Increasing evidence suggests that reactions to lipopolysaccharide (LPS), particularly in the gut, can be partly or completely mitigated by colostrum- and milk-derived oligosaccharides. Confirmation of this hypothesis could lead to the development of new therapeutic concepts.
To demonstrate the influence of equine colostral carbohydrates on the inflammatory response in an in vitro model with equine peripheral blood mononuclear cells (PBMCs).
Carbohydrates were extracted from mare colostrum, and then evaluated for their influence on LPS-induced inflammatory responses in PBMCs isolated from the same mares. mRNA expression of tumour necrosis factor-α, interleukin-6 and interleukin-10 was measured as well as the protein levels of tumour necrosis factor-α (TNF-α) and interleukin-10 (IL-10).
Equine colostral carbohydrates significantly reduced LPS-induced TNF-α protein at both times measured and significantly reduced LPS-induced TNF-α, IL-6 and IL-10 mRNA expression by PBMCs. Moreover, cell viability significantly increased in the presence of high concentrations of colostral carbohydrates.
Carbohydrates derived from equine colostrum reduce LPS-induced inflammatory responses of equine PBMCs.
Colostrum and milk-derived carbohydrates are promising candidates for new concepts in preventive and regenerative medicine.
For several decades, colostral compounds have been thought to have protective effects on neonates. In foals, the importance of colostral intake is well recognised by equine breeders and practitioners. Insufficient intake of colostrum by neonatal foals increases the risk of developing septic illnesses . Recent investigations demonstrate that next to the essential passive transfer of maternal immunoglobulins via colostrum during the first day of the foal's life , mammalian colostrum contributes significantly to the maturation of the developing gut-associated immune system and probably to the total immune competence in the first period of life. For example, human breast milk contains soluble Toll-like receptor 2 (TLR2) and CD14 (a protein essential for TLR4 signalling), which could inhibit signalling through TLR2 and TLR4 in the gastrointestinal tract of neonates by occupying binding sites for these signal molecules [3-5]. This may be important during early bacterial colonisation of the intestines and the development of tolerance to commensal bacterial flora. Furthermore, the presence of anti-inflammatory cytokines in human breast milk are purported to enhance the tight junction related epithelial barrier function [6, 7]. Recently, the colostral transfer of tumour necrosis factor-α (TNF-α) was documented to occur in newborn foals, eliciting a possible immunomodulatory role of this proinflammatory cytokine in early phases of life . The results of comparable studies in calves had previously indicated that stimulation of the immune system is mediated by proinflammatory cytokines present in bovine colostrum  and that lactoferrin, another component of bovine colstroum, has immunomodulatory and antimicrobial capabilities [10-12].
In light of the aforementioned immunomodulatory effects of colostrum, the results of recent mechanistic studies illustrate that bovine colostrum decreases NF-κB (nuclear factor kappa B) activation and subsequent production of proinflammatory cytokines in intestinal epithelial cells [13, 14]. In a study performed with human peripheral blood mononuclear cells (PBMCs), immune responses to several stimuli were modulated by bovine colostrum [15, 16]. Furthermore, distinct human milk-derived oligosaccharides were shown to exhibit immunomodulatory as well as prebiotic properties . Moreover, glycans in human milk have been proven to protect neonates against bacterial invasion and are thought to play an important role in the optimal bacterial colonisation of the neonatal gastrointestinal tract .
Limited research has been conducted with equine colostrum, despite the known differences in chemical composition and pattern of individual oligosaccharides between different animal species and man. Chemical analysis of equine colostrum reveals both similarities and differences in oligosaccharide patterns of man and horses [19-21]. In this study, we investigated the effects of carbohydrates derived from equine colostrum on the lipopolysaccharide (LPS)-induced inflammatory response in equine PBMCs.
Materials and methods
Animals and sample collection
Six healthy adult mares (NRPS [Nederlands Rijpaarden en Pony Stamboek or the Dutch Ridinghorse and Pony Studbook] and New Forest breed) participated in this study. Within 1 h post partum, 50 ml of colostrum was collected and immediately stored at -20°C. Within 12 h post partum, 150 ml of blood was collected by jugular venipuncture directly into a sterile blood collection bag containing citrate phosphate dextrose adenine as an anticoagulant1. Blood samples were kept cooled during transport to the laboratory, where PBMC isolation started within 2 h after collection. Colostrum was processed simultaneously with PBMC isolation, and colostral extracts from each horse were tested on PBMCs from the same horse (homologous testing). All experimental procedures were approved by the committee on ethical considerations in animal experiments of Utrecht University (DEC Utrecht).
