrous behavior and are checked for progesterone by RIA must wait another 21 d before being rebred. These missed breeding oppor- tunities result in longer intervals from parturi- tion to rebreeding, resulting in lower economic returns to the dairy producer (17).

We developed an enzyme immunoassay (EIA) for determining progesterone content of milk (5). For a single day sample on d 21 after breeding, EIA was 76% accurate for predicting pregnancy and 95% accurate for predicting nonpregnancy.

Objectives of the present study were 1) to modify the EIA further so that results could be obtained in a few hours without using a centri- fuge or spectrophotometer without loss of accuracy and 2) to evaluate the effectiveness of the modified milk progesterone EIA under field conditions for determining the reproductive status of cows on d 21 after breeding.

MATERIALS AND METHODS

All materials used have been previously de- scribed (5) except the antibody. The antibody (Ab) was produced in a rabbit against 11~-

1986 J Dairy Sci 69:1115-1121 11 15

1116 WIMPY ET AL.

hydroxyprogesterone conjugated to bovine serum albumin. Crossreactivity was less than 12% for desoxycorticosterone, testosterone, 20a-hydroxy-4-pregnen-3-one, 17a-hydroxyprogesterone,, 20~hydroxy-4-pregnen-3-one, preg- nenolone, estrone, and corticosterone. As expected from the antigen preparation, l la- hydroxyprogesterone had a higher crossreactivity (78%). The accuracy, sensitivity, and precision of the EIA are the same as previously reported (5).

The modified EIA method used in the field trial made use of paper-Ab that had been conjugated at pH 7.0. About 1.8 mg of paper-Ab was placed into a test tube with .5 ml of phos- phate-buffered saline (PBS) containing gelatin and thimerosal (Sigma Chemical Co., St. Louis, MO). Duplicate .2-ml aliquots of the milk sample, which had been warmed to 37°C and thoroughly mixed, were added and followed by .1 ml of a 1:2000 dilution of progesterone- horseradish peroxidase conjugate (5). The con- tents were gently shaken and incubated at room temperature of 22°C for 3 h. The liquid con- tents were decanted and the solid phase paper- Ab was washed twice with 2.4 ml of PBS. The enzyme substrate (2.0 ml) was added to the paper-Ab and incubated in the dark for 1 h at room temperature. To stop the reaction and intensify the color, .5 ml of 8 N sulfuric acid was added.

Visual comparison of unknown samples to milk samples from known pregnant and nonpregnant cows was used to classify the cows as pregnant, nonpregnant, or questionable. All cows classified as questionable were not in- cluded in any further analysis. Producer records of rectal palpation by consulting veterinarians or dates of subsequent breedings or estrus were obtained. A comparison of the EIA results to these records was made to establish the accuracy of the EIA.

Modification and Testing of the Enzyme Immunoassay

Incubations of paper-Ab with progesterone- horseradish peroxidase for 3 h at 17°C, 3 h at 22°C (room temperature), 3 h at 28°C, and 2 h at 37°C were tested and compared with incuba- t ion overnight at 4°C (original method, 5). The 3 h at 22°C (modified method) was satisfactory (Figure 1).

Progesterone in 80 postmilking stripping samples from Holstein cows in various stages of reproduction were analyzed using both the original and modified methods. The results were correlated between the two methods.

Postmilking strippings were obtained from other Holstein cows (n=67) 21 d after artificial insemination and frozen. Reproductive status was determined by the milk progesterone EIA using the original and modified methods and compared with subsequent rectal palpation or return to estrus. Visual observation of color intensity and calculations of milk progesterone based on absorbance (OD) at 492 nm were used to classify the milk samples as being from pregnant (light in color, >6 ng/ml) or nonpregnant (dark in color, <4 ng/ml) cows. From the results obtained, it was concluded that the modified EIA method was suitable for use in the field trial.

Field Trial

Forty dairy producers in three areas of

EFFECT OF INCUBATION TIME AND TEMPERATURE

1"6 L Overnight 4"C e - - - - - - e 1.4~-~ 3h 17"C D - - - - - - D

I~L 3h 22"C 1.2 I;-amIN 3h 28"C LI- ' - - - - -~

E I ~ ' ~ 2h 35"C O - - - - - - O

ff 0 •

. 2

I • I

0 2.5 5.0 10 20

PROGESTERONE (ng/mi) Figure 1. Effect of incubation time and tempera-

ture on the enzyme immunoassay.

