Introduction

In neurodegenerative disease, prions and protein aggregates can transfer among cells of the nervous system to promote pathology spread ^1,2. Determination of the factors that enhance or deter pathological transfer is therefore a central goal in the effort to clinically address neurodegenerative disease. Study of aggregate transfer in the context of the mammalian brain is a major experimental challenge as events are rare, sporadic and transiently apparent, and tissue is not easily accessible for in vivo observation. We model aggregate transfer by proteostressed ALMR touch receptor neurons in the living C. elegans nervous system ^3–5, an experimental system that enables molecular and genetic manipulation and evaluation in physiological context, directly through the transparent cuticle ⁶.

More specifically, C. elegans adult neurons can extrude large vesicles called exophers (∼5µm, 100X larger than exosomes) that carry potentially deleterious proteins and organelles out of the neuron ^3–5. Disrupting proteostasis via diminished chaperone expression, autophagy, proteasome activity, or over-expressing aggregating proteins like human Alzheimer’s disease associated fragment Aβ_1-42, expanded polyglutamine Q128 protein, or high concentration mCherry fluorophore, increases exopher production from the affected neurons. Neurons that express proteotoxic transgenes maintain higher functionality if those neurons produce exophers as compared to those that do not, suggesting that exopher-genesis can be neuroprotective, at least in young adult neurons. Extruded exopher contents can be transferred to neighboring glial-like hypodermal cell for content degradation in the lysosomal network ⁷. Several mammalian models feature similar biology ^8–12, and thus we speculate that the basic transfer biology represents a conserved process that can be recruited for animal-wide proteostasis balance.

Neuronal exophers are generated only in adult animals, with an unexpected pattern of a peak at young adult days 2-3 and then later in age with more variable onset ³. While using chemical regent 5-fluoro-2’-deoxyuridine (FUdR) to block progeny production for aging studies, we found that blocking reproduction strongly limited the early adult peak of exopher production. Here we report data that support that early adult exopher production is sensitive to the load of eggs in the reproductive tract. We document that uterine expansion, rather than chemical signals emanating from fertilized eggs, correlates strongly with level of exopher production and suggest a model in which mechanical signaling, normally induced across generations from egg to parent via uterine stretch, is a license for proteostressed neurons to release potentially toxic materials in large extracellular vesicles.

Mechanical signaling exerts a profound impact on virtually all cell types, and has been implicated in traumatic brain injury and neurodegenerative disease, yet remains poorly understood ¹³. Our observations direct enhanced experimental attention to studies on how mechanical force can influence extracellular vesicle formation and aggregate transfer in the living brain and in neurodegenerative disease.

Results

The six C. elegans gentle touch receptor neurons (AVM, ALML and ALMR located in the anterior body and PVM, PLML and PLMR located in the posterior body) can be readily visualized in vivo by expression of fluorescent proteins under the control of the touch neuron-specific mec-4 promoter (Fig. 1A). We commonly monitor exophers extruded by the ALMR neuron, which typically produces more exophers than the other touch neurons ^3,14, using a strain in which fluorophore mCherry is highly expressed in touch neurons and is avidly eliminated (strain ZB4065 bzIs166[P_mec-4::mCherry], hereafter referred to as mCherryAg2 for simplicity). ALMR exopher production occurs with a distinctive temporal profile, such that at the L4 larval stage ALMR rarely, if ever, produces an exopher, but in early adult life, the frequency of exopher events increases, typically reaching a peak of 5-20% of ALMR neurons scored at adult day 2 (Ad2); exopher detection then falls to a low baseline level after Ad3 ^3,14, a pattern that parallels adult reproduction (Fig. 1B). Late in life, exophers can reappear with variable frequency, but here we focus on young adult exopher generation.

Figure 1.

Sterility-inducing drug FUdR suppresses ALMR exophergenesis

In experiments originally designed to study exopher events in aging animals, we sought to generate age-synchronized populations by blocking progeny production from early adult stages using DNA synthesis inhibitor 5-fluoro-2’-deoxyuridine (FUdR). Unexpectedly, we found that 51μM FUdR strongly suppressed exopher events as quantitated at Ad2 (Fig. 1C). FUdR is commonly used in C. elegans to inhibit the proliferation of germline stem cells and developing embryos, but has also been noted to impair RNA metabolism ¹⁵ and improve adult proteostasis ¹⁶. To probe which FUdR outcome might confer exopher suppression, we first addressed whether disruption of progeny production by alternative genetic means might suppress exophergenesis, which would implicate a viable germline as a factor in exopher modulation.

Germline elimination and germline tumors suppress ALMR exopher production

In the assembly line-like hermaphrodite C. elegans gonad, germline stem cells close to the signaling distal tip cell (DTC) proliferate by mitosis (cartoon of C. elegans gonad and germ cell development in Fig. 1D). We disrupted germ cell production using the glp-4(bn2ts) valyl-tRNA synthetase 1 mutant. At the restrictive temperature (25°C), glp-4(bn2ts) causes germ cell arrest during the initial mitotic germ cell divisions, effectively eliminating the germline ¹⁷. We constructed an mCherryAg2; glp-4(bn2ts) strain and quantitated ALMR exopher production in animals grown at the restrictive temperature, scoring exophers during early adulthood (Fig. 1E). ALMR neurons in the germline-less glp-4 mutant produced significantly fewer exophers on Ad1-4 as compared to age- and temperature-matched controls.

Both oogenesis and spermatogenesis are critical for early adult exopher production

Given that lack of functional germ cells impaired neuronal exopher production, we sought to test whether oocytes or sperm might be specifically required for exophergenesis, taking advantage of C. elegans genetic reagents available for the manipulation of gamete development.

