The RCAN1 inhibits NF-κB and suppresses lymphoma growth in mice

Nuclear factor-κB (NF-κB) has a vital role in cell survival. Inhibition of NF-κB has been proven to be an efficient therapeutic pathway for various cancers. Activation of NF-κB is mainly through serine residues' phosphorylation of inhibitor of κBα (IκBα) by IKK complex. Phosphorylation at tyrosine 42 is an alternative pathway in regulation of IκBα and NF-κB signaling, though little is known about the underlying mechanism. Here we identified regulator of calcineurin 1 (RCAN1) as a novel endogenous inhibitor of NF-κB signaling pathway. RCAN1 can interact with IκBα and affect the phosphorylation of IκBα at tyrosine 42. Overexpression of RCAN1 by adenovirus reduced cell viability in lymphoma Raji cells and restrained the growth of lymphoma transplants in mice. We further found that N terminus 1–103aa of RCAN1 is sufficient to inhibit NF-κB and reduce cell viability of lymphoma cells. Our study implicated a novel therapeutic approach for lymphoma by RCAN1 through inhibition of NF-κB signaling.

Nuclear factor-κB (NF-κB) is a family of transcription factors that share homology to the retroviral onco-protein v-Rel. NF-κB is an important regulator of cell survival, proliferation and differentiation. 1 NF-κB signaling pathway has a critical role in carcinogenesis, thereby considered to be an important therapeutic target in cancer. 2 Abnormal levels of constitutively activated NF-κB have been detected in various solid tumors, 3 leukemias, and either in cells of lymphoma patients or lymphoma cell lines. 4,5 NF-κB could resist TRAIL-induced apoptosis in Burkitt's lymphoma (BL) cell lines infected by Epstein-Barr virus. 6 All these results indicate that NF-κB is an interesting therapeutic target in lymphoma.
The mammalian NF-κB families contain five members including RelA (p65), RelB, c-Rel, p50 (p105 precursor) and p52 (p100 precursor), and share a highly conserved DNAbinding domain termed Rel homology region that enables them to form various homo-or hetero-dimeric complexes. 7 Activation of NF-κB is tightly modulated by inhibitory subunit known as inhibitor of κB (IκB). 8 IκBα binds to NF-κB dimers and sterically blocks their nuclear localization sequences, thereby causing their cytoplasmic retention in rest cells. When cells are stimulated by activators, such as tumor necrosis factor α (TNFα) and lipopolysaccharide, IκBα are phosphorylated at S32 and S36, which subsequently leads to its degradation by ubiquitin-proteasome pathway. 7 Then NF-κB dimers were released and translocated into the nucleus, where they regulate NF-κB target genes' transcription. An alternative mechanism for NF-κB activation involves the phosphorylation of tyrosine (Y) residue 42 of IκBα, with or without inducing its degradation. 9,10 Though there were extensive studies in the serine phosphorylation by IKK complex, little is known about the regulation of IκBα Y42 phosphorylation that also affects NF-κB signaling.
The regulator of calcineurin 1 (RCAN1) is located on chromosome 21 in the region of 21q22.12. RCAN1 inhibits calcineurin-nuclear factor of activated T-cell (NFAT) pathway in vitro and in vivo by interacting with calcineurin subunit A via its C-terminal 140-197aa. 11 Conversely, NFAT can activate RCAN1 isoform 4 gene expression via several responsive elements in the promoter region of − 350-166 bp, 12,13 thus forming a negative feedback loop. Recent studies demonstrated involvement of RCAN1 in cancer. RCAN1 attenuates endothelial cell proliferation and angiogenesis. 13 In the transgenic mouse model of xenografted tumors, overexpressing RCAN1 gives rise to a broad cancer protection via suppressing tumor angiogenesis by inhibiting the calcineurin pathway in the vascular endothelium. 14 A single extra copy of RCAN1 is sufficient to suppress tumor angiogenesis in mice. 15 Coincidently, the incidence of many cancer types, such as solid tumors, is obviously reduced in Down's syndrome patients. 16 Our recent reports showed overexpression of RCAN1 induced neuronal apoptosis. 17 Though we recently showed NF-κB signaling regulated RCAN1 isoform 4 gene transcription through a NF-κB responsive element in the region of − 576-554 bp, 18 it is unknown if RCAN1 can conversely affect NF-κB signaling activity. We were interested to know if RCAN1 can affect NF-κB signaling and subsequent tumorigenesis.
