Zac1 regulates IL-11 expression in osteoarthritis

Interleukin (IL)-11, a member of the IL-6 family of cytokines, exerts pleiotropic effects under normal and various disease conditions. We assessed IL-11 expression regulation and the IL-11/IL-6 ratio in osteoarthritis (OA) to better guide clinical therapeutic decision-making. Our findings suggest that Zac1, a zinc finger protein that regulates apoptosis and cell cycle arrest, is a transcription factor regulating IL-11 expression. Zac1 overexpression or knockdown respectively induced or suppressed IL-11 expression in HeLa cells. Zac1 acted synergistically with AP-1, human papillomavirus E2, and hypoxia inducible factor 1 alpha (HIF1α). IL-11 expression under various conditions, including hypoxia or treatment with phorbol 12-myristate 13-acetate or copper sulfate. Recombinant IL-11-induced phosphorylation of signal transducer and activator of transcription 3 at tyrosine 705 was reduced in a dose-dependent manner in HeLa cells. Cross-talk between Zac1, IL-11, p53, and suppressor of cytokine signaling 3 was differentially affected by copper sulfate, digoxin, and caffeine. Finally, aggressive vs. conventional treatment of OA patients was primarily determined by IL-6 levels. However, we suggest that OA patients with higher IL-11 levels may respond well to conventional treatments, even in the presence of high IL-6.


INTRODUCTION
Interleukin (IL)-11 is a member of the IL-6 family of cytokines and exerts pleiotropic effects, conferring protection in the intestine, promoting tumorigenesis in gastrointestinal, breast, and thyroid tumors, regulating macrophage differentiation, and stimulating hemopoiesis and thrombopoiesis; however, IL-11 regulatory mechanisms in vivo remain unclear. While recent progress has been made toward elucidating the IL-6 and IL-11 modes of action [1][2][3][4], the effects of these molecules on certain human diseases, including osteoarthritis (OA), are not well understood. OA is characterized by joint pain and varying degrees of functional limitation in the peripheral joints, especially the knee [5], resulting from cartilage degradation and erosion [6][7][8][9][10].
IL-11 shares characteristics with both immuneregulatory (IL-6) and neuro-protective (leukemia inhibitory factor and ciliary neurotrophic factor) members of this cytokine family. IL-11 forms a hexameric signaling complex similar to that of IL-6, but the IL-11 receptor complex contains a single glycoprotein 130 (gp130) chain and a cytokine-specific receptor α chain. Janus kinase-signal transducer and activator of transcription Research Paper www.oncotarget.com (JAK-STAT) signaling, including dissociation of receptorassociated JAK molecules, endocytosis of the receptor complex, and nuclear export of activated STAT molecules. Suppressor of cytokine signaling 3 (SOCS3) limits gp130mediated signaling in a negative-feedback loop by binding to a tyrosine residue at position 757 in mice and 759 in humans [11][12][13].
IL-6 and IL-11 are largely absent from body fluids of healthy individuals [39,41]. However, a wide variety of cell types produce these cytokines following an appropriate stimulus. In contrast to the plethora of cell types that can produce cytokines, expression of their respective receptors is much more restricted. This limits the spectrum of cells that can be directly activated by IL-11 and IL-6. Dysregulation of IL-6 and IL-11 signaling contributes to several diseases, such as inflammatory bowel disease, osteoporosis, rheumatoid arthritis, and various types of cancer [11,12]. In particular, the relationship between IL-6 and IL-11 in human articular tissues remains unclear. This study assessed IL-11 regulatory mechanisms and compared clinical IL-6 and IL-11 levels to better elucidate the value of the IL-6/ IL-11 ratio in OA patients. Our findings provide novel insights into therapeutic strategies for treating IL-6related disorders.

