Eb

La

b

a

ARAA

KA�LRMP

1

taccfsStc

rfcaa

p

S

0h

Peptides 51 (2014) 54– 58

Contents lists available at ScienceDirect

Peptides

jo ur nal home p age: www. elsev ier .com/ locate /pept ides

nhanced levels of immunoreactive �-casomorphin-8 in milk ofreastfeeding women with mastitis

ennart Righarda, Anna Carlsson-Jonssonb, Fred Nybergb,∗

Department of Paediatrics, University Hospital, Malmö, University of Lund, Lund, SwedenDepartment of Pharmaceutical Biosciences, Uppsala University, Uppsala, Sweden

r t i c l e i n f o

rticle history:eceived 25 September 2013ccepted 25 October 2013vailable online 1 November 2013

eywords:typical opioids-Casomorphins

a b s t r a c t

An incorrect, superficial suckling technique in breastfeeding frequently leads to milk congestion andsometimes mastitis. In the present study we have examined whether milk congestion may affect levelsof the atypical opioid �-casomorphin-8 in milk and in plasma. We also investigated whether the rate ofacute psychosis during the first half year after parturition has declined in Sweden over the years. Milk andplasma samples were collected for peptide analysis from 14 women with mastitis and 10 controls. Wefound that in a group of 14 late cases of mastitis (median 48 days post partum) the detected mean levelof �-casomorphin-8 in milk was significantly higher and somewhat higher in plasma at the acute stage

actationadioimmunoassayastitis

ostpartum psychosis

compared with 2–3 weeks later, after recovery when the symptoms had disappeared, as well as comparedto the control subjects. Swedish official statistics show that the incidence of acute psychosis in the firstmonth and in the first half year after birth has declined by a half during the last 30 years. A relationshipbetween postpartum psychosis and elevated �-casomorphin-8 levels in CSF has been suggested fromearlier studies. In this study, milk congestion led to enhanced levels of �-casomorphin-8 in milk, whichmay be related to postpartum psychosis and probably also to ‘the postnatal blues’.

© 2013 Elsevier Inc. All rights reserved.

. Introduction

Soon after the discovery of the endogenous opioid peptides,hese compounds received interest for a possible role in psychi-tric disorders. Biochemical and pharmacological studies on theerebrospinal fluid (CSF) levels of opioid peptides led to the dis-overy of atypical opioids, the so-called �-casomorphins, in CSFrom women with postpartum psychosis but also from healthyubjects at term pregnancy and the early puerperium [11,12,14].everal studies have shown increased levels of opioid peptides inhe CSF of postpartum psychosis and the peptides could finally beharacterized as related to �-casomorphin-8 [13].

The presence of opioid active peptides in milk was originallyeported by Brantl and co-workers [1]. They isolated a peptideragment from an enzymatic digest of the bovine milk protein �-

asein. This fragment was found to interact with opioid receptorsnd was shown to produce opioid effects in the guinea-pig ileumssay (GPI-assay). This finding together with its origin made it rea-

Abbreviations: CSF, cerebrospinal fluid; RIA, radioimmunoassay; HPLC, highressure liquid chromatography.∗ Corresponding author at: Department of Pharmaceutical Biosciences, Box 591,

-75124 Uppsala, Sweden.E-mail address: fred.nyberg@farmbio.uu.se (F. Nyberg).

196-9781/$ – see front matter © 2013 Elsevier Inc. All rights reserved.ttp://dx.doi.org/10.1016/j.peptides.2013.10.027

sonable to name the peptide �-casomorphin. Sequence analysis of�-casomorphin revealed that its structure diverges from that of theclassical endogenous opioids; the milk-derived peptide contains aproline residue next to its N-terminal tyrosine (Table 1). In studiesthat followed, �-casomorphin was found to exist in bovine plasmaboth in truncated and extended forms of the original peptide [22]but also in human body fluids. Indeed, human �-casomorphin-8-like immunoreactivity was measured in human plasma [7,13],in human CSF [13] but also in human milk [16]. In plasma fromnewborn calves, immunoreactive �-casomorphin was detected fol-lowing their first milk intake after birth [22].

