Trib1 and Trib2 are leukemia disease genes. Trib2 was first identified as a downregulated gene upon γ-secretase inhibitor treatment of Notch-dependent mouse T-ALL cell lines.(17) Retrovirus-mediated gene transfer of Trib2 into mouse BM cells induced monocytic differentiation, inhibited granulocytic differentiation and accelerated self-renewal activity. Transplantation of Trib2-expressing BM cells into lethally irradiated mice resulted in AML development at 100% penetrance. Subsequently, Trib1 was identified as a collaborator of Hoxa9 and Meis1 in myeloid leukemogenesis.(30) Cooperative genes for Hoxa9/Meis1 were identified as common targets of retroviral integration when retrovirus was used as an insertional mutagen,(39) and Trib1 was identified as the most frequent common integration site of the Hoxa9/Meis1 retrovirus in AML. Trib1 is by itself a transforming gene for myeloid cells but also significantly accelerates progression of Hoxa9/Meis1 AML. Specific genetic interaction between Trib1 and Hoxa9/Meis1 was further confirmed by identifying Trib1-retroviruses in Trib1-AML at Hoxa7, Hoxa9 or Meis1 loci. Consistent with the results, Keeshan et al.(40) confirmed that Trib2 also cooperates with Hoxa9.
Argiropoulos et al.(41) identified Trib2 as a downstream target gene of Meis1 in Meis1/NUP98-HOXD13 leukemia cells and they speculated that Trib2 replaces Meis1 function in leukemogenesis. However, this is an interesting issue given the functional similarity between Trib1 and Trib2, and specific genetic interaction between Trib1 and Meis1; further study is needed to define the relationship between tribbles and homeodomain proteins in leukemogenesis.
The molecular mechanism of tribbles-induced leukemogenesis is an important issue for understanding tribbles function as well as the biological nature of leukemia. First, phosphorylation of ERK1/2 is enhanced in Trib1-transfected HeLa cells and BaF3 cells as well as leukemia cells derived from Trib1-induced AML upon cytokine stimulation.(30) The MEK1 binding motif ILLHPWF in the C-terminal region of Trib1 was then identified (Figs 2C,3A). A Trib1 mutant lacking the motif was unable to enhance phosphorylation of ERK1/2 or prolong self-renewal of BM cells.(14) More importantly, the mutant that lacked the ILLHPWF motif could neither induce AML nor accelerate Hoxa9/Meis1-induced AML. These results indicate that MEK interaction and enhancement of the MEK/ERK pathway is required for the leukemogenic activity of Trib1.
Second, downregulation of C/EBPα p42 (the full-length isoform) was observed in Trib2-induced AML as well as in 32D and U937 myeloid cells overexpressing Trib2. As discussed above, it was proposed that Trib2 induces downregulation of C/EBPα through proteasome-dependent degradation. An E3 ubiquitin ligase COP1 that is required for degradation of ACC by Trib3 also mediates C/EBPα degradation by Trib2.(42) The COP1 binding motif is required for both degradation of C/EBPα and leukemogenic activity of Trib2. All three tribbles members could interact with COP1 through the C-terminal DQXVP sequence (Figs 2C,3A); however, while Trib1 and Trib2 could induce degradation of C/EBPα, Trib3 could not.(43) These data suggest degradation of C/EBPα is causally related to the leukemogenic activity of Trib1 and Trib2. Importantly, the Trib1 mutant lacking the MEK1 binding motif also failed to degrade C/EBPα,(14) suggesting degradation of C/EBPα by Trib1 is dependent on activation of the MEK1/ERK pathway, which was further confirmed by the finding that MEK inhibitor U0126 suppressed degradation of C/EBPα by Trib1.
