|Home | About | Journals | Submit | Contact Us | Français|
Absent in Melanoma 2 (AIM2) is a member of the HIN-200 family of hematopoietic, IFN-inducible, nuclear proteins, associated with both, infection defense and tumor pathology. Recently, AIM2 was found to act as a DNA sensor in innate immunity. In addition, we and others have previously demonstrated a high frequency of AIM2-alterations in microsatellite unstable (MSI-H) tumors. To further elucidate AIM2 function in colorectal tumors, we here addressed AIM2-responsive target genes by microarray based gene expression profiling of 22244 human genes. A total of 111 transcripts were significantly upregulated, whereas 80 transcripts turned out to be significantly downregulated in HCT116 cells, constitutively expressing AIM2, compared with AIM2-negative cells. Among the upregulated genes that were validated by quantitative PCR and western blotting we recognized several interferon-stimulated genes (ISGs: IFIT1, IFIT2, IFIT3, IFI6, IRF7, ISG15, HLA-DRA, HLA-DRB, TLR3 and CIITA), as well as genes involved in intercellular adhesion and matrix remodeling. Expression of ISGs correlated with expression of AIM2 in 10 different IFN-γ treated colorectal cancer cell lines. Moreover, small interfering RNA-mediated knock-down of AIM2 resulted in reduced expression of HLA-DRA, HLA-DRB and CIITA in IFN-γ-treated cells. IFN-γ independent induction of HLA-DR genes and their encoded proteins was also demonstrated upon doxycyclin-regulated transient induction of AIM2. Luciferase reporter assays revealed induction of the HLA-DR promoter upon AIM2 transfection in different cell lines. STAT-signaling was not involved in IFN-γ independent induction of ISGs, arguing against participation of cytokines released in an autostimulating manner. Our data indicate that AIM2 mediates both IFN-γ dependent and independent induction of several ISGs, including genes encoding the major histocompatibility complex (MHC) class II antigens HLA-DR-α and -β. This suggests a novel role of the IFN/AIM2/ISG cascade likewise in cancer cells.
Interferon-inducible Absent in Melanoma 2 (AIM2), a member of the HIN-200 family of hematopoietic, IFN-inducible, nuclear proteins, was recently found to act as a DNA sensor in innate immunity (Schroder et al., 2009). By direct binding to foreign double stranded DNA in infected macrophages, AIM2 triggers the assembly of an AIM2 inflammasome resulting in caspase-1-mediated inflammatory responses and cell death (Burckstummer et al., 2009; Fernandes-Alnemri et al., 2009; Hornung et al., 2009; Roberts et al., 2009). Moreover, the AIM2 inflammasome appears to be essential for host defense against DNA viruses and cytosolic bacteria, such as Francisella tularensis and Listeria monocytogenes (Fernandes-Alnemri et al., 2010; Rathinam et al., 2010; Sauer et al., 2010). Although this activity places AIM2 function in the context of the cell's defense against infection, the role of AIM2 in carcinogenesis is less clear.
Both, tumor suppressive and tumor promoting functions of AIM2 have been reported. Exogenous AIM2 expression was shown to reduce human breast cancer cell proliferation by inhibition of NF-κB transcriptional activity and to suppress mammary tumor growth in a mouse model (Chen et al., 2006). In addition, AIM2 is frequently affected by genetic and epigenetic alterations in different human tumor entities: In pre-neoplastic and neoplastic microsatellite unstable (MSI) lesions, a coding 10-bp polyadenine region in Exon 6 of the AIM2 gene appears to be positively selected for frameshift mutations (Michel et al., 2010a; Woerner et al., 2003, 2005). Finally, we identified additional mutations and epigenetic silencing of the AIM2 promoter, the latter conferring insensitivity to IFN-γ-induced AIM2 expression in MSI colorectal cancers (Woerner et al., 2007). How this aberrant or missing AIM2 expression affects tumor cell fate and tumor immunity is unknown. However, recent evidence suggests that the immune response modulates the fate of human tumors. Type, density and location of immune cells within human colorectal tumors were shown to predict clinical outcome (Galon et al., 2006). Furthermore, the high local lymphocyte infiltration of MSI colorectal tumors, which is associated with a low frequency of distant metastases (Buckowitz et al., 2005) suggests that inflammatory mechanisms, including induction of IFN-γ-regulated target genes are involved in tumorigenesis. Thus, considering the role of AIM2 in innate immunity and intracellular DNA recognition, we hypothesized that AIM2 expression in tumor cells may bridge cytokine release from tumor-infiltrating immune cells with tumor cell fate.
By restoration of AIM2 expression in AIM2-deficient colorectal cancer cell lines, we recently demonstrated that AIM2-mediated inhibition of cell proliferation is associated with accumulation of cells at late S-phase, resulting in G2/M arrest (Patsos et al., 2010). Remarkably, persistent AIM2 expression affected adhesion of colorectal cancer cells to fibronectin and stimulated invasion through extracellular matrix-coated membrane in transwell assays (Patsos et al., 2010). To further elucidate the immunomodulating AIM2 function in colorectal tumors, we here addressed AIM2-regulated target genes by microarray based gene-expression profiling of colorectal cancer cells. Our data indicate that AIM2 mediates IFN-γ-induced induction of several Interferon-stimulated genes (ISGs) including genes encoding the major histocompatibility complex (MHC) II antigens HLA-DR-α and -β.
