KCTD15 deregulation is associated with alterations of the NF-κB signaling in both pathological and physiological model systems

Like other KCTD proteins, KCTD15 is involved in important albeit distinct biological processes as cancer, neural crest formation, and obesity. Here, we characterized the role of KCTD15 in different physiological/pathological states to gain insights into its diversified function(s). The silencing of KCTD15 in MLL-rearranged leukemia models induced attenuation of the NF-κB pathway associated with a downregulation of pIKK-β and pIKB-α. Conversely, the activation of peripheral blood T cells upon PMA/ionomycin stimulation remarkably upregulated KCTD15 and, simultaneously, pIKK-β and pIKB-α. Moreover, a significant upregulation of KCTD15 was also observed in CD34 hematopoietic stem/progenitor cells where the NF-κB pathway is physiologically activated. The association between KCTD15 upregulation and increased NF-κB signaling was confirmed by luciferase assay as well as KCTD15 and IKK-β proximity ligation and immunoprecipitation experiments. The observed upregulation of IKK-β by KCTD15 provides a novel and intriguing interpretative key for understanding the protein function in a wide class of physiological/pathological conditions ranging from neuronal development to cancer and obesity/diabetes.


NF-κB
Nuclear factor kappa-light-chain-enhancer of activated B cells IKB-α Inhibitor α of NF-κB IKK-β IKB-α kinase HSPC Hematopoietic stem and progenitor cell MLL Mixed-lineage leukemia ALL Acute lymphoid leukemia AML Acute myeloid leukemia PLA Proximity ligation assay PBMC Peripheral blood mononuclear cells 2′F-ANA-scrambled 2′-Deoxy, 2′Fluroarabino Nucleic Acids scramble sequence 2′F-ANA-KCTD15 2′-Deoxy, 2′Fluroarabino Nucleic Acids against KCTD15 mRNA IPA Ingenuity pathway analysis FCM FlowCytoMetry. KCTD (K)potassium channel tetramerization domain containing protein Leukemic cells originate from the malignant transformation of undifferentiated myeloid or lymphoid hematopoietic progenitors normally residing in bone marrow 1 . Then, according to the immunological features of the malignant progenitors, acute leukemias are classified as acute lymphoblastic/lymphoid leukemia (ALL) or acute myeloid leukemia (AML). About 80% of ALL occurs in children whereas the majority of AML cases are diagnosed in adult patients 2 . The different types of leukemias are caused by either genetic or environmental alterations, although the precise molecular mechanisms underlying these heterogeneous diseases are yet to be Results KCTD15 deregulation influences the NF-κB signaling. We have recently shown that KCTD15 is strongly upregulated in B cell ALL (B-ALL), although the molecular mechanism underlying a potential action of this protein in the pathology is yet to be elucidated 9 . The involvement of KCTD15 in leukemia was initially uncovered by performing a comparative analysis of the transcriptome profile of the peripheral blood of 3 B-ALL patients and 3 healthy subjects (Bio project: PRJNA601326) 9 . To gain insights into the role played by the protein in the disease, here we re-interrogated this dataset looking at other genes that were upregulated/downregulated in this comparative analysis and clustered them according to their biological functions. In particular, the Ingenuity Pathway Analysis (IPA, release 2019) (see the "Methods" section for details) on this dataset unraveled the occurrence of a total of 873 differentially expressed genes. As shown in Supplementary Figure 1, the IPA analysis indicates that genes of the NF-κB activation pathway are the most affected ones. Indeed, 16 out of the 87 annotated genes within IPA Knowledgebase for this pathway were shown to be differently expressed by applying stringent statistical criteria. This finding is not surprising since the NF-κB pathway is frequently found to be hyper-activated in ALL [22][23][24][25][26] . This very preliminary observation prompted us to set up experiments aimed at evaluating the possible role of KCTD15 in the NF-κB pathway. To this scope, we decided to silence the kctd15 