Differentiation and function of cardiac valves in the adult Drosophila heart

In all circulatory systems, valves control and regulate the direction of fluid flow. In mammals, two valves separate the upper atria from the lower ventricles, and two valves regulate blood flow at the entrances to the arteries leaving the heart. In addition, one-way valves in blood and lymphatic vessels prevent the backflow of fluids. In a closed circulatory system, as in many metazoan taxa, the unidirectional blood flow in the heart and vessels is essential for continuous oxygen delivery to all organs and tissues. However, in insects, oxygen transport is uncoupled from the flow of the body fluid. Thus, they possess an open circulatory system for facilitating the distribution of haemolymph within the open body cavity and the transport of components contained therein. Open circulatory systems are low-pressure systems where, in addition to the pumping activity of the heart, the circulation of haemolymph is supported by body movements and the contractile activity of body muscles. The Drosophila heart contains cardiac valves for regulating flow directionality as well. The larval valve separates the heart tube into two distinguishable, consecutive chambers and is formed by only two specialised cardiac valve cells facing each other (Fig. 1). In the adult fly, however, the heart contains three valves, each formed by two cells. The Drosophila heart undergoes partial restructuring during metamorphosis. This remodelling process converts the linear heart tube with one terminal wide-lumen heart chamber into a linear four-chambered heart tube (Fig. 1). The adult heart differs from the larval heart in several histological and physiological aspects described and analysed in numerous studies, some of which we mention here as representatives (Curtis et al., 1999; Lehmacher et al., 2012; Molina and Cripps, 2001; Monier et al., 2005; Rizki, 1978; Sellin et al., 2006; Zeitouni et al., 2007).

The differentiation of the valve cells has been largely overlooked for many years. Recent studies demonstrated how the cardiac valves function in the larval Drosophila heart and how they regulate haemolymph flow. In particular, their striking histology with large intracellular valvosomes, which arise from membrane tubulation and are maintained by redirected recycling endosomes towards the valvosome membrane, make them an attractive model to study intracellular vesicle trafficking (Lammers et al., 2017; Meyer et al., 2022).

Upon metamorphosis, the heart undergoes several remodelling processes, which include changes in the ultrastructure of the contractile cardiomyocytes, i.e. number, orientation and size of myofilaments (Cannon et al., 2017; Lehmacher et al., 2012; Petersen et al., 2020), the differentiation of functional ostia from embryonic precursors in each heart chamber (Cripps and Olson, 2002; Gajewski et al., 2000; Molina and Cripps, 2001; Popichenko and Paululat, 2004; Reim and Frasch, 2005; Sláma, 2010; Wasserthal, 2007), the transdifferentiation of larval alary muscles to ventrally located longitudinal muscles (Bataillé et al., 2020; Boukhatmi et al., 2014; Schaub et al., 2015; Shah et al., 2011), the addition of an abdominal-oriented reversal heartbeat (Dulcis and Levine, 2005; Sláma, 2010; Sláma and Farkaš, 2005; Wasserthal, 2007) and finally, the segmental innervation by peripheral neurons that release cardioactive peptides, ablation of which affects anterograde heartbeat (Dulcis and Levine, 2003, 2005; Dulcis et al., 2005).

Most of the structural changes that the heart undergoes during metamorphosis are undoubtedly linked to the lifestyle of the imago, to help the heart achieve its physiologically required pumping performance. Because the adult heart tube is divided into four chambers with three valves, more complex control of haemolymph flow within the heart compartments is required. In this study, we focus on the differentiation and function of the three cardiac valves in the heart of adult flies. To better understand the role of cardiac valves, we analysed their origin, differentiation, histology and function. We were particularly interested in answering the following questions. (1) Is the single cardiac valve of the larvae abandoned and reformed during metamorphosis, and if yes, from which cellular resource? Or is it retained during metamorphosis and becomes part of the adult heart? (2) When do the second and third cardiac valves develop during metamorphosis, and from which cells? (3) Is the cellular organisation of larval valve cells with large vesicular compartments the same in all adult cardiac valve cells? (4) Are adult valves able to close the heart lumen, and what is their mode of function? To answer these questions, we used live-cell reporters that permit visualisation and analyses of cardiac valve cells in the living animal, including metamorphosis. The reporter lines used in this study derived from regulatory elements from the tailup and hand genes as previously described (Boukhatmi et al., 2014; Jenett et al., 2012; Lammers et al., 2017; Meyer et al., 2022; Sellin et al., 2006; Tao et al., 2007). By combining live imaging with histological and ultrastructural analyses, we found that the larval valve is maintained into adulthood with several changes at the ultrastructural levels. The two newly formed valves of the adult heart start to differentiate within a short time window, identified as 24–25 h after puparium formation (APF), from existing pairs of cardiomyocytes, giving rise to three valves in the imago. Furthermore, all cardiac valve cells displayed the characteristic large intracellular vesicular compartments – valvosomes – which allow the cell to change its shape dynamically upon continuous heartbeat in larvae. We investigated how the activity of the three valves is synchronised and how they open and close the heart lumen upon heart contraction in living animals. We present our current model of adult valve cell function and mode of action based on our analyses of the calcium ion dynamics in living heart valve cells using GCaMP3-reporters.

