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

1. Introduction

1.1. Wnt Secretion

Contrary to the signaling phenomena, the secretion phenomena is about the release and transportation of the WNT protein/ligand in and out of the cell, respectively. Briefly, the WNT proteins that are synthesized with the endoplasmic reticulum (ER), are known to be palmitoyleated via the Porcupine (PORCN) to form the WNT ligand, which is then ready for transportation Tanaka et al. [1]. It is believed that these ligands are then transported via the EVI/WNTLESS transmembrane complex out of the cell (Banziger et al. [2], Bartscherer et al. [3] and Goodman et al. [4]). The EVI/WNTLESS themselves are known to reside in the Golgi bodies and interaction with the WNT ligands for the later’s glycosylation Kurayoshi et al. [5] & Gao and Hannoush [6]. Once outside the cell, the WNTs then interact with the cell receptors, as explained in the foregoing paragraph, to induce the Wnt signaling. Of importance is the fact that the EVI/WNTLESS also need a transporter in the from of a complex termed as Retromer.

Voloshanenko et al. [7] show that WLS and WNT3 are highly expressed in colon carcinomas and EVI/WLS is required for high levels of WNT pathway activation. Further, Chua et al. [8] demonstrate that clinical osteosarcoma samples show high WLS and $\beta$-catenin expression. In colorectal cancer cells treated with ETC-1922159, WLS was found to be up regulated along with other genes. Some of the WLS-X (X, a particular gene/protein) combinations have been experimentally validated, while many remain untested/unexplored. I use the machine learning based search engine (the next section) to rank/prioritize these combinations to reveal untested/unexplored combinations.

1.2. Combinatorial Search Problem and a Possible Solution

In a recently published work Sinha [9], 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 [10]. The work uses SVM package by Joachims [11] 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. WLS Related Synergies

2.1.1. WLS-SEC

Sun et al. [12] show that WLS forms a complex with SEC12. Binding of mature WNT to WLS increases WLS-SEC12 interaction and promotes association of WLS with SAR1 (a key activator of the COPII machinery). Mutant WLS that fail to communicate with the COPII machinery cannot effectively support WNT secretion. In colorectal cancer cells treated with ETC-1922159, SEC family and WLS, were found to be up regulated and recorded independently. I was able to rank 2^nd order combination of SEC family and WLS, that were up 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 SEC family w.r.t WLS. SEC24C - WLS shows low ranking of 1560 (laplace) and 622 (rbf). This low ranking points to the fact that the combination is not relevant after ETC-1922159 treatment of CRC, however, it might be prevalent in CRC, before treatment.

Further, SEC24D - WLS show high ranking of 1644 (laplace), 2092 (linear) and 1679 (rbf). SEC31A - WLS show high ranking of 1856 (laplace) and 1793 (linear). This high ranking points to the fact that the combination is relevant after ETC-1922159 treatment of CRC, however, it might not be prevalent in CRC, before treatment.

One can also interpret the results of the Table 1 graphically, with the following influences - • SEC family w.r.t WLS with WLS $->$ SEC-24C/24D/31A.

2.1.2. WLS-RAB

Das et al. [13] show that WNT secretion is dependent on RAB8A mediated transport of GPR177 (WNTLESS). GPR177 was found to bind with RAB8A, depletion of which compromised GPR177 traffic, thereby weakening the secretion of multiple WNTs. Further, Sun et al. [12] tabulate proteomic identification of RAB family in wild-type WLS and ${\mathrm{WLS}}^{1-491}$ immunoprecipitates. In colorectal cancer cells treated with ETC-1922159, RAB family and WLS, were found to be up regulated and recorded independently. I was able to rank 2^nd order combination of RAB family and WLS, that were up regulated.

