Non-muscle myosin II-dependent cumulus cell migration at ovulation is required for sperm to reach the egg in mice

Abstract

At ovulation in mammals, epidermal growth factor (EGFR) signaling in the cumulus granulosa cells that enclose the egg causes them to generate an extracellular matrix and become dispersed within it. This process, termed expansion, is required for sperm to reach the egg. The extracellular signal-related kinase (ERK) pathway mediates matrix production, but the mechanism responsible for cell dispersion is poorly understood. We show that EGFR signaling activates non-muscle myosin II (NMII) in the cumulus cells and that NMII activity is required for full expansion. NMII activation does not require ERK signaling, but instead depends on the Rho-associated coiled-coil containing kinases (ROCK) and the myotonic dystrophy kinase-related CDC42-binding kinases (MRCK). Blocking ROCK or MRCK impairs the ability of the cumulus cells to migrate and reduces the number of sperm that penetrate through the cumulus layer to reach the oocyte. EGFR thus promotes cumulus layer expansion by integrating ERK-dependent production of the matrix with NMII-dependent migration of the cumulus cells within it.

Activation by the epidermal growth factor receptor signaling pathway of non-muscle myosin II within the cumulus granulosa cells that surround the oocyte enables their migration away from it thereby facilitating sperm access and fertilization.

Introduction

Fertilization in mammals depends on a process termed cumulus layer expansion, during which the cumulus granulosa cells that have enclosed and nourished the oocyte throughout its development within the ovarian follicle generate an extracellular matrix (ECM) and become dispersed within it^1,2. The displacement of the cumulus cells away from the germ cell relieves a physical barrier to sperm penetration, and a failure of cumulus layer expansion is associated with impaired ovulation and reduced fertility^3–7. Expansion is initiated by the surge of luteinizing hormone (LH) that also triggers meiotic maturation and ovulation. Binding of LH to its receptor on mural granulosa cells causes release of amphiregulin and other epidermal growth factor (EGF)-related ligands, which diffuse to and bind the EGF receptor located on the cumulus (as well as mural) granulosa cells^8–10. Genetic and pharmacological studies have unequivocally established that EGFR signaling is the proximate trigger for expansion^11–13. Expansion also requires SMAD2/3 signaling, which is activated independently of the LH surge by TGFβ family members secreted by the oocyte^14–16.

The mechanism by which the ECM is produced is well-understood. The LH-induced EGFR activation and constitutive SMAD2/3 signaling together increase transcription of genes encoding proteins that generate the matrix, including hyaluronan synthase (Has), which catalyzes production of hyaluronan, a glycosaminoglycan commonly present in ECM, that is the major component of the cumulus matrix^17,18. Tumor necrosis factor α-induced protein 6 (Tnfaip6) helps to assemble the hyaluronan filaments into a mesh-like network¹⁹, and pentraxin 3 (Ptx3) helps to stabilize the network⁵. Prostaglandin-endoperoxide synthase (Ptgs2), which maintains expression of EGFR ligands^3,20, is also up-regulated. The increased expression of these genes and consequently expansion are blocked when the ERK MAP kinase pathway is blocked pharmacologically or genetically in the cumulus granulosa cells^21,22, indicating that EGFR signaling triggers production of the matrix through this well-known effector pathway.

Less is known of the process by which the cumulus cells become displaced within the matrix, even though this displacement is needed to provide sperm with access to the oocyte². Other cell types use multiple mechanisms to move within a matrix^23,24. Under conditions of weak interactions with the ECM, cells may exploit a mechanism whereby the plasma membrane becomes locally detached from underlying actin cortex, generating transient or stable blebs^25,26. The blebs are thought to mediate propulsion of the cell. Under conditions of stronger ECM interaction, cells use an adhesion-based mode of movement, based on the elaboration of lamellipodia that pull the cell forward^23,24. Cells are able to switch between modes of movement, such as in response to changes in ECM adhesivity^26,27. Importantly, both rely on the activity of non-muscle myosin II (NMII), which by regulating actomyosin contractility and therefore changes in cell shape is implicated in multiple aspects of migration, including adhesion of cells to the ECM, formation of blebs and lamellipodia, and detachment from the substrate of the trailing edge^24,28. NMII activity depends on phosphorylation of its regulatory subunit, myosin light chain 9 (MYL9, also known as MLC2/MRLC1/MYRL2)^29,30, typically via the Rho-associated coiled-coil containing kinases 1 and 2 (ROCK1/2) or the myotonic dystrophy kinase-related CDC42-binding kinases (MRCK)^31–33. These enzymes are downstream effectors of the small Rho GTPases RhoA³⁴ and CDC42²⁹, respectively. ROCK1/2 also inactivates the MYL9 phosphatase (MYPT1)^35,36. High and moderate levels of NMII activity are associated with bleb- and adhesion-based modes of migration, respectively²⁴.

Several lines of evidence suggest that migration could contribute to the displacement of the cumulus cells away from the oocyte. Hyaluronan supports cell migration in other contexts³⁷ and cumulus cells of the pig and cow express the hyaluronan-binding receptor, CD44^38,39. Near the time of ovulation, cumulus cells develop blebs^2,40,41 and acquire the ability to adhere to and migrate through ECM in vitro⁴². Displacement of the cumulus cells requires calpain proteases, which become activated in vivo following hCG injection and in vitro by EGFR ligands via an ERK-dependent pathway⁴¹. Calpain activity is thought to enable cells to detach from each other, thereby enabling them to move within the hyaluronan matrix⁴¹. Yet, little else is known of the mechanism responsible for cumulus cell displacement during expansion. Notably, although the RhoA-ROCK pathway can be activated through EGFR signaling in other cell types^43,44, implying a potential role for NMII in cumulus cell displacement, a previous study reported that activation of the RhoA-ROCK-NMII axis disrupts the morphology of the expanded cumulus layer and impairs sperm penetration through it egg⁴⁵. This suggests that the movement of the cumulus cells during expansion is not mediated through NMII and thus differs essentially from cell migration in other contexts.

