Precise regulation of stem cell activity is crucial for tissue homeostasis and necessary to prevent overproliferation. In the Drosophila adult gut, high levels of reactive oxygen species (ROS) has been detected with different types of tissue damage, and oxidative stress has been shown to be both necessary and sufficient to trigger intestinal stem cell (ISC) proliferation. However, the connection between oxidative stress and mitogenic signals remains obscure. In a screen for genes required for ISC proliferation in response to oxidative stress, we identified two regulators of cytosolic Ca2+ levels, transient receptor potential A1 (TRPA1) and ryanodine receptor (RyR). Characterization of TRPA1 and RyR demonstrates that Ca2+ signaling is required for oxidative stress-induced activation of the Ras/MAPK pathway, which in turns drives ISC proliferation. Our findings provide a link between redox regulation and Ca2+ signaling and reveal a novel mechanism by which ISCs detect stress signals.https://doi.org/10.7554/eLife.22441.001
Multipotent intestinal stem cells (ISCs) are responsible for tissue homeostasis in the adult Drosophila midgut (Jiang and Edgar, 2011; Micchelli and Perrimon, 2006; Ohlstein and Spradling, 2006). Similar to the mammalian digestive epithelium, the activity of ISC proliferation and differentiation increases upon tissue damage (Amcheslavsky et al., 2009). Although signaling pathways controlling ISC activity, such as Ras/MAPK (Jiang et al., 2011), Hippo/Yki (Karpowicz et al., 2010; Shaw et al., 2010), and JAK/Stat (Jiang et al., 2009), play critical roles in gut homeostasis, the mechanisms modulating these pathways under different physiological and pathological conditions are not well understood. In particular, it is unclear how ISCs sense microenvironment signals under tissue damage conditions to upregulate the activity of mitogenic pathways such as Ras/MAPK. The simplicity of the Drosophila midgut and the ease of performing genetic screens provide a powerful system to study how stem cells sense microenvironment cues and adjust their activities accordingly.
The concentration of reactive oxygen species (ROS) is emerging as a critical microenvironment signal that regulates stem cell activity in neuronal and glial progenitors (Liu et al., 2009; Smith et al., 2000; Tsatmali et al., 2005), as well as in mammalian hematopoietic stem cells (HSCs) (Tothova et al., 2007) and Drosophila hematopoietic progenitors (Owusu-Ansah and Banerjee, 2009). In most cases, high ROS concentration stimulates, at least in the short term, the activity of stem cells by driving them out of an otherwise quiescent status. In mammals, oxidative stress is associated with intestine damage and inflammatory bowel disease (Rezaie et al., 2007). In the Drosophila midgut, high ROS levels are produced in response to different types of tissue damage (Ha et al., 2005; Hochmuth et al., 2011; Lee et al., 2013). Moreover, oxidants such as paraquat and hydrogen peroxide are commonly used to induce tissue damage (Chatterjee and Ip, 2009). Although high ROS levels in ISCs are sufficient to trigger ISC proliferation (Hochmuth et al., 2011), the connection between oxidative stress and mitogenic signals in ISCs remains obscure.
We performed an in vivo RNAi screen to identify novel components required for ISC proliferation in response to tissue damage. In the screen, we identified two cation channels, transient receptor potential A1 (TRPA1) and ryanodine receptor (RyR), which are required for ISC self-renewal and damage-induced proliferation. We show that these channels are required for Ca2+ homeostasis in ISCs and mediate increases in cytosolic Ca2+ in response to oxidative stress and ISC proliferation. Our results are consistent with a recent study (Deng et al., 2015) showing that high cytosolic Ca2+ in ISCs can stimulate proliferation. Finally, we characterize the signaling events downstream of cytosolic Ca2+ increase and demonstrate that Ras/MAPK mediates the effects of Ca2+ on ISC proliferation.
From an RNAi screen for new transmembrane proteins required for ISCs proliferation, we identified TRPA1 and RyR as candidate hits required for ISC pool size maintenance and damage-induced ISCs proliferation (see Materials and methods). We confirmed that the lines producing a consistent phenotype (v37249, BL31504 for TRPA1; BL29445 for RyR) effectively knockdown the corresponding transcripts, while the lines that did not score as hits from our screen (BL31384, BL36780 for TRPA1; BL31540, BL31695 for RyR) cannot achieve efficient knockdown (Figure 1—figure supplement 1C–D).
TRPA1 is a Ca2+-permeable cation influx channel (Wang et al., 2013). RyR is an endoplasmic reticulum (ER) cation channel that releases ER Ca2+ into the cytosol (Santulli and Marks, 2015). RNAi-mediated knockdown of either trpA1 or RyR in adult ISCs completely blocked proliferation induced by the tissue-damaging agent paraquat or bleomycin (Figure 1A–C and A’-C’, G). Depletion of trpA1 or RyR causes a gradual loss of the stem cell population, with a significant decrease in esg+ cell density undetectable until 7 days of RNAi expression (Figure 1—figure supplement 1E). The RNAi lines do not cause stem cell apoptosis, because cell death cannot be detected with anti-cleaved-caspase 3 staining in the midgut (Figure 1—figure supplement 2A–C), and the anti-apoptotic gene p35 (Hay et al., 1994) cannot rescue the mitosis defect caused by trpA1 RNAi or RyR RNAi (Figure 1—figure supplement 2D–I,D’–I’). In contrast, mitosis activity in stem cells expressing trpA1 RNAi or RyR RNAi can be regained by forced cytosolic calcium influx with RNAi against SERCA (Figure 1D–G), an ER Ca2+ ATPase whose inhibition triggers Stim/Orai channel (SOC)-mediated Ca2+ entry into the cytosol (Clapham, 2007). Altogether, these results place cytosolic calcium signaling genetically downstream of TRPA1 and RyR during ISC proliferation. Consistent with the results obtained with RNAi, CRISPR/Cas9-mediated knockout of trpA1 or RyR in adult ISCs suppressed damage-induced proliferation (Figure 1H). In addition, at the organ level, inhibition of ISC proliferation by trpA1 RNAi is associated with a shortening in midgut length (Figure 1I). Further, while wild-type midguts maintain their proper length after feeding paraquat or bleomycin for 2 days, midguts with trpA1 RNAi expression in the ISCs exhibit accelerated shortening following tissue damage. This damage-induced midgut shortening phenotype is reminiscent to the gut phenotype observed after depletion of the esg+ cell population when the cell death gene reaper (rpr) is expressed in ISCs for 4 days (Figure 1—figure supplement 3), suggesting the importance of ISC activity for maintaining tissue size.
To determine whether TRPA1 and RyR act in a cell-autonomous manner, we generated GFP+ clones originating from ISCs that are deficient for either trpA1 or RyR using the MARCM method (Wu and Luo, 2006). MARCM clones generated from ISCs expressing trpA1 RNAi (Figure 2A’), or homozygous for the loss-of-function allele trpA11 (Kwon et al., 2008) (Figure 2B’), or RyRk04913 (Sullivan et al., 2000) (Figure 2C’) had a smaller average size compared to wild-type clones 10 days after clonal induction (Figure 2A–C, trpA1 RNAi and RyRk04913 clone sizes are quantified in Figure 2D), suggesting that TRPA1 and RyR are required autonomously for clonal expansion. Although smaller than wild-type clones, MARCM clones expressing trpA1 RNAi or losing RyR can survive and differentiate into Pdm1+ enterocytes (ECs) (Figure 2—figure supplement 1A–D). We also performed lineage tracing with the EGT F/O system (Jiang and Edgar, 2009), which relies on induced expression of Flp in the stem cells to excise transcriptional STOP cassettes flanking act::Gal4, and found that stem cells expressing trpA1 RNAi can still differentiate into mature Pros+ enteroendocrine cells (EEs) or Pdm1+ ECs (Figure 2—figure supplement 1E–H). Furthermore, while Dl+ cells (Figure 2D–F and L) are dramatically decreased by trpA1 RNAi or RyR RNAi expression for 5 days, there was no significant change in the density of Su(H)Gbe+ or Pros+ cells (Figure 2G–I and M). The Notch pathway reporter Su(H)GbeGFP is indicative of ISC differentiation activity because Notch signaling is high when ISCs differentiate, especially for the EC lineage (Micchelli and Perrimon, 2006; Ohlstein and Spradling, 2007). Altogether, our data suggest that trpA1 RNAi or RyR RNAi mainly affects ISC self-renewal and proliferation in the adult gut, rather than ISC differentiation or cell death.
