Matrine-chitosan hydrogels for treating subclinical bovine mastitis by intramammary infusion—effect on milk microbiome and metabolites

Background The metabolic processes of cows undergo signi�cant changes during subclinical mastitis, but their molecular mechanisms have not been clearly elucidated. This study investigated the changes in milk metabolites after intramammary infusion of matrine, a plant alkaloid with anticancer properties, in the form of a chitosan hydrogel into bovine mammary glands with subclinical mastitis. Infusions were continued for seven days, and milk samples were collected on day 1 and day 7 for analysis of the microbiome by 16S rRNA gene sequencing and metabolites by liquid chromatography-mass spectrometry. Results


Background
The udder microbiota plays an important role in the host-pathogen interactions of the innate and adaptive immune system.This is especially true of pathogens that trigger in ammation that is detrimental to mammary tissue and bovine physiology [1].Mastitis is an in ammation of the udder most commonly from bacterial infection, and the pathological changes not only affect milk quality but also quantity [2].The increasing demand for animal protein has led to substantial growth in the use of antibiotics in food animal production [3], which can cause antimicrobial resistance that is a global concern for treating human as well as farm animal diseases.Developing an effective drug for mastitis that is safe for use in food animals and for which no resistance exists in the microbial population would be highly desirable for the worldwide dairy industry.
Matrine is a polycyclic alkaloid isolated from the plant, Sophora avescens that has been used in Chinese traditional medicine because of its anticancer, anti-in ammatory, cardioprotective, and opioid effects [4].
Previous research [5] found that matrine inhibited the virulence of Staphylococcus aureus, one of the main pathogens of mastitis.Furthermore, matrine attenuated the lipopolysaccharide-induced immune response by downregulating IL-1 and IL-17 and inhibiting production of the proin ammatory compound, malondialdehyde, reducing in ammation and oxidative stress, and enhancing CCR7 expression [6,7].
Chitosan is a natural polymer produced by acetylation and enzymatic cleavage of chitin, which is the most abundant animal polysaccharide in nature.[8].The applications for chitosan have continued to increase because of the abundance of its source, its biocompatibility and ease of chemical modi cation.
Studies have also shown that chitosan has anti-in ammatory, antimicrobial, cholesterol-lowering, immunomodulatory, and antitumor properties [5].Chitosan can be bene cial at the cellular or molecular level [9], by reducing intracellular material leakage and triggering an antimicrobial response to enhance the therapeutic effects of mastitis treatment.Chitosan can be formulated to be injectable at room temperature but change into a biodegradable hydrogel at body temperature [10].It has been used in drug formulations and delivery vehicles for decades; however, the in uence of matrine and chitosan on the udders of dairy cows has received little research attention.Therefore, it is worthwhile to explore the effectiveness of intramammary infusion of matrine-loaded chitosan hydrogels (MCHs) on the udders of dairy cows with mastitis in terms of positively changing the microbial population and metabolic pro le of milk.We hypothesized that intramammary injection of MCHs would favor the growth of bene cial microorganisms and cause a shift in the milk microbiota in normal udder quarters.
DNA sequencing has become routine for many labs and the advanced software and bioinformatics databases available have created great opportunities for studying pathogenic mechanisms [11,12].By understanding the host responses to microbial attack, we hope to be able to identify the most effective targets for antimicrobial intervention, to develop new treatments for bovine mastitis and to nd sensitive biomarkers for early detection and diagnosis.Metabolomics can also be used for quantitative measurement of the metabolic state of milk, including the biomarkers of lactation [13], the variation in metabolites associated with mastitis [14], and differences in metabolism after antibacterial therapy [15].
Little is known about the changes in milk metabolites in response to MCH treatment of mastitis, and one of the goals of this study was to obtain detailed information on the effect of matrine delivery by chitosan hydrogel injection into bovine mammary glands on milk microbiota and metabolites.We investigated the effects MCHs on subclinical-mastitis and the relationship between milk microbiota diversity and metabolite pro les, to probe the mechanisms of the antibacterial bene ts of matrine and chitosan for mastitis, and to provide new knowledge for the development of novel, safe and effective prophylactic and therapeutic compounds for dairy cow operations.

