Koumiss and immunity: A thorough investigation of fermentation parameters and their impact on health benefits

The aim of this study was to determine the components and cytokine and immunoglobulin levels of koumiss during different fermentation periods, and to reveal the interrelation between these parameters. For achieving this objective, 10 samples of koumiss were prepared and randomly divided into 2 groups: the first group was sampled at 0, 1, 5, 12, and 24 h of incubation at room temperature for analysis. The second group was stored at +4°C, and samples were taken on d 5, 10, 15, and 20. The counts of Enterobacteriaceae spp., Staphylococcus , and Micrococcus spp. progressively decreased with the pe­ riod of fermentation until becoming undetectable in the final samples of both groups. We fond positive or nega­ tive correlations between cytokine and immunoglobulin levels and the physicochemical and microbiological parameters in the koumiss samples in both groups. How­ ever, the levels of IFN-γ, IL-2, TNF-α, and IgG did not change significantly over time in both groups. Overall, it is clear that traditionally prepared koumiss under different fermentation times and temperatures does not show any differences in cytokine and immunoglobulin concentrations.


INTRODUCTION
Koumiss is a drink fermented from mare milk that has been produced for at least 5,000 years (Oleinikova et al., 2024).Koumiss is a traditional drink that is more common among Central Asian countries, and it is known to have beneficial health and nutritional value (Aydemir Atasever et al., 2021;Rakhmanova et al., 2021;Istan bullugil et al., 2023b).Recent years have brought in vi tro evidence of the properties of koumiss for supporting healing from bacterial, viral, and parasitic diseases, as well as evidence of its contribution to strengthening the immune system (Istanbullugil and Atasever, 2019;Li et al., 2022;Ren et al., 2022Ren et al., , 2023;;Yan et al., 2022;Mar tuzzi et al., 2024).
It is estimated that 30 million people all over the world drink mare milk (Martuzzi et al., 2024).However, the production and market availability of koumiss is limited due to the low supply of mare milk (Rakhmanova et al., 2021).
The demand for koumiss and its industrialscale pro duction have both been on the rise for years, with the discovered potential health benefits being a subject of recent examination (Zhao et al., 2024).
Interferons were the first group of cytokines to be discovered and described as antivirals for viral infec tions.They are divided into 3 classes: type I, type II, and type III.Type I IFN mainly include IFN-α, IFN-β, IFN-ε, IFN-κ, and IFN-ω variants, and IFN-γ is the only known type II IFN.Type III IFN also have antiviral and immunomodulatory functions (de Weerd and Nguyen, 2012;Dembic, 2015).Unlike type I IFN, IFN-γ does not have a direct antiviral effect.However, it is reported to have important regulatory functions during the cellular immune response (Diker, 2005;de Weerd and Nguyen, 2012).Interleukin2 is expressed as a growth factor for T cells, but it is also reported to subsequently aid in the growth of many hematopoietic cells (Dembic, 2015).Its most important function is facilitating the exchange and development of helper and cytotoxic T lymphocytes, B lymphocytes, and natural killer cells.It stimulates mac rophage activation when it stimulates Th1 cells and helps antibody synthesis when it stimulates Th2 cells (Diker, 2005).One of the most important cytokines secreted by Th1 cells is IL-2, which is critical in stimulating cellular immunity against intracellular pathogens (Alnakip et al., 2014).Tumor necrosis factor (TNF) has 2 forms: TNF-α and TNF-β; TNF-α is mostly secreted from macrophages, whereas TNF-β is secreted from T cells.The spectrum of action of these 2 cytokines is similar.They promote apoptosis in tumors and virusinfected cells and have a systemic pyrogenic effect that causes high fever.How ever, overstimulation by these cytokines can result in severe tissue destruction and even shock (Diker, 2005).There is an important relationship between nutrition and the immune system.Kefir, koumiss, and other probiotics obtained by fermenting milk are recommended foods for a healthy life.It has been revealed that probiotics have many effects on the immune system due to the interaction of the molecules they produce with human cells (Tegin and Gönülalan, 2014).Some publications have focused on the cytokine levels of mare milk (Istanbullugil et al., 2023a,b).However, there are not enough studies on the cytokine levels of koumiss.Previous research on koumiss has reported its therapeutic potential on the digestive, cir culatory, nervous, and immune systems in humans (Rong et al., 2015;Tasdemir and Sanlier, 2020;Afzaal et al., 2021).These studies have mainly focused on probiotic strains isolated from koumiss and their immunomodula tory effects on the body.It is widely acknowledged that the health effects of these probiotic strains are multiple and speciesspecific.The Lactobacillus strains found in koumiss are potent inducers of IL-10, IL-12, and TNF-α (Foligné et al., 2006;Lu et al., 2008).
This study aims to determine the components and cy tokine and immunoglobulin levels of koumiss at different fermentation times and to reveal the relationship between these parameters.

