Reciprocating intestinal flows enhance glucose uptake in C. elegans

Despite its physiological and pathological importance, the mechanical relationship between glucose uptake in the intestine and intestinal flows is unclear. In the intestine of the nematode Caenorhabditis elegans, the defecation motor program (DMP) causes reciprocating intestinal flows. Although the DMP is frequently activated in the intestines, its physiological function is unknown. We evaluated the mechanical signature of enhanced glucose uptake by the DMP in worms. Glucose uptake tended to increase with increasing flow velocity during the DMP because of mechanical mixing and transport. However, the increase in input energy required for the DMP was low compared with the calorie intake. The findings suggest that animals with gastrointestinal motility exploit the reciprocating intestinal flows caused by peristalsis to promote nutrient absorption by intestinal cells.

Nutrient absorption is indispensable for animal survival and offspring prosperity. In most animals, the intestine is the major site for digestion of ingested food and nutrient absorption 1,2 ; it is one of the oldest organs in the history of evolution. The intestinal functions of digesting food and absorbing nutrients are driven by intestinal peristalsis, which is mediated by periodic muscular contractions and relaxations 3 . Experimental studies and numerical simulations have revealed relationships among intestinal movement, intestinal flow, and mixing of gut contents [4][5][6][7][8][9][10][11][12][13][14][15][16][17][18][19][20] . For example, peristalsis in the intestine of zebrafish larva contributes to transport and mixing of the gut contents 4,5 . Higher intestinal flow rates and larger intestinal volumes enhance the absorption of nutrients such as glucose, vitamins, and minerals in rat intestine in vivo [6][7][8][9][10][11][12] . Experiments involving human participants, rats, and pigs in vivo, as well as rat and pig intestines in vitro, have shown that glucose uptake and insulin secretion after feeding decrease as the viscosity of gut contents increases [13][14][15][16][17][18][19][20] . However, the relationships between intestinal movements and nutrient uptake in the intestine are unclear because of difficulties in the visualization and reproduction of nutrient absorption in the intestines during digestion.
As a model organism, Caenorhabditis elegans has been used for studies of food digestion and nutrient absorption [21][22][23][24][25] . C. elegans is a small (approximately 1-mm-long) multicellular terrestrial nematode and a filter feeder 21,[26][27][28] . C. elegans worms have several advantages for biological and biomedical research. Their small body size and short life span (approximately 3 weeks) facilitate their maintenance 28 . The C. elegans genome has 60-80% homology with the human genome, despite the invertebrate nature of the animal 29,30 . The transparent C. elegans body allows visualization of tissues and cells by microscopy 28 . The only glucose transporter identified in C. elegans is FGT-1 25 , a homolog of the mammalian GLUT-2, which mediates transmembrane glucose transport by concentration gradient-mediated passive diffusion 31 . FGT-1 transporters are located on the basolateral membrane of intestinal cells facing the pseudocoel 25 . Although the glucose transporters that mediate glucose transport into intestinal cells from the intestinal lumen have not been identified in C. elegans 25 , whose transports are performed by SGLT in humans 31 , those transporters are considered to mediate glucose in C .elegans like SGLT. Therefore, the homology of genes and glucose transporters between humans and C. elegans makes this animal useful for experimental studies of intestinal nutrient uptake. In C. elegans, food bacteria ingested by the pharynx are digested and absorbed by microvilli in the intestinal lumen [32][33][34] (Fig. 1A). Subsequently, C. elegans executes defecation by the periodic action of a stereotyped motor program that involves contractions of three sets of muscles; this program is regarded as the defecation motor program (DMP) 35 . The DMP functions as follows 36 ( Fig. 1B-K). The DMP begins with contractions of body wall muscles near the tail, which causes retrograde flow of gut contents toward the pharynx (Fig. 1G, H). Next, while the posterior body wall muscles relax, the anterior body wall muscles contract, causing anterograde flow of gut contents toward the anus (Fig. 1I). Finally, while the anterior body wall muscles relax, the muscles that open the anus contract, expelling through the anus some of www.nature.com/scientificreports/ the gut contents as feces (Fig. 1J). The total duration of the DMP is approximately 5 s; it is repeated at intervals of 45 ± 3 s (mean ± standard deviation) in healthy wild-type worms 35,36 . Because digestion and nutrient uptake associated with microbiota in C. elegans has been evaluated as a model experiment of pathogenic bacteria on symbiotic microbiota 37 , the mechanical role of DMP might have contributions involving nutrient uptake and host health 38 .