Here, we report that MYL9 becomes di-phosphorylated in the cumulus cells after EGF treatment of mouse cumulus-oocyte complexes (COCs) in vitro or induction of ovulation in vivo, and that inhibiting NMII activity impairs cumulus cell displacement. Moreover, stimulating RhoA activity in COCs in the absence of EGF is sufficient to trigger MYL9 di-phosphorylation. Consistent with these results, exposing EGF-treated COCs to inhibitors of ROCK or MRCK partially impairs MYL9 di-phosphorylation and cumulus cell displacement, and simultaneous inhibition of both substantially impairs these events. Strikingly, however, these treatments do not prevent up-regulation of genes encoding the matrix components, and inhibiting ERK activity does not prevent MYL9 phosphorylation. Functionally, blocking the activity of NMII, ROCK, or MRCK reduces the ability of the cumulus cells to migrate across a collagen membrane and through a synthetic hyaluronan gel, and reduces the number of sperm that can reach the oocyte following incubation of COCs with capacitated sperm. These results suggest that EGFR signaling initiates cumulus layer expansion by activating two downstream pathways—on one hand, ERK-dependent production of the hyaluronan matrix and cell detachment from each other, and on the other hand, NMII-dependent migration of the individual cells within the newly generated matrix.

Results

We began by collecting COCs from large antral follicles of eCG-primed mice and incubating them in medium supplemented with EGF to induce expansion. After 4 h of incubation, no change in the volume of the COCs was apparent, as assessed by measuring the surface area of optical images (Fig. 1A, B). By 8 h of incubation, the surface area had increased by a factor of 3.18 ± 0.15 (mean, sem, p < 0.0001) as compared to the 0 h group. Thus, EGF-triggered expansion of COCs begins 4–8 h after addition of the ligand. This is similar to the timing of expansion in vivo after injection of human chorionic gonadotropin, which activates the LH receptor^46,47 and in vitro after adding EGF⁴⁸. We therefore focused the subsequent studies within this time range.

Fig. 1. EGF activates non-muscle myosin II (NMII) during cumulus layer expansion.

COCs were collected from eCG-primed mice and incubated with recombinant EGF for 8 h. At 0, 4, and 8 h after initial EGF exposure, brightfield pictures (A) were taken to assess cumulus cell displacement by measuring the COC area (B). COCs were fixed and stained with an antibody raised against di-phosphorylated MYL9 (C) to assess NMII activation (D). COCs were also stained with phalloidin (c) to assess cell morphology and the percentage of membrane blebbing cells (E). Ovary sections obtained 0, 4 or 8 h post hCG injections were stained with anti- di-phosphorylated MYL9 (F) to assess in vivo NMII activation (G). DNA is shown in light blue, filamentous actin (F-actin) in red, and ppMYL9 foci in yellow. White arrowheads indicate membrane blebs. The number of values recorded for each group was B Fresh: 36, 4 h EGF: 60, 8 h EGF: 45. D Fresh: 33, 4 h EGF: 43, 8 h EGF: 27. E Fresh: 3, 4 h EGF: 3, 8 h EGF: 3. F 0 h: 18, 4 h post hCG: 23, 8 h post hCG: 32.

To study the potential role of NMII in cumulus cell displacement during expansion, COCs were treated with EGF as above, fixed and stained with an antibody previously used to detect the active di-phosphorylated form of MYL9 (ppMYL9) by immunofluorescence in numerous cell types including mouse embryos^49–51, as well as with phalloidin to label the cell cortex, then examined using confocal microscopy. Prior to EGF treatment, a small number of fluorescent foci were visible in some cumulus cells (Fig. 1C, D). No change was observed after 4 h. By 8 h post-EGF, however, the number of foci had increased almost three-fold (2.81 ± 0.33, p < 0.0001) compared to the 0 h group. This increase was substantially attenuated in the presence of a chemical inhibitor of EGFR signaling, AG1478, confirming that it was dependent on activation of EGFR (Supplemental Fig. S1A, B). We also observed that the percentage of cells containing membrane blebs remained statistically unchanged between 0 h and 4 h post-EGF (3.61 ± 1.81 vs 13.8 ± 2.74, p = 0.2) but had increased substantially by 8 h (68.06 ± 5.37, p < 0.0001) (Fig. 1C, E). A focal pattern of MYL9 immunostaining in blebbing cells has previously been reported⁵².

To test whether NMII became activated during cumulus layer expansion in vivo, we examined histological sections of ovarian antral follicles following injection of human chorionic gonadotropin (hCG, an LH analogue) into equine chorionic gonadotropin-primed mice. We found that, as observed in vitro in response to EGF, the number of foci remained unchanged between 0 h and 4 h post-hCG and then increased by 8 h (Fig. 1F, G). In contrast, no increase was detectable in the mural granulosa cells (Supplemental Fig. S1C, D). Thus, NMII activity in the cumulus cells of COCs increases between 4 and 8 h after exposure to EGF in vitro or injection of hCG in vivo.

We then tested whether NMII activity was required for cumulus cell displacement. COCs were treated with EGF in the presence of para-amino blebbistatin (PAB), which specifically binds and inhibits the ATPase subunit of NMII, preventing actomyosin contraction⁵³. As PAB does not block MYL9 phosphorylation, its effectiveness could not be tested using anti-ppMYL9 immunofluorescence. We instead examined cell blebbing as a readout of NMII activity. Blebs were apparent in most cumulus cells of COCs exposed to EGF alone, but in very few of those exposed to EGF and PAB (Fig. 2A, B). This result implies that, as expected, PAB inhibited NMII activity. The cumulus cells of COCs exposed to PAB also often manifested an elongated ‘tail’ (Fig. 2A, dotted lines). We quantified this using a morphological descriptor, the cell elongation index (CEI), defined as the ratio between cell length and width (Fig. 2C). While the cumulus cells from freshly collected COCs and the ones from COCs exposed exclusively to EGF presented a CEI around 1 (0.86 ± 0.21 vs. 1.16 ± 0.21, p = 0.2156), the cumulus cells of COCs exposed to the inhibitor reached a CEI greater than 2, denoting an elongated morphology. We then assessed expansion, by measuring the surface area in images of COCs treated with EGF and PAB for 8 h. Whereas the area increased about 3-fold in the control group (2.91 ± 0.17 vs. 1.00 ± 0.01, p < 0.0001), the increase was reduced in a dose-dependent manner in the presence of PAB, declining to 1.46 ± 0.07 at 100 µM (p < 0.0001 vs EGF alone) (Fig. 2D, E). Thus, NMII activity is required for full expansion. Because expansion depends on the production of the hyaluronan matrix that provides the substrate for cell displacement, we tested whether PAB affected expression of the genes whose activation is required for matrix production. No change in the quantity of the encoded mRNAs (Fig. 2F) or TNFAIP6 expression (Supplemental Fig. S2) was detectable, implying that the drug did not impair the ability of the cells to produce the matrix. Taken together, these results indicate that NMII activity is required for cumulus cell displacement during expansion.

