1
HIGH MATERNAL PERCEIVED STRESS IS ASSOCIATED WITH CHANGES
IN CORTISOL CONCENTRATIONS IN HUMAN MILK
ABSTRACT
Human milk is still considered the most optimal source of nourishment as it provides bioactive compounds like glucocorticoids to prepare new-borns for challenging postnatal conditions. The concentration of glucocorticoids in human milk is influenced by many factors, however the effect of stress on the concentration is unknown. The aim of the present study was to determine how maternal perceived stress influences the composition of the human milk with emphasis on the glucocorticoids cortisol and cortisone.
It was hypothesised that high perceived stress would be associated with alterations of the HPA-axis with a less prominent surge of glucocorticoids in human milk during the first hours after awakening and with a lower or normal basal concentration of glucocorticoids in human milk during the nadir hours compared to mothers with less stress, following the diurnal rhythm of the HPA-axis measured in blood serum. For the present study the Perceived Stress Scale (PSS) scores of breastfeeding mothers were compared to the concentration of cortisol en cortisone in human milk.
Results showed that higher PSS-scores were associated with lower concentrations of cortisol in human milk during the morning peak. Meanwhile, no association could be found between PSS-scores and the concentration cortisol in human milk during the nadir hours. Also, no association could be found between PSS-scores and the concentration cortisone in human milk during both the morning peak and nadir hours. Lastly, higher PSS-scores were associated with smaller changes of cortisol concentrations during the day. However, this relation was not found between PSS-scores and the change in cortisone concentration during the day. In conclusion, high maternal perceived stress is associated with changes in concentrations of cortisol in human milk.
Total word count: 4544
Bachelor Project BSc Biomedical Sciences Daily and senior supervisor
Anthony Lee Dr. E.F.G. Naninck
11641479 18 EC
2
HIGH MATERNAL PERCEIVED STRESS IS ASSOCIATED WITH CHANGES
IN CORTISOL CONCENTRATIONS IN HUMAN MILK
INTRODUCTION
In the 21st century breastfeeding is still considered the most optimal source of nourishment for new-borns. Breastfeeding provides favourable nutrition as it contains immune factors and microbiota, which are crucial for the development of a healthy immune system and gut microbiome (Kau et al., 2011; Mueller et al., 2015). Also, human milk regulates the growth, development and metabolism of new-borns by providing hormones, neuropeptides and growth factors through human milk (Ballard & Morrow, 2013). For mothers, breastfeeding is associated with improved health in the short-term by aiding the recovery after delivery and in the long-term by lowering the risk of developing future disease regarding metabolism, cardiovascular health and cancer (Godfrey & Lawrence, 2010). Thus, breastfeeding is associated with improved health outcomes for mothers and new-borns (Dieterich et
al., 2013) and therefore, the exclusive breastfeeding of new-borns for the first six months is
considered the normative standard for infant feeding and considered the ‘gold standard’ of infant feeding by which alternative feeding methods are compared to (Victora et al., 2016).
Unlike formula milk, human milk is a complex and dynamic biofluid where the composition of human milk is influenced by many factors, as it changes in response to the specific requirements of new-borns and the environment (Andreas et al., 2015). As earlier stated, human milk contains a wide variety of nutrients and bioactive compounds that could affect new-borns. This vertical transmission of bioactive compounds could prepare the new-born for the extra-uterine environment by changing the body’s structure and function (Hollanders et al., 2017). This process is known as lactational programming and could pose as a mechanism in which breastfeeding mothers could help their new-born child cope with a challenging postnatal environment (Langley-Evans, 2015). However, the effects of maternal conditions on human milk composition are currently not well understood. Therefore, the Amsterdam Mother Milk Study (AMS) was set up. The AMS researches the effect of maternal stress on the composition of human milk and the effect of the change in composition on the development of new-borns.
One of the bioactive compounds present in human milk, possibly attributing to lactational
programming are glucocorticoids (Ballard & Morrow, 2013). The primary glucocorticoids in human milk are cortisol and cortisone. Glucocorticoids are produced by the adrenal gland in response to physiological and psychological stress and in a circadian manner. Glucocorticoids are involved in regulating the stress response (Nicolaides et al., 2015), additionally, glucocorticoids play a key role in regulation of metabolism, the development of the brain and stimulation of surfactant production in the lungs (Fowden & Forhead, 2015; Miranda & Sousa, 2018). The concentration of glucocorticoids is regulated by the hypothalamus-pituitary-adrenal (HPA) axis and chronic or severe stress could lead to alterations of the axis, especially during sensitive periods of development (Ebner & Singewald, 2017). Earlier research has shown that adversities in early life could predispose to later
cardiovascular, metabolic and neuropsychiatric disorders (Gluckman et al., 2005). For example, it has been shown that alterations of the HPA-axis by prenatal use of glucocorticoids could lead to
increased risks for type 2 diabetes mellitus and cardiovascular disease in new-borns (Harris & Seckl, 2011). Also, it was shown in earlier research that high stress is associated with flattening of the circadian rhythm of cortisol in blood serum with most prominently a lower concentration of cortisol during the morning, whereas the concentration of cortisol during the nadir hours was not disrupted of only slightly blunted (Gunnar & Vazquez, 2001).
