ABSTRACT
Syndromic congenital tufting enteropathy (CTE) is a life-threatening recessive human genetic disorder that is caused by mutations in SPINT2, encoding the protease inhibitor HAI-2, and is characterized by severe intestinal dysfunction. We recently reported the generation of a Spint2-deficient mouse model of CTE. Here, we show that the CTE-associated early-onset intestinal failure and lethality of Spint2-deficient mice is caused by unchecked activity of the serine protease matriptase. Macroscopic and histological defects observed in the absence of HAI-2, including villous atrophy, luminal bleeding, loss of mucin-producing goblet cells, loss of defined crypt architecture and the resulting acute inflammatory response in the large intestine, were all prevented by intestinal-specific inactivation of the St14 gene encoding matriptase. The CTE-associated loss of the cell junctional proteins EpCAM and claudin 7 was also prevented. As a result, inactivation of intestinal matriptase allowed Spint2-deficient mice to gain weight after birth and dramatically increased their lifespan. These data implicate matriptase as a causative agent in the development of CTE and may provide a new target for the treatment of CTE in individuals carrying SPINT2 mutations.
This article has an associated ‘The people behind the papers’ interview.
INTRODUCTION
Congenital tufting enteropathy (CTE, OMIM #613217) is an early-onset severe intestinal insufficiency characterized by epithelial dysplasia, villous atrophy and a compromised intestinal epithelial barrier, leading to chronic watery diarrhea, dehydration and failure to thrive in the absence of parenteral feeding (Sivagnanam et al., 2008). More than 70% of individuals with CTE carry bi-allelic mutations in the EPCAM gene, which encodes the epithelial cell adhesion molecule (EpCAM), a highly conserved cell surface glycoprotein involved in regulation of epithelial cell physiology and barrier function (Huang et al., 2018; Maetzel et al., 2009; Maghzal et al., 2013; Salomon et al., 2014). A syndromic form of CTE, also termed congenital sodium diarrhea (OMIM #270420), affects about 20% of individuals with CTE and is caused by loss-of-function mutations in the SPINT2 gene, encoding the transmembrane Kunitz-type serine protease inhibitor HAI-2 that is widely expressed in epithelial compartments of most mouse and human tissues (Heinz-Erian et al., 2009; Holt-Danborg et al., 2019; Kataoka et al., 2018; Salomon et al., 2014; Szabo et al., 2008).
In mice, HAI-2 is essential for the completion of embryonic development and, depending on the genetic background, HAI-2 deficiency is associated with an early or mid-gestational embryonic lethality, high frequency of neural tube defects, including exencephaly, spina bifida and curly tail, and placental failure (Mitchell et al., 2001; Szabo et al., 2009, 2012). All of these developmental defects are rescued by the inactivation of either of the two membrane-anchored serine proteases matriptase (ST14/MT-SP1/epithin) or prostasin (CAP-1/PRSS8), demonstrating the crucial role of HAI-2 in the regulation of biological function of the two proteases during prenatal development in mice (Szabo et al., 2009, 2012). Furthermore, studies using genetically modified mouse or rat strains carrying inactivating mutations in the St14 and Prss8 genes, encoding, respectively, matriptase and prostasin, revealed that the two proteases also play crucial roles in epithelial development and function in a number of tissues, including skin, salivary gland, intestines, lungs, placenta and thymus (Frateschi et al., 2012; Keppner et al., 2016; List et al., 2002, 2009; Malsure et al., 2014; Planès et al., 2010; Szabo et al., 2014). Presence of loss-of-function mutations in the ST14 gene in humans with autosomal recessive ichthyosis and hypotrichosis (ARIH)/ichthyosis, follicular atrophoderma and hypotrichosis (IFAH), and in horses with naked foal syndrome indicate that the function of matriptase in epithelial development and homeostasis may be evolutionarily conserved across mammalian species (Alef et al., 2009; Avrahami et al., 2008; Basel-Vanagaite et al., 2007; Bauer et al., 2017; Neri et al., 2016).
The overlapping phenotypes observed in matriptase- and prostasin-deficient mice, and their shared contribution to embryonic lethality observed in mice lacking either HAI-2 or its close homolog HAI-1, suggest that, at least in some tissues, the two proteases are part of a single proteolytic cascade that is fundamental for regulation of epithelial development and function. Indeed, several studies document the ability of prostasin to directly bind matriptase and to stimulate matriptase activity in cell culture as well as in tissues, suggesting that prostasin is likely to act upstream of matriptase (Buzza et al., 2013; Camerer et al., 2010; Friis et al., 2013; Szabo et al., 2012). Interaction with prostasin appears to be crucial for matriptase function in some tissues, including skin, but not in others, such as intestine, as documented by a complete loss of intestinal epithelial integrity and postnatal viability in mice deficient in intestinal matriptase, but not prostasin (Kosa et al., 2012; Malsure et al., 2014).
Identification of loss-of-function SPINT2 mutations in individuals with the syndromic form of CTE suggests that the disease develops in response to an unchecked serine protease activity. In in vitro and in cell culture-based systems, HAI-2 displays potent inhibitory activity towards a large number of membrane-bound and soluble trypsin-like serine proteases, including hepsin, matriptase, matriptase-2, prostasin, HGFA, coagulation factors IXa, Xa, XIa and XIIa, as well as several members of kallikrein family, presenting a considerable challenge in identification of physiologically relevant inhibitory targets (Beckmann et al., 2016; Delaria et al., 1997; Herter et al., 2005; Kirchhofer et al., 2005; Maurer et al., 2013; Qin et al., 1998; Solis-Calero and Carvalho, 2017; Szabo et al., 2008). However, genetic silencing of HAI-2 in cultured human intestinal epithelial cells was recently shown to increase cleavage of EpCAM by matriptase, causing premature degradation of the tight junction protein claudin 7, thus implicating matriptase hyperactivity in the etiology of the syndromic form of CTE (Wu et al., 2017). Similarly, increased cleavage of EpCAM in Spint2-deficient mouse intestinal organoids was alleviated by shRNA targeting of matriptase expression or by the matriptase-selective synthetic inhibitor ZFH7185-8 (Kawaguchi et al., 2019). In contrast, a recent study by Holt-Danborg and co-workers reported that, in a cell culture-based system, CTE-associated missense mutations in SPINT2 reduce the ability of the mutant HAI-2 proteins to inhibit prostasin but not matriptase, proposing that unopposed activity of prostasin or prostasin-like activity, rather than that of matriptase, is responsible for development of syndromic CTE (Holt-Danborg et al., 2019).
In this report, we use an in vivo model to directly test the role of matriptase in the development of intestinal abnormalities associated with the loss of HAI-2 function. We show that loss of HAI-2 leads to an increased activation of the matriptase zymogen, and that intestinal inactivation of matriptase in HAI-2-deficent mice restores normal macroscopic and histological appearance of intestinal tissue, expression of CTE-associated markers EpCAM and claudin 7, as well as the ability to gain weight, and leads to a substantial increase in postnatal survival. The study identifies matriptase as a causative agent in the development of early-onset intestinal failure resulting from loss of HAI-2 function.