Fig. 2. NMII activity is required for cumulus layer expansion.

COCs were collected from eCG-primed mice and incubated with recombinant EGF for 8 h with or without the molecular inhibitor of NMII ATPase, para-amino Blebbistatin (PAB), at a final concentration of 10, 50 or 100 µM. COCs were fixed and stained with phalloidin (A) to assess the percentage of blebbing cumulus cells (B, insets in A) and their morphology by measuring the Cell Elongation Index (CEI, ratio between cell length and cell width) (C). DNA is shown in light blue, and filamentous actin (F-actin) in red. White arrowheads show membrane blebs, and dotted lines highlight cell morphology. Cumulus cell displacement was assessed based on brightfield pictures obtained before and after the culture (D) by measuring the COC area (E). In (F), COCs were collected from eCG-primed mice and incubated with recombinant EGF for 3 h with or without PAB at a final concentration of 100 µM. COCs were harvested for quantitative PCR analysis to assess mRNA quantity of cumulus matrix-related genes Has2, Tnfaip6, Ptx3, and Ptgs2. The number of values recorded for each group was B Fresh: 3, 8 h EGF: 3, 8 h EGF + PAB 10 μM: 3, 8 h EGF + PAB 50 μM: 3, 8 h EGF + PAB 100 μM: 3. C Fresh: 110, 8 h EGF: 63, 8 h EGF + PAB 10 μM: 142, 8 h EGF + PAB 50 μM: 98, 8 h EGF + PAB 100 μM: 127. E Fresh: 73, 8 h EGF: 29, 8 h EGF + PAB 10 μM: 52, 8 h EGF + PAB 50 μM: 31, 8 h EGF + PAB 100 μM: 55. F Fresh: 5, 8 h EGF: 5, 8 h EGF + PAB 10 μM: 5, 8 h EGF + PAB 50 μM: 5, 8 h EGF + PAB 100 μM: 5.

The ERK pathway is a key downstream effector of EGFR signaling and cumulus expansion is blocked when the pathway is pharmacologically inhibited or genetically disabled^21,22. We therefore tested whether ERK signaling mediated the activation of NMII. U0126, a widely used inhibitor of ERK signaling, inhibited cumulus layer expansion (Fig. 3A, B) and blocked transcriptional up-regulation of the expansion-enabling genes (Supplemental Fig. S3) in EGF-treated COCs, confirming these well-established functions^21,22. In contrast, the number of ppMYL9 foci per cell increased in the presence of EGF and U0126 as compared to pre-treatment controls (2.67 ± 0.24 vs 1.00 ± 0.08, p < 0.0001), although it remained below the number observed in COCs exposed to EGF alone (3.51 ± 0.23, p = 0.02) (Fig. 3C, D). Similarly, the percentage of cells that contained blebs increased in COCs treated with EGF and U0126 compared to pre-treatment controls (32.89 ± 6.39 vs 8.64 ± 1.10, p = 0.01), although not to the same extent as EGF alone (73.16 ± 1.70, p < 0.0001) (Fig. 3C, E). These results indicate that, in contrast to its essential role in up-regulating expression of the expansion-related genes, ERK signaling is not required for EGFR-dependent activation of NMII. Therefore, we sought other pathways that might mediate this response.

Fig. 3. EGF-triggered activation of NMII does not require ERK signaling.

COCs were collected from eCG-primed mice and incubated with recombinant EGF for 8 h with or without the ERK signaling inhibitor U0126 at a final concentration of 25 µM. Brightfield images of the COCs (A) were acquired to assess cumulus cell displacement by measuring the COC area (B). COCs were fixed and stained with anti- di-phosphorylated MYL9 (C) to assess NMII activation (D). COCs were also stained with phalloidin (C) to assess cell morphology and the percentage of membrane-blebbing cells (E). DNA is shown in light blue, filamentous actin (F-actin) in red, and ppMYL9 foci in yellow. White arrowheads show membrane blebs. The number of values recorded for each group was B Fresh: 99, 8 h EGF: 51, 8 h EGF + U0126: 68. D Fresh: 28, 8 h EGF: 34, 8 h EGF + U0126: 32. E Fresh: 3, 8 h EGF: 3, 8 h EGF + U0126: 3.

As discussed in the Introduction, RhoA and its downstream effectors, ROCK1/2, can be activated by EGF and trigger activation of NMII. We tested whether EGF acts through RhoA to activate NMII during expansion in two ways. First, we collected and cultured COCs for 8 h in the presence of EGF and Y27632 (Y27), an inhibitor of ROCK activity. Whereas the number of ppMYL9 foci increased by a factor of 5.13 ± 0.43 in the COCs treated with EGF alone, the fold-increase was only 3.26 ± 0.32 (p = 0.0009 vs EGF alone) and 3.17 ± 0.24 (p = 0.0013 vs EGF alone) in the COCs also exposed to Y27 at a concentration of 10 µM and 50 µM, respectively (Fig. 4A, B). Consistent with this result, virtually none of the cells exposed to Y27 at these concentrations generated blebs (Fig. 4A, C). Rather, many cells were elongated with the long axis oriented radially with respect to the COC, a morphology that strikingly resembles that of COCs exposed to PAB (Fig. 2A). Moreover, the extent of cumulus layer expansion was significantly, albeit modestly, reduced in the presence of Y27 (EGF alone: 1.76 ± 0.04; 50 µM: 1.46 ± 0.03, p < 0.0001) (Fig. 4D, E).

