Sirtuin-1 regulates organismal growth by altering feeding behavior and intestinal morphology in planarians

Sirtuins are a group of highly conserved protein deacetylases that require oxidized nicotinamide adenine dinucleotide (NAD⁺) as a cofactor for their enzymatic function. The ratio of NAD⁺ to reduced nicotinamide dinucleotide (NADH) in response to the nutrient state of cells links sirtuin function to metabolism. Sirtuin proteins are cellular sensors that function in response to metabolic inputs and that regulate stem cells, growth factor signaling, and food intake (Arul Nambi Rajan et al., 2018; Calvanese et al., 2010; Dietrich et al., 2010; Fujitsuka et al., 2016; Hisahara et al., 2008; Igarashi and Guarente, 2016; Li et al., 2007; Martins et al., 2012; Ou et al., 2011; Prozorovski et al., 2008; Saunders et al., 2010; Velasquez et al., 2011). Recent evidence implicates sirtuins in a myriad of human conditions including obesity related disease, aging and cancer, but the mechanistic roles that sirtuins play in these conditions remain elusive (DeBerardinis et al., 2008; Houtkooper et al., 2012; Kurylowicz, 2016). Studying the role of sirtuins in metabolism at the systemic level is challenging due to the complex regulatory mechanisms involved, and the fact that downregulation of sirtuin often results in developmental defects and decreased lifespan (Dang, 2014; Lemieux et al., 2005; Li et al., 2007; McBurney et al., 2003; Mostoslavsky et al., 2006; Vakhrusheva et al., 2008). Despite widespread interest in understanding the links between metabolism, homeostasis and disease, the mechanisms by which sirtuin regulates body growth and animal behavior are mostly unknown.

To gain insight into sirtuin function in the regulation of body homeostasis, we focus on the planarian flatworm Schmidtea mediterranea, which is an emerging model organism that has the unique capability of responding to nutrient availability by either increasing or decreasing body size. Changes in planarian size are noticeable within a few weeks and result from variations in cell number across the body rather than changes in cellular size (Baguna and Romero, 1981; Oviedo et al., 2003; Takeda et al., 2009; Thommen et al., 2019). Consequently, smaller animals will contain fewer cells than their larger counterparts, but their proportions are maintained despite differences in body size. Independently of their size or metabolic input, planarians are constantly renewing differentiated tissues, fueled by a large pool of adult stem cells called neoblasts. The neoblast is the only cell type with capacity to divide during tissue renewal and in response to injury repair (Wagner et al., 2011; Zeng et al., 2018). Therefore, the maintenance of body proportionality in planarians results from finely balanced regulation of cell division and cell death that is greatly influenced by the availability of nutrients. Recent research has revealed evolutionary conservation in planarians of signaling pathways associated with metabolism (such as TOR and insulin signaling) (Gonzalez-Estevez et al., 2012; Miller and Newmark, 2012; Peiris et al., 2012; Roberts-Galbraith et al., 2016), but it remains unknown whether sirtuins influence the cellular flux that occurs in response to tissue turnover and nutrients in planarians.

Our results show that sirtuins are molecularly conserved in planarians. We discovered that homologs for six of the seven mammalian sirtuins are present in the S. mediterranea genome (Smed-Sirt-1–6). This is intriguing because evolutionarily related invertebrate organisms such as Caenorhabditis elegans and Drosophila melanogaster possess fewer sirtuin orthologs than mammals and planarians (Frye, 2000; Vassilopoulos et al., 2011). Furthermore, we found that it is possible to modulate sirtuin function in planarians using RNAi and treatment with pharmacological compounds used in vertebrate experimental models and humans (Bhatt et al., 2012; Brasnyó et al., 2011; Knop et al., 2013; Liu et al., 2014; Nguyen et al., 2009). Disrupting Sirtuin-1 in planarians led to reduced body size, an unexpected increase in stem cell division and a reduction in levels of cell death. These effects were accompanied by a reduced branching of the gut. Our analysis revealed that Sirtuin-1 plays an integral role in feeding behavior, regulating not only the length of time taken to locate food and begin feeding but also overall food consumption. Taken together, our findings provide evidence for regulation of cellular turnover by Sirtuin-1, and introduce planarians as a tractable experimental model to investigate regulation of metabolic inputs and cellular behavior in the adult body.

Sirtuins are a group of NAD⁺-regulated deacetylase enzymes that are highly conserved from yeast to humans (Frye, 2000; Satoh et al., 2011). In the genome of S. mediterranea (Rozanski et al., 2019), we identified homologs for six of the seven human sirtuins, Smed-Sirt-1–6 (Fig. S1A). Although most of the S. mediterranea sirtuins appear evolutionarily conserved, the protein encoded by Smed-Sirt-1 showed the highest molecular similarity to its counterparts in humans and other species, sharing 60% amino acid sequence identity with human sirtuin-1 (also known as SIRT1) in the SIRTUIN domain (Fig. S1A–C).

To functionally characterize the role of sirtuins in planarians, we performed RNAi of the individual Smed-Sirt-1–6 transcripts. We developed an RNAi strategy based on multiple microinjections with double-stranded RNA (dsRNA) over ten days, and animals were fixed 15 days after the first injection (Fig. 1A). We confirmed that the protocol was effective in knocking down the gene expression for two weeks without noticeable abnormalities (Fig. S2A).

