Senescent cells perturb intestinal stem cell differentiation through Ptk7 induced noncanonical Wnt and YAP signaling

Cellular senescence and the senescence-associated secretory phenotype (SASP) are implicated in aging and age-related disease, and SASP-related inflammation is thought to contribute to tissue dysfunction in aging and diseased animals. However, whether and how SASP factors influence the regenerative capacity of tissues remains unclear. Here, using intestinal organoids as a model of tissue regeneration, we show that SASP factors released by senescent fibroblasts deregulate stem cell activity and differentiation and ultimately impair crypt formation. We identify the secreted N-terminal domain of Ptk7 as a key component of the SASP that activates non-canonical Wnt / Ca2+ signaling through FZD7 in intestinal stem cells (ISCs). Changes in cytosolic [Ca2+] elicited by Ptk7 promote nuclear translocation of YAP and induce expression of YAP/TEAD target genes, impairing symmetry breaking and stem cell differentiation. Our study discovers secreted Ptk7 as a factor released by senescent cells and provides insight into the mechanism by which cellular senescence contributes to tissue dysfunction in aging and disease.

Cellular senescence and the senescence-associated secretory phenotype (SASP) are implicated in aging and age-related disease, and SASP-related inflammation is thought to contribute to tissue dysfunction in aging and diseased animals. However, whether and how SASP factors influence the regenerative capacity of tissues remains unclear. Here, using intestinal organoids as a model of tissue regeneration, we show that SASP factors released by senescent fibroblasts deregulate stem cell activity and differentiation and ultimately impair crypt formation. We identify the secreted N-terminal domain of Ptk7 as a key component of the SASP that activates non-canonical Wnt / Ca 2+ signaling through FZD7 in intestinal stem cells (ISCs). Changes in cytosolic [Ca 2+ ] elicited by Ptk7 promote nuclear translocation of YAP and induce expression of YAP/TEAD target genes, impairing symmetry breaking and stem cell differentiation. Our study discovers secreted Ptk7 as a factor released by senescent cells and provides insight into the mechanism by which cellular senescence contributes to tissue dysfunction in aging and disease. Cellular senescence is recognized as an evolutionarily conserved process in tumor suppression, wound healing and regeneration. However, the accumulation of senescent cells during aging can contribute to age-related diseases [1][2][3][4][5][6][7][8] . Cellular senescence is characterized by an essentially permanent cell cycle arrest, as well as by impaired apoptosis [9][10][11][12] . A critical feature of senescent cells is the senescenceassociated secretory phenotype (SASP): the secretion of numerous proinflammatory cytokines, chemokines, proteases, growth factors, and lipids [13][14][15][16][17] . SASP factors have been implicated in promoting cellular plasticity and stem cell gene expression 18 , but can also cause inflammation and negatively impact the local tissue environment and function. Whether and how the SASP influences stem cell function and thus impacts tissue regeneration in the aging organism, however, remains unclear.
Organoid systems provide an efficient and convenient model of tissue regeneration to dissect the effects of SASP components on stem cell function. Intestinal organoids, for example, can be established from single stem cells that develop into budding organoids with defined crypt and villus compartments 19,20 . During organoid development, stem cells proliferate and form a multicellular sphere, which breaks symmetry and self-organizes into a crypt-containing organoid 21 . Fully mature intestinal organoids recapitulate cell types and features present in the intestinal epithelium, providing an attractive system for disease modeling and drug screening [22][23][24][25][26][27][28][29][30] .
The Wnt signaling pathway is critical for intestinal stem cell proliferation and maintenance [31][32][33][34][35] . Mutations causing constitutive activation of Wnt signaling are associated with epithelial hyperplasia and tumor formation, whereas reduced Wnt signaling, for example due to elevated expression of the negative Wnt regulator Notum in intestines of old mice, leads to decreased stem cell activity and impaired regeneration [36][37][38] .
Canonical β-Catenin-dependent Wnt signaling is triggered by Wnt ligand-mediated formation of a ternary complex between Fzd receptors and LRP5 or 6 co-receptors, resulting in inhibition of the β-catenin destruction complex 32,39 . β-catenin-independent Wnt signaling, sometimes referred to as noncanonical Wnt signaling, can be triggered by Wnt ligand binding to Fzd receptors with co-receptors Ror2, Ryk, and Protein Tyrosine Kinase 7 (PTK7), leading to activation of JNK and Ca 2+ signaling [40][41][42][43] . Noncanonical Wnt signaling results in cytoskeletal rearrangements and influences cell migration and planar cell polarity. While the role of canonical β-Catenin-dependent Wnt signaling in intestinal stem cell activity and regeneration is well established, it remains unclear whether β-catenin-independent Wnt signaling influences intestinal stem cells.
Ptk7, also called colon cancer Kinase 4 (CCK4), is a member of the receptor tyrosine kinase (RTK) family, and contains N-terminal Immunoglobulin-like repeats (Ig), a transmembrane region, and an inactive tyrosine kinase domain [55][56][57] . The extracellular domain of Ptk7 can be shed after cleavage by metalloproteases, and the kinase domain then translocates to the nucleus after subsequent cleavage by γsecretase 58,59 . Ptk7 influences cellular processes such as migration and planar cell polarity (PCP) [60][61][62][63][64][65][66][67] . In Drosophila, the Ptk7 orthologue Otk regulates intestinal stem cell migration after epithelial injury. Injuryinduced expression of matrix metalloproteinases promotes shedding of the N-terminal domain of Otk from enteroendocrine cells, inducing noncanonical Wnt signaling in ISCs, and promoting their migration towards the wound 68 .
Here, we report the identification of the N-terminal domain of Ptk7 as a SASP factor elevated in aging intestines, and released from senescent cells. In mouse intestinal organoids, N-terminal Ptk7 causes defective differentiation of intestinal epithelial cells. We find that secreted Ptk7 acts through Fzd receptors on intestinal stem cells (ISCs) to activate Wnt/Ca 2+ signaling, which in turn triggers nuclear translocation of YAP. Changes in YAP/TEAD-dependent gene expression are responsible for Ptk7-induced defects in ISC differentiation. Our work identifies Ptk7 secreted by senescent cells as a critical mediator of ISC dysfunction.