Several studies have shown that the MAPK pathway affects C/EBPα function. MAPK activation induced by the BCR-ABL fusion protein upregulates hnRNP E2 and downregulates miR-328.(44)C/EBPα mRNA is one of the targets of hnRNP E2 transcriptional repression, and the translational activity of C/EBPα is restored when MAPK activity is blocked (Fig. 3A). Furthermore, using AML cell lines that possess FLT3 mutations showing constitutive activation of the MAPK pathway, phosphorylation of C/EBPα at serine 21 was induced and granulocytic differentiation was suppressed.(45) Conversely, inhibition of the MAPK pathway with a specific inhibitor PD98059 or overexpression of the C/EBPα S21A mutant induced granulocytic differentiation. These results support the role of MAPK signaling in C/EBPα function, and Trib1 and Trib2 could mediate that pathway as adaptors. Another important motif for C/EBPα degradation was observed as a conserved LRDLKLRK motif within the pseudokinase domain.(42)
A potential oncogenic role of TRIB1 has been suggested in human AML. On chromosome 8q24 TRIB1 is located 1.5 Mb from c-MYC. It has been believed that the target of 8q24 amplification in human AML is c-MYC. However, Storlazzi et al. have shown that TRIB1 is overexpressed at least in some cases of AML with 8q24 amplification while c-MYC expression is not detected, and similar results were obtained in AML with ring chromosome 8.(46,47) Nevertheless, upregulation of MYC is undoubtedly important in AML, suggesting cooperative roles between TRIB1 and MYC.(48)
In search of genetic abnormalities that involve tribbles in human leukemia, we recently identified a somatic point mutation of TRIB1 in a case of Down syndrome (DS)-related acute megakaryocytic leukemia (AMKL) (Yokoyama T, Toki T, Aoki Y, Kanezaki R, Park M, Kanno Y, Takahara T, Yamazaki Y, Ito E, Hayashi Y and Nakamura T, unpublished data, 2011). The G : T point mutation was observed within the pseudokinase domain resulting in amino acid conversion from arginine to leucine (R107L). When the R107L mutant was expressed in mouse BM cells and transferred into lethally irradiated recipients, acceleration of AML development and further enhancement of MAPK phosphorylation was observed, suggesting that the mutation is a gain-of-function mutation. This case exhibited a loss of function mutation of GATA1 in its blastic phase.(49) Interestingly, the TRIB1 mutation preceded the GATA1 mutation at the hematopoietic stem cell level and emerged at a very early stage of leukemogenesis, and the TRIB1 mutation is necessary but not sufficient for leukemogenesis.
In DS-related leukemogenesis, GATA2 upregulation contributes to dysregulated megakaryocytic proliferation in the absence of GATA1.(50) In accordance with this result, we identified Gata2 as a common retroviral integration site in Trib1-induced AML (Table 2). The data strongly suggest cooperation between Trib1 and the Gata transcriptional network (Fig. 3B). Moreover, the same study identified Bcl11a as the second most frequent integration site (Table 2). Recently, Yin et al.(51) showed that Bcl11a cooperates with Nf1 mutations. The data raised the possibility that Bcl11a is a cooperative partner of the MAPK pathway where both Nf1 and Trib1 are important components (Fig. 3B).
Table 2. Common integration sites and candidate cooperative genes in Trib1-induced AML| Candidate gene | Chromosome locus | Incidence, % (n = 37) | Products | Known common site in RTCGD |
|---|
|
| Hoxa9, Hoxa7 | 6B3 | 35.1 | Transcription factor | Yes |
| Bcl11a | 11A3.2 | 24.3 | Transcription factor | Yes |
| Sox4 | 13A3-A5 | 13.5 | Transcription factor | Yes |
| Anp32b | 4B2 | 8.1 | Histone chaperone | Yes |
| Gata2 | 6D1 | 5.4 | Transcription factor | No |
| AK044786 | 2A1 | 5.4 | Unknown | No |
| 1600014C10Rik | 7B1 | 5.4 | Unknown | No |
Tribbles involvement in non-hematopoietic malignancies. Bowers et al.(25) found that TRIB3 is highly expressed in primary breast and colorectal cancer. In addition, expression of TRIB3 in human colon or prostate cancer cell lines was upregulated under hypoxic conditions, suggesting that Trib3 could be induced by HIF1α and might play a role in cellular survival in vivo. High levels of TRIB3 expression in lung, esophagus and colon cancer were also reported.(52) Another study indicated that a high level of TRIB3 expression was correlated with high metastatic activity and a poor prognosis of human colorectal cancer, and that TRIB3 knockdown suppressed growth of colorectal cancer cells.(53)
In a series of human skin cancer, TRIB2 was found overexpressed in melanoma.(54) Interestingly, the expression levels of TRIB2 correlated with cytoplasmic localization of FOXO3a, and TRIB2 knockdown restored nuclear localization of FOXO3a. The TRIB2-mediated downregulation of FOXO3a increased cell proliferation, colony formation, wound healing and in vivo tumor growth, indicating the oncogenic nature of TRIB2 in certain non-hematopoietic neoplasms.
It remains to be determined which genetic pathways and target molecules of tribbles are involved in non-hematopoietic cancers. In this respect, FOXO family transcription factors are good candidates, although more comprehensive studies in other tumor types are needed. Moreover, given the fact that TRIB3 is induced in ER stress and hypoxic conditions, ATF4/CHOP and AKT pathways are important in the treatment for tribbles-expressing cancer.