The trpA1 gene is expressed as multiple mRNAs (Figure 1—figure supplement 1A), four of which have been characterized previously (Zhong et al., 2012). Compared to the TRPA1-A and -B isoforms, expressed in the fly head, TRPA1-C and -D isoforms are expressed in the midgut (Figure 1—figure supplement 1B) and have different N-terminal residues that might confer sensitivity to different stimuli. Based on anti-TRPA1 staining, TRPA1 appears to be expressed in stem cells (Figure 3A–B). Using a Gal4 fused to the C terminal of trpA1 transcripts via CRISPR/Cas9-mediated knock-in at the endogenous locus (Figure 3—figure supplement 1), we could also detect trpA1 expression in stem cells in the anterior region of the posterior midgut (Figure 3C). Consistently, stem cell depletion reduces trpA1 total mRNA expression in the midgut by more than 50% (Figure 3—figure supplement 2A). Because the signal-to-noise ratio for anti-TRPA1 staining is low, and trpA1Gal4CP2A might not fully recapitulate the endogenous expression pattern, we cannot exclude the possibility that trpA1 is also expressed in other cell types. In a recent study, for example, TRPA1-C has been identified to regulate defecation of food-borne pathogens in the EEs (Du et al., 2016). However, TRPA1 is mainly required in ISCs for proliferation, because knockdown of trpA1 in self-renewing ISCs, rather than in differentiating ISCs (Su(H)Gbe+ cells, also called enteroblasts (EBs)) or visceral muscles, blocks damage-induced proliferation (Figure 3—figure supplement 2B). Further, an additional piece of evidence that TRPA1 protein functions in the stem cells comes from the ex vivo calcium-imaging experiments with GCamP6f reporter, for which the intensity of green fluorescence indicates cytosolic calcium levels (Chen et al., 2013). While the TRPA1 agonist AITC (allyl isothiocyanate (Bautista et al., 2005) can induce calcium influx in ISCs carrying a wild-type allele of trpA1, such response is abolished in flies losing both alleles of wild-type trpA1 (Figure 3E–F).
RyR appears to be ubiquitously expressed at low levels in the midgut epithelium, based on the expression pattern of RyRGal4R14G09 (Figure 3D), where Gal4 is under the control of 2.92 kb putative enhancer region sequence of RyR (Pfeiffer et al., 2008).
Oxidative stress has previously been observed in tissue damage conditions involving JNK activation (Hochmuth et al., 2011) and pathogen infection (Ha et al., 2005; Lee et al., 2013). The source of oxidative stress is generally attributed to an imbalance in the formation of ROS and impaired antioxidant machinery in unhealthy tissues (Hochmuth et al., 2011; Lih-Brody et al., 1996). Staining with dihydro-ethidium (DHE) (Owusu-Ansah et al., 2008) confirmed that ROS levels in the midgut increase following tissue damage induced by feeding flies with paraquat (Figure 4—figure supplement 1B), a widely used oxidant that interferes with electron transfer to catalyze formation of superoxide-free radicals (Bus and Gibson, 1984), or bleomycin (Figure 4—figure supplement 1C), a chemotherapy drug that cleaves DNA via an unresolved mechanism potentially related to ROS production (Hecht, 2000). In addition, induction of midgut stress by JNK pathway activation or massive EC apoptosis by rpr expression can also lead to increased ROS and ISC expansion (Figure 4—figure supplement 1F–J).
Both TRPA1 and RyR are associated with redox sensing (Guntur et al., 2015; Hamilton and Reid, 2000; Sun et al., 2011). Guntur et al. found that TRPA1-C and TRPA1-D isoforms are H2O2-sensitive and that their endogenous expressions in the ring gland, or ectopic expression in other tissues, confer UV light sensitivity via sensing UV light-induced H2O2 production. To directly measure whether gut-specific TRPA1 isoforms can respond to paraquat, we used two-electrode voltage clamp recordings to measure the channel activities of TRPA1-C and -D (Wagner et al., 2000). In this experiment, we expressed similar amount of mRNA for TRPA1-C or -D isoform in the Xenopus oocytes, clamped the membrane potential at −80 mV, and measured the current flowing through the ion channels in the oocyte membrane. The current value is proportional to the opening probability of the ion channels. We added paraquat and AITC (used as the positive control) into the bath solution of oocytes sequentially to trigger the opening of the TRPA1-C or -D channel and compare their strength and dynamics. TRPA1-C barely displayed any response to paraquat (Figure 4A), whereas expression of TRPA1-D induced a slow increase in current (Figure 4B). Consistently, overexpression of TRPA1-D, but not TRPA1-C, stimulated ISCs expressing trpA1 RNAi to proliferate in response to paraquat (Figure 3C–F).
Previously, it has been reported that knockdown of antioxidant protein CncC in ISCs can induce mitosis (Hochmuth et al., 2011). Our observation that trpA1 RNAi suppresses the hyper-proliferation status induced by CncC RNAi (Figure 4G) places TRPA1 genetically downstream of the pathway controlling ROS for ISC proliferation. Further, overexpression of TRPA1-D or conventional TRPA1-A isoform (active at 29°C) in stem cells can induce mitosis significantly, suggesting that increased TRPA1-calcium signaling is sufficient to drive ISC proliferation (Figure 4H).
Interestingly, TRPA1 and RyR are not critically required in the midgut epithelium for development, because flies expressing trpA1 RNAi or RyR RNAi with esgGal4 throughout development [that is, in adult midgut progenitors (AMPs) during larval stages and in adults ISCs, see (Jiang and Edgar, 2009) can survive to adulthood without gross defects when they are fed with normal food]. For young adult flies of these genotypes, we used the ultrasensitive GCaMP6s fluorescent reporter (Chen et al., 2013) to examine cytosolic Ca2+ concentrations in their ISCs. While the number of GFP+ (i.e. high cytosolic Ca2+ concentration) ISCs in wild-type midguts increases significantly 10 min after adding paraquat to a final concentration of 4 mM (Figure 5A, A’ and D), expression of trpA1 RNAi or RyR RNAi in ISCs (driven by esgGal4 throughout development) not only reduced the basal levels of cytosolic Ca2+ in stem cells (Figure 5A–C) but also inhibited the ROS-mediated increase in cytosolic Ca2+ (Figure 5A’–C’). The observed reduction of high Ca2+ ISCs is not due to decrease in the total number of ISCs, as total ISC number or density in young adult flies is not yet significantly affected by trpA1 RNAi or RyR RNAi (Figure 5—figure supplement 1).
Defining the average number of high Ca2+ stem cells before drug treatment as a baseline, we used the confocal time-lapse GCaMP6s imaging (Videos 1–6) and analysis to identify the kinetics of high Ca2+ in stem cells. Consistent with the TRPA1 channel kinetics as measured in the voltage clamp experiments in oocytes (Figure 4A–B), treatment with 4 mM paraquat caused a slow and continuous increase in the average number of high Ca2+ ISCs (Figure 5E), whereas the TRPA1 agonist AITC triggered a sharp increase followed by a quick drop in the number of high Ca2+ ISCs (Figure 5F). Finally, responses to both paraquat and AITC were dependent on TRPA1 and RyR (Figure 5E–F), whereas treatment with the SERCA inhibitor thapsigargin, as a positive control, led to increased Ca2+ in ISCs expressing trpA1 RNAi or RyR RNAi (Figure 5G).
Using GCaMP5G as the Ca2+ reporter and tdTomato as an internal control, we found that trpA1 RNAi blocked the increase of cytosolic Ca2+ following treatment with 2 mM paraquat (Figure 5—figure supplement 2A,A’, B,B’). Similar results were observed with another reporter GCaMP6f (data not shown).