SASP impairs cell differentiation in mouse intestinal organoids
To ask whether and how SASP factors can influence ISC function and regeneration, we exposed mouse intestinal organoids derived from isolated crypts of the small intestine (jejunum) to conditioned media (CM) from mouse embryonic fibroblasts (MEFs) that were induced to senesce by X-ray irradiation or doxorubicin treatment (Fig. 1a 69,70 ,). Senescence-associated beta-galactosidase (SA-β-gal) staining confirmed senescence in the MEFs (Supplementary Fig. 1a). As a control, we used CM from quiescent (serum deprived) MEFs, which were growth arrested, similar to senescent cells. Organoids from fresh crypts acquired a distinct spherical cyst morphology after exposure to senescent CM (SCM), while organoids exposed to quiescent CM (QCM) exhibited normal crypt budding (Fig. 1b, c). Cystic organoids lacked defined crypt domains (Fig. 1d). Immunostaining for Ki67 shows significant proliferative activity in organoids exposed to both QCM and SCM. Mitotic figures (pH3+ cells) were observed mainly in crypt domains of QCM organoids, but were distributed throughout all areas of SCM organoids (Fig. 1d). These observations suggest that exposure to SCM did not cause secondary senescence in organoids (as no sign of cell cycle arrest was observed based on pH3 and Ki67 staining), but influence ISC function in different ways ( Supplementary Fig. 1b, c). To assess if the cystic organoid phenotype entails changes in cell composition, we compared transcript levels of various cell-type specific markers in QCM and SCM organoids. We observed a strong reduction in OLFM4 (intestinal stem cell), CHGA (enteroendocrine cell) and MUC2 (goblet cell) transcript levels in SCM organoids, while ALPI (enterocyte), MKI67 (proliferating cell), LGR5 (stem cell) and LYZ (Paneth cell) transcripts were not significantly changed (Fig. 1e). Whole mount immunostaining confirmed these findings: secretory cells such as goblet, enteroendocrine and Tuft cells (Dclk), and Olfm4+ SCs were reduced, while other cell types, including enterocytes, were not ( Fig. 1f-m, Supplementary Fig. 1d).
We confirmed these findings using single-nucleus RNA sequencing of organoids exposed to QCM and SCM. Using a droplet-based single-nucleus method, we sequenced 38154 nuclear transcriptomes from QCM and SCM-exposed organoids. Unsupervised graph-based clustering identified 15 distinct clusters, which were annotated using known cell markers (Fig. 1n, Supplementary Fig. 1e). The relative contribution of cells from QCM and SCM exposed organoids to each cluster revealed differences in cell-type composition between QCM and SCM exposed organoids (Fig. 1o, p, and Supplementary Fig. 1f): Compared to QCM exposed organoids, SCM exposed organoids consist of an increased proportion of stem cells and of cells expressing genes of a fetal signature, but a decreased proportion of enteroendocrine progenitors and of differentiated cells of the secretory lineage such as enteroendocrine, Tuft, and Paneth cells.
Critically, the morphological change caused by SASP factors was reversible (Fig. 1q, Supplementary Fig. 1g). When organoids were switched from QCM to SCM 4 days after seeding (when crypt buds become visible in QCM), they were smaller and exhibited less budding and cystic crypt domains. Budding organoids did not become completely cystic, suggesting that SASP factors influence stem cells primarily at early stages of symmetry breaking. On the other hand, cystic organoids switched from SCM to QCM 4 days after seeding, started budding and lost cystic morphology. Consistently, we observed no long-term effect of SCM on stem cell activity. When organoids exposed to SCM were passaged and maintained in regular organoid culture media, colony-forming efficiency remained comparable to those exposed to QCM ( Supplementary Fig. 1h, i). Together, these data indicated that SCM causes specific and reversible defects in cell differentiation in the intestinal stem cell lineage.
Secreted Ptk7 is a SASP factor causing the cystic organoid phenotype SASP factors include proinflammatory cytokines, chemokines and growth factors [9][10][11][12] . To identify factor(s) in SCM responsible for the morphological changes in organoids, we analyzed proteins in SCM by mass spectrometry (MS). SCM from a large-scale culture of senescent MEFs was dialyzed, concentrated and fractionated by ion exchange and size exclusion columns (Fig. 2a). Fractions categorized as containing small, medium and large proteins were tested for cystic organoid inducing activity, which were present only in the medium fraction ( Supplementary Fig. 2a). MS analysis of the small and medium fractions identified and quantified a total of 148 proteins (39 differentially present proteins with the following cutoffs: q value <0.05, absolute fold change >1.5) (Fig. 2b, Supplementary Fig. 2b, Supplementary Data 1 and 2). Proteins previously described as SASP components were Representative images of organoids after 5 d of culture in conditioned media. Enlarged images (boxed area) are shown at the right. Scale bar, 200 μm. c Quantification of data from (b). The percentage of cystic organoids in organoids cultured in conditioned media is shown (mean ± s.d.; two-tailed t-test). Each dot represents a result from organoids established from a mouse, 3 biological replicates per experiment (n = 3). d H&E staining (left column) and immunostaining of paraffin-embedded organoids. Immunostaining for Ki67 (green) and phospho-Histone H3 (red) in intestinal organoids cultured in conditioned media and stained with Hoechst (blue). Scale bar, 50 μm. e Gene expression in organoids exposed to SCM determined by qPCR and represented as a ratio relative to organoids exposed to QCM (mean ± s.d.; n = 4; two-tailed t-test). f-m Numbers of cells expressing the listed markers in organoids exposed to QCM or SCM (mean ± SEM; n = 10; twotailed t-test). n UMAP of pooled snRNA-seq data of nuclei from QCM and SCM exposed organoids colored by cell-type annotation (right). o UMAP showing contribution of nuclei from QCM and SCM exposed organoids labeled in blue and orange, respectively. p Relative contribution of cells from QCM and SCM exposed organoids to each cluster. Cells from QCM and SCM exposed organoids were labeled in blue and orange, respectively. q Representative images of organoids taken every 2 days starting on day 4. From day 0 to 4, organoids were cultured in quiescent conditioned media (left panel) and senescent conditioned media (right panel). On day 4, quiescent-conditioned media was changed to senescentconditioned media (left bottom panel), whereas senescent-conditioned media was changed to quiescent-conditioned media (right bottom panel). In all conditions, media was changed every other day. Scale bars, 100 μm. See also Fig. S1. Source data are provided as a Source Data file.
One of the proteins identified as differentially present in the medium fraction was the N-terminal extracellular domain of Ptk7 ( Fig. 2b-e, Supplementary Data 4). Ptk7 transcripts are robustly expressed in MEFs, and can also be detected in intestinal organoids ( Supplementary Fig. 2c). The N-terminal domain of Ptk7 is known to be shed after cleavage by either Matrix Metalloprotease 14 (MMP14) or A Disintegrin And Metalloprotease 17 (ADAM17) (Fig. 2d) 58,59,71 , and MMPmediated shedding of N-terminal Otk influences ISC migration in Drosophila 68 . Consistently, peptides detected in SCM mapped to the N-terminal extracellular domain (Fig. 2e). We confirmed the selective presence of the soluble/shed form of Ptk7 (sPtk7) in the medium fraction of SCM using a commercially available antibody against the N-terminal domain of Ptk7 ( Supplementary Fig. 2d). Conditioned media (CM) from both doxorubicin-and irradiation-induced senescent b Volcano plot representing the statistical significance and protein abundance ratios between medium (active) and small (inactive) fractions. Dashed lines depict a q = 0.05 cutoff (horizontal) and 1.5-fold difference cutoffs (vertical). Proteins that are present highly in medium (active) and small (inactive) fractions are represented in red and blue, respectively. Proteins with q < 0.01 and more than 1.5-fold difference are labeled. c Graphs showing peptide peak area profiles for Ptk7. Each Ptk7 peptide detected is shown as a black line, indicating the peak area measurements for the inactive and active fractions. The average of all Ptk7 peptides is presented as a red line. d Schematic representation of the Ptk7 protein domain structure. Ig: Immunoglobulin-like loops (lilac), TM: a transmembrane region (purple), and an inactive tyrosine kinase domain (blue). e A graphic representation of the list of peptides detected by MS. Ptk7 peptide sequences detected by MS in lilac and the transmembrane domain in purple. f Western blot for Ptk7 in conditioned media collected from quiescent and doxorubicin-or irradiation-induced senescent MEFs. Conditioned media corresponding to the same number of cells was loaded. Increased shedding of Ptk7 (soluble form of Ptk7; sPtk7) was detected in senescent conditioned media. g Immunofluorescence staining of Ptk7 (red) and stromal cell marker Vimentin (green) in the large intestine. Nuclei were counter stained with DAPI (blue). Scale bar: 50 μm. h The relative PTK7 expression in intestinal organoids and fibroblasts (mean ± s.d.; n = 3; two-tailed t-test). i Western blots of Ptk7 in the large intestine extract from young and old mice. Blot images of two different exposure times are shown. Each lane presents tissue lysate from one mouse (n = 3 for young mice, n = 5 for old mice). Actin was used as loading control. j Quantification of data from (i). Ptk7 and Actin protein levels were analyzed by ImageJ. Ptk7 protein levels normalized to Actin are shown as mean ± s.d. Two-tailed t-test was used for statistical analysis. See also Fig. S2. Source data are provided as a Source Data file. cells contained higher levels of sPtk7 compared to CM from quiescent cells ( Fig. 2f; loading was normalized by the number of cells cultured in the corresponding media), and increased sPtk7 was detected in the SASP Atlas of senescent human lung fibroblasts (Supplementary Data 3). These observations indicate that secreted Ptk7 is a common component of the SASP.

