Introduction

The main function of the intestinal tract is to digest and absorb ingested nutrients. The intestinal mucosa is in direct contact with the external environment; therefore, it needs to act as a semipermeable barrier that enables efficient nutrient absorption and to protect the epithelium from potentially harmful agents present in the luminal content1. In this environment, the epithelium and the underlying immune system work together to maintain homeostasis and intestinal health. The epithelium forms a physical barrier restricting microbial access to the underlying tissues and secretes mucus that further limits interactions between the luminal content and the tissue2,3,4,5. Innate immune cells in the lamina propria survey and sample their local environment and respond with tolerogenic or pro-inflammatory responses depending on the nature of the antigens they encounter6. During the past decade, the goblet cell has emerged as a central player in the regulation of the intestinal barrier during health and disease7,8,9,10. Goblet cells are known for their role in providing the protective mucus barrier that covers the intestine, but it is becoming clear that they also have an essential role in regulating gut immune responses by sampling and delivering luminal antigens to the immune system for induction of adaptive immune responses9,11.

In this Review, we explore the role of intestinal goblet cells in regulating and maintaining the intestinal barrier via mucus secretion and trans-epithelial antigen delivery to the immune system. We discuss age-specific, organ-specific and location-specific adaptations of the functions of goblet cells and mucus, and summarize what is known about mechanisms regulating these functional adaptations. Finally, we discuss how dysregulation of goblet cells and perturbations of mucus barrier properties contribute to infectious and inflammatory diseases and how microorganisms take advantage of the normal functions of goblet cells to invade host tissues.

The intestinal mucus barrier

The intestinal tract performs vital functions ranging from nutrient digestion and absorption in the small intestine to bacterial fermentation and storage of waste products in the colon. While performing these functions, the epithelium is continuously exposed to various hazards, including reactive and toxic chemicals, degrading enzymes, microorganisms and mechanical forces1. To manage these challenges, the epithelium and its mucus cover form a selective barrier2,5,12 (Box 1).

Organization and composition

Mucus is formed by the material secreted from the epithelium, with its core components produced by goblet cells: mucin-2 (MUC2), Fcγ binding protein (FCGBP, also known as IgGFc-binding protein) and calcium-activated chloride channel regulator 1 (CLCA1)13,14,15,16. In addition to these core proteins, mucus contains other secreted molecules, serum contaminants, such as albumin and haemoglobin, products from detached epithelial cells, and substances originating from the luminal content, such as microbial proteins13,14,17. The mucus core comprises the densely glycosylated MUC2 that creates a multimeric, crosslinked network18,19,20,21. Biosynthesis of the mucus proteins demands specific machinery with chaperones assisting assembly and control in the endoplasmic reticulum (ER) and additional posttranslational modifications, such as O-glycosylation and crosslinking, which occur in the Golgi apparatus and late in the secretory pathway. The best-studied mucus protein in this context is MUC2, and its biosynthesis involves several proteins, such as ERN2, AGR2 and PDIA family proteins22,23 (Fig. 1). The mature mucus core proteins produced by goblet cells are packed and stored in secretory vesicles under low pH and high calcium conditions18. The stored vesicles can be released either under baseline conditions or in response to stimulation, often referred to as ‘compound exocytosis’24. Upon release from the goblet cell, the MUC2 multimers expand, a process dependent on simultaneous bicarbonate secretion25, and assemble into stacked networks to form the structural core of the mucus layer18. The secreted MUC2 can be further stabilized by intramolecular crosslinks20,21 or expanded via a proteolysis process that generates a more loose mucus gel that is easier to transport15,26 (Fig. 1).

Fig. 1: Mucin biosynthesis and mucus formation.
figure 1

The biosynthesis of mucin-2 (MUC2) involves an adapted endoplasmic reticulum (ER) system in which folding, carboxy-terminal disulfide-mediated dimerization and initial N-glycosylation occurs. MUC2 is then extensively O-glycosylated in the Golgi apparatus and packed as multimers in secretory vesicles, and it expands upon release, which is dependent on bicarbonate secretion that changes the milieu. The mucus layer can be reinforced by additional crosslinks, whereas proteolysis mediates further expansion and release that are important for distal transport.

Historically, the intestinal mucus layer has been described as an unstirred layer, indicating a zone with low diffusion27. It forms a gel with viscoelastic properties responsive to sheer forces that are important for many of its functions28. The overall mucus structure, as visualized by electron microscopy, reveals a mostly uniform net-like three-dimensional structure with small pores29,30, suggesting that the mucus restricts diffusion and penetration of substances through its structural arrangement and electrostatic charge. The mucus also provides numerous binding sites that promote specific and non-specific interactions with other molecules and bacteria31. It also reduces friction and mechanical stress of the epithelium and encloses much of the luminal content, including faecal matter, creating separate but interacting compartments inside the intestinal lumen4,5,10,32,33. Continuous peristalsis transports the luminal contents and mucus through the intestine, which requires fast renewal of mucus to maintain an intact layer34,35,36. Thus, mucus can act as a lubricant to trap and transport the luminal contents, and it enables selective passage of molecules mediated through different types of interactions, which are all important functions in fulfilling the physiological requirements of the small and large intestine (Box 1 and Fig. 2).

Fig. 2: The role of the various goblet cells in the adult small and large intestine.
figure 2

Different goblet cells mediate different functions in the intestine. In the small intestine, crypt goblet cells secrete a dense mucus that acts as a diffusion barrier for antimicrobial proteins and peptides that keep bacteria at a distance. Villus goblet cells produce and secrete mucus rapidly, and the mucus is easily removable and enables nutritional uptake. Goblet cells in the crypt and villus form goblet cell-associated antigen passages that deliver antigen to mononuclear phagocytes in the lamina propria. In the distal colon, crypt goblet cells secrete a dense mucus layer that protects the crypt, whereas inter-crypt goblet cells secrete a more permeable mucus. The crypt and inter-crypt mucus mix and form a stratified layer that keeps the microbiota at a distance. Mucus also coats the faecal matter reducing mechanical stress upon the epithelium. AMP, antimicrobial products; GAP, goblet cell-associated antigen passage; MAMPs, microorganism-associated molecular patterns; senGC, sentinel goblet cell.

Mucus in the small intestine

In the small intestine, the key function of the epithelium is to digest and absorb nutrients. This function is facilitated by a mucus layer loosely attached to the epithelium4,5. The mucus layer is partially penetrable, enabling smaller digested nutrients in the lumen to diffuse down to the epithelium, which could also be beneficial for the delivery of orally ingested drugs31,37,38. Mucus secretion is integrated with the digestive system as parasympathetic stimulation is initiated by feeding. A coordinated secretory response induced by acetylcholine (ACh) triggers the secretion of fluid, digestive enzymes and mucus, and activates peristalsis that propagates the intestinal contents in the distal direction1. Mucus protects the epithelium by trapping the chyme containing gastric and pancreatic digestive enzymes and by being resistant to chemical and proteolytic challenges such as acid, bile and the endogenous gastric and pancreatic digestive enzymes31,39,40,41,42,43,44,45 (Box 1). In the distal small intestine, increasing microbial densities pose new challenges. The ileal mucus is penetrable, a phenotype induced by bacterial colonization46,47,48, but bacteria are kept from infiltrating between neighbouring villi by secretion of fluids and antimicrobial products that probably form gradients in the mucus5,38,49,50,51,52 (Fig. 2).