IL-11 and IL-11 receptor α (IL-11Rα) were induced in Zac1 stably-expressing HeLa cells
We overexpressed Zac1 in the HPV type 18 infected cervical carcinoma cell line, HeLa, to identify its potential target genes. Our mRNA expression array revealed that 890 genes were upregulated and 385 were downregulated. Table 1 lists the top 20 up-and downregulated genes. IL-11 and IL-11Rα were dramatically induced by Zac1 in HeLa cells (Table 2). We then addressed whether or not Zac1 directly targeted IL-11.
In the tetracycline-inducible Zac1 expression system, we observed that IL-11 expression was dependent on doxycycline (Dox) concentration ( Figure 1A). IL-6 was not induced under these experimental conditions. We previously showed that the two SUMO-binding lysine residues in Zac1, K237 and K424, repress Zac1 transactivation activity [32]. We examined the importance of these two sites with respect to IL-11 regulation. The Zac1 K237/K424A double mutant did not induce IL-11 expression in HeLa cells ( Figure 1B). K237A or K424A single mutants reduced Zac1-induced IL-11 expression by half. RT-PCR data were consistent with the reduced Zac1 transactivation activity and no effect on IL-6 expression.
We observed different responsive patterns for wildtype and mutant Zac1 in the IL-11 and IL-6 promoter reporter systems ( Figure 1C). Double and single mutant Zac1 had similar effects as wild-type Zac1 on IL-11 promoter activity, and all Zac1 constructs had differential transactivation activities on IL-6 promoter activity in HeLa cells. The consistent enhancement on gene expressions by mouse Zac1 was higher than that of human Zac1 [14,16].
In addition to regulating IL-11, AP-1 also binds the IL-6 promoter region [39]. Our previous work revealed that Zac1 might regulate AP-1 function via enhancement of c-Jun and c-Fos transactivation activities [29]. Here, we transiently transfected various c-Jun family proteins, including c-Jun, JunB, and JunD, with c-Fos, Zac1, or both into HeLa cells to examine their effects on IL-11, IL-6, and SOCS3 promoter activities. Our data revealed that Zac1 worked synergistically with c-Jun/c-Fos to enhance IL-11 and SOCS3 promoter activities (Figure 2A and 2C), whereas Zac1 suppressed the effects of c-Jun/c-Fos on IL-6 promoter activity ( Figure 2B). Zac1 selectively synergistically worked with JunD/c-Fos to enhance IL-6 promoter activity ( Figure 2B).
Our previous work showed that Zac1 physically and functionally interacts with AP-1 and HPV E2 proteins [26,29]. Thus, the effects of c-Jun/c-Fos and PMA on endogenous IL-11 expression prompted us to examine whether c-Jun/c-Fos might work with Zac1 and HPV E2 proteins in the IL-11 promoter system. We observed that the c-Jun/c-Fos complex worked synergistically with hZac1, HPV E2, and both in combination in the IL-11 promoter reporter system ( Figure 4A, histograms 1 and 17-24). In general, there was no apparent effect on the IL-6 promoter reporter system, except for the combination of hZac1 and HPV E2 proteins ( Figure 4B, histograms 1,4,8,12,16,20,24). We further examined the effects of the combination of mouse or human Zac1 and HPV18 E2 on IL-11, IL-6, and SOCS3 levels using RT-PCR and western blotting. Our data showed that both mouse and human Zac1 acted synergistically with E2 to induce IL-11 transcription, but not translation, in HeLa cells ( Figure 4C, 4D). No apparent effect on the IL-6 mRNA was observed, but the combination randomly affected IL-6 proteins.
Multiple studies have shown that the IL-11/IL-11R/ STAT3/NF-κB signaling axis plays important roles in tumor growth, angiogenesis, and metastasis [44][45][46][47][48][49][50][51][52]. We examined the functional role of IL-11 in HeLa cells using The transfectants were then harvested for luciferase reporter assays. The numbers above the columns (A-C) indicate the luciferase activity relative to an index of 1 assigned to the reporter with empty vector. Results are representative of three independent experiments. www.oncotarget.com a recombinant IL-11 (rIL-11) protein. Our western blotting results showed that p-Stat3 (Y705) and the p-Stat3(Y705)/ Stat3 ratio were decreased in a rIL-11 dependent manner in HeLa cells ( Figure 5C and 5D). Zacl and SOCS3 were not affected by the addition of rIL-11.

ZAC1 silencing downregulated IL-11 expression
We stably silenced ZAC1 expression in HeLa cells. The silencing efficacy of clone #262361 was better than that of #262362 ( Figure 6A). Our data suggested that IL-11, not SOCS3 or IL-11Rα, was suppressed in HeLa cells. Caffeine, digoxin, TSA, cobalt chloride, and rIL-11 were employed separately to examine their impacts on ZAC1induced IL-11 expression in HeLa cells. We observed no apparent effects on ZAC1 target mRNA expression ( Figure 6B). We further examined the effects of copper sulfate, digoxin, and caffeine on p53, SOCS3, and IL-11 proteins in shZAC1 HeLa cells. All three of these agents suppressed IL-11 protein expression in these cells ( Figure 6C-6E), including copper sulfate decreased IL-11 expression, digoxin selectively decreased SOCS3 and IL-11 expression, and caffeine decreased IL-11 expression. Consistent with our previous work, digoxin and caffeine increased the β form of p53 in shZAC1 HeLa cells [53,54].
We further assessed ZAC1-regulated IL-11 expression using the chromatin immunoprecipitation (ChIP) assay in HeLa cells. Previous studies showed that the IL-11 promoter region contains two AP-1 sites adjacent to a GC-rich region, a potential Zac1-binding site ( Figure  6F). Our ChIP analysis demonstrated that ZAC1 failed to directly bind this 211 bp fragment, but c-Fos did ( Figure  6G). However, Zac1 might primarily bind the AP-1 site of the IL-11 promoter in complex with AP-1 in HeLa cells ( Figure 6H).