Increased shedding of sPtk7 expresses in the aging mouse intestine
Since senescent cells increase with age in many tissues 1,2,8 , we asked whether secreted Ptk7 could be detected at a higher frequency in intestines of old mice. Ptk7 transcripts and proteins were detected in both epithelial and stromal cells of the mouse intestine, with higher expression in stromal cells (Supplementary Fig. 2e and Fig. 2g, h). We did not detect any differences in Ptk7 transcript levels between young and old intestine ( Supplementary Fig. 2f). Consistently, total Ptk7 protein levels did not change with age. However, Ptk7 shedding was strongly increased in the old intestine, as indicated by the increased levels of shed N-terminal Ptk7 (Fig. 2i, j).
ADAM and MMP expression increase in senescent cells [72][73][74][75] , which may explain the increased release of sPtk7 into the SCM. To test this possibility, we exposed senescent MEFs to metalloprotease inhibitors previously shown to block Ptk7 shedding 59 and collected SCM. Metalloprotease inhibition reduced the levels of sPtk7 in SCM (Fig. 3a). Organoids exposed to this SCM (after dialysis to remove metalloprotease inhibitors) showed reduced cystic phenotypes (Fig. 3b, c). To validate that sPtk7 is responsible for the cystic organoid phenotype, we exposed organoids to SCM collected from Ptk7-deficient fibroblasts.
SCM collected from Ptk7-deficient fibroblasts contained reduced levels of sPtk7 (Fig. 3d), and organoids exposed to this SCM showed reduced cystic organoid phenotypes compared to its counterpart (Fig. 3e, f). We further blocked Ptk7 directly in SCM by adding anti-Ptk7 antibodies targeting the N-terminal domain to the organoid culture, which suppressed the cystic organoid phenotype in a dose-dependent manner (Fig. 3g, h).
We asked whether co-culturing organoids with intestinal fibroblasts recapitulates what we observed using conditioned media. Cellular quiescence and senescence of intestinal fibroblasts were induced by serum deprivation and doxorubicin treatment, respectively. SA-β-gal staining confirmed senescence in the intestinal fibroblasts (Supplementary Fig. 2g). Senescent intestinal fibroblasts showed increased Ptk7 shedding, which coincides with increased levels of MMP14 (Supplementary Fig. 2h and Fig. 3i). While organoids co-cultured with quiescent intestinal fibroblasts displayed normal budding organoid morphology, organoids co-cultured with senescent intestinal fibroblasts exhibited cystic morphology that resembled SCM exposed organoids (Fig. 3j, k). Inhibition of Ptk7 suppressed this cystic phenotype caused by co-culture with senescent intestinal fibroblasts (Fig. 3j, k).
Since exogenous Wnt3a can also cause a cystic phenotype in intestinal organoids 80 , we tested for a direct effect of Wnt secretion by senescent cells. We treated senescent MEFs with a Porcupine inhibitor to block Wnt secretion 81 . Organoids exposed to SCM from these MEFs i Quantification of data from (h). mean ± s.d. The percentage of cystic organoids in organoids cultured in conditioned media is shown. In (c), (e), (g), and (i), data are presented as means ± s.d. One-way ANOVA followed by Tukey's post-hoc multiple comparison tests. Each dot represents a result from organoids established from a mouse, n = 3. j Representative ELISA binding curves showing binding of Ptk7 or FZD7 to Wnt ligands. k Representative ELISA binding curves showing binding of Ptk7 to Wnt ligands in the absence or presence of 500 nM FZD7. In (j) and (k), data are presented as means ± s.d. n = 3 (l) Relative Ptk7 and FZD7 binding to Wnt ligands from (k). Relative Ptk7 binding was calculated by normalization to Ptk7 and Wnt ligand binding at 1500 nM in the absence of FZD7 (black striped bar). Relative FZD7 binding was calculated by normalization to FZD7 binding to Wnt ligands in the absence of Ptk7 (white striped bar). Data are presented as means ± s.d. See also Fig.  S3. Source data are provided as a Source Data file.
(after dialysis to remove the inhibitor) still exhibited cystic phenotypes, confirming that Ptk7-but not Wnt ligands-in the SCM causes morphological alterations in organoids (Supplementary Fig. 3a-d). Furthermore, a Wnt reporter assay indicated no activation of β-Catenin-dependent Wnt signaling by SCM ( Supplementary Fig. 3e).
Combined inhibition of FZD5, 7 and 8, as well as selective inhibition of FZD7 or FZD8, but not of FZD5, strongly suppressed the cystic phenotype induced by SCM, confirming a role for these FZD receptors in the SASP-induced phenotype (Fig. 4b, c). Inhibition of LRP5/6, in turn, failed to rescue the cystic phenotype, suggesting that sPtk7 acts through a FZD-dependent but LRP5/6-independent Wnt signaling pathway.
To further examine the effect of Wnt signaling on organoids, we added Wnt3a and Wnt5a recombinant proteins to crypt cultures. While both Wnt3a and Wnt5a dose-dependently increased organoids with cystic morphology, there was a noticeable difference in organoid size (Fig. 4d, e). Crypts treated with Wnt5a developed cystic organoids that were comparable to organoids exposed to SCM, but larger than those treated with Wnt3a. Wnt11, another noncanonical Wnt ligand that acts through FZD7 85,86 , failed to induced cystic organoids ( Supplementary  Fig. 3o, p). Similar to SCM, FZD receptor inhibition reduced cystic organoids generated by Wnt5a, but inhibition of LRP5/6 did not (Fig. 4f, g).
The percentage of cystic organoids caused by Wnt5a was reduced by Ptk7 inhibition (Fig. 4h, i), suggesting an interaction between Wnt5a and Ptk7 that contributes to cystic organoid morphology. To test this possibility, we developed an ELISA to probe binding between purified recombinant proteins -Ptk7 extracellular domain (sPtk7), FZD7 CRD domain and Wnt ligands ( Supplementary Fig. 3q). sPtk7 and FZD7 bind both Wnt3a and Wnt5a, but sPtk7 bound Wnt5a with higher affinity, whereas FZD7 showed the opposite (Fig. 4j). Since both sPtk7 and FZD7 bind Wnt ligands, we asked whether Ptk7 and FZD7 compete for binding to Wnt ligands. sPtk7 binding to either Wnt3a or Wnt5a was unaffected by a saturating amount of FZD7 CRD, and the interaction between FZD7 CRD and these Wnt ligands was not affected by Ptk7 (Fig. 4k, l), suggesting that Ptk7 and FZD7 bind different sites on Wnt. Together with the observation that inhibition of either FZD or Ptk7 rescued the cystic organoid phenotype caused by Wnt ligands, these results suggest that Ptk7/FZD7/Wnt form a ternary complex to activate Wnt signaling.

Cytosolic Ca 2+ oscillations in intestinal organoids
Wnt5a-mediated noncanonical signaling elicits cytoskeletal changes through activation of Rho and Rac GTPases and modulation of cytosolic Ca 2+ levels [87][88][89][90] . We and others have shown that intestinal stem cells in Drosophila exhibit regular cytosolic and mitochondrial Ca 2+ oscillations, and that modulation of these oscillations in response to mitogenic signals control ISC proliferation and metabolic reprogramming [91][92][93] . Whether Ca 2+ signaling also influences intestinal stem cell activity in the mammalian gut remains unclear. To assess possible effects of noncanonical Wnt signaling on Ca 2+ levels in mammalian ISCs, we expressed the fluorescent Ca 2+ indicator GCaMP6f under control of Lgr5-Cre and performed Ca 2+ imaging in live intestinal organoids (Fig. 5a). In the absence of stimuli, Ca 2+ oscillations were observed in multiple cells of organoid crypts (Fig. 5b, Supplementary Movie 1). In unchallenged organoids, baseline cytosolic Ca 2+ levels and the frequency and intensity of oscillations were variable and did not reveal an obvious pattern (Fig. 5c, d, Supplementary Movie 2) When exposed to Wnt5a, however, cytosolic Ca 2+ levels and oscillations increased (Fig. 5c-e, Supplementary Fig. 4a, Supplementary Movie 3). This increase was distinct from Wnt3a treatment, which resulted in decreased cytosolic Ca 2+ levels ( Supplementary  Fig. 4b), but also occurred when organoids were exposed to SCM. SCM, but not QCM, increased the frequency and magnitude of cytosolic Ca 2+ oscillations (Fig. 5f, g; Supplementary Fig. 4c, d). Addition of SCM in the presence of anti-Ptk7 failed to induce any changes in Ca 2+ oscillations, suggesting that sPtk7 stimulates the Wnt/Ca 2+ pathway in ISCs (Fig. 5h, Supplementary Fig. 4e). Ptk7 inhibition in unchallenged organoids decreased cytosolic Ca 2+ oscillations, suggesting a role for endogenous Ptk7 in modulating Ca 2+ signaling (Fig. 5h, Supplementary Fig. 4e).

