Tumor hypoxia impairs NK cell cytotoxicity through SHP-1-mediated attenuation of STAT3 and ERK signaling pathways

Natural killer (NK) cells are innate immune effectors with potent anti-tumor activity. Nonetheless, tumor cells have the ability to create an immunosuppressive microenvironment, thereby escaping from immune surveillance. Although accumulating evidence indicates that microenvironmental hypoxia plays an important role in favoring tumor development and immune evasion, it is still unclear how hypoxia directly impairs NK cell anti-tumor activity. In this study, we confirmed that hypoxic NK cells show significantly lower cytotoxicity against tumor cells. Consistent with this, we also found that the reduction in NK cell cytotoxicity resulting from hypoxia is related to the lower expression of granzyme B, IFN-γ, degranulation marker CD107a, as well as killer activation receptors including NKp30, NKp46, and NKG2D on NK cells. More importantly, we further demonstrated that a reduction in the phosphorylation levels of ERK and STAT3 secondary to hypoxia are tightly associated with the attenuated NK cell cytotoxicity. Focusing on the mechanism responsible for reducing phosphorylation levels of ERK and STAT3, we revealed that the activation of protein tyrosine phosphatase SHP-1 (src homology region 2 domain-containing phosphatase-1) following hypoxia may play an essential role in this process. When knocking down SHP-1 or blocking its activity using a specific inhibitor TPI-1, we were able to partially restore NK cell cytotoxicity under hypoxia. Taken together, we demonstrated that hypoxia can impair NK cell cytotoxicity by decreasing the phosphorylation levels of ERK and STAT3 in a SHP-1 dependent manner. Therefore, targeting SHP-1 could provide an approach to enhance NK cell-based tumor immunotherapy. that SHP-1 plays an important role in hypoxia-impaired NK cell cytotoxicity. Our finding supports that targeting SHP-1 may provide an important approach for improving NK cell-based tumor immunotherapy.


Background
Natural killer (NK) cells are cytotoxic innate lymphoid cells involved in the immune suggests that the tumor microenvironment is involved in tumor evasion from NK cellmediated killing through cellular and non-cellular mechanisms [12]. In this regard, a clear demonstration of how the tumor microenvironment can affect the phenotype and function of NK cells has not yet been fully elucidated.
Hypoxia is a prominent feature of solid tumors and one of the hallmarks of the tumor microenvironment [13]. More recently, it has also been demonstrated that hypoxia is a prevalent feature of the bone marrow microenvironment in acute myeloid leukemia and multiple myeloma, but not the normal bone marrow [14][15][16][17]. It is well established that tumor hypoxia is an adverse prognostic and predictive factor, and is involved in fostering many aspects of tumor development. Of note, emerging evidence also indicates that tumor hypoxia is a key factor regulating the loss of immune reactivity either by decreasing tumor cell sensitivity to cytotoxic immune effectors or by promoting immunosuppressive mechanisms [18,19].
In this study, focusing on the roles of hypoxia in NK cell-mediated immune surveillance, we investigated the resultant effects of hypoxia on NK cell cytotoxicity machinery and the underlying mechanisms. We demonstrated that hypoxia can directly impair NK cell cytolytic function by decreasing the expression of the killer activation receptors (KARs) on the NK cell surface. More importantly, we revealed that impaired NK cell cytotoxicity secondary to hypoxia was mediated by increased expression of phosphatase SHP-1, which can catalyze dephosphorylation at the tyrosine sites of ERK and STAT3, thereby attenuating NK cell activation signaling. Our findings demonstrate a previously unknown role for hypoxia in dysfunction of NK cell-mediated tumor surveillance, and suggest that SHP-1 may represent a novel target for preserving NK cell function in tumor patients and improving NK cell-based immunotherapy. Methods 7 CO 2 . Normoxic or hypoxic cell culture conditions were obtained by culturing cells in a sealed incubator flushed with the mixture of 20% O 2 , 5% CO 2 , and 75% N 2 , or the mixture of 1% O 2 , 5% CO 2 , and 94% N 2 , respectively.

