Background
Human breastmilk has been historically regarded as a source of nutrition for infants. Recent studies have evidenced that breastmilk is a complex and dynamic tissue that provides newborns with components involved with functions beyond nutrition, such as the breastmilk microbiota and mother-derived cytokines and leukocytes [
1]. The production and composition of breastmilk are partially modulated by external parameters such as the mother’s diet, stress levels, or health status [
2‐
4].
Obesity is an expanding public health problem worldwide that can be regarded as a state of low-grade systemic inflammation where larger adipocytes secrete proinflammatory mediators and recruit leukocytes [
5,
6]. Individuals suffering from obesity exhibit altered peripheral blood cell counts with increased risks of leukocytosis, and modulations in the phenotypes of lymphocyte subpopulations [
7,
8]. Obesity directly hampers breastfeeding by various mechanisms, including delayed lactogenesis, decreased milk supply, and issues in adequately positioning the infant, all of which have been previously described [
9‐
11]. In addition, obesity impacts micro- and macronutrient breastmilk composition and selectively regulates its abundance in soluble immune components. Higher levels of immunoglobulin A (IgA) and secretory IgA (sIgA) concentrations were found in obese serum and colostrum, respectively, while IgG and IgM concentrations were unaffected by maternal body mass index (BMI) [
12]. The relevance of increased sIgA in obese colostrum remains to be elucidated. It may be a consequence of the observed disruption of the microbiota in these samples. In Mexico, the colostrum from obese mothers overall contains a microbiota with more bacterial species (increased richness) and more diversity between species abundances (decreased evenness) compared to colostrum microbiota from lean mothers [
13,
14]. Obese colostrum microbiota also include more potentially pathogenic bacteria genus such as
Staphylococcus [
14]. This may also partly explain the regulation of immune soluble factors described in obese breastmilk, including decreased TGF-β and sCD14, while IL-1 β, IL-6, IL-8, IL-10, and TNF- α concentrations were not significantly altered [
15‐
17].
While historical empirical observations have associated breastfeeding with a moderately decreased risk of suffering from obesity later in life, in these studies, the weight status of breastfeeding mothers was not investigated and may be a confounding factor skewing conclusions [
18,
19]. Indeed, overall maternal obesity is associated with multiple immune-mediated negative outcomes for infants, including neurodevelopmental disorders and increased morbidity [
20‐
22]. To date, these outcomes have been discussed as consequences of epigenetic regulations and gut microbiota alterations in early life [
22]. Recent evidence suggests breastmilk-transferred immune factors may impact infant health as the transfer of immune factors through breastmilk may also promote the development of autoimmune conditions [
23,
24]. However, the consequences of obesity on the majority of breastmilk leukocyte populations have not been reported to date [
25].
As alterations in breastmilk immune components could impact infants’ future health, we sought to investigate the consequence of obesity on breastmilk leukocytes in a cross-section observational study [
25]. The primary objective of this study was to explore possible variations in leukocyte proportions in the colostrum of mothers suffering from obesity. A secondary objective of this work was to compare the proportions and characteristics of leukocytes depending on the tissue of origin: colostrum or peripheral blood.
Methods
Study design and participants
We conducted a cross-sectional study of leukocyte subpopulations in blood and colostrum of mothers with BMI <25 and BMI >30. This study was approved by the Ethics Committee of the Hospital Regional Materno Infantil, Servicios de Salud de Nuevo León, Mexico, and the Institutional Review Board at Escuela de Medicina y Ciencias de la Salud, TecSalud, in Monterrey, Mexico, with the ID CarMicrobioLHum-2018. All samples were collected and used following signed informed consent and anonymization, between October 2020 and March 2021 at the Hospital Regional Materno Infantil, in Nuevo León, Mexico. Briefly, adult mothers between 18 and 34 years of age were invited to participate during the first obstetric consultation occurring during the first trimester of gestation. Participants were allocated to the obese cohort (BMI >30) or lean cohort (BMI <25), according to declared pre-pregnancy weight during the first visit and in accordance with the World Health Organization classification guidelines [
26]. Eligibility to participate in the study was determined based on (1) mother’s age between 18 and 34 years, (2) adequate prenatal visits without any adverse event during pregnancy, (3) pre-pregnancy BMI <25 or >30, (4) term infant, and (5) willingness to participate. Exclusion criteria included (1) having received antibiotics anytime during the 3-month period before birth, or having received a prolonged antibiotic treatment (>3 months) anytime during pregnancy; (2) having received immunosuppressive doses of steroids during pregnancy; (3) previous monoclonal antibody treatment; (4) history of chronic disease (outside of obesity); (5) suffering from any dietary disease; (6) episodes of diarrhea during the last 2 weeks of pregnancy; (7) history of surgery within 12 months prior to pregnancy; and (8) history of antineoplastic treatment. Elimination criteria included (1) having received antibiotics for >24 h post-birth, (2) necessity of intensive care unit admission of the neonate, and (3) any additional cause impeding sample collection. Oxytocin was not used during the first stages of labor. However, oxytocin was prescribed in 34 of 41 subjects (84%), during the first 8 h after delivery, as per international recommendations [
27].
