Sub-chronic, low dose co-exposure to Aflatoxin B1 and Microcystin-LR in C57BL/6 mice significantly alters the cytokine response in serum and liver

ABSTRACT In this study, we examined the expression of 11 cytokines in liver and serum of C57BL/6 mice upon single or co-exposed to Aflatoxin B1 and Microcystin-LR at low-dose and sub-chronic conditions. The objective is to elucidate the potential combined hepatotoxic effects induced by AFB1+MC-LR, and we aim to provide references for further immune impacts in the liver. Luminex-based cytokine examination was applied, and we found that most of the cytokine expression increased with time. The level of IL-18, GM-CSF, IFN-γ, IL-1β and IL-4 in the liver showed a significant rise. However, some cytokines (IL-13) showed an upward trend after being suppressed during the middle stage of exposure. The level of IFN-γ and IL-12p70 generally represented the immune response in the combined exposure group. Conclusionally, there are many uncertainties due to sub-chronic and low-dose exposure. Therefore, the occurrence of subsequent irreversible liver injury and liver cancer is a probabilistic event.


Introduction
Approximately 600,000 people worldwide died of hepatic cell carcinoma (HCC) annually (Gomaa et al., 2008;Kucukcakan & Hayrulai-Musliu, 2015). The mortality rate of HCC is as high as 14.56% in China (Edite Bezerra da Rocha et al., 2014;Palencia et al., 2010). Factors that have been proven to induce HCC are hepatitis B virus infection, unhealthy lifestyle (e.g. smoking, alcoholism, etc.) and dietary exposure to various mycotoxins. Aflatoxins have been extensively linked to growth stunting during early childhood. Specifically, aflatoxin B 1 (AFB 1 ) acts synergistically with hepatitis B virus infection and contributes to the occurrence of human primary liver cancer (Misihairabgwi et al., 2017). Noticeably, it is often overlooked that the microcystins (MCs) may accumulate in aquatic species at high concentrations and the presence of dissolved MCs in the source of drinking water is often loosely tracked (Song et al., 2007).
Moreover, the study found increased genotoxicity in hepatic cell lines followed by the combined exposure of AFB 1 +MC-LR (Liu et al., 2018). Therefore, it can be speculated that coexposure to grain-oil contaminated with Aflatoxin B 1 (AFB 1 ); drinking water, aquatic products contaminated with Microcystin LR (MC-LR) can cause potential combined toxic effects and may lead to subsequent liver injury thus accelerate the outcome of HCC.
The co-expression of cytokines, chemokines and growth factors plays a crucial role in liver injury and immune regulations of the body, as they participate in the whole process of liver inflammation, subsequent liver injury, irreversible liver cirrhosis and eventually the outcome of liver cancer (Girón-González et al., 2004;Nakagaki et al., 2018). Related reports have shown that single exposure to AFB 1 or MC-LR may lead to alternations in the expression of proinflammatory cytokines such as IFN-γ, IFN-α, IL-6, TNF-α, and TNF-β in the liver (Bird et al., 1990;Lacour et al., 2005;Lu et al., 2014) and accelerate the formation of platelet activating factor and activated cyclooxygenase, thereby promoting inflammation (Tlsty, 2001). During these processes, the roles of AFB 1 act on liver immunity are not only pro-inflammatory, but alternations of exacerbation and inhibition of immune response also occur (Chung et al., 1998). These phenomena, as some studies speculated, were caused by coupling of different exposure doses and different immune responses at different times (Almeida et al., 1996;Ishikawa et al., 2017). However, further related reports are lacking. Research elaborated on co-exposure to AFB 1 +MC-LR inducing liver inflammation or liver injury based on cytokine expression has not been reported extensively.
Based on these backgrounds, we mainly examined the expression of 11 key cytokines (including IL-6, etc.) in the liver and serum of C57BL/6 male wild-type mice induced by single or co-exposed to AFB 1 +MC-LR under low-dose and sub-chronic conditions. By combining with Principal Component Analysis to demonstrate the strength and trend of combined effects (antagonistic or synergistic) of AFB 1 +MC-LR on liver injury, thus will provide support and reference for further revealing the reversibility and regularity of the outcome of coexposure to AFB 1 +MC-LR.