Flow cytometry
The expression of NK cell cytotoxicity effector molecules and KARs was analyzed by flow cytometry. For membrane staining, 5×10 5 cells were collected and washed with staining buffer (PBS containing 0.1% NaN 3 and 0.1% BSA) three times. The cells were then incubated for 30 min on ice, according to the instructions provided with the respective antibodies. After washing 3 times with cell stain buffer, the cells were resuspended in 300 µL staining buffer in the presence of Sytox Green or 7-AAD, which were used to gate out dead cells. Acquisition of 10,000 cells per reaction was performed using a CytoFLEX Cytometer (Beckman Coulter Life Sciences). Data were analyzed with Flowjo v7.6.2 (Tree Star). For intracellular staining, 5×10 5 cells were collected and fixed with 1 mL 1% paraformaldehyde in PBS for 15 minutes at room temperature. After washing 3 times with cell stain buffer, the fixed cells were then resuspended in 2 mL permeabilization buffer (0.1% saponin in cell staining buffer) and incubated for 30 min at room temperature. The cells were collected again by centrifugation, stained with antibody at an optimal working concentration in permeabilization buffer for 15 min on ice. After washing three times with permeabilization buffer, the cells were resuspended cells in 300 uL cell staining buffer for final flow cytometric analysis.

CD107a degranulation assay
Degranulation of cytotoxic contents from NK cells was measured by analysis of the degranulation marker CD107a by flow cytometry. Briefly, NK cells and tumor cells were individually pre-incubated for 14-16 h at 20% or 0% O 2 and after that combined in different effector-to-target (E:T) ratios at either 20% or 0% O 2 for 4 h. APC labeled anti-CD107a was added to the wells within 5-10 minutes after combining NK and tumor cells.
As a positive control, PMA (100 ng/mL, Sigma-Aldrich) and ionomycin (1 mg/mL; Sigma-Aldrich) were added to the NK cells during the 4 h degranulation assay.

Flow cytometric cytotoxicity assay
NK cells and tumor cells were individually pre-incubated for 24 h at 20% or 1% O 2 first. NK and target cells were then coincubated under comparable conditions in different E:T ratios in a 24 well plate. After the 4 h incubation at 37°C and 5% CO2, samples were labeled with CD2-APC to distinguish effector from target cells. Cell death was detected with Annexin V-FITC and Sytox Green and analyzed on a flow cytometer. A minimum of 10,000 target events were collected per sample and the results were analyzed using Flowjo v7.6.2.

Western blotting
For western blotting, treated and untreated KHYG-1 and NK92 cells were lysed in buffer containing 50 mM Tris, 150 mM NaCl, 1% Triton X-100, 1% sodium deoxycholate, 0.1% SDS and protease inhibitors on ice for 30 min. Lysates were centrifuged at 12,000 rpm for 15 min and supernatants were collected. Protein concentration was determined by the BCA protein assay kit (HEART Biotech, #WB003). Equal amounts of protein were loaded and separated on sodium dodecyl sulfate-polyacrylamide gel electrophoresis gel, and transferred onto a PVDF membrane (Millipore, #IPVH00010). After blocking for 1 h with 5% non-fat milk in PBS with 0.1% Tween-20 at room temperature, the membrane was incubated with primary antibody at 4℃ overnight. Immunoblots were visualized using HRPconjugated secondary antibodies and the ECL Western Blot Detection kit (Phygene Life Sciences, #PH0353).

siRNA-mediated gene silencing in NK cells
KHYG-1 cells were kept in the above-mentioned condition. Before siRNA transfection cells were washed in pre-warmed Opti-MEM medium (Life Technologies, Carlsbad, CA, USA) and resuspended in the same medium. 10 6 cells were electroporated with 2 μg of siRNA in 100 μL Opti-MEM medium in 0.2 cm cuvettes with an electroporator CUY21EDIT (BEX Co. Ltd, Japan). The electroporation program was set as follows: PpV=200V, Pp on 10 ms, Pp off 10 ms, PdV=25V, Pd on 50 ms, Pd off 50 ms; Pd N=10, capacity=940 μF, and exponential decay wave type. Following electroporation, cells were resuspended in 2 mL complete media. 16-24 h after electroporation, the cells were used for western blotting or killing assay. Transfection efficiency and viability were analyzed by flow cytometry 2-6 h after electroporation, to quantitatively measure the expression of fluorescein isothiocyanate (FITC)-labeled siRNA and 7-AAD. SHP-1 mRNA was silenced by using a gene-specific siRNA pool from GenePharma (see Supplementary Table 1).

