The rete mirabile: a possible control site for swimbladder function

In a recent study, a large number of transport proteins was detected in the transcriptome and proteome of saline perfused rete mirabile tissue of the European eel. In this study, the data set was reanalyzed for the presence of receptor proteins and proteins involved in intracellular signaling pathways. A large number of expressed receptor proteins and proteins involved in intracellular signal transduction was detected. Several G-protein-coupled receptor signal pathways were significantly enriched in their expression level, in particular receptors and signaling pathways involved in the control of blood flow. The enriched signaling pathways also include pathways involved in trafficking of crucial transport proteins like, monocarboxylate transporters, V-ATPase, and aquaporin. The data, therefore, suggest that the rete mirabile has the capacity to control swimbladder function by regulating blood flow and by modifying countercurrent multiplication. Supplementary Information The online version contains supplementary material available at 10.1007/s00360-023-01486-5.


Introduction
Countercurrent systems are widespread in the animal kingdom and used to concentrate heat or gases (Stevens 2011). Countercurrent heat exchange allows for the concentration of heat in the body core, while peripheral tissues stay cold, as observed, for example, in some fish with elevated body temperatures, in some birds or in mammals (Scholander and Schevill 1955;Dickson and Graham 2004;Stevens 2011;McCafferty et al. 2013). It may also be used for selective brain cooling (Jessen 2001;Strauss et al. 2017). In fish, a countercurrent system is used for the concentration of gases in the eye and in the swimbladder. In the choroid rete of the eye oxygen is concentrated by countercurrent concentration to supply the avascular retina (Pelster 2001;Berenbrink et al. 2005;Waser 2011). In physoclist fish a countercurrent system, the rete mirabile, is used to concentrate gases for the filling of the swimbladder under elevated hydrostatic pressure. Presence of fish with gas filled swimbladder at a depth of several hundred or even more than 1000 m demonstrates that extremely high gas partial pressures can be generated (Marshall 1960;Pelster 1997Pelster , 2001Pelster , 2023. The European eel is physostome, but the connection between the esophagus and the swimbladder is closed soon after opening the swimbladder in the glass eel stage, therefore, the eel is functionally physoclist (Dorn 1961;Zwerger et al. 2002). While in many fish the rete mirabile is intimately connected to gas gland cells of the swimbladder, in the eel the rete mirabile is bipolar, allowing access to larger blood vessels immediately in front and behind the rete, and the eel thus became a model for the analysis of swimbladder function (Steen 1963;Fänge 1983;Pelster 2001).
In the European eel rete mirabile about 58.000 arterial capillaries are surrounded by about 44.000 venous capillaries (Krogh 1929) for a distance of a few millimeters, and the diffusion distance between arterial and venous vessels is in the range of only 1-2 µm (Stray-Pedersen and Nicolaysen 1975;Wagner et al. 1987). Thus, the rete provides a large surface area with very short diffusion distance for the backdiffusion of gases. Blood samples collected in front of and immediately behind the rete of a European eel swimbladder preparation revealed a seven-to eightfold increase in PO 2 and PCO 2 during arterial passage of the rete, resulting from back-diffusion and countercurrent concentration in the rete (Kobayashi et al. 1990).
Retia mirabilia are typically considered to allow for the countercurrent concentration of gases or of heat (Kuhn et al. 1963;Stevens 2011). In the European eel measurements of lactate, water and hemoglobin concentration, however, indicated that lactate may also be concentrated in the rete by back-diffusion from venous to arterial capillaries (Kobayashi et al. 1989a). A recent study of the transcriptome and the proteome of rete mirabile tissue of the European eel indeed demonstrated the presence of a large number of transport proteins and membrane ATPases (Schneebauer et al. 2021). Membrane transport proteins, however, are not always constitutively active, their activity may be adjusted in relation to cellular activity. This appears especially important for the rete as model calculations revealed that countercurrent concentration of lactate in the rete mirabile would enhance countercurrent concentration of gases (Kobayashi et al. 1989b). The presence of membrane transport proteins in rete mirabile membranes with adjustable activity thus could significantly affect the gas concentrating ability of the rete. Moreover, some receptor proteins have been detected in the study of Schneebauer et al. (Schneebauer et al. 2021), and receptor proteins require intracellular signaling cascades, but this was not in the focus of that study. This study, therefore, was set out to reanalyze the transcriptome and proteome data published by Schneebauer et al. (Schneebauer et al. 2021) searching in particular for receptor proteins and proteins involved in intracellular signaling. The results revealed expression of a large number of receptor proteins, and of several intracellular signaling pathways, connected to these receptors. The results provide evidence that blood flow through the swimbladder may be controlled at the level of the rete mirabile, and signaling pathways involved in the insertion of crucial transport proteins into the cell membranes are enriched in their expression level.