gene in SEM cells, a B-cell precursor cell line established from the peripheral blood of a 5-year-old girl at B-ALL relapse and featured by the t(4;11) KMT2A-AFF1 (MLL-AFF1; MLL-AF4) translocation 27 . In this pathological model system, we tested the expression levels of the proteins associated with the NF-κB signaling after KCTD15 silencing using the 2′F-ANA methodology (see "Methods" for details). In particular, we performed flow cytometry (FCM) experiments to evaluate the expression levels of phosphorylated IKB-α (pIKB-α) and IKK-β (pIKK-β) (Fig. 1). Notably, we found a significant reduction of the proteins that play key roles in NF-κB signaling, such as of pIKK and pIKB-α at day 7 of the silencing. It is important to note that these analyses were performed specifically on live cells selected according to the FSC and SSC parameters. This selection was essential considering the significant cell mortality observed at day 7 (Ref. 9 and Supplementary Figure 2). On the other hand, the unphosphorylated forms of these proteins were unchanged (Fig. 1), suggesting that their reduced phosphorylation does not have the time to impact their total content. The association between KCTD15 downregulation and NF-κB signaling reduction was also explored in the RS4;11 cell line, a second B-cell precursor featured with KMT2A-AFF1 (MLL-AFF1; MLL-AF4) fusion gene 27 . Also in this model system, the downregulation of KCTD15 is associated with reduced expression levels of the proteins involved in NF-κB signaling (Supplementary Figure 3). Again, considering the significant cell death in lymphoid cell lines induced by KCTD15 signaling (Ref. 9 and Supplementary Figure 2), the overexpression levels of these proteins were specifically measured on live cells.
To overcome some limitations of these silencing analyses, as the relatively long time taken to obtain measurable effects (7 days for SEM and 13 days for RS4;11) and the significant cell lethality, and to corroborate/extend  Figure 4). In SKBR3 cells, the successful inactivation of KCTD15 gene by CRISP/CAS9 recombinase (Fig. 1A) did not influence the cellular viability (Supplementary Figure 5). Interestingly, as for leukemia cell lines, this KCTD15 inactivation caused a decrease of the expression levels of both pIKB-α and pIKK-β compared to the controls. In this case, we also observed a global increase of IKB-α as a consequence of its reduced phosphorylation and, likely, degradation (Fig. 2B). As these findings suggest that the KCTD15 downregulation could lead to a decrease of the NF-κB signaling, its status was monitored by evaluating the localization of Rel-A, a subunit of NF-κB, in SKBR3 cells. In these cells, in which the pathway is activated, Rel-A was localized in the cytoplasm and, partly, in the nucleus (Fig. 2C). However, upon KCTD15 silencing (SKBR3 KCTD15− cells), we observed a primary cytoplasmic localization of Rel-A in a manner that is similar to that observed after the treatment of these cells with JSH-23 (Fig. 2C), a potent inhibitor of the NF-kB nuclear translocation 29 . Of note, the inhibition of NF-κB shuttling, associated with KCTD15 downregulation, was also detected in SEM and RS4;11 cells (Supplementary Figure 6). Indeed, in these cell lines, it was possible to observe an accumulation of NF-κB in the subtle cytoplasmic rim of the two-cell line after their treatment with 2′F-ANA-KCTD15. These findings clearly indicate that KCTD15 downregulation impairs the NF-κB signaling, likely though the downregulation of pIKK-β, the active form of the kinase.
The association between KCTD15 expression levels and NF-κB signaling was also monitored in an additional model system, the NF-κB reporter (Luc)-HEK293 cell line, that produces the luciferase enzyme under the control of NF-κB transcription factor after its stimulation with either TNF-alpha or PMA/Ionomycin. As shown in Fig. 3, we found that the transient over-expression of KCTD15 did not affect the luciferase activity. However, when cells are stimulated with PMA/Ionomycin the transient over-expression of KCTD15 increased the luciferase activity.