Blood samples were 1:1 diluted in fresh phosphate buffered saline (PBS)2 containing 2 mmol/l EDTA3 and subsequently layered over Ficoll-Paque plus4. After centrifugation (400 xg, 30 min at room temperature) PBMCs were pipetted from the Ficoll layer and washed twice in PBS/EDTA. PBMCs were resuspended in RPMI 1640 Medium2 containing 2 mmol/l glutamine2, 100 iu/ml penicillin2, 100 μg/ml streptomycin2 and 10% horse serum (prepared in our own laboratory according to standard procedures). PBMCs were counted using trypan blue and resuspended to a density of 4 x 106 cells/ml medium. Following storage overnight at 4°C to attenuate possible stimulatory effects of the Ficoll, PBMCs were seeded in 24 well plates at a density of 4 x 106 cells/ml medium/well.
The carbohydrate fraction of the colostrum samples (equine colostral carbohydrates, eCC) was extracted as described by Fukuda et al. . A colostrum sample of 50 ml was thawed and mixed with 210 ml chloroform:methanol 2:1 v/v. This emulsion was centrifuged in glass tubes for 30 min at 4°C and 4000 ×g and the lower chloroform layer and denatured protein discarded. Methanol was removed from the upper layer using a vacuum centrifuge. The residue was dissolved in 35 ml fortified RPMI 1640 medium (as described above) and filtered through a 0.2 μm filter. This solution was called 1.00 eCC (stock solution) and dilutions of 0.50 eCC and 0.25 eCC prepared in RPMI 1640 medium were used in the following experiments.
Cell culture experiments
After seeding the PBMCs in 24 well plates, the plates were incubated for 2 h at 37°C and 5% CO2. The plates were then centrifuged for 10 min at 400 ×g to refresh the medium without removing PBMCs. Before starting the experiments, PBMCs were preincubated with 0, 0.25, 0.50 or 1.00 eCC medium solutions as described above. After preincubation, the experiments were started (t0) by replacing the medium with medium containing 0 or 1 μg/ml LPS (Escherichia coli O111:B4)3 and 0, 0.25, 0.50 or 1.00 eCC. The concentration of LPS was selected based on the results obtained from preliminary experiments performed in a similar manner. Plates were placed in the incubator and samples for PCR and ELISA were collected at 2 and 4 h (t2 and t4). All different combinations and time points were investigated in triple for each horse. For the ELISAs, supernatants were collected and stored at -80°C. For PCR, the PBMCs were lysed using RNA lysis buffer5 and stored at -80°C until RNA isolation was resumed.
Cell viability assessment
To investigate possible influence of the colostral extract on cell viability, 2 assays were performed using Alamar Blue3 and CCK-83. Cell viability assays were performed according to manufacturer's instructions. Before viability assessment, PBMCs were incubated with 0, 0.0625, 0.125, 0.25, 0.50, 0.75 and 1.00 eCC for 6 h in total (similar to the maximal exposure to eCC in the experiments; 2 h of preincubation and 4 h of experiment).
To measure protein levels of TNF-α and interleukin-10 (IL-10), ELISAs were performed on the supernatants using Duoset ELISA Development System for equine TNF-α and equine IL-106. Standard operating procedures of the manufacturer were followed, with all required buffers and solutions being used in the form provided by the manufacturer6. The detection limits of the ELISAs were 15.625 ng/ml for TNF-α and 156.25 ng/ml for IL-10.
RNA was isolated from PBMCs using the SV Total RNA isolation system6 according to the manufacturer's instructions. Isolated fractions were dissolved in 50 μl RNAse free water and stored at -80°C. Quality and quantity of RNA was determined spectophotometrically (Nanodrop).
Real-time PCR analysis
cDNA was generated using the iScriptTM cDNA Synthesis Kit7 according to the manufacturer's protocol. For reverse transcriptase reactions, either 1000 or 500 ng RNA was used per sample. Expression of mRNA was assessed by real-time PCR using a Biorad iQ5 Multicolor Real-time PCR detection system and iQ SYBR Green Supermix7. Specific primer pairs were designed and tested for efficiency and accuracy (before testing specificity was investigated using the NCBI-BLASTN search programme). For this study, expression of IL-6, IL-10, TNF-α and GAPDH mRNA was determined using the following primer pairs:
IL-6; F 5'-TGGCTGAAGAACACAACAACT-3’ R 5'-GAATGCCCATGAACTACAACA-3’,
IL-10; F 5'-GAGAACCACGGCCCAGACATCAAG-3’ R 5'- GACAGCGCCGCAGCCTCACT-3’,
TNF-α; F 5'-TCCAGACGGTGCTTGTGC-3’ R 5'- GGCCAGAGGGTTGATTGACT-3’,
GAPDH (HK gene); F 5'-TGGCATGGCCTTCCGTGTCC-3’ R 5'-GCCCTCCGATGCCTGCTTCAC-3’.
