Machine Learning Discoveries of BUB1-X Synergy in ETC-1922159 Treated Colorectal Cancer Cells

1. Introduction

1.1. Spindle Assembly Checkpoint

Zirkle [1] first observed that, when just one chromosome is retarded to arrive at the metaphase plate, anaphase onset is postponed until some minutes after its arrival. Later, Hoyt et al. [2] identified mutant strains of S. cerevisiae that fail to properly arrest their cell cycles at mitosis in response to the loss of microtubule function. Their studies defined three genes (BUB) required for normal cell cycle arrest. The multidomain protein kinases BUB1 and BUBR1 are central components of the mitotic checkpoint for spindle assembly (SAC). In most colorectal cancers, a chromosomal instability (CIN) leading to an abnormal chromosome number (aneuploidy) is observed Strand et al. [3]. Cahill et al. [4] show that CIN is associated with the loss of function of a mitotic checkpoint. In some cancers displaying CIN, the loss of this checkpoint was associated with the mutational inactivation of a human homologue of the yeast BUB1 gene. The normal mitotic checkpoints of cells displaying microsatellite instability become defective upon transfer of mutant hBUB1 alleles from either of two CIN cancers.

The missegregation of sister chromatids during mitosis results in the loss or gain of chromosomes in daughter cells (aneuploidy). To avoid this disastrous outcome, BUB works in synergy with multiple proteins (Bolanos-Garcia and Blundell [5]). Many of the different synergies represented as combinations of genes/proteins have been experimentally tested and published. However, there are combinations that have yet to be explored. I address the problem of identifying these unexplored combinations via use of a machine learning based search engine in the next section.

1.2. Combinatorial Search Problem and a Possible Solution

In a recently published work Sinha [6], a frame work of a search engine was developed which can rank combinations of factors (genes/proteins) in a signaling pathway. Readers are requested to go through the adaptation of the above mentioned work for gaining deeper insight into the working of the pipeline and its use of published data set generated after administration of ETC-1922159, Sinha [7]. The work uses SVM package by Joachims [8] in https://www.cs.cornell.edu/people/tj/svm_light/svm_rank.html. I use the adaptation to rank 2^nd order gene combinations.

2. Results & Discussion

2.1. BUB1 Related Synergies

2.1.1. BUB1-MAD2/INCENP/ZWILCH/AURKB/ZWINT/BUB1B/BUB3

Johnson et al. [9] show that BUB1 plays a key role in the assembly of checkpoint proteins at the kinetochore. It is required for the subsequent localization of CENPF, BUBR1 (or BUB1B), CENPE and MAD2. Carvalhal et al. [10] state that proper kinetochore-microtubule attachments are supported by BUB1 dependent centromeric H2A-T120 phosphorylation, which promotes the recruitment of the chromosomal passenger complex, consisting of Aurora B (AURKB), inner centromere protein (INCENP), survivin, and borealin. BUB1 contributes to the recruitment of ROD-ZWILCH-ZW10 (RZZ) in human cells (Cheeseman and Desai [11] , Foley and Kapoor [12]). Wang et al. [13] show that AURKB phosphorylates HASPIN to promote generation of H3T3ph and that AURKB activity is required for normal chromosomal localization of the chromosomal passenger complex. This indicates an intimate linkage between AURKB and HASPIN functions in mitosis. They also propose that AURKB activity triggers a CPC-HASPIN-H3T3ph feedback loop that promotes generation of H3T3ph on chromatin and provide evidence that the BUB1-SGO-CPC pathway supplies a signal that boosts the CPC-HASPIN-H3T3ph feedback loop specifically at centromeres to produce the well-known accumulation of the CPC in these regions. Kiyomitsu et al. [14] show that the kinetochore target of BUBs in human cells is a kinetochore protein (blinkin), which interacts with ZWINT-1 (or ZWINT) and the two subunits hMis13 and hMis14 of the hMis12 complex. Bolanos-Garcia and Blundell [5] also state that BUBR1 (or BUB1B) forms part of the mitotic checkpoint complex (MCC) that contains BUB3, MAD2 and cell division cycle 20 (CDC20).

All the above experimental validations have been done and each of them were found to be down regulated in ETC-1922159 treated CRC. Individual recordings of these down regulations have been documented. Using the adaptation of the above mentioned search engine, I was able to rank 2^nd order combination of MAD2 / INCENP / ZWILCH / AURKB / ZWINT / BUB1B / BUB3 with BUB1, that were down regulated.