AFB 1 and MC-LR standards
AFB 1 (>90%, analytical purity) was purchased from Sigma Aldrich, MC-LR (>90%, analytical purity) were purchased from Tianjin First standard, and the standards were provided in powder. The maximum gavage volume is 0.4 mL for each mouse per administration. AFB 1 and MC-LR were dissolved in tricaprylin, respectively, as previously described (Jeannot et al., 2012). The concentration was reasonably calculated according to daily administration concentration, the maximum volume can be administered according to mouse body weight. The control group (blank) was only gavage with the same volume of tricaprylin. Meanwhile, MC-LR is easy to degrade and needs to be reconstituted once a week. Both AFB 1 and MC-LR solutions were kept away from light in the refrigerator at 4°C.

Animals
Because male mouse is more sensitive to hepatotoxic cytokine expression than female, this study took C57BL/6 wildtype male mouse aged 3 weeks as the experiment subject. The weight is 17.5 ± 3.5 g, and the difference is no more than 20%.

Experimental dose design
To simulate the dietary exposure scenario as much as possible, a previous study designed the dose of AFB 1 by oral gavage in mouse at 44 µg/kg.bw with reference to the maximum residue level (MRL) of AFB 1 in grains and grain-oil, taking into consideration the human/mouse body surface area conversion (Jha et al., 2013). The determination of the dose ratio of AFB 1 and MC-LR follows the combined effect evaluation model -Median-effect Principle of the Nonconstant Ratio Model (Chou et al., 2005;Chou, 2006). The ToxNet database system showed that oral exposure LD 50 of AFB 1 in mouse is 9.0-9.5 mg/kg.bw (Bedard et al., 2005), while the MC-LR is 5.0 mg/kg.bw (Mrdjen et al., 2018). Therefore, in this study, we followed AFB 1 :MC-LR= LD50: LD50 = 9:5 to set the combined ratio, and the final dose of AFB 1 was 44 µg/kg.bw, and MC-LR was 24 µg/kg.bw. Four groups were set up, including control group, AFB 1 single group, MC-LR single group and AFB 1 +MC-LR co-exposure group (AFB 1 : 22 µg/kg.bw + MC-LR: 12 µg/kg.bw), with a total of six time points, the blank group was set to 21 mice, and the average number of three experimental groups was 33 mice, with a total of 120 mice, administered orally once a day. After each administration, whole cage animals were being monitored, and clinical symptoms were recorded.

Animal experiments
Three-week C57BL/6 mice were chosen and were uniquely marked. The experiment started (week 0) after the mice were pre-raised for 2 weeks. Sampling at 3, 5, 7, 9, 11 and 13 weeks refer to the third week of continuous administration after 3-week-old mice were pre-raised for 2 weeks, that is, the corresponding mice were 8, 10, 12, 14, 16 and 18 weeks old, respectively. The administration was by oral gavage between 7:30 and 9:30 am. The animals were euthanized upon isoflurane anesthesia at weeks 3, 5, 7, 9, 11 and 13 accordingly. Blood was collected into the EP tube via cardiac puncture (also between 7:30 and 9:30 in the morning) and centrifuged (volume of blood about 0.8-1 ml), and then the animals were fixed with pins on the dissection tray. The abdominal cavity was cut open and flushed with saline. The liver tissue was taken out using a tweezer together with intact marginal membranous structure and gallbladder, which were then placed in a dry EP tube. Samples were frozen at −80°C, and after sample collection at the 13th week, all were processed at one time and tested on the computer.

Luminex-based cytokine examination
Fresh liver and serum sample dilution, standard preparation and calibration curve, sample microsphere and antibody incubation, etc., were completed according to the instruction of the cytokine detection kit. Protein quantification kit was used to quantify protein in the supernatant before the on-machine test to determine whether the sample concentration met the determination requirements. Samples that met the concentration requirements were tested on a liquid chip instrument. Results are based on a five-parameter nonlinear regression fitting standard curve, including scale, heat map, median fluorescence intensity and concentration.

