If you don't remember your password, you can reset it by entering your email address and clicking the Reset Password button. You will then receive an email that contains a secure link for resetting your password
If the address matches a valid account an email will be sent to __email__ with instructions for resetting your password
Enteric Neuroscience Program and Department of Physiology and Biomedical Engineering, Mayo Clinic, Rochester, MinnesotaGastroenterology Research Unit, Division of Gastroenterology and Hepatology, Department of Medicine, Mayo Clinic, Rochester, Minnesota
Enteric Neuroscience Program and Department of Physiology and Biomedical Engineering, Mayo Clinic, Rochester, MinnesotaGastroenterology Research Unit, Division of Gastroenterology and Hepatology, Department of Medicine, Mayo Clinic, Rochester, Minnesota
Enteric Neuroscience Program and Department of Physiology and Biomedical Engineering, Mayo Clinic, Rochester, MinnesotaGastroenterology Research Unit, Division of Gastroenterology and Hepatology, Department of Medicine, Mayo Clinic, Rochester, Minnesota
Enteric Neuroscience Program and Department of Physiology and Biomedical Engineering, Mayo Clinic, Rochester, MinnesotaGastroenterology Research Unit, Division of Gastroenterology and Hepatology, Department of Medicine, Mayo Clinic, Rochester, Minnesota
Enteric Neuroscience Program and Department of Physiology and Biomedical Engineering, Mayo Clinic, Rochester, MinnesotaGastroenterology Research Unit, Division of Gastroenterology and Hepatology, Department of Medicine, Mayo Clinic, Rochester, Minnesota
Enteric Neuroscience Program and Department of Physiology and Biomedical Engineering, Mayo Clinic, Rochester, MinnesotaGastroenterology Research Unit, Division of Gastroenterology and Hepatology, Department of Medicine, Mayo Clinic, Rochester, Minnesota
Enteric Neuroscience Program and Department of Physiology and Biomedical Engineering, Mayo Clinic, Rochester, MinnesotaGastroenterology Research Unit, Division of Gastroenterology and Hepatology, Department of Medicine, Mayo Clinic, Rochester, Minnesota
Enteric Neuroscience Program and Department of Physiology and Biomedical Engineering, Mayo Clinic, Rochester, MinnesotaGastroenterology Research Unit, Division of Gastroenterology and Hepatology, Department of Medicine, Mayo Clinic, Rochester, Minnesota
Enteric Neuroscience Program and Department of Physiology and Biomedical Engineering, Mayo Clinic, Rochester, MinnesotaGastroenterology Research Unit, Division of Gastroenterology and Hepatology, Department of Medicine, Mayo Clinic, Rochester, Minnesota
Signal Transduction and Molecular Nutrition Laboratory, Kogod Aging Center and Department of Anesthesiology and Perioperative Medicine, Mayo Clinic, Rochester, Minnesota
Department of Internal Medicine, University of Texas Southwestern Medical Center, Dallas, TexasDivision of Anti-aging Medicine, Center for Molecular Medicine, Jichi Medical University, Shimotsuke, Tochigi, Japan
Signal Transduction and Molecular Nutrition Laboratory, Kogod Aging Center and Department of Anesthesiology and Perioperative Medicine, Mayo Clinic, Rochester, Minnesota
Enteric Neuroscience Program and Department of Physiology and Biomedical Engineering, Mayo Clinic, Rochester, MinnesotaGastroenterology Research Unit, Division of Gastroenterology and Hepatology, Department of Medicine, Mayo Clinic, Rochester, MinnesotaCenter for Individualized Medicine, Mayo Clinic, Rochester, Minnesota
Gastric dysfunction in the elderly may cause reduced food intake, frailty, and increased mortality. The pacemaker and neuromodulator cells interstitial cells of Cajal (ICC) decline with age in humans, and their loss contributes to gastric dysfunction in progeric klotho mice hypomorphic for the anti-aging Klotho protein. The mechanisms of ICC depletion remain unclear. Klotho attenuates Wnt (wingless-type MMTV integration site) signaling. Here, we examined whether unopposed Wnt signaling could underlie aging-associated ICC loss by up-regulating transformation related protein TRP53 in ICC stem cells (ICC-SC).
Methods
Mice aged 1–107 weeks, klotho mice, APCΔ468 mice with overactive Wnt signaling, mouse ICC-SC, and human gastric smooth muscles were studied by RNA sequencing, reverse transcription–polymerase chain reaction, immunoblots, immunofluorescence, histochemistry, flow cytometry, and methyltetrazolium, ethynyl/bromodeoxyuridine incorporation, and ex-vivo gastric compliance assays. Cells were manipulated pharmacologically and by gene overexpression and RNA interference.
Results
The klotho and aged mice showed similar ICC loss and impaired gastric compliance. ICC-SC decline preceded ICC depletion. Canonical Wnt signaling and TRP53 increased in gastric muscles of klotho and aged mice and middle-aged humans. Overstimulated canonical Wnt signaling increased DNA damage response and TRP53 and reduced ICC-SC self-renewal and gastric ICC. TRP53 induction persistently inhibited G1/S and G2/M cell cycle phase transitions without activating apoptosis, autophagy, cellular quiescence, or canonical markers/mediators of senescence. G1/S block reflected increased cyclin-dependent kinase inhibitor 1B and reduced cyclin D1 from reduced extracellular signal-regulated kinase activity.
Conclusions
Increased Wnt signaling causes age-related ICC loss by up-regulating TRP53, which induces persistent ICC-SC cell cycle arrest without up-regulating canonical senescence markers.
Aging-associated depletion of interstitial cells of Cajal arises from persistent cell cycle arrest of precursors occurring without an increase in canonical senescence markers or apoptosis. The blockade of precursor self-renewal is initiated by unopposed Wnt signaling and mediated by TRP53.
Aging is associated with a progressive decline in the functions of most organs including of the gastrointestinal tract.
reduced fundal compliance and accommodation leading to prolonged antral distention may contribute to early satiety and increased satiation in the elderly.
Indeed, body weight decreases steadily by an average of 0.5% per year when older than 65 years of age, reflecting a decline in food intake termed anorexia of aging.
In contrast, interstitial cells of Cajal (ICC), electrical pacemakers, mediators of nitrergic and cholinergic neuromuscular neurotransmission, and regulators of smooth muscle membrane potential and tone,
we previously reported a profound decrease in gastric ICC accompanying impaired fundal nitrergic inhibitory neuromuscular neurotransmission, which occurred without a reduction in neuronal nitric oxide synthase expression or enteric neuron numbers.
Therefore, ICC loss may be central to age-related gastric dysfunction.
Cellular senescence is an irreversible state of cell growth arrest induced by cellular stress and an important driver of aging and age-related diseases.
a multifunctional protein with well-established roles in DNA damage response (DDR), apoptosis, metabolism, autophagy, cell cycle inhibition/arrest, cellular senescence, aging, and cancer.
A similar mechanism may also affect ICC-SC. However, the function of Wnt signaling in the ICC lineage has not been characterized. Here, we investigated the hypothesis that aberrant activation of Wnt signaling leads to ICC depletion by triggering ICC-SC senescence via TRP53 up-regulation. Our findings in cultured ICC-SC, progeric klotho and naturally aged mice, in APCΔ468 mice with genetic up-regulation of canonical Wnt signaling,
and in human gastric tissues obtained from young and middle-aged donors identify a novel role for canonical Wnt signaling in ICC-SC proliferation and establish a link between overactive Wnt and TRP53 signaling and ICC-SC/ICC aging. Our data also reveal a role for TRP53-induced persistent cell cycle arrest occurring without apoptosis, autophagy, cellular quiescence, or the up-regulation of canonical mediators of senescence in aging-associated ICC-SC dysfunction.
Results
Aging-related Interstitial Cell of Cajal and Interstitial Cell of Cajal Stem Cell Decline Is Associated With Impaired Gastric Compliance
To establish the organ-level significance of these findings and extend their validity to naturally aged mice, we first measured gastric compliance ex vivo and determined ICC and ICC-SC frequencies and levels of v-kit Hardy-Zuckerman 4 feline sarcoma viral oncogene homolog (KIT) (stem cell factor receptor, a key ICC marker) protein by flow cytometry and Western immunoblotting (WB), respectively. Gastric compliance was reduced in both klotho and naturally aged mice (18–24 months old) vs age-matched wild-type (WT) and 4- to 8-week-old controls (Figure 1A), indicating impaired ability of the stomach to relax in response to filling. These changes were associated with a decrease in KIT protein and an increase in the DDR-associated histone modification γ-H2A.X (H2A.X phosphorylated at Ser139), a marker of aging
γ-H2A.X was also up-regulated, whereas KIT, anoctamin 1 (ANO1) (a calcium-activated chloride channel and functionally significant ICC marker), and ETS variant 1 (ETV1) (a transcription factor important for the development of most ICC including all gastric ICC classes
) were reduced in gastric corpus muscles of a 49-year-old male patient and a 51-year-old female patient compared with 19-year-old and 23-year-old sex-matched controls (Figure 1C). Next, we enumerated ICC and ICC-SC in the hematopoietic marker-negative fraction (to exclude KIT+ mast cells) in the gastric corpus + antrum of a mixture of C57BL/6 and BALB/c mice between 1 and 107 weeks of age using previously established and validated protocols.
