Comparison with previous data on molluscan nephrogenesis
Previous reports on nephrogenesis are restricted to only four of eight molluscan classes: Gastropoda, Bivalvia, Cephalopoda and Polyplacophora.
Data on Gastropoda are scarce and differ considerably: The anlagen of kidneys and pericardia are reported to be paired (e.g.: Patella) or unpaired (: Marisa), to be separate (e.g.: Viviparus) or common (e.g.[22, 24]: Marisa). The descriptions agree on the fact, that initially solid anlagen next to the hindgut give rise to kidney and pericardium. In gastropods the renopericardial complex is differentiated prior to metamorphosis[21, 25, 26].
Renopericardial development of Bivalvia has been investigated by a number of studies, but they date back between 80 and 130 years (e.g.[27–32]). These studies can be summarized as follows: Paired, solid anlagen give rise to kidneys, pericardium and (later) gonads. Kidney differentiation with formation of a lumen precedes that of the heart; the kidney becomes tube-like and gets in contact with the epidermis and, thereby, forms a porus to the outside; most of the remaining initial anlage gives rise to the prospective pericardium; the two sides subsequently fuse in the sagittal plane to form the unpaired pericardium and heart.
Cephalopoda show a highly derived mode of development. Data on renopericardial ontogeny are limited to Sepia, Loligo and Octopus[33–37], where kidney development with formation of a lumen distinctly precedes that of the pericardium. From the onset of pericardium differentiation, there is a detectable connection to the kidney, which becomes the future renopericardial duct.
Information about nephrogenesis of the Polyplacophora is scarce. In a study on general development of Acanthochiton discrepans Hammarsten and Runnström report the kidneys to be formed as outgrowth of the preformed pericardium; they subsequently get in contact with the epidermis and form a nephroporus. The more recent studies by Bartolomaeus and Salvini-Plawen & Bartolomaeus on Lepidochitona cinerea added ultrastructural details on pericardium and heart development. In this study observations are restricted to three stages (250, 550 and 1000 μm length). Whereas we can confirm the principle mode of development in Lepidochitona (i.e. outgrowth of the releasing duct from the pericardium with minimal support from the ectoderm), we note a number of differences in the details of nephrogenesis: Concerning the first anlage at about 550 μm body length, Bartolomaeus and Salvini-Plawen & Bartolomaeus described a paired anlage of the pericardium epithelial tubes left and right from the rectum with already differentiated podocytes and cilia. In contrast our study shows an unpaired mass of cells which differentiate towards an epithelial cavity lacking ultrafiltration cells (podocytes). In a specimen with about 1 mm body length Bartolomaeus described a large pericardium with heart and caudally emerging pericardioducts with blind endings in the body wall. In our specimens of this size the whole metanephridial system is fully developed and shows nephropores, and the pericardioducts emerge relatively far anterior on the sides of the pericardium.
Both species, Lepidochitona cinereus and L. corrugata, are closely related, nevertheless heterochronic effects, i.e. shifts in the timing of nephrogenesis, cannot be fully excluded. On the other hand we based our study on 16 specimens of varying stages of nephrogenesis, whereas the former study investigated specimens of different stages only. Accordingly, studies on the nephrogenesis of further species, in particular of representatives of the Lepidopleurida, appear necessary to proceed towards inference of a ground pattern of nephrogenesis in the Polyplacophora.
Inference of homology always is a matter of probabilities and even identical genetic background of a phenotypic subject does not solve the matter i.e. "homocracy" structures;. All homologization is comparison, and various hierarchical levels need to be clearly distinguished[40, 41]. Concerning the excretory systems the levels to be considered are; (1) Iterative homology of ultrafiltration cells, i.e. the protonephridial cyrtocyte with the metanephridial podocyte; (2) Serial homology of larval protonephridia with adult metanephridial system; (3) Supraspecific homology of both systems among the Mollusca and beyond. Current understanding of the modular organization of the genome makes it possible to consider homologies (similarities based on shared gene expression) independently from synapomorphies (appearance of characters at the phylogenetic tree).
Proto- and metanephridial system in Lepidochitona corrugata:
The (developing) metanephridial system resembles the protonephridia of Lepidochitona corrugata in several aspects: (1) Both organ systems start most internally with ultrafiltration sites. These are followed by a transportation duct, then a modification portion, and finally an efferent duct, which opens via a pore to the outside. Nematodes or arthropods have principally different excretory structures, so this similarity is not trivial. (2) In both cases epithelial ultrafiltration cells filter molecules from the primary (haemocoel) to a secondary (endothelial) body cavity. (3) The fine structure of the ultrafilter itself is identical and complex; it is composed of meandering slits, which form gaps between pedicle elements, and these gaps are interconnected by a thin diaphragm. (4) The transportation duct following the ultrafiltration portion shows dense downstream ciliation. (5) Both organs exhibit an absorptive portion, the kidney, which has identical cellular organization with basal infoldings, apical microvillous border and strongly vacuolized cytoplasm. Again the case of vertebrate kidney shows that this similarity is not trivial.