MATERIALS AND METHODS
The study was conducted in the Department of Food Hygiene and Food Technology Laboratory, Faculty of Veterinary Medicine at KyrgyzstanTurkey Manas University (Bishkek, Kyrgyzstan).The experiment was carried out on 10 different samples of koumiss that had been prepared in the laboratory according to traditional methods (Istanbullugil, 2017).To prepare the koumiss samples, milk was collected between May and August 2022 from Kyrgyz mares ranging from 3 to 5 years old on the same farm with due care and feeding.The fresh milk used in the production of koumiss came from mares with clinically healthy udders and in good general condition.Before milking, udders were subjected to the California Mastitis Test (CMT) and milk was collected from CMTnegative animals (i.e., animals without sub clinical mastitis).After the mares' teats were wiped with a clean cloth, they were milked by hand.The mare milk was brought to the laboratory in cold chain within 2 h for the preparation of koumiss.Koumiss was purchased from the same facility as the mare milk, placed in ster ile plastic bottles inside cold boxes, and brought to the laboratory.

Experimental Koumiss Production
Once at the laboratory, the mare milk was allowed to reach the fermentation temperature of 20°C to 25°C.We added 10% to 15% koumiss to unpasteurized fresh mare milk and mixed thoroughly.The mixing process was re peated 3 to 5 times during the day.The samples prepared in this way were placed in 10 bottles.A total of 5 koumiss samples, designated as group 1, were incubated at room temperature (20°C-25°C), and samples were taken at 0, 1, 5, 12, and 24 h to analyze the components and cy tokine and immunoglobulin levels.The other 5 samples (group 2) were stored at +4°C, and samples were taken on d 5, 10, 15, and 20 for examination of components and cytokine and immunoglobulin levels (Figure 1).

Cytokine and Immunoglobulin Analysis
All the chemicals and reagents used in this study were of analytical reagent grade and purchased from com mercial sources.The levels of IFN-γ, IL-2, TNF-α, and IgG in the collected koumiss samples were determined using a commercial ELISA kit (Can Şahna and Rişvanlı, 2015) and read in an ELISA reader (BioTek Instruments, Winooski, VT).The catalog numbers of the commercial kits (all from USCN Life Science Inc., Wuhan, China) used were as follows: IFN-γ (catalog no.L220525434, SEA133Eq), IL-2 (catalog no.L220525459, SEA073Eq), TNF-α (catalog no.L220525477, SEA133Eq), and IgG (catalog no.L220525439, CEA544Eq).The ELISA test was carried out following the manufacturer's instruc tions, and after all steps were completed, cytokine levels were measured by reading the plates in an ELISA reader (BioTek Instruments) at 450 nm.
Various parameters were determined in the koumiss samples, including pH, titration acidity, ash, moisture, fat, alcohol content, and protein amounts.The pH was measured using a pH meter (Orion 3Star, Thermo Scientific, Waltham, MA; Tekinşen and Nizamlıoğlu, 2001).The titration acidity was determined according to the Association of Official Analytical Chemists method 947.05 (AOAC International, 2016) and expressed as the percentage of lactic acid (AOAC International, 2000).The amount of alcohol in koumiss was deter mined according to the method outlined by the Federal Agency for Technical Regulation and Metrology (Rus sian state standards) for dairy products, including kefir and koumiss (ГОСТ 3629-47).The amount of fat was determined by the Gerber method (the fat column was observed, which represents the percentage of fat in milk; AOAC 989.05), and the protein amount was determined by the micro Kjeldahl method (AOAC 939.02;AOAC International, 2000).The micro Kjeldahl procedure was executed in 3 steps: sample digestion, ammonia distil lation, and titration.Milk samples were digested with concentrated sulfuric acid, followed by distillation with a 40% NaOH solution.The ash amount was determined according to AOAC 942.05, and the moisture content was determined according to AOAC 948.12 (AOAC International, 2000).Samples were dried in an oven at 100°C for 2 h.

