Nucks1 synergizes with Trp53 to promote radiation lymphomagenesis in mice

NUCKS1 is a 27 kD vertebrate-specific protein, with a role in the DNA damage response. Here, we show that after 4 Gy total-body X-irradiation, Trp53+/− Nucks1+/− mice more rapidly developed tumors, particularly thymic lymphoma (TL), than Trp53+/− mice. TLs in both cohorts showed loss of heterozygosity (LOH) of the Trp53+ allele in essentially all cases. In contrast, LOH of the Nucks1+ allele was rare. Nucks1 expression correlated well with Nucks1 gene dosage in normal thymi, but was increased in the majority of TLs from Trp53+/− Nucks1+/− mice, suggesting that elevated Nucks1 message may be associated with progression towards malignancy in vivo. Trp53+/− Nucks1+/− mice frequently succumbed to CD4- CD8- TLs harboring translocations involving Igh but not Tcra/d, indicating TLs in Trp53+/− Nucks1+/− mice mostly originated prior to the double positive stage and at earlier lineage than TLs in Trp53+/- mice. Monoclonal rearrangements at Tcrb were more prevalent in TLs from Trp53+/− Nucks1+/− mice, as was infiltration of primary TL cells to distant organs (liver, kidney and spleen). We propose that, in the context of Trp53 deficiency, wild type levels of Nucks1 are required to suppress radiation-induced TL, likely through the role of the NUCKS1 protein in the DNA damage response.


INTRODUCTION
Ionizing radiation (IR) is an environmental carcinogen, and exposure to IR is associated with negative effects on health, such as reduced hematopoietic cell function and an elevated risk for cancer. These malignancies are considered to also result from the direct induction of mutations due to insufficient or imprecise repair of DNA damage after IR. IR induces a variety of DNA lesions, of which DNA double-strand breaks (DSBs) are considered to be the most detrimental [1]. To sense and repair DSBs, cells have evolved numerous highly efficient repair pathways, and the two main pathways for DSB repair in eukaryotes are classical non-homologous end joining (NHEJ) and homologous recombination (HR). Defects in either DNA repair pathway can cause genome instability and tumorigenesis [2,3].
In addition to choosing the right DSB repair pathway, the capacity of cells to sense DNA damage and to signal to downstream effectors in the DNA damage response (DDR) network is crucial for genome stability and cancer avoidance. For example, when the Ataxia Telangiectasia Mutated Serine/Threonine-Protein Kinase (ATM), one of the key components of the DDR, is mutated, patients develop Ataxia-Telangiectasia (A-T), an autosomal recessive syndrome characterized by progressive neurodegeneration, radiosensitivity, immune dysfunction, cell cycle checkpoint defects and an increased predisposition to cancer [4]. ATM