Journal of Dairy Science Vol. 69, No. 4, 1986

MILK PROGESTERONE ENZYME IMMUNOASSAY 11 17

Washington cooperated by collecting postmilking strippings on selected cows over an l l-d period in each area. Calving date, previous breeding dates, and reproductive treatments given during the current lactation were recorded. Milk samples were collected without preservative from cows bred by artificial insemination 21 d before the collection date except for cows that returned to estrus in 18 or fewer d. A technician analyzed the refrigerated samples daily in a central laboratory in each of the three areas. A total of 622 milk samples were analyzed from 548 Holstein, 54 Jersey, and 20 Guernsey cows.

Retrospectively, EIA classification of cows was compared with results of palpation of reproductive organs per rectum or return to estrus. The percentage classified correctly by EIA was considered the accuracy of diagnosing the cow's reproductive state.

In one area, 233 milk samples were also analyzed for percent fat, percent protein, and somatic cell counts by the Washington State Dairy Herd Improvement Association (DHIA) laboratory. The DHIA data and other informa- t ion supplied by the producers were used to evaluate possible interactions with the accuracy of the EIA.

Statistical Analysis

The modified method was compared to the original method using correlations and linear regressions. Chi-square analysis was used to determine if percent fat, percent protein, somatic cell counts, times bred, postpartum interval, or reproductive treatments given during the current lactation altered the accuracy of the EIA (20).

RESULTS

Modification of the Enzyme Immunoassay

Incubation for 3 h at room temperature (22°C) yielded results most similar to the original method (Figure 1). Other times and tem- peratures showed less overall change in OD for the standard curve. The 17°C incubation temperature yielded much lower binding as shown by the low OD at zero progesterone (Figure 1). To give good visual discrimination, it is desirable to have a change in OD of .5 or more between 0 and 5 ng/ml of progesterone. The

modified method (3 h at 22°C) fits this cri- terion and was tested further.

The relationship of the concentrations of milk progesterone found by the modified and the original methods in 80 samples is shown in Figure 2. The linear regression equation between these two methods shows that the original method measures slightly more progesterone than the modified method (Y = 1.3x - .63), but the correlation between them was high (r=.94, P<.05).

Accuracy of the original method for detect- ing pregnancy in the 67 cows sampled was 91% for the spectrophotometric and 86% for visual determinations. The modified E1A resulted in 89% (spectrophotometric) and 82% (visual) accuracy for the same milk samples. Accuracy for nonpregnant diagnosis was 100% in all methods. It was concluded that the modified method using visual determinations was accept- able for use in a field trial.

Field Trial

The field trial accuracy of the EIA for deter-

24

I Y = 1.30X - .63 r = .94

20 n = 8 0 • • • • co ; •

E 16

¢-

,< t.,u

12

z

Ilg 0 8

• - y ./. I B • •

N •

41- / •

4 8 t2 16

MODIFIED EIA (ng/ml)

Figure 2. Relationship between the concentrations of progesterone as measured by the original and modified enzyme immunoassay methods.


1118 W[MPY ET AL.

mining pregnancy status is shown in Table 1. Of the 359 cows called pregnant by EIA o f milk proges terone , 254 (71%) were conf i rmed pregnant by palpat ion and of the 208 cows called nonp regnan t by EIA, 168 (81%) were conf i rmed nonpregnant . The 55 cows called ques t ionable by EIA were de te rmined to be 69% nonpregnan t and 31% pregnant .

Milk samples (n=233) were g rouped arbi- trarily into upper , middle, and lower thirds

for percent fat, pe rcen t protein , and somat ic cell coun t by DHIA analysis. Accuracy o f the EIA was lowered (P< .05) for pregnancy diagnosis in the upper thi rd for somat ic celI counts (Table 2). Other groups were no t significantly d i f fe ren t f rom the overall results wi thin each classification.

When the pos tpa r tum interval (calving to breeding) was used to group cows, accuracy of pregnancy evaluat ion by EIA was no t a f fec ted

TABLE 1. Accuracy of the modified milk progesterone enzyme immunoassay (EIA) during the field trial.