KCTD15 is important for the activation of NF-κB in physiological contexts. The functional role of KCTD15 in conjunction with the NF-κB pathway was additionally investigated in physiological contexts using peripheral blood circulating lymphocytes and CD34 positive hematopoietic progenitors. Specifically, in the case of PB lymphocytes, the NF-κB pathway is critical for the activation of T cells in response to external stimuli for the production of proinflammatory cytokines, such as TNF-α. To evaluate the role, if any, that KCTD15 could play in the early stages of T-cells activation, we stimulated peripheral blood T-cells for 3 h using Duractive™ activation tubes containing an activation mix composed of phorbol 12-myristate 13-acetate (PMA), Ionomycin, and Brefeldin A. The mix is specifically designed for inhibiting cytokine degranulation. As expected, the application of the protocol induced the production of TNF-α cytokine in stimulated T-lymphocytes ( Supplementary Figure 7). Since the production of the cytokine in these lymphocytes is induced by the canonical NF-κB signaling, we monitored the phosphorylation levels of IKK-β, and IKB-α and non-phosphorylated form of IKB-α before www.nature.com/scientificreports/ and after the stimulation. As shown in Fig. 4, the addition of the mix upregulated both pIKK-β and pIKB-α. This observation corroborates the notion that the stimulation occurs through the activation of the NF-κB signaling.
Interestingly, the analysis of the KCTD15 levels in response to the PMA/Ionomycin activation highlights a clear upregulation of the protein (Fig. 4). Successively, to evaluate the possible role of KCTD15 in the signal transduction for the production of the TNF-alpha cytokine, we decided to transiently inhibit KCTD15 mRNA with 2′F-ANA antisense nucleotides. To this aim, Peripheral Blood Mononuclear Cells (PBMCs) were pulse chased overnight with anti 2′F-ANA-KCTD15 oligo at + 4 °C for allowing cellular uptake. The day after, unstimulated lymphocytes, were used to define the gate boundary and to determine the percentages of antigen positivity. The treatment of PBMC with the 2′F-ANA-KCTD15 oligo was able to moderately reduce the upregulation of KCTD15 after 3 h of PMA-Ionomycin stimulation (Fig. 5). While no significant differences were detected for IKB-α expression, the KCTD15 reduction was associated with lower TNF-alpha (34.58% vs 54.70%) and pIKK-β (49.52% vs 58.63%) production compared to the controls. Finally, since NF-κB activation is physiologically important in the ontogenesis of hematopoietic stem cells (HSC) 30 , we decided to investigate KCTD15 expression levels in this cellular compartment. To this aim, we selected the CD34 positive cells from cryopreserved BM samples of four pediatric B-ALL patients with undetectable minimal residual disease (day + 78 of chemotherapy treatment) and analyzed them by FCM. Remarkably, we found that KCTD15 levels were significantly higher in CD34 pos /CD45 dim hematopoietic stem cells when compared to mature lymphocytes (Fig. 6, and Supplementary Figure 8). Once again, these findings highlight a close connection between KCTD5 upregulation and NF-κB activation and suggest that KCTD15 could be important for sustaining the NF-κB signaling during the development of hematopoietic progenitors in physiological conditions.

KCTD15 as a novel interactor of IKK-β.
