Abstract
Background Given the growing interest in human exploration of space, it is crucial to identify the effect of space conditions on biological processes. The International Space Station (ISS) greatly helps researchers determine these effects. However, the impact of the ISS-introduced potential confounders (e.g., the combination of radiation and microgravity exposures) on the biological processes are often neglected, and separate investigations are needed to uncover the impact of individual conditions.
Results Here, we analyze the transcriptomic response of Caenorhabditis elegans to simulated microgravity and observe the maintained transcriptomic response after return to ground conditions for four, eight, and twelve days. Through the integration of our data with those in NASA GeneLab, we identify the gravitome, which we define as microgravity-responsive transcriptomic signatures. We show that 75% of the simulated microgravity-induced changes on gene expression persist after return to ground conditions for four days while most of these changes are reverted after twelve days return to ground conditions. Our results from integrative RNA-seq and mass spectrometry analyses suggest that simulated microgravity affects longevity regulating insulin/IGF-1 and sphingolipid signaling pathways.
Conclusions Our results address the sole impact of simulated microgravity on transcriptome by controlling for the other space-introduced conditions and utilizing RNA-seq. Using an integrative approach, we identify a conserved transcriptomic signature to microgravity and its sustained impact after return to the ground. Moreover, we present the effect of simulated microgravity on distinct ceramide profiles. Overall, this work can provide insights into the sole effect of microgravity on biological systems.
Background
Significant efforts have being made for the exploration of space. As such, NASA and SpaceX have planned round-trip landings by 2030s and placement of a million people in 40 years to Mars. However, space conditions present different types of environmental stress such as increased levels of radiation and reduced gravity (i.e., microgravity) that negatively affect biological processes by causing changes in the transcriptome leading to bone loss, muscle atrophy, and immune system impairment, to name a few examples. To overcome this problem, the affected biological mechanisms in space conditions should be uncovered to take precautions before the colonization of Mars.
The International Space Station (ISS) has provided a great opportunity to study the impact of space conditions on biological systems. However, potential confounders at the ISS are generally neglected. For example, experiments in the ISS introduces elevated radiation and reduced gravity along with the hypergravity during the launch. These mixed stress factors make it highly challenging to detect what precisely causes the observed phenotypes. To identify how the space conditions affect the biological systems, each of the factors require a separate investigation.
Microgravity is one of the major stress factors causing detrimental health effects on humans. Simulation of microgravity is a cost-effective way to study the impact of microgravity on biological systems. For this purpose, various platforms including drop towers, parabolic flights and space flights are available. Because of their advantages to enable long term simulations in an effective and economical manner, clinostats are widely preferred to study gravitational response in biological systems, mainly plants, for decades [1–5]. Clinostats rotate a sample around a horizontal axis, thereby exposing the sample to a rotating gravitational vector. This can reduce and even possibly remove gravitational bias in the development of an organism [3, 6–10].
C. elegans is an ideal model organism for space biology studies with its completely documented cell lineage, high reproduction rate, short lifespan, and high similarity to the human genome [11–16]. In the public bioinformatics repository of NASA (GeneLab; https://genelab.nasa.gov), four studies from three missions to the International Space Station (ISS) are available from C. elegans (Table 1). However, the experimental design of these studies highly varies. For example, these studies adopted different food source (e.g., axenic or monoxenic liquid media), duration of study, developmental stage, or the worm strain while the usage of microarray technology was common for all. We and others have reported substantial variations in the transcriptomic and phenotypical responses to different liquid cultivations of the worms [11, 17]. We additionally showed that two strains of C. elegans (N2 and AB1) exhibit highly different transcriptomic profiles under liquid cultivations [11, 16]. Overall inconsistency among the space biology experiments on C. elegans highlights the necessity of a more organized and uniform experimental design for future studies where each stress factor is investigated individually.
This study investigates the sole effect of simulated microgravity on C. elegans transcriptome and the sustained impacts after return to ground control conditions. We report a group of putative microgravity-responsive genes, we define as the gravitome, through the integration of our in-house data with the re-analyzed C. elegans data from GeneLab. Through the RNA-sequencing (RNA-seq) and mass spectrometry analyses, we reveal the downregulation of the sphingolipid signaling pathway and its potential function in longevity under simulated microgravity.
Results
To dissect the biological processes affected under simulated microgravity and sustained after the exposure, we first cultured C. elegans in CeHR for three weeks on ground control condition. Then, we exposed the worms to clinostat-simulated microgravity for four days (Fig. 1a; Additional file 2: Movie S1) and observed the maintained impacts at four, eight, and twelve days after placing the worms back to ground conditions (Fig. 1b). RNA-seq was performed for each condition with three replicates (Additional file 3: Dataset S1). Because RNA-seq can detect low abundance transcripts and achieve less noise in the data, unlike previous space biology studies on C. elegans, we preferred RNA-seq over microarray [18].