Regarding their functional relevance the �-casomorphins havebeen suggested to be involved in the interplay between motherand child during lactation. Studies of milk-derived opioid pep-tides have been directed to both mothers and infants. A previousreport described that at full term of the pregnancy, the levels of�-casomorphin-8 was increased both in plasma and CSF comparedto non-pregnant women [13]. Moreover, studies also revealed apositive correlation between plasma concentration and CSF levelsof the milk-derived peptide. This observation was considered toindicate that �-casomorphin might cross the breast parenchyma-

blood barrier to enter the circulation and subsequently penetratethe blood–brain barrier to reach the CNS. From this observation itwas speculated that the peptide may have a function in maternalbehavior [13]. Studies carried out by others have suggested that the

L. Righard et al. / Peptides 51 (2014) 54– 58 55

Table 1Amino acid sequences of �-casomorphins, enkephalins and endomorphins.

Peptide Amino acid sequence

�-Casomorphin-5 (bovine) H-Tyr-Pro-Phe-Pro-Gly-OH�-Casomorphin-8 (bovine) H-Tyr-Pro-Phe-Pro-Gly-Pro-Ile-Pro-OH�-Casomorphin-5 (human) H-Tyr-Pro-Phe-Val-Glu-OH�-Casomorphin-8 (human) H-Tyr-Pro-Phe-Val-Glu-Pro-Ile-Pro-OHMet-Enkephalin H-Tyr-Gly-Gly-Phe-Met-OHLeu-Enkephalin H-Tyr-Gly-Gly-Phe-Leu-OH

mauiiosIam

icppifAppoloja[

bTwTdowowofi

2

2

wwwaamhfi

Fig. 1. Levels of �-casein in milk samples from three healthy lactating women (a, c,

Endomorphin 1 H-Tyr-Pro-Trp-Phe-NH2Endomorphin 2 H-Tyr-Pro-Phe-Phe-NH2

ilk-derived opioid may also reach the brain of the infant. Applyingn extraction procedure combined with high performance liq-id chromatography (HPLC) and radioimmunoassay rostrocaudally

ncreasing levels of immunoreactive �-casomorphin-8 in humannfants were detected [21]. This increase was obvious in brain areasf relevance for various behaviors suggested for the actual peptide,uch as mesencephalon, pons cerebri, and medulla oblongata [15].ndeed, it has been discussed that the �-casomorphins also mayffect the child during lactation and it has been assumed that theilk-opioids may induce sedation [12].Several studies have implicated �-casomorphin-like peptides

n postpartum psychosis [8–10]. Although CSF levels of �-asomorphin-8 immunoreactivity at late pregnancy and in theuerperium seem to be enhanced at late pregnancy and in theuerperium compared to non-pregnant women [13], a further

ncrease of �-casomorphin-like material has been detected in CSFrom puerperal women diagnosed for postpartum psychosis [9].lso, an analyze of milk samples from women with postpartumsychosis revealed an evident reduction of the �-casein levels com-ared to control subjects, possibly due to an increased conversionf �-casein to �-casomorphins (Fig. 1) [8]. Furthermore, in milk col-ected from women with postpartum psychosis the concentrationf �-casomorphin-8 was also increased compared to control sub-ects [19]. In fact, the presence of high levels of a peptide, identifieds �-casomorphin-8, has been confirmed in postpartum psychosis16].

Puerperal complications, like postpartum psychosis, have alsoeen suggested to have a link to the technique for breastfeeding.hus, incorrect suckling may cause milk congestion and mastitisith an inflamed aching breast, fever and often a restless sleep.

he breast with mastitis is insufficiently emptied and the milk thatoes not come out the natural way must be taken care of and metab-lized by the woman’s own body. The aim of the present studyas to examine whether milk congestion gives rise to higher levels

f �-casomorphin-8 in milk and plasma than in lactating womenithout problems. We also raised the question of whether the rate

f psychosis that develops during the first month up to six monthsollowing birth has declined over time, possibly due to the babies’ncreased ability to suckle more effectively.

. Methods

.1. Study subjects

Throughout 1990, we followed up all cases of mastitis thatere referred to the Malmö University Hospital Clinic for womenith breastfeeding problems. Mastitis is caused by milk congestion,hich causes milk to leak under pressure through the distended

lveoli walls into the surrounding tissues and circulation. There is lump, tenderness, redness and fever, which in the first stages of

astitis will disappear as soon as the milk can be released, with the

elp of the baby’s correct suckling or a breast pump. The criterionor mastitis in this study was that 20% or more of one breast wasnflamed and aching. During the stay on the maternity ward there

e) and three women with postpartum psychosis (b, d, f). The second peak from theright (marked with an arrow) corresponds to the fraction of �-casein. Samples wereanalyzed on SDS-polyacrylamide gel electrophoresis, data are adapted from [8].

were some cases of milk congestion as the milk “came in” but nomastitis with a lump.