Considering the association between KCTD15 deregulation and the NF-κB signaling highlighted in the above sections, we wondered whether the KCTD15 expression could be under the control of this transcription factor. In particular, we performed an in-silico analysis for predicting NF-κB binding sites (TFBSs) in KCTD15 promoter gene, at 500 bp and 1000 bp upstream transcription start position (TSS), by Contra V3 software. According to Supplementary Figure 9, we found two possible NF-κB binding sites, such as: NFKB1 JASPAR_CORE_2016, MA0105.4 (IC 14.2, consensus: AGG GGA WTC CCC T) and MA0778.1 (IC11.4, consensus: AGG GGA WTC CCC Y), with an information content (IC) of the positional weight matrices (PSWs) higher than 5 (ranging from 5 to 35). However, no any predicted NF-κB binding sites on KCTD15 promoter were statistically significant (q > 0.25). This finding makes unlikely a NF-κB RelA regulation on KCTD15 at transcriptional level. In this scenario, we evaluated the possibility of physical interaction between KCTD15 and the NF-κB activator IKK-β. To this aim, we performed the experiment presented in Fig. 7 where IKK-beta-protein immunoprecipitation in RS4;11 and SEM were screened for KCTD15 association. According to the WB analysis, we found that the immunoprecipitation for IKK-β was able to enrich the signal for the endogenous KCTD15 protein. This enrichment is qualitatively similar to that observed for NF-κB after protein  www.nature.com/scientificreports/ RS4;11 and SEM, respectively) was not very different from that observed for NF-κB/IKB-α (78.50 and 62.51% in RS4;11 and SEM, respectively), which are well known functional partners 28 and then used as a positive control in these experiments. To further validate the data obtained using the endogenous proteins, we decided www.nature.com/scientificreports/ to test the IKK-β /KCTD15 interaction in the HeLa cells transiently transfected with IKK-β (FLAG conjugated) and KCTD15 (at 24 h). As shown in Fig. 7D, immunoprecipitation of IKK-β-FLAG protein was significantly enriched for the KCTD15 when compared to non-transfected HeLa cells. Collectively, these results provide a strong indication that KCTD15 interacts with IKK-β or with the IKK complex. www.nature.com/scientificreports/

Discussion
The NF-κB signaling is fundamental in a plethora of physiological and pathological manifestations. In its canonical version, the activated (phosphorylated) kinase IKK phosphorylates the NF-κB inhibitor marking it for degradation 31 . These events liberate the NF-κB transcription factor that can translocate to the nucleus where it induces the transcription of a myriad of genes including anti-apoptotic proteins (cFLIP, BCL-2, and BCL-xL), growth factors, cytokines (IL-1 and IL-6), cell adhesion molecules, and chemokines 32 . In this scenario, it is not surprising that the NF-κB signaling is crucial in a countless number of biological processes that span from inflammation and immune response to cell growth/survival and development 33,34 . A full understanding of the many factors that can regulate this pathway may have immense implications for the development of innovative therapeutic applications 35 . Starting from our recent observation that KCTD15 is upregulated in both B-ALL 9 and AML cell lines 10 and patient-derived samples, we here found an intriguing link between the upregulation of the protein and the activation of the NF-κB pathway that was observed in different contexts and conditions. In a first instance, the transient silencing of KCTD15 in MLL-rearranged leukemia model systems (RS4;11 and SEM) induced cell death and apoptosis with a significant downregulation of pIKK-β, the kinase deputed to the activation of the NF-κB pathway 9 . Here we initially demonstrated a down regulation of proteins involved in the NF-kB pathway upon silencing in these leukemia models systems. This finding is particularly important in light of the observations of Kuo et al. proving that in MLL-rearranged leukemias the cellular survival and proliferation were constitutively dependent by the NF-κB pathway activation 27 . An extension of these results was achieved by silencing KCTD15 in breast cancer cell line (SKBR3), where this protein is remarkably upregulated (Supplementary Figure 4). In www.nature.com/scientificreports/ this case, we applied the CRISP/CAS9 methodology and were able to obtain viable cells in which KCTD15 was knocked down. Despite the