Oogenesis is required for the peak of early adult exophergenesis

In the C. elegans hermaphrodite, the differentiation of germline stem cells begins during the L4 larval stage with spermatogenesis, after which sperm production is shut down and oogenesis begins ¹⁸. Genetic mutants that produce only sperm or only oocytes have been well characterized. To test a mutant that produces sperm but no oocytes, we employed fem-3(q20ts), a temperature sensitive gain-of-function mutant that causes germ cells to exclusively differentiate into sperm ^19,20. We constructed an mCherryAg2; fem-3(q20ts) line, reared animals at non-permissive temperature 25°C, and scored ALMR exophers in mutant and control animals during early adulthood. In the sperm-only, oocyte-deficient fem-3(q20ts) background, exophergenesis is mostly diminished over the first five days of adulthood (Fig. 1F). We infer that oocytes must be present for early adult ALMR exopher production and conclude that sperm alone are not sufficient to drive exopher elevation.

Spermatogenesis is required for the peak of early adult exophergenesis

The presence of functional sperm can stimulate ovulation to trigger oocyte maturation. To assess neuronal exopher production in the reciprocal reproductive configuration in which animals have oocytes but no sperm, we disrupted spermatogenesis using two approaches. First, we used a temperature sensitive fem-1 mutation to produce animals with oocytes but no sperm ²¹. We crossed fem-1(fc17ts) into the mCherryAg2 strain and examined exopher production from ALMR neurons in hermaphrodites at the restrictive temperature of 25°C. We found that ALMR exopher production in fem-1(fc17) mutants was suppressed at 25°C (Fig. 1G).

Second, we used the auxin-inducible degron system (AID ^22,23) to degrade SPE-44, an essential protein required for spermatid differentiation ^24,25. In the AID system, addition of auxin to the culture induces rapid degradation of proteins genetically tagged with a specific auxin-dependent degron sequence. AID targeting of SPE-44-degron is highly effective in disrupting sperm maturation ²⁴. We treated strain ZB4749 (fxIs1[P_pie-1::TIR1::mRuby] zdIs5[P_mec-4::GFP]; bzIs166[P_mec-4::mCherry]; spe-44(fx110[spe-44::degron]) with auxin during larval developmental stages to block sperm maturation and then transferred Ad1 animals to NGM plates without auxin, a condition that disrupted sperm maturation but allowed oocyte generation. We find that blocking sperm maturation by targeting SPE-44 for degradation abolished ALMR exopher production (Fig. 1H) even though oogenesis is not significantly impacted ²⁴.

We infer that functional sperm must be present for early adult ALMR exopher production and conclude that oogenesis alone is not sufficient to drive exopher elevation in early adult life.

Fertilization and early embryonic divisions are required for the early adult exophergenesis peak

The requirement of sperm and oocytes for neuronal exopher production raises the obvious question as to whether fertilized eggs/embryos are required for ALMR exophergenesis (Fig. 1I). C. elegans genes impacting fertilization and embryonic development have been studied in exquisite detail ^26–29. We tested embryonic development genes known for roles at particular stages for impact on neuronal exopher levels using RNAi knockdown approaches.

In C. elegans fertilization, as the mature oocyte encounters the sperm-filled spermatheca, a single sperm enters the mature ovulated oocyte (fertilization time 0), triggering the rapid events of egg activation, which include polyspermy block, eggshell formation, completion of meiotic divisions and extrusion of polar bodies. The multi-step eggshell formation (Fig. 2A) is completed within 5 minutes of sperm entry, by which time the zygote has been passed into the uterus, where the maternal chromosomes execute both meiotic divisions (meiosis accomplished by ∼20 minutes after fertilization). After establishment of egg polarity cues, the first mitotic cell division occurs ∼40 minutes after fertilization ^27,29. Eggs are held in the mother’s uterus until they are laid at the ∼30 cell stage (gastrulation). In wild type, ∼8 fertilized eggs occupy the uterus when egg laying begins at ∼6 hours of adult life ³⁰.

Figure 2:

Our data revealed an unexpected theme: we found that disruption of tested early-acting genes essential for eggshell assembly (chitin binding domain protein cbd-1 ³¹ (Fig. 2B), chitin precursor synthesis gene gna-2 ³² (Fig. 2C, D), permeability barrier-required CDP-ascarylose synthesis gene perm-1 ³³ (Fig. 2E)), and/or needed for the progression through the 1-cell or 2-cell stages of embryonic development ^34–37 (profilin pfn-1 and coronin-related pod-1 (Fig. 2F), cause potent suppression of ALMR exophergenesis (Fig. 2G). In contrast, RNAi knockdowns of genes playing important roles for later stage embryonic development (4 cell to 8 cell stages mex-3, mom-2, gastrulation genes end-1/-3 or gad-1) (Fig. 2F) confer negligible impact on ALMR exophergenesis (Fig. 2H). Our data support a fertilization requirement for neuronal exopher stimulation and suggest that the exophergenesis signal/condition that promotes the early adult wave of exophergenesis is associated with the very earlest stages of embryonic development. Of note, disruption of eggshell biosynthesis has the immediate downstream consequence of disruption of early polarity establishment and first divisions, such that genetic separation of eggshell production from first division proficiency is not possible, for example, pod-1 RNAi has been reported to be associated with eggshell deficits ³⁵, although no eggshell deficits are oberseved for pfn-1 RNAi ³⁸. We thus conclude that either eggshell integrity or biochemical events associated with the first embryonic cell divisions are required for neuronal exophergenesis.

Restoring fertilized eggs later in life can extend ALMR exophergenesis

Having found that fertilized eggs are necessary for the early adult elevation in exopher production, we asked whether the presence of eggs is sufficient for stimulation of exopher production by introducing fertilized eggs later in adult life when they are not normally present. In C. elegans, sperm are made in the L4 stage and are stored in the spermatheca to fertilize oocytes that mature in the adult. Sperm are limiting for hermaphrodite self-fertilization, with more oocytes made than can be fertilized by self-derived sperm.