Sirtuin function is linked to regulation of metabolism; therefore, we analyzed how nutrient availability impacted animal size over time. First, we prolonged the RNAi protocol for about one month in the absence of nutrients and evaluated the effects on the reduction in animal size (Fig. S2B). These experiments demonstrated that disturbing Smed-Sirt-1–6 function does not affect the reduction in the overall size of planarians, termed ‘degrowth’ (Fig. S2C). Second, animals subjected to RNAi were exposed to nutrients, and their growth was evaluated for almost two months. This protocol alternated six feedings with RNAi of the individual Smed-Sirt-1–6 transcripts and was effective in reducing expression for approximately two months (Fig. 1B; Fig. S2D). The analysis shows that control and Smed-Sirt-2–6(RNAi) animals increased in size at a similar rate, whereas animals subjected to Smed-Sirt-1(RNAi) displayed a 30% reduction in growth (Fig. 1C,D; Fig. S2E). Additionally, Smed-Sirt-1 expression was increased upon feeding, suggesting a potential role in planarian metabolism (Fig. S2F). Based on these observations, we concluded that Smed-Sirt-1 is required for animal growth when nutrients are available. Therefore, we focused our attention on Smed-Sirt-1.

The spatial expression of Smed-Sirt-1 was determined by whole-mount in situ hybridization. The signal indicated that Smed-Sirt-1 is ubiquitously expressed, but is enriched in the prepharyngeal area (Fig. 1E). This finding was validated in silico by analyzing the expression of Smed-Sirt-1 along the planarian anteroposterior (AP) axis (Rozanski et al., 2019; Stückemann et al., 2017) (Fig. S2G). We also analysed single-cell gene expression data (Fincher et al., 2018) and confirmed that Smed-Sirt-1 expression is scattered among different cell types, including neoblasts and differentiated cells (Fig. S2H). Taken together, our results reveal that Smed-Sirt-1 is evolutionarily conserved in planarians and its expression pattern is broadly detected across the planarian body.

The activity of sirtuin proteins can be modulated by treatment with pharmacological compounds. Specifically, planarians were treated with resveratrol (RESV), a compound known to enhance sirtuin-1 function in across vertebrate and invertebrate species (Wood et al., 2004; Chao et al., 2017; Ma et al., 2017; Yu et al., 2017), which increases the rate of sirtuin-1 enzymatic activity by raising the substrate binding affinity (lowering K_m) (Bonkowski and Sinclair, 2016; Howitz et al., 2003). To inhibit sirtuin function, animals were treated with nicotinamide (NAM), which impairs the hydrolysis of NAD – a process essential for sirtuin activity (Avalos et al., 2005; Bitterman et al., 2002; Fischer et al., 2012; Strømland et al., 2019). First, we determined concentrations at which RESV and NAM did not lead to any macroscopic or behavioral defects. This was achieved by adding RESV (in DMSO) and NAM at different concentrations directly into the water where planarians live (Fig. S3A). Following this strategy, we determined that treating animals with concentrations of RESV greater than 20 µM occasionally led to anterior tissue loss known as head regression (Fig. S3B). Treatment with NAM using concentrations above 100 µM affected planarian locomotion, manifested by inching movement and lack of motility. At higher concentrations of both RESV and NAM, animals displayed lesions leading to lysis, and ultimately death shortly after. Therefore, we chose to treat planarians with 20 µM RESV and 100 µM NAM, because these concentrations did not produce evident abnormalities. Next, we evaluated the growth of animals exposed to treatment with RESV or NAM. Under this protocol, drugs were replaced every other day for 28 days and animals were fed once a week, starting five days after the first drug exposure (Fig. 2A). These experiments revealed that animals treated with RESV grew 37% more than the respective control (DMSO treated). Conversely, planarians exposed to NAM displayed ∼40% reduced growth (Fig. 2B,C). In the absence of nutrients, RESV and NAM treatment did not affect planarian ‘degrowth’ (Fig. S3C). In addition, we combined Smed-Sirt-1(RNAi) with a feeding schedule that lasted over a month, while animals were challenged with pharmacological compounds (Fig. 2D). We found that combining treatment with NAM and Smed-Sirt-1(RNAi) in the presence of nutrients led to a similar size reduction to that observed following NAM treatment alone, suggesting they both consistently limit growth by disturbing sirtuin function (Fig. 2E). These results demonstrate that it is possible to pharmacologically modulate planarian growth in the presence of nutrients by targeting Sirtuin-1. Importantly, neither RESV nor NAM appear to regulate changes in body size in the absence of nutrients, which is consistent with the results of Smed-Sirt-1(RNAi) treatment.

To assess whether RESV requires Sirtuin-1 for its effect on growth in the presence of nutrients, we simultaneously treated animals with RESV and inhibited Sirtuin-1 function with either RNAi of Smed-Sirt-1 or NAM treatment. We found that RESV treatment did not lead to increased growth when Smed-Sirt-1 expression was downregulated (Fig. 2F). Furthermore, RESV failed to enhance growth when Sirtuin-1 function was inhibited using NAM (Fig. 2G). Additionally, we found that RESV treatment was accompanied by a reduction in Smed-Sirt-1 expression (Fig. S3D). Our findings suggest that RESV-induced animal growth is the result of Smed-Sirt-1 enhancement at the post-translational level.

To address potential mechanisms involved in Smed-Sirt-1 regulation of animal size, we evaluated feeding behavior. Specifically, we implemented an experimental strategy that allowed us to consider the time taken to begin feeding and overall food consumption. First, we designed a chamber in which individual worms could be placed at similar distances from a food source and the time taken by each worm to find food and begin feeding, as evidenced by protrusion of the pharynx and continuous contact with the food source, could be tracked. This setup was effective in tracking up to 20 worms in each individual experiment (Fig. 3A). We found that animals subjected to Smed-Sirt-1 inhibition, either by RNAi or treatment with NAM, took double the time to locate food and begin feeding compared to untreated animals (Fig. 3B,C). Conversely, RESV treated animals were ∼37% faster to begin feeding (Fig. 3D). To rule out the possibility of impaired locomotion in Smed-Sirt-1(RNAi) animals, we measured the gliding distance over time (speed) and found no difference between control and experimental groups. Furthermore, qualitative evaluation of ventral cilia revealed no noticeable differences between the two groups (Fig. S4A,B). These results demonstrate that inhibition and enhancement of Smed-Sirt-1 affects the time taken to begin feeding without disrupting locomotion.