In C. elegans, tracer particles pass through the intestine in 3-10 min 21 . The mean gut residence time is reportedly < 2 min; approximately 43 ± 10% of the maximum volume of the intestinal lumen is expelled through the anus during each activation of the DMP 22 . The estimated time constant needed for digestion of bacteria is 14 ± 4 s 24 . We previously reported that a C. elegans worm must eat at least three bacterial cells per second to survive for 3 days, which implies the consumption of hundreds of thousands of bacteria daily 39 . The intestinal flows caused by the DMP have been proposed to facilitate nutrient uptake by intestinal cells 23 , but the mechanical relationship between gut-content movements and nutrient uptake in the intestine has not been elucidated. Therefore, the effects of intestinal flows on nutrient uptake by intestinal cells are unclear.
Here, we used fluorescent glucose and scaling analysis to investigate the effects of intestinal flows on glucose uptake by intestinal cells in C. elegans. We employed fluorescence microscopy to measure the distribution and amount of fluorescent glucose entering intestinal cells. We also used a high-speed camera and differential interference contrast (DIC) microscopy to observe intestinal flows caused by the DMP. Finally, we quantitatively evaluated the effects of intestinal flows on glucose uptake in the intestine.

Results
Distribution and uptake of glucose in intestinal cells. To evaluate the effects of intestinal flows on glucose uptake by intestinal cells, three defecation-defective mutants were used: unc-16(e109), which was isolated in genetic screens for Unc mutants 35 and is defective in contractions of body wall muscles around the anterior intestine, and therefore the anterograde flow caused by the DMP is slow and weak; egl-8(sa47), which was isolated in genetic screens for DMP mutants and is defective in contractions of body wall muscles around the posterior intestine, and thus the retrograde flow caused by the DMP is slow and weak; and exp-1(sa6), which  www.nature.com/scientificreports/ was isolated in genetic screens for DMP mutants and is defective in contractions of anus muscles, and therefore feces are expelled approximately once every six DMP cycles 35 . First, we measured the distribution of glucose in intestinal cells using fluorescent glucose and a high-sensitivity camera. The wild-type strain N2 and the defecation-defective mutants unc-16, egl-8, and exp-1 accumulated 0.13% (v/v) Escherichia coli OP50-1 labeled with fluorescent glucose in the intestines after 30 and 60 min ( Fig. 2A-H). In the N2 intestine, the fluorescence intensity was highest in the most anterior part of the intestine; the intensities were similar among other parts ( Fig. 2A, B). In unc-16 and egl-8 intestines, the glucose distribution was similar to the distribution in N2 (Fig. 2C-F). In the exp-1 intestine, the fluorescence intensity was higher at both ends than in other parts (Fig. 2G, H). www.nature.com/scientificreports/ Next, to quantify the effects of intestinal flows caused by the DMP on glucose uptake by intestinal cells, we analyzed fluorescence images of the intestine using ImageJ software (Fig. 2I, J). To measure glucose fluorescence in intestinal cells, the fluorescence intensity in the intestinal lumen was subtracted from the total fluorescence in the intestine. At 30 min, the fluorescence intensity was highest in exp-1; the fluorescence intensity in N2 was similar to the intensities in unc-16 and egl-8 (Fig. 2I). In contrast, at 60 min, glucose fluorescence was higher in N2 than in the mutants (Fig. 2J). Moreover, we calculated the amount of glucose uptake per unit length of the intestine by applying the trapezoidal rule to plotted curves of the fluorescence of glucose in each image (Fig. 2K, L). At 30 min, there was no difference in intestinal glucose uptake between N2 and unc-16 or egl-8 ( Fig. 2K). At 60 min, although glucose uptake by the three mutants decreased or was unchanged, glucose uptake by N2 increased and became higher than uptake by the mutants (Fig. 2L).
Accumulation of fluorescent particles in the intestine. We next measured the fluorescence intensity of 0.5-μm-diameter fluorescent particles in the intestine after feeding for 5, 15, 30, 40, and 60 min. At 5 min, fluorescence was observed only in unc-16 (Fig. 3A1, B1, C1, D1). At 15 min, fluorescence was distributed throughout the intestine in N2 and unc-16, while it was localized to the anterior or posterior part in exp-1 and egl-8 (Fig. 3A2, B2, C2, D2). At 30 min, the particles were distributed throughout the intestine, except in egl-8 (Fig. 3A3, B3, C3, D3). At 40 and 60 min, the particles were distributed throughout the intestine in all strains (Fig. 3A4, A5, B4, B5, C4, C5, D4, D5). Although the timing and rate of increase in particle accumulation differed among strains, the total fluorescence of accumulated particles in the worm was similar among the four strains ( Fig. 3E-H).
