GRAMD1B regulates cell migration in breast cancer cells through JAK/STAT and Akt signaling

Dysregulated JAK/STAT signaling has been implicated in breast cancer metastasis, which is associated with high relapse risks. However, mechanisms underlying JAK/STAT signaling-mediated breast tumorigenesis are poorly understood. Here, we showed that GRAMD1B expression is upregulated on IL-6 but downregulated upon treatment with the JAK2 inhibitor AG490 in the breast cancer MDA-MB-231 cells. Notably, Gramd1b knockdown caused morphological changes of the cells, characterized by the formation of membrane ruffling and protrusions, implicating its role in cell migration. Consistently, GRAMD1B inhibition significantly enhanced cell migration, with an increase in the levels of the Rho family of GTPases. We also found that Gramd1b knockdown-mediated pro-migratory phenotype is associated with JAK2/STAT3 and Akt activation, and that JAK2 or Akt inhibition efficiently suppresses the phenotype. Interestingly, AG490 dose-dependently increased p-Akt levels, and our epistasis analysis suggested that the effect of JAK/STAT inhibition on p-Akt is via the regulation of GRAMD1B expression. Taken together, our results suggest that GRAMD1B is a key signaling molecule that functions to inhibit cell migration in breast cancer by negating both JAK/STAT and Akt signaling, providing the foundation for its development as a novel biomarker in breast cancer.


GRAMD1B inhibition causes morphological changes of breast cancer cells.
To examine the function of GRAMD1B in JAK/STAT signaling-mediated biological processes such as cell invasion and migration in breast cancer cells, we first assessed the potency of si-Gramd1b, and found that it can efficiently knockdown Gramd1b, as evidenced by the significant reduction of both mRNA (Fig. 2a) and protein levels (Fig. 2b,b'). Interestingly, we noticed that the parental MDA-MB-231 cells which exhibit a spindle-shaped morphology (Fig. 2c), become rounded and flattened in shape on Gramd1b knockdown, (Fig. 2d), with the occurrence of cell membrane protrusions (Fig. 2e, arrow). In support of this, phalloidin staining of the Gramd1b knockdown cells revealed the presence of F-actin-rich membrane protrusions (arrow) accompanied by membrane ruffling (dotted line) (Fig. 2f-h), hallmarks of cell motility 25 , suggesting the putative function of GRAMD1B in cell migration. To exclude possible off-target results of si-Gramd1b, we used another siRNA targeting Gramd1b (si-Gramd1b-2) and confirmed a change in morphology of MDA-MB-231 cells, with formation of membrane protrusions (Supplementary Figs S1 and S2).

GRAMD1B inhibition promotes cell migration in breast cancer cells.
To further explore the potential role of GRAMD1B in cell migration, we conducted the transwell migration assay. Cells with reduced GRAMD1B activity showed significantly higher migration rates as compared to control cells (Fig. 3a,a'). This pro-migratory phenotype was further confirmed by the wound healing assay. While Gramd1b knockdown   cells almost entirely covered the wound by 36 hours, only ~40% of the wound had been covered by control cells (Fig. 3b,b'), suggesting that GRAMD1B negatively regulates cell migration in breast cancer cells. Since the Rho family of GTPases have been implicated in modulating the dynamics of actin cytoskeleton, thereby controlling directional migration 26 , we next assessed whether GRAMD1B can modulate their expression. Interestingly, we detected an increase in both mRNA and protein levels of Rac1, RhoA and Cdc42 upon Gramd1b knockdown (Fig. 3c,d). Furthermore, inhibitors targeting the members of the Rho family of GTPases efficiently suppressed changes in cell morphology observed upon Gramd1b knockdown ( Supplementary Fig. S3). These findings suggest that GRAMD1B inhibition-mediated morphological changes of the cells, followed by increased migratory rates, may be associated with the Rho family of GTPases. GRAMD1B negates JAK/STAT signaling. The IL-6/JAK/STAT3 feedback loop has been well documented in breast tumor growth and metastasis 27 . Notably, increased expression levels of IL-6 and p-STAT3 have been detected at the leading edge of breast tumors and linked to advanced disease, suggesting a mechanistic role of the JAK/STAT cascade in promoting breast tumor progression 8,28 . Since GRAMD1B is positively regulated by the JAK/STAT pathway, and its inhibition facilities cell migration, we hypothesized that cell migration facilitated by the loss of GRAMD1B activity might be a result of JAK/STAT signaling activation, as some of STAT transcriptional targets such as SOCS negatively feedback to suppress JAK/STAT signaling 23,24 . Interestingly, we observed a dramatic induction of p-JAK2 and its downstream target p-STAT3 upon Gramd1b knockdown (Fig. 4a). This suggests that the JAK/STAT transcriptional target GRAMD1B negates the signaling by regulating JAK2 activity and that the increased migratory rates observed in Gramd1b knockdown cells are due to JAK/STAT signaling activation. To verify this hypothesis, we then examined the effects of GRAMD1B inhibition on cell morphology and migration in the presence of the JAK2 inhibitor AG490. We observed that JAK/STAT signaling inhibition almost completely reversed the flattened cells back into their original spindle-shaped morphology (Fig. 4b) and significantly suppressed the enhanced cell migration (Fig. 4c,c'). Furthermore, our qRT-PCR analysis showed that AG490 treatment efficiently blocked the increase in the expression of Rac1 and RhoA induced by Gramd1b knockdown (Fig. 4d).