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 RAB family w.r.t WLS. RAB24 - WLS shows low ranking of 249 (laplace), 739 (linear) and 1344 (rbf). RAB3B - WLS shows low ranking of 642 (laplace), 459 (linear) and 828 (rbf). RAB22A - WLS shows low ranking of 757 (laplace), 657 (linear) and 1078 (rbf). RAB5A - WLS shows low ranking of 771 (laplace) and 267 (rbf). RAB9A - WLS shows low ranking of 858 (laplace), 1562 (linear) and 371 (rbf). RAB4B - WLS shows low ranking of 1273 (laplace) and 572 (linear). RAB8A - WLS shows low ranking of 919 (linear) and 917 (rbf). RAB1A - WLS shows low ranking of 658 (linear) and 1535 (rbf). This low ranking points to the fact that the combination is not relevant after ETC-1922159 treatment of CRC, however, it might be prevalent in CRC, before treatment.

Further, RAB25 - WLS show high ranking of 1939 (linear) and 2498 (rbf). RAB3GAP1 - WLS show high ranking of 2438 (linear) and 2103 (rbf). RAB11FIP1 - WLS show high ranking of 1561 (laplace) and 1653 (linear). RAB7A - WLS show high ranking of 1870 (laplace) and 2465 (rbf). RAB11A - WLS show high ranking of 2135 (laplace) and 2148 (linear). RAB1B - WLS show high ranking of 2237 (laplace), 2002 (linear) and 1730 (rbf). This high ranking points to the fact that the combination is relevant after ETC-1922159 treatment of CRC, however, it might not be prevalent in CRC, before treatment.

One can also interpret the results of the Table 3 graphically, with the following influences - • RAB family w.r.t WLS with WLS $->$ RAB-24/3B/22A/5A/9A/4B/8A/1A (before ETC-1922159 treatment of CRC) and WLS $->$ RAB-25/3GAP1/11FIP1/7A/11A/1B (after ETC-1922159 treatment of CRC).

2.1.3. WLS-VPS

Belenkaya et al. [14] examined the role of VPS35 (a retromer subunit), in WNT signaling. They provide compelling evidence that VPS35 colocalizes in endosomes and interacts with WLS and WLS becomes unstable in the absence of retromer activity. Further, Sun et al. [12] tabulate proteomic identification of VPS family in wild-type WLS and ${\mathrm{WLS}}^{1-491}$ immunoprecipitates. In colorectal cancer cells treated with ETC-1922159, VPS family and WLS, were found to be up regulated and recorded independently. I was able to rank 2^nd order combination of VPS family and WLS, that were up regulated.

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 VPS family w.r.t WLS. VPS37C - WLS shows low ranking of 1305 (laplace), 1381 (linear) and 1208 (rbf). VPS33B - WLS shows low ranking of 967 (linear) and 604 (rbf). This low ranking points to the fact that the combination is not relevant after ETC-1922159 treatment of CRC, however, it might be prevalent in CRC, before treatment.

Further, VPS37B - WLS show high ranking of 1715 (laplace), 2182 (linear) and 1821 (rbf). VPS28 - WLS show high ranking of 2091 (laplace) and 2453 (rbf). VPS4B - WLS show high ranking of 2487 (laplace) and 1872 (rbf). This high ranking points to the fact that the combination is relevant after ETC-1922159 treatment of CRC, however, it might not be prevalent in CRC, before treatment.

One can also interpret the results of the Table 5 graphically, with the following influences - • VPS family w.r.t WLS with WLS $->$ VPS-37C/33B (before ETC-1922159 treatment of CRC) and WLS $->$ VPS-37B/28/4B (after ETC-1922159 treatment of CRC).

