Unveiling the Roles of Low-Density Lipoprotein Receptor-Related Protein 6 in Intestinal Homeostasis, Regeneration and Oncogenesis

Intestinal epithelial self-renewal is tightly regulated by signaling pathways controlling stem cell proliferation, determination and differentiation. In particular, Wnt/β-catenin signaling controls intestinal crypt cell division, survival and maintenance of the stem cell niche. Most colorectal cancers are initiated by mutations activating the Wnt/β-catenin pathway. Wnt signals are transduced through Frizzled receptors and LRP5/LRP6 coreceptors to downregulate GSK3β activity, resulting in increased nuclear β-catenin. Herein, we explored if LRP6 expression is required for maintenance of intestinal homeostasis, regeneration and oncogenesis. Mice with an intestinal epithelial cell-specific deletion of Lrp6 (Lrp6IEC-KO) were generated and their phenotype analyzed. No difference in intestinal architecture nor in proliferative and stem cell numbers was found in Lrp6IEC-KO mice in comparison to controls. Nevertheless, using ex vivo intestinal organoid cultures, we found that LRP6 expression was critical for crypt cell proliferation and stem cell maintenance. When exposed to dextran sodium sulfate, Lrp6IEC-KO mice developed more severe colitis than control mice. However, loss of LRP6 did not affect tumorigenesis in ApcMin/+ mice nor growth of human colorectal cancer cells. By contrast, Lrp6 silencing diminished anchorage-independent growth of BRafV600E-transformed intestinal epithelial cells (IEC). Thus, LRP6 controls intestinal stem cell functionality and is necessary for BRAF-induced IEC oncogenesis.


Introduction
Intestinal stem cells, located at the base of intestinal crypts, play a central role in the establishment and maintenance of the intestinal epithelium. These cells undergo asymmetric division leading to migration and differentiation along the intestinal crypt axis [1]. Thus, the intestinal epithelium self-renewal is tightly regulated by interacting intracellular signaling pathways controlling stem and progenitor cell expansion and differentiation. In particular, the Wnt/β-catenin pathway controls proliferation and survival and is required for the maintenance of intestinal stem cells. In the absence of Wnt ligand, β-catenin is destabilized by a destruction complex composed of casein kinase 1 alpha (CK1α), GSK3β, Axin and adenomatous polyposis coli (APC) proteins. β-catenin binds to the destruction complex and is phosphorylated by CK1α and then by GSK3β, which marks β-catenin for ubiquitin-mediated degradation by the proteasome [2]. Wnt, its receptor and co-receptors LRP5/LRP6 form a trimeric complex which recruits the scaffold protein Dishevelled (DVL) and Axin, inducing LRP5/LRP6 aggregation. These aggregates promote phosphorylation of LRP5/LRP6 by CK1γ and GSK3β, releasing β-catenin. β-catenin then accumulates into the cytosol and translocates into the nucleus to interact with nuclear DNA-binding T-cell factor (TCF) and lymphoid enhancer-binding protein (LEF), forming a functional transcription factor promoting target gene expression. and phosphorylated ERK1/2 (M8159) from Millipore Sigma (Oakville, ON, Canada), ERK2 (sc-154) from Santa Cruz Biotechnology (Santa Cruz, CA, USA), non-phospho (active) βcatenin (D13A1), LRP5 (D80F2), LRP6 (C5C7), LRP6 (C47E12) and phospho-LRP6 (Ser1490) from Cell Signaling Technology (Danvers, MA, USA) and β-catenin (C14) and BrdU (B44) from BD Biosciences (San Jose, CA, USA). Horseradish peroxidase antibodies were obtained from GE Healthcare Life Sciences (Mississauga, ON, Canada) and alkaline phosphataseconjugated antibodies from Promega Corporation (Madison, WI, USA). The specific ERK inhibitor SCH772984 was obtained from Cayman Chemical Company (Ann Arbor, MI, USA). Recombinant mouse Wnt3a was obtained from Abcam (Cambridge, MA, USA). Other materials were purchased from Millipore Sigma unless stated otherwise.

Colitis Induction with DSS and Clinical Evaluation
Ten to twelve-week-old Lrp6 IEC-KO mice and control littermates were administered dextran sodium sulfate 1.75-2% (DSS, colitis grade; MP Biomedicals, Solon, OH, USA) in drinking water for 7 days. Clinical parameters (weight loss, rectal bleeding and diarrhea) were monitored every day. The disease activity index (DAI) was measured at euthanasia at day 7 [13]. Histological damage scoring was assessed on hematoxylin and eosin-stained longitudinal sections from the entire colon length based on the destruction of normal mucosal architecture, presence and degree of cellular infiltration, extent of muscle thickening, presence or absence of crypt abscesses and the presence or absence of goblet cell depletion, as described in the supplementary Table S1. Clinical scorings were performed in a blinded manner.