remarkable differences between SKBR3 and RS4;11/SEM model systems as well as the methodologies applied, we essentially replicate the results obtained in leukemia cells by measuring a reduction of the phosphorylated forms of IKK-β and IKBA. Notably, in both leukemia and breast cancer model systems, the downregulation of this key players in the NF-κB signaling was also associated with an increased cytoplasmic localization of the transcription factor where it is generally inactivated by IKBA. The association of KCTD15 deregulation and NF-kB signaling was here detected in a variety of other contexts as for example the NF-κB reporter (Luc)-HEK293 cell line 36 and circulating T-lymphocytes upon stimulation with PMA/Ionomycin. The PMA/Ionomycin activation in T-cell bypasses the T cell membrane receptor complex and will leads to activation of several intracellular signaling pathways, which favor the production of a variety of cytokines, including the TNF-alpha that is under the transcriptional control of NF-κB 37 . As expected, the stimulation with the PMA/Ionomycin mitogen led to an increased phosphorylation of IKBA and IKK as well as production of the TNF-alpha cytokine that was concomitant with a remarkable KCTD15 upregulation. Moreover, the silencing of KCTD15 in these cells led to a reduction of TNF-alpha production. The possible link between KCTD15 expression and NF-κB signaling in normal immune/hematopoietic cells was also explored taking into consideration the CD34 compartment. Indeed, in CD34 HSPC the activation of NF-κB signaling has a positive regulation of the transcription of genes involved in the maintenance and homeostasis [38][39][40] . Once more, we found a significant upregulation of KCTD15 in the CD34 positive compartment that is suggestive of a possible role of KCTD15 in the physiology of HSPC mediated by the NF-κB activation. This converging indications of a functional link between the upregulation of KCTD15 and the upregulation of the NF-κB pathway both in physiological and pathological states are corroborated at the molecular level by the observation of the physical interaction of KCTD15 and IKK-β as highlighted by PLA and immune-precipitation analyses. It is worth mentioning, however, that, although the picture that emerged from these results fits in the canonical view of an NF-κB activation by KCTD15 that favors the dissociation of the cytoplasmatic NF-κB/ IKB-α complex through the upregulation of pIKK-β, it cannot be excluded that KCTD15 influences IKK-β also in other compartments. Specifically, according to the human protein atlas (https:// www. prote inatl as. org/ search/ kctd15) KCTD15 can be found at the nuclear level too. This consideration could be of relevance according to recent findings by Armache et al. showing the ability of IKK-alpha to phosphorylate Histone H3.3 and enhance stimulation-induced transcription 41 . It cannot be excluded that in pathological and physiological model systems different from those considered in this paper, KCTD15 could interact with proteins of the IKK complex also in the cell nucleus 42 .
The association of KCTD15 with the NF-κB pathway opens a new perspective for the interpretation of the mechanism of action of KCTD15 in diversified biological contests. It has been recently reported that IKK-β is a β-catenin kinase as it phosphorylates the degron motif of β-catenin to prime it for ubiquitination/degradation mediated by the E3 ubiquitin ligase β-transducin repeat-containing protein β-TrCP. Therefore, based on the present results, the upregulation of KCTD15 could downregulate the β-catenin thus decreasing the Wnt/βcatenin pathway. This hypothesis perfectly fits with the observation that KCTD15 inhibits neural crest formation by attenuating Wnt/β-catenin signaling 43,44 . The upregulation of IKK-β is a property that KCTD15 likely shares with the close homolog KCDT1 that also suppresses the canonical Wnt/β-catenin pathway by enhancing β-catenin degradation through β-TrCP 19 . By extending these considerations, it also possible to make a tentative but intriguing functional link between KCTD15 and obesity whose connection has been found at the genetic level [43][44][45] . Indeed, it has been reported that IKK-β as critical for adipocyte survival and adaptive adipose remodeling in obesity 45 and that it could serve as a key molecular switch that triggers the adipogenic differentiation of mesenchymal stem cells 46 .