RESULTS
Genetic inactivation of intestinal matriptase improves postnatal survival of HAI-2-deficient mice
Recent reports using cell culture and organoid culture models have suggested unchecked activity of the serine protease matriptase as the likely cause of severe defects in intestinal development and function that are associated with loss-of-function mutations in SPINT2, which encodes the serine protease inhibitor HAI-2 in both humans and mice (Kawaguchi et al., 2019; Wu et al., 2017). We have previously developed a mouse model that recapitulates main histological and molecular characteristics of CTE by inactivating Spint2 in mice that carry a point mutation in a Prss8 gene encoding the serine protease prostasin (Spint2−/−;Prss8R44Q/R44Q), thereby bypassing the embryonic lethality associated with HAI-2 deficiency (Szabo and Bugge, 2018). To investigate whether loss of HAI-2 affects the intestinal activity of matriptase, protein levels of matriptase zymogen and the activated double-chain form of matriptase were assayed in intestinal tissues from 5-day-old HAI-2- and matriptase-expressing (control), and HAI-2 deficient (Spint2−/−;Prss8R44Q/R44Q) mice by western blot (Fig. 1A, intestinal matriptase-deficient Vil-Cre+/0;St14fl/fl tissues were used as negative control). Interestingly, the absence of HAI-2 lead to a considerable decrease in levels of matriptase zymogen (20% of control, P=0.021, band intensities normalized to average signal observed in matriptase-deficient tissues). However, western blot of whole-cell lysates failed to conclusively detect any active form of matriptase, either due to low abundance of this form, or because of an overlap of the activated serine protease domain with a non-specific band showing at ∼30 kDa. In order to circumvent this problem, we performed immunoprecipitation using antibody against the endogenous inhibitor of matriptase, HAI-1, that specifically binds the activated double-chain, but not the single-chain, zymogen form of matriptase (Lin et al., 2008; Szabo et al., 2012). Subsequent detection of matriptase by western blot showed a dramatic increase in the amount of the activated matriptase serine protease domain in tissues from HAI-2-deficient mice (Fig. 1B, P=0.003). This indicates that loss of HAI-2 does indeed lead to a substantial increase in matriptase activation and, presumably, activity in intestinal tissues.
Regulation of matriptase activation by HAI-2 is crucial for early postnatal survival. (A,B) Western blot detection of single-chain zymogen (A) and activation site-cleaved serine protease domain (B) of matriptase in intestinal tissues from HAI-2- and matriptase-expressing (Spint2+/+;St14+/+;Prss8R44Q/R44Q, lanes 1-3), HAI-2-deficient (Spint2−/−;St14+/+;Prss8R44Q/R44Q, lanes 4-6) or matriptase-deficient (Vil-Cre+/0;St14fl/fl, lanes 7-9) mice 5 days after birth. Loss of HAI-2 leads to an 80% decrease in the amount of matriptase zymogen, but more than a 4.5-fold increase in levels of active double-chain matriptase. (C) Western blot detection of matriptase in intestinal lysates from Vil-Cre0;St14fl/fl (St14+/+, lanes 1 and 2), and Vil-Cre+/0;St14fl/fl (St14−/−, lanes 3 and 4) mice in a Prss8R44Q/R44Q background. Conditional inactivation of the St14 gene resulted in a near complete loss of matriptase protein. (A-C) Positions of matriptase zymogen (black arrowhead), activated matriptase catalytic domain (grey arrowhead) and GAPDH control (open arrowhead) are indicated on the right. Positions of molecular weight markers (kDa) are shown on the left. (D) Distribution of matriptase genotypes among newborn offspring from Vil-Cre+/0;St14fl/+;Prss8R44Q/R44Q×St14fl/fl;Prss8R44Q/R44Q breeding pairs. Matriptase deficiency (Vil-Cre+/0;St14fl/fl;Prss8R44Q/R44Q) was observed in 74 out of 336 newborn pups, consistent with the expected Mendelian frequency of 25% (P=0.21, χ2). (E) Postnatal survival of Spint2-deficient mice carrying two (St14+/+, blue), one (St14+/−, red) or no (St14−/−, green) functional allele(s) of intestinal matriptase. Median postnatal survival increased with decreasing gene dose of matriptase. (F,G) Total postnatal body weight (F) and overall outward appearance (G) of HAI-2- and matriptase-expressing (control), Spint2 single-deficient (Spint2−/−;St14+/+), and Spint2- and intestinal St14 double-deficient (Spint2−/−;St14−/−) mice. Inactivation of matriptase restored the ability of HAI-2-deficient mice to grow after birth. (H,I) Outward appearance (H) and length (I) of small and large intestines from 5-day-old HAI-2- and matriptase-expressing (Control), Spint2 single-deficient (Spint2−/−;St14+/+), and Spint2- and intestinal St14 double-deficient (Spint2−/−;St14−/−) mice. Loss of HAI-2 leads to severe shortening of intestinal tissues that is rescued by simultaneous inactivation of matriptase. *P<0.05, **P<0.01, ***P<0.001. Scale bars: 1 cm.
Regulation of matriptase activation by HAI-2 is crucial for early postnatal survival. (A,B) Western blot detection of single-chain zymogen (A) and activation site-cleaved serine protease domain (B) of matriptase in intestinal tissues from HAI-2- and matriptase-expressing (Spint2+/+;St14+/+;Prss8R44Q/R44Q, lanes 1-3), HAI-2-deficient (Spint2−/−;St14+/+;Prss8R44Q/R44Q, lanes 4-6) or matriptase-deficient (Vil-Cre+/0;St14fl/fl, lanes 7-9) mice 5 days after birth. Loss of HAI-2 leads to an 80% decrease in the amount of matriptase zymogen, but more than a 4.5-fold increase in levels of active double-chain matriptase. (C) Western blot detection of matriptase in intestinal lysates from Vil-Cre0;St14fl/fl (St14+/+, lanes 1 and 2), and Vil-Cre+/0;St14fl/fl (St14−/−, lanes 3 and 4) mice in a Prss8R44Q/R44Q background. Conditional inactivation of the St14 gene resulted in a near complete loss of matriptase protein. (A-C) Positions of matriptase zymogen (black arrowhead), activated matriptase catalytic domain (grey arrowhead) and GAPDH control (open arrowhead) are indicated on the right. Positions of molecular weight markers (kDa) are shown on the left. (D) Distribution of matriptase genotypes among newborn offspring from Vil-Cre+/0;St14fl/+;Prss8R44Q/R44Q×St14fl/fl;Prss8R44Q/R44Q breeding pairs. Matriptase deficiency (Vil-Cre+/0;St14fl/fl;Prss8R44Q/R44Q) was observed in 74 out of 336 newborn pups, consistent with the expected Mendelian frequency of 25% (P=0.21, χ2). (E) Postnatal survival of Spint2-deficient mice carrying two (St14+/+, blue), one (St14+/−, red) or no (St14−/−, green) functional allele(s) of intestinal matriptase. Median postnatal survival increased with decreasing gene dose of matriptase. (F,G) Total postnatal body weight (F) and overall outward appearance (G) of HAI-2- and matriptase-expressing (control), Spint2 single-deficient (Spint2−/−;St14+/+), and Spint2- and intestinal St14 double-deficient (Spint2−/−;St14−/−) mice. Inactivation of matriptase restored the ability of HAI-2-deficient mice to grow after birth. (H,I) Outward appearance (H) and length (I) of small and large intestines from 5-day-old HAI-2- and matriptase-expressing (Control), Spint2 single-deficient (Spint2−/−;St14+/+), and Spint2- and intestinal St14 double-deficient (Spint2−/−;St14−/−) mice. Loss of HAI-2 leads to severe shortening of intestinal tissues that is rescued by simultaneous inactivation of matriptase. *P<0.05, **P<0.01, ***P<0.001. Scale bars: 1 cm.