Research Paper
is one of the six members of the phosphoinositide 3-kinaserelated protein kinase (PIKK) family that includes other DDR sensors such as AT and Rad3-related protein kinase (ATR) and DNA-dependent protein kinase catalytic subunit (DNA-PKcs).
More than 1,300 proteins are phosphorylated in response to DNA damage, as shown by the results from several studies using quantitative proteomics [5][6][7][8][9]. One such protein is Nuclear Casein Kinase and Cyclindependent Kinases Substrate 1 (NUCKS1), a nuclear and highly phosphorylated protein [10][11][12], which also is acetylated, methylated, ubiquitylated and formylated [13] (http://www.phosphosite.org/). Specifically, in phosphoproteomic screens, NUCKS1 Ser14 was identified as a substrate of either ATM or ATR after IR [5], and Ser54 and Ser181 were identified as ATM-dependent phosphorylation sites after neocarzinostatin [6]. In addition, our own results show that the DNA damageinduced phosphorylation of NUCKS1 at Ser54 is ATMdependent and that it occurs in HeLa cells both after exposure to IR and after mitomycin C [14]. Collectively, the results from phosphoproteomic screens [5][6][7][8][9] and of our previous investigation [14] show that the NUCKS1 protein is an important new player in the DDR, although its precise functions still remain to be elucidated.
Albeit there are little functional cancer-related data on NUCKS1, several studies suggest that there are some links, particularly between NUCKS1/NUCKS1 expression and breast cancer [13,[15][16][17]. NUCKS1 also was identified as a colorectal cancer prognostic marker [18], as a biomarker for recurrence-free survival in cervical squamous cell carcinoma [19], as a risk factor for poor prognosis and recurrence in endometrial cancer [20], as an immunodiagnostic marker in hepatocellular carcinoma [21], and as aberrantly low expressed in adult T-cell leukemia-lymphoma [22] and in childhood acute lymphoblastic leukemia [23]. However, how expression of NUCKS1/NUCKS1 is linked to initiation and/or progression towards malignancy is currently unknown.
Using gene-specific knockdown of NUCKS1 in human cells, we have shown that NUCKS1 is a chromatin-associated protein with a novel role in the DDR and in HR, a DNA repair pathway critical for tumor suppression [14]. However, whether functional loss of NUCKS1 in mice would lead to an increased susceptibility to cancer had not been explored. Here we show that Trp53+/− Nucks1+/− mice more rapidly developed tumors than Trp53+/− mice after exposure to 4 Gy total-body X-irradiation (TBI). Notably, in Trp53+/− Nucks1+/− mice IR-induced thymic lymphomas (TLs) were more prevalent and arose at earlier lineage than in Trp53+/− mice, frequently with concomitant with an upregulated expression of Nucks1 transcript. We propose that, in the context of Trp53 deficiency, wild type levels of murine NUCKS1 are required to suppress radiation carcinogenesis, in line with an important role for NUCKS1 in the DDR.
To test if NUCKS1 protein levels change in wild type mice exposed to IR, five-week old wild type 129S1/ SvImJ mice were either sham-irradiated or exposed to 4 Gy TBI, and thymi were harvested and fixed for immunohistochemistry (IHC) at 2 h, 4 h, 6 h, 24 h and 6 weeks post TBI. Cell death was observed within defined areas of the cortex as early as 4-6 h post treatment ( Figure  4). Viable T-cells in the cortex overall expressed higher   levels of NUCKS1 protein at 4-6 h post IR exposure than T-cells in sham-irradiated thymi, or in thymi that were recovered 6 weeks post IR exposure (Figure 4b). At 24 h post exposure, the vast majority of T-cells in the thymus cortex were dead, and viable T-cells with increased NUCKS1 were distributed mostly throughout the medulla ( Figure 4a). These results suggest that NUCKS1 protein expression increases at early times post IR exposure in thymocytes of irradiated mice.
PCR was used to test for Dβ-Jβ clonality at Tcrb. In Trp53+/− mice, 2/9 TLs (TL-6, TL-21) showed monoclonal expansion and 3/9 TLs (TL-29, TL-32 and TL-102) showed oligoclonal expansion (Figure 6a and 6b; Supplementary Table S1). As TL-29, TL-32 and TL-102 showed more than two predominant PCR products, likely these TLs originated from more than two cells. In contrast,  Figure S5), but clonal expansion of Dβ-Jβ events was not detected (see Figure  6a). The reasons for this are unclear at this point. However, it is possible, that complex rearrangements within Tcrb may have occurred, not detectable by the primers used for Dβ1-Jβ1 and Dβ2-Jβ2 recombination events here.
We Array-based CGH analysis shows that CNVs in TLs from Trp53+/− and Trp53+/− Nucks1+/− mice are both specific and overlapping Array-based CGH analysis was used to examine genome-wide CNVs in TLs from Trp53+/− and Trp53+/− Nucks1+/− mice. The overall patterns of genome-wide CNVs were similar between TLs from both cohorts, including changes at Igh and Tcrg ( Figure  7b, Supplementary Figure S6B; data not shown). In addition, chromosome regions that contain the Ikaros [31,32] and Pten genes [33,34] were lost frequently in TLs from both Trp53+/− and Trp53+/− Nucks1+/− mice (data not shown). CNVs, other than at Tcra/d and Tcrb, that were more specific to one but not the other cohort were also detected. For example, gains in the telomeric arm of chromosome 1, reported to correlate tightly with cancer occurrence [35], were observed in ~50% of the TLs from   Table 2).