We generated a dendrogram for the transcriptomic profile in each condition tested and found significant differences during and after exposure to simulate microgravity compared to the ground control conditions (Fig. 1c). Thus, this result suggests that exposure to simulated gravity induces highly distinct gene expression patterns which are maintained even after 12 days return to ground conditions (approximately three generations of C. elegans in axenic medium [19]).
Simulated microgravity triggers differential expression of hundreds of genes
To identify the genes with the most distinctive expression levels after the exposure, we determined the differentially expressed genes (DEGs). The genes with over two-fold log2 expression difference with the FDR-adjusted p-value ≤ 0.05 were considered as DEGs. Hundreds of genes demonstrated differential expression during exposure to simulated microgravity and up to eight days after return to ground conditions (Fig. 1d). Twelve days after the return, the gene expression levels started to resemble the ones in the ground control with only 91 upregulated and 13 downregulated genes.
The determination of the spatial expression of the DEGs is of great importance to identify the potentially affected tissues. Thus, we performed tissue enrichment analysis for the DEGs in simulated microgravity (Fig. 2a). The downregulated genes were overrepresented in neuronal and epithelial tissues, and intestine while the upregulated genes were enriched in the reproductive system-related tissues.
Next, we conducted pathway enrichment analysis for the DEGs in simulated microgravity and return to ground conditions to identify the altered pathways and whether they are remained to be altered after return to ground conditions (Fig. 2b). Our results suggested an upregulation of dorso-ventral axis formation and downregulation of lysosome during the exposure, and these expression patterns were maintained for four and eight days after the return, respectively.
Simulated microgravity-induced gene expression differences are highly maintained for eight days after return to ground conditions
We then identified the number of simulated microgravity-induced DEGs that preserved their expression patterns after the return (Fig. 2c). The majority of the DEGs (approximately 75%) from the exposed animals maintained their expression patterns after the return for four days. The shared number of DEGs decreased drastically at 12 days after the return to 16% for commonly upregulated and <1% downregulated genes with the exposed animals.
To elucidate the genes showing altered expression under simulated microgravity and maintaining these expression patterns after return to ground conditions for short (4 days) and long-term (12 days), we categorized the DEGs in Venn diagrams (Fig. 2d, e; Additional file 4: Dataset 2). The genes solely upregulated in the exposed animals did not exhibit enrichment for any gene ontology (GO) term. The genes upregulated during the exposure and four days after the return, however, showed an overrepresentation for reproduction-related processes (Fig. 2d). Chitin metabolic process-related genes were induced at four days after return to ground conditions, and the expression profiles were conserved for eight days after the return.
We found that the downregulated genes were enriched for the biological processes which are affected in space conditions. In particular, genes functioning in body morphology, collagen and cuticulin-based cuticle development, defense response, and locomotion were downregulated under simulated microgravity and up to eight days after return to ground conditions (Fig. 2e). In simulated microgravity and spaceflight studies, both small movement defects [20] and no difference in locomotory behavior [21, 22] have been noted. The reason behind these discrepancies is unclear, but our results indicated differences in the expression of the locomotion genes in close to weightless environment. Similarly, neuropeptide signaling pathway genes (flp-22, flp-8, flp-26, nlp-10, nlp-12, nlp-17, nlp-20, nlp-24, nlp-25, nlp-26, nlp-39, nlp-33, nlp-28, nlp-29, nlp-30, and flp-24) were downregulated during the exposure, and this expression profile was lost after return to ground conditions (Fig. 2e). These neuropeptide signaling pathway genes also exhibit enrichment for movement variant phenotypes (Additional File 1: Fig. S1) indicating altered locomotion in response to simulated microgravity.
During the spaceflight, the neuromuscular system and collagen have been negatively affected [20]. In agreement with these findings, collagen genes were downregulated in our experiment during the exposure, and the downregulation was sustained for eight days after return to ground conditions. Previously, Higashibata et al. (2006) reported decreased expression for the myogenic transcription factors and myosin heavy chains in space-flown worms [20]. We did not observe differential expression for these muscle-related genes. Since their reported flight to ground gene expression ratio for the downregulated genes was less than one, and we only consider the genes with the log2 ratio greater than two as DEG, it is expected not to observe those genes in our group of DEGs. To test whether other muscle-related genes are differentially expressed under simulated microgravity, we compared our DEGs to 287 genes reported in WormBase for involvement in muscle system morphology variant or any of its transitive descendant terms via RNAi or variation. Only T14A8.2, col-103, and sqt-3 among the 287 genes showed downregulation under simulated microgravity (hypergeometric test, p-value = 0.99). Thus, our results did not suggest a significant change in the expression of the genes known to function in muscle morphology.
