Project
title / Automatic Ovulation Prediction for Dairy Cows
/ MAFF
project code / LS2401

ministry of agriculture, fisheries and food CSG 15

Research and Development

Final Project Report

(Not to be used for LINK projects)

Two hard copies of this form should be returned to:
Research Policy and International Division, Final Reports Unit
MAFF, Area 6/01
1A Page Street, London SW1P 4PQ
An electronic version should be e-mailed to
Project title / Automatic Ovulation Prediction for Dairy Cows
MAFF project code / LS2401
Contractor organisation and location / Silsoe Research Institute
Wrest Park,
Silsoe
BEDFORD
MK45 4HS
Total MAFF project costs / £ 353,480
Project start date / 01/04/99 / Project end date / 31/03/02
Executive summary (maximum 2 sides A4)
To tab in this section press the tab key and the Control key together
Press the DOWN arrow once to move to the next question.

CSG 15 (1/00) 3

Project
title / Automatic Ovulation Prediction for Dairy Cows
/ MAFF
project code / LS2401

An automated biosensor system for detecting pregnancy and predicting ovulation in dairy cows has been built and tested. The system automates the sampling of milk, and the analysis of the milk by a deployment of a screen printed carbon electrode biosensor for progesterone. Test and calibration benchmarks to validate performance have been developed. The biosensor system measured concentrations of 5-20 ng/ml progesterone in fresh milk with an R2=0.965 in the process detecting pregnancy in 2 of 2 test cows and the ovulation cycles of five cows. Subsequent field tests are currently in process to predict ovulation cycles of dairy cows which were postponed due to health precautions associated with Foot and Mouth Disease.

The system was constructed based on a screen printed carbon electrode using a proprietary ink printed onto a pvc substrate, on which were deposited antibody layers to bind to progesterone in solution. Milk was applied to the sensor by a system of pumps and valves to allow the manual steps in operating the assay to be computer controlled. This was interfaced to a milking machine using a sampler loaned from a commercial collaborator. The whole system was built into a box that could be interfaced to any milking machine. The system deployed an automated competitive assay technique in a target time of 10 minutes. A number of problems were identified and overcome. A series of tests were developed to ensure that each stage of the process operated successfully. The quality of the antibody surface was tested with a colorimetric test. The electro-chemistry was tested by injecting various concentrations of 1-napthol into the chamber. A calibration solution can be used automatically to periodically test the accuracy of the system.

In field operations it was found that the progesterone concentration in milk changed as milking progressed, this was due to the lipid soluble progesterone being associated with the fat concentration that we showed increased from 1% in foremilk to over 6% in strippings. By fixing the time of sampling to a point at which 2 litres of milk has been drawn we have demonstrated a good correlation with ELISA tests on whole milk samples. Similarly in field tests we found variable electrochemical activity in milk that bound to ligands on the sensor partly independently of the concentration of progesterone we overcame this.

As a spin-off from the main project we have also found that with minor adaptation the system will measure N-acetyl- glutamase (Nagase) in concentrations found in the milk of cows with clinical mastitis.

We are currently negotiating exploitation of the IP with a number of parties. We are currently preparing a business case with a view to a joint venture to launch a commercial version of the system suitable for installation in any modern milking system.

We have now completed this automated ovulation prediction project for DEFRA but anticipate that the results will form a major part of the development of an integrated management system for dairy cows. The aim of the integrated management system will be to use sensors to develop a feedback control system enabling reduced methane and ammonia emissions from dairy cows at the same time enhancing efficiency of feed utilisation.

Publications


MOTTRAM, T., VELASCO-GARCIA, M., BERRY, P., RICHARDS, P., GHESQUIERE, J., MASSON, L.
(2002)Automatic on-line analysis of milk constituents(Urea, Ketones, Enzymes and Hormones) using biosensors. Comparative Clinical Pathology 11, 50-58. (032465)

PEMBERTON, R. M., HART, J. P., MOTTRAM, T. T.
(2001)An electrochemical immunosensor for milk progesterone using a continuous flow system. Biosensors and Bioelectronics 16 (9-12), 715-723. (032430)

PEMBERTON, R. M., HART, J. P., MOTTRAM, T. T.
(2001)An assay for the enzyme N-acetyl-b-D-glucosaminidase ( NAGase) based on electrochemical detection using screen-printed carbon electrodes (SPCEs). Analyst 126 (11), 1866-1871. (032133)

VELASCO-GARCIA, M. N., MOTTRAM, T. T.
(2001)Biosensors in the livestock industry: an automated ovulation prediction system for dairy cows. Trends in Biotechnology 19 (11), 433-434. (031975)


METZ, J. H. M., MALTZ, E., MOTTRAM, T. T. (1999)Monitoring health and welfare in practice. Cattle Practice 7 (1), 123. (031488)