Intestinal flows during the DMP. We next visualized intestinal flows during the DMP by DIC microscopy ( Fig. 4A-D, see Supplementary moves). In the N2 intestine, when the DMP began, the particles in the posterior part of the intestine experienced retrograde flow caused by the DMP; thus, they were stored in the anterior part ( Fig. 4A1-A3). Next, they were transported to the anus through anterograde flow caused by the DMP; they were stored in the posterior part of the intestine (Fig. 4A4, A5). Finally, some particles in the posterior part of the intestine were expelled through the anus (Fig. 4A6). In the unc-16 intestine, when the DMP began, particles in the posterior part of the intestine experienced retrograde flow and were stored in the anterior part and first half of the posterior part of the intestine (Fig. 4B1-B3). Next, these stored particles were transported to the anus through anterograde flow and stored in the posterior part of the intestine (Fig. 4B4, B5). Finally, some particles in the posterior part of the intestine were expelled to the outside of the body (Fig. 4B6). In the egl-8 intestine, when the DMP began, particles in the first half of the posterior part of the intestine experienced retrograde flow and were stored in the anterior part of the intestine (Fig. 4C1-C3). Next, they were transported to the anus through anterograde flow and were stored in the posterior part of the intestine (Fig. 4C4, C5). Finally, some particles in the posterior part of the intestine were expelled through the anus (Fig. 4C6). In the exp-1 intestine, the particle flows were similar to the flows in N2 (Fig. 4D). Therefore, intestinal flows during the DMP differed according to whether the DMP was defective and the nature of any defect.
To quantify the differences in intestinal flows during the DMP, we measured the flow velocities of particles flowing from the posterior to the anterior part of the intestine by DMP-induced retrograde flow, as well as the flow velocities of particles flowing from the anterior to the posterior part of the intestine through anterograde flow; we also measured the migration distances of the group of particles. There was no significant difference between N2 and exp-1 in terms of anterograde flow velocity (P > 0.05 by Welch's t-test) (Fig. 4E, F). Therefore, anterograde flow velocity is not dependent on the ability to expel gut contents. However, the anterograde flow velocity differed between N2 and unc-16 (Student's t-test), retrograde flow velocity differed between N2 and egl-8 (Welch's t-test) and between N2 and exp-1(Welch's t-test), and anterograde flow differed between N2 and unc-16 (Welch's t-test) and between N2 and egl-8 (Welch's t-test). Thus, intestinal flow velocities during the DMP differed according to defects in retrograde and anterograde flow caused by the DMP. Furthermore, the retrograde and anterograde flow velocities were significantly higher in N2 than in unc-16 or egl-8.
There was no significant difference between N2 and exp-1 in terms of migration distance by retrograde or anterograde flow (P > 0.05; Welch's t-test or Student's t-test) (Fig. 4G, H). Therefore, migration distances by retrograde and anterograde flow do not depend on the ability to expel gut contents. However, migration distances by retrograde and anterograde flow significantly differed between N2 and unc-16 (Welch's t-test), while migration distances by retrograde and anterograde flow significantly differed between N2 and egl-8 (Welch's or Student's t-test). Thus, retrograde or anterograde particle migration distance caused by the DMP was dependent on the presence of a functional retrograde or anterograde flow mechanism. Additionally, the gut content migration distances by retrograde and anterograde flows were significantly greater in N2 than in unc-16 or egl-8.
Relationship between glucose uptake and intestinal flows. We next evaluated the relationship between glucose uptake and flow velocity or migration distance. There was an association between glucose uptake by intestinal cells at 60 min and the mean intestinal flow velocity (Fig. 5A). Glucose uptake tended to increase with increasing flow velocity. There was an association between glucose uptake by intestinal cells at 60 min and the mean particle migration distance (Fig. 5B). Therefore, glucose uptake increased with increasing migration distance.
