In view of the increase of antibiotic resistance in animal husbandry, the ongoing restriction of antimicrobials and growing public skepticism about the use of antibiotics and other medications, interest in alternatives of plant origin is steadily increasing ^[1]. Phytogenic feed additives (PFAs) are defined as substances extracted from plants by technical processes (e.g., steam distillation or cold pressing) and integrated into the diets of livestock in order to improve their performance and health, among other benefits ^[2]. Whether antimicrobial, antioxidant or performance-enhancing effects, numerous in vitro studies have demonstrated a positive impact of various bioactive herbal additives in pigs and chickens, which is why a wide range of different PFA are already used in animal nutrition today [3,4,5,6,7]^{[3][4][5][6][7]}. Evidence of efficacy is primarily provided by in vivo feeding trials, but there is a growing interest in the use of different in vitro methods in the development of PFA. Although in vitro methods, unlike animal studies, cannot map the complete metabolism, they allow the determination of several defined parameters such as barrier function ^[8] or antioxidant potential ^[9] and are useful to study host–pathogen interactions ^[10] much more efficiently, cost-effectively and in line with 6R principles ^[11]. The range of available in vitro models is large and constantly expanding. In addition to other methods such as chemical assays to analyze antioxidant potential ^[9] or microbial assays such as the broth microdilution method to determine inhibitory potential on microorganisms [12^[12][13][14],13,14], the use of different cell lines enables comprehensive studies to investigate new potential feed additives. Since intestinal epithelial cells are in direct contact with the digesta, displaying their interaction with PFA is of great interest. Cultivated epithelial cell lines such as the IPEC-J2 cell line are important in vitro tools for this purpose.

2. Intestinal Porcine Epithelial Cell Lines

Several different established intestinal porcine epithelial cell lines have already been used to investigate the influence of PFA on the epithelium of pigs. IPEC-1 cells were isolated from the ileum and jejunum and IPEC-J2 cells from the mid-jejunum of an unsuckled piglet less than 12 h old ^[15]. Both cell lines, as with many other common cell lines such as the human Caco-2 cell line, are immortalized, but non-cancerous and non-transformed, being morphologically and functionally closer to the epithelium of living pigs. In contrast, transformed cell lines are considered to be less vulnerable to stress and cytotoxicity ^[16]. Beside the lacking ability to form a polarized monolayer, that probably contributes to the fact, that the intestinal porcine epithelial cell line IPI-2I which was isolated from the ileum of an adult boar ^[17] have not been used for studying PFA so far. The cell line ZYM-SIEC02, similarly to IPEC-1 and IPEC-J2 cells, was isolated from the small intestine of neonatal unsuckled piglets ^[18], but its use is still limited to a small number of studies. This also applies to the cell line PSI cl 1 which was derived from the small intestine of an adult pig ^[19]. Thus, these cell lines, as well as the IPI-2I cell line, will not be discussed further in this revisewarch. When cultured, IPEC-J2 and IPEC-1 form a polarized monolayer with tight junctions (TJ), and microvilli located on their apical side ^[20] and moreover develop extensive metabolic functions. Those features allow measurements of barrier function and nutrient transport ^[21] as well as investigations on the impact of external challenges as mycotoxins ^[22], microbial pathogens ^[10], and their toxins [23^[23][24],24], heavy metals such as zinc ^[25] and copper ^[26]. Gene expression analysis of both cell lines revealed that IPEC-J2 is higher differentiated in morphology e.g., developing a two-fold higher number of microvilli and villin, as compared to IPEC-1, which indicates a more efficient crosslinking between microvilli ^[21]. Furthermore, whereas IPEC-1 cells seem to be in a non-proliferative state, IPEC-J2 cells show an active aerobic and anaerobic glucose-consumption characterized by a much higher rate of oxidative phosphorylation, intracellular ATP-content, O2- and glucose-consumption as well as lactate production and are therefore more suitable for metabolism-focused investigations ^[21].

