The small intestine: dining table of host–microbiota meetings

Abstract Growing evidence suggests the importance of the small intestinal bacteria in the diet–host–microbiota dialogue in various facets of health and disease. Yet, this body site is still poorly explored and its ecology and mechanisms of interaction with the host are just starting to be unraveled. In this review, we describe the current knowledge on the small intestinal ecology, its composition and diversity, and how the intestinal bacteria in homeostatic conditions participate in nutrient digestion and absorption. We illustrate the importance of a controlled bacterial density and of the preservation of absorptive surface for the host’s nutritional status. In particular, we discuss these aspects of the small intestinal environment in the framework of two disease conditions, namely small intestinal bacterial overgrowth (SIBO) and short bowel syndrome (SBS). We also detail in vivo, ex vivo, and in vitro models developed to simulate the small intestinal environment, some applied for (diet–)host–bacteria interaction studies. Lastly, we highlight recent technological, medical, and scientific advances applicable to investigate this complex and yet understudied body environment to broaden our knowledge in support of further progress in the medical practice, and to proceed towards the integration of the (small)intestinal bacteria in personalized therapeutic approaches.


Introduction
For millions of years, resident microbes have been coevolving with their host, establishing highly specialized ecological niches and a fine-r egulated cr oss-talk in distinct bod y sites, and contin uously shaping homeostasis for both the host and the gut ecosystem. In this way, the gut microbiota is closely associated to human health, and became the aim of intense scientific studies.
Among the body sites colonized by micr obes, the digestiv e tr act, and particularly the small intestine, is a crucial interface where the dialogue between host, microbes , and en vironmental factors is complex and profound. As the small intestine is the main site of nutrient digestion and absorption, it is crucial to understand how the complex cross-talk between gut physiology, dietary factors, and the small intestinal microbiota may affect host health status.
Anatomical or pathological alterations in disease conditions may alter this fine dialogue between nutrition, host, and microbe and disrupt homeostasis.
To untangle this complex interaction in health and disease, most studies relied on fecal samples to characterize the intestinal ecology, adv anta geous for the non-inv asiv e collection, although unable to ca ptur e the div erse micr obial phylogen y and functionalities along the gastrointestinal tract. The small intestine is in fact a poorly accessible body site, which makes the direct sampling c hallenging and inv asiv e. Consequentl y, the micr obial ecology of the different small intestinal regions remained undescribed for a long time . T he expansion of omics tec hniques, high-thr oughput sequencing, and metagenomic and metabolomic a ppr oac hes, hav e gr eatl y expanded our knowledge in functionality and microbial composition of the small intestine. Yet, the accurate quantification and c har acterization of its ecology still r emains limited by technique-dependent sampling bias. Additionally, studying cellular signaling mechanisms that govern the hostmicrobiota-diet dialogue in vivo is challenging and hinders the acquisition of novel insights. In this vie w, div erse ex vivo and in vitro a ppr oac hes hav e been de v eloped to obtain a mor e mec hanistic understanding of host-microbe dialogue in the small intestine that can further complement or e v en support in vivo observations. This liter atur e r e vie w, ther efor e, aims at discussing recent insights in the description of the small intestinal ecology, epithelium, and its interaction with dietary constituents. We put particular emphasis on the impact from the host-microbiota-diet interplay in the duoden um, jejun um, and ileum on n utrient digestion and absorption under normal healthy conditions and how this is affected when ecological and/or epithelial homeostasis is disrupted in small bo w el syndrome and small intestinal bacterial ov er gr owth (SIBO). We will cov er r ecent br eakthr oughs, innov ativ e in vivo , in vitro , and ex models and discuss potential novel scientific routes to address mechanisms of interaction that are currently poorly understood.

The small intestine: anatomy and function in digestion and absorption
Food digestion and nutrient absorption are fine-regulated processes allowing to extract energy from the diet and contribute to the maintenance of the vital functions in the human body. The main organ devoted to these functions is the small intestine, a curved tubular structure with average length of 690.1 ± 93.7 cm, forming the longest organ in the body (Tacchino 2015 ). The small intestine begins at the pylorus and ends at the ileocecal v alv e and comprises three parts: duoden um, jejun um, and ileum (Fig. 1 ). In the small intestine, the median transit time, guaranteed by motility patterns, such as peristalsis, segmentation and mixing, varies between 196 and 287 minutes depending on the studied group (age, gender, environmental factors, i.e . smoke) and measurement tec hnique (de vice and marker) (Camilleri et al. 1991, Degen and Phillips 1996, Graff et al. 2001, Worsøe et al. 2011, Nandhra et al. 2020, Tominaga et al. 2020. During this time, the food (or partly digested chyme) is exposed to diverse secreted pancreatic and intestinal enzymes and physicochemical parameters existing along the gastrointestinal tract. Specific cellular receptors are also differentially expressed on the surface of the intestinal epithelium, allowing the uptake of the nutrients , a vailable in the lumen. Ov er all, the differ ent secr etions and r eceptors define functional-specialized intestinal segments.
In these complex processes, the host is supported by the metabolic activity of the intestinal microbiota, able to break down nutrients, otherwise inaccessible to the host's digestive enzymes. In this section, the k e y host's digestive and absorption processes are explained.

Physico-chemical parameters in the small intestine
In the small intestine, the acid c hyme r esulting fr om the passa ge thr ough the stomac h is disc har ged, then neutr alized by intestinal secretion in the duodenum (Agrawal and Aoun 2014 ). As a result, the pH r anges fr om v ery acidic in the stomach to slightly basic (pH 7.3-7.7) in the ileum and varies from pH 5.7 to 6.4 in the duodenum and jejunum (Table 1 ).
Along with the pH, the presence of intestinal bacteria as well as the metabolism by host cells determine a concentr ation gr adient of different gases (e.g . oxygen, carbon dioxide, nitrogen, and hydr ogen) acr oss the gastr ointestinal tr act. In particular, the concentration of carbon dioxide increases in the duodenum, before being reabsorbed in the colon (Cormier 1990 ). Conv ersel y, oxygen concentr ation decr eases fr om high le v els in the or al compartment, to micr oaer ophilic in the small intestine and complete anoxic conditions in colon. In humans, the oxygen tension in the small intestine and colon, measured at the serosal surface by intraoperati ve tissue o ximetry, ranges from 36.0 ± 9.7 mmHg in the midileum, to 33.5 ± 11.5 mmHg in the terminal ileum, to a minimum of 29.3 ± 11.0 mmHg in the descending colon (Sheridan et al. 1990 ). Mor eov er, an oxygen gradient exists when moving from the micr oaer ophilic small intestinal lumen to w ar ds the highly vascularized oxygen-rich subepithelial m ucosa (r e vie wed in Espey 2013 ) ( Fig. 1 ), determining different niches for the intestinal microbes.

Nutrient digestion in the small intestine
Through the pancreatic and bile ducts, digestive enzymes produced in the pancreas and bile from the liver, are released in the lumen of the duoden um. Here, pe ptides, starch, and triglycerides ar e br oken do wn in smaller subunits b y the action of pancreatic proteases (Ross et al. 2013 ), pancreatic amylases (Goodman 2010 ), and pancr eatic lipases. Additionall y, pancr eatic nucleases (DNase and RNase) allow some of the nucleotide bases to be recycled and used as building blocks for human DNA and RNA synthesis (Hoard and Goad 1968 ). Lipids ar e em ulsified by bile to facilitate their solubilization and further absorption in the small intestine. A minor fraction of bile salts is subjected to chemical modifications by intestinal microbiota, pivotal for the generation of secondary bile acids. As suc h, micr obial bile salt metabolism in the small intestine and more distally in the colon may indirectly impact lipid metabolism. For example, following deconjugation by microbial bile salt hydrolase (BSH), bile salts will be less able to emulsify lipids lo w ering the accessibility to lipase that are essential for lipid digestion. The lipid metabolism is, thus closely linked to the small intestinal microbiota activity, for generating secondary bile acids.
Additionally, the small intestinal microbiota is involved in the c holecystokinin (CCK) hormone r egulation (Martinez-Guryn et al. 2018 ), crucial for lipid digestion, by inducing the release of bile salts from the gallbladder and the secretion of pancreatic lipase in the small intestine. Coher entl y, in germ-fr ee animals, impair ed lipid digestion and downregulation of CCK ar e observ ed, compared to conventional mice (Martinez-Guryn et al. 2018 ). Upon administration of two bacterial strains (either live bacteria and conditioned media) in germ-free mice, an incr eased expr ession of CCK r eceptor ( Cc kar ) in the pancr eas was observ ed, further confirming the bacterial impact on the regulation of lipid metabolism (Martinez-Guryn et al. 2018 ). Lastly, LPS has been identified to bind Toll-like receptors (TLRs) expressed in enteroendocrine cells (EECs), which in turn triggers the release of CCK in vitro (Bogunovic et al. 2007 ).
The bacterial metabolic activity on complex carbohydrates pr oduces, as end by-pr oducts, short-c hain fatty acids (SCFAs) of whic h acetate, butyr ate, and pr opionate ar e the most abundant in the human small intestine. In sudden-death victims, the total SCFA concentr ation incr eased fr om 0.6 to 13 mmol/kg of intestinal content in the jejunum and ileum, r espectiv el y (Cummings et al. 1987 ), while, in ileostomy effluent, the total SCFA ranged between 51.9 and 119 mM, in slightly different ratios compared to sudden-death victims (Zoetendal et al. 2012 ) (Table 2 ). Besides SCFA, lactate (2.5-11.6 mM in ileostomy effluents), formate (0-26 mM in ileostomy effluents) (Zoetendal et al. 2012 ), and succinate (3.7 mmol/kg and 8.3 mmol/kg in jejunum and ileum, respectiv el y) (Cummings et al. 1987 ) were detected. Lastly, the chyme is transported to the large intestine, whose primary function is to absorb water and electr ol ytes, and wher e the abundant resident microbiota continues the degradation of the nondigestible dietary fibers initiated in the terminal small intestine.

Figure 1.
Ov ervie w of small intestinal anatomy , histology , and k e y pr ocesses for food digestion and absor ption. Physico-c hemical par ameters, and bacterial load (CFU/ml) are indicated for each segment of the small intestine. pH values are based on Ibekwe et al. ( 2008 ) in fasted patients. Cell types present on the small intestinal epithelium are re presented. Ke y digesti ve processes taking place in the small intestine are summarized in the top left-hand panel and, below, the major transport pathway for nutrient absorption. IESC: intestinal epithelial stem cell; IECs: intestinal epithelial cells; CBC: crypt base columnar cell; + 4 LRC: + 4 label retaining cell; EEC: enteroendocrine cells; IgA: immunoglobulin A; AMPs: antimicrobial peptides; Gal: galactose; Glu: glucose; Fru: fructose; SCFA: short-chain fatty acids; and GPCRs: G-protein coupled receptors. Created with BioRender.com.