Unmated hermaphrodites thus cease egg laying around adult day 4 because they run out of self-supplied sperm. However, if males mate with hermaphrodites, increased progeny numbers can be produced due to increased sperm availability. More germane to our experiment, if males are mated to the reproductively senescing hermaphrodite to replenish sperm, hermaphrodites can produce fertilized eggs for a few additional days (Fig. 3A).

Figure 3:

To determine if the presence of eggs might be sufficient to drive exopher production after Ad3, we mated males to Ad3 reproductively senescing hermaphrodites. We found that restored egg production is associated with increased and extended exopher production, provided that hermaphrodites were mated to fertilization-proficient males (Fig. 3B). We conclude that adult ALMR exophergenesis is driven by the presence of fertilized eggs and that the older age decrease in exopher production (∼Ad4) under standard culture conditions is more likely attributed to the lack of fertilized egg accumulation at this lifestage, rather than to the existence of a chronological limit on the biochemical capacity to elevate exopher levels at older ages.

Genetically induced egg retention elevates ALMR exopher production

Fertilized eggs might release a signal or create a condition that stimulates exophergenesis. If such an influence were limiting, increasing the egg concentration in the body might increase exopher numbers. To address whether elevating the young adult egg load can increase exophers, we took multiple distinctly-acting genetic approaches to limit egg laying and promote egg retention in the body (Fig. 3C). We tested four genetic conditions that lower or block egg-laying: a null mutation of prolyl hydroxylase egl-9, for which disruption induces a mild egg laying defect ⁴³; a null allele of proprotein convertase subtilisin/kexin type 2 egl-3 ⁴⁴ that perturbs neuropeptide processing and confers a severe egg-laying defective phenotype ⁴³; a reduction-of-function mutation in SOX transcription factor SEM-2 (sem-2(n1343)) that eliminates production of the sex myoblasts needed to generate egg laying muscle; and RNAi directed against the lin-39 homeobox transcription factor HOXA5 ortholog required for vulval cell fate specification ^45,46.

We found that the associated massive egg retention correlated with a dramatic elevation of exopher numbers for each genetic impediment to egg laying (egl-9 (Fig. 3D); egl-3 (Fig. 3E); sem-2 (Fig. 3F), lin-39(RNAi) (Fig. 3G)). For example, under the treatment of lin-39(RNAi), ∼60% of the wild type strain expressing mCherry in ALMR by a multi-copy transgene produced ALMR exophers on Ad1, compared to only ∼7% of the same strain treated with empty vector (EV) control (Fig. 3G). We conclude that regardless of the genetic strategy employed to trap eggs in the body, egg retention can lead to high ALMR exopher production.

In complementary studies, we examined the impact of egg retention on ALMR exophergenesis when we expressed fluorescent protein mKate from a single copy transgene (i.e., in the absence of an over-expressed reporter). We found that the empty vector control treatment without induction of egg retention is associated with no ALMR exophergenesis in the single copy mKate transgenic strain (0% in all three trials). However, when we treated with lin-39(RNAi) to induce egg retention, we measured ∼60% exophergenesis. Thus regardless of the expression levels of exogenous proteins in the ALMR neuron, the egg retention condition can induce high exopher production.

Hyperactive egg-laying and consequent low egg retention is associated with low exopher production

Neurotransmitters octopamine (OA) and serotonin (5-HT) have been well documented to play opposing roles in C. elegans egg-laying behavior ⁴⁷. Feeding C. elegans octopamine strongly suppresses egg-laying to induce egg retention, while supplementing with 5-HT causes hyperactive egg-laying ⁴⁸. Consistent with outcomes in animals physically blocked for egg-laying, treatment with egg retention-promoting OA enhanced ALMR exophergenesis (Fig. 3I).

To test the outcome of 5-HT-induced enhanced egg laying, we measured the impact of 5-HT consequent to 6 hour food withdrawal, a condition that we previously found markedly enhances ALMR exophergenesis ⁴, and therefore is expected to increase the dynamic range for scoring exopher suppression. We find that 5-HT treatment, which lowers egg retention, strongly suppresses fasting-associated ALMR exophergenesis (Fig. 3I). Although perturbing neuronal signaling holds complex consequences for whole-animal physiology, our findings are consistent with a model in which high egg load increases neuronal exophergenesis, and low egg retention decreases exophergenesis.

The egg laying circuit Is controlled in part by Goα inhibition—null allele goa-1(n1134) removes this inhibition such that eggs are often laid at very early developmental stages (2-4 cell stage) rather than being retained in the uterus until gastrulation (∼30 cell stage) ⁴⁹. We introduced goa-1(n1134) into the mCherryAg2 background and scored ALMR exopher events at Ad2. We confirmed that goa-1(n1134) retains few eggs in the uterus and is associated with a significantly lower number of ALMR exopher events as compared to the age-matched wild type control (Fig. 3J).

In sum, manipulation of uterine egg occupancy is strongly correlated with the extent of ALMR neuronal exophergenesis—high egg retention promotes high exophergenesis and low egg retention is associated with low exophergenesis.

ALMR neuron proximity to the egg zone correlates with exophergenesis frequency

The ALML and ALMR anterior touch neurons share developmental, morphological and functional similarities ⁵⁰, yet paradoxically, ALMR consistently produces exophers at higher levels than ALML ³ (also Fig. 4G). The ALM neurons are embedded within the C. elegans hypodermis, but the ALMR soma is situated in vicinity of the gonad and its resident eggs, whereas the ALML soma, on the opposite side of the animal, is positioned closer to the intestine (Fig. 4A, 4B) ⁵¹.

Figure 4.