To quantify the effects of DMP-induced intestinal flows on glucose uptake by intestinal cells, we used two dimensionless numbers: the Péclet number (Pe) and the Sherwood number (Sh) 40,41 . Pe is defined as the ratio of transport caused by advection (i.e., flow velocity) to transport caused by molecular diffusion. Pe is given by:   There was an association between Pe and glucose uptake by intestinal cells (F G ) at 60 min (Fig. 5C). Therefore, F G increases in proportion to the square of Pe (F G = 9.80 × 10 −2 Pe 2 ). Sh is defined at an interface as the ratio of the mass transfer rate by advection and diffusion relative to the mass transfer rate by diffusion alone. Sh is given by: where U m is the permeation velocity at the interface, L is the characteristic length, and D is the diffusion coefficient. By assuming that mass absorption occurs on the inner surface of a circular tube of diameter R and length L, U m (averaged over the surface πRL) can be expressed as: where m s is the mass flow velocity through the surface. In this study, m s was the amount of glucose taken up by intestinal cells per unit time; it was assumed to be proportional to the fluorescence intensity of glucose in intestinal cells (F G ). R was determined to be 15 µm in this study and a previous study 43 .
By combining (2) and (3), and using the relations m s ∝ F G ∝ Pe 2 under constant R and D conditions, Sh can be rewritten as: In Fig. 5C, Sh is plotted on the vertical axis; it is proportional to the square of Pe. To quantify DMP energetics, we estimated the energy input required to pump the gut contents during the DMP. The input energy E can be derived by applying Darcy-Weisbach's formula: www.nature.com/scientificreports/ where P is the pressure drop in the intestinal lumen, Q is the flow rate, and T is the duration of retrograde or anterograde flow during the DMP. By assuming that the intestine is a straight tube of circular cross-section and that the Hagen-Poiseuille law is applicable 39,44 , P is given by: where Re is the Reynolds number, L m is the particle migration distance by DMP-induced retrograde or anterograde flow, R is the inner diameter of the intestinal lumen, U is the velocity of retrograde or anterograde flow during the DMP, ρ is the density of the gut contents, and μ is the viscosity of the gut contents. Thus, Q is given by: Therefore, Eq. (5) can be rewritten as: There was an association between the glucose fluorescence intensity in intestinal cells at 60 min, F G , and the input energy E during the DMP obtained from Eq. (8) (Fig. 5D). Thus, the amount of glucose taken up tended to increase with increasing energy input.

Discussion
We investigated glucose uptake in the C. elegans intestine. When using a high glucose concentration 1.3% (v/v), we found no significant difference in glucose uptake among the four strains (see Supplementary Information). The lack of a difference in glucose uptake between the wild-type and mutant strains is likely because of saturated glucose fluorescence. In contrast, when using glucose at 0.13% (v/v), glucose uptake at 60 min was considerably higher in N2 than in the mutants (Fig. 2J, L). Therefore, the difference in glucose uptake between N2 and the mutants might be caused by differences in the amounts of gut contents or in the characteristics of DMP-induced intestinal flow. Note that at a glucose concentration of 0.13%, hyperglycemia did not appear to have a significant effect on the regulation of glucose uptake by worms. This is because the worms have the capacity to absorb more glucose, since they absorbed nearly three times as much glucose at a glucose concentration of 1.3%.
At 40 and 60 min, there was minimal variation in the amounts of gut contents among N2 and the mutants (Fig. 3E-H). This suggests that a defective DMP does not affect the amounts of gut contents. Thus, the differences in glucose uptake between N2 and the three mutants were not caused by differences in the amounts of gut contents.
Next, we analyzed the intestinal flows induced by the DMP (Fig. 4). The gut retrograde-and anterograde-flow velocities were significantly higher in N2 than in unc-16 or egl-8 (Fig. 4E, F). The retrograde and anterograde migration distances of the gut contents were significantly greater in N2 than in unc-16 or egl-8 (Fig. 4G, H). Moreover, glucose uptake by intestinal cells tended to increase with increasing flow velocities or migration distances (Fig. 5A, B). Therefore, the difference in glucose uptake between N2 and the three mutants resulted from different DMP-induced intestinal flow characteristics.
Glucose uptake increased with the square of Pe (F G = 9.80 × 10 −2 Pe 2 ) (Fig. 5C). Importantly, mass transport into a wall induced by flow in a tube with peristalsis increases with the square of Pe in a large Pe regime 45 (see Supplementary Information). A similar scaling (i.e., Taylor dispersion) has been reported 46 , in which apparent diffusivity in a tube is enhanced by flow with a factor of Pe 2 . Considering that F G and Sh both depend on Pe, the DMP-induced intestinal flows significantly promote glucose uptake by intestinal cells, compared with molecular diffusion alone.