In addition, we used TRIC (transcriptional reporter of intracellular Ca2+) (Gao et al., 2015) to examine cytosolic Ca2+ concentration in ISCs in vivo on fixed midguts. With TRIC lexGFP and a conventional UAS-RFP reporter, we found that the percentage of high Ca2+ ISCs among all stem cells increased from 23% to 63% (Figure 5—figure supplement 2C,C’, F) when flies were fed 2 mM paraquat for 1 day. In adult midguts expressing trpA1 RNAi or RyR RNAi in adult ISCs for 6 days, high Ca2+ ISCs showed little, if any, increase following paraquat feeding (Figure 5—figure supplement 2D,D’, E,E’, F). Unlike GCamP reporters, a reduction in basal levels of high Ca2+ ISCs by trpA1 RNAi or RyR RNAi was not observed with TRIC reporter, which might be due to differences in the stability and detection threshold of reporters as well as the time frame of knockdown for the two sets of experiments.
The biological functions of TRPA1, RyR, and cytosolic Ca2+ signaling have been extensively studied in neurons, where it has been established that membrane depolarization and Ca2+ influx can activate Ras/MAPK (Randlett et al., 2015a; Rosen et al., 1994). Thus, we next asked whether high Ca2+ in ISCs similarly activates Ras/MAPK. To do this, we stained midguts with antibodies against diphospho-extracellular signal-regulated kinase (dpErk), which is a measure of Drosophila Ras/MAPK activity (Gabay et al., 1997). Adult midguts expressing SERCA RNAi in ISCs for 2 days, which at that stage do not present signs of ISCs expansion, showed induced MAPK activity (Figure 6A–B), and expression of trpA1 RNAi (Figure 6C) or RyR RNAi (Figure 6D) decreased the levels of MAPK activity. Consistent with our observation that ISC Ca2+ induction by SERCA inhibition does not depend on trpA1 or RyR (Figure 5G), SERCA RNAi could override trpA1 RNAi (Figure 6G) or RyR RNAi (Figure 6H) to activate MAPK. To test whether the increased dpErk signal we observed with SERCA RNAi merely reflects an hyper-proliferative status for ISCs, we combined SERCA RNAi with knockdown of Yorkie (Yki), the Hippo pathway transcriptional factor required for damage-induced ISC proliferation (Karpowicz et al., 2010; Ren et al., 2010; Shaw et al., 2010; Staley and Irvine, 2010). Although Yki RNAi could inhibit ISC proliferation caused by SERCA RNAi (Figure 7—figure supplement 1F–G), it could not block MAPK activity (Figure 6F). On the other hand, Ras1 (Drosophila Ras) RNAi could block SERCA RNAi-induced MAPK activity (Figure 6E). Altogether, our data suggest that Ras/MAPK activity is likely the direct target of high cytosolic Ca2+ in ISCs, rather than the consequence of cross-activation by other proliferation-related pathways such as Yki.
To further examine the relationship between Ca2+ levels and MAPK activity, we performed dpErk staining of midguts expressing the TRIC reporter in ISCs. Paraquat feeding for 1 day led to an increase in the number of cells with high dpErk signals, a majority of which are high Ca2+ ISCs (Figure 6I,I’). Moreover, paraquat induction of MAPK activity appears to require high Ca2+ in ISCs, because such response could be inhibited by trpA1 RNAi expression in ISCs (Figure 6J,J’). It should be noted that the best timing to detect autonomous activation of dpErk by high Ca2+ in the ISCs is before they enter the hyper-proliferative stage. Once ISCs start to expand massively following prolonged high Ca2+ induction, pErk returns to normal levels in many ISCs and is non-autonomously induced in some ECs, resulting in a diffusive activation pattern (Figure 6—figure supplement 1A–C).
In neurons, a common mechanism by which Ca2+ regulates Ras/MAPK is through Ca2+-sensitive Src kinases (Cullen and Lockyer, 2002), in a process faster than any transcriptional response (Randlett et al., 2015b; Rosen et al., 1994). Interestingly, overexpression of Src (Src42A or Src64B in Drosophila) in ISCs triggers ectopic MAPK activation and proliferation (Figure 6—figure supplement 2A–C), even in the presence of trpA1 RNAi (Figure 6—figure supplement 2E–H). Moreover, we could observe significant inhibition of SERCA RNAi-induced ISC proliferation by Src64B RNAi (Figure 6—figure supplement 2I) despite the likelihood of functional redundancy between Src64B and Src42A (Ma et al., 2013; Tateno et al., 2000). These data suggest that Src might mediate the activation of Ras/MAPK by cytosolic Ca2+ in ISCs.
As a universal intracellular signal, increased Ca2+ might trigger a wide range of responses. Consistent with previous results (Deng et al., 2015; Jiang et al., 2011), ISCs expressing the active forms of many Ca2+ effector proteins, such as Ras1, Raf, phosphatase Calcineurin A1 (CanA1), CREB-regulated transcriptional co-activator (CRTC), and cyclic-AMP response element binding protein B (CrebB), displayed an hyper-proliferation phenotype (Figure 7A–E). To determine which Ca2+ effector proteins are sufficient for Ca2+-induced ISC proliferation, we expressed trpA1 RNAi in ISCs together with each putative Ca2+ effector protein. Strikingly, MAPK activation via expression of active Ras1 (Ras1A), constitutively active Raf (Rafgof), or secreted form of spi is epistatic to trpA1 RNAi for ISCs proliferation (Figure 7A–B and A’–B’, 7F). Consistently, MAPK activation is epistatic to RyR RNAi (Figure 7—figure supplement 3). By contrast, trpA1 RNAi effectively blocked the mitogenic activity of constitutively active CanA1 (CanA1ca) (Dijkers and O'Farrell, 2007), CRTC (Kim et al., 2016), or active CrebB (CrebBact) (Ganguly-Fitzgerald et al., 2006) (Figure 7C–E and C’–E’, F). Furthermore, while Ras1 RNAi could suppress SERCA RNAi-induced ISC proliferation, neither CanA1 RNAi nor CrebB RNAi could (Figure 7—figure supplement 1C–G), and CRTC RNAi could only confer moderate suppression of ISC hyper-proliferation (Figure 7—figure supplement 1E–F). Results of these genetic epistasis analyses, together with examination of changes in pErk signal intensity (Figure 6), indicate that MAPK activity is both necessary and sufficient for high Ca2+-triggered ISC proliferation.
Previous studies have documented that receptor tyrosine kinase (RTK) EGFR is required for MAPK activity and ISC proliferation (Jiang et al., 2011). We found that EGFR RNAi completely inhibited ISCs proliferation caused by SERCA RNAi (Figure 7—figure supplement 1B,F), indicating that MAPK activation by Ca2+ requires EGFR signaling. Interestingly, as transcriptional regulation of ligands and receptors by their own pathways is commonly observed (see for example Ammeux et al., 2016), we examined by RT-qPCR the expression of EGFR ligands, as well as other RTK ligands, in midguts with cytosolic Ca2+ levels altered in ISCs. Interestingly, transcription of the EGFR ligand spi and the Pvr ligand pvf1 (Bond and Foley, 2012) were suppressed by trpA1 RNAi and up-regulated in midguts expressing SERCA RNAi in ISCs for 1 day to mimic early stages of Ca2+ induction (Figure 7—figure supplement 2). Altogether, these results suggest that positive RTK signaling feedback loops may play a role in the activation of Ras/MAPK in the context of Ca2+ signaling.
We have found that the two cation channels TRPA1 and RyR are critical for cytosolic Ca2+ signaling and ISC proliferation. Under homeostatic conditions, the basal activities of TRPA1 and RyR are required for maintaining cytosolic Ca2+ in ISCs to ensure their self-renewal activities and normal tissue turnover (Figure 8A). Agonists, including but not limited to low levels of ROS, could be responsible for the basal activities of TRPA1 and RyR. Under tissue damage conditions, increased ROS stimulates the channel activities of TRPA1 to mediate increases in cytosolic Ca2+ in ISCs (Figure 8B). As for RyR, besides its potential to directly sense ROS, it is known to act synergistically with TRPA1 in a positive feedback mechanism to release more Ca2+ from the ER into the cytosol upon sensing the initial Ca2+ influx through TRPA1 (Pan et al., 2016).
Our findings on the role of Ca2+ signaling for ISC proliferation are consistent with a recent study showing that high levels of cytosolic Ca2+ correlate with ISC proliferation, and that a prolonged increase in cytosolic Ca2+ concentration is sufficient to trigger ISC proliferation (Deng et al., 2015). Here, we demonstrate that TRPA1 senses damage-induced ROS; that knockdown of trpA1 or RyR efficiently reduces Ca2+ induction in the ISCs caused by ROS; and that cytosolic Ca2+ induction is required for damage-induced ISC proliferation.