To ask if proximity to the gonad is correlated with exopher production, we randomly selected 128 Ad2 hermaphrodites that expressed mCherry in the touch receptor neurons, imaged in brightfield to visualize the egg zone of each animal, and imaged again in the red channel to visualize the touch neuron, recording the relative positions of ALMR and the egg zone (Fig. 4C). We also assessed whether the ALMR neuron had produced an exopher or not.

We found that 37/128 ALMRs examined had produced an exopher (Exopher+), and that 36/37 (95%) of the Exopher+ ALMR neuronal somas that had produced exophers were positioned within the visualized egg zone (Fig. 4D). For the Exopher-ALMRs that had not produced exophers, 63/91 (70%) had neuronal somas located in the egg zone. Thus, although ALMR soma positioning in the egg zone does not guarantee exophergenesis in the mCherryAg2 strain, the neurons that did make exophers were nearly always in the egg zone (p < 0.01 in Chi-Square test, Fig. 4D).

To further test for association of egg zone proximity to ALMR and exopher production, we genetically shifted ALM position. During development, the ALM neurons migrate posteriorly to near the mid-body ⁵² and most commonly, ALM somas are situated posterior to AVM. ALM soma positions, however, can be influenced by migration and specification cues. In particular, transgenic introduction of a mec-3 promoter fragment bearing an internal deletion (fusion of the −1 to −563 sequences to the −1898 to −2372 mec-3 promoter fragment, plasmid pJC4) can induce anterior ALM migration during development, sometimes resulting in final ALM positions anterior to AVM ⁵³ (Fig. 4E). We took advantage of the partially deleted mec-3 promotor sequences in pJC4 to manipulate ALM position. In these studies we introduced pJC4 with the co-transformation marker pRF4 (rol-6(su1006)) that disrupts the cuticle to induce rolling of transgenic animals into the P_mec-4::mCherry background. Rol hermaphrodites have a strikingly high baseline of ALMR exophergenesis (∼40% exophers in rollers vs. ∼20% in the wild type). Strikingly, we found that when ALMs are situated anterior to AVM, ALMR exophergenesis drops to ∼5% (4/73) vs. 71% for posterior position (55/78) (Fig. 4F). Although we cannot exclude that physiological changes in differently-positioned touch neurons underlie reduced exophergenesis, data are consistent with a model in which proximity to the egg zone correlates with exophergenesis.

Another way to increase egg proximity to ALMR is to disrupt egg laying capacity, which confers egg retention and uterine expansion. We hypothesized that in the sem-2(rf) mutant, which is associated with considerable internal egg accumulation, additional touch neurons should experience increased proximity to eggs in the blocked uterus. Indeed, we find that in the sem-2(rf) background, every touch neuron that is positioned in the general region of the expanded uterus (ALML, AVM, PVM) increases exopher production, but the posterior PLM neurons, which cannot be approached by the gonad, do not produce exophers (Fig. 4G). Thus, touch neurons can be stimulated to produce exophers if the egg domain is artificially brought closer to them. Data are consistent with a model in which ALMR normally makes most exophers because of its closest natural positioning near the egg-filled uterus. Possibilities are that a diffusible signal may eminate from the filled uterus or that mechanical pressure associated with a filled uterus might signal enhanced exophergenesis.

Uterine expansion associated with high egg load correlates with high exophergenesis

How might the presence of eggs signal to the maternal neurons to induce exophers? We consider two main possibilities: 1) eggs filling the uterus might exert physical pressure that activates essential stretch-signaling for young adult neuronal exopher release. This mechanical stretch signal might act directly (for example introducing chronic and/or dynamic pressure on the touch neurons), or indirectly (possibly inducing the stretched uterus/somatic gonad to release chemical signals that promote neuronal exopher formation); 2) early fertilized eggs might release a short-range chemical signal that contributes to young adult proteostasis reorganization ^54,55 to promote exophergenesis.

To begin to dissect the role of egg pressure in promoting exophergenesis, we analysed the physical relationships of neurons, eggs and uterine shape. Eggs can readily be observed to distort tissue structure in young adult C. elegans. For example, in a strain that expresses GFP to label the hypodermis and expresses mCherry to label the touch neurons, the distortion of the hypodermis by eggs can be easily visualized as dark non-fluorescent eggs project through the observation plane of the hypodermis (Fig. 5A). Thus, the uterus can approach and pressure surrounding tissue, including touch neurons.

Figure 5.

We quantitated the absolute uterine length as an indicator for stretch in relation to exopher production levels under representative condition of high and low exophergenesis. We found that egg retention mutants that exhibit high exophergenesis, egl-3(τι) (Fig. 5B), egl-9(τι) (Fig. 5C), and sem-2(rf) (Fig. 5D), had significantly longer egg zones (i.e., uterus length) as compared to wild type. In contrast, cbd-1(RNAi) (Fig. 5E), sperm-less induction with SPE-44 AID (Fig. 5F), and hyperactive egg-laying mutant goa-1(τι) (Fig. 5G), which we find to be strong exophergenesis suppressors, are all associated with short uterine egg zones. Knocking down either the 4-cell stage gene mex-3 or the gastrulation gene gad-1, which are associated with neither egg retention nor exopher elevation, does not have an extended egg zone/uterine length (Fig. 5H), so not all developmental compromises are associated with uterine extension.

Even when eggshell production and early embryonic divisions are disrupted, forced uterine expansion can elevate exopher levels

Our initial studies suggest extended uterine length is correlated with high exopher levels, but high egg retention is also a feature of an extended uterus. To begin test a requirement for fertilized eggs per se in the exopher influence, we asked whether egg viability is essential for promoting early adult exophergenesis. We manipulated egg integrity/uterine contents in egg retention mutants by egg/embryo perturbation, testing for impact on exophergenesis.