Fig. 4. RhoA-ROCK signaling contributes to activation of NMII and cumulus layer expansion.

COCs were collected from eCG-primed mice and incubated with recombinant EGF for 8 h with or without Y27632 (Y27), a ROCK inhibitor, at final concentrations of 1, 10, or 50 µM. COCs were fixed and stained with anti-di-phosphorylated MYL9 (A) to assess NMII activation (B). COCs were also stained with phalloidin (A) to assess the percentage of membrane-blebbing cells (C). DNA is shown in light blue, filamentous actin (F-actin) in red, and ppMYL9 foci are shown in yellow. White arrowheads show membrane blebs and dotted white lines show cumulus cell morphology. Cumulus cell displacement was assessed based on brightfield pictures obtained before and after the culture (D) by measuring the COC area (E). In (F, G, H), COCs were incubated with a RhoA activator (RhoA act) at final concentrations of 0, 0.25 or 1 µg/ml and then stained with anti-di-phosphorylated MYL9 (F) to assess NMII activation levels (G). COCs cultured for 3 h either in plain media, with recombinant EGF or with RhoA activator (1 µg/ml) alone were submitted to quantitative PCR analysis to assess mRNA quantity of the matrix-related genes Has2, Tnfaip6, Ptx3, and Ptgs2 (H). The number of values recorded for each group was B: Fresh: 15, 8 h EGF: 18, 8 h EGF + Y27 1 μM: 19, 8 h EGF + Y27 10 μM: 18, 8 h EGF + Y27 50 μM: 18. C: Fresh: 3, 8 h EGF: 3, 8 h EGF + Y27 1 μM: 3, 8 h EGF + Y27 10 μM: 3, 8 h EGF + Y27 50 μM: 3. E Fresh: 89, 8 h EGF: 28, 8 h EGF + Y27 1 μM: 32, 8 h EGF + Y27 10 μM: 43, 8 h EGF + Y27 50 μM: 49. G Fresh: 31, 3 h: 34, 3 h 0.25 μg/ml: 31, 3 h 1 μg/ml: 23. H Fresh: 4, 3 h EGF: 4, 3 h RhoA act: 4.

Next, COCs were cultured for 3 h in the presence of a chemical activator of RhoA in the absence of EGF. This brief period of exposure was chosen to favor detection of direct effects of the activator. The results revealed a dose-dependent increase in the number of ppMYL9 foci in the presence of the activator (1 µg/ml: 4.47 ± 0.55 vs 1.00 ± 0.07, p < 0.0001) (Fig. 4F, G). Notably, the magnitude of the increase at 1 µg/ml was similar to the increase that is induced by EGF (Fig. 1D). At this concentration, almost all cumulus cells displayed blebs (Fig. 4F). Importantly, in contrast to its ability to activate NMII and cell blebbing, the RhoA activator did not increase expression of the expansion-associated genes (Fig. 4H). Because the hyaluronan matrix was not produced in the presence of the RhoA activator alone, it was not possible to assess displacement of the cumulus cells. These results support the idea that RhoA mediates the EGF-induced activation of NMII in the cumulus cells. Conversely, it does not mediate EGF-induced up-regulation of the expansion-enabling genes.

Although Y27 reduced EGF-induced MYL9 di-phosphorylation and impaired cumulus layer expansion, both remained elevated compared to pre-treatment COCs (Fig. 4B, E). MYL9 can also be phosphorylated by MRCK, which is activated by the Rho family member CDC42. To investigate its role, COCs were treated with EGF in the presence of BDP9066 (BDP), a highly selective inhibitor of MRCK⁵⁴. Whereas the number of ppMYL9 foci increased about four-fold after 8 h in the presence of EGF alone compared to the 0 h control (3.81 ± 0.38 vs 1.00 ± 0.09, p < 0.0001), this increase was reduced to two-fold in the presence of BDP (10 µM: 1.89 ± 0.16, p < 0.0001 vs EGF alone; 50 µM: 2.08 ± 0.23, p < 0.0001 vs EGF alone) (Fig. 5A, B). BDP also attenuated the increase in the percentage of cells showing blebs following EGF treatment (0 h: 6.62 ± 3.73; EGF alone: 58.69 ± 1.09; 10 µM: 18.59 ± 4.71; 50 µM: 2.22 ± 2.22; p < 0.0001 vs EGF alone) (Fig. 5A, C), and impaired cumulus layer expansion (EGF alone: 2.99 ± 0.09; 10 µM: 2.27 ± 0.06; 50 µM: 1.84 ± 0.04; p < 0.0001 vs EGF alone) (Fig. 5D, E). We then examined the effect of blocking both ROCK and MRCK. The increase in the number of ppMYL9 foci observed after addition of EGF (2.82 ± 0.28 vs 1.00 ± 0.10, p < 0.0001) was largely prevented in the presence of Y27 and BDP (1.35 ± 0.13 vs 2.82 ± 0.28, p < 0.0001) (Fig. 5F, G). Cell blebbing was completely blocked (Fig. 5F, H) and cumulus layer expansion was considerably attenuated (3.39 ± 0.13 vs 1.72 ± 0.05, p < 0.0001) (Fig. 5I, J).

Fig. 5. RhoA-ROCK and MRCK signaling cooperatively activate NMII and cumulus layer expansion.