To address the potential involvement of matriptase in the development of a CTE-like phenotype in Spint2-deficient mice, we genetically inactivated epithelial matriptase by crossing Spint2−/−;Prss8R44Q/R44Q mouse strain to mice carrying a conditional knockout allele of matriptase (St14fl/fl) and a transgene expressing Cre recombinase under control of the intestinal epithelium-specific villin promoter (Vil-Cre+/0). The resulting Vil-Cre+/0;St14fl/fl;Prss8R44Q/R44Q mice show little to no detectable matriptase protein expression within small or large intestines (Fig. 1C). Importantly, loss of intestinal matriptase did not affect embryonic development and prenatal survival in the Prss8R44Q/R44Q genetic background, as Vil-Cre+/0;St14fl/fl;Prss8R44Q/R44Q mice were born in the expected Mendelian ratio and did not exhibit any obvious developmental abnormality (Fig. 1D), thus enabling analysis of the effect of expression of intestinal matriptase on development and survival of Spint2−/−;Prss8R44Q/R44Q mice after birth. As all the mice used in this study from this point onwards, including all littermate controls, were homozygous for the prostasin/Prss8 R44Q mutation (Prss8R44Q/R44Q), for simplicity hereafter we will refer to these mice only as Spint2-deficient, or Spint2−/−.
Consistent with our previous study (Szabo and Bugge, 2018), analysis of postnatal survival revealed that all of the Spint2−/− mice carrying two functional alleles of matriptase (Spint2−/−;St14fl/fl or fl/+;Vil-Cre0) showed gross growth retardation and died typically within the first week after birth, with a median survival of 6.5 days (Fig. 1E-G). In contrast, initial postnatal growth was largely normalized in HAI-2-deficient mice lacking functional alleles of the matriptase gene (Spint2−/−;St14fl/fl;Vil-Cre+/0) (Fig. 1F, P<0.001), although these mice generally could be distinguished from their littermate controls based on their smaller size at or after the end of their first week of life (Fig. 1G). Matriptase inactivation within intestinal epithelium also dramatically improved the lifespan of Spint2−/− mice, increasing median postnatal survival to 19.5 days (Fig. 1E, P<0.0001). Interestingly, postnatal survival was also markedly improved by inactivation of only one allele of the matriptase gene (Spint2−/−;St14fl/+;Vil-Cre+/0), although not to an extent observed in complete absence of matriptase [Fig. 1E, P<0.0001 (St14+/−//St14+/+), P<0.005 (St14+/−//St14−/−)]. Collectively, these data suggest that the detrimental effects of the loss of HAI-2 on postnatal development and survival in Spint2−/− mice can be largely attributed to intestinal matriptase activity.
Defects in intestinal development in Spint2−/−;Prss8R44Q/R44Q mice correlate with gene dose of St14/matriptase
Our analysis of intestinal development in Spint2−/−;Prss8R44Q/R44Q mice revealed significant shortening of large intestine, and progressive deterioration of epithelial architecture in both small and large intestines as the likely cause of postnatal lethality (Szabo and Bugge, 2018). To evaluate the contribution of matriptase activity to the observed defects, we first examined macroscopic and histological appearance of intestinal tissues from 5-day-old HAI-2-deficient mice expressing at least one (Spint2−/−;St14+) or no (Spint2−/−;St14fl/fl;Vil-Cre+/0) functional allele of intestinal matriptase. Consistent with our previous data, intestines from HAI-2-deficient mice expressing matriptase appeared distended and considerably shorter (Fig. 1H,I, P<0.0005 for small intestines; P<0.0001 for colons) (Szabo and Bugge, 2018). Histologically, intestines of matriptase-expressing Spint2−/− mice exhibited widespread mild to moderate villous atrophy, crowding and occasional tufting of epithelial cells in small intestine, as well as severe disorganization of colonic epithelium associated with greatly enhanced exfoliation of epithelial cells and a complete loss of crypt structure in large intestine (Fig. 2A, compare middle to left panels). In contrast, intestines of Spint2−/−;St14fl/fl;Vil-Cre+/0 mice did not show any obvious developmental abnormality at this age, and were macroscopically and histologically indistinguishable from their wild-type littermate controls (Figs 1H,I and 2A).
Matriptase inactivation restores normal organization and differentiation of intestinal epithelium in Spint2-deficient mice. (A,B) Hematoxylin and Eosin (A) and Alcian Blue/PAS (B) staining of small (top panels) and large (bottom panels) intestines from control (left), HAI-2 single-deficient (Spint2−/−;St14+, middle), and HAI-2 and matriptase double-deficient (Spint2−/−;St14−/−, right) mice at postnatal day 5. Loss of HAI-2 was associated with: increased vacuolization (A, arrows); epithelial crowding and dyslocalization of nuclei along the base of villous epithelium (A, black arrowheads); tufting (A, upper panels, open arrowheads); general villous atrophy of small intestine and an increased epithelial erosion (A, lower panels, open arrowheads); a near-complete loss of crypt structure in large intestines; and a substantial decrease and a near complete loss of mucin-producing goblet cells (B, arrows) in small and large intestine. All defects were rescued by genetic ablation of intestinal matriptase. (C) Quantification of Alcian Blue/PAS staining of small and large intestines from control (blue), HAI-2 single-deficient (red), and HAI-2 and matriptase double-deficient (green) mice at postnatal day 5. Graphs show individual values, and mean±s.d. based on five mice per genotype. *P<0.05, ***P<0.001. Scale bars: 50 µm in A; 25 µm in B.