In our previous work, we revealed that the expression of many ncRNA molecules is triggered in response to environmental changes [16]. To identify whether a similar pattern occurs under simulated microgravity, we determined the expressed ncRNA molecules in ground control, simulated microgravity, and return. Since the housekeeping gene pmp-3 presents consistent expression patterns [16], we considered its minimum expression level (FPKM = 24.7) in our experiment as the cutoff for the expression of ncRNA molecules. The number of expressed ncRNA molecules slightly decreased in response to simulated microgravity (Additional file 1: Fig. S2a). Interestingly, this number increased drastically in the eighth day of return. We found that the expression of 126 ncRNA molecules is induced in simulated microgravity and for 12 days after the return (Additional file 1: Fig. S2b) while the expression of 16 ncRNA molecules is lost during and after simulated microgravity (Additional file 1: Fig. S2c). Among the classified ncRNA molecules, mostly snoRNA and lincRNA molecules are induced while other sets of snoRNA and asRNA molecules lose expression in simulated microgravity. For instance, asRNA molecules anr-33, K12G11.14, and ZK822.8 are induced whereas anr-2, anr-9, and Y49A3A.6 are silenced during and 12 days after the exposure.
We next sought to identify the putative transcriptional regulators (i.e., transcription factors) of the simulated microgravity-induced genes and whether these regulators maintain their impact after return to ground conditions. That is, we determined the putative transcription factor (TF) genes [23] that are upregulated under simulated microgravity and after the return. Our results have revealed that 20 TF genes are upregulated during the exposure and 90%, 75%, and 15% of these genes maintained their upregulation after the return for four, eight, and twelve days, respectively (Additional File 1: Fig. S3). These TFs play a role in a variety of mechanisms such as double strain break repair or sex determination. For example, tbx-43 is upregulated during the exposure and for at least 12 days after the return. The best human BLASTP-match of tbx-43 in WormBase, TBR, functions in developmental process and is required for normal brain development (E-value = 3e-33; identity = 68.5%). However, many of the upregulated TFs do not have a known function. Hence, our list of simulated microgravity-induced TF gene expressions can be a rich source for the discovery of the microgravity-related transcriptional regulators.
Longevity regulating pathways are affected under simulated microgravity
Spaceflight affects the mechanisms involved in delayed aging in worms. For instance, Honda et al. found that age-dependent increase of 35-glutamine repeat aggregation is suppressed and seven longevity-controlling genes are differentially expressed in spaceflight [24, 25]. To examine whether simulated microgravity induces such changes, we mapped the DEGs to longevity regulating pathway in the Kyoto Encyclopedia of Genes and Genomes (KEGG). We used a less stringent criterion for differential expression by including the genes with log2(FPKM) > 1.5 in this analysis. We determined 11 DEGs in the longevity regulating pathway genes most of which are involved in the insulin/insulin-like growth factor signaling pathway (insulin/IGF-1)(Fig. 3). Interestingly, the transcription factor DAF-16 gene did not exhibit differential expression unlike its targets sod-3, lips-17, and ctl-1. Because the translocation of DAF-16 into the nucleus activates or represses target genes functioning in longevity, metabolism, and stress response, this finding was unexpected [26–29]. The nuclear localization of DAF-16 is antagonized by transcription factor PQM-1 [30], and thus an upregulation in pqm-1 may indicate inhibition of DAF-16 translocation. Our data, however, did not present a differential expression for pqm-1 and thus it does not suggest inhibition of the DAF-16 translocation. It is possible that the DAF-16 targets are differentially expressed due to translocation state of DAF-16 or involvement of other factors (e.g., other transcriptional regulators).
The upregulation of the longevity regulating pathway transcriptional targets generally contributes to increased lifespan. Since many of these targets were downregulated under simulated microgravity, we wished to identify the potential impact of their downregulation by acquiring their RNAi phenotype from WormBase [31]. Lifespan variant (WBRNAi00063155 and WBRNAi00063156) for hsp-12.6 and extended lifespan (WBRNAi00064044) for sod-3 have been reported. Follow-up experiments are needed to fully understand the role of these genes in the longevity regulation under simulated microgravity.
To further investigate the involvement of the longevity genes in simulated microgravity, we compared the DEGs to the DAF-16-responsive genes [30]. We found 144 DAF-16-induced genes were downregulated and this overlap was statistically significant (hypergeometric test, p-value < 0.0001). The number of shared genes between the simulated microgravity-induced genes and the genes induced or repressed by DAF-16 are 11 and 35, respectively (hypergeometric test, p-value > 0.05 for both). Similarly, the number of downregulated DAF-16-repressed genes were insignificant with a total of 57 shared genes (hypergeometric test, p-value > 0.05) (Additional File 5: Dataset S3). Together, our results suggest that the longevity regulation genes are affected under simulated microgravity and that along with the DAF-16-regulation, other mechanisms have an important function in this process.