Second, because the time taken to locate food does not necessarily relate directly to the amount of food ingested, we implemented an additional assay to estimate food consumption. We mixed a known amount of planarian food (liver paste) with a non-toxic dye commonly used in Drosophila behavioral experiments (0.2% Erioglaucine disodium salt) (Aditi et al., 2016; Debban and Dyer, 2013). The food–dye mix was offered overnight, before the animals were homogenized and dye absorbance readings of the homogenates used to determine the relative amounts of food consumed (Fig. 3E). Because differences in planarian size may be a reflection of the amount of food ingested, we plotted a standard curve of dye absorbance against animal size. As expected, we found that food consumption was proportional to animal size (Fig. 3F). Using this assay, we identified that Smed-Sirt-1(RNAi)- and NAM-treated animals consumed less food than control animals (12% and 35%, respectively) (Fig. 3G,H). Conversely, RESV-soaked animals consumed ∼16% more food than the DMSO-treated control group (Fig. 3I). Taken together, our findings suggest that Smed-Sirt-1 influences both the time taken to begin feeding and the food intake of planarians.

Planarians are able to sense food through chemosensory signaling mediated by the nervous system. Thus, we explored whether Smed-Sirt-1 activity influences feeding behavior by affecting the mechanisms underlying this chemosensory response. To this end, we performed a gene expression analysis of putative markers commonly associated with feeding behavior mediated by neural chemosensory signaling in planarians (Collins et al., 2010; Miller and Newmark, 2012; Roberts-Galbraith et al., 2016; Shimoyama et al., 2016). We devised an assay in which planarians can sense food without physically ingesting it, to avoid intestinal distention that is known to activate metabolic pathways. The chemosensory response was stimulated by placing starving planarians in water containing a highly diluted food source (liver paste, 0.1%) for either 3–5 min or 30 min. This was considered the stimulated group, and as a control we used a similar condition without liver (baseline group). Changes in gene expression were determined by comparing the stimulated and baseline groups. Experiments were carried out using starving animals subjected to 15 days of Smed-Sirt-1(RNAi) or RESV-treated animals, and gene expression changes in each treatment compared to gene expression changes in untreated animals (Fig. 4A).

Upon chemosensory stimulation only a few genes in the control group were upregulated (4 of 27). However, treatment with RESV led to upregulation of 59% of the genes tested. Interestingly, some of the genes upregulated in the control (such as those encoding the neuropeptide Y homolog Smed-npy3 and the insulin-like growth factor-binding protein subunit Smed-ALS) were further enhanced by RESV treatment and downregulated in Smed-Sirt-1(RNAi) animals (Fig. 4B, Fig. S5A). Extending the evaluation to 30 min post-stimulation revealed an increased number of genes upregulated in the control group (15 of 27) and about the same number of genes (14 of 27) downregulated in Smed-Sirt-1(RNAi) animals (Fig. 4C). RESV treatment led to an enhanced expression of the genes upregulated in the control group (such as those encoding the neuropeptides Smed-npp22 and Smed-spp5; Fig. 4D, Fig. S5B). These results suggest that Smed-Sirt-1 regulates the expression of genes associated with chemosensory stimulation and feeding responses.

It is well established that the number of cells, rather than cell size, determines organismal dimensions in planarians (Baguna and Romero, 1981; Oviedo et al., 2003; Takeda et al., 2009; Thommen et al., 2019). Size is maintained by a constant renewal of tissues, which relies on a fine balance between cell proliferation and death (Pellettieri and Sánchez Alvarado, 2007). Based on this premise, we evaluated cell division and cell death to determine how Smed-Sirt-1 affects organismal growth at the cellular level. First, we analyzed mitotic events throughout the animal using immunostaining with an anti-phosphorylated histone 3 (H3P) antibody (Fig. 5A). Intriguingly, Smed-Sirt-1(RNAi) animals had on average ∼37% more dividing cells than the control group. An increase in cell division in the experimental group of ∼27% was observed as early as 15 days post-RNAi (Fig. 5B). At this stage of the assay the animals were starving, and the overall size of animals in the control and experimental groups was similar. The increase in cell division in the Smed-Sirt-1(RNAi) animals was greater after 55 days RNAi treatment (48%), by which time the RNAi-treated animals were markedly smaller than the control group (Fig. 5A,B). Upon feeding, planarians undergo a ‘mitotic burst’ occurring between 6 and 24 h after feeding begins (Baguna, 1974; Kang and Sánchez Alvarado, 2009). The rapid influx of dividing cells significantly influences the total number of cells in the body, thus leading to changes in body mass. Unexpectedly, Smed-Sirt-1(RNAi) animals displayed an even larger ‘mitotic burst’ after feeding (Fig. 5C). Lastly, we found that Smed-Sirt-1(RNAi) led to an alteration of cell cycle dynamics that was characterized by a 7% increase in the number of cells in the G2-M phases (Fig. S6A). These results suggest Smed-Sirt-1 regulates cell division, which is restricted to a subpopulation of neoblasts.

