Autism Spectrum Disorder (ASD) is a polygenic disorder with a complex underlying genetic etiology (Bourgeron, 2015). Advances in whole genome sequencing and genome-wide association studies have dramatically expanded our understanding of the genetic architecture of ASD. In particular, the identification of rare de novo mutations that confer high risk for ASD has generated new molecular insight (De Rubeis et al., 2014; Iossifov et al., 2014; Sanders et al., 2015). Yet, even in cases where a rare de novo mutation confers risk for ASD, additional processes are likely to contribute to the ASD phenotype (Leppa et al., 2016; Peter et al., 2019) including the engagement of adaptive physiological mechanisms (Gaugler et al., 2014; Gibson, 2009; Hartman et al., 2001; Hou et al., 2019; Kitano, 2007; Plomp et al., 1992; Sackton and Hartl, 2016; Sardi and Gasch, 2018; Bourgeron, 2015).

Homeostatic plasticity, in particular, has garnered considerable attention as an adaptive physiological process that might be relevant to the phenotypic penetrance of ASD mutations (Antoine et al., 2019; Bourgeron, 2015; Mullins et al., 2016; Nelson and Valakh, 2015; Ramocki and Zoghbi, 2008). Yet, very little is known at a mechanistic level regarding the interface of homeostatic plasticity and ASD genetics. There remains ongoing debate regarding whether homeostatic plasticity is normally induced or whether it is impaired in the context of rare de novo mutations that confer risk for ASD (Antoine et al., 2019; Bourgeron, 2015; Ramocki and Zoghbi, 2008). And, there is no mechanistic information regarding how rare de novo mutations that confer risk for ASD might be connected to the signaling mechanisms that are essential for the induction and expression of homeostatic plasticity.

It is well established that homeostatic signaling systems function throughout the central and peripheral nervous systems to stabilize neural function following a perturbation that can be of genetic, immunological, pharmacological or environmental origin (Davis, 2006; Marder, 2011; Turrigiano, 2011). Evidence for this has accumulated by measuring how nerve and muscle respond to the persistent disruption of synaptic transmission, ion channel function or neuronal firing. In systems ranging from Drosophila to human, cells have been shown to restore baseline function in the continued presence of these perturbations by rebalancing ion channel expression, modifying neurotransmitter receptor trafficking and modulating neurotransmitter release (Davis, 2013; Hengen et al., 2013; Maffei and Fontanini, 2009; Watt and Desai, 2010). There is evidence that homeostatic signaling systems function at the level of individual cells and synapses (Davis, 2013). There is also evidence that homeostatic signaling systems influence the function of neural circuitry (Deeg and Aizenman, 2011; Hengen et al., 2013; Maffei and Fontanini, 2009; Nelson and Valakh, 2015).

We set out to determine whether there exists a molecular interface between mutations in ASD gene orthologs in Drosophila and the induction or expression of presynaptic homeostatic plasticity. Presynaptic homeostatic plasticity (PHP) is an evolutionarily conserved form of homeostatic plasticity, observed in Drosophila, mice and humans (Davis, 2013). PHP has been documented at both central and peripheral synapses in response to differences in target innervation (Liu and Tsien, 1995) altered postsynaptic excitability (Davis, 2006; Marder and Goaillard, 2006; Mullins et al., 2016), following chronic inhibition of neural activity (Kim and Ryan, 2010; Zhao et al., 2011) and following disruption of postsynaptic neurotransmitter receptor function (Henry et al., 2012; Jakawich et al., 2010). The mechanisms of PHP have a remarkable ability to modulate and stabilize synaptic transmission, with an effect size that can exceed 200% (Müller and Davis, 2012; Ortega et al., 2018).

Many of the rare de novo mutations that confer high risk for ASD are considered to be heterozygous loss of function (LOF) mutations (Bourgeron, 2015; De Rubeis et al., 2014; Iossifov et al., 2014; Sanders et al., 2015). Therefore, we examine the phenotype of heterozygous LOF mutations in five different ASD gene orthologs. We make several fundamental advances. First, we demonstrate that these individual heterozygous LOF mutations have no overt effect on baseline transmission or PHP. However, we demonstrate that PHP is sensitized to failure. Next, we sought to define the molecular mechanisms that connect ASD gene orthologs to the mechanisms of PHP. A genome-scale screen and subsequent systems-genetic analyses yielded unexpected insight. We do not simply identify genes that, when mutated, enhance the phenotype of individual ASD gene mutations. We discovered genes that, when their function is diminished by heterozygous LOF mutations, commonly modify multiple ASD gene orthologs, causing a selective failure of homeostatic plasticity. Thus, we define the first class of common phenotypic modifiers of ASD genes in any system. Finally, we do not stop with the identification of a novel class of ASD gene modifiers. We proceed to characterize how homeostatic plasticity fails in one such condition. The mechanism we discovered is also unexpected and illuminates the complexity by which double heterozygous gene-gene interactions can generate a cellular or organismal phenotype. We demonstrate maladaptive, enhanced expression of a gene known as Cellular Repressor of E1A Stimulated Genes (CREG), a gene that is conserved from Drosophila to human and expressed in the brain (Yang et al., 2011).

Taken together, we define a novel, unexpected genetic architecture that connects heterozygous LOF mutations in ASD-associated gene orthologs with the mechanisms of homeostatic plasticity. In particular, the observation that PHP is commonly sensitized by multiple, different ASD genes, and the fact that we identify and characterize common phenotypic modifiers of five different ASD genes, defines a novel means by which a diversity of ASD-associated risk genes may converge to affect synaptic transmission. We propose that this information may be relevant to new therapeutic approaches that might someday modify ASD phenotypic severity, regardless of the underlying genetic mutation(s) that confer risk for ASD.

We began an investigation of ASD gene orthologs in Drosophila by acquiring heterozygous null mutations in five genes; RIMS1, CHD8, CHD2, WDFY3 and ASH1L (Figure 1A; Note: throughout we use the human nomenclature). Heterozygous null mutations were analyzed, as opposed to use of RNAi-mediated gene knockdown, in order to more precisely reflect the proposed genetic perturbations in human.

Figure 1 with 2 supplements see all

Download asset Open asset

All five of these genes are considered high confidence ‘category 1’ ASD-associated genes based on SFARI Gene (Simons Simplex Collection, 2020). All five of these genes have clear Drosophila orthologs. Further, we demonstrate that all five genes are expressed in Drosophila third instar motoneurons based on a Patch-Seq analysis of gene expression (Figure 1—figure supplement 1). The five ASD gene orthologs were also chosen to reflect a broad range of biological activities that are associated with the numerous ASD-associated genes identified to date. The RIMS1 gene is a synaptic scaffolding protein that localizes to and organizes sites of neurotransmitter release, termed active zones. The CHD8 and CHD2 genes encode chromatin remodeling factors that localize to the cell nucleus. WDFY3 encodes a phosphatidylinositol 3-phosphate-binding protein and regulator of autophagy and intracellular signaling. ASH1L encodes a member of the trithorax group of transcriptional activators and is found in the cell nucleus. A supplemental table (Supplementary file 1, Supplemental Table S1) includes known disease associations for each of these five human genes, and links to web-based genetic and genomic resources. A survey of biochemical and genetic interaction networks in Drosophila demonstrates no known interactions among these five genes (Flybase). In humans, there appear to be no known direct biochemical interactions among these genes. Yet, heterozygous LOF mutations in each of these genes are associated with risk for ASD in humans.

Forward genetic screen for altered baseline transmission and PHP

The screen that we performed is diagrammed in Figure 2A. We took advantage of a collection of small chromosomal deficiencies (5–50 genes per deficiency, each with known chromosomal breakpoints) that tile the 3^rd chromosome, uncovering approximately 6000 genes in total. For every double heterozygous combination of RIMS1/+ with a heterozygous deficiency, we performed multiple (n = 3–15) intracellular recordings, quantifying mEPSP amplitude, EPSP amplitude, quantal content (EPSP/mEPSP), resting membrane potential and input resistance. Recordings were made in the presence of PhTx to induce PHP. If the baseline EPSP is normal and quantal content is increased compared to wild type, then we can conclude that PHP is normally expressed. In these instances, we expect that baseline transmission was also normal in the absence of PhTx. However, if EPSPs are diminished in a given genetic combination (RIMS1/+, Df/+) and quantal content is not increased compared to wild type, then there are two possible origins: 1) the double mutant impairs baseline transmission or 2) baseline transmission is normal and PHP is selectively impaired. In these instances, the double heterozygous mutant combinations were re-assessed in the absence of PhTx to test for altered baseline transmission.