Inhibition of Ca 2+ signaling suppresses the cystic phenotype
To explore the mechanistic link between increased Ca 2+ signaling and the cystic organoid phenotype, we asked whether inhibition of Ca 2+ signaling components suppresses SCM or Wnt5a-induced cystic organoid phenotypes. Trifluoperazine (TFP), an antagonist of the multifunctional Ca 2+ binding protein Calmodulin (CaM) 94 , reduced the percentage of cystic organoids induced by SCM or Wnt5a (Fig. 5i, j, Supplementary Fig. 5a, b). Inhibition of Calcineurin by FK506 or cyclosporin A had no effect on this phenotype, but inhibition of CaMKK and CaMKII by STO-609 and KN-62, respectively, suppressed the cystic organoid phenotype (Fig. 5k, l, Supplementary Fig. 5c-e). These data suggest that cytosolic Ca 2+ release and subsequent activation of Calmodulin / CaMK signaling lead to the cystic organoid morphology.
YAP/TEAD target gene enrichment in SASP-exposed organoids We performed transcriptome analysis to gain further insight into changes elicited in intestinal organoids by SCM and Ptk7. Intestinal organoids were cultured in complete media, QCM or SCM (with or without anti-Ptk7), and complete media with Wnt5a (with or without anti-Ptk7) for 2 days, and subjected to RNAseq analysis ( Fig. 6a; all cultures were exposed to either anti-Ptk7 or IgG control). Principal component analysis (PCA) of 36 organoid transcriptomes revealed a marked similarity between the transcriptomes of organoids cultured in complete media and those in QCM, but significant differences between organoids cultured in other conditions (Fig. 6b). Differential gene expression analysis revealed that close to half the genes induced by Wnt5a were also induced by SCM, while the majority of SCM-induced genes were also induced by Wnt5a, consistent with a major contribution of noncanonical Wnt signaling to the SCM-induced cystic organoid phenotype. Of genes differentially regulated by Wnt5a and SCM, about half were Ptk7 dependent, confirming a critical role for Ptk7 in the SCM-induced phenotype (Fig. 6c).
A role for YAP in SCM-mediated effects in organoids was supported by our snRNA-seq analysis. While we did not see any clear changes in ISC/CBC specific genes, we observed an overall increase of YAP target gene expression across different clusters ( Supplementary  Fig. 6e-h). Increase of YAP target gene score compared to ISC/CBC signature was particularly noticeable in the fetal signature gene expressing cell cluster ( Supplementary Fig. 6i, j). YAP activation promotes expression of genes constituting a fetal signature in intestinal organoids and in the injured intestinal epithelium 53,[97][98][99] . Consistently, we observed expression of several fetal signature genes in cells with high YAP target gene expression ( Supplementary Fig. 6k).
Altogether, these data suggest that the cystic organoid phenotype caused by Ptk7 or Wnt5a may result from elevated YAP/TEAD activity. This hypothesis was supported by the relative nuclear localization of YAP in SCM or QCM exposed organoids. Active YAP translocates to the nucleus to form a complex with TEADs 48 . Whole mount immunostaining showed that YAP localizes mainly to the cytosol in budding organoids in QCM (Fig. 7a, top). However, cystic organoids in SCM or after Wnt5a treatment displayed increased nuclear localization of YAP (Fig. 7a, middle and bottom). Interestingly, organoid morphology correlated strongly with YAP localization. In rare cystic organoids in QCM, YAP also localized to the nucleus, while the few normally budding organoids after SCM or Wnt5a treatment exhibited cytosolic YAP ( Supplementary Fig. 7a, b). Additionally, the morphological changes observed after switching CM corresponded to the nuclear localization of YAP ( Supplementary Fig. 7c). This finding is consistent with the reported need for heterogeneous YAP activation for symmetry breaking in organoids 21 , and suggests that constitutive YAP activation and YAP/TEAD target gene expression may cause the cystic organoid phenotype.

Inhibition of YAP reverses cystic phenotype
To test this hypothesis, we asked whether inhibition of nuclear YAP rescues the cystic organoid phenotype. Verteporfin is a small molecule that inhibits YAP 100 . Verteporfin strongly reduced the cystic organoid percentage due to QCM and SCM ( Supplementary Fig. 7d, Fig. 7b, c). We also used an irreversible covalent TEAD inhibitor, TED347, which blocks TEAD binding to YAP 101 . Similar to verteporfin, TED347 reduced the percentage of cystic organoids (Fig. 7d, e).
Ca 2+ signaling has been implicated in the regulation of YAP, but the effect of perturbing cytosolic [Ca 2+ ] on YAP nuclear translocation may be context-dependent 102 . To test whether Ca 2+ signaling regulates YAP localization downstream of SCM/sPtk7/Wnt5a in organoids, we inhibited CaM, CaMKK and CaMKII using TFP, STO-609 and KN-62, respectively. Inhibition of these Ca 2+ signaling components reversed SCM-or Wnt5a-induced nuclear translocation of YAP (Fig. 7f), as well as the cystic organoid phenotype (Fig. 5j-m). Together, our results demonstrate that noncanonical Wnt-Ca 2+ signaling promotes YAP/ TEAD activation in intestinal stem cells, resulting in the cystic morphologies adopted by organoids exposed to SCM or Wnt5a.