Previously, Deng et al. identified L-glutamate as a signal that can activate metabotropic glutamate receptor (mGluR) in ISCs, which in turn modulates the cytosolic Ca2+ oscillation pattern via phospholipase C (PLC) and inositol-1,4,5-trisphosphate (InsP3). Interestingly, L-glutamate and mGluR RNAi mainly affected the frequency of Ca2+ oscillation in ISCs, while their influence on cytosolic Ca2+ concentration was very weak. Strikingly, the number of mitotic cells induced by L-glutamate (i.e. an increase from a basal level of ~5 per midgut to ~10 per midgut) is far less than what has been observed in tissue damage conditions (depending on the severity of damage, the number varies from ~20 to more than 100 per midgut following damage) (Amcheslavsky et al., 2009; Chatterjee and Ip, 2009; Jiang et al., 2011; Karpowicz et al., 2010). Consistent with this, in our screen for regulators of ISC proliferation in response to tissue damage, we tested three RNAi lines targeting mGluR (BL25938, BL32872, and BL41668, which was used by Deng et al.), and none blocked the damage response in ISCs, suggesting that L-glutamate and mGluR do not play a major role in damage repair of the gut epithelium.
We find that ROS can trigger Ca2+ increases through the redox- sensitive cation channels TRPA1 and RyR under damage conditions. In particular, we demonstrate using voltage-clamp experiments that the TRPA1-D isoform, which is expressed in the midgut, is sensitive to the oxidant agent paraquat. In addition, the results of previous studies have demonstrated the direct response of RyR to oxidants via single channel recording (Terentyev et al., 2008) and showed that RyR could amplify TRPA1-mediated Ca2+ signaling through the Ca2+-induced Ca2+ release (CICR) mechanism. Interestingly, expression of oxidant- insensitive TRPA1-C isoform in the ISCs also exhibits a tendency to induce ISC proliferation (Figure 4H), indicating that ROS may not be the only stimuli for TRPA1 and RyR under physiological conditions. Possible other activators in the midgut may be irritant chemicals, noxious thermal/mechanical stimuli, or G-protein-coupled receptors (Fowler and Montell, 2013; Shen et al., 2011; Zhong et al., 2012).
Altogether, the concentration of cytosolic Ca2+ in ISCs appears to be regulated by a number of mechanisms/inputs including mGluR and the ion channels TRPA1 and RyR. Although mGluR might make a moderate contribution to cytosolic Ca2+ in ISCs, TRPA1 and RyR have a much stronger influence on ISC Ca2+ levels. Thus, it appears that the extent to which different inputs affect cytosolic Ca2+ concentration correlates with the extent of ISC proliferation.
Although, as a universal intracellular signal, cytosolic Ca2+ controls a plethora of cellular processes, we were able to demonstrate that cytosolic Ca2+ levels regulate Ras/MAPK activity in ISCs. Specifically, we found that trpA1 RNAi or RyR RNAi block Ras/MAPK activation in stem cells, and that forced cytosolic Ca2+ influx by SERCA RNAi induces Ras/MAPK activity. Moreover, Ras/MAPK activation is an early event following increases in cytosolic Ca2+, since we observed increased dpErk signal in stem cells expressing SERCA RNAi before they undergo massive expansion, and when we co-expressed Yki RNAi to block proliferation. It should be noted that a more variable pattern of pErk activation was observed with prolonged increases of cytosolic Ca2+, suggesting complicated regulations via negative feedback, cross-activation, and cell communication at late stages of Ca2+ signaling. This might explain why Deng et al. failed to detect pErk activation after 4 days induction of Ca2+ signaling (Deng et al., 2015). Previously, Ras/MAPK activity was reported to increase in ISCs, regulating proliferation rather than differentiation, in regenerating midguts (Jiang et al., 2011), which is consistent with our findings about TRPA1 and RyR.
The CanA1/CRTC/CrebB pathway previously proposed to act downstream of mGluR-calcium signaling (Deng et al., 2015) is not likely to play a major role in high Ca2+-induced ISC proliferation, as we tested multiple RNAi lines targeting CanA1 or CrebB and none of them suppressed SERCA RNAi-induced ISC proliferation. In support of this model, we also found that the active forms of CanA1/ CRTC/ CrebB cannot stimulate mitosis in ISCs when their cytosolic Ca2+ levels are restricted by trpA1 RNAi, whereas mitosis induced by the active forms of Ras or Raf is not suppressed by trpA1 RNAi.
Prior to our study, it has been shown that paracrine ligands such as Vn from the visceral muscle, and autocrine ligands such as Spi and Pvf ligands from the stem cells, can stimulate ISC proliferation via RTK-Ras/MAPK signaling (Bond and Foley, 2012; Jiang et al., 2011; Xu et al., 2011). We found that multiple RTK ligands in the midgut are down-regulated by trpA1 RNAi expression in the ISCs, including spi and pvf1 that can be induced by SERCA RNAi. Further, we demonstrated that high Ca2+ fails to induce ISC proliferation in the absence of EGFR. As spi is induced by EGFR-Ras/MAPK signaling in Drosophila cells (Ammeux et al., 2016), and DNA binding mapping (DamID) analyses from our lab and others indicate that spi might be a direct target of transcriptional factors downstream of EGFR-Ras/MAPK in the ISCs (Jin et al., 2015) (Doupé and Perrimon, unpublished), the autocrine ligand Spi might therefore act as a positive feedback mechanism for EGFR-Ras/MAPK signaling in ISCs.
In summary, our study identifies a mechanism by which ISCs sense microenvironment stress signals. The cation channels TRPA1 and RyR detect oxidative stress associated with tissue damage and mediate increases in cytosolic Ca2+ in ISCs to amplify and activate EGFR-Ras/MAPK signaling (Figure 8). In vertebrates, a number of cation channels (Monteith et al., 2007), including TRPA1 and RyR (Abdul et al., 2008; Büch et al., 2013; Du et al., 2014), have been associated with tumor malignancy. Our findings, unraveling the relationship between redox-sensing, cytosolic Ca2+, and pro-mitosis Ras/MAPK activity in ISCs, could potentially help understand the roles of cation channels in stem cells and cancers, and inspire novel pharmacological interventions to improve stem cell activity for regeneration purposes and suppress tumorigenic growth of stem cells.