3 Past research has already shown that the glucocorticoid concentration in human milk is influenced by the diurnal pattern of the HPA-axis (Voorn et al., 2016). This study revealed that the concentration of cortisol and cortisone in human milk fluctuated during the day with a peak in the first hours after awakening and a decrease further throughout the day reaching a nadir in the evening. Other research has shown that the concentration of glucocorticoids in human milk is dependent on the gestational age of the new-born (Braun et al., 2013). This research revealed the role of placental corticotropin-releasing hormone, which increased the concentration of glucocorticoids during the last trimester of pregnancy. Also research has shown that maternal characteristics influence the amount of glucocorticoids in human milk (Pundir et al., 2019). The concentration of glucocorticoids in human milk was shown to be associated with maternal weight, pre-term birth and educational status of the mother.
The relation between maternal perceived stress and the concentration of glucocorticoids in human milk is currently unknown. The aim of the present study was to determine how maternal perceived stress influences the composition of the human milk with emphasis on the glucocorticoids cortisol and cortisone. It was hypothesised that high perceived stress would be associated with alterations of the HPA-axis with a less prominent surge of glucocorticoids in mother milk during the first hours after awakening and with a lower or normal basal concentration of glucocorticoids in mother milk during the nadir hours compared to mothers with less stress, following the diurnal rhythm of the HPA-axis measured in blood serum. For the present study the Perceived Stress Scale (PSS) scores of
breastfeeding mothers were compared to the concentration of cortisol and cortisone in human milk. The PSS-scores and human milk samples used for this study were collected for the AMS.
4
METHODS
Participants
For this study, which is part of the AMS, mothers were primarily recruited at the maternity
departments of Amsterdam UMC location VUmc and location AMC in Amsterdam, the Netherlands. The study was approved by the medical research ethics committee of the Amsterdam UMC location AMC (METC-AMC). Written informed consent was obtained from all participants.
As the AMS is still recruiting new participants, only a small part of participants was included in this research. For this research, a total of 60 participants was included, primarily gestational age >36.6 weeks. Mothers who breastfed their new-born infant were eligible for inclusion. Exclusion criteria were infant age >10 days, mother age <18 years, use of systemic glucocorticoid medication during since delivery, use of SSRI medication during pregnancy and/or after delivery, pre-existing or gestational diabetes and/or major congenital anomalies of the neonate.
Additional exclusion criteria were used after completion of measurement of cortisol and cortisone in human milk and further analysis of cortisol/cortisone concentrations. Milk samples were only included for further analysis when milk samples were complete with time of collection, measured cortisol/cortisone concentration and maternal PSS-score. Ultimately, a total of 53 participants were included for data-analysis
Sample collection
Milk samples were collected weekly at postpartum day 10, 17 and 21. For each collection day, the participants were asked to collect ±15 mL milk in total for that day. This total was acquired by collecting ±5 mL milk on three separate occasions, preferably during the morning, noon and evening. There was no distinction made between fore and hindmilk as leftover milk at separate feeding occasions could be used. The mothers breastfed their children and collected milk on demand and therefore were asked to write down the exact time of sample collection. Additional information about the method of collection and feeding status of the infant during the collection were also acquired. After collection, samples were stored in polypropylene vials at -21°C until aliquoting. After thawing, aliquots were made by mixing equal parts of the three separate samples of one collection for a sample representative for the collection day. The remaining milk of each sample was further divided. So for each separate sample collection day multiple samples were available: morning samples, noon samples, evening samples and mixed samples. After aliquoting, samples were stored in Eppendorf tubes at -21°C until analysis. It should be noted that only milk samples of postpartum day 10 were used in this study.
In addition to the milk samples, saliva samples were also collected weekly at postpartum day 10, 17 and 24. For each collection day, participants were asked to collect two saliva samples, one
immediately after waking up and one 30 minutes afterwards. Participants were instructed to collect saliva before brushing their teeth, eating and drinking. In addition, they were asked to write down the exact time of sample collection. Saliva samples were collected with Salivettes (Sarstedt). After each collection day, samples were immediately sent by post. Hereafter, saliva samples were collected according to the manufacturer’s instructions. After collection, samples were stored in Eppendorf tubes at -21°C until analysis. It should be noted that the saliva samples themselves were not used for this study, as only the time of awaking was used.
5
Measurement of cortisol and cortisone concentrations in human milk
For each milk sample, 0.5 mL of human milk was used to determine the concentrations of cortisol and cortisone by utilising an isotope dilution liquid chromatography-tandem mass spectrometry (LC-MS/MS) method as described in earlier research (van der Voorn et al., 2015). Analysis was performed at the AMC Department of Chemical Endocrinology.