Mucus in the colon

In the colon, the requirement for protection of the epithelium from the expanded microbiota is drastically increased, and the compact luminal material exerts strong mechanical forces on the epithelium. The colonic mucus is adapted to overcome these challenges by increasing in density and abundance and by adhering to the epithelium4,32,53 (Box 1). In mice, bacteria in the caecum and proximal colon are often found close to the epithelium5. However, constant mucus secretion keeps the crypts largely free from bacteria5. The folds in the proximal mouse colon create specific reservoirs in which resident bacteria can stably thrive, as more dynamic bacterial communities reside in the lumen54. Little is known about mucus properties in the human proximal colon, but it might differ from that observed in mice as their physiology is markedly different55. In mice, the proximal colon mucus encapsulates the faecal matter10,33 (Fig. 2). In humans, the faecal matter is also covered by a thin mucus layer, but what region of the colon this mucus originates from is unknown56. Mucus covering of the faecal matter is probably a continuous process occurring in parallel with propulsion along the intestine. Although interactions between the host and the encapsulated microbiota are limited, bacteria also reside outside the enclosed faecal matter and are prevented from accessing the epithelium by a stratified mucus layer in the distal colon in humans and mice32,42,57 (Fig. 2). The colonic mucus layer is formed by intermixing mucus secreted from crypt goblet cells and mucus secreted from goblet cells in the inter-crypt epithelium, named inter-crypt goblet cells present in both humans and mice57,58. The mucus secreted from crypt goblet cells is impenetrable to small particles and bacteria, whereas the mucus secreted from inter-crypt goblet cells is penetrable to small particles but not to bacteria57. How the colonic goblet cells can produce mucus with such varying properties using the same set of core proteins is unclear, but it might be explained by different processing, such as multimerization, crosslinking and cleavages of the mucus components15,20,21. Proteolysis gradually expands the stratified mucus, which gives the bacteria improved access to the mucus and enables distal transport15. Microbial compartmentalization in the mucus is further promoted by the aggregation of bacteria by host proteins such as zymogen granule membrane protein 16 (ZG16) and secreted IgA51,59. Some bacteria are adapted to feed on mucin carbohydrates and provide cooperative benefits and substrates for other bacteria60. This cooperation between bacterial species could be a way to maintain a stable bacterial community independent of variations in food intake61. However, in mice fed a diet low in dietary fibre, the lack of dietary fibre results in enhanced bacterial degradation of the mucus and subsequent breakdown of the mucus barrier, demonstrating that although several bacterial species can survive on mucins as their sole nutritional source, the mucus barrier is not adapted to be the sole nutritional source for the microbiota47,62,63.

Mucus during development

Concerning the establishment of the intestinal mucus layers, insights into the timing and mechanisms regulating the maturation process are limited due to a lack of functional studies of mucus properties during early life in both human and animal models. In rats, the only species in which mucus barrier function has been studied during early life, the small intestine and colonic mucus layers develop during the first week of life (Fig. 3b). In the small intestine, a gradual increase in mucus coverage of the intestinal villi was observed, with the entire villi being covered by mucus by postnatal day (PND) 9. In the colon, a developed mucus layer was observed on PND9, but the exact timing remains to be explored64. The establishment of the intestinal mucus layers coincides with colonization resistance against pathogenic Escherichia coli. In the absence of small intestinal mucus, E. coli can more easily access the tissue, suggesting that the mucus contributes to protecting the neonate from infectious disease64.

Fig. 3: Mucus and goblet cells during prenatal and postnatal development.
figure 3

a | In the human small intestine, goblet cells and enteroendocrine cells appear during the first trimester of prenatal development, whereas Paneth cells and tuft cells appear during the second trimester. Vacuolated fetal enterocytes are present during the second and third trimester and gradually decrease towards the end of the third trimester. In the human colon, the transition from a stratified epithelium to a columnar epithelium occurs around postconceptional week (PCW) 12, and the epithelium forms a villus-like structure. Around PCW 30, the villus epithelium transforms into a crypt epithelium. b | In the mouse and rat small intestine, immature goblet cells are observed in the days prior to birth. During the first week of life, small intestine crypts start to develop, and Paneth cells appear at the end of the second week of life. Vacuolated fetal enterocytes are present during postnatal life and gradually disappear around the time of weaning. Goblet cell-associated antigen passage (GAP) formation is initiated during the second week of life and persists throughout adulthood. In the mouse colon, GAP formation is inhibited during the first week, is initiated at the beginning of the second week of life and peaks around the time of weaning, after which GAP formation is reduced by goblet cell intrinsic sensing of the microbiota. E, embryonic day; P, postnatal day.

Conventionalization experiments of germ-free mice, which in contrast to conventionally raised mice, have attached mucus in the small intestine and penetrable mucus in the colon, demonstrated that the microbiota has a vital role in regulating both mucus adhesion and mucus penetrability46. These experiments, however, were performed in adult mice in which normalization of the mucus took several weeks. It is likely that maturation of the mucus occurs more rapidly during the normal colonization process following birth in conventionally raised mice. The importance of the mucus layer in protecting the colon from the microbiota became apparent around the time of weaning when Muc2−/− mice lacking intestinal mucus or C1galt1fl/flVillinCre mice with an impaired mucus barrier developed spontaneous colitis originating from the distal colon32,65,66. Notably, the absence of mucus production resulted in more severe inflammation in the distal colon than in the proximal colon. In the proximal colon, this was presented as an increase in crypt length at 12 weeks of age and a loss of crypt architecture at 16 weeks of age, whereas in the distal colon, such changes were observed at 5 weeks of age67. In the small intestine of these mice, the absence of mucus production resulted in an increased crypt length at 16 weeks of age. In both the small intestine and colon, the absence of mucus production increased the risk of adenocarcinoma development65,67,68. The lower abundance and diversity of bacteria in the small intestine, than in the colon, combined with antimicrobial products, might provide adequate protection to prevent inflammation in the small intestine, and increased expression of markers associated with immunosuppression, such as FOXP3, could be of importance in restricting inflammation in the proximal colon65. The timing and severity of disease in the Muc2−/− mice are also influenced by the genetic background of the mice.