We suspected that it would be difficult to find women with mas-titis early after birth, as all mothers were informed about correctsuckling and consequently all babies were suckling correctly at dis-charge by the time of the study. Our suspicions were confirmed, as

5 Peptides 51 (2014) 54– 58

waowtehwsdm(vsf

cb

2

otpoc

2

waAc0Tw5Swt(drSw0R

2

atar0IbgwAmbd

Fig. 2. Levels of �-casomorphin-8-like immunoreactivity in milk from lactating

6 L. Righard et al. /

e did not find any women with mastitis during the first monthfter birth. However, after one year of observation we had 14 casesf late onset mastitis, appearing between 1 and 6 months after birthith a median of 48 days. Nine women came to our clinic within

he first day after the first symptoms had appeared, while five hadxperienced symptoms for a few days. All cases were cured withinours or a few days. We took samples of milk and plasma from eachoman with mastitis in the acute stage and in 10 cases we took a

econd sample two to three weeks later when the symptoms hadisappeared. We also collected blood and milk samples about oneonth after birth from 10 women with normal lactation as controls

mean 34 days). Altogether we had 24 healthy women with normalaginal deliveries and healthy term newborn babies enrolled in thetudy. The median age of women with mastitis was 29.6 years andor the controls 27.3 years.

The study was approved by the Ethical Committee of the Medi-al Faculty of Lund University. Each subject gave informed consentefore being studied.

.2. Peptides and chemicals

Synthetic human �-casomorphin-8 (h�-casomorphin-8) wasbtained from Novabiochem (Laufelfingen, Switzerland). The pep-ide was labeled with 125I by the chloramine-T procedure andurified by reverse phase HPLC, as described previously [4]. Allther chemicals and solvents were of analytical reagent grade fromommercial sources.

.3. Sample preparation

Prior to radioimmunoassay (RIA), 1.0 ml of each plasma sampleas acidified (100 �l of 1 M HCl/ml) and separated on dispos-

ble reverse phase silica gel cartridges (Sep-Pak C-18, Watersssociates, Milford, MA). The cartridge was preconditioned by suc-essive washing with 5 ml methanol, 5 ml methanol containing.4% trifluoroacetic acid (TFA), and 10 ml distilled water with 0.04%FA. The acidified sample (1.1 ml) was applied, followed by 10 mlater with 0.04% TFA. The cartridge was subsequently diluted with

ml methanol/0.04% TFA and the diluent was evaporated in apeed-Vac concentrator (Savant, Hicksville, NY). The dried residuesere reconstituted in 200 �l buffer B (see below) before RIA. For

he milk analyses, the samples were de-fatted by centrifugationtwice, 20 min each; 15,000×g; 4 ◦C) before Sep-Pak separation asescribed above. The initial Sep-Pak diluents were evaporated andeconstituted in 1 ml buffer B and subjected to a second run on theep-Pak cartridges. The material diluted with 5 ml methanol-TFAas collected and evaporated. The residues were reconstituted in

.5 ml buffer B and further diluted (1:100) in the same buffer beforeIA.

.4. Radioimmunoassay

The antiserum against h�-casomorphin-8 was raised in rabbitsccording to the procedure described previously [6,7,13], andhe iodinated peptide was used as tracer. Cross-reaction of thentiserum with h�-casomorphin-7 and -9 was 6.3% and 16%,espectively, and with shorter h�-casomorphin fragments less than.01%. With �-endorphin, Leu-enkephalin, ACTH, and angiotensinI it was also less than 0.01% [6]. The buffers used in the assay were:uffer A, 0.05 mol/l sodium phosphate, pH 7.4, containing 1 mg/mlelatin, 8.2 mg/ml NaCl, and 9.3 mg/ml EDTA; buffer B, buffer Aith 1 mg/ml BSA and 1 mg/ml Triton X-100; and buffer C, buffer

with 1 mg/ml BSA. Triplicate samples (50 �l) in buffer B wereixed with 100 �l antiserum (diluted 1:100,000 in buffer C), 200 �l

uffer A, and 50 �l labeled h�-casomorphin-8 (4000–5000 cpm)issolved in buffer B. The mixtures were incubated overnight at

women. Each bar represents the average ± S.E.M from 14 individuals (mastitisgroup) or 10 individuals (control group). ***P < 0.001, **P < 0.01 versus acute mastitisgroup.