Matriptase inactivation restores normal organization and differentiation of intestinal epithelium in Spint2-deficient mice. (A,B) Hematoxylin and Eosin (A) and Alcian Blue/PAS (B) staining of small (top panels) and large (bottom panels) intestines from control (left), HAI-2 single-deficient (Spint2−/−;St14+, middle), and HAI-2 and matriptase double-deficient (Spint2−/−;St14−/−, right) mice at postnatal day 5. Loss of HAI-2 was associated with: increased vacuolization (A, arrows); epithelial crowding and dyslocalization of nuclei along the base of villous epithelium (A, black arrowheads); tufting (A, upper panels, open arrowheads); general villous atrophy of small intestine and an increased epithelial erosion (A, lower panels, open arrowheads); a near-complete loss of crypt structure in large intestines; and a substantial decrease and a near complete loss of mucin-producing goblet cells (B, arrows) in small and large intestine. All defects were rescued by genetic ablation of intestinal matriptase. (C) Quantification of Alcian Blue/PAS staining of small and large intestines from control (blue), HAI-2 single-deficient (red), and HAI-2 and matriptase double-deficient (green) mice at postnatal day 5. Graphs show individual values, and mean±s.d. based on five mice per genotype. *P<0.05, ***P<0.001. Scale bars: 50 µm in A; 25 µm in B.
Mucin-producing goblet cells are responsible for the formation and maintenance of the protective mucus layer throughout the small and large intestine, and originate from stem cells at the bottom of intestinal crypts (Specian and Oliver, 1991). A significant reduction in the number of goblet cells was one of the prominent histological changes observed in the intestinal tissues from HAI-2-deficient mice (Kawaguchi et al., 2019; Szabo and Bugge, 2018). Histological examination of intestines from 5-day-old matriptase-expressing (Spint2−/−) or matriptase-deficient (Spint2−/−;St14fl/fl;Vil-Cre+/0) HAI-2 knockouts, and their littermate controls, revealed that inactivation of matriptase was able to completely restore differentiation of goblet cells and mucin production in both small and large intestines (Fig. 2B,C). These data strongly indicate that defects in differentiation and general organization of intestinal epithelium resulting from the loss of HAI-2 are caused by unchecked matriptase activity.
Inactivation of matriptase prevents inflammatory response to loss of HAI-2
The intestinal epithelium and mucosal surface form a crucial barrier separating gut bacteria and other luminal content from tissue stroma that contains one of the most active immune systems in our body. The aberrant structure of intestinal epithelium, including reduction in mucin-producing goblet cells and loss of an organized epithelial monolayer observed in the large intestines of HAI-2-deficient mice (Szabo and Bugge, 2018; see above), can lead to an increased leakage of material present in the lumen into the underlying tissue, affecting tissue homeostasis and triggering an immune response, thus contributing to the disease progression. To assess the inflammatory response to the intestinal damage invoked by loss of HAI-2 in mice, we examined stromal infiltration of immune cells into small and large intestines of 5-day-old mice. Immunostaining for CD11b, a general marker of innate leukocytes and phagocytic cells, including monocytes, macrophages, granulocytes and NK cells, showed a relatively modest 1.8-fold increase in the number of positive cells within the small intestine from 5-day-old Spint2−/− mice (Fig. 3A,B, P<0.05). Similarly, loss of HAI-2 was associated with about a twofold increase in the number of cells positive for the general marker of both native and adaptive immune cells, CD45 (Fig. 3C and D, P<0.01), consistent with a relatively mild effect of HAI-2 deficiency on goblet cell number, structure and continuity of epithelial layer in small intestines observed in this study. In contrast, loss of HAI-2 led to a substantial increase in infiltrating immune cells in large intestines, with a 7.7-fold increase in number of CD11b-positive and 4.6-fold increase in number of CD45-positive cells (P<0.001 and <0.005, respectively) in colons of Spint2−/− mice compared with their HAI-2-expressing littermate controls (Fig. 3A-D). Inactivation of matriptase prevented enhanced recruitment of immune cells into Spint2−/− tissues, as evidenced by the numbers of CD11b- or CD45-positive cells in the intestines from Spint2−/−;St14fl/fl;Vil-Cre+/0 mice that were comparable with those observed in HAI-2- and matriptase-expressing littermate controls (Fig. 3A-D).
HAI-2 deficiency triggers matriptase-dependent immune infiltration. (A,C) Representative images of anti-CD11b (A, arrowheads) and anti-CD45 (C, arrowheads) immunostaining of small (upper panels) and large (lower panels) intestines from control (left), HAI-2 single-deficient (Spint2−/−;St14+, middle), and HAI-2 and matriptase double-deficient (Spint2−/−;St14−/−, right) mice at postnatal day 5. (B,D) Quantification of CD11b-positive (B) and CD45-positive (D) cells in postnatal tissues from control (blue), HAI-2 single-deficient (red), and HAI-2 and matriptase double-deficient (green) mice at postnatal day 5 based on five mice per genotype. Loss of HAI-2 led to an increased infiltration of immune cells into both small and large intestines that was reversed by inactivation of matriptase. *P<0.05, **P<0.01, ***P<0.001. Scale bars: 25 µm.
HAI-2 deficiency triggers matriptase-dependent immune infiltration. (A,C) Representative images of anti-CD11b (A, arrowheads) and anti-CD45 (C, arrowheads) immunostaining of small (upper panels) and large (lower panels) intestines from control (left), HAI-2 single-deficient (Spint2−/−;St14+, middle), and HAI-2 and matriptase double-deficient (Spint2−/−;St14−/−, right) mice at postnatal day 5. (B,D) Quantification of CD11b-positive (B) and CD45-positive (D) cells in postnatal tissues from control (blue), HAI-2 single-deficient (red), and HAI-2 and matriptase double-deficient (green) mice at postnatal day 5 based on five mice per genotype. Loss of HAI-2 led to an increased infiltration of immune cells into both small and large intestines that was reversed by inactivation of matriptase. *P<0.05, **P<0.01, ***P<0.001. Scale bars: 25 µm.
Absence of matriptase restores expression of epithelial junctional proteins
Increased turnover of epithelial junctional proteins EpCAM and claudin 7 has been implicated in the etiology of intestinal failure in individuals with CTE who carry SPINT2 mutations (Wu et al., 2017). Furthermore, a significant decrease in the expression of the two proteins in the intestines from two independently generated Spint2-deficient mouse models has been reported recently (Kawaguchi et al., 2019; Szabo and Bugge, 2018). To investigate whether the observed decrease in the expression of EpCAM and claudin 7 can be also attributed to the lack of regulation of matriptase activity by HAI-2, we next analyzed expression of the two proteins in intestines of 5-day-old HAI-2 single-deficient (Spint2−/−;St14+) or HAI-2 and matriptase double-deficient (Spint2−/−;St14fl/fl;Vil-Cre+/0) mice. Consistent with our previous study, tissues from HAI-2 single-deficient mice showed normal levels of EpCAM and claudin 7 transcripts but significantly reduced expression of both EpCAM and claudin 7 proteins compared with HAI-2-expressing littermate controls (Fig. 4A-C, P<0.001) (Szabo and Bugge, 2018). Loss of HAI-2 did not affect expression of the intestine-specific epithelial marker villin, suggesting that loss of the two junctional proteins is not a result of a decreased number of epithelial cells due to altered tissue structure or increased infiltration of immune cells (Fig. 4A-C). Importantly, the expression of both EpCAM and claudin 7 proteins was restored in HAI-2-deficient mice lacking intestinal matriptase (Fig. 4A,B, P>0.05 [control versus Spint2−/−;St14fl/fl,Vil-Cre+/0) for both EpCAM and claudin 7].