Macroadenoma Count, Histological Staining, Immunofluorescence and Immunohistochemistry
Polyps were stained with methylene blue and visualized under an SZ51 stereomicroscope (Olympus, Tokyo, Japan). Polyp sizes were measured with a digital caliper (Thermo Fisher Scientific, Waltham, MA, USA) and polyp numbers were counted from the duodenum to the rectum. For BrdU staining, mice were injected with BrdU (10 µL/g of body weight; Thermo Fisher Scientific) 90 min before euthanasia. Tissues were fixed with paraformaldehyde 4%, paraffin-embedded, sectioned and stained [14]. Immunohistochemistry staining was performed with the EnVision+System Kit (Dako, Burlington, ON, Canada). Slides were visualized with a NanoZoomer slide scanner and NDP.view2 software (Hamamatsu, Boston, MA, USA). Cell counts were performed on well-oriented crypts in a blinded manner.

Enteroids
Crypts were isolated from the jejunum of 8-to 12-week-old mice [15]. Jejunum was opened longitudinally and rinsed with cold PBS. The intestine was then cut into pieces of approximately 5 mm and fragments washed with PBS multiple times by stirring in a 50 mL conical tube. Intestinal fragments were transferred into a new 50 mL conical tube containing 20 mL of 30 mM EDTA in PBS and incubated for 5 min on ice. EDTA was replaced followed by a 20 min incubation on ice with gentle stirring. EDTA was then replaced with 40 mL of PBS and the tube was shaken vigorously until dissociated crypts were visualized in the solution under a microscope. Residual villi were removed by filtering the solution through a 70 µm cell strainer with crypts going through as opposed to residual villi being retained by the filter. Crypts were then centrifuged at 150× g for 5 min, crypt pellet was washed twice with 25 mL of Advanced DMEM/F-12 medium, resuspended in phenol red free Matrigel basement membrane (Corning, Glendale, AZ, USA) and plated in a 48-well plate (Corning Costar) (20 µL/well). Enteroids were cultured in Advanced DMEM/F-12 culture medium (Gibco, Waltham, MA, USA) supplemented with 1 mM N-acetylcysteine, 50 ng/mL murine EGF (Life Technologies, Waltham, MA, USA), B27 supplement 1× and N2 supplement 1× (Life Technologies), 10% R-spondin and 10% Noggin-conditioned media. Noggin-Fc and R-spondin 1-Fc were produced in HEK293T cell lines stably expressing R-spondin 1-Fc (C. Kuo, Stanford University, Stanford, CA, USA) or Noggin-Fc (G. R. van den Brink, Hubrecht Institute, Utrecht, The Netherlands). Phase contrast images were taken using a Celldiscoverer 7 live Imaging Station equipped with ZEN software (Carl Zeiss, Toronto, ON, Canada). Enteroid proliferation was evaluated with the Click-it 5-Ethynyl-2 -deoxyuridine (EdU) Alexa Fluor 555 imaging kit (Thermo Fisher Scientific) and visualized with an LSM Olympus FV1000 confocal microscope equipped with FV10-ASW 3.1 software (Olympus; Coopersburg, PA, USA).

RNA Extraction and qRT-PCR
RNA was isolated from scraped jejunum mucosa and enteroids with the RNeasy minikit (Qiagen, Toronto, ON, Canada). RT-PCR was performed using the Transcriptor reverse transcriptase with random hexamer primers (Roche Diagnostics, Laval, QC, Canada). Quantitative polymerase chain reaction (qPCR) was performed by the RNomics Platform at Université de Sherbrooke. Target expression was quantified relative to Pum1, Psmc4 and Tbp expression. PCR conditions and primer sequences are available upon request.

Clonogenic Assay
Inducible shRNA cell lines were treated with 50 ng/mL doxycycline 48 h before seeding. Approximately 1000 cells were seeded into six-well plates and cultured in media containing 50 ng/mL doxycycline for 14 days. Colonies were then stained with a PBS solution containing 0.05% crystal violet, 1% formaldehyde and 1% methanol. Phase contrast images were taken with a Zeiss Cell Discoverer 7 live Imaging Station equipped with ZEN software. Colonies were counted using ImageJ software.