To investigate the modulation of gene expression in response to constitutive AIM2 expression we used Affymetrix microarray analysis for studying transcriptional changes in the AIM2-deficient MSI-H tumor cell line HCT116. Two recently characterized AIM2-expressing subclones (D1 and B8; Patsos et al., 2010) were chosen for comparison with MOCK-transfected HCT116 cells (subclone D3). Cluster analysis of the data revealed that D1 and B8 subclones are similar to each other but distinct from the control ((D3), data not shown). Among the 22244 genes tested on arrays, 111 transcripts were significantly upregulated, whereas, 80 transcripts were significantly downregulated more than 1,4-fold in both AIM2-expressing cell subclones (Table 1 and Supplementary Tables 2 and 3). As expected, AIM2 was found to be the most upregulated gene, with a slightly higher expression level in subclone B8 (Supplementary Table 2). A number of induced genes appear to be linked to the immunomodulating function of AIM2, such as the IFN-stimulated genes (ISGs) HLA-DRA, HLA-DRB, IFIT1, IFIT2, IFIT3, IFI6, IRF7, ISG15 and TLR3 (Table 1 and Supplementary Table 2). Moreover, genes involved in intercellular adhesion and matrix remodeling (for example, COL9A3, COL12A1, COL13A1, CSPG2, CTHRC1, ECM1, LOXL4, MCAM, NANOS1, TIMP2, VIM and others) were found to be upregulated in response to constitutive AIM2 expression (Supplementary Table 2), which is in line with our previous findings of AIM2 affecting cell migration and invasion (Patsos et al., 2010).
The gene expression profile in AIM2-transfected cells was validated by quantitative real-time PCR of selected genes from independent samples of subclones D1 and B8 versus the control. As shown in Figure 1a, upregulation of VIM, CSPG2, MCAM, CTGF, FGF9, BMP4 and CPA4 was confirmed in AIM2-expressing cells, whereas ANXA10, GPR110, CDH1, NRIP, S100A14, GPR87, HMGCS1 and CK20 were confirmed to be downregulated (Figure 1b and Supplementary Table 3). Moreover, increased expression of the ISGs was demonstrated by both, real-time PCR (Figure 1c) and by immunoblotting (Figure 1d) in AIM2-positive cells.
As transcription of genes encoding MHC class II antigens HLA-DR-α and -β is known to be tightly controlled by the CIITA and the essential RFX complex comprising RFX5, RFXAP and RFXANK (Harton and Ting, 2000; Krawczyk and Reith, 2006; Satoh et al., 2004), we were interested in expression of CIITA and its cofactors in response to AIM2. Although no induction was detected on the microarrays (data not shown), real-time PCR revealed a 15- and 20-fold induction of CIITA transcripts in AIM2-positive cell subclones D1 and B8, respectively, when a primer pair that corresponds to a region shared by all CIITA transcripts was used (Figure 1e). In contrast, expression of RFX5, RFXAP and RFXANK was unchanged. This suggests that CIITA expression might likewise be regulated by AIM2, thereby triggering induction of HLA-DR-α and -β.
On the basis of our observation that AIM2 mediates upregulation of a subset of ISGs, we assumed that IFN-γ-stimulated expression of these ISGs will also correlate with AIM2 expression. To determine the optimal time frame for IFN-γ treatment we performed a time response analysis in HT-29 cells, which revealed maximum transcript expression of AIM2 and most ISGs within 24h of treatment (Supplementary Figure 1A). HLA-DR-α and HLA-DR-β protein expression could be detected after 15h, the level increasing with time (Supplementary Figure 1B). Figure 2 shows fold induction of AIM2 and ISG transcripts in 10 colorectal cancer cell lines upon treatment with IFN-γ for 48h versus untreated cells. Consistent with previous findings (Woerner et al., 2007), the absolute AIM2 transcript level (data not shown) and the magnitude of induction were very low in Vaco-432, SW48 and Caco-2 cells; no AIM2 induction was observed in (parental) HCT116 and RKO cells. In fact, expression of the ISGs was induced by IFN-γ only in cell lines where AIM2 was induced as well, whereas none of the genes was upregulated in AIM2-negative cell lines RKO and HCT116. To further elucidate whether upregulation of ISGs and HLA-DR genes is AIM2 dependent and does not result from a more general unresponsiveness of AIM2-null cells to IFN-γ-signaling, we analyzed IFN-γ-induction of an interferon-inducible gene, IFI16, that is thought to be independent from AIM2 expression. As shown in Figure 2 (last panel), IFI16 was clearly induced by IFN-γ to a similar degree in 9 of the 10 cell lines, including the AIM2-deficient HCT116 cells. In contrast to these cell lines, IFI16 was strongly expressed in RKO cells even in the absence of IFN-γ (data not shown), therefore no further induction was detected upon IFN-γ treatment, here. We thus conclude that IFN-γ-signaling is basically intact in AIM2-null cells. This finding is in line with our recent study (Woerner et al., 2007) that revealed IFN-γ-responsiveness of AIM2-deficient HCT116 cells. Taken together, our data further support the idea that AIM2 may have a role in IFN-γ-stimulated induction of target genes.