Nutrient absorption in small intestine
Once the dietary components are broken down into monomers, an estimated 85% of them penetrate the small intestinal epithelium by diffusion, whereas the remaining 15% is taken-up by transcytosis (Szefel et al. 2015 ). Ho w e v er, numer ous r eceptors and transport systems are localized on the epithelium, specific for the different class of molecule to absorb. Then, through underlying blood and l ymphatic ca pillaries in the subm ucosa, the absorbed nutrients transit into the bloodstream. Upon digestion in the lumen and cell intak e, pe ptides may under go intr acellular digestion b y c ytosolic enzymes such as aminotripe ptidases, dipe ptidases, or by lysosomal and microsomal enzymes. Within minutes, virtually all the last dipeptides and tripeptides are digested in the cytosol to single amino acids, which then pass on through to the basal side of the enterocyte and then into the blood. The transfer of amino acids into and out of cells or cellular organelles is ensured by transmembrane proteins (amino acid transporters -AAT) (reviewed in Kandasamy et al. 2018 ).
The fate of products of fat digestion (fatty acids, monoglycerides, gl ycer ol, c holester ol, and fat-soluble vitamins) proceeds via the cellular intake on the intestinal epithelium in form of micelles. Once in the cell, short-and medium-chain fatty acids and gl ycer ol can be absorbed into the bloodstr eam. Conv ersel y, long-chain fatty acids and other digestion products need to reassemble into c hylomicr ons befor e passing into the lymph vessels and, from there, being delivered in the bloodstream. In the terminal ileum, 95% of the bile acids are reabsorbed (De Aguiar Vallim et al. 2013 ), through a combination of passive absorption and active transport in the proximal and distal small intestine, r espectiv el y. Unconjugated bile acids are activ el y tr ansported in the terminal ileum through the apical sodium dependent transporter (ASBT) and r eac hes the portal vein by a basolateral heter odimeric or ganic solute tr ansporter (OST) (r e vie wed in Dawson et al. 2009 ). The small intestinal microbiota has also an important role in host fatty acid absorption mainly via some Clostridiaceae str ains suc h as Clostridium bifermentans , able to affect the gene expr ession of trigl ycerides r e-esterification enzymes, diacylgl ycer ol O-acyltr ansfer ases (Dgat1 and Dgat2) in mice small intestine (Martinez-Guryn et al. 2018 ).
From the host's enzymatic break down of digestible carbohydr ates, the r esulting glucose and galactose are internalized into the enterocytes by active carrier transport, through the apical sodium-glucose cotransporter 1 (SGLT1), whereas fructose pass by facilitated diffusion through the apical glucose transporter GLUT5. The absorption into the blood circulation is ensured by the tr ansporter GLUT2, expr essed on the basal side of the enter ocytes (r e vie wed in Koepsell 2020 ) Conv ersel y, fr om the microbial activity on nondigestible carbohydrates, SCFAs are formed. Dependent on the luminal concentration, SCFAs enter the epithelium by diffu-sion or through active transport systems in the apical membrane of intestinal epithelial cells (Iwanaga et al. 2006 ). Monocarboxylate transporter 1 (MCT1), whose expression is higher in the colon than in the small intestine (Gill et al. 2005 ) and sodium-coupled monocarboxylate transporter 2 (SMCT2 or SLC5A12), exclusiv el y expressed in the small intestine, support the uptake of SCFAs and monocarboxylates such as lactate, respectively (Sivaprakasam et al. 2018 ). Additionall y, G-pr otein coupled r eceptors (GPCRs) were also identified as receptor for SCFAs and thus, named also free fatty acid receptors (FFARs), differently binding to diverse SCFAs. In particular, GPR41 (FFAR3) and GPR43 (FFAR2), recognize acetate, butyr ate, and pr opionate, while GPR109a (HCAR2) is solely activated by butyrate (Brown et al. 2003, Thangaraju et al. 2009 ). After absorption, the metabolic fate of the SCFAs differs. In mammals, butyrate is the primary energy source for colonocytes, oxidized via β-oxidation and tricarboxylic acid cycle. Propionate and acetate ar e tr ansported to the liv er and peripher al tissues, r espectiv el y and both are used as substrates for energy metabolism and lipid synthesis (Wong et al. 2006 ). Around 6%-9% of the total energy intake for humans, accounts from SCF A absorption (W ong et al. 2006 ). Ho w e v er, the biological effects of SCFAs are not restricted to their sole role as energy substrates for the epithelial cells, but they also contribute to water and electr ol yte absor ption in the colon, modulate the mucosal immune system and aid in the maintenance of the mucosal barrier (reviewed in Martin-Gallausiaux et al. 2021 ).

Histology
As the major absor ptiv e site, the small intestine has se v er al arc hitectur al modifications of the mucosa and submucosa to increase its surface, namely folds or plicae circularis , arranged circularl y ar ound the lumen, and villi and micr o villi, co vering the apical surface of the small intestine (Fig. 1 ). The plicae , villi, and micr ovilli ar e long and numer ous in the duoden um and jejun um and decrease in abundance and thickness to w ar ds the proximal ileum. Micr ovilli ar e pr esent on the surface of eac h epithelial cell, packed together to form the so-called brush bor der, dev oted to the secretion of digestive enzymes (brush border enzymes), absorption, and cellular adhesion (reviewed in Walton et al. 2016 ).
In the small intestinal epithelium, enterocytes are dedicated to perform terminal digestion of pol ysacc harides and peptides and absorb nutrients present in the intestinal lumen. This cell type, comprising for about 80% of the small intestinal epithelial cells (Van Der Flier and Cle v ers 2009 ) is c har acterized by a specialized portion of the cell membrane on the luminal surface, the microvillar membrane, or brush border, bearing digestive enzymes and specific carrier proteins. Along with its role in digestion and absor ption, enter ocytes participate in the formation of a biochemical barriers, to pr e v ent the diffusion of pathogens, toxins, and al- Table 3. Characteristics and distribution of mucins in the small intestine of adult humans.

MUC3
No Duoden um, jejun um, ileum Goblet cells and absor ptiv e cells, predominance on villi Cao et al. ( 1997 ), Williams et al. ( 1999a , b ), Buisine et al. ( 2001), Paulsen et al. ( 2006, Audie et al . ( 1993 )  ler gens fr om the lumen to the mucosa (Peterson and Artis 2014 ). This selectiv el y permeable barrier depends on the interaction of se v er al barrier components , including mucus , immunoglobulin A (IgA), and antimicrobial peptide secretion, to segregate microorganisms and allergens in the lumen. In particular, enterocytes secr ete se v er al micr obicidal or antivir al a gents and tr ansfer imm unoglobulins fr om the m ucosal plasma cells to the lumen. An important component of this chemical barrier is also the mucus layer, a single thin layer covering the small intestinal epithelium, constituted by m ucins, expelled thr ough a distinctive mode of secretion referred as 'expanding secretion' (Dolan et al. 2022 ). The secreted mucus provides protection, lubrication, and hydration of the human epithelial tissues exposed to the environment (Andrianifahanana et al. 2006 ). In the human genome, 21 mucin (MUC) genes are known, encoding for secreted or membr ane-bound m ucins (Boltin et al. 2013 ) (summarized in Table 3 ). Among the secr eted m ucins, MUC2 is the most predominant in the small intestinal epithelium. Besides protection for the e pithelium, mucins offer n utrient support for adhering bacteria, promoting their colonization of the outer part of the mucus layer (Liévin-Le Moal and Servin 2006 ).
Exposed to constant environmental stimuli, the intestinal epithelium has e volv ed sensing strategies to detect the passage of food or the presence of potential harmful compounds and micr oor ganisms. EECs detect luminal content and coordinate the response of the gastr ointestinal tr act to food ingestion, through the secretion of hormones. Present in the small intestine at a density of a ppr oximatel y 1%, EECs r epr esent a famil y of cell subtypes classified according to their localization, shape, and hormonal secretion pr ofile (r e vie wed in Guo et al. 2022 ). Hormones ar e sorted into secr etory gr an ules by carbo xype ptidase E (Hosaka et al. 2005, Mc-Girr et al. 2013 ) along with gr anins, including secr etogr anin III and c hr omogr anin A, the former being extensiv el y used as a specific marker of EEC in intestinal epithelium. EEC subpopulations express a wide range of receptors, enabling the detection of luminal content or the response to paracrine stimulation (Raybould 2010, Reimann et al. 2012 ), among whic h G-pr otein-coupled r eceptors (GPCRs). For example, through a mechanism mediated by GPR43 and inhibition of histone deacetylases , butyrate , and propionate have been described to stimulate peptide YY (PYY) expression in human EEC cell lines (Larraufie et al. 2018 ). Furthermore, both in vitro and in vivo studies demonstrated that EECs also express functional TLRs and respond to bacterial TLR ligands (Larraufie et al. 2017 ), supporting the role of EECs as sensor of gut microbiota.
Along with these cell types, sentinels driving type 2 immune mec hanisms, in r esponse to pathogens (Gerbe et al. 2012 ), ar e spor adicall y distributed on the epithelium and termed Tuft cells . T he pr e v alence of this cell type in the human small intestinal epithelium has not been reported yet, but, in the human sigmoid colon Tuft cells are present at a density of ∼100 cells per square millimeter tissue (Kjaergaard et al. 2021 ). Yet, the presence of Tuft cells in the mouse small intestine has been documented (Banerjee et al. 2021 ). Additionally, it has been reported that succinate, derived fr om intestinal bacteria, driv es the expansion of a subpopulation of Tuft cells (ATOH1-independent tuft cells) exclusiv el y pr esent in the small intestine, and ultimately participating in the reduction of c hr onic intestinal inflammation in mice (Banerjee et al. 2021 ).
Inv a ginations of the mucosa from the bases of the villi and into the lamina propria are called crypts of Lieberkühn, more prominent in the proximal small intestine compared to the distal part (Helander and Fändriks 2014, Parker and Hohenberger 2019, Agarwal et al. 2021, acting as glands that secr ete antimicr obial a gents and hormones. To this scope, Paneth cells, highly specialized secr etory cells, ar e located in the crypts . T hey contain prominent eosinophilic granules in their cytoplasm (Lueschow and McElroy 2020 ), composed of antimicrobial peptides and immunomodulating proteins, that once released at the apical surface into the lumen, regulate the composition and abundance of the intestinal micr obiota and pr otect fr om pathogens (Luesc how and McElr oy 2020 ). Lastl y, specialized micr ofold cells (M cells) cov er or ganized lymphoid follicles in the ileum, called P ey er's patches . T hey pla y a centr al r ole in initiating m ucosal imm une r esponse by tr ansport antigens and micr oor ganisms to the underl ying l ymphoid tissue. In fact, ablation of M cells in mice results in delayed maturation of P ey er's patches and inefficient induction of secretory IgA (Rios et al. 2016 ). In ad dition to these protecti ve barriers, a physical barrier is guaranteed by the presence of cell-cell junction complexes. These protein complexes are involved in cell-cell adhesion, prev enting par acellular diffusion of micr oor ganisms and antigens while r egulating par acellular tr ansport of molecules . J unctional complexes include tight junctions, the most apical component of intracellular junctions.
To support the protection and digestive function of the epithelium, submucosal Brunner's glands are located in the first and distal portion of the duodenum, and secrete several products, such as a bicarbonate-rich alkaline secretion to neutralize the acid chyme, a m ucinous secr etion, to lubricate the m ucosa, bactericidal factors, epidermal growth factor, and surface-active lipids (Gelberg 2014 , Bass and Wershil 2015 ). Mor eov er, by conv eying a ric h network of blood v essels, l ymphatics, and nerv es, the subm ucosa supports the mucosa in nutrient, fluid, and electrolyte absorption.
The absor ptiv e and pr otectiv e functions of the gut are dependent on an intact and functional epithelium, maintained by constant cell r ene wal. In adult mammals, the intestinal epithelium undergoes continuous turnover every 2-5 days (Darwich et al. 2014 ) from the pool of multipotent stem cells, residing at the base of the small intestinal crypts . T hese cells have been wellc har acterized and ar e known to expr ess stem cell markers such as a Leu-ric h r epeat-containing G-pr otein-coupled r eceptor (LGR5) (Barker et al. 2007 ), fundamental for intestinal homeostasis (Tan et al. 2021 ). In fact, the mature cell type that constitutes the epithelium, originates from stem cells and differentiate during the migration a wa y from the re plicati ve zone at the bottom of the crypt, along the crypt-villus axis (Barker 2014, Agarwal et al. 2021. Active intestinal stem cells, also known as crypt base columnar cells (CBCs), spaced alternativ el y to Paneth cells, undergo constant pr olifer ation, and giv e rise to tr ansit-amplifying cells . T hese cells differentiate into absorptive lineage, giving rise to mature enterocytes and secretory lineages, from which goblet cells , EECs , Tuft cells, and M cells mature. To guarantee the preservation of the epithelium, both pr olifer ativ e pr ogenitors and terminall y differentiated cells can 'revert' to an intestinal stem cell phenotype, following depletion of the Lgr5 + population, to support tissue r egener ation (Tetteh et al. 2016 ). Additionally, a quiescent stem cell population commonly referred to as + 4 label retaining cells ( + 4 LRC), is also present in the crypt and is able to r estor e the LGR5 + CBC stem cells, when depleted (Tian et al. 2011 ).