Statistical Analyses
Assays were carried out at least in triplicate. Results are representative of three independent experiments if not stated otherwise. Data were analyzed with Student's t-test for populations following a normal curve or Mann-Whitney U test in other cases. Results were considered statistically significant at p ≤ 0.05. Graphs and statistics were generated using GraphPad Prism software (Prism, San Diego, CA, USA).

Lrp6 Deletion Does Not Alter Intestinal Architecture and Homeostasis but Impairs Wnt Target Gene Expression
To address the importance of LRP6 in intestinal development and architecture, we bred Lrp6 loxP/loxP mice with Villin-Cre mice that express the Cre recombinase in intestinal epithelial cells [19] leading to specific loss of LRP6 expression in these cells ( Figure 1A). Because LRP6 is a co-receptor involved in the activation of the Wnt/β-catenin signaling pathway, we verified the activation status of this pathway by analyzing the levels of total and activated β-catenin protein. As shown in Figure 1A, total and active β-catenin levels were not significantly different in mutant and control mucosae. Although we did not observe a difference in β-catenin expression levels, we cannot exclude that subtle changes might occur in its transcriptional activity. We therefore analyzed the expression levels of Wnt/β-catenin target genes. Intriguingly, while Cd44, Sox9 and Ascl2 gene expression remained unaffected by loss of epithelial LRP6, Axin2, Lgr5 and EphB3 gene expression was significantly decreased in Lrp6 IEC-KO mucosal enrichments ( Figure 1B). Histological analyses of Lrp6 IEC-KO mouse jejunal sections revealed, however, no difference in intestinal architecture, villus length, crypt depth and proliferation in comparison to control littermates ( Figure 1C,D). No histological change was observed in the colonic epithelium as well ( Figure S1A,B). Expression levels of genes encoding mucin 2, lysozyme and sucrase-isomaltase, the main markers of goblet, Paneth and absorptive differentiated cells, respectively, were also comparable between control and mutant mice ( Figure 1E). The decreased expression of Lgr5 in Lrp6 KO mucosal enrichments suggests a change in the number of active cycling stem cells, also named crypt base columnar (CBC) cells. We therefore performed immunohistochemistry against OLFM4, a robust marker for these cells. As shown in Figure 1F, similar numbers of OLFM4-positive cells were found in Lrp6 IEC-KO and control mice. Thus, while loss of LRP6 in intestinal epithelial cells reduced the expression of some Wnt target genes, including the stem cell marker Lgr5, it was not sufficient to perturb homeostatic maintenance of the intestinal epithelium under normal conditions.

LRP6 Expression Is Required for Ex Vivo Intestinal Crypt Regeneration
Since the expression of some Wnt target genes is reduced upon Lrp6 deletion, we next examined the regenerative abilities of intestinal stem cells by preparing enteroids from control and Lrp6 IEC-KO mice. Notably, development of Lrp6 IEC-KO enteroids was compromised in comparison to control enteroids. Indeed, the number of protrusions was significantly reduced in enteroids lacking Lrp6 expression ( Figure 2A). Three days after seeding, proliferation of Lrp6-deficient enteroids was significantly decreased in comparison to control enteroids, as visualized by a reduced number of EdU-positive cells ( Figure 2B). After 5 days, Lrp6-deficient enteroids degenerated and died ( Figure 2C) and increased expression of the pro-apoptotic Noxa gene was noticed ( Figure 2D). To examine the effect of Lrp6 deletion on Wnt signaling in enteroids, we determined the mRNA expression level of Wnt target stem cell genes by quantitative RT-PCR. There was a significant reduction in the expression of Axin2, Lgr5, Ascl2 and Cd44 in Lrp6 IEC-KO enteroids, 3 days after seeding ( Figure 2E). Since LRP5 expression was also detected in enteroids ( Figure 2F), we tested the ability of exogenous Wnt3a to rescue the growth defect observed in Lrp6 IEC-KO enteroids.
As shown in Figure S2A,B, Lrp6-deficient enteroids still degenerated and died despite the addition of Wnt3a to culture medium. These results suggest that LRP6 is much more important than LRP5 for the control of intestinal stem cell renewal and crypt regeneration, at least ex vivo.

