Unphosphorylated STAT3 in heterochromatin formation and tumor suppression in lung cancer

As has been previously reported, uSTAT3 is predominantly detected in the nucleus of various types of serum-starved, unstimulated cells (e.g., Fig. 1 in [12]). We examined the subcellular localization of endogenous STAT3 and HP1α in in serum starved cells using several lines of lung cancer cells including A549, H226, H441, H460, H520, as well as HeLa and HEK293T cells. When observed with a laser confocal microscope in thin optical sections, we found that endogenous STAT3 in all these cell lines can be detected in the nucleus as unevenly distributed particulates, some of which are colocalized with HP1α (Fig. 1a). Some of these cancer cell lines show very low or no detectable pSTAT3 under unstimulated conditions, as previously reported (e.g., A549) [44]. The pattern of nuclear STAT3 distribution in these cells does not correlate with its phosphorylation status, consistent with previous reports that STAT3 is prominently detected in the nucleus regardless of its phosphorylation status [12, 13, 16]. In addition, we found that endogenous STAT3 and HP1α co-immunoprecipitate (Fig. 1b), consistent with the idea that they may physically interact.

To investigate if the partially colocalized STAT3 and HP1α in the nucleus indeed physically interact, we employed fluorescence resonance energy transfer (FRET), which is based on transfer of energy from one fluorescent molecule (donor) to another (acceptor) only when the two are in close proximity (< 10 nm apart), which occurs when the two molecules directly interact [45, 46]. In order to investigate the proximity of unmodified endogenous STAT3 and HP1α in cells, we used fluorophore-conjugated secondary antibodies and FRET [46, 47] in fixed cells. We found high FRET efficiency from anti-HP1α-Alexa488 (donor) to anti-STAT3-Alexa546 (acceptor) only in discrete regions of the nucleus (Fig. 1c), suggesting that STAT3 and HP1α physically interact in these regions of the nucleus. Importantly, when treated with the cytokine IL6, which activates STAT3 by phosphorylation, increasing pSTAT3 and decreasing uSTAT3, we found that FRET efficiency was significantly reduced, suggesting a loss of STAT3-HP1α interaction (Fig. 1d, e). The FRET results suggest that the colocalized uSTAT3 and HP1α physically interacts.

To investigate if by binding to HP1α, uSTAT3 is able to promote heterochromatin formation, we carried out the following experiments. First, using Western blotting, we found that expressing an unphosphorylatable mutant STAT3 (STAT3^Y705F), representing uSTAT3, causes a notable increase in the levels of the heterochromatin mark, H3K9me3, whereas depleting endogenous STAT3 leads to moderately reduced H3K9me3 levels (Fig. 2a). Second, by using immunostaining, we found that expressed STAT3^Y705F increased H3K9me3 levels, whereas expressing the non-HP1-interacting STAT3^V462A had no effects (Fig. 2b). Third, we employed a FRET-based heterochromatin sensor to investigate the effects of HP1α and uSTAT3 on global heterochromatin levels. The heterochromatin FRET sensor consists of peptides of H3 and HP1 fused to YFP and CFP, respectively, which has been previously used as a robust readout of H3K9me3 levels when expressed by transfection [48‐50]. We found that, as expected, altering HP1α levels, by overexpressing HP1α or expressing a small hairpin (sh) HP1α RNA, affect global heterochromatin levels, as indicated by the FRET sensor (Fig. 2c, d). Importantly, we found that expressing STAT3^Y705F increased, whereas knocking down STAT3 decreased global heterochromatin levels in A549 cells (Fig. 2c, d). A hallmark of non-canonical STAT function is that expressing unphosphorylatable mutant STAT has the opposite effect to knocking down STAT, whereas in the canonical pathway, they have the same effect. These results suggest that STAT3 also has non-canonical function in regulating heterochromatin globally.