We further expanded our analysis by measuring levels of gene expression of the pan-neoblast marker smedwi-1 (also known as Piwi) and a recently identified marker for the clonogenic neoblast, tetraspanin-1 (Tspan-1), which labels the Nb2 subclass (Reddien et al., 2005; Zeng et al., 2018). We found ∼28% overexpression of both Smed-Piwi and Smed-Tspan-1 in Smed-Sirt-1(RNAi) animals over the observational period (Fig. 5D,E). Our findings demonstrate that Smed-Sirt-1 regulates cellular division and the expression of neoblast markers.

Next, we evaluated whether Smed-Sirt-1 regulates the rate of cell death. The apoptotic maker caspase-3 is cleaved in the final stages of apoptosis, leading to programmed cell death. We found a significant reduction in cleaved caspase-3 activity at 15 and 55 days of Smed-Sirt-1(RNAi) treatment (87% and 32%, respectively) (Fig. 5F–H). We confirmed the reduction in cell death in the experimental group at day 15 post RNAi using a terminal deoxynucleotidyl transferase dUTP nick-end labeling assay (TUNEL) (Fig. 5I). To explore other means of cell loss, we attempted to detect putative autophagy by measuring expression levels of genes critical for autophagosome formation (Smed-Atg5, Smed-Atg8 and Smed-Atg12), which have been established in various models (Levine and Kroemer, 2019; Mizushima et al., 2011), as well as the planarian autophagy marker Smed-DAP1 (González-Estévez et al., 2007; Levine and Kroemer, 2019; Mizushima et al., 2011). We identified that Smed-Sirt-1(RNAi) animals had ∼20% lower expression of the autophagy markers Smed-Atg8 and Smed-DAP1 compared to expression in control animals (Fig. 5J,K). To validate these findings, we analyzed a pro-survival marker, and found the expression of Smed-Bcl-2 to be upregulated by 32% in Smed-Sirt-1(RNAi) animals 15 days after RNAi (Fig. 5L). Because planarians are capable of regenerating any tissue in their body by activating neoblast division and cell death, we explored whether Sirtuin-1 was required for large scale tissue repair. Both control and Smed-Sirt-1(RNAi) animals were amputated in anterior and posterior regions, and the regenerating fragments were evaluated over seven days (Fig. S7A). The results demonstrate that both the control and experimental groups developed blastemas with equal dynamics and of equal size, suggesting that Sirtuin-1 function is more relevant in the context of animal growth than in regeneration (Fig. S7B,C). The data suggest that Smed-Sirt-1 is required to maintain the balance between cell division and cell death during adult tissue turnover.

Our results demonstrate that Smed-Sirt-1(RNAi) animals exhibit reduced growth despite excessive cellular proliferation, and lower levels of cell death. Next, we asked whether cellular differentiation was also part of the cellular imbalance. We considered a range of markers linked to terminal differentiation of distinct cell populations in intestinal, excretory, muscle, and epithelial cells (Zeng et al., 2018; Zhu et al., 2015). The results evidenced an initial reduction in the expression of markers associated with gut and parapharyngeal cells, which extended to other cell types as the phenotype progressed. Conversely, markers of protonephridia and epidermal tissue showed no change in expression, and markers of epithelial tissue showed increased expression after 55 days of RNAi treatment (Fig. 6A,B). Whole-mount in situ hybridization confirmed that expression of a marker for protonephridia differentiation is not reduced after Smed-Sirt-1(RNAi) (Oviedo and Levin, 2007), but suggested that there may be dysregulation in the expression of markers corresponding to different segments of protonephridia (Fig. 6C). To test this hypothesis, we measured the expression levels of marker genes expressed in protonephridia tubules (proximal and distal), and collecting ducts (Fig. 6D,E). The results demonstrate that there was a tendency for upregulation of markers for the proximal tubules upon RNAi treatment. This would be consistent with downregulation of Smed-Egr5 expression, because loss of Smed-Egr5 function has been shown to induce proliferation of neoblasts accompanied by an increase in the expression of protonephridia markers (Tu et al., 2015). We found that Smed-Sirt-1(RNAi) animals displayed a 22% and 27% decrease in the expression of Smed-Egr5 at 15 and 55 days of treatment, respectively (Fig. 6F). Taken together, Smed-Sirt-1 may regulate tissue-specific transcriptome dynamics in the protonephridia.

To confirm that, in addition to protonephridia, other tissues such as intestine were also compromised after Smed-Sirt-1(RNAi), we evaluated expression of the putative cytoskeletal regulators Smed-tpm-1 and Smed-rho-A. We found that Smed-Sirt-1(RNAi) animals displayed ∼30% decrease in the expression of both Smed-tpm-1 and Smed-rho-A (Fig. 6G,H), which are critical regulators of gut branching (Forsthoefel et al., 2012). To address the possibility that reduced expression of the intestinal differentiation markers might affect the gut architecture, we used fluorescent in situ hybridization with the intestinal marker Smed-inx-9 (Oviedo and Levin, 2007). This approach allowed us to appreciate gut morphology in detail, using the main branch length as a reference for quantification of the number of sub-branches within the anterior and posterior regions (Fig. 6I) (Barberán et al., 2016; Forsthoefel et al., 2011). Our results show that Smed-Sirt-1(RNAi) animals had fewer tertiary intestinal branches (Fig. 6I–K). We also observed that the primary posterior inner sub-branches in the posterior branch length were missing or severely reduced in early and late phases of the Sirt-1(RNAi) phenotype (Fig. 6J,K). These results suggest that Smed-Sirt-1 is required for maintaining proper morphology of the planarian digestive system.