Data analysis and statistical methods
Data were analysed by means of linear regression. Independent variables were LPS (1 μg/ml), different concentrations of colostrum extract, and time (t2 and t4), allowing for interactions among all 3. Control samples from t0 were left out as they were not informative for the model (i.e. there were no other t0 measurements). For the PCR measurements, the dependent variable was time until PCR gave a positive signal, expressed as the number of PCR cycles (decimals allowed). Parameter estimates (other than the intercept) represented the difference in number of cycles between the reference and another experimental setup (for example, LPS vs. control). Consequently, 2 to the power of any parameter estimate represented the relative expression of a gene in a sample compared with the reference, being the blank measurements of the same time point (blanks at t4 were compared with blanks at t2). For the ELISA measurements, the dependent variable was the log-transformed concentration. Both dependent variables were assumed to be normally distributed.
Parameters were estimated using a generalised linear mixed model. Clustering of data was modelled through random intercepts at the level of horses/plates (48 samples for each of the 6 horses, with samples from each horse on one ELISA or PCR plate). Any ELISA measurements below the detection limit were assumed to be equal to the detection limit (TNF-α: 15.625 ng/ml; IL-10: 156.25 ng/ml). Models were fitted by means of residual maximum likelihood (REML) approach in R+ (version 2.14.0)8 using the function ‘lmer’ from the package ‘lme4’ (http://lme4.r-forge.r-project.org/). REML-ratio tests indicated that addition of aforementioned random intercept significantly improved the model (specifically, there was significant interhorse variation). P values and confidence intervals (highest posterior density intervals) were estimated by means of Markov chain Monte Carlo sampling from the posterior distribution of parameter values (10,000 iterations).
Cytokine production by PBMCs is reduced by eCC
Results of the ELISAs for TNF-α and IL-10 are presented in Figure 1. Incubation of PBMCs with LPS resulted in significant (P<0.01, compared with blank controls) increases in TNF-α production at both t2 (40.09-fold) and t4 (216.41-fold). In contrast, IL-10 production after LPS challenge was not changed at t2 and only mildly (1.53-fold), but significantly, increased at t4 (P<0.01). The addition of eCC to the medium did not influence basal TNF-α production by PBMCs. However, at both t2 and t4 the 1.00 solution of eCC significantly reduced LPS-induced production of TNF-α response by 88% at t2 and by 92% at t4 (P<0.01). The IL-10 concentrations were significantly decreased (P<0.01) at all tested concentrations at both time points in both the presence and absence of LPS (compared with blank controls in the absence of LPS and compared with the response to LPS alone in LPS treated cells).
Induction of cytokine mRNA expression is reduced by eCC
Results of the RT-PCR measurements for TNF-α, IL-6 and IL-10 are presented in Figure 2. For all 3 cytokines, mRNA expression was increased significantly for cells incubated with LPS at both time points (P<0.01). These responses to LPS were significantly reduced by 1.00 eCC at both time points (P<0.01). In cells incubated in the absence of LPS, only IL-10 mRNA expression was decreased by eCC; this effect was dependent on the concentration of eCC (significant at t4 for 0.25, 0.50 and 1.00 eCC, P<0.01).
In these experiments, mRNA expression of the housekeeping gene GAPDH was not significantly influenced under all investigated circumstances, confirming its suitability as a reference gene.
Cell viability is improved by eCC
Results of both cell viability assays (Alamar Blue reduction and CCK-8) are presented in Figure 3. A concentration-dependent increase in cell viability was identified, with cell viability being significantly higher (P<0.05) for 0.50 eCC (CCK-8), 0.75 eCC (CCK-8) and 1.00 eCC (CCK-8 and Alamar Blue) than in control samples.