In the present study, we have solid data to show that RCAN1 inhibited NF-κB signaling activity. The underlying mechanism is that RCAN1 interacted with IκBα and affected its Y42 phosphorylation. Overexpression of RCAN1 by adenovirus in vitro markedly reduced lymphoma Raji cell viability via activation of caspases. Overexpression of RCAN1 by adenovirus in vivo suppressed xenografted tumor growth in severe combined immunodeficiency (SCID) mouse model. The inhibition of NF-κB activity was mediated by N terminus of RCAN1 (1-103aa), which is distinct from its C terminus (140-197aa) inhibition on calcineurin. Our studies provided a novel therapeutic pathway for lymphoma by activating RCAN1 to inhibit NF-κB signaling.
To investigate if RCAN1 affected NF-κB expression, total levels of NF-κB/p65 were detected in cells overexpressing RCAN1. The results showed that RCAN1 increased total NF-κB/p65 expression (Figures 1b and c), precluding its effect on total levels of NF-κB. NF-κB activity is mainly controlled by its nuclear translocation. To examine if RCAN1 affected NF-κB translocation, nuclear proteins were extracted, and NF-κB/p65 western blot showed that RCAN1.1 and RCAN1.4 significantly reduced the nuclear NF-κB/p65 to 34.91 ± 0.28% and 38.4 ± 1.16%, respectively, compared with controls (lanes 2 and 3 compared with lane 1 of Figures 1d-f, Po0.0001). The increase of NF-κB/p65 in cytosol corresponded to the decrease in nucleus fraction, indicating the decrease of NF-κB translocation in RCAN1 transfected cells.
To further determine the DNA-binding activity of NF-κB affected by RCAN1, electrophoretic mobility shift assay (EMSA) was performed using a 20-bp double-strand IRDye 800-labeled consensus NF-κB/p65 oligonucleotide (Lincoln, NE, USA) as the probe. As shown in Figure 1k, the DNAbinding activity of NF-κB was elevated in cells with RCAN1 knockdown (144.3 ± 0.03% relative to control) and reduced in cells transfected with RCAN1.1 and RCAN1.4 (69.03 ± 0.04% and 48.53 ± 0.02%, respectively, relative to control) (Figures 1k and l). These results clearly indicate that RCAN1 can inhibit the transcriptional activity of NF-κB.
RCAN1 increased the endogenous IκBα levels. As the activity of NF-κB signaling was mainly regulated by IKB, endogenous IκBα levels were detected in cells overexpressing Both the serine phosphorylation at 32 and 36 residues (mainly phosphorylated by IKK) and tyrosine phosphorylation at residue 42 will affect the degradation of IκBα and NF-κB activity. 7,9 To further examine which phosphorylation sites were involved, IKK inhibitor Bay 11-7085 and tyrosine kinase inhibitor Genistein were added to HEK293 cells transfected with si-RCAN1 and pNF-κBluc. Figure 2e clearly showed that inhibition of tyrosine phosphorylation by Genistein reduced NF-κB activity (121.9 ± 12.88% compared with 478.6 ± 46.66%, P = 0.0002, Figure 2e To further exclude the effects on S32/36 phosphorylation, Raji cells were treated with TNFα to induce phosphorylation of IκBα at Ser32/36. IκBα-S32/36 phosphorylation antibody showed increased IκBα-S32/36 phosphorylation after TNFα treatment; however, RCAN1 expression did not affect the IκBα-S32/36 phosphorylation (Figures 2f and g). These results suggest that RCAN1 affected NF-κB activity by affecting the tyrosine phosphorylation of IκBα, which subsequently affected the NF-κB activity.
RCAN1 effected on phosphorylation of IκBα at tyrosine 42. It was previously reported that oxidative stress could activate NF-κB through Y42 phosphorylation of IκBα via spleen tyrosine kinase (SYK). 19,20 To further confirm the effect of RCAN1 on IκBα Y42 phosphorylation, hydrogen peroxide (H 2 O 2 ) was used to induce Y42 phosphorylation of IκBα. To further confirm that the effect of RCAN1 was through the Y42 phosphorylation of IκBα, the IκBα phospho-Y42 antibody (Bioworld Tech, Nanjing, Jiangsu, China) was used to immunoprecipitation (IP) and anti-IκBα was used in immunoblotting. The results showed knockdown of RCAN1 increased the ratio of Y42 phosphorylated IκBα ( Figure 3e) and RCAN1 expression decreased the Y42 phosphorylated IκBα (Figure 3f). These results demonstrated that RCAN1 inhibited NF-κB signaling by interacting with IκBα and affecting its phosphorylation at Y42.