Under conditions of cbd-1(RNAi), eggshell development and embryonic development are blocked; eggs that form are fragile and can be malformed consequent to passing through the spermatheca into the uterus; embryonic development does not proceed and egg remnants tend to be sticky ⁵⁶. As shown in Fig. 2B (and again in Fig. 6A), treating WT reproductive animals (that have functional egg laying capacity) with cbd-1(RNAi) to kill embryos exerts a potent block on ALMR exophergenesis. We proceeded to test the consequence of cbd-1(RNAi) in mutants that cannot extrude eggs or their remnants, and therefore would retain defective eggshell/dead embryos in the uterus.

Figure 6.

We subjected the sem-2(rf) egg-laying defective mutant to cbd-1(RNAi) so that the uterus would fill with eggshell/dead embryo remains. Strikingly, we found that cbd-1(RNAi)/dead embryo retention in the sem-2(rf) background is still associated with significantly elevated levels of ALMR exophergenesis (∼45%), while in egg-laying proficient wild type, barely any ALMR exophergenesis is observed under cbd-1(RNAi) conditions (<1%) (69/150 for sem-2 egg-laying blocked cbd-1(RNAi) vs. 1/150 for egg-laying proficient cbd-1(RNAi), Fig. 6A).

Importantly, under conditions of defective eggshell/dead embryo retention associated with cbd-1(RNAi); sem-2(rf), the uterine egg zone is expanded (Fig. 6B), extending the correlation of exopher production with uterine length. We conclude that intact eggshells and earliest embryonic divisions are not required for the boost in exopher production observed when uterine contents are forced to accumulate—-uterine retention of dead eggs and egg remnants is sufficient for exopher elevation if the egg laying apparatus is defective. Expansion of the uterine compartment, rather than eggshell/embryo integrity, tracks with exopher elevation.

Forced uterine expansion via oocyte accumulation can elevate exopher levels

Although uterine retention of malformed inviable embryos is sufficient to elevate neuronal exophers when egg laying is blocked, the defective cbd-1 embryos or debris might still release egg-associated chemical signals. To test for a requirement of any fertilization-dependent egg signals in the egg-laying compromised mutants, we asked whether uterine filling with only oocytes can suffice to promote neuronal exopher elevation. Unfertilized oocytes cannot initiate embryonic development or egg-shell biosynthesis; nor can oocytes elevate ALMR exophergenesis in hermaphrodites that are proficient at egg-laying (Fig. 1G&H).

We tested two distinct uterine retention conditions—-sem-2(rf) and lin-39(RNAi)—-in which we used the auxin inducible degron system to disrupt sperm maturation such that only oocytes filled the gonad. Note that in the absence of sperm, oocytes do not mature but are “ovulated” at a much reduced rate; oogenesis continues such that ∼25 oocytes are typically found stacked in the gonad (Fig. s6) ⁵⁷. For both genetic retention strategies, we found that build-up of retained oocytes in egg-laying blocked animals was sufficient to elevate exophers (Fig. 6C&E) and expand the uterus (Fig. 6D& F; Fig. s6).

Moreover, the oocyte retention was similarly efficacious in elevating exopher production to egg retention, increasing ALMR exophergenesis to approximately 80% in the sem-2(rf) mutant (Fig. 6C). We conclude that fertilization, egg shells and egg remnants are not essential for the early adult exopher peak. Expansion of the uterus with unfertilized oocytes can suffice to elevate neuronal exopher formation.

Lack of a functional egg-laying apparatus does not induce exopher elevation when the uterus is not filled

The above-described experiments left open the possibility that the lack of a functional egg-laying apparatus itself might be causative in the elevation of exopher production. To address this possibility, we compared disruption of sperm (permissive for oocyte accumulation) to disruption of oogenesis (effectively empties the uterus) when egg-laying capacity was compromised by lin-39(RNAi) (Fig. 6G). lin-39(RNAi) + oocyte retention promotes exopher formation, but eliminating oocytes (fem-3(gf)) eliminates exopher elevation even when egg-laying is blocked by lin-39(RNAi). That is to say, although oocyte accumulation with uterine expansion suffices to elevate exophers, removing the oocytes and uterus occupancy eliminates the exopher boost. We observe the same outcome of suppressed exopher formation when cbd-1(RNAi)-induced dead embryo retention in the sem-2(rf) egg-laying defective mutant (which is exopher-inducing) is prevented from oocyte production by fem-3(gf) (Fig. 6H). Thus, disruption of egg laying on its own is not the driving factor in high exophergenesis; rather, uterine filling is required.

We revisited the relationship of uterine length and exopher level by adding data from the studies with oocyte retention to reinforce the conclusion that ALMR exophergenesis is strongly correlated with the level of uterus stretching caused by the accumulation of uterine contents (Fig. 6I).

Sustained physical distortion of the gonad by fluid injection can rapidly elevate exopher production

To independently test for a role for physical stretch/filling of the uterus in exopher induction, we distorted the gonad compartment by injecting dye-containing M9 buffer into a very young adult. The animal subjects we tested were vulva-less (lin-39(RNAi)) and also subjected to spe-44 AID to block sperm production. These vulva-less + sperm-less hermaphrodites normally exhibit high ALMR exophergenesis at late Ad1 and Ad2 (Fig. 6C&D) due to oocyte accumulation. To avoid oocyte influence, we conducted our physical expansion studies just as animals reach Ad1, a time when ALMR exophergenesis is typically not observed (Fig. 7A).

Figure 7.

We picked the L4 stage vulva-less (lin-39(RNAi)) + sperm-less (SPE-44 AID) hermaphrodites for age synchronization (20°C, grown for 12 hours after L4 selection) and then injected dye containing M9 buffer into the uteri of these very young adult day 1 hermaphrodites, scoring for exopher production ∼10 minutes after the injection (Fig. 7B).