GRAMD1B inhibition increases Akt signaling.
The PI3K/Akt signaling pathway has been found to be frequently dysregulated in breast tumors where it promotes cell migration and invasion, as well as contributes to chemoresistance [9][10][11] . Since GRAMD1B inhibition facilitates cell migration, we asked whether it interacts with the PI3K/Akt pathway in cell migration. We first assessed the inhibitory effects of GRAMD1B on PI3K and Akt activation, and found that cells transfected with si-Gramd1b showed no change in p-PI3K levels but exhibited a significant increase in p-Akt, whereas total PI3K and Akt levels remained unchanged (Fig. 5a). This finding suggests that GRAMD1B inhibition-enhanced cell migration is associated with Akt activation and that inhibition of Akt may block the enhanced cell migration. Interestingly, the effects of GRAMD1B inhibition on cell morphology ( Fig. 5b) and migration (Fig. 5c,c') were efficiently suppressed in the presence of MK-2206, an oral pan-Akt inhibitor. Furthermore, co-treatment of cells with si-Gramd1b and MK-2206 significantly inhibited the expression of Rac1 and Cdc42 induced by Gramd1b knockdown (Fig. 5d). These findings suggest that GRAMD1B may act downstream to PI3K, but upstream or parallel to Akt in the regulation of cell migration.

GRAMD1B plays a key role in the linkage between JAK/STAT and Akt signaling. Since
GRAMD1B is involved in the regulation of both JAK/STAT and Akt signaling, we next asked whether JAK/ STAT and Akt signaling mutually regulate each other. Inhibitory effect of the JAK/STAT cascade on PI3K/Akt signaling was examined by treating cells with increasing concentrations of AG490. As expected, AG490 decreased p-STAT3 levels. Interestingly, AG490 treatment resulted in a dose-dependent increase in p-Akt levels, whereas it failed to affect p-PI3K levels (Fig. 6a). This finding suggests that JAK/STAT signaling acts downstream to PI3K, but upstream or parallel to Akt which is similar to the mechanism of action of GRAMD1B on PI3K/Akt signaling. On the other hand, inhibition of Akt signaling by MK-2206 did not cause any drastic alteration in p-STAT3 levels (Fig. 6b). To understand the molecular mechanism underlying the increase in p-Akt levels by JAK/STAT signaling inhibition, we performed epistasis analysis in cells treated with si-Gramd1b in the presence of AG490 or MK-2206. As expected, knockdown of Gramd1b alone caused an increase in both p-STAT3 and p-Akt levels. The increased p-STAT3 levels were diminished by AG490 treatment, but were not altered in the presence of MK-2206. However, interestingly, the elevated p-Akt levels by GRAMD1B inhibition were further enhanced in the presence of AG490 but completed suppressed by MK-2206 (Fig. 6c). Considering the fact that both JAK/STAT signaling and GRAMD1B act upstream or parallel to Akt, these findings may suggest that inhibitory effects of JAK/STAT signaling on Akt activation is through the regulation of GRAMD1B expression (Fig. 6d).