The small intestinal microbiota
In the small intestine, the main phyla described are Firmicutes, Pr oteobacteria, Bacter oidetes, Fusobacteria, and Actinobacteria, r ecentl y r enamed as Bacillota, Pseudomonadota, Bacteroidota, Fusobacteriota, and Actinomycetota, r espectiv el y. Ho w e v er, along the small intestinal segments, differences exist in bacterial composition and abundance . T he small intestinal microbiota can be considered an open ecosystem receiving an influx of microorganisms from proximal locations of the digestive tract. An important determinant of the small intestinal microbiota composition is the or al cavity. Dail y, about 1-1.5 l of saliva is sw allo w ed, (Humphrey and Williamson 2001 ) resulting in the ingestion of about 10 12 bacteria per day. When ingested, these oral bacteria need to conquer m ultiple c hemical and physical barriers, gastric acid, and bile acids before colonizing further along the gastrointestinal tract (Martinsen et al. 2005, di Gregorio et al. 2021 ). In healthy individ-uals, 89% of the taxa present in the duodenum are also found in pair ed saliv a samples , indicating a huge impact of the oral-intestinal transfer in the determination of micr obial composition. Like wise, the micr obial comm unity fr om the jejunum resembles that of the duodenum microbial community (Nagasue et al. 2022 ), and hence also ov erla ps with the oral community, including Prevotella , Veillonella , Haemophilus , and Fusobacterium (Sundin et al. 2017 ). Conv ersel y, the ileum, which shows significant differ ences fr om the jejunum composition, clusters between the upper and lower gastrointestinal tract (Nagasue et al. 2022 ). The transfer of oral-like bacteria to the ileum is a certainty as e v en comparison between saliv a and stool, shows or al-fecal transmission for members of oral Streptococcus , Veillonella , Actinomyces , and Haemophilus , while members of the Prevotella genus are onl y occasionall y tr ansmitted (Sc hmidt et al. 2019 ). When comparing the bacterial α-diversity of the upper, lo w er intestinal tract and fecal samples, the small intestine is reported to have the lowest α-diversity (Seekatz et al. 2019, Vuik et al. 2019, Kashiwagi et al. 2020. Additionall y, when compar ed with saliv a, the jejunum was also found to have a lower diversity than saliva (Sundin et al. 2017 ), possibly due to the drastic bacterial reduction in the stomach. In this section, the bacterial community in the adult small intestinal segments (duoden um, jejun um, and ileum), is described.

Duodenum
The duodenal microbial load in a healthy individual is considered lo w er than or equal to 10 3 CFU/ml of duodenal aspir ate, whic h is also the threshold to define a disease condition termed SIBO (detailed in a dedicated section) . T he o verall duodenum luminal (aspirates) and mucosal (biopsies) bacterial community is dominated by Bacillota and Pseudomonadota accounting together for more than 70%, while Bacteroidota, Actinomycetota, and Fusobacteriota are present at lo w er levels (Li et al. 2015, Vuik et al. 2019, Leite et al. 2020b, Nagasue et al. 2022 ). Leite and colleagues identified Actinomycetota as second dominant phyla after Bacillota and in other studies some individuals pr esent high le v els of Bacter oidota, making Bacillota, Pseudomonadota, Actinomycetota, and Bacteroidota the dominant phyla in the duodenum, follo w ed b y Fusobacteriota and TM7 (Li et al. 2015, Seekatz et al. 2019, Kashiwagi et al. 2020, Leite et al. 2020b ). At genus le v el, most studies r eport Streptococcus ( Streptococcaceae ) as one of the dominant bacteria in the duodenal lumen and mucus (Li et al. 2015, Seekatz et al. 2019, Vuik et al. 2019, Kashiwagi et al. 2020, Nagasue et al. 2022. Other occurring gener a, r e ported in duoden um in healthy conditions, are summarized in Table 4 . In a study by Li et al. ( 2015 ), the microbial composition of duodenal biopsies and duodenal fluid was compared, and they observ ed dominant micr obes differing between both samples. In particular, while the biopsies were dominated by Acinetobacter , Bacteroides , and Prevotella , in the duodenal fluid Prevotella , Stenotrophomonas , and Streptococcus wer e abundant. Yet, r eports comparing the microbial composition in the mucosal and luminal nic he ar e limited and additional r esearc h is needed to full y a ppreciate how this niche-specific community varies in the human duodenum, not only at interindividual, but also at intraindividual le v el.

Jejunum
The jejunum load ranges from 5.8 × 10 3 to 8.0 × 10 6 CFU/ml when sampled during enteroscopy (Sundin et al. 2017 ), yet when sampled during surgery, lo w er bacterial levels < 1.6 × 10 3 were de-   tected in the median population (Villmones et al. 2022 ). In terms of oxygen resistance, the jejunal luminal microbiota are primarily aer obes, facultativ e and obligate anaerobes and oxygen-tolerant bacteria (Hayashi et al. 2005, Sundin et al. 2017 ). The jejunal lumen and mucosa are dominated by Bacillota and Pseudomonadota, follo w ed b y Bacteroidota, Actinomycetota, and Fusobacteriota in v arying le v els, depending on the study (Wang et al. 2005, Dlugosz et al. 2014, Sundin et al. 2017, Vuik et al. 2019, Leite et al. 2020b. In comparison to the duodenum, jejunal biopsies and aspir ates pr esent lo w er le v els of Bacter oidota, among whic h the genus Prevotella (Seekatz et al. 2019, Leite et al. 2020b, Nagasue et al. 2022. While Prevotella was reported within the top three most abundant genera in the jejunal mucosa and lumen of healthy individuals, no comparison to duodenum samples was made in these studies (Dlugosz et al. 2014, Sundin et al. 2017 ). Similar to the duoden um, the jejun um at gen us le v el is dominated by Streptococcus in most studies (Hayashi et al. 2005, Dlugosz et al. 2014, Sundin et al. 2017, Vuik et al. 2019, Villmones et al. 2022, while the presence of other genera greatly differs between reports (Table 4 ). A study by Sundin et al. ( 2017 ) on jejunal aspirates, identified six core species, defined as abundant in more than 50% of the subjects, namely: Streptococcus mitis , Veillonella atypica , Haemophilus parainfluenzae , Fusobacterium periodonticum , Streptococcus vestibularis , and Prevotella melaninogenica . Dlugosz et al. ( 2014 ) observed a clustering of about 24% of the jejunal mucosal samples dominated by Prevotella , the remaining samples were distributed along a gradient between a high Streptococcus or Escherichia abundance. In ad dition, the y re ported patterns of codependence between Prevotella and Veillonella and mutual exclusivity between Esc heric hia and Rothia .

Ileum
When describing the ileum microbiota, a distinction is made between the proximal and terminal part, usually sampled in a differ ent manner, whic h possibl y intr oduces cr oss-contaminations from upper or lo w er gastrointestinal tract, respectively. The proximal ileum mucosa is dominated by Bacillota and Pseudomonadota (Vuik et al. 2019, Nagasue et al. 2022, while the terminal ileum mucosa has increased Bacteroidota levels compared to the proximal ileum (Wang et al. 2005, Vuik et al. 2019, Nagasue et al. 2022. Ho w e v er, when sampling the distal ileum (and sometimes the proximal too) a retrograde endoscopy method is used with possible cross-contamination from the lo w er gastrointestinal tr act, whic h harbours incr eased Bacter oidota le v els. To limit the cross-contamination, Villmones et al. ( 2018 ) sampled the terminal ileum dir ectl y during sur gery and r eported that Bacillota was predominant, followed by Actinomycetota, Candidate division TM7, Pseudomonadota, and Fusobacteriota while Bacteroidota was only found in 40% of the subjects. Ho w e v er, conflicting results are described in a study on terminal ileum biopsies, where Fusobacteriota dominates, follo w ed b y Pseudomonadota, Bacillota, Bacteroidota, and Actinomycetota (Fan et al. 2020 ). These contradictory results might be due to the different sampling method, the demogr a phic par ameter of the studied population (e.g . age) and other possible confounders, such as diet and pathologies (Booijink et al. 2010. In contrast to the upper small intestinal sites, the Verrucomicrobia phylum seems to primarily appear in the terminal ileum (Wang et al. 2005, Nagasue et al. 2022. At genus le v el, the ileum displays increased Bacteroides and Esc heric hia le v els, but lo w er Prevotella le v els, compar ed to the duodenum and jejunum (Nagasue et al. 2022 ).
In terms of diversity, ileum biopsies were found more diverse than jejunal biopsies (Nagasue et al. 2022 ). Indeed, intraindividual differences in ileum effluent are described to be , o verall, higher than in fecal samples, and show daily fluctuations, possibly impacted by diet or other confounders. Over a period of 9 da ys , about 44% similarity was observed in ileostomy effluent, while fecal samples are found to have about 92% similarity over a period of minimal 2 months (Rajili ć-Stojanovi ć et al. 2009, Booijink et al. 2010 ). The Streptococcus population shows high diversity in ileostomy effluent as se v en genetic linea ges (not all within one sample) wer e observ ed closel y r elated to S. saliv arius , S. thermophilus ( S. salivarius species group), and S. parasanguinis ( S. mitis species group). In contrast, the Veillonella gen us re presents less diversity, as all belonged to the same genetic lineages (Van den Bogert et al. 2013 ).