The expression of class II MHC antigens is considered a crucial step in immune response towards cancer cells including colorectal carcinomas (Matsushita et al., 1996, 2006; Oshita et al., 2006). We therefore focused our further analysis on the role of AIM2 in regulation of HLA-DRA and HLA-DRB expression. To elucidate whether AIM2 is compulsory for IFN-γ-mediated induction of MHC class II transcripts, we transiently knocked down AIM2 in IFN-γ-treated HT-29 colorectal cancer cells by RNA interference (siAIM2-6, siAIM2-249 and siAIM2-500) targeting independent sequences in the AIM2 transcript. The efficiency of knockdown was verified by real-time RT–PCR, showing that AIM2 expression was reduced by three AIM2-specific small interfering RNAs (siRNAs) to 35%, 25% and 47% of the original IFN-γ-induced expression level (Figure 3a). Using a mix of the three siRNAs (Mix siAIM2 A), AIM2 expression was reduced to 19% of the original IFN-γ-induced expression level, whereas nonsense siRNA and siRNA targeting an unrelated gene (siUnrelated; Figure 3e) did not affect AIM2 expression (Figure 3a). In the consequence of AIM2 knockdown, the expression of HLA-DRA and HLA-DRB mRNAs was specifically decreased (Figures 3b and c), indicating that AIM2 is indeed required for IFN-γ-induced expression of these MHC class II genes. Likewise, IFN-γ-stimulated CIITA induction was diminished, albeit one of the siRNAs (siAIM2-249) was ineffective, here (Figure 3d).
Considering the upregulation of CIITA in consequence of persistent AIM2 expression, we assumed that AIM2 indirectly mediates induction of HLA-DR genes via CIITA. The CIITA gene is controlled by several distinct promoters, two of which direct constitutive expression in immune cells, whereas promoter IV mediates IFN-γ-stimulated expression of a specific CIITA transcript in epithelial cells (Muhlethaler-Mottet et al., 1997). Using transcript-specific primers we could demonstrate that induction of CIITA in IFN-γ-stimulated HT-29 indeed results from the CIITA type IV product (CIITA-p4; Figure 4a). Moreover, IFN-γ-stimulated CIITA-p4 expression was specifically reduced on knockdown of AIM2 transcripts via RNA interference (Figure 4b), which points to involvement of AIM2 in CIITA-p4 induction as well. Like it was observed for CIITA (all transcripts) siAIM2-249 was ineffective, here.
To elucidate whether short-term expression of AIM2 can trigger HLA-DR transcription in the absence of IFN-γ, we used a subclone of HCT116 expressing AIM2 under the control of a tetracycline-inducible promoter (HCT116-tetON-AIM2), in which high amounts of AIM2 transcripts and protein are detectable within 6–12h after addition of doxycycline (Figures 5a and b). In contrast to the D1 and B8 subclones, which persistently express low levels of AIM2 and CIITA, neither CIITA nor CIITA-p4 transcripts were detectable in HCT116-tetON-AIM2 cells in response to AIM2 expression (Figure 5a and data not shown). However, HLA-DRA and HLA-DRB transcription increased 2.8 and 5.1-fold, respectively after 24h (Figure 5a). Furthermore, HLA-DR antigen expression was clearly detectable after 48h of doxycycline-treatment and disappeared when the amount of AIM2 decreased again (Figure 5b).
We next investigated activation of the HLA-DRA promoter region in response to AIM2 by luciferase reporter assays in different cell lines. 293T, HCT116 and HT-29 cells were co-transfected with a luciferase reporter that is driven by a fragment of the HLA-DRA promoter region (pDRA-luc) and an expression vector encoding either EGFP or an EGFP–AIM2 fusion protein. The activity of the empty vector pGL3-basic was used for comparison. In AIM2-transfected cells, the relative activation of the HLA-DRA promoter was 6.1-, 3.1- and 2.6-fold higher compared with EGFP-transfected cells (Figure 6), consistent with the observation that AIM2 is able to induce HLA-DR. In accordance to our previous data, the HLA-DRA promoter was not activated by IFN-γ in HCT116 cells, whereas it was strongly induced in HT-29 and 293T cells. Although the magnitude of HLA-DR induction without CIITA participation is lower in AIM2-transfectants than upon IFN-γ induction, the above experiments argue for a CIITA-independent mechanism of AIM2-mediated HLA-DR induction.