Classification by EIA Confirmed Non- Ques- status 1 pregnant Pregnant tionable Total

Nonpregnant Percent 81 29 69 Number 168 105 38 311

Pregnant Percent 19 71 31 Number 40 254 17 311

1 Confirmed status of the cow was determined from records of rectal palpation or return to estrus.

TABLE 2. Accuracy of the modified milk progesterone enzyme immunoassay (E1A) for predicting nonpregnancy and pregnancy when samples are grouped into upper, middle, and lower thirds for percent fat, percent protein, or somatic cell counts.

Accuracy of the EIA

Item Nonpregnant Pregnant

(%) (n) (%) (n)

Percent fat Upper (9.3-20) 68 22 73 48 Middle (6.5-9.3) 95 20 70 56 Lower (.7--6.5) 78 32 76 29

Percent protein Upper (3.2-5.6) 83 23 75 44 Middle (2,8--3.2) 75 24 63 46 Lower (1.6-2.8) 82 27 79 43

Somatic cell counts Upper (>323,000) 75 24 51 a 45 Middle (134,000- 318,000) 91 22 88 48 Lower (<134,000) 75 28 78 40

aChi-square analysis (P<.05). Comparison of overall accuracy to group accuracy within E1A classification.


MILK PROGESTERONE ENZYME IMMUNOASSAY 1119

(Table 3). Accu racy of n o n p r e g n a n c y evalua- t ion b y EIA was increased ( P< .05 ) for cows w i th p o s t p a r t u m intervals grea ter t h a n 90 d.

Analys is of the da ta f rom cows rebred wi th in 3 d o f t h e tes t ing date (n=100) showed t h a t 80% were called n o n p r e g n a n t b y EIA, 11% were called p regnan t , and 9% were called ques- t ionab le , wh ich is n o t d i f fe ren t f rom t he overall accuracy. The n u m b e r of t imes a cow was b red (once or m o r e t h a n once ) also had no effect o n t h e accuracy of the EIA.

R e p r o d u c t i v e t r e a t m e n t s given dur ing the p o s t p a r t u m interval t h a t lowered ( P < . 0 5 ) t he accuracy of p regnancy diagnosis were h u m a n chor ion ic g o n a d o t r o p i n , gonado t rop in - re l eas ing h o r m o n e or its analogs, and es t rogens (55, 50, and 45% accuracy, respect ively, c o m p a r e d wi th

74% for t he u n t r e a t e d cows) (Table 4). The low n u m b e r of animals receiving some t r e a t m e n t s makes it d i f f icul t to draw conc lus ions f rom the ef fec t of these t r e a t m e n t s o n the accuracy of t he EIA. A l t h o u g h increases in the accuracy of diagnosis o f n o n p r e g n a n c y b y EIA were ev ident f rom some t r e a t m e n t s (Table 4), these changes were no t s ignif icant .

DISCUSSION

The field tr ial gave resul ts lower in accuracy t h a n t he work d o n e in o u r l a b o r a t o r y (5) and also lower t h a n o t h e r r epor t s (11, 15, 19). One d i f fe rence in the field trial c o m p a r e d wi th previous l a b o r a t o r y work is t h a t the samples col- lec ted in the field trial were k e p t at 4°C and

TABLE 3. Effect of postpartum interval on the accuracy of pregnancy and nonpregnancy evaluation by enzyme immunoassay (EIA).

Accuracy of the EIA Postpartum interval 1 Nonpregnant Pregnant

(%) (n) (%) (n) <90 d 74 89 73 146 >90 d 87 a 108 70 204

aChi-square analysis (P<.05). Comparison of overall accuracy to group accuracy within EIA classification.

t Postpartum interval is calculated as the days from calving to breeding.

TABLE 4. Effect of reproductive treatments given during the current lactation on the accuracy of pregnancy and nonpregnancy evaluation by the enzyme immunoassay (EIA).

Accuracy of the EIA Treatment Nonpregnant Pregnant

(%) (n) (%) (n) None 79 97 74 191 Aborted 71 7 67 6 Prostaglandins 92 24 70 56 Human chorionic gonadotropin 91 11 55 a 20 Gonadotropin-releasing hormone or its analogs 76 25 50 a 28 Prebreeding intrauterine infusions 83 54 69 97 Postbreeding intrauterine infusions 100 7 83 6 Estrogens 78 2 $ 46 a 22

aDifferent from overall percent accuracy of pregnancy diagnosis by chi-square (P<.05). Other accuracies were not statistically different from the overall percent accuracy of nonpregnancy or pregnancy diagnosis.