Regarding the variables of the study, the main independent variable and hypothesized predictor was the BMI, calculated from self-reported pre-pregnancy weight and size. Additional variables collected or measured in this work included participant’s age, primiparity (yes/no), infant gender, gestational age at birth, type of delivery (vaginal/C-section), weight of infant at birth, volume of colostrum obtained, and frequency of leukocyte subpopulations in blood and colostrum samples. Additional details are included in the study’s STROBE statement (Sup. Table
1) [
26,
28‐
31].
Participants provided blood and colostrum samples on a single occasion, within 2 days of giving birth. Briefly, 3–4 ml of peripheral blood was collected in K2-EDTA vacutainers and placed on ice until processing. Following washing of the breast and nipple area using soap and water, 1–3 ml of colostrum per donor was obtained through pump-assisted milk extraction and immediately stored on ice. All samples were processed and acquired on the flow cytometer within 3 h of collection.
Leukocyte enrichment from colostrum
Around 1 ml of colostrum was processed for cell enrichment prior to staining for flow cytometry. Briefly, samples were centrifuged at 400 g for 15 min at 4°C. The supernatant was discarded, and the cell pellet was washed twice with PBS/2% FBS. Cells were counted using trypan blue for viability assessment and aliquoted for flow cytometry staining.
Colostrum-enriched cell staining
Depending on availability, between 200,000 and 1 × 106 cells were used for staining, and the same number of cells per sample was kept as unstained control. The same antibody lots were used to stain both types of tissues and antibody titration optimizations were performed for each tissue type to optimize resolution of fluorescence intensity over background. Cells were resuspended in the antibody master mix, which consisted of 2.5 μL mouse anti-human CD2-APC (BD® cat. 560642), 5 μL mouse anti-human CD16-APC-H7 (BD® cat. 560195), 5 μL mouse anti-human CD19-V450 (BD® cat. 560353), 2.5 μL mouse anti-human CD36-PE (BD® cat. 555455), 5 μL mouse anti-human CD45-V500 (BD® cat. 560777), and 1.25 μL rat anti-human CD294-Alexa Fluor 647 (BD® cat. 558042), in a final 100-μL staining volume with PBS + 2% FBS per 106 cells. Samples were then incubated for 30 min on ice in the dark, then washed and resuspended in PBS/2% FBS. Ten min before acquisition, propidium iodide (BD® cat. 556463) was added to the tube as per the manufacturer’s recommendations.
Peripheral blood staining
Fifty microliters of anticoagulated peripheral blood was stained using 2.5 μL CD2-APC, 1.25 μL CD16-APC-H7, 5μL CD19-V450, 2.5 μL CD36-PE, 1.25 μL CD45-V500, and 1.25 μL CD294-Alexa Fluor 647, in a final 100-μL staining volume with PBS + 2% FBS for 30 minutes on ice, in the dark. Samples were then subjected to erythrocyte lysis using BD® Pharm Lyse solution (BD® cat. 555899) as per the manufacturer’s instructions. Ten minutes before acquisition propidium iodide was added to tubes as per the manufacturer’s recommendations.
Flow cytometry data acquisition
Samples were analyzed on a BD® FACSCelesta flow cytometer fitted with 405-nm, 488-nm, and 633-nm lasers and operated through the BD® FACSDiva software v.8. Cytometer settings were checked prior to all acquisition using CS&T beads (BD® cat. 642412) according to manufacturer’s instructions. Compensation controls were prepared using compensation beads (anti-mouse Ig, K Neg Control compensation, BD® cat. 552843) following the manufacturer’s recommendations. At least 30,000 uncompensated events were recorded from every sample, with the forward scatter (FSC) event threshold adjusted to 35,000 for peripheral blood and 28,000 for colostrum samples.