Methods
For this study, transcriptome (accession number GSE172092, https:// www. ncbi. nlm. nih. gov/ geo/ query/ acc. cgi? acc= GSE17 2092) and proteome data (accession number PXD025435, https:// www. ebi. ac. uk/ pride/ archi ve/ proje cts/ PXD02 5435), published by Schneebauer et al. (2021), were reanalyzed focusing on the presence of transcripts coding for receptor proteins and for proteins connected to intracellular signaling, and on the presence of the appropriate proteins. Transcripts and proteins were selected by searching the 'Description' and the 'GO biological function' of annotated transcripts and proteins for the terms 'receptor' and 'signaling'. Extracted transcripts and proteins were then used to identify signaling pathways that are expressed more than would be expected by chance in rete tissue by performing a GO enrichment analysis using the GO Ontology database https:// doi. org/ 10. 5281/ zenodo. 67997 22, released 2022-07-01. Based on this enrichment analysis signaling pathways were identified that may be of functional importance for rete tissue. GO pathways are defined based on mammalian data, but meanwhile a large number of non-mammalian species have been added, including the zebrafish, Danio rerio (see http:// geneo ntolo gy. org/), and a remarkable number of orthologue genes and of overlap has been detected between different vertebrates, including mammals and fish (Howe et al. 2013).

Results
The description of identified transcripts of rete mirabile tissue revealed presence of 1808 mRNA transcripts coding for a receptor protein. Searching the GO biological function of identified transcripts for the term 'signaling' identified 7206 transcripts. Analyzing the proteome for the same terms revealed presence of 136 receptor proteins and 1291 proteins connected to the biological function 'signaling'. 299 of these proteins were detected in the transcriptome as well as in the proteome (Suppl. Tab. 1).
In Fig. 1 for all genes detected in the transcriptome as well as in the proteome the relative mRNA expression value is plotted versus the relative protein abundance. Relative mRNA expression values usually were between 10 0 and 10 5 , while the relative protein abundance covered the range between 10 5 and 10 10 .
The GO enrichment analysis confirmed a more than fivefold enrichment of various G-protein-coupled receptor  Relative mRNA expression plotted versus the relative protein abundance of genes coding for a receptor protein or a protein connected to intracellular signal transduction detected in the proteome as well as in the transcriptome of European eel rete mirabile tissue. Regression line intercept: 10 6.02 (0.15); log 10(×) 0.25 (0.05); R 2 0.07; p < 0.001; calculated with 'R'. Broken lines show 95% confidence interval signaling pathways (G-protein-coupled adenosine receptor signaling; G-protein-coupled purinergic receptor signaling; G-protein-coupled acetylcholine receptor signaling; G-protein-coupled glutamate receptor signaling). In addition, pathways including cAMP signaling, or including phospholipase C and inositol trisphosphate signaling were more than fivefold enriched, and this was highly significant (Table 1). An endothelin signaling pathway was almost sixfold enriched, and a cGMP signaling pathway was significantly enriched.
A more detailed analysis revealed expression of a number of receptor proteins involved in blood flow regulation, like atrial natriuretic peptide receptor (anpra), beta-adrenergic receptor kinase (arbk2), and endothelin receptor (ednra; ednrb). At the mRNA transcript level additional transcripts coding for muscarinic acetylcholine receptor m2-m5 have been detected (acm2-acm5), but the proteins have not been found in the proteome.
As already indicated by the GO enrichment analysis, in the transcriptome 114 transcripts connected to G-protein coupled receptor signaling could be detected. Downstream of G-protein coupled receptors 53 transcripts connected to adenylate cyclase signaling were identified. Several guanine nucleotide-binding proteins and inositol-trisphosphate receptor (itpr1, itpr2) were also expressed at the protein level. In addition, calcium calmodulin-dependent protein kinase (kcc2b) was detected in the transcriptome as well as in the proteome.