The mucus-producing goblet cell

Mucus is produced by goblet cells, named so in 1866 based on their morphological shape with an apical cup-like profile. These are one of the cell types in the secretory lineage originating from stem cells located at the bottom of the intestinal crypts69,70. Goblet cell differentiation is dictated by the inhibition of Notch and Wnt signalling pathways, which ultimately results in fully differentiated cells on the small intestinal villi and colonic surface epithelium70,71. The transcriptional factor ATOH1 regulates the transcriptional control of the early secretory cell commitment, and subsequent goblet cell differentiation is mediated by several transcription factors, including GFI1 and SPDEF72,73,74,75,76. Single-cell RNA sequencing and proteomics data from both mouse and human intestinal samples have identified several additional goblet cell-enriched markers and transcription factors, such as BCAS1, SPINK4, REP15, CREB3L1 and FOXA3, but their specific role in the function and regulation of goblet cell differentiation and maturation remains to be explored57,77. The frequency of goblet cells increases in the proximal to the distal direction, with the lowest frequency in the duodenum and the highest in the rectum11. Similar to enterocytes, goblet cells are short-lived cells that migrate along the crypt–villus and crypt–surface axis, and at the tip of the villus and the colonic surface, where cells produced by adjacent crypts meet, goblet cells are exfoliated78. The high turnover of the epithelium results in the entire goblet cell population in mice being replaced in approximately 7 days, as observed in Math1fl/flVillinERT2Cre mice in which secretory cells can be deleted by tamoxifen injections11.

Goblet cell subpopulations

The ability to profile gene expression in individual cells using single-cell RNA sequencing has remarkably improved our knowledge regarding the expression signatures of goblet cells and identified distinct goblet cell populations present in humans and mice58,77,79,80,81,82,83,84,85. The main goblet cell features identified in such studies are genes encoding transcription factors driving goblet cell differentiation, core mucus proteins and goblet cell-specific proteins involved in biosynthesis and protein folding58,77,79,80,82,85. In addition, a subset of goblet cells in the lower crypt expresses proliferative markers, and some goblet cells express enterocyte markers such as SLC26A3 and DMBT1 (refs.77,80,81). Goblet cell markers are also expressed in some cells clustering with enterocytes77,80,81. Nyström et al. analysed Muc2-expressing cells in mouse small intestine and colon, and revealed a comprehensive view of distinct goblet cell populations and their interrelation during differentiation57. The study revealed a dynamic system composed of distinct types of goblet cells, termed canonical and non-canonical goblet cells, which diverged during differentiation (Fig. 2). Canonical goblet cells identified in this study had a classic goblet cell signature with high expression of Atoh1, Muc2, Fcgbp and Clca1, whereas non-canonical goblet cells were characterized by a more enterocyte-like expression profile with higher expression levels of Hes1, Dmbt1, Muc17, and ion channels57, which was in concordance with previously mentioned studies58,77,80,84. Coordinating mucus production and release with specific ion transporters in the same cell might influence the properties of the secreted mucus as the ionic environment has been shown to regulate mucus expansion in mouse small intestine43. The contribution and properties of the mucus secreted by the different cell types were not easily dissected; however, the canonical goblet cells were probably the main contributors to the secreted mucus barrier57. The non-canonical goblet cells expressed digestive enzymes, such as maltase–glucoamylase and sucrose–isomaltase, and additional metabolically related genes, suggesting that they contribute to food digestion57,77.

The different goblet cell populations identified in single-cell RNA sequencing mouse data are probably formed during differentiation, but interlinks between populations in the trajectories indicated a potential for plasticity57,86. The dynamics of goblet cell differentiation and their abilities to be reprogrammed by dedifferentiation or trans-differentiation87 have not been investigated in detail, but it will be of interest for future developments of strategies to manipulate cell differentiation in a controlled way. Reprogramming and plasticity of goblet cells induced by injury have been studied both in vitro and in vivo; whether such processes are also active during steady-state is not clear88. A specific cell transition event of interest is the major differentiation step occurring at the crypt–villus junction in the small intestine, similar to the switching event, in which the large upper crypt goblet cells switch to the morphologically different small surface-located inter-crypt goblet cells in the distal colon57,89. Whether the inter-crypt goblet cells are transformed from the upper crypt goblet cells or other cell types in the crypts is not known. These two goblet cell populations (that is, the upper crypt and inter-crypt goblet cells) are, however, responsible for producing the core mucus proteins that, following secretion, form the structural basis of a mucus layer with different properties depending on its location; that is, if it covers the crypt opening or the inter-crypt regions, as described earlier57. The heterogeneity with distinct expression patterns of goblet cells indicates a higher transformation potential as suggested for other cell types, which could be important in responding appropriately to environmental changes for the intestinal stem cells and trans-amplifying cells as these also express goblet cell markers such as MUC2 and TFF3 (refs.57,58). A dynamic system of cell-producing mucus or substances affecting the mucus can facilitate a better balance to maintain the protective mucus system and enable rapid transient responses to a constantly changing environment.

Goblet cells, during differentiation, can often be observed in the mid-upper crypt, and acquire features related to microbial defence such as expression of defensins and lysozyme normally attributed to the Paneth cells located at the bottom of the crypts of the small intestine57,58,77. Lysozyme expression was noted in the human colon crypt base cells that support the stem cells77,90. Goblet cell expression of genes related to microbial defence is not set to a specific programme as expression of defensins was observed both with and without lysozyme57. The expansion of cells with a defence profile is likely to be a response to a more hostile environment, which can enhance the bactericidal capacity of mucus in the crypts.

Mucus production and secretion

As mentioned above, goblet cell morphology differs along the crypt–villus and crypt–surface axis, and so does the mucus biosynthesis. In vivo studies of the small intestine have demonstrated that villi goblet cells produce and secrete mucus more rapidly than crypt goblet cells34. Similar in vivo studies in the colon have demonstrated that surface goblet cells produce and secrete mucus at a higher rate than crypt goblet cells35. All goblet cells secrete mucus at a baseline rate, but in contrast to crypt goblet cells, this represents the main mode of secretion for the villi and surface-located goblet cells as determined in vivo and ex vivo34,91. A focus on baseline secretion in these cells is supported by higher expression of the gene Stxbp1, which is involved in baseline mucus secretion in the airways92. Several goblet cell-specific SNARE and RAB proteins, such as RAB27A, RAB27B and SYTL2, involved in vesicle organization and transport have been identified, indicating that goblet cells have specific machinery adapted to their regulated secretory pathway57. However, many functions related to secretion are shared with other cell types, as exemplified by the goblet cell-enriched gene, Vamp8, which in mice was shown to be important for mucus release but is not exclusively expressed in goblet cells93. The expression of a mechanosensitive ion channel, PIEZO1, in human goblet cells is interesting, as pressure activation could potentially be involved in the activation of mucus release94. In addition to baseline release, mucus secretion can be induced by different stimuli such as cholinergic agents, histamine, prostaglandins and cytokines; however, differences in responsiveness have been observed between the small and large intestine, as well as between villus or surface and crypt cells in several animal models24,91,95,96. Cholinergic agonists stimulate mucus secretion primarily from crypt goblet cells in the small intestine, and proximal and distal colon in mice91,97, whereas histamine has been shown to induce mucus secretion in mouse colon but not in the small intestine24. On the contrary, studies have demonstrated that prostaglandin E2 induces mucus secretion in the mouse small intestine but not in the colon95,96. Receptor expression mediating this differential responsiveness at a cellular level has not yet been elucidated. The differentiated goblet cells in the upper parts of the distal colon crypts produce and store large amounts of mucus that can be expelled upon stimulation by the microbiota, as observed for the sentinel goblet cells present in both mice and humans as discussed below7,14,24.