The following strains were obtained from the Bloomington Drosophila Stock Center: Dl-lacZ (BL11651), y v; attP2 (landing site only, BL36303), UAS-Luc RNAi (BL31603), UAS-trpA1 RNAi (BL31504, BL31384, BL36780), UAS-RyR RNAi (BL29445, BL31540, BL31695), UAS-SERCA RNAi (BL25928, BL44581), trpA11 (BL26504), UAS-trpA1-A (BL26263, also called UAS-TrpA1.K), UAS-Ras1 RNAi (BL29319), UAS-Yki RNAi (BL31965), UAS-Src42A RNAi (BL44039, validated by Sopko et al. (2014)), UAS-Src64B RNAi (BL51772), UAS-EGFR RNAi (BL25781, BL31525, BL31526), UAS-CanA1 RNAi (BL25850, phenotypically validated by Lee et al. (2016)), UAS-CrebB RNAi (BL63681, BL29332), UAS-CRTC RNAi (BL28886), UAS-rpr (BL5823), UAS-Hepca (BL6406), UAS-Src42Aca (BL6410), UAS-Src64B (BL8477), UAS-GCaMP6s (BL42749), UAS-GCaMP6f (BL42747), UAS-RFP LexAop2-GFP; UAS-MKII::nlsLexADBD, UAS-p65AD::CaM; UAS-p65AD::CaM, tubGal80ts (TRIC lexGFP reporter, BL62828), tubGal80ts (on the X chromosome, BL7016), tubGal80ts (BL7108), tubGal80ts (BL7019), tubGal80ts (BL7018), UAS-mCherry (BL52268), UAS-mCherryCAAX (BL59021), RyRGal4R14G09 (with cloned enhancer region, BL48662), FRT19A (BL1744), FRT42D; FRT82B (BL8216), FRT42D w+ (BL1928), FRT42D; ry605 (BL1802), FRT2A (BL1997). Strains from Vienna Drosophila RNAi Center: yw; attP (landing site only, v60100), UAS-trpA1 RNAi (v37249) (Zhong et al., 2012), UAS-Ras1 RNAi (v106642), UAS-Yki RNAi (v104523). Strains from Perrimon lab stock: w1118, UAS-Luciferase, UAS-nlsGFP, UAS-Ras1A, UAS-Rafgof, esgGal4 UAS-GFP tubGal80ts (EGT), esgGal4, tubGal4, 24BGal4, DaGal4, esgGFP (generated by David Doupé), Su(H)GbeGFP (generated by Li He), UAS-spi (secreted form, validated in Ghiglione et al. (2002)). trpA1Gal4CP2A was generated by Junjie Luo. FRT42D RyRk04913 was generated from DGRC111172 (Kyoto Stock Center). trpA11 FRT2A and the stock with a deletion of trpA1, Df4415, were obtained from C. Montell lab, DlGal4 and Su(H)Gal4 from Steven Hou, UAS-Cas9.P2 from Fillip Port, UAS-2xEGFP [m5B29]; UAS-2xEGFP [m6B1] (UAS-ultraGFP) from Fabio Demontis, UAS-CncC RNAi from Dirk Bohmann, UAS-trpA1-C (also referred to as trpA1(A)10b) and UAS-trpA1-D (also referred to as trpA1(A)10a) (Guntur et al., 2015) from Paul Garrity, UAS-CanA1 RNAi fb5 and UAS-CanA1ca (constitutively active) (Wong et al., 2014) from Patrick H. O'Farrell, UAS-pvf1 from Edan Foley, UAS-CRTC and UAS-tdTomato-P2A-GCaMP5 from Henri Jasper, UAS-CrebBact from Jerry Yin. EGT F/O system is from Bruce Edgar: EGT; UAS-Flp, Act>Stop>Gal4. For MARCM induction, hsFlp tubGal80 FRT19A/ FM7; tubGal4 UAS-mCD8::GFP/ TM3 was obtained from Ben Ohlstein, hsFlp tubGal4 UAS-GFP-myc-nls;; tubGal80 FRT2A/ TM6B from Gary Struhl, and yw hsFlp UAS-GFP tubGal4; FRT42D tubGal80 from Huaqi Jiang.
Flies were reared on standard cornmeal/agar medium in 12: 12 light: dark cycles at 25°C unless noted otherwise. Fly food was changed every 2 days to keep fresh. Conditional expression in adult flies using tubGal80ts was achieved by maintaining flies at 18°C until 4–7 days after eclosion, and then shifting young adults to 29°C. For MARCM experiments, flies were maintained at 18°C until 3–5 days after eclosion, heat-shocked at 37°C for 1 hr, and then maintained back at 18°C before dissection and analysis.
See Supplementary file 1 for detailed genotypes of flies used in each figure.
Standard cornmeal/agar medium were melted at 70 ~ 80°C and mixed with a final concentration of 25 µg/ml bleomycin (Calbiochem #203408) or 2 mM paraquat (Sigma Aldrich #856177) to prepare food that can induce tissue damage in the fly midgut. For pathogen infection, Pseudomonas entomophila (from Tony IP) was grown in LB media at 29°C until reaching OD600 = 1, and Pseudomonas aeruginosa 14 (from Chrysoula Pitsouli) was grown in LB media at 37°C until reaching OD600 = 3. Flies were transferred to vials containing a cotton ball impregnated with 5 ml of water solution composed of 20% Pseudomonas culture media and 4% sucrose.
With a luciferase reporter that we have previously developed as a sensitive surrogate for stem cell pool size (Markstein et al., 2014), we performed an RNAi screen to identify novel components regulating ISC activity. In brief, midguts expressing luciferase together with RNAi targeting individual trans-membrane and receptor proteins (Hu et al., 2015) in adult ISCs for 9 days, with the last 2 days feeding either normally or on bleomycin food to induce tissue damage. Ten midguts were collected per genotype/ treatment, snap frozen, and stored at −80°C. After samples from all experiment groups were collected, the midguts were homogenized at 4°C using bead blender in Glo Lysis Buffer (Promega) with Halt protease inhibitor (Pierce 87786) and 2 mM trypsin inhibitor benzamidine. The lysate were centrifuged at 8000 rpm for 5 min to sediment the debris. And the supernatants were measured for luciferase activity using the Steady-Glo luciferase assay system (Promega). Luciferase activities were normalized to the value of spiked-in control groups (yv; attP2 for TRiP stocks, v60100 for Vienna stocks, and w1118 for NIG stocks). In particular, we noticed that trpA1 RNAi (BL31504), trpA1 RNAi (v37249), and RyR RNAi (BL29445) reduced EGT>luciferase activity by 74%, 88%, and 80%, respectively under normal feeding condition, and blocked damage-induced EGT>luciferase expression. We validated the top candidates from the screen by fluorescence imaging.
Drosophila midguts were dissected in PBS, fixed in 4% paraformaldehyde PBS solution for 30 min, rinsed and kept for 1 hr in PBST solution (PBS with 0.1% BSA, 0.3% Triton X-100) containing 5% Normal Donkey Serum. Midguts were stained overnight at 4°C with the following primary antibodies in PBST solution: rabbit anti-pH3 (Millipore #06–570; 1:3000), mouse anti-GFP (Invitrogen A11120; 1:300), rabbit anti-GFP (Invitrogen A6455; 1:500), rabbit anti-RFP (Life Technologies R10367; 1:500), rabbit anti-β-galactosidase (Cappel; 1:6000), rabbit anti-dpErk1/2 (Cell Signaling #4370; 1:500), rabbit anti-Pdm1 (from Xiaohang Yang; 1:500), rat anti-TRPA1 (from Paul Garrity; 1:100). Fly midguts were then washed three times in PBST and incubated with secondary antibodies and DAPI (1:2000) in PBST for 2 hr in the dark at room temperature. Secondary antibodies were donkey anti-rabbit, anti-mouse, and anti-rat IgGs conjugated to Alexa488, Alexa594 and Alexa647 (used at 1:1000, Molecular Probes). Fly guts were then washed three times in PBST, mounted in Vectashield (Vector Laboratories). For anti-TRPA1 staining, primary antibody incubation was done for 1 day and secondary antibody incubation overnight at 4°C. Images of posterior midgut area were captured with a Zeiss LSM780 confocal microscope equipped with 40x oil lens. A z-stack of 10–20 images covering the epithelium from the apical to the basal side were acquired and shown in max projection. All images were adjusted and assembled using NIH ImageJ.
For counting mitotic cells or MARCM clone size, an epifluorescence microscope was used to examine the stained midguts. For the quantification of esg+, Pros+, Dl+ or Su(H)Gbe+ cell density, as well as GCamP+ cell number, images of the majority of posterior midgut region were acquired on a Keyence microscope.
For measuring midgut length, images of the whole midgut were captured with an epifluorescence microscope. The middle line of the midgut was measured using NIH ImageJ, starting from the end of the proventriculus to the border between the posterior midgut and hindgut, where the malphighian tubules connect to the gut.
For detection of ROS via DHE staining, we followed the previously described protocol (Hochmuth et al., 2011). In brief, midguts were dissected and handled throughout in Schneider’s medium (HyClone). After incubation in 30 µM DHE (Invitrogen) for 5 min in the dark at room temperature, midguts were washed three times and mounted in Schneider’s medium. Images were captured immediately with a Zeiss LSM780 confocal microscope (543 nm excitation, 550–610 nm detection). All images were adjusted and assembled using NIH ImageJ.