Matriptase inactivation restores expression of EpCAM and claudin 7 in HAI-2-deficient intestines. (A) Western blot analysis of the expression of EpCAM, claudin 7, villin and GAPDH in intestines from HAI-2- and matriptase-expressing control (lanes 1-3), HAI-2 single-deficient (Spint2−/−;St14+, lanes 4-6), and HAI-2 and matriptase double-deficient (Spint2−/−;St14−/−, lanes 7-9) mice at postnatal day 5. Expected positions of the protein bands are indicated on the right (arrowheads). Positions of molecular weight markers (in kDa) are indicated on the left. (B,C) Quantification of the relative expression of EpCAM, claudin 7 and villin proteins (B) and mRNA (C). Inactivation of intestinal matriptase largely restored expression levels of both EpCAM and claudin 7 in Spint2−/− mice. Loss of HAI-2 did not affect expression of genes encoding EpCAM and claudin 7 or gene or protein expression of an intestinal epithelial marker villin. (D,E) Representative images of anti-EpCAM (D) and anti-claudin 7 (E) immunostaining of small (upper panels) and large (lower panels) intestines from control (left), HAI-2 single-deficient (Spint2−/−;St14+, middle), and HAI-2 and matriptase double-deficient (Spint2−/−;St14−/−, right) mice at postnatal day 5. Loss of HAI-2 lead to a prominent decrease in levels of membrane-associated EpCAM in small and large intestines, to a corresponding decrease and segmentation of claudin 7 expression in villous epithelium of small intestine, and to a complete loss of membrane association in large intestines (E, arrows). Simultaneous inactivation of matriptase restored the expression level and cell surface localization of both proteins. *P<0.05, **P<0.01, ***P<0.001. Scale bars: 20 µm.
Matriptase inactivation restores expression of EpCAM and claudin 7 in HAI-2-deficient intestines. (A) Western blot analysis of the expression of EpCAM, claudin 7, villin and GAPDH in intestines from HAI-2- and matriptase-expressing control (lanes 1-3), HAI-2 single-deficient (Spint2−/−;St14+, lanes 4-6), and HAI-2 and matriptase double-deficient (Spint2−/−;St14−/−, lanes 7-9) mice at postnatal day 5. Expected positions of the protein bands are indicated on the right (arrowheads). Positions of molecular weight markers (in kDa) are indicated on the left. (B,C) Quantification of the relative expression of EpCAM, claudin 7 and villin proteins (B) and mRNA (C). Inactivation of intestinal matriptase largely restored expression levels of both EpCAM and claudin 7 in Spint2−/− mice. Loss of HAI-2 did not affect expression of genes encoding EpCAM and claudin 7 or gene or protein expression of an intestinal epithelial marker villin. (D,E) Representative images of anti-EpCAM (D) and anti-claudin 7 (E) immunostaining of small (upper panels) and large (lower panels) intestines from control (left), HAI-2 single-deficient (Spint2−/−;St14+, middle), and HAI-2 and matriptase double-deficient (Spint2−/−;St14−/−, right) mice at postnatal day 5. Loss of HAI-2 lead to a prominent decrease in levels of membrane-associated EpCAM in small and large intestines, to a corresponding decrease and segmentation of claudin 7 expression in villous epithelium of small intestine, and to a complete loss of membrane association in large intestines (E, arrows). Simultaneous inactivation of matriptase restored the expression level and cell surface localization of both proteins. *P<0.05, **P<0.01, ***P<0.001. Scale bars: 20 µm.
Similarly, immunohistological analysis of intestinal tissues confirmed decreased levels of both EpCAM and claudin 7 in HAI-2 single-deficient mice, but not in mice simultaneously lacking both HAI-2 and matriptase (Fig. 4D,E). Interestingly, loss of HAI-2 did not appear to interfere with normal cell surface-associated localization of remaining EpCAM protein, although it did lead to a predominantly intracellular localization of claudin 7 signal in large intestine (Fig. 4D,E). These data suggest that the abnormal expression and localization of the epithelial junctional proteins previously described in HAI-2-deficient mice after birth can be attributed to matriptase activity.
Inactivation of matriptase is not sufficient to rescue late developmental defects in Spint2-deficient mice
Despite an apparent complete or near-complete rescue of intestinal defects in Spint2-deficient mice lacking intestinal matriptase at macroscopic, histological and molecular levels in the first week of postnatal life, the Spint2−/−;Vil-Cre+/0;St14fl/fl mice still suffer from apparent growth retardation that becomes particularly noticeable in the third week of life. These mice uniformly die before they reach 4 weeks of age (Figs 1E and 5A). Diarrhea and occasional rectal bleeding were also observed in mice older than 2 weeks, indicating persisting intestinal malfunction. Indeed, histological examination of small intestines from 20-day-old Spint2−/−;Vil-Cre+/0;St14fl/fl mice revealed widespread severe atrophy of both the intestinal mucosa and the underlying stromal tissue (Fig. 5C). Even more strikingly, colons of these mice presented with a complete loss of crypt structure and a detectable epithelial layer, associated with massive inflammatory reaction evidenced by highly increased infiltration of CD45-positive cells (Fig. 5C,D), suggesting intestinal failure as the likely cause of death in these animals.
HAI-2 is crucial for late intestinal development and survival independent of matriptase. (A) Outward appearance of control (left) and Spint2−/−;St14−/− (right) mice (all Prss8R44Q/R44Q) 20 days after birth. (B) Body weight of control (Spint2+;Prss8+, blue, n=31), HAI-2 and matriptase double-deficient (Spint2−/−;St14−/−, green, n=10), and HAI-2-expressing matriptase-deficient (Spint2+;St14−/−, red, n=31) mice after birth. HAI-2-deficient mice exhibit strong growth retardation in the third week of life. (C-F) Representative images of H&E (C,E) and anti-CD45 (D,F) staining of small (upper panels) and large (lower panels) intestines from control (C,D, left panels), HAI-2 and matriptase double-deficient (Spint2−/−;St14−/−, C,D, right panels), and matriptase single-deficient (Spint2+;St14−/−, E,F) mice 20 days after birth. Normal epithelial architecture in both small and large intestines is severely affected and is associated with highly increased infiltration of CD45-postitive immune cells (D,F, arrowheads) in HAI-2-deficient, but not in HAI-2-expressing, mice that lack intestinal matriptase. Scale bars: 50 µm.