COCs were collected from eCG-primed mice and incubated with recombinant EGF for 8 h with or without BDP9066, an MRCK inhibitor, at final concentrations of 1, 10, or 50 µM. COCs were fixed and stained with anti-di-phosphorylated MYL9 (A) to assess NMII activation (B). COCs were also stained with phalloidin (A) to assess the percentage of membrane blebbing cells (C). DNA is shown in light blue, filamentous actin (F-actin) in red, and ppMYL9 foci are shown in yellow. White arrowheads show membrane blebs, and dotted white lines show cumulus cell morphology. Cumulus cell displacement was assessed based on brightfield pictures obtained before and after the culture (D) by measuring the COC area (E). In (F–J), COCs were incubated with recombinant EGF for 8 h with or without Y27632 and BDP9066 at a final concentration of 50 µM each. COCs were fixed and stained with anti-di-phosphorylated MYL9 (F) to assess NMII activation (G). COCs were also stained with phalloidin (F) to assess the percentage of membrane blebbing cells (H). Cumulus cell displacement was assessed based on brightfield pictures obtained before and after the culture (I) by measuring the COC area (J). The number of values recorded for each group was B Fresh: 25, 8 h EGF: 20, 8 h EGF + BDP 1 μM: 24, 8 h EGF + BDP 10 μM: 19, 8 h EGF + BDP 50 μM: 25. C Fresh: 3, 8 h EGF: 3, 8 h EGF + BDP 1 μM: 3, 8 h EGF + BDP 10 μM: 3, 8 h EGF + BDP 50 μM: 3. E Fresh: 176, 8 h EGF: 83, 8 h EGF + BDP 1 μM: 90, 8 h EGF + BDP 10 μM: 86, 8 h EGF + BDP 50 μM: 103. G Fresh: 29, 8 h EGF: 35, 8 h EGF + Y27 + BDP: 53. H Fresh: 3, 8 h EGF: 3, 8 h EGF + Y27 + BDP: 3. J Fresh: 91, 8 h EGF: 76, 8 h EGF + Y27 + BDP: 90.

The experiments described so far established that NMII activity is required for displacement of the cumulus cells in response to EGF, but did not demonstrate that it conferred the ability to actively migrate. We tested this in two ways. First, COCs were submitted to a transwell migration assay, where the ability of cells to migrate across a membrane is quantified^42,55. COCs were incubated for 4 h in the presence of CNP to maintain the non-expanded condition or EGF to induce cumulus layer expansion. They were then transferred to the upper surface of cell membrane inserts placed in wells containing media supplemented with either CNP or EGF and incubated overnight (about 16 h), after which the inserts were fixed and stained. The migration index of the EGF group was about 10-fold greater than the control CNP group (11.02 ± 1.54, p < 0.0001) (Fig. 6A, B), establishing that cumulus cells acquired the ability to migrate in response to EGF. Blocking the activity of NMII using PAB decreased the relative migration index to 1.46 ± 0.57 (p < 0.0001 vs EGF), and blocking its activation (ie, phosphorylation of MYL9) by targeting either or both of ROCK and MRCK reduced the migration index to the same extent.

Fig. 6. NMII activity is required for cumulus cell migration.

COCs were collected from eCG-primed mice and incubated for 4 h in one of the following conditions: CNP (100 nM), recombinant EGF alone or with the addition of inhibitors targeting NMII (PAB, 100 µM), ROCK (Y27632, 50 µM), MRCK (BDP9066, 50 µM), or ROCK + MRCK (Y27632 + BDP9066, 50 µM each). Transwell migration assays were then performed in the presence of the inhibitors, with either CNP or EGF in the bottom chamber. After an overnight incubation, the membranes were fixed, stained (A) and the images used to quantify the relative migration index for each condition (B). COCs collected from eCG-primed mice were also embedded in commercial hyaluronan hydrogels, and exposed to either CNP, EGF or RhoA act (RhoA activator). Cumulus cell migration was assessed based on brightfield pictures obtained at 0, 4, and 8 h after the start of the culture (C) by measuring the COC area (D). The number of values recorded for each group was B: EGF: 3, PAB: 3, Y27: 3, BDP: 3, Y27 + BDP: 3. D: CNP 0 h: 27, CNP 4 h: 33, CNP 8 h: 37, EGF 0 h: 44, EGF 4 h: 48, EGF 8 h: 34, RhoA act 0 h: 42, RhoA act 4 h: 53, RhoA act 8 h: 58.

Next, we reasoned that if the COCs were embedded in a substrate that supported cell migration, the RhoA activator might trigger this process even in the absence of the endogenous hyaluronan matrix. Freshly isolated COCs were embedded in a commercially supplied hyaluronan matrix and incubated for 8 h in the presence of CNP, EGF, or RhoA act (Fig. 6C, D). No change in the diameter of the COCs occurred in the presence of CNP, whereas it had increased significantly by 8 h after addition of EGF (1.71 ± 0.07 vs 1.00 ± 0.02 at 0 h, p < 0.0001), demonstrating that expansion could occur under these experimental conditions. Strikingly, the COC diameter also increased 8 h after addition of the RhoA activator (1.28 ± 0.03 vs 1.04 ± 0.02 at 0 h, p < 0.0001). The reduced expansion triggered by RhoA activator as compared to EGF may indicate that the commercial hyaluronan matrix is less supportive of migration than the endogenous matrix, which also contains other components. Thus, activation of RhoA in intact COCs is sufficient to trigger migration of the cumulus cells.

Finally, we examined whether blocking NMII impaired the ability of sperm to penetrate through the cumulus cell layers to reach the oocyte. COCs were incubated for 8 h in the presence of EGF, in the presence or absence of the inhibitors of NMII activity. During this period, sperm were collected and incubated in the presence of Hoechst 22378, a fluorescent dye that stains the DNA of living cells, but without the NMII inhibitors to avoid potential effects on sperm motility. After washing, sperm and COCs were incubated together in the absence of the inhibitors for 1 h. The COCs were then fixed, imaged, and the number of sperm that had reached the neighborhood of the oocyte, as defined by a circle whose diameter was 15% larger than the oocyte, thus approximating the thickness of the zona pellucida, was counted. In each experiment, the values were normalized to the value obtained for COCs treated with EGF alone, whose mean was set to one.