The expressional profile of goblet cells is directed towards mucus production, a demanding and adaptable secretory machinery57,58,77,79,80,82,85. The observed regional differences in the rate of mucus biosynthesis indicate that goblet cells can respond to local stimuli and challenges and adapt their protein synthesis and secretory machinery to fit the needs of their specific position in the epithelium4,5,58. Mucus biosynthesis is a demanding process involving the translation and folding of long mucins, protein stabilization by numerous disulfide bonds, protein multimerization and additional modifications such as isopeptide bonds, and extensive O-glycosylation18,19,20,98. Some of the other core mucus components such as FCGBP, an extended molecule composed of a chain of von Willebrand D domains, are probably also difficult to produce due to their length and modifications such as disulfide bonds and numerous cleavages16,99. Thus, mucus biosynthesis requires tight regulation of when and where intermolecular and intramolecular interactions of these large proteins occur. This complex biosynthetic programme is supported by the expression of several goblet cell-specific proteins, such as the chaperone AGR2 and the stress sensor IRE1β (also known as ERN2), and if dysregulated reduces mucus production resulting in subsequent intestinal inflammation100,101,102,103,104. The high demand for the goblet cells to continuously produce large amounts of mucus proteins during steady-state is probably involved in the association between ER stress, goblet cell dysfunction and the development of chronic inflammation observed in humans and mice105,106. The relationship between mucus secretion and mucus production might require adapted ER stress responses as indicated by the expression repertoire of the colonic inter-crypt goblet cells that differs from that of upper crypt goblet cells, with reduced chaperone expression and suppression of several ER stress-related genes, that probably enables their high secretion rate at homeostasis57,58. Mucus production is an energy-consuming process in which the use of energy sources provided by the microbiota has an important role in maintaining intestinal homeostasis107. Thus, the varying demands for mucus secretion at specific locations require goblet cells to adapt to their local environment to provide the niched function.

Prenatal intestinal development

In mammals, intestinal development during fetal and postnatal life follows a similar general pattern, starting with the formation of a tube covered by a stratified epithelial layer which transitions into a single columnar layer organized into villus and crypt structures in the small intestine and into crypts separated by a flat epithelium in the colon70. The maturation process of the intestine proceeds in a cranial–caudal direction, starting in the duodenum and gradually progressing towards the colon108. Despite following a similar developmental pattern, the timing of the maturation process differs greatly between humans and the most commonly used laboratory animal species, mouse and rat109,110 (Fig. 3).

In humans, intestinal differentiation is initiated in the first trimester, around week 9 after conception, when the stratified epithelial layer that covers the immature intestine transitions into a single columnar epithelial layer111. In the small intestine, primitive villus structures appear around week 10, and crypts start to develop the following week112. In the fetal colon, the developmental pattern is initially similar to the conversion of the stratified layer into a villus structure with crypts emerging around weeks 12–14. The fetal colon undergoes a final stage of remodelling around week 30 when the villi disappear, and the mature crypt epithelium appears108 (Fig. 3a). Early studies using in situ hybridization on intestinal tissue specimens identified immature MUC2-expressing goblet cells already in the stratified layer, and differentiated goblet cells that were present in the small intestine and colon around week 12 (refs.113,114). These observations were confirmed by single-cell RNA sequencing data of embryos ranging in age from 6 to 10 weeks, that demonstrated the presence of secretory cells (goblet cells and enteroendocrine cells) along the entire length of the small and large intestine during this stage of development112 (Fig. 3a). Although secretory cells were observed in both the small and large intestine, the largest proportion of secretory cells were observed in the duodenum and jejunum of the small intestine, whereas the smallest proportion of secretory cells was observed in the colon, demonstrating that secretory cells also develop in a cranial–caudal direction112. Although secretory cells can be found in the stratified layer, they are in low abundance prior to week 12, when progenitor cells and/or transit-amplifying cells are the predominant cell types115. After week 12, the cellular composition of the developing epithelium is similar to that in the adult115. A single-cell RNA sequencing study evaluating 14 embryos ranging in age from 6 to 25 weeks identified goblet cells expressing MUC2, TFF3, CLCA1 and SPDEF around weeks 11–12 in both the small intestine and colon, but the frequency of goblet cells increased towards the end of the time window116. Together these studies demonstrate that in humans, secretory cell differentiation is complete or nearly complete during the second trimester (Fig. 3a).

In contrast to the human small and large intestine that are fully developed at term, in mice and rats that have a much shorter gestational period (19–21 and 21–22 days, respectively), intestinal differentiation is initiated later in embryonic development, continues in the postnatal period, and finishes around the time of weaning (approximately 3 weeks after birth)109. The transition from a stratified to a columnar epithelial layer starts on embryonic day 14.5 in mice and embryonic day 17 in rats. In rats, immature goblet cells appear on embryonic day 17, and differentiated goblet cells are observed on embryonic day 18, correlating with villus formation in the small intestine110,117,118,119. In both mice and rats, small intestine crypt formation is initiated around PND3 and finishes around PND14120 (Fig. 3b). Additional differences in the timing of the maturation process between humans and rodents include the presence of vacuolated fetal enterocytes (VFEs). In rodents (and larger mammals such as piglets), VFEs transport colostral milk proteins, growth factors and immunoglobulins from the lumen to the circulation and contribute to intracellular digestion of luminal nutrients during the first 2 weeks of life, after which VFEs are replaced by mature enterocytes121,122,123. In primates, VFEs are frequent during fetal life and are replaced by mature cells prior to term111,124. Thus in humans, VFEs are not considered to contribute to nutrient digestion and absorption, but they contribute to the uptake of growth factors and immunoglobulins present in the amniotic fluid125. It is essential that these species-specific differences are taken into consideration when using animal models to study human physiology and pathophysiology. One additional important aspect to keep in mind is that human intestinal development occurs mainly in utero, in a largely sterile environment126,127,128, whereas intestinal development in rodents proceeds after birth in parallel with the establishment of the intestinal microbiota.

Postnatal goblet cell development

Postnatal goblet cell differentiation is regulated both by growth factors and the developing microbiota. Epidermal growth factor (EGF), which is present at a high concentration in maternal milk during the first 4 weeks after birth depending on species, has a stimulatory effect on both goblet cell proliferation and Muc2 expression in mice and pigs9,129. In mice, this translates into a transient increase in Muc2 expression in the small intestine and distal colon that peaks around 2 weeks after birth and decreases to levels similar to that observed at birth around the time of weaning. In the proximal colon, Muc2 expression increases during the first 3 weeks of life and stays elevated throughout adulthood130,131. The mechanisms regulating these organ-specific and segment-specific expression patterns are not known but might involve the development of Paneth cells that occurs around PND10 and differences in microbial load and composition in the proximal and distal colon109.