Sphingolipid signaling pathway is suppressed in response to simulated microgravity
Previous studies have suggested that the sphingolipid signaling pathway plays a role in the expression of DAF-16/FOXO-regulated genes [32, 33]. In line with this, our results showed that the sphingolipid signaling pathway is downregulated under simulated microgravity. That is, putative glucosylceramidase 4 gene (gba-4) and putative sphingomyelin phosphodiesterase asm-3 are downregulated and these downregulation patterns were sustained for eight days after return to ground conditions (Fig. 4a). This downregulation pattern indicates an attenuation in the ceramide levels through a potential decrease in the degradation of sphingomyelin and glucosylceramide to ceramide. Along with the other biological functions such as autophagy, senescence, and apoptosis, the sphingolipid signaling pathway has critical functions in ageing and longevity regulation [32, 34, 35]. The inhibition of this conserved pathway results in an extension of lifespan in animals from worms to humans [34, 35]. For example, the inactivation of asm-3 in C. elegans causes translocation of DAF-16 into the nucleus, promotion of DAF-16 target gene expression, and extension of lifespan by 14% - 19% [33].
To validate the ceramide and sphingosine levels are indeed decreased in response to the aforementioned gene downregulations, we performed mass spectrometry analysis in ground control and simulated microgravity-exposed worms concurrently (Additional file 6: Dataset S4; Fig. 4b,c). Our results revealed that total ceramide, total hexosylceramide (HexCer), and d18:1 sphingosine are lower (1.5-, 3.6-, and 1.4-fold, respectively) under simulated microgravity (t-test, p-value < 0.05). Similarly, sphingoid base (SB) levels decreased 1.9-fold under simulated microgravity, but this decrease was not significant (t-test, p-value = 0.055) (Fig. 4b). Interestingly, the levels of total lactosylceramide (LacCer) increased by 27-fold (t-test, p-value < 0.05). Increase in the LacCer and HexCer levels have been determined as biomarkers of aging in humans and murine [36] while LacCer C18:1 was unaffected during aging in C. elegans [34]. It is unclear why LacCer levels are elevated when the HexCer levels are decreased, but the overall pattern indicates a downregulation in the sphingolipid signaling pathway under simulated microgravity.
The levels of long acyl-chain ceramides (≥C24) are elevated with advancing age and in age-related diseases such as diabetes and cardiovascular disease while C20 and C26 ceramides are unaffected with aging or developmental stage in C. elegans [34, 37, 38]. To further decipher the potential effect of simulated microgravity on aging-related mechanisms, we quantified the levels of different acyl-chain ceramides (Additional file 6: Dataset S4; Fig. 4c). We observed a decrease in d18:1-C24, and d18:1-C24:1 ceramides (1.2-, and 3.2-fold, respectively) under simulated microgravity (t-test, p-value < 0.05). It has been found that the inhibition of C24-C26 ceramide-inducing ceramide synthase HYL-1 causes improvements in neuromuscular function and age-dependent hypoxia and stress response [39, 40]. Hence, the lower levels of C24 and C24:1 ceramides in the simulated microgravity-exposed worms may indicate their positive impact on the stress response and aging of the worms.
Along with their roles in aging and longevity, different acyl-chain ceramides have other distinctive functions. For example, C16-ceramide induces germ cell apoptosis [41] and C20-C22 ceramide functions in resistance to hypoxia [40]. We determined that ceramides with different acyl-chain showed altered levels in response to simulated microgravity. While d18:1-C16 and d18:1-C18 exhibited a decrease (2.8- and 1.6-fold, respectively), d18:1-C20 and d18:1-C22 ceramides exhibited an increase (3.7- and 1.5-fold, respectively) under simulated microgravity (t- test, p-value < 0.05). Collectively, our findings suggest that sphingolipid and insulin/IGF-1 pathways play a role in simulated microgravity response.
Identification of the “gravitome” through comparison of the microgravity-exposed worm gene expression profiles
To investigate the “gravitome”, we compared the results from our simulated microgravity experiment to those from previous spaceflight experiments on C. elegans reported in NASA’s GeneLab database (https://genelab.nasa.gov/). For this analysis, we used GEO2R to determine the microarray DEGs (|log2 fold change (logFC)| > 2 and adj. p-value < 0.05) from the first International C. elegans Experiment in Space (ICE-First), C. elegans RNA Interference in Space Experiment (CERISE) four and eight days, and PRJNA146465 conducted by UKM. We reasoned that the genes that are commonly differentially expressed in all the studies should be the ones responding to microgravity. Surprisingly, the number of shared DEGs were highly limited among the studies (Fig. 5a). For instance, there were no shared DEGs between ICE-First and the other previous studies, and only five shared DEGs between four and eight days of CERISE. The experimental design of these studies has substantial differences including food source, developmental stage, strain, and exposure time (Table 1). These differences can be the underlying reason causing the highly distinct DEG profiles. For example, in the four-day culture of CERISE, approximately 9,000 L1 worms were used whereas in the eight-day culture, the number of worms was 30 to 50 L1 [42]. Additionally, because the worms become gravid in four days in axenic culture [19, 43], the eight-day culture probably has mixed developmental stage worms as oppose to the synchronized worms in the four-day culture. As expected, the four- and eight-day CERISE cultures showed profound difference in the overall gene expression profiles, as well (Fig. 5b).