The number of sperm that were able to penetrate through the cumulus cell layers was decreased to 0.19 ± 0.02 in the COCs expanded in the presence of PAB (p < 0.0001 vs EGF) (Fig. 7). The inhibitors of ROCK and MRCK similarly reduced the number of penetrating sperm compared to the EGF group – to about half when applied individually and to 0.14 ± 0.02 when applied together (p < 0.001 vs EGF for all treatment conditions). Notably, the number of sperm that were associated with the COCs remained unchanged among the different experimental groups (Supplemental Fig. S4), implying that the sperm ability to reach the COCs was not impaired. These results indicate that NMII-dependent cumulus cell migration is required for efficient penetration through the cumulus matrix by sperm.

Fig. 7. Sperm penetration through the cumulus layer requires NMII activation.

COCs were collected from eCG-primed mice and incubated for 8 h in one of the following conditions: recombinant EGF alone or with the addition of inhibitors targeting NMII (PAB, 100 µM), ROCK (Y27632, 50 µM), MRCK (BDP9066, 50 µM), and ROCK + MRCK (Y27632 + BDP9066, 50 µM each). COCs were washed and incubated with nuclei-stained and capacitated sperm. After 1 h, the complexes were fixed, stained with phalloidin and imaged (A). The images were used to quantify the number of sperm close to the oocyte (i.e., inside drawn dotted circle) (B). Sperm nuclei are shown in green and filamentous actin (F-actin) in red. The number of values recorded for each group was B: EGF: 58, PAB: 51, Y27: 28, BDP: 31, Y27 + BDP: 42.

Discussion

Previous studies have reported that the cumulus cells of COCs acquire a migratory phenotype during cumulus layer expansion^41,42. We show here that they migrate within the expanding matrix by a mechanism that requires NMII activity. This activity is regulated by both the RhoA-ROCK and MRCK pathways, acting downstream of EGFR signaling. These results suggest that cumulus cells exploit the same actomyosin-based mechanism to move within the matrix that is used by migrating cells in other physiological and pathological contexts. A previous study indicated, however, that the cumulus cells do not migrate via a ROCK/NMII-based mechanism⁴⁵. The apparent inconsistency between the two reports may be due to differences in timing. Whereas we studied COCs during the period of expansion, the earlier study examined post-ovulatory COCs that had completed expansion. In view of evidence that cumulus cells lose their migratory phenotype following ovulation⁴², the differences may be reconciled by proposing that NMII becomes active in the cumulus cells during the pre-ovulatory expansion phase and subsequently becomes inactive following ovulation.

Mechanistically, NMII can promote cell migration through several mechanisms, including through cellular blebbing or cell adhesion to the ECM. Our observation that the cumulus cells developed blebs 4–8 h after EGF treatment confirms previous results^40,41 and coincides both with the time of MYL9 di-phosphorylation and when they began to migrate. Treatments that blocked cell migration also blocked the appearance of blebs, consistent with migration via a bleb-based mechanism. On the other hand, relatively low concentrations of chemical inhibitors that blocked the appearance of blebs did not detectably impair cell migration, whereas higher concentrations blocked both. Since moderate levels of NMII activity support adhesion-based migration, whereas high levels are required for bleb-based migration, it may be that cumulus cells are able to exploit both mechanisms. Consistent with this, during maturation, cumulus cells of the mouse increase expression of CD44 (Supplemental Fig. S5), as previously reported in other species^38,39, which could mediate adhesion-based migration during expansion.

It may be speculated whether the migration of the cumulus cells during expansion is directed away from the egg. In the transwell assay, EGF was present in the lower chamber only, towards which the cells migrated. When EGF was added to the upper chamber as well, the cumulus cells did not migrate towards the lower chamber (Supplementary Fig. S6). These results support the idea the cells sense and move towards a higher concentration of EGF. In vivo, EGFR ligands are released from the mural granulosa cells of the pre-ovulatory follicle and diffuse towards the cumulus cells, and conceivably this could generate a concentration gradient of within the COC. Alternatively, it may be that non-directed migration of cells initially located within a cluster leads to their subsequent dispersion.

Previous studies have established that EGFR signaling initiates expansion through the ERK pathway^15,21,22. Yet, pharmacological inhibition of this pathway did not block NMII activation, as assayed by MYL9 di-phosphorylation. Conversely, inhibiting RhoA-ROCK signaling in EGF-treated COCs did not impair up-regulation of the ERK-responsive genes. Similarly, chemical activation of RhoA could induce migration of the cumulus cells but not expression of these genes. These results taken together suggest that EGFR signaling drives expansion by activating multiple effector pathways (Fig. 8). On one hand, signaling through the ERK pathway triggers production of the matrix. This also requires SMAD signaling in the cumulus cells. On the other hand, signaling through RhoA-ROCK and CDC42-MRCK, both of which can be activated through EGFR^43,44,56, enables the cumulus cells to efficiently migrate through the matrix. Additionally, Ca-activated calpain activity is required to detach the cumulus cells from each other, facilitating their migration⁴¹. The integrated effect of these independent pathways—each activated through EGFR—enables expansion of the cumulus layer and ultimately efficient ovulation and fertilization.

Fig. 8. Schematic depicting how EGFR signaling promotes cumulus layer expansion.

EGFR activity triggers expansion by integrating (A) ERK-dependent production of the cumulus matrix and (B) NMII-dependent migration of the cumulus cells, which requires the small Rho GTPases RhoA and CDC42 along with their respective downstream effectors ROCK and MRCK that phosphorylate and activate NMII, promoting membrane blebbing. Cumulus cells are depicted in blue, cumulus matrix in pink, oocyte cytoplasm in yellow, and sperm in black.

Methods

Animals

All experiments were conducted in accordance with the regulations and policies of the Canadian Council on Animal Care and were approved by the Animal Care Committee of the Research Institute of the McGill University Health Centre (protocol #7783). CD-1 mice were obtained from Charles River (St-Constant, QC) and housed at 21 °C with 40–60% relative humidity under a 12 h/12 h light/dark regime.