Growth factor signaling appeared to be of importance in rete mirabile endothelial cells, as 177 transcripts connected to growth factor signaling could be detected, including, for example, epidermal growth factor (eps15), insulin like growth factor (igf1r), hepatocyte growth factor (met), and transforming growth factor (tgbr3; tgfa1; tgfr2), which were also detected in the proteome. In addition, 7 tyrosineprotein kinase receptor proteins and 10 receptor-type tyrosine phosphatase proteins were found. Moreover, tyrosinekinase receptor proteins connected to angiogenesis have been detected like ephrin type-b receptor (ephb3), tie1 and tie 2 receptor (tie1; tie2), and vascular endothelial growth factor receptor (vegfr2; vegfr4).
Small G-protein signaling also appeared to play an important role, as 67 transcripts connected to rho-protein signal transduction were identified, and 32 transcripts connected to rac protein signal transduction. Numerous small G-proteins (rho; rac; ras) were also present in the proteome.

Discussion
The number of transcripts coding for a receptor protein or a protein involved in intracellular signaling by far exceeded the number of proteins detected. This at least in part can be explained by the whole genome duplication that occurred in teleost evolution (Glasauer and Neuhauss 2014;Meyer and Van de Peer 2017). Also, a number of genes detected in the transcriptome was not detected in the proteome and vice versa. As repeatedly observed in different species including humans the correlation between protein and mRNA is often modest (Liu et al. 2016;Wang et al. 2019;Schneebauer et al. 2021). There are in fact several phenomena that may contribute to a discrepancy between the transcriptome and the protein expression level, like, for example, stored transcripts, a delay between transcription and translation, post-translational regulation, or post-transcriptional regulation. The main focus of this study, therefore, was on genes present in the transcriptome as well as in the proteome. A large number of members of the family of G-proteincoupled receptors has been detected in the transcriptome and in the proteome of the rete mirabile of the European eel. The family of G-protein-coupled receptors is a very large and very diverse family of receptor proteins (Lefkowitz 2004;Sanders et al. 2008;Calebiro et al. 2021). It includes adrenergic receptors and muscarinic cholinergic receptors, operating with adenylyl cyclase and protein kinase A (Pierce et al. 2002). The central components of these signaling pathways were detected in the transcriptome as well as in the proteome. Adrenaline has been shown to reduce swimbladder perfusion (Pelster 1994), and the influence of the vagosympathetic trunk on swimbladder blood flow is well documented (Schwerte et al. 1997;Nilsson 2009;Smith and Croll 2011). This suggests that control of blood flow through the swimbladder may be established at the level of the rete mirabile, and blood flow is a crucial parameter determining the rate of gas secretion. In the European eel, an increase in swimbladder perfusion has been shown to result in a proportional increase in the rate of gas secretion into the swimbladder (Pelster and Scheid 1992). Smooth muscle cells at the entrance of the rete could adjust peripheral resistance and thus control swimbladder perfusion. Presence of the endothelin receptor supported the conclusion that blood flow regulation may be possible at the site of the rete mirabile, endothelin being a potent vasoconstrictor (Davenport et al. 2016).
Another signaling pathway significantly enriched was the G protein-coupled glutamate receptor signaling pathway, which includes the calcium-sensing receptor (Pierce et al. 2002). The importance of Ca 2+ as a signaling molecule was also supported by the presence of calcium calmodulindependent protein kinase, detected in the proteome, and of the inositol-trisphosphate receptor. Inositol-tris-phosphate is produced by the activity of phospholipase C, activated by a G protein-coupled receptor, and triggers the release of Ca 2+ from the intracellular stores (like endoplasmic reticulum). Ca 2+ is essential for activation of smooth muscle cells, which would be consistent with the conclusion that smooth muscle cells at the rete mirabile may modify perfusion resistance and thus control blood flow through the swimbladder, a crucial parameter for swimbladder function. The rate of gas secretion into the swimbladder is correlated to blood flow (Pelster and Scheid 1992), and blood flow influences countercurrent concentration in the rete mirabile (Stevens 2011). In Fig. 2, different G protein-coupled receptors detected are depicted together with possible intracellular signaling molecules, identified in the transcriptome and proteome of the rete.