As demonstrated by our data, CIITA was induced in HCT116 subclones D1 and B8 that persistently express AIM2, but not upon transient AIM2 expression in the absence of IFN-γ. This suggested that persistent AIM2 expression might result in release of factors such as IFNs and other cytokines into the medium, thereby activating expression of CIITA and ISGs in an autostimulating manner. We therefore analyzed Janus kinase-signal transducers and activators of transcription (JAK-STAT)-signaling in the D1 and B8 subclones in comparison with HT-29 on treatment with IFN-γ by immunoblotting. The JAK-STAT pathway is activated in response to numerous cytokines via specific receptors resulting in a phosphorylation cascade and STAT homo/hetero dimerization (Aaronson and Horvath, 2002). STAT phosphorylation thus indicates activation of the JAK-STAT pathway by the majority of cytokines, including IFNs. As shown in Figure 7a, AIM2 expression did not result in phosphorylation of STATs 1, 2, 3, 5 and 6 in untreated HCT116 subclones D1 and B8. This indicates that production and autostimulating mechanisms of cytokines are not involved in AIM2-mediated activation of CIITA and ISGs. The treatment of the cells with IFN-γ resulted in a transient phosphorylation of STAT1 and STAT3 in each of the cell lines, HT-29, HCT116 parental cells and sublcones D1 and B8, irrespective of AIM2 expression, with a maximum phosphorylation after 1h of treatment (Figures 7a and b). Interestingly, the total amount of STAT1 increased upon IFN-γ treatment in each of the cell lines, reaching a peak at 24h, whereas the total amount of STAT3 remained unaffected (Figure 7b). However, again these effects were independent from AIM2 expression. In addition, STAT phosphorylation/activation were not detectable in doxycyclin-inducible HCT116-tetON-AIM2 cells (data not shown). In summary, this argues for an AIM2-mediated mechanism of induction that proceeds independently from cytokine-mediated induction of ISGs.
Considering the recently identified role of AIM2 in activation of inflammasome-mediated cell death in response to cytoplasmic DNA (Burckstummer et al., 2009; Fernandes-Alnemri et al., 2009; Hornung et al., 2009; Roberts et al., 2009), we investigated activation of caspase-1 and its substrate interleukin 1β (IL-1β) in response to AIM2 expression in colorectal cancer cells. Western blot analysis revealed induction of pro-Caspase-1 expression in HT-29 cells, HCT116 parental cells and AIM2-expressing subclones D1 and B8 upon IFN-γ treatment (Figure 8). In agreement with previous findings (Patsos et al., 2010), pro-Caspase-1 was not detected in untreated colorectal cancer cell lines. No difference was observed between the pro-Caspase-1 expression level in parental HCT116 cells and AIM2-transfected subclones D1 and B8, indicating that this induction is not caused by AIM2, but rather results from other IFN-γ-mediated mechanisms. Moreover, no activation of caspase-1, as indicated by emergence of cleaved IL-1β (p17), was observed, neither in response to IFN-γ nor as a result of AIM2 expression (Figure 8). In fact, IL-1β precursor was not expressed in any of the cell lines independent from treatment or AIM2-status. We therefore conclude that AIM2 function does not involve activation of caspase-1 or IL-1β in colorectal cancer cells and in this regard apparently behaves differently in macrophages.
The impact of induced AIM2-expression on colorectal tumors is largely unknown. Using a microarray expression profiling approach and subsequent functional analysis in colorectal cancer cells, we here present an evidence that AIM2 mediates induction of genes that were previously shown to be stimulated by IFN-γ. The idea that AIM2 may have a role in IFN-γ action was originally stated by De Young et al. who detected increased AIM2 mRNA levels in HL-60 promyelocytic leukemia cells after IFN-γ treatment (DeYoung et al., 1997). Response to IFNs and subsequent expression of ISGs, in particular expression of class II MHC antigens (HLA-DR, -DQ, -DP) is considered a crucial step in immune response towards colorectal carcinomas (Matsushita et al., 1996, 2006; Pfizenmaier et al., 1985). Accordingly, our finding that AIM2 is involved in induction of these genes adds important information to understanding the impact of inflammatory cytokines on tumor cells. Until now, AIM2 has been primarily identified to act as a DNA sensor in macrophages that induces a machinery of defense mechanisms against cytosolic bacteria and DNA viruses via innate immunity (Burckstummer et al., 2009; Fernandes-Alnemri et al., 2009; Hornung et al., 2009; Jones et al., 2010; Rathinam et al., 2010; Roberts et al., 2009; Tsuchiya et al., 2010). For the first time we present the evidence that AIM2 may likewise have a role in epithelial cells responding to inflammatory processes during tumor progression. Most notably, we show that AIM2 is required for IFN-γ-dependent induction of HLA-DR-α and -β on colorectal cancer cells, thus linking AIM2 function with adaptive immunity as well.
