Overexpressed Gαi1 exerts pro-tumorigenic activity in nasopharyngeal carcinoma

The current study tested the expression and potential functions of Gαi1 in nasopharyngeal carcinoma (NPC). The Cancer Genome Atlas (TCGA) database results demonstrate that Gαi1 transcripts’ number in NPC tissues is significantly higher than that in the normal nasal epithelial tissues. Its overexpression correlates with poor survival in certain NPC patients. Moreover, Gαi1 is significantly upregulated in NPC tissues of local primary patients and in different primary human NPC cells. Whereas its expression is relatively low in cancer-surrounding normal tissues and in primary nasal epithelial cells. Genetic silencing (via shRNA strategy) or knockout (via CRISPR-sgRNA method) of Gαi1 substantially suppressed viability, proliferation, cell cycle progression, and migration in primary NPC cells, causing significant caspase-apoptosis activation. Contrarily, ectopic Gαi1 expression exerted pro-tumorigenic activity and strengthened cell proliferation and migration in primary NPC cells. Gαi1 is important for Akt-mTOR activation in NPC cells. Akt-S6K phosphorylation was downregulated after Gαi1 shRNA or KO in primary NPC cells, but strengthened following Gαi1 overexpression. In Gαi1-silenced primary NPC cells, a S473D constitutively-active mutant Akt1 (caAkt1) restored Akt-S6K phosphorylation and ameliorated Gαi1 shRNA-induced proliferation inhibition, migration reduction and apoptosis. Bioinformatics analyses proposed zinc finger protein 384 (ZNF384) as a potential transcription factor of Gαi1. In primary NPC cells, ZNF384 shRNA or knockout (via CRISPR-sgRNA method) decreased Gαi1 mRNA and protein expression, whereas ZNF384 overexpression upregulated it. Importantly, there was an increased binding between ZNF384 protein and the Gαi1 promoter in human NPC tissues and different NPC cells. In vivo studies showed that intratumoral injection of Gαi1-shRNA-expressing adeno-associated virus (AAV) impeded subcutaneous NPC xenograft growth in nude mice. Gαi1 downregulation, Akt-mTOR inactivation, and apoptosis induction were detected in Gαi1-silenced NPC xenograft tissues. Gαi1 KO also effectively inhibited the growth of NPC xenografts in nude mice. Together, overexpressed Gαi1 exerts pro-tumorigenic activity in NPC possibly by promoting Akt-mTOR activation.


INTRODUCTION
Nasopharyngeal carcinoma (NPC), a common malignancy in nasopharynx epithelia, demonstrates diverse etiopathogy and histopathology [1,2].NPC is one of most malignant cancers, causing over 72,000 death each year globally [2][3][4].Its incidence is still rising and can exceed two cases per ten thousand people in certain regions (i.e.South East Asia) [5][6][7].Epstein-Barr virus (EBV) infection is one primary risk factor of NPC [2][3][4] and cigarette smoking/exposure is another important factor [2][3][4].NPC has three histology types, including squamous cell carcinoma (SCC), nonkeratinizing carcinoma and undifferentiated carcinoma [5][6][7].The current standard therapies for NPC include radiotherapy plus conventional chemotherapy, molecularly-targeted therapy and immunotherapy, which have increased overall survival of a large proportion of NPC patients [1,2].Yet, for the NPC patients with recurrent or metastatic carcinomas, the prognosis is still poor and survival is often short.Thus, there are urgent needs to explore new oncotargets and to develop corresponding therapeutics for NPC [8].
Recent studies have proposed a possible pro-tumorigenic role of Gαi1 in cancer cells.Liu et al. reported that Gαi1 expression is elevated in human glioma tissues possibly owing to the downregulation of an anti-Gαi1 microRNA, miR-200a [20].Contrarily, overexpression miR-200a or genetic silencing of Gαi1 potently inhibited glioma cell growth in vitro and in vivo [20].Overexpression of YME1L, a mitochondrial protein, enhanced transcription and expression of Gαi1, thereby promoting glioma cell growth in vitro and intracranial glioma xenograft growth in mice [23].Lv et al., discovered that upregulation of Gαi1 was important for gastric cancer growth [24].PINK1-AS, a novel long non-coding RNA, promoted Gαi1 expression and gastric cancer cell growth possibly by sponging miR-200a [24].The current study tested the expression and potential functions of Gαi1 in NPC and explored the possible underlying mechanisms.