Table 1 shows rankings of these combinations. Followed by this is the unexplored combinatorial hypotheses in Table 2 generated from analysis of the ranks in Table 1. The Table 1 shows rankings of individual members w.r.t BUB1. MAD2L1 - BUB1 shows low ranking of 227 (laplace) and 17 (rbf). MAD2L2 - BUB1 shows low ranking of 977 (laplace) and 219 (linear). INCENP - BUB1 shows low ranking of 959 (laplace), 842 (linear) and 1188 (rbf). ZWILCH - BUB1 shows low ranking of 345 (laplace) and 1561 (linear). AURKB - BUB1 shows low ranking of 97 (laplace) and 491 (linear). ZWINT - BUB1 shows low ranking of 219 (laplace), 187 (linear) and 27 (rbf). BUB3 - BUB1 shows low ranking of 1657 (laplace) and 1042 (rbf). These rankings point to the synergy existing between the two components, which have been down regulated after the drug treatment. AURKA and BUB1B did not show synergistic down regulation with BUB1.

One can also interpret the results of the Table 1 graphically, with the following influences - • Individual members w.r.t BUB1 with BUB1 $->$ MAD2-L1/L2; BUB1 $->$ INCENP; BUB1 $->$ ZWILCH; BUB1 $->$ AURKB; BUB1 $->$ ZWINT; and BUB1 $->$ BUB3.

2.1.2. BUB1-CENP Family

Johnson et al. [9] show that BUB1 plays a key role in the assembly of checkpoint proteins at the kinetochore. It is required for the subsequent localization of CENPF, BUBR1 (or BUB1B), CENPE and MAD2. Johnson et al. [9] also mention a number of proteins, including CENPA, CENPC and CENPI, that localize to the centromere region throughout the cell cycle and many other proteins only localize to kinetochores transiently during mitosis, while pointing to relevant literature. All the above experimental validations have been done and each of them were found to be down regulated in ETC-1922159 treated CRC. Individual recordings of these down regulations have been documented. I was able to rank 2^nd order combination of CENP family members with BUB1, that were down regulated. Rankings confirm some of the existing combinations while giving insight about unexplored combinations also.

Table 3 shows rankings of these combinations. Followed by this is the unexplored combinatorial hypotheses in Table 4 generated from analysis of the ranks in Table 3. The Table 3 shows rankings of CENP family w.r.t BUB1. CENPH - BUB1 shows low ranking of 43 (laplace), 261 (linear) and 482 (rbf). CENPA - BUB1 shows low ranking of 67 (laplace), 41 (linear) and 31 (rbf). CENPM - BUB1 shows low ranking of 90 (laplace), 1492 (linear) and 1105 (rbf). CENPJ - BUB1 shows low ranking of 142 (laplace), 69 (linear) and 1484 (rbf). CENPU - BUB1 shows low ranking of 342 (laplace), 632 (linear) and 852 (rbf). CENPE - BUB1 shows low ranking of 385 (laplace) and 105 (linear). CENPF - BUB1 shows low ranking of 606 (laplace), 430 (linear) and 1419 (rbf). CENPK - BUB1 shows low ranking of 607 (laplace), 3 (linear) and 1371 (rbf). CENPI - BUB1 shows low ranking of 672 (laplace), 170 (linear). CENPW - BUB1 shows low ranking of 730 (laplace), 64 (linear) and 304 (rbf). CENPN - BUB1 shows low ranking of 917 (laplace), 482 (linear) and 447 (rbf). These rankings point to the synergy existing between the two components, which have been down regulated after the drug treatment. CENPO, CENPL and CENPV did not show synergistic down regulation with BUB1.

One can also interpret the results of the Table 3 graphically, with the following influences - • CENP family w.r.t BUB1 with BUB1 $->$ CENP-H/A/M/J/U/E/F/K/I/W/N.

2.1.3. BUB1-CDC Family

Goto et al. [15] suggest that CDC28 mediates phosphorylation of BUB1 at T566. This plays an important role for the degradation of BUB1 in anaphase, and the phosphorylation is required for adaptation of the spindle checkpoint to prolonged mitotic arrest. Bolanos-Garcia and Blundell [5] depict a model where BUB1 KEN boxes mediate BUB1 degradation by the APC/C–CDH1 in the G1 phase of the cell cycle and BUB1–CDC20 binding. Phosphorylation of CDC20 S153 by BUB1 is required for efficient SAC activity. All the above experimental validations have been done and each of them were found to be down regulated in ETC-1922159 treated CRC. Individual recordings of these down regulations have been documented. I was able to rank 2^nd order combination of CDC family members with BUB1, that were down regulated. Rankings confirm some of the existing combinations while giving insight about unexplored combinations also.