Our results show that Sirtuin-1 is evolutionarily conserved in S. mediterranea and that it regulates organismal growth by affecting feeding behavior and gut morphology. At the cellular level, we found that Smed-Sirt-1 is critical to maintain the balance between cell division and cell death during planarian growth. We also reveal that Smed-Sirt-1 may have a role in adult stem cell biology and the maintenance of intestinal function, as summarized in Fig. 7. Taken together, our findings highlight the planarian S. mediterranea as a tractable model for analysis of the behavioral and molecular mechanisms integrating sirtuin activity and nutrient availability.

The extended number of sirtuin homologs in S. mediterranea offers the opportunity to analyze evolutionary aspects of this conserved signaling pathway in metazoans. Unlike in other invertebrate model organisms (such as C. elegans and D. melanogaster), we demonstrate that the number of sirtuins in planarians is almost equivalent to the number in humans. Furthermore, the SIRTUIN domains of human SIRT1 and S. mediterranea Sirtuin-1 showed a similar degree of conservation (∼60%) as that shared between human SIRT1 and the sirtuins of C. elegans and D. melanogaster, studies of which have expanded our understanding of sirtuin function (Frye, 2000). It is expected that further analysis using S. mediterranea, a member of the Lophotrocozoa clade, will complement work in other species and provide unique insights about the evolution of sirtuins in metazoans.

Our results demonstrate that it is possible to modulate sirtuin function in planarians by genetic and pharmacological approaches. Methods for gain of function studies are limited in planarians, but the use of RESV treatment as an enhancer of Sirtuin-1 function allowed us to overcome this hurdle. Conversely, the specific effects achieved with RNAi could also be induced with NAM treatment, which is commonly used to inhibit sirtuin function in other experimental models (Avalos et al., 2005; Bitterman et al., 2002; Green et al., 2008; Sauve and Schramm, 2003). NAD is required for various cellular functions, such as ADP-ribosylation by PARPs and CD38-mediated signaling, and is not limited to being a cofactor for sirtuins (Okabe et al., 2019). Therefore, it is not surprising that NAM treatment led to a more dramatic effect than Smed-Sirt-1(RNAi) alone. Taken together, our findings introduce a simplified experimental platform amenable for in vivo systemic analysis of Sirtuin-1 function. Given the ever growing use of RESV as supplement in humans, this experimental platform could be relevant for gaining precise knowledge about the potential side effects and risks of RESV treatment (Bhatt et al., 2012; Brasnyó et al., 2011; Knop et al., 2013; Liu et al., 2014; Nguyen et al., 2009; Sanchez-Fidalgo et al., 2012).

The function of sirtuins in planarians can be analyzed at the organismal, cellular and molecular level. Studying the effects of sirtuins in the complexity of the adult body is advantageous because it allows comprehensive analysis integrating behavioral, neural, and metabolic inputs. This was evident in the behavioral effects associated with food consumption and the resulting decrease in body size observed after disrupting Smed-Sirt-1 function. The underlying mechanisms regulating food consumption in planarians remain largely unknown. Studies in different organisms consistently show that feedback between the nervous and digestive systems influences food intake (Cummings et al., 2001; Kohno et al., 2003; Mikani et al., 2012). For example, in humans and other vertebrates, orexigenic molecules such as ghrelin are secreted from the stomach and trigger a hunger response through the arcuate nucleus of the hypothalamus, stimulating release of neuropeptide Y (NPY) and agouti-related peptide (AgRP) from NPY/AgRP neurons (Inui et al., 2004; Klok et al., 2007). In planarians, NPY homologs exist, and some of these NPY homologs, along with various other neuropeptides, appear to regulate feeding behavior (Collins et al., 2010; Shimoyama et al., 2016). Recent studies also revealed the transcription factor friend leukemia integration 1 transcription factor (Smed-FLI1) to be a potential regulator of the planarian chemosensory response (Roberts-Galbraith et al., 2016). The newly developed protocols to evaluate feeding stimulation reported in this study showed that within a few minutes of sensing food, the expression of genes encoding the NPY homolog Smed-npy3 and the insulin-like growth-factor binding protein subunit Smed-ALS was increased. The expression of these genes was further enhanced with RESV treatment, and decreased in Smed-Sirt-1(RNAi) animals, suggesting these components might participate in a primary feeding response that is regulated by Smed-Sirt-1. The feeding response appears to be incremental over time because we found similar trends in expression of Smed-npp22 and Smed-spp5 after 30 min of feeding stimulation. The mechanisms underlying the observed incremental changes in gene expression are not readily evident, but our data might suggest a feedback loop involving Smed-npp22 and Smed-spp5. Although these results validated the efficacy of our feeding stimulation assay, further biochemical and genetic studies will be required to determine the epistatic regulation of these components and to identify how Smed-Sirt-1 and the neurohumoral response are integrated to regulate the hunger response.

Use of S. mediterranea as a model provides alternative opportunities to analyze Sirtuin-1 function in the context of tissue renewal. For example, our findings revealed that inhibition of Smed-Sirt-1 leads to an imbalance between cell division and cell death during tissue turnover. There is very limited information available about how Sirtuin-1 affects cell fate and the flux of cells between progenitor and differentiated cell populations during the constant renewal of adult tissues. The mechanisms underlying the cellular imbalance induced by inhibition of Smed-Sirt-1 are not clear; but, based on our results, we speculate that continuous neoblast proliferation may be promoted by the deficient differentiation of cells toward the gut lineage. This possibility is supported by findings in mice demonstrating that Sirtuin-1 regulates stem cell function in the intestinal epithelium (Igarashi and Guarente, 2016; Igarashi et al., 2019). Likewise, Smed-Sirt-1 regulation of cell cycle dynamics may alter timely differentiation of specific sets of progenitors in multiple tissues such as intestine, protonephridia and epithelia.