To evaluate the changes in cellular phenotypes from altering heterochromatin, we carried out the following two experiments. First, we investigated cellular senescence, a state of irreversible cell cycle arrest associated with formation of specialized facultative heterochromatin, which is hypothesized to repress proliferation-promoting genes and suppress tumor development [35, 51]. Senescent cells are commonly detected by histochemical staining for senescence-associated β-galactosidase (SA-βgal) [52]. Using stable A549 cells, we found that expressing uSTAT3 or HP1α led to more senescent cells, whereas knocking STAT3 or HP1α resulted in fewer senescent cells (Fig. 3a, b). Second, heterochromatin is known to repress transcription from satellite repeats, whose derepression is found in many cancers [34]. Using quantitative real-time polymerase chain reaction (qRT-PCR), we found that uSTAT3 and HP1α repressed major satellite transcripts, whereas knocking down endogenous STAT3 or HP1α caused a 4-fold increase in major satellite transcripts (Fig. 3c).

To investigate whether uSTAT3 and HP1α play roles in cancer growth, we assessed the effects of overexpressing these genes on lung cancer cell growth using stably transfected A549 lung cancer cells. We first used a clonogenic assay to assess the ability of single lung cancer cells to grow into colonies [54]. We found that, compared with control, A549 cells stably expressing STAT3-shRNA had more colonies, whereas cells expressing uSTAT3 (STAT3^Y705F) or HP1α had fewer colonies (Fig. 4a, b). We then used the soft-agar assay [55] to assess the effects of uSTAT3 and HP1α on anchorage-independent growth of lung cancer cells. We found that knocking down STAT3 or HP1α in A549 cells by expressing shRNAi constructs resulted in more and larger colonies, whereas overexpressing uSTAT3 or HP1α led to fewer and smaller colonies in soft agar (Fig. 4c, d). These results are consistent with the notion that uSTAT3 and HP1α suppress cancer cell growth by increasing heterochromatin levels. The discrepancy for HP1α-shRNA expression in the two different assays may involve differences in cell survival and will be further investigated.

Finally, we investigated the effects of uSTAT3 and HP1α in cancer cell growth using mouse xenografts. We established lung cancer A549 cell lines stably expressing HP1α, STAT3^WT, STAT3^Y705F, and sh RNAi constructs targeting STAT3 and HP1α, respectively, as well as controls. We implanted these cells into immunodeficient (nude) mice subcutaneously and measured the growth of these cells as tumors in mice during the 4 weeks period after implantation. As shown at 4 weeks after implantation, we found that A549 cells expressing either HP1α or STAT3^Y705F exhibited reduced capacity to grow as tumors, as indicated by the small tumor volumes at week four compared with A549 cells expressing vector control (Fig. 4e). Conversely, knocking down HP1α or STAT3 by stably expressing RNAi constructs HP1α-sh or STAT3-sh resulted in larger tumor volumes (Fig. 4e). These results are consistent with a previous report that knocking down STAT3 in A549 cells promotes the growth of these lung cancer cells as mouse xenografts [38]. Taken together, our results suggest that uSTAT3 and HP1α play roles in tumor suppression.

Human NSCLC cell lines A549, H226, H441, H460, and H520; and other human cancer cell lines HeLa and HEK293T were from American Type Culture Collection (ATCC, Manassas, VA 20110 USA).

Human NSCLC cell lines A549, H226, H441, H460, and H520; and other human cancer cell lines HeLa and HEK293T cells were maintained in RPMI medium (Gibco) supplemented with 10% (v/v) FBS and antibiotics at 37 °C and with 5% CO₂ in water-jacketed, humidified incubators. Transient transfection with plasmid DNA was done by using FuGENE 6 (Life Sciences, Inc) according to the manufacturer’s protocol. Cells stably transfected with the indicated cDNAs or shRNAs were selected as puromycin (5 μg/ml)-resistant colonies and several colonies were pooled. Cell culture and transfection procedure was approved by UCSD BUA R1347.

Plasmid DNA constructs with Myc-tagged human STAT3 and STAT3^Y705F were kindly provided by Dr. Pradipta Ghosh (UCSD). DNA construct for sh-HP1α was from Open BioSystems. The following DNA constructs were acquired from Addgene (Cambridge, MA): human HP1α (17652), sh-STAT3 (26596), and H3K9me3 FRET reporter (22866).