Figure 2

Download asset Open asset

Double heterozygous combinations that strongly affected muscle resting potential or input resistance were not observed. We uncovered two instances where a mutant combination (RIMS1/+, Df/+) caused a specific deficit in baseline transmission, without altering PHP. We did not isolate any double heterozygous conditions with dramatic defects in mEPSP amplitude. The majority of double mutant combinations specifically affected the expression of PHP. This was unexpected.

Double mutant combinations were determined to disrupt PHP by satisfying one of two criteria. First, the average quantal content (+PhTx) had to be more than one standard deviation below the population mean of all genotypes (Figure 2B, solid horizontal black line). Second, average quantal content had to reside outside a boundary that encompasses 95% of all individual recordings made in the RIMS1/+ mutant alone (Figure 2B, black dashed lines). Two example ‘hits’ are shown in red (Figure 2B; dark red point shows data in the absence of PhTx and light red point shows data recorded in the presence of PhTx, and the red lines simply connect the points for a given genotype for the purposes of data display). We also present a complete data set for a single hit from the screen as a standard format bar graph with representative traces (Figure 2D,E). Note that the heterozygous deficiency has normal baseline transmission and PHP (p>0.1 One-way ANOVA with post-hoc Tukey multiple comparison), but when combined with a heterozygous mutation in RIMS1, PHP completely and selectively fails.

In total, our screen identified, and we subsequently confirmed, 20 small deficiencies that cause PHP to fail when combined with RIMS1/+. To achieve a final list of 20 hits, each potential hit was re-validated in a second set of experiments, increasing sample sizes (generally 7–16 NMJ). During the process of re-validation, we rule out approximately one third of the potential hits selected from the screen. The identified deficiencies are randomly distributed across the 3^rd chromosome (Figure 2C). The screen was empirically validated by the identification, blind to genotype, of deficiencies that uncovered the RIMS1 locus, as well as the Pi3K68D locus (not included in hit list), previously shown to interact as a double heterozygous mutant with RIMS1/+ (Hauswirth et al., 2018). Furthermore, the rim binding protein (RBP) locus was not identified as disrupting PHP, consistent with the previously published observation that a rbp/+ mutant does not interact with RIMS1/+ for PHP (Müller et al., 2015). However, rbp/+ did interact with RIMS1/+ for baseline neurotransmitter release as expected based on previously published data (not shown) (Müller et al., 2015). No other genes previously implicated in the mechanisms of PHP were present in the deficiencies isolated in our screen. It is important to note that, according to a formal genetic analysis, no strong conclusion can be made regarding the negative result of a double heterozygous genetic interaction (see Supplemental Tables S2-S5 for further detailed information on the screen results).

Finally, we assessed whether there was any relationship between the number of genes that were deleted within a given deficiency and the robustness of PHP. One hypothesis is that the additive effects of multiple, heterozygous gene mutations would increase for larger deficiencies and PHP would be increasingly compromised. That was not the case (Figure 3). There was no correlation between the number of genes uncovered by a given deficiency and EPSP amplitude recorded in the presence of PhTx (R² = 0.003; Figure 3A). Thus, impaired PHP cannot be accounted for by a simple additive accumulation of genetic mutations within a given deficiency.

Figure 3

Download asset Open asset

Identification of common phenotypic enhancers of multiple unrelated ASD genes

The results of our forward genetic screen, encompassing approximately one third of the Drosophila genome, might identify genetic interactions specific to RIMS1. However, we reasoned that because RIMS1 also showed a strong genetic interaction with CHD8, as well as CHD2 and ASH1L (Figure 1; Figure 1—figure supplement 2), a portion of the hits from our screen might also interact with these genes. When initial experiments confirmed that this was the case, we expanded our analysis to encompass all five of the ASD-associated gene orthologs from Figure 1. Thus, we performed a systems-genetic test of all possible double heterozygous genetic interactions, using wild type and the five ASD gene orthologs introduced in Figure 1 combined with wild type and five hits (deficiencies) randomly selected from our forward genetic screen. In total, we tested 36 genetic combinations for baseline transmission and PHP, recording every genotype in the presence and absence of PhTx (Figure 4).

Figure 4 with 1 supplement see all

Download asset Open asset

To facilitate visual interpretation, genetic interaction data are presented as a heat map superimposed on a matrix representing all genetic combinations, in the presence or absence of PhTx (Figure 4A). All but one genotype responded to the application of PhTx with decreased mEPSP amplitudes (Figure 4A, mEPSP; compare top left matrix with top right matrix, the transition from blue to red indicates diminished average mEPSP for each genetic combination). Thus, we induced homeostatic pressure in 35 out of 36 genetic combinations (CHD8/+ with Df(3)7562/+ being the exception). Next, we demonstrate that all heterozygous deficiencies (x-axis) or heterozygous ASD-associated gene mutations (y-axis), when crossed to the wild type strain (w1118) showed normal EPSP amplitudes in the absence and presence of PhTx, demonstrating robust induction of PHP (Figure 4A, EPSP, bottom right matrix). Next, nearly all (23 out of 25) of the double heterozygous combinations show normal EPSP amplitudes in the absence of PhTx, demonstrating normal baseline neurotransmission (Figure 4A, bottom left matrix). However, a majority (16 of 25) of the double heterozygous genetic combinations showed a failure of PHP in the presence of PhTx (Figure 4A, red and light-red boxes, bottom right matrix). In Figure 4B, we also plot the induction of PHP for each double heterozygous combination by calculating the percent change in quantal content following PhTx application (Figure 4B, top matrix). Here, if quantal content does not change (<15% change; gray), then PHP is impaired or blocked. Moderate increases in quantal content (15–30% change; light orange) suggest suppression of PHP, in some instances being statistically significant suppression (see below).

We performed statistical analyses for each double mutant combination, asking whether there was a statistically significant increase in quantal content for a given double mutant in the presence of PhTx compared to that same double mutant combination in the absence of PhTx (Figure 4B, bottom matrix.) Note that we are testing whether PHP is induced in a given double heterozygous mutant combination (an individual square in the matrix), comparing quantal content in the absence and presence of PhTx. We do not compare PHP expression among different double heterozygous mutant combinations. The colors gray and ‘light pink’ each reflect a complete block of PHP, an effect that is observed in the majority of double mutant combinations. As a complementary statistical analysis, we tested the differences between individual genotypic conditions (quantal content in each box) versus the wild type quantal content (One-way ANOVA with Dunnett’s multiple comparisons) (Figure 4—figure supplement 1). In this case, if PHP is blocked, then there will be a statistically significant difference compared to wild type. Again, 18 of 25 comparisons are significantly different. The genotypic comparison against wild type verified the analysis based on individual genotypic comparisons (Figure 4B). It should be noted that, in a few instances, minor differences were observed caused by a change in quantal content that was significant (-/+PhTx), but which remained smaller compared to wild type and therefore became significant. Thus, comparisons within genotypes (-/+PhTx) seem to assess the presence or absence of PHP most accurately (Figure 4B).

Our data demonstrate that four out of five deficiencies, isolated in our forward genetic screen as interacting with RIMS1, also cause PHP to fail when combined with any one of four different heterozygous ASD-associated gene mutations (Figure 4B, bottom). The pattern of PHP blockade is not uniform. WDFY3, CHD8 and CHD2 show a common pattern of interactions with the same three deficiencies. However, ASH1L interacts with only two out of the five tested deficiencies. To our knowledge, this is the first demonstration, in any system, of common phenotypic enhancement for multiple, independent and unrelated ASD gene orthologs.

In Figure 4C–F, we elaborate on three of the genetic interactions with data presentations that are more detailed. We show evidence of normal PHP in CHD8/+ (Figure 4C, replicated from Figure 1). In the adjacent graph (Figure 4C, right), we show evidence of a strong disruption of PHP in the double heterozygous combination of CHD8/+ with a heterozygous deficiency (Df(3)24410/+) isolated in genetic screen. A similar analysis is presented for the ASH1L/+ heterozygous gene mutation and the interaction with a different heterozygous deficiency (Figure 4D). Finally, a third genetic interaction is presented in a format that is standard for the field of homeostatic plasticity (Figure 4E,F), inclusive of representative traces (Figure 4E) and bar graphs with associated statistical analyses (Figure 4F). Note that values for all recordings are presented (Supplementary file 1 Supplemental Table S6). Several additional controls were performed to validate and extend the findings reported for our genetic interaction data set. First, we note that all double heterozygous mutant combinations are adult viable. Thus, it was possible to inspect adult animals for phenotypes that might indicate altered signaling. Inspection of the compound eye and wings (bristles, wing veins and size) demonstrate wild type tissue morphogenesis (data not shown).