Several differentiation markers have been identified in planarians, including for lineages of different cell types (Barberán et al., 2016; Forsthoefel et al., 2012; Fraguas et al., 2011; Tasaki et al., 2011; Tu et al., 2015; Zeng et al., 2018; Zhu and Pearson, 2018; Zhu et al., 2015). We found Smed-Sirt-1(RNAi) reduces expression of the gut differentiation marker Smed-GST1, which was accompanied by formation of fewer intestinal branches. These effects in intestinal morphology appeared to be independent of Smed-egfr-1 and its ligand Smed-nrg-1, which are critical for gut branching (Barberán et al., 2016). Nonetheless, our findings suggested that defective gut morphology in Smed-Sirt-1(RNAi) animals might be associated with a reduction in expression of two putative cytoskeletal regulators, Smed-tpm-1 and Smed-rho-A, that are known to regulate branching formation in intact and regenerating animals (Forsthoefel et al., 2012). These findings, taken together, imply that Smed-Sirt-1 might regulate gut branching through regulation of both differentiation and cytoskeletal organization. We speculate that expansion of the gut may interfere with total food consumption, limiting growing capacity. Collectively, our findings reveal novel opportunities to study the systemic effects of Sirtuin-1 activity on the integration of long-range intercellular communication (for example, neuronal and stem cell signaling) that affects homeostasis and organismal behavior. Future studies would be guided toward dissecting the genetic regulatory networks modulating Sirtuin-1 in the adult body.

The inhibition of sirtuin-1 in different model organisms often leads to decreased lifespan, making it challenging to study the systemic effects of such inhibition in the adult body (Aström et al., 2003; Dang, 2014; Kaeberlein et al., 1999; Lemieux et al., 2005; Li et al., 2007; McBurney et al., 2003; Sun and Dang, 2016; Viswanathan et al., 2005). Intriguingly, disrupting Smed-Sirt-1 with RNAi or NAM treatment does not seem to affect lifespan nor induce early death in planarians. The underlying mechanism controlling lifespan in planarians is unknown, but recent findings link metabolic inputs with the extraordinary capacity of neoblasts to regenerate their telomeres by suppressing TOR signaling (Iglesias et al., 2019). We found that long-term downregulation of Smed-Sirt-1 results in suppression of Smed-TOR expression (Fig. S7E). Sirtuin-1 has previously been identified as a negative regulator of mTOR signaling (Ghosh et al., 2010). These observations could be analyzed further in planarians by testing whether Smed-Sirt-1 modulates neoblast telomere elongation through indirect regulation of TOR signaling.

In invertebrates, insulin/insulin growth factor signaling (IIS) through insulin-like peptides, plays major roles in regulating growth during development (Dabour et al., 2011; Murphy and Hu, 2013; Vallejo et al., 2015). IIS and its downstream components AKT and TOR are known regulators of growth in the planarian model (Gonzalez-Estevez et al., 2012; Miller and Newmark, 2012; Peiris et al., 2016b; 2012; Tu et al., 2012). Our results revealed that Smed-Sirt-1 is also a regulator of organismal growth in S. mediterranea. This finding is consistent with impaired growth in sirtuin-1 null mice mediated by insulin growth factor signaling (Lemieux et al., 2005; Li et al., 2007; McBurney et al., 2003). We also found that animals subjected to Smed-Sirt-1(RNAi) display an important reduction in insulin signaling components, particularly Smed-IGFBP3, Smed-IGFBP5 and Smed-ALS (Fig. S7D,E). In Drosophila, ALS plays a role not only in growth, but also in metabolism (Arquier et al., 2008). Similarly, humans with mutations in ALS were shown to have a reduction in birth weight and postnatal growth, followed by a delay in puberty (Domené et al., 2009). It is possible that ALS acts as a metabolic sensor or regulator of growth in planarians under the control of Smed-Sirt-1, and that loss of Smed-Sirt-1 leads to impaired ALS production, disrupting growth in the early part of the Smed-Sirt-1(RNAi) phenotype. However, further analyses using planarian-specific antibodies would provide a complementary view to confirm the current gene expression studies.

The planarians used were Schmidtea mediterranea, asexual CIW4 strain. Cultures were maintained as previously described (Oviedo et al., 2008a).

Sirtuins were identified using a BLAST search where human sirtuin protein sequences were queried against available genomic resources for S. mediterranea (Brandl et al., 2016; Robb et al., 2015; Rozanski et al., 2019). Identified sequences went through a six-frame translation using the Pfam protein domain database (http://pfam.xfam.org/) and domain conservation was confirmed using both UNIPROT (https://www.uniprot.org) and PROSITE (https://prosite.expasy.org/). The sequences were further validated by BLASTn and BLASTp in NCBI (http://blast.ncbi.nlm.nih.gov/Blast.cgi). The confirmed sequences were aligned by CLUSTALW (MEGA 7 software) using sequences obtained using HomoloGene (www.ncbi.nlm.nih.gov/homologene). Presentation of protein alignment was done using SeaView software (http://doua.prabi.fr/software/seaview). A predictive phylogenetic tree was built using MEGA7 software (www.megasoftware.net).

Sirtuin gene expression analysis was obtained via the Planaria single-cell database hosted by the Reddien Lab at the Whitehead Institute for Biomedical Research (https://radiant.wi.mit.edu/app/) (Wurtzel et al., 2015) and https://digiworm.wi.mit.edu/ (Fincher et al., 2018).

The synthesis of dsRNA was performed as previously described (Oviedo et al., 2008b). Microinjections were performed following the schedule in Figs 1A,B, 2D, and Fig. S2B. All animals were starved for at least one week prior to RNAi treatment and at least 5 days after any RNAi experiments.