Table 5 shows rankings of these combinations. Followed by this is the unexplored combinatorial hypotheses in Table 6 generated from analysis of the ranks in Table 5. The Table 5 shows rankings of CDC family w.r.t BUB1. CDC45 - BUB1 shows low ranking of 12 (laplace) and 228 (rbf). CDC25C - BUB1 shows low ranking of 53 (laplace), 303 (linear) and 188 (rbf). CDC20 - BUB1 shows low ranking of 253 (laplace) and 362 (rbf). CDC7 - BUB1 shows low ranking of 305 (laplace), 34 (linear) and 853 (rbf). CDC6 - BUB1 shows low ranking of 647 (laplace), 776 (linear) and 444 (rbf). CDC25A - BUB1 shows low ranking of 853 (laplace) and 447 (linear). These rankings point to the synergy existing between the two components, which have been down regulated after the drug treatment. CDC23 and CDC123 did not show synergistic down regulation with BUB1.

One can also interpret the results of the Table 5 graphically, with the following influences - • CDC family w.r.t BUB1 with BUB1 $->$ CDC-45/25C/20/7/6/25A.

2.1.4. BUB1-TGF$\beta$ Family

Nyati et al. [16] identified BUB1 as a mediator of TGF$\beta$ signaling. BUB1 interacted with TGFBRI in response to stimulation with TGF$\beta$ and promoted the heterodimerization of TGFBRI and TGFBRII. Nyati et al. [17] demonstrate that TGFBR2 phosphorylates BUB1 at Serine-318. All the above experimental validations have been done and each of them were found to be down regulated in ETC-1922159 treated CRC. Individual recordings of these down regulations have been documented. I was able to rank 2^nd order combination of TGF$\beta$ family members with BUB1, that were down regulated.

Table 7 shows rankings of these combinations. Followed by this is the unexplored combinatorial hypotheses in Table 8 generated from analysis of the ranks in Table 7. The Table 7 shows rankings of TGF$\beta$ family w.r.t BUB1. TGFBR3 - BUB1 shows low ranking of 1066 (laplace), 890 (linear) and 1525 (rbf). TGFB1 - BUB1 shows low ranking of 1350 (linear) and 1000 (rbf). These rankings point to the synergy existing between the two components, which have been down regulated after the drug treatment. TGFBRAP1 did not show synergistic down regulation with BUB1.

One can also interpret the results of the Table 7 graphically, with the following influences - • TGF$\beta$ family w.r.t BUB1 with BUB1 $->$ TGF$\beta$-R3/1.

2.1.5. BUB1-TOP Family

Chromosome segregation in mitosis requires the removal of catenation between sister chromatids. Timely decatenation of sister DNAs at mitotic centromeres by topoisomerase IIa (TOP2A) is crucial to maintain genomic stability. Zhang et al. [18] show that BUB1 generated modification, that is histone H2A Thr-120 phosphorylation (H2ApT120), is necessary and sufficient for the centromeric localization of TOP2A. Individual recordings of these down regulations have been documented. All the above experimental validations have been done and each of them were found to be down regulated in ETC-1922159 treated CRC. I was able to rank 2^nd order combination of TOP family members with BUB1, that were down regulated.

Table 9 shows rankings of these combinations. Followed by this is the unexplored combinatorial hypotheses in Table 10 generated from analysis of the ranks in Table 9. The Table 9 shows rankings of TOP family w.r.t BUB1. TOP2A - BUB1 shows low ranking of 392 (laplace), 423 (linear) and 984 (rbf). TOP1MT - BUB1 shows low ranking of 884 (laplace) and 476 (linear). These rankings point to the synergy existing between the two components, which have been down regulated after the drug treatment. TOP2B and TOPBP1 did not show synergistic down regulation with BUB1.

One can also interpret the results of the Table 9 graphically, with the following influences - • TOP family w.r.t BUB1 with BUB1 $->$ TOP-2A/1MT.

2.1.6. BUB1-SGO Family

Salic et al. [19] demonstrate that the vertebrate SGO localizes to kinetochores and is required to prevent premature sister centromere separation in mitosis. Further, Tang et al. [20] show that human BUB1 protects centromeric sister-chromatid cohesion through SGO during mitosis. All the above experimental validations have been done and each of them were found to be down regulated in ETC-1922159 treated CRC. Individual recordings of these down regulations have been documented. I was able to rank 2^nd order combination of SGO family members with BUB1, that were down regulated.

Table 11 shows rankings of these combinations. Followed by this is the unexplored combinatorial hypotheses in Table 12 generated from analysis of the ranks in Table 11. The Table 11 shows rankings of SGO family w.r.t BUB1. SGOL1 - BUB1 shows low ranking of 336 (laplace) and 123 (linear). SGOL2 - BUB1 shows low ranking of 537 (laplace) and 115 (linear). SGOL1.AS1 - BUB1 shows low ranking of 637 (laplace), 4 (linear) and 1494 (rbf). These rankings point to the synergy existing between the two components, which have been down regulated after the drug treatment. SGO2B and SGOBP1 did not show synergistic down regulation with BUB1.