The relationship between HLA-DR antigen expression on cancer cells and prognosis has been pointed out repeatedly. Although HLA-DR antigen expression is barely detectable in normal colorectal epithelium, strong HLA-DR antigen expression on cancer cells relates to better prognosis of colorectal cancer patients (Andersen et al., 1993; Diederichsen et al., 2003; Lovig et al., 2002; Matsushita et al., 2006; Walsh et al., 2009). Consequently, regulation and impairment of HLA class II antigen expression in colorectal tumors is subject of intensive investigations (Garrity-Park et al., 2009; Michel et al., 2010b; Satoh et al., 2004). Constitutive expression of HLA class II antigens is restricted to ‘professional antigen-presenting cells', but can be induced on various tissues by IFN-γ. This tissue-specific expression pattern is mainly determined by a highly specialized molecular machinery that involves interaction between ubiquitous DNA-binding transcription factors and the non-DNA-binding coactivator CIITA. CIITA comprises three cell-type specific protein isoforms, depending on which of three alternative promoters is used for transcription (Harton and Ting, 2000). In epithelial cells, IFN-γ activates transcription of CIITA from promoter p4 through JAK/STAT signaling pathway. Upon interaction with several co-factors, including the essential RFX complex, CIITA then activates the HLA class II promoter (Krawczyk and Reith, 2006).
In accordance with this, we could demonstrate the concerted induction of CIITA and HLA-DRA, and B transcripts in 8 out of 10 colorectal cancer cell lines stimulated with IFN-γ, whereas neither of these genes was induced in IFN-γ-responsive, AIM2-null cells. The IFN-γ-stimulated induction was significantly decreased when AIM2 was downregulated by RNA interference. In addition, CIITA and HLA-DR transcription appeared to be induced by AIM2 that was derived from constitutive overexpression in colorectal cancer cells in the absence of IFN-γ. AIM2 may thus be considered as a co-factor of CIITA for transcriptional activation of HLA-DRA and B. It remains to be demonstrated whether AIM2 in cooperation with CIITA directly activates transcription of HLA class II genes or rather indirectly by regulating CIITA gene expression, which in turn activates transcription of HLA class II genes. Our data showing that AIM2 downregulation results in reduced IFN-γ-stimulated CIITA transcription argues for a role of AIM2 in transcription of the CIITA gene. On the other hand, we demonstrated a direct induction of HLA-DR-α and -β expression without participation of CIITA, when AIM2 was transiently induced in HCT116-tetOn-AIM2 cells. Furthermore, the HLA-DR promoter was activated in AIM2-transfected cells, even in the absence of IFN-γ. As the magnitude of HLA-DR induction without CIITA participation was much lower than after IFN-γ-induction in IFN-γ-responsive cells, we assume that full activation of HLA-DR transcription requires the presence of both, CIITA and AIM2.
Interestingly, CIITA and AIM2 both appear to be evolutionarily related to a group of receptors for the recognition of intracellular pathogens. CIITA shares structural features with the nucleotide-binding oligomerization domain family of proteins (Krawczyk and Reith, 2006), which function as cytosolic sensors for the induction of apoptosis, as well as for innate recognition of microorganisms and regulation of inflammatory responses in mammalian cells (Inohara and Nunez, 2003). AIM2 was recently shown to share the functions of nucleotide-binding oligomerization domain-like receptors by forming inflammasomes, which mediate caspase-1-activation in response to intracellular pathogens (Brennan and Bowie, 2010). So far it is unknown, whether CIITA and AIM2 likewise share some functional roles in adaptive immunity, such as activation of HLA class II-mediated antigen expression and presentation to CD4+ T cells. Unfortunately, the demonstration of protein–protein interaction or co-expression of endogenous AIM2 and CIITA is currently hindered by the lack of experimental tools, such as specific antibodies detecting either protein via immunoblots or immunohistochemistry (Supplementary Figure 2 and data not shown).
Regardless of the precise mechanism of HLA-DR induction mediated by AIM2, our data add to understanding alterations of HLA class II antigen expression in colorectal tumors. Deficiency of HLA class II antigen expression despite the presence of HLA-class II-inducing factors, such as IFN-γ, had previously been reported to occur in about one third of primary MSI-H colorectal carcinomas and in more than half of microsatellite stable colorectal carcinomas (Lovig et al., 2002; Michel et al., 2010b). Moreover, reduced or absent IFN-γ-stimulated induction of HLA class II transcripts and/or protein had been demonstrated in colorectal cancer cell lines HCT116, RKO and Caco2 (Michel et al., 2010b; Satoh et al., 2004), which is in line with our findings. The underlying mechanisms of this deficiency, however, have been elucidated only in part. The lack of HLA class II expression in primary colon tumors has been attributed to mutations in several regulatory genes, such as RFX5 (26.5 %) and CIITA (2.9%) (Michel et al., 2010b). In addition, epigenetic silencing of CIITA by promoter VI methylation was identified in gastric cancers (Satoh et al., 2004) and in cell lines HCT116, RKO and Caco2 (Michel et al., 2010b; Satoh et al., 2004). Here, we present the evidence for an additional mechanism interfering with HLA class II antigen expression in cancer cells, namely alterations in AIM2 expression. We have previously shown that AIM2 promoter hypermethylation conferred insensitivity to IFN-γ-induced AIM2 expression of some MSI-H colon cancer cell lines. Complete silencing of the AIM2 promoter region analyzed was observed in HCT116, whereas considerably reduced AIM2 transcript expression and partial silencing of the AIM2 promoter was found in SW48, Vaco432, RKO and Caco2 cells ((Woerner et al., 2007) and unpublished data). Defects in AIM2 expression, as confirmed in the present study, may thus contribute to the low HLA-DRA transcript expression in IFN-γ-stimulated cells. Immunohistochemical analysis may help to support these findings as a mechanism that is also relevant in primary colorectal cancer tissue, as soon as specific AIM2 antibodies will be available.