LRP6 Expression Is Required for Ex Vivo Intestinal Crypt Regeneration
Since the expression of some Wnt target genes is reduced upon Lrp6 deletion, we nex examined the regenerative abilities of intestinal stem cells by preparing enteroids from control and Lrp6 IEC-KO mice. Notably, development of Lrp6 IEC-KO enteroids was compro mised in comparison to control enteroids. Indeed, the number of protrusions was signifi cantly reduced in enteroids lacking Lrp6 expression (Figure 2A). Three days after seeding proliferation of Lrp6-deficient enteroids was significantly decreased in comparison to con trol enteroids, as visualized by a reduced number of EdU-positive cells ( Figure 2B). After 5 days, Lrp6-deficient enteroids degenerated and died ( Figure 2C) and increased expres sion of the pro-apoptotic Noxa gene was noticed ( Figure 2D). To examine the effect of Lrp6 expression of Axin2, Lgr5, Ascl2 and Cd44 in Lrp6 IEC-KO enteroids, 3 days after seeding (Fi ure 2E). Since LRP5 expression was also detected in enteroids ( Figure 2F), we tested t ability of exogenous Wnt3a to rescue the growth defect observed in Lrp6 IEC-KO enteroid As shown in Figure S2A,B, Lrp6-deficient enteroids still degenerated and died despite t addition of Wnt3a to culture medium. These results suggest that LRP6 is much more im portant than LRP5 for the control of intestinal stem cell renewal and crypt regeneratio at least ex vivo.

Lrp6 IEC-KO Mice Are More Sensitive to DSS-Induced Epithelial Damage and Colitis
We next evaluated the susceptibility of control and Lrp6 IEC-KO mice to develop colonic inflammation in response to dextran sodium sulfate (DSS) administration. As shown in Figure 3A,B, DSS administration induced acute colitis in control mice after 7 days, as shown by significant weight loss, diarrhea and occasional bleeding, resulting in a relatively high Disease Activity Index (DAI). Interestingly, Lrp6 IEC-KO mice exhibited more weight loss and significantly higher DAI associated with more severe diarrhea, rectal bleeding and bloody stools. Hematoxylin and eosin staining was also performed on colonic tissue to score damage according to the extent of mucosal architecture destruction, the presence of cellular infiltration, crypt abscesses, goblet cell depletion and the extent of muscle thickening; typical alterations observed during colitis. As shown in Figure 3C,D, Lrp6 IEC-KO mice exhibited a significantly higher histological score than control mice, indicating that Lrp6 IEC-KO intestinal mucosa was more affected following DSS treatment. experiments, each with at least 40 organoids, is summarized in the graph. Chi-square, *** p ≤ 0.001. (B) Enteroid proliferation was evaluated 3 days after seeding by EdU incorporation and immunofluorescence to evaluate the ratio of proliferative cells per protrusion. The average of 4 independent experiments, each with at least 8 protrusions, is summarized in the graph. Bars: 50 µ m. Unpaired t-test, * p ≤ 0.05. (C) Enteroid viability was determined 5 days after seeding by phase contrast microscopy. The average of 3 independent experiments, each with at least 50 organoids, is summarized in the graph. Unpaired t-test, * p ≤ 0.05. (D,E) Noxa, Ascl2, Lgr5, Axin2 and Cd44 expression was evaluated by qPCR by comparing mutant and control enteroid RNA isolated 3 days after seeding (n ≥ 3). Unpaired t-test, * p ≤ 0.05. (F) Enteroid LRP6 and LRP5 expression was analyzed by Western blot 3 days after seeding. Data are expressed as mean ± SD.

Lrp6 IEC-KO Mice Are More Sensitive to DSS-Induced Epithelial Damage and Colitis
We next evaluated the susceptibility of control and Lrp6 IEC-KO mice to develop colonic inflammation in response to dextran sodium sulfate (DSS) administration. As shown in Figure 3A,B, DSS administration induced acute colitis in control mice after 7 days, as shown by significant weight loss, diarrhea and occasional bleeding, resulting in a relatively high Disease Activity Index (DAI). Interestingly, Lrp6 IEC-KO mice exhibited more weight loss and significantly higher DAI associated with more severe diarrhea, rectal bleeding and bloody stools. Hematoxylin and eosin staining was also performed on colonic tissue to score damage according to the extent of mucosal architecture destruction, the presence of cellular infiltration, crypt abscesses, goblet cell depletion and the extent of muscle thickening; typical alterations observed during colitis. As shown in Figure 3C,D, Lrp6 IEC-KO mice exhibited a significantly higher histological score than control mice, indicating that Lrp6 IEC-KO intestinal mucosa was more affected following DSS treatment.