One can also interpret the results of the Table 11 graphically, with the following influences - • SGO family w.r.t BUB1 with BUB1 $->$ SGO-L1/L2/L1.AS1.

2.1.7. BUB1-KIF Family

Jiang et al. [21] in their study, found that BUB1 expression was increased in human bladder cancer (BCa). They identified that BUB1 interacted directly with STAT3 and mediated the phosphorylation of STAT3 at Ser727. Finally, they also found that the BUB1/STAT3 complex promoted the transcription of STAT3 target genes. To confirm this conclusion, they performed qRT–PCR by knocking down BUB1 with siRNA, and the results showed that the transcriptional activity of STAT3 was significantly reduced and that the expression of STAT3 target genes, like E2F3, KIF18B, etc. , was diminished. Individual recordings of these down regulations have been documented. All the above experimental validations have been done and each of them were found to be down regulated in ETC-1922159 treated CRC. I was able to rank 2^nd order combination of KIF family members with BUB1, that were down regulated.

Table 13 shows rankings of these combinations. Followed by this is the unexplored combinatorial hypotheses in Table 14 generated from analysis of the ranks in Table 13. The Table 13 shows rankings of KIF family w.r.t BUB1. KIF18B - BUB1 shows low ranking of 14 (laplace), 1575 (linear) and 310 (rbf). KIF15 - BUB1 shows low ranking of 19 (laplace), 92 (linear) and 238 (rbf). KIF20A - BUB1 shows low ranking of 224 (laplace) and 117 (linear). KIF22 - BUB1 shows low ranking of 252 (laplace), 121 (linear) and 424 (rbf). KIF2C - BUB1 shows low ranking of 254 (laplace), 110 (linear) and 1197 (rbf). KIF20B - BUB1 shows low ranking of 258 (laplace) and 824 (linear). KIF18A - BUB1 shows low ranking of 285 (laplace), 47 (linear) and 10 (rbf). KIF4A - BUB1 shows low ranking of 362 (laplace) and 120 (linear). KIF23 - BUB1 shows low ranking of 399 (laplace) and 1111 (rbf). KIF9 - BUB1 shows low ranking of 497 (laplace), 580 (linear) and 1530 (rbf). KIF11 - BUB1 shows low ranking of 541 (laplace) and 1362 (linear). KIF7 - BUB1 shows low ranking of 1063 (laplace), 1171 (linear) and 616 (rbf). KIF27 - BUB1 shows low ranking of 1362 (linear) and 633 (rbf). These rankings point to the synergy existing between the two components, which have been down regulated after the drug treatment. KIF14 and KIF13A did not show synergistic down regulation with BUB1.

One can also interpret the results of the Table 13 graphically, with the following influences - • KIF family w.r.t BUB1 with BUB1 $->$ KIF-18B/15/20A/22/2C/20B/18A/4A/23/9/11/7/27.

2.1.8. BUB1-CDK Family

Vazquez-Novelle et al. [22] show that CDK1 inactivation terminates mitotic checkpoint surveillance and stabilizes kinetochore attachments in anaphase. Ji et al. [23] show that phosphorylation of BUB1 by CDK1 is required for the spindle checkpoint. All the above experimental validations have been done and each of them were found to be down regulated in ETC-1922159 treated CRC. Individual recordings of these down regulations have been documented. Using the adaptation of the above mentioned search engine, I was able to rank 2^nd order combination of CDK family members with BUB1, that were down regulated.

Table 15 shows rankings of these combinations. Followed by this is the unexplored combinatorial hypotheses in Table 16 generated from analysis of the ranks in Table 15. The Table 15 shows rankings of CDK family w.r.t BUB1. CDK20 - BUB1 shows low ranking of 1070 (laplace) and 577 (rbf). CDK6 - BUB1 shows low ranking of 1005 (linear) and 426 (rbf). These rankings point to the synergy existing between the two components, which have been down regulated after the drug treatment. CDK5-RAP1/RAP2 and CDK4 did not show synergistic down regulation with BUB1.

One can also interpret the results of the Table 15 graphically, with the following influences - • CDK family w.r.t BUB1 with BUB1 $->$ CDK-20/6.

3. Conclusion

Presented here are a range of multiple synergistic BUB1 2^nd order combinations that were ranked via a machine learning based search engine. Via majority voting across the ranking methods, it was possible to find plausible unexplored synergistic combinations of BUB1-X that might be prevalent in CRC cells after treatment with ETC-1922159 drug.