1120 WIMPY ET AL.

were analyzed within 18 h after collection. All previous work in the laboratory was conducted on frozen samples. Studies to ascertain any differences from the handling of the milk are currently underway, although handling of milk in either manner had no effect on milk progesterone concentrations when using RIA tech- niques (8). Another possible reason for the lowered accuracy could be from sampling the wrong cows, but the extent of this error cannot be determined, since each producer sampled his own cows.

Accuracy of pregnancy diagnosis by EIA was slightly lower than the original report (71 vs. 76%). Errors in this category are from variations in cycle length, embryonic mortality, and breeding cows during the luteal phase. It has been reported that only 91% of unbred cows have low progesterone on d 21 after ob- served estrus because of variations in cycle length (18). Embryonic mortality is reported to be 10 to 28% between d 14 to 70 (4, 6, 19). In both aberrant cycling and embryonic mortality, milk samples from d 21 could be expected to be high in progesterone, and yet the cow would be nonpregnant when palpation is performed later. Breeding cows in the luteal phase of the estrous cycle has been reported to occur in 5 to 54% of all breedings (1, 4, 6, 11, 18, 19). Such cows have a very low conception rate and would be expected to be nonpregnant although high progesterone would occur on d 21 postbreeding as a result of the luteal phase of their next cycle. Because of embryonic mortality, aberrant cycling, and improper timing of artificial insemination, reducing the errors in the accuracy of determining pregnancy by EIA may be impossible.

One factor that lowered the accuracy of pregnancy diagnosis by EIA was high somatic cells (Table 2). Highest somatic cell count (>323,000) is within the range of a trace (T) or above by a California Mastitis Test or a 5 or above on the DHIA scale. The reason for high somatic cell count samples giving high progesterone values by EIA from nonpregnant cows has not yet been determined.

Reproductive treatments that lowered the accuracy of pregnancy diagnosis by EIA were gonadotropin-releasing hormone, human chorionic gonadotropin, and estrogens (Table 3). Because no dates were recorded on these treatments, any direct effect on the EIA results can-

not be determined, but those treatments are commonly used for treating retained corpora lutea and cystic ovaries. It is possible that these cows had elevated progesterone from retained corpora lutea or luteal cysts, and this was re- flected by the EIA results.

Lowered efficiency of detection of nonpregnancy in the field trial is more difficult to rationalize. Some possible errors other than method error could have been the result of sampling the wrong cow or extremely low milk fat. Ginther et al. (9) have shown that fat content can effect the progesterone in milk. Although not statistically significant, milk with extreme fat amounts did have lower accuracy for predicting nonpregnancy (Table 2), which could account for some missed classifications by our EIA. A transitory decrease in progesterone at 3 wk postbreeding in pregnant cows has been reported by Claus et al. (6). Lower progesterone from such a transitory decrease may be suffi- cient to cause error in our EIA determination of some cows actually pregnant.

In conclusion, the modified EIA method provides the ability to test a milk sample on day of collection and have the results within a few hours. Possible uses for the measurement of milk progesterone include testing on the day of breeding, d 21 postbreeding, and thrice weekly (4, 6, 11, 19). Thrice weekly analysis can be used to determine postpartum cyclicity and ovarian dysfunction with a consulting veterinarian (2). Current work in our laboratory is focusing 1) on further modifications of the EIA so that the producer can analyze his own milk samples and 2) on ways to standardize the influence of such factors as percent fat, somatic cell counts, and reproductive treatments the cows receive. The end result should be a simple, fast, and accurate test for measuring progesterone in milk.

ACKNOWLEDGMENTS

The authors would like especially to thank extension agents Eddie Thomason, Dick Mathews, and Jack Crawford and the 40 dairy producers for their involvement in the field trial. Appreciation is also expressed to the Pilchuck Veterinary Hospital, Yakima County DHIA, and the Washington State DHIA Labora- tory for allowing use of their facilities to analyze the field trial samples.


MILK PROGESTERONE ENZYME IMMUNOASSAY 1 121

REFERENCES

1 Appleyard, W. T., and B. Cook. 1979. The detect ion of oestrus in dairy cattle. Vet. Rec. 99:253.

2 Ball, P.J.H., and N. W. Jackson. 1979. The fertility o f dairy cows inseminated on the basis of milk progesterone measurements . Br. Vet. J. 132:537.