LRP6 Is Dispensable for Adenoma Formation in Apc Min/+ Mice and for Growth of Human Colorectal Cancer Cells
Because initiation of most colorectal cancers (CRC) is often due to activating mutations in Wnt/β-catenin signaling, we verified the role of LRP6 in Apc Min/+ mice, which are heterozygous for an Apc mutation frequently found in human CRC, and which spontaneously develop intestinal adenomas [20]. Lrp6 IEC-KO mice were crossed with Apc Min/+ mice

LRP6 Is Dispensable for Adenoma Formation in Apc Min/+ Mice and for Growth of Human Colorectal Cancer Cells
Because initiation of most colorectal cancers (CRC) is often due to activating mutations in Wnt/β-catenin signaling, we verified the role of LRP6 in Apc Min/+ mice, which are heterozygous for an Apc mutation frequently found in human CRC, and which spontaneously develop intestinal adenomas [20]. Lrp6 IEC-KO mice were crossed with Apc Min/+ mice and intestinal tumor load was analyzed after 3 months. Unexpectedly, the absence of LRP6 did not significantly affect tumor development in Apc Min/+ mice ( Figure 4A,B). We next assessed the role of LRP6 in two established human CRC cell lines exhibiting activated Wnt signaling, namely HCT116 (CTNNB1 mutation) and SW48 (APC mutation) cells, by using a doxycycline-inducible shRNA expression vector targeting LRP6 mRNA, to knockdown LRP6 protein expression. LRP6 knockdown ( Figure 4C) was not sufficient to alter clonogenic potential nor anchorage-independent growth of HCT116 and SW48 cancer cells ( Figure 4D-F).
next assessed the role of LRP6 in two established human CRC cell lines exhibiting ac vated Wnt signaling, namely HCT116 (CTNNB1 mutation) and SW48 (APC mutatio cells, by using a doxycycline-inducible shRNA expression vector targeting LRP6 mRN to knockdown LRP6 protein expression. LRP6 knockdown ( Figure 4C) was not sufficie to alter clonogenic potential nor anchorage-independent growth of HCT116 and SW cancer cells (Figure 4D-F).

ERK-Dependent Phosphorylation of LRP6 in KRAS-and BRAF-Mutated Colorectal Cancer Cells-A Role for LRP6 in BRAF-Induced Intestinal Oncogenesis
Previously, we reported that oncogenic KRAS and BRAF signaling in rodent intestinal epithelial cells (IEC-6 cells) activates the Wnt/β-catenin pathway by increasing LRP6 phosphorylation [7]. Accordingly, treatment of KRAS (HCT116) and BRAF (HT-29)-mutated human CRC cells with the ERK1/2 inhibitor, SCH7729846, reduced LRP6 phosphorylation on serine 1490, a residue required for Wnt signal transduction. Interestingly, LRP6 phosphorylation was not or only barely affected by the SCH7729846 inhibitor in KRAS and BRAF wild-type CRC cells (Caco-215 or SW48) ( Figure 5A). These data suggest that LRP6 phosphorylation may be instrumental for oncogenesis induced by abnormal ERK signaling. We therefore knocked down LRP6 expression in IEC-6 cells transformed by the 4-hydroxytamoxifen-inducible BRAF V600E :ER fusion protein ( Figure 5B). As shown in Figure 5C, LRP6 silencing clearly reduced the ability of oncogenic BRAF to induce anchorage-independent growth in IEC-6 cells.
Cells 2021, 10, x FOR PEER REVIEW 10 of 16 HCT116 cells were grown for 14 days and SW48 cells for 21 days in soft agar before MTT staining. The number of colonies (E) and their size (F) were evaluated. Three independent experiments are summarized in the graph. Data are expressed as mean ± SD. Paired t-test.

ERK-Dependent Phosphorylation of LRP6 in KRAS-and BRAF-Mutated Colorectal Cancer Cells-A Role for LRP6 in BRAF-Induced Intestinal Oncogenesis
Previously, we reported that oncogenic KRAS and BRAF signaling in rodent intestinal epithelial cells (IEC-6 cells) activates the Wnt/β-catenin pathway by increasing LRP6 phosphorylation [7]. Accordingly, treatment of KRAS (HCT116) and BRAF (HT-29)-mutated human CRC cells with the ERK1/2 inhibitor, SCH7729846, reduced LRP6 phosphorylation on serine 1490, a residue required for Wnt signal transduction. Interestingly, LRP6 phosphorylation was not or only barely affected by the SCH7729846 inhibitor in KRAS and BRAF wild-type CRC cells (Caco-215 or SW48) ( Figure 5A). These data suggest that LRP6 phosphorylation may be instrumental for oncogenesis induced by abnormal ERK signaling. We therefore knocked down LRP6 expression in IEC-6 cells transformed by the 4-hydroxytamoxifen-inducible BRAF V600E :ER fusion protein ( Figure 5B). As shown in Figure 5C, LRP6 silencing clearly reduced the ability of oncogenic BRAF to induce anchorageindependent growth in IEC-6 cells. or a control shRNA (shCtrl) and compared to non-transfected cells. (C) IEC6 BRAF V600E :ER cells stably expressing shLrp6, shCtrl or that were non-transfected were treated with 250 nM 4OH-tamoxifen 16 h before soft agar seeding. Cells were grown for 3 to 5 weeks in soft agar before MTT staining and colony counting. Three independent experiments are summarized in the graph. Data are expressed as mean ± SD. Paired t-test, ** p ≤ 0.01. :ER fusion protein, and shRNA against Lrp6 (shLrp6) or a control shRNA (shCtrl) and compared to non-transfected cells. (C) IEC6 BRAF V600E :ER cells stably expressing shLrp6, shCtrl or that were non-transfected were treated with 250 nM 4OH-tamoxifen 16 h before soft agar seeding. Cells were grown for 3 to 5 weeks in soft agar before MTT staining and colony counting. Three independent experiments are summarized in the graph. Data are expressed as mean ± SD. Paired t-test, ** p ≤ 0.01.