All treatment experiments were done via soaking. Drugs were diluted into 50 ml of planarian water (Tasaki et al., 2016) and administered every other day for the length of the experiment. All nicotinamide (Acros Organics, Cat.# 128271000) experiments were performed at a 100 µM concentration, with the control group receiving planarian water alone. All resveratrol (TCI America, cat. # R0071-1G) experiments were performed at a 20 µM concentration, with the control group receiving planarian water containing 0.04% DMSO. All animals were soaked in drugs for 5 days prior to assays, unless otherwise noted.

Only intact animals were used for all growth and degrowth (i.e. decrease in size upon starvation) experiments. During all growth experiments, animals approximately 2 mm in length were fed liver paste once a week. Feedings were carried out following the schedules in Figs 1B and 2A,D. For RNAi experiments, the initial feeding began 15 days following the first injection. In the case of drug-soaking experiments, animals were soaked 5 days prior to the first feeding, unless otherwise noted. Live images were taken either every 5 days for RNAi experiments or every week for drug soaking experiments. During degrowth experiments, animals approximately 8 mm in length were starved for the entirety of the timecourse. Drug-soaking starved animals were soaked on the first day of the experiment. Injected animals were injected on the first day of the experiment. Live images were taken every 5 days, beginning 15 days following the first injection. All live images were taken using a Nikon AZ-100 multizoom microscope and NIS Elements AR 3.2 software. Animal area measurements were calculated using ImageJ (NIH, MD) and the difference in animal size was determined as fold change in reference to the control group at each time point.

A custom-made dish was used containing ten lanes approximately 1 cm wide by 10 cm long that were filled to half their depth with planarian water. In the middle of each lane two drops of liver paste (Oviedo et al., 2008a), 10 µl each, were placed. Video recording began before the addition of animals. Individual animals were placed into each end of the lane, for all lanes, with two animals total per lane. Lanes 1–5 contained the control group and lanes 6–10 contained the experimental group. Animals were fed for 1 h and then removed from the lanes. The time taken to begin feeding was quantified as the time, in seconds, between the animal being placed into the lane, and the extension of the pharynx out onto the liver to ingest the food. This space and filming equipment was provided by Dr Fred Wolf UC Merced, CA.

Surface area measurements were taken for all animals prior to feeding. Animals were fed liver paste containing 0.2% Erioglaucine disodium salt dye (Sigma, Cat.# 861146-5G) overnight. The same results could be obtained by feeding animals for 1 h (data not shown). Prior to homogenization, dishes containing animals were cleaned and animals were placed into 2 ml centrifuge tubes with 1.5 ml of PBS and a small number of glass beads (Chemglass Life Sciences, cat. # CLS-1835-BG5). Samples were homogenized using a Bead Ruptor homogenizer (Omni international, cat. # 19-040) (provided by Dr Aaron Hernday and Dr Clarissa Nobile, UC Merced, CA) for 30 s. Samples were centrifuged at 20,817 g for 1 min. After centrifugation, 1 ml of supernatant was used to measure activity at 620 nm absorbance. Food consumption was quantified by taking the absorbance reading for the sample and dividing it by the animal area (mm²) of the individual animals in each group. These values were normalized and expressed as fold change between groups.

Homogenized liver paste (50 µl) was added to 50 ml of planarian water and repeatedly vortexed. The mixture was left to settle for 10 min at room temperature. Using a serological pipette, 40 ml of the 0.1% liver water was taken, avoiding anything that sank to the bottom of the tube, and was transferred to clean dishes. Animals that underwent stimulation were carefully transferred to these dishes containing 0.1% liver water for the described length of time. Non-stimulated control animals were transferred to clean dishes containing 40 ml of planarian water. Following this incubation period, animals were removed, and placed into TRIzol (Invitrogen, Cat.# 15596026) for RNA extraction.

All animals were starved for at least 5 days prior to any experiment. For staining with H3P antibody, animals were killed in 5.7% HCl solution for 10 min on ice. Animals were washed twice with 0.05% PBSTx (1× PBS containing 0.05% Triton X-100) on ice and then fixed in 3% formamide and 6% H₂O₂ in 0.05% PBSTx for 30 min at room temperature. Fixative was removed and worms were bleached overnight in a solution containing 6% H₂O₂ in 0.05% PBSTx. Primary antibody, anti-H3P (Millipore cat. # 05-817R), was used at 1:500. Secondary antibody, HRP-conjugated goat anti-rabbit (Millipore cat. # 12-348), was used at 1:1000. Neoblasts were counted and normalized to the total area of each animal (mm²) using ImageJ. For staining with acetylated tubulin, animals were killed in 5.7% HCl solution for 5 min on ice. HCl was replaced with Carnoy's solution (6:3:1 ethanol:chloroform:acetic acid) for 2 h on ice. Animals were transferred into 100% methanol and stored at −20°C for 1 h. Animals were bleached overnight in 6% H₂O₂ in methanol. Primary antibody, anti-acetylated tubulin (Sigma, clone 6-11B-1) was used at 1:500. Secondary antibody, HRP-conjugated goat anti-mouse (Invitrogen, cat. # G21040), was used at 1:500. Images were taken using a Nikon AZ-100 multizoom microscope and NIS Elements AR 3.2 software.