MATERIALS AND METHODS Reagents
Fluorescence dyes, including TUNEL, DAPI, EdU, and propidium iodide (PI) were provided by Thermo-Fisher Invitrogen (Suzhou, China).All the antibodies were provided by Dr. Cao at Soochow University [20].Other reagents were reported in our previous study [25].

Human tissues
As reported early [25], NPC tumor tissues and matched adjacent normal nasopharynx epithelial tissues were obtained from twenty (n = 20) primary NPC patients.All patients were administrated at the First affiliated Hospital of Soochow University (Suzhou, China) and each provided writteninformed consent.Tissues were obtained freshly at the time of surgery and were immediately stored in liquid nitrogen.The protocols of testing human tissues were approved by the Ethics Committee of Soochow University, according to the principles of the Declaration of Helsinki.

Immunohistochemistry (IHC) staining
Briefly, paraffin-embedded tissue sections were baked at 60 °C for 2 h.After dewaxing and hydration, the tissue slices were covered with citric acid buffer for 15 min at 95 °C.Next, 3% hydrogen peroxide was added, followed by incubation with the primary antibody at 4 °C overnight.Slices were then re-warmed and washed.Slices were incubated with biotinlabeled sheep anti-rabbit IgG for 1.5 h, washed and incubated with the horseradish-labeled streptavidin.Then, tissue slices were stained with diaminobenzidine and were re-dyed with Hematoxylin for 5 min, covered with 1% hydrochloric acid alcohol (75%) solution and rinsed.Finally, slices were dehydrated, sealed, and scanned.

Tissue immunofluorescence staining
Briefly, paraffin-embedded tissue sections were baked at 60 °C for 2 h.After dewaxing and hydration, the tissue slices were covered with citric acid buffer for 15 min at 95 °C and were washed with PBS for three times.Goat serum was used to block the tissue slices for 20 min at 37 °C.The tissue slices were thereafter incubated with the described TUNEL and DAPI fluorescence dyes, washed and visualized under a confocal microscope (ZEISS).

Cells
The detailed protocols for culturing primary human NPC cells were described early [25,26].The primary cells, pNPC-1, pNPC-2, pNPC-3, and pNPC-4, were derived four different patients.The primary human nasal epithelial cells (HNEpC) from two donors (pHNEpC-1 and pHNEpC-2) were provided by Dr. Chen at Jiangsu University [26] and were reported in our previous study [25].All the primary human cells were verified routinely.The protocols were approved by the Ethics Committee of Soochow University and were in according with Declaration of Helsinki.All patients provided written-informed consent.

CRISPR-Cas9-induced gene knockout (KO)
The primary NPC cells were first infected with the lentivirus encoding the Cas9-expressing construct (from Dr. Cao), and stable cells formed after selection using puromycin.The Cas9-expresing NPC cells were then infected with lentivirus with the CRISPR-Gαi1-KO construct [containing small guide RNA (sgRNA) against human GNAI1] (from Dr. Cao [17]) and stable cells were established after six passages of puromycin selection.The stable cells were then distributed into 96-well plates and Gαi1 KO was verified using PCR.At last, the single stable Gαi1 KO primary NPC cells were formed and these cells were named as "Gαi1 KO" cells.ZNF384 KO in NPC cells was through the same procedure using the verified sgRNA from Genechem.For the control treatment, the Cas9-expresing primary human NPC cells were infected with the CRISPR-KO control construct ("Cas9-C").