In summary, internal standards of 2H
4-labeled cortisol and 2H8-labeled cortisone were mixed with 200 μL of thawed milk. Lipids were removed by washing the milk in 2 mL hexane three times in total. After each mixing, the mixture was centrifuged (2 min, 19 °C, 1900 g) and frozen (-60 °C CO2 ice bath), enabling the liquid hexane to be removed from the milk.
40 μL of washed milk was injected in an online solid phase extraction system with C8 cartridges for extraction and further purification. Sample elusion was performed by using methanol-water and this was further focused on a Synergi Hydro RP C18 guard column with a linear binary gradient from 55 to 61% methanol containing formic acid (0.1%) and ammonium acetate (2 mmol/L). Elusion time for respectively cortisol and cortisone were 5.2 and 4.8 minutes with a total run time of 7.5 minutes. Detection was performed with a Quattro Premier XE tandem mass spectrometer in electrospray positive ionisation mode with argon as collision gas. Further, detection was performed under a capillary voltage of 0.5 kV and a source temperature of 120 °C. Cortisol and cortisone were
respectively measured using the transitions (Q1 > Q3) m/z 363.2 → 121.1 and m/z 361.2 → 124.1.
Determination of perceived stress
The maternal perceived stress was determined by using the Perceived Stress Scale (PSS) as it measures the degree to which situations in one’s life were experienced as stressful during the past month (Cohen et al., 1983). For this study, a Dutch version of the 14-item, 5-point scale PSS
questionnaire was used. The questionnaire was completed right after the last sample collection day (postpartum day 24), approximately one month after delivery.
Data-analysis
To study relation between the concentration cortisol and concentration cortisone in human milk, the cortisol concentration in nmol/L was plotted against the cortisone concentration in nmol/L for regression line analysis. Student t-test was used to determine whether the slope of the regression line was significantly different from zero.
For the corrected time analysis, waking time extracted from the saliva samples were used as t=0 instead of 0:00, relating the time of sample collection to the moment of awaking. Original milk sample times were corrected by subtracting the waking time from the milk sample times. Samples with negative times were excluded, also samples with no awaking time provided.
To determine the peak and basal time slots, the time slots were arbitrary determined by visual determination of the peak. Peak concentrations were defined as the highest concentrations of respectively cortisol and cortisone concentrations during a 24-hour period. Basal concentrations were defined as lower concentrations of respectively cortisol and cortisone concentrations during a 24-hour period after peak. Both the concentrations of cortisol as the concentrations of cortisone were divided in peak and basal time slots. Data-analysis of whether the peak and basal time slot
6 concentrations of cortisol and cortisone were different was done by comparing the median
concentration of respectively cortisol and cortisone during peak and basal time slots. Mann-Whitney U test was used to determine whether the median concentration of cortisol/cortisone was
significantly different between peak and basal time slots.
To study relation between the maternal perceived stress and the glucocorticoid concentration, the PSS-score was plotted against the glucocorticoid concentration in nmol/L for both the peak and basal time slots. Student t-test was used to determine whether the slope of the regression line was
significantly different from zero.
To study relation between the maternal perceived stress and the change in glucocorticoid concentrations during the day, the PSS-score was plotted against the change in glucocorticoid concentration in nmol/L. The change in glucocorticoid concentration was calculated by subtracting the lowest concentration glucocorticoids from the highest concentration of glucocorticoids. Student t-test was used to determine whether the slope of the regression line was significantly different from zero. Participants with only one cortisol or cortisone measurement were excluded.
For all analysis, p <0.05 was defined as being significant. Statistical analysis was performed with GraphPad Prism 8.3.
7
RESULTS
Positive correlation found between the concentration cortisol and cortisone in human milk
0 10 20 30 40 0 20 40 60 80 100
Milk cortisol, [nmol/L]
M il k c o rt is o n e , [n m o l/ L ]
Figure 1. The relation between the concentration cortisol and cortisone in human milk
Displayed are the concentrations cortisol in human milk in [nmol/L] against the concentrations cortisone in human milk in [nmol/L] for each individual milk sample. n = 60 mothers and 154 milk samples. t-test of regression slope, slope = 1.956 and p<0.0001 with R2 = 0.5660.
Figure 1 presents the relation between the concentration cortisol and the concentration cortisone in human milk. t-test of the regression slope showed that the cortisol concentrations in human milk were positively correlated with cortisone concentrations in human milk (slope 1.9956, p<0.0001) with R2 0.5660 as shown in figure 1.
n = 60; 154 slope = 1.956 p<0.0001 R2 = 0.5660
8
Both cortisol and cortisone concentration in human milk showed a pattern suggestive for
diurnal rhythmicity
0 4 8 12 16 20 24 0 10 20 30 40Time during day, h
M il k c o rt is o l, [ n m o l/ L ]
Figure 2. The cortisol concentration in human milk during a 24-hour period with uncorrected time
Displayed are the concentrations cortisol in human milk in [nmol/L] against the time during day in hours. n = 53 mothers and 138 milk samples. The shown curve represents a spline best fitted to data.
