Skip to main page content
Access keys NCBI Homepage MyNCBI Homepage Main Content Main Navigation
. 2015 Jun 3;4(6):1140-1155.
doi: 10.1016/j.stemcr.2015.04.010. Online ahead of print.

Transcriptome-wide Analysis Reveals Hallmarks of Human Intestine Development and Maturation In Vitro and In Vivo

Affiliations
Free PMC article

Transcriptome-wide Analysis Reveals Hallmarks of Human Intestine Development and Maturation In Vitro and In Vivo

Stacy R Finkbeiner et al. Stem Cell Reports. .
Free PMC article

Abstract

Human intestinal organoids (HIOs) are a tissue culture model in which small intestine-like tissue is generated from pluripotent stem cells. By carrying out unsupervised hierarchical clustering of RNA-sequencing data, we demonstrate that HIOs most closely resemble human fetal intestine. We observed that genes involved in digestive tract development are enriched in both fetal intestine and HIOs compared to adult tissue, whereas genes related to digestive function and Paneth cell host defense are expressed at higher levels in adult intestine. Our study also revealed that the intestinal stem cell marker OLFM4 is expressed at very low levels in fetal intestine and in HIOs, but is robust in adult crypts. We validated our findings using in vivo transplantation to show that HIOs become more adult-like after transplantation. Our study emphasizes important maturation events that occur in the intestine during human development and demonstrates that HIOs can be used to model fetal-to-adult maturation.