Microbiological Analysis
The total mesophilic aerobic bacteria count (TMAB) was determined by sowing the samples on plate count agar (1.05463, Merck, Darmstadt, Germany) using the cast plate technique, incubating at 32°C for 48 to 72 h and then counting the colonies (Halkman and Sağdaş, 2011).The Lactobacillus spp.count was determined by sowing on deMan, Rogosa, and Sharpe agar (Merck 1.10660) using the pour plate technique, incubating at 30°C for 72 h, and counting the colonies (Halkman and Sağdaş, 2011).The Enterobacteriaceae spp.count was determined by sowing on violet red bile on dextrose agar (Merck 1.10275) using the pour plate method, incubating at 37°C for 24 h, and counting the colonies (Halkman and Sağdaş, 2011).The yeast and mold count were determined by sowing on yeast extract glucose chloramphenicol agar (Merck 1.16000) us ing the cast plate method, incubating at 22°C for 5 d, and counting the colonies (Halkman and Sağdaş, 2011).The Staphylococcus and Micrococcus spp.count was deter mined by sowing on BairdParker agar (Merck 1.05406) with egg yolk tellurite emulsion (Merck 1.03785) added using the cast plate method, incubating at 37°C for 48 h, then identifying and counting the 1 to 3 mm diameter shiny black colonies (Bridson, 1998;Turkish Standards Institute, 2001).Finally, all the data obtained from the koumiss samples were compared based on hours and days.

Statistical Analysis
All the variables were loaded into SPSS 22.0 (IBM) for Windows, and descriptive statistics were computed.The normal distribution of the data was checked, and nonparametric tests were used because the data did not meet the assumptions of normal distribution.The Fried man twoway analysis of variance method was used to analyze the difference between the time and day groups for nonparametric dependent samples, and the Bonfer roni test was used to compare the subgroups.
The correlations between milk components and vari ables examined at different times and days were tested using the Spearman correlation analysis.A Pvalue less than 0.05 was considered statistically significant to indi cate a difference in the mean values.