The following fly stocks were used in this study: handC-GFP, handC-Gal4 (Sellin et al., 2006), handC-mCherry (Paululat and Heinisch, 2012) and 76E11-Gal4 (Dionne et al., 2018). From the Bloomington Drosophila Stock Center (Bloomington, IN, USA), we obtained the following fly stocks: UAS-nGFP (BL4776), UAS-CD4td-GFP (BL35836) and UAS-GCaMP3 (BL32116, BL32236). We obtained UAS-Trol-GFP from the Kyoto Drosophila Stock Center (K110836) and tinC-Gal4 from Manfred Frasch, University of Erlangen, Germany. Finally, white¹¹¹⁸ was used as a control strain.

Larvae and adults were anaesthetised before dissection. The animals were then mounted on Sylgard 184 silicone elastomer plates filled with PBS buffer and dissected from the ventral side. After removing viscera and other unwanted body parts, their tissue was further processed for immuno-staining using standard protocols or direct imaging.

We used a Celena S Digital Imaging System microscope to monitor the formation and development of the adult valve cells (Logos Biosystems, Anyang, South Korea). In brief, selected pupae (stage P1) were cleaned from food and other particles using PBS and a small brush. Afterwards, a single specimen was placed upside down onto a Sylgard Petri dish, equipped with a small piece of wet paper tissue surrounding the specimen and covered by a fitting plate cover. The prepared sample was placed onto the microscope sample holder and covered with the onstage incubation system, which was turned off during time-lapse imaging. For time-lapse imaging, pictures were taken automatically every minute and processed to a 30 frames s⁻¹ video sequence using the microscope's onboard software. Movies were further analysed using Fiji software (Schindelin et al., 2012).

Ready-to-hatch flies were used to capture movies of the heartbeat of late pupal stages (pupal stage 15). The puparium was cleaned with tap water and a fine brush before being placed on an objective slide. To ensure proper imaging of the three valve cells simultaneously, the posterior part of the pupa was placed on top of two cover glasses on an objective slide. Third-instar larvae were cleaned as described above and placed with their dorsal side up on a Sylgard plate. Larvae were pinned on their anterior and posterior ends with a fine needle or gently locked in place by a cover slide. Heartbeat was captured using a Zeiss LSM 800 confocal microscope equipped with a Zeiss Plan-NEOFLUAR 20×/0.5 objective, 0.5× zoom and an appropriate pinhole with an acquisition speed of ∼54 ms per frame. Data were further processed using Fiji software. In brief, the mean heartbeat (Hz) was calculated by measuring the time between two closed states of each valve, the anterior, middle and posterior pairs. Measurements were taken for three animals and two heartbeat cycles.

For the calculation of the number of valvosomes per cell, Technovit^® cross-sections were used (n=39). In total, three animals were analysed. For each of the respective valves, at least six Technovit^® sections were made, and valvosomes were defined by the expression of membrane-bound green fluorescent protein (GFP) or anti-Spectrin antibody staining. Mean numbers of valvosomes per cell were calculated and statistically examined using a two-tailed Student's t-test.

The analysis of heartbeat periodicity was performed by measuring the time for each valve to close after the posterior valve – or anterior in case of reversal heartbeat – is closed. For schematic illustration, a sigmoid curve was chosen and adjusted given the calculated parameters: heartbeat (ms), maximum luminal distance (µm) and valve periodicity (ms). The mean maximum luminal distance was calculated from measurements on the three valves of three animals for two heartbeats. Contraction speed was calculated by measuring the distance between the posterior and anterior valves in relation to the time it takes the anterior valve to close after the posterior valve is closed – or vice versa in the case of reversal heartbeat (n=9 heart cycles). Valve dynamics were analysed in three animals and three heartbeat cycles for each valve (equals to n=18 for each condition). The maximum width and length of the valve cells were compared between the closed and open states.