ICC decreased sharply between 1 and 15 weeks of age and reached minimum at 58 weeks of age (Figure 1D). ICC-SC proportions declined more sharply, reaching a plateau before 15 weeks of age, and decreased further to a minimum at 107 weeks. Minima for both ICC and ICC-SC corresponded well to frequencies detected in klotho mice.
Thus, ICC-SC loss observed in klotho mice also occurs during natural aging and likely contributes to ICC depletion and its functional consequences. Our results also indicate that aging-associated changes in ICC can be identified in ∼50-year-old humans.
Figure 1Age-related ICC and ICC-SC decline is associated with impaired gastric compliance. (A) Reduced gastric compliance of intact stomachs excised from 3 50- to 70-day-old klotho and 4 18- to 24-month-old C57BL/6 mice relative to age-matched WT (n = 4) and 4- to 8-week-old controls (n = 4), respectively (average traces). Stomachs were infused with 1 mL Krebs solution
at 37°C at a rate of 0.1 mL/min while recording luminal pressure. P values are from Mann-Whitney rank sum tests. (B) Reduced KIT protein and increased γ-H2A.X (H2AXS139p) protein in the gastric tunica muscularis of klotho vs WT mice (n = 5–9) and old (18–24 months) vs young (4–8 weeks) mice (n = 7–9/group). GAPDH was used as a loading control. P values are from Mann-Whitney rank sum tests. (C) Down-regulated ICC-related proteins (KIT, ANO1, ETV1) and up-regulated γ-H2A.X in gastric corpus muscles of 49-year-old male patient and 51-year-old female patient vs 19-year-old male and 23-year-old female controls. (D) Gastric ICC (KIT+CD34− subpopulation; green) and ICC-SC (KITlowCD34+ subpopulation; red) frequencies assessed by flow cytometry in a mixture of male and female C57BL/6 (n = 31) and BALB/c (n = 20) mice between 1 and 107 weeks of age (6–14 mice/time point). Representative projections (left) and time course data (right) depicting age-associated reduction in ICC and ICC-SC frequencies are shown. Data points in black are ICC and ICC-SC frequencies in 7- to 10-week-old klotho mice shown for comparison.
The Canonical Wnt Signaling Pathway Is Enriched in the Interstitial Cell of Cajal Lineage and Overactivated in the Aging Gastric Tunica Muscularis
To determine the molecular mechanisms underlying aging-related ICC-SC/ICC depletion, we first analyzed Wnt signaling-related gene expression in transcriptome profiles obtained by RNA sequencing (RNA-seq) (Illumina mRNA- and total RNA-seq)
were studied (footnotes to Table 1, Table 2, Table 3, Table 4, Table 5). The MetaCore Signal transduction_WNT signaling network (Supplementary Datasets 1) and canonical and noncanonical Wnt pathway-related gene sets from the Molecular Signatures Database (MSigDB) 6.2
were significantly represented in the ICC lineage and specifically in all ICC-SC lines, whereas the Wnt calcium-modulating and polar cell planarity pathways were not enriched (Figure 2A, Tables 1 and 2, and Supplementary Table 1). GSEA also revealed a significant expression of stemness-related gene sets in the ICC-SC lines (Figure 2B, Table 3, and Supplementary Table 2). Immunohistochemistry indicated the presence of the key Wnt-induced transcription factor catenin beta 1 (CTNNB1) in KIT+ ICC and KIT− interstitial cells and enteric neurons in young WT mice (Figure 2C). These data provide evidence for the presence and activity of the canonical Wnt pathway in the ICC lineage.
Table 1Representation of the “Signal Transduction_WNT Signaling” Network (Metacore) in Mouse Gastric ICC-SC, Mouse Small Intestinal and Colonic ICC, and Human Gastric ICC
Cell lines: mouse gastric ICC-SC lines 2xSCS2F10 (C57BL/6) and 2xSCS70 (Immortomouse)21,22; FACS-purified native cells: colonic and jejunal ICC from Kit+/copGFP mice (background: 129S6Sv/Ev-C57BL/6J),38 BALB/c mouse small intestinal ICC associated with the deep muscular plexus (ICC-DMP) and the myenteric plexus (ICC-MY),39 and human gastric ICC.40
Analysis platforms included mRNA- and total RNA-sequencing (Illumina HiSeq 2000), as well as Affymetrix Mouse Genome 430 2.0 Arrays (MG430.2) and Affymetrix Human Genome U133 Plus 2.0 Arrays (HGU133+2).
RNA-seq data subjected to network analysis were normalized expression values (RPKM). Microarray data subjected to MetaCore network analysis were log2 fold changes (FC) and false discovery rates (FDR, Benjamini-Hochberg Q values) from differential gene expression analysis (DGE) vs unfractionated tunica muscularis source tissues as follows: 2xSCS2F10 cells (n = 3) vs C57BL/6J gastric corpus + antrum (n = 2), 2xSCS70 cells (n = 3) vs Immortomouse gastric corpus + antrum (n = 3), ICC-DMP (n =3 ) and ICC-MY (n = 3) vs BALB/c small intestines (n = 2), and human ICC (n = 6) vs human gastric corpus + antrum (n = 4).
RNA-seq data subjected to network analysis were normalized expression values (RPKM). Microarray data subjected to MetaCore network analysis were log2 fold changes (FC) and false discovery rates (FDR, Benjamini-Hochberg Q values) from differential gene expression analysis (DGE) vs unfractionated tunica muscularis source tissues as follows: 2xSCS2F10 cells (n = 3) vs C57BL/6J gastric corpus + antrum (n = 2), 2xSCS70 cells (n = 3) vs Immortomouse gastric corpus + antrum (n = 3), ICC-DMP (n =3 ) and ICC-MY (n = 3) vs BALB/c small intestines (n = 2), and human ICC (n = 6) vs human gastric corpus + antrum (n = 4).
Ranking of the Signal transduction_WNT signaling network among biological process networks detected by MetaCore analysis of gene sets meeting the specified cutoffs. See top 50 process networks for each cell type in Supplementary Datasets 1.
b Analysis platforms included mRNA- and total RNA-sequencing (Illumina HiSeq 2000), as well as Affymetrix Mouse Genome 430 2.0 Arrays (MG430.2) and Affymetrix Human Genome U133 Plus 2.0 Arrays (HGU133+2).
c,d RNA-seq data subjected to network analysis were normalized expression values (RPKM). Microarray data subjected to MetaCore network analysis were log2 fold changes (FC) and false discovery rates (FDR, Benjamini-Hochberg Q values) from differential gene expression analysis (DGE) vs unfractionated tunica muscularis source tissues as follows: 2xSCS2F10 cells (n = 3) vs C57BL/6J gastric corpus + antrum (n = 2), 2xSCS70 cells (n = 3) vs Immortomouse gastric corpus + antrum (n = 3), ICC-DMP (n =3 ) and ICC-MY (n = 3) vs BALB/c small intestines (n = 2), and human ICC (n = 6) vs human gastric corpus + antrum (n = 4).
d Ranking of the Signal transduction_WNT signaling network among biological process networks detected by MetaCore analysis of gene sets meeting the specified cutoffs. See top 50 process networks for each cell type in Supplementary Datasets 1.
e P values and FDR Q values from the network analysis.
f Ratio of genes meeting cutoff criteria and all genes in the Signal transduction_WNT signaling network.
NOTE. Cell lines: mouse (m) gastric corpus + antrum (ca) ICC-SC lines 2xSCS2F10 (C57BL/6), D2211B, and 2xSCS70 (Immortomouse). D2211B cells were treated with nutlin 3b (30 μmol/L, 72 hours) used as control for nutlin 3a (Table 4). Analysis platforms included mRNA-seq and total RNA-seq (RS) (Illumina HiSeq 2000 and 4000; GSE60854 and GSE139539). RNA-seq data subjected to GSEA
Preranked analysis were log2 normalized expression values (RPKM). Gene set matrix was assembled by searching the Molecular Signatures Database (MSigDB) 6.2
for “WNT AND FZD” without restrictions. Genes assigned to the indicated gene sets are listed in Supplementary Table 1. aThese gene sets were significantly enriched in the indicated cell lines and sorted cell populations (FDR Q <0.25). Only positively correlated gene sets are shown for clarity. Gene sets without data were rejected based on the basis of GSEA analysis criteria applied.
D2211B cells were treated with nutlin 3b (30 μmol/L, 72 hours) used as control for nutlin 3a (Table 4). Analysis platforms included mRNA-seq and total RNA-seq (RS) (Illumina HiSeq 2000 and 4000; GSE60854 and GSE139539). RNA-seq data subjected to GSEA
Preranked analysis were log2 normalized expression values (RPKM). Gene set matrix was assembled by searching the Molecular Signatures Database (MSigDB) 6.2
for “Pluripoten∗” without restrictions. Genes assigned to the indicated gene sets are listed in Supplementary Table 2. aThese gene sets were significantly enriched in the indicated cell lines and sorted cell populations (FDR Q <0.25). Only positively correlated gene sets are shown for clarity. Gene sets without data were rejected on the basis of the GSEA analysis criteria applied.