Host-bacterial-diet interaction in nutrient digestion and absorption
The complex microbial community residing in the small intestine encompasses diverse metabolic activities, pivotal for the digestion of nutrients, otherwise not accessible for host absorption. In this section, we describe how the host digestive processes are supported by microbial metabolic functions for the digestion of carbohydr ates, pr oteins, lipids, and some micronutrients, hence contributing to nutrient absorption.
The distal small intestinal microbiota undergoes the hydrolysis of nondigestible carbohydr ates, r esistant to host's digestive enzymes, thr ough carbohydr ate-activ e enzymes (CAZymes), with SCFAs as end-products. To date, in the human gut microbiome, 15 882 different CAZyme genes have been identified (Kaoutari et al. 2013 ), classified based on amino-acid sequence similarities, into five families: (i) glycoside hydrolases (GHs), the most pr e v alent among the gut bacteria, responsible for the hydrolysis and/or tr ansgl ycosilation of the glyosidic bonds; (ii) glycotr ansfer ases (GTs), catal yzing the gl ycosidic bond formation by transferring a moiety from an activated donor molecule to specific donor molecules; (iii) pol ysacc haride l yases (PLs), a gr oup of 31 enzymes whic h cleav e ur onic acid-containing pol ysacc haride c hains; (iv) carbohydr ate ester ases (CEs), whic h r emov e esterbased modification in mono-, oligo-and pol ysacc harides, hence facilitating the action of GHs on complex pol ysacc harides; and (v) carbohydrate-binding modules (CBMs), often associated to other CAZyme and without enzymatic activity per se , are dedicated to facilitate the enzyme-substrate interaction and potentiate the enzymatic activity (Davies et al. 2005, Cantarel et al. 2009 ) ( http: //www.cazy.or g ), extensiv el y r e vie w ed in War dman et al. ( 2022 ).
By functional metagenomic screening of a fosmidic library constructed from ileal mucosa, Patrascu et al. ( 2017 ) demonstrated that the ecosystem in the human ileal mucosa, harbours the fibr ol ytic potential to catabolize complex and diversified plant cell wall pol ysacc harides. In particular, they identified 25 enzymes dedicated to carbohydrate metabolism (21 GH, 2 CE, and 1 GT) from Bacteroides and Eubacterium related species, mainly responsible for plant-cell-wall degradation, but also starch and fructosebased saccharide degradation in the ileum (Patrascu et al. 2017 ). Mor eov er, by compar ativ e functional anal ysis, se v er al pathways and functions for carbohydrate uptake and metabolism are described as enriched in the small intestinal metagenome, compared with those of fecal metagenomes, suggesting that uptake and fermentation of av ailable carbohydr ates in the small intestinal lumen contributes to the maintenance of the resident microbiota. In particular, genes involved in the expression of several sugar phosphotr ansfer ase systems (PTS), enzymes r elated to central metabolism (e.g . pentose phosphate pathway), and fermentation pathwa ys (e .g . lactate and propionate fermentation) are highl y enric hed in the small intestinal micr obiome (Zoetendal et al. 2012 ). Zoetendal et al. ( 2012 ) also re ported that genes link ed with PTS transcription are mainly expressed by streptococci, suggesting that these bacteria are the main utilizers of available carbohydrates in the small intestinal lumen. Coher entl y, genes assigned to the butyrate fermentation pathway are reported in the human small intestinal microbiome (Zoetendal et al. 2012 ), although limited compared to the enrichment described in the metagenome of the large intestine, which is in line with the higher abundance of butyrate producers in colon (Pryde et al. 2002 ).
The microbial utilization of amino acids starts in the small intestine and the bacterial composition associated with protein metabolism has been described. Indeed, a shift in ileal microbiota composition is reported in response to the intake percentage (Qiu et al. 2018 ) and sources (Kar et al. 2017 ) of dietary proteins, in animal models. In particular, bacterial richness and SCFAs concentration in the ileum decrease with the reduction of protein intake (Qiu et al. 2018 ). Additionall y, the pr oportion of Clostrid-ium_sensu_stricto and Esc heric hia-Shigella decr eases and increases, r espectiv el y, with the reduction of protein intake (Qiu et al. 2018 ). As well, the source of protein (peptides or amino acids) impact the Lactobacillus colonization dominance in the pig small intestine, leading to the pr e v alence of Lactobacillus am ylovorus in peptide-ric h environment (Jing et al. 2022 ). Furthermore, through culturing appr oac hes, it was reported that Klebsiella spp ., Streptococcus spp ., E. coli , and Mitsuokella spp. from the porcine small intestine utilize amino acids at a rate of 50%-90% over 24 h, suggesting a potentiall y r ele v ant impact on the ov er all small intestinal ecology (Dai et al. 2010 ). Although similar studies on differ ential pr otein intake have not yet been conducted on humans, to our knowledge, these results suggest an important role of small intestinal microbiota in protein utilization. Indeed, in human ileal aspirates, enzymes related to amino acid metabolism are highly enriched, compared to fecal samples (Zoetendal et al. 2012 ). It is possible that, considering the r a pid host uptake of peptides and amino acids present in the small intestinal lumen, the de novo synthesis of amino acid by the small intestinal microbiome is stimulated (Zoetendal et al. 2012 ).
The lipid metabolism in the human gut is indir ectl y linked to the small intestinal microbiota activity, that convert conjugated primary bile acids from the host into deconjugated analogues and subsequentl y conv ert the primary into secondary bile acids . T he deconjugation of primary bile acid reaction is catalyzed by the activity of the BSH.
From a construct metagenomic dataset of sequences from different cohorts worldwide, Song and collea gues r eported that BSH sequences are distributed in 591 intestinal bacterial strains (Song et al. 2019 ). Indeed, BSH activity provide an ecological adv anta ge by enhancing the resistance to the conjugated bile acids and promotes the survival and colonization in the intestine (Jones et al. 2008 ). Those unconjugated primary bile acids are converted into secondary bile acids following a C-7-epimerization and a 7-alphadehydroxylation, encoded by the bile acid-inducible (bai) baiB gene (Ye et al. 1999 ) by Bacteroides , Eubacterium , and Clostridium genera. Within the Actinomycetota , Bifidobacterium species possess two major BSH enzyme types: A and C with a highest specificity for the glycine-conjugated bile salts over taurine-conjugated forms (Kim et al. 2004 ).
Upon lipid digestion, fatty acids are taken up by enterocytes via both protein-mediated and protein-independent transport.
Along with their role in the digestion of carbohydrates, proteins, and bile salts metabolism, intestinal microbiota can synthesize certain vitamins, notably B group vitamins and vitamin K (Hill 1997 ). These vitamins, important for bacterial metabolisms, also have a metabolic and physiological significance in humans. In fact, humans exposed to low vitamin K diet during 3-4 weeks did not de v elop vitamin deficiency, in contrast to those administered with a large spectrum antibiotic (Frick et al. 1967 ). Ho w ever, the majority of the studies focus on the ov er all gut and fecal micr obial comm unity and, to our knowledge, no r eport specificall y focused on bacteria isolated from small intestine . Nonetheless , by genome annotation of 256 human gut bacteria, the biosynthesis pathways for eight B-vitamins (B8, B12, B9, B3, B5, B7, B2, and B1) was predicted in 40%-65% of the analyzed genomes and the majority of these pr edictions matc hed published experimental data (Magnúsdóttir et al. 2015 ).
Animals are incapable of synthesizing cobalamin (vitamin B12), and thus r el y on dietary sources of cobalamin. In humans, cobalamin uptak e tak es place in the ileum. As suc h, micr obial vitamin B12 produced further along the intestine is not absorbed by the host but, instead, used to synthesize other corrinoids, not used by the human. The bacterial synthesis of cobalamin can be performed either aer obicall y or anaer obicall y but human gut microbiota pr efer entiall y uses the anaer obic r oute (Ma gnúsdóttir et al. 2015 ). It has been shown that cobalamin biosynthetic pathways involv e nearl y 30 differ ent enzymes, including hemBCD , cbi , and cob genes (Taranto et al. 2003, Piw o w arek et al. 2018. Within the human gut microbiome, the synthesis of cobalamin was predicted in most of Fusobacteriota, r ar e in Actinomycetota and Pseudomonadota while missing in half of the genomes in the Bacteroidota and Bacillota phyla (Magnúsdóttir et al. 2015 ). Besides bacterial vitamin B12 production, some bacteria also utilize vitamin B12, essential for enzyme cofactors and gene regulations (Degnan et al. 2014, Wexler et al. 2018. Folate (vitamin B9) in the gut, involved in major metabolic pathways such as amino acid conversion and nucleotide synthesis, mostl y deriv es fr om tw o sour ces: one from the dietary pr oducts, whic h is absorbed by the small intestine enterocytes, and another, which is the by-product of dietary fibers fermentation by the gut microbiota and, is absorbed then in the colon. De novo folate biosynthesis involves both 6-h ydroxymeth yl-7,8dihydr opterin pyr ophosphate (DHPPP) and par a-aminobenzoic acid (pABA) as precursors (Rossi et al. 2011 ). By systematic analysis of the ov er all human microbiota genome, it is reported that folate biosynthesis pathway, while r ar e in Actinomycetota and Bacillota genomes, is present in almost all Bacteroidota, Fusobacteriota, and Pseudomonadota genomes (Magnúsdóttir et al. 2015 ). This suggest a r ele v ant impact on host's folate metabolism, although report focusing specifically on human small intestinal mi-crobiome and folate synthesis are not yet a vailable . Lastly, biotin (vitamin B7) is r equir ed for normal cellular function and de v elopment, yet humans and other mammals are not able to synthesize it. Exogenous biotin needed to satisfy the nutritional r equir ement is provided through two sources: the diet and the microbial production. In the human small intestinal microbiome, biotin synthesis genes are described, phylogenetically linked primarily to Pseudomonadota but also associated with Bacillota and Bacteroidota (Zoetendal et al. 2012 ). Mor eov er, since biotin absorption by epithelia takes place in the intestine (Said 2009 ), it is plausible that small intestinal bacteria may contribute to the host's biotin supply (Zoetendal et al. 2012 ).

Disruption of small intestinal homeostasis in nutrient balance
It is evident that the small intestine is pivotal in nutrient digestion and absor ption. As suc h, an y disruption of its homeostasis can lead to an alter ed micr obial ecology and metabolic activity to w ar d the dietary component present in the lumen. Consequentl y, the alter ed small intestinal envir onment may compr omise nutrient absorption and ultimately result in malnutrition. In this context, an expanded knowledge on small intestinal ecology in healthy and disease conditions is crucial to define or adapt ther a peutic a ppr oac hes to impr ov e nutritional status . T he conditions impacting small intestinal homeostasis and subsequent nutritional status can be div erse, r anging fr om micr obial dysbiosis and pathogen infections to inflammatory diseases and anatomic changes. In the next section we will describe two small intestinal conditions, namely short bo w el syndrome (SBS) and small intestinal bacterial ov er gr owth (SIBO), as examples to highlight the drastic impact from physical resection or bacterial dysbiosis on host nutrient status.