3 Bishop, C. A., C. P. Bond, and C. Roberts. 1976. Early diagnosis o f nonpregnancy in cattle: the first eighteen m o n t h s of a commercial service. Br. Vet. J. 132:529.

4 Bulman, D. C., and G. E. Lamming. 1979. The use of milk progesterone analysis in the s tudy of oestrus detection, herd fertility and embryonic mortal i ty in dairy cows. Br. Vet. J. 135:559.

5 Chang, C. F., and V. L. Estergreen. 1983. Develop- men t o f a direct enzyme immunoassay of milk progesterone and its application to pregnancy diagnosis in cows. Steroids 41:173.

6 Claus, R., H. Karg, D. Zwiauer, I. von Butler, F. Pirchner, and E. Rattenberger. 1983. Analysis of factors influencing reproductive performance of the dairy cow by progesterone assay in milk fat. Br. Vet. J. 139:29.

7 Darling, J.A.B., R. W. Kelly, A. H. Laing, and R. A. Harkness. 1972. The isolation and identification of progesterone obtained from cows' milk during pregnancy. J. Endocrinol. 54: 347.

8 Foote, R. H., E.A.B. Oltenacu, H. L. Kummerfe ld , R. D. Smith, P. M. Riek, and R. K. Braun. 1979. Milk progesterone as a diagnostic aid. Br. Vet. J. 135:550.

9 Ginther, O. J., L. C. Nuti, M. C. Garcia, B. C. Wentworth, and W. J. Tyler. 1976. Factors affect- ing progesterone concentrat ion in cow's milk and dairy products. J. Anim. Sci. 42:155.

10 Ginther, O. J., L. Nuti, B. C. Wentworth, and W. J. Tyler. 1974. Progesterone concentrat ion in milk and blood during pregnancy in cows. Proc. Soc. Exp. Biol. Med. 146:354.

11 Gunzler, O., E. Rattenberger, A. Gorlach, R. Hahn,

P. Hocke, R. Claus, and H. Karg. 1979. Milk progesterone determinat ions as applied to the confir- mat ion of oestrus, the detection of cycling and as an aid to veterinarian and biotechnical measures in the cow. Br. Vet. J. 135:541.

12 Heap, R. B., M. Gwyn, J. A. Laing, and D. E. Waiters. 1973. Pregnancy diagnosis in cows; changes in milk progesterone concentra t ion during the oestrous cycle and pregnancy measured by a rapid radioimmunoassay. J. Agric. Sci. Camb. 81: 151.

13 Heap, R. B., A. Henville, and J. L. Linzell. 1975. Metabolic clearance rate, product ion rate and m a m m a r y uptake and metabol ism of progesterone in cows. J. Endocrinol. 66:239.

14 Heap, R. B., R. J. Holdsworth, J. E. Gadsby, J. A. Laing, and D. E. Waiters. 1976. Pregnancy diagnosis in the cow from milk progesterone concen- tration. Br. Vet. J. 132:445.

15 Hoffman, B., O. Gunzler, R. Hamburger, and W. Schmidt. 1976. Milk progesterone as a parameter for fertility control in cattle: methodological ap- proaches and present status of application in Germany. Br. Vet. J. 132:469.

16 Laing, J. A., and R. B. Heap. 1971. The concentra- t ion of progesterone in the milk o f cows during the reproductive cycle. Br. Vet. J. 127:xix.

17 Pelissier, C. L. 1972. Herd breeding problems and their consequences. J. Dairy Sci. 55:385.

18 Pennington, J. A., S. L. Spabr, and J. R. Lodge. 1976. Pregnancy diagnosis in dairy cattle by progesterone concentra t ion in milk. J. Dairy Sci. 59: 1528.

19 Reimers, T. J., R. D. Smith, and R. H. Foote. 1980. Milk progesterone test ing to determine reproductive status of cows. Page 906 /n XI Int. Congr. Dis. Cattle. E. Mayer, ed. Bregman Press, Haifa, Israel.

20 Steel, R.G.D., and J. T. Torrie. 1980. Principles and procedures of statistics. McGraw-Hill Book Co., New York, NY.


milk progesterone enzyme immunoassay: modifications and a field trial for pregnancy detection in...

Documents