Table 4Changes in Stemness-, Senescence-, Autophagy-, and DREAM Complex-related Gene Sets in D2211B ICC-SC Treated With the MDM2 Antagonist Nutlin 3a (Control: Nutlin 3b) by Gene Set Enrichment Analysis
Matrix
Gene set
D2211B (m, ca, RS)
Nutlin 3a vs 3b
Size
NES
FDR Q
Stemness
CONRAD_GERMLINE_STEM_CELL
23b
-1.37b
0.1000b
CONRAD_STEM_CELL
KORKOLA_CORRELATED_WITH_POU5F1
16b
-1.35b
0.0974b
KORKOLA_EMBRYONAL_CARCINOMA_DN
KORKOLA_EMBRYONAL_CARCINOMA_UP
32b
-1.50b
0.0530b
KORKOLA_SEMINOMA_DN
KORKOLA_SEMINOMA_UP
33b
-1.59b
0.0305b
KORKOLA_TERATOMA_UP
KORKOLA_YOLK_SAC_TUMOR_UP
18b
-1.35b
0.0840b
LEE_NEURAL_CREST_STEM_CELL_DN
71
-1.00
0.4883
LEE_NEURAL_CREST_STEM_CELL_UP
78b
-1.74b
0.0031b
MIKKELSEN_PLURIPOTENT_STATE_DN
MIKKELSEN_PLURIPOTENT_STATE_UP
MUELLER_PLURINET
265
-1.49b
0.0468b
Senescence
BIOCARTA_TEL_PATHWAY
17
-1.18
0.3004
COURTOIS_SENESCENCE_TRIGGERS
DEMAGALHAES_AGING_DN
DEMAGALHAES_AGING_UP
36b
-1.36b
0.2360b
FRIDMAN_SENESCENCE_DN
FRIDMAN_SENESCENCE_UP
63b
-1.82b
0.0042b
GO_CELL_AGING
58
-0.85
0.7943
GO_MULTICELLULAR_ORGANISM_AGING
21
-0.78
0.8038
KAMMINGA_SENESCENCE_(DN)
29a
-1.27a
0.2484a
KEGG_P53_SIGNALING_PATHWAY
60b
-1.30b
0.2447b
KUMAMOTO_RESPONSE_TO_NUTLIN_3A_DN
KUMAMOTO_RESPONSE_TO_NUTLIN_3A_UP
ONGUSAHA_TP53_TARGETS
33a
2.59a
0.0000a
TANG_SENESCENCE_TP53(inhibition)_TARGETS_DN
48b
-1.73b
0.0088b
TANG_SENESCENCE_TP53(inhibition)_TARGETS_UP
17
-1.19
0.3416
Autophagy
GO_NEGATIVE_REGULATION_OF_AUTOPHAGY
40b
-1.50b
0.0655b
GO_POSITIVE_REGULATION_OF_AUTOPHAGY
67b
-1.42b
0.0692b
GO_REGULATION_OF_AUTOPHAGY
211b
-1.41b
0.0509b
KEGG_REGULATION_OF_AUTOPHAGY
19
1.05
0.3680
MIZUSHIMA_AUTOPHAGOSOME_FORMATION
18
-0.50
0.9933
DREAM
BIOCARTA_DREAM_PATHWAY
11
-0.74
0.8980
FISCHER_DREAM_TARGETS
850b
-1.47b
0.0363b
REICHERT_MITOSIS_LIN9_TARGETS
28b
-1.82b
0.0010b
NOTE. Mouse (m) gastric corpus + antrum (ca) ICC-SC from the line D2211B
were treated with nutlin 3a or its 150-fold less potent enantiomer nutlin 3b (30 μmol/L, 72 hours) used as control (n = 3/group). Total RNA-seq (RS) was performed on Illumina HiSeq 4000 platform (GSE139539). RNA-seq data subjected to GSEA
for the appropriate terms. Genes assigned to the indicated gene sets are listed in Supplementary Tables 2 and 3. a,bThese gene sets were significantly enriched (FDR Q <0.25), showing functional aup-regulation and bdown-regulation, respectively, of the pathway in nutlin 3a–treated cells. (For example, genes in the Kamminga_Senescence_(DN) set were down-regulated on serial passage of mouse embryonic fibroblast; thus their relative reduced expression in nutlin 3a–treated D2211B cells indicates up-regulation of senescence-related genes.) Gene sets without data were rejected on the basis of GSEA analysis criteria applied.
Table 5Changes in Apoptosis/Cell Death–related Gene Sets in D2211B ICC-SC Treated With the MDM2 Antagonist Nutlin 3a (Control: Nutlin 3b) by Gene Set Enrichment Analysis
were treated with nutlin 3a or its 150-fold less potent enantiomer nutlin 3b (30 μmol/L, 72 hours) used as control (n = 3/group). Total RNA-seq (RS) was performed on Illumina HiSeq 4000 platform (GSE139539). RNA-seq data subjected to GSEA
for “Apoptosis OR Apoptotic OR Death” (in the title or description). Genes assigned to the indicated gene sets are listed in Supplementary Table 4. a,bThese gene sets were significantly enriched (FDR Q <0.25), showing functional aup-regulation and bdown-regulation, respectively, of the pathway in nutlin 3a–treated cells. (For example, genes in the Yan_Escape_from_Anoikis set were down-regulated; thus their relative reduced expression in nutlin 3a–treated D2211B cells indicates functional up-regulation of anoikis-related genes). Gene sets without data were rejected on the basis of the GSEA analysis criteria applied.
Figure 2The canonical Wnt signaling pathway is enriched in the ICC lineage and overactivated in the aging gastric tunica muscularis. (A) Enrichment of canonical and noncanonical Wnt pathways but not the calcium-modulating or planar cell polarity pathways in 2xSCS2F10 and D2211B ICC-SC lines analyzed by mRNA-seq and GSEA. Vertical lines indicate genes ranked by RPKM values. Gene sets with FDR Q value <0.25 and positive NES were considered significantly enriched. See further data and gene sets analyzed in Table 2 and Supplementary Table 1. (B) Enrichment of stemness-related genes in 2xSCS2F10 and D2211B ICC-SC lines by GSEA. See further data and gene sets analyzed in Table 3 and Supplementary Table 2. (C) Immunoreactivity for CTNNB1 (red), KIT (green), and DAPI (blue) in 5-μm cryosections of gastric tunica muscularis tissues from young and old mice (n = 3/group). Scale bars, 25 μm. CTNNB1 was expressed in ICC (arrows), KIT− interstitial cells (arrowheads), and enteric neurons (asterisks) in both young and old mice. Note reduced KIT+ ICC in the old mouse. (D) CTNNB1 phosphorylation on Ser33/Ser37/Thr41 (P-CTNNB1) was reduced and nuclear CTNNB1 was increased in gastric corpus + antrum tissues of klotho mice vs WT controls and in old mice vs young controls (n = 4–8/group), indicating age-related overactivation of Wnt signaling. P values are from Mann-Whitney rank sum tests. (E) Up-regulated nuclear CTNNB1 protein in gastric corpus muscles of 49-year-old and 51-year-old patients vs 19-year-old and 23-year-old controls. CM, circular muscle; LM, longitudinal muscle.
To examine whether Wnt signaling increases in mouse and human gastric tunica muscularis with age, we analyzed levels of unstable and nuclear CTNNB1 proteins in the gastric tunica muscularis of klotho and naturally aged mice as well as in human gastric muscles. CTNNB1 phosphorylation on Ser33/Ser37/Thr41, which targets CTNNB1 for proteasomal degradation, rendering it unstable, was reduced in both klotho and naturally aged mice (Figure 2D), suggesting an activated state. Indeed, nuclear CTNNB1, a hallmark of active Wnt signaling, was increased in both klotho and aged mice (Figure 2D). We also detected similar increases in nuclear CTNNB1 in a 49-year-old patient vs a 19-year-old patient and a 51-year-old patient vs a 23-year-old patient (Figure 2E). Taken together, these findings indicate that Wnt/CTNNB1 signaling is activated in gastric tissues with increasing age.
Prolonged Overactivation of Canonical Wnt Signaling Causes Paradoxical Inhibition of Interstitial Cell of Cajal Stem Cell Proliferation
Overactive Wnt signaling from reduced Klotho levels can lead to cancer
To establish a mechanistic link between the observed age-related activation of Wnt/CTNNB1 signaling in gastric muscles and ICC-SC depletion, we cultured D2211B ICC-SC with 30 ng/mL mouse recombinant Wnt3a as an inducer of canonical Wnt signaling. Indeed, we found increased ICC-SC growth after 8 days of treatment but suppressed growth after 15 days by counting cell numbers and by immunostaining for Ki-67, a cell proliferation marker (Figure 3A and B). WB and immunofluorescence analysis showed that the up-regulation of CTNNB1 in response to 15-day exposure of ICC-SC to Wnt3a was also associated with increased levels of the DDR marker γ-H2A.X and the DDR response mediator protein TRP53 (Figure 3C). These findings indicate that prolonged exposure of ICC-SC to high concentration of a canonical Wnt ligand can induce some form of ICC-SC growth arrest after stimulation of proliferation.