Akt1 mutation
The lentiviral particles containing the S473D constitutively active mutant Akt1 (caAkt1) were provided by Dr. Chen [28] and were added to cultured NPC cells.Stable caAkt1-expressing cells were established with selection by puromycin.Control cells were infected with lentiviral particles with empty vector ("Vector").
siRNA Verified siRNAs against different transcription factors (ZNF460, ZNF384, EWSR1-FLI1, ZNF680, and EIf5) were provided by Genechem (Shanghai, China) and each (at the concentration of 200 nM) was individually transfected to NPC cells using Lipofectamine 3000.The transfection was repeated once at 24 h and stopped at 48 h.At least 70% reduction of targeted mRNA was achieved by each utilized siRNA.The control cells were transfected with non-sense control siRNA (siC).

Chromatin immunoprecipitation (ChIP) assay
As described [29], fresh tissue lysates or total cellular lysates were homogenized using a homogenizer [30] and were diluted in ChIP dilution buffer (from Dr. Cao [29]).Lysates were further immunoprecipitated with an anti-ZNF384 antibody and ZNF384-associated DNA was eluted by protein A/ G agarose (Santa Cruz Biotech) containing NaCl.The proposed Gαi1 promoter in the JASPAR database was thereafter tested by quantitative PCR (qPCR).

Other assays
Western blotting, quantitative real-time PCR (qRT-PCR), cell viability CCK-8 assay, nuclear EdU staining of cell proliferation, "Transwell" in vitro migration, the Caspase-3 activity assay, JC-1 assaying of mitochondrial depolarization, Histone DNA ELISA and nuclear TUNEL staining of cell apoptosis were described in detail in our previous study [25].

Xenograft studies
The nude mice were reported previously [25].At six million cells per mouse, the primary human NPC cells, pNPC-1, were subcutaneously (s.c.) injected into the flanks of the nude mice and pNPC-1 xenografts were established after three weeks.The xenograft-bearing mice were then assigned into two groups randomly with ten mice in each group.The mice were then subject to intratumoral injection of the designated adenoassociated virus (aav).Virus was injected twice (48 h between intervals).The protocols were approved by the Ethics Committee and Institute Animal Ethics Review Board of Soochow University.

Statistical analyses
The numerical data in this study were distributed normally and were always shown as mean ± standard deviation (SD).The Student's t-test (Excel 2013) was utilized when comparing two groups.Otherwise, one-way ANOVA plus a Scheffe' and Tukey Test (SPSS 23.0) were utilized for multiple groups' comparison.P < 0.05 indicates significant difference.

Gαi1 overexpression in NPC tissues
We first searched The Cancer Genome Atlas (TCGA) database and Gαi1 (GNAI1) expression data in NPC tissues was retrieved.As shown, the number of Gαi1 mRNA transcripts in NPC tissues ("Tumor") was significantly higher than that in normal nasopharynx epithelial tissues ("Normal") (Fig. 1A).The subgroup analyses further demonstrated that Gαi1 overexpression in human NPC tissues correlated with higher T-stages (Fig. 1B).Moreover, Gαi1 overexpression in NPC tumor tissues was correlated with poor overall survival of a large portion of NPC patients, including patients with T3-stage NPCs (Fig. 1C), N2-N3 NPCs (Fig. 1D) and pathological stage III-IV NPCs (Fig. 1E).

Gαi1 upregulation in NPC tissues of local patients and in different primary NPC cells
Next we tested expression of Gαi1 in NPC tissues of locallyadministrated patients.We obtained NPC tumor tissues ("T") and matched adjacent normal nasopharynx epithelial tissues ("N") from twenty (n = 20) primary NPC patients.All patients were administrated at authors' institution [25].As shown, Gαi1 mRNA expression in tumor tissues was over three folds higher than that in adjacent normal tissues (Fig. 2A).Western blotting assay data in Fig. 2B confirmed upregulation of Gαi1 protein in NPC tumor tissues of six representative patients ("Patient-1#" to "Patient-6#").The quantified results of all twenty sets of blotting data demonstrated that Gαi1 protein upregulation in NPC tissues was significant (P < 0.05 vs. "N" tissues) (Fig. 2C).