Control animals (mock injected animals that were stabbed without fluid delivery) exhibited no ALMR exophers. By contrast, we found that when we held injection pressure continuous for 2 or more minutes, ∼20% of the ALMs scored exhibited an exopher event shortly after injection (Fig. 7C&D). Injection experiments using animals with functional vulvae (in which injected material is rapidly extruded through the vulval opening) failed to induce ALMR exophers (Fig. 7E), supporting that ALMR exophergenesis caused by 2-minute injection of dye-containing M9 is due to the physical distortion of gonad rather than the chemical impact of the dye-containing M9.

In sum, exopher induction by uterine accumulation of eggs, malformed eggs, dead embryos, oocytes, or fluid-induced expansion support a model in which early adult ALMR exophergenesis is elevated by physical distortion of the uterus that occurs with reproduction (Fig. 7F).

Discussion

Exopher production by proteo-stressed C. elegans touch neurons occurs with a striking temporal pattern that features an early adult peak of exophergenesis coincident with the period of maximal egg production. Here we report that the early exophergenesis peak is dependent on uterine occupation, which normally is conferred by fertilized egg accumulation prior to the deposition of eggs. Uterine expansion that is associated with the filling of a blocked uterus with unfertilized oocytes or by fluid alone can also induce high neuronal exopher production, supporting a model in which the physical distention associated with uterine occupancy, rather than chemical signals derived from fertilized eggs per se, is a required component of the signaling relay between reproduction and young adult neuronal debris elimination. Trans-tissue cross talk to the maternal nervous system thus appears accomplished via mechanical force transduction. Our findings hold implications for mechano-biology in neuronal proteostasis management.

The mechanical landscape of reproduction that influences neuronal exophergenesis

C. elegans reproduction features physical expansion and contraction of multiple tissue/cell types--the gonad houses the expanding germline, the spermatheca expands and retracts vigorously as each mature oocyte enters and exits. The uterus also stretches to house eggs and can contract locally as eggs transit or are expelled via action of the vulval and uterine muscles. The filled reproductive apparatus can thus clearly exert both constitutive and sporadic pressure on surrounding tissues as it enacts its essential functions.

What is the source of the force that promotes exopher production? Elegant work on the egg laying circuit (comprising the somatic gonad, the HSN and VC neurons, and the vulval and uterine muscle) has provided evidence for mechanical signaling within the egg laying circuit that regulates initiation, promotion and termination of egg laying ^30,58–60. Working backward, exopher-promoting force seems unlikely to derive from the vulva or vulval muscle contractions, since when these cells are genetically disrupted in lin-39(RNAi) or in sem-2 mutants, high levels of exophers still are generated. Changes in spermatheca volume, which expands and contracts dramatically as mature oocytes enter via valve opening/closure ⁶¹, might be sensed, but spermatheca contractions are reported to be normal in hyperactive egg laying goa-1 ⁶², which is an exopher-suppressing background. Under no-sperm conditions, oocyte transit rates are lower than for fertilized eggs, and sperm-derived signals influence spermathecal valve opening ^57,63,64, but if the egg laying apparatus is genetically compromised and oocytes accumulate in the absence of sperm, exopher levels are high, suggesting deficits in spermatheca operations or sperm signals per se do not drive exophergenesis. Given that under normal reproductive conditions of egg-laying proficiency, correctly shelled eggs are required for the early peak in exopher production, a plausible hypothesis was that fertilized eggs might produce an essential diffusible factor that stimulates neuronal exopher-genesis. However, exophers can be produced abundantly in the absence of fertilized eggs when the vulva is unable to open and release uterine contents, resulting in uterine distention due to debris filling. Thus, the simplest model we envision for the reproductive cues that influence maternal neuronal exophergenesis is that a filled uterus (under normal conditions the consequence of hard-shelled eggs that occupy it) is sensed and required for the early adult peak in exopher production.

How might force be sensed and transduced?

Mechanotransduction is the sensing of a mechanical signal, such as pressure or stretch, and conversion into a cellular response. Members of several ion channel families have been implicated in sensing of touch, hearing, shear stress, and pressure, including Piezo, TRP, and DEG/ENaC families ⁶⁵. These are rational candidates for mechanosensors in neurons, the uterus, or other cells that might act in a relay between the uterus and neuron.

At the same time, classic mechanosensory channels have extremely rapid gating and might not be the best suited candidates for acting in the sustained and locally dynamic forces anticipated for the reproductive uterine environment. Adhesion G-Protein Coupled Receptors, which have extracellular adhesion motifs and 7 transmembrane domains characteristic of the GPCR class (lat-1, lat-2, cdh-6 in C. elegans ^66–68), or components of the YAP/TAZ transcriptional program ⁶⁹ may integrate responses to forces transmitted via the cytoskeleton and could be considered as potential players in the required signaling.

Determination of the identity of mechanotransducers and assignment of site of action to the neuron, the uterus, or an intermediate relay cell type remains for future studies. Modeling will also need to incorporate the fact that fluid injections, which required 2 minute long sustained application of the filling stimulus to induce exophers, could provoke exopher production on a rapid timeframe, typically recorded only 10 minutes after injection period. Thus the proteostressed touch neurons appear poised to eliminate contents upon mechanical stimulation.

Mechanical signaling in reproduction across species

Uterine stretch may be a more prevalent mechanism for inducing maternal nervous system response than currently appreciated. Distention of the female fly reproductive tract by egg passage through the tract (normal biology) or by artificial means (experimental fluid injection) can induce behavioral attraction of the mother to acetic acid, thought to signal a favorable food environment for offspring ⁷⁰. In this case, DEG/ENaC channel family member PPK1 expressed in a subset of mechanosensitve neurons that tile the reproductive tract and respond to its contraction/distention is required. The pathway to the behavioral change remains to be determined. Uterine stretch in mammals has also been reported influence maternal behavior ⁷¹.

Why link exophers to reproduction?