The contraction status of valves was analysed in three animals expressing tinC-Gal4>UAS-GCaMP3 or tinC-Gal4;handC-Gal4>UAS-GCaMP3 and for three heartbeat cycles. The luminal distance was measured as the shortest distance between opposing valve cells in the open and closed states. The heart diameter was measured at the position of the valves. The valve cell area was calculated by encircling each cell with the polygon selection tool, and mean grey values were calculated for each valve with ImageJ. Colour-coded images were generated by applying a LUT in ImageJ. All data were statistically examined using a two-tailed Student's t-test.

Mouse anti-alpha-Spectrin (1:25, 3A9, supernatant, Developmental Studies Hybridoma Bank, IA, USA), rabbit anti-GFP (1:1000, Abcam-ab6556, Abcam, UK) and mouse anti-GFP (1:500, A1120, Invitrogen, Carlsbad, CA, USA) were used for tissue immunostaining. Secondary antibodies were anti-mouse Cy3 and anti-rabbit Cy2 (1:200, Dianova GmbH).

Specimens were stained after standard protocols, dehydrated throughout a series of ethanol steps (30%, 50%, 70%, 80%, 95% and, 100%, 10 min on ice) and embedded in Technovit 8100, following the protocol provided by the manufacturer (Kulzer, Hanau, Germany). Sections (5 µm) were cut on a Leica EM UC6 microtome (Leica Microsystems, Vienna, Austria) with a glass knife, placed onto a microscope slide and examined with a Zeiss LSM800.

Stained tissue was embedded in Fluoromount and imaged with a Zeiss Pascal 5 or LSM800. Image processing and contrast and brightness adjustment were performed with Affinity Photo (Serif Europe Ltd) or Fiji (Schindelin et al., 2012).

Specimens were prepared as previously described (Lehmacher et al., 2012; Meyer et al., 2022; Psathaki et al., 2018; Psathaki and Paululat, 2022). Briefly, flies were dissected in PBS, fixed for 4 h at room temperature in a fixative composed of 2% glutaraldehyde (Sigma-Aldrich) and 4% paraformaldehyde (Merck) in PBS, and post-fixed for 1 h at 4°C in 1% osmium tetroxide in PBS (Science Services, München, Germany). Next, samples were dehydrated stepwise in a graded ethanol series followed by 100% acetone. Specimens were embedded in Epon 812 (Merck) and polymerised for 48 h at 60°C. Ultrathin sections (70 nm) were cut on an ultramicrotome (UC7, Leica) and mounted on formvar-coated copper slot grids. Sections were stained for 30 min in 2% uranyl acetate (Science Services) and 20 min in 3% lead citrate (Roth). All samples were analysed at 80 kV with a Zeiss 902 transmission electron microscope.

Data are shown as box plots with lower and upper whiskers representing minimum and maximum distribution, respectively. Boxes represent 50% of data (quartile groups 2 and 3). Lines within boxes represent the median. Lower and upper outliers are shown as single data points and are defined as exceeding a distance of 1.5 times the interquartile range (IQR) below the first quartile or above the third quartile, respectively.

The 76E11-Gal4 driver line used in this study is post-embryonically active in the larval intracardiac valve and all three valves of the imago (Meyer et al., 2023). To follow larval valve cells during metamorphosis and to monitor the emergence of the two new heart valves, we crossed 76E11-Gal4 to UAS-nGFP- or UAS-CD4td-GFP-bearing animals. We used an automated microscopic imaging system to continuously acquire fluorescence images of the heart throughout metamorphosis, from the first pupal stage (P1) to the hatch-ready adult fly (P15) (Bainbridge and Bownes, 1981). Images were analysed individually and combined into time-lapse movies (Fig. 2 and Movie 1). We found that the larval valve cells, residing at position 34 of the larval heart tube (see Fig. 1, scheme), maintain the expression of the reporter gene without any discontinuity, indicating that this pair of valve cells perpetuates its identity and differentiation state during metamorphosis. We also observed that during metamorphosis, the reporter gene expression frequently expands and spreads to the immediately adjacent cells, such that two to three heart cells are expressed in a row, the middle cell being the valve cell. This transient expression in valve cell neighbours sometimes passes through metamorphosis to adulthood. Approximately 24 h APF (stage P6) (Fig. 2E,F), we detected reporter activity in cardiac cells at positions 22 and 28 (counted from anterior, see Fig. 1). These cardiomyocytes eventually differentiate into the two additional valves of the adult heart. In all individuals analysed, reporter gene activity in the cells at positions 22 and 28 appeared simultaneously. Signal intensities continuously increased during further pupal development (see also Movie 1). In summary, we found that the larval valve remains structurally intact during metamorphosis and is maintained as an integral component of the adult heart, and that the two additional valves of the imago originate from existing cardiomyocytes and display the first signs of differentiation at approximately 24 h APF.