Collection and culture of cells

Cumulus-oocyte complexes (COCs) containing fully grown oocytes were collected from post-natal day (PD) 19–21 mice injected intraperitoneally with 5 IU of equine chorionic gonadotropin (Sigma, Windsor, ON) 44 h previously. COCs were collected in HEPES-buffered minimal essential medium (MEM) containing C-type natriuretic peptide (CNP; 100 nM; Sigma N8768) to prevent meiotic resumption. To induce cumulus layer expansion, COCs were washed to eliminate CNP and then incubated in bicarbonate-buffered MEM supplemented with recombinant human EGF (10 ng/ml; BD Biosciences 354052) and Fetal Bovine Serum (FBS, 1%; Gibco A3840001) at 37 °C in a 5% CO₂/95% air atmosphere. Alternatively, freshly isolated COCs were embedded in a tridimensional hyaluronan hydrogel (Advanced Biomatrix, #GS311F), prepared according to the manufacturer’s standard instructions and allowed to polymerize for 2 h at 37 °C 5% CO₂/95% air atmosphere before culture media addition. Depending on the experiment, the following supplements were added to the indicated final concentration: AG1478 (1 µM; Calbiochem 658552), U0126 (25 μM; Sigma U120), PAB (10, 50, or 100 µM; Cayman Chemical 22699), Y27632 (1, 10, or 50 µM; Tocris 1254), Rho Activator II (0.25 or 1 µg/ml; Cytoskeleton CN03), BDP9066 (1, 10, or 50 µM; MCE HY-111424).

Immunofluorescence

COCs were fixed for 30 min in freshly prepared 2% (w/v) para-formaldehyde (Fisher Scientific 04042) in phosphate-buffered saline (PBS, pH 7.2) containing 0.1% Triton X-100 (PBST; ACROS 9002-93-1). Samples were then washed in PBST containing bovine serum albumin (BSA; 3%; Sigma). The samples were incubated overnight in primary antibody diluted in BSA-supplemented PBST at 4 °C, washed twice in PBST, then incubated for 1 h at room temperature in the secondary antibody as well as phalloidin and DAPI in BSA-supplemented PBST, then washed in PBST. Samples were then incubated for 30 min in a 1:1 solution of Vectashield (Vector Laboratories, H-1900) in PBS and mounted in 0.5 µl drops of Vectashield on a 35 mm glass bottom dish (VWR, 75856-742). Dishes were imaged using an LSM 880 confocal microscope (Zeiss, Toronto, ON). The following reagents were used for fluorescent detection of specific proteins: Alexa Fluor 555-conjugated Phalloidin (1:1000; Thermo Fisher, A34055), DAPI (1:100, Roche 10236276001), anti-ppMYL9 (1:100; Cell Signaling, #3674), anti-rabbit IgG-Alexa 488 (1:100, Thermo Fisher, A11008).

Immunohistochemistry

Mice on PD 19–21 were injected intraperitoneally with 5 IU of equine chorionic gonadotropin (Sigma, Windsor, ON), followed by a 5 IU injection of human chorionic gonadotropin (Sigma, Windsor, ON) 46 h after. Harvested ovaries were dissected and then fixed in freshly prepared 4% (w/v) para-formaldehyde (Fisher Scientific 04042) in phosphate-buffered saline (PBS, pH 7.2) at 4 °C. After 16 h, samples were submitted to a dehydrating ethanol gradient series (70, 90 and 100%), cleared in xylene and embedded in paraffin. Paraffin sections were then rehydrated in a decreasing ethanol series (100, 90, 70%), boiled in 10 nM sodium citrate solution (pH 6) for 20 min and blocked with CAS universal blocking solution (Thermo Fisher, 008120) for 1 h at room temperature to reduce nonspecific binding. After overnight incubation with the primary antibody diluted in CAS solution, the slides were washed twice in PBS, then incubated for 1 h at room temperature in the secondary antibody as well as DAPI in CAS solution, then thoroughly washed in PBST. To stain and image the extracellular matrix, COCs were collected as described above and embedded in small drops of tridimensional hyaluronan hydrogel (Advanced Biomatrix, #GS311F) that were allowed to polymerize inside upsidedown plates, to form domes. After polymerization, domes were incubated in bicarbonate-buffered MEM supplemented with recombinant human EGF (10 ng/ml; BD Biosciences 354052) and Fetal Bovine Serum (FBS, 1%; Gibco A3840001) at 37 °C in a 5% CO₂/95% air atmosphere. After 8 h, domes were fixed in freshly prepared 4% (w/v) para-formaldehyde in PBS for 1 h. After rinsing with PBS, the domes were carefully embedded in 2% low-melting-point agarose (Thermo Scientific, FERR0801). Agarose blocks were subsequently submitted to a sucrose gradient treatment (15 and 30% (w/v)) for 1 h each step before being frozen in OCT (Fisher Scientific, 23-730-571). OCT blocks were cut with the help of a cryostat (Leica CM1950). Slides containing COC sections were washed with PBS, boiled in 10 nM sodium citrate solution (pH 6) for 20 min, blocked with CAS solution, incubated with the primary and secondary antibodies, as described above. Slides were then mounted with Vectashield (Vector Laboratories, H-1900) and imaged using an LSM 880 confocal microscope (Zeiss, Toronto, ON). The following reagents were used for fluorescent detection of specific proteins: DAPI (1:100, Roche, 10236276001), anti-ppMYL9 (1:100; Cell Signaling, #3674), anti-TNFAIP6 (1:100, Invitrogen, PA5-47300), anti-rabbit IgG-Alexa 488 (1:100, Thermo Fisher, A11008), anti-goat IgG-Alexa 488 (1:100, Thermo Fisher, A-11055).

Confocal image analysis

To quantify the number of ppMYL9 foci, confocal tiles of 2 × 2 were obtained at the oocyte’s equatorial plane. Using Fiji software (National Institutes of Health, Bethesda, MD), the ‘Analyze particles’ tool was used to count the number of ppMYL9 foci above a set threshold that eliminated background noise. This value was divided by the number of cumulus cells present in the image to obtain the average number of ppMYL9 foci per cell per COC and enable comparison between different samples. The number of cumulus cells per COC was calculated by counting the number of nuclei stained by DAPI in each image. To quantify the Cell Elongation Index (CEI), which is the ratio between cell length and width, the length of cumulus cells was measured along the oocyte radial axis, and the width was obtained by tracing a line perpendicular to it, at the height of the nucleus.