Discussion
The aging intestinal epithelium displays reduced regenerative capacity. Several age-dependent changes such as reduced Wnt signaling, elevated Cdc42 activity, and metabolic changes have been suggested to cause reduced intestinal stem cell activity in the intestinal crypt 36,37,[103][104][105][106] . At the same time, aging is associated with dramatically increased colorectal cancer incidence, likely resulting from intestinal stem cell deregulation. Whether and how senescent cells contribute to these changes in intestinal homeostasis remains unclear.
Our data suggest a possible role for the SASP in the age-related deregulation of ISC function and differentiation (Supplementary    . 6l, m). The SASP (including factors like sPtk7, which are not technically secreted but are shed as a consequence of senescent cell surface remodeling) is believed to be a critical part of the contribution of senescent cells to age-related disease, primarily by influencing the tissue microenvironment and spreading senescence through a "bystander effect" [107][108][109][110][111] . Accordingly, selective elimination of senescent cells improves many aging symptoms and disease phenotypes 2,3,8,9,112 . Many SASP components have been identified using proteomic and transcriptomic methods 13,16,17,109,[113][114][115] . However, how individual SASP factors influence the tissue microenvironment and contribute to tissue dysfunction is incompletely understood. Our study identifies sPtk7 as a critical SASP factor that has a direct and reversible impact on intestinal stem cell proliferation and differentiation (Fig. 7g, h).
Ptk7 has previously been identified as a surface marker of human colonic stem cells, with cells expressing high levels of Ptk7 displaying the highest self-renewal and re-seeding capacity in 3D organoid culture 116 . Our data show that Ptk7 is also expressed in fibroblasts and epithelial cells of the mouse small intestine, and that shedding of the N-terminal domain of Ptk7 is increased in the gut of old mice. Of note, we did not detect sPtk7 in the serum of mice at any age, suggesting that sPtk7 acts locally rather than systemically. Accordingly, we have found in a recent study that the N-terminal domain of the Ptk7 orthologue Otk is shed from enteroendocrine cells after intestinal epithelial injury, and acts locally to attract migrating stem cells to the site of damage 68 .
Our co-culture experiments of intestinal organoids with senescent intestinal fibroblasts further show that fibroblast-derived Ptk7 impacts differentiation of intestinal stem cells, causing a cystic organoid phenotype. How this effect on intestinal stem cells influences epithelial homeostasis and regeneration remains to be established.
Previous studies have observed cystic intestinal organoid morphology after activation of β-catenin-dependent canonical Wnt signaling [117][118][119][120] . Our data indicate a distinct difference between Wnt3a and Wnt5a. Wnt5a can induce the cystic organoid phenotype by activating noncanonical Wnt/Ca 2+ signaling. sPtk7 phenocopies these Wnt5a-mediated effects, acting through Ca 2+ and YAP/TEAD signaling. Mechanistically, our data provide evidence for a possible ternary complex between sPtk7, Wnt5a and FZD7, where Wnt5a mediates the interaction between FZD7 and cell surface Ptk7. The ternary complex is required for cystic organoid formation, as inhibition of either FZD or Ptk7 suppresses the Wnt5a-induced cystic organoid phenotype. Wnt5a is known to play a role in intestinal development and wound healing [121][122][123] . Additional studies to explore if Ptk7 similarly influences intestinal development and wound healing will be of interest in the future.
At the molecular level, it is conceivable that sPtk7 acts similarly to Wnt5a, a notion that is supported by the fact that the fly orthologue of Ptk7 can form homo-oligomers and bind to FZD1/2 via its N-terminal domain 79 . It is also possible that a yet to be identified cell surface coreceptor inhibits ternary Wnt/FZD7/Ptk7 complex formation by sequestering cell surface Ptk7, and that sPtk7 binds this co-receptor to liberate cell surface Ptk7 and enable or promote complex formation with Wnt/FZD7 (especially at low endogenous Wnt concentrations). In such a scenario, recombinant Wnt5a (high Wnt) might force ternary complex formation without the need for secreted Ptk7.
Our observation of cytosolic [Ca 2+ ] oscillations in mammalian intestinal epithelial cells is reminiscent of [Ca 2+ ] oscillations found to regulate stem cell quiescence in the fly intestinal epithelium 91 . Supporting the proposed role of sPtk7 in modulating Wnt5a-mediated signaling, we find that sPtk7 and Wnt5a modulate cytosolic Ca 2+ oscillations, while Wnt3a does not. Interestingly, inhibition of Ptk7 strongly suppressed baseline Ca 2+ oscillations, suggesting a role for Ptk7 in maintaining homeostatic Ca 2+ signaling. Further studies are needed to fully explore the role of Ptk7 in Ca 2+ signaling in stem cells during homeostasis and regeneration.
In intestinal organoids, our data indicate a critical role for Ca 2+ / CaM/CaMKK/CaMKII signaling in promoting constitutive nuclear translocation of YAP and YAP/TEAD target gene expression after Wnt5a/sPtk7 exposure. It remains to be established, however, whether this influence of Ca 2+ signaling on YAP nuclear translocation is direct or whether it is a consequence of more general cellular changes triggered by elevated cytosolic [Ca 2+ ]. The constitutive activation of YAP/TEAD target expression causes the cystic phenotype, consistent with reports from early organoid development in which abnormal YAP activation prevents symmetry breaking and results in spheroid organoids 21 .
We propose an impact of YAP activation on epithelial cell differentiation, as exposure to SCM resulted in reduced numbers of differentiated cell types, especially of cells in the secretory lineage, as well as the induction of a fetal gene signature. Accordingly, loss of differentiated cell types and induction of fetal genes has been reported when YAP is activated in the mouse intestine 51,52,54 . This activation of YAP is important for tissue regeneration, and we propose that the observed sPtk7-mediated YAP activation also regulates epithelial regeneration in young animals. Such a role for sPtk7 would be consistent with the recently described function of the Drosophila orthologue of Ptk7, Otk 68 . After tissue injury, the N-terminal domain of Otk is shed from enteroendocrine cells and promotes noncanonical Wnt signaling in nearby intestinal stem cells, promoting their migration to the site of injury. Our data suggest that such an evolutionarily conserved role for transient release of sPtk7 during regeneration results in tissue dysfunction in aging animals, where chronic exposure to elevated sPtk7 (as observed in our study) would prevent proper cell differentiation of the epithelium. Future studies exploring the effects of sPtk7 during tissue regeneration in the young and old mouse intestine will be of interest.
The specific loss of Olfm4+ SCs over Lgr5+ SCs we observed is interesting in this context. While both Lgr5 and Olfm4 are stem cell markers, Lgr5 is a downstream target of canonical Wnt signaling, whereas the expression of Olfm4 is induced by Notch signaling in intestinal stem cells 124 . While this would indicate a repression of Notch activity in SCM exposed crypts, the reduced differentiation of secretory cells is reminiscent of activated Notch signaling [125][126][127] . Clearly, the interaction between Ptk7/FZD/Ca 2+ /YAP signaling and Notch signaling in intestinal stem cells and their daughter cells is an interesting topic for further study. It will be important to evaluate whether and how such an interaction causes differentiation defects or skewed cell fate decisions in the ISC lineage.
The SASP has been postulated as a main culprit for tissue dysfunction in aging and disease, in part because proinflammatory factors that are part of the SASP can cause excessive immune cell recruitment/ activation and consequential tissue damage. Notably, however, disruption of normal differentiation in the ISC lineage by SASP factors in our model occurs in the absence of immune cells. Our data thus reveal that factors secreted by senescent cells can also have direct effects on epithelial regeneration and stem cell function, effects that potentially contribute to age-related tissue dysfunction and oncogenic transformation.

Mice
All mice were used in accordance with protocols approved by Genentech's Institutional Animal Care and Use Committee and adhere to the NRC Guidelines for the Care and Use of Laboratory Animals. C57BL/6 (JAX:000664), GCaMP6f (JAX:030328) 128 , and ROSA26-tdTomato (JAX:007914) 129 mice were purchased from the Jackson Laboratory. Three-to four-month-old male C57BL/6 mice (CRL:027) and 16-18 months old male C57BL/6 mice (CRL:701) were purchased from Charles River laboratories and further aged until 28-30 months of age. Ptk7 fl/fl (Gene ID: 71461) mice were purchased from genOway. Lgr5 DTR/EGFP and Lgr5 CreERT2 mice were generated previously previously 130 , and GCaMP6f; ROSA26-tdTomato mice were crossed with Lgr5 CreERT2 mice to enable expression in Lgr5+ intestinal stem cells and their progeny.

Induction of quiescence and senescence
MEFs were split a day before induction of quiescence or senescence. To induce quiescence, MEFs were cultured in low serum (DMEM supplemented with 0.5% FBS and P/S at 37°C, 5% CO 2 , and 3% O 2 for 2 days. Cellular senescence was induced by X-ray irradiation (total 10 Gy) or 250 nM doxorubicin treatment for 3 days (doxorubicin; Sigma-Aldrich, #44583). After X-ray irradiation and change to fresh culture media (DMEM supplemented with 10% FBS and 1% P/S), MEFs were cultured at 37°C, 5% CO 2 , and 3% O 2 for 9 days. For doxorubicin treatment, cells were maintained at 37°C, 5% CO 2 during the 3-day doxorubicin treatment. Then, MEFs were washed with serum-free DMEM twice and cultured in DMEM supplemented with 10% FBS and 1% P/S at 37°C, 5% CO 2 , and 3% O 2 for 6 days. Culture media was changed every 2-3 days.

Collection of conditioned media
Twenty-four hours before collecting conditioned media (CM), MEFs were washed with serum-free DMEM twice, and serum-free DMEM was added. After 24 h in serum-free DMEM, this CM was collected, filtered using Steriflip filters (Millipore, #SE1M179M6) to remove cell debris, and concentrated using centrifugal filter units (Amicon ultra-15 centricon; Millipore, # UFC900324). To collect CM without sPtk7, TAPI-1 (Millipore, #59053) or GM6001 (Sigma-Aldrich, #364206) containing serum-free DMEM instead of serum-free DMEM was added to the cells. To inhibit Wnt secretion, IWP-L6 (Sigma-Aldrich, #504819) containing serum-free DMEM instead of serum-free DMEM was added to the cells. CM was filtered and dialyzed for 24-48 h to remove the inhibitors before concentration. Concentrated CM from equal numbers of MEFs was mixed with IntestiCult Organoid Growth Medium (Stem Cell Technologies, #06005) and added to the organoid cultures.

