Table 1.

Overview of intestinal models used to study interactions between the intestine and microbes, and discussed in this review.

model	commensal bacteria	pathogenic bacteria	parasite
tissue explant. Portions of organs maintained in culture by means of specialized media, substrates and atmospheres these models are commonly used to study human cells and parasite interaction, principally in human obligate pathogen advantages: the presences of mucus layer, immunological, neuronal and vascular responses. Models that recapitulate microenvironment observed during human pathologies disadvantages: difficulty in obtaining tissue samples and also, they show short duration of viability. There are genetic variations among the samples, which in some cases affect the interpretation of the data	—study of microbiome. Demonstration of the protective effect of lectin-like protein, ZG16, as protector against Gram-positive bacterial infection [73] —study of probiotic and commensal Escherichia coli strains. Demonstration that outer membrane vesicles (OMVs) secreted by E. coli Nissle 1917 and E. coli ECOR12, is an effective system to communicate human host and bacteria to trigger specific responses [74] —study of commensal bacteria. In ileal explants from Crohn's disease (CD) patients, incubation with commensal bacteria E. coli or Lactobacillus casei the alterations observed in CD samples were partially recovered [75]	—study of early events of Shigella flexneri infection. Interaction between human tissue and the invasive strain of S. flexneri induce desquamation and reduction of epithelial height and neuronal alteration of submucosal neurons [76] —evaluation of cholera toxin B subunit (CT-B) from Vibrio cholerae. Mucosal explants from CD patients incubated with recombinant CT-B demonstrated the effects of CT-B as an inhibitor of mucosal Th1 cell signalling [81]	—study of virulent factors of Entamoeba histolytica. Demonstration of cysteine proteinase A5 activity as a modulator of human metalloproteinase-3 to trigger trophozoite invasion [77]. Transcriptome comparisons between virulent (HM1:IMSS) and non-virulent (Rahman) strains, to provide a global view of the gene expression profile of each strain [78]. In different contexts, was carried out the transcriptome analysis of gene expression of E. histolytica. The major transcriptome changes were observed in human explants [79] —study of gene expression of Cryptosporidium parvum or Cryptosporidium hominis during intestinal invasion. During tissue invasion, both strains showed overexpression of osteoprotegerin (OPG) and inhibition in the OPG ligand tumour necrosis factor-related apoptosis-inducing ligand (TRAIL) b via OPG. [80]
organoid. Miniaturized intestine that derives from proliferation of the intestinal stem cells (ISC) and show tissue structures such as villus-like domains with cellular polarization and crypt-like domains with proliferative regions these models are currently used as a platform in pharmacology studies and also as host–bacteria and host–virus interaction models principally, in cases that lack a suitable animal model advantages: their major, strengths are the presence of 3D structures, cellular polarization and presences of several epithelial-cell lineages organoids can be developed from biopsy derivate from healthy samples or from patient samples in two weeks disadvantages: the biggest disadvantage is the need to manipulate the structure to establish the interaction between the organoid and pathogens. At the moment, there are three major methods to stablish host–microbe interaction studies:	—evaluation of microbiota-derived molecules. In organoids deriving from duodenal biopsies from gluten-intolerant patients, the barrier function was improved after the microinjection of microbiota-derived molecules [92] —study of regulatory effects of Lactobacillus reuteri in damaged mucosal barrier. By co-culture of mouse intestinal organoids and lamina propria lymphocytes, it was demonstrated that the presence of L. reuteri D8 was effective to repair the damaged epithelium caused by TNF-α treatment [93] —evaluation of non-pathogenic strain of E. coli in immature human epithelium. In human organoids, it was shown that microinjection of non-pathogenic E. coli generated a close host–microbe interaction in naive epithelium. This interaction increases antimicrobial peptide production, maturation of mucus layer, and improve of barrier function [94]	—study of earliest stages of infection of enterohaemorrhagic E. coli. Human colonoids, were used to generate a differentiated monolayer in trans-wells inserts. Their infection with enterohaemorrhagic E. coli showed high levels of the serine protease EspP. This protein sequentially targets the Mucin-2 and protocadherin-24 to allow the bacterial attachment to the epithelium [95] —infection by Clostridium difficile. Microinjection of C. difficile showed reduction of MUC2 production, but no changes in mucus oligosaccharide composition [96]. Likewise, a marked epithelial disruption and loss of paracellular barrier promoted by the toxin A were observed [97]. Colonoids derived from Wnt receptor Frizzled 7 (FZD7) KO mice, show that the toxin B is targeted by FZDs receptors in the colonic epithelium [98] —infection with Salmonella enterica serovar Typhimurium. Microinjection of S. Typhimurium induced changes in the human transcriptional pattern and sustained bacterial invasion [100]	—study of the life cycle of Cryptosporidium spp. Organoids derived from human small intestine were microinjected with Cryptosporidium. This model showed substantial physiological relevance by its ability to complete the complex life cycle of parasite [107,108]