Diversity and relative numbers of milk microbiota
A total of 1,117,565 high-quality sequences from twelve samples passed quality control and could be used for testing.The sequences averaged 308 bp, and there was greater than ninety-nine percent depth coverage; therefore, the amount of data was su cient to show differences in all bacterial species.The αdiversity indices of the microbiota are shown in table 1.Although ACE showed a tendency to decrease, no different was observed in the Chao indices, which were all representative of bacterial community richness.Moreover, the Simpson index showed that the bacterial diversity at day 7 was signi cantly lower than at day 1 (p 0.05), and there was an overall tendency to decrease (p = 0.06), but this is variable and depends on if the calculations are based on abundance or biomass.No signi cant differences were seen among the groups for the other α-diversities.
Principal component analysis (PCA) on the microbial populations present on day 1 vs day 7 showed clear differences, and the weighted UniFrac values were also determined (Fig. 2).Principal coordinates one and two were found to account for 51.76 and 26.2% of total variance, con rming that the milk microbiome after seven days of MCH infusion was remarkably distinct from that on day one.
The taxonomic changes in milk samples from day 1 to day 7 were determined.Proteobacteria, Firmicutes, Actinobacteria and Bacteroidetes were the three predominant phyla (Fig. 4A).The abundance of Firmicutes was extremely signi cantly decreased on day 7 (p = 0.01), whereas the numbers of Proteobacteria increased (p = 0.01) (Fig. 3A, B).Taxa with a relative abundance of one percent in at least one sample were identi ed, and the ten most abundant genera are presented (Fig. 3C, D).On day 7, there was a signi cantly lower abundance of Aerococcus (p = 0.01), Corynebacterium_1 (p = 0.08) and Staphylococcus (p = 0.03).In contrast, the relative abundances of Pseudomonas (p = 0.01) and Ralstonia (p = 0.02) were signi cantly increased.

Identi cation and comparison of milk metabolites
A non-targeted metabolome method was used to evaluate milk metabolites after MCH treatment.Reproducible metabolite data pro les (1001) were obtained and differences between day 1 and day 7 were characterized using VIP metrics from OPLS-DA.Seventy-four milk metabolite signals were obtained that were signi cantly different between MCH-treated and untreated cows (VIP > 1 and p < 0.05) (table 2).
They consisted mainly of sphingolipids, oxanes, glycerophospholipids, avonoids and fatty acyls.Milk from animals receiving MCHs for seven days had higher levels of steroids and steroid derivatives, prenol lipids, oxanes, macrolides and analogues, hydroxy acids and derivatives, carboxylic acids and derivatives and Benzene and substituted derivatives than milk on day 1.
PCA and OPLS-DA were employed to evaluate the differences in metabolomes.The PCA scores showed clear differences between day 1 and day 7 milk samples (Fig. 4A).Principal components one and two were found to account for 31.70% and 16.10% of the variation.The variables for assessing OPLS-DA model quality are shown in validation plots (Fig. 4B).The Q2 value of the OPLS-DA model was 0.939, and the R2Y value was 0.963.Q2 represents the model's prediction ability and the closer these 3 indicators are to 1, the more stable and reliable it is.Q2 > 0.5 indicates that the model's prediction ability is good, while Q2 < 0.5 indicates poor predictability.The OPLS-DA data con rmed that the two groups had signi cantly different types and levels of metabolites (Fig. 4C).OPLS-DA model integrity is shown by validation plots.The day 7 metabolite pro les were signi cantly different from those of day 1 indicating that the PCA and OPLS-DA results were valid for assessing variations in milk metabolomes between the two days.

Differences in metabolites resulting from changes in metabolic pathways
Comparison of metabolism via changes in pathway activation from MCH treatment on day 1 to 7 was performed using the Kyoto Encyclopedia of Genes and Genomes (KEGG) for determining enrichment of pathways involved in metabolite production (table 3; Fig. 5).Enrichment analysis showed that sphingolipid metabolism, the sphingolipid signaling pathway, glycerophospholipid metabolism and ABC transporters, were signi cantly affected by MCH injection.