PDPK1 and PPP2R5D are common phenotypic enhancers of multiple ASD gene orthologs

We isolated the causal single gene mutations within two of the deficiencies isolated from our screen. To do so, we tested smaller sub-deficiencies that mapped within the originally isolated deficiencies. Sub-deficiency mapping either identified the causal gene, or a limited number of candidates. We subsequently tested individual gene candidates with established single gene mutations or RNAi. The process of mapping to single genes, therefore, included several rounds of independent phenotype verification. The first two instances in which we have isolated single causal genes are presented. Each candidate gene was tested individually against all five ASD gene orthologs, using previously published mutations (Figure 5). For both genes, we confirmed the same set of genetic interactions that occurred when analyzing the deficiency that included the identified gene (Figure 5).

Figure 5 with 3 supplements see all

Download asset Open asset

The first gene that we identified encodes a serine threonine kinase encoded by the PDPK1 gene (PDK1 in Drosophila). PDPK1 is a master controller of cellular metabolism, as well as cellular and synaptic growth at the Drosophila NMJ (Cheng et al., 2011). The second gene encodes a regulatory subunit of the PP2A phosphatase encoded by the PPP2R5D gene (wrd in Drosophila) (Viquez, 2006). PPP2R5D is also a master controller of cellular metabolism (Bernal et al., 2014), as well as cellular and synaptic growth at the Drosophila NMJ (Viquez, 2006). Both proteins are present at the Drosophila NMJ (Cheng et al., 2011; Viquez et al., 2009; Viquez, 2006) Single-cell Patch-Seq experiments confirmed the expression of these genes in Drosophila motoneurons at third-instar (Figure 1—figure supplement 1). Intriguingly, the PPP2R5D gene has recently been associated with intellectual disability and autism in human (Loveday et al., 2015; Shang et al., 2016).

Next, we demonstrate that both genes, PDPK1 and PPP2R5D, are common modifiers of multiple heterozygous ASD-associated gene mutations (Figure 5). To underscore the specificity of the double heterozygous genetic interactions (Figure 5B,E), we restored the expression of the CHD2 gene to wild type levels in the CHD2/+; PPP2R5D/+ double heterozygous combination. This was achieved using a previously published translocation of the CHD2 gene locus that allowed us to put back one copy of the CHD2 gene (CHD2^WT/+) in the background of the CHD2/+; PPP2R5D/+ double mutant. We demonstrate that PHP is fully restored (Figure 5C). An identical series of experiments was performed with a previously characterized PDPK1 mutation (Figure 5E,F). Thus, PDPK1 and PPP2R5D represent the first known common phenotypic modifiers of ASD gene orthologs, causing PHP to fail (see also Supplementary file 1 Supplemental Table S7 for values and statistics). The data also underscore that deficiencies isolated in our forward genetic screen can be resolved to the activity of single genes. As such, the screen may have identified a novel class of common phenotypic modifier.

Another series of control experiments were performed. We asked whether neuronal morphology was substantially altered in the heterozygous ASD-associated gene mutations and in select double heterozygous genetic interactions (Figure 5—figure supplement 1). We do find evidence that the heterozygous CHD8/+ mutation predisposes the NMJ to modest overgrowth, consistent with CHD8 influencing brain development in other systems (Gompers et al., 2017). But, this effect does not become more severe when combined with either the PDPK1 or PPP2R5D mutation as double heterozygotes. Thus, we conclude that altered synaptic growth is not highly correlated with the block of PHP in these double heterozygous combinations.

Ultrastructural correlate of impaired PHP: altered presynaptic membrane trafficking

Next, we turned to electron microscopy to determine whether the genetic interaction of PPP2R5D/+ with CHD8/+ affects the presynaptic release site. Ultrastructural changes have previously been linked to impaired PHP (Harris et al., 2018). Thin section transmission electron microscopy (EM) was used to examine the synapse, defined as a characteristic increase in pre- and postsynaptic membrane electron density, opposing clustered presynaptic vesicles and a characteristic presynaptic density, termed a T-bar. We find that the ultrastructure of CHD8/+ alone was wild type (Figure 6B,D,E). The ultrastructure of PPP2R5D/+ alone was wild type (Figure 6A,D,E). However, the double heterozygous mutant showed evidence of large membrane structures surrounding the presynaptic release site and apparent stalled endocytic events, appearing adjacent to sites of neurotransmitter release where compensatory synaptic vesicle endocytosis occurs (Figure 6C, insets). Quantification of vesicle size reveals a large increase in average intracellular vesicle diameter for all vesicles within 150 nm of the base of the presynaptic release site, defined by the T-bar structure (Figure 6D,E), again selective to the double heterozygous mutant. These data provide a striking visual confirmation of the genetic interaction between PPP2R5D/+ and CHD8/+. And, this is further evidence linking the action of a chromatin-remodeling factor (CHD8) to the stability of synaptic transmission.

Figure 6 with 1 supplement see all

Download asset Open asset

Given the appearance of enlarged vesicles at or near the presynaptic release site, we repeated our ultrastructural analysis of the double heterozygous mutant, fixing the synapse immediately (~1–5 s) after strong stimulation of presynaptic release (50 Hz stimulation, 10 s). In wild type, there was no change in the number or appearance of presynaptic vesicles when fixed immediately following the stimulus. However, in the double heterozygous mutant condition (CHD8/+; PPP2R5D/+) we found that intracellular vesicles were further increased in size and took on a crenulated appearance (Figure 6F–I). These data are consistent with the enlarged vesicles being endosomal intermediates, arguing that the process of vesicle recycling is altered in the double heterozygous mutant. In further support of this idea, we demonstrate enhanced synaptic depression in response to high frequency (50 Hz) stimulation (Figure 6—figure supplement 1). Recently, homozygous CHD8 loss of function mutations have been linked to defects in synaptic vesicle endocytosis at the Drosophila NMJ (Latcheva et al., 2019). Our data underscore that that an endocytosis phenotype can be uncovered in the heterozygous CHD8/+ mutant in the context of the PPP2R5D/+ mutant. Regardless of the underlying molecular mechanism leading to this EM phenotype and associated physiological deficits (a topic for future study), these data present a striking, visual confirmation of a strong synthetic genetic interaction between PPP2R5D/+ and the CHD8/+ heterozygous mutations. Furthermore, these data link the activity of a chromatin remodeling factor, present in the nucleus (CHD8), to a profound synaptic defect. Experiments detailed below, including genetic rescue, confirm the specificity of this EM phenotype.

Differential gene expression analyses

One possible reason that genes isolated from our screen are common modifiers of diverse ASD genes is that each modifier is a direct transcriptional target of the ASD mutants. It is possible to assess this by RNAseq. To our knowledge, side-by-side differential gene expression analysis has yet to be performed for multiple heterozygous ASD-associated gene mutant backgrounds. We performed whole genome RNAseq analysis for wild types and the four heterozygous ASD mutants (four biological replicates) (Figure 7A). We asked whether any of the genes contained within the 20 deficiencies identified in the screen (37 genes; a number arrived at following sub-deficiency mapping and sub-selection based on gene expression in nerve or muscle) are commonly altered in all four of ASD-associated mutants (Figure 7A, orange data points). None were commonly differentially regulated (p-value=0.096 for ASH1L/+; p-value=0.636 for WDFY3/+; p-value=0.392 for CHD2/+; p-value=0.112 for CHD8/+; Wilcoxon sign rank test two-sided). We conclude that common down-regulation of identified genetic modifiers cannot account for the common impairment of PHP that we observe electrophysiologically.