SA-β-Gal staining
Staining was performed with the Senescence Detection Kit (Abcam, #ab65351) following the manufacturer's protocol. Briefly, MEFs were fixed for 15 min with 1× fixation solution, washed with PBS, then stained overnight at 37°C in 1× staining solution. Brightfield images were captured on the Zeiss Axio Imager M2 microscope.

Mouse intestinal organoid culture
The jejunum part of the small intestine was removed, flushed with PBS, opened longitudinally, then cut into 0.5 cm pieces. The collected tissues were washed with ice cold PBS multiple times by vortexing and incubated in PBS containing 2.5 mM EDTA for 5 min at 37°C. After gentle shaking and aspiration, tissues were incubated in PBS containing 5 mM EDTA for 7 min at 37°C. Tissues were then washed with PBS to remove EDTA and vigorously shaken to release crypts. The released crypts were passed through 70 μm strainers, washed with PBS, and pelleted by centrifugation at 200 × g. Resuspended crypts were mixed with Matrigel (Corning, #356231; Lot#8218007) and pipetted into culture plates. After mixture of crypts and Matrigel solidified, Intesti-Cult Organoid Growth Medium (Stem Cell Technologies, #06005; Lots #1000021927, #1000034230, #1000037372) was added. Seven days after seeding, organoids were resuspended in dissociation buffer (Gentle Cell Dissociation Reagent; Stem Cell Technologies, #07174), dissociated by pipetting, and seeded for experiments.

Brightfield imaging of organoids and quantification
3-4 biological replicates of organoids from different mice were used per condition. Organoids were cultured in the indicated media, inhibitors, antibodies and/or recombinant proteins for 5 d and fixed in 4% formaldehyde in PBS. Organoids were imaged with a 4× Plan Fluor objective (NA: 0.13, Nikon) on a Nikon Ti-E inverted microscope equipped with a Neo scMOS camera (Andor, Oxford Instruments), a linear encoded automated stage (Applied Scientific Instrumentation), all run by NIS Elements software (Nikon). Organoids were imaged in brightfield and images covering the whole well were stitched and focused into one image projection with an extended depth of focus module (EDF; Nikon). One hundred to 200 organoids per replicate per condition were counted and used for quantification. The number of budding organoids in each condition was normalized by the number of budding organoids in untreated or DMSO treated controls to calculate cystic organoid percentage: Cystic organoid percentage = 100 × 1 À Number of budding organoids f rom experiment Number of budding organoids f rom control ð1Þ Induction of quiescence and senescence of mouse intestinal fibroblasts and co-culture with intestinal organoids Mouse intestinal fibroblasts were purchased from Cell Biologics (#C57-6231) and cultured in Complete Fibroblast Medium (Cell Biologics, M2267) at 37°C and 5% CO 2 . Cellular quiescence and senescence in mouse intestinal fibroblasts were induced by the similar methods described earlier for MEFs with minor modifications. Briefly, cellular quiescence was induced by culturing the fibroblasts in low serum (0.5% FBS), and cellular senescence was induced by 500 nM doxorubicin treatment for 5 days (doxorubicin; Sigma-Aldrich, #44583). Once cellular quiescence or senescence was induced, MEFs were maintained at 37°C, 5% CO 2 , and 3% O 2 for another 5 days. The quiescent or senescent fibroblasts were dissociated using 0.05% Trypsin-EDTA (Gibco, #25300-054) and counted for seeding. The same number of quiescent and senescent fibroblasts (5.2 × 10 4 cells) were mixed with intestinal organoids making total volume of 20 μl of Matrigel: organoid culture media mix, which was mounted on a 48 well plate.

Colony forming efficiency assay
Two millimolar EDTA in PBS was added to organoids to break Matrigel. Collected organoids were pelleted by centrifugation, washed with PBS, pelleted again, and incubated in prewarmed TrypLE Express (Gibco, #12604-013) for 8 min. The digestion was stopped by adding serum containing DMEM/F12, and the number of cells was assessed. 1 × 10 4 cells were seeded in total volume of 20 μl of Matrigel: organoid culture media mix per well in a 48 well plate. Once Matrigel was solidified, organoid culture media was added.

H&E and immunofluorescence stains
Organoids were paraformaldehyde-fixed and paraffin-embedded using standard procedures. The H&E Stain Kit (Abcam, #ab245880) was used following the manufacturer's manual. For Ki67 and pH3 antibody stains, sections were incubated for 15 min in citrate buffer, blocked in normal donkey serum, and incubated with primary antibodies overnight. Sections were then incubated with fluorescent secondary antibodies and applied with ProLong Diamond Antifade Mountant (Invitrogen, # P36970). Images were acquired with a Zeiss LSM 710. Information on antibodies is listed separately.

Whole mount imaging of organoids and quantification of different cell types
For whole mount staining, freshly passaged crypts were mixed with Matrigel (Corning, #356231) and mounted to µ-Slide 8 well chamber slides (ibidi, #80826). Media were changed every 2-3 d. 5 days later, organoids in the chamber slides were then fixed in 4% formaldehyde in PBS for 30 min, permeabilized with 0.5% Triton X-100 for 1 h, and incubated in blocking solution (3% normal donkey serum (Sigma-Aldrich, #566460) and 5% BSA (Sigma-Aldrich, #A7979) in 0.1% Triton X-100 + PBS) for 1 h. Primary and secondary antibodies were diluted in blocking solution and added to fixed organoids. Organoids were incubated with primary antibodies overnight at 4°C and with secondary antibodies for 2 h. ProLong Diamond Antifade Mountant (Invitrogen, # P36970) was added. Images were acquired with a Yokogawa CSU-W1/Zeiss 3i Marianas spinning disk confocal microscopy system. For each organoid, Z-planes up to 100 μm with 1 or 1.5 μm z-steps were acquired to quantify the number of cell types.

Quantitative PCR (qPCR)
RNA was isolated using the RNeasy Plus Mini Kit (Qiagen, #74134) following the manufacturer's instructions. cDNA was synthesized using iScript cDNA Synthesis Kit (Bio-Rad, #1708890) and used in standard qPCR reactions utilizing TaqMan assays (Thermo Fisher Scientific, probes listed separately) for target genes and TaqMan Universal PCR Master Mix (Thermo Fisher Scientific, #4326708). qPCR reactions were run on the QuantStudio 6 Real-Time PCR System (Applied Biosystems, Thermo Fisher Scientific) at the following thermal cycling conditions: 30 min at 48°C followed by 10 min at 95°C and 40 cycles of 10 s at 95°C and 1 min at 60°C. Data were normalized to the housekeeping gene GAPDH.

Immunofluorescence staining and in situ hybridization of the mouse intestine
Mouse large intestine tissues were harvested and fixed in 10% neutral buffered formalin for 24 h. Fixed tissues were then washed and processed for paraffin embedding. Formalin-fixed paraffin-embedded tissues were sectioned at 5 μm thickness and used for immunofluorescence staining and in situ hybridization. For immunofluorescence staining, sections were deparaffinized and processed using Target retrieval solution (Dako, #S1700). The sections were then blocked in 3% normal donkey serum, incubated with primary and secondary antibodies, and mounted using ProLong Diamond Antifade Mountant (Invitrogen, # P36970). In situ hybridization was performed using RNAscope mouse Ptk7 probe (ACD, #429191) and reagents (RNAscope 2.5 HD Reagent Kit; ACD, #322350) following the manufacturer's instructions.