2.1.4. WLS-SNX

Harterink et al. [15] found that SNX3 has an evolutionarily conserved function in WLS recycling and WNT secretion. Brown et al. [16] show similar findings about SNX3 and WLS in mammalian neural tube closure. In colorectal cancer cells treated with ETC-1922159, SNX family and WLS, were found to be up regulated and recorded independently. I was able to rank 2^nd order combination of SNX family and WLS, that were up 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 SNX family w.r.t WLS. SNX9 - WLS shows low ranking of 229 (laplace), 19 (linear) and 1501 (rbf). SNX11 - WLS shows low ranking of 1159 (laplace) and 1353 (rbf). This low ranking points to the fact that the combination is not relevant after ETC-1922159 treatment of CRC, however, it might be prevalent in CRC, before treatment.

Further, SNX33 - WLS show high ranking of 1948 (laplace) and 2378 (linear). This high ranking points to the fact that the combination is relevant after ETC-1922159 treatment of CRC, however, it might not be prevalent in CRC, before treatment.

One can also interpret the results of the Table 7 graphically, with the following influences - • SNX family w.r.t WLS with WLS $->$ SNX-9/11 (before ETC-1922159 treatment of CRC) and WLS $->$ SNX-33 (after ETC-1922159 treatment of CRC).

2.1.5. WLS-ARF

Yu et al. [17] show that WLS Golgi-to-ER retrieval requires the COPI regulator ARF as well as ERGIC2. In colorectal cancer cells treated with ETC-1922159, ARF family and WLS, were found to be up regulated and recorded independently. I was able to rank 2^nd order combination of ARF family and WLS, that were up 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 ARF family w.r.t WLS. ARFGAP3 - WLS shows low ranking of 480 (laplace) and 1075 (rbf). ARF3 - WLS shows low ranking of 866 (laplace), 736 (linear) and 516 (rbf). ARF6 - WLS shows low ranking of 1291 (laplace) and 855 (rbf). ARF1 - WLS shows low ranking of 1526 (laplace) and 1170 (linear). This low ranking points to the fact that the combination is not relevant after ETC-1922159 treatment of CRC, however, it might be prevalent in CRC, before treatment.

Further, ARF4 - WLS shows high ranking of 2040 (laplace) and 2414 (linear). This high ranking points to the fact that the combination is relevant after ETC-1922159 treatment of CRC, however, it might not be prevalent in CRC, before treatment.

One can also interpret the results of the Table 9 graphically, with the following influences - • ARF family w.r.t WLS with WLS $->$ ARF-GAP3/3/6/1 (before ETC-1922159 treatment of CRC) and WLS $->$ ARF-4 (after ETC-1922159 treatment of CRC).

2.1.6. WLS-UBE2

Wolf et al. [18] found that EVI/WLS is ubiquitylated and degraded in cells irrespective of their level of WNT production. This ubiquitylation is mediated by the E2 ubiquitin-conjugating enzymes UBE2K, UBE2J2 and UBE2N. In colorectal cancer cells treated with ETC-1922159, UBE2 family and WLS, were found to be up regulated and recorded independently. I was able to rank 2^nd order combination of UBE2 family and WLS, that were up 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 UBE2 family w.r.t WLS. UBE2H - WLS shows low ranking of 868 (laplace), 1409 (linear) and 1051 (rbf). UBE2J1 - WLS shows low ranking of 957 (laplace) and 1346 (linear). UBE2F - WLS shows low ranking of 1379 (laplace), 356 (linear) and 221 (rbf). This low ranking points to the fact that the combination is not relevant after ETC-1922159 treatment of CRC, however, it might be prevalent in CRC, before treatment.

Further, UBE2A - WLS shows high ranking of 1865 (laplace) and 2340 (linear). UBE2Z - WLS shows high ranking of 2032 (laplace) and 2265 (rbf). UBE2B - WLS shows high ranking of 2353 (laplace) and 1925 (rbf). This high ranking points to the fact that the combination is relevant after ETC-1922159 treatment of CRC, however, it might not be prevalent in CRC, before treatment.

One can also interpret the results of the Table 11 graphically, with the following influences - • UBE2 family w.r.t WLS with WLS $->$ UBE2-H/J1/F (before ETC-1922159 treatment of CRC) and WLS $->$ UBE2-A/Z/B (after ETC-1922159 treatment of CRC).