Figure 7

Download asset Open asset

Next, we asked whether the ASD-associated gene mutations might cause common changes in gene expression, with potential relevance to a common disruption of PHP. We define all differentially expressed genes common to at least two ASD mutations (Figure 7B,C). While there are individual genes that are commonly differentially regulated, a GO database analysis of differentially expressed genes did not reveal any consistent change in a gene category across all four genotypes. The patterns of gene dysregulation do not predict any pattern of genetic interactions documented in our systems-genetic analysis. Finally, while there are genes that are commonly dysregulated in multiple ASD gene orthologs, there are only two genes that are commonly down-regulated in all four ASD mutants (FBgn0027578 [Nepl21] and FBgn0037166 [CG11426]) (Figure 7C). FBgn0027578 encodes a metalloprotease of the Neprilysin family, with homology to endothelin converting enzyme one in human, of unknown function in the nervous system. FBgn0037166 encodes phosphatidic acid phosphatase type 2, which is expressed in the Drosophila nervous system, but of unknown function. There is no obvious means to connect the down regulation of these two genes to impaired homeostatic signaling, although future experiments will explore these genes in greater depth. Furthermore, there is no clear connection between FBgn0027578 or FBgn0037166 and the roles of either PDPK1 or PPP2R5D in the nervous system. Thus, a transcriptional analysis of heterozygous ASD gene mutations alone did not allow us to make clear progress toward understanding the mechanisms of impaired PHP.

Candidate mechanisms for impaired PHP based on differential gene expression analysis

Next, we continued with our focus on characterizing the homeostasis defect in the CHD8/+; PPPR25D/+ double heterozygous mutant combination. We repeated the RNAseq differential gene expression analysis comparing the double heterozygous condition to three control conditions, inclusive of wild type and each single heterozygous mutant alone. In this manner, we sought to identify synergistic effects on gene expression that could not be accounted for in either single heterozygous mutant alone (Figure 7D,E). As expected, many of the differentially expressed genes documented in the double heterozygous mutant, when compared to wild type, could be accounted for by subsequent comparisons to each single heterozygous mutant. However, a small number of genes (14 genes; 5 upregulated and nine downregulated) appear to be synergistically differentially expressed in the double heterozygous mutant compared to all three control conditions (Figure 7D,E). We successfully replicated altered expression of four genes in the double heterozygous mutant combination by quantitative RT-PCR (Figure 8A,B). Of these genes, CREG stood out as being robustly and dramatically up-regulated. Upon closer inspection, CREG showed a slight, but significant, up-regulation in the CHD8/+ mutant (Figure 7A), and this was enhanced by the presence of the heterozygous PPP2R5D/+ mutation (Figure 7E, left). Next, we confirmed the up-regulation of CREG in the third instar larval central nervous system by quantitative RT-PCR (Figure 8—figure supplement 1B). Finally, we took advantage of a previously published gene expression data set (Parrish et al., 2014) and document CREG expression in motoneurons throughout embryonic and larval development. CREG is strongly expressed in embryonic motoneurons (20–24 hr after egg laying – AEL), after which expression levels plummet (Figure 8—figure supplement 1A).

Figure 8 with 3 supplements see all

Download asset Open asset

In this study, we make several fundamental advances. First, we provide evidence that mutations in multiple different ASD-associated genes sensitize homeostatic plasticity to fail (Figure 9A,B). Second, using genome-scale forward genetics and subsequent systems-genetic analyses, we identify the first phenotypic modifiers that commonly enhance five different ASD-associated gene mutations, causing a specific failure of PHP (Figure 9A). Third, we identify PDPK1 and PPP2R5D as common phenotypic modifiers of multiple ASD-associated genes and, thereby, define a mechanistic link between synaptic transmission, PHP and chromatin remodeling complexes in the neuronal nucleus (Figure 9A). Finally, we define how PHP fails at the intersection of an ASD-associated gene mutation and phenotypic modifier. The mechanism is unexpected, involving the maladaptive up-regulation of a novel repressor of homeostatic plasticity (CREG) (Figure 9C). We demonstrate that up-regulation of CREG cannot explain other gene-gene interactions, underscoring the potential complexity of gene-gene interactions and the common failure of PHP. Regardless of potential mechanistic complexity, our data argue that impaired PHP may be a common pathophysiological effect downstream of LOF mutations in five different ASD-associated genes. If our data can be extended to additional ASD genes, and to other experimental systems including human neurons, then it may be possible to use this information to advance therapeutic approaches that modify ASD phenotypic severity regardless of the underlying genetic mutation(s) that confer risk for ASD.

Figure 9

Download asset Open asset

The loss or impairment of PHP could contribute to the phenotypic penetrance of an ASD gene mutation in multiple ways. Impaired PHP is expected to render the nervous system less robust to perturbation including the effects of environmental stress, immunological stress, or genetic mutation (Davis, 2013; Davis, 2006). If an ASD-associated gene mutation leads to neural developmental defects, then loss of PHP would be expected to exacerbate the functional consequences. According to the same logic, loss of PHP might enhance the adverse effects of environmental or immunological stress, both of which are thought to contribute to ASD pathophysiology (Beversdorf et al., 2018; Modabbernia et al., 2017). Finally, loss of PHP could be relevant to the appearance or severity of ASD comorbidities, including epilepsy.

It should be emphasized that failed homeostatic plasticity cannot be determined by simply assessing the phenotype of a heterozygous ASD-associated gene mutation. The observation of a phenotype, such as altered E/I balance or impaired neurotransmission, could reflect failure of homeostatic plasticity, or it could reflect the outcome of successful homeostatic mechanisms that constrained a phenotype that might otherwise have been more severe (Davis, 2013; Kulik et al., 2019). Ultimately, the loss or impairment of homeostatic plasticity can only be determined by a direct test of homeostatic robustness; specifically referring the ability of a neuron, synapse or neural circuit to respond to a perturbation and sustain normal function in the continued presence of the perturbation (Davis, 2013; Davis, 2006). Thus, our data set the stage for similar analyses in other model organisms, potentially extending the connection between ASD-associated gene mutations and the robustness of PHP or other forms of homeostatic plasticity.

Sequencing data have been deposited in GEO under accession code GSE153225. Analysis code is available via Github https://github.com/joonan30/Genc2020_RNAseq (copy archived at https://github.com/elifesciences-publications/Genc2020_RNAseq).

1. 1. Genç O
   2. An J-Y
   3. Fetter RD
   4. Kulik Y
   5. Zunino G
   6. Sanders SJ
   7. Davis GW

   (2020) NCBI Gene Expression Omnibus

   ID GSE153225. Transcriptomics analysis of heterozygous mutant and wild-type flies for presynaptic homeostatic plasticity.

   http://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE153225

1. 1. Alessi DR
   2. James SR
   3. Downes CP
   4. Holmes AB
   5. Gaffney PR
   6. Reese CB
   7. Cohen P

   (1997) Characterization of a 3-phosphoinositide-dependent protein kinase which phosphorylates and activates protein kinase balpha

   Current Biology 7:261–269.

   https://doi.org/10.1016/S0960-9822(06)00122-9

   • PubMed
   • Google Scholar
2. 1. Antoine MW
   2. Langberg T
   3. Schnepel P
   4. Feldman DE

   (2019) Increased Excitation-Inhibition ratio stabilizes synapse and circuit excitability in four autism mouse models

   Neuron 101:648–661.

   https://doi.org/10.1016/j.neuron.2018.12.026

   • PubMed
   • Google Scholar
3. 1. Ashworth A
   2. Lord CJ
   3. Reis-Filho JS

   (2011) Genetic interactions in Cancer progression and treatment

   Cell 145:30–38.

   https://doi.org/10.1016/j.cell.2011.03.020

   • PubMed
   • Google Scholar
4. 1. Baetz KK
   2. Krogan NJ
   3. Emili A
   4. Greenblatt J
   5. Hieter P

   (2004) The ctf13-30/CTF13 genomic haploinsufficiency modifier screen identifies the yeast chromatin remodeling complex RSC, which is required for the establishment of sister chromatid cohesion

   Molecular and Cellular Biology 24:1232–1244.

   https://doi.org/10.1128/MCB.24.3.1232-1244.2003

   • PubMed
   • Google Scholar
5. 1. Bergquist S
   2. Dickman DK
   3. Davis GW

   (2010) A hierarchy of cell intrinsic and target-derived homeostatic signaling

   Neuron 66:220–234.

   https://doi.org/10.1016/j.neuron.2010.03.023

   • PubMed
   • Google Scholar
6. 1. Bernal M
   2. Zhurinsky J
   3. Iglesias-Romero AB
   4. Sanchez-Romero MA
   5. Flor-Parra I
   6. Tomas-Gallardo L
   7. Perez-Pulido AJ
   8. Jimenez J
   9. Daga RR