Gαi1 silencing inhibits viability, proliferation, cell cycle progression, and migration in primary NPC cells
To understand the possible function role of Gαi1 in NPC cells, the shRNA strategy was utilized to knockdown Gαi1.Specifically, to the pNPC-1 primary NPC cells (reported previously [25]) Gαi1 shRNA-packed lentivirus was added.Stable cells were thereafter established after puromycin treatment.Two different shRNAs against Gαi1 [20], Gαi1-shRNA-s1 and Gαi1-shRNA-s2 (with nonoverlapping sequences), were utilized and each resulted in substantial silencing of Gαi1 mRNA (Fig. 3A) and protein Fig. 1 Gαi1 overexpression in NPC tissues.The Cancer Genome Atlas (TCGA) database shows Gαi1 mRNA transcripts in NPC tissues ("Tumor") and normal nasopharynx epithelial tissues ("Normal") (A).Subgroup analyses of Gαi1 mRNA expression in NPC tissues with the described pathologic stages (B).The Kaplan-Meier Survival analyses of Gαi1-low and Gαi1-high NPC patients with the described pathological grades (C-E)."TPM" stands for "Transcripts Per Million" ."HR" stands for "Hazard Ratio" .The numerical values were mean ± standard deviation (SD).*P < 0.05.***P < 0.01.
To study whether Gαi1 exert similar activity in other primary NPC cells, the Gαi1-shRNA-s2-expressing lentivirus was added to primary human NPC cells that were derived from other primary patients, including pNPC-2, pNPC-3, and pNPC-4 (see Fig. 2).Puromycin was again employed to select stable cells.As shown, Gαi1 mRNA expression was downregulated in the primary NPC cells with Gαi1-shRNA-s2 (Fig. 3G).Similar to the functional results in pNPC-1 cells, knocking down of Gαi1 by Gαi1-shRNA-s2 decreased cell viability (Fig. 3H), suppressed cell proliferation (EdU incorporation, Fig. 3I) and inhibited cell migration (Fig. 3J) in these primary NPC cells.
We also tested the potential effect of Gαi1 silencing in nasal epithelial cells.To this aim, pHNEpC-1 and pHNEpC-2 cells, the primary nasal epithelial cells from two healthy donors (see Fig. 2), were infected with Gαi1-shRNA-s2-expressing lentivirus and were then treated with puromycin to establish stable cells.The applied shRNA resulted in substantial Gαi1 mRNA silencing (Fig. 3K).Unlike the functional consequence in NPC cells, shRNAinduced silencing of Gαi1 failed to inhibit viability (CCK-8 OD, Fig. 3L) and proliferation (EdU incorporation, Fig. 3M) in the HNEpCs.

Gαi1 knockout results in anti-cancer activity in primary NPC cells
To exclude the possible off-target effect of the applied shRNAs and to completely knockout (KO) Gαi1, the CRSIPR/Cas9 strategy was employed.Specifically, to the Cas-9-expressing pNPC-1 cells, the CRISPR-Gαi1-KO construct-expressing lentivirus was added.Single stable cells were then formed via puromycin selection and Gαi1 KO screening and these cells were named as Gαi1-KO cells.As shown, the protein expression of Gαi1 was depleted in Gαi1-Fig.2 Gαi1 upregulation in NPC tissues of local patients and in different primary NPC cells.Gαi1 mRNA (A) and protein (B, C) expression in the described NPC tumor tissues ("T", n = 20) and matched adjacent normal nasopharynx epithelial tissues ("N", n = 20) was shown, with results quantified.Gαi1 mRNA (D) and protein (E) expression in the described primary human NPC cells and primary human nasal epithelial cells ("pHNEpC-1" and "pHNEpC-2") was shown.The numerical values were mean ± standard deviation (SD, n = 5).Scale Bar = 100 μm.* P < 0.05 vs. "N" tissues or "pHNEpC-1" cells.