Figure 3Prolonged overactivation of canonical Wnt signaling causes paradoxical inhibition of ICC-SC proliferation. (A) 50,000 D2211B ICC-SC were cultured in the presence or absence of 30 ng/mL mouse recombinant Wnt3a (n = 12/group). Cell counts were determined when Wnt3a-treated cells reached confluence at 8 days of culturing (P1). Then 50,000 cells were re-plated (12 cultures/group) and counted when the controls reached confluence at 15 days (P2). ICC-SC counts in the Wnt3a-treated cultures were reduced after an initial increase. P values are from Mann-Whitney rank sum tests. (B) Reduced proliferation of D2211B ICC-SC detected by Ki-67 immunofluorescence after 15-day stimulation with Wnt3a (30 ng/mL). Nuclei were counterstained with DAPI. P value is from Mann-Whitney rank sum test. (C) Fifteen-day exposure of D2211B cells to 30 ng/mL Wnt3a up-regulated CTNNB1 and the DDR-associated proteins TRP53 and γ-H2A.X by WB and γ-H2A.X by immunofluorescence (n = 6/group). P values are from Mann-Whitney rank sum tests. (D) Reduced EdU+ proliferating ICC-SC and ICC detected in the gastric tunica muscularis of klotho vs WT mice by flow cytometry (n = 5/group). P values are from Mann-Whitney rank sum tests. (E) Up-regulated CTNNB1 and reduced ICC-SC and ICC in gastric tunica muscularis of APCΔ468 vs WT mice (n = 6–8/group). P values are from Mann-Whitney rank sum tests.
To demonstrate reduced ICC-SC proliferation in the context of aging in vivo, we analyzed, after daily exposure for 2 weeks, the incorporation of the modified deoxyribonucleoside and DNA synthesis marker 5-ethynyl-2′-deoxyuridine (EdU) into ICC-SC and ICC of the gastric corpus + antrum of klotho and WT mice by flow cytometry. Indeed, the numbers of EdU+ ICC-SC that have undergone DNA replication at least once during the 2-week period of injections were reduced in klotho mice compared with WT mice (Figure 3D), indicating reduced ICC-SC proliferation with age. EdU+ ICC were less consistently reduced, suggesting that this population may have mainly contained cells that had been labeled as ICC-SC during the early days of the EdU treatment protocol before differentiating into ICC. Indeed, although the total numbers of EdU+ gastric ICC-SC and ICC were very similar, they represented 27% ± 12% (mean ± standard deviation) of all ICC-SC but only 3% ± 1% of all ICC (P = .008, Mann-Whitney rank sum test).
Next, to demonstrate a causal role for chronically elevated canonical Wnt signaling in age-related ICC loss in vivo, we enumerated ICC-SC and ICC in the gastric corpus + antrum APCΔ468 mice, which lack functional adenomatous polyposis coli (APC) protein, an essential component of the CTNNB1 destruction complex and an endogenous inhibitor of the Wnt-CTNNB1 pathway.
The proportions of ICC-SC and particularly differentiated ICC were reduced in the stomach of APCΔ468 mice with chronically elevated CTNNB1 levels (Figure 3E). These findings provide in vivo mechanistic evidence for canonical Wnt signaling causing aging-associated ICC-SC growth arrest leading to ICC decline.
Transformation Related Protein 53 Is Up-regulated in the Aging Gastric Tunica Muscularis and Inhibits Interstitial Cell of Cajal Stem Cell Growth
To establish a causal relationship between excess CTNNB1 and TRP53 up-regulation,
we first verified that overexpression of constitutively active Ctnnb1 increased TRP53 protein levels in 2xSCS2F10 ICC-SC (Figure 4A). Conversely, small interfering RNA (siRNA)-mediated Ctnnb1 knockdown down-regulated TRP53 protein levels in D2211B cells (Figure 4A). These results strongly support a role for CTNNB1 in regulating TRP53 in ICC-SC. By WB and immunohistochemistry, we found increased TRP53 protein levels in gastric tunica muscularis of both klotho and naturally aged mice and in the 49-year-old and 51-year-old patients vs their controls (Figure 4B and C). mRNA for sestrin 2 (Sesn2), sestrin 3 (Sesn3), and cyclin-dependent kinase 1a (Cdkn1a; also known as p21Waf1/Cip1), which are established transcriptional targets of TRP53,
were robustly increased in klotho mice and moderately in aged mice by real-time quantitative reverse transcription polymerase chain reaction (RT-qPCR) (Figure 4D). To investigate whether elevated TRP53 signaling could recapitulate Wnt/CTNNB1-induced ICC-SC loss, we exposed D2211B cells to 10–30 μmol/L nutlin 3a, an inhibitor of the mouse double minute 2 E3 ubiquitin-protein ligase (a negative regulator of TRP53), which belongs to a drug class currently in early clinical trials.
Forty-eight-hour nutlin 3a treatment of ICC-SC up-regulated TRP53 protein, Sesn2, Sesn3, and Cdkn1a mRNA, and γ-H2A.X protein levels (Figure 4E) and dose-dependently reduced ICC-SC growth (Figure 4F). These results indicate that TRP53 likely plays an important role in Wnt-induced, aging-associated ICC-SC and ICC depletion.
Figure 4TRP53 is up-regulated in the aging gastric tunica muscularis and inhibits ICC-SC growth. (A) Overexpression of constitutively active Ctnnb1 in 2XSCS2F10 ICC-SC by nucleofection increased TRP53 protein levels (n = 5/group). The siRNA-mediated knockdown of Ctnnb1 down-regulated TRP53 protein expression in D2211B cells (n = 8/group). Efficacy of RNA interference against Ctnnb1 was verified by WB (n = 8/group). P values are from Mann-Whitney rank sum tests. (B) Left panels: increased TRP53 protein in klotho mouse gastric lysates compared with WT controls and in old vs young mice (n = 8–9/group). Right panels: immunoreactivity for TRP53 (red) and nuclear DAPI (blue) in cryosections of gastric tissues from klotho and WT control (n = 3/group) and young and old mice (n = 3/group). Note increased TRP53 protein in both the tunica mucosa and the tunica muscularis. P values are from Mann-Whitney rank sum tests. (C) Up-regulated TRP53 protein in the gastric corpus tunica muscularis of 49-year-old and 51-year-old patients vs 19-year-old and 23-year-old patients. (D) Increased expression of TRP53 transcriptional target genes Sesn2, Sesn3, and Cdkn1a (n = 3–4/group) by RT-qPCR in gastric tissues of klotho mice compared with WT controls. More modest changes were seen in old mice compared with young controls (n = 3–4/group). P values are from t tests. (E) Treatment with nutlin 3a, inhibitor of the TRP53-degrading E3 ubiquitin ligase MDM2 (30 μmol/L), of D2211B cells increased TRP53 and γ-H2A.X protein levels by WB (72 hours; n = 4/group) and the expression of TRP53 target genes Sesn2, Sesn3, and Cdkn1a by RT-qPCR (48 hours) vs treatment with the 150-fold less potent enantiomer nutlin 3b (n = 5–6/group). P values are from Mann-Whitney rank sum tests. (F) Nutlin 3a dose-dependently reduced D2211B proliferation by MTS assay (n = 7–8/group). P values are from Kruskal-Wallis one-way ANOVA (ANOVA on ranks). Groups not sharing the same superscript are different by multiple comparisons (P < .05; Tukey tests).
Transformation Related Protein 53 Activation Inhibits Interstitial Cell of Cajal Stem Cell Growth by Reducing Cell Proliferation and Clonogenicity via Down-regulation of Self-renewal Genes and Cell Cycle Arrest
To investigate the spectrum of TRP53 effects that may underlie the observed inhibition of ICC-SC growth, we performed total RNA-seq in D2211B cells maintained under conditions nonpermissive for the expression of the tsA58-mutant SV40 large T antigen (tsTAg), which was undetectable by immunofluorescence and WB,
D2211B cultures were treated with 30 μmol/L nutlin 3a or nutlin 3b for 72 hours (n = 3/group). Differential enrichment of MSigDB 6.2 gene sets assembled into matrices related to TRP53 functions—senescence, autophagy, cell cycle arrest by the DREAM (dimerization partner, RB-like, E2F and multi-vulval class B) complex, and apoptosis/cell death
—was analyzed by GSEA (Tables 4 and 5, Supplementary Tables 3 and 4). GSEA verified the up-regulated expression of canonical TRP53 target genes in nutlin 3a–treated cells vs nutlin 3b–treated controls (Figure 5A). Autophagy-related gene sets were mainly down-regulated (3/5). Unexpectedly, most gene sets covering apoptosis/cell death (43/106), cellular senescence (4/10), and quiescence (DREAM complex targets; 2/3), which are typically up-regulated with age, were also down-regulated in nutlin 3a–treated ICC-SC. Up-regulated gene sets were restricted to 12/106 apoptosis/cell death-related and 2/10 senescence-related sets (including the set containing canonical TRP53 target genes). Consistent with GSEA analysis, WB indicated significantly reduced cleaved caspase 3 (CL.CASP3), a marker of apoptosis, and light chain 3B isoform II, a marker for autophagy, in nutlin 3a–treated D2211B cells (Figure 5B). Histochemical analysis of senescence-associated β-galactosidase (SA-β-gal) activity
in aged and young mouse stomachs only revealed increased cellular senescence in the gastric mucosa but not in the tunica muscularis of 18- to 24-month-old mice (Figure 5C). A similar, mucosa-restricted increase in SA-β-gal activity compared with WT mice was previously reported in klotho small intestines.