   (2014) Proteome-wide search for PP2A substrates in fission yeast

   Proteomics 14:1367–1380.

   https://doi.org/10.1002/pmic.201300136

   • PubMed
   • Google Scholar
7. 1. Beversdorf DQ
   2. Stevens HE
   3. Jones KL

   (2018) Prenatal stress, maternal immune dysregulation, and their association with autism spectrum disorders

   Current Psychiatry Reports 20:76.

   https://doi.org/10.1007/s11920-018-0945-4

   • PubMed
   • Google Scholar
8. 1. Bharucha N
   2. Chabrier-Rosello Y
   3. Xu T
   4. Johnson C
   5. Sobczynski S
   6. Song Q
   7. Dobry CJ
   8. Eckwahl MJ
   9. Anderson CP
   10. Benjamin AJ
   11. Kumar A
   12. Krysan DJ

   (2011) A large-scale complex haploinsufficiency-based genetic interaction screen in Candida Albicans: analysis of the RAM network during morphogenesis

   PLOS Genetics 7:e1002058.

   https://doi.org/10.1371/journal.pgen.1002058

   • PubMed
   • Google Scholar
9. 1. Bourgeron T

   (2015) From the genetic architecture to synaptic plasticity in autism spectrum disorder

   Nature Reviews Neuroscience 16:551–563.

   https://doi.org/10.1038/nrn3992

   • Google Scholar
10. 1. Brusich DJ
    2. Spring AM
    3. Frank CA

    (2015) A single-cross, RNA interference-based genetic tool for examining the long-term maintenance of homeostatic plasticity

    Frontiers in Cellular Neuroscience 9:107.

    https://doi.org/10.3389/fncel.2015.00107

    • PubMed
    • Google Scholar
11. 1. Chan DA
    2. Giaccia AJ

    (2011) Harnessing synthetic lethal interactions in anticancer drug discovery

    Nature Reviews Drug Discovery 10:351–364.

    https://doi.org/10.1038/nrd3374

    • PubMed
    • Google Scholar
12. 1. Chen R
    2. Shi L
    3. Hakenberg J
    4. Naughton B
    5. Sklar P
    6. Zhang J
    7. Zhou H
    8. Tian L
    9. Prakash O
    10. Lemire M
    11. Sleiman P
    12. Cheng WY
    13. Chen W
    14. Shah H
    15. Shen Y
    16. Fromer M
    17. Omberg L
    18. Deardorff MA
    19. Zackai E
    20. Bobe JR
    21. Levin E
    22. Hudson TJ
    23. Groop L
    24. Wang J
    25. Hakonarson H
    26. Wojcicki A
    27. Diaz GA
    28. Edelmann L
    29. Schadt EE
    30. Friend SH

    (2016) Analysis of 589,306 genomes identifies individuals resilient to severe mendelian childhood diseases

    Nature Biotechnology 34:531–538.

    https://doi.org/10.1038/nbt.3514

    • PubMed
    • Google Scholar
13. 1. Cheng L
    2. Locke C
    3. Davis GW

    (2011) S6 kinase localizes to the presynaptic active zone and functions with PDK1 to control synapse development

    The Journal of Cell Biology 194:921–935.

    https://doi.org/10.1083/jcb.201101042

    • PubMed
    • Google Scholar
14. 1. Davis GW

    (2006) Homeostatic control of neural activity: from phenomenology to molecular design

    Annual Review of Neuroscience 29:307–323.

    https://doi.org/10.1146/annurev.neuro.28.061604.135751

    • PubMed
    • Google Scholar
15. 1. Davis GW

    (2013) Homeostatic signaling and the stabilization of neural function

    Neuron 80:718–728.

    https://doi.org/10.1016/j.neuron.2013.09.044

    • PubMed
    • Google Scholar
16. 1. De Rubeis S
    2. He X
    3. Goldberg AP
    4. Poultney CS
    5. Samocha K
    6. Cicek AE
    7. Kou Y
    8. Liu L
    9. Fromer M
    10. Walker S
    11. Singh T
    12. Klei L
    13. Kosmicki J
    14. Shih-Chen F
    15. Aleksic B
    16. Biscaldi M
    17. Bolton PF
    18. Brownfeld JM
    19. Cai J
    20. Campbell NG
    21. Carracedo A
    22. Chahrour MH
    23. Chiocchetti AG
    24. Coon H
    25. Crawford EL
    26. Curran SR
    27. Dawson G
    28. Duketis E
    29. Fernandez BA
    30. Gallagher L
    31. Geller E
    32. Guter SJ
    33. Hill RS
    34. Ionita-Laza J
    35. Jimenz Gonzalez P
    36. Kilpinen H
    37. Klauck SM
    38. Kolevzon A
    39. Lee I
    40. Lei I
    41. Lei J
    42. Lehtimäki T
    43. Lin CF
    44. Ma'ayan A
    45. Marshall CR
    46. McInnes AL
    47. Neale B
    48. Owen MJ
    49. Ozaki N
    50. Parellada M
    51. Parr JR
    52. Purcell S
    53. Puura K
    54. Rajagopalan D
    55. Rehnström K
    56. Reichenberg A
    57. Sabo A
    58. Sachse M
    59. Sanders SJ
    60. Schafer C
    61. Schulte-Rüther M
    62. Skuse D
    63. Stevens C
    64. Szatmari P
    65. Tammimies K
    66. Valladares O
    67. Voran A
    68. Li-San W
    69. Weiss LA
    70. Willsey AJ
    71. Yu TW
    72. Yuen RK
    73. Cook EH
    74. Freitag CM
    75. Gill M
    76. Hultman CM
    77. Lehner T
    78. Palotie A
    79. Schellenberg GD
    80. Sklar P
    81. State MW
    82. Sutcliffe JS
    83. Walsh CA
    84. Scherer SW
    85. Zwick ME
    86. Barett JC
    87. Cutler DJ
    88. Roeder K
    89. Devlin B
    90. Daly MJ
    91. Buxbaum JD
    92. DDD Study, Homozygosity Mapping Collaborative for Autism, UK10K Consortium

    (2014) Synaptic, transcriptional and chromatin genes disrupted in autism

    Nature 515:209–215.

    https://doi.org/10.1038/nature13772

    • PubMed
    • Google Scholar
17. 1. Deeg KE
    2. Aizenman CD

    (2011) Sensory modality-specific homeostatic plasticity in the developing optic tectum

    Nature Neuroscience 14:548–550.

    https://doi.org/10.1038/nn.2772

    • PubMed
    • Google Scholar
18. 1. Dobin A
    2. Davis CA
    3. Schlesinger F
    4. Drenkow J
    5. Zaleski C
    6. Jha S
    7. Batut P
    8. Chaisson M
    9. Gingeras TR

    (2013) STAR: ultrafast universal RNA-seq aligner

    Bioinformatics 29:15–21.

    https://doi.org/10.1093/bioinformatics/bts635

    • PubMed
    • Google Scholar
19. 1. Eaton BA
    2. Fetter RD
    3. Davis GW

    (2002) Dynactin is necessary for synapse stabilization

    Neuron 34:729–741.

    https://doi.org/10.1016/S0896-6273(02)00721-3

    • PubMed
    • Google Scholar
20. 1. Frank CA
    2. Kennedy MJ
    3. Goold CP
    4. Marek KW
    5. Davis GW

    (2006) Mechanisms underlying the rapid induction and sustained expression of synaptic homeostasis

    Neuron 52:663–677.

    https://doi.org/10.1016/j.neuron.2006.09.029

    • PubMed
    • Google Scholar
21. 1. Frank CA
    2. Pielage J
    3. Davis GW

    (2009) A presynaptic homeostatic signaling system composed of the eph receptor, ephexin, Cdc42, and CaV2.1 calcium channels

    Neuron 61:556–569.

    https://doi.org/10.1016/j.neuron.2008.12.028

    • PubMed
    • Google Scholar
22. 1. Friend SH
    2. Schadt EE

    (2014) Clues from the resilient

    Science 344:970–972.