Turek et al. report that exophers produced by C. elegans muscle cells follow a similar time course of highest production at adult day 2, and demonstrated a dependence of the temporal muscle exophergenesis pattern on eggs, and commonly higher close to the uterus ⁷² (muscle exophers may be released to supply nutrients to developing progeny). Together with our observations, data raise the possibility that the onset of reproduction and the initial filling of the uterus triggers, or generates a ‘license” for EV/exopher production across tissues. In the case of stressed touch receptor neurons that have been our focus, evidence suggests that deleterious protein aggregates and/or excess proteins and organelles are handed off to neighboring glial-like hypodermal cells for degradation ⁷. Clearing the nervous system (and other organs) for optimal function might confer a selective advantage for successful maternal reproduction.

In this regard it is fascinating that the peak exopher production period is coincident with a proteostasis reconfiguration that has been well documented to accompany reproduction onset in young adult C. elegans. In brief, during larval development C. elegans exhibits high activity of HSF-1 and consequently HSF-1-dependent chaperone expression, but HSF-1 activity is turned down in adult life ^54,55,73,74. At the same time in early adult life, proteasome activity is relatively enhanced (at least as measured in the hypodermis) ⁷⁵. These measures may reflect a general proteostasis reorganization (chaperone activity, proteosome activity) that occurs in early adult life in response to reproduction ^54,73. Our observations on neuronal exophers suggest that exopher-mediated content elimination may constitute another co-regulated branch of this proteostasis reconfiguration. Importantly, the HSF-1 turn-down in young adult life is blocked by cbd-1(RNAi) (and additional early eggshell/development gene RNAi) ⁷⁴. Thus the presence of eggs can signal across tissues to turn-down hsf-1 proteostasis-related activities in the mother. We speculate that this young adult reconfiguration of proteostasis might reflect a mechanism to optimize successful reproduction, possibly both fine tuning nervous system function and shifting resources balance to favor progeny as suggested by the disposable soma theory of aging proposed by Kirkwood ⁷⁶.

Of note, we do not observe exophers in larval stages ³. We speculate that young adult physiology might be temporally tweaked such that some tissues have optimized capacity to manage/degrade large aggregates and organelles at an early adult developmental “clean up” time, possibly analogous to how a town service for bulky oversized garbage pick-up might be limited to particular times during the year. As exopher production appears generally beneficial for neuronal function and survival ^3,77, the early life extrusion phase appears a positive feature of reproductive life. More broadly, proteome “clean up” phases may be programmed as key steps at specific transitions during development and homeostasis, for example, as occurs in the temporal lysosome activation that clears aggregate debris in C. elegans maturing oocytes ⁷⁸ or in the maturation of mouse adult neuronal stem cells via vimentin-dependent proteaseome activity during quiescence exit ⁷⁹.

Across species, production of exopher-like vesicles may also be enhanced by mechanical signals anchored outside of reproduction. For example, mice cardiomyocytes that are constantly under mechanical stress due to contraction activities produce exopher-like vesicles ⁸. Mouse kidney proximal tubular epithelia cells (PTEC) under constant mechanical stress due to both fluid shear stresses and absorption-associated osmotic pressure, also release exopher-like vesicles ⁸⁰.

Large vesicle extrusion, mechanobiology and neurodegenerative disease

The impact of mechanical force on in vivo production of extracellular vesicles has not been a major focus of the EV field, although a range of studies have considered force consequences (such as fluid shear responses, stretch) in cultured cells. Overall, however, EV biogenesis and uptake appear to be markedly influenced by biomechanical force type, magnitude, and duration ⁸¹. At the same time, the neurodegeneration field has generated myriad studies linking Alzheimer’s disease susceptibility and AD pathology signatures such as extracellular accumulation of amyloid-β protein and/or intracellular accumulation of tau as outcomes of mechanical stress-based stimuli such as traumatic brain injury, arterial hypertension, and normal pressure hydrocephalus ^82,83. Mechanical stress may trigger or promote protein misfolding, aggregation, and extrusion. Examples of recent implication of mechanical stimuli in AD-related outcomes include that stretch in the brain vascular system can increase APP and B-secretase expression to increase Aβ production ⁸⁴ and that microglial mechanosensing via the Piezo1 mechanotransducing channel limits progression of Aβ pathology in mouse models ⁸⁵. Our study reveals a capacity of mechanical force to influence neuronal release of large vesicles containing neurotoxic species, inviting more serious consideration of the roles of mechanobiology in maintaining proteostasis and influencing aggregate transfer within the context of a living nervous system.

Methonds and materials

Strains and maintenance

All strains used in this study carry the transgene bzIs166[P_mec-4::mCherry] to mark the six touch receptor neurons: ALMR, ALML, AVM, PVM, PLMR and PLML. The genotype of C. elegans strains used in this study are listed in Table 1. We maintained all C. elegans strains on nematode growth media (NGM) seeded with OP50-1 Escherichia coli in a 20°C or 15°C incubator. We kept all animals on food for at least 10 generations before using them in test.

Table 1:

Age synchronization

For the majority of experiments, we used a bleaching protocol or an egg-laying protocol for age synchronization, otherwise, we picked L4 animals for synchronization.

Temperature sensitive mutants

We maintained the age-synchronized temperature sensitive mutations in a 15°C incubator. For either fem-1 or fem-3 mutants, we directly placed the isolated eggs into a 25°C incubator in each experimental test. Since the egg-hatching of glp-4 mutant is out of sync at 25°C, we placed the isolated eggs in a 15°C incubator for 24 hours before transferring them into the 25°C incubator for experimental tests.