In aged mice, CL.CASP3 only showed a small increase relative to young animals by WB analysis (Figure 5D), and immunofluorescent microscopy only revealed increased CL.CASP3 levels in the gastric mucosa but not in the tunica muscularis (Figure 5E). Together, these findings suggest no significant involvement of apoptosis and autophagy or canonical markers/mediators of cellular senescence or quiescence in gastric ICC-SC depletion with age.
Figure 5Cell death–, apoptosis-, autophagy-, senescence-, and quiescence-related pathways are not stimulated by TRP53 up-regulation in ICC-SC. (A) GSEA of total RNA-seq data showing enrichment of TRP53 target genes and depletion of cell death–, apoptosis-, autophagy-, senescence-, and quiescence (DREAM complex)-related genes in D2211B ICC-SC treated with nutlin 3a vs cells treated with nutlin 3b (30 μmol/L, 72 hours; n = 3/group). Vertical lines indicate genes. Negative NES and FDA Q <0.25 indicate significant depletion. See all data and gene sets analyzed in Tables 4 and 5 and in Supplementary Tables 3 and 4. (B) Nutlin 3a (30 μmol/L, 72 hours) in D2211B cells reduced CL.CASP3 (n = 12/group) and light chain 3B isoform II (LC3B-II) levels vs nutlin 3b (n = 7/group). Staurosporin (3 μmol/L, 24 hours) was used as a positive control for apoptosis induction. P values are from Mann-Whitney rank sum tests. (C) Senescent cells detected by SA-β-Gal activity were increased in the gastric mucosa of old (18–24 months) vs young mice (4–8 weeks). No SA-β-gal activity was evident in the gastric muscle layers of either old or young mice. (D and E) Apoptosis detected in the gastric corpus + antrum tunica muscularis by CL.CASP3 immunoblotting (D) and in the gastric corpus + antrum (full thickness) by immunohistochemistry (E) in old (18–24 months) vs young mice (4–8 weeks). CL.CASP3 activity was low in the gastric muscle layers of both old and young mice and did not increase with age. Positive controls were exposed to staurosporin (3 μmol/L, 24 hours; n = 6/group). Data in D were analyzed by Mann-Whitney rank sum test.
However, the best definition of a senescent cell is an essentially permanent growth arrest, because molecular senescence markers are neither exclusive to this state nor universally applicable to different cell types.
Therefore, we next analyzed the effects of TRP53 induction on ICC-SC proliferation and the persistence of these effects in the absence of the initiating stimulus. GSEA of total RNA-seq data revealed depletion of stemness-related gene sets in nutlin 3a–treated D2211B cells (Figure 6A, Table 4, Supplementary Table 2). Kruppel-like factor 4 (KLF4) and myelocytomatosis oncogene (MYC), stemness genes consistently expressed by ICC-SC, were reduced by RT-qPCR and WB (Figure 6A). MetaCore Biological Process Network Analysis of differentially expressed genes (Supplementary Dataset 2) indicated a predominance of cell cycle-related gene networks (G1/S and G2/M; Supplementary Dataset 3). We also detected significant reduction in the incorporation of the halogenated deoxyribonucleoside 5-bromo-2-deoxyuridine (BrdU), Ki-67 immunolabeling, and the expression of proliferating cell nuclear antigen (a protein associated with the DNA replication fork), which was also down-regulated in both klotho and aged mice (Figure 6B). Cell cycle analysis by propidium iodide (PI) flow cytometry in D2211B cells treated with 30 μmol/L nutlin 3a revealed arrest in the G2/M phase with concomitant reduction of cells in G0/G1 and S phases (Figure 6C). In contrast, in response to 10 μmol/L nutlin 3a, a reduced proportion of cells in the S phase was the most prominent change. Collectively, these findings indicate that TRP53 activation reproduces Wnt-induced ICC-SC loss by down-regulating self-renewal genes and causing cell cycle arrest. To examine the persistence of these effects, we analyzed clonal growth in 2xSCS2F10 ICC-SC lentivirally transduced with the pGreenFire-p53-mCMV-EF1α-Puro Transcriptional Reporter and sorted singly into 96-well plates after 3 days of treatment with 10 μmol/L nutlin 3a or vehicle. After an additional 14 days of culturing in the absence of the drug, previously nutlin 3a–treated cells displaying high levels of TRP53-induced fluorescence showed significantly reduced clonogenicity as indicated by fewer positive wells and significantly lower level of methyltetrazolium salt (MTS) signal per well than vehicle-treated or nutlin 3a–treated cells with low or no reporter fluorescence (Figure 6D). Thus, even short-term (3-day) up-regulation of TRP53 increased the proportion of cells that lacked the ability to proliferate and also reduced the proliferative capacity of the cells that did not undergo cell cycle arrest likely because of some degree of stress relief.
Figure 6TRP53 activation inhibits ICC-SC growth by reducing cell proliferation and clonogenicity via down-regulation of self-renewal genes and cell cycle arrest. (A) Left panel: GSEA of total RNA-seq data showing down-regulation of stemness-related genes in D2211B ICC-SC treated with nutlin 3a vs cells treated with nutlin 3b (30 μmol/L, 72 hours; n = 3/group). Vertical lines indicate genes ranked by signal-to-noise ratios. Negative NES and FDR Q <0.25 indicate significant depletion in nutlin 3a–treated group. See all data and gene sets analyzed in Table 4 and Supplementary Table 2. Right panels: reduced mRNA and protein levels of stemness genes KLF4 and MYC. P values are from Mann-Whitney rank sum tests. (B) Nutlin 3a (30 μmol/L, 72 hours) reduced D2211B proliferation by BrdU incorporation (n = 12/group), Ki-67 immunofluorescence (n = 10/group), and PCNA protein expression (n = 8/group). PCNA expression was also reduced in klotho and old mice vs their respective controls (n = 6/group). P values are from Mann-Whitney rank sum tests. (C) Cell cycle arrest detected by PI flow cytometric assay in D2211B cells treated with 30 μmol/L or 10 μmol/L nutlin 3a for 72 hours (n = 3/group). P values are from t tests applied to arcsine square root transformed data. (D) Nutlin 3a–treated (10 μmol/L, 3 days), clonally sorted, TRP53high 2xSCS2F10 ICC-SC lentivirally transduced with the pGreenFire-p53-mCMV-EF1α-Puro Transcriptional Reporter displayed reduced clonogenicity and proliferation by MTS assay than vehicle-treated or nutlin 3a–treated, TRP53low cells (192 wells/group). In the left and right panels, P values are from Kruskal-Wallis one-way ANOVA (ANOVA on ranks). Groups not sharing the same superscript are different by multiple comparisons (left panel, P < .05, Tukey tests; right panel, P < .05, Dunn’s method). Proportions in the middle panel were analyzed by χ2 test. PCNA, proliferating cell nuclear antigen.
Transformation Related Protein 53 Inhibits Interstitial Cell of Cajal Stem Cell Proliferation via the ERK-CDKN1B-CCND1 Pathway
We next investigated the mechanisms that may underlie the TRP53-induced inhibition of S phase entry, because this mechanism, which is directly regulated by mitogen-activated protein kinases,
may be more conducive to future pharmacologic targeting than G2/M arrest. Progression through G1 and entry into S phase require the induction, in mid-G1 phase, of cyclin D1 (CCND1) by extracellular signal-regulated mitogen-activated protein kinases ERK1 and ERK2. Formation of the CCND1-CDK4/6 (cyclin-dependent kinase 4/6) complex results in the sequestration of cyclin-dependent kinase inhibitors CDKN1A and CDKN1B (p27Kip1), leading to the activation of cyclin E/A-CDK2 complexes and further cell cycle progression including S phase entry.