Short Bo w el Syndrome
Short Bo w el Syndrome (SBS) is a r ar e and se v er e condition defined by an extensive loss of small intestinal surface . T he most frequent causes of SBS, in adults, are mesenteric ischemia, Crohn's disease, r adiation enteritis, postsur gical intr a-abdominal adhesions, and postoper ativ e complications (Pironi et al. 2006 ). In adults, where normal small intestinal length is a ppr oximatel y 600 cm, SBS is defined by a remaining small bo w el in continuity of less than 200 cm and it is classified based on anatomical, pathophysiological, and postoper ativ e e volution criteria (Pir oni et al. 2015 ). According to anatomical criteria, three types of SBS are defined: (i) type I, endjejunostomy with no colon in continuity; (ii) type II, jejuno-colic anastomosis, where the remnant jejunum is in continuity with part of the colon; and (iii) type III, jejuno-ileal anastomosis with ileo-cecal v alv e and the intact colon in contin uity (Je ppesen 2014 ) ( Table 5 ). The extensive removal of small intestinal surface results in intestinal failure, defined as the insufficient digestion of nutrients and hence requiring parenteral nutrition to sustain the metabolism and pr e v ent malnutrition and dehydr ation (Pir oni et al. 2015 ). The real incidence and prevalence of SBS is unclear, due to the lack of reliable patient databases but, based only on the patients receiving home parenteral nutrition (5-80 per million population in Europe), SBS is estimated to contribute for 75% of them (O'Keefe et al. 2006, Pironi et al. 2006, Jeppesen 2014. After surgery and first acute intestinal failure stage, a spontaneous process termed adaptation follo ws, characterized b y morpholog ical, histolog ical, and metabolic changes in the intestinal mucosa to compensate for the reduced absorptive area (Jeppe-   Billiauws et al. 2018 ). The most notable response occurs in the residual ileum, with increased enterocyte pr olifer ation as soon as 8 h after resection in a SBS animal model (Dahly et al. 2003 ), indicating the importance of preserving residual ileum and colon in continuity. When the distal ileum and colon are removed, a loss of gastric emptying inhibition (ileal brake) is observed immediately after the intestinal resection, because of the extensive anatomical alteration and the crucial contribution of the small intestine in gastric emptying (through the production of peptide hormones ghrelin and GLP-1), resulting in fast transit of food, gastric acid, and bile acid (Nightingale et al. 1993 ). Conv ersel y, in patients with colon in continuity, gastric emptying and transit time are comparable with nonresected condition (Nightingale et al. 1993 ). Suc h alter ations in tr ansit, along with the incr ease in oxygen le v el, gastric and bile acids concentrations in the remnant bo w el, str ongl y influence the intestinal ecology, and result in a significant decrease in OTU counts and α-diversity and, overall, a diverse micr obial comm unity structur e between healthy and SBS conditions and among different SBS types (Huang et al. 2017, 2020, Zeichner et al. 2019, Hu et al. 2021 ) (Summarized in Table 6 ). Inter estingl y, a gr eater pr oportional abundance of fecal Enterobacteriaceae and lactobacilli were correlated with a longer and shorter parenteral nutrition dur ation, r espectiv el y (Huang et al. 2017 ), suggesting that the intestinal ecology upon extensive small bo w el resection may contribute to the evolution of SBS and the ac hie v ement of nutritional autonomy. Coher entl y, type III SBS patients, for whom ileocecal v alv e and colon ar e pr eserv ed, hav e r elativ el y quic ker intestinal adaptation and better clinical outcome. Ne v ertheless, overload of Lactobacillus in SBS can also be deleterious, in pr e v enting implantation of other bacteria such as Clostridium clusters (Joly et al. 2010 ). It is noteworthy that the roles of Lactobacillus in SBS patients are complicated and variable. While lactobacilli abundance is associated with shorter par enter al nutrition duration in human (Huang et al. 2017 ), the capacity of certain lactobacilli to produce D -lactic acid, coupled with the reduced absorption potential, exposes SBS patients to a high risk of D -lactic acidosis (Mayeur et al. 2013 ) and neurological disorders (Mayeur et al. 2013 , Kowlgi andChhabra 2015 ).
In SBS patients, the shifted fecal microbial ecology is reflected in altered functional profiles of carbohydrate and amino acid metabolisms, along with the depletion in anaerobic Bacillota and pr e v alence of facultativ e anaer obic Pseudomonadota (Huang et al. 2017 ). Coher entl y, dietary supplementation with the pyrimidine precursor orotate and uracil stimulated jejunal ada ptiv e gr owth in a SBS animal model (Evans et al. 2005 ). As well, other microbial metabolic pathways reduced in SBS patients include methane metabolism and o xidati ve phosphorylation, suggesting insufficient energy harvest (Huang et al. 2017 ).
Furthermor e, the c hanges in the intestinal environment after intestinal resection expose SBS patients to the development of SIBO, a condition c har acterized by excessiv e number of bacteria in the small intestine (further detailed in the next section).
Ov er all, studies on SBS animal models and patients indicated that ac hie ving intestinal ada ptation and n utritional autonom y is not only influenced by the extensiveness of the r emov ed small intestine (Berlin et al. 2019, Sun et al. 2020 ) but also, by the remaining colon and early plasma citrulline concentrations (Amiot et al. 2013 ). Ho w e v er, considering the e vidences on the r ole of small intestinal microbiota in nutrient digestion and absorption, we cannot exclude that the ecological alterations in the remnant intestine have an impact also on host's adaptation, although not yet elucidated, to our knowledge. Coher entl y, par enter al nutrition supplemented with SCFAs and, in particular, butyrate at physiological concentrations, is described to improve structural and metabolic adaptation (i.e . increased villus height, number of crypt  --cells and plasma GLP-2 concentrations) in the small intestine, acceler ates ada ptation and potentiall y shortens the period of full adaptation, in a SBS piglet model (Jeppesen et al. 2000, Bartholome et al. 2004, Filippi et al. 2021 ). Furthermore, after reinfusing proximal jejunostomy output into the distal part of the small intestine Liu et al. ( 2016 ) observed a shorter par enter al nutrition period, protected integrity of the intestinal mucosa and increased nutrient absorption, in particular when the ileocecal valve was pr eserv ed. Although the mec hanisms for this amelior ation wer e not fully elucidated, it is plausible that by reinfusing jejunostomy output, also jejunum bacteria are added into the distal intestine, leading to a modulation of the ov er all ecology. In terms of nutritional r equir ement of SBS patients, the gastrointestinal anatomy after resection is crucial to personalize the nutritional needs. For example, while jejuno-colonic anastomosis patients should r eceiv e about 30-35 kcal/kg/day of complex carbohydrates with soluble fibers, this is not necessary for patients without remnant colon, but relevant is the supplementation in long-c hain trigl ycerides (Buc hman et al. 2003 ). Along with ensuring the a ppr opriate n utrient intak e, the curr ent ther a peutic a ppr oac h for SBS patients r elies on antisecr etory, antidiarrheal, and antimotility (e.g . loperamide) drugs and somatostatin to reduce intestinal loss (Vílchez-López et al. 2021 ), proton-pump inhibitors and hormonal ther a pies with GLP-2 analogues (i.e . teduglutide) (Vorre et al. 2022 ), to maximize absorption and effectiv el y r educe par enter al nutrition r equir ements (Sc hwartz et al. 2016, Lam et al. 2018, Joly et al. 2020, although not univ ersall y effectiv e (Billiauws and Joly 2019 ). In particular, besides the intestinotrophic effect on the intestinal epithelium, GLP-2 treatment is described to partially ameliorate also the intestinal bacterial dysbiosis of SBS rats by significantl y downr egulating the r elativ e abundance of Proteus genus and increasing the relative abundance of Clostridium genus in SBS rats (Hu et al. 2021 ).
In this challenging context, it emerges the need for a personalized and multidisciplinary approach for SBS management. To this aim, one European (INTENS, ID 668294) and one US (Clinical trial ID NCT03530852) innov ativ e pr ojects ar e exploring alternativ e str ategies to impr ov e the quality of life of SBS patients by creating a functional small bo w el from patients' own cells for autologous transplants and designing a fat predigestion device, respectiv el y. As suc h, this c hallenging and innov ativ e a ppr oac hes have the potentiality to drive the SBS management to w ar d more patient-based ther a pies, in the near futur e.

Small Intestinal Bacterial Overgrowth
Small Intestinal Bacterial Ov er gr owth (SIBO) manifests in the small intestine and, as the name indicates, is associated with an excessi ve n umber of bacteria in the small intestine causing gastrointestinal complaints . Despite the fact that natur al micr obial counts in the duodenal-jejunal ar ea hav e been reported to be approximately 10 3 CFU/ml, the cut-off to define SIBO differs between studies. In the past, SIBO was defined by a microbial load ≥ 10 5 CFU/ml, but recent validations based on healthy controls urged to consider a lo w er cut-off for the diagnosis (Khoshini et al. 2008, Rezaie et al. 2017. SIBO patients usually experience bloating, diarrhea, abdominal discomfort, and in more severe cases steatorrhea, weight loss, anemia, nutritional deficiencies (e.g . vitamin B12 deficiency), and/or mucosal inflammation (Lema et al. 2020, Quigley et al. 2020 ). Yet, no association was found between the ov er all incr eased micr obial load and typical SIBO gastrointestinal symptoms, in a human SIBO cohort  ).The pr e v alence of SIBO v aries among r eports between 2% and 22%, depending on the studied disease and the diagnostic method (Lakshmi et al. 2010, Fasano et al. 2013, Ierardi et al. 2016, Niu et al. 2016, Ricci et al. 2018, Wu et al. 2019 , Ko w alski and Mulak 2022 ).
Predisposing factors for SIBO development can be anatomical, pharmacological, or pathological abnormalities (i.e . intestinal stasis, decreased gastric acid production, pancreatic or biliary secretion deficiency, and an increased ileocecal valve reflux) or a malfunctioning gastrointestinal immune response resulting in a reduced microbial barrier or microbial clearance (Chander Roland et al. 2017, Quigley 2019. Additionall y, a ging, female gender, and proton pump inhibitors have also been proposed as predisposing factors, although not consistentl y (Duk owicz et al. 2007, Choung et al. 2011, Erdogan and Rao 2015, Su et al. 2018, Shin et al. 2019, Ghoshal et al. 2022. A complicating feature for the diagnosis is that these risk factors often appear in other diseases, making SIBO intertwined with other pathologies among which gastr oenter ological disorders (e.g . inflammatory bowel disease, irritable bo w el syndr ome, SBS, nonalcoholic fatty liv er disease, cirrhosis , diabetes , and cystic fibr osis (Ier ardi et  SIBO diagnosis is usually based on direct or indirect quantitativ e measur ements, thr ough aspir ate culturing a ppr oac h or br eath tests, r espectiv el y. In clinical pr actice, the simplest and most widely used and available tests are hydrogen and methane breath tests, used as indirect read-outs for the microbial gas production in the small intestine, instead of in the colon under normal conditions. At histological le v el, no r emarkable differ ence was observed in biopsies, besides a lo w er villous to crypt ratio ( < 3:1) in SIBO individuals compared to controls (Lappinga et al. 2010 ). Labor atory r esults can indicate SIBO by increased folate, due to increased bacterial synthesis, or decreased vitamin B12 levels. Vitamin B12 deficiency can be the result of increased bacterial consumption by the higher bacterial load, competitive binding with cobalamin-like bacterial metabolites or damage of binding sites (Quigley, Murray and Pimentel 2020 ).
Aside fr om quantitativ e dia gnosis, qualitativ e a ppr oac hes c haracterizing the small intestinal community in SIBO might impr ov e SIBO management as gastrointestinal symptoms in SIBO were associated to a high load of so-called disruptor taxa but not to an ov er all high microbial load  ). Instead, a high microbial load might be confounded by dietary habits. Indeed, in a small study cohort, nonsymptomatic healthy individuals consuming a high fiber diet positiv el y corr elated with SIBO, based on positive duodenal cultures, but only the switch to a low fiber high simple-sugar diet evoked gastrointestinal complaints (Saffouri et al. 2019 ).
To date, few studies characterized the small intestinal community in these patients and reported the associations of SIBO with changes in the small intestinal ecology (summarized in Table 7 ). By comparing duodenal aspirates of SIBO and non-SIBO individuals, Barlow et al. ( 2021 ) identified se v en disruptor bacterial taxa, containing human pathogenic str ains, that a ppear to displace common strict anaerobes . T he taxa pointed out were Enterobacteriaceae , Esc heric hia-Shigella , Clostridium sensu stricto_1 , Enterococcus , Romboutsia , Aeromonas , and Bacteroides. Among those, Enterobacteriaceae and Esc heric hia-Shigella wer e most commonl y found in SIBO samples  ) and, when present, negatively affected the network connectivity in SIBO individuals (Valiente-Banuet et al. 2020 ).
Ho w e v er, the efforts in c har acterizing the small intestinal ecology in SIBO patients may result in diverging results due to the difference in sampling locations across different study cohorts (Leite et al. 2020a (Table 7 ). Additionally, inconsistencies among SIBO studies might also be related to the type of SIBO, often not indicated. Shin et al. ( 2019 ) described two subtypes of SIBO, based on the type of bacteria ov er gr owing: upper aer odigestiv e tract SIBO and coliform SIBO, related to upper gut and oropharynx or colon-like bacteria, r espectiv el y (Table 7 ). When comparing these SIBO groups a significant difference in β-diversity is described. Additionall y, when compar ed to non-SIBO individuals, significant taxonomical differences were described only between coliform SIBO type and non-SIBO individuals (Shin et al. 2019 ). These results indicate the need for a better microbial characterization of SIBO patients, including a more tailored discrimination between the two aforementioned SIBO groups, potentially determining differ ent mana gement str ategies, as also suggested by other authors (Saffouri et al. 2019. Current methods to manage SIBO are mainly empirical and include antibiotics and a change in diet. The goal of a conventional ther a peutic a ppr oac h with antibiotics is to modulate, and at least partially inhibit the microbial community to eventually improve symptoms (Quigley et al. 2020 ). Rifaximin, a broad-spectrum nonabsorbable antibiotic, often prescribed for SIBO, shows in vitro efficacy against bacteria often associated with SIBO among which Klebsiella spp., E. coli , and Enterobacter spp. (Pistiki et al. 2014 ) making it possibly more effective in coliform SIBO. Although proven in vitro and in vivo efficacy, with SIBO eradication in about 70.8% of patients, (Gatta et al. 2017, Barkin et al. 2019, r eoccurr ence of SIBO after antibiotic treatment is common. In a study group treated with rifaximin, about 44% r ela psed within 9 months (Lauritano et al. 2008 ). The high r ela pse number might be due to the underl ying cause, whic h is not al ways possible to cur e. Alternativ el y, dietary intervention with a reduced intake of poorly absorbable short-c hain carbohydr ates (fermentable oligosacc harides, disacc harides, monosacc harides, and pol y ols) as in lo w-FODMAP diets, used to treat SIBO patients, ho w e v er, mor e e vidences on its effectiv eness and mec hanisms ar e r equir ed (Srisukthav eer at et al. 2021, Biesiekierski and Tuck 2022, Wielgosz-Gr oc howska et al. 2022. Additionally, the use of probiotics in SIBO was shown effective in SIBO eradication and symptom relief according to a metaanalysis (Zhong et al. 2017 ), yet there is no consensus and little detail regarding the used probiotic strains and doses. In an attempt to a ppr oac h SIBO tr eatments in a mor e holistic and ecosystemoriented a ppr oac h, fecal micr obial tr ansplants ar e also consider ed for modulating the gastrointestinal community. This has already been pr ov ed as effectiv e str ategy to tr eat r ecurr ent Clostridium difficile infections (Fuentes et al. 2014 ). In SIBO, fecal microbial transplants resulted in an improvement of gastrointestinal symptoms (Xu et al. 2021 ), though mor e r esearc h is needed to confirm the result of this trial.