Flow cytometry data analysis
Cytometry data were analyzed using FlowJo software v.10 (Treestar LLC). Automatic compensation was performed prior to analysis, with a compensation matrix generated at each acquisition. A strict quality control workflow was established to ensure the exclusion of suboptimal quality samples that may artificially skew the final analysis (Fig. S
1) [
32]. Briefly, samples had to exhibit a stable flow stream (Side Scatter (SSC) vs. time), debris was excluded on SSC/FSC plot, sample viability >85% from the singlet gate, and >10,000 leukocytes (CD45
+ cells) acquired, to be included in the final analysis and comparisons [
33]. The gating strategy applied to discriminate the leukocyte populations has been previously described and is summarized in Table
1 [
31,
34]. Fluorescence minus one (FMO) controls of colostrum and blood samples were used to adjust gates, which were then applied to all samples.
Table 1
Flow cytometry qualitative thresholds considered to identify the investigated leukocyte populations
Granulocytes | |
Neutrophils | SSCbright, CD45+, CD16+ |
Eosinophils | SSCbright, CD45+, CD16-, CD2 / CD294+ |
Basophils | SSCint, CD45+, CD16-, CD2 / CD294+ |
Lymphoid lineage cells | |
Noncytotoxic T lymphocytes | SSCdim, CD45+, CD16-, CD2 / CD294+ |
Cytotoxic T/NK lymphocytes | SSCdim, CD45+, CD16+, CD2 / CD294+ |
B lymphocytes | SSCdim, CD45+, CD16-, CD2 / CD294-, CD19+ |
Monocytes | |
CD16- (classical) monocytes | SSCint, CD45+, CD16-, CD2 / CD294-, CD19-, CD36+ CD16- |
CD16+ (non-classical) monocytes | SSCint, CD45+, CD16+, CD2 / CD294-, CD19-, CD36+, CD16+ |
Precursors/Immature cells | |
Myeloid precursors | SSCdim, CD45+, CD19-, CD2/ CD294- |
Immature granulocytes | SSCbright, CD45+, CD16-, CD2 / CD294- |
Statistical analysis
Proportions of leukocyte subsets were calculated as % of CD45+ live cells per sample. Shapiro-Wilk tests were used to investigate data normality with α=0.05. Wilcoxon matched-pairs tests were used to compare intra-individual leukocyte proportions in paired blood-colostrum samples. Mann-Whitney U tests were used to compare leukocyte % and median fluorescence intensity (MFI) of surface markers in colostrum between study groups. Student t-tests were used to compare leukocyte proportions in blood samples. All statistical analyses were performed using GraphPad Prism v. 9, or SPSS® v. 26, IBM Corporation, Armonk, NY, USA. Graphs are showing discrete data and mean with SD and p values in the APA format.
Discussion
Here, we applied a 7-color panel for flow cytometry to investigate 10 major leukocyte subpopulations in peripheral blood and colostrum from mothers presenting lean or obese BMI [
31,
34]. In answering the primary objective of the study, we evidenced a reshaping of the colostrum B lymphocyte compartment in obesity, with less B cells present in the colostrum from mothers suffering from obesity, while all other leukocyte populations remained unaltered in the colostrum between groups. In answering the secondary objective of this study, we identified considerable cell-specific phenotypic alterations of all leukocyte subtypes investigated between blood and colostrum. The alterations evidenced included regulation of cell size, internal complexity, and surface expression CD45 and CD16. Altogether, this report informs for the first time on regulated processes in colostrum leukocytes possibly involved in activation and trafficking from human blood to colostrum and evidences regulations correlated to maternal obesity.
Neutrophil average proportions in colostrum were 1.5 to 5 times higher than previously reported using flow cytometry (medians >65% in both groups, versus less than 15% in [
31]), but similar to previously measured in colostrum using a blood hematology analyzer [
31,
36]. Blood-circulating neutrophils have a lifespan of a few hours only, which is significantly shorter compared to other leukocytes [
37]. Reducing the time between collection and analysis to < 3 h may have allowed increased detection of live neutrophils, compared to longer wait periods in earlier studies. Proportions measured in blood were higher than expected in this tissue, which is consistent with the literature reporting leukocytosis and impaired neutrophil apoptosis during pregnancy and labor [
38].