    https://doi.org/10.1126/science.1255648

    • Google Scholar
23. 1. Gaugler T
    2. Klei L
    3. Sanders SJ
    4. Bodea CA
    5. Goldberg AP
    6. Lee AB
    7. Mahajan M
    8. Manaa D
    9. Pawitan Y
    10. Reichert J
    11. Ripke S
    12. Sandin S
    13. Sklar P
    14. Svantesson O
    15. Reichenberg A
    16. Hultman CM
    17. Devlin B
    18. Roeder K
    19. Buxbaum JD

    (2014) Most genetic risk for autism resides with common variation

    Nature Genetics 46:881–885.

    https://doi.org/10.1038/ng.3039

    • PubMed
    • Google Scholar
24. 1. Genç Ö
    2. Dickman DK
    3. Ma W
    4. Tong A
    5. Fetter RD
    6. Davis GW

    (2017) MCTP is an ER-resident calcium sensor that stabilizes synaptic transmission and homeostatic plasticity

    eLife 6:e22904.

    https://doi.org/10.7554/eLife.22904

    • PubMed
    • Google Scholar
25. 1. Ghobrial G
    2. Araujo L
    3. Jinwala F
    4. Li S
    5. Lee LY

    (2018) The structure and biological function of CREG

    Frontiers in Cell and Developmental Biology 6:136.

    https://doi.org/10.3389/fcell.2018.00136

    • PubMed
    • Google Scholar
26. 1. Gibson G

    (2009) Decanalization and the origin of complex disease

    Nature Reviews Genetics 10:134–140.

    https://doi.org/10.1038/nrg2502

    • PubMed
    • Google Scholar
27. 1. Gompers AL
    2. Su-Feher L
    3. Ellegood J
    4. Copping NA
    5. Riyadh MA
    6. Stradleigh TW
    7. Pride MC
    8. Schaffler MD
    9. Wade AA
    10. Catta-Preta R
    11. Zdilar I
    12. Louis S
    13. Kaushik G
    14. Mannion BJ
    15. Plajzer-Frick I
    16. Afzal V
    17. Visel A
    18. Pennacchio LA
    19. Dickel DE
    20. Lerch JP
    21. Crawley JN
    22. Zarbalis KS
    23. Silverman JL
    24. Nord AS

    (2017) Germline Chd8 haploinsufficiency alters brain development in mouse

    Nature Neuroscience 20:1062–1073.

    https://doi.org/10.1038/nn.4592

    • PubMed
    • Google Scholar
28. 1. Harris N
    2. Braiser DJ
    3. Dickman DK
    4. Fetter RD
    5. Tong A
    6. Davis GW

    (2015) The innate immune receptor PGRP-LC controls presynaptic homeostatic plasticity

    Neuron 88:1157–1164.

    https://doi.org/10.1016/j.neuron.2015.10.049

    • PubMed
    • Google Scholar
29. 1. Harris N
    2. Fetter RD
    3. Brasier DJ
    4. Tong A
    5. Davis GW

    (2018) Molecular interface of neuronal innate immunity, synaptic vesicle stabilization, and presynaptic homeostatic plasticity

    Neuron 100:1163–1179.

    https://doi.org/10.1016/j.neuron.2018.09.048

    • PubMed
    • Google Scholar
30. 1. Hartman JL
    2. Garvik B
    3. Hartwell L

    (2001) Principles for the buffering of genetic variation

    Science 291:1001–1004.

    https://doi.org/10.1126/science.291.5506.1001

    • PubMed
    • Google Scholar
31. 1. Hauswirth AG
    2. Ford KJ
    3. Wang T
    4. Fetter RD
    5. Tong A
    6. Davis GW

    (2018) A postsynaptic PI3K-cII dependent signaling controller for presynaptic homeostatic plasticity

    eLife 7:e31535.

    https://doi.org/10.7554/eLife.31535

    • PubMed
    • Google Scholar
32. 1. Hengen KB
    2. Lambo ME
    3. Van Hooser SD
    4. Katz DB
    5. Turrigiano GG

    (2013) Firing rate homeostasis in visual cortex of freely behaving rodents

    Neuron 80:335–342.

    https://doi.org/10.1016/j.neuron.2013.08.038

    • PubMed
    • Google Scholar
33. 1. Henry FE
    2. McCartney AJ
    3. Neely R
    4. Perez AS
    5. Carruthers CJ
    6. Stuenkel EL
    7. Inoki K
    8. Sutton MA

    (2012) Retrograde changes in presynaptic function driven by dendritic mTORC1

    Journal of Neuroscience 32:17128–17142.

    https://doi.org/10.1523/JNEUROSCI.2149-12.2012

    • PubMed
    • Google Scholar
34. 1. Hou J
    2. Tan G
    3. Fink GR
    4. Andrews BJ
    5. Boone C

    (2019) Complex modifier landscape underlying genetic background effects

    PNAS 116:5045–5054.

    https://doi.org/10.1073/pnas.1820915116

    • PubMed
    • Google Scholar
35. Software

    1. Hunter S

    (2019) HTStream, version 1.1.0

    GitHub.

    https://github.com/s4hts/HTStream
36. 1. Iossifov I
    2. O'Roak BJ
    3. Sanders SJ
    4. Ronemus M
    5. Krumm N
    6. Levy D
    7. Stessman HA
    8. Witherspoon KT
    9. Vives L
    10. Patterson KE
    11. Smith JD
    12. Paeper B
    13. Nickerson DA
    14. Dea J
    15. Dong S
    16. Gonzalez LE
    17. Mandell JD
    18. Mane SM
    19. Murtha MT
    20. Sullivan CA
    21. Walker MF
    22. Waqar Z
    23. Wei L
    24. Willsey AJ
    25. Yamrom B
    26. Lee YH
    27. Grabowska E
    28. Dalkic E
    29. Wang Z
    30. Marks S
    31. Andrews P
    32. Leotta A
    33. Kendall J
    34. Hakker I
    35. Rosenbaum J
    36. Ma B
    37. Rodgers L
    38. Troge J
    39. Narzisi G
    40. Yoon S
    41. Schatz MC
    42. Ye K
    43. McCombie WR
    44. Shendure J
    45. Eichler EE
    46. State MW
    47. Wigler M

    (2014) The contribution of de novo coding mutations to autism spectrum disorder

    Nature 515:216–221.

    https://doi.org/10.1038/nature13908

    • PubMed
    • Google Scholar
37. 1. Jakawich SK
    2. Nasser HB
    3. Strong MJ
    4. McCartney AJ
    5. Perez AS
    6. Rakesh N
    7. Carruthers CJL
    8. Sutton MA

    (2010) Local presynaptic activity gates homeostatic changes in presynaptic function driven by dendritic BDNF synthesis

    Neuron 68:1143–1158.

    https://doi.org/10.1016/j.neuron.2010.11.034

    • Google Scholar
38. 1. Kim SH
    2. Ryan TA

    (2010) CDK5 serves as a major control point in neurotransmitter release

    Neuron 67:797–809.

    https://doi.org/10.1016/j.neuron.2010.08.003

    • PubMed
    • Google Scholar
39. 1. Kitano H

    (2007) Towards a theory of biological robustness

    Molecular Systems Biology 3:137.

    https://doi.org/10.1038/msb4100179

    • PubMed
    • Google Scholar
40. 1. Kowalewski-Nimmerfall E
    2. Schähs P
    3. Maresch D
    4. Rendic D
    5. Krämer H
    6. Mach L

    (2014) Drosophila melanogaster cellular repressor of E1A-stimulated genes is a lysosomal protein essential for fly development

    Biochimica Et Biophysica Acta (BBA) - Molecular Cell Research 1843:2900–2912.

    https://doi.org/10.1016/j.bbamcr.2014.08.012

    • Google Scholar
41. 1. Kulik Y
    2. Jones R
    3. Moughamian AJ
    4. Whippen J
    5. Davis GW

    (2019) Dual separable feedback systems govern firing rate homeostasis

    eLife 8:e45717.

    https://doi.org/10.7554/eLife.45717

    • PubMed
    • Google Scholar
42. 1. Latcheva NK
    2. Delaney TL
    3. Viveiros JM
    4. Smith RA
    5. Bernard KM
    6. Harsin B
    7. Marenda DR
    8. Liebl FLW

    (2019) The CHD protein, kismet, is important for the recycling of synaptic vesicles during endocytosis

    Scientific Reports 9:19368.

    https://doi.org/10.1038/s41598-019-55900-6

    • PubMed
    • Google Scholar
43. 1. Leppa VM
    2. Kravitz SN
    3. Martin CL
    4. Andrieux J
    5. Le Caignec C
    6. Martin-Coignard D
    7. DyBuncio C
    8. Sanders SJ
    9. Lowe JK
    10. Cantor RM
    11. Geschwind DH