The IL-6/IL-11 ratio is clinically relevant for guiding OA treatment options
Cartilage degradation and erosion are important pathogenetic mechanisms in OA [6]. Chemokines and cytokines play key roles in OA pathogenesis. We analyzed the abundance of various cytokines, including IL-1β, IL-6, and IL-11 from the synovial fluids of 77 OA patients using the Cytometric Bead Array (CBA) flex sets (Figure 7). We compared IL-6 and IL-11 median values from the CBA data to define high vs. low IL-6 or IL-11 levels in this study (Table 3). We analyzed 65 OA cases to identify trends in IL-6 level, which was the primary determinant for aggressive vs. conventional clinical treatment ( Table  4). In general, high IL-6 level of OA patients received the aggressive treatment, whereas high IL-11 level had the possibility to receive the conventional treatment. Most of low IL-6 level of OA patients received the  conventional treatment, whereas low IL-11 level had the possibility to receive the aggressive treatment. Based on the analysis, we further found one OA case with high IL-6 and low IL-11 levels (10731 ng/ml IL-6 vs. 1255 ng/ml IL-11) in the synovial fluid of the right knee associated with recurrent joint effusion and progressive OA joint damage. In contrast, a patient with high IL-11 and low IL-6 levels (538 ng/ml IL-11 vs. 15 ng/ml IL-6) in the left knee responded well to conventional treatment over two years. Thus, the use of high vs. low IL-11 or IL-6 levels to determine OA treatment options requires additional clinical study.

DISCUSSION
In this work, we verified that Zac1 induced IL-11 expression in HeLa cells. Zac1 might indirectly bind the IL-11 promoter as a transcription factor or/and cofactor complexed with AP-1, HIF-1α, or other transcription factors (Figure 9). Zac1 may also induce IL-6 expression under specific conditions. Finally, we observed that high or low IL6 or IL-11 levels might serve as indicators to determine which treatment strategy would best benefit a given OA patient. Our work provides novel insight into IL-11 expression regulation and the clinical relevance of the IL-11/IL-6 ratio in OA patients.
A functional IL-11Rα subunit might trigger signal transduction by IL-11 in human cancer cells cultured under hypoxic conditions [43]. Zac1 also induced IL-11Rα, and recombinant IL-11 proteins failed to activate STAT3 phosphorylation in this study. Most of current experimental conditions might not be performed in the hypoxic situation and the expression of IL-11 mRNA is inconsistent with a hypoxia-inducible gene in HeLa cells. One study showed that IL-11 and IL-11Rα are coexpressed in rheumatoid arthritis synovitis fluid fibroblasts and endothelial cells and thus interconnect the function of these two cell types, whereas macrophages are not IL-11 responder cells because of no IL-11Rα expression [57]. Therefore, activation of the IL-11/IL-11Rα complex under hypoxic conditions must be addressed with respect to tumorigenesis. Previous studies demonstrated that copper sulfate induced HIF-1α expression via the c-Fos and hypoxia reduced some AP-1 protein expression [58,59].
However, our results suggested that copper sulfate might positively or negatively regulate HIF-1α expression depending on hypoxic conditions. Cartilage consists of two main extracellular matrix (ECM) macromolecules: type II collagen (COL2A1) and a large aggregating proteoglycan, aggrecan [10]. COL2A1, the major cartilage collagen, and types IX and X encoding so-called cartilage-specific collagens are the predominant ECM structural proteins involved into OA pathogenesis. COL9A3 is associated with knee OA in the Japanese population [60] and was among the top five Zac1-induced genes in our study. Zac1 expression patterns correlated  with COL2A1 distribution during cartilage development [61]. Hence, ZAC1 might promote OA progression through regulation of cytokines, such as IL-11, or the collagens, such as COL2A1 or COL9A3, although verification of this hypothesis will require further study. IL-6 and its family of proteins, including oncostatin M, are among the most prominently elevated cytokines in the OA inflammatory response [62]. Many studies have correlated of IL-6 levels in synovial fluid with cartilage pathology and associated outcomes in knee OA. Unlike IL-6, IL-11 is well studied with respect to OA [7][8][9][10]. MMP-9 activity is crucial for the destruction of articular cartilage in OA. Our data suggest that ZAC1 upregulates both IL-6 and IL-11 via AP-1 family proteins. Zac1 may directly induce expression of MMP-related genes through AP-1-and Sp1-dependent pathways. Thus, we propose that high vs. low IL-6 or IL-11 levels can be indicators for OA clinical treatments. However, there are likely additional factors within synovial fluid that could impact treatment decisions, and these require further study.