Statistical analysis
The significant differences in the expression level of 11 cytokines in the liver and serum of single and co-exposed to AFB 1 +MC-LR mice were all based on Tukey test of paired comparison, where p < 0.05 was considered statistically significant. The results indicated by lowercase letters (a, b, c, d, etc.), and the above statistical analysis was all performed using Origin 2019b software (OriginLab Corporation, USA) and Microsoft Excel.

Differential analysis of cytokines in serum and liver induced by single and mixtures of AFB 1 and MC-LR at different times
The overall trend of expression of IFN-γ, IL-12p70, IL-13, IL-1β, IL-2, IL-4, IL-5, IL-6, TNF-α, GM-CSF and IL-18 in serum and liver of C57BL/6 mice based on results obtained at week 0, 3, 5, 7, 9, 11, and 13: (1) The expression of majority cytokines increased with the exposure time (Tables 1-6); (2) The expression levels of IL-18, GM-CSF, IFN-γ, IL-1β and IL-4 in the liver and their rising trends were significant, and the cytokine storm was typically characterized by the high expression of IL-18 in the liver; (3) During the 5th to 11th week of exposure, the expression of some cytokines decreased. For example, the level of IL-13 in the liver of combined exposure group decreased with the prolongation of exposure time and the increase in exposure frequency: from 18.14 ± 5.75 pg/ml (Wk 3) to 12.21 ± 6.61 pg/mg (Wk 5) and to 11.16 ± 1.23 pg/mg (Wk 7), and then gradually increased to 18.65 ± 3.65 pg/mg (Wk 9) until 43.03 ± 10.29 pg/mg (Wk 13). Similarly, IL-2 also had this trend (Table 6).

Evaluation of the potential combined effects of cytokines in serum and liver
Based on principal component analysis (PCA), assuming 11 cytokines were linearly correlated in the time series, the comprehensive scores of principal components of different exposure groups were obtained by dimensionality reduction principal component analysis to determine the combined effects and the trend of single and co-exposure group (Tables 7 and 8 and Figures 1-2): (1) The main components (PC1 and PC2) in serum and liver are IFN-γ and IL-12p70, respectively. Their expression levels can approximately represent the overall cytokine expression of the samples; (2) The dose ratio of AFB 1 and MC-LR was set to satisfy the unequal-ratio model of the median effect equation, therefore theoretically the expression level of AFB 1 and MC-LR in single groups may be consistent, while in serum the expression level (comprehensive principal component score) of AFB 1 single group was the highest overall at week 3, 5, 9 and 13, followed by co-exposure group and the MC-LR single group. Therefore, the combined effects generally drifted between additive, weak synergistic or weak antagonistic effects. The specific trend was as follows: Wk 3: additive -Wk 5: Additive -Wk 7: slight synergism/synergism -Wk 9: slight antagonism/antagonism -Wk 11: slight synergism/synergsim -Wk 13: slight antagonism/antagonism (Table 7 and Figure 1); (3) The specific trend of liver expression was as follows: Wk 3: additive -Wk 5: slight synergism/synergism -Wk 7: slight synergism/synergism -Wk 9: additive -Wk 11: additive -Wk 13: slight antagonism/ antagonism; (4) In terms of overall expression trend, with the increase in exposure time and the number of administrations, the combined exposure group entered the plateau period during week 3-5: the additive effect, weak antagonism or weak synergism were alternated. However, there were subtle differences in the combined effects of the liver and serum, that is, the cytokines in the liver showed synergistic or weak synergistic effects at week 7 and week 9, while the serum showed synergistic or weak synergistic effects at week 5 and week 7. This phenotype possessed a time difference in the dose-response relationship ( Figure 1).