Two pairs of short guide RNAs (sgRNAs) targeting the coding region of trpA1 or RyR were designed using the ‘Find CRISPRs' online tool (http://www.flyrnai.org/crispr2/) (Housden et al., 2016), and cloned into the double sgRNA vector pCFD4 (Port et al., 2014) for site- specific insertion into attP2 on the third chromosome. We obtained transgenic flies ubiquitously (driven by U6 promoters) expressing sgRNAs with the following seed sequences:
U6-sgtrpA1: CACACGATCCACGTCACTGT & CTGTAGACTCCGTTGTAGAG
U6-sgRyR: TTCGTGTTCCATCTGTACAA & ACAAGAACGTGCCGCCGGAT
Total RNA was extracted from 15 to 20 midguts or 15 heads or 5 larvae (all females) by TRIZOL reagent (Thermo Fisher), converted to cDNA template after DNase I treatment and purification by QIAGEN RNeasy kit. The cDNA was analyzed by real-time quantitative PCR using SYBR Green or by regular PCR (for trpA1 isoform typing) with GAPDH and rp49 as an internal control. RT-qPCR data were analyzed with Bio-Rad CFX Manager software. Primers used for RT-qPCR or regular PCR (for trpA1 isoform typing) are shown below:
TRPA1 currents in Xenopus laevis oocytes were recorded as previously described (Kwon et al., 2010). Xenopus ovaries were dissected and treated with 1.5 mg/ml collagenase A (Roche) and 1 mg/ml trypsin inhibitor (Roche) in OR2 Ringers (100 mM NaCl, 2 mM KCl, 1 mM MgCl2, 5 mM HEPES, pH = 7.5). The deflocculated oocytes were recovered at 18°C for 12 hr in OR3 medium (pH = 7.5): 50% Leibovitz’s media L-15 (Sigma L1518), 13 mM HEPES, 90 μg/ml gentamicin, 90 μg/ml Fungizone, 90 μg/ml penicillin/streptomycin. trpA1-C and trpA1-D (Kang et al., 2011) mRNAs were generated from pCS2 or pOX expression vectors with the mMessage mMachine kit (Ambion) and diluted in DEPC-treated water at the concentration of 1 μg/μl. Stage V-VI oocytes were injected with 50 nl prepared mRNA or water, and incubated at 18°C for 3 days for optimal expression. Two electrode voltage clamp recordings were performed at −80 mV with TURBO TEC-03X system and Cellwork software (NPI Electronics, Tamm, Germany), using ND96 buffer (96 mM NaCl, 1 mM MgCl2, 4 mM KCl, and 5 mM HEPES, pH = 7.6) with 1.8 mM CaCl2. In each experiment, we used 4 mM paraquat to treat the oocytes expressing TRPA1-C or -D, and added 0.01% AITC near the end of the recordings as a positive control.
Cytosolic Ca2+ in stem cells was monitored ex vivo using UAS-GCaMP6f (Figure 3E,E’, F,F’), UAS-GCaMP6s (Figure 5 and Supplementary Movies) or UAS-tdTomato-P2A-GCaMP5G (Figure 5—figure supplement 2A,A’, B,B’). In the latter case, tdTomato was used as an internal control. Young adult flies (4–7 days after eclosion) were first incubated at 29°C for 3 days to boost reporter expression before experiments. Intact midguts were dissected and handled in adult hemolymph-like (AHL) buffer. 1–2 intact midguts were placed in single wells of eight-well clear bottom cell culture chamber slides, gently oriented with Nylon mesh, and immersed in 200 µL AHL buffer for imaging. Images of the posterior midgut area (the region close to the middle midgut) were captured with a Zeiss LSM780 confocal microscope equipped with 10x lens. A z-stack of eight images covering one layer of the epithelium was recorded every 10 s for UAS-GCaMP6s, or every 20 s for UAS-tdTomato-P2A-GCaMP5G. Images before and after drug treatment are shown in z-max projection in Figure 5—figure supplement 2, while the time-lapse z-max projection movies for GCaMP6s are shown in Videos 1–6. Paraquat (Sigma Aldrich #856177), AITC (Sigma Aldrich #377430), and thapsigargin (Sigma Aldrich T9033) were dissolved in AHL buffer as 5x stock, and 50 µL drug stock was added to 200 µL imaging buffer to obtain desired final concentration. To quantify stem cells with high Ca2+ levels in most of the posterior region, Z-stack images were acquired on a Keyence microscope (Figure 5). All images were adjusted, assembled, and analyzed using NIH ImageJ.
Statistical analyses were performed using Microsoft Excel (except for qPCR experiments, which were analyzed with Bio-Rad CFX Manager software). Average value and standard error of the mean (SEM) were calculated and shown in the figures. For comparison of data groups, p values were determined by Student’s two-tail t-test with unequal variances, except for the comparison before and after drug treatment in Figure 5D, where the p value was determined by two-tail paired t-test. Double asterisks (**) label a p value that is less than 0.01. The p value of more than 0.1 is labeled as ‘not significant’ (n.s.) or not labeled. Sample sizes were chosen empirically based on the observed effects and listed in the figure legends.
Ryanodine receptor expression correlates with tumor grade in breast CancerPathology & Oncology Research 14:157–160.https://doi.org/10.1007/s12253-008-9045-9
Autocrine platelet-derived growth factor-vascular endothelial growth factor receptor-related (Pvr) pathway activity controls intestinal stem cell proliferation in the adult Drosophila midgutJournal of Biological Chemistry 287:27359–27370.https://doi.org/10.1074/jbc.M112.378018
Pathogenic stimulation of intestinal stem cell response in DrosophilaJournal of Cellular Physiology 220:664–671.https://doi.org/10.1002/jcp.21808
Drosophila calcineurin promotes induction of innate immune responsesCurrent Biology 17:2087–2093.https://doi.org/10.1016/j.cub.2007.11.001
A transcriptional reporter of intracellular Ca(2+) in DrosophilaNature Neuroscience 18:917–925.https://doi.org/10.1038/nn.4016
Bleomycin: new perspectives on the mechanism of actionJournal of Natural Products 63:158–168.https://doi.org/10.1021/np990549f
Design and generation of Drosophila Single Guide RNA expression constructsCold Spring Harbor Protocols 2016:pdb prot090779.https://doi.org/10.1101/pdb.prot090779
GLAD: an online database of gene list annotation for DrosophilaJournal of Genomics 3:75–81.https://doi.org/10.7150/jgen.12863
Intestinal stem cells in the adult Drosophila midgutExperimental Cell Research 317:2780–2788.https://doi.org/10.1016/j.yexcr.2011.07.020
The calcineurin inhibitor Sarah (Nebula) exacerbates aβ42 phenotypes in a Drosophila model of alzheimer's diseaseDisease Models & Mechanisms 9:295–306.https://doi.org/10.1242/dmm.018069
Increased oxidative stress and decreased antioxidant defenses in mucosa of inflammatory bowel diseaseDigestive Diseases and Sciences 41:2078–2086.https://doi.org/10.1007/BF02093613
Systematic screen of chemotherapeutics in Drosophila stem cell tumorsProceedings of the National Academy of Sciences 111:4530–4535.https://doi.org/10.1073/pnas.1401160111
Synergistic effect of Ferulic Acid and Z-Ligustilide, Major components of A. sinensis, on regulating Cold-Sensing protein TRPM8 and TPRA1 In VitroEvidence-Based Complementary and Alternative Medicine 2016:3160247.https://doi.org/10.1155/2016/3160247
Niche appropriation by Drosophila intestinal stem cell tumoursNature Cell Biology 17:1182–1192.https://doi.org/10.1038/ncb3214
Whole-brain activity mapping onto a zebrafish brain atlasNature Methods 12:1039–1046.https://doi.org/10.1038/nmeth.3581
Whole-brain activity mapping onto a zebrafish brain atlasNature Methods 12:1039–1046.https://doi.org/10.1038/nmeth.3581
Oxidative stress and pathogenesis of inflammatory bowel disease: an epiphenomenon or the cause?Digestive Diseases and Sciences 52:2015–2021.https://doi.org/10.1007/s10620-006-9622-2
Essential roles of intracellular calcium release channels in muscle, brain, metabolism, and agingCurrent Molecular Pharmacology 8:206–222.https://doi.org/10.2174/1874467208666150507105105
The use of xenopus laevis oocytes for the functional characterization of heterologously expressed membrane proteinsCellular Physiology and Biochemistry 10:1–12.https://doi.org/10.1159/000016341
Expression of hsp22 and hsp70 transgenes is partially predictive of Drosophila survival under normal and stress conditionsThe Journals of Gerontology Series A: Biological Sciences and Medical Sciences 64:828–838.https://doi.org/10.1093/gerona/glp054
Bruce EdgarReviewing Editor; University of Utah, United States
In the interests of transparency, eLife includes the editorial decision letter and accompanying author responses. A lightly edited version of the letter sent to the authors after peer review is shown, indicating the most substantive concerns; minor comments are not usually included.