RESULTS
The analysis results of the koumiss samples in the first group of the study, which were incubated at room temperature (20°C-25°C) for different durations, are summarized in Table 1.The pH values decreased as time progressed depending on adding the starter culture to the mare milk.The lowest pH was observed at 24 h (3.96 ± 0.07).The effect of incubation time on pH value was found to be significant (P < 0.007).
The analysis results of the parameters examined on different days in the koumiss samples kept at +4°C after adding the starter culture to mare milk (the second study group) are presented in Table 2.It was observed that the pH of the koumiss samples decreased with longer storage time.The lowest pH value (4.22 ± 0.02) was observed on the d 20.The effect of storage time on pH value was found to be significant (P < 0.013).
The analysis of koumiss samples which were kept for a different duration showed no significant differences in terms of IFN-γ, IL-2, TNF-α, and IgG levels, fat, ash, moisture, and protein content, and yeast numbers.How ever, Enterobacteriaceae spp.were below the detection limit in all the cases analyzed, and Staphylococcus and Micrococcus spp.were also below the detection limit on d 10, 15, and 20.
The analysis showed that the titration acidity ratio of the koumiss samples increased after the first hour and continued increasing with longer incubation times.The highest titration acidity was observed at 24 h (1.56 ± 0.01).The effect of incubation time on titration acidity was determined as significant (P < 0.001, Table 1).It was shown that the titration acidity ratio of the analyzed koumiss samples increased with longer storage time.The highest titration acidity was detected on d 20 (0.93 ± 0.04%).The effect of storage time on titration acidity was found to be significant (P < 0.017, Table 2).
The sample analysis showed an increase in the alco holic content of koumiss with the duration of incubation; the increase was relatively high at 24 h (1.81 ± 0.11%).We found a significant effect of the period of incubation on the alcoholic content (P < 0.001; Table 1).Alcohol concentration in the koumiss samples increased with time.The highest concentration of alcohol was recorded on d 20 (1.81 ± 0.10%), and the lowest concentration was observed on d 5 (0.06 ± 0.01%).The effect of storage time on the amount of alcohol was significant (P < 0.003, Table2).
The analysis of the koumiss samples between the time points from one to another indicated that the lowest con tent of protein took place at 24 h (2.12% ± 0.01).This was a very marked difference regarding the effect of incubation time on protein content (P < 0.011, Table 1).
It was analyzed and estimated that at the 0 h, the popu lation of Enterobacteriaceae spp. was at a maximum (4.95 ± 0.33 log cfu/mL), which further decreased with the increase in incubation time.Similarly, the effect of a-c Within a row, the difference between groups with different superscripts is significant (P < 0.05).
incubation time on the number of Enterobacteriaceae spp. was also significant (P < 0.011, Table 1).The analysis showed that the counts of Staphylococcus and Micrococcus spp.were highest at 0 h (3.83 ± 0.63 log cfu/mL) and then reduced to reach the undetectable level at 24 h in koumiss samples.The effect of incubation time on the counts of Staphylococcus and Micrococcus spp. was significant (P < 0.003, Table 1).The analysis of koumiss samples stored for different durations revealed the lowest TMAB count on d 20 (5.37 ± 0.28; P < 0.015, Table 2).Additionally, the lowest number of Lactobacillus spp. was observed on d 20 (4.73 ± 0.27 log cfu/mL; P < 0.01, Table 2).
Examining correlations between the properties ana lyzed at different time points revealed a negative correla tion between protein content at the 0th hour and fat and titration acidity, as well as a positive correlation between yeast numbers and protein content.
A positive correlation was observed in the first hour between protein content and Staphylococcus and Micrococcus spp.numbers, ash ratio and IL-2 levels, moisture content and titration acidity, TNF-α levels and titration acidity, Enterobacteriaceae spp.numbers and pH, as well as between IFN-γ and IL-2 levels.
A negative correlation was observed between titration acidity and fat content at 5 h, whereas a positive cor relation was found between ash ratio and IFN-γ levels, humidity ratio and Enterobacteriaceae spp.numbers, as well as between titration acidity and IFN-γ levels.
We identified a negative correlation at 12 h between pH and moisture and fat ratios, TNF-α level and titration acidity, IFN-γ level and protein ratio, moisture and ash ratio and pH, as well as between IL-2 level and ash ratio.
We observed a positive linear relationship with the amount of alcohol and titration acidity, moisture and fat ratios, IgG level, and titration acidity at 24 h.
Focusing on the correlation of the features analyzed, per day, it was established that the ash ratio versus al cohol gives a negative correlation on d 5, whereas the ratio of Staphylococcus versus Micrococcus spp.shows a positive correlation with the amount of ash.We observed a positive relationship between the number of Lactobacillus spp.and titration acidity on d 10.On d 15, a posi tive correlation was found between TNF-α levels and pH, Lactobacillus spp.number, alcohol content, TMAB, and yeast numbers, as well as IL-2 and IgG levels.Moreover, a negative correlation was observed between Lactobacillus spp.numbers and titration acidity.On d 20, a positive correlation was observed between pH and TMAB and yeast counts, and between TMAB numbers and moisture content, and a negative correlation was observed between TMAB numbers and alcohol content.