DISCUSSION
Environmental factors, including exposure to IR, are recognized as exogenous risk factors that induce genetic changes to drive carcinogenesis. Mice harbor many single-nucleotide polymorphisms and CNVs similar to those observed in humans [36,37], and mouse models are powerful tools for the identification of alleles associated with susceptibility or resistance to carcinogenesis.
The effects of Nucks1 inactivation on carcinogenesis in mice had not been investigated. In human cells, however, NUCKS1 deficiency after siRNA-mediated knockdown impairs DNA damage signaling, DSB repair and genome stability [14], suggesting that this protein may function as a tumor suppressor. Here, we investigated the phenotypic consequences of partial Nucks1 deficiency and the susceptibility of these mice to radiation carcinogenesis. As the constitutive inactivation of HR genes in mice frequently leads to embryonic lethality [24,25,[38][39][40], we chose to conduct our studies in a Nucks1+/− Trp53+/− double heterozygous context. Using a Trp53+/− mouse model  that is prone to radiation-induced TL [32,34,41,42], we found that heterozygous Nucks1 synergistically accelerated radiation lymphomagenesis and also led to other changes in the associated tumor spectrum. Importantly, IR-induced TLs that arose in double heterozygous Nucks1+/− Trp53+/− mice developed at earlier lineage stage and more frequently led to the infiltration of distant organs than TLs in single heterozygous Trp53+/− mice. We speculate that, in our mouse model, Trp53 loss itself (as observed in both cohorts to essentially full extent) does not suffice to fully drive invasive migration, in accord with what has been reported by others [43,44]. Our findings for Trp53+/− mice also are consistent with our unpublished results (J.-H. Mao) that show that IR-induced TLs from Trp53+/− mice rarely infiltrate distant organs (e.g., liver, kidney and spleen). In contrast, TLs from Nucks1+/− Trp53+/− mice show a much greater propensity to infiltrate distant organs, and it will be important in the future to investigate the additional molecular determinants involved in this difference.
Many tumors contain chromosomal translocations and deletions (resulting from mis-repaired or unrepaired DSBs) leading to the activation of oncogenes and/or inactivation of tumor suppressor genes [45]. As such, DSBs likely play a major role in driving malignant transformation. Notably, NUCKS1 deficiency in human cells has been linked to a DSB repair defect [14], and inactivation of NHEJ and HR factors in mice leads to increased tumor burden, frequently resulting from immature T-cell lymphomas [30,46,47]. Defective repair of IR-induced DSBs can lead to TLs associated with translocations, deletions and CNVs [30,[48][49][50][51], similar to the findings reported here. Specifically, we show that TLs in Trp53+/− mice predominantly exhibited CNVs indicative of chromosomal translocations at Tcra/d, as reported previously [52], whereas TLs in Trp53+/− Nucks1+/− mice predominantly exhibited CNVs indicative of chromosomal translocations at Igh. Combined with the observation that the majority of TLs from Trp53+/− mice were CD4+ CD8+ DP or CD8+ SP, we suggest that IRinduced TLs in Trp53+/− mice largely originated at the DP lineage stage, the stage during which Tcra/d recombination occurs. In contrast, in Trp53+/− Nucks1+/− mice, 5 of the 8 analyzed TLs contained mixed CD4-CD8-DN TL cell populations or were purely CD4-CD8-DN, suggesting that a large proportion of TLs in Trp53+/− Nucks1+/− mice originated earlier than the DP lineage stage. Among the TLs that contained CD4-CD8-DN TL cell populations or that were purely CD4-CD8-DN, clonal translocations involving Tcra/d were not detected. Interestingly, the presence of CD4-CD8-DN cells detected in this study in TLs from Trp53+/− Nucks1+/− mice is similar to that reported in TLs derived from mice with somatic mutation of Trp53 in hematopoietic stem cells [53], and could potentially be related to the disruption of the T cell checkpoint pathways through PD-1 expression [54].
Our results suggest that ~ 50% of TLs in Trp53+/− Nucks1+/− mice originated from one or two cells, while TLs in Trp53+/− mice largely appear to have originated from several pre-cancerous cells. Moreover, TLs in Trp53+/− Nucks1+/− mice frequently developed from less mature thymocytes than those in Trp53+/− mice. The heterogeneous expression of CD4 and CD8 in TL-79 and TL-131 from two double heterozygous mice suggests that these TLs originated from single thymocytes and accumulated subsequent changes during tumor progression.
It has been well established that IR almost exclusively induces LOH of the Trp53 wild type allele in TLs derived from Trp53+/− mice [41,55], in line with the results presented here. However, the exact course of events taken during TL development for loss of the wild type Trp53 allele to occur is unclear. Notably, LOH can be caused by several mechanisms, including mitotic recombination, mitotic non-disjunction and also by multi-locus deletion events. We speculate that partial NUCKS1 deficiency may lead to the accelerated loss of the Trp53 wild type allele in TLs from Trp53+/− Nucks1+/− mice, that compared to TLs from Trp53+/− mice arise at an earlier stage during T-cell development (i.e. show increased DN T-cell populations), predominate in Tcrb mono-and bi-clonality, and show relatively higher fractions of Igh-than of Tcra/d-associated translocation events, suggestive of their overall dependency on fewer genetic changes for tumor formation. Interestingly, Trp53+/− Nucks1+/− mice also developed primary lung epithelial tumors that were never observed in Trp53+/− single heterozygous mice, in accord with a previous report investigating Trp53+/− mice only [52]. In contrast, loss of the remaining Nucks1 wild type allele was rare in Trp53+/− Nucks1+/− mice, suggesting linkage of murine Nucks1 to essential loci on chromosome 1.
NUCKS1 expression is reduced in adult T-cell leukemia-lymphoma and in childhood acute lymphoblastic leukemia [22,23]. However, the human NUCKS1 gene also belongs to a group of co-expressed genes located on chromosomal region 1q32.1 that is amplified in some breast cancers [16,17,56,57], and in other cancers [18,19,21,23,35,58,59]. Interestingly, in mouse lung adenocarcinoma, amplification of a similar region on mouse chromosome 1, which results in elevated expression of several genes including Nucks1, correlates with tumor invasiveness [58]. It is interesting to note that we find increased expression (relative to normal thymi from Nucks1+/− mice) of the wild type Nucks1 transcript in 9/13 of the TLs isolated from Trp53+/− Nucks1+/− mice. We speculate that, within a subset of IR-induced TLs in Trp53+/− Nucks1+/− mice, increased expression of Nucks1/NUCKS1 may be associated with a selective advantage during tumor initiation and development, enabling early neoplastic cells to overcome replication stress during lymphomagenesis (Figure 9), thereby limiting DSBs resulting from collapsed replication forks leads to a defect in DNA double-strand break (DSB) repair and to genome rearrangements (i.e., translocations) after ionizing radiation (IR) exposure. c. Tumor initiating cells experience increased replication stress, as described previously [60][61][62], which potentially may be ameliorated by up-regulation of NUCKS1 expression. d. High levels of NUCKS1 provide a selective advantage and promote TL growth, and e. invasion of the tumor border leading to f. intravasation of TL cells into the circulatory system. www.impactjournals.com/oncotarget [60,61]. Of note, NUCKS1 plays an important role in mitigating replication stress in human cells, as we [14] and others [21] have shown. It is also possible that, in TLs from Trp53+/− Nucks1+/− mice, Nucks1 expression is upregulated as part of the activated DDR network during tumorigenesis, as described previously for clinical specimens from different stages of human tumors [62].
We have shown that there is a direct link between Nucks1 status and radiation carcinogenesis in mice. The results of our investigation are relevant to many published reports [16-19, 21-23, 35, 56-59] that indirectly have described an association between NUCKS1/NUCKS1 and several human cancer types. In a different genetic background, Nucks1−/− mice with wild type Trp53 were reported to exhibit decreased insulin signaling and increased body weight/fat mass along with impaired glucose tolerance and reduced insulin sensitivity [63], related to the role of NUCKS1 in the hypothalamus [64]. These studies demonstrate that NUCKS1 can function as a tissue-specific transcriptional regulator of the insulin receptor, critical for insulin signaling and consequent peripheral metabolic activities. Interestingly, the DDR intersects the insulin-IGF1-PI3K-AKT pathway at many points [5]. While Qiu et al. [63,64] did not expose their Nucks1−/− mice to DNA damaging agents and did not investigate the direct phenotypic consequences of persistent or mis-repaired DNA damage in these mice, persistent and repetitive DNA damage has been proposed to alter insulin-IGF1 signaling, thereby contributing to diabetes and other age-associated metabolic disorders [5]. In future investigations it therefore will be important to further dissect the role of the NUCKS1 protein in metabolic syndrome and in cancer avoidance to improve both diagnosis and targeted therapy of these prevalent ailments in humans.