Discussion
Although LRP6 is involved in Wnt/β-catenin signal transduction, its role in the maintenance of intestinal homeostasis remains unclear [5]. Herein, we demonstrated that IEC-specific Lrp6 deletion in mice does not affect intestinal development nor crypt cell proliferation and differentiation. Because the maintenance of Wnt signaling in the intestinal epithelium is absolutely required to maintain homeostasis, total and active β-catenin expression was analyzed as a readout of Wnt pathway activation. Levels of total and active β-catenin proteins remained similar between control and Lrp6-deficient mucosae. Nonetheless, expression of some Wnt target genes, namely Lgr5, Axin2 and Ephb3, was decreased in Lrp6 IEC-KO mice, suggesting that intestinal crypt cells might be somehow less responsive to Wnt stimulation in the absence of the LRP6 protein. One of the reduced genes, Lgr5, was previously recognized as a marker of active cycling stem cells, located at the crypt base of the crypt-villus axis, and whose self-renewal is dependent on Wnt signaling. Hence, impaired activation of Wnt pathway activation may result in decreased LGR5-positive active stem cell numbers, as observed after β-catenin deletion in the murine intestinal epithelium [21]. However, immunohistochemistry against OLFM4 protein, a robust marker for these stem cells [1], indicates that Lrp6 IEC-KO and control crypts retain similar numbers of active stem cells. Additionally, expression of Ascl2, a Wnt-responsive master transcription factor that controls LGR5-positive intestinal stem cell gene expression programs [22], remains unaffected without LRP6 in the murine intestinal epithelium. Thus, although Lgr5 gene expression is reduced, the number of actively cycling stem cells and proliferative cells in Lrp6-deficient intestinal crypts remain comparable to what is observed in control crypts. This could be at least explained by the fact that LRP5, a LRP6 ortholog, is also expressed in murine crypt epithelial cells. Indeed, LRP5 and LRP6 can play compensatory functions as reported during murine intestinal epithelial development [5]. A modest increase in LRP5 protein expression was observed in Lrp6 IEC-KO intestines in comparison to controls ( Figure S3A). This slight LRP5 increase could be sufficient to sustain enough Wnt signaling and maintain intestinal homeostasis and renewal in mice. Alternatively, additional factors such as fibroblast growth factors (FGF2 and FGF9) [23], glucagon-like pertide-2 [24], heparin-binding epidermal growth factor-like growth factor [25] and epidermal growth factor (EGF) [26] have been also shown to stimulate β-catenin signaling in intestinal crypts and thus, could compensate for the absence of LRP6 for the maintenance of intestinal homeostasis.
Interestingly, Lrp6 IEC-KO jejunal crypt enteroids showed hindered proliferation and died a few days after seeding, probably by increased apoptosis, as suggested by the increased expression of the pro-apoptotic gene Noxa. Decreased Lrp6-deficient enteroid proliferation and viability may be due to Wnt signaling inhibition in stem and progenitor cells, as revealed by the substantially decreased expression of the Wnt target genes Lgr5, Ascl2, Axin2 and Cd44. Additionally, Olfm4 was also significantly decreased in Lrp6-deficient enteroids. Decreased expression of Olfm4, Ascl2 and Lgr5 strongly indicates that functionality of active stem cells in enteroids was impaired in the absence of LRP6. Since such alterations were not observed in vivo, one could speculate that underlying mesenchymal niche cells likely maintain stem cell homeostasis in Lrp6 IEC-KO mice. Moreover, our findings are reminiscent of those observed with Math1and Wnt3-inducible-IEC-KO mouse models. While no alteration in stem cell numbers and Wnt signaling, as assessed by measuring active β-catenin levels, was observed in vivo, enteroids derived from these KO mice could not grow [27,28]. Of note, MATH1 is an essential driver of differentiation of secretory cells, including Paneth cells, which produce growth factors such as EGF, Notch and Wnt3, which collectively promote the growth and maintenance of LGR5-positive stem cells. Therefore, the absence of MATH1, specifically in intestinal epithelial cells, leads to crypt enteroid growth arrest, resulting from the inhibition of Paneth cell differentiation and of the production of growth factors, especially Wnt3 [28]. Thus, these studies have revealed that Paneth cell-secreted Wnt3 is not the only niche factor involved in stem cell maintenance in vivo, suggesting that redundant Wnt signals exist in mice to ensure intestinal homeostasis. Indeed, intestinal stromal cells act as important regulators of the stem cell niche, by producing and secreting several Wnt factors, notably Wnt2b and Wnt4 [29][30][31][32][33]. Although Wnt2b or Wnt4 gene expression was not modulated in Lrp6 IEC-KO intestines ( Figure S3B), other mesenchymal-derived growth factors may contribute to the maintenance of intestinal homeostasis in Lrp6 IEC-KO mice, as described above. That being said, Lrp6 IEC-KO enteroids continued to degenerate and die despite the addition of Wnt3a ( Figure S2), indicating that LRP6, but not LRP5, is required for Wnt signaling transduction in intestinal crypt cells. Likewise, LRP6 has been reported to be much more potent than LRP5 in activating Wnt signaling in other cell types [34,35].