0 4 8 12 16 20 24 0 20 40 60 80
Time during day, h
M il k c o rt is o n e , [n m o l/ L ]
Figure 3. The cortisone concentration in human milk during a 24-hour period with uncorrected time
Displayed are the concentrations cortisone in human milk in [nmol/L] against the time during day in hours. n = 53 mothers and 139 milk samples. The shown curve represents a spline best fitted to data.
Figure 2 and 3 presents respectively the fluctuations in cortisol and cortisone in human milk during a day with hours representing the natural hours of the day. Both the cortisol as the cortisone
concentration in human milk fluctuated during the day with a peak concentration during the morning hours and slowly diminished during the rest of the day, suggesting a diurnal rhythm as shown
respectively in figure 2 and 3.
n = 53; 138
9
Diurnal rhythmicity of cortisol and cortisone was not better visualised by correcting the time
to hours after waking
0 4 8 12 16 20 24 0 10 20 30 40
Corrected time after waking, h
M il k c o rt is o l, [ n m o l/ L ]
Figure 4. The cortisol concentration in human milk during a 24-hour period with corrected time
Displayed are the concentrations cortisol in human milk in [nmol/L] against the time during day in hours. n = 48 mothers and 119 milk samples. The shown curve represents a spline best fitted to data.
0 4 8 12 16 20 24 0 20 40 60 80
Corrected time after waking, h
M il k c o rt is o n e , [n m o l/ L ]
Figure 5. The cortisone concentration in human milk during a 24-hour period with corrected time
Displayed are the concentrations cortisone in human milk in [nmol/L] against the time during day in hours. n = 48 mothers and 120 milk samples. The shown curve represents a spline best fitted to data.
Figure 4 and 5 presents respectively the fluctuations in cortisol and cortisone in human milk during a day with hours representing the corrected time after waking of the day. Similar to the results with the uncorrected time, both the cortisol as the cortisone concentration with the corrected time after waking showed a peak when waking up and slowly diminished during the rest, also suggesting a diurnal rhythm as shown in figure 4 and 5. However, this rhythmicity was less evident than the concentration curves with uncorrected times for both cortisol as cortisone concentrations. For this reason, the uncorrected time graphs were used for further analysis.
n = 48; 119
10
Peak and basal time slots were determined to be 6:00 till 10:00 and 15:00 till 19:00
respectively
Peak Basal 0 5 10 15 M il k c o rt is o l, [ n m o l/ L ] ✱Figure 6. The difference in median cortisol in human milk between peak and basal time slots
Displayed are the median concentrations cortisol [nmol/L] during peak (6:00-9:59) and basal (15:00-18:59) time slots. For the peak: n = 29 mothers and 29 milk samples with median 8.4 nmol/L. For the basal: n = 25 milk and 25 milk samples with median 2.0 nmol/L. Mann-Whitney U test of median difference, * p<0.0001.
Peak Basal 0 10 20 30 40 50 M il k c o rt is o n e , [n m o l/ L ] ✱
Figure 7. The difference in median cortisone in human milk between peak and basal time slots
Displayed are the median concentrations cortisone [nmol/L] during peak (6:00-9:59) and basal (15:00-18:59) time slots. For the peak: n = 29 mothers and 29 milk samples with median 44.5 nmol/L. For the basal: n = 25 milk and 25 milk samples with median 19.5 nmol/L. Mann-Whitney U test of median difference, * p<0.0001.
Figure 6 and 7 presents the differences in median concentration cortisol and cortisone between peak and basal time slots. For both the cortisol as the cortisone concentrations in human milk, the peak was determined to be between 6:00 and 9:59 and basal was determined to be between 15:00 and 18:59. These time slots were determined to have approximately the same sample size. The median cortisol and cortisone concentrations during peak and basal were significantly different using the Mann-Whitney U test (both p<0.0001) as shown in respectively figure 6 and 7. These time slots were used for further analysis.
