GENERAL AND COMPARATIVE

ENDOCRINOLOGY

General and Comparative Endocrinology 132 (2003) 315–320

www.elsevier.com/locate/ygcen

Communication in Genomics and Proteomics

Clarification of emu serum for peptide hormone assayusing polyethylene glycol precipitation

J.K. Van Cleeff,* M.A. Blackberry, D. Blache, and G.B. Martin

School of Animal Biology, Faculty of Natural and Agricultural Sciences, The University of Western Australia, 35 Stirling Highway,

Crawley 6009, WA, Australia

Accepted 24 January 2003

Abstract

Interference in radioimmunoassays (RIAs) was frequently encountered during endocrinological studies of the emu (Dromaius

novaehollandiae). Interference was greatest when serum was cloudy or opaque. Such samples appeared seasonally, in spring and

summer during the phase of fat deposition, and in the winter when females were laying. These poor quality samples did not allow

accurate measurement by RIAs of several peptide hormones for a full year. To prepare them for assay, these sera were clarified using

a polyethylene glycol (PEG) solution at a final concentration of 7.5%. This treatment was effective in most cases. After treatment

with the PEG, recoveries of LH, glucagon, and prolactin were greater than 75% and that for insulin was 40%. Regardless of the level

of recovery, there was a high correlation of assay results between non-opaque native and PEG-treated sera. Serum samples con-

taining large amounts of interfering high molecular weight components, such as lipoproteins, can be clarified with PEG, enabling

their accurate measurement by RIA.

� 2003 Elsevier Science (USA). All rights reserved.

Keywords: Radioimmunoassay; Sample preparation; Polyethylene glycol; emu

1. Introduction

Seasonal changes in reproductive activity, feed intake

and fat deposition in the emu (Dromaius novaehollan-

diae) are a major theme of research in our laboratory.Our primary tool to investigate those changes is

measurement of metabolic (insulin, glucagon) and re-

productive hormones (LH, prolactin) using radioim-

munoassay (RIA). However, the quality of serum

collected from emus is very variable and some samples

can be cloudy or opaque in appearance. In those sera,

RIA of peptide hormones gives concentrations that are

10–100-fold higher than the range expected. Theseaberrant readings are mainly due to poor precipitation

of the bound hormone. It is likely that high concen-

trations of large blood proteins, such as lipoproteins,

* Corresponding author. Animal Sciences Laboratory, University

of Illinois, 1207 West Gregory Drive, MC630, Urbana, IL 61801,

USA. Fax: 1-217-333-8286.

E-mail address: jkvc@uiuc.edu (J.K. Van Cleeff).

0016-6480/03/$ - see front matter � 2003 Elsevier Science (USA). All rights

doi:10.1016/S0016-6480(03)00088-1

are responsible for this interference because the cloudy

samples seem to be most common during spring and

summer, when emus are in their fat deposition phase,

and during egg laying. These large molecular weight

molecules can be precipitated using polyethylene glycol(PEG; Atha and Ingham, 1981), an approach that has

been used successfully to prepare canine serum for as-

says that use spectrophotometric analysis (Thompson

and Kunze, 1984). PEG precipitation is also compatible

with RIA of serum samples because PEG is water sol-

uble and does not denature proteins or chemically in-

teract with them, even at high concentrations (Ingham,

1978, 1984).In the present study, we show that treatment with

PEG greatly improves the quality of emu serum for RIA

without preventing the measurement of insulin, gluca-

gon, prolactin, and LH. We also show that, when sera

are classified according to the appearance of the precip-

itate following treatment with PEG, serum quality varied

with sex and with season in association with the stage of

the metabolic and reproductive cycles of the emu.

reserved.

316 J.K. Van Cleeff et al. / General and Comparative Endocrinology 132 (2003) 315–320

2. Materials and methods

2.1. Samples

Samples (n ¼ 2021 obtained from 51 emus) were

collected via jugular venipuncture from adult male and

female emus between 13 May 1998 and 13 January 2001.

Each bird was represented multiple times and through

several seasons. They were housed outdoors either singlyor in breeding pairs in pens located at the Field Station

of the University of Western Australia, near Perth

(31�56.90S, 115�47.60E). Animals were provided ad libi-tum with water and standard pelleted commercial feed.

All experimental protocols conformed to the Code of

Practice formulated by the National Health & Medical

Research Council of Australia and implemented by the

Fig. 1. Examples of emu sera illustrating the effects of PEG treatment. A sa

sample after treatment with PEG is shown on the right, with its precipitat

compact precipitate, floating fatty layer (not shown); (C) floating fatty layer

phous floating fat layer, poor separation from supernatant.