The two systems also exhibit differences:
(1) Ultrafiltration of protonephridia is done by terminal cells, while it is carried out by podocytes in the case of metanephridia. Essentially, the difference between these two cell types is presence (terminal cells) and absence (podocytes) of ciliation. Nevertheless, a homology between these cell types appears likely , since intermediate forms have been reported, e.g. the "cyrtopodocytes" in the amphioxus Branchiostomma and do occur also in early juvenile bivalves (Ruthensteiner et al. unpubl.). In addition, both types do occur subsequently during ontogeny in certain polychaetes (e.g. ). The finding of a cilium in a podocyte of Lepidochitona corrugata – remarkably the first report of a podocyte bearing a cilium in a mollusc – lends substantial support to this concept.
(2) In the metanephridial system filtration pressure is (partly) generated by heart beat, while it is (partly) generated by the ciliation of terminal cells in protonephridia. However, this difference seems to be of limited significance, since in both systems a major part of negative pressure generation is facilitated by the same structures, the downstream ciliation of the excretory duct (see [43, 44] for review of molluscan metanephridial systems).
In conclusion, we assume that there is substantial homology between the protonephridial and metanephridial system in Lepidochitona at least at the cellular level. For Lepidochitona it seems possible that the metanephridial system as a whole is a serial homologue of the protonephridial system. However, the overall data are still fragmentary and do not permit a final clarification of that question. For example, it remains unclear if the absorptive portions (kidneys) of protonephridia are homologous throughout the Mollusca.
Proto- and metanephridial systems in molluscs and related taxa:
The current evolutionary understanding of nephridial systems in the Mollusca is as follows:
(1) Protonephridia: There is wide agreement that a single pair of anteriorly placed, larval protonephridia ("head-kidneys" sensu , "archinephridia" sensu  of polychaetes) is a synapomorphic organ system at least for Trochozoa (i.e. Entoprocta, Mollusca, Sipuncula, and Annelida; doubtfully also Nemertinea) [4, 5, 9, 47] and thus plesiomorphic for the Mollusca (see also ).
(2) Metanephridial system: Salvini-Plawen (e.g. [11–13, 48–50]) considered metanephridial nephroducts ("kidneys") to be evolved within the Mollusca at the evolutionary level of Testaria (Polyplacophora and Conchifera). These "emunctoria" are interpreted as specialized portions of previously undifferentiated "pericardioducts", which originally lack a function in excretion. Salvini-Plawen provides two main reasons for his hypothesis: (a) the absence of an ultrafiltration/reabsorption (metanephridial) system in aplacophoran molluscs; (b) the non-homology with excretory organs of other phyla, such as the metanephridia of annelids. However, both arguments seem doubtful: (ad a) TEM findings suggest regular metanephridial function also among aplacophorans: Podocytes were described from the pericardium of both aplacophoran taxa, Solenogastres  and Caudofoveata , providing evidence for ultrafiltration. Morphological evidence for reabsorption has been reported in the caudofoveate Falcidens crossotus: Cells of the lower pericardial ducts (releasing also the gametes) exhibit all ultrastructural characteristics of absorptive cells including a distinct basal infolding system (Figure 23 G of ). Accordingly, a metanephridial system probably also exists in Solenogastres. (ad b) Salvini-Plawen (e.g. ) and Salvini-Plawen & Bartolomaeus ( , also ) also rejected a homology between molluscan and annelid/sipunculan metanephridia by assumed different germ layer origin and different mode of formation: While the molluscan kidneys are a mesodermal outgrowth of the pericardium, the annelid kidneys are formed as ingrowth of ectodermal epidermis. However, there are molluscs showing metanephridial anlagen without a connection to a pericardium  and there are certain annelids (e.g. [55, 56]) where metanephridia descend from the coelothelium of the previously formed coelomic pouches (see also ). Thus, because of the variability in nephrogenesis of both, Mollusca and Annelida this reasoning cannot be upheld. Also the second argument against homology of molluscan and annelid nephridia is weak at least: Molluscs lack a funnel at the beginning of the nephroduct as found in annelids . However, a funnel is just an outbulging opening of the metanephridial duct into a voluminous coelomic cavity. Again, annelids vary in this respect, and particularly those with a restricted coelomic cavity, such as some leeches, lack a funnel too .
Accordingly, and based on substantial similarities in fine-structure, we think that a metanephridial system consisting of terminal podocytes at the pericardial walls (ultrafiltration site), a renopericardial duct (transport and generation of pressure by cilia), and a nephroduct ("kidneys": reabsorption) belongs to the basic pattern of the Mollusca and is homologous with those of other trochozoan phyla.
Admittedly, the data basis for these assumptions is still poor. Further comprehensive studies with a 4-dimensional approach (3D-analysis of successive development stages) on nephridial organs of more molluscan taxa and other invertebrates are required for a better understanding of the questions on homology and synapomorphy raised above and thus on the framework of evolution of filtration excretory organs among Metazoa.