To further confirm the effect of RCAN1 is through the Y42 residue of IκBα, two mutant IκBα, IκBα-Y42E and IκBα-Y42F, were constructed and co-transfected with RCAN1. Compared with the increase of IκBα-WT by RCAN1 expression after addition of H 2 O 2 (Figures 3g and h), mutation of Y42 residue abolished the increasing effects of RCAN1 on IκBα (Figures 3i  and j). The weak bands in IκBα-Y42E indicated a faster degradation compared with IκBα-Y42F (Figures 3i and j), suggesting that the Y42 phosphorylation can lead to its degradation. Proteasome inhibitor lactacystin ( (lac) was used to stabilize and visualize the IκBα-Y42E expression. Plasmids pIκBα-Y42E/F were transfected into HEK293 cells and 0.5 μM lac was added to the cells after 48 h transfection. About 4 h after lac was added, the cells were chased with 50 ng/ml cycloheximide. IκBα-Y42E turnover rate was much faster than the Y42F mutant (Figures 3k and l), suggesting the Y42 phosphorylation indeed affected the IκBα stability. These results demonstrated that RCAN1 could affect Y42 phosphorylation and subsequently IκBα protein stability.   (Figures 4c and  d). The decrease of cell viability detected by Cell Counting Kit-8 (CCK8) can be either due to increased cell death or cell cycle arrest. The flow cytometry test of cell cycle showed RCAN1 had no effect on cell cycle. We previously showed RCAN1 expression induced neuronal apoptosis. To elucidate if RCAN1 can induce Raji cell apoptosis, caspase-3/7 activity was examined in Raji cells infected with RCAN1 adenovirus. The caspase-3/7 activity assay showed RCAN1 increased caspase-3/7 activity in Raji cells (Figure 4e, Po0.01). Activation of caspase-3 can also be indicated by its cleavage from 32 to 17 kD. RCAN1.1 and RCAN1.4 increased the level of cleaved caspase-3 in Raji cells to 202.24 ± 6.08% (P = 0.0004) and 159.2 ± 4.36% (P = 0.0021), respectively, (Figures 4f and g). Furthermore, the flow cytometry of Annexin V also showed that RCAN1 increased Raji cell apoptosis to 246.6 ± 1.19% (Po0.0001) and 142.5 ± 0.78% (P = 0.0008) (Figures 4h and i). These data suggest that RCAN1 induced the apoptosis of Raji cells via inhibiting the NF-κB signaling pathway.
RCAN1 suppressed the growth of lymphoma xenografts in SCID mice. To further confirm that RCAN1 inhibited NF-κB signaling in vivo, Raji cells infected with RCAN1.1 or GFP adenovirus were s.c. injected into SCID mice. Lymphoma formation was observed~2 weeks after Raji cells transplantation. The tumor occurrence time was delayed in the RCAN1.1 group compared with GFP control group (Figure 5a). Furthermore, the tumor volumes were also smaller in RCAN1.1 group compared with GFP controls (Figure 5b, P = 0.0008). About 80% (8/10) of mice for each group had tumors formed after 38 days post transplantation.
The experiment was ended 42 days after transplantation. The tumors were dissected and the endogenous IκBα levels were examined in tumor homogenates. Western blot using the tumor tissues showed IκBα was significantly elevated in the RCAN1.1 group (Figures 5c and d, P = 0.0013). The log rank test showed that the tumor growth in RCAN1 overexpressed group is slower than the control (Figure 5e, P = 0.015). The results indicated that RCAN1 reduced the incidence, as well as the growth of lymphoma xenografts in mice.