Our hypothesis that AIM2 has an important role in cytokine mediated immunomodulation of colorectal cancer cells was further supported by the observation that AIM2 mediates induction of a subset of ISGs in different colorectal cancer cell lines. IFIT1, IFIT2, IFIT3, IFI6, IRF7, ISG15 and TLR3 transcripts were upregulated upon constitutive AIM2 expression and/or in eight colorectal cancer cells lines, in which AIM2 was induced by IFN-γ. The precise mechanism of this gene induction remains to be elucidated, as we cannot conclude from our data whether the upregulation is directly induced by AIM2 or whether it involves intermediate mechanisms, such as participation of CIITA or other factors. Preliminary analysis of HCT116-tetON-AIM2 cells revealed a maximum induction of ISG expression at 24–48h after doxycyclin-stimulated expression of AIM2 in the absence of CIITA (data not shown), thereby correlating with the here reported time course of HLA-DR expression in these cells.
The physiological significance of AIM2-mediated upregulation of IFIT1, IFIT2 and IFIT3 is difficult to evaluate, as the function of these genes is poorly understood. The IFIT1/ISG56 gene family is known to be activated in response to interferons, as well as by a multitude of viruses through involvement of IFN-regulatory factors (IRFs), with and without functional JAK/STAT signaling and IFN-γ action (Fensterl and Sen, 2011). The latter is in good agreement with our observation that AIM2 was capable to induce ISGs in the presence or absence of IFN-γ and without using JAK/STAT signaling. The function of IFI6, also known as G1P3, is likewise largely unknown. It was reported to be overexpressed in human senescent fibroblasts and to inhibit mitochondrial-mediated apoptosis in cancer cells (Cheriyath et al., 2007; Tahara et al., 2005). IRF-7 belongs to a family of interferon regulatory factors, which interact with IFN-stimulated response elements in promoters of IFN-responsive target genes (reviewed in (Zhang and Pagano, 2002)). Among these target genes, there are many additional members of the IFNA gene family, encoding type I interferons. The IRF7 gene has previously been shown to be activated from a basal level by endogenous or exogenous IFN-α, bacterial lipopoysaccharide and some viruses. In contrast to previous studies, in which IRF7 gene expression was restricted to induction by IFN-α/β in lymphoid cell types (Marie et al., 1998), we clearly demonstrated IRF7 upregulation in colorectal cancer cells in response to IFN-γ and AIM2 expression, and may thus expand IRF-7 function to cancer cells. ISG15 is known as one of the earliest ISGs induced by type I INFs. The encoded ISG15 protein is an ubiquitin-like protein that targets many components of the antiviral innate immune response (reviewed in (Pitha-Rowe and Pitha, 2007)). Recently, ISG15 conjugation was shown to be involved in ubiquitination of β-catenin, thereby suppressing Wnt/β-catenin signaling activity, the ‘driving force' of tumorigenesis in colorectal cancer tissues (Lee et al., 2010). In addition to being a post-translational protein modifier, ISG15 is also a cytokine that is secreted from immune cells and epithelial cancer cell lines and was shown to activate monocyte-derived dendritic cells and cytolytic activity of NK cells (D'Cunha et al., 1996a, 1996b). Finally, TLR3 encoding toll-like receptor 3, serves as an important signaling receptor for the recognition of dsRNA for the triggering of antiviral and inflammatory responses to combat viral infections. Whether these ISG functions are important in the context of AIM2 expression in cancer cells remains to be investigated.
In conclusion, our data place AIM2 in a signaling cascade, mediating the biological effects of interferons that result in expression of immunomodulating target genes. Although the AIM2-regulated ISGs and AIM2 itself were previously associated with innate immune response to pathogens, our findings suggest a novel role of the IFN-AIM2-ISG-machinery in cancer cells. Our data provide a functional evidence that AIM2 is required for induction of some ISGs, as well as HLA-DRA, HLA-DRB and the master regulator CIITA in colorectal tumor cells. In contrast to macrophages, AIM2 action appears to not involve Caspase-1 or IL-1β activities, here. Further in vivo and in situ studies will help to understand the biological significance of these interactions for the susceptibility of tumor cells towards anti-tumoral immune response.