Quantitative PCR was performed as previously described (Peiris et al., 2012). TATA-box-binding protein domain was used as an internal control. Each individual experiment consisted of triplicates for each condition, and experiments were independently repeated twice. RNA was extracted from intact animals (≥20 per condition) using TRIzol (Invitrogen, Cat.# 15596026) with a small number of glass beads (Chemglass Life Sciences, Cat# CLS-1835-BG5). Samples were homogenized using a Bead Ruptor homogenizer (Omni international, cat. #19-040) for 20 s and converted into cDNA using the Verso cDNA synthesis kit (Thermo Scientific, Cat.#AB1453A). Gene expression was expressed as fold change in comparison to the control group. See Table S1 for details of oligonucleotides used.

Riboprobes for in situ hybridization were synthesized using T3 or T7 polymerases (Thermo Fisher, Cat.# FEREP0103 and FEREP0111) and digoxigenin-labeled ribonucleotide mix (Roche, Cat.# 11277073910) using specific PCR templates as previously described (Pearson et al., 2009). Whole mount in situ hybridization and fluorescence in situ hybridization were performed as previously described (King and Newmark, 2013). See Table S2 for details of probe sequences used. Images were taken using a Nikon AZ-100 multizoom microscope and NIS Elements AR 3.2 software.

Animals were placed into 2 ml centrifuge tubes with 250 µl of 1X RIPA buffer (Cell Signaling Technologies, Cat.# 9806) and a small number of glass beads (Chemglass Life Sciences, Cat.# CLS-1835-BG5). 1X RIPA buffer was supplemented with Complete Mini Protease Inhibitor Cocktail (Roche, cat. # 04693124001), 1 mM PMSF and 1 mM DTT. Samples were homogenized using a Bead Ruptor homogenizer (Omni international, cat. # 19-040) for 20 s. Samples were incubated on ice for 30 min and then centrifuged at 20,817 g for 20 min at 4°C. The supernatant was transferred to a new tube and placed on ice. A 10 µl aliquot of the supernatant was set aside to measure protein concentration using a Bradford protein assay (VWR, cat. # E530-1L). The remaining supernatant was mixed with equal volumes of 2X Laemmli buffer (4% SDS, 10% β-mercaptoethanol, 20% glycerol, 0.004% Bromophenol Blue and 0.125 M Tris–HCl) and incubated at 94°C for 10 min to denature and reduce. Protein lysates were stored at −20°C.

Protein lysate aliquots of 20 µg were heated at 94°C for 5 min and loaded into a 15% SDS–PAGE gel along with a molecular weight marker (Thermo Scientific, cat. # 26619). Samples were transferred to a PVDF membrane that had been activated with methanol for 1 min (Bio-Rad, cat. # 162-0175), overnight at 30 V in 1× Tris–glycine transfer buffer [25 mM Tris base, 192 mM glycine and 20% (v/v) methanol] at 4°C. The membrane was blocked with 5% BSA in TPBS [Tris-buffered saline, pH 7.6, 20 mM Tris-base, 140mM NaCl, 0.1% Tween 20] for 2 h and incubated in the primary antibodies overnight at 4°C on a rocker. Primary antibodies used were anti-β-tubulin E7 (1:10,000; Developmental Studies Hybridoma Bank, cat.# AB2315513), anti-caspase (1:500; Abcam, cat.# ab13847). The membrane was washed four times for 20 min prior to the addition of secondary antibodies: IRDye 800RD goat-anti-rabbit IgG (1:2000; LI-COR, Cat.# 925-32211) for anti-caspase and IRDye 680RD goat-anti-mouse IgG (1:10,000; LI-COR, cat. # 925-68070) for anti-β-tubulin E7. Blots were imaged using a LI-COR 9120 Odyssey Infrared Imaging System. Band quantification was performed by taking band intensities and computing the area under the curve, using ImageJ. Cleaved caspase activity was normalized to β-tubulin.

The Tunel assay was performed as previously described using the ApopTag Red In Situ Apoptosis Detection Kit (Millipore, Cat.# S7165) (Pellettieri et al., 2010). Images were taken using a Nikon AZ-100 multizoom microscope and NIS Elements AR 3.2 software.

Planarians were dissociated in a calcium-magnesium-free (CMF) solution and FACS was carried out as previously described (Thiruvalluvan et al., 2018; Peiris et al., 2016a,b). Samples were run using LSRII flow cytometer with BD FACSDiva software (BD Biosciences). Analysis of data was performed using FlowJo software (version 8) (Thiruvalluvan et al., 2018; Peiris et al., 2016a,b).

One-way ANOVA, two-way ANOVA or t-test statistics were performed and data are shown as the mean±s.e.m. or fold change±s.e.m., unless otherwise noted. All statistics were performed by pooling biological replicates from each technical replicate. together for analysis. One-way ANOVA tests were performed using multiple comparisons to compare the mean of each column with the mean of a control column. Two-way ANOVA tests were performed within each row with compared columns (simple effects within rows), and compared each cell mean with the control cell mean of that row. ANOVA tests did not use matching or pairing of data in rows and data was assumed to have a Gaussian distribution. Dunnett's test was used to correct for multiple comparisons. All statistical analysis was performed using Prism7, Graphpad software Inc. (http://www.graphpad.com).

We thank Edelweiss Pfister for technical assistance and members of the Oviedo laboratory for comments on the manuscript. We would like to thank Dr Fred Wolf from UC Merced for sharing filming equipment and space for behavioral feeding assays, as well as Dr Aaron Hernday and Clarissa Nobile from UC Merced for sharing the bead rupture for RNA and protein extractions, and feeding dye assays. Results and discussion segments in this paper are reproduced from the PhD thesis of Benjamin Ziman (University of California, Quantitative and System Biology Graduate Program 2019).

The peer review history is available online at https://jcs.biologists.org/lookup/doi/10.1242/jcs.239467.reviewer-comments.pdf