Discussion
During the process of liver injury, inflammatory reaction, subsequent irreversible liver cirrhosis and development of liver cancer, one of the typical features is the increase in concentration of cytokines in the liver and serum, such as IL-6, TNF-α, TNF-β, etc.. Related research has shown that mouse fed with AFB 1 at a dose of 0.01-1.6 mg/kg.bw, once every 4 weeks, may induce a significant immune response after 12 weeks at 0.4 mg/kg.bw and 1.6 mg/kg.bw dose levels, respectively (Bird et al., 1990). However, studies have also shown that the intragastric dose of AFB 1 as low as 5-75 μg/ kg.bw or intraperitoneal injection at a dose of 25-50 μg/kg. bw can initiate immune response and the expression of cytokines such as IL-4 and others after 4 weeks, which was positively correlated with the exposure time (Watzl et al., 1999). Meanwhile, MC-LR can induce the expression of IL-1β, IL-6 and TNF-α (Hinton et al., 2003). These were confirmed in our study. Regardless of the single or co-exposure group, the exact expression profile was inconsistent; nevertheless, the expression level of majority cytokines increased significantly with the prolongation of exposure time and administration frequency, including IFN-γ, IL-12p70, IL-18, etc. However, some cytokines, including IL-6, TNF-α, etc., were not highly expressed.
Both the single and co-exposure group were characterized by a cytokine storm of IL-18 expression, especially the expression level of that in the liver was significantly higher than that of the serum. IL-18 is a pro-inflammatory cytokine that can stimulate the production of vascular endothelial growth factor VEGF and fibroblast growth factor FGF, therefore has a strong pro-angiogenesis effect. It is a key regulator during tumor immune escape. In addition, IL-18 is a proinflammatory cytokine that promotes type 1 immune response and plays an important role in the differentiation of T cells as well as the process of mediating immune responses. It can also stimulate Th1-directed polarization and, therefore, increase IFN-γ expression (Qian et al., 2014). Noticeably, without the presence of IL-12p70, IL-18 can prime NK cells, Th1 cells, etc. In synergy with IL-2, this will lead to production of IL-3, IL-9, and IL-13. In addition, IL-18, together with IL-3, may stimulate secretion of IL-4, IL-13, etc., by mast cells and basophils (Yasuda et al., 2019). IL-18 plays Table 1. Expression of 11 cytokines in serum of mice exposure to AFB 1 at different times.
Cytokine concentration (pg/ml) (X ± SD), n=3 Time ( 1. ≤0 means that the detected MFI is not within the range of the standard curve or the concentration value cannot be calculated; 2. Difference analysis is the comparison of the expression levels of the same cytokines at different times. The statistical method adopts the Turkey test of paired comparison. Significant differences (p < 0.05) are indicated by lowercase letters. 1. ≤0 significa que el MFI detectado no está dentro del rango de la curva estándar o que no se puede calcular el valor de la concentración. 2. El análisis de diferencias es la comparación de los niveles de expresión de las mismas citocinas en diferentes momentos. El método estadístico adopta la prueba de Tukey de comparación por pares. Las diferencias significativas (p<0,05) se indican con letras minúsculas. a role in acute liver injury, chronic inflammation, and autoimmune disease. Its performance and mechanism of action may be different, as it plays a dual regulatory role in inflammatory response (Debets et al., 2000). Studies have shown that the serum level of IL-18 in HBV+ HCC patients (n = 15374.40 ± 45.20 ng/L) was significantly lower than that in the control group (n = 165112.48 ± 68.65ng/L; p < 0.001) (Ghayur et al., 1997). In this study, the serum and liver coexposed groups exhibited additive, weak antagonistic and antagonistic effects in the early and middle exposure periods; weak synergistic or synergistic effects in the middle and late periods; and weak synergistic and antagonistic effects in the later period. Due to the unique dual regulatory role of IL-18, it is possible that high expression of which during Table 7. Principal component analysis of cytokines in serum and liver of mice, AFB1+MC-LR single and combined exposure groups at different times.
Tabla 7. Análisis de los componentes principales de las citocinas en el suero y el hígado de ratones; grupos de exposición única y combinada a AFB1+MC-LR en diferentes momentos. network interaction of immune response as an upstream event of T cell differentiation into Th1 cells assists in effective immune response initially, and is also involved in subsequent quenching of immune response. Ultimately, the outcome of alternations between additive, weak antagonistic and weakly synergistic effects was formed. INF-γ and IL-12p70 were the main component factors. Among which, IFN-γ is one of the most important mediators of inflammatory response. Unlike INF-α and INF-β, INF-γ binds different receptors. This is partially explained by the stimulation of the secretion of INF-γ by mediators IL-18 and IL-12p70, which can synergistically or individually maintain and enhance the production of INF-γ before the early adaptive immune response (including innate and acquired immunity). In this study, the overexpression of IL-18 in the liver and serum of mice induced by single and co-exposure of toxin may participate in stimulating the expression of INF-γ, which became a key inducer in the upstream of the signaling network, and thus, the production of INF-γ can be regulated. The expression of cytokines such as IL-4, IL-10, TGF-β, etc., can therefore be characterized as downstream events.