4 ◦C. Separation of antibody-bound and free peptide was carried outby charcoal adsorption. The samples were incubated for 15 min onan ice bath with 300 �l dextran-coated charcoal (20 mg/l charcoaland 2 mg/l Dextran T-70 in buffer A). After centrifugation for 1 minin a Beckman Microfuge B (Beckman, Palo Alto, CA), 500 �l super-natant was counted in a Packard gamma counter (Packard, DownersGrove, IL). The detection limit of the RIA was about 5 fmol/tube, and50% inhibition of tracer binding occurred at about 25 fmol/tube.The mean interassay coefficient of variation of 10 assays was 6%.The identity of the measured immunoreactivity was confirmed byHPLC as reported previously [13,16].

2.5. Statistics

Statistics were calculated using GraphPad Prism version 5.0 forMac OS X, GraphPad Software, San Diego, CA, USA. A Shapiro–Wilksnormality test was performed in order to determine whether thesamples were normally distributed. Since not all groups displayeda normal distribution, non-parametric t tests (Mann–Whitney testand Wilcoxon signed rank test) were performed between relevantgroups. P < 0.05 was considered significant. P-values have not beenadjusted for multiplicity.

3. Results

We analyzed purified milk and plasma samples from lactat-ing women with mastitis in the acute phase and two to threeweeks after mastitis as well as samples from the control group.Results from the radioimmunoassay showed that the levels of h�-casomorphin-8 in milk were in the pico-molar range among allgroups studied. The h�-casomorphin-8 levels were significantlyhigher in the subjects suffering from mastitis compared to the con-trol group (P = 0.0002). The symptoms from mastitis disappearedafter approximately 24 h. About 2–3 weeks after hospital treat-ment, when the patients were recovered from the mastitis, theh�-casomorphin-8 levels in milk were significantly reduced com-pared to the acute state (P = 0.0092) (Fig. 2). In plasma, the levels ofh�-casomorphin-8 were in the femto-molar range, thus markedlylower than in the milk samples, as expected. However, althoughthe h�-casomorphin-8 levels in plasma samples from the lactatingwomen with mastitis tended to be higher than in the control group,the difference was not statistically significant. In similarity to thelevels in milk samples, �-casomorphin-8 levels returned to controllevels 2–3 weeks after the treatment (Fig. 3).

To investigate the incidence of postpartum psychosis over time,

we have investigated data collected from the Swedish official statis-tics provided by the National Board of Health and Welfare. Aspostpartum psychosis has not been classified as a distinct dis-ease entity, we had to follow the incidence of all psychosis during

L. Righard et al. / Peptid

Ft

stUr4tyaptai

4

seofgaibecbSo

ftmdemttiptd

sil[it

ig. 3. Levels of �-casomorphin-8-like immunoreactivity in blood plasma from lac-ating women. Each bar represents the average ± S.E.M from 10 individuals.

ix months post partum from 1973 to 2008. During this period,he classification system was changed twice, in 1986 and 1996.ntil 1986 chronic psychosis was separated from the others and

epresented 58% of all psychosis until 30 days post partum and4% within six months post partum. As postpartum psychosis cer-ainly is not chronic, this percentage was withdrawn. During theears 1973–1982, the incidences of acute psychosis were 0.15nd 0.51 per 1000 births during the first month and six monthsost partum, respectively. The corresponding incidences duringhe period 1999–2008 were 0.09 and 0.24 per 1000 births. Therere about 100,000 births each year in Sweden and the decreasedncidence of acute psychosis is of course very convincing.

. Discussion

In the 1990s the knowledge about correct deep and incorrectuperficial suckling in breastfeeding was widely spread in west-rn countries among midwives and lactation consultancies. Thebservation that a superficial nipple suckling from the breast is inef-ective and leads to breastfeeding problems is obvious [2,3,17,18],iving rise to complications such as engorgement, milk congestionnd mastitis. Due to the inefficient suckling, the milk accumulatesn the breast and has to be taken care of by the woman’s ownody. The milk protein �-casein is partially hydrolyzed to gen-rate �-casomorphins with opioid activity [1]. In our study, weould not find any women with mastitis within the first month afterirth. This is probably a result from the well-established routines atwedish hospitals regarding suckling techniques and a correctionf faulty breastfeeding.