Models of the small intestinal environment
To untangle the complexity of the intestinal environment, in vivo , ex vivo , and in vitro a ppr oac hes hav e been designed and emplo y ed to address the microbial ecology, intestinal mucosa, and the hostbacteria-nutrition interactions. In this section, we discuss the models used to mimic the adult healthy small intestinal environment and emplo y ed to study host-bacteria-nutrition interaction (Fig. 2 ).

Human studies
In humans, medical practices like biopsies, luminal brush, catheter aspiration, and intelligent wireless capsules, allow to access and sample the small intestine in vivo . Ho w e v er, when sampling in in vivo conditions, the inv asiv eness of the medical practice, might pr e v ent sampling from healthy indi viduals. Ad ditionally, study subjects are often of higher age or with an underlying disease, whic h ur ges them to perform an endoscopy , surgery , or colectomy (r emov al of the colon) giving r esearc hers the opportunity to sample. Consequently, some bias linked to the sampling method and study population, is introduced and makes the comparison of results from different cohorts challenging. Alternativ el y, samples fr om sudden-death victims and ileostomy patients offer valuable information to investigate in vivo features of the small intestinal environment with limited cross-contamination. Individuals with an ileostomy (i.e . no colon) allow repeated sampling of small intestinal content over time (Booijink et al. 2010, Zoetendal et al. 2012, Jonsson 2013, Van den Bogert et al. 2013, Van Trijp et al. 2020. Ho w e v er, in these cases, the risk of inaccur ate r epr esentation of the small intestinal microbiota exists, due to the abnormal anatomy of the gastr ointestinal tr act (i.e . absence of colonic reflux) and suggested increased oxygen penetration (Hartman et al. 2009, Booijink et al. 2010, Zoetendal et al. 2012 ). Yet, a comparison of the small intestinal microbiota from healthy and ileostom y indi viduals sho w ed ileostomy effluent clustered closely to jejunal samples, and ileum samples from healthy adults positioned between the ileostomy effluent and fecal samples (Zoetendal et al. 2012 ). Mor eov er, the penetr ation of oxygen is suggested to be limited as strict anaer obes, suc h as Ruminococcus gnavus and Coprococcus eutactus , are still detected in ileostomy effluents (Booijink et al. 2010, Zoetendal et al. 2012. Sampling sudden death victims is also prone to samplingbias (Hayashi et al. 2005 ). In particular, these studies often consist of elderly individuals, and aging has been shown to impact the small intestinal microbiota with reduced α-diversity and incr eased Pseudomonadota le v els ( Enterobacteriaceae , Esc heric hia , and Klebsiella )  ). In addition, post mortem microbial changes can occur as a result of, among others, hypoxia and nutrient depletion (Tuomisto et al. 2013 ). Yet, a study b y P echal et al.
( 2018 ) demonstrated that microbiomes from mouth and rectum can still r epr esent the ante mortem health conditions within 24-48 h of death, but, to our knowledge, this has not been validated for small intestinal composition. As such, a timely sampling (within hours after death, as suggested by Hayashi et al. 2005 ), is crucial to limit the variability in the microbial composition.
An alternative to the aforementioned sampling conditions, is offer ed by dir ectl y accessing the small intestine, during gastric surgery or a cystectomy (Villmones et al. 2018(Villmones et al. , 2022. In patients undergoing a cystectomy, the distal ileum community was found to be similar to the oral community (Villmones et al. 2018 ). These findings are in line with what was described in sudden death victims (Hayashi et al. 2005 ), but contradicting other studies sampled thr ough colonoscopy, whic h found incr eased pr esence of Bacteroidota, with higher similarity to the colon community (Wang et al. 2005, Vuik et al. 2019, Nagasue et al. 2022 ). Additionally, cross-contamination during sampling, can also explain for these differences. Indeed, while in Hayashi and Villmones reports they accessed the small intestine dir ectl y during sur gery (Hayashi et al. 2005, Villmones et al. 2018, in other studies the authors made use of antegrade or retrograde endoscopy (Wang et al. 2005, Zoetendal et al. 2012, Dlugosz et al. 2014, Li et al. 2015, Sundin et al. 2017, Saffouri et al. 2019, Seekatz et al. 2019, Vuik et al. 2019, Fan et al. 2020, Kashiwagi et al. 2020, Vaga et al. 2020, Leite et al. 2020b, Nagasue et al. 2022, with higher c hance of cr oss-contamination by the densel y populated mouth or colon envir onment. Furthermor e, while pr epar ation for an antegrade endoscopy usually requires only an overnight fasting period, for r etr ogr ade endoscop y (colonoscop y), bo w el pr epar ation by use of a laxative is often demanded to enable a clear view during colon inspection (Dlugosz et al. 2014, Sundin et al. 2017, Saffouri et al. 2019, Vuik et al. 2019, Fan et al. 2020, Kashiwagi et al. 2020, Vaga et al. 2020. Bo w el pr epar ation is a w ell-documented sour ce of bias for the description of intestinal microbial composition in vivo in humans (Shobar et al. 2016, Nagata et al. 2019. Ho w e v er, the extent to which it affects the small intestinal ecology has not yet been reported. As described, all above mentioned techniques are not suitable for healthy controls, due to the inv asiv eness of the pr ocedur e and the associated ethical restrictions. To overcome this limitation, intelligent wireless capsules have been developed to monitor parameters in the intestinal environment and, in some cases, sample the lumen content, in healthy individuals without the need for inv asiv e pr ocedur es (Tang et al. 2020 ). Yet, the high costs, the potential cross-contamination with sampling sites up-or downstream and the sample preservation until fecal excretion, still limit their applicability. To preserve the sample during the multiple hours the pills remain at body temperature, Rios-Morales et al. ( 2021 ) de v eloped a quenc hing a gent to stabilize microbial composition, fibers fermentation, and SCFAs production for 48 h, enabling studying the small intestinal microbial composition, yet limiting downstream culturomics approach.

Animal studies
To study physiological processes, animal models are often considered a suitable alternative to human in vivo studies, offering the possibility to anal yze m ultiple endpoints and the impact of genetic modifications. Additionall y, germ-fr ee animals pr ovide the opportunity to study specific microbial populations in hostmicr obiota a pplications. Adv anta ges and limitations of monogastric animal models, such as nonhuman primates , pigs , and rodents for the study of small intestinal ecology and host-bacteria inter action, ar e discussed here.

Nonhuman primates
Nonhuman primates are our closest r elativ es and, ther efor e, a highl y r ele v ant r esearc h animal model. Ho w e v er, the expensiv e and difficult housing and husbandry of nonhuman primates, along with strict ethical regulations, limit a more widespread employment as in vivo model (Walker and Eggel 2020 ). In the context of host-micr obe inter action r esearc h, differ ent nonhuman primate models exist: (i) wild nonhuman primates, with distinct micr obial ecology, differ ent to humans; (ii) ca ptiv ated nonhuman primates (Firrman et al. 2019, Yuan et al. 2020, Yan et al. 2022, showing micr obiota r esembling mor e to human micr obiome than wild nonhuman primates (Clayton et al. 2016 ). To assess the impact of multiple diets on the human micr obiome, ca ptiv ated nonhuman primates fed a human-like diet during the study, are of-ten emplo y ed (Amato et al. 2015, Na gpal et al. 2018, Ne wman et al. 2021. Ho w e v er, despite the genetic proximity with humans, not all nonhuman primate species are suitable for extrapolation to human gut micr obial r esearc h. For example, compar ed to humans ther e ar e marked differ ences in the intestinal microbial community for ca puc hin monk e ys ( Cebus apella ), (Firrman et al. 2019 ), c har acterized by lower r elativ e abundance of Streptococcus in small intestine. In baboons as well, Spirochaetes , not commonly found in the human gut (Angelakis et al. 2019 ), Bacillota, Bacteroidota, and Pseudomonadota (in consecutive order) are described to be the main species in the small intestine (duoden um, jejun um, and ileum) (Yuan et al. 2020 ). Conv ersel y, in rhesus macaques, there is a lo w er pr esence of Spiroc haetes in the m ucosa and a pr edominance of Bacteroidota, Bacillota, and Pseudomonadota in the gut (Yasuda et al. 2015 ). Nonetheless, the r ele v ance of nonhuman primates for direct testing host-microbe-nutrition interactions was questioned, as similar diets are described to differ entl y affect the human and nonhuman primate gut microbiota based on the rectal microbiota (Amato et al. 2015 ).

Pigs
Pigs are an alternative in vivo model, with high relevance to human gut anatomy , physiology , nutritional r equir ements, and immune system (Meurens et al. 2012 ). One of the advantages of the pig model is the possibility for cannulation (often performed in the terminal ileum), enabling multiple sampling points over time within the gastr ointestinal tr act, without need for euthanasia (Metzler-Zebeli et al. 2010, Shen et al. 2020. To study the host-microbe-nutrition inter action, m ultiple a ppr oac hes in the pig exist, among which: (i) conventional pigs (i.e . with conventional pig feed and microbiome), as beside a relevant model for human health, there is a concern in pigs health by their economic value (Zhao et al. 2015, Crespo-Piazuelo et al. 2018 ); (ii) germ-free pigs which offer the possibility to study specific bacterial communities including the complex human microbiota by fecal microbial transfer (human-microbiota associated pigs), though no comparison to the small intestinal microbiota has yet been published to our knowledge (Aluthge et al. 2017, 2020, Fischer et al. 2017. To study microbenutrition interaction in the human context, pigs can be fed with a human-like diet, comprising foods commonly consumed by humans (Hooge v een et al. 2020, Shen et al. 2020, Xu et al. 2020 ). In con ventional pigs , the small intestine (duodenum, jejunum, and ileum) is dominated by Bacillota and Pseudomonadota (Zhao et al. 2015, Crespo-Piazuelo et al. 2018, while the presence of Pseudomonadota increases throughout the small intestine reaching maxim um pr esence in the ileum, similar to what was observed in humans. Yet, at lo w er taxonomic le v els discr epancies with the human micr obial comm unity r emain pr esent (Cr espo-Piazuelo et al. 2018, Nagasue et al. 2022 ). One element that should be taken into account regarding intestinal stressors that impact resident small intestinal microbiota, is the difference in bile salt profiles between pigs and humans. Indeed, the two major constituents of human bile salts , cholic , and deoxycholic acids , ar e pr esent at lo w er proportion in pigs . Con v ersel y, hyoc holic and ursodeoxyc holic acids, minor bile salts in humans, r epr esent a higher fraction in pigs (Spinelli et al. 2016 ).