Thank you for submitting your article "Oxidative stress induces stem cell proliferation via TRPA1/RyR mediated Ca2+ signaling in the Drosophila midgut" for consideration by eLife. Your article has been favorably evaluated by Jonathan Cooper (Senior Editor) and three reviewers, one of whom, Bruce Edgar (Reviewer #1), is a member of our Board of Reviewing Editors. The following individual involved in review of your submission has agreed to reveal their identity: W. Daniel Tracey (Reviewer #2).
The reviewers have discussed the reviews with one another and the Reviewing Editor has drafted this decision to help you prepare a revised submission.
Xu et al. demonstrate that two conserved regulators of intracellular calcium, TrpA1 and RyR, are required for stem cell stress responses in the Drosophila gut, and provide evidence that one of these effectors, the Ca2+ channel TRPA1, may be activated by reactive oxygen species (ROS). They present an extensive and coherent set of tests supporting a model in which ROS trigger Ca2+ release via TRPA1 in ISCs, and that the Ca2+ in turn activates Ras/Mapk signaling to promote intestinal stem cell (ISC) divisions. The reviewers agreed that the experiments were generally performed to a high standard and that most of the conclusions were convincingly supported. However, no mechanistic insight was offered as to how ROS might activate TRPA1, or how Ca2+ might activate Ras, though similar effects have been reported before in other cell types. Given that other studies of the fly gut have concluded that Ras/Mapk activation in ISCs is effected non-cell autonomously, via the EGFR and ligands produced by other cells, the authors' proposal that Ras is activated cell autonomously by Ca2+ is novel, interesting, and provocative.
1) Two reviewers noted that authors' proposal that the effect of ROS via Ca2+ on ERK and ISC division is cell autonomous, is not well supported and runs counter to many previous publications suggesting that ERK is activated by stress-dependent EGFR ligands non-cell autonomously supplied from other cells in the epithelium. The authors should either provide additional data supporting their cell-autonomous mechanism and ruling out non-autonomous mechanisms, or they should alter the model they present in Figure 7. In any case the autonomous vs. non-autonomous models should be carefully discussed.
2) Along the same lines, reviewer #3 notes that further data should be provided to rule out the role of CrebB and CanA as effectors of Ca2+, and to support the authors proposal that ERK is the principal Ca2+ effector.
3) Reviewer #2 notes a substantial body of relevant published work that was not cited or discussed (i.e. Du and Guntur papers). The manuscript should be revised to address these deficiencies.
4) Two reviewers commented on the weakness of data showing the TRPA1 activation is sufficient to promote ISC divisions, and reviewer #2 suggests an experiment to test this using a heat-activatable TRPA1 variant. If this experiment is possible, the authors should try it, or present some other type of data further supporting the sufficiency of TRPA1 mediated Ca2+ release to promote ISC division. This might also address point #1 above.
5) Reviewers 1and 3 discussed several issues with conclusions that could be clarified by counting ISCs and other cell types (EEs), from experiments that were already performed. This additional analysis should be provided if possible.
In this paper Xu et al. demonstrate that two conserved regulators of intracellular calcium, TrpA1 and RyR, are required for stem cell stress responses in the Drosophila gut, and provide evidence that one of these effectors, the Ca2+ channel TRPA1, may be activated by reactive oxygen species (ROS). They present an extensive and coherent set of tests supporting a model in which ROS trigger Ca2+ release via TRPA1 in ISCs, and that the Ca2+ in turn activates Ras/Mapk signaling to promote ISC divisions. No mechanistic insight is offered as to how ROS might activate TRPA1, or how Ca2+ might activate Ras, though similar effects have been reported before in neurons (also without mechanistic explanation). The results are consistent with a recent report in Nature from Jasper's group (Deng, 2016). Given that other studies of the fly gut have concluded that Ras/Mapk activation in ISCs is effected non-cell autonomously, via the EGFR and ligands produced by other cells, the authors' proposal (Figure 7) that Ras is activated cell autonomously by Ca2+ is novel and interesting. However, this study does not definitively rule out non-cell autonomous signaling as a mechanism of ROS sensing or Ca2+ release, and so I think the model should be presented with reference to the rather different conclusions of previous work. Otherwise, apart from a few relatively minor issues with data analysis and interpretation, the paper is a nice story in principle appropriate for eLife. I list issues that need to be addressed in more detail below:
1) Subsection “TRPA1 and RyR are required for stem cell damage response and self-renewal”, third paragraph, Figure 2E-H: The authors show that the trpA1 and RyR RNAi's deplete Δ+ stem cells in 5 days. Nevertheless, they conclude that these RNAi's are not killing stem cells. This is not very convincing, and I think the number of stem cells should be counted using several assays, at several timepoints, since the issue of cell viability affects other critical results and the overall interpretation. For instance, in Figure 3 we see decreases in ISC mitoses. These results would be confounded if the total number of ISCs decreases during the experiment.
2) Similarly, in Figure 4 we see increases in high Ca2+ cells after paraquat treatment in controls, but not in animals expressing RNAi against TRPA1. (Subsection “TRPA1 and RyR are required for normal and ROS-induced cytosolic Ca2+ 146 levels in ISCs”, last paragraph, Figure 4—figure supplement 1). But these increases could be due to division of cells with high Ca2+, rather than to an increase in the fraction of ISCs with high Ca2+. This would not occur in the RNAi treated ISCs, since the RNAi blocks their division. To address this potentially confounding issue the authors should count total ISC numbers in these experiments, as well as numbers of high Ca2+ cells, and display the fraction of ISCs that are Ca2+ high.
3) As alluded to above, the regulation of Ras activity by Ca2+ is not explained mechanistically, and the authors' model that the effect of ROS via Ca2+ is autonomous to stem cells is based on correlations that don't rule out other explanations. Although TrpA1 and RyR are clearly necessary for ISC activation, and although other treatments that increase Ca2+ are sufficient to activate ISCs (e.g. SERCA-RNAi), this paper doesn't provide strong evidence that activation of TrpA1 or RyR (in this case by ROS) is sufficient to activate ISCs. Moreover the data on ROS activation of TrpA1, though extensive and essential to the model (Figure 7), is the weakest in the paper. Thus, although the authors' model is coherent and consistent with their data, it seems possible that the increase of Ca2+ and activation of ERK in ISCs actually involves signals from other cells in the epithelium, rather than a direct effect of ROS on TrpA1, as depicted in Figure 7. It would be interesting, for instance, to test whether the EGFR is required in this context (it is required for ISC responses to many stresses, but should not be required here). In the absence of experiments ruling out roles for other cell types, I feel the model figure (Figure 7) and Discussion should be modified to include the possibility that the oxidative damage signals are received by cells other than stem cells.
4) The dpERK stainings in Figure 5A-F should be repeated with anti-TrpA1 RNAi's. Only tests with RyR-RNAi are shown.
5) For most of the study, only TrpA1 is tested in detail. The manuscript would be better if the authors had carried through with a deeper analysis of RyR as well
The study of Xu et al. finds an important role for the Drosophila TRPA1 channel and Ryanodine Receptor channels in Reactive Oxygen Species (ROS) induced proliferation of intestinal stem cells. The combined evidence in support of the role of these two channels in this biological process as presented in the study is very strong. With respect to dTRPA1, the study points to a specific isoform (dTRPA1-D) in the sensing of ROS and the study also finds that Ras/Mapk function downstream of Ca influx/release that are triggered by dTRPA1 and Ryr respectively. The strength of the advance in the study is that it ties calcium signaling to ISC proliferation specifically in the context of ROS triggered proliferation and the study also specifically ties the calcium to dTRPA1 and Ryr. The authors do discuss the fact that a prior study has tied calcium to ISC proliferation through mGluRs and dietary glutamate. But some of the prior literature on dTRPA1 channels (in the context of ROS signaling) has not been adequately credited by the current study. For example, the dTRPA1-D isoform has been previously shown to be activated by ROS (Drosophila TRPA1 isoforms detect UV light via photochemical production of H2O2.