RCAN1 N-terminal domain 1-103aa directly interacted with IκBα. To investigate if there is an interaction between RCAN1 and IκBα, a co-IP assay was performed. HEK293 cells were transfected with pRCAN1.1-6myc. Anti-myc (9E10) was used to pull down RCAN1. Our results showed that anti-IκBα detected a band in 9E10 IPs that migrated together with IκBα protein in the cell lysate (Figure 6a). And, vice versa, the 9E10 antibody detected RCAN1 protein in the anti-IκBα antibody pull down (Figure 6b). The endogenous RCAN1 and IκBα interaction was also confirmed by Co-IP using DCT3 antibody, 17 targeting the RCAN1 C terminus (Figure 6c). These results indicated that there is a direct interaction between RCAN1 and IκBα proteins.
To further elucidate the domain of RCAN1 interacting with IκBα, a series of deletion mutants was constructed. Co-IP assay showed that only 1-103aa from N terminus of RCAN1.1 can interact with IκBα, while 51-103aa, 51-172aa, 141-172aa and 141-197aa did not interact with IκBα (Figure 6d). Luciferase assay also confirmed that the 1-103aa of RCAN1.   Inhibition of NF-κB by RCAN1 was independent of its inhibition on calcineurin. RCAN1 is previously known to be an inhibitory regulator of calcineurin. Knockdown of RCAN1 would increase calcineurin activity, in that calcineurin overexpression and knockdown of RCAN1 would show similar effects if inhibition of NF-κB by RCAN1 was through its regulation of calcineurin. To investigate whether the inhibition of NF-κB by RCAN1 is associated with calcineurin, the calcineurin expression plasmid and si-RCAN1 were cotransfected with pNF-κBluc or pNFATluc reporter into HEK293 cells. Luciferase assay showed overexpression of calcineurin or knockdown of RCAN1 increased luciferase activity of NFATluc to 149.7 ± 3.35 and 139.3 ± 5.96% relative to control (lanes 4 and 5 of Figure 6g), indicating that RCAN1 regulated NFAT activity by its effect on calcineurin. However, in pNF-κBluc-transfected cells, calcineurin significantly inhibits NF-κB-controlled luciferase activity to 9.51 ± 0.495% (lane 4 of Figure 6h), while knockdown of RCAN1 increased NF-κB luciferase activity to 211.2 ± 20.2% (lane 3 of Figure 6h). The N-terminal domain of RCAN1 1-103aa can inhibit the NF-κB activity (Figure 6e), while it has no effect on NFAT activity (lane 3 of Figure 6g). The opposite effect of calcineurin overexpression and RCAN1 knockdown on NF-κB activity indicate that the inhibition of NF-κB by RCAN1 is independent upon its inhibition of calcineurin, the former is related to the N terminus (1-103aa) and the latter is associated with its C terminus (140-197aa). 11

Discussion
Our study here identified RCAN1 as a novel inhibitor of NF-κB signaling pathway. The underlying mechanism is that N terminus 1-103aa of RCAN1 can physically interact with IκBα and affect its Y42 phosphorylation. The RCAN1 overexpression in cells leads to Raji cell death in vitro. And RCAN1 expression in vivo reduced tumor growth in xenografted lymphomas in SCID mice. Sustained activation of NF-κB is prevalent in cell lines and tumor tissue specimens and contributes to malignant progression and therapeutic resistance in most of the major forms of human cancer. 21 As a newly approved drug for multiple myeloma, the proteasome inhibitor bortezomib (VELCADE) can inhibit the degradation of IκBα and subsequently inhibit NF-κB signaling pathway. 22 Since NF-κB is a popular cancer drug target, the identification of RCAN1 as an NF-κB inhibitor provides a potential treatment for cancers in which NF-κB signaling is aberrantly activated. Our data here showed that RCAN1 affected IκBα Y42 phosphorylation. Previous reports have shown that RCAN1 overexpression stabilized IκBα 23 and deficiency of RCAN1 led to increased NF-κB activity. 24,25 Our study here further provided the molecular mechanism of RCAN1 effect on IκBα Y42 phosphorylation. High glucose and reactive oxygen species have been shown to be able to activate NF-κB signaling via IκBα Y42 phosphorylation by spleen tyrosine kinase (SYK). 26 Continuous studies will be needed to elucidate if RCAN1 has interactions with IκBα tyrosine kinases such as SYK.