Proteomic sample preparation
MEFs were cultured as above, with the exception that CM was prepared by incubating MEFs with DMEM without phenol red (Gibco, #31053-028) for 24 h. Filtered CM was snap-frozen and stored at −80°C after addition of complete protease inhibitors (Roche, #11836170001). Samples thawed on ice were additionally filtered using 70 kDa cutoff centrifugal filters (Millipore, #UFC700308). The filtrate was concentrated using 3 kDa cutoff Centricon-Plus 70 centrifugal filters to 60 ml. Additional complete protease inhibitors were added and then the concentrate was dialyzed against three changes of 4 L of 20 mM Tris, 100 mM NaCl, 5 mM EDTA, pH 8.0 at 4°C. Samples were then loaded onto a HiTrap Q HP column (Millipore, #GE29-0513-25) equilibrated with 20 mM Tris, 5 mM EDTA, pH 8.0. The flowthrough fraction was buffer exchanged via repeat dilution with PBS and concentration on 3 kDa cutoff filters, and loaded onto an albumin precoated Superdex-200 gel filtration column (Millipore, #GE28-9909-44). Sets of pooled fractions were tested for ability to promote the cystic organoid phenotype. The separated SASP fractions, containing 2-4 µg of proteins, were concentrated from 400 µL to 25 µL using 0.5 mL centrifugal filters, after which the samples were loaded on a 1D SDS-PAGE system. The gel bands were diced and collected in tubes, and vortexed several times in dehydration buffer (25 mM ammonium bicarbonate in 50% acetonitrile and water). The samples were concentrated to dryness (speedvac), reduced with 10 mM dithiothreitol and incubated for 1 h at 56°C with agitation. They were subsequently alkylated with 55 mM iodoacetamide and incubated for 45 min at room temperature in the dark. The diced gels pieces were washed with 25 mM ammonium bicarbonate in water, and then dehydrated once again with the dehydration buffer. The samples were then concentrated to dryness (speedvac) after which they were incubated with 250 ng trypsin for 30 min at 4°C, then digested overnight at 37°C with agitation. The digestions were extracted first with water and subsequently with 50% acetonitrile and 5% formic acid in water. After each step the supernatant was collected into a new tube, and pooled peptide extractions were concentrated for 2 h to dryness (speedvac), then re-suspended in 0.2% formic acid. The re-suspended samples were desalted with C18 Zip Tips, concentrated (speedvac) and re-suspended in a solution containing "Hyper Reaction Monitoring" (Biognosys) indexed retention time peptide standards (iRT) and 0.2% formic acid in water.

Mass spectrometry
Samples were analyzed by reverse-phase HPLC-ESI-MS/MS using an Eksigent Ultra Plus nano-LC 2D HPLC system (Dublin, CA) with a cHiPLC system (Eksigent) directly connected to a quadrupole time-offlight (QqTOF) TripleTOF 6600 mass spectrometer (SCIEX, Concord, CAN). After injection, peptide mixtures were loaded onto a C18 precolumn chip (200 µm × 0.4 mm ChromXP C18-CL chip, 3 µm, 120 Å, SCIEX) and washed at 2 µl/min for 10 min with the loading solvent (H 2 O/0.1% formic acid). Subsequently, peptides were transferred to the 75 µm × 15 cm ChromXP C18-CL chip, 3 µm, 120 Å, (SCIEX), and eluted at a flow rate of 300 nL/min with a 3 h gradient using aqueous and acetonitrile solvent buffers. Data-dependent acquisition (for spectral library building): The mass spectrometer was operated in data-dependent acquisition (DDA) mode, where the 30 most abundant precursor ions from the survey MS1 scan (250 ms) were isolated at 1 m/z resolution for collisioninduced dissociation tandem mass spectrometry (CID-MS/MS, 100 ms per MS/MS, 'high sensitivity' product ion scan mode) using the Analyst 1.7 (build 96) software with a total cycle time of 3.3 sec, as described 131 .
Data-independent acquisitions: For quantification, all peptide samples were analyzed by data-independent acquisition (DIA, e.g. SWATH) using 64 variable-width isolation windows 132,133 . The window width was adjusted according to the complexity of the typical MS1 ion current observed within a certain m/z range using a DIA 'variable window method' algorithm (more narrow windows were chosen in 'busy' m/z ranges, wide windows in m/z ranges with few eluting precursor ions). DIA acquisitions produce complex MS/MS spectra, which are a composite of all the analytes within each selected Q1 m/z window. The DIA cycle time of 3.2 s included a 250 ms precursor ion scan, followed by 45 ms accumulation time for each of the 64 variable SWATH segments.
Mass-spectrometric data processing, quantification, and bioinformatics MS data-dependent acquisitions (DDA) were analyzed using the database search engine ProteinPilot (SCIEX Beta 4.5, revision 1656) using the Paragon algorithm (4.5.0.0,1654) (specific results tables are uploaded to the data repository MassIVE). Using these results, a MS/MS spectral library was generated in Spectronaut (Biognosys). The DIA/SWATH data was processed for relative quantification comparing peptide peak areas from various different time points (Supplementary Data 1). For DIA/SWATH MS2 data sets, quantification was based on XICs of 6-10 MS/MS fragment ions, typically y-and b-ions, matching to specific peptides present in the spectral libraries. Differential protein abundance analysis was performed using Student's t-test, and p values were corrected for multiple testing, specifically applying group-wise testing corrections using the Storey method 134 . Significantly changed proteins were accepted at a 5% FDR (q value <0.05), also see Supplementary Data 2.

Western blotting
Conditioned media corresponding to the same number of cells were loaded for SDS-PAGE. Total cell and tissue lysates of mouse fibroblasts and intestine were prepared using RIPA lysis buffer (Thermo Fisher Scientific, #89900) with protease/phosphatase inhibitor (Thermo Fisher Scientific, #1861281). Protein concentration was measured with BCA protein assay kit (Thermo Fisher Scientific, #23227), and equal amount of proteins were loaded for SDS-PAGE. Proteins were transferred to PVDF membranes using iBlot 2 Dry Blotting System (Thermo Fisher Scientific). The membranes were blocked in 5% skim milk in Trisbuffered saline with 0.1% Tween20 (TBST) for 1 h and incubated in primary antibodies overnight at 4°C followed by secondary antibodies for 1 h at room temperature. Blots were developed using the Pierce ECL Plus Western Blotting Substrate (Thermo Fisher Scientific, #32132). The uncropped and unprocessed scans of the blots were provided in a Source data file.

Phage display selection of peptide ligands against FZD5-CRD-Fc and FZD8-CRD-Fc
Phage pools of Linear-lib 82,135 were cycled through rounds of binding selections with FZD5-CRD-Fc or FZD8-CRD-Fc following established protocols 136 . We included 10 μM Herceptin in all rounds to block binders to the Fc-domain. After four rounds of binding selection, individual phage clones were analyzed in a high-throughput phage ELISA using plate-immobilized FZD-CRD-Fc as target or off-target for specificity as described previously 136 . The binding signal of the same phage particle to BSA was detected as non-specific binding. Clones with phage-binding signal to target over 0.5 and signal/noise ratio > 10 were considered positive clones and subjected to DNA sequence analysis.
Generation of inhibitor peptides and assessment of specificity and affinity of dFz5-43 and dFz8-80 Several inhibitor peptide candidates were synthesized and, for those that contained single cysteines, oxidized to their disulfide bridged dimers. Peptides were screened by surface plasmon resonance (SPR) against immobilized FZD7-Fc, FZD5-Fc and FZD8-Fc to assess their specificity and affinity. dFz5-43, the best hit from the FZD5 campaign showed binding to FZD5 in the 50 nM range, but at least 15-fold higher affinity (~800 nM) to FZD8 and no detectable binding to FZD7. The top hits demonstrated exquisite selectivity to either FZD5/8 or FZD8, with affinities in the nM to pM range. dFz8-80, the best hit obtained in the FZD8 campaign showed \selective binding to only FZD8 and a dissociation constant of~600 pM.