    (2016) Rare inherited and de novo CNVs reveal complex contributions to ASD risk in multiplex families

    The American Journal of Human Genetics 99:540–554.

    https://doi.org/10.1016/j.ajhg.2016.06.036

    • PubMed
    • Google Scholar
44. 1. Liu G
    2. Tsien RW

    (1995) Properties of synaptic transmission at single hippocampal synaptic boutons

    Nature 375:404–408.

    https://doi.org/10.1038/375404a0

    • PubMed
    • Google Scholar
45. 1. Loveday C
    2. Tatton-Brown K
    3. Clarke M
    4. Westwood I
    5. Renwick A
    6. Ramsay E
    7. Nemeth A
    8. Campbell J
    9. Joss S
    10. Gardner M
    11. Zachariou A
    12. Elliott A
    13. Ruark E
    14. van Montfort R
    15. Rahman N
    16. Childhood Overgrowth Collaboration

    (2015) Mutations in the PP2A regulatory subunit B family genes PPP2R5B, PPP2R5C and PPP2R5D cause human overgrowth

    Human Molecular Genetics 24:4775–4779.

    https://doi.org/10.1093/hmg/ddv182

    • PubMed
    • Google Scholar
46. 1. Maffei A
    2. Fontanini A

    (2009) Network homeostasis: a matter of coordination

    Current Opinion in Neurobiology 19:168–173.

    https://doi.org/10.1016/j.conb.2009.05.012

    • PubMed
    • Google Scholar
47. 1. Mair B
    2. Moffat J
    3. Boone C
    4. Andrews BJ

    (2019) Genetic interaction networks in Cancer cells

    Current Opinion in Genetics & Development 54:64–72.

    https://doi.org/10.1016/j.gde.2019.03.002

    • PubMed
    • Google Scholar
48. 1. Manning BD
    2. Toker A

    (2017) AKT/PKB signaling: navigating the network

    Cell 169:381–405.

    https://doi.org/10.1016/j.cell.2017.04.001

    • PubMed
    • Google Scholar
49. 1. Marder E

    (2011) Variability, compensation, and modulation in neurons and circuits

    PNAS 108 Suppl 3:15542–15548.

    https://doi.org/10.1073/pnas.1010674108

    • PubMed
    • Google Scholar
50. 1. Marder E
    2. Goaillard JM

    (2006) Variability, compensation and homeostasis in neuron and network function

    Nature Reviews Neuroscience 7:563–574.

    https://doi.org/10.1038/nrn1949

    • PubMed
    • Google Scholar
51. 1. Modabbernia A
    2. Velthorst E
    3. Reichenberg A

    (2017) Environmental risk factors for autism: an evidence-based review of systematic reviews and meta-analyses

    Molecular Autism 8:13.

    https://doi.org/10.1186/s13229-017-0121-4

    • PubMed
    • Google Scholar
52. 1. Müller M
    2. Liu KS
    3. Sigrist SJ
    4. Davis GW

    (2012) RIM controls homeostatic plasticity through modulation of the readily-releasable vesicle pool

    Journal of Neuroscience 32:16574–16585.

    https://doi.org/10.1523/JNEUROSCI.0981-12.2012

    • PubMed
    • Google Scholar
53. 1. Müller M
    2. Genç Ö
    3. Davis GW

    (2015) RIM-binding protein links synaptic homeostasis to the stabilization and replenishment of high release probability vesicles

    Neuron 85:1056–1069.

    https://doi.org/10.1016/j.neuron.2015.01.024

    • PubMed
    • Google Scholar
54. 1. Müller M
    2. Davis GW

    (2012) Transsynaptic control of presynaptic Ca2+ influx achieves homeostatic potentiation of neurotransmitter release

    Current Biology 22:1102–1108.

    https://doi.org/10.1016/j.cub.2012.04.018

    • Google Scholar
55. 1. Mullins C
    2. Fishell G
    3. Tsien RW

    (2016) Unifying views of autism spectrum disorders: a consideration of autoregulatory feedback loops

    Neuron 89:1131–1156.

    https://doi.org/10.1016/j.neuron.2016.02.017

    • PubMed
    • Google Scholar
56. 1. Nelson SB
    2. Valakh V

    (2015) Excitatory/Inhibitory balance and circuit homeostasis in autism spectrum disorders

    Neuron 87:684–698.

    https://doi.org/10.1016/j.neuron.2015.07.033

    • PubMed
    • Google Scholar
57. 1. Neuner SM
    2. Heuer SE
    3. Huentelman MJ
    4. O'Connell KMS
    5. Kaczorowski CC

    (2019) Harnessing genetic complexity to enhance translatability of Alzheimer's Disease Mouse Models: A Path toward Precision Medicine

    Neuron 101:399–411.

    https://doi.org/10.1016/j.neuron.2018.11.040

    • PubMed
    • Google Scholar
58. 1. O'Neil NJ
    2. Bailey ML
    3. Hieter P

    (2017) Synthetic lethality and cancer

    Nature Reviews Genetics 18:613–623.

    https://doi.org/10.1038/nrg.2017.47

    • PubMed
    • Google Scholar
59. 1. Ortega JM
    2. Genç Ö
    3. Davis GW

    (2018) Molecular mechanisms that stabilize short term synaptic plasticity during presynaptic homeostatic plasticity

    eLife 7:e40835.

    https://doi.org/10.7554/eLife.40385

    • Google Scholar
60. 1. Parrish JZ
    2. Kim CC
    3. Tang L
    4. Bergquist S
    5. Wang T
    6. Derisi JL
    7. Jan LY
    8. Jan YN
    9. Davis GW

    (2014) Krüppel mediates the selective rebalancing of ion channel expression

    Neuron 82:537–544.

    https://doi.org/10.1016/j.neuron.2014.03.015

    • PubMed
    • Google Scholar
61. 1. Penney J
    2. Tsurudome K
    3. Liao EH
    4. Elazzouzi F
    5. Livingstone M
    6. Gonzalez M
    7. Sonenberg N
    8. Haghighi AP

    (2012) TOR is required for the retrograde regulation of synaptic homeostasis at the Drosophila neuromuscular junction

    Neuron 74:166–178.

    https://doi.org/10.1016/j.neuron.2012.01.030

    • PubMed
    • Google Scholar
62. 1. Peter B
    2. Dinu V
    3. Liu L
    4. Huentelman M
    5. Naymik M
    6. Lancaster H
    7. Vose C
    8. Schrauwen I

    (2019) Exome sequencing of two siblings with sporadic autism spectrum disorder and severe speech sound disorder suggests pleiotropic and complex effects

    Behavior Genetics 49:399–414.

    https://doi.org/10.1007/s10519-019-09957-8

    • PubMed
    • Google Scholar
63. 1. Plomp JJ
    2. van Kempen GT
    3. Molenaar PC

    (1992) Adaptation of quantal content to decreased postsynaptic sensitivity at single endplates in alpha-bungarotoxin-treated rats

    The Journal of Physiology 458:487–499.

    https://doi.org/10.1113/jphysiol.1992.sp019429

    • PubMed
    • Google Scholar
64. 1. Ramocki MB
    2. Zoghbi HY

    (2008) Failure of neuronal homeostasis results in common neuropsychiatric phenotypes

    Nature 455:912–918.

    https://doi.org/10.1038/nature07457

    • PubMed
    • Google Scholar
65. Software

    1. Roehr J

    (2019) seqan/flexbar, version 3.5.0

    GitHub.

    https://github.com/seqan/flexbar
66. 1. Sackton TB
    2. Hartl DL

    (2016) Genotypic context and epistasis in individuals and populations

    Cell 166:279–287.

