It might happen due to a poor connection of the kinetochore to the microtubule, which won't be able to pull the chromosome to one of the poles correctly. This could generate 3n/n or even 4n/0n clones.

Or a failure of the duplicated chromosomes to separate properly (the enzyme seperase hydrolyses cohesin which holds the chromosomes together).

Yes, it is possible. This situation is possible, especially in cancer cells, via binuclear step after pathological cytokinesis. Next divisions of each nucleus could be in different time.

I agree with Vera, improper segregation with a pathological cytokinesis(PC).  Possibly endoreduplication at one pole after incomplete segregation and PC or asynchronous replication after PC.  In one of my papers I found a chromosome that would replicate independently of the normal diploid set of chromosomes, I don't see why this couldn't happen  to specific "haploid" set after mitosis and PC.  I can forsee many steps in mitosis that could cause this interesting phenomenon.

Thank you all! But I still don't understand it: all of you suggest that the replication happens regularily. At the metaphase, the chromosomes will align at the center of the cells (46 pairs - 2 of which belong together, each somehow accidentially positioned). By the possible mechanisms you have suggested, HOW could it be, that accidentially, one set of chromosomes keep aggragated and exactly the other set get seperated, through PC, that exactly one set in one pole and 3 sets in another pole, generating a 3N/1N? For me, it could be anything instead of 3N/1N. Would you guys there give me more details? It would be great if you can make it in a picture.

Yes. it can happen. In cancer cells, there is often overexpression/mutation of "PERICENTRIN" (a centrosomal protein) or respective gene. Overexpression of pericentrin  would lead to multipolar  spindle formation due to aberrant centrosomal copies and . Thereby, higher chances of aneuploidy (3N) due to non-disjunction of sister chromatids leading to mutlipolar segregation  of chromosomes.

Refer the follwing link:

Hope this answers your doubts.

http://www.nature.com/leu/journal/v17/n11/full/2403142a.html

If I understand your question, you're not asking about specific mechanisms (there are plenty of ways for cell division to fail). I think the key to understanding how this happens is that typically, massive cell division defects cause polyploidy first, often repeatedly. So you might have 4 or more copies of a particular chromosome in the same mother cell. Then, chromosome loss occurs, and you end up with 3 copies. There are some studies that show this is probably how cancer and stem cells end up selecting for chromosomes that give them a growth advantage, while discarding extra copies of chromosomes that contain genes that would limit cell division. see for example http://bit.ly/1qaoznK (stem cells tend to have extra copies of chromosomes 12 and 17) or http://www.nature.com/nrm/journal/v5/n1/full/nrm1276.html. 

Triploid or para-triploid tumours occur rather often, these karyotypes are stable. Theoretically they form through tetraploidy by elimination a haploid set, which means that at some stage haploid sets are existing in a tetraploid multi-nucleated cell (this is disputed by some authors). Why triploid tumours happen so often and are relatively stable is a non-answered question. But it is interesting that in macroevolution tetraploidy in plants can return to diploidy through several generations by so called 'triploid bridge' (reviewed in Nature by L.Comai several years ago). For tumours this may mean the presence of a similar cycle which is very slow. This is my hypothesis, nearly nothing more, however we had some data by FISH for three different chromosomes showing that hyper-triploid HeLa after irradiation (48 hs) increase the number of 1C and 2C-4C cells on account of decrease of triploidy (not much but statistically very sure). 

We and other scientists have shown that spontaneous canine mammary tumors frequently are hypodiploid. Cell cycling of this cell population gives triploid tumor cells. We have shown this both in primary mammary carcinomas, their metastases and in cultured tumor cells.

Dear Dan, you insist: HOW could it be, that accidentially, one set of chromosomes keep aggregated and exactly the other set get separated, through PC, that exactly one set in one pole and 3 sets in another pole, generating a 3N/1N?

 I understand that you are not satisfied with explanations where random effects are supposed. Haploid sets can exist, you are right and this is not a random but programmed effect. Just now I am preparing a small article where I provide a guess, how it happens and why tumours are so often triploid. A bit of patience, it will be ready soon! 

Dear Jekaterina, thanks! It sounds facsinating! It will be very nice of you, if you can send me a copy of your publication then. I am not from the area of cytology, so I have a lot of questions based on our recent observations. Would you please give me your contact information, so that I can ask you several more questions? Best, Dan

To Eva Hellmen - I have looked through your articles. Have you ever done karyotypes for the canine breast cancers? Have you noticed any regularities between the change of a karyotype and triploidy? Do you consider now the mixed epithelial-sarcomal phenotypes typical for canine as epithelial to mesenchymal transition and do you see any change of ploidy and karyotypes associated with these parts of breast cancer?

To Dan - I shall be happy to correspond with you. [email protected]

To Vera B Dugina. It is a very right observation, in particular if to add that the next asynchronic division of each nucleus of a bi-nuclear cell is often with incomplete cytotomy, and in this way the multinuclearity is progressing (see our observations attached). The interesting question still remains by which process a haploid subnucleus  appears to make a triploid set.

Jekaterina, thank you for attached file of your paper. The role of Aurora B is very interesting for me and my students.

Dear All,

There is a nice piece of music entitled in French "Les Cavaliers d'Offenbach" ...

More seriously, I sincerely apologize to be so in late with my comments but I became "active" on RG in the last days of December 2015.

I would therefore be delighted if I could receive comments / explanations about the "tridivision" observed in the "bottom-third center" of the attached film with a SKMEL-28 melanoma cell.

Best regards

Robert 

Dear all, please get a look at our article 

Histochem Cell Biol. 2016 Feb 9. 

Disentangling the aneuploidy and senescence paradoxes: a study of triploid breast cancers non-responsive to neoadjuvant therapy. If you want more than Abstract, I shall send to you. A puzzle is why triploidy