Postpartum psychosis is a rare but serious disease that may ariseollowing birth. Although the disease has been known for a longime, with well-described symptoms, it is still difficult to deter-

ine the frequency of postpartum psychosis. Most studies do notistinguish between postpartum psychosis and other psychotic dis-ases since postpartum psychosis does not have its own entity inost classification systems. A recent study shows that hospitaliza-

ions for postpartum psychosis in Sweden are low, 0.04%, whereashe risk of being afflicted with postpartum psychosis is greatlyncreased (9.24%) if the woman has a history of psychosis beforeregnancy [5]. Here we show that the incidence of psychosis inhe first month and six months post partum has declined by a halfuring the past 30 years.

The levels of opioid peptides are elevated in psychiatric patientsuffering from acute schizophrenia and affective disorder, but alson postpartum psychosis [20]. Studies have revealed cases with high

evels of �-casomorphin-like materials in samples from both CSF9] and milk [19] in puerperal women with postpartum psychosis,ndicating the importance of �-casomorphin-8 and related pep-ides in this disorder. In this study, we investigated the levels of

es 51 (2014) 54– 58 57

�-casomorphin-8 like reactivity in milk and plasma from womenwith late mastitis leading to milk congestion. We found that the lev-els of �-casomorphin-8-like immunoreactivity were significantlyhigher in milk from women with mastitis. Furthermore, we couldsee a slight tendency of elevated plasma levels of the peptide frag-ment compared to the control group. Interestingly, the elevatedlevel of immunoreactive �-casomorphin-8 in milk and plasma wasreversed 2–3 weeks after the mastitis and milk congestion wasgone. This indicates that �-casomorphins are an important factorin the development of mastitis.

As previously mentioned, the �-casomorphins are peptidesderived from the milk protein �-casein [1]. To our knowledge,there are no reports indicating either synthesis or processingof these peptides from any other precursor protein within thecentral nervous system or in the periphery. Thus, the theorythat �-casomorphins are able to cross physiological barriers isstill valid [7,15]. �-casomorphin-8 seems to be able to passthe breast parenchyma-blood barrier and blood–brain barrierin pregnant and lactating women [13] and thereby reach thecentral nervous system. In infants, �-casomorphins seem to betransported over the permeable gastrointestinal wall allowingan uptake of �-casomorphins to the infant body [21]. Transportover the blood–brain-barrier, indicated by �-casomorphin-8 likeimmunoreactivity in certain brain regions in infants, has beenreported [15], believed to exert physiological effects. Therefore,it is tempting to believe that �-casomorphins, originating from amilk-derived protein, may possess properties to affect the CNS inchildren as well as pregnant and lactating women.

Previous studies have implied increased levels of �-casomorphin-8 in the CSF in women with postpartum psychosis[9]. That finding, together with our results that show elevated�-casomorphin-8-like immunoreactivity in the milk of lactatingwomen suffering from mastitis, may suggest a possible connectionbetween mastitis, elevated �-casomorphin levels and postpartumpsychosis.

5. Conclusion

Our study suggests that an improved suckling techniquehas reduced the number of women suffering from mastitis.In the women that eventually developed mastitis and wereincluded in our study, we saw an increased �-casomorphin-8-likeimmunoreactivity in milk compared to controls. In addition, similartendencies to the milk samples were seen in the plasma samples.Data from the Swedish National Board of Health and Welfare indi-cate the incidence of postpartum psychosis or acute psychiatricillnesses has declined by a half over the past 30 years. These diseaseshave been linked with increased �-casomorphin levels in milk,plasma and CSF. Therefore, a possible connection between incor-rect suckling, mastitis, �-casomorphins and postpartum psychosishas been suggested. Finally, elevated �-casomorphin-8 levels mayalso be one of the reasons behind the “postnatal blues”.

Acknowledgement

This study was supported by the Swedish Medical ResearchCouncil (Grant 9459).

References

[1] Brantl V, Teschemacher H, Blasig J, Henschen A, Lottspeich F. Opioid activities

of beta-casomorphins. Life Sci 1981;28:1903–9.

[2] Fisher C. Breastfeeding, a midwife’s view. Matern Child Health J 1981;6:52–7.[3] Frantz KB, Fleiss PM. Ineffective suckling as a frequent cause of failure to thrive

in the totally breastfed infant. In: Freier S, Eidelman A, editors. Human milk:its biological and social value. Amsterdam: Excerpta Medica; 1980. p. 318–21.