DISCUSSION
Koumiss contains all indispensable amino acids required in the human system, with an especially high content of proline, lysine, tyrosine, valine, and leucine.Thus, koumiss is potentially capable of fulfilling all hu man amino acid requirements (Ha et al., 2003).Koumiss has a lactose content that is variable between 1.4% and 4.5%.It includes the fatty essential acids, such as lin oleic and linolenic acid.Generally, mare milk contains a-c Within a row, the difference between groups with different superscripts is significant (P < 0.05).
According to the Turkish Food Codex, the amount of titration acidity in koumiss should not be less than 0.7, and the amount of ethanol should not be less than 0.5.In the first group, the acidity in titration was 0.67 ± 0.01, and the ethanol content was 0.53 ± 0.06 at 5 h.In the same group, at 12 h, the titration acidity was 0.85 ± 0.03, and the content of ethanol was 0.91 ± 0.03.It could be said that after 5 h, the formed koumiss in the first group conformed to the Turkish Food Codex about Fermented Dairy Products.In the second group, the titration acidity was 0.67 ± 0.07, and the ethanol content was 0.81 ± 0.07 on d 5. Based on this, it could be said that the koumiss produced in the second group meets the criteria of the Turkish Food Codex for fermented milk products after d 5.
The Kyrgyz Republic National Standard, under its Technical Regulation on the Processing of Milk and Dairy Products (Republic of Kyrgyzstan, 2013), stipulates 1% maximum milk fat content, 2.8% protein content, and 7.8% to 9.5% nonfat dry matter.In this research, the fat content of the milk varied from 0.91%, and protein content was 2.45% in all the groups analyzed.It has been determined that in the koumiss prepared in both groups complies with the National Standards of the Kyrgyz Re public Technical Regulation on Milk and Dairy Product Processing.
According to the Turkish Food Codex (2022), koumiss should contain a minimum of 10 7 cfu/g of the specific microorganism, the total number of additional microor ganisms indicated should not be less than 10 6 cfu/g, and the yeast content should not be less than 10 4 cfu/g.The study revealed that the general microorganism total and the specific total yeast count in the koumiss samples pre pared from raw mare milk in both groups meet the Turk ish Food Codex standards (Turkish Food Codex, 2022).
The present study showed a decrease in the pH of kou miss produced from mare milk after fermentation in both groups of the experiment.The time of fermentation had a very significant effect on the pH (P < 0.05).A decrease in pH was observed in all experimental groups, which was also in accordance with previous studies (Küçükçetin, 2003;Danova et al., 2005;Wang et al., 2011;Akpinar, 2015).The decreased level of pH is a common result of the fermentation process.
In the experimental koumiss samples prepared, the titration acidity increased until the end of the fermenta tion period.The effect of fermentation time on titration acidity was therefore significant (P < 0.05).This increase observed in the experimental groups was also reported in previous studies (Küçükçetin, 1999(Küçükçetin, , 2003;;Topuz, 2005;Wu et al., 2009;Tegin and Gönülalan, 2014;Akpinar, 2015;Macitova, 2016).The increase in titration acidity is due to the fermentation of lactic acid bacteria.
We found a significant effect of fermentation time on the amount of protein in the koumiss samples prepared in the first group (P < 0.05).Many researchers have also reported a decrease in protein during this fermentation process (Danova et al., 2005;Akpinar, 2015;Macitova, 2016).The decrease in protein content is thought to be due to proteolysis resulting from fermentation.However, we found no statistical relevance of change in the ratio of proteins in the second group (P > 0.05) because proteoly sis occurs much faster at room temperature.
The alcohol content developed gradually during the incubation time in both groups (P < 0.05).The increase in alcohol content with time in both groups supports the observation of several other previous works (Küçükçe tin, 1999;Karaçil and Tek, 2013;Macitova, 2016).This increase in the alcohol content results from the yeast fermentation in the starter culture (Mu et al., 2012;Te gin and Gönülalan, 2014;Macitova, 2016).It was found that the alcohol content in the samples of koumiss had increased due to the yeast fermentation process.
The results of the physicochemical analysis results showed that the pH ratio decreased, the titration acidity increased, and the amount of alcohol increased as fermen tation time increased in both groups of koumiss samples experimentally prepared from mare milk.As shown in Tables 1 and 2, it is noticeable that these changes were greater in the first group.This excess is due to the fact that the starter culture flora at room temperature pro gresses more rapidly and, therefore, its activity exceeds the desired level.