Figures

Figure 1
Figure 1
HIOs Resemble Fetal Intestine (A) PCA was conducted on undifferentiated H9 ESCs (n = 3 independent experiments), definitive endoderm (n = 3 independently differentiated from H9s), HIOs (n = 3 independently differentiated from H9s), adult distal small intestine (n = 4 independent biological specimens), adult duodenum (n = 2 independent biological specimens), adult colon (n = 3 independent biological specimens), adult stomach (n = 3 independent biological specimens), fetal stomach (n = 6 independent biological specimens), and fetal small intestine (n = 6 independent biological specimens) (see also Table S1). The top two principal components account for 68.3% of the variation in the data and show that HIOs group with adult and fetal tissues of the intestine, stomach, and colon, all of which are clearly distinct from endoderm and ESCs. (B) A dendrogram was generated based on hierarchical clustering of the gene sets using Canberra distance. Fetal tissues are more similar to each other than to their respective adult tissues, and HIOs cluster within the same clade as the fetal tissues. Branch lengths indicate the level of dissimilarity between samples, with longer branch lengths indicating a greater level of dissimilarity. Red labels at branch point correspond to the approximately unbiased (AU) p value. AU > 95 indicates that a given branch assignment is strongly supported by the data. Green labels at each branch point correspond to the bootstrap probability (BP) of a cluster, defined as the frequency of a given relationship among the bootstrap replicates. ES, ESCs; HuSI, human small intestine; HuSI Duo, human small intestine duodenum; HuSI Dist, human distal small intestine; HuSto, human stomach. (C) Spearman ranking was used to cluster samples and generate a heatmap of similarity, with dark red indicating the highest level of similarity between samples. The dendrogram indicates that HIOs are most similar to human small intestine. (D) Differential gene expression analysis was carried out to identify genes that are upregulated in adult intestine compared to either HIOs (blue circle) or fetal intestine (yellow circle). The gene lists were then compared to determine how many genes overlap between those two comparisons (green), and the results are shown as a Venn diagram. (E) The list of genes from each of the three categories of the Venn diagram were subjected to GO analysis to identify enriched GO terms. The results are shown as REVIGO (Supek et al., 2011) scatterplots in which similar GO terms are grouped in arbitrary two-dimensional space based on semantic similarity. Each circle indicates a specific GO term and circle sizes are indicative of how many genes are included in each term, where larger circles indicate a greater number of genes that are included in that GO term. Colors indicate the level of significance of enrichment of the GO term. Asterisks indicate terms that are hidden under other circles. (F and G) Heatmaps were generated for gene sets belonging to the specific GO terms (F) Digestive Tract Development (GO:0048565) and (G) Defense Response (GO:0006952), and they further demonstrate that HIOs are more similar to fetal intestine with respect to these biological functions of the intestine. (H) A curated heatmap was generated to examine particular genes that were chosen as being some of the most highly differentially expressed genes and were also of biological interest. See also Figure S1.
Figure 2
Figure 2
HIOs Develop Mature Architecture after In Vivo Transplantation (A) H&E sections of an HIO, tHIO, and human intestine demonstrate that HIOs lack the characteristic crypt-villus units that are found in the adult small intestine and that transplanted HIOs (tHIOs) have a morphology that resembles the adult intestine. (B) Villi morphology was quantified by counting the number of projections with a vimentin-positive core. tHIOs contain a similar number of villi per millimeter of intestinal length as adult intestine, whereas HIOs contain less projections and even fewer true villi. Three independently derived biological replicates for each tissue type with five fields of view for each sample were analyzed. Averages were compared using unpaired t tests. p value 0.01–0.05, ∗∗p value 0.001–0.01. (C) Ki67 staining demonstrates that the crypt-like domains are also the site of localized proliferation in both tHIOs and adult tissue. (D) Staining for human nuclear antigen confirms that the tissue of intestinal morphology found in tHIOs is of human origin and not derived from the mouse host.
Figure 3
Figure 3
HIOs Express Brush Border Enzymes after In Vivo Transplantation (A) HIOs have sparse and short microvilli compared to the microvilli present in tHIO cells, as determined by electron microscopy. (B) Quantification of microvilli length indicates tHIOs have microvilli that are approximately two times longer than microvilli on HIO cells. Six independently derived HIOs and six tHIOs were analyzed with five measurements taken per sample. Averages were compared using unpaired t tests. ∗∗p value 0.001–0.01. (C) Normalized FPKMs were plotted for HIOs (n = 3 independently differentiated from H9s), fetal small intestine (n = 6 independent biological specimens), and adult small intestine (n = 6 independent biological specimens), with each sample represented by a black dot. Red bars indicate the average FPKM value for each group. Brush border enzymes involved in carbohydrate digestion are expressed at low levels in both HIOs and fetal intestine, whereas they are highly expressed in adult intestine. p value 0.01–0.05, ∗∗p value 0.001–0.01, ∗∗∗p value 0.0001–0.001. All samples are biological replicates. (D) Transplantation of HIOs into the kidney capsule results in an increase in the expression of brush border enzymes. qRT-PCR analysis was carried out by normalizing gene expression to the housekeeping gene GAPDH and plotting values as a.u. (n = at least 3 biological replicates for each tissue). (E) Immunofluorescence shows that sucrase isomaltase protein expression is not detectable in HIOs, but is evident in tHIOs and resembles the expression seen in adult intestine. See also Figures S2 and S3.
Figure 4
Figure 4
HIOs Form Crypts In Vivo Containing Mature Paneth Cells (A) Normalized FPKMs were plotted for HIOs (n = 3 independently derived H9-HIOs), fetal small intestine (n = 6 independent biological specimens), and adult small intestine (n = 6 independent biological specimens), with each sample represented by a black dot. Red bars indicate the average FPKM value for each group. Lysozyme (LYZ) is expressed in HIOs at a similar level to adult intestine; however, mature Paneth cell markers DEFA5 and REG3A are both expressed at high levels in adult intestine and expressed at very low levels in both HIO and fetal intestine. p value 0.01–0.05, ∗∗p value 0.001–0.01, ∗∗∗p value 0.0001–0.001. All samples are biological replicates. (B) Immunofluorescence staining shows that lysozyme is diffusely expressed in HIOs, but is localized to punctate granules in transplanted HIOs and adult intestine. (C) Transplantation of HIOs (tHIOs) into the kidney capsule results in an increase in expression of Paneth cell genes. qRT-PCR analysis was carried out by normalizing gene expression to the housekeeping gene GAPDH and plotting values as a.u. (n = at least 3 biologically independently differentiated HIOs and tHIOs, or independent biological intestinal specimens). (D) Immunofluorescence staining shows that α-defensin 5 is not detectable in HIOs, but shows a similar staining pattern in both tHIOs and adult intestine.
Figure 5
Figure 5
OLFM4 Is a Marker of Mature Intestine (A) Normalized FPKMs were plotted for HIOs (n = 3 independently differentiated from H9s), fetal small intestine (n = 6 independent biological specimens), and adult small intestine (n = 6 independent biological specimens), with each sample represented by a black dot. Red bars indicate the average FPKM value for each group. LGR5 is expressed to a similar level in both HIOs and adult human intestine; however, there is higher expression detected in human fetal intestine. In contrast, OLFM4 is expressed at a very low level in both HIOs and fetal intestine, but is dramatically higher in adult intestine. All samples are biological replicates. (B) OLFM4 expression is enhanced in tHIOs and is similar to the level of expression in adult intestine, whereas the level of expression in HIOs is much lower. qRT-PCR analysis was carried out by normalizing gene expression to the housekeeping gene GAPDH and plotting values as a.u. (n = at least 3 biological replicates for each tissue). (C) Immunohistochemistry demonstrates that HIOs do express OLFM4 protein; however, OLFM4 is sporadically expressed throughout HIOs and cannot be detected in all histological sections. After transplantation, OLFM4 is readily detectable in all histological sections, and it is localized to crypt-like domains in a pattern similar to adult human intestine. (D) In situ hybridization confirms that tHIOs express a higher level of LGR5 and that HIOs express a level similar to adult intestine. (E) Epithelium-only cultures were used to functionally analyze the responsiveness of enterocyte and stem cell gene expression to changes in culture conditions. Mature human enteroids express a higher level of IAP than immature HDEs when grown in complete media, which favors stem cells (gray bars, left). An increase in IAP expression in basal media for both enteroids and HDEs confirms that removal of pro-stem cell factors results in increased differentiation as expected (black bars, left). LGR5 is expressed to similar levels in both enteroids and HDEs grown in complete media (gray bars, middle). In contrast, OLFM4 is highly expressed in enteroids, but is barely detected in HDEs (gray bars, right). Growth of both enteroids and HDEs in basal media results in a reduction of LGR5 expression (black bars, middle); however, OLFM4 expression is only reduced in enteroids, but not HDEs grown in basal media (black bars, right). n = 3 biological replicates for all tissues. p value 0.01–0.05. (F) To confirm the developmental pattern of OLFM4 seen in human tissues, we also examined histological sections of mouse intestine at different developmental ages. OLFM4 was only detected in the adult intestine and not at any of the embryonic ages examined. See also Figures S4 and S5.
Figure 6
Figure 6
HIOs Express Mesenchymal Cell Markers but Become Organized Like Adult Intestine after Transplantation (A) Normalized FPKMs were plotted for HIOs (n = 3 independently differentiated from H9s), fetal small intestine (n = 6 independent biological specimens), and adult small intestine (n = 6 independent biological specimens), with each sample represented by a black dot. Red bars indicate the average FPKM value for each group. VIM is expressed to a similar level across HIOs and fetal and adult human intestines. HIOs express lower levels of ACTG2 than adult tissue, with fetal tissue expressing a range of levels. Most of the HIO and fetal intestine samples express a higher level of PDGFRa than adult tissue. (B) The expression trends are confirmed by immunofluorescence staining with most notable changes in the spatial arrangement of positive cells for all three markers. All samples are biological replicates with an n. p value 0.01–0.05, ∗∗p value 0.001–0.01, ∗∗∗p value 0.0001–0.001.