Goblet cells as luminal sensors

In addition to the classic goblet function, the production and secretion of gel-forming mucins, it is becoming clear that intestinal goblet cells have an important role in conveying signals from the intestinal lumen to the immune system by forming so-called goblet cell-associated antigen passages (GAPs) that deliver luminal antigens to innate immune cells in the lamina propria132 (Fig. 4). GAP formation is a regulated process that is present along the entire length of the small intestine of the adult mouse, is inhibited in the adult proximal colon and is present again in the adult distal colon, although at a lower frequency than the small intestine11. Intestinal GAPs deliver antigens to lamina propria mononuclear phagocytes (MNPs), some of which traffic to draining lymph nodes in which they induce adaptive immune responses that promote induction of tolerance to the luminal content11. In addition to the immune regulatory role of antigen sampling, a subset of goblet cells at the top of the distal colon crypts of mice and humans, termed sentinel goblet cells, endocytose microorganism-associated molecular patterns, and at a certain threshold concentration of bacterial ligands initiate mucus secretion from surrounding goblet cells aimed at reinforcing the mucus barrier7. Together these findings demonstrate that the physiological role of intestinal goblet cells extends beyond their role as mucus-producing cells and includes sensing the local environment and regulation of gut immunity.

Fig. 4: Antigen sampling by goblet cells in the small intestine and colon.
figure 4

In the small intestine and colon, goblet cells form goblet cell-associated antigen passages (GAPs) that deliver antigens to mononuclear phagocytes in homeostasis coupled with secretion (left panel). This process is stimulated by acetylcholine (ACh) binding to muscarinic ACh receptors (mAChRs) and IL-13 via its receptor IL-4R–IL-13R and inhibited by microbial signals via toll-like receptors (TLRs) 2, 4 and 5 and by epidermal growth factor receptor (EGFR) signalling. The ability of goblet cells to form GAPs can be utilized by pathogens, such as Escherichia coli and Salmonella spp., to invade the host (middle panel). If the mucus barrier loses its integrity, an increasing concentration of bacterial products reaches the epithelium and triggers mucus release induced by endocytosis of bacterial ligands by sentinel goblet cells protecting the crypt openings (right panel). Activation of sentinel goblet cells is in contrast to GAP formation stimulated by TLR2, TLR4 and TLR5 in an inflammasome-dependent manner, involving the propagation of Ca2+ signals to neighbouring cells via gap junctions, triggering mucus release to restore the mucus barrier. LP-MNP, lamina propria mononuclear phagocytes; MAMPs, microorganism-associated molecular patterns.

Although observations of goblet cells sampling luminal antigens can be traced back to the 1980s117,133, the seminal study by McDole et al., that demonstrated the antigen sampling capacity of goblet cells and subsequent delivery of acquired antigens to MNPs in mice, provided compelling evidence of the immune regulatory role of goblet cells132. Although all goblet cells produce and secrete mucus, GAP formation is a regulated process under the influence of growth factors9, neurotransmitters132, Toll-like receptor (TLR) ligands134 and cytokines secreted by T helper 2 cells135. During steady-state, ACh acting on muscarinic receptor 4 in the small intestine and muscarinic receptor 3 in the colon is the main driver of GAP formation in mice, whereas activation of the EGF receptor (EGFR) and/or TLRs inhibits GAP formation134,136. In response to ACh, goblet cells that form GAPs undergo a bulk endocytic event that shuttles fluid phase cargo through the cell for delivery to MNPs. In addition to triggering GAP formation, ACh triggers mucus secretion primarily from the small intestine and colonic crypts97. By using different subsets of muscarinic receptors, goblet cells can differentiate GAP formation from mucus secretion136, which might explain why the proximal mouse colon responds to ACh with a mucus secretory response without GAP formation5,134.

Due to spatial and temporal differences in the expression levels and/or localization of muscarinic receptors, TLRs and EGFR and their respective ligands, GAP formation follows a strict temporal and spatial pattern which regulates when and where the immune system is exposed to the luminal contents9. During the first week of life in mice, GAP formation is inhibited in the small and large intestine by the high levels of EGF present in the dam’s milk. As EGF concentrations in milk decrease after birth, GAPs start to form, first in the colon and later on in the small intestine, probably due to the lower amounts of EGF that reach the colon9. Once small intestinal GAPs form, the process remains active throughout adulthood. In the colon, a transient increase in GAP formation is observed when GAPs appear during the second week of life and peak around the time of weaning, after which GAP numbers drop and remain low throughout adulthood9 (Fig. 4). In the proximal colon, inhibition of GAP formation during adulthood is mediated by goblet cell-intrinsic sensing of the microbiota via TLR2, TLR4 and TLR5 and their major adaptor MyD88 (ref.134). Steady-state GAP numbers differ between the proximal and distal colon, with GAPs being more prevalent in the distal than in the proximal colon9,137. Mechanisms regulating these differences are yet to be explored.

The temporal pattern of GAP formation in the mouse colon creates a window in time when the colonic immune system is exposed to large quantities of luminal antigens. The functional role of GAP formation during this period is to induce and maintain oral tolerance to dietary and microbial antigens via induction of regulatory T (Treg) cells9,138. Alterations to the window of time when GAPs are formed either by prematurely opening GAPs or delaying GAP formation results in impaired Treg cell induction during early life9. This impairment has long-lasting effects on gut immunity, predisposing mice to chronic inflammatory diseases later in life, including impaired induction of tolerance to dietary antigens and increased susceptibility to the development of colitis9,138. Interestingly, the temporal pattern of GAP formation coincides with a transient increase in the pro-inflammatory cytokines tumour necrosis factor and interferon-γ that, similar to colonic GAPs, are regulated by EGF and the microbiota139. This transient increase in pro-inflammatory cytokine expression imprints the developing immune system towards tolerance, and inhibition of this response during early life predisposes mouse pups to the development of chronic disease later in life139.

During adulthood, mouse small intestine and distal colon GAPs continue to sustain tolerance to the luminal content by maintaining local Treg cell populations, promoting proliferation and differentiation of de novo Treg cells in draining lymph nodes11,132,134. In addition, small intestine GAPs imprint MNPs with properties promoting tolerance by stimulating IL-10 production in macrophages and retinoic acid activity in dendritic cells11. Furthermore, sampling of the endogenous goblet cell protein Muc2 by MNPs is associated with improved Treg cell induction, and culturing of isolated MNPs in the presence of Muc2 promotes the development of a tolerogenic MNP phenotype in mice140,141. Similar to what has been observed in the intestine, mouse ocular goblet cells induce and maintain tolerance in the eye142,143. Whether or not goblet cells at other mucosal surfaces, such as the airways and the urogenital tract, have functions similar to those of the intestine and the ocular surface remains to be explored.