To compare the morphology of larval and adult valve cells, we used immunostaining on fixed hearts, semi-thin sections and ultrastructural transmission electron microscopy analysis (Figs 3 and 4). The larval valves displayed several typical cellular characteristics directly linked to their function, including very large valvosomes, which are membrane-enclosed compartments (Fig. 4J). Other morphological hallmarks of valve cells are the lateral localisation of the cell nucleus, the distinct distribution of phosphoinositides on the valvosome membrane, and the unique orientation of the myofilaments, allowing the contraction of the valve cells in conjunction with the neighbouring ‘normal’ cardiomyocytes (Lammers et al., 2017; Lehmacher et al., 2012; Meyer et al., 2023). To unravel the histology of adult valve cells and compare them with their larval counterparts, we analysed histological cross-sections and longitudinal semi-thin sections. Before embedding and sectioning, we stained the heart of 2-week-old adult flies either for membrane-targeted GFP (76E11-Gal4> UAS-CD4td-GFP) and anti-alpha Spectrin or for Trol::GFP and anti-Spectrin (Fig. 3). The cells forming the three cardiac valves in the adult heart all displayed identical and characteristic histology (Figs 3 and 4). They all possess large intracellular valvosomes, like their larval counterparts. The number of valvosomes is very similar in all valve cells (Fig. S1) and comparable to the number of valvosomes in third-instar larvae (Meyer et al., 2022). In addition, the adult valvosomes also constitute vesicle-like membranous compartments, as indicated by membrane-bound GFP signal (Fig. 3B–G) and ultrastructural data (Fig. 4C,L). Moreover, alpha-Spectrin decorates the adult valvosomal membrane (Fig. 3B–J). In addition, the inner side of the valvosomal membrane is covered with extracellular matrix (ECM), similar to larval valvosomes (Fig. 3H–J; Lammers et al., 2017). Our results indicate that the general morphology of the adult cardiac valves is very similar to the characteristic features seen in the larval counterparts (Lehmacher et al., 2012). However, ultrastructural analysis of adult valve cells also revealed some morphological differences. Unlike in the third-instar larvae, where the nucleus faces the heart's luminal side, the nucleus of adult valves locates more towards the inside of the cell and is surrounded by valvosomal compartments and cytoplasm (Fig. 4D,G). In addition, under the valve cell membrane facing the luminal side of the heart, we found several patches of muscle fibres and cytoplasm, whereas only a small strand of cytoplasm without patches of musculature is present in the larvae (Fig. 4E,H,K). ECM is present on valve cell membranes and inside valvosome compartments at both developmental stages. However, ECM appears to be abundant and more disorganised in adult valve cells (Fig. 4). In histological cross-sections, longitudinal muscle fibres appear more present in adult valves than in their larval counterparts, where circular muscle fibres can also be found in addition to longitudinal ones (Lehmacher et al., 2012). In summary, all valve cells in the adult heart share similar histology and ultrastructure. However, although harbouring all characteristic features, they differ in several details from the larval heart valve cells, including the orientation and number of myofibrils, which we discuss later in the context of the possible mode of action of the adult valve cells.

Valve cells of third-instar larvae oscillate between an elongated and roundish shape upon heartbeat, thereby effectively closing or opening the heart lumen. The adult valve cells share some histological and ultrastructural similarities but also display structural differences. Therefore, we wondered whether the three valves in the adult heart act in the same manner as their larval counterpart and, for example, also deform their shape upon heartbeat. We used handC-GFP to non-invasively capture high-speed movies for cardiac contraction periods in pharate adults. The cell width to cell length ratio was determined during cardiac contraction cycles to measure a possible change in valve cell shape (Lammers et al., 2017; Meyer et al., 2023). If the cell width to cell length ratio is identical in open and closed states, the ratio of the two values at these two time points would equal 1. Interestingly, the calculated ratio between the closed and open state is approximately 0.96 for all three valves, indicating an almost absence of change in the cell's shape during a heartbeat. This result starkly contrasts the ratio determined for larval valves of approximately 1.8 (Fig. 5A–E; see also Movie 2). However, by measuring the distance between the opposing valve cells, we found that all three valves in the adult heart effectively seal the heart lumen upon the heart beating (Fig. 5F–H,L). This observation suggests that adult valves, just like the single heart valve of the larval heart, are responsible for controlling the directionality of the haemolymph flow. However, the biomechanical properties of the cells seem to differ, thus displaying almost no deformation.