Finally, glucose uptake increased with increasing energy input (Fig. 5D). Moreover, we previously reported that the energy obtained by digestion of a single OP50-1 bacterium was 7.0 × 10 −7 kcal (1 kcal = 4.184 kJ); the number of bacteria required to survive for 3 days was approximately three cells per second 39 . Using these values, we estimate the minimum calorie intake of a worm in 1 h to be 31.6 J. Therefore, minimal energy (in the order of picojoules) was used for glucose uptake compared to the energy obtained from ingested bacteria. Because glucose uptake increased with increasing intestinal flow velocities, N2 likely had the greatest capacity for glucose uptake among the four strains.
In summary, glucose uptake by intestinal cells was higher in N2 than in unc-16, egl-8, or exp-1. Glucose uptake tended to increase increasing intestinal flow velocities during the DMP. Glucose uptake and the Sherwood number increased with the square of Pe. Glucose uptake also tended to increase with increasing energy input, despite far greater calorie intake. We conclude that C. elegans exploits DMP-induced intestinal flows to promote glucose uptake by intestinal cells, which may be the primary function of the DMP.

Materials and methods
Worm preparation. The wild-type strain N2 was obtained from Dr. Shohei Mitani (Tokyo Women's Medical University School of Medicine, Tokyo, Japan). The mutant strains unc-16(e109), exp-1(sa6), and egl-8(sa47) were obtained from the Caenorhabditis Genetics Center (University of Minnesota, Minneapolis, MI, USA), and had been used without further outcrossing to the wild-type strain. The worms were maintained on nematode growth medium plates at 20 °C using standard protocols 26,28 . Worms were removed from nematode growth medium by washing with 700 µL of M9 buffer 28 . OP50-1 cells were cultured in liquid Luria-Bertani (LB) medium in an incubator at 37 °C with shaking at 200 rpm for 12-14 h 28 .  49 . We measured the distributions of 0.13% (v/v) OP50-1 labeled with fluorescent glucose in the intestine at 30 and 60 min, respectively. Note that the intestinal lumen was manually distinguished using the brightness gap between the lumen and the intestinal cells. The glucose uptake was evaluated by subtracting the total fluorescence intensity in the lumen from that in the whole worm. The background noises in the image were initially subtracted. Visualization of intestinal flows. L4 or young adult worms (n = 5-10) were transferred to 200 μL of M9 buffer containing 0.013% (v/v) fluorescent particles (diameter, 1.0 µm; excitation wavelength, 505 nm; fluorescence wavelength, 514 nm; Thermo Fisher Scientific) mixed with 2% (v/v) OP50-1 in a chamber (surface area, 9 × 9 mm 2 ; Bio-Rad Laboratories) on a microscope slide (surface area, 24 × 60 mm 2 ; thickness, 0.12-0.17 mm; Matsunami). The suspension was covered by a coverslip (surface area, 24 × 24 mm 2 ; thickness, 0.17 mm; Matsunami) to minimize dehydration. Particle movement in the intestine during the DMP was observed using an upright DIC microscope (BX51WI, Olympus) with a high-speed camera (SA3, Photron, Japan), a 20 × dry objective lens (UPlanSApo 20 × ; NA, 0.75; WD, 0.6 mm; Olympus), a halogen lamp (U-LH100HGAPO, Olympus), a power supply unit (U-RFL-T, Olympus), and a mirror (excitation, 479-495 nm; emission, 510-550 nm; U-MWIBA3, Olympus) 39,50 . High-speed image sequences of approximately 45 s duration at 60 fps were recorded using a high-speed video camera (see Supplementary moves). We manually traced the positions of the particles in the intestine during the DMP using ImageJ software 49 . From the frame numbers and pixel numbers, we calculated the time-averaged particle flow velocities and the mean time-averaged particle velocity through division of the particle migration distance by the number of frames.

Statistical analysis. Statistical analysis was performed using Microsoft Excel 2019 (Microsoft Corporation,
Redmond, WA, USA) on macOS Sierra (Apple Inc., Cupertino, CA, USA). A value of n indicates the number of the worms, not the number of particles tracked in total. A value of p < 0.05 was considered to indicate statistical significance.

Data availability
The data that support the findings of this paper are available from the corresponding author upon request. Supplementary information is provided with the paper to support the experimental results.