Peak n = 29; 29 median = 8.4 nmol/L Basal n = 25; 25 median = 2,0 nmol/L * p<0.0001 Peak n = 29; 29 median = 44.5 nmol/L Basal n = 25; 25 median = 19.5 nmol/L * p<0.0001
11
Negative correlation found between PSS-score and cortisol concentration during peak
0 10 20 30 40 0 10 20 30 40
PSS-score
M il k p e a k c o rt is o l, [ n m o l/ L ]Figure 8. The relation between the PSS-score and the concentration cortisol in human milk during peak
Displayed are the PSS-scores against the concentrations cortisol in human milk in [nmol/L] during the peak time slot for each individual milk sample. n = 28 mothers and 28 milk samples. t-test of regression slope, slope = −0.4412 and p=0.0480 with R2 = 0.1412. 0 10 20 30 40 0 20 40 60 80 PSS-score M il k p e a k c o rt is o n e , [n m o l/ L ]
Figure 9. The relation between the PSS-score and the concentration cortisone in human milk during peak
Displayed are the PSS-scores against the concentrations cortisone in human milk in [nmol/L] during the peak time slot for each individual milk sample. n = 28 mothers and 28 milk samples. t-test of regression slope, slope = −0.4211 and p>0.05 (n.s.) with R2 = 0.06223.
Figure 8 and 9 presents the relation between PSS-score and respectively the concentration cortisol and cortisone in human milk during the peak time slot (6:00-9:59). t-test of the regression slope showed that the PSS-score was negatively correlated with the cortisol concentration in human milk during the peak time slot (slope −0.4412, p=0.0480) with R2 = 0.1412 as shown in figure 8. However, no correlation was found between the PSS-score and the cortisone concentration in human milk during the peak time slot (slope −0.4211, p>0.05) with R2 = 0.06223 as shown in figure 9.
n = 28; 28 slope = −0.4412 p=0.0480 R2 = 0.1420 n = 28; 28 slope = −0.4211 p>0.05 (n.s.) R2 = 0.06223
12
No correlation found between PSS-score and glucocorticoid concentration during basal
0 10 20 30 40 0 5 10 15
PSS-score
M il k b a s a l c o rt is o l, [ n m o l/ L ]Figure 10. The relation between the PSS-score and the concentration cortisol in human milk during basal
Displayed are the PSS-scores against the concentrations cortisol in human milk in [nmol/L] during the basal time slot for each individual milk sample. n = 24 mothers and 24 milk samples. t-test of regression slope, slope = −0.05628 and p>0.05 (n.s.) with R2 = 0.02022. 0 10 20 30 40 0 20 40 60 80 PSS-score M il k b a s a l c o rt is o n e , [n m o l/ L ]
Figure 11. The relation between the PSS-score and the concentration cortisone in human milk during basal
Displayed are the PSS-scores against the concentrations cortisone in human milk in [nmol/L] during the basal time slot for each individual milk sample. n = 25 mothers and 25 milk samples. t-test of regression slope, slope = −0.1366 and p>0.05 (n.s.) with R2 = 0.005211.
Figure 10 and 11 presents the relation between PSS-score and respectively the concentration cortisol and cortisone in human milk during the basal time slot (15:00-18:59). t-test of the regression slope showed that the PSS-score was not significantly correlated with the cortisol concentration in human milk during the basal time slot (slope −0.05628, p>0.05) with R2 = 0.02022 as shown in figure 10. Similarly, no correlation was found between the PSS-score and the cortisone concentration in human milk during the basal time slot (slope −0.1366, p>0.05) with R2 = 0.005211 as shown in figure 11.
n = 24; 24 slope = −0.05628 p>0.05 (n.s.) R2 = 0.02022 n = 25; 25 slope = −0.1366 p>0.05 (n.s.) R2 = 0.005211
13
Negative correlation found between PSS-score and change in cortisol concentration
0 10 20 30 40 0 10 20 30 40 PSS-score M il k c o rt is o l c h a n g e , [n m o l/ L ]
Figure 12. The relation between the PSS-score and the change in cortisol concentration in human milk
Displayed are the PSS-scores against the change in cortisol concentration in human milk during the day in [nmol/L] for each individual milk sample. n = 52 mothers and 52 milk samples. t-test of regression slope, slope = −0.3269 and p=0.0122 with R2 = 0.1192. 0 10 20 30 40 0 10 20 30 40 50 PSS-score M il k c o rt is o n e c h a n g e , [n m o l/ L ]
Figure 13. The relation between the PSS-score and the change in cortisone concentration in human milk
Displayed are the PSS-scores against the change in cortisone concentration in human milk during the day in [nmol/L] for each individual milk sample. n = 52 mothers and 52 milk samples. t-test of regression slope, slope = −0.4809 and p>0.05 (n.s.) with R2 = 0.07333.