Tetracycline-induced Zac1 and short-hairpin Zac1 RNA in HeLa cells
We established tetracycline-induced Zac1 expression in HeLa cells using the standard protocols for Tet-On Inducible Gene Expression System (Clontch, USA). Two vectors, pTet-On and pTRE-Zac1 (or control vector pTRE-LacZ), were screened and selected using G418 and hygromycin, respectively (Clontch, USA). Positive clones were expanded and sorted for Zac1 expression induced by doxycycline.
Zac1shRNA-containing lentiviral vectors were purchased from the National RNAi Core Facility (Academia Sinica, Taiwan, Republic of China) and prepared using standard protocols. Cells were infected with the indicated lentiviruses in selection medium containing 2 μg/ml polybrene. Forty-eight hours after infection, the cells were treated with 8 μg/ml puromycin for selection of resistant clones.

Total RNA isolation, mRNAs expression profiles, and first strand cDNA synthesis
Total RNA was isolated from HeLa cells using TRIzol reagent (Invitrogen) following the instructions provided by the manufacturer. Gene expression profiles were determined using the HOA OneArray method and HmiOA v5. RNA quantitation was provided by the Phalanx Biotech Group (HsinChu, Taiwan, ROC). The pass criteria indicating acceptable RNA purity using NanoDrop ND-1000 were the following absorbance ratios A260/A280 ≥ 1.8 and A260/A230 ≥ 1.5. RNA integrity (RIN) values were determined using an Agilent RNA 6000 Nano assay. The pass criterion was a value ≥ 6. Contamination by genomic DNA was evaluated using gel electrophoresis. Data were processed though the Rosetta Resolver ® System (Rosetta Biosoftware). Reverse transcription for first strand cDNA synthesis was carried out using a ThermoScript™ RT-PCR system (Invitrogen) with 1 μg of total RNA. GAPDH was used as an internal control for normalization.

Luciferase reporter assay
HeLa cells were plated in 24-well plates and transfected using jetPEI (Polyplus Transfection Inc., New York, NY, USA) following the manufacturer's protocol (Promega luciferase assay kit and DLR2 model). The total DNA was adjusted to 1.0 μg using empty vector. Cells were harvested for luciferase reporter assays using a Promega Luciferase Assay Kit. Values are expressed numerically as relative light units. Luciferase activity is presented as the mean ± SD of three transfected wells and is representative of at least three independent experiments.

Reverse transcription-polymerase chain reaction (RT-PCR)
One microgram of total RNA was subjected to the reverse transcription using MMLV reverse transcriptase for 60 min at 37° C (Epicentre Biotechnologies, USA). The PCR reactions were run in a Veriti Thermal Cycler (Applied Biosystems, MA, USA). PCR was performed in the linear range (30 cycles) with primers specific for hZac1, IL-11, IL-6, SOCS3, IL-11Rα, cadherin-11, aggrecan, MMP-2, MMP-9, and GAPDH. The sequences of the primers for the amplification of target genes are shown in Table 5. The thermocycling conditions were as follows: one cycle at 95° C for 5 mins followed by 30 cycles of 95° C for 45 s, 55° C for 30 s, and 72° C for 40 s. Amplified products were subjected to 1.2 % agarose gel electrophoresis and visualized by staining with ethidium bromide.