The predominantly transcriptionally controlled cyclin A/E-CDK2 inhibitor Cdkn1a was up-regulated by nutlin 3a in D2211B ICC-SC and klotho mice but increased only moderately in aged mice (Figure 4D). In nutlin 3a–treated D2211B ICC-SC, we also detected up-regulation of CDKN1B and down-regulation of CCND1, and phosphorylation of ERK1/2, a critical signaling mechanism for ICC survival and maintenance,
Platelet-derived growth factor receptor-alpha regulates proliferation of gastrointestinal stromal tumor cells with mutations in KIT by stabilizing ETV1.
was reduced (Figure 7A). The same changes were also evident in the gastric muscles of klotho vs WT and old vs young mice (Figure 7B), as well as in 49- and 51-year-old patients vs 19- and 23-year-old controls (Figure 7C). Treatment with selumetinib (72 hours), a selective inhibitor of ERK1/2 phosphorylation, dose-dependently inhibited ERK1/2 phosphorylation, reduced CCND1 protein, up-regulated CDKN1B protein, and reduced ICC-SC viability (Figure 8A and B). ERK1/2 may facilitate the proteasomal degradation of CDKN1B directly or indirectly via up-regulation of CCND1.
To investigate the contribution of CDKN1B to TRP53-mediated inhibition of ICC-SC viability, we performed siRNA-mediated knockdown of Cdkn1b in D2211B cells in the presence of 30 μmol/L nutlin 3a or nutlin 3b. These experiments showed a significant, albeit modest, mitigation of nutlin 3a–induced reduction in ICC-SC viability by Cdkn1b siRNAs (Figure 8C). These results indicate that inhibition of the ERK-CDKN1B-CCND1 pathway also contributes to Wnt/TRP53-mediated ICC-SC depletion in aging (Figure 8D).
Figure 7Increased CDKN1B, reduced CCND1, and reduced ERK activation in nutlin 3a–treated ICC-SC and during aging. (A) Nutlin 3a treatment (30 μmol/L, 72 hours) in D2211B cells up-regulated CDKN1B protein levels, down-regulated CCND1 protein levels, and reduced ERK1/2 phosphorylation by WB (n = 4–7/group). P values are from Mann-Whitney rank sum tests. (B) Increased CDKN1B protein, reduced CCND1 protein, and reduced ERK1/2 phosphorylation in klotho mouse gastric lysates compared with WT controls and in old vs young mice (n = 5–8/group). P values are from Mann-Whitney rank sum tests. (C) Up-regulated CDKN1B protein, reduced CCDN1 protein, and reduced ERK1/2 phosphorylation in the gastric corpus muscles of 49-year-old and 51-year-old patients vs 19-year-old and 23-year-old controls.
Figure 8TRP53 inhibits ICC-SC proliferation via the ERK-CDKN1B-CCND1 pathway. (A) The mitogen-activated protein kinase kinase 1/2 (MEK1/2) inhibitor selumetinib applied for 3 days dose-dependently reduced ERK phosphorylation, increased CDKN1B protein levels, and reduced CCND1 protein by WB (n = 7–8/group) in D2211B ICC-SC. P values are from Kruskal-Wallis one-way ANOVA (ANOVA on ranks). ∗P < .05 vs vehicle control by post hoc multiple comparisons (Dunn’s method). (B) Selumetinib also dose-dependently inhibited ICC-SC viability by MTS assay (n = 15/group). P value is from Kruskal-Wallis one-way ANOVA (ANOVA on ranks). ∗P < .05 vs vehicle control by post hoc multiple comparisons (Dunn’s method). (C) siRNA-mediated knockdown of Cdkn1b, verified by WB, mitigated the nutlin 3a–induced reduction in D2211B growth by MTS assay (n = 20/group). P values are from Kruskal-Wallis one-way ANOVA (ANOVA on ranks). Groups not sharing the same superscript are different by multiple comparisons (P < .05, Tukey tests). (D) Proposed mechanisms of ICC-SC depletion leading to ICC loss during aging. Aging causes overactive Wnt signaling in part by reducing Klotho protein levels. Increased Wnt signaling increases TRP53 protein levels by stabilizing CTNNB1 and promoting its nuclear localization and by inducing DDR. TRP53 causes ICC depletion in part by inhibiting ICC-SC proliferation via G2/M arrest. TRP53 also inhibits ERK1/2 phosphorylation (P), decreasing CCND1 and increasing CDKN1B protein levels. Reduced CCND1, likely together with other G1/S cyclins, down-regulates ICC-SC proliferation and self-renewal by interfering with S phase entry.
Here, we generalized these observations by demonstrating similarly reduced ICC and ICC-SC and impaired compliance in the stomach of aged and klotho mice and by showing that the decline of ICC-SC precedes that of ICC, reflecting reduced ICC-SC proliferation. Furthermore, we offer a mechanistic explanation for the age-related depletion of the ICC lineage (Figure 8D). Our results support a role for overactive canonical Wnt signaling, possibly arising from reduced levels of the Wnt chaperone protein Klotho,
and TRP53 in the inhibition of ICC-SC self-renewal. Up-regulated canonical Wnt signaling from genetic loss of the Wnt inhibitor protein APC led to gastric ICC depletion in mice, providing in vivo validation of our concept. Downstream of TRP53, we detected repression of stemness genes, G2/M arrest, and ERK inhibition causing G1/S transition block via increased CDKN1B protein levels and down-regulation of CCND1. Three-day up-regulation of TRP53 in cultured ICC-SC led to persistent growth arrest without activating apoptosis, autophagy, cellular quiescence, or, surprisingly, canonical markers/mediators of cellular senescence. We have confirmed our key findings by studying klotho and aged mice and gastric tunica muscularis tissues from middle-aged humans.
Reduced protein consumption, likely from lower overall food intake,
Consistent with a role for this age-related ICC loss in anorexia of aging, in klotho mice we previously reported net caloric deficit accompanying impaired nitrergic neuromuscular neurotransmission from ICC depletion,
and in KIT ligand-deficient Sl/Sld mice, primary loss of intramuscular ICC was associated with early satiation, decreased meal size, and reduced body weight.
ICC populations likely have a significant reserve because a more subtle—but, as our present data confirm, clearly detectable—decline in middle-aged individuals does not appear to result in a frank decrease in food intake. Whether the more modest ICC loss could still reduce ability to maintain nutritional status in response to metabolic challenge associated with various diseases remains unclear and will require further clinical studies.
In this study, we extended our previous finding of ICC loss-associated impaired nitrergic neuromuscular neurotransmission
to the organ level by demonstrating impaired ex vivo gastric compliance in both klotho and naturally aged mice. Reduced compliance has also been reported in excised guinea pig stomachs and in anesthetized rats after exposure to nitric oxide synthesis inhibitors,
supporting the notion that impaired compliance in our models reflected reduced nitrergic signaling. However, in isolated mouse stomachs, the same type of pharmacologic intervention appeared to increase compliance,
whereas we focused on local reflexes because our approach did not permit the assessment of vagal mechanisms. Because vagal intramuscular arrays, which are structures presumed to function as stretch receptors,
age-related ICC loss may also impair vagovagal reflexes that contribute to distention-induced relaxation. This intriguing possibility requires further investigation.
indicate that the main physiological consequence of aging-associated ICC loss is impaired gastric compliance due to reduced nitrergic neuromuscular neurotransmission. However, ICC also mediate cholinergic excitatory mechanisms.
Although it is possible that a more subtle decline in cholinergic excitation was masked by a more robustly impaired nitrergic relaxation, our previous electrophysiological and pharmacologic studies in the klotho stomach did not reveal a reduction in cholinergic responses,
found a significant change in either nitrergic or total neurons in the stomach of klotho mice and aged rats, respectively. Therefore, at the present time we are unable to explain why nitrergic inhibitory neuromuscular signaling is selectively affected in our models showing age-related ICC depletion.
Klotho protein levels are reduced during aging. Klotho mitigates/delays stem cell senescence by binding to Wnt family members.
we hypothesized that overactive Wnt signaling may also underlie age-related ICC loss, and TRP53, a major target of Wnt/CTNNB1 signaling in murine embryonic stem cells,
may mediate this effect. By using multiple, orthogonal approaches of transcriptome analysis, immunohistochemistry, Wnt3a treatments, overexpression of constitutively active CTNNB1, and studying mice deficient in the Wnt signaling inhibitor APC, we established a role for canonical Wnt signaling in the ICC lineage and showed that its activation can both stimulate and, after prolonged exposure, inhibit ICC-SC proliferation. Overactive canonical Wnt signaling was clearly demonstrable in the gastric musculature of klotho and aged mice and middle-aged humans and associated with DDR and TRP53 up-regulation. These effects could be reproduced by pharmacologic stabilization of TRP53 and linked to G2/M and ERK-CDKN1B-CCND1–mediated G1/S transition blockade occurring without activation of apoptosis, autophagy, or cellular quiescence pathways. Although TRP53 activation caused persistent cell cycle arrest, we found no convincing evidence of up-regulation of canonical markers/mediators of cellular senescence. At the present time, we cannot explain this finding. A limitation of our study is that the D2211B and 2xSCS70 cell lines were originally isolated from mice expressing a temperature-sensitive mutant SV40 tsTAg. However, both lines have been maintained in the verified absence of tsTAg,
the most widely used senescence marker, in the gastric tunica muscularis of klotho or aged mice. Therefore, considering that molecular senescence markers are neither exclusive to this state nor universally applicable to different cell types,
we propose that in the gastrointestinal tunica muscularis, persistent cell cycle arrest and other aspects of cellular senescence may be mediated by mechanisms different from those observed in the rapidly proliferating cells of the mucosa and other tissues. Identification of these mechanisms and associated biomarkers will require further studies.