Statistical analysis
Statistical analyses were performed using the Prism software package 5.0 (GraphPad Software, San Diego, CA, USA). Data are expressed as the mean ± SEM of at least three independent experiments. Statistical significance was evaluated by two-tailed paired Student's t-test. A p < 0.05 ( * ), < 0.01( ** ), or < 0.001( *** ) was considered statistically significant.

Hypoxic NK cells show decreased cytotoxicity against tumor cells
We first investigated whether hypoxia impairs NK cell-mediated lysis of tumor cells. To this end, two NK cell lines KHYG-1 and NK92 were cultured in the presence of IL-2 under hypoxic (1% O 2 ) or normoxic (20% O 2 ) conditions for 24 h, and subsequently incubated with the tumor cell lines K562 or MM.1S at different E:T ratios for 4 h to evaluate the cytotoxicity by flow cytometry. As shown in Fig. 1a and b, it revealed that the NK cell cytotoxicity was significantly decreased in 1% compared to 20% O 2 in both tumor cell lines tested. Meanwhile, we observed a marked accumulation of the hypoxia marker HIF-1α in hypoxic NK cells, whereas it was weak in normoxic NK cells monitored by western blotting (Fig. 1c). Moreover, we excluded the possibility that the decreased cytotoxicity in hypoxia was caused by reduced NK cell viability, since we did not observe increased NK cell death by hypoxia (Fig. 1d).

Hypoxia decreases the expression of cytotoxic effectors and activating receptors on NK cells
To further explore how hypoxia reduces NK cell killing ability, we measured the expression level of cytotoxic effector granzyme B and perforin. As shown in Fig. 2a, hypoxiatreatment led to decreased secretion of both granzyme B and perforin. Additionally, we observed a reduced expression of the cytokine IFN-γ in hypoxic NK cells compared to normoxic condition (Fig. 2b). Importantly, CD107a, which is a degranulation marker of natural killer cell activity, was also diminished by hypoxia (Fig. 2c). Given that a range of receptors that can trigger cytolytic programs, as well as cytokine or chemokine secretion tightly regulates NK cell function, we next evaluated the effects of hypoxia on the expression of the main receptors capable of triggering cytolytic activity. Surface expression of the KARs including NKp46, NKp30, and NKG2D was measured by flow cytometry on both normoxic and hypoxic NK cells. As shown in Fig. 2d, it confirmed that hypoxia could decrease the expression of KARs on the NK cell surface.

Hypoxia attenuates ERK and STAT3-mediated NK activation
It is known that intracellular signals activating NK cell cytokine production and cytotoxic activity are propagated primarily through protein phosphorylation of ERK (extracellular signal-regulated kinase) and STAT3 (signal transducer and activator of transcription 3). Therefore, we further investigated whether hypoxia could affect the activation of ERK and STAT3, and revealed that hypoxia markedly diminished the phosphorylation level at the tyrosine sites of ERK and STAT3 in the two NK cell lines (Fig. 3a, b). To further validate the effects of the phosphorylation of ERK and STAT3 on the expression of KARs and NK cytotoxicity under hypoxic conditions, we used specific small molecule inhibitors U0126 and cryptotanshinone to block ERK and STAT3 signaling, respectively. As shown in Fig. 3c, inhibition of ERK and STAT3 significantly reduced the expression of NKp30 and NKG2D.
Importantly, we found that inhibition of ERK or STAT3 results in significantly impaired cytotoxicity against K562 cells (Fig. 3d).

Hypoxia-decreased phosphorylation level of STAT3 and ERK was mediated by the activation of protein tyrosine phosphatase SHP-1 other than SHP-2
Cell surface receptors harboring intracytoplasmic tyrosine-based activation motifs (ITAMs) or intracytoplasmic tyrosine-based inhibitory motifs (ITIMs) are often phosphorylated by Src family protein tyrosine kinase (PTK), which in turn creates docking sites for the protein tyrosine phosphatase SHP-1 and SHP-2. Recruitment and activation of the SHP-1 and/or SHP-2 have been demonstrated to be a dominant inhibitory mechanism to prevent the induction of the stimulatory signaling cascade [20,21]. In this regard, we further investigated whether phosphatase SHP-1 and SHP-2 are involved in the decrease of ERK and STAT3 phosphorylation by hypoxia. As shown in Fig. 4a and b, hypoxia triggers a significant increase of SHP-1 and SHP-2 phosphorylation. When using a specific SHP-1 inhibitor TPI-1, we observed it could reverse the decrease of the phosphorylation of both ERK and STAT3 (Fig. 4c). Moreover, we also observed that pretreatment with p-SHP1 inhibitor TPI-1 significantly could restore the NK cell cytotoxicity under hypoxia (Fig. 4d).
However, we did not observe the same effects when using a specific SHP-2 inhibitor SHP099, which had no effect on the phosphorylation levels of ERK and STAT3, as well as NK cell cytotoxicity (Fig. 5).