Chromatin Immunoprecipitation (ChIP) assays
ChIP assays were performed according to the manufacturer's protocol (Cell Signaling Technology, #9003S). Briefly, chromatin/DNA proteins complexes were prepared from HeLa cells transfected for 24 h with 1 μg of pSG5-HA (empty vector), HA-mZac1, HA-Jun or/ and HA-HIF-1α. Chemical cross-linking of DNA-proteins was carried out using 1% formaldehyde for 10 min at room temperature. The Cross-linking was quenched by addition of glycine for 5 min at room temperature and followed by two washes with ice-cold PBS. Cells were then scraped into PBS containing Protease Inhibitor Cocktail (PIC). The resultant cell suspension was centrifuged, and the pellet was mixed by inverting the tube every 3 min in buffer A + DTT+ PIC and then incubated for 10 min on ice. The pellet (nuclei) was resuspended in 100 µl of Buffer B +DTT + 0.5 μl of micrococcal nuclease and incubated for 20 min at 37° C with frequent mixing to digest the DNA to lengths of approximately 150-900 bp. The nuclei were then completely lysed by sonication using Q125 sonicator (Qsonica, NY, USA). The lysate was incubated overnight at 4° C with rotation with appropriate ChIPgrade anti-HA, anti-c-Fos, or anti-HIF-1α antibody (Cell Signaling Technology) to immunoprecipitate chromatin. This was followed by incubation 2 h at 4° C with rotation with ChIP-grade protein G magnetic beads. The magnetic beads were washed using buffers supplied with the kit. The eluted DNA was purified using ChIP DNA Clean & Concentrator Kits (Zymo Research, CA, USA) and analyzed using PCR to assess binding of c-Jun, HIF-1α, or Zac1 to the IL-11 promoter. The PCR primers were designed using criteria described in the kit. A region of the human IL-11 promoter bearing putative binding sites for both AP-1 and HRE and a GC-rich potential Zac1 binding site were amplified using the following primers: 5′-AGC CTG AGT GTC TGC TCC G-3′ (forward) and 5′-TGA CAC ATC CTG ACT CAC CCT CC-3′ (reverse). The control primers for human RPL30 exon 3 were provided with the kit. End-point PCR amplification was performed using EmeraldAmp MAX HS PCR Master Mix (Clontech, CA, USA) and using the Applied Biosystems Veriti Mastercycler.

Patient recruitment and collection of synovial fluid
Synovial fluids were collected from patients who underwent medical examinations at the National Defense Medical Center (Taipei, Taiwan, ROC). The Institutional Review Board of Tri-Service General Hospital, Taipei, Taiwan approved the study protocol, and signed informed consent was obtained from all subjects before study participation (TSGHIRB No:1-104-05-066). Radiographs were reviewed to determine the size and stage of progression of the OA lesions. Synovial fluid was collected from OA patients using a 22-gauge needle and then stored at −80° C until assayed.

IL-1β, IL-6, and IL-11 measurement
IL-1β, IL-6, and IL-11 levels were measured using a BD™ CBA Human Soluble Protein Flex Set System (Becton Dickson, San Jose, CA) according to the manufacturer's instructions. For CBA analysis, diluted synovial fluid or standards were mixed with Capture Beads, vortexed, centrifuged, and transferred to tubes containing PE Detection Reagents for incubation and analysis. The complexes were then analyzed using flow cytometry (FACScan argon laser cytometer) to identify particles with fluorescence characteristics of both the bead and the detector. Data were analyzed using FCAP Array software (No. 652099).

Cultures of chondrocytes and synoviocytes
To culture chondrocytes, samples of articular hyaline cartilage were collected from patients who underwent total knee replacement to treat OA. The cartilage was minced into 1-5 mm 3 pieces and digested in 5-8 ml of 0.2% collagenase II (Sigma-Aldrich) for 12-16 h at 37° C under 5% CO 2 . The dissociated chondrocytes were then centrifuged, after which the pellets were resuspended in 5 ml of DMEM/F12 (Invitrogen) containing 15% FBS and 1% penicillin/streptomycin solution, transferred to a culture flask, and incubated for 24 h at 37° C under 5% CO 2 . After refreshing the growth medium, the remaining adherent cells were cultured for an additional 2 weeks, changing the growth medium every 3 days. Chondrocyte passages 1 and 2 were used for experiments.
To culture synoviocytes, samples of synovial tissues were treated with trypsin (10% v/v) for 15 min in a 37° C water bath, and the resultant tissue fragments were transferred to DMEM containing 5% FCS, penicillinstreptomycin-Fungizone and 2 mg/ml clostridial collagenase type IV (Sigma-Aldrich). The fragments were then digested on a gyratory shaker for approximately 3 h to dissociate the synoviocytes. The cells were washed and cultured in T175 flasks for 24 h. After refreshing the medium, the adherent synovial cells were further cultured until experiments were performed.

Light microscopic analysis of cell morphology
Human chondrocytes (or synoviocytes) grown in 24well plates to 80% confluence were examined using an Zeiss Axiovert 40 C light microscope (Göttingen, Germany).

Statistical analysis
Student's t-test was used to compare protein amounts derived from Western blot analysis. The Mann-Whitney U test was used to compare levels of IL-11 and IL-6 in samples from recruited OA patients. All values are expressed as means ± SEM. Values of P < 0.05 were considered significant.