Figure 12 and 13 presents the relation between PSS-score and respectively the change in cortisol and cortisone concentration in human milk. t-test of the regression slope showed that the PSS-score was negatively correlated with the change in cortisol concentration in human milk during the day (slope −0.3269 and p=0.0122) and R2 = 0.1192 as shown in figure 12. However, no correlation was found between PSS-score and the change in cortisone concentration in human milk during the day (slope −0.4809, p>0.05) with R2 = 0.07333 as shown in figure 13.
n = 52; 52 slope = −0.3269 p=0.0122 R2 = 0.1192 n = 52; 52 slope = −0.4809 p>0.05 (n.s.) R2 = 0.07333
14
DISCUSSION
In summary, the results showed that the concentration cortisol in human milk was highly correlated to the concentration cortisone in human milk. We also showed that the concentration cortisol and cortisone both follow a clear diurnal rhythm. This is reflective for the circadian rhythm of HPA-axis activity, showing a peak concentration cortisol and cortisone during the waking hours by using uncorrected time measurement. However, correcting the time by the moment of waking did not provide better visualisation of the diurnal rhythm. Finally, higher PSS-scores were associated with lower concentrations of cortisol in human milk during the peak time slot. Meanwhile, no association could be found between PSS-scores and the concentration cortisol in human milk during the basal time slot. Also, no association could be found between PSS-scores and the concentration cortisone in human milk during both the peak and basal time slots. Lastly, higher PSS-scores were associated with smaller changes of cortisol concentrations during the day. However, this relation was not found in the change in cortisone concentration during the day. With these findings it could be concluded that high maternal perceived stress is associated with changes in concentrations of cortisol in human milk.
The finding that the concentration cortisol in human milk was associated with the concentration in cortisone in human milk was consistent with the results of earlier research (Pundir et al., 2019). Similar to the same study conducted by Pundir et al. and other studies, it has also been found in this study that the concentration of cortisone in human milk was higher than the concentration of cortisol in human milk. This is unlike the situation in the maternal blood plasma, where cortisol is the predominant glucocorticoid and cortisone the less abundant one (Meulenberg & Hofman, 1990). Multiple studies have suggested the role of 11β-HSD (hydroxysteroid dehydrogenase) type-2
isoenzyme as explanation for the higher concentration cortisone in human milk, as 11β-HSD converts cortisol to cortisone in breast tissue or milk. Therefore inactivating the biological active cortisol to the less active metabolite cortisone (Shams et al., 1998). However, the exact mechanism of this process is still unknown. Another possibility could be that serum cortisone penetrates human milk better than serum cortisol. This could be researched with gene KO of 11β-HSD and glucocorticoid measurement in blood serum and human milk. Nevertheless, as the concentration cortisone is correlated to the concentration cortisol in human milk, measurement of cortisone concentrations could be used as a measure of the cortisol concentration in human milk, as cortisol is less prominent in human milk.
Furthermore, the present study has showed diurnal rhythmicity of cortisol and cortisone
concentrations in human milk, which was also consistent with the results of past research (van der Voorn et al., 2016). As the procedure for the measurement of the cortisol and cortisone used in this study was the same as the one performed in the research of van der Voorn et al., the results are similar. Furthermore, both studies were performed at the same location (Amsterdam UMC) and therefore have similar participant populations. As a result, the peak concentration in both cortisol as cortisone concentrations in this study was around 7:00, identical to the peak time previously found. As the diurnal rhythmicity mimics the circadian rhythm of glucocorticoid activity in blood serum, it could be hypothesised that cortisol in human milk originates from blood serum.
Also, in the present study it was attempted to correct the time of milk sample by relating it to actual awaking time, but this have not resulted in better visualisation of the diurnal rhythm of both
glucocorticoids as the peak concentration cortisol and cortisone was less prominent. To the best of our knowledge, the present study was the first that has attempted to correct the time of sample collection by using the actual moment of awaking. The expectation was that by performing this
15 correction, the variance of the moment of awaking between participants could be reduced, as the morning cortisol peak is associated with the moment of awaking (Debono et al., 2009). However, despite the correction, the visualisation of the diurnal rhythm was not improved. This could be explained by the fact that indirect data for the moment of awaking was used, namely the time provided by the first saliva sample of the collection day. Also it should be noted that many participant have not collected saliva immediately after waking up, as many of the collected milk samples were collected before the saliva samples. Lastly, in the first month postpartum, maternal sleep patterns are disrupted with awakenings throughout the night (Insana et al., 2013). This could lead to more variance of the moment of awaking between participants. In future studies, this correction method could be improved by explicitly asking the moment of waking up instead of indirectly assuming the moment of waking.
Additionally, as mentioned earlier the peak of both cortisol and cortisone was similar to earlier research. However, the peak and basal time slots were arbitrary chosen by visual determination of the peak. As this method is not robust, future research could use generalized estimating equations (GEE) to determine the peak concentrations as performed by other studies (van der Voorn et al., 2016). Better assessment of the peak concentrations of cortisol and cortisone could also be achieved by requiring more samples on each milk collection day, for example every 3 hours, resulting in 5 samples per day. For this study, three samples for each milk collection date was acquired as more samples for each collection could influence the acquisition of samples negatively (smaller quantities of milk for each sample, participants less inclined to participate). However, benefits of more samples are better visualisation of the diurnal rhythm and less time variance between participants.