The conclusion of this study also confirmed and suggested the possibility of this hypothesis. During the early stage of exposure (up to 9 weeks), the expression level of IL-4 and others were relatively low and the apparent upward trend until the 9th week suggested that AFB 1 +MC-LR initially induced cytokines such as IL-18, the expression of which led to T cells' polarization into Th1 subtype. As a result, the pro-inflammatory response occurred, and then, the positive immune response shifted T cells' differentiation to Th2 subtype, thus the anti-inflammatory effects gradually became dominated, as indicated by the expression of cytokines such as IL-4. The shift of Th1/Th2 accompanied the entire process. Study found that C57BL/6 mice administered with 442 µg/ kg.bw and 663 µg/kg.bw of AFB 1 exhibited liver damage, mainly hepatocyte necrosis and the increase in megakaryocyte compared to the dose at 44 µg/kg.bw. The expression of IL-4, IFN-γ and body weight of mouse treated with the highest dose of AFB 1 were significantly increased (p = 0.002), but the expression of IL-17 cytokine was not significantly (p > 0.05) different compared with control group (Jha et al., 2013). IFN-γ mainly functions through the JAK-STAT    2. Mecanismo hipotético de la respuesta inmunitaria inducida por AFB1+MC-LR, que muestra la activación de la vía Th1 de diferenciación de las células T en la fase inicial de la respuesta y el proceso de sustitución de la inmunosupresión en la fase posterior. signaling pathway, but activation or inhibition of MAP kinase, PI3-K, and NF-κB pathways can also induce different regulatory mechanisms for immune responses, depending on cell types and its microenvironment.
Similarly, IL-12p70 is also a key regulatory cytokine found in this study, which mainly induces inflammation and Th1directed polarization. In response to IL-12p70, NK cells release a certain amount of IFN-γ, which then involved in Th1-mediated immunity, including promoting T cell polarization into Th1 subtype, increasing the secretion of IL-12p70 and further enhancing the type 1 immune response (Gracie et al., 1999). Therefore, AFB 1 +MC-LR was likely to stimulate NK cells to induce INF-γ expression during early adaptive immune response (pro-inflammatory response), promoting T cell polarization into Th1 subtype which acts as an upstream signal to subsequently regulate the secretion of IL-12p70 by DC cells. This process may become a key molecular event in the regulation showing a weak antagonistic effect during the period of pre-exposure to AFB 1 +MC-LR. With the accumulation of exposure dose (due to the superposition of exposure frequency) and exposure time (Wk 7-13), the ability of acquired immune response was enhanced, Th2-directed polarization became dominant and even additive or weak synergistic effect became the main molecular event. Studies have shown that the effect of AFB 1 on the immune system may be stimulatory or inhibitory, depending on coupling of critical point for exposure dose and time (Bao et al., 2015); thus, the differences of consistency lay between "time-effect" and "quantity-effect". The coupling of exposure time and dose is one of the most critical prerequisites for the different combined effects of AFB 1 +MC-LR.
In addition, AFB 1 +MC-LR single and co-exposure group did not show significant differences in the expression of certain cytokines, including TNF-α and IL-6, as the exposure time and frequency increased. The study also showed that 10-week intraperitoneal injection of 1 mg/kg AFB 1 male SD rats compared with control group (only tricaprylin injection), the expression of TNF-α was two times higher than that of the blank (P < 0.01) and the IL-1α was three times higher than that of the blank (P < 0.01). TNF-α plays an important role during the occurrence and development of tumor. It has been reported that rats lacking this factor can avoid the occurrence of tumor (Kalinski et al., 2013). The lack of TNF-α/ IL-1α cytokine can inhibit the rapid proliferation of liver cancer cells, while the induction of ROS through the TNFα/IL-1α pathway is one of the main mechanisms by which AFB 1 promotes tumor formation in rats (Chung et al., 1998). In this study, regardless of AFB 1 +MC-LR single group or the co-exposed group, the expression of TNF-α both showed an upward trend; however, it was not significant. This indicated that mouse macrophages may still play an active role in the positive regulation of immune response under the coupling conditions of conducted dose level and exposure time. IL-6 family cytokines also have positive and negative regulatory functions, the high expression of which in the liver indicates the occurrence of liver injury, mediated by Kupffer cells in the liver in a paracrine manner. The anti-inflammatory activity was mainly achieved through membrane binding of IL-6 R canonical signal transduction (Qin et al., 2016). Studies have shown that MC-LR can affect the expression of IL-2 and IL-6 in human (Lankoff et al., 2004). Mice exposed to MC-LR at a dose of 7, 12, 24 and 36 mg/kg.bw both resulted in a significantly reduced mRNA level of TNF-α, IL-1β, IL-4, IL-2, and IL-10 (Th1/Th2-related cytokines) after 8 h, while IL-6 expression did not alter significantly (Cornell et al., 1990). In addition, IL-6 in peripheral blood can be induced by TNF-α and IL-1β (Rose-John, 2018); thus, the level of IL-6 can also indirectly reflect the expression of these factors. In this study, the expression level of IL-6 and TNF-α did not increase significantly. This was in correlation with the previous study (Shi et al., 2004), suggesting that induction of IL-6 may be a downstream event of TNF-α, which led the antiinflammatory process during the early stages of immune response.