Animal Ethics Committee of The University of WesternAustralia.

Blood was allowed to clot at room temperature, re-

frigerated overnight, then centrifuged at 1800g for 25

min. Serum was divided into two aliquots and either

frozen immediately or treated with PEG by adding

200 ll of 45% PEG-6000 (Merck, Melbourne, Australia)

in water per milliliter of serum, giving a final concen-

tration of 7.5%. PEG-treated samples were incubated atroom temperature for 15min, then centrifuged (1800g at20�C for 10min). Precipitated sample was assigned a

quality type based on an arbitrary scale, and clarified

serum was aspirated and stored frozen until assay. The

quality types were based on readily distinguishable dif-

ferences among types of precipitates formed after PEG

treatment (Fig. 1).

mple of native serum is shown on the left of each panel, and the same

e type. Precipitate types are as follows: (A) compact precipitate; (B)

only; (D) flocculent precipitate; and (F) flocculent precipitate, amor-

J.K. Van Cleeff et al. / General and Comparative Endocrinology 132 (2003) 315–320 317

2.2. Hormone recovery and RIA quality control

We used three tests for quality control. First, the ef-

fect of PEG treatment on hormone recovery was de-

termined by incubating samples ðn ¼ 5Þ of native serumof each of the different types with known amounts of

iodinated hormone. Activities in aspirated clarified sera

were then compared to the initial dose. Second, a known

amount of unlabelled hormone was added to a set ofserum samples of each type, and these samples were also

treated with PEG and assayed. Third, in the same assay,

concentrations of hormone were compared to those

measured in both non-treated samples and samples

treated without addition of hormones. Only clear sam-

ples were used in the last test because cloudy samples

would not give a reliable estimate of hormone concen-

trations in the absence of PEG treatment.Assays were standard double-antibody RIAs. LH

was analysed in 50 ll replicates using the chicken LHassay adapted from procedures of Follett and Sharp

(Follett et al., 1971; Sharp et al., 1987). Prolactin was

analysed in 50 ll replicates using an assay based on re-combinant chicken prolactin (Talbot and Sharp, 1994).

Hormones and antibodies (both raised in rabbits) were

supplied by Dr. Peter Sharp (Roslin Institute, Edin-burgh, Scotland). Insulin was analysed in 100 ll repli-cates according to the procedures of McMurtry et al.

(1983), with highly purified pancreatic chicken insulin

and guinea pig anti-insulin supplied by Dr. John

McMurtry (Livestock and Poultry Research Labora-

tory, USDA, Beltsville, MD). Glucagon was analysed in

50 ll replicates using commercial kit reagents (guineapig anti-glucagon antibody, iodinated tracer and humanglucagon standards) from Linco Research (MO, USA).

All of these assays have been validated and utilized for a

variety of avian species, including the emu (LH, Blache

et al., 2001; prolactin, Malecki et al., 1998; insulin and

glucagon, Van Cleeff et al., 1997, 1998). For each hor-

mone, parallelism was estimated by assaying a serially

diluted, pooled sample of emu serum.

Fig. 2. Categorization and frequency of each category of precipitate

type in emu sera collected between 13 May 1998 and 13 January 2001.

For abbreviations see Fig. 1.

3. Results

3.1. Sample quality

The type of PEG-induced precipitate depended on the

sex and reproductive status of the bird (Fig. 2; v2,P < 0:0001) and differed according to the time of year(month) for each reproductive status (Fig. 3; v2,P < 0:0001). Compact precipitate (Type A) and floatinglayer (Type C) appeared in the sera of all categories of

emus at some time of the year, but Type F, which did

not precipitate fully with 7.5% PEG, was most common

in females that did not lay eggs (Figs. 2, 3). Flocculent

precipitate (Type D) was rare in male sera but common

in females during the breeding season, even in female

sera that did not lay eggs (Fig. 2). Overall, the floating

precipitate resulting from PEG treatment (Type C)

ranged in thickness from 1 to 11mm/ml of serum in a

standard 12� 75 mm test tube. Generally, the thick-

ness of the layer increased as spring progressed, and

Fig. 3. Distribution of serum quality types (%) over the year for various groups of emus. Both ‘‘Incubating males’’ and ‘‘Nonincubating males’’ were

paired with ‘‘Laying females.’’ Incubation occurred in June, July and August for Incubating males. Eggs were laid between mid-April and mid-

August. ‘‘Unpaired males’’ were isolated from females. For abbreviations see Fig. 1.

318 J.K. Van Cleeff et al. / General and Comparative Endocrinology 132 (2003) 315–320

narrowed towards autumn. Different types of serum

precipitate were observed in the same emu at different

times of the year but, at a given time, only one type of

precipitate was observed.