    https://doi.org/10.1016/j.cell.2016.06.047

    • PubMed
    • Google Scholar
67. 1. Sanders SJ
    2. He X
    3. Willsey AJ
    4. Ercan-Sencicek AG
    5. Samocha KE
    6. Cicek AE
    7. Murtha MT
    8. Bal VH
    9. Bishop SL
    10. Dong S
    11. Goldberg AP
    12. Jinlu C
    13. Keaney JF
    14. Klei L
    15. Mandell JD
    16. Moreno-De-Luca D
    17. Poultney CS
    18. Robinson EB
    19. Smith L
    20. Solli-Nowlan T
    21. Su MY
    22. Teran NA
    23. Walker MF
    24. Werling DM
    25. Beaudet AL
    26. Cantor RM
    27. Fombonne E
    28. Geschwind DH
    29. Grice DE
    30. Lord C
    31. Lowe JK
    32. Mane SM
    33. Martin DM
    34. Morrow EM
    35. Talkowski ME
    36. Sutcliffe JS
    37. Walsh CA
    38. Yu TW
    39. Ledbetter DH
    40. Martin CL
    41. Cook EH
    42. Buxbaum JD
    43. Daly MJ
    44. Devlin B
    45. Roeder K
    46. State MW
    47. Autism Sequencing Consortium

    (2015) Insights into autism spectrum disorder genomic architecture and biology from 71 risk loci

    Neuron 87:1215–1233.

    https://doi.org/10.1016/j.neuron.2015.09.016

    • PubMed
    • Google Scholar
68. 1. Sardi M
    2. Gasch AP

    (2018) Genetic background effects in quantitative genetics: gene-by-system interactions

    Current Genetics 64:1173–1176.

    https://doi.org/10.1007/s00294-018-0835-7

    • PubMed
    • Google Scholar
69. 1. Shang L
    2. Henderson LB
    3. Cho MT
    4. Petrey DS
    5. Fong CT
    6. Haude KM
    7. Shur N
    8. Lundberg J
    9. Hauser N
    10. Carmichael J
    11. Innis J
    12. Schuette J
    13. Wu YW
    14. Asaikar S
    15. Pearson M
    16. Folk L
    17. Retterer K
    18. Monaghan KG
    19. Chung WK

    (2016) De novo missense variants in PPP2R5D are associated with intellectual disability, macrocephaly, Hypotonia, and autism

    Neurogenetics 17:43–49.

    https://doi.org/10.1007/s10048-015-0466-9

    • PubMed
    • Google Scholar
70. Software

    1. Shtatland T
    2. Langhorst B

    (2018) nebnext-single-cell-rna-seq, version 3

    GitHub.

    https://github.com/nebiolabs/nebnext-single-cell-rna-seq
71. Website

    1. Simons Simplex Collection

    (2020) SFARI gene

    Accessed October 1, 2019.

    https://www.sfari.org/resource/sfari-gene/
72. 1. Spring AM
    2. Brusich DJ
    3. Frank CA

    (2016) C-terminal src kinase gates homeostatic synaptic plasticity and regulates fasciclin II expression at the Drosophila neuromuscular junction

    PLOS Genetics 12:e1005886.

    https://doi.org/10.1371/journal.pgen.1005886

    • PubMed
    • Google Scholar
73. 1. Turrigiano G

    (2011) Too many cooks? intrinsic and synaptic homeostatic mechanisms in cortical circuit refinement

    Annual Review of Neuroscience 34:89–103.

    https://doi.org/10.1146/annurev-neuro-060909-153238

    • PubMed
    • Google Scholar
74. 1. Velmeshev D
    2. Schirmer L
    3. Jung D
    4. Haeussler M
    5. Perez Y
    6. Mayer S
    7. Bhaduri A
    8. Goyal N
    9. Rowitch DH
    10. Kriegstein AR

    (2019) Single-cell genomics identifies cell type-specific molecular changes in autism

    Science 364:685–689.

    https://doi.org/10.1126/science.aav8130

    • PubMed
    • Google Scholar
75. 1. Viquez NM

    (2006) The B' Protein phosphatase 2A regulatory subunit well-rounded regulates synaptic growth and cytoskeletal stability at the Drosophila neuromuscular junction

    Journal of Neuroscience 26:9293–9303.

    https://doi.org/10.1523/JNEUROSCI.1740-06.2006

    • Google Scholar
76. 1. Viquez NM
    2. Füger P
    3. Valakh V
    4. Daniels RW
    5. Rasse TM
    6. DiAntonio A

    (2009) PP2A and GSK-3beta act antagonistically to regulate active zone development

    Journal of Neuroscience 29:11484–11494.

    https://doi.org/10.1523/JNEUROSCI.5584-08.2009

    • PubMed
    • Google Scholar
77. 1. Watt AJ
    2. Desai NS

    (2010) Homeostatic plasticity and STDP: keeping a neuron's Cool in a Fluctuating World

    Frontiers in Synaptic Neuroscience 2:5.

    https://doi.org/10.3389/fnsyn.2010.00005

    • PubMed
    • Google Scholar
78. 1. Yang G
    2. Han Y
    3. Tian X
    4. Tao J
    5. Sun M
    6. Kang J
    7. Yan C

    (2011) Pattern of expression of the CREG gene and CREG protein in the mouse embryo

    Molecular Biology Reports 38:2133–2140.

    https://doi.org/10.1007/s11033-010-0340-7

    • PubMed
    • Google Scholar
79. 1. Yeung KS
    2. Tso WWY
    3. Ip JJK
    4. Mak CCY
    5. Leung GKC
    6. Tsang MHY
    7. Ying D
    8. Pei SLC
    9. Lee SL
    10. Yang W
    11. Chung BH

    (2017) Identification of mutations in the PI3K-AKT-mTOR signalling pathway in patients with macrocephaly and developmental delay and/or autism

    Molecular Autism 8:66.

    https://doi.org/10.1186/s13229-017-0182-4

    • PubMed
    • Google Scholar
80. 1. Yook KJ
    2. Proulx SR
    3. Jorgensen EM

    (2001)

    Rules of Nonallelic noncomplementation at the synapse in Caenorhabditis elegans

    Genetics 158:209–220.

    • PubMed
    • Google Scholar
81. 1. Zhao C
    2. Dreosti E
    3. Lagnado L

    (2011) Homeostatic synaptic plasticity through changes in presynaptic calcium influx

    Journal of Neuroscience 31:7492–7496.

    https://doi.org/10.1523/JNEUROSCI.6636-10.2011

    • PubMed
    • Google Scholar

Author details

1. Özgür Genç

   Department of Biochemistry and Biophysics Kavli Institute for Fundamental Neuroscience University of California, San Francisco, San Francisco, United States

   Contribution

   Data curation, Formal analysis, Validation, Investigation, Visualization, Writing - review and editing

   Competing interests

   No competing interests declared

   "This ORCID iD identifies the author of this article:" 0000-0003-0635-3192

2. Joon-Yong An

   1. Department of Psychiatry UCSF Weill Institute for Neurosciences University of California, San Francisco, San Francisco, United States
   2. School of Biosystem and Biomedical Science, College of Health Science, Korea University, Seoul, Republic of Korea

   Contribution

   Data curation, Formal analysis, Writing - review and editing

   Competing interests

   No competing interests declared

3. Richard D Fetter

   Department of Biochemistry and Biophysics Kavli Institute for Fundamental Neuroscience University of California, San Francisco, San Francisco, United States

   Contribution

   Data curation, Formal analysis, Methodology, Writing - review and editing

   Competing interests

   No competing interests declared

4. Yelena Kulik

   Department of Biochemistry and Biophysics Kavli Institute for Fundamental Neuroscience University of California, San Francisco, San Francisco, United States

   Contribution

   Data curation, Formal analysis

   Competing interests

   No competing interests declared

5. Giulia Zunino

   Department of Biochemistry and Biophysics Kavli Institute for Fundamental Neuroscience University of California, San Francisco, San Francisco, United States

   Contribution

   Data curation, Formal analysis

   Competing interests

   No competing interests declared

6. Stephan J Sanders

   Department of Psychiatry UCSF Weill Institute for Neurosciences University of California, San Francisco, San Francisco, United States

   Contribution

   Methodology, Writing - review and editing

   Competing interests

   No competing interests declared

7. Graeme W Davis

   Department of Biochemistry and Biophysics Kavli Institute for Fundamental Neuroscience University of California, San Francisco, San Francisco, United States

   Contribution

   Conceptualization, Resources, Supervision, Funding acquisition, Visualization, Methodology, Writing - original draft, Project administration, Writing - review and editing

   For correspondence

   graeme.davis@ucsf.edu

   Competing interests

   Reviewing editor, eLife

   "This ORCID iD identifies the author of this article:" 0000-0003-1355-8401

• 2,678

  views

• 412

  downloads

• 13

  citations

Views, downloads and citations are aggregated across all versions of this paper published by eLife.