Single extra transgenic copy of Dscr1 is sufficient to confer significant suppression of tumor growth in mice, and that such resistance has thought to be a consequence of a deficit in tumor angiogenesis arising from suppression of the calcineurin pathway. 14,27 Hence, it is reasonable to speculate that cyclosporin A and FK506 (tacrolimus), immunosuppressive drugs that specifically inhibit calcineurin, 28 would also suppress tumor angiogenesis. Surprisingly, numerous clinical studies indicate that a significant increase in cancer incidence is a serious complication of transplant recipients receiving long-term immunosupressive therapy. 29 The mechanism behind this increased rate of cancer is not yet understood; however, such studies indicate another oncogenic pathway other than inhibition of calcineurin in which RCAN1 may be involved. RCAN1's inhibition of NF-κB is independent of its inhibition on calcineurin, thus providing a novel mechanism for tumor suppressive effect of RCAN1. Further, we found that the N terminus of RCAN1 is sufficient in inhibiting the NF-κB signaling and reducing the lymphoma cell viability, while it has no effect on NFAT signaling.
RCAN1 not only can repress NFAT signaling pathway via inhibition of calcineurin, 30 but also can be activated by activators of calcineurin-NFAT pathway such as calcium ionophore, VEGF, angiotensin II, TNF-α and so on, 12,13 thereby forming a negative feedback loop in RCAN1 gene regulation. Our recent study showed NF-κB can activate RCAN1 isoform 4 gene promoter through a NF-κB responsive element in the region of − 576-554 bp. 18 Our study here showed inhibition of NF-κB by RCAN1, thus forming a second negative feedback loop in regulating RCAN1 gene expression by NF-κB. The tight regulation of RCAN1 gene expression by negative feedback loops involving NFATand NF-κB implies the crucial role of RCAN1 in regulation of cellular functions.
Generation and purification of recombinant adenovirus. About 5 μg Pac I-digested pAd-RCAN1.1 and pAd-RCAN1.4 were transfected into HEK293 cells using lipofectamine 2000 transfection reagent (Invitrogen, Waltham, MA, USA) in a 60-mm cell culture plate. Seven days later, cells were harvested and lysed with three consecutive cycles of freezing-thawing in a methanol/dry ice bath. Then 50% of the cell lysates were used to infect HEK293 cells in a 60-mm culture plate. Recombinant adenoviruses were amplified with the re-infection of HEK293 cells twice more. Figure 6 RCAN1 directly interacted with IκBα through its N-terminal 1-103aa. (a and b) RCAN1 interacted with IκBα. Co-IP was performed using HEK293 cells transfected with pRCAN1.1-6myc. 9E10 was used as the pull-down antibody and anti-IκBα antibody was used as detection antibody in a, and vice versa in b. (c) Endogenous RCAN1 interacted with IκBα. Co-IP was performed using HEK293 cells. DCT3 antibody targeting RCAN1 C terminus was used as the pull-down antibody and anti-IκBα antibody was used as detection antibody. (d) RCAN1.1 deletion mutants were transfected into HEK293 cells. Anti-myc beads (Sigma, E6654, Shanghai, China) were used as IP and anti-IκBα was used in western blot. The numbers above indicate the fragments of RCAN1.1. In, input. The antibody 9E10 was used to detect the expression of RCAN1 mutants. (e) RCAN1-103 inhibited the transcriptional activity of NF-κB. Plasmids expressing RCAN1.1, RCAN1-103, RCAN141-172 and RCAN141-197 were co-transfected with pNF-κBLuc into HEK293 cells. Renilla luciferase activity was used to normalize transfection efficiency. Dual luciferase assay was performed 48 h after transfection. Values represent mean ± S. E.M.; n = 3; *Po0.0001 by Student's t-test. (f) RCAN1-103 reduced the Raji cell viability. Cells were infected with adenovirus expressing RCAN1 1-103aa. Cell viability assay was performed 4 days after infection. Values from GFP control were arbitrarily set to 100%. Values represent mean ± S.E.M.; n = 4; *Po0.0001 by Student's t-test. (g and h) Inhibition of NF-κB by RCAN1 is independent of its inhibition on calcineurin. The plasmids RCAN1-103 containing 1-103aa from N terminus of RCAN1.1, expression vector of calcineurin, as well as pRCAN1.1-6myc and pSi-RCAN1 were co-transfected with pNFATluc (g) or pNF-κBluc (h) reporter plasmid. Dual luciferase assay was performed 48 h after transfection. Values represent mean ± S.E.M.; n = 3; *Po0.01, by Student's t-test