A variant of the human LRP6 gene was recently found to be frequently expressed in early onset ileal Crohn's Disease, a chronic intestinal inflammatory disorder [36]. We therefore analyzed the sensitivity of Lrp6 IEC-KO mice to intestinal inflammation. Even though no difference in Lrp6 IEC-KO colons was observed compared to controls in basal conditions ( Figure S1A-C), DSS-treated Lrp6 IEC-KO mice displayed increased DAI and more severe mucosal histological damage than control mice. The mechanisms explaining this increased susceptibility to colitis in the absence of Lrp6 still need to be determined. However, although LRP6 was dispensable for the maintenance of intestinal homeostasis under homeostatic conditions, LRP6 may be important to ensure adequate intestinal mucosal recovery through crypt regeneration, following epithelial damage. Indeed, histological analysis of Lrp6-deficient colonic mucosae after DSS treatment revealed the presence of large areas still denuded of epithelium, as opposed to DSS-treated control mice. Increased epithelial cell proliferation is necessary to allow intestinal mucosal recovery after DSS-induced injury [37]. In this regard, IBD patient mucosa exhibits enhanced epithelial proliferation, which decreases during resolution of inflammation [38]. Altogether, these data suggest that, by promoting crypt regeneration, LRP6 may protect the intestinal epithelium against damage and inflammation.
Most CRC develop via adenomatous polyps initiated by mutations activating the Wnt/β-catenin pathway, commonly in the adenomatous polyposis coli (APC) gene, or less frequently in the β-catenin (CTNNB1) gene. Interestingly, LRP6 overexpression in CRC cells activates Wnt/β-catenin signaling and cell migration [9]. However, whether LRP6 expression contributes to Wnt signaling and CRC tumoral properties, especially in CRC with APC or β-catenin mutations, remains unclear. Chen and He (2019) did not observe changes in Wnt signaling activation in APC-deficient CRC cells after Lrp6 knockout by CRISPR/Cas9 genome editing [39]. In contrast, Saito-Diaz et al. (2018) demonstrated that LRP6 knockdown in CRC cells, even those exhibiting APC mutations, significantly reduced Wnt signaling [40] and this was recently confirmed by single cell analyses of endogenous β-catenin levels [41]. Unfortunately, in their study, Saito-Diaz et al.
(2018) did not analyze the impact of LRP6 knockdown on CRC cell growth. Herein, we show that significantly decreased expression of LRP6 did not affect CRC cell clonogenic potential nor ability to form colonies in soft agar. Of note, the two cell lines analyzed, namely HCT116 and SW48 cells, exhibit mutations in CTNNB1 (β-catenin) and APC genes, respectively. Thus, our results suggest that CRC cells harboring activating mutations in Wnt signaling do not need LRP6 expression for their growth. Likewise, epithelial Lrp6 deletion does not affect intestinal tumor load and size in Apc Min/+ mice. Taken together, these findings support the notion that epithelial LRP6 expression is dispensable not only for the maintenance of intestinal homeostasis but also for tumorigenesis induced by aberrant β-catenin signaling.
Serrated adenocarcinomas, which account for 20-30% of CRC, follow an alternative pathway independent of APC mutations, in which serrated polyps replace traditional adenoma as the CRC precursor lesion [42]. This pathway involves early BRAF mutations, excess CpG island methylation and DNA microsatellite instability (MSI). Previous reports have demonstrated that the expression of constitutively active mutants of BRAF or KRAS in non-transformed IEC, such as IEC-6 cells, is sufficient to promote tumoral transformation [7,43]. We previously reported that expression of BRAF V600E or KRAS G12D oncogenes in IEC activates β-catenin nuclear transcriptional activity and target gene expression [7]. Notably, LRP6 was phosphorylated on serine-1490 and threonine-1572 in a MEK-dependent manner, in IEC transformed by oncogenic KRAS or BRAF, thus providing a mechanism integrating KRAS/MAPK and canonical Wnt/β-catenin signaling during intestinal transformation. Indeed, both the serine-1490 and threonine-1572 residues are localized within the PPPS/TP motifs of the LRP6 co-receptor, motifs required for Wnt signal transduction.
Herein, we observed ERK-dependent LRP6 phosphorylation in human CRC cell lines with KRAS or BRAF mutations. This suggests that LRP6 phosphorylation is involved in oncogenic signaling induced by KRAS or BRAF mutations. However, LRP6 silencing in HCT116 cells which harbored a KRAS mutation did not significantly alter anchorageindependent growth nor clonogenic potential. It is important to note that, in addition to KRAS mutation, HCT116 cells exhibit several other genetic alterations including CTTNB1 mutation, PIK3CA mutation, CIMP and microsatellite instability that may also stimulate their growth. To circumvent this situation, we used a simpler model, namely IEC-6 stably expressing an inducible oncogenic form of human BRAF, BRAF V600E :ER. Following stimulation with 4-hydroxytamoxifen, BRAF V600E induced ERK1/2 signaling, cell transformation and anchorage-independent growth [43]. LRP6 knockdown in this model significantly reduced the capacity of BRAF V600E to induce colony formation in soft agar, which is a measure of anchorage-independent growth potential correlating with tumorigenic growth in vivo. Nonetheless, it will be important in future studies to confirm this function of LRP6 in a more relevant model of human cancer such as patient-derived CRC organoids (e.g., with or without BRAF mutation). Indeed, fewer genetic or epigenetic events are required to induce a malignant transformation in murine cells compared with human cells [10], and patient-derived tumoroids, which grow as irregular compact structures, maintain hallmarks of the isolated CRC tumor type [44]. They thus represent a more relevant model for drug screening in cancer [45].