The presence of atrial natriuretic peptide receptor 1 in the transcriptome and the proteome supported the conclusion that the rete is involved in the control of blood flow through the swimbladder. ANP signaling is involved in the regulation of blood pressure and body fluid volume, and is antagonistic to the renin angiotensin system (Nakagawa et al. 2019). Intracellular signaling includes cGMP signaling, and  Fig. 2 Selected receptor proteins, proteins involved in intracellular signal transduction and assumed signaling pathways in rete mirabile endothelial cells. GF-(growth factor) signaling, several growth factor receptors have been detected including tie1,2; vegf2,4; ephrin receptor. ACh, acetylcholine; ANP, atrial natriuretic peptide; PKA, protein kinase A; PKC, protein kinase C, CAMK, calmodulin dependent protein kinase; ß-arbk2, beta-adrenergic receptor kinase; TRK, receptor tyrosine kinase members of the cGMP mediated signaling pathway were significantly enriched in the transcriptome (Fig. 2). A remarkable number of growth hormone receptor proteins have been identified, like the receptor tyrosine kinases tie1, tie 2, and vegfr2 and vegfr4, potent angiogenic growth factor receptors (Simons et al. 2016;Akwii et al. 2019). In addition, ephrin type-b receptor, involved in morphogenesis and cell differentiation (Wilkinson 2014), but also in angiogenesis and vasculogenesis (Zhang and Hughes 2006), has been detected (Fig. 2). Schneebauer et al. compared yellow and silver eels and found only a small number of genes differentially expressed in the rete, so that the two different developmental stages were combined in their analysis focusing on transport protein expression in the rete (Schneebauer et al. 2021). Studies on the American eel Anguilla rostrata, however, revealed an increase in rete length (Kleckner and Krueger 1981), suggesting that angiogenesis should be activated in silver eels. Elongation of the rete increases surface area available for countercurrent exchange and, therefore, significantly improves the capacity for countercurrent multiplication (Kobayashi et al. 1989b). The transcript of angiopoietin-related protein 4 was indeed elevated in the expression level in silver eels, but the according protein was not detected in the proteome (Schneebauer et al. 2021;this study). Based on the silvering and ocular index, the silver eels used by Schneebauer et al. were far beyond the threshold for the transition from yellow to silver eels (Schneebauer et al. 2021). Therefore, it appears possible that elongation of the rete was close to complete and elevated angiogenic signaling to stimulate further elongation of the rete was no longer necessary.
With respect to membrane transport proteins V-ATPase and aquaporin have been shown to contribute to acid backdiffusion in the rete, significantly enhancing the countercurrent concentration of oxygen and CO 2 , whereas monocarboxylate carriers allow for a back-diffusion of lactate, facilitating countercurrent concentration of inert gases by supporting the salting out effect (Kobayashi et al. 1989b(Kobayashi et al. , 1990Schneebauer et al. 2021). V-ATPase and also aquaporin are well known to be introduced into cell membranes by trafficking of vesicles between cytoplasm and cell membranes. Both transport proteins are inserted into the cell membrane following hormonal activation of G-proteincoupled receptors, followed by activation of adenylyl cyclase and protein kinase A (Brown et al. 1998;Nedvetsky et al. 2009;Mcguire et al. 2017;Collins and Forgac 2020). As already discussed, the G-protein-coupled signaling pathway was significantly enriched and crucial components of these signaling pathways were expressed in rete tissue. It, therefore, appears quite possible that regulated insertion of aquaporin and V-ATPase support back-diffusion and countercurrent concentration in the rete. The same appears to be true for monocarboxylate transporters. cAMP dependent trafficking of monocarboxylate carriers has been shown in rat brain endothelial cells (Smith et al. 2012;Uhernik et al. 2014). The anxiliary proteins basigin and embigin have been shown to be involved in monocarboxylate trafficking (Nakai et al. 2006;Felmlee et al. 2020), and basigin has been detected in the transcriptome, but not in the proteome. The proteome and transcriptome data therefore reveal that the rete mirabile is equipped with the required signaling pathways to allow for a regulated insertion of transport proteins into rete membranes. The number of transport proteins affects back-diffusion of metabolites, and thus influences countercurrent multiplication achieved in the rete.
Taken together the data suggest that the rete mirabile is equipped with the capacity to regulate blood flow through the rete, and to regulate back-diffusion of metabolites in the rete. Both parameters are crucial components for the effectiveness of the rete as a countercurrent multiplier. The rete, therefore, appears to be a crucial component significantly influencing the secretory activity of the swimbladder.
Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http:// creat iveco mmons. org/ licen ses/ by/4. 0/.