Guntur AR, Gu P, Takle K, Chen J, Xiang Y, Yang CH.Proc Natl Acad Sci U S A. 2015 Oct 20;112(42):) through a dual oxidase pathway in the ring gland. As well, a role for dTRPA1 in sensing ROS in the gut (also through dual oxidase) has been previously reported: TrpA1 Regulates Defecation of Food-Borne Pathogens under the Control of the Duox Pathway. Du EJ, Ahn TJ, Kwon I, Lee JH, Park JH, Park SH, Kang TM, Cho H, Kim TJ, Kim HW, Jun Y, Lee HJ, Lee YS, Kwon JY, Kang K.PLoS Genet. 2016 Jan 4;12(1)). This latter study also finds that it is the dTRPA1-D isoform which is activated by ROS and the study has not been cited in the current manuscript. Although Guntur et al. is briefly cited in the current study, the analysis of ROS activation by dTRPA1in heterologous oocytes do not significantly extend those that were initially reported by the studies of Guntur and Du. As such the new heterologous expression analyses should thus be presented as important confirmatory findings (even if investigating different chemical compounds to generate the ROS) with a bit more credit/discussion given to the prior studies. The new biological role for this function- in promoting ISC proliferation in response to ROS- is what is important here, and this will be of great interest to the readers of eLife.
Experimental points to be addressed:
The anti-dTRPA1 staining does not provide a clear description of where dTRPA-C/D is actually expressed in the gut. Reporters for dTRPA1-C/D are available from several labs and these could be used to more clearly illustrate the expression pattern.
In this manuscript, Xu et al. describe a novel mechanism controlling intestinal stem cell proliferation in Drosophila. They found that oxidative stress stimulates Ca2+ signaling by regulating TRPA1 and RyR. While the role of calcium signaling has previously been reported, this new study describes a different pathway causing changes in [Ca2+] and a different downstream effector cascade (Ras/ERK).
Experiments reported in this paper are generally well designed and properly controlled. They address an important question in the field, regarding the mechanisms that allow the integration of multiple signaling pathways in ISCs to control proliferation rates in response to stress (oxidative stress and tissue damage).
This study has the potential to be of interest for the large audience of eLife.
In my opinion, however, a few important points would need to be addressed to fully support the authors' main conclusions:
The authors show that TRPA1 and RyR knock-down results in a loss of Dl+ stem cells. Because they argue that these cells do not die of apoptosis and do not commit to the enteroblast lineage, I am confused about what the fate of these ISCs is, after these genetic manipulations. The authors should at least test whether it promotes enteroendocrine cells fate by monitoring the number of prospero-positive cell in these epithelia. Also, it is possible that the expression of Δ is lost in these stem cells? Other ISC markers such as Sanpodo would be useful to test this hypothesis. Finally, are the single cell trpA and RyR mutant clones Dl-positive?
Altogether, these experiments would greatly help clarifying the fate of ISCs in which trpA or RyR activities is impaired.
The authors use paraquat exposure as an oxidative stress in most of their experiments. While the results are compatible with a regulation of Ca2+ signaling by ROS via a cell autonomous mechanism (including the control of TRPA1 activity), relatively simple experiments could be performed to rule out a non-cell autonomous mechanism following paraquat-induced tissue damage.
Does the expression of antioxidant proteins in ISCs prevent ROS-induced Ca2+ signaling? Do ISC-specific genetic manipulations that have been shown to induce ROS production in ISCs and promote cell proliferation (such as ND42 RNAi) increase Ca2+ signaling?
In the last part of this manuscript, the authors try to exclude the role of CanA1 and CrebB downstream of Ca2+ release, arguing that Ras/ERK is the main Ca2+ effector in ISCs. This is an important point as it has been previously published that these factors are part of the signaling pathway that regulate ISC proliferation in response to [Ca] changes, and this new study challenges this view. I believe that this conclusion needs to be supported by additional data.
The experiment in Figure 6H needs to be quantified and the other genetic manipulations (multiple RNAi lines targeting CanA1 and CrebB mentioned as "not shown" in the Results and Discussion) should be included in the main figures. It would be essential here to demonstrate that CanA1 and CrebB are dispensable for ROS- and SERCARNAi-mediated proliferation as suggested by the authors' model.
[Editors' note: further revisions were requested prior to acceptance, as described below.]
Thank you for resubmitting your work entitled "Oxidative stress induces stem cell proliferation via TRPA1/RyR mediated Ca2+ signaling in the Drosophila midgut" for further consideration at eLife. Your revised article has been favorably evaluated by Jonathan Cooper (Senior Editor), a Reviewing Editor and two reviewers.
The consensus is that the manuscript has been improved, and we appreciate the new data you've included as well as your editorial efforts. However, there are two remaining issues that are quite significant, and really need to be addressed before acceptance, as outlined below:
1) One reviewer and the reviewing editor still believe the manuscript still falls short of establishing how, mechanistically, [Ca2+] activates ERK. A key issue is whether this happens directly, cell-autonomously, or through non-autonomous interactions between different cells in the gut epithelium. The old and new data do not resolve this issue, despite requests from reviewers. The model figure tries to cover both (or all possible) mechanisms and as such is a confusing conclusion – in fact the model as presented will only serve to obscure whatever the real mechanism is, not to resolve it. We would very much like to see a clear resolution of this issue, so that this paper can be seen as an obvious advance over the recent similar publication from Deng et al. A reviewer's specific comments on this point follow: "[…] the authors provide new data demonstrating that [Ca2+] controls the expression of Spitz and other RTK ligands. However, the authors appear to favor a model where [Ca2+] controls ERK activation directly through Src, but such activation is not demonstrated and the mechanism is unclear. As an example of this, the revised model still highlights a direct regulation of Ras/MAPK by Ca2+, and the authors discuss a scenario where Spitz induction is downstream of an initial direct activation of ERK rather than an integral part of the ERK activation mechanism. I think the two models (direct or autocrine/paracrine) could be easily be distinguished by testing whether EGFR and/or PVR are required for ISC proliferation induced by SERCARNAi and TRPA1-D overexpression, as suggested in the initial review. I believe that confirming that Ca2+ can activate ERK directly in ISCs (independently of a RTK signal) would greatly increase the impact of this study." Conversely, demonstrating the EGFR is essential for deregulated proliferation after Ca2+ signaling dysfunction would at least show that the mechanism is not simple and not cell autonomous.
2) A second, and related, issue is how your results and conclusions relate to the Deng et al. publication. We feel that you should show the relevant data you have and clarify how your conclusions are different from Deng's and why. On this issue, a reviewer commented: "some of the data required is not presented and some of the conflicting data not discussed. For example, the genetic interaction between SERCARNAi and CrebBRNAi (CrebBRNAi not being able to block SERCARNAi-induced proliferation) is still reported as "data not shown" in the revised version. This result is used as an important piece of data to support the authors' conclusion that ERK is the major effector of Ca2+ in the control of ISC proliferation. Deng et al. reported that CrebBRNAi expression causes ISC loss: do the authors here find something different? Also, Deng et al. reported that dpERK levels are not affected by SERCARNAi expression. This inconsistency should be at least noted and discussed." Although we feel these issues are less pressing than #1 above, we would appreciate some revisions to make it clear how your results and conclusions differ from those presented by Deng et al.https://doi.org/10.7554/eLife.22441.039
- Craig Montell
- Craig Montell
- Norbert Perrimon
- Norbert Perrimon
The funders had no role in study design, data collection and interpretation, or the decision to submit the work for publication.
We thank the Transgenic RNAi Project (TRiP) and Bloomington Drosophila Stock Center (BDSC) for providing fly stocks; we thank Drs. Paul Garrity, Patrick O'Farrell, Henri Jasper, Jerry Yin, Ben Ohlstein, Gary Struhl, Huaqi Jiang, Dirk Bohmann, and Xiaohang Yang for sharing reagents and fly stocks; we thank the members of the Perrimon lab, especially Kwon Young, Yanhui Hu, David Doupé, Stephanie Mohr, Patrick Jouandin, and Leonard Rabinow, for advice on experiments and comments on the manuscript. Work in the Montell lab is supported by the NIDCD (DC007864) and the NEI (EY010852). Work in the Perrimon lab is supported by NIGMS (GM067761) and HHMI. N.P. is an Investigator of the Howard Hughes Medical Institute.
- Bruce Edgar, Reviewing Editor, University of Utah, United States
- Received: October 16, 2016
- Accepted: May 23, 2017
- Version of Record published: May 31, 2017 (version 1)
© 2017, Xu et al.
This article is distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use and redistribution provided that the original author and source are credited.