Discussion
The results of this study revealed that MCHs markedly decreased somatic cell count after 7 days of treatment.Although we did not observe any changes in bacterial community richness by the Chao indices, the Simpson index showed a signi cant decrease in the bacterial diversity.In particular, on day 7, the relative abundance of Firmicutes was signi cantly decreased, and that of Proteobacteria was signi cantly increased compared to levels on day 1.It has been reported that the richness and diversity of the milk microbiome re ects the health of the cows and functional performance of the organs [16][17][18].Antibacterials are commonly used to treat bovine mastitis, but their e cacy is unclear [19].That is why we chose to test the plant-derived compound, matrine, as an intramammary treatment loaded into chitosan hydrogels to see if it would alter the milk microbiota and improve mammary gland function.
Measurement of the number of somatic cells in milk is an internationally recognized test for the detection of mastitis.A SCC of more than 200,000 per ml is considered to be evidence of disease [20].Recent investigations have shown that the milk microbiota of dairy cows are usually closely associated with the SCC [21,22].This suggests that modulation of the udder ecosystem through the microbiota could help to maintain homeostasis and enhance the mammary gland's defense mechanism [23,24]Consistent with these reports, we observed that the udder SCC was signi cantly increased by in mastitis by common pathogens like Staphylococcus, but was markedly decreased after perfusion of the udder with matrinechitosan hydrogel.This is in line with previous studies showing that chitosan hydrogels decreased the SCC in milk from dairy cows [25].The therapeutic synergy achieved by loading the chitosan hydrogel with matrine suggests potential application in mastitis protection and treatment.The mechanism by which the milk microbiota in uences in ammatory and immune responses to decrease the SCC still needs further investigation.
The next generation of high-throughput DNA sequencing methods together with updated bioinformatics software is now being used for the in-depth assessment of microbial communities to elucidate how bacterial activity and metabolites affect human and animal health [26].We found that the microbial diversity in milk was dramatically decreased after one week of intramammary infusions of matrinechitosan hydrogels.The relative abundance of Proteobacteria was signi cantly increased, while Staphylococcus numbers decreased signi cantly.Staphylococci are internationally recognized as one of the most prevalent mastitis pathogens, accounting for about 70% of the cases [2]..Moreover, it has been reported that Proteobacteria are signi cantly more abundant in healthy cows than those with mastitis [2].Previous studies also showed that the numbers of Proteobacteria after non-antibiotic hydrogel treatment increased signi cantly [27], indicating that hydrogel therapy can effectively return the milk microbial environment to normal diversity and defend against disease.Our investigation provides mechanistic insights into how the microbiota respond to MCH may help in the development of novel prophylactic and therapeutic products as alternatives to antibiotics in dairy cows.Our data are also important for understanding how the regulation of biosynthesis by milk microbiota in uences udder health and defense against mastitis.The milk microbiota are usually closely associated with the SCC [21,22].
suggesting that the udder microbiota plays an important part in modulation of the ecosystem and maintenance of homeostasis to resist mammary gland infections [23,24] Metabolomics is a comparatively new research method that has been widely used in the detection of mastitis in recent years due to its more comprehensive test results.Previous studies suggested that metabolomics could provide a more complete understanding of an animal's physiology and biochemistry .[28].The metabolomics data from this study highlight the potential function of matrine-chitosan hydrogels in modulating the metabolite levels, which were found to be enriched in sphingolipid metabolic pathways, phenylalanine, glycerophospholipid, and starch and sucrose metabolism.A previous report suggested that sphingolipid metabolites acted as signaling molecules to regulate a diverse group of cellular processes, particularly those related to immunity, in ammation and in ammatory disorders .[29].We speculate that this activity might be associated with the decrease in SCC caused by matrine-chitosan hydrogel injection.The metabolic differences we observed gave us a further insight into how MCH affects metabolite levels before and after mastitis treatment, and showed that alterations in the milk metabolome could be used to reveal the therapeutic mechanism of matrine for mastitis treatment and in the recovery of milk production in dairy cows.
We found signi cant changes in concentration of a number of sphingolipids, glycerolipids, fatty acyls, glycerophospholipids, and organo-oxygen compounds that signi cantly differed from day 1 to day 7.
Sphingolipids are major components of cell membranes and are considered to be widely involved in important processes such as cell aging and apoptosis .[30].Glycerolipids [30] and fatty acyls [32] are both important components of cell membranes.Previous research found that bacterial invasion can induce in ammatory reactions and cause oxidative damage during the development of the breast [33], accompanied by apoptosis .[34,35].The reduction in sphingolipids, glycerolipids, and fatty acyls seen in this study indicates that apoptosis caused by in ammatory reactions may have been weakened.The decrease of organooxygen compounds suggests that the oxidative damage caused by in ammatory reactions may also have been reduced.The metabolites were enriched in sphingolipid and glycerophospholipid metabolism [36], which are both related to apoptosis pathways.[37,38].Therefore, we conclude that MCHs can inhibit apoptosis by blocking oxidative damage, thereby reducing in ammatory changes to the udder.
It is worth noting that L-phenylalanine is involved in many pathways that are signi cantly changed in this study (table 3).L-phenylalanine is an essential amino acid and a precursor for the synthesis of antiviral and anticancer drugs as well as new sweeteners.It can also be a marker of in ammatory reactions and affect the body's immune response [33].Here, the content of L-phenylalanine was signi cantly increased on day 7, which means that the immune defenses in the mammary gland may have been effectively improved.Previous study [35] found that tryptophan and phenylalanine are used in the body to synthesize serotonin, L-DOPA derivatives and 5, 6, 7, 8-tetrahydrobiopterin (BH4), which compounds can affect the severity of in ammation.In the study of Smith , [39], tyrosine and phenylalanine participated in the regulation of TLR4 signaling pathways, thereby in uencing the degree of in ammation.Phenylalanine also is part of the glycolytic and liposynthetic pathways.[40,41].It has been reported that phenylalanine de ciency can severely affect protein metabolism in the breast and compromise the health of the organ.[42].Phenylalanine is an essential amino acid in milk and is also the precursor of tyrosine, which is one of the main amino acids in milk.[43].We know that untreated bovine mastitis results in reduced milk fat and milk protein content; therefore, the increase in L-phenylalanine by MCH injection could be effective in restoring normal milk production during lactation.
The milk metabolites produced by MCH injection were found to have an intimate relationship with sphingolipid and glycerophospholipid metabolism.A previous study suggested that the sphingolipid metabolic pathways participate in a variety of immune-related signal transductions, in ammation and in ammatory disorders [29].The sphingolipid metabolism-related pathways have many functions, such as regulating cell adhesion and cellular immunity, activating cancer repressors, regulating apoptosis, and modulating immune function and the in ammatory response [29] We speculate that this activity might be associated with the decrease in SCC caused by MCH injection.Furthermore, Cowart et al. [44] showed that an imbalance in the pathway for sphingolipid metabolism can cause ketosis, mastitis, and metritis in cows.It has been proved that sphingolipid metabolites, especially ceramide and sphingosine-1phosphate, can regulate a variety of biochemical processes important in immune, in ammation and in ammatory diseases [29].In the present study, we found that MCHs could signi cantly regulate the sphingolipid metabolism-related pathways to improve in ammation; thus, it is reasonable to believe that MCHs could effectively treat bovine mastitis.
Despite the documented health-promoting properties of matrine and chitosan, scienti c evidence for the e cacy of MCHs in dairy cows is limited.Our study affords many useful insights, but the effect of MCHs on immune regulation still requires further study.Another limitation relates to the mechanistic links involved in the observed signi cant changes in the microbiota and biomarker metabolites as a result of MCH injection that will need further exploration to be clearly understood.Previous studies suggested that the antimicrobial activity of chitosan can stimulate the innate immune response and hasten the involution process of the mammary gland [25].It is not known whether matrine-chitosan complexes could be used in the circulation to regulate the immune system and inhibit in ammation, but our data indicate that future research in this area is strongly warranted.