As mentioned above, in addition to sampling luminal antigens for delivery to the immune system, a subset of distal colon goblet cells at the top of the crypts in mice act as sentinel cells that, upon endocytosis of large quantities of microbial components, induce mucus secretion from neighbouring cells by increasing intracellular levels of Ca2+, resulting in reinforcement of the mucus barrier7 (Fig. 4). Notably, the secretory response induced by these sentinel cells is triggered by the TLR ligands P3CSK4 and lipopolysaccharide acting on TLR1, TLR2 and TLR4, in a MyD88-dependent manner; these are the same pathways that inhibit GAP formation in the mouse proximal colon7,134. Combined, these studies demonstrated that in mice, goblet cell-intrinsic sensing of bacterial molecules protects the colon from the microbiota by restricting trans-epithelial antigen delivery to the immune system and reinforcing the mucus barrier in case of a breach. Although the initial signalling events triggering the activation of sentinel goblet cells and inhibition of GAP formation are shared, downstream signalling pathways regulating the two processes differ as sentinel goblet cell activation relies on inflammasome activation mediated by Nlrp6 and caspase 1 and caspase 11, whereas GAP formation is independent of caspase 1 and caspase 117,144 (Fig. 4). Nlrp6 has also been demonstrated to have a role in steady-state mucus layer formation as Nlrp6−/− mice were shown to lack a functional mucus barrier145. However, a study by Volk et al. could not confirm a mucus defect in Nlrp6−/− mice, suggesting that the previously observed mucus defect in Nlrp6−/− mice might be related to external factors, such as differences in microbiota composition, as some of these experiments were not performed using littermate controls146. At this point, most of the knowledge regarding both sentinel goblet cells and intestinal GAPs is based on experimental model systems, that is, mice; however, both GAP formation and the sentinel function of goblet cells are present in the human intestine14,132.

Mucus in inflammation and infection

The mucus layer is a dynamic system, and disruptions might occur regularly by mechanical or infectious agents. The mucus system is adapted to rapidly recover following insults, as with mucus release in response to elevated levels of microbial ligands or after ischaemia7,147,148,149. The intestinal mucus system of mice is also altered with age, resulting in a thinner mucus layer that is more penetrable to bacteria causing increased susceptibility to infections150. Impaired mucus protection is probably an early event in the development of chronic inflammatory conditions, and results in increased bacterial penetration of the mucus, and subsequent increased interactions between the microbiota and the intestinal epithelium that trigger the onset of disease, as shown in mouse models42,67,151,152 (Fig. 5).

Fig. 5: Impaired mucus barrier function and altered goblet cell composition in relation to inflammation.
figure 5

In a healthy colon, the mucus forms a layer that separates the microbiota from the epithelium (left). During periods of colitis, mucus impairment resulting in enhanced bacterial penetration is common, as well as an altered goblet cell composition with loss of sentinel goblet cells (senGC) and inter-crypt goblet cells (icGC) (right). Loss of mucus barrier function is also associated with enhanced epithelial cell proliferation resulting in crypt elongation. ER, endoplasmic reticulum; GAP, goblet cell-associated antigen passage; MAMPs, microorganism-associated molecular patterns.

Defects in chronic inflammation

The two main chronic intestinal inflammatory diseases are Crohn’s disease and ulcerative colitis. Both are relapsing inflammatory diseases with a multifactorial cause involving immunological, microbial and environmental factors. A dysfunctional mucus barrier and/or altered goblet cell functions are probably involved in the pathogenesis of both diseases, but in Crohn’s disease, genome-wide association studies (GWAS) have identified altered bacterial sensing, ER stress, and autophagy as having a role in the pathogenesis, and Paneth cell dysfunction has been associated with disease onset in patients with Crohn’s disease153,154. Despite the clear association between Crohn’s disease and Paneth cell dysfunction, many Paneth cell features are shared with goblet cells, and the potential contribution of goblet cell dysfunction to the pathogenesis of Crohn’s disease should be taken into consideration155,156,157. Increased mucus thickness in areas at a distance from the inflammatory site was noted in patients with Crohn’s disease, further supporting altered goblet cell function in this disease158,159. A goblet cell-expressed gene, ITLN1, has been indicated as a susceptibility locus for Crohn’s disease in several GWAS, but the identified single-nucleotide polymorphisms might represent gene variants rather than disease-causing variations160. ITLN1 is a bacterial glycan-binding protein and upregulated ITLN1 expression was noted in immature goblet cells in patients with non-inflamed ulcerative colitis, but an altered expression was observed in patients with Crohn’s disease or ulcerative colitis with active disease77,84,85,160.

The role of an impaired mucus barrier in disease development is more intuitive for ulcerative colitis as it exclusively involves and starts in the distal colon, in which mucus protection limiting bacterial exposure and mechanical stress is important42. A thinner mucus layer was observed in the distal colon of patients with ulcerative colitis with increasing disease activity, but the mucus thickness was not altered in patients in remission158,159,161. Goblet cell depletion is frequently reported as a common feature in colitis, which probably reflects a combination of increased mucus release, reduced mucus storage, altered goblet cell differentiation and increased apoptosis162. The active phase of the disease is associated with the loss of mature goblet cells, as shown by the lack of sentinel goblet cells, which probably occurs as a result of decreased differentiation rate and enhanced cell death14. However, in patients in non-inflamed remission, the total number of goblet cells is mostly preserved, but reduced numbers of specific goblet cell populations can be observed42. For example, the inter-crypt goblet cells are prematurely shed regardless of the state of inflammation57 (Fig. 4). Defects in the inter-crypt mucus due to loss of inter-crypt goblet cells were observed in patients with ulcerative colitis in remission, suggesting that it contributes to the initiation of the disease and/or disease progression. The mechanisms driving the increased shedding of inter-crypt goblet cells are not known but might involve induced cell death. As these cells are fast mucus producers, ER stress can be a trigger that has also previously been associated with disease development102,104,105,154,156.

The unfolded-protein response (UPR) is important for goblet cells to cope with normal mucus production, but enhanced expression, misfolding mutations or defects in ER function can induce stress105,156,163,164. This response is evident as the murine goblet cell expression signature is enriched in many of the genes involved in UPR, especially in the IRE1α–XBBP1 pathway, such as Xbp1, Hspa5 and Ern1 (refs.57,80). However, this expression is lower in inter-crypt goblet cells, indicating altered and less demanding biosynthesis machinery or a way to control stress57,165. Specific regulation of the goblet cell-enriched protein IRE1β, which can control ER stress by inhibiting the IRE1α–XBBP1 pathway and potentially degrading specific mRNAs, such as Muc2, is probably important in goblet cell-specific ER functions104. Unresolved ER stress induces inflammation and cell death, defence mechanisms that in some cases can persist and cause reduced mucus production104,156,163. In mice, an impaired mucus barrier with enhanced bacterial–epithelial contact can cause a vicious cycle where both ER stress and bacterial exposure trigger immune reactions, but both inflammation and ER stress can be resolved by anti-inflammatory treatments166,167. To dissect the influence of microbial triggers in regulating the inflammatory response in the context of ER stress, a study in germ-free mice of a MUC2 misfolding model caused by a missense mutation revealed a major role for microorganisms in driving UPR and inflammation166. Under germ-free conditions, some effects remained, such as an elevated expression of chemokines, such as Cxcl2, and a slight increase in proliferation, indicating that these effects were directly triggered by mucin misfolding166. In this MUC2 misfolding mouse model, an altered expression of transcription factors, such as Hes1 and Spdef, was observed166. As these transcription factors are involved in the differentiation of the different types of goblet cells identified with single-cell RNA sequencing, their altered expression could indicate changes in the goblet cell composition57.