2.1. Methods to Study Barrier Function in IPEC

The barrier function is one of the most important basic properties of the intestinal epithelium due to its direct contact to the intestinal lumen and the digesta contained therein. Damage to the epithelium integrity alters its permeability, which in normal conditions ensures selective passage of essential substances such as nutrients. As a consequence, microorganisms or their toxins could cross the epithelium and cause local or systemic damage to the host. Some exhaustive reviews about the influence of food components and PFA on intestinal barrier function in animal experiments but also ex vivo and in vitro have recently been published [8,27]^[8][27]. Various methods are available to determine the integrity of the barrier function. One important measurement is the cell viability, i.e., the intactness of the enterocytes, which is commonly measured by 3-[4, 5-dimethylthiazole-2-yl]-2, 5-diphenyltetrazolium bromide (MTT) reduction assay, neutral red method or lactate dehydrogenase release. Moreover, propidium iodide is often used for cell viability assessment, not only for epithelial cells but also for bacteria viability [24,28]^[24][28]. In addition, the electrical measurement of the transepithelial electrical resistance (TEER), which is largely determined by the integrity of the TJ, as well as the paracellular permeability, usually analyzed using fluorescein isothiocyanatedextran are significant indicators of barrier function intactness ^[16]. The measurements of TEER and epithelial permeability are also influenced by the culture medium. Specifically, the addition of fetal bovine serum (FBS) leads to a lower resistance and higher epithelial permeability in comparison to the use of porcine serum (PS). While a higher TEER induced by FBS is more suitable for investigating the impact of substances expecting to impair barrier integrity, the lower TEER is more comparable to the in vivo situation ^[16]. Furthermore, TJ ultrastructure as well as the morphology of enterocytes better mimic the normal physiological porcine epithelium when cultured in PS. Polarization markers glucose transporter 2 (GLUT2) and sodium–potassium adenosine triphosphatase (Na+/K+-ATPase) in basolateral membrane are expressed in cells cultured in either PS or FBS but apical membrane polarization markers ezrin and sodium/glucose cotransporter (SGLT) 2 could not be found in FBS-cultured IPEC-J2 cells, probably because of their low cell height ^[29]. Despite these supposed advantages of PS, the addition of FBS to the medium is far more common, which might be due to the original protocol as well as the lower cost of purchase.

2.2. Cellular Transport and Permeability

As part of the regulation of the barrier function of intestinal epithelium, the impact of PFA on transport mechanisms forms another field of interest. In general, the epithelial cell barrier can be passed by compounds through two main routes: The paracellular pathway between cells or the transcellular pathway throughout the cell ^[30]. Due to the expression of various transport proteins and the formation of tight junctions, as well as the measurability of paracellular permeability, studies of both para- and transcellular transport are also possible in an IPEC-J2 cell model ^[16]. In IPEC-J2 cells challenged with enterotoxigenic Escherichia coli (ETEC K88), for example, expression of Na+/H+ exchanger 3 (NHE3) as well as water channel aquaporins were decreased ^[31], which eventually could be counteracted by PFA which are already known for their ability to reduce ETEC K88 adhesion. The Ussing chamber is a common ex vivo method for determining absorption/transport processes of the epithelium. In most of the protocols, intestinal explants are mounted into the chambers, but in the course of reducing and replacing animal experiments, animal-free alternatives are also being sought here. Mounting IPEC-J2 monolayers into the Ussing chamber demonstrated a positive influence of epidermal growth factor on TEER and the absorption of glucose and glutamine in lipopolysaccharide (LPS)-challenged cells ^[32], which is why combining IPEC-J2 cells with the Ussing chamber may also offer further information on the influence of PFA on transport processes.