Rodents
The use of rodents (mostly mice and rats) in gut microbiome studies is well-established thanks to their reduced housing and husbandry costs (in comparison to nonhuman primates and pigs), high r epr oductiv e r ates , short life cycle , and possibility for ge-netic manipulations . Moreo ver, the a vailability of well-established transgenic mouse strains for several human diseases, has driven the incr eased inter est of this model for r esearc h, during the last decade (Flemer et al. 2017 ). In contrast, the bigger size of rats permits increased sampling (tissue and feces) and they are reported to sustain human-like fecal micr obiota pr ofiles better than mice (Flemer et al. 2017, Lleal et al. 2019. In host-bacteria-nutrition interaction studies, multiple rodent models can be considered, by adjusting genetics (e.g . gene knockouts), gut microbiota composition (e .g. germ-free , mono-colonized gnotobiotic , humanized gnotobiotic , specific pathogen free, or con ventional) or en vironment (e .g . diet, antibiotics) (Martinez-Guryn et al. 2018, Todorov et al. 2020, Tuganbaev et al. 2020, Escoto et al. 2021. Choosing the adequate model and housing condition for the scientific question to address, is crucial. Indeed, animals born and raised in total sterile conditions, show altered physiology, including extensive deficit in the de v elopment of mucosal immunity (Round and Mazmanian 2009 ), reduced number of Paneth cells in (jejunal) mucosa associated with decreased concentration of antimicrobial peptide secretion (Schoenborn et al. 2019 ), and a major susceptibility to chemically induced epithelial damages (Hayes et al. 2018 ). Conv ersel y, the antibiotic-induced germ-free condition, may result in only a partial depletion of bacterial species, induces antibiotic-resistance, and impacts epithelial cell metabolisms (r e vie w ed b y Kennedy et al. 2018 ).
Instead of regular fecal microbial transplants for humanization of germ-free mice, Li et al. ( 2020a ) explored the introduction of whole intestinal micr obial tr ansplants, comprising not only fecal bacteria but also microbiota from jejunum, ileum, cecum, and colon (derived from pigs' intestine) and observed an increased colonization of small intestinal related bacteria in the small intestine. Based on this study, whole intestinal microbial transplants would be a mor e r epr esentativ e alternativ e for r odent humanization in view of small intestinal research (Li et al. 2020a ).
For small intestinal microbial research, the use of rodents has been questioned by their habit of copr opha gy described to shift the small intestinal microbiota closer to the colon composition, with incr eased pr esence of Clostridiales and Bacteroidales , while noncopr opha gic mice, pr e v ented b y w earing tail-cups, w ere dominated by lactobacilli, a taxon also r egularl y found in the humans small intestine (Hayashi et al. 2005, Li et al. 2015, Seekatz et al. 2019, Bogatyr e v et al. 2020. This questions the r ele v ance of mouse models in small intestinal microbial research and, to our knowledge, has not yet been addressed in current studies. Ho w ever, tailcups can induce stress to the animals (observed by a decreased weight) and it is described to influence the gut microbial community (Gao et al. 2018 ), resulting in the introduction of an additional bias for microbiome studies. In the study of Bogatyr e v et al. ( 2020 ), the microbial community and bile acid profiles between mocktail-cup (i.e . a tail cup which does not prevent coprophagy) and copr opha gic animals was similar, suggesting possible stress induced by the tail-cup does not influence the upper gastrointestinal tract comm unity significantl y. No compar able studies inv estigating the impact of copr opha gy on the small intestinal micr obiota of r ats are published, though the fecal community of rats wearing a tailcup shown a decr eased pr esence of lactobacilli and increased enterococci and coliforms (Fitzgerald et al. 1964 ). Nonetheless, a distinction in the microbial community of rodents between the small intestine and lo w er gastr ointestinal tr act is still found (Gu et al. 2013. Another possible confounder in the use of rodents as in vivo model is that mice and rats are largely herbivorous, with an intestinal anatomy adapted to this feature (i.e . large cecum). This may determine different ecological selection pr ocesses, r esulting in a pr oportionall y higher abundance of fiberdegr ading micr obiota in the rodent gut as opposed to the human gut.
The extr a polation of host-micr obiota findings fr om animals to humans should be performed with care . T he cospeciation between mammals and gut microbiota, whether or not as a result of coe volution, host biogeogr a phy, or allopatry (Gr oussin et al. 2020 ), results in symbiotic interactions, likely resulting in different responses or decreased microbial colonization if studied in other animals (Chung et al. 2012, Amato et al. 2015, Lundberg et al. 2020. In this view, when stud ying host-bacteria-n utrition interactions, the model should be chosen based on the scientific question and ultimately, if possible, the best model remains humans, despite increased ethical constraints and sampling difficulties.

Ex vivo models
One of the ways to stud y host-microbe-n utrient interaction in the small intestine is to make use of ex vivo intestinal explant tissues, deriv ed fr om either human or animal small intestinal biopsies and exhibiting a high resemblance to in vivo tissue complexity and morphology (Rozehnal et al. 2012, Roeselers et al. 2013. Ex vivo intestinal tissue segments from the different regions of the small and large intestine allow the investigation of regional absorption and immune responses (Rozehnal et al. 2012 ).
Different ex vivo approaches have been developed during the years to study host-microbiota interaction in small intestine, the most prominent being the Ussing chamber and the InTESTine TM System. The Ussing chamber has become a useful ex vivo tool widely used to assess the transport of several materials and nutrients (i.e . glucose, amino acids, and miner als) acr oss differ ent segments of the intestinal tract. This model has also been applied to study host-microbe interactions largely in the context of alter ed or dama ged permeability, suc h as in the case of exposure to bacterial toxins but also with live intact bacteria. Ho w ever, the application of Ussing chamber to study host-microbiota interaction in healthy condition is scarce and mostly restricted to colonic epithelium. Although being an excellent tool to study intestinal permeability, a major limitation for the application of Ussing chamber is that the epithelial layer alone is not capable of fully recapitulating the in vivo complexity . Additionally , the difficult preservation of the tissue viability through the experimental period (restricted to a maximum of 90 min, as tested with human ileum tissue in Söderholm et al. 1998 ) limits its applications to shorttime measur ements. Lastl y, another major dr awbac k of the Ussing chamber is its low-throughput, since it does not allow simultaneous pr epar ation and anal yses of a lar ge set of segments of epithelial tissues, limiting its applicability in the context of compound scr eening. Nonetheless, Ussing c hamber hav e been emplo y ed to study intestinal transport and barrier function after lipopolysacc haride exposur e (Albin et al. 2007), nutrient supplementations (Wo y engo et al. 2012 ) in animal small intestine and the impact of bacterial invasion (Isenmann et al. 2000, Jafari et al. 2016, and probiotics (Chen et al. 2010, Shi et al. 2014, in colon. Alternativ el y, The Netherlands Or ganization for Applied Scientific Research (TNO) has recently developed an ex vivo tissue model called the InTESTine™. Like in the Ussing c hamber, fr esh intestinal tissue (duoden um, jejun um, ileum, and colon) of human or porcine origin are mounted in the two-compartment system creating an apical and basolateral side (Westerhout et al. 2014, Stevens et al. 2019 ). Compared to Ussing chamber, the InTESTine™ system provides a higher throughput and easy horizontal setup in standard 6-or 24-well plates via which up to 96 ex vivo intestinal tissue can be used per day to test for intestinal absorption.

In vitro models
The host-microbiome research field has greatly benefit from the de v elopment of in vitro models. By simulating the physiological conditions in a contr olled envir onment and limited confounding factors, in vitro models have been widely developed and used to addr ess differ ent scientific questions, among whic h studying the complexity of the intestinal envir onment. Mor eov er the versatility, r epr oducibility, cost, and time-efficiency of in vitro models makes them well-suited for mechanistic studies and adaptable (most of the time) for high-thr oughput a ppr oac hes. Ho w e v er, the fitness of use of in vitro models depends on the scientific question. In the next section we describe in vivo models , their limitations , and adv anta ges, when used to study small intestinal ecology and mucosa.

In vitro models for small intestinal microbiota
In vitro culturing of the small intestinal microbiota was firstly addressed by using static models, such as the small intestinal model de v eloped by Schantz et al. ( 2010 ). Through a batch set-up inoculated with ileostomy effluent in airtight ano xic tubes, shak en and incubated at 37 • C for 24 h, the authors studied the impact of ileal microbial digestion on the green tea catechins and described interindividual differences between donors in both aerobic ( E. coli , Proteus sp ., and Enterococcus sp.) and anaerobic bacterial species ( Bacteroides sp., Bifidobacterium sp., and Lactobacillus sp.) and fungi ( Geotric hum sp.). Batc h models ar e suitable in vitro models to screen v arious tr eatments in small r eactors during a short time-span and with limited costs, but they remain an oversimplification of the in vivo situation, often without pH regulation or transit time simulation. In addition, such an in vitro method is mainly limited by the inv asiv eness of ileostomy, which cannot be performed on healthy volunteers.
To overcome these limitations, dynamic (semi-)continuous long-term cultur e a ppr oac hes wer e de v eloped, r epr oducing the main physico-chemical parameters of the human gut (i.e . pH, temper atur e, r etention time, nutrient availability, and anoxic conditions). Existing dynamic ( (Aidy et al. 2015 ), some of these wer e ada pted to study the small intestinal ecology.
The study of the ileal ecosystem with dynamic in vitro models is, so far, primaril y ac hie v ed thr ough thr ee bior eactor models: (i) The Smallest Intestine (TSI) (Cieplak et al. 2018 ); (ii) the SHIME ® (Roussel et al. 2020 ), and (iii) the TIM2 ® (Stolaki et al. 2019 ). These three dynamic models implement continuous pH regulation, maintain anoxic conditions, and simulate the transit time of each studied part. They mainly differ in what part of the intestinal tract is studied, and how the ileal microbiota is introduced into the model (summarized in Table 8 ). In the context of SHIME ® experiments, fecal micr obiota fr om human origin wer e first ada pted to pr oximal colon conditions . T hen, the r etr ogr ade micr obial colonization through the ileocecal sphincter, was simulated by a diluted feedback inoculation from proximal colon to the terminal ileum compartment (Laird et al. 2013, Roussel et al. 2020. This a ppr oac h resulted in lo w er di versity and metabolic acti vity in the ileal compartment (assessed by SCFA quantification) compared to the ascending colon, yet some k e y taxa c har acteristic of the small intestinal micr obiota, suc h as Streptococcus , r emained absent or wer e present in low numbers (Roussel et al. 2020 ).
Efforts to compare ileostomy effluents and fecal samples as inoculum in the ileum compartment, on a TIM de vice, r esulted in the description of similar comm unities de v eloped independentl y from the used sample but distinct from the original inoculum, and mainly constituted of Bacteroidota, Pseudomonadota, Bacilli, Clostridium , and Actinomycetota (Stolaki et al. 2019 ). Additionally, in this report, the authors described the presence of Streptococcus , in low amounts. In contrast, to have a straightforw ar d and easily r epr oducible ileal microbiota, TSI does not require ileal or fecal samples to inoculate, and is rather inoculated with a consortium of se v en bacteria, known to be pr esent in the human small intestine: E. coli , S. salivarius , S. luteinensis , Enterococcus faecalis , Bacteroides fragilis , Veillonella parvula , and Flavonifractor plautii (Cieplak et al. 2018, Jakobsen et al. 2022. By simplifying the complexity of the ileal microbiota, this model avoids the study of interindividual v ariability, ther efor e, limiting extr a polation to the physiological condition. In addition, this model r equir es se v er al days to prepare the consortium with different incubation times of the bacteria and does not r epr oduce the complex interactions occurring within the small intestinal microbiota ecosystem. Nevertheless, one of the main adv anta ges of the TSI model is to allow a better throughput by using five low volume ileal vessels in parallel.
Despite the existence of these in vitro models to mimic the small intestine, the de v elopment and the validation of physiologic-like bioreactors able to simulate the small intestinal micr obial comm unity r emains c hallenging. This is mostl y due to ethical constr ains, r estricting the accessibility for an a ppr opriate inoculum from healthy donors to validate the in vitro model against in vivo condition, along with technical issues to manage oxygen le v els, impacting obligate aer obe and facultativ e anaer obe in the ileum comm unity. Additionall y, the small intestine is a low micr obiota biomass envir onment with low div ersity and quantity compared to the high stool microbial biomass . T his increasing load of bacteria along the small intestine, as well as changes in microbial profiles are challenging to reproduce in an in vitro model, as well as the human interindividual differences. In order to address interindividual differences, some studies on colonic microbiota indicate parameters such as transit time and nutrient load are important drivers for microbial community development (Vandeputte et al. 2017, Minnebo et al. 2021, Pr oc házk ová et al. 2022.
To date, v ery fe w information exists on in vitro micr obial models simulating the ileal ecology and, to our knowledge, none for the duodenum or jejunum. Increased validation and research are essential to unlock the potential of in vitro systems in this still understudied area.

