Performing a mutagenesis screen in Drosophila, Heisenberg and colleagues identified several mutants with structural brain defects, including five mutants showing severe neurodegeneration as seen by the formation of vacuoles in the brain [67]^[19]. A complementation analysis showed that they all affected the same gene which based on the appearance of the mutant phenotype was named swiss-cheese or sws (Figure 1). Cloning and sequencing revealed that two of the alleles contained point mutations in the sws gene ^[1]. Another mutation generated a stop codon early in the sequence, resulting in no protein being detectable in Western blots. This allele (sws¹) therefore appears to result in a complete loss of SWS. The mutations in the remaining two alleles could not be identified. The phenotypic analyses showed that all sws alleles exhibit age-dependent neuronal degeneration due to apoptosis, glial cell death, and a reduced lifespan ^[1]. In addition, sws mutants develop locomotion deficits, disruptions in the ER and ER stress [31,68]^[20][21]. Specifically knocking down SWS in neurons also caused age-related neurodegeneration, locomotion deficits, reduced life span and a reduction of mitochondria in the CNS, whereas mitochondrial transport was reduced in the PNS [69]^[22]. Consistent with reduced mitochondrial function, these flies also showed an increase in reactive oxygen species, in addition to an accumulation of lipid droplets. A knockdown in all glial cells resulted in glial wrapping defects in the CNS and abnormal glial layers and fragmented glial nuclei in the PNS [54,70]^[23][24]. Furthermore, the pan-glial knockdown of SWS expression induced defects in neuronal transmission and locomotion, while more restricted knockdowns in particular glial subtypes revealed that SWS is specifically required in ensheathing and surface glia [70]^[24].

To determine the functional importance of the different domains identified in SWS, several constructs were tested for their ability to rescue the phenotypes seen in the sws mutant. Using a wildtype form of SWS, revealed that the neuronal degeneration could be rescued by selective expression in neurons, whereas the glial phenotype could be rescued by expression in glia, showing that SWS is autonomously required in both cell types [31]^[20]. In addition, expression in either neurons or glia restored esterase activity and for the neuronal rescue it was also demonstrated that it reduced the abnormal increase in PC seen in sws flies. In contrast, as mentioned above, a SWS construct with a mutation in the active site serine of the phospholipase domain did not restore esterase activity in sws mutants and it only partially rescued the neurodegeneration phenotype [9,31]^[9][20]. Similarly, a mutation in the PKA-BD domain only partially rescued the degenerative phenotype ^[9], suggesting that both domains contribute equally to the neuronal function of SWS. Although the PKA-BD mutant protein has an intact phospholipase domain, it could not restore the esterase activity in sws mutants; however, it doubled the level of esterase activity when expressed in the presence of wildtype SWS. These findings support the hypothesis that the PKA-regulatory function of SWS is required for its phospholipase activity, though not directly within the same molecule. In glia, the phospholipase function seems to play a more prominent role than the PKA-regulatory function, because the mutation in the active site serine completely failed to rescue the glial phenotypes in the glia-specific knockdown of SWS [70]^[24]. In contrast, expression of the PKA-BD mutant resulted in a partial rescue. Lastly, to confirm the functional conservation of SWS and mammalian PNPLA6/NTE, constructs encoding either the mouse or human PNPLA6/NTE were expressed in sws mutants, and both could rescue the behavioral and degenerative phenotypes, as well as restore esterase activity [31,71,72]^[20][25][26].

Effects of the loss of PNPLA6/NTE have also been studied in mammalian models. A knock-down in zebrafish embryos caused various abnormalities, with a curled tail being the most prominent phenotype [73]^[27]. In addition, the embryos showed defects in eye development, midbrain-hindbrain boundary abnormalities, and a reduced number of motor neurons with short and abnormally branched axons. Addressing the importance of the esterase domain, the authors co-injected PNPLA6/NTE-specific morpholinos with different PNPLA6/NTE mRNAs. Whereas expression of wildtype PNPLA6/NTE rescued the curly tail and motor neuron phenotype of the knock-down, three mRNAs that contained mutations in the esterase domain, including one with a mutation in the active site serine, did not. Similar experiments by Hufnagel et al. confirmed the curly tail phenotype in the PNPLA6/NTE knockdown and also demonstrated that wildtype PNPLA6 rescued this phenotype but PNPLA6/NTE with mutations in the esterase domain did not [59]^[28]. In combination, these studies provide further evidence that the esterase/phospholipase activity is important for the biological function of PNPLA6.

Mice lacking PNPLA6/NTE show severe growth retardation and die around day 9 of embryonic development [74]^[29]. A histological analysis revealed defects in placental development (with no placental labyrinth forming), in addition to a breakdown of yolk sac circulation, leading to enlarged pericardia and dilated blood vessels in the embryo. A role early in embryonic development is also supported by findings that PNPLA6/NTE is expressed in mouse embryonic stem cells, with increased levels during differentiation, and that silencing caused changes in the expression of genes involved in neuronal development and the formation of the respiratory and vascular system [75,76]^[30][31]. Silencing of PNPLA6/NTE in a human pluripotent cell line induced similar changes in the transcriptome, suggesting that the function of PNPLA6/NTE during development is conserved in humans [77,78]^[32][33]. Interestingly, treatment with OPs did not induce these transcriptional changes, although it inhibited the esterase function, suggesting that this developmental function of PNPLA6/NTE is not dependent on its esterase activity [77,79]^[32][34]. Whereas the loss of PNPLA6/NTE and its developmental functions causes lethality during embryogenesis, heterozygous knock-out animals develop normally but have reduced esterase activity, increased susceptibility to certain OPs, and they appear hyperactive in open field tests [74,80]^[29][35]. A conditional, brain-specific knock-out did not affect development but these mice showed defects in motor coordination and neuronal degeneration when aged to 4.5 month [81]^[36]. Further analyses revealed a loss of Purkinje cells, disruptions of the ER, and abnormal reticular aggregates. In the spinal cord, axonal degeneration was first noticed in the distal parts of the longest spinal axons at 1 month of age, followed by axonal swelling that increased with aging in both, ascending and descending tracts ^[7]. These defects were accompanied by progressive hindlimb dysfunction, first detectable by clenching of the digits, and by 4 months the animals were unable to fully support their lower body when walking on a beam. Comparable axonal degeneration and progressive swelling of spinal cord axons was detected when treating animals with TOCP ^[7]. A glial-specific knock-out resulted in incomplete ensheathment of Remak fibers in the sciatic nerve whereas myelination was not affected [60]^[37], confirming that the loss of PNPLA6/NTE in glia also has functional consequences in mammals. Lastly, a selective knock-down of PNPLA6/NTE in testis demonstrated a role in the proliferation of spermatogonial stem cells and a reduction in sperm count was also observed when treating male mice with TOCP [82]^[38].