Higashibata et al., have reported reproducible gene expressions for myosin, paramyosin, unc-87, dim-1, sirtuin and its target genes from four-day CERISE and ICE-First [44]. We wished to evaluate the expression of these genes in eight-day CERISE, PRJNA146465, and our study, as well. Additionally, we assessed the expression levels of these genes when a different analysis method (i.e., GEO2R) is used and noted the reported expression levels in different manuscripts for the same studies (Table 2). None of the genes were in our list of DEGs and most of them showed logFC < 1 between the conditions. Furthermore, some genes had high differences in the logFC or discrepancy for up- or downregulation among the experiments. For example, pqn-5 was reported to be 4.18-fold upregulated in the four-day CERISE while it was 0.9-fold downregulated in the ICE-First. The same gene was upregulated 0.25-fold in PRJNA146465 (Table 2). These findings highlight the necessity of uniformity in the experimental design and a detailed evaluation of the currently used techniques for achieving more reproducible and accurate results.
We analyzed our RNA-seq data both with the Tuxedo pipeline [46] and DEseq2 [47] to confirm that the results are not dependent upon the analysis method. The results from both the methods were clustered together indicating their similarity (Fig. 5b). The highest number of DEG overlap was between our experiment and four-day CERISE, and the overlap was significant for both the methods (hypergeometric test, p-value < 0.05) (Fig. 5c). Since the overlap between the four-day CERISE and our results from the Tuxedo pipeline is higher, we decided to use the results from this pipeline throughout the manuscript. The results from DEseq2 are reported in the supplementary data (Additional File 1: Fig. S4; Additional File 7: Dataset S5).
We determined a total of 134 common DEGs between our experiment and four-day CERISE, eight of them also being common to PRJNA146465, and categorized these 134 genes as the putative gravitome (Fig. 5a, c; Additional File 8: Dataset S6). We discarded 16 DEGs showing conflicting patterns between the experiments (i.e., upregulated in one experiment and downregulated in the other) for further analyses. We reasoned that if the remaining genes are collaboratively involved in a biologically relevant process, they might have protein-protein interactions (PPIs) with each other. To test this hypothesis, we examined the enrichment for PPI of these genes by using the STRING database [48]. Among 118 DEGs, 58 had known PPIs with each other (FDR < 1.0e-16) indicating their collaborative involvement in a biological process (Fig. 5d). The GO analysis revealed that the gravitome genes play a functional role in locomotion, body morphogenesis, and collagen and cuticulin-based cuticle development (Fig. 5e). Similarly, the gravitome gene products exhibit enrichment for nematode cuticle collagen N- terminal and collagen triple helix repeat domains (Fig. 5f). Together, our findings suggest that mainly the collagen genes are affected under microgravity and this effect is reproducible between the studies.
Next, we asked whether the human orthologs of the gravitome genes have relevant functions which might be affected in astronauts. We identified 64 human orthologs to 44 gravitome genes in worms (Additional File 9: Dataset S7). The human orthologs exhibited overrepresentation for protein domains including calpain large subunit domain III and calpain family cysteine protease (Additional file 1: Fig. S5). Given the reported upregulation of calpain under microgravity and its function in muscle atrophy [49, 50], these results provide an additional support for a conserved gravitome in worms and humans.
Our analyses revealed that 20 putative TF genes are upregulated under simulated microgravity conditions. We wondered whether these upregulation patterns are shared in the four-day CERISE. Among the 20 upregulated putative TF genes, Y56A3A.28 and T24C4.2 showed upregulation during the four-day CERISE, indicating a potential transcriptional regulation role for these two in response to microgravity. Our results additionally suggested that the upregulation pattern of these TFs are maintained for eight and twelve days after the return, respectively. We could not find a human ortholog for T24C4.2 while the WormBase-suggested human ortholog of Y56A3A.28 is PRDM16. PRDM16 regulates the switch between skeletal muscle and brown fat cells by inhibiting the skeletal muscle development and gene expression and stimulating brown adipogenesis [51]. Therefore, Y56A3A.28 can be a strong candidate as a negative regulator of the muscle development under microgravity conditions.
Discussion
In this study, we examined the simulated microgravity responsive gene expression patterns and their maintained levels for four, eight, and twelve days after return to ground conditions. Longevity regulating pathways such as insulin/IGF-1 and sphingolipid signaling pathway were affected under simulated microgravity. We identified a putative gravitome by determining the common microgravity responsive transcriptomic signatures by incorporating our study to the previous ones. Moreover, we revealed that the gravitome is potentially conserved between worms and humans by performing function analysis for the human orthologs of the worm gravitome and identifying some of their potential transcriptional regulators.