In vitro models for intestinal mucosa
In order to study the regulation of intestinal epithelial homeostasis and the mechanisms of this fine regulating hostenvir onmental cr oss-talk, differ ent epithelial models hav e been de v eloped. Depending on the r esearc h question (study nutrient absorption, permeability, and host-microbe interactions) different le v els of complexity to these models are needed.
The challenges of developing an in vitro model suitable to study host-micr obiota-diet inter actions r esides in the combination of the k e y components constituting such models: (i) the epithelial  (Hidalgo et al. 1989, Hilgers et al. 1990, to more complex 3D models with coculturing techniques that enable more than one type of intestinal cell (i.e . mucosal epithelia and a submucosal cell type) to be incubated sim ultaneousl y with select bacterial populations (Bernardo et al. 2012 ). Indeed, it is well known that the 3D physical en vironment pla ys a major r ole in the mor phology, bioc hemistry, and metabolism of mammalian cells (Anselme andBigerelle 2006 , Bettinger et al. 2009 ). Yet, mimicking the complex 3D crypt-villus arc hitectur e r emains c hallenging.
To simulate the cellular component in vitro , well-established immortalized cell lines, primary cell cultures, or induced pluripotent stem cells (iPSCs) can be used. In the next section, we describe se v er al in vitro cell models de v eloped for studying nutrients absorption and host-microbiota interaction in the small intestine.

Immortalized cell lines
Immortalized cells are derived from a population of cells that have e v aded normal cellular senescence due to a mutation and are ther efor e able to pr olifer ate indefinitel y.
To mimic enterocytes, human Caco-2, HT-29 (both deriv ed fr om colon adenocarcinoma) and HuTu-80 (deriv ed fr om duodenum adenocarcinoma) cell lines are generally used in vitro models to pr edict absor ption, to study permeability and diffusion of compounds through the epithelium, and describe mechanisms of host-micr obe inter actions (Hidalgo et al. 1989, Smetanová et al. 2011, Brosnahan and Brown 2012, Kavanaugh et al. 2013, Takenaka et al. 2016. In particular, Caco-2 cells are a well-established model for enterocytes and are , therefore , incorporated in many in vitro models. After seeding, Caco-2 cells spontaneously differentiate to form confluent monolayer of polarized cells, structur all y and functionall y r esembling the small intestinal epithelium and expressing several morphological and functional properties c har acteristic of small bo w el enteroc ytes. In particular, Caco-2 cells have been found to express a large number of enzymes and transporter proteins present in normal human small intestinal epithelium. For example, exposing Caco-2 cells to fluidic shear or periodic contraction, induces the expression of metabolic enzymes, m ucus pr oteins, as well as formation of villus-like structures Ingber 2013 , Lindner et al. 2021 ). Ho w e v er, r ecent studies suggest that variations exist between gene expression profiles of transformed epithelial cell lines, like Caco-2, and normal human intestinal epithelium (Bourgine et al. 2012 ), limiting their physiological r ele v ance. Indeed, although widel y used, immortalized cells are generally derived from cancer cells and, as such, metabolically differ from intestinal cells in healthy condition.
Se v er al m ucin-pr oducing immortalized cell lines hav e been established and used to mimic goblet cells, in the context of the small intestinal r esearc h. For example, the colonic cancer tissue isolated LS174T cell line shows high MUC2 mRNA expression, in line with the secr etory pr ofile of small intestinal goblet cells, along with low MUC5AC expression and are , therefore , often used as a model for goblet cells (Lesuffleur et al. 1990, van Klinken et al. 1996, Martínez-Maqueda et al. 2013.
Besides LS174T cells, the HT29-MTX cell line is commonly used in in vitro models to simulate goblet cells phenotype . T his cell type originates from the gradual exposure of the human colorectal adenocarcinoma HT-29 cell line to increasing concentrations of methotrexate (MTX), resulting in their transformation into m ucus-secr eting differ entiated cells (Lesuffleur et al. 1990 ). The major secreted mucin in this cell line is MUC5AC, along with the in lesser amounts-secreted MUC2 (Martínez-Maqueda et al. 2013, Elzinga et al. 2021. To address the enteroendocrine function of the small intestinal epithelium, mammalian EEC lines, such as human NCI-h716 cell line, are widely used to study gut hormone response to envir onmental stim uli, including food and micr obiota. NCI-h716 cell line exhibits lymphoblast morphology and enteroendocrine differ entiation, including the expr ession of secr etory gr anules and c hr omogr anin A (De Bruine et al. 1993 ). This cell line produces GLP-1, GLP-2 (Kuhre et al. 2016 ), making it suitable to study the secretion of these hormones. Ho w ever, it does not produce CCK nor peptide-YY (PYY) (Kuhre et al. 2016 ), limiting its resemblance with normal L-cells. Alternativ el y, another widel y used human cell line expressing PYY and precursor of glucagon/GLP-1, is the HuTu-80, isolated from healthy human duodenal cells (Nevé et al. 2010 ). Ho w e v er, this cell line also expresses the glucose-dependent insulinotr opic pol ype ptide (GIP), described as exclusi v el y pr oduced by K cells, in vivo (Rozengurt et al. 2006 ).

Immortalized cell lines in host-microbiota interaction studies
To better simulate the cellular complexity of the small intestinal e pithelium and stud y the host-bacteria inter action, cocultur e of different immortalized cell lines, with bacterial samples, is frequently used.
For example, bacterial-epithelial interactions have been studied using a Caco-2 cell monolayer system seeded on a porous membrane (Cruz et al. 1994 ) or through triple coculture of epithelial (Caco-2), goblet (HT29-MTX), and immune-like cells (THP-1) in physical contact and in combination with a synthetic microbial consortium of eight bacterial strains resembling the small intestine microbiome (Calatayud et al. 2019 ). Howe v er, one major dr awbac k of this a ppr oac h is that these experiments can be carried out only over a relatively short period of time before bacterial ov er gr owth leads to cell injury and death.
An alternative approach for studying host-microbiota interaction using cell lines, has been proposed by Marzorati and colleagues in 2014 with the development of the human −microbiota interaction module (HMI TM module). By combining a SHIME reactor with Caco-2 cells, the HMI TM module permits to e v aluate the effect of both aerobic and anaerobic microorganisms on enterocyte-like cells under an oxygen gradient. Because of the absence of cell-derived mucus in this model, a nanopor ous membr ane with an artificial m ucus lay er w as ad ded to se par ate the micr obial comm unity fr om the compartment hosting the cells, and, the cocultur es wer e maintained for 48 h (Marzorati et al. 2014 ). Although applied for studying colonic bacteria, this set-up offers the possibility of coupling cell models with a continuous simulator of intestinal ecology and simulate both bacterial adhesion and indirect host-bacteria interaction.
Despite the available tools and the rapidly expanding research field, in vitro studies on host-microbe interactions in the small intestine, combining immortalized cell lines with bacterial samples, still r emains limited, compar ed to colonic models, possibl y due to the slo w er gain of kno wledge on the ecology of this body site, for long time.
tween the microbiota and the central nervous system. In this context, two innov ativ e Eur opean-funded pr ojects, IMBIBE (ID 723951 'Innov ativ e tec hnology solutions to explor e effects of the microbiome on intestine and brain pathophysiology') and MIN-ERVA (ID 724734, 'MIcr obiota-Gut-Br aiN EngineeRed platform to eVAluate intestinal micr oflor a impact on brain functionality') have been designed to study the complex mechanisms of gutbrain axis on multiorgan-on-chip models . T he IMBIBE project aims at exploiting conductiv e pol ymer 3D tubular scaffolds for continuous monitoring of cell activity and integrity of an integr ated gastr ointestinal and blood-br ain barrier/neur ov ascular units (Pitsalidis et al. 2018 , Moysidou andOwens 2021 ). Conv ersel y, the MINERVA project makes use of five organ-on-chip devices to elucidate the effect of microbiota secretome on brain function, in a context of Alzheimer's disease. In particular, the MINERVA platform is based on miniaturized or gan-on-c hip de vices connected sequentially and designed to represent (i) the gut microbiota, (ii) the gut epithelial barrier, (iii) the immune system, (iv) the bloodbrain barrier, and (v) the brain.
One of the gr eat adv anta ges of these or gan-on-c hip models is the possibility to adapt the model to different scientific questions and making them versatile for several applications , e .g. including the c har acterization of micr obiota-small intestine-m ultior gan axis.
Yet, besides these pr omising tec hnological in vivo , in vitro , and in silico advances, collecting exhaustive information about compounds in the diet and the exact composition of macro-and micr onutrients, might be c hallenging. To fill this ga p, Blasco et al. ( 2021 ) de v eloped AGREDA, an extended reconstruction of diet metabolism in the human gut micr obiota, pr edicting diet-specific output metabolites from gut microbiota. As such, this tool has the potential to establish r ele v ant metabolic interactions between diet and gut microbiota and it could potentially be the starting point for novel prediction models for diet-host-microbiota interaction.
Altogether, the novel sampling methods, the technological advances in the field of intestinal in vivo and in vitro studies, and the r a pid expansion of computational-based techniques, could unloc k undiscov er ed potentials of the small intestinal micr obiota. In addition, this may lead to an expanded knowledge on diet-hostmicr obiota inter actions, a pplicable for ther a peutic pur poses, dev elopment and scr eening of nov el healthy food supplements or the impact of food pathogens, personalized nutrition, and the design of effective dietary strategies in cases of small intestinal dysbiosis and nutrient deficiencies.

Concluding remarks
Over the past decade, numerous studies explored the composition and broad metabolic activities harboured within the small intestinal micr obial comm unity. For example, some food components, such as dietary fibers, ar e r esistant to human host enzymes, hence their transformation into SCFAs, a major energy source for colonocytes, is exclusiv el y performed by intestinal microbiota. Additionall y, it emer ged that distinct ecological niches exist along the small intestinal tr act, sha ped by the diverse physicochemical, biochemical, and physiological parameters in the duodenum, jejunum, and ileum and the nutrient availability. Despite the fact that the limited accessibility of this body site has gr eatl y r estricted for long time our knowledge on the small intestinal environment, its ecology has become a novel attractive target in food digestion and absorption research. Nevertheless, conflicting data on bacterial composition along the small intestine have limited the translation of current host-bacteria interaction knowledge into medical pr actice. Her e, we hav e pr ovided an updated ov ervie w of diverse studies aiming at c har acterizing the small intestinal environment and discussed how different sampling techniques, a ppr oac hes, study gr oups, pathologies, and interindividual variabilities influences the microbial description. Additionally, as the host-micr obiota inter action in the context of nutrition has become a major r esearc h tar get, numer ous and div erse a ppr oac hes have been designed for a deeper mechanistic understanding of this fine-r egulated cr oss-talk. Her e, we r e vie wed differ ent in vivo , ex vivo , and in vitro a ppr oac hes used to this aim, and discussed their adv anta ges and limitations.
Along with the role of microbiota in nutrient digestion and absorption under healthy conditions, the implications for disease etiology or disease pr ogr ession hav e also become mor e e vident, yielding an enticing ther a peutic tar get for nutrition-related conditions, as well. Indeed, when an extended area of the small intestine is r esected, suc h as in SBS, or the small intestinal ecology is dr asticall y alter ed, as in SIBO, the consequences for human health are severe. Ho w ever, despite the numerous studies, our current knowledge on small intestinal environment and its modulation in healthy and pathological conditions, in the context of nutrition, is still limited. Novel technological advances will surely bring valuable contributions to the field. Yet, taking into consideration the strong interindividual variability, the wider translation to clinical practices needs to be carefully considered. In light of these, we belie v e that a personalized a ppr oac h to study the dialogue between the host, microbiota, and diet in the small intestine in both healthy individuals and patients is needed to move to w ar ds nov el patient-tailor ed ther a peutic a ppr oac hes with a potential impr ov ed impact on the quality of life.

Conflict of interest.
None to declare.