Ectopic Gαi1 overexpression causes cancer-promoting activity in primary NPC cells
The results above clearly showed that Gαi1 knockdown or KO resulted in significant anti-cancer activity in primary human NPC cells.We next hypothesized that ectopic Gαi1 overexpression might exert opposite functions and could promote NPC cell growth.Therefore, the lentivirus with the Gαi1-expressing construct ("OE-Gαi1", from Dr. Cao [13][14][15][16]20]) were added to pNPC-1 cells.After   3 Gαi1 silencing inhibits viability, proliferation, cell cycle progression and migration in primary NPC cells.pNPC-1 cells with the designated Gαi1 shRNA (Gαi1-shRNA-s1 and Gαi1-shRNA-s2, representing two different sequences) or scramble non-sense control shRNA ("shC"), were cultured and expression of Gαi1/2/3 (both mRNA and protein) was tested (A,B).The exact same number of the above cells were cultivated for designated hours, cellular functions, including cell proliferation (EdU-incorporated nuclei percentage, C), viability (CCK-8 OD, D), cell cycle progression (E) and in vitro cell migration ("Transwell" assays, F) were measured, with results quantified.Other stable primary NPC cells (pNPC-2, pNPC-3 and pNPC-4, derived from three primary patients) or the primary human nasal epithelial cells (pHNEpC-1 and pHNEpC-2, derived from two donors), expressing shC or Gαi1-shRNA-s2 were formed, with Gαi1 mRNA expression tested (G, K).The exact same number of the above cells were cultivated for designated hours, cell viability (H, L), proliferation (I, M) and migration (J) were examined, with results quantified.The numerical values were mean ± standard deviation (SD, n = 5)."pare" indicates the parental control cells.* P < 0.05 vs. "shC" cells."N.S." stands for non-statistical difference (P > 0.05).Experiments in this figure were repeated five times, with similar results obtained.Scale Bar = 100 μm.Fig. 5 Gαi1 KO results in anti-cancer activity in primary human NPC cells.pNPC-1 cells with the Cas9-expressing construct plus the CRISPR-Gαi1-KO construct ("Gαi1-KO") were established.Control cells were with the Cas9-expressing construct plus the CRISPR-KO control construct ("Cas9-C").Expression of Gαi1/2/3 protein was tested, with results quantified (A, B).The exact same number of above cells were cultivated for designated hours, cell proliferation (EdU-incorporated nuclei percentage, C), in vitro cell migration ("Transwell" assays, D) as well as cell apoptosis (TUNEL assays, E) were tested, with results quantified.The numerical values were mean ± standard deviation (SD, n = 5).*P < 0.05 vs. "Cas9-C" cells."N.S." stands for non-statistical difference (P > 0.05).Experiments in this figure were repeated five times, with similar results obtained.Scale Bar = 100 μm.Fig. 6 Ectopic Gαi1 overexpression causes cancer-promoting activity in primary NPC cells.pNPC-1 cells with the Gαi1-expressing construct (OE-Gαi1-L1 and OE-Gαi1-L2, representing two stable cell selections) or the empty vector ("Vec") were formed and expression of Gαi1/2/3 (both mRNA and protein) was tested (A, B).The exact same number of the above cells were cultivated for designated hours, cellular functions, including cell proliferation (EdU-incorporated nuclei percentage, C) and viability (CCK-8 OD, D), and in vitro cell migration ("Transwell" assays, E) were measured, with results quantified.Other primary NPC cells (pNPC-2, pNPC-3 and pNPC-4) with the Gαi1-expressing construct (OE-Gαi1) or the empty vector ("Vec") were formed as well, and expression of Gαi1 mRNA measured (F); Cells were cultivated for indicated times, cell proliferation (G) and migration (H) were tested similarly, with results quantified.The numerical values were mean ± standard deviation (SD, n = 5).*P< 0.05 vs. "Vec" cells."N.S." stands for non-statistical difference (P > 0.05).Experiments in this figure were repeated five times, with similar results obtained.Scale Bar = 100 μm.