Reduced proliferative capacity of ICC-SC can be expected to lead to diminished pools of both ICC-SC and ICC because the latter require constant replacement because of natural attrition.
However, ICC-SC and ICC frequencies ran approximately parallel courses during both the period of initial rapid decline of ICC, which by lineage tracing we previously attributed to a combined effect of ICC transdifferentiation and death,
and their subsequent stabilization. These results suggest that changes in the demand for ICC replacement may ultimately determine the rate of ICC-SC loss through the regulation of the speed of ICC-SC self-renewal and resultant exhaustion. This proposed mechanism is consistent with the effects of increased Wnt signaling on other stem cell types.
However, the signals communicating the size of the ICC pool to the ICC-SC remain to be identified. It is also important to note that ICC at some stages of maturity may also be able to proliferate.
Indeed, we counted very similar numbers of EdU+ cells among ICC-SC and ICC after daily administration of this labeled nucleoside for 2 weeks. However, EdU+ ICC-SC were more consistently reduced in klotho mice, and the EdU+ cells represented ∼9 times greater percentage of ICC-SC than ICC populations. These findings suggest that at least a part of the EdU+ ICC may have originally been labeled as ICC-SC during the early days of the EdU treatment regimen. Further studies using a pulse-chase paradigm
are needed to fully understand the dynamic aspects of ICC differentiation from their precursors.
In conclusion, our findings identify a novel role for canonical Wnt signaling in ICC-SC proliferation and establish a link between overactive Wnt signaling, TRP53 activation, and persistent cell cycle arrest in aging-associated ICC-SC and ICC depletion. Age-related ICC loss leads to gastric dysfunction predominantly by impairing gastric compliance through reducing nitrergic neuromuscular neurotransmission, as we demonstrated previously.
Age-related ICC-SC/ICC depletion could potentially be countered by the inhibition of the Wnt/CTNNB1 and/or stimulation of ERK-mediated signaling pathways.
Materials
Recombinant mouse Wnt3a was from Calbiochem (EMD Millipore, Billerica, MA). Nutlin 3a and nutlin 3b were from Cayman Chemical (Ann Arbor, MI). Dimethyl sulfoxide (DMSO), Staurosporin, and Triton-X were from Sigma-Aldrich (St Louis, MO). Selumetinib (AZD6244) was from Selleckchem (Houston, TX).
Animal Experiments
Homozygous klotho mice hypomorphic for α-Klotho and age-matched WT littermates (both sexes) were obtained from our heterozygous breeders, and their genotype was verified by PCR as reported previously.
Experiments were performed between 50 and 70 days of age. At 4–5 weeks of age, klotho and WT mice were intraperitoneally injected with EdU (50 mg/kg body weight) for 2 weeks to label proliferating cells. APCΔ468 mice with overactivated CTNNB1 signaling due to disruption of the Apc gene were developed by Khazaie and colleagues.
WT controls for genetically modified mice were co-housed with their mutant siblings. Male 18- to 24-month-old C57BL/6 mice were from the National Institute on Aging (Bethesda, MD). One- to 107-week-old WT mice of C57BL/6 or BALB/c background and of either sex were from Charles River Laboratories (Wilmington, MA), the Jackson Laboratory (Bar Harbor, ME), and our colonies. None of the mice were used in any previous experiments.
Mice were housed maximum 5/cage using an Allentown, Inc (Allentown, NJ) reusable static caging system in the Mayo Clinic Department of Comparative Medicine Guggenheim Vivarium under a 12-hour light/12-hour dark cycle. Bedding material was irradiated one-fourth-inch corn cob with the addition of Bed-r’Nest (4 g; The Andersons, Inc, Maumee, OH) irradiated paper-twist nesting material as enrichment. Mice were kept on irradiated PicoLab 5058 Mouse Diet 20 (≥20% protein, ≥9% fat, ≤4% fiber, ≤6.5% ash, ≤12% moisture; LabDiet, Inc, St Louis, MO). Food and water were available ad libitum. Before gastric compliance studies, mice were fasted overnight in a metabolic cage with free access to water. Animals were handled during the light phase.
Mice were killed by CO2 inhalation anesthesia or by decapitation performed under deep isoflurane (Baxter Healthcare, Deerfield, IL) inhalation anesthesia.
Tissues and Cell Lines
Gastric corpus + antrum muscles were prepared as described.
Pieces of human gastric corpus tunica muscularis were prepared by cutting away the mucosa and submucosa. Isolation and maintenance of the ICC-SC cell lines D2211B, 2xSCS70, and 2xSCS2F10 were described previously.
with minor modifications. Briefly, intact stomachs were excised, placed in a heated water bath, and connected via the esophagus to a syringe pump (Model 975 Compact Infusion Pump; Harvard Apparatus, Ltd, Cambridge, MA) and a pressure transducer (MP100A-CE; BIOPAC Systems, Inc, Goleta, CA; amplifier: Transbridge 4M; World Precision Instruments, Sarasota, FL) through the pylorus. The stomachs were then filled with Krebs solution
(37°C) to 1 mL at a rate of 100 μL/min while recording pressure using ClampFit 10.7.0 software (Molecular Devices, LLC, San Jose, CA).
Western Immunoblotting
Tissue and cell lysates were prepared and subjected to sodium dodecyl sulfate–polyacrylamide gel electrophoresis and immunoblotting as described previously
(antibodies in Table 6). Target and reference proteins were detected simultaneously by using LI-COR Biosciences (Lincoln, NE) secondary antibodies tagged with near-infrared and infrared fluorescent dyes (IRDye700, red pseudocolor; IRDye800CW, green pseudocolor). For nuclear β-catenin detection, nuclear faction was isolated according to manufacturer’s instruction (EMD Millipore).
(Tables 7 and 8). EdU+ proliferating in ICC-SC and ICC were detected by Click-iT Plus EdU Alexa Fluor 488 Flow Cytometry Assay Kit (Thermo Fisher Scientific, Waltham, MA) according to the manufacturer’s protocol (see details under Cell cycle analysis)
Table 7Antibodies Used for Flow Cytometry Analysis of Cells Freshly Dissociated From Murine Gastric Muscles
Transcriptome Analysis by Total Stranded RNA Sequencing
Total RNA was isolated from nutlin 3a– and nutlin 3b–treated D2211B cells and purified by using the Qiagen (Valencia, CA) RNeasy Mini Kit. Sequencing libraries were constructed by using 100 ng total RNA and the TruSeq Stranded Total RNA LT (with Ribo-Zero Human/Mouse/Rat) Set A kit (Illumina, San Diego, CA). Libraries were sequenced at 60–110 million fragment reads per sample following Illumina’s standard protocol using the Illumina cBot and HiSeq 3000/4000 PE Cluster Kit. The flow cells were sequenced as 101 × 2 paired-end reads on an Illumina HiSeq 4000 using HiSeq 3000/4000 SBS Kit (150 cycles) and HCS v 3.3.52 collection software. Base-calling was performed using Illumina’s RTA version 2.7.3. The transcriptome data have been deposited in a public database (National Center for Biotechnology Information, U.S. National Library of Medicine Gene Expression Omnibus (GEO), https://www.ncbi.nlm.nih.gov/gds/) as series GSE139539.
Briefly, the pipeline generates detailed quality control data to estimate the distance between paired-end reads, evaluate the sequencing depth for alternate splicing analysis, determine the rate of duplicate reads, and to evaluate coverage of reads across genes by using the RSeQC software.
Differentially expressed genes (Benjamini-Hochberg false discovery rate [FDR] Q <0.05) were analyzed for biological process networks by using MetaCore software. Biological process network analysis was also applied to the average reads per kilobase of transcript per million mapped reads (RPKM) (gene count) values obtained for the individual ICC classes and ICC-SC cell lines.
Previously published transcriptome data generated by using Affymetrix Mouse Genome 430 2.0 Arrays (MG430.2) (GEO gene sets GSE7809
Briefly, probe-level data were pre-processed by robust multiple-array analysis and analyzed for differential gene expression vs unfractionated tunica muscularis source tissues by the empirical Bayes approach with Benjamini-Hochberg adjustment using software packages in Bioconductor.
applied to the normalized gene expression values (RPKM). Average RPKM data from individual cell types or cell lines were subjected to GSEA Preranked analysis. The gene set matrices interrogated were assembled by searching the Molecular Signatures Database (MSigDB) 6.2
for the terms specified in the footnotes to Table 2, Table 3, Table 4, Table 5. Genes assigned to the indicated gene sets are listed in Supplementary Tables 1–4. Both the standard (differential) and Preranked GSEA analyses were performed by using default parameters. Table 2, Table 3, Table 4, Table 5 list the number of genes in the datasets that belong to the individual gene sets (size), the normalized enrichment score (NES), and the FDR Q value (cutoff for significance, 0.25).