The ability of KCTD15, emerged from the present analyses, play a role in the NF-κB pathway in both pathological and physiological contexts holds interesting implications on the etiology of leukemia that could also apply to other carcinogenic processes. Present findings also highlight hitherto unknown functionalities of this protein that may be shared by the other members of the KCTD family. Finally, the observed upregulation of IKK-β by KCTD15 provides a novel and intriguing interpretative key for understanding a wide and diversified class of physiological and pathological states ranging from neuronal development to obesity and diabetes. Studies aimed at putting these considerations on more solid grounds are in progress. www.nature.com/scientificreports/ Cell lines. RS4;11,SEM, SKBR3, HeLa and NF-κB reporter (Luc)-HEK293 cell lines were used for the present study. NF-κB reporter (Luc)-HEK293 cell line was purchased from BPS Bioscience (#60650) while the other model systems were authenticated at DSMZ for short tandem repeat (STR) profile. For RS4;11 and SEM culture media (Sigma-Aldrich, MO, USA) was composed of Iscove's Modified Dulbecco's Medium supplemented with 2 mmol/L l-Glutamine (Sigma-Aldrich) and 10% heat-inactivated FBS (ThermoFisher, GIBCO). For HeLa and NF-κB reporter (Luc)-HEK293 culture media were Dulbecco's Modified Medium (DMEM, Gibco) supplemented with 2 mmol/L l-Glutamine (Sigma-Aldrich) and 10% heat-inactivated FBS (ThermoFisher GIBCO). For SKBR3 culture medium was McCoy's (Gibco) supplemented with 2 mmol/L l-Glutamine (Sigma-Aldrich) and 10% heat-inactivated FBS (ThermoFisher, GIBCO). All cell lines were cultured at 37 °C in a humidified atmosphere with 5% CO 2 . Mycoplasma contamination was routinely (monthly) checked using the PCR Mycoplasma Detection KITfrom ABM (Richmond, BC, Canada).
Proximity Ligation Assay (PLA) employs a pair of oligonucleotide-conjugated antibodies (called PLA probes) with an affinity for the primary antibodies that will be used. If two targeted proteins can be recognized by the primary antibodies from two different species and can interact, the PLA probes will remain nearby. In this way, the PLA probes serve as templates for the hybridization of two additional DNA oligonucleotides, guiding their ligation into DNA circles. The circles will be locally amplified by rolling circle amplification to generate intracellular fluorescent products that will be studied by FCM 52 . For these experiments, 5 × 10 5 RS4;11 and SEM cells were fixed and permeabilized using PerFix Expose Kit. Protein interaction were detected using the following antibody pairs: (i) Anti IKB-α (662402, Biolegend, USA, Mouse) with Anti NF-κB (622602 Biolegend, USA, Rabbit,); (ii) anti-IKK-β (GTX107970, Genetex International, USA, Rabbit) with Anti KCTD15 (GTX50002, Genetex International, USA, Mouse). Primary monoclonal antibodies were added at a concentration of 1:50 in Buffer 3. After 30 min of incubation 2 wash steps in PBS 1 × were performed and PLA protocol was applied according to the manufacturer's instructions (DUO94002, DUO92001, DUO92005, Sigma Aldrich, Germany). PLA assays were acquired on Cytoflex cytofluorimeter and fluorescence was recorded in the FITC channel.
Statistical analysis and reproducibility. p-values were calculated as described in individual figure legends using GraphPad Prism 7 (GraphPad Software). Numbers of biological and/or technical replicates as well as a description of the statistical parameters are stated in the figure legends. All experimental images are representative of at least two independent experiments. RNA-sequencing and in silico analyses. Gene expression levels from previous RNA-Seq results 9 , were retrospectively used for functional genomic analyses using the Ingenuity Pathway Analysis software (IPA, QIA-GEN Inc., https:// www. qiage nbioi nform atics. com/ produ cts/ ingen uityp athway-analy sis). The top statistically significant IPA canonical pathways with enrichment score threshold (− log adj p-value) ≥ 5, by using Benjamini-Hochberg approach for Multiple Testing Correction 50 were ranked for the relative ratio of differentially expressed genes overlapping the total number of molecules within each Ingenuity Knowledgebase pathways (n. 7). We carried out in silico prediction analyses of the KCTD15 promoter region and RELA/NF-κB TF binding sites by NCBI database 51 and Contra V3 web server 52 . We analyzed the KCTD15 promoter region (1000 bp and 500 bp upstream) with respect to the reference sequences (RefSeq) "NM_001129994.2" and "NM_001129995.2". We carried out a positional weight matrices (PWMs) approach 53 with stringency parameters core = 0.95, similarity matrix = 0.85.

Data availability
The datasets analyzed during the current study are available in the Bio project: PRJNA601326 (more information at reference number 9).