2.3. Pathogen Infections

With regard to host–microbial interactions, many similarities between the IPEC-J2 cell line and the porcine epithelium in vivo have been detected. Invasion with the bacterium Salmonella enterica ser. Typhimurium in IPEC-J2 cells is comparable to that in porcine ileal mucosal explants ^[33] and secretion of interleukin (IL) -8 and macrophage inflammatory protein (MIP) 3 alpha increases in both IPEC-J2 cells and orally infected pigs ^[34]. IPEC-J2 therefore seem to be a suitable cell line for studying antimicrobial effects in vitro. In studies using IPEC-J2 cells and porcine mucosal explants interactions with E. coli also showed a high correlation; however, some enterohemorrhagic E. coli (EHEC) mutants differed in their adherence ^[35]. Moreover, there is disagreement as to whether IPEC-J2 cells support the adhesion of the pathogenicity factor F18 of ETEC through the expression of a fimbriae-specific receptor. Some authors stated that IPEC-J2 cells do not express this receptor [36^[36][37],37], but the expression of the F18 receptor has been the subject of several recent studies [38^[38][39],39], suggesting that IPEC-J2 cells do express the F18 receptor. It should be furthermore noted that the IPEC-J2 cell line has been isolated from the small intestine, so the influence of specific pathogens or its antigens located in the colon may not be adequately represented. Nevertheless, receptors for some specific toxins physiologically acting in colon are also present in IPEC-J2 cells, as a reaction of IPEC-J2 cells to toxins from Clostridioides difficile was recently reported ^[24]. Certain mechanisms affecting colonic cells in vivo therefore can also be studied in IPEC-2 cells. At present, moreover, no non-cancerous colon cell line (human- or animal-derived) is available ^[40]. Although co-infections of IPEC-J2 cells, for example with porcine bocavirus and porcine circovirus 2 ^[41] have been performed, to date, culturing of bacterial community dynamics is not possible, therefore, studying of bacterial–host interactions is limited to single bacterial strains ^[40]. In vitro digestion models such as the continuous fermentation model PolyFermS give the opportunity to display the interaction of e.g., probiotics and bacterial communities derived from feces ^[42]. Connecting IPEC-J2 cells to such a digestion model could enable further studies concerning the interplay between host cells and entire heterogeneous bacterial communities and therefore act as an interesting tool for also studying PFA. Due to its similarity in gene expression to the human intestine, the IPEC-J2 cell line offers the opportunity to study interactions with zoonotic pathogens [10,34]^[10][34]. When researching foreign pathogens, on the other hand, it is always important to bear in mind that the response may differ from that in the actual host. For example, Shiga toxin 2e-producing E. coli (STEC) strains isolated from humans did not bind to IPEC-J2 cells, unlike isolates from pigs ^[43], and also rotavirus strains derived from pigs showed a significant higher infectivity in IPEC-J2 cells than those derived from humans ^[44]. Thus, host specificity plays an important role and should be considered when designing studies on the interactions of epithelial cells with bacteria.

2.4. Oxidative Stress

A further field of interest concerning PFA is their anti-oxidative potential, not only for extending the shelf life of feed, but especially for preventing endogenous cells from cell death due to oxidative stress ^[45]. A favored applicant for experimentally inducing oxidative stress in IPEC-J2 cell line is H₂O₂ due to its capability of negatively affecting distribution of TJ proteins e.g., zonula occludens 1 (ZO-1), deteriorating barrier function, decreasing cell viability and triggering intracellular production of reactive oxidative species (ROS), an indicator for oxidative stress ^[46]. When examining antioxidant properties of PFA, the cell model cannot automatically be transferred to the situation in vivo. Although exogenous oxidative stress can be successfully simulated by adding H₂O₂, it should be noted that in addition to exogenous stress, epithelial cells in vivo are exposed to stress of endogenous origin (mainly ROS originating from mitochondrial oxidation), which is not simulated in vitro, nor is the multifactorial oxidation induction by other stimulants such as enterotoxins ^[47]. The use of the cell model nevertheless offers the possibility of investigating substances in general for their antioxidant activity. Most of the studies considered in Section 2.5.5. (“Antioxidative effects”) used H₂O₂ for inducing oxidative stress in IPEC-J2 cells. Interestingly, a study from 2016 suggested a xanthine/xanthine oxidase (X/XO)-induced oxidative stress model when investigating apoptosis in weaned piglets. The reseauthorchers detected, that expression of apoptosis-related genes in (X/XO) challenged cells is more highly correlated to 21-day-old pigs than those in H₂O₂-induced models ^[47].