We show that the abundance of CD16 on the surface of neutrophils and of CD16
+ monocytes is significantly regulated by tissue type, and depending on the cohort. In lean cohort blood, neutrophils express significantly more CD16 while CD16
+ monocytes express significantly less CD16, than in colostrum. CD16 is a Fc gamma III receptor (FcgIIIR) for the constant fraction of IgG antibodies. CD16 is abundant on the surface of phagocytic cells and its presence correlates with the phagocytic capacity of opsonized pathogens [
39]. It is interesting to measure such a regulation for FcR of IgG in colostrum, as the main immunoglobulin isotypes present in colostrum are IgA and IgM, which are not recognized by CD16 [
40].
Downregulation of CD16 on colostrum neutrophils could be the result of ectodomain shedding caused by activation or apoptosis. While apoptosis is also generally marked by a decrease in cell size, here no variation was observed in neutrophil relative size between tissue, casting doubt on apoptosis being the cause of CD16 downregulation on neutrophil surfaces in colostrum. Neutrophil activation is a rational alternative in the light of the well-described colostrum microbiota [
14,
41,
42]. Finally, CD16 downregulation from colostrum neutrophils may be caused by internalization after cross-linking IgG Fc, in contrast to shedding proposed earlier. Overall, additional experiments are necessary to conclude on the cause of neutrophil CD16 downregulation in colostrum.
Contrasting from findings in neutrophils, in mothers from the lean cohort, the relative abundance of CD16 on CD16-expressing monocytes was higher in colostrum compared to blood. Of note, the present flow cytometry panel was not designed to further subclassify CD16+ monocytes between non-classical and intermediate populations, the latter known to express relatively less CD16 than the former.
Therefore, the observed difference could have various origins. There could be an expansion of the higher CD16-expressing non-classical monocytes population, as observed in peripheral blood during infections [
43,
44]. A possible alternative could be the upregulation of CD16 from the intermediate population, as previously described [
45]. Interestingly, this difference between tissues was not recapitulated in the obese cohort, because CD16 was significantly increased on blood monocytes compared to the lean cohort, to levels that were similar to that of CD16 in colostrum monocytes. This is consistent with obesity involving systemic low-grade inflammation and highlights the relevance of investigating CD16 expression levels on monocytes in addition to other monocyte characteristics known to be modulated by obesity [
46]. In the present study, the blood proportion of CD16
+ monocytes was not perturbated by obesity. It is a possibility that the distinctive post-partum immune profile is causing this discrepancy compared to the obesity-mediated modulations of blood monocytes described in the literature [
47]. Overall, it will be necessary to investigate further monocyte subpopulations in colostrum.
CD45 upregulation has been described on granulocytes upon exposure to pathogenic microbes and physiological activators such as fMLP [
48‐
50]. However, the implications of this regulation on the development of the immune response remain unclear, and conflicting results have been described. For example in neutrophils, upregulation of CD45 is consistent with their activation [
50]. CD45 is also partially involved in regulating various neutrophil immune functions like cell adhesion, phagocytosis, and ROS production [
51]. However, CD45 was also shown to downregulate neutrophil chemotaxis, and in turn, neutrophil ROS production was shown to inhibit CD45 [
52,
53]. Therefore, the present results warrant future in-depth analyses of the activation status of granulocytes present in colostrum using functional assays.
Breastmilk is the recommended source of nutrition for infants globally. At present, only exceptional conditions warrant a healthcare professional to consider discouraging this practice, including specific substance abuse but also treatments affecting the immune system of the mother [
54‐
56]. The presented results indicate that suffering from obesity significantly reduces the B lymphocyte compartment in the colostrum, without affecting peripheral blood. Much remains to be investigated about colostrum B lymphocytes in obesity. In peripheral blood, B lymphocytes from obese individuals are more inflammatory and less efficient at switching to memory B cells upon antigen exposure [
8]. Here, the features of colostrum B lymphocytes hint toward a phenotype of antibody-secreting cells, with increased cell size, although this remains to be confirmed. Infants born with an immature immune system benefit from the passive transfer of antibodies from their mothers through breastfeeding. This includes immunologically relevant concentrations of immunoglobulins in breastmilk over a long period of time and vaccine-induced antigen-specific IgA and IgG into breastmilk 2-6 weeks post-vaccination [
57,
58]. Unvaccinated infants therefore benefit from antibody-mediated protection against infectious diseases, in addition to training of their immature immune system by exposure to these components [
59]. Interestingly, a previous study described increased sIgA in obese colostrum [
12]. Although more studies are necessary to confirm these findings globally, it is possible that breast-tissue resident plasma cells secrete more sIgA in obesity to compensate for less B cells present in colostrum. The present results therefore suggest obesity may impact the quantity and quality of passive immunity provided to nursing infants.