Next, we measured the temporal sequence in which the three heart valves close or open the heart lumen upon the heart beating. We expect that during a myogenic contraction wave running from posterior to anterior, the valves open and close one after the other. Therefore, the motion of the three valve cell pairs was evaluated in relation to each other. For this purpose, the distance between the opposing valve cells at each time frame was measured and graphically displayed. Using this analysis, we examined the periodicity with which the three valves open and close the heart lumen (Fig. 5I–L). The single heartbeat originates from the posterior end of the heart tube. Thereby, the contraction wave travels with a speed of 4 µm ms⁻¹ along the heart. Our observations indicated that the three cardiac valves close and open sequentially from the posterior to the anterior end upon anterograde beating (Fig. 5L). On average, the middle pair of valves closes 72 ms and the anterior one 101 ms after the posterior valve closes the heart (Fig. 5J,K). In our video analyses of intact animals, we regularly observed a spontaneously occurring reversal of the heartbeat (compare Movies 3 and 4); however, we could not find a trigger or a cause for this phenomenon. On average, the middle pair of valves closes 83 ms and the posterior one 131 ms after the anterior valve closes the heart. Upon reversal, the heart beats at 2.3 Hz and the relative contraction speed is approximately 3.3 µm ms⁻¹ (Fig. S2). The phenomenon called heartbeat reversal has been described previously for many insect species, including Drosophila spp. (Wasserthal, 1976, 2007, 2014).

We found that adult cardiac valves do not significantly change their shape upon the heart beating and lumen closing (Fig. 5E,L). To analyse the mode of operation of the adult valves in depth, we used the GCaMP3 reporter gene encoding a fluorescent Ca²⁺ sensor for live imaging of beating hearts of pharate adults, expressing the GCaMP3 construct in heart cells (Movie 5). The calcium-sensitive GCaMP3 reporter (Sun et al., 2013) displays the highest fluorescent intensity upon calcium ion release from the sarcoplasmic reticulum into the cytoplasm, directly correlating with maximal muscle contraction (Senneff and Lowery, 2021). This tool allowed us to visualise the real-time calcium dynamics in valve cells during cardiac contraction cycles. The open and closed states of the valve were identified by measuring the luminal distance between opposing valves simultaneously. In the closed state, the luminal membranes of the valve cells are in contact with each other (Figs 5L and 6C). At the same time, the outer diameter of the heart (heart tube diameter) is the smallest (Fig. 6D). We calculated the mean pixel intensity of the fluorescence emitted by the GCaMP3 sensor for each valve in the open and closed states. We found that the valve cells display the highest pixel intensity when closed, corresponding to a higher amount of intracellular calcium in the closed valves (Fig. 6E). Calcium triggers muscle contraction, so we concluded that the valves close the heart lumen due to muscle contraction. We compared valve cell areas in the closed and open states to support our results. Valve cell areas are constant between the two functional states, indicating that the calculated mean pixel intensities do not result from cell compression (Fig. 6F). Furthermore, this result indicates the absence of volume change of valve cells upon heartbeat, comparable to the observations in larval valve cells (Meyer et al., 2022).

These observations illustrate that adult valve cells seal the heart lumen without cell deformation through an active myogenic contraction. In contrast, the larval heart valve cells change their shape dynamically and alternate between an elongated and a round shape. We have previously assumed that muscle contraction in the valve cell causes an elongated cell shape, leading to an opening of the heart lumen. According to our previous hypothesis, when the cells enter a relaxed state, they adopt a spherical shape and close the heart tube. Because our current results for adult heart valves suggest a different mechanism, we reevaluated larval valve cell activity using GCaMP3 reporters to measure calcium dynamics in living animals. The calcium signal was highest when the larval valve cells adopted a spherical cell shape and closed the heart lumen, indicating that muscle contraction also leads to heart lumen closure (Fig. 6G–I). Therefore, we will present a new model for the functioning of the heart valve cells.