Colorectal cancer cell lines LoVo, KM12, HCT116, SW48, LS174T, LS180, RKO, Vaco-432, Caco-2 and HT-29 have been described previously (Woerner et al., 2001, 2007). Construction of HCT116-tet-AIM2 (subclones D1 and B8) and HCT116tTA-Hyg (subclone D3) has been described in (Patsos et al., 2010). All cell lines were grown in RPMI 1640 (Invitrogen, Life Technologies, Karlsruhe, Germany) supplemented with 10% FCS, 100U/ml penicillin and 100mg/ml streptomycin using standard conditions. INF-γ was added as indicated in the figures.
RNA was extracted from the indicated cell lines using RNeasy Kit (Qiagen, Hilden, Germany). Samples of two independent HCT116-tet-AIM2 subclones (D1 and B8) and Mock-transfected, AIM2-deficient HCT116-tTA-hyg cells (subclone D3) were subjected to oligonucleotide microarray analysis using GeneChip HG-U133 Plus 2.0 (Affymetrix, Santa Clara, CA, USA) according to manufacturer's recommendation. Raw data obtained from microarray were reported to GEO (http://www.ncbi.nlm.nih.gov/geo/), sample number GSE24155. Microarray data were annotated using a Unigene based customer CDF-file (http://brainarray.mbni.med.umich.edu/
Brainarray/Database/CustomCDF7CDF_download.aspand, version 10) and raw signals were quantile normalized. A mixed model of ANOVA for differential gene expression was performed with the JMP genomics (SAS) software (SAS, Cary, NC, USA).
For cDNA synthesis, 1–2μg of total RNA was reverse transcribed using oligo-dT primers and SuperScript II reverse transcriptase (Invitrogen, Life Technologies, Karlsruhe, Germany) following the manufacturer's instructions. PowerSYBR Green Master Mix (Applied Biosystems, Darmstadt, Germany) was added to appropriate cDNA samples and primers (Supplementary Table 1). Samples were loaded onto 96-well PCR plates and analyzed in an ABI Prism 7300 thermo-cycler (Applied Biosystems, Darmstadt, Germany) with the following cycling program: 95°C for 10min, followed by 40 cycles of 95°C for 15s and 60°C for 60s. Quantitative analysis of gene expression was performed relative to expression of 18s RNA in corresponding samples by using individual standard amplification curves of each transcript. For undetectable levels of transcripts, we used a baseline of 40 cycles for calculation.
Cells were harvested on ice in cell lysis buffer (25m Tris-PO4 pH 8,0; 2m DTT; 2m CDTA; 10% Glycerin; 1% Triton-X-100) supplemented with protease inhibitor cocktail (Complete EDTA-free, Roche, Mannheim, Germany) and 1m Na3VO4. Cell lysis was performed by two freeze-thaw cycles. For immunoblotting, equal amounts of cell lysates were separated on a 10–12.5% SDS–polyacrylamide gel and blotted on nitrocellulose by semi-dry electroblotting. Equal loading was verified by staining with Ponceau S solution, and blots were blocked for 1h in blocking buffer (Tris-buffered saline, 0.1% Tween-20 (TBST); 5% non-fat dry milk) before primary antibodies were added in blocking buffer (containing 5% non-fat milk for mouse antibodies and 5% BSA for rabbit-antibodies) for overnight incubation at 4°C. After four times washing in TBST, appropriate peroxidise-coupled secondary goat or rabbit antibodies (Dianova, Hamburg, Germany) were added in blocking buffer for one hour. After six times washing in TBST, signals were detected by chemoluminescence (Western Lightning Plus ECL, PerkinElmer LAS, Rodgau, Germany).
Anti-FLAG (mouse monoclonal, clone M2, Sigma-Aldrich, Taufkirchen, Germany); Anti-AIM2 (purified MaxPAb polyclonal antibody B01P, Abnova, Biozol, Eching, Germany), dilution 1:500. This antibody can detect overexpressed AIM2 but not endogenous AIM2 ((Fernandes-Alnemri et al., 2009) and our own unpublished observations). Anti-IFIT1 (mouse polyclonal, Abcam, Cambridge, UK), dilution 1:1000; Anti-IFIT2 (MaxPab mouse polyclonal antibody, Abnova, Biozol Eching, Germany), dilution 1:1000; Anti-actin (mouse monoclonal antibody, clone C4; MP Biomedicals, Aurora, OH, USA), dilution 1:5000. Anti-HLA class II (monoclonal antibody LGII-612.14) was kindly provided by Soldano Ferrone, Department of Microbiology and Immunology, New York Medical College, NY, USA (Temponi et al., 1993). The Phospho-Stat Antibody sampler kit (Cell Signaling Technology, NEB, Frankfurt, Germany) was used for detection of Phospho Stat1 (Tyr701), Phospho Stat2 (Tyr690), Phospho Stat3 (Tyr705), Phospho Stat3 (ser727), Phospho Stat5 (Tyr694) and Phospho Stat6 (Tyr641) as recommended by the manufacturer. Anti-STAT1, anti-STAT3 and anti-IL-1β, dilution 1:1000, were purchased from Cell Signaling Technology, NEB, Frankfurt, Germany; Anti-caspase-1, dilution 1:500, was derived from Merck Biosciences (Darmstadt, Germany).