5 Peptid

[

[

[

[

[

[

[

[

[

[

[

[

8 L. Righard et al. /

[4] Glamsta EL, Marklund A, Hellman U, Wernstedt C, Terenius L, Nyberg F. Isolationand characterization of a hemoglobin-derived opioid peptide from the humanpituitary gland. Regul Pept 1991;34:169–79.

[5] Harlow BL, Vitonis AF, Sparen P, Cnattingius S, Joffe H, Hultman CM. Incidenceof hospitalization for postpartum psychotic and bipolar episodes in womenwith and without prior prepregnancy or prenatal psychiatric hospitalizations.Arch Gen Psychiatry 2007;64:42–8.

[6] Koch G, Wiedemann K, Drebes E, Zimmermann W, Link G, Teschemacher H.Human �-casomorphin-7 immunoreactive material in the plasma of womenduring pregnancy and after parturition. In: Schowen RL, Barth A, editors. Pep-tides and proteases: recent advances. Oxford: Pergamon Press; 1987. p. 35–40.

[7] Koch G, Wiedemann K, Drebes E, Zimmermann W, Link G, Teschemacher H.Human beta-casomorphin-8 immunoreactive material in the plasma of womenduring pregnancy and after delivery. Regul Pept 1988;20:107–17.

[8] Lindstrom L, Nyberg F, Terenius L. Postpartum psychosis – new discoveries inan old disease. Lakartidningen 1987;84:755–8.

[9] Lindstrom LH, Nyberg F, Terenius L, Bauer K, Besev G, Gunne LM, et al. CSF andplasma beta-casomorphin-like opioid peptides in postpartum psychosis. Am JPsychiatry 1984;141:1059–66.

10] Lindstrom LH, Terenius L, Nyberg F. Opioid peptides in psychiatric disorders.Clin Neuropharmacol 1992;15(Suppl. 1 Pt A):58A–9A.

11] Nyberg F. CSF opioids in pathophysiology. Handbook Exp Pharmacol1993;104:653–72.

12] Nyberg F. Casomorphins and hemorphins – opioid active peptides releasedby partial hydrolysis of structural proteins. In: Kastin A, editor. Handbook ofbiologically active peptides. San Diego: Academic Press; 2006. p. 1339–44.

13] Nyberg F, Lieberman H, Lindstrom LH, Lyrenas S, Koch G, Terenius L.Immunoreactive beta-casomorphin-8 in cerebrospinal fluid from pregnant

[

es 51 (2014) 54– 58

and lactating women: correlation with plasma levels. J Clin Endocrinol Metab1989;68:283–9.

14] Nyberg F, Terenius L. Identification of high molecular weight enkephalinpercursor forms in human cerebrospinal fluid. Neuropeptides 1985;5:537–40.

15] Pasi A, Mahler H, Lansel N, Bernasconi C, Messiha FS. beta-Casomorphin-immunoreactivity in the brain stem of the human infant. Res Commun ChemPathol Pharmacol 1993;80:305–22.

16] Renlund S, Erlandsson I, Hellman U, Silberring J, Wernstedt C, LindstromL, et al. Micropurification and amino acid sequence of beta-casomorphin-8 in milk from a woman with postpartum psychosis. Peptides 1993;14:1125–32.

17] Righard L. Are breastfeeding problems related to incorrect breastfeeding tech-nique and the use of pacifiers and bottles. Birth 1998;25:40–4.

18] Righard L, Alade MO. Sucking technique and its effect on success of breastfeed-ing. Birth 1992;19:185–9.

19] Terenius L, Lyrenas S, Lutsch H, Lindstrom L, Nyberg F, Lindberg B. Opioid pep-tides at term pregnancy in the early puerperium and in postpartum psychosis.Adv Biochem Psychopharmacol 1987;43:201–9.

20] Terenius L, Nyberg F. Opioid peptides in the cerebrospinal fluid of psychiatricpatients. Prog Brain Res 1986;65:207–19.

21] Teschemacher H, Koch G, Brantl V. Milk protein-derived atypical opioidpeptides and related compounds with opioid agonist activity. In: Brantl V,

Teschemacher H, editors. �-Casomorphins and related peptides: recent devel-opments. New York: VCH Publishers; 1993.

22] Umbach M, Teschemacher H, Praetorius K, Hirschhauser R, Bostedt H. Demon-stration of a beta-casomorphin immunoreactive material in the plasma ofnewborn calves after milk intake. Regul Pept 1985;12:223–30.