3.2. Hormone recovery and RIA quality control

For all hormones except insulin, recoveries after PEG

extraction were greater than 75%, with both the labelled

Table 1

Hormone recoveries and comparison of native and PEG-treated emu sera su

Hormone Recovery (%) n

LH 75.1�3.4 19

Prolactin 87.0�3.3 36

Insulin 40.0�2.2 67

Glucagon 89.2�1.8 4

Kendall�s rank correlation was applied to RIA data.

and unlabelled hormone tests (Table 1). Variability in

recovery between samples of different types was mini-

mal, judging from the small standard deviations.

Treatment with PEG did not increase the non-specificbinding in the assays (less than 4% in all cases) because

the final concentration of PEG in the assay tube was

only 0.65% (glucagon) to 1.5% (LH) after addition of

the second antibody. The readings for the serially di-

luted samples treated with PEG plotted parallel to the

bjected to assays for four protein hormones

Tau P

0.807 <0.0001

0.654 <0.0001

0.411 <0.0001

1 <0.0005

Fig. 4. Standard curves (circles) for radioimmunoassays for (A) chicken LH, (B) chicken prolactin, (C) human glucagon and (D) chicken insulin

showing parallelism with serial dilutions of emu plasma samples (squares). Note: for the insulin test, only a limited range of dilutions was feasible

because emus have low absolute concentrations of the hormone.

J.K. Van Cleeff et al. / General and Comparative Endocrinology 132 (2003) 315–320 319

standard curves (Fig. 4). With samples that were clear

and thus did not require PEG treatment, the correlationbetween measurements obtained from untreated and

PEG-treated samples was high (Table 1).

4. Discussion

The quality of serum collected from emus varied ac-

cording to the physiological status of the bird, buttreatment with PEG was able to clarify most of the

samples and allow the measurement of peptide hor-

mones using RIA.

Treatment with PEG reduced peptide hormone con-

centrations in serum. This reduction was more marked

for insulin than for LH, prolactin or glucagon. Serum

concentrations of insulin might have been affected be-

cause insulin can bind to binding proteins (Zhvirblene

et al., 1988) or can form aggregates (Jeffrey and Treacy,

1992), both such molecular associations being large en-ough to be precipitated by the PEG treatment.

The decrease in hormone concentration after PEG

treatment did not affect the discriminative power of the

four RIAs tested, as shown by the high coefficient of

correlation between the measurements of peptide hor-

mones in native and PEG-treated serum. However, PEG

treatment could greatly interfere with the measurement

of hormones that are bound in large proportion tobinding proteins, such as thyroid hormones (Bartalena,

1990) or with the assay of lipophilic hormones such as

steroids. In both cases, PEG precipitation would be

expected to decrease the serum concentrations to below

the limit of detection of the assay. It is not possible to

generalise from the four assays tested in this paper, so

each assay system should be validated for PEG sample

preparation.

320 J.K. Van Cleeff et al. / General and Comparative Endocrinology 132 (2003) 315–320

The nature of the interfering compounds was not in-vestigated in our study but it seems likely that the opacity

was due to high concentrations of lipoprotein or other

very large proteins. During spring and summer, emus are

in their fat deposition phase and serum lipoproteins,

produced in large quantity by the liver (Blem, 1976;

Stevens, 1996), are abundant in the blood stream as nu-

trients are transported to the fat depot. Moreover, serum

collected frommales was usually clear during the autumnand winter breeding season when food intake is at its

lowest (O�Malley, 1996). Apparently, mobilisation of fatreserves, involving release of fatty acids and triacylgly-

cerols from the fat depot, has less effect on serum quality

than does fat deposition. After treatment with PEG, se-

rum collected from females during the breeding season

yielded a large pellet that did not float. It is likely that this

precipitate mainly contains the yolk precursor vitelloge-nin, a large glycophospholipoprotein with a molecular

weight of 480,000 in domestic hens (Griffin et al., 1984)

that is likely to be precipitated by PEG. In particular,

Type F serum, mainly observed in females that did not

lay eggs, could be due to an accumulation of yolk pre-

cursors in the circulation that are produced as in laying

females but not used because no egg yolk is formed.

Accurate measurements by RIA of peptide hormonesin emu serum were made possible by sample preparation

using PEG precipitation, allowing us to analyse samples

from all times of year, regardless of the reproductive or

metabolic state of the birds. This simple procedure

eliminates interference caused by large quantities of li-

poprotein in the RIA of glucagon, insulin, prolactin and

LH. While every hormone assay must be validated be-

fore PEG precipitation could be used, this method willprovide a quick, simple and inexpensive method for

clarification of lipemic serum in other species.