Chaves -López et al., (2011) studied Colombian koumiss derived from cow milk and found lactic acid bacteria and yeast levels ranging between 7.05 and 9.53 log10 cfu/mL and 6.26 to 8.65 log10 cfu/mL, respectively.However, in this study, the lactic acid bacteria and yeast counts were at the same level as the koumiss samples that were prepared.This difference might have resulted from the use of mare milk, different incubation temperatures, or the initial culture used in fermentation.Mu et al., (2012) reported that koumiss samples in China contained yeast at rates of 5 to 7 log10 cfu/mL, which was similar to the yeastmold counts observed in the koumiss samples prepared in this study.
This study found that the numbers of Staphylococcus and Micrococcus spp. in both groups decreased dur ing the incubation period, with a value of 0.31 ± 0.27 log cfu/mL at 12 h of incubation and falling below the detectable level in the analyses performed at 24 h.On the other hand, the number of Enterobacteriaceae spp.for the first group decreased below the detectable level after 12 h, whereas in the second group, it was under the detectable level in all the tested time periods.The presence of Staphylococcus and Micrococcus spp.and Enterobacteriaceae spp. in koumiss is thought to be due to the traditional use of unpasteurized mare milk in pro duction.Failure to adhere to hygienic conditions during milking, such as for the udder of the animal, the hands of the milking person, and the containers and strainers being used, might contaminate the mare milk.Com petitive attributes of the starter cultures, which conduct fermentation, bring about the decrease in the count of these microorganisms during the incubation period and, after that, their suppression.Similar results have been reported in many studies on koumiss (Yinfeng et al., 2002;Zhang et al., 2007;Wang et al., 2011;Chen et al., 2015Chen et al., , 2017;;Istanbullugil, 2017).Additionally, studies have shown that other fermented foods, such as yogurt and kefir, inhibit pathogenic bacteria (Bachrouri et al., 2002;Gülmez and Guven, 2003).Istanbullugil et al. (2023b) found positive correlations between IL-5 level and protein ratio on the 10th postnatal day, between IL-2 level and fat ratio on the 20th day, and between IL-2 level and fat-free dry matter ratio on the 20th day in mare milk.In the same study, cytokine levels in milk, foal, and mare sera were compared, and it was determined that the highest IL-2 level was detected in the milk samples 0, 5, 10, and 20 d postpartum, with the highest value of IL-2 in the milk samples on d 20 (17.97 ± 24.88 pg/mL).Additionally, the highest IFN-γ levels were detected in postpartum milk samples on d 5 (45.31 ± 6.98 pg/mL) and d 10 (67.0.8 ± 9.72 pg/mL), and the highest TNF-α level was found in milk samples on d 5 (83.77 ± 13.86 pg/mL) and d 10 (99.67 ± 3.12 pg/mL).Despite this, no publications were found on the correlations between koumiss in cytokine levels and its composition.The present study found no difference be tween IFN-γ, IL-2, TNF-α, and IgG levels of koumiss samples analyzed at different hours and days.

CONCLUSIONS
The results of the study showed that the koumiss samples' decrease in pH ratio, with the increase in titra tion acidity and alcohol content during the fermentation period, was observed to have an inhibitory effect on pathogenic microorganisms that may be present in mare milk.In both groups, the numbers of Staphylococcus and Micrococcus spp.decreased below the detectable level during the fermentation period, indicating that koumiss, which is formed as a result of fermentation, is safer for consumption than unpasteurized mare milk.The cyto kine and immunoglobulin levels in the koumiss samples remained unchanged during the fermentation period.Incubation at 20 to 25°C resulted in faster fermentation and earlier formation of koumiss, whereas incubation at +4°C prolonged the shelf life of the product, making it suitable for commercialization and contributing to the country's economy.However, it was concluded that it would be beneficial to support the obtained results with more comprehensive studies and projects.

NOTES
This study was supported the Kyrgyzstan Turkish Ma nas University (Bishkek, Kyrgyzstan) Scientific Research Projects Unit (project no.KTMUBAP2022.FB.04).No human or animal subjects were used, so this analysis did not require approval by an Institutional Animal Care and Use Committee or Institutional Review Board.The au thors have not stated any conflicts of interest.
Figure 1.Preparation steps for the koumiss samples.