Surface plasmon resonance
All measurements were done on a Biacore S200 instrument (GE Healthcare) with 10 mM Hepes pH7.2, 150 mM NaCl and 0.001% P20 as running buffer. Fc-tagged FZD target proteins (R&D Systems) were immobilized on sensor spots of a Protein A Sensor chip (GE Healthcare). A concentration series (3-fold dilution) across 6-dilutions of the indicated peptides were injected for single cycle kinetics. For Fab measurements, studies were performed with a CM5 chip and 50 mM Tris pH 8.0, 300 mM NaCl, 5% Glycerol and 0.05% Triton X as running buffer. Fabs were immobilized on chips using the Amine coupling kit (GE Healthcare, #BR100050) according to the manufacturer's instructions. A dilution series of the extracellular domain of recombinant mouse LRP5 (R&D Systems, #7344-LR) was injected. All data were double-referenced by subtracting the signal from a blank sensor spot and the signal from a buffer injection. Referenced data sets were fitted using the instrument's software. Peptides showing high specificity to either FZD5 and/or FZD8 were synthesized as disulfide bridged dimers (CSBio).

Live Ca2+ imaging of mouse intestinal organoids and data analysis
Organoids were live-imaged at identical-sized z stacks (2 micron intervals) using a 20X PlanFluor air objective attached to a 3i W Spinning Disc Confocal possessing a Zeiss AxioObserver M1 inverted microscope, Yokogawa W1 spinning disc unit and Photometrics Evolve EMCCD and Hamamatsu FLASH 4.0 sCMOS cameras. Z stacks were acquired every 10 s. Z-stack images were converted to mean intensity projections and, where required; background was subtracted using a rolling ball algorithm. Segmentation was conducted on a max-projection over-time of the GCaMP6f channel using Cellpose's pre-trained "Cytoplasm" classifier 137 . Mean intensity traces were measured in each ROI over the time course in both the GCaMP6f and TdTomato channels using scikit-image 138 . Ca 2+ oscillation patterns were calculated in an unsupervised, automated manner using MATLAB. Mean intensity traces were smoothed using a Savitzky-Golay filter and peaks were identified using the "findpeaks" algorithm from the MATLAB Signal Processing Toolbox. Graphs for GCaMP6f/tdTomato fluorescence ratio, peak numbers, peak height, peak width and baselined area under the curve were generated using Python matplotlib and seaborn packages. Statistical analysis was performed using Python SciPy and scikit-post-hoc packages.

RNAseq library preparation and sequencing
Total RNA was isolated using the RNeasy Plus Mini Kit (Qiagen, #74134) per manufacturer's instructions. Total RNA was quantified with Qubit RNA HS Assay Kit (Thermo Fisher Scientific, #Q32852) and quality was assessed using RNA ScreenTape on 4200 TapeStation (Agilent Technologies, #5067-5576). cDNA library was generated from 2 ng of total RNA using Smart-Seq V4 Ultra Low Input RNA Kit (Takara, #634894). 150 pg of cDNA was used to make sequencing libraries by Nextera XT DNA Sample Preparation Kit (Illumina, #FC-131-1024). Libraries were quantified with Qubit dsDNA HS Assay Kit (Thermo Fisher Scientific, #Q32851) and the average library size was determined using D1000 ScreenTape on 4200 TapeStation (Agilent Technologies, #5067-5582). Libraries were pooled and sequenced on NovaSeq 6000 (Illumina) to generate 30 millions single-end 50-base pair reads for each sample.

RNAseq analysis
Sequencing reads were filtered and aligned using HTSeqGenie v4.2.2. GSNAP v2013-11-01 was used for alignment, through the HTSeqGenie wrapper, against the mouse reference genome (NCBI Build 38). Reads within coding sequences were counted to determine an estimate of gene expression. Read counts were scaled by library size, quantile normalized and precision weights were calculated using the voom R package 139 . Differential expression analysis was performed using the limma package 140 . Differential gene expression was calculated using voom+limma with multiple-hypothesis correction of P values performed using the Benjamini-Hochberg method. Cutoff thresholds of fold change >1.5 and adjusted p value <0.05 were used. Transcription factor binding motif analysis was performed using RcisTarget 141 .

single-nucleus RNAseq (snRNA-seq)
Organoids were harvested, washed in PBS, pelleted, and frozen on dry ice for snRNA-seq. Nuclei were isolated following the protocol described previously 142 . Briefly, frozen samples were resuspended in TST buffer (146 mM NaCl, 10 mM Tris, 1 mM CaCl 2 , 21 mM MgCl 2 , 0.03% Tween20, and BSA) and pelleted by centrifugation. The nuclei pellets were resuspended in ST buffer (146 mM NaCl, 10 mM Tris, 1 mM CaCl 2 , 21 mM MgCl 2 ), passed through a 35 µm filter, and counted using a hemocytometer. Generation of gel Beads-in-emulsion, cDNA amplification, and library preparation were performed with Chromium Next GEM Single Cell 3ʹ Reagent Kits v3.1 (10x Genomics) following the manufacturer's instructions. 8000-10,000 nuclei were loaded per channel of a 10x Genomics chip.

single-nucleus RNAseq analysis
Reads were aligned to the mouse genome (mm10/Ensembl 98) using the 10x CellRanger 'count' function with the resulting matrix imported into the scanpy framework 143 . For each sample, we discarded cells falling more than 3 MAD (mean absolute deviation) above/below the population in any of three metrics: Fraction of mitochondrial reads (above), total number of reads (below), or total number of features detected (below). Library sizes per cell were normalized to a target sum of 1e4 using the 'normalize_total' function from scanpy followed by log-transformation. For projection and cluster identification, data were restricted to highly variable genes using the default parameters in scanpy's 'pp.highly_variable_genes' module and the data scaled to unit variance and zero mean after correcting data for differences in mitochondrial read counts and total counts using linear regression and followed by dimensionality reduction using PCA. Data were then integrated at the level of individual sample/library using Harmony Clusters were identified using the Leiden algorithm implemented in scanpy following calculation of the neighborhood graph of the observed cell data using the top 25 PCs and a local neighborhood of 10 for manifold approximation.UMAP and tSNE projections were subsequently calculated using default parameters 144 .
The above processing resulted in 14 clusters with no evidence of sample-level batch effects. These clusters were subsequently annotated by examining the expression of markers from major intestinal cell types (see Figure S1e). To further support this annotation, we calculated marker genes using the scoreMarkers function from the scran package in R as applied to the log-transformed, normalized count data 145 . To calculate gene scores for gene sets with many genes (e.g. in the case of the YAP target genes 99 ) we made use of the AddModuleScore function from Seurat 146 . The list of genes used for ISC/CBC signature -Lgr5, Olfm4, Ascl2, Axin2, Hmgb2, Smoc2, Lrig1. This function calculates the average expression level of the genes in a gene set normalized by the expression of an expression-level matched set of control genes 147 . All subsequent statistical analyses were carried out in R as described in the text.

Statistics and reproducibility
Data were presented as means ± SD, and the n value means biological replicates with individual values representing a mouse, unless otherwise indicated. Statistical analyses were performed using GraphPad Prism 9.0 (GraphPad Software, Inc., La Jolla, CA, USA). For comparisons between two independent groups, two-tailed unpaired Student's t test was used. Statistical significance of differences between multiple groups was determined using one-way ANOVA with Tukey's post-hoc multiple comparison tests. Experiments were independently repeated with similar results. For Figs. 1d, q, 2g, i, 3a, d, i and Supplementary Figs. 2a, d-e, 3b, and 7c the experiments were repeated twice. For Figs. 1a, 2f, 5c, 7b-f and Supplementary Fig. 5e, the experiments were repeated 3 times. For Fig. 7a and Supplementary Fig. 7a, the experiment was repeated 4 times. For Fig. 5b, the experiment was repeated five times.  Table 4) used in this study.

Reporting summary
Further information on research design is available in the Nature Portfolio Reporting Summary linked to this article.

Data availability
The bulkRNA-seq and snRNA-seq raw data generated in this study have been deposited in the GEO database under accession code GSE219128. The mass-spectrometric raw data are deposited at ProteomeXchange with the ID PXD028176. Other data generated in this study are provided in the Supplementary Information and Source Data file. Source data are provided with this paper.