Sphingolipid signaling pathway plays crucial biological roles such as lifespan regulation, apoptosis, and oxidative-stress response [32, 34, 52]. Our results revealed an overall downregulation of the sphingolipid pathway which is generally related to an increase in longevity [34]. A recent study identified decreased ASM-1 and ASM-2 levels in the ISS-housed worms indicating a similar downregulation in response to microgravity [53]. Mutant strains or RNAi knockdown of acid sphingomyelinase ASM-1, ASM-2, and ASM-3 genes increase the lifespan, ASM-3 being the most prominent one. Overall, both simulated and the ISS-introduced microgravity seem to contribute to a potential increase in longevity through the downregulation of the sphingolipid signaling pathway. Follow-up experiments are needed to delimitate the contribution of sphingolipids on longevity under microgravity.
In the Twins Study of NASA, Scott Kelly was sent to the ISS for 340 days while his identical twin Mark Kelly remained on Earth. The preliminary results from this investigation showed that Scott Kelly’s telomere length significantly increased during the mission. The telomere length of Mark Kelly, however, remained relatively stable [54]. Telomere length and lifespan have a strong relationship. That is, longer telomeres are linked to longer lifespan and increased resistance to environmental stress while shorter telomeres are linked to accelerated aging and reduced longevity [55–58]. Ceramide functions in the regulation of telomere length [35, 59]. Considering our findings on decreased ceramide levels and the observations on increased telomere in the ISS, investigation of the ceramide and telomere length relationship might prove important to uncover the mechanisms affecting longevity in space.
Most of the space flight-triggered physiological changes are reverted after return to Earth [60]. Similarly, we observed that the majority of the simulated microgravity-induced DEGs (over 84%) are reverted 12-days after return to ground conditions. Our results have revealed that while some biological processes are only affected during the exposure, others can be affected for at least eight days (approximately two generations) after the return. For example, neuropeptide signaling pathway genes were differentially expressed only during the exposure whereas defense response genes remain downregulated for eight days after the return. Approximately 50% of the astronauts from Apollo mission experienced minor bacterial or viral infections at the first week of their return. Later studies reported that space flight-induced reactivation of latent herpes viruses which lasted for a week after the return to Earth (reviewed in [61]). These consistent observations indicate that the detrimental effects of space conditions (i.e., microgravity) on immune system carry over even after the return. Future investigations could provide insight into whether the duration of the exposure has an impact on the duration of the lasting effects.
The limitation of our study is the inclusion of mixed stage worms and the collection of samples from whole animals. This limitation is common among the aforementioned worm experiments conducted in the ISS. Considering the significantly different gene expression patterns among developmental stages and across cell types [62], it is possible to retrieve biased results from the averaged gene expression levels. It is essential to address these issues in future studies.
Our integrative approach to define the gravitome has shown that the gene expression profiles widely vary among the experiments. Similarly, the comparison of spaceflight responsive genes in C. elegans and Drosophila melanogaster demonstrated only six common genes from the European Soyuz flights to the ISS [63]. The reason for these discrepancies can be the batch effects introduced due to the differences in the experimental designs. For instance, different exposure times to the space conditions or the usage of different food source or strain can induce highly distinct responses in the worms. Furthermore, during the missions to the ISS, the organisms experience drastic environmental changes including exposure to radiation, different temperature, and gravitational forces (i.e., hypergravity and microgravity). It is crucial to examine methods used for achieving the most accurate and reproducible results and to identify the effects of individual factors (e.g., diet and gravitational force). Towards that end, we previously reported the impact of liquid cultivation of two C. elegans strains on gene expression and phenotype [11, 16] and here, studied the impact of simulated microgravity.
Conclusions
Given the elevated interest in the human exploration of space, it is crucial to determine the detrimental effects of the space conditions on biological systems. Our study demonstrates that C. elegans is an efficient model to identify the impact of space on biological systems. We believe that our results will be a valuable reference for the studies on transcriptomic response to microgravity by controlling for the renowned batch effects from space conditions, allowing the worms to acclimate to liquid cultivation before the experiment, presenting the sustained transcriptomic responses after return to ground along with their putative regulators, integrating the results from the previous studies conducted at the ISS, and providing the response of different ceramide profiles to simulated microgravity.