Discussion
The JAK/STAT pathway functions as a key regulator of a wide variety of physiological and biological processes 5,29,30 . Thus, dysregulated signaling is often associated with various pathological conditions, including cancer. In breast tumors, persistently-active STAT3 has been found to promote breast tumor progression by facilitating cancer cell proliferation, angiogenesis and EMT 6 . Particularly, the IL-6/JAK/STAT3 autocrine activation loop is a key driver of cancer progression and metastasis in breast cancer 8,27 . However, there is a gap in our knowledge of the downstream effectors and signaling mechanisms underlying JAK/STAT-mediated breast carcinogenesis. Here, we showed that JAK/STAT signaling positively regulates GRAMD1B expression, which in turn negates the signaling in breast cancer cells, signifying the existence of a negative feedback mechanism. Moreover, we provided evidence that GRAMD1B modulates breast cancer cell migration through the regulation of both JAK/ STAT and Akt signaling. Lastly, our epistasis analysis suggested the pivotal function of GRAMD1B in mediating the inhibitory effect of JAK/STAT signaling on Akt activity in breast cancer cells.
Intriguingly, loss of GRAMD1B activity transformed the parental spindle-shaped MDA-MB-231 cells into rounded and flattened cells (Fig. 2c-e), which are often associated with extensive cell migration 31 . During the process of metastasis, cancer cells also undergo transformation into rounded and flattened cells, conferring a cell migratory advantage to metastasize from the primary tumor to secondary sites 32 . Furthermore, si-Gramd1b knockdown cells exhibited F-actin-rich protrusions at the cell leading edges, as well as membrane ruffle formation ( Fig. 2f-h). F-actin-rich membrane cell protrusions at the leading edge of motile cells are known to serve as one of the key driving forces in cell migration and extension 25 . In addition to this, membrane ruffling which often precedes the formation of lamellipodia is essential among other events for cell motility 33 . These observations led us to postulate that the GRAMD1B-associated cell morphology changes may play an important role in breast cancer cell migration. As expected, we observed a dramatic increase in cell migration rates (Fig. 3a,b) on Gramd1b knockdown, and detected increased expression of Rac1, RhoA and Cdc42 (Fig. 3c,d), which play a central role in the regulation of the actin cytoskeleton and cellular migration [34][35][36] . Increased Rac1 activity promotes ruffling and stimulates actin polymerization to generate lamellipodia through the activation of the ARP2/3 complex 26,37 . On the other hand, RhoA activation is required for the formation of stress fibres, which function in cellular contractility that is essential for driving cell migration 38 , and Cdc42 is responsible for the formation of filopodia that is often found to be active at the front of migrating cells, resulting in increased directional migration 39 . Hence, it is conceivable that GRAMD1B functions to suppress cell migration by regulating the expression of the Rho family of GTPases. Nonetheless, further analyses are required to determine the mechanism underlying the interaction between GRAMD1B and Rho GTPases in modulating breast cancer cell migration.
The JAK/STAT signaling pathway is one of many control pathways that promotes cell motility by regulating actin dynamics and activating key metastasis-promoting genes 40 . In particular, STAT3 activation induced by interleukin family of cytokines can promote migration and invasion via the regulation of downstream target molecules such as Vimentin, Twist, MMP-9 and MMP-7 6 . Moreover, JAK2 activation mediates WASF3 upregulation, which subsequently promotes the formation of lamellipodia and increases cell migration via recruitment of ARP2/3 41 . We showed that IL-6-induced JAK/STAT activation increases GRAMD1B expression, whereas blockage of JAK2 activity decreases its expression (Fig. 1a,b). Notably, inhibition of GRAMD1B resulted in the induction of JAK2 activity (Fig. 4a), suggesting that a negative feedback mechanism exists. Several STAT downstream target genes are known to feedback into the circuitry and affect activity 24 . For instance, cytokine stimulation induces the production of SOCS proteins, which in turn negate JAK/STAT signaling by either directly inactivating JAK, blocking access of the STAT molecules to receptor binding sites, or promoting ubiquitination of JAK and/or STAT 23,42 . Considering the essential roles of JAK/STAT signaling in cell migration, our study suggests that GRAMD1B inhibition-enhanced cell migration is mediated by the activation of JAK/STAT signaling. In support of this, we detected that treatment of Gramd1b knockdown cells with AG490 almost completely bypassed the inhibitory effects of GRAMD1B on cell morphology (Fig. 4b) and migration (Fig. 4c,c'), as well as blocked the increase in the expression of Rac1 and RhoA induced by Gramd1b knockdown (Fig. 4d). Notably, the CDC42-like GTPase 1 RhoU was found to be transcriptionally regulated by STAT3, with its mRNA levels induced on gp130-mediated cytokine stimulation and a significant reduction of its protein level in Stat3-null cells 43 , suggesting the possible role of STAT signaling in the transcriptional regulation of the Rho family of GTPases. Particularly, the GRAM domain in myotubularin was shown to be required for the endosomal trafficking of endocytosed EGFR to regulate its activation 44,45 . In support of this, the GRAM domain shares similarities to the GLUE (GRAM-like ubiquitin-binding in Eap45 domain), which binds to ubiquitin that functions to mark endocytosed receptors for lysosomal degradation 46 . In the JAK/STAT signaling pathway, the ligand-receptor internalization and trafficking to the early endosome were reported to be associated with the signaling intensity 47 , suggesting that GRAMD1B may function to promote ligand-receptor decay, causing the de-activation of JAK/STAT signaling.