HT-29 cells were plated at 1.5 × 105 cells in 24-well plates and immediately transfected with siRNA duplexes targeting AIM2 as indicated in the figures at a final concentration of 100n using 9μl/well HiPerFect transfection reagent (Qiagen, Hilden Germany). siRNAs were (sense strand) siAIM2-6: 5′-CCCGAAGAUCAACACGCUUCA-3′, siAIM2-249: 5′-GAGAGUAAAUACAAGGAGA-3′, si-AIM2-500: 5′-GGAGAAAGUUGAUAAGCAA-3′. AllStars Negative Control (‘nonsense') siRNA was derived from Qiagen, Hilden Germany; AIM2 siRNA (Santa Cruz Biotechnology, Heidelberg, Germany), AIM2 siRNA (h), representing a pool of three target-specific 19–25nt siRNA duplexes (Santa Cruz Biotechnology, Heidelberg, Germany) was used as a positive control. In addition, siRNA targeting an irrelevant gene (‘siUnrelated') was used for control. After 24-h transfection, cells were stimulated for 48h with 100U/ml IFN-γ.
A multistep cloning strategy was pursued to generate a doxycycline-inducible flag-tagged AIM2 expression construct (pTRE-Tight-BI-DsRedExpress-FH-AIM2(wt)). The AIM2 cDNA was amplified by PCR using primers that harbor NheI or NotI restriction sites (sense primer AIM2-NheI: 5′-CCCCCGCTAGCATGGAGAGTAAATACAAGGAGAT-3′ antisense primer AIM2-NotI: 5′-CCCCCGCGGCCGCGAGAGGAGCCTGTGAACTGC-3′). The NheI/NotI digested AIM2 cDNA amplicon was subsequently cloned into NheI/NotI digested plasmid pFH-IRESneo (Robert Roeder, The Rockefeller University, New York, NY, USA), thereby generating plasmid pFH-AIM2-IRESneo that contains the Flag-HA-AIM2 expression cassette. Bacterial clones that harbor recombinant plasmids were identified by colony PCR and sequence analysis of the expression cassette. After KpnI/NotI digestion of pFH-AIM2-IRESneo, the 1257bp Flag-HA-AIM2 expression cassette was cloned into KpnI/NotI digested pTRE-Tight-BI-DsRed-Express plasmid (Takara Bio Europe/Clontech, Saint-Germain-en-Laye, France). Recombinant bacterial clones were identified by colony PCR and sequencing across the insert cloning sites. For construction of double-stable cell lines inducibly expressing the above construct, the Tet-On colorectal cancer cell line HCT116-AWE17 (Welman et al., 2005) was used as a background. 107 cells were transfected with 10μg of pTRE-Tight-BI-DsRed-FH-AIM2 expression plasmid using electroporation (AMXA Nucleofector I, solution V). Cells were seeded onto 10-cm dishes and subjected to G418 selection (PAA Laboratories, 600μg/ml) for 3 weeks. Single-cell clones were obtained by limiting dilution of pooled G418-resistant clones from each plate. Double stable DsRed-positive clones were identified upon doxycycline induction (1μg/ml; 24h) by fluorescence microscopy.
Cells were grown at a density of 2 × 105 cells in 24-well plates for 24h. For reporter assays, cells were transfected with equal amounts (300ng/well) of pGL3-basic or pDRA-luc, respectively, using Fugene HD transfection reagent (ROCHE Diagnostics, Mannheim, Germany) according to the manufacturer's instructions. pDRA-luc containing a 6.1kb fragment of the HLA-DRA-promoter region driving a luciferase gene (Osborne et al., 2001), was kindly provided by George Blanck, University of South Florida, Tampa, FL, USA. Expression plasmids encoding EGFP or EGFP-AIM2, respectively, (Patsos et al., 2010) were added as indicated in the figures. For normalization of transfection efficiency, 60ng of a beta-galctosidase-encoding reporter (pRSV-lacZ) was included in each sample. Non-coding DNA was used to fill all transfection samples to a total of 600ng per well. 100U/ml of IFN-γ was added as indicated in the figures. Transfected cells were incubated for 24h, lysed and collected for reporter assays as described previously (Dihlmann et al., 1999) with the following modification: for beta-galactosidase reporter assays, 15μl of each lysate was added to 100μl of assay buffer (60m Na2HPO4, 40m NaH2PO4, 10m KCl, 1m MgSO4, 2.5m EDTA and 3μl/ml beta-Mercaptoethanol) and 25μl of Chlorophenol Red-β--galactoside (CPRG, Sigma-Aldrich, Taufkirchen, Germany) Colorimetric detection was performed at 570nm using a photometer. For comparison of promoter activity, basic induction of pGL3-Basic was used as a reference and set as 1.0.
The work was funded by a grant of the Deutsche Forschungsgmeinschaft, DFG to Dr S Dihlmann (Di-933 4/1).
The authors declare no conflict of interest
Supplementary Information accompanies the paper on the Oncogene website (http://www.nature.com/onc)