Materials and Methods
C. elegans strain and growth conditions
Wild-type N2 strain was obtained from the Caenorhabditis Genetics Center (CGC). The worms were grown at 21°C. Stocks of C. elegans were acclimated to CeHR medium for three weeks prior to microgravity experiments. The ground control animals were maintained in CeHR medium in 20 mL scintillation vials as described [16]. Approximately 100,000 live worms (P0 generation) were cleaned by sucrose floatation, their embryos (F1 generation) were harvested by bleaching and placed in ∼40 mL of CeHR culture in sterile VueLife culture bags (“AC” Series 32-C, Saint-Gobain). Culture bags were mounted in the clinostat and rotated at 1 rad/sec (∼ 10 RPM) for four days, which is the measured time to reach L4/adulthood [19]. The “F2” and subsequent generations were approximated by using four-day intervals. Prior to RNA extraction, only live worms are retrieved using the sucrose floatation method. We note that “F2” generation and beyond will not maintain synchronicity, and all will contain a mixture of generations, but recovering only the live animals enriches the RNA pool for the expected (majority) generation as each adult will produce ∼30 offspring. The reason for the usage of mixed stage worms were the difficulties experienced (e.g., contamination of the liquid medium) during the synchronizing of the animals.
RNA isolation, Illumina sequencing
RNA isolation and RNA-seq was performed with triplicates on Illumina HiSeq 2500 in 50bp single-end reads as described previously [16]. As we achieved high correlation for gene expressions with both RNA-seq and qRT-PCR in our previous work [16], we did not perform a separate qRT-PCR for this study.
Gene expression analysis
Quality control on the RNA-seq data was done with FASTQC (version 0.11.2) [64]. All the reads were in the “very good” quality range. The DEGs were identified by using two separate approaches. First, the Tuxedo pipeline [46] was used with default parameters. Second, the DESeq2 software (version 1.16.1)[47] was utilized after quantifying the expression of transcripts with Salmon (version 0.8.2)[65]. The reference genome (WBCel235) along with the annotation file were retrieved from Ensembl [66]. We considered genes as differentially expressed if FDR adjusted p-values <□0.05, and log2 fold change >□2 unless otherwise stated. The results from the Tuxedo pipeline was used throughout the paper as explained in the results and the results from DESeq2 were reported in the supplementary data (Additional File 1: Fig. S4; Additional File 7: Dataset S5). We discarded miRNA, piRNA, and rRNA molecules from the analysis. The ncRNA molecules (WS250) and unconfirmed genes were obtained from WormBase [31]. Dendrogram was generated with Ward’s clustering criterion by using R software.
Functional analysis of the genes
Gene ontology (GO) enrichment for biological processes and protein domains were assessed with Database for Annotation, Visualization and Integrated Discovery (DAVID) (v6.8) [67]. The GO network for the common DEGs between our simulated microgravity experiment and CERISE4 was created in Cytoscape (version 3.4.0) [68] by using the BINGO plugin [69]. Pathway analysis was done using KEGG Mapper (https://www.genome.jp/kegg/tool/map_pathway1.html). Tissue enrichment analysis was performed by using WormBase Gene Set Enrichment Analysis tool [70], and FDR adj. p-value□<□0.05 considered significant and top ten tissues were reported.
Comparison of DEGs to the previous studies in C. elegans
NASA’s GeneLab (https://genelab.nasa.gov/) platform was used to acquire the Gene Expression Omnibus (GEO) accession numbers for the previous studies on microgravity related gene expressions in C. elegans. We identified four such studies with the following GEO accession numbers: GSE71771 (Expression Data from International C. elegans Experiment 1st), GSE71770 (CERISE, 4 days), GSE27338 (CERISE, 8 days), and GSE32949 by UKM (PRJNA146465) from the query made on June 30, 2016 with “C. elegans” and “microgravity” terms on GeneLab. We analyzed these publicly available microarray datasets with GEO2R (https://www.ncbi.nlm.nih.gov/geo/geo2r/) for gene expressions under microgravity at the ISS versus 1G control in four and eight days CERISE and versus ground control conditions for the other two. The genes with logFC > 2 and Benjamini & Hochberg corrected p-value < 0.05 were considered as differentially expressed.
STRING database (version 10.5) [48] was used to determine whether the commonly DEGs from our simulated microgravity experiment and CERISE4 show enrichment for protein-protein interactions (PPI). We only considered the interactions only from experiments, gene-fusion, databases, and co-expression and selected “high-confidence (0.7)” results. PPI enrichment showed significance at p-value < 1.0e-16 with 58 interacting nodes and 155 edges. The PPI network was visualized with Cytoscape (version 3.4.0) [68]. Edge width was determined based on the “combined score” for the interactions obtained from the STRING database. We identified the human orthologs of the gravitome with OrthoList 2 with default settings [71].