Another interesting aspect of the effect of inflammation on mucus layer integrity is that the structural properties of the mucus seem to be perturbed, with increased penetrability associated with inflammation as observed in a study of 28 patients with ulcerative colitis42. The altered properties of the mucus in inflamed tissue are potentially influenced by changes to the biosynthesis influencing multimerization, protein crosslinking, and reduced glycosylation, as shown in a study in 15 patients with active ulcerative colitis who were found to have shorter O-glycans on MUC2 (ref.168). In a study of 60 patients with ulcerative colitis, the composition of the mucus layer was also found to be affected, being characterized by reduced amounts of core mucus proteins, including MUC2 and FCGBP14. However, these alterations also occurred in the non-inflamed tissue in patients with active ulcerative colitis, which argues against inflammation being the main driver14. One explanation for this might be that the observed differences are on the protein level rather than the transcript level, as MUC2 expression was shown to be independent of disease state in patients with ulcerative colitis, whereas protein amounts changed, indicating post-transcriptional control169. Furthermore, the inability to form a functional mucus barrier is also influenced by the ionic environment and pH25,43. In the latter case, the chloride anion exchanger, SLC26A3, which has been shown to be important for mucus layer formation as its malfunction leads to chloride-rich diarrhoea, was downregulated in patients with ulcerative colitis14,85,170. Mucus expansion can either occur via an increase in cystic fibrosis transmembrane conductance regulator (CFTR)-mediated bicarbonate secretion43 or reduced fluid absorption mediated by sodium uptake involving sodium–hydrogen exchangers. Malfunction of sodium–hydrogen exchangers has been associated with mucus impairments and bacterial translocation leading to colitis171.

Studies of epithelial cells isolated from endoscopic biopsy samples from patients with inflammatory bowel disease revealed an altered gene expression profile in some goblet cells84,85. The total goblet cell number was not altered, but the number of goblet cell progenitors was reduced during active inflammation85. Concerning the observed reduction in progenitor cells, a novel inflammation-associated goblet cell population was identified84, characterized by upregulation of genes such as DMBT1, implying a shift in the canonical versus non-canonical goblet cell balance. In addition, the inflammatory process induced an altered bactericidal profile with upregulation of LYZ84. In contrast, ZG16, which mediates bacterial aggregation through binding peptidoglycan59, and WFDC2, a potential protease inhibitor with selective antimicrobial function, were downregulated in patients with ulcerative colitis84. In mice, Wfdc2 expression was mainly found in the non-canonical goblet cell lineage57. The expression pattern of genes usually confined to goblet cells was observed to expand to other cell types during inflammation (for example, the expression of proteins in the kynurenine pathway)57,85.

Although the pathogenesis of ulcerative colitis is unknown, chronic inflammation is thought to be driven by the gut microbiota, as many spontaneous colitis models do not develop disease under germ-free conditions, and altered microbial composition has frequently been reported in both mouse models of colitis and patients with ulcerative colitis151,172,173,174. The mucus has a central role in interactions between the gut microbiota and the host as the development of a functional mucus layer is dependent on the microbiota, as the mucus selects and supports specific bacteria such as Bacteroidaceae and Deferribacteraceae species175,176,177,178. Inflammation-induced changes to the microbiota can in part be explained by the enhanced proliferation, with increased use of aerobic glycolysis generating a more aerobic extracellular environment that contributes to the altered microbial composition107,173. A study including biopsies from 18 patients with ulcerative colitis in remission suggested that the metabolic shifts associated with inflammation impair goblet cell differentiation due to reduced mitochondrial β-oxidation179. How this increasingly aerobic environment and altered goblet cell metabolism affect the mucus layer is not known. Although the steady-state microbiota is not extensively invasive, increased bacterial translocation occurs during active disease180. Dissemination of bacteria outside the intestine, combined with inflammatory reactions (that is, the production of pro-inflammatory cytokines and chemokines) and altered composition of bacterial metabolites, can have implications for the development of distant inflammatory diseases, such as primary sclerosing cholangitis181.

To identify potential causative alterations in a disease such as ulcerative colitis, it is preferable to conduct investigations during disease development, but this is not easy to accomplish as patients do not seek medical care before disease onset. Although ulcerative colitis is a relapsing disease with periods characterized by epithelial restoration, sustained changes to goblet cell and mucus functions could either reflect chronic changes induced by previous periods of inflammation or predisposing defects182. Larger studies are required to establish a causative relationship between mucus defects and the pathogenesis of ulcerative colitis.

Host defence against infections

In addition to ensuring that we can live in symbiosis with our microbiota, the intestinal mucus has an important role in host defence against enteric pathogens. The continuous renewal of the mucus layer and its ability to support commensal bacteria, hold antimicrobial proteins and entrap microorganisms all assist in protecting the host from colonization by pathogens183. Thus, the ability of enteric pathogens to colonize their host largely depends on their ability to subvert the mucus barrier. The importance of goblet cells and mucus in host defence against enteric pathogens has been extensively studied in the context of parasite infections. Due to the large number of reviews covering this research area184,185,186, we focus on the role of goblet cells in host defence against bacteria and viruses.

Bacterial infections

Pathogenic Escherichia coli (enteropathogenic, enterotoxigenic and enterohaemorrhagic E. coli), Vibrio cholerae, Salmonella spp., Shigella spp., Yersinia spp. and Listeria spp. are all bacteria that infect the small intestine or colon, triggering periods of diarrhoea and subsequent inflammation. To colonize the intestine, the bacteria have to cross the mucus barrier to access the intestinal epithelium. To facilitate movement through the mucus, most enteric pathogens are equipped with flagella that enable them to propel themselves through the mucus towards the underlying epithelium187. To further facilitate access to the epithelium, pathogenic bacteria often express mucolytic enzymes that degrade the mucus gel188,189,190,191. The importance of the mucus layer in protecting the colon from pathogenic infection was demonstrated in mouse studies using the mouse-specific pathogen Citrobacter rodentium as a model of human E. coli infection. In wild-type mice, infection with C. rodentium resulted in a transient disease that cleared within approximately 2 weeks, whereas in Muc2−/− mice lacking intestinal mucus, C. rodentium infection was lethal183.

Pathogenic E. coli species, C. rodentium and V. cholerae, primarily reside in the intestinal lumen, in which they establish themselves on the epithelial surface, forming biofilms, whereas Salmonella spp., Shigella spp., Yersinia spp., and Listeria spp. are invasive species that cross the epithelial barrier192,193,194,195,196,197,198,199. It is well established that in mouse models of infection with Salmonella spp., Shigella spp. and Yersinia spp., the bacteria use M cells in the follicle-associated epithelium (FAE) covering Peyer’s patches and isolated lymphoid follicles to cross the intestinal barrier200,201,202,203,204. The sparsity of goblet cells in the FAE, and the subsequent thin mucus layer covering the FAE facilitate bacterial translocation to the tissue5. However, it is becoming clear that outside of the FAE these pathogens target and use goblet cells as ports of entry to the host.