HAI-2 is crucial for late intestinal development and survival independent of matriptase. (A) Outward appearance of control (left) and Spint2−/−;St14−/− (right) mice (all Prss8R44Q/R44Q) 20 days after birth. (B) Body weight of control (Spint2+;Prss8+, blue, n=31), HAI-2 and matriptase double-deficient (Spint2−/−;St14−/−, green, n=10), and HAI-2-expressing matriptase-deficient (Spint2+;St14−/−, red, n=31) mice after birth. HAI-2-deficient mice exhibit strong growth retardation in the third week of life. (C-F) Representative images of H&E (C,E) and anti-CD45 (D,F) staining of small (upper panels) and large (lower panels) intestines from control (C,D, left panels), HAI-2 and matriptase double-deficient (Spint2−/−;St14−/−, C,D, right panels), and matriptase single-deficient (Spint2+;St14−/−, E,F) mice 20 days after birth. Normal epithelial architecture in both small and large intestines is severely affected and is associated with highly increased infiltration of CD45-postitive immune cells (D,F, arrowheads) in HAI-2-deficient, but not in HAI-2-expressing, mice that lack intestinal matriptase. Scale bars: 50 µm.
Mice lacking intestinal matriptase have previously been reported to suffer from compromised intestinal epithelial barrier function and loss of epithelial integrity within the colon, leading to severe colitis and diminished lifespan (Kosa et al., 2012). Therefore, it was possible that the intestinal defects and loss of viability in late development observed in Spint2−/−;Vil-Cre+/0;St14fl/fl mice could result from the loss of matriptase rather than HAI-2. To distinguish between the two possibilities, we performed littermate-controlled macroscopic and histological analysis of intestinal tissues from HAI-2-expressing and HAI-2-deficient mice lacking intestinal matriptase (Spint2+/+;Vil-Cre+/0;St14fl/fl and Spint2−/−;Vil-Cre+/0;St14fl/fl, respectively). As late as postnatal day 20, tissues from HAI-2-expressing Vil-Cre+/0;St14fl/fl mice did not reveal any obvious developmental abnormality in the villous epithelium within the small intestine (Fig. 5E, compare to C). As reported previously by Kosa et al., the large intestine was more visibly affected by loss of matriptase, with less organized crypt and surface epithelia, loss of mucin-producing goblet cells and widespread inflammatory infiltrates (Fig. 5E,F, compare C with D) (Kosa et al., 2012). However, unlike HAI-2 and matriptase double-deficient mice, Spint2+/+;Vil-Cre+/0;St14fl/fl mice still maintained general colonic tissue architecture with well-formed crypts and continuous surface epithelium, and had a much less extensive inflammatory reaction. Furthermore, although their lifespan was generally limited to 3-4 months, mice lacking intestinal matriptase did not display any loss of viability and were indistinguishable in their size, weight and overall appearance from their wild-type littermate controls within the first 3 weeks of life (Fig. 5B). These data indicate that, in addition to its crucial role in regulating intestinal matriptase activity during the first week of life, HAI-2 continues to play an essential role in later stages of intestinal development that is, at least in part, independent of matriptase.
DISCUSSION
The discovery of mutations in SPINT2 in individuals with the syndromic form of CTE, also known as congenital sodium diarrhea (CSD), as well as several recent mouse model studies, firmly established the role of the transmembrane serine protease inhibitor HAI-2 as an essential regulator of intestinal development in humans and in mice (Heinz-Erian et al., 2009; Kawaguchi et al., 2019; Salomon et al., 2014; Szabo and Bugge, 2018). SPINT2 mutations so far identified in individuals with CTE/CSD all result either in loss of HAI-2 expression or reduction of its protease inhibitory activity (Faller et al., 2014; Heinz-Erian et al., 2009). This implies that a trypsin-like serine protease activity is likely to be involved in etiology of intestinal malfunctions observed in individuals with CTE/CSD and in Spint2-deficient mice. Indeed, proteolytic processing of epithelial junctional protein EpCAM, loss-of-function mutations of which are associated with the non-syndromic from of CTE, has been recently proposed as likely mechanism for compromised epithelial barrier and disease progression in affected individuals (Wu et al., 2017). Here, we show that loss of HAI-2 leads to an increased activation of matriptase zymogen in mouse intestines, and that tissue-specific inactivation of matriptase in Spint2-deficient mice restores normal macroscopic and histological appearance of intestinal mucosa, as well as the expression of CTE-associated markers EpCAM and claudin 7 in early postnatal development. Spint2-deficient mice lacking intestinal matriptase also reacquire the ability to gain weight after birth, resulting in a substantial increase in postnatal survival. These data strongly indicate that matriptase activity plays a crucial role in the development of the CTE-like phenotypes observed in Spint2-deficient mice and are consistent with two recent studies that report efficient cleavage of EpCAM protein in cell or tissue organoid cultures, leading to destabilization of claudin 7 and increased intestinal permeability (Kawaguchi et al., 2019; Wu et al., 2017). However, this is contrasted by a recent biochemical analysis by Holt-Danborg and colleagues, showing that CTE/CSD-associated HAI-2 variants inhibit matriptase with efficiency comparable with the wild-type HAI-2, while diminishing the ability of the inhibitor to regulate another reported physiological target of HAI-2, the GPI-anchored serine protease prostasin/PRSS8 (Holt-Danborg et al., 2019). This observation appears to support a role for prostasin, or for a prostasin-like protease, rather than matriptase, in the etiology of CTE/CSD. The apparent contradiction may in part lie in the unusual, and not yet fully understood, functional relationship between prostasin and matriptase. Several studies show that prostasin acts as a crucial activator of matriptase biological activity (Buzza et al., 2013; Camerer et al., 2010; Friis et al., 2013; Szabo et al., 2012). Loss of prostasin inhibition by HAI-2, as proposed by Holt-Danborg and colleagues, would therefore be expected to lead to a stimulation of matriptase activity that, consistent with the data presented here, could lead to increased EpCAM degradation and subsequent defects in intestinal development. Indeed, it has been reported that prostasin, unlike matriptase, does not have the ability to directly cleave EpCAM protein in cell culture-based system (Wu et al., 2017). However, two recent observations seem to raise doubts of even an indirect role of prostasin in the etiology of intestinal defects. First, mice expressing only the zymogen form of prostasin (Prss8R44Q/R44Q, used as littermate controls also in this study), which has been shown to be unable to form stable inhibitory complexes with HAI-2 and therefore is expected to be refractory to HAI-2 inhibition, do not exhibit any of the phenotypes observed in our Spint2−/− mice (Szabo and Bugge, 2018). Even more importantly, genetic inactivation of prostasin did not rescue intestinal defects in Spint2-deficient mice generated by Kawaguchi and colleagues (Kawaguchi et al., 2019). This is seemingly at odds with an increased activation of intestinal matriptase observed in HAI-2-deficient mice and may suggest that another, yet unknown, enzyme can substitute for prostasin by facilitating proteolytic conversion of matriptase zymogen to two-chain matriptase in this tissue. Indeed, activated matriptase has been observed in skin from prostasin knockout mice (Friis et al., 2016; Szabo et al., 2012), indicating that activation of matriptase zymogen can occur in a prostasin-independent manner in vivo. However, unlike Prss8 conditional knockouts employed by Kawaguchi et al., Prss8R44Q/R44Q mice used in this study still express a zymogen-locked variant of prostasin that has been shown to promote the activation of matriptase zymogen in a cell culture model (Friis et al., 2013). Therefore, we cannot formally exclude the possibility that zymogen-locked prostasin contributes to the observed increase in conversion of matriptase zymogen into its two-chain form. Furthermore, in the study by Kawaguchi et al., lack of prostasin did rescue cellular and molecular phenotypes, including increased EpCAM cleavage, associated with HAI-2 deficiency in intestinal organoid cultures, clearly underscoring the importance of interpreting the data within the limits of any given experimental system. Thus, despite the data presented here and elsewhere that do not support a role for prostasin in loss of HAI-2-mediated intestinal malfunction in mice, we cannot exclude the possibility that prostasin or other, as yet unidentified, proteases as being relevant targets for HAI-2 inhibition in the context of human tissues and CTE etiology.