Conclusions
Our data indicate that matrine-chitosan hydrogels signi cantly decrease the somatic cell counts and affect the structures of bacterial communities in the udder, especially the relative abundances of Corynebacterium_1, Aerococcus and Staphylococcus.These ndings show signi cant changes in metabolites and metabolic pathways as a result of intramammary MCH infusion, and some of the resulting 74 milk metabolites may be used as indicators of the response to MCH treatment.The results from the milk metabolic pathway analysis are promising for the investigation of matrine-chitosan's antimicrobial and anti-in ammatory properties that are closely associated with sphingolipid and glycerophospholipid metabolism.These insights into the complex mechanisms and corresponding biological responses highlight the bene cial action of matrine-chitosan hydrogels and justify continued investigations to identify the immunoregulatory mechanisms for treatment of mastitis dairy cows.

Preparation of matrine-loaded chitosan hydrogels
Matrine (98% purity) was purchased from Sigma-Aldrich Corp., St. Louis, MO, USA, and chitosan was purchased from Shanghai Sunny Biotech Co., Ltd., China.All solutions were prepared with nonpyrogenic products and materials under aseptic conditions in a laminar ow hood.The procedures were similar to those of Lanctôt et al. [25].Matrine (0.05 g) and 4 g of chitosan were added to 95 mL of deionized water and heated at 37 °C until the chitosan dissolved to obtain the matrine-loaded chitosan complexes.Ten ml of matrine-chitosan and 5 ml of 4% hydroxyethyl cellulose (2:1 ratio) were put into disposable plastic dishes (60 mm diameter) and mixed thoroughly.The water used was nonpyrogenic with < 0.005 endotoxin units/mL (Lonza, Walkersville, MD).For intramammary infusion, plastic syringes were lled with the 10 ml of MCH, sealed with a cap, and stored at room temperature.