Transdifferentiation, or reprogramming, is a process by which already differentiated somatic cells are converted into another cell type without passing through an intermediate pluripotent form (Frasch, 2016; Graf and Enver, 2009). Several cells in the Drosophila larvae have been shown to undergo transdifferentiation, as in the immune system, where plasmatocytes transdifferentiate into encapsulating cells, the lamellocytes (Anderl et al., 2016; Cevik et al., 2019; Csordas et al., 2021). Another example is the transdifferentiation of syncytial alary muscles into the ventral longitudinal muscles that underlie the adult heart (Schaub et al., 2015). Here, we show that two of the three adult cardiac valves arise during metamorphosis by transdifferentiation of regular cardiomyocytes. The new valves originate from existing cardiomyocytes that are an integral part of the larval heart tube. The contractile, lumen-forming cardiomyocytes at positions 22 and 28 (counted from the anterior end, see Fig. 1) start to differentiate before or at 24 h APF (Fig. 2). Moreover, the putative valve cells develop intracellular valvosomes, a highly characteristic feature of the larval valve cells. The onset of valve cell differentiation at approximately 24 h APF roughly corresponds to when remodelling of the posterior ventricle begins and approximately coincides with the latest and major peak of ecdysone release during Drosophila metamorphosis (Monier et al., 2005; Riddiford, 1993).

The larval valve remains intact during metamorphosis, thus forming the third adult cardiac valve (see Fig. 1). However, the cells that form this valve also undergo some histological refinements so that, at the end of metamorphosis, they have the same histology and ultrastructure as the newly formed valve cells. The histology of all adult valve cells is similar but not identical to larval valve cells. They are defined by the presence of valvosomes, abundant amounts of ECM and a prominent Spectrin cytoskeleton. However, the specific circular orientation of myofibrils seems more pronounced than in larval valve cells, probably resulting in stronger contractions, a higher cellular stiffness and less deformability (Figs 3 and 4).

Our current data allow us to postulate an updated model for the functioning of cardiac valve cells. The contraction wave of the heart proceeds from its posterior to its anterior end. When the cardiac valve cells contract (highest concentration of free calcium, Fig. 6), the heart tube narrows and valve cells with their spherical shape projecting towards the lumen close the heart lumen. In the relaxation phase, the heart tube dilates and the lumen reopens. In larvae, the cardiac valves work in the same way. However, there are some fundamental differences. The larval valve cells are highly flexible in their cell shape and change it with every heartbeat (Lammers et al., 2017). The dramatic change in cell shape in larval valve cells could result from the required higher elasticity and flexibility of the larval heart tube. The heart tube needs elastic properties for two reasons. First, larvae grow from a size of 1 mm at hatching to 5 mm at the third-instar larval stage. During larval development, the heart tube grows exclusively by cell growth, not cell proliferation. Second, in the larva, which constantly changes its body's shape and length owing to the somatic musculature's peristaltic activity as it crawls on the substrate, the heart tube needs to be flexible and elastic to accommodate the changes in body shape without damage. In contrast, the adult fly has a relatively rigid cuticle and stable body shape, and the heart hardly changes in length and is supported by massive somatic muscles on the ventral side, the ventral longitudinal muscles. This could explain why the adult cardiac cells are less elastic and no changes are observed in the shape of valve cells during a contraction cycle of the adult heart.

In summary, a contraction wave runs over the heart chamber in the imago, resulting in a sequential closing of the valves. The four-chambered heart acts like a squeezing tube and valves close the heart lumen owing to muscle contraction. In contrast, the larval heart acts as a suction pump and haemolymph is pushed out by the overall contraction of the heart chamber. Muscle contraction is needed to close the heart lumen during diastole, enabling a proper haemolymph inflow into the posterior heart chamber.

Two new heart valves are formed in early metamorphosis and exhibit morphological similarities. The third posterior valve in the adult is identical to the larval valve and is maintained during metamorphosis. Unlike the larval valve cells, adult cardiac valve cells do not deform upon heartbeat and act sequentially, meaning they close one after another. In addition, there is evidence that the valve cells close the heart lumen owing to muscle contraction. It will be of great interest to study how the filling phase is regulated in adult hearts and how haemolymph streaming properties of the animal are regulated by cardiac valve cells. Next, we will apply life-imaging methods to measure haemolymph streaming in the heart tube and the body cavity in the presence and absence of functional cardiac valve cells to understand their role in circulation.

We thank Martina Biedermann, Kerstin Etzold and Mechthild Krabusch for excellent technical assistance and the Bloomington Drosophila Stock Center and Kyoto Drosophila Stock Center for providing stocks essential for this work. We received further support from the Center of Cellular Nanoanalytics (CellNanOs), Osnabrück University.