microinjection, mechanical shearing and plating, and by making monolayers	—evaluation of non-pathogenic strain and pathogenic strain of E. coli. Microinjection of commensal strains and pathogenic strain O157:H7 of E. coli showed rapid growth. The commensal strain did not cause damage suggesting the positive effect of mucus production. Conversely, a loss of actin and epithelial integrity was observed with O157:H7 [99]	—infection with Helicobacter pylori. Microinjection of H. pylori was used in a model to study bacterial infection [102]. Also, bacterial infection resulted in rapid association of CagA with the c-Met receptor and the induction of epithelial proliferation [101]. Infection induces epithelial proliferation and c-Met phosphorylation [103] —infection with Shigella flexneri. Human enteroids were used to generate epithelial layer in trans-well inserts. Their infection showed basolateral invasion and disruption of tight junctions. Finally, bacterial replication, actin tails and increase in proinflammatory signals were observed [104] —infection with pathogenic E. coli. Polarized organoids were disaggregated and grown on trans-well inserts. They demonstrated novel adherence phenotypes in various strains of pathogenic E. coli [106]
scaffold-based 3D models. Matrices made with synthetic or natural materials for the deposition of cells to mimic the architecture of the intestine and promote the acquisition of tissue characteristics these models are commonly used to study human cells and bacterial or parasite interactions, principally in human obligate pathogens advantages: their major strengths are the presence of 3D structures, cellular polarization and presence of several cell lineages in some models, the major advantages are their versatility. Components can be varied independently according to the study objectives. There is the possibility of combining 3D scaffolds and cells derived from human intestinal explants or organoids disadvantages: the biggest disadvantage is their lack of physical forces presents in tissues	—study of Lactobacillus gasseri and E. coli Nissle 1917 as protective agents of the intestinal tract. The therapeutic potential of L. gasseri and E. coli Nissle 1917 against Salmonella typhimurium and Pseudomonas aeruginosa was demonstrated [125] —culture of probiotic, Lactobacillus rhamnosus GG. Development of bioengineered human intestinal tissues that mimic in vivo luminal oxygen levels. This system supports the growth of dominant anaerobic probiotic bacteria, Lactobacillus rhamnosus GG [128] —study of commensal bacterial strains. Development of a scaffold-based 3D model that contained enterocytes, goblet and immune-like cells. The cells were exposed to synthetic commensal microbial community and LPS from E. coli O111:B4 strain. This interaction promotes the adhesion of specific bacterial strains, Veillonella parvula, to stimulate the epithelium barrier function and interleukin production [134]	—infection with Salmonella typhimurium. The human intestinal villi epithelium was mimicked in a hydrogel scaffold. Using this model it was possible to establish the importance of MUC17 during bacterial infection [126] —infection with Yersinia pseudotuberculosis. The human intestinal lumen was replicated in a 3D porous scaffold. In to evaluate the importance of mucus and the establishment of low oxygen to study bacterial interactions/colonization [128] —infection with Campylobacter jejuni. A decellularized extracellular matrix scaffold and reseeded with human Caco-2 cells was designed to study host–pathogen interactions. The infection with C. jejuni replicated some pathogenic processes previously observed in animal models and showed new virulent factors involved [132]	—infection with Cryptosporidium parvum. It was possible to reproduce the microenvironmental conditions of the intestinal tract to support the life cycle of C. parvum in a bioengineered 3D human intestinal tissue model. This system has opened the possibility to evaluate host–parasite interactions and identify new drug targets [129,130] —study of the early steps of Entamoeba histolytica infection. The general aspects of the human colon were reproduced in a 3D model, which highlighted the importance of several virulence markers previously observed and also, described new molecules and regulatory factors involved in the amoebic invasive process [136]
organ-on-a-chip. These are micropatterned synthetic surfaces that support the correct spatial arrangement of the cells and help to control gradients of biomolecules by means of microfluidic applications these models are currently used as platform to study host–microbe interaction models for long periods of time. Those models are used principally to study bacterial interaction in cases that lack a suitable animal model advantages: their major advantage is the ability to reproduce the physiology of the human gastrointestinal tract in 3D context, in the presence of flux fluid, peristaltic motion and oxygen gradient disadvantages: their principal weakness lies in the implementation of highly complex structures requiring qualified personnel for their design, and PDMS, the material most commonly used for chip construction, which absorbs small hydrophobic molecules that could interfere with some drug-screening studies	—infection with Lactobacillus rhamnosus GG. Organ-on-a-chip support long cultivation of L. rhamnosus GG with improved barrier functions and without compromising epithelial viability [142] —iIntestinal interaction among commensal and pathogenic bacteria and intestinal cells. The co-cultivation of commensal microbes and human intestinal epithelial cells suppress villus injury induced by enteroinvasive E. coli (O124:NM) [145] —study of complex communities of anaerobic and aerobic commensal bacteria and intestinal epithelium. The control of physiological oxygen gradients was possible in a gut-on-a-chip model, which grew aerobic and anaerobic human microbiota. The model established the importance of oxygen gradient in barrier function and maintaining relevant levels of microbial diversity [149] —study of Lactobacillus rhamnosus GG and Bacteroides caccae under anaerobic conditions. A microfluidics-based model demonstrated that the co-culture of both bacterial strains promotes the transcriptional response in epithelial cells, which is distinct from that of a co-culture solely comprising L. rhamnosus [150]	—infection with Shigella flexneri. A gut-on-a-chip model demonstrated the positive impact of peristalsis during the epithelial invasion of S. flexneri [144] —infection with enterohaemorrhagic E. coli. An organ-on-a-chip model demonstrated that during infection by enterohaemorrhagic E. coli, administration of human microbiome metabolites promote the expression of virulent bacterial factors and increased epithelial injury [148]	no information available