Gene expression analysis by qPCR

Total RNA was extracted using the PicoPure RNA isolation kit, and cDNA was synthesized using the SuperScript IV Reverse Transcription kit, all following the manufacturers’ instructions. cDNA samples were diluted at 1:10 in nuclease-free water and used as a template to assess the relative mRNA quantity of Has2, Tnfaip6, Ptx3 and Ptgs2, using Rpl19 and Atpb5 as internal controls. The quantitative PCR (qPCR) reactions were run in an LC480 II LightCycler (Roche) with the SsoAdvanced Universal SYBR Green qPCR Master Mix (Biorad, 1725271). Fold changes were calculated using the 2^–ΔΔCt method⁵⁷. The following primer pairs were used Has2 (F: 5’ GGCCGGTCGTCTCAAATTCA 3’, R: 5’ ACAATGCATCTTGTTCAGCTCC 3’), Tnfaip6 (F: 5’ GATGGCCAAGGGTAGAGTCG 3’, R: 5’ ACACTCCTTTGCATGTGGGT 3’), Ptx3 (F: 5’ CTGCCCGCAGGTTGTGAA 3’, R: 5’ TGGTCTCACAGGATGCACG 3’), Ptgs2 (F: 5’ AGCCAGGCAGCAAATCCTT 3’, R: 5’ CAGTCCGGGTACAGTCACAC 3’), Rpl19 (F: 5’ TCAGGCTACAGAAGAGGCTTGC 3’, R: 5’ ATCAGCCCATCCTTGATCAGC 3’), and Atpb5 (F: 5’ GAGGTCTTCACGGGTCACAT 3’, R: 5’ CCCACCATGTAGAAGGCTTG 3’).

Transwell migration assay

COCs were collected as described and initially incubated in bicarbonate-buffered MEM for 4 h at 37 °C in a 5% CO₂/95% air atmosphere. Depending on the experimental group, the initial incubation media was supplemented with FBS (1%) and either CNP (100 nM), EGF (10 ng/ml), or EGF plus PAB (100 µM), Y27632 (50 µM), BDP9066 (50 µM) or Y27632 + BDP9066 (50 µM each). After this period, groups of 20 COCs were thoroughly washed and transferred to the top of 12 µm pore polycarbonate transwell inserts (Millipore, PIXP01250) placed in 24-well cell culture plates and incubated overnight at 37 °C and a 5% CO₂ atmosphere. The bottom well contained media supplemented with the same components previously added during the initial incubation, according to the experimental group. The next morning, the remaining cells on the top of the inserts were scraped with a cotton swab, and the membranes were fixed in ethanol 70% for 5 min and stained using Crystal Violet (2%) for 5 min at room temperature. After three washing steps in distilled water, the membranes were removed from the plastic inserts, mounted on slides, and imaged with a Leica DMi6000 brightfield microscope. The area covered by the crystal violet staining was divided by the total membrane area to obtain the migration index for each experimental group.

Sperm penetration assay

COCs were collected as described and incubated for 8 h at 37 °C and a 5% CO₂ atmosphere. Depending on the experimental group, the bicarbonate-buffered MEM incubation media was supplemented with FBS (1%) and either CNP (100 nM), EGF (10 ng/ml), or EGF plus PAB (100 µM), Y27632 (50 µM), BDP9066 (50 µM) or Y27632 + BDP9066 (50 µM each). Approximately 6 h into the COC incubation, epididymides of two 5-week-old CD-1 male mice were collected, dissected and had their contents squeezed into 300 µl drops of bicarbonate-buffered MEM supplemented with BSA (0.9%) but no EGF or drugs and incubated at 37 °C for capacitation. After 1 h, 30 µl of HOECHST (0.01 mg/ml) was added to the drops with motile sperm to stain the DNA, and the plates were incubated for 10 additional minutes. For insemination, groups of 15 COCs were washed three times in 400 µl each time of medium without drugs, then transferred to 50 µl drops of bicarbonate-buffered MEM supplemented with BSA (0.4%). Ten µl of the capacitated and stained sperm was added to each drop, and the plates were incubated for an hour. This protocol was designed to ensure that the sperm were not exposed to the drugs. The COCs were then fixed, stained with phalloidin and mounted as previously described. Detailed Z-stacks were obtained for each COC. To quantify the number of penetrating sperm, a maximal intensity projection image was obtained for each channel (red – F-actin, blue – sperm nuclei) using the Fiji software. A circle with a diameter 15% bigger than the oocyte was drawn around it and used as a reference on the blue channel maximal intensity projection to count the number of sperm that had penetrated the cumulus layer and reached the surroundings of the zona pellucida. To verify if sperm motility remained unchanged among the different experimental groups, the maximal intensity projection of pictures containing an isolated COC was used to quantify the total number of sperm per COC. This was done by counting the total number of sperm nuclei around and within the COC in the projections.

Statistics and reproducibility

Data for each experiment were collected from a minimum of three biological replicates. Quantitative data were analyzed using GraphPad Prism (10.1.0) employing one-way ANOVA and the Tukey multiple-comparison tests. In all figures, error bars represent the standard error of the mean (SEM). Differences in means were considered statistically significant when P < 0.05 and are denoted by different letters above the relevant bars of each histogram.

Reporting summary

Further information on research design is available in the Nature Portfolio Reporting Summary linked to this article.

Author contributions

K.F.C. conceived and developed the project, performed all experiments and co-wrote the manuscript. H.J.C. helped to develop the project, obtained funding and co-wrote the manuscript.

Peer review

Peer review information

Communications Biology thanks Darryl L Russell and the other, anonymous, reviewer(s) for their contribution to the peer review of this work. Primary Handling Editors: Edwina McGlinn and Mengtan Xing. [A peer review file is available.]

Data availability

No large data sets were generated. All data supporting the findings of this study are available within the paper and its Supplementary Information. Source data can be found in Supplementary Data 1.

Competing interests

The authors declare no competing interests.

Supplementary information

The online version contains supplementary material available at 10.1038/s42003-025-08736-y.