Auxin inducible degradation

We dissolved auxin (indole-3-acetic acid, 98+%, A10556, Alfa Aesar) in 95% ethanol to make a 400 mM stock, then we prepared a 40 mM auxin solution by diluting the 400mM stock solution in M9 medium (which contains 1mM MgSO4) and applied 200 ul of the 40 mM solution on to the NMG-agar plate (which is 60 mm in diameter and contains ∼8-9 g medium). We left the plate on an open bench at room temperature for one or two days to dry out the auxin solution, then seeded the plate with 200 ul OP50-1 E. coli and waited for another two or three days before storing the plates in a 4°C environmental room. We only used the plates which had been stored in the 4°C environmental room for at least a week, so the concentration of auxin can be equilibrated into 1 mM. We prepared the ethanol control plates under the same procedure, and the final concentration of ethanol in the control plates should be ∼0.25%.

To proceed auxin inducible degradation of SPE-44, we placed the isolated eggs on auxin treated plates or the ethanol control plates. After ∼72 hours of culture in a 20°C incubator, the animal reached the age of adult day 1 and the auxin treated worms became sterile. Then we transferred the worms into a regular NGM-agar plate without auxin or ethanol.

Microscopy and image processing

We took the DIC or fluorescent pictures with a Zeiss compound microscope or spinning disc confocal microscope driven by MetaMorph software, then processed the pictures with Fiji ImageJ software.

Exopher scoring

Exopher numbers vary in experiments between 5-30%, mostly 10-15% range, and typically reach peak at adult day 2 ³. Exphers can remain intact for approximately 2 hours, but the vesicle form of exophers is mostly identifiable in the following 24 hours. Therefore, the exopher count includes both the intact form and the fragmented degraded form, also known as “starry night” ⁷. We scored the exopher count with the protocol published in JoVE ¹⁴. We age-synchronized the animal via egg-laying, bleaching, or L4 picking. Then, we examined at least 50 animals for each genotype or treatment with the FBS10 Fluorescence Biological Microscope (KRAMER scientific), repeated for at least 3 trials.

RNAi treatment

All RNAi clones used in this study are coming from the Ahringer RNAi library. The NGM-agar RNAi plate contains 1 mM IPTG and 25 μg/ml carbenicillin. The food on top of the medium is HT115 bacteria expressing dsRNA against targeted gene or carrying an empty vector (EV, L4440) as the control. The treatment is from eggs to the last day of each test.

FUDR treatment

The concentration of FUdR (5-fluorodeoxyuridine) on NGM-agar plate is ∼51 mM. The treatment started from adult day 1.

Male mating experiment

In each trial, we prepared ∼2000 age synchronized adult day 1 hermaphrodites and did exopher counting for 50 worms from adult day 1 to day 3. In adult day 3, we divided the worms into three groups (∼400 worms in each group, and ∼100 worms per plate). Group 1 is the control worms without males. We added ∼100 sterile males into group 2 and ∼100 normal males into group 3 for each plate.

Electron microscopy

We prepared mCherry animals for TEM analysis by high pressure freezing and freeze substitution (HPF/FS), and followed a standard to preserve ultrastructure. After HPF in a Baltec HPM-010, we exposed animals to 1% osmium tetroxide, 0.1% uranyl acetate in acetone with 2% water added, held at −90°C for 4 days before slowly warming back to −60°C, −30°C, and 0°C, over a 2 day period. We rinsed the samples several times in cold acetone and embedded the samples into a plastic resin before curing them at high temperatures for 1–2 days. We collected serial thin sections on plastic-coated slot grids and post-stained them with 2% uranyl acetate and then with 1:10 Reynold’s lead citrate, and examined with a JEOL JEM-1400 Plus electron microscope. By observing transverse sections for landmarks such as the 2nd bulb of the pharynx, it was possible to reach the vicinity of the ALM soma before collecting about 1500 serial thin transverse sections.

The microinjection experiment

We mounted animals on coverslips with dried 2% agarose pads covered in halocarbon 700 oil. We then placed coverslips onto an Axiovert S100 TV Inverted Microscope (Carl Zeiss) and injected animals with capillary needles filled with M9 + 10% red dye under 40 psi of pressure. Needles were pulled from borosil capillary tubing (1.0mm OD, 0.5mm ID; FHC) using a P-97 Micropipette Puller (Sutter Instrument). Puller parameters were as follows: Heat 474 (Ramp-25); Pull 90; Vel 100; Time 180. For the control mock injection, we transiently poked the animal in the uterine region with a needle and then transferred the animal onto an NGM-agar plate. For the injected group, we applied injection pressure for 2 minutes with an estimated flow rate of approximately 25 fL/s and then transferred the animal onto an NGM-agar plate. After each injection or mock injection, we immediately examined the exopher status of the ALMR neuron in less than 10 minutes.

Statistics

The exopher phenotype (+ or -) for each animal is a nominal variable, so we analyzed the exopher data by Cochran–Mantel–Haenszel test.There is no power analysis for each experiment, but each experiment has at least three independent trials. For analyzing progeny count, we used 2way ANOVA with Šídák’s multiple comparisons test. For analyzing the uterus length, we used unpaired two-tail t-test.

Author contributions

GW, RG, SS, EC, AM, AJS, DJ, SS, EF, IM, and KCN did the experiment. GW, RG, SS, EC, AM, AJS, DHH, BDG, and MD designed the experiment. GW, RG, SS, DHH, BDG, and MD wrote the manuscript.

Declaration of interests

The authors declare no competing interests.

Acknowledgements

We thank Dr. Andrew Singson, Dr. Amber Krauchunas, and Dr. Xue Mei for sharing reagents and providing comments in experimental design. We also thank the Caenorhabditis Genetics Center (CGC, founded by National Institutes of Health - Office of Research Infrastructure Programs (P40 OD010440)) for providing some strains. Funding sources: NIH R24 OD010943 to DHH; NIH 5R01GM135326 to BDG; NIH R37AG56510 to MD; and NIH R01AG047101 to MD and BDG.