DITCHAM, W. G. F., AL-OBAIDI, A. H. R., MCSTAY, D., MOTTRAM, T. T., BROWNLIE, J., THOMPSON, I.
(2001)An immunosensor with potential for the detection of viral antigens in body fluids based on surface second harmonic generation. Biosensors and Bioelectronics 16 (3), 221 - 224. (031486)

VELASCO-GARCIA, M. N., MOTTRAM, T. T. F.
(2002)Biosensors: technology and opportunities in livestock production. Landwards 57 (2), Information Engineering for Agriculture and Horticulture, SRI 15th May 2001, 16-20. (031446)

MOTTRAM, T. T., HART, J., PEMBERTON, R.
(2000)Biosensing techniques for detecting abnormal and contaminated milk. Robotic milking: proceedings of the International Symposium Lelystad, The Netherlands, 17-19 August 2000, edited by HOGEVEEN, H. & MEIJERING, A. Netherlands: Wageningen Pers, 108-113. (030793)

CSG 15 (1/00) 3

Project
title / Automatic Ovulation Prediction for Dairy Cows
/ MAFF
project code / LS2401
Scientific report (maximum 20 sides A4)
To tab in this section press the tab key and the Control key together
Press the DOWN arrow once to move to the next question.

CSG 15 (1/00) 3

Project
title / Automatic Ovulation Prediction for Dairy Cows
/ MAFF
project code / LS2401

The scientific report is based on a paper describing our final experiment and this includes the results of earlier work.

1. Introduction

At present the principal method of detecting oestrus in dairy cows is by observation of changed behaviour combined with regular recording of fertility events, which is a labour-intensive process. Monitoring progesterone levels in whole milk is a more effective method for fertility monitoring but is only used by a small percentage of farmers. Existing enzyme linked immunosorbent assay (ELISA) test kits require manual sampling, time and skills, which are not abundant on farm. Livestock farmers need a non-invasive automated system that will predict ovulation with high specificity. The device has to operate automatically and provide rapid information at the farm to daily identify cows ready for artificial insemination (Mottram & Frost, 1997).

The sequence of events associated with the ovarian cycle in dairy cows is well known (Peters & Ball, 1994), (Gordon, 1996). In the immediate period post partum, the concentration of progesterone is low but begins to rise after about 15 days. After the first ovulation, a pattern emerges with a period of 21 days. The ovulation cycle is characterised by a slow rise in progesterone concentration to approximately 20 ng/ml, when a plateau of concentration is reached. At 15-17 days post ovulation the concentration drops to below 5 ng/ml. A few hours after this drop, oestrus occurs and 24-48 hours later the cow is ovulating and in an optimum condition for insemination (Foulkes et al., 1982). If a cow becomes pregnant the progesterone concentration rises to a plateau and remains at that level for the duration of the oestrus cycle. Consequently, on-line monitoring of progesterone in milk could be used to predict ovulation, to detect pregnancy and to diagnose ovarian failure.

Several biosensor strategies have been developed by researchers to determine progesterone in milk. Koelsche et al. (1994) coated a piezoelectric crystal with antibodies specific for progesterone and measured the change in resonant frequency of the crystal as progesterone bound to the active surface. However a limitation to the piezoelectric transducer was the inconsistency of the crystal oscillation in liquid phase. In contrast, Claycomb and Delwiche (1998) developed an on-line bovine progesterone biosensor employing an enzyme immunoassay for molecular recognition connected to an optical transducer. The endogenous progesterone present in the milk competed with progesterone labelled with horseradish peroxidase (HRP) for binding to the active biosensing surface (a nitro-cellulose membrane). The substrate was a solution of tetramethylbenzidine and the transduction was based on the oxidation of the substrate, generating a blue product; the absorbance measured at 650 nm was inversely proportional to the concentration of progesterone in milk.

An alternative approach was reported by Pemberton et al. (1998). The device was based on a disposable screen-printed amperometric progesterone biosensor, operated in a competitive immunoassay. The biosensor comprised a monoclonal anti-progesterone antibody (mAb) immobilised on the working area of a screen-printed carbon electrode (SPCE). It relied upon a reduction in the binding of alkaline phosphatase-labelled progesterone in the presence of endogenous milk progesterone. The enzyme substrate was naphthyl phosphate and the 1-naphthol generated in the enzymatic reaction was electrochemically oxidised, producing a signal inversely proportional to the concentration of unlabelled progesterone in milk. This SPCE-based immunosensor for progesterone was incorporated into a thin-layer flow cell offering advantages such as on-line analysis and improved fluid handling with the possibility of future automation (Pemberton et al, 2001). Our team has chosen this electrochemical technology to further develop a prototype automatic ovulation prediction system for dairy cows. This approach also has the benefit that the biosensors can be mass fabricated by screen-printing technology at low cost, enabling the SPCEs to be single use devices.