Another signaling cascade often implicated in breast tumorigenesis is the PI3K/Akt signaling pathway, where it is considered a potential therapeutic target due to its role in tumor initiation and progression 10,48,49 . For example, Ca 2+ -dependent Akt activation was implicated in TRPV4-mediated breast cancer cell migration and metastasis 50 , and Twist-Akt2 signaling axis was shown to be essential in promoting the invasive ability and survival of breast cancer cells 13 . Furthermore, Akt has been found to be required for the formation of membrane ruffles and lamellipodia through interaction with actin filaments and co-localization with the ARP2/3 complex 51,52 . Notably, our data suggested that GRAMD1B acts downstream to PI3K, but upstream/parallel to Akt. The spatial localization of PI3K and its products at the leading edge of motile cells was shown to be crucial for effective cell migration. Specifically, PI3K and its interacting PH domain-containing protein binding partners such as Akt have been reported to translocate to the plasma membrane in response to a chemoattractant stimulus 53,54 . At the leading edge, the membrane translocated Akt interacts with the PI3K products PI(3, 4, 5)P 3 and PI(3, 4)P 2 , triggering well-coordinated cell movement 53 . Since GRAM domain has been implicated in membrane-coupled lipid/ protein-binding, it is conceivable that GRAMD1B may play a role in translocation of Akt and/or de-stabilization of the interaction between Akt and the PI3K products at the leading edge of motile cells. In support of this, in another GRAM domain-containing protein MTMR2, the GRAM domain has been recognized to play a role in PI(3, 5)P 2 and PI(5)P substrate recognition 55 , thereby highlighting the possible function of GRAM domain in phosphoinositide recognition.
Several reports have indicated a close interaction between the JAK/STAT and PI3K/Akt signaling cascades in promoting metastasis in breast cancer 4,16 . Specifically, co-inhibition of the PI3K/mTOR and JAK2 signaling cascades was found to synergistically reduce breast tumor growth and metastasis, as well as improve overall survival in vivo 16 . Therefore, we undertook epistasis analysis to further elucidate the regulatory hierarchy between JAK/ STAT and PI3K/Akt signaling cascades on loss of GRAMD1B activity. Interestingly, we detected a dose-dependent increase in p-Akt levels on JAK/STAT inhibition by AG490 (Fig. 6a), however, no alteration in p-STAT3 levels was observed on Akt inhibition by MK-2206 (Fig. 6b), suggesting the inhibitory effect of the JAK/STAT cascade on Akt activity in MDA-MB-231 cells. Importantly, this induction of p-Akt levels was further enhanced in cells transfected with si-Gramd1b (Fig. 6c), thereby providing new knowledge about GRAMD1B being the central player in regulating the inhibitory effect of JAK/STAT signaling on Akt activity in breast tumorigenesis.