Mass spectrometry for ceramides
The worms were chopped manually with razor blades on ice-cold glass. Three replicates for ground control and simulated microgravity conditions with mixed stage worms were used simultaneously. Each sample was sent to Avanti Polar Lipids Analytical Services as a frozen extract in glass tubes for liquid chromatography with tandem mass spectrometry (LC/MS/MS) experiment for detection of Ceramide levels. Each sample was extracted by modified Bligh and Dyer extraction. The samples were dried and resolvated in 50:50 Chloroform:Methanol and diluted prior to analysis. The resolvated sample was used for analyses and stored at −20°C until assayed. Samples were diluted as needed for analysis with internal standards for Ceramides for quantization by injection on LC/MS/MS. The individual molecular species for each sphingolipid group were measured by reversed phased liquid chromatography tandem mass spectrometry methods which separate the compounds by multiple reaction monitoring m/z to fragments and retention time. Quantification was performed by ratio of analyte to internal standard response multiplied by ISTD concentration and m/z response correction factor. The results are expressed as ng/rnl. The consistency among the replicates were examined with Pearson correlation. All but one replicates in both the experiments showed high correlation (R2 > 0.9). The ground control replicate which did not show a strong correlation (around 60%) with the other replicates was excluded from the analysis. The data were analyzed with t-test (two-tailed), and p-value < 0.05 was considered significant. Equal variances were confirmed with Levene’s test at α = 0.05 for all the ceramides tested.
Microgravity simulation with clinorotation
We used a lab-built clinostat with integrated microscope to observe the motion of small objects including microspheres, yeast cells and C. elegans embryos during clinorotation. The trajectory of a spherical particle in a rotating vessel have been described previously in literature [2, 8, 72, 73]. Dedolph and Dipert (1971) separated the motion of a spherical particle into two parts: a circular motion under the influence of gravity, the radius of which is inversely proportion to the angular velocity (ω, rad/sec), and a radial motion due to centrifugal forces, the velocity of which is proportional to ω2 [72]. The equation of motion that describes the orbital trajectories is given by:
In the above equation, λ is the viscous drag coefficient, m is the mass of the object, mw is the mass of the displaced water, t is time, g is the acceleration due to gravity, and q(x, y, z, t) is the rotating reference frame defined as:
As a test of the performance of the clinostat we recorded the motion of fluorescent melamine-formaldehyde microspheres (5.6 µm diameter, 1.51 g/cm3 density; Corpuscular Inc., Cold Harbor, NY) suspended in a 3% bovine serum albumin (BSA) solution (Additional file 10: Movie S2). This solution was injected into an 8-mm diameter, 100-µm deep, coverglass topped chamber mounted on the clinostat microscope. Some of the melamine microspheres adhere to coverglass of the chamber, conveniently providing fiducial markers which we were able to use to compensate for the mechanical noise introduced into the micrographs by the rotation of the clinostat. We measured the terminal velocity of the suspended microspheres to be 3.2 ± 0.2 µm/s (mean ± SEM, n = 8). Figure S6 shows a plot of the measured radius of circular motion of suspended microspheres as a function of inverse angular velocity (black dots), and the calculated radius [2] using the measured mean terminal velocity (blue line). The shaded area encompasses calculated radii for terminal velocities within one standard deviation of the mean.
Funding
This work was supported by NASA grants NNX12AR59G, NNX10AN63H, and NNX13AM08G, awarded to CRS and Sigma Xi grant G2016100191301836, awarded to İÇ. The BIOMIX computing cluster and the Bioimaging center are supported by Delaware INBRE grant (NIH/NIGMS GM103446) and Delaware EPSCoR grants (NSF EPS-0814251 and NSF IIA- 1330446). None of these funding bodies were involved in the design, data collection, analysis, interpretation of data, or in writing the manuscript.
Authors’ contributions
İÇ and CRS designed the study. AJ developed the clinostat. İÇ and JHD conducted the experiments. İÇ performed the analyses and interpreted the results. İÇ, AJ, and CRS wrote the manuscript. All authors read and approved the final manuscript.
Ethics approval and consent to participate
Not applicable as C. elegans do not require ethical permits or consent to participate.
Consent for publication
Not applicable.
Competing interests
The authors declare that they have no competing interests.
Acknowledgements
We are grateful to Michael Moore for his help developing the clinostat. İÇ thanks the colleagues from Comparative and Experimental Approaches to Aging Biology course in the MDI Biological Laboratory for helpful discussions. İÇ acknowledges the University of Delaware for Dissertation Fellowship and Graduate Fellow Award.
Footnotes
Contact information of the authors: İrem Çelen: irem{at}udel.edu, Aroshan Jayasinghe: aroshan{at}udel.edu, Jung H. Doh: jdoh{at}udel.edu, Chandran R. Sabanayagam: chandran{at}udel.edu
Abbreviations
- CeHR
- C. elegans Habitation and Reproduction
- CeMM
- C. elegans Maintenance Medium
- DAVID
- Database for Annotation, Visualization and Integrated Discovery
- DEG
- Differentially expressed gene
- FDR
- False discovery rate
- FPKM
- Fragments Per Kilobase of transcript per Million mapped reads
- GO
- Gene ontology
- ncRNA
- Non-coding RNA
- RNA-seq
- RNA-sequencing
- GC
- Ground control
- MG
- Simulated microgravity
- R1_GC
- Four days after return to ground conditions
- R2_GC
- Eight days after return to ground conditions
- R3_GC
- Twelve days after return to ground conditions