Similarly, although a role for EpCAM cleavage and subsequent internalization of claudin 7 in the etiology of syndromic CTE appear to be well accepted, it may require further inspection. If loss of EpCAM function, owing to its excessive cleavage, is to lead to intestinal malfunction in humans and mice that lack functional HAI-2, one would expect that HAI-2-deficiency would phenocopy defects observed in EpCAM-deficient tissues. That does not appear to be the case. Whereas individuals with non-syndromic CTE that carry EPCAM loss-of-function mutations typically display an isolated intestinal disease, individuals with mutations in SPINT2 present with symptoms that affect a number of tissues other than those of the digestive tract, including skin, oral and nasal cavities, and bone (Salomon et al., 2014). Similarly, despite common intestinal malfunction and nearly identical lifespan, Spint2-deficient mice and intestinal-specific Cldn 7 knockouts suffer from severe defects in colon development and function that are not evident in EpCAM-deficient mice (Guerra et al., 2012; Lei et al., 2012; Mueller et al., 2014; Xu et al., 2019) (see Table 1 for phenotypic comparison of currently available mouse models of CTE). This suggests that, at the very least, the contribution of the loss of HAI-2-mediated regulation of matriptase activity to intestinal development and destabilization of claudin-based tight junctions may not be limited to increased cleavage and inactivation of EpCAM protein. However, it is also conceivable that this might be an indication of distinct molecular mechanisms leading to apparently similar histological and functional aberrations in mice and possibly humans lacking HAI-2 and EpCAM proteins. It is worth noting that unlike previously published studies using cell-culture and organoid models, our analysis of Spint2-deficient mouse tissues failed to reveal any obvious increase in cleaved EpCAM, despite an apparent increase in the amounts of active matriptase and a dramatic decrease in levels of EpCAM protein (Szabo and Bugge, 2018; this study), although this can also be explained by a rapid turnover of cleaved form in vivo. As the cleavage site in EpCAM molecules has recently been identified (Wu et al., 2017), use of mice expressing uncleavable variants of EpCAM may be well suited to conclusively address the issue of the role of EpCAM cleavage in the etiology of intestinal defects observed in humans and mice lacking functional HAI-2.
Intestinal phenotypes observed in mice with genetic modification in the CTE-associated genes Epcam, Cldn7 and Spint2

Finally, it should be noted that, whereas inactivation of matriptase rescued intestinal defects associated with early-onset intestinal failure and postnatal lethality in HAI-2-deficient mice (Szabo and Bugge, 2018), the resulting increase in lifespan unmasked subsequent defects in later stages of intestinal development. That these defects manifest in Spint2−/− mice lacking intestinal matriptase indicates that functions of HAI-2 may not be limited to regulation of matriptase activity. Further studies to identify these additional inhibitory targets will be required, but may prove challenging, considering the genetic complexity of existing HAI-2-deficient mouse models, apparent discrepancies between in vivo and in vitro models, and the considerable promiscuity of HAI-2 towards membrane-bound and soluble trypsin-like serine proteases.
In conclusion, we have identified membrane-anchored epithelial serine protease matriptase as the culprit behind the early intestinal malfunction caused by the loss of HAI-2. This may open new avenues for targeting of intestinal symptoms in individuals with syndromic CTE by controlling excessive matriptase activity. However, as loss of matriptase is detrimental for proper development and functionality of multiple tissues in both mice and in humans, a complete inhibition of matriptase function might not be desirable. One plausible approach is the use of reagents that only limit rather than completely eliminate protease activity, such as blocking antibodies that selectively bind and inhibit fully active, double-chain form of matriptase, thus leaving less active zymogen to provide sufficient function to sustain normal development and tissue homeostasis (Farady et al., 2008; Sun et al., 2003).
MATERIALS AND METHODS
Mouse strains
All experiments involving mice were performed in an Association for Assessment and Accreditation of Laboratory Animal Care International-accredited vivarium following Institutional Guidelines and Standard Operating Procedures as approved by the NIDCR Institutional Animal Care and Use Committee. Mice carrying HAI-2 knockout (Spint2−/−), matriptase conditional knockout (St14fl/fl), zymogen-locked prostasin knock-in (Prss8R44Q/R44Q) and intestinal-specific expression of Cre recombinase (Vil-Cre+/0) alleles have all been described in detail previously (Friis et al., 2016; List et al., 2009; Szabo et al., 2009). All studies used mice of mixed 129S6/Sv;NIH BlackSwiss;FVB/NJ;C57Bl/6J genetic background and were littermate controlled. Mice were genotyped by PCR using ear or tail clips of newborn to 2-week-old mice as described elsewhere (Friis et al., 2016; Kosa et al., 2012; Szabo et al., 2009).
Protein extraction and western blot analysis of mouse intestinal tissues
Small and large intestines were collected at postnatal day 5 or 20, snap-frozen in liquid nitrogen and stored at −80°C until further use. For western blot analysis, the tissues were homogenized in 2% SDS and 10% glycerol in 62.5 mM Tris/Cl (pH 6.8) containing Protease Inhibitor Cocktail (Sigma-Aldrich). The lysates were cleared by centrifugation at 16,000 g for 10 min at 4°C to remove the tissue debris and the protein concentration in supernatant was determined by BCA assay (Pierce). For western blot detection, 60 µg of total protein was mixed with 4×SDS sample buffer (NuPAGE, Invitrogen, Carlsbad, CA) containing 7% β-mercaptoethanol (Sigma-Aldrich), boiled for 5 min at 99°C, and run on 4-12% BisTris NuPAGE gels using 1×MOPS running buffer (both Invitrogen). Separated proteins were transferred to PVDF membranes (0.2 μm, Invitrogen) and blocked with 5% nonfat dry milk in Tris-buffered saline (TBS) and 0.05% Tween 20 (TBS-T). Membranes were incubated with primary antibody overnight at 4°C, followed by incubation with secondary antibody conjugated to alkaline phosphatase for 1.5 h at room temperature (see Table S1 for further information on antibodies). Alkaline phosphatase activity was visualized using nitro-blue tetrazolium and 5-bromo-4-chloro-3′-indolylphosphate substrates (Sigma-Aldrich). Data shown are representative of at least two independent western blot experiments run on tissue lysates from three separate mice per each indicated genotype. Band intensities were quantified using ImageJ.