Lastly, the result that lower concentrations of cortisol during the morning peak in human milk was associated with high stress was similar to the result of earlier findings that lower cortisol
concentrations were found in participants with high stress compared to participants with low stress in blood serum (Gunnar & Vasquez, 2001). Similarly, the results that changes in cortisol
concentrations in human milk during the day are smaller with high stress are also similar to earlier research. As earlier described, it is hypothesised that the cortisol concentration in human milk could reflect the concentration cortisol in blood plasma. If this is the case, downregulation of HPA-axis activity measured in blood serum would also be visible in human milk. Past research have showed that high stress could lead to hyperactivation of the HPA-axis in the short term, however in the long term it could lead to hypoactivity of the HPA-axis, characterised by lower concentrations of cortisol during the morning peak in high stress people (Fries et al., 2005). Several mechanisms have been proposed to clarify this phenomenon, namely downregulation of number or sensitivity of target receptors in the HPA-axis; reduction of synthesis or release of hormones on different levels of the HPA-axis; reduction of the availability of free cortisol and/or upregulation of the sensitivity of negative feedback in the HPA-axis (Heim et al., 2000; Raison & Miller, 2003). As the cortisol
concentrations in serum are reflected in human milk, this ‘hypocortisolism’ could explain the lower concentrations of cortisol in high stress participants and also the small changes in cortisol
concentrations in human milk. As cortisone could also be synthesised by converting cortisol to cortisone in human milk, the effect of stress is less likely to be reflected on the concentrations of cortisone in mother, as presumably this conversion is independent of stress. This was also supported by the results of this as study, as no association was evident between cortisone concentrations and stress in both peak and basal time slots and between change in cortisone concentrations and stress. As mentioned earlier, the concentration cortisol in human milk is influenced by many factors and this study has found another factor, maternal stress. Although maternal stress could lower the amount of glucocorticoids in human milk during peak hours, at the moment it is unknown what the effect of this
16 change is. Glucocorticoids in human milk could play an important role in the adjustment of the HPA-axis in the new-born, possibly leading to a predisposition for disease regarding cardiovascular, metabolic and neuropsychiatric health later in life (Ebner & Singewald, 2017; Gluckman et al., 2015). If this is the case, human milk of mothers with high stress could be supplemented with
glucocorticoids. However before this is possible, more research on the effect of this change in glucocorticoids in human milk should be done and what the optimal amount of glucocorticoids in human milk is.
Although this study is limited by the small sample size, it was however adequately powered and significant results were found. Further research could be done by replicating this research with a larger sample size, which is part of future research performed by the AMS. As the sample size of the AMS is larger (n=160) refinement of the current research is possible. For example, it could be possible to study the changes of glucocorticoid concentrations in human milk during the first three weeks after delivery. Also it is possible to differentiate groups (pre-term versus full-term) and take in account factors that could influence the concentration glucocorticoids (maternal weight, pre-term birth, educational status mother) to determine whether the found effect in this study could actually be attributed to stress (Braun et al., 2013; Pundir et al., 2019). In addition, it is also interesting to look at the fat composition and other metabolites in the human milk as the concentration of glucocorticoids could depend on the fat composition of milk (Sullivan et al., 2011). Lastly, as it is unknown what the effect of changes in glucocorticoid concentrations in human milk will pose in the long term, longitudinal studies are needed. Fortunately, most of these questions will be addressed by the AMS.
In conclusion, high maternal perceived stress could be associated with changes in concentrations of cortisol in human milk. Although the results of this study are promising, little is known about the effect of this change in glucocorticoid concentrations on the developing new-born. As the AMS is still collecting new data, it is possible that the effect of glucocorticoid disturbances on new-borns could soon be revealed. Nevertheless, further research is needed to elucidate the effect of glucocorticoids on the new-born, but also the effect of stress on the general composition of human milk.
17
REFERENCES
Andreas, N. J., Kampmann, B., & Le-Doare, K. M. (2015). Human breast milk: A review on its composition and bioactivity. Early human development, 91(11), 629-635.
Ballard, O., & Morrow, A. L. (2013). Human milk composition: nutrients and bioactive factors. Pediatric Clinics, 60(1), 49-74.
Braun, T., Challis, J. R., Newnham, J. P., & Sloboda, D. M. (2013). Early-life glucocorticoid exposure: the hypothalamic-pituitary-adrenal axis, placental function, and long-term disease risk. Endocrine reviews, 34(6), 885-916.
Cohen, S., Kamarck, T., & Mermelstein, R. (1983). A global measure of perceived stress. Journal of health and social behavior, 385-396.
Debono, M., Ghobadi, C., Rostami-Hodjegan, A., Huatan, H., Campbell, M. J., Newell-Price, J., ... & Ross, R. J. (2009). Modified-release hydrocortisone to provide circadian cortisol profiles. The Journal of Clinical Endocrinology & Metabolism, 94(5), 1548-1554.