Notably, the JAK/STAT signaling cascade is an important regulatory pathway mediating breast tumor growth and survival 6 . There is evidence supporting a direct correlation between STAT3 activation and increased Cyclin D1 expression in primary breast tumors and breast cancer-derived cell lines 6,56 . Consistently, loss or depletion of STAT3 in breast carcinoma cells has been shown to result in tumor inhibition and induction of apoptosis 57,58 . Several reports also support a pivotal function of the Akt pathway in mediating breast cancer cell proliferation 9,59 , such that its inhibition impedes cell cycle progression and promotes cell death 60 . Since our findings revealed a novel central role of GRAMD1B in negatively regulating the JAK/STAT and Akt pathway, it is conceivable that GRAMD1B may function to inhibit breast cancer cell proliferation and promote cell death. Further analyses exploring the role of GRAMD1B in these cancer hallmarks is necessary to provide a deeper understanding into its exact function in regulating breast tumorigenesis.
Given that the understanding of the mechanisms underlying cell migration can provide crucial insights for the development of anticancer therapeutic agents, the identification of molecules that play an important role in cell motility is therefore imperative in the fight against cancer metastasis. Due to the novelty of GRAMD1B, more studies need to be carried out to further understand the mechanisms by which it regulates breast tumorigenesis through modulating both JAK/STAT and Akt signaling. Nonetheless, our study suggests that GRAMD1B is a key signaling molecule that functions to inhibit cell migration in breast cancer, providing the foundation for its development as a novel biomarker in breast tumors. Cruz Biotechnology) were reconstituted in DMSO. Small interfering RNA (siRNA) targeting Gramd1b (5′GCUCUUAGAGUCCCAACAATT3′; 3′TTCGAGAAUCUCAGGGUUGUU5′) was designed and synthesized by Singapore Advanced Biologics Pte. Ltd. (SABio, Singapore). si-Gramd1b-2 (Ambion, AM16708) was used to rule out off-target effects of siRNA for Gramd1b. Non-targeting siRNA (Ambion) was used as a negative control. Cells were transfected with siRNA using the transfection reagent Lipofectamine 3000 (Invitrogen, USA) as per the manufacturer's instructions.

RNA extraction and quantitative real-time PCR (qRT-PCR). The RNeasy mini kit (Qiagen GmbH,
Germany) was used to extract total RNA, which was subsequently converted to cDNA using the Revert Aid First Strand cDNA Synthesis Kit (Thermo Fisher Scientific, USA). The samples were loaded in triplicates in a 96-well plate for each sample set, and quantitative PCR was carried out using the FAST SYBR green cocktail (Applied Biosystems, USA) in the HT7900 FAST Realtime PCR system (Applied Biosystems,USA). Sequence of the primers (IDT technologies) used for this study are listed in Supplementary Table S1 Immunofluorescence staining. Following 72 hours of transfection with siRNA, cells were fixed using 4% paraformaldehyde, and then permeabilized with 0.1% Triton X-100. For F-actin cytoskeleton staining, the cells were incubated with TRITC-conjugated Phalloidin (Merck Millipore, USA) at 1:500 dilution at room temperature for 1 hour. The slides were imaged under the Confocal Laser Scanning Microscope (Olympus Fluoview FV 1000).
Transwell migration assay. Following 48 hours of transfection with siRNA, cells were treated with AG490 at 50 μm or MK-2206 at 10 μm for an additional 24 hours. Following the treatment, cells were re-suspended in serum free RPMI-1640 medium and seeded into polycarbonate membrane transwell inserts (Corning Inc., USA). RPMI-1640 medium containing 10% FBS was used as a chemoattractant, and cells were incubated at 37 °C for 18 hours to allow migration. Following incubation, the migrated cells were stained using 0.5% crystal violet and visualized using the Nikon SMZ1500 microscope. The average number of migrated cells per insert was calculated by imaging five different fields.
Wound healing assay. A linear scratch or wound was made across the confluent monolayer of the 72 hours post transfected cells using a fine 10 μl pipette tip. Three random fields were marked out, and images were subsequently taken at 12 hour intervals to monitor cell migration. The average gap width and average number of migrated cells across the three marked fields were then measured and calculated at each time point. Statistical analysis. The GraphPad prism 6 software (GraphPad Prism, USA) was used to carry out statistical analysis. For comparing means between two groups, a two-tailed student T-test was used. A one-way ANOVA was used for tests involving more than two groups. For wound healing assay, the two-way ANOVA statistical test was adopted. Data is represented as means ± SEM, and results are considered statistically significant if P < 0.05.