For immunofluorescence, transfected cells were fixed in 4% paraformaldehyde for 10 min, permeabilized in 0.2% Triton X-100 in phosphate-buffered saline (PBS) for 15 min, blocked with 5% bovine serum albumin (BSA) in PBS for 30 min, and incubated overnight at 4 °C in primary antibodies. Primary antibodies used in this study include rabbit anti-STAT3 (1:250; Santa Cruz, sc-482), mouse anti-HP1 (CBX5) (1:200; Life Sciences, 730,019), goat anti-c-Myc (1:500, Fisher, NB600–335). Slides were washed four times in PBS and then immunostained with Alexa Fluor® 546-conjugated secondary antibody (1:250; Molecular Probes) at 37 °C for 1 h.

To detect changes in histone methylation (H3K9me3) levels in living cells, cells were transfected with a CFP/YFP histone methylation FRET reporter construct [48]. Donor and acceptor bleed through was corrected using donor and acceptor only samples. FRET measurements were carried out using a Zeiss Axio Observer fluorescence microscope equipped with FRET setup and software.

To detect FRET in fixed cells, cells were fixed on a coverslip and immunostained with anti-HP1α-Alexa488 (donor) to anti-STAT3-Alexa546 (acceptor) primary and secondary antibody pairs, as described in [47]. Donor and acceptor bleed through was corrected using donor and acceptor only samples. FRET was detected and processed using a Leica confocal microscope with built-in FRET software.

Cells were plated on 24 well plates. Nearly confluent cells were fixed with 3.7% Paraformaldehyde in PBS. 500ul of β-gal staining solution (0.1% X-Gal, 5 mM 5 mM potassium ferrocyanide, 5 mM potassium ferricyanide, 150 mM Sodium chloride, and 2 mM Magnesium chloride in 40 mM citric acid/sodium phosphate solution, pH 6.0) was added and plates were incubated at 37 °C overnight. After removal of staining solution, 70% glycerol was added and visualized with 20X bright field microscope. Senescence cells show blue staining in the cytosol [53].

Major Sattellite Forward: AGGGAATGTCTTCCCATAAAAACT.

Major Satellite Reverse: GTCTACCTTTTATTTCAATTCCCG.

Gapdh forward: CATGGGTGTGAACCATGAGA.

Gapdh reverse: CAGTGATGGCATGGACTGTG.

Stable A549 cell lines with the indicated transgene were maintained in RPMI medium supplemented with 10% fetal bovine serum. Cells grown to 60% confluency were resuspended in 0.4% Noble agar (Sigma, St. Louis, MO) in RPMI, and were seeded at a density of 1.5 × 10⁵ cells/well in 6-well culture plates on top of a 2 ml underlayer composed of 0.8% agarose in RPMI. Media were refreshed twice per week for 3 weeks, and then were stained with p-iodonitrotetrazolium violet (Sigma, St. Louis, MO) and photographed on an inverted compound microscope with phase contrast optics.

Stable A549 cell lines with the indicated transgene were maintained in RPMI medium supplemented with 10% fetal bovine serum. Cells grown to 60% confluency were harvested and diluted in fresh media, and seeded at 100 cells/plate in 6-well plates. Cells were maintained in RPMI medium (Gibco) supplemented with 10% (v/v) FBS and antibiotics at 37 °C and with 5% CO₂ in a water-jacketed, humidified incubator until large colonies (in excess of 50 cells) were visible in control plates. Medium was removed and cells were fixed with a solution of 6% glutaraldehyde and 0.5% crystal violet in PBS, and the number of colonies was counted.

Tumor formation was assayed by xeno-implantation of genetically perturbed cells. Prior to xeno-implantation, transfected cells were grown for 48 h under standard culture conditions without selective drugs. 5 × 10⁵ cells were subcutaneously injected into the left and right flanks of 4–6 month old female CD-1 nude mice (Crl:CD-1-Foxn1^nu, Charles River Laboratories). Tumor volumes were measured weekly using a caliper for 4 weeks. Following final tumor measurements, mice were euthanized by CO2 inhalation per IACUC protocol. Tumor volume was calculated using the average of 3 measurements of the tumor radius and the formula Volume = (4/3)πr³. The statistical significance of differences in tumor size was determined by Student’s t-test. The vertebrate animal protocol has been approved by UCSD Institutional Animal Care and Use Committee (IACUC) (Protocol Number: S13282).