In a humanized mouse model of Listeria monocytogenes infection, the bacteria infected the host cells by binding to E-cadherin via internalin A198. It was proposed that E-cadherin, an adherence junction protein, is exposed during mucus secretion allowing the bacteria to access the otherwise inaccessible protein and thereby specifically invade goblet cells198. Following binding to E-cadherin, the bacteria use the recycling machinery of the goblet cell to translocate through the cell, and live bacteria leave the goblet cell by exocytosis205. L. monocytogenes express the internalins B, C and J that all bind to Muc2, which might provide an additional mechanism for invading goblet cells206. A similar process was described for a mouse model of Salmonella infection in which S. typhimurium, in addition to translocating via M cells207, used goblet cells to cross the intestinal barrier144. Translocation of S. typhimurium through goblet cells depended on the bacterial virulence factor invasin and host factors, as invasin-deficient bacteria failed to translocate through goblet cells and translocation of wild-type bacteria only occurred in the presence of GAP formation144. Shigella flexneri, which infects the colon, targets goblet cells by glycan–glycan interactions208. Whether Yersinia uses goblet cells to cross the epithelium is not known; however, Y. pseudotuberculosis infection of the small intestine is invasin-dependent, suggesting that Yersinia spp. might also use goblet cells to cross the intestinal epithelium200,209.

The relationship between GAP formation and the ability of the bacteria to use goblet cells to cross the epithelial barrier was also observed in a mouse model of late-onset neonatal sepsis210. In this study, colonic GAP formation was prematurely activated by inhibition of EGFR signalling, followed by oral administration of commensal and pathogenic E. coli strains isolated from patients with late-onset sepsis210. In the absence of GAP formation, neither commensal nor pathogenic E. coli strains were able to cross the colonic barrier, whereas, in the presence of colonic GAPs, both strains used goblet cells to cross the barrier, but only the pathogenic E. coli strains caused disease210.

These studies demonstrate that several bacterial species target and use goblet cells to cross the intestinal barrier. These findings do not exclude the possibility that pathogenic bacteria invade other cell types such as enterocytes. However, it seems that translocation through goblet cells provides an efficient way for the bacteria to cross the epithelial barrier without being targeted for degradation. One explanation might be related to the high rate of regulated exocytosis of the goblet cells, which require continuous compensatory endocytosis to recycle secretory granule membrane, which pathogens can leverage to translocate through the cell211,212.

Viral infections

In addition to pathogenic and commensal bacteria using goblet cells to cross the epithelial barrier, studies have demonstrated that a number of human and mouse viruses also use goblet cells to cross the epithelial barrier and specifically target goblet cells for replication. Astroviruses (RNA viruses) are a dominant cause of childhood diarrhoea, and a study using a mouse model of astrovirus infection demonstrated that the virus primarily infects goblet cells that undergo active secretion and replicate specifically within goblet cells213,214. Enterovirus 71, which causes hand, foot and mouth disease in children, was also shown to preferentially infect and replicate within goblet cells when added to primary human epithelial monolayers215. Human adenoviruses (HAdVs) are a group of DNA viruses that cause a wide variety of diseases, including respiratory infection, conjunctivitis, cystitis and gastroenteritis. Although not all HAdVs cause gastritis, most of them replicate in the intestine, resulting in virus shedding and pathology at sites outside the intestine216. Of the HAdVs, HAdV-5p was shown to primarily infect and replicate in goblet cells in a human enteroid model216. Why and how these viruses specifically target goblet cells and the physiological consequences of specifically targeting goblet cells are not known. Astrovirus infection of goblet cells resulted in upregulation of Muc2, suggesting that a viral infection could positively affect the mucus barrier213. Further information on this topic can be found in a review by Cortez and Schultz-Cherry217.

Analysis of mucus and goblet cells

Histology has traditionally been used to study goblet cell and mucus morphology and structure. However, as the hydrated mucus shrinks dramatically during fixation of histological samples, there are limitations to its usefulness when it comes to translating findings from fixed tissue specimens to the properties of the native mucus gel. To overcome some of these limitations, in vivo, ex vivo and in vitro methods have been developed that enable more complex studies of goblet cell function and studies of the native mucus gel4,48,53,152,218 (Table 1). In vivo methods based on anaesthetized animals have the advantage of studying goblet cells and mucus in intact tissues4,48. This approach has been used to study mucus thickness and adhesiveness, but with current developments in intra-vital imaging techniques and the availability of mucin reporter mice (mCherry-Muc2 mice), in vivo analysis of mucus secretion and mucus properties is an area with great potential7. Concerning goblet cell function, intra-vital imaging has been used to study GAP formation and goblet cell interactions with MNPs132, an area that can easily be expanded to include interactions with other immune cells as well as stromal cells.

Table 1 Comparison of methods to analyse intestinal goblet cells and mucus

Although in vivo methods provide the most physiologically relevant environment, they are limited to animal models, which creates a need for ex vivo and in vitro methods to allow studies of goblet cell and mucus properties in human specimens. Ex vivo methods have been used to successfully study mucus secretion and mucus properties in the human colon and have proved to be a useful tool for studying mucus properties also in animal tissues42,53. Such systems also enable interventions that are not applicable in the in vivo setting, such as the use of substances affecting respiration and/or cardiac activity.

Despite providing a useful tool for studies of mucus properties, ex vivo methods are limited by the restricted survival time of the tissue (1–2 h)53, and this is where primary cell-based in vitro assays excel. Organoids, spheroids, 2D monolayers and organs on a chip provide useful tools in studies of goblet cell differentiation, mucus secretion and evaluation of mucus properties over time as these cultures can be followed and imaged over long periods219,220,221,222. By growing cells on scaffolds with a crypt–villus structure, the in vivo topography of the intestine can be mimicked. How well these systems resemble the in vivo setting of the small intestine and the colon remains to be explored, but they have the potential to enable long-term studies of the regulation of goblet cell differentiation, mucus secretion and mucus properties along the crypt–villus axis. Thus, with respect to the analysis of goblet cells and mucus properties, a large number of methods are available. It is important to consider the limitations and advantages of the respective methods and adapt them to the research question investigated.

Conclusions

During the past decade, intestinal goblet cells have emerged as central players in the regulation of intestinal health owing to their role in providing the protective mucus layer that covers the intestine, sensing changes in the local environment, and shaping gut immunity. Functional studies of goblet cell and mucus properties have revealed a highly adaptable system that is fine-tuned to meet the temporal and spatial needs of the small and large intestine throughout life. Alterations of goblet cells and mucus function are associated with the development of chronic inflammatory diseases, including colitis, and various pathogens have been shown to target goblet cells specifically to gain access to host tissues. Despite these advances in understanding goblet cell and mucus function, many fundamental questions remain to be answered (Box 2).