Dieterich, C. M., Felice, J. P., O’Sullivan, E., & Rasmussen, K. M. (2013). Breastfeeding and health outcomes for the mother-infant dyad. Pediatric Clinics of North America, 60(1), 31.
Ebner, K., & Singewald, N. (2017). Individual differences in stress susceptibility and stress inhibitory mechanisms. Current Opinion in Behavioral Sciences, 14, 54-64.
Fowden, A. L., & Forhead, A. J. (2015). Glucocorticoids as regulatory signals during intrauterine development. Experimental physiology, 100(12), 1477-1487.
Fries, E., Hesse, J., Hellhammer, J., & Hellhammer, D. H. (2005). A new view on hypocortisolism. Psychoneuroendocrinology, 30(10), 1010-1016.
Gluckman, P. D., Hanson, M. A., & Pinal, C. (2005). The developmental origins of adult disease. Maternal & child nutrition, 1(3), 130-141.
Godfrey, J. R., & Lawrence, R. A. (2010). Toward optimal health: the maternal benefits of breastfeeding. Journal of women's health, 19(9), 1597-1602.
Gunnar, M. R., & Vazquez, D. M. (2001). Low cortisol and a flattening of expected daytime rhythm: Potential indices of risk in human development. Development and psychopathology, 13(3), 515-538.
Harris, A., & Seckl, J. (2011). Glucocorticoids, prenatal stress and the programming of disease. Hormones and behavior, 59(3), 279-289.
Heim, C., Ehlert, U., & Hellhammer, D. H. (2000). The potential role of hypocortisolism in the pathophysiology of stress-related bodily disorders. Psychoneuroendocrinology, 25(1), 1-35.
Hollanders, J. J., Heijboer, A. C., van der Voorn, B., Rotteveel, J., & Finken, M. J. (2017). Nutritional programming by glucocorticoids in breast milk: targets, mechanisms and possible implications. Best Practice & Research Clinical Endocrinology & Metabolism, 31(4), 397-408.
18 Insana, S. P., Williams, K. B., & Montgomery-Downs, H. E. (2013). Sleep disturbance and
neurobehavioral performance among postpartum women. Sleep, 36(1), 73-81.
Kau, A. L., Ahern, P. P., Griffin, N. W., Goodman, A. L., & Gordon, J. I. (2011). Human nutrition, the gut microbiome and the immune system. Nature, 474(7351), 327.
Langley‐Evans, S. C. (2015). Nutrition in early life and the programming of adult disease: a review. Journal of Human Nutrition and Dietetics, 28, 1-14.
Meulenberg, P. M., & Hofman, J. A. (1990). Differences between concentrations of salivary cortisol and cortisone and of free cortisol and cortisone in plasma during pregnancy and postpartum. Clinical Chemistry, 36(1), 70-75.
Miranda, A., & Sousa, N. (2018). Maternal hormonal milieu influence on fetal brain development. Brain and behavior, 8(2), e00920.
Mueller, N. T., Bakacs, E., Combellick, J., Grigoryan, Z., & Dominguez-Bello, M. G. (2015). The infant microbiome development: mom matters. Trends in molecular medicine, 21(2), 109-117.
Nicolaides, N. C., Kyratzi, E., Lamprokostopoulou, A., Chrousos, G. P., & Charmandari, E. (2015). Stress, the stress system and the role of glucocorticoids. Neuroimmunomodulation, 22(1-2), 6-19.
Pundir, S., Mäkelä, J., Nuora, A., Junttila, N., Wall, C. R., Linderborg, K., ... & Lagström, H. (2019). Maternal influences on the glucocorticoid concentrations of human milk: The STEPS study. Clinical Nutrition, 38(4), 1913-1920.
Raison, C. L., & Miller, A. H. (2003). When not enough is too much: the role of insufficient glucocorticoid signaling in the pathophysiology of stress-related disorders. American Journal of Psychiatry, 160(9), 1554-1565.
Shams, M., Kilby, M. D., Somerset, D. A., Howie, A. J., Gupta, A., Wood, P. J., ... & Stewart, P. M. (1998). 11Beta-hydroxysteroid dehydrogenase type 2 in human pregnancy and reduced expression in intrauterine growth restriction. Human Reproduction (Oxford, England), 13(4), 799-804.
van der Voorn, B., Martens, F., Peppelman, N. S., Rotteveel, J., Blankenstein, M. A., Finken, M. J., & Heijboer, A. C. (2015). Determination of cortisol and cortisone in human mother’s milk. Clin Chim
Acta, 444(0), 154-5.
van der Voorn, B., de Waard, M., van Goudoever, J. B., Rotteveel, J., Heijboer, A. C., & Finken, M. J. (2016). Breast-milk cortisol and cortisone concentrations follow the diurnal rhythm of maternal hypothalamus-pituitary-adrenal axis activity. The Journal of nutrition, 146(11), 2174-2179.