Knockdown of SHP-1 rescues of NK cell cytotoxicity in hypoxia
To further validate the role of SHP-1 in regulating NK cell cytotoxicity, we silenced the gene expression of SHP-1 in KHYG-1 cells and found that knockdown of SHP-1 could increase the phosphorylation level of ERK and STAT3 under hypoxia (Fig. 6a). More importantly, we also confirmed that NK cells with SHP-1 silencing showed enhanced cytotoxicity against K562 cells than control cells in hypoxic condition (Fig. 6b).

Discussion
In the current study, we investigated how hypoxia affects NK cell activity against tumor cells. We first confirmed that hypoxia could directly impair NK cell cytotoxicity against tumor cells, validated by the decreased expression of granzyme B, perforin, CD107a, and KARs. Additionally, we demonstrated that the inhibition of NK cytotoxicity induced by hypoxia is tightly related to a reduction in the phosphorylation level of ERK and STAT3.
More importantly, we further found that activation of the tyrosine phosphatase SHP-1 is responsible for the hypoxia-induced decrease in ERK and STAT3 phosphorylation, thereby controlling NK cytotoxicity (Fig. 6c). that is dependent on IL-2. When blocking ERK and STAT3 signaling, NK cell cytotoxicity was significantly reduced. Meanwhile, we also found that the expression of KARs on the NK cell surface was also decreased, suggesting an essential role of ERK and STAT3 signaling in regulating NK cytotoxicity. Of note, we also observed that hypoxia could inactivate the phosphorylation of ERK and STAT3, in line with the lower cytotoxicity of hypoxic NK cells.
With respect to hypoxia-mediated inactivation of ERK and STAT3, our findings further support that tyrosine phosphatase SHP-1 is a key player in this process. It is now accepted  41].Thirdly, the phosphatase activity of SHP-1 and SHP-2 is dependent on the C-terminal phosphorylation at different tyrosine or serine residues [38,[42][43][44][45]. On the contrary, while the importance of the SHP-1 and SHP-2, as well as the ERK and STAT3 signaling in NK cell function are well known, it is still unclear whether SHP-1 or SHP-2 can directly dephosphorylate ERK and STAT3. Herein, we demonstrated that hypoxia could activate both SHP-1 and SHP-2 by inducing phosphorylation, suggesting a possibility that SHP-1 and SHP-2 dephosphorylate ERK and STAT3. However, using specific blocking inhibitors of SHP-1 and SHP-2, we observed that targeting SHP-1 with TPI-1 alone could increase the phosphorylation of ERK and STAT3, and enhance NK cell cytotoxicity. However, SHP-2 inhibition had no effect on the phosphorylation level of ERK and STAT3. Furthermore, by silencing the gene expression of SHP-1 in KHYG-1 cells, we further confirmed the dephosphorylation effect on ERK and STAT3. Collectively, our findings show that hypoxia could impair NK cell cytotoxicity through decreasing the phosphorylation of ERK and STAT3 in a SHP-1 dependent manner.
Numerous studies have demonstrated that hypoxia could affect a large number of genes in many aspects of cells including proliferation, apoptosis, migration and metabolism, in multiple cell types. Although little is known about the effect of hypoxia on NK cells, a recent study based on comprehensive transcriptome analysis of hypoxic NK cells showed significant changes in genes coding for proinflammatory cytokines, chemokines, and chemokine-receptors [19]. Along these lines, it may suggest more complex mechanisms in hypoxia-mediated NK cell dysfunction against tumors. Therefore, further study is still needed to clarify more details in NK cell response to hypoxia.
In conclusion, we demonstrated that SHP-1 plays an important role in hypoxia-impaired NK cell cytotoxicity. Our finding supports that targeting SHP-1 may provide an important approach for improving NK cell-based tumor immunotherapy.