Acknowledgments

We thank Mr. Peter Cowl for caring for the birds and

for his help with blood sampling. This work was funded

by the Australian Research Council (D.B., M.A.B.,

G.B.M.), and an Overseas Postgraduate ResearchScholarship from The University of Western Australia

(J.K.V.C.).

References

Atha, D.H., Ingham, K.C., 1981. Mechanism of precipitation of

proteins by polyethylene glycols. Analysis in terms of excluded

volume. Journal of Biological Chemistry 256, 12108–12117.

Bartalena, L., 1990. Recent achievements in studies on thyroid

hormone-binding proteins. Endocrine Reviews 11, 47–64.

Blache, D., Talbot, R.T., Blackberry, M.A., Williams, K.M., Martin,

G.B., Sharp, P.J., 2001. Photoperiodic control of the secretion of

luteinizing hormone, prolactin and testosterone in the male emu

(Dromaius novaehollandiae), a bird that breeds on short days.

Journal of Neuroendocrinology 13, 998–1006.

Blem, C.R., 1976. Patterns of lipid storage and utilization in birds.

American Zoologist 16, 671–684.

Follett, B.K., Scanes, C.G., Cunningham, F.J., 1971. A radioimmu-

noassay for avian luteinizing hormone. Journal of Endocrinology

51, v–vi.

Griffin, H.D, Perry, M.M., Gilbert, A.D., 1984. Yolk Formation. In:

Freeman, B.M. (Ed.), Physiology and Biochemistry of the Domes-

tic Fowl, vol. 5. Academic Press, pp. 345–380.

Ingham, K.C., 1978. Precipitation of proteins with polyethylene glycol:

characterization of albumin. Archives of Biochemistry and Bio-

physics 186, 106–113.

Ingham, K.C., 1984. Protein precipitation with polyethylene glycol.

Methods in Enzymology 104, 351–356.

Jeffrey, P.D., Treacy, G.B., 1992. Chicken insulin (native form) self-

associates like mammalian self-association of kangaroo and

chicken insulins, chicken despentapeptide insulin and deamidated

insulin. Biochimica et Biophysica Acta 1119 (1), 107.

Malecki, I.A., Martin, G.B., O�Malley, P.J., Meyer, G.T., Talbot,R.T., Sharp, P.J., 1998. Endocrine and testicular changes in a

short-day seasonally breeding bird, the emu (Dromaius novaehol-

landiae), in southwestern Australia. Animal Reproduction Science

53, 143–155.

McMurtry, J.P., Rosebrough, R.W., Steele, N.C., 1983. A homolo-

gous radioimmunoassay for chicken insulin. Poultry Science 62,

697–701.

O�Malley, P., 1996. An estimate of the nutritional requirements of

emus, In: Improving our understanding of ratites in a farming

environment, Manchester, UK, pp. 92–108.

Sharp, P.J., Dunn, I.C., Talbot, R.T., 1987. Sex differences in the LH

responses to chicken LHRH-I and -II in the domestic fowl. Journal

of Endocrinology 115, 323–331.

Stevens, L., 1996. Avian Biochemistry and Molecular Biology.

Cambridge University Press, Cambridge.

Talbot, R.T., Sharp, P.J., 1994. A radioimmunoassay for recombinant-

derived chicken prolactin suitable for the measurement of prolactin

in other avian species. General and Comparative Endocrinology

96, 361–369.

Thompson, M.B., Kunze, D.J., 1984. Polyethylene glycol-6000 as a

clearing agent for lipemic serum samples from dogs and the effects

on 13 serum assays. American Journal of Veterinary Research 45,

2154–2157.

Van Cleeff, J., Williams, K.M., Blache, D., Malecki, I.A., Martin,

G.B., 1997. Relationships between incubation behaviour, body

weight, photoperiod, appetite and insulin and glucagon secretion in

male emus (Dromaius novaehollandiae). Proceedings, 28th Annual

Conference, Australian Society for Reproductive Biology, Canb-

erra, p. 47.

Van Cleeff, J., Blache, D., Martin, G.B., 1998. Effect of daylength on

insulin, glucagon, and corticosterone in emus (Dromaius novaehol-

landiae), 41st Annual Scientific Meeting, Endocrine Society of

Australia, Perth, Australia, pp. 82.

Zhvirblene, A.A., Barbashov, S.F., Severin, E.S., 1988. Isolation and

study of the properties of an insulin-binding protein from the blood

serum. Problemy Endokrinologii 34, 64–68.