This study was performed to test the hypothesis that the inflammatory response induced in equine PBMCs by challenging these cells in vitro with LPS, could be suppressed by biologically active carbohydrates in colostrum. In the LPS challenged cells, mRNA expression of proinflammatory cytokines (TNF-α and IL-6) as measured by RT-PCR and production of TNF-α as measured by ELISA were significantly increased as expected. In support of the hypothesis, eCC suppressed the proinflammatory responses to LPS, as confirmed by reductions in TNF-α and IL-6 mRNA expression and TNF-α protein concentrations. Furthermore, the effect was dose-dependent. Nonspecific binding of TNF-α protein by the colostral extract could be largely excluded due to the differences in the suppression of TNF-α concentrations at different time points (88% reduction at t2 and 92% at t4) and the fact that ELISA and RT-PCR results were comparable. Moreover, neither 0.25 nor 0.50 eCC reduced TNF-α protein levels significantly. In the case of nonspecific binding, lower eCC fractions would probably have had a suppressive effect as well (given the great impact of eCC at a fraction of 1.00). In contrast to TNF-α, all concentrations of eCC tested in this study caused a reduction in IL-10 protein levels in comparison with cells incubated in the absence of eCC. This finding could indicate that IL-10 binds to eCC or that eCC components interfere with antibody binding in the ELISA. However, the significant and dose-dependent reduction in IL-10 mRNA expression by eCC is in agreement with the ELISA results, thereby making nonspecific binding of IL-10 by eCC components less plausible.
Alternatively, it is possible that the reduction of proinflammatory signals identified using this model is not modulated through IL-10. If that was the case, IL-10 production should increase in the presence of eCC to regulate the inflammatory response to LPS. In contrast, we observed a suppression of IL-10 production by eCC, concomitant to the suppression of proinflammatory cytokines. A possible explanation for this observation is the short time-span of the experiments. The functional response by IL-10 protein as a reaction to proinflammatory signals is possibly limited in this short-term model. This assumption is supported by the fact that LPS induced a clear increase in IL-10 mRNA expression, whereas at these time points only limited protein levels were measured (also in PBMCs not incubated with eCC). This could be caused by post transcriptional regulation of IL-10 mRNA but it can not be excluded that comparable experiments performed using longer incubation periods would result in upregulation of IL-10 protein levels.
Because in the presence of eCC the concentrations of all investigated cytokines decreased simultaneously in LPS challenged PBMCs, there probably is a direct influence of eCC in this model. This could be a direct influence of eCC on TLRs, as similarities exist in both the structure and immunomodulatory capacities of colostral oligosaccharides and known TLR ligands [17, 18, 23]. Among other ligands, TLRs are activated by bacterial constituents (such as LPS, which activates TLR4) and subsequently induce an intracellular signalling cascade leading to the production of various cytokines . In this study, while cytokine protein and expression levels were reduced by eCC (particularly in LPS stimulated cells), transcription of TNF-α was slightly upregulated in the presence of 1.00 eCC (t4). It cannot be excluded that this effect is caused by slight contamination of the natural colostrum with LPS. On the other hand, this finding may also indicate that similar to known immunomodulatory agents such as TLR agonists, eCC alone may trigger intracellular signalling pathways and, in the case of parallel activation of the immune system, modulate the cytokine response in favour of the host.
To exclude possible cytotoxic effects of the colostral extracts (and possible traces of chloroform and/or methanol), 2 cell viability assays were performed. Instead of solely excluding cytotoxic effects, we detected a significant increase of PBMC viability at the high concentrations of eCC in both assays. Furthermore, this increase in cell viability was especially evident at the concentrations of colostral carbohydrates that suppressed the inflammatory responses to LPS.
Using isolated equine PBMCs, we have primarily demonstrated that mixtures of colostrum-derived oligosaccharides exert dose-dependent immunomodulatory effects. Fingerprint-analysis (data not shown) showed clear differences between equine and bovine colostral fractions but, at present, the exact composition and quantity of galacto-oligosaccharides in these samples await further characterisation. The results of the present study warrants further research of milk- and colostrum-derived oligosaccharides, with the aims of identifying the most active compounds and providing additional insight into the molecular processes involved in their anti-inflammatory effects.
The results of the present study indicate that equine colostral carbohydrates are capable of altering inflammatory responses of equine PBMCs stimulated with LPS in vitro. Moreover, cell viability of PBMCs was increased under the same experimental conditions. Despite the fact that more research is needed to elucidate the immunomodulatory mechanisms of specific constituents of the tested carbohydrate fractions, these compounds are promising candidates for the development of new strategies for prevention and treatment of intestinal inflammation in horses.
Authors’ declaration of interests
None of the authors of this paper has a financial or personal relationship with other people or organisations that could inappropriately influence or bias the content of the paper.
Source of funding
This research was funded institutionally by Utrecht University.
The authors thank the owners of the studfarm that kindly participated in this study.
All stated authors have contributed to the study design, data collection and study execution, data analysis and interpretation, and preparation of the manuscript.
Macopharma, Mouvaux, France.
Lonza, Basel, Switzerland.
Sigma-Aldrich, St. Louis, USA.
GE Healthcare, Waukesha, USA.
Promega, Madison, USA.
R&D Systems, Minneapolis, USA.
Biorad, Hercules, USA.
R Foundation for Statistical Computing, Vienna, Austria.