Conclusions
In conclusion, although LRP6 does not play an essential role in the regulation of intestinal homeostasis and Apc-induced tumorigenesis in mice, LRP6 promotes crypt regeneration after epithelium damage and contributes to oncogenesis induced by BRAF oncogene. Our results suggest that LRP6 could represent a novel target for colorectal tumorigenesis associated with KRAS or BRAF mutations.

Supplementary Materials:
The following are available online at https://www.mdpi.com/article/ 10.3390/cells10071792/s1. Table S1: Histological damage scoring description, Figure S1: Loss of LRP6 expression does not alter colonic architecture. (A) Crypt depth was determined on H&Estained colonic sections from Lrp6 IEC-KO mice and control littermates (n ≥ 5). Scale bars, 50 µm. (B) Proliferative cells were observed by BrdU immunohistochemistry on colonic sections from mutant mice compared with control littermates (n = 6). Scale bars, 50 µm. (C) Colon weight (mg), colon length (cm) and body weight (g) were measured in 3-month-old mice (n ≥ 5). Data are expressed as mean ± SEM. Mann-Whitney U test, Figure S2: Wnt3a does not rescue impaired Lrp6 IEC-KO enteroid development. (A) Enteroid viability was determined 5 days after seeding by phase contrast microscopy on Lrp6 IEC-KO organoids treated or not with Wnt3a (100 ng/mL) from day 1 to day 5. The average of 3 independent experiments, each with at least 50 organoids, is summarized in the graph. Unpaired t-test. (B) Protrusions per enteroid were counted 5 days after seeding on Lrp6 IEC-KO organoids treated or not with Wnt3a (100 ng/mL) from day 1 to day 5. The average of 3 independent experiments, each with at least 40 organoids, is summarized in the graph. Chi-square test. Data are expressed as mean ± SD, Figure S3  Informed Consent Statement: Not applicable.

Data Availability Statement:
No new data were created or analyzed in this study. Data sharing is not applicable to this article.