Animals and experimental design
Animals selected in this study were provided by the Beijing Sunlow Livestock Dairy Farming Center (Beijing, China), our animal protocols were reviewed and accepted by the Animal Care Committee of Beijing University of Agriculture in concordance with the guidelines for the use of bovines in research studies of the SSTCC (The State Science and Technology Commission of the P.R. of China, 2017).The dairy farm owner was consent to collect samples from cows for the presnet study.Samplings of untreated milk (n = 580) were received from commercial dairies in Beijing from June to August 2019 with the California mastitis rapid detection reagent (CMT).Six mid-lactation cows with high somatic cell counts (SCC, average 250,000 cells/ml) indicative of subclinical-mastitis were selected according to their milk yield and parity.There were few initial differences in milk yield (27.2 ± 1.8 kg/d), DIM (114.6 ± 7.5 d), parity (2.6 ± 0.4), or body weight (BW; 670 ± 24 kg).The cows were fed a standard basal diet (table 4), had free access to water and were kept in a tie-stall barn.They received daily intramammary infusions of MCHs after each milking, twice daily, (7 AM and 7 PM) for 7 days.Milk samples were collected on the rst day (D 1) and the last injection day (D 7) and totaled 15 ml per animal, with about equal volumes from each lactating udder quarter.Somatic cell counts were determined using an automatic cell counter (DeLaval International AB, Tumba, Sweden), and the remaining milk was stored in liquid nitrogen for later analysis.After the period of experimental, all dairy cows were back to the herd of cows after the veterinary examination and ensure they were healthy.

Analytical bioinformatics of milk microbiotas
Analyses were conducted with FLASH version 1.2.11 and quantitative insights into microbial ecology (QIIME) version 1.9.1.These programs gave data similar to that published by Tong et al. [14].Sequences were assigned to taxa by BLASTing the ribosomal database project (RDP) dbase using a 97% identity cutoff.Operational taxonomic units (OTUs) were normalized to relative abundance and bacterial composition was determined by Majorbio I-Sanger.Within-sample diversity (α-diversity) was measured as bacterial community enrichment (ACE and Chao indices) and diversity (Shannon and Simpson indices) that were measured in a stochastic subset of the OTUs.Between-sample microbial diversity (β-diversity) was determined by phylogenetically-based weighted UniFrac distances [47].A more detailed picture of the diversity of the most abundant evolutionary clades in the bovine microbiota was obtained by ltering the OTUs to yield those with a relative abundance of ≥ 1% for at least one sample.

Milk metabolome determination
Milk samples were assessed by Majorbio Bio-Pharm using the LC-MS AB Sciex Triple TOF 5600TM (AB SCIEX, U.S.) according to published procedures [14] The liquid chromatography conditions were similar to those of a previous study [48].Quality controls were run by combining milk samples and injecting them periodically during experimental measurements.The results were analyzed with XCMS (ver.3.4.4).The retention times, MZ, observations and peak intensities were normalized with Excel.The differentially expressed metabolites were analyzed with the public database (https://metlin.scripps.edu/)on the Majorbio I-Sanger platform (www.i-sanger.com)and the KEGG pathway software for differential metabolite pro les (www.metaboanalyst.ca/).

Multivariate statistics
Comparisons were validated by Student's t and p < 0.05 was de ned as signi cant.Hierarchical clustering was conducted using the Bray-Curtis similarity index and the unweighted pair-group method with arithmetic averages.The SPSS software v.21.0 (IBM, Armonk, NY) was used.The α-diversity indices are given as mean ± SD.A p < 0.05 was considered statistically signi cant, and a p < 0.10 suggested a trend.Principal component analysis and orthogonal partial least-squares-discriminant analysis (OPLS-DA) were carried out to show the metabolism changes among the experimental groups after mean centering and unit variance scaling.Parameters with variable importance in the projection (VIP) values > 1.0 were allowed for group discriminant testing.Our OPLS-DA model was con rmed by 7-fold permutation testing.Signi cantly different metabolites among groups were assessed and identi ed by Wilcoxon rank-sum tests.
FXW wrote the manuscript.JJT, HZ, LSJ and BHX revised the paper.All authors carefully read the manuscript and agree to be held accountable for all aspects of the work.

Supplementary Files
This is a list of supplementary les associated with this preprint.Click to download. renamede61c7.docx

Figures
Figures

Figure 4 Principal
Figure 4

Table 2 .
Milk metabolites on day 1 compared to day 7 after MCH treatment

Table 3 .
Differences metabolites enriched from specific pathways in milk of dairy cows receiving MCH infusions (7 days).

Table 4 .
Components and nutritional analysis of the total mixed rations (dry matter