2.5. The IPEC-J2 Cell Line as Model for Investigating PFA

Due to the above mentioned advantages in function and morphology over the IPEC-1 cell line, IPEC-J2 cells in particular have been used so far for investigations of epithelial reactions to PFA. Cytokine secretion and protein expression of TJ is mainly consistent with those of jejunocytes in vivo ^[16] and even the inflammatory pathway markers Nuclear factor kappa B (NF-kB) and myeloid differentiation marker 88 (MYD88) are expressed ^[48]. Nevertheless, the model presents some limitations when comparing with the porcine jejunum. Claudin (CLDN) -2 and -15 are not expressed in IPEC-J2, resulting in a higher ion permeability than in vivo [20,29]^[20][29]. Moreover, although expression of mucin and MUC3 ^[20] is assumed, it remains unclear if IPEC-J2 cells express MUC2, a major mucus-producing mucin in humans and other animals. Thus, the lack of goblet cells or MUC2 expression have been repeatedly reported [20,49]^[20][49] but very recent studies have detected and measured messenger ribonucleic acid (mRNA) expression of MUC2 [50,51]^[50][51]. Goblet cells could not be detected ^[20]. Finally, no brush border enzymes have been determined yet, making it impossible to study the effects of pathogen agents or feed materials on those enzymes ^[16]. For reproducible research with in vitro cell lines, a strictly standardized setting is essential. Differences in seeding density influence the TEER, proliferation capacity and functionality of the cultured IPEC-J2 cells ^[16]. Therefore, cell density should always be considered in order to reduce variability. The passage number of the culture is a frequently underestimated parameter; it is often not even stated in experiments ^[16]. When acquiring cell cultures from cell banks, this important information often remains unclear. Although there is little literature on the influence of the passage number on IPEC-J2 cell physiology, it is known that Caco-2 human cell line is clearly affected by the number of passages [52,53]^[52][53]. It has been shown that increasing passage number can lead to a reduced secretion of signaling agents by IPEC-J2 cells, but this might be counteracted by adding PS in the culture media ^[29]. Thus, in order to obtain reproducible values, the number of passages should always be provided if possible. Furthermore, it is particularly important to work under strictly sterile conditions in order to prevent contamination of the cells. Important contaminants are Mycoplasma spp., which can remain unnoticed for a long time and require complex and cost-intensive disinfection processes when diagnosed ^[54]. Regular rapid tests for the detection of mycoplasma are therefore advisable as well as examining the identity of cell lines when purchasing from cell banks as it, especially in animal cell lines, occurs to be incorrect ^[55]. Since PFA also pose a great risk of contamination, they should always be sterile filtered before being tested in cell models. Essential oils in excessively high concentration can have a cytotoxic effect on epithelial cells and the positive effect can thus be transformed into damage to the epithelium. Finding the appropriate concentration of the tested substances is therefore crucial ^[8]. When it comes to investigations of slightly soluble extracts, dissolving of samples can be a challenging task and although dimethyl sulfoxide (DMSO) or ethanol can be used for bringing them into solution, a concentration above one percent is not recommended ^[16].

2.6. The IPEC-1 Cell Line as Model for Investigating PFA

Only few studies have been performed using the IPEC-1 cell line for investigating the impact of PFA on epithelium in vitro. This might be due to lower differentiation in morphology and metabolic functions of IPEC-1 compared to the IPEC-J2 cell line ^[21], making it less attractive. Nevertheless, a barrier integrity enhancing effect as well as antimicrobial activity could be determined using the IPEC-1 cell line. Cinnamaldehyde could improve barrier function characterized by increased TEER, decreased paracellular permeability and promoted localization of TJ proteins CLDN-1 and CLDN-3 to the plasma-membrane. Furthermore, protein expression of neutral and basic amino acid transport protein rBAT (rBAT), cystine/glutamate transporter xCT (xCT) and L-type amino acid transporter 2 (LAT2) were elevated by cinnamaldehyde [96]^[56]. In ETEC-challenged IPEC-1 cells, yeast extract, daidzein, bromelain and allicin contributed to barrier integrity by an increase of TEER and decrease of paracellular permeability. Yeast nucleotides, unsaturated oligo-mannuronic acid, ulvan, Chlorella vulgaris extract, cinnamaldehyde and carvacrol, on the contrary, could not affect those parameters [97]^[57].