ZNF384 is an important transcription factor of Gαi1 in NPC cells
Considering that both mRNA and protein expression of Gαi1 was increased in NPC tissues and cells (Figs. 1, 2), we hypothesized that there could be a transcriptional mechanism of Gαi1 upregulation in NPC tissues.Few studies reported the definite transcription factors of Gαi1 in human cells, we therefore searched JASPAR transcription factor database [33].Five transcription factors with the highest possible binding affinity to Gαi1 were identified, including ZNF460, ZNF384, EWSR1-FLI1, ZNF680 and EIf5 (Fig. 8A).We next designed siRNAs targeting each of the five different transcription factors.These siRNAs were individually transfected to pNPC-1 cells and their efficiency on Gαi1 mRNA expression was analyzed.As shown, only ZNF460 siRNA and ZNF384 siRNA resulted in significant Gαi1 silencing in pNPC-1 cells (Fig. 8B).Silencing of other transcription factors was ineffective on Gαi1 expression in pNPC-1 cells (Fig. 8B).ZNF384 siRNA was more potent than ZNF460 siRNA in downregulating Gαi1 in pNPC-1 cells (Fig. 8B).
To further support our hypothesis, the lentivirus encoding the ZNF384-expressing construct was added to pNPC-1 cells.Stable cells, "oeZNF384", were formed following puromycin-mediated selection.These cells showed substantial ZNF384 mRNA (Fig. 8E) and protein (Fig. 8F) upregulation.With ZNF384 overexpression, Gαi1 mRNA (Fig. 8E) and protein (Fig. 8F) expression was upregulated as well.Importantly, ChIP assay results showed that the binding between ZNF384 protein and the proposed Gαi1 promoter region (in the JASPAR database) was significantly increased in NPC tissues of local patients (Fig. 8G).Moreover, the binding between the two was significantly increased in different primary NPC cells (pNPC-1, pNPC-2, pNPC-3, and pNPC-4) (Fig. 8H).ZNF384-Gαi1 promoter binding affinity was relatively low in normal nasopharynx epithelial tissues ("N") (Fig. 8G) and also in HNEpC cells (Fig. 8H).These results supported that the increased binding between the ZNF384 and the Gαi1 promoter region could be a primary mechanism of Gαi1 upregulation in NPC.
supported that shGαi1-aav injection inhibited NPC xenograft growth in nude mice.
To further corroborate the role of Gαi1 in NPC cell growth in vivo, we conducted subcutaneous injections of Gαi1-KO pNPC-1 cells and control Cas9-C pNPC-1 cells into the flanks of nude mice.After an eight-week period, we isolated and measured the resulting xenografts.Gαi1-KO pNPC1 xenografts exhibited significantly smaller sizes and lower weights compared to the Cas9-C pNPC1 xenografts (Fig. S2A, B).The body weights of the mice in Fig. 9 Gαi1 depletion inhibits NPC xenograft growth in nude mice.pNPC-1 xenograft-bearing nude mice were intratumorally injected Gαi1-shRNA-s2-expressing adeno-associated virus ("shGαi1-aav") or scramble control shRNA adeno-associated virus ("shC-aav"), the volumes of pNPC-1 xenografts (A) and animal body weights (D) were recorded every six days.The estimated daily pNPC-1 xenograft growth was calculated (B); At day-42, all pNPC-1 xenografts were isolated and weights (C).The tissue lysates of the described pNPC-1 xenografts were obtained and expression of the described mRNAs and proteins was tested (E, F, G, H, and K), with Caspase-3 activity tested as well (J).Alternatively, the described pNPC-1 xenograft slices were subjected to IHC staining p-Akt (Ser-473) (I), or subjected to immuno-fluorescence staining of TUNEL/DAPI (L).Values were mean ± standard deviation (SD).In (A-D), ten mice were in each group (n = 10).For (E-L), five random tissue pieces in each xenograft were tested (n = 5).*P < 0.05 vs. "shC-aav" group.Scale bar = 100 μm. the two groups were not significantly different (Fig. S2C).In addition, Gαi1 protein expression was substantially reduced in Gαi1-KO pNPC1 xenografts (Fig. S2D), while the expression levels of Gαi2 and Gαi3 remained unchanged (Fig. S2D).These findings provide strong support for the notion that Gαi1 KO effectively inhibited the growth of pNPC1 xenografts in nude mice.