HAI-1 immunoprecipitation
Intestines of 5-day-old mice were homogenized in ice-cold 50 mM Tris/HCl (pH 8.0), 1% NP-40 and 500 mM NaCl buffer and incubated on ice for 10 min. The lysates were centrifuged at 20,000 g for 10 min at 4°C to remove the tissue debris and the supernatant was pre-incubated with 100 µl GammaBind G Sepharose beads (GE Healthcare Bio-Sciences) for 30 min at 4°C with gentle agitation. The samples were spun at 1000 g for 5 min to remove the beads, and the supernatant was then incubated with 3 µg goat anti-mouse HAI-1 antibody (R&D Systems) and 50 µl of GammaBind G Sepharose beads for 16 h at 4oC. The samples were spun at 1000 g for 1 min, the supernatant was removed and the beads were washed three times with 1 ml ice-cold 50 mM Tris/HCl (pH 8.0), 1% NP-40 and 500 mM NaCl buffer. The beads were then mixed with 30 µl of 1×SDS loading buffer (Invitrogen) with 0.25 M β-mercaptoethanol, incubated for 5 min at 99°C and cooled on ice for 2 min. The samples were spun at 1000 g for 1 min and the released proteins were resolved by SDS-PAGE (4-12% polyacrylamide gel) and analyzed by western blot as described above.
RNA preparation and quantitative RT-PCR
Small and large intestines collected from postnatal day (P)5 mice were homogenized in Trizol reagent (Life Technologies) and total RNA was extracted according to the manufacturer's instructions. RNA (2 µg) was reverse transcribed with oligo dT primer using the RevertAid H Minus First Strand cDNA Synthesis Kit (Thermo Fisher Scientific Baltics). Real-time PCR was conducted on 1 µl of cDNA template using iQ SYBR Green Supermix (Bio-Rad Laboratories) and 7500 Real-Time PCR System with 7500 Software v2.3 (Applied Biosystems). Primer sets used for cDNA amplification are listed in Table S2. The expression of each gene was normalized to expression of GAPDH using the ΔΔCt method. The assay was performed in triplicate. Results shown are mean and standard deviations from three independent mice per genotype.
Histological analysis
Pups were euthanized on postnatal days 5 or 20, and intestinal tissues were extracted and immediately fixed in aqueous-buffered zinc formalin fixative (Z-Fix, Anatech) for 24 h at room temperature prior to paraffin embedding and sectioning (Histoserv). Sections (5 µm) were stained using Hematoxylin and Eosin (H&E) and Alcian Blue/PAS (both performed by Histoserv) or used for immunohistochemistry as described below. All histological quantifications were performed by a manual count of cells of interest in five non-overlapping fields per tissue sample and at least five independent biological samples per genotype.
Immunohistochemistry
Sections (5 µm) from formalin-fixed, paraffin-embedded mouse tissues were immunostained after antigen retrieval by incubation for 20 min at 100°C in 0.01 M citric acid for detection of CD11b and CD45, and in 0.01 M sodium citrate buffer (pH 6.0) for detection of EpCAM and claudin 7, essentially as described previously (Szabo et al., 2016). Briefly, the sections were blocked with 2.5% bovine serum albumin (Fraction V, MP Biomedicals) in PBS and incubated overnight at 4°C with goat anti-hEpCAM (R&D Systems), rabbit anti-hClaudin 7 (Life Technologies), rat anti-CD45 (Abcam) or rabbit anti-CD11b (Novus Biologicals) primary antibodies (see Table S1 for further information on antibodies). Bound antibodies were visualized using biotin-conjugated anti-rat, -rabbit or -goat secondary antibodies (all Vector Laboratories) and a Vectastain ABC Kit (Vector Laboratories), using 3,3′-diaminobenzidine as the substrate (Sigma-Aldrich). All microscopic images were acquired on Olympus BX43 microscope using Olympus DP74 digital camera with cellSens Entry system.
Statistical analysis
All statistical analysis was carried out using GraphPad Prism software (ver 8.0.2). Differences in protein expression relative to GAPDH expression observed on western blots were statistically evaluated using a two-sample Student's t-test, two-tailed in experiments with only two experimental groups, or by using one-way ANOVA if more than two groups were compared. Band intensities from at least two independent western blot experiments run on tissue lysates from three separate mice per each indicated genotype, corresponding to at least six independent biological samples, were used for the analysis.
To evaluate the effect of intestinal matriptase deficiency on the embryonic survival of prostasin zymogen-locked mice, Chi-square analysis was performed on the observed versus the expected distribution of intestinal matriptase-deficient (Vil-Cre+/0;St14fl/fl;Prss8R44Q/R44Q) and their matriptase-expressing littermates from Vil-Cre+/0;St14fl/+;Prss8R44Q/R44Q×St14fl/fl;Prss8R44Q/R44Q breeding pairs on postnatal day 1. Postnatal survival of the Spint2−/−;Prss8R44Q/R44Q mice expressing two (Vil-Cre0;St14fl/fl or Vil-Cre0;St14fl/+), one (Vil-Cre+/0;St14fl/+) or no (Vil-Cre+/0;St14fl/fl) functional alleles of intestinal matriptase (N≥13 for each genotype) was analyzed using the Mantel-Cox log-rank test. Differences in median survival were analyzed for significance using one-way ANOVA. Postnatal growth in the first week of life was analyzed by comparing linear regression slopes of the average total body weights of mice of different genotypes measured daily.
Numbers of PAS-, CD11b- and CD45-positive cells on histological stains were determined by manually counting positive cells in five non-overlapping fields of either small or large intestines in five samples per genotype. The observed values, relative to the longitudinal length of the counted area, were statistically evaluated using a one-way ANOVA.
The absolute lengths of small and large intestines were measured immediately after the extraction of the tissues from the mice after euthanasia and were statistically evaluated using a one-way ANOVA.
Acknowledgements
We thank Dr Mary Jo Danton for critically reviewing this manuscript.
Footnotes
Author contributions
Conceptualization: R.S., T.H.B.; Methodology: R.S.; Validation: R.S.; Formal analysis: R.S., L.K.C.; Investigation: R.S., L.K.C.; Writing - original draft: R.S.; Writing - review & editing: R.S., L.K.C., T.H.B.; Visualization: R.S.; Supervision: R.S., T.H.B.; Funding acquisition: T.H.B.
Funding
Funding for this study was provided by the Intramural Research Program at the National Institute of Dental and Craniofacial Research (T.H.B.), and by the National Institute of Dental and Craniofacial Research Gene Transfer Core (ZIC DE000744-04) and Veterinary Resources Core (ZIC DE000740-05). Deposited in PMC for release after 12 months.
Peer review history
The peer review history is available online at https://dev.biologists.org/lookup/doi/10.1242/dev.183392.reviewer-comments.pdf
References
Competing interests
The authors declare no competing or financial interests.