Similar articles

See all similar articles

Cited by 66 articles

See all "Cited by" articles

References

    1. Bell S.M., Schreiner C.M., Wert S.E., Mucenski M.L., Scott W.J., Whitsett J.A. R-spondin 2 is required for normal laryngeal-tracheal, lung and limb morphogenesis. Development. 2008;135:1049–1058. - PubMed
    1. Brown B.N., Badylak S.F. Extracellular matrix as an inductive scaffold for functional tissue reconstruction. Transl. Res. 2014;163:268–285. - PMC - PubMed
    1. Cash H.L., Whitham C.V., Behrendt C.L., Hooper L.V. Symbiotic bacteria direct expression of an intestinal bactericidal lectin. Science. 2006;313:1126–1130. - PMC - PubMed
    1. Chen Y.-J., Finkbeiner S.R., Weinblatt D., Emmett M.J., Tameire F., Yousefi M., Yang C., Maehr R., Zhou Q., Shemer R. De novo formation of insulin-producing “neo-β cell islets” from intestinal crypts. Cell Rep. 2014;6:1046–1058. - PMC - PubMed
    1. Clevers H.C., Bevins C.L. Paneth cells: maestros of the small intestinal crypts. Annu. Rev. Physiol. 2013;75:289–311. - PubMed

LinkOut - more resources

Feedback