This paper describes work carried out in the laboratory to develop and calibrate an automatic electrochemical biosensor system for progesterone in whole fresh milk. The authors also report the field tests conducted at a dairy farm over a 6-week period, which have demonstrated the ability of this prototype biosensor to track the ovulation cycles of dairy cows by on-line measurement of progesterone during milking. The biosensor data were compared and validated with standard oestrus detection methods (ELISA test kits).

2. Experimental

2.1. Chemicals and reagents

The purified IgG fraction of polyclonal rabbit antibody against sheep IgG was purchased from Sigma-Aldrich. Ridgeway Science Ltd. (Gloucestershire, UK) kindly provided sheep monoclonal antibody (mAb) against progesterone and solution containing progesterone labelled with alkaline phosphatase (AP-prog). Unlabelled progesterone, a-naphthyl-phosphate, a-naphthol and p-nitrophenylphosphate were purchased from Sigma-Aldrich. Precoated microtitre plates for the progesterone ELISA were kindly provided by Ridgeway Science Ltd. Plain microtitre plates were purchased from Becton Dickinson Labware, New Jersey.

All solutions were prepared in Milli-Q water. Buffer solutions: carbonate coating buffer, pH 9.6 containing 0.015 M Na2CO3 and 0.035 M NaHCO3; 100 mM diethanolamine-HCl buffer pH 7.0 and 100 mM diethanolamine-HCl buffer pH 9.8 containing 10mM MgCl2 as a cofactor for alkaline phosphate (Merck, Poole, Dorset). Wash solution: 0.05 % Tween 20 in diethanolamine buffer pH 7.0 (Merck, Poole, Dorset).

Milk samples from eight different dairy cows at various stages of their oestrus cycles were obtained from Boltons Park Farm (Hatfield, Hertfordshire, UK). One Lactab MkIII tablet (Thompson & Capper Ltd., Runcorn, Cheshire, UK) containing 30 mg potassium dichromate was added to every 25 ml milk sample as a preservative.

2.2. Fabrication of biosensors

Screen-printed carbon electrodes (SPCEs) were prepared at Gwent Electronic Materials Ltd. (Mamhilad, Gwent, UK). They comprised a carbon working electrode (code D14 ink) and a dielectric layer, both printed onto polyester card. Before antibody coating the circular working area (9 mm2), each individual sensor was cut from the card with a guillotine into a 20x30-mm rectangle to match the internal size of the sensor chamber. The fabrication of biosensors required the deposition onto the active area of a first layer of rabbit anti-sheep IgG (to capture and orientate the sheep anti-progesterone antibodies) and a second layer of monoclonal antibody against progesterone, as described previously (Hart et al, 1997). Each week a batch of 100 biosensors was made up, from which 20 were retained for the batch calibration with standard progesterone solutions prepared in cow’s milk. The biosensors were stored at 4oC in a moist atmosphere before use.

2.3. Colorimetric test

As quality control checks on the storage, stability and performance of biosensors from different fabrication batches, colorimetric tests, based on the same principle as the ELISA, were run. The biosensors were incubated for 20 minutes with a mixture of 0 ng/ml progesterone standard and labelled progesterone (3:5). After washing to remove surplus material, the working area was cut from each SPCE and placed in a well of a plain 96-well microtitre plate. Each well then received 150 ml of p-nitrophenylphosphate, prior to 60 min incubation at RT. Following removal of the biosensors, measurement of absorbance at 405 nm with a microplate reader (Molecular Devices Limited, West Sussex, UK) indicated the efficiency of the biosensor at capturing enzyme-labelled progesterone. If the active biosensor surface (mAb-SPCEs) deteriorated, then the absorbance decreased.

2.4. Continuous-flow biosensing system

The disposable biosensor was placed in the measuring cell, which consisted of a Perpex block with flow paths machined through it; a reference and a counter electrode were also positioned in close proximity within the cell. The reference electrode was a BAS RE-6 type (Ag/AgCl). The working electrode was a screen-printed carbon biosensing electrode and a stainless steel exit tube served as a counter electrode. The measuring cell was connected to a fluid handling system, which employed miniature valves and a pump. In addition, a mini 19-inch enclosure was attached to the cell, and this housed the entire sensor electronics, power supplies and embedded PC (see Figure 1). A computer program controlled all the sequences of events: the pumping of milk samples and reagents into the flow-cell, the incubation times, the application of a constant voltage between the working and the counter electrodes and the measurement of the resulting oxidation current. The data were transferred directly into a spreadsheet.