DISCUSSION
NPC is considered as a poor-prognosis malignancy, mainly due to the lack of characteristic clinical symptoms [1,8,9].The failure of current treatment methods is also one main reason for patients' poor survival [1,8,9].To improve NPC patients' prognosis, exploring novel signaling targets is urgently needed and an in-depth understanding of molecular events associated with pathogenesis and progression of the cancer is critical [1,8,9].Discovering novel therapeutic targets is also essential for developing effective targeted therapies [1,8,9].
The results of the present study supported a key pro-cancerous role of Gαi1 in NPC.TCGA results demonstrate that Gαi1 transcripts' number in NPC tissues is significantly higher than that in the nasal epithelial tissues.Its overexpression correlates with poor survival of a large number of NPC patients.Moreover, Gαi1 mRNA and protein expression is significantly elevated in NPC tissues of local patients and in primary NPC cells.Whereas its expression is relatively low in cancersurrounding normal tissues and in nasal epithelial cells.Genetic silencing or KO of Gαi1 substantially suppressed viability, proliferation, cell cycle arrest, and migration, and induced apoptosis activation in primary NPC cells.Contrarily, ectopic Gαi1 expression in primary NPC cells exerted pro-tumorigenic activity and augmented cell proliferation and migration.In vivo studies showed that intratumoral injection of Gαi1-shRNA-expressing aav impeded subcutaneous pNPC-1 xenograft growth in nude mice.Gαi1 KO also effectively inhibited the growth of pNPC1 xenografts in nude mice.Therefore, overexpressed Gαi1 is a novel and promising therapeutic target of NPC.
Here we found that Gαi1 is important for Akt-mTOR activation in NPC cells.Akt-S6K phosphorylation was downregulated with Gαi1 shRNA or KO in primary NPC cells, but was augmented after Gαi1 overexpression.In Gαi1-silenced primary NPC cells, caAkt1 restored Akt-S6K phosphorylation and largely ameliorated Gαi1 shRNA-induced proliferation inhibition, migration reduction and apoptosis.Akt-mTOR inactivation was also detected in Gαi1silenced NPC-1 xenograft tissues.Thus, mediating Akt-mTOR activation could be a primary mechanism of Gαi1-driven NPC cell growth in vitro and in vivo (see Fig. 10).
Studies have explored a possible cancer-promoting activity of the transcription factor ZNF384 in different cancers.Meng et al. showed that ZNF384 is upregulated in breast cancer and is important for cancer cell metastasis [36].ZNF384 is possible transcription factor for the histone acetyltransferase HBO1, the latter promoted osteosarcoma cell growth in vitro and in vivo [27].Yan et al. demonstrated that ZNF384 overexpression promoted colorectal cancer metastasis by upregulating MMP2 [37].
One important discovery of this study is that ZNF384 could be a primary transcription factor of Gαi1 in NPC cells (see Fig. 10).In primary NPC cells, ZNF384 shRNA or KO decreased Gαi1 mRNA and protein expression, whereas ZNF384 overexpression upregulated Gαi1 expression.Importantly, there was an increased binding between ZNF384 protein and the proposed Gαi1 promoter in both human NPC tissues and cells.Increased binding between the two could be the primary mechanism of Gαi1 upregulation in NPC (see Fig. 10).

CONCLUSION
Overexpressed Gαi1 exerts pro-tumorigenic activity in NPC possibly by promoting Akt-mTOR activation.Fig. 10 The proposed signaling cartoon of this study.The increased binding between the transcription factor ZNF384 and the Gαi1 promoter causes Gαi1 overexpression, which exerts pro-tumorigenic activity in nasopharyngeal carcinoma (NPC) possibly by promoting Akt-mTOR activation.