Irradiations
Five week-old F2 Trp53+/− and Trp53+/− Nucks1+/− littermates (Supplementary Materials and Methods) were exposed to 4 Gy TBI using a 360 kVp X-ray machine (Precision X-ray Inc., North Branford, CT, USA). Necropsies were performed when mice appeared morbid, or at the end of the study (45 weeks post TBI). The Kaplan-Meier method was used to compare the tumor development between different genotypes. The study was carried out in strict accordance with the Guide for the Care and Use of Laboratory Animals of the National Institutes of Health.

Immunophenotyping of primary TL cells by flow cytometry
Primary TL cells were isolated from dissected murine TLs and cultured in RPMI-1640 medium with 10% FBS and 1% antibiotics/antimycotics. Freshly isolated or sequentially expanded TL cells were used for cell surface marker staining. The following fluorochrome-conjugated primary antibodies (BD Biosciences, San Jose, CA, USA) were used: PE-Cy7-CD3e, PE-CD4, and APC-CD8a. A FACSCalibur flow cytometer (BD Biosciences) with CellQuest Software (BD Biosciences) was used for data acquisition. Live cells were gated based on forward and side scatter. Data analyses were done using FCS Express 4 software package (De Novo Software, Los Angeles, CA, USA).

PCR detection of Tcrb DJ rearrangements
Tcrb gene DJ rearrangements were analyzed as described [65]. Five and 50 ng of genomic DNA were used for semi-quantitative PCR to detect Tcrb D-J recombination. Primer sequences were as listed in Supplementary Table S1.