2.1.7. WLS-ATPases

McGough et al. [19] demonstrate the role of SNX3 for WNTLESS transport and report that SNX3 associates with a membrane remodelling complex composed of MON2, DOPEY2 and the putative aminophospholipid translocase, ATP9A. ATP9A comes under the category of one of the P-type ATPases (under the general category of ATPases). In colorectal cancer cells treated with ETC-1922159, ATP family and WLS, were found to be up regulated and recorded independently. I was able to rank 2^nd order combination of ATP family and WLS, that were up 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 ATP family w.r.t WLS. ATP1B1 - WLS shows low ranking of 302 (laplace) and 241 (rbf). ATP2B4 - WLS shows low ranking of 303 (laplace), 236 (linear) and 335 (rbf). ATP2A2 - WLS shows low ranking of 340 (laplace), 1116 (linear) and 76 (rbf). ATP13A2 - WLS shows low ranking of 657 (laplace), 306 (linear) and 1173 (rbf). ATP2B1 - WLS shows low ranking of 778 (laplace) and 346 (rbf). This low ranking points to the fact that the combination is not relevant after ETC-1922159 treatment of CRC, however, it might be prevalent in CRC, before treatment.

Further, ATP6V1E1 - WLS shows high ranking of 1660 (laplace) and 1924 (rbf). ATP10B - WLS shows high ranking of 1810 (laplace) and 1527 (rbf). ATP6V1D - WLS shows high ranking of 1897 (laplace) and 1807 (linear). ATP11B - WLS shows high ranking of 1929 (laplace), 2336 (linear) and 2421 (rbf). ATP6V0D1 - WLS shows high ranking of 2435 (laplace), 1724 (linear) and 2289 (rbf). This high ranking points to the fact that the combination is relevant after ETC-1922159 treatment of CRC, however, it might not be prevalent in CRC, before treatment.

One can also interpret the results of the Table 13 graphically, with the following influences - • ATP family w.r.t WLS with WLS $->$ ATP-1B1/2B4/2A2/13A2/2B1 (before ETC-1922159 treatment of CRC) and WLS $->$ ATP-6V1E1/10B/6V1D/11B/6V0D1 (after ETC-1922159 treatment of CRC).

2.1.8. WLS-TMEM

Li and Niswander [20] discovered a novel regulator of WNT pathway called TMEM132A. They show physical evidence and functional interaction of TMEM132A with WLS. In colorectal cancer cells treated with ETC-1922159, TMEM family and WLS, were found to be up regulated and recorded independently. I was able to rank 2^nd order combination of TMEM family and WLS, that were up 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 TMEM family w.r.t WLS. TMEM61 - WLS shows low ranking of 141 (laplace), 107 (linear) and 748 (rbf). TMEM45A - WLS shows low ranking of 149 (laplace), 1276 (linear) and 534 (rbf). TMEM120B - WLS shows low ranking of 166 (laplace), 53 (linear) and 169 (rbf). TMEM40 - WLS shows low ranking of 172 (laplace), 114 (linear) and 879 (rbf). TMEM86A - WLS shows low ranking of 203 (laplace), 664 (linear) and 614 (rbf). TMEM50B - WLS shows low ranking of 224 (laplace), 1411 (linear) and 464 (rbf). TMEM63B - WLS shows low ranking of 300 (laplace) and 236 (rbf). TMEM171 - WLS shows low ranking of 435 (laplace), 139 (linear) and 484 (rbf). TMEM82 - WLS shows low ranking of 441 (laplace) and 334 (linear). TMEM253 - WLS shows low ranking of 494 (laplace) and 896 (linear). TMEM45B - WLS shows low ranking of 506 (laplace), 14 (linear) and 347 (rbf). TMEM92 - WLS shows low ranking of 672 (laplace), 248 (linear) and 17 (rbf). TMEM127 - WLS shows low ranking of 764 (laplace), 1117 (linear) and 1219 (rbf). TMEM120A - WLS shows low ranking of 1000 (laplace), 868 (linear) and 43 (rbf). TMEM164 - WLS shows low ranking of 1149 (laplace) and 1048 (rbf). TMEM150B - WLS shows low ranking of 1165 (laplace) and 1360 (linear). TMEM106A - WLS shows low ranking of 1166 (laplace), 162 (linear) and 1095 (rbf). TMEM62 - WLS shows low ranking of 1173 (laplace) and 796 (rbf). TMEM159 - WLS shows low ranking of 1205 (laplace) and 1551 (rbf). TMEM139 - WLS shows low ranking of 1235 (laplace) and 1127 (linear). TMEM54 - WLS shows low ranking of 1302 (laplace), 1080 (linear) and 1371 (rbf). TMEM217 - WLS shows low ranking of 1432 (laplace), 530 (linear) and 291 (rbf). TMEM176B - WLS shows low ranking of 892 (linear) and 313 (rbf). TMEM229B - WLS shows low ranking of 1392 (linear) and 83 (rbf). TMEM51-AS1 - WLS shows low ranking of 1389 (linear) and 1109 (rbf). TMEM8A - WLS shows low ranking of 832 (linear) and 1151 (rbf). This low ranking points to the fact that the combination is not relevant after ETC-1922159 treatment of CRC, however, it might be prevalent in CRC, before treatment.