Mouse gastric tissues were fixed with 4% paraformaldehyde for 2 hours at 4°C or with cold acetone (10 minutes). Five-μm cryosections were blocked with 1% bovine serum albumin (Sigma-Aldrich), incubated with primary antibodies (Table 9) at 4°C overnight, washed, and incubated with fluorochrome-tagged secondary antibodies (Table 9) at room temperature for 30 minutes. Nuclei were counterstained with 4′,6-diamidino-2-phenylindole (DAPI). Wide-field fluorescence images of 5-μm cryosections were captured with either a Nikon (Melville, NY) Eclipse TS-100F microscope equipped with a Modulation Optics (Glen Cove, NY) HMC ELWD Plan Fluor 40×, 0.60 NA air objective and a Jenoptik (Brighton, MI) MFCool CCD camera or an Olympus (Center Valley, PA) Magnafire camera mounted on an Olympus BX51 microscope equipped with a UPlanFl 40×, 0.75 NA air objective. Specificity of immunolabeling was verified by omitting the primary antibodies and by examining the samples with filter sets not designed for the fluorochrome used.
Table 9Antibodies Used in the Mouse Immunohistochemistry and Immunocytochemistry Studies
D2211B cells were plated onto no. 1 coverslips coated with rat-tail collagen and maintained with complete growth media. After Wnt3a or nutlin 3a treatment, the cells were washed, fixed with 4% paraformaldehyde for 10 minutes at room temperature, permeabilized with 0.3% Triton X-100 for 10 minutes at room temperature, and blocked with 1% bovine serum albumin in phosphate-buffered saline (pH 7.4) overnight at 4°C. After labeling with anti-Ki-67 or anti-γ-H2A.X antibodies (Table 9) at 4°C overnight, the cells were washed and incubated with fluorochrome-tagged secondary antibodies (Table 9) at room temperature for 30 minutes. Nuclei were counterstained with DAPI. Images were captured with a Nikon Eclipse TS-100F microscope equipped with a Modulation Optics 20× HMC ELWD Plan Fluor 0.45 NA air objective and a Jenoptik ProgRes MFCool CCD camera.
Beta-Catenin Overexpression
In 2xSCS2F10 ICC-SC, overexpression of a constitutively active β-catenin mutant lacking the glycogen synthase kinase 3 phosphorylation sites required for proteasomal degradation (Ctnnb1ΔGSK-KT3; Addgene plasmid #14717, a gift from Tannishtha Reya
) was performed using Cell Line Nucleofector Kit L (Lonza, Allendale, NJ). The 106 2xSCS2F10 cells were resuspended in 100 μL nucleofector solution L. The 100 μL cell suspension was combined with 2 μg plasmid or 2 μg pmaxGFP vector used as control. Program T-030 was used for the electrical settings. Expression was verified by WB.
RNA Interference
Cdkn1b or Ctnnb1 knockdown was performed by using Dhamacon ON-TARGETplus SMARTpool siRNA or corresponding scrambled sequences (25 nmol/L) and DharmaFECT1 Transfection Reagent (Thermo Fisher Scientific) according to the manufacturer’s protocol. Treatment was applied after 1-day culturing in antibiotic-free and antimycotic-free media. Knockdown efficacy was assessed after 72 hours by WB.
The cDNA was amplified on a Bio-Rad CFX96 (Bio-Rad Life Science Research, Hercules, CA) real-time PCR detector using the SYBR GreenER qPCR SuperMix (Thermo Fisher Scientific).
Three thousand cells per well were plated in complete media in 96-well flat-bottom plates. After 72 hours, cells were incubated as indicated. Viable cell counts and proliferation were evaluated by MTS assay (CellTiter 96 AQueous Non-Radioactive Cell Proliferation Assay; Promega, Madison, WI) and 5-bromo-2-deoxyuridine cell proliferation assay (Cell Signaling Technology, Danvers, MA), respectively.
Gastric tissues from naturally aged mice and young mice as well as klotho and age- and sex-matched WT mice were fixed with 4% PFA for 2 hours at 4°C. SA-β-gal activity was detected in 5-μm cryosections of the stomachs by using SA-β-gal staining Kit (Cell Signaling Technology) following the manufacturer’s protocol.
Cell Cycle Analysis
The Click-iT EdU Alexa Fluor (AF) 647 Flow Cytometry Assay Kit (Thermo Fisher Scientific) was used according to the manufacturer’s protocol with minor modifications. Briefly, EdU was added to cell culture medium to a final concentration of 10 μmol/L for 60 minutes. After a wash, cells were harvested, pelleted at 500g for 5 minutes, and fixed for 15 minutes at room temperature with Click-iT fixative containing 4% paraformaldehyde. After washing and centrifugation, the cells were permeabilized with Click-iT saponin-based permeabilization buffer and incubated with 500 μL Click-iT reaction cocktail containing the AF 647 fluorochrome for 30 minutes at room temperature in the dark. After a wash with the Click-iT permeabilization buffer, the cells were incubated with 20 mg/mL ribonuclease A and PI staining solution (50 μg/mL) for 45 minutes at room temperature in the dark. Samples were analyzed by using a Becton Dickinson LSR II flow cytometer (Table 8 for configuration) and FlowJo software (Treestar, Woodburn, OR). Chicken erythrocyte nuclei (used for checking instrument linearity) and calf thymocyte nuclei (BioSure, Grass Valley, CA) were used as controls.
Clonogenicity Assay
The 2xSCS2F10 ICC-SC were lentivirally transduced with the pGreenFire-p53-mCMV-EF1α-Puro Transcriptional Reporter purchased as a virus (System Biosciences, Palo Alto, CA). The 0.5 × 105 2xSCS2F10 cells were plated in a 24-well plate. The next day the media were replaced with 0.5 mL media containing 5 mg/mL Polybrene, and the cells were infected with the pseudovirus. On day 3 the media were changed to Polybrene-free media, and the cells were incubated overnight. On day 4 the cells were split, incubated for additional 48 hours, and then treated with nutlin 3a (10 μmol/L) or DMSO vehicle for 3 days. The cells were clonally plated on the basis of green fluorescent protein fluorescence reporting TRP53 activity (high vs low) by fluorescence-activated cell sorting. Cell viability was quantified by MTS assay after 14 days of culturing in the absence of nutlin 3a.
Statistical Analyses
Data were expressed as mean ± standard deviation or median and interquartile range with 5th and 95th percentiles shown as appropriate. Each graph also contains an overlaid scatter plot showing all independent observations. The “n” in the figure legends refers to these independent observations. Hypothesis testing was performed by nonparametric methods including the Mann-Whitney rank sum test and the Kruskal-Wallis one-way analysis of variance (ANOVA) on ranks followed by appropriate post hoc tests (Tukey’s test and Dunn’s method). The t tests were only applied to arcsine square root transformed proportional data and when the sample size equaled 3. Proportions were analyzed by the χ2 test. P < .05 was considered statistically significant. Methods used for the analysis of transcriptome data are described under “Transcriptome Data Analysis”.
Methods
All authors had access to the study data and reviewed and approved the final manuscript.
Regulatory Approvals
De-identified gastric corpus tissues were collected from nondiabetic patients undergoing bariatric surgery (Institutional Review Board protocol 13-008138). Animal experiments were performed in accordance with the National Institutes of Health Guide for the Care and Use of Laboratory Animals. The protocols were approved by the Mayo Clinic Institutional Animal Care and Use Committee (A64812, A48315). Protocols for work with biohazardous agents, recombinant DNA, and synthetic nucleic acids including gene expression plasmids, reporter lentivectors, and siRNA were approved by the Mayo Clinic Institutional Biosafety Committee (Bios00000076.01). Work with these agents was performed at biocontainment level BSL2 or BSL2þ according to standard operating procedures covering personal protective equipment use, decontamination, and waste handling.
Acknowledgments
The authors thank Merry J. Oursler, PhD (Robert and Arlene Kogod Center on Aging, Mayo Clinic, Rochester, MN) for providing stomachs from old mice. They also thank Pritha Chanana, MS (Division of Biostatistics and Informatics, Department of Health Sciences Research) for the initial bioinformatic analysis of RNA-sequencing data and Zhenqing Ye, PhD (Division of Biostatistics and Informatics, Department of Health Sciences Research) for his help with transcriptome data archiving.
Platelet-derived growth factor receptor-alpha regulates proliferation of gastrointestinal stromal tumor cells with mutations in KIT by stabilizing ETV1.
30117-X&id=fx1.jpg){kind=link}
30117-X&id=gr1.jpg){kind=link}
30117-X&id=gr2.jpg){kind=link}
30117-X&id=gr3.jpg){kind=link}
30117-X&id=gr4.jpg){kind=link}
30117-X&id=gr5.jpg){kind=link}
30117-X&id=gr6.jpg){kind=link}
30117-X&id=gr7.jpg){kind=link}
30117-X&id=gr8.jpg){kind=link}