Conclusion
As shown in Figure 2, which summarizes and reviews the mechanistic hypothesis of the combined effects of cytokine immune response regulatory network. With the hypothesis of Th1/Th2 balance as the core mechanism, IL-18, IFN-γ and IL-4 participated in the complex network related to immune response and hepatotoxicity induced by AFB 1 +MC-LR mixtures.
The detailed conclusions are as follows: (1) AFB 1 +MC-LR induced inflammatory response initiated starting the early stage (0 5Wks). IFN-γ receptor signaling is activated by the restricted transcription factors IL-12 R, T-bet, etc., which enhances the Th1 phenotype and induces IFN-γ, IL-2 and other cytokine expression. IFN-γ also induces a macrophageactivating factor and upregulates the expression of MHC class I and MHC class II molecules as well as co-stimulatory molecules to promote antigen presentation, while IL-18 expression as an upstream event promotes Th1-directed differentiation and induces differentiated Th1 cells producing IFN-γ. This process inhibits the combined effects and manifests as a weak antagonistic or antagonistic effect (or additive effect). (2) During the early and middle stages (5 -7Wks), with the increase in exposure time, dose and frequency, the antigen level increased, immune response entered a plateau drift phase, weak antagonistic or antagonistic effect gradually shifted to an additive effect; (3) In the middle and late stages (7-11Wks), the "time-quantity" exposure of AFB 1 +MC-LR further enhanced the oppression, initiated Th2 mediated immune response, controlled the excessive inflammatory response. IL-4 in the middle and late stages significantly increased in expression was a key molecular event or an outcome of Th2 mediated immune response. C-Maf and others help T cells differentiate into Th2 phenotype by promoting IL-4 and inhibiting IFN-γ. IL-4 can also act as a precursor to induce Th2 differentiation and inhibit excessive immune inflammation. Conclusionally, AFB 1 +MC-LR exhibited a weak synergistic or synergistic combined effect during middle and late stages; (4) In the later stage (11-13Wks), the immune response induced by AFB 1 + MC-LR was enhanced, given the overall low-dose condition and could be described as a process of repeated alternation, thus may lead to liver inflammatory response, the possible outcome of system collapse, characterized as irreversible liver damage and eventually hepatocellular carcinoma. There are many uncertainties during the whole process due to low doses and dietary scenarios of subchronic exposure. Therefore, irreversible liver damage and eventually the occurrence of hepatocellular carcinoma is a probabilistic event.

Disclosure statement
No potential conflict of interest was reported by the author(s).

Funding
This study is supported by the National Natural Science Foundation of China [No.31972187]

Data availability statement
Data are available within the article or can be obtained from the authors.