Further, TMEM79 - WLS shows high ranking of 2278 (linear) and 1822 (rbf). TMEM184A - WLS shows high ranking of 1545 (laplace) and 2070 (rbf). TMEM65 - WLS shows high ranking of 1581 (laplace) and 2121 (linear). TMEM185A - WLS shows high ranking of 1666 (laplace) and 1561 (linear). TMEM234 - WLS shows high ranking of 1736 (laplace) and 1678 (rbf). TMEM57 - WLS shows high ranking of 2005 (laplace), 2106 (linear) and 2254 (rbf). TMEM44 - WLS shows high ranking of 2179 (laplace), 2154 (linear) and 2454 (rbf). TMEM176A - WLS shows high ranking of 2255 (laplace), 1699 (linear) and 1588 (rbf). TMEM220 - WLS shows high ranking of 2292 (laplace), 2468 (linear) and 2428 (rbf). TMEM30B - WLS shows high ranking of 2297 (laplace) and 2108 (rbf). TMEM184B - WLS shows high ranking of 2308 (laplace), 2381 (linear) and 1706 (rbf). TMEM140 - WLS shows high ranking of 2342 (laplace), 1862 (linear) and 2347 (rbf). TMEM2 - WLS shows high ranking of 2472 (laplace), 1865 (linear) and 1978 (rbf). TMEM31 - WLS shows high ranking of 2513 (laplace), 2481 (linear) and 2305 (rbf). This high ranking points to the fact that the combination is relevant after ETC-1922159 treatment of CRC, however, it might not be prevalent in CRC, before treatment.

One can also interpret the results of the Table 15 graphically, with the following influences - • TMEM family w.r.t WLS with WLS $->$ TMEM-61 / 45A / 120B / 40 / 86A / 50B / 63B / 171 / 82 / 253 / 45B / 92 / 127 / 120A / 164 / 150B / 106A / 62 / 159 / 139 / 54 / 217 / 176B / 229B / 51-AS1 / 8A (before ETC-1922159 treatment of CRC) and WLS $->$ TMEM-79 / 184A / 65 / 185A / 234 / 57 / 44 / 176A / 220 / 30B / 184B / 140 / 2 / 31 (after ETC-1922159 treatment of CRC).

3. Conclusion

Presented here are a range of multiple synergistic WLS 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 WLS-X that might be prevalent in CRC cells after treatment with ETC-1922159 drug.