Additionally, this work provides insights into the regulation of leukocyte trafficking between blood and colostrum since various significant trends were equally recapitulated in both cohorts. Overall, the data indicate minimal regulation of the lymphoid compartment between tissues while myeloid cells were significantly altered morphologically and on the cell surface in colostrum. Mechanisms of leukocyte recruitment to the alveolar lumen during lactation remain largely unknown. Leukocytes are thought to reach breastmilk through the paracellular pathway from a mammary gland origin, crossing tight junctions (TJ), and not by direct extravasation from blood vessels. As TJ are tightly sealed during lactation, it has been suggested that leukocytes are recruited before initiation of lactation [
60,
61]. However, a recent study showed increasing numbers of post-mitotic plasma cells in the mouse mammary gland during lactation, suggesting some recruitment may actually take place during lactation [
62]. Mouse breastmilk T lymphocytes express TJ proteins, possibly to maintain TJ integrity during leukocyte transmigration during lactation [
63]. On the other hand, extravasation is the reported process by which the mammary gland undergoes the initial leukocyte recruitment during pregnancy [
64]. In the context of infections, transmigration cause leukocytes to modulate membrane expression of various markers and overall exhibit a proinflammatory profile. This includes enhanced survival for granulocytes and lymphocytes, and increased phagocytosis for neutrophils and monocytes, among other features described in [
65]. Transcriptional analysis of the mammary gland throughout gestation, lactation, and weaning showed an upregulation of immune-related function during the involution of the tissue post-weaning, compared to earlier timepoints including lactation [
66]. Overall, this suggests a largely unknown complex process physiologically distinct from infection-induced leukocyte transmigration and calls for further investigations into breastmilk leukocyte recruitment.
Early literature has speculated active immunity transfer from breastmilk to neonates [
67]. More recently, breastmilk was shown to be significantly enriched in regulatory T cells compared to peripheral blood [
68]. This scattered literature implicates a regulation of leukocytes in breastmilk with potential outcomes in the suckling neonate. Here, providing a differential description of leukocyte phenotypes in both tissue types helps to start dissecting this complex and selective recruitment process. We describe that the mothers’ BMI impacts B lymphocyte proportions in colostrum, suggesting a mother’s health status may in turn affect neonatal health.
A possible limitation of this work was relying on BMI to organize cohorts. Various reports demonstrate that BMI alone may not be a sufficient indicator for obesity and % body fat should be used instead [
69,
70]. In addition, recruitment and allocation to cohorts were performed during the first trimester of pregnancy, without later revisions of weight gain. We argue that overall first-trimester weight gain has been previously reported as minimal and that mothers suffering from obesity have a lower weight increase compared to lean mothers during this stage of pregnancy [
71,
72]. Therefore, the present results may be minimally confounded by differential weight gain during the development of the pregnancy.
Technically, while reporting leukocyte proportions in colostrum provide novel insights, it would be ideal to also measure absolute numbers of cells in colostrum. While earlier work has described absolute counts using BD TruCount tubes, the necessary pre-processing of colostrum samples may challenge the validity of the obtained results. Unfortunately, there is presently no alternative to estimate leukocyte absolute counts in breastmilk, while the physical properties of this tissue hamper their unprocessed use with TruCount tubes. Furthermore, we could not identify all of the leukocytes present in samples, as shown by events outside of population-calling gates, which is nonetheless consistent with the literature [
31]. While CD45
+ leukocytes make up the large majority of nucleated blood-circulating cells, rare CD45
− cells such as erythroid precursors or CD45
− megakaryocytes were recently reported in healthy individuals which could participate in explaining the < 3.5% CD45
− fraction identified in these samples [
73,
74].
This report highlights multiple key questions regarding active immunity in human colostrum, that require further study. First, what are the causes of the reduced B cell compartment in obese mothers’ colostrum, and what are the short- and long-term consequences in suckling infants? Why and how are leukocytes trafficked to colostrum, and is the altered phenotype in colostrum a requisite for, or a consequence of trafficking? Finally, the presented data hint toward activation of the innate immune system in colostrum, accentuating the need to investigate colostrum as a complex system, together with its microbiota. Host-microbe crosstalk should be considered in future studies to shed light on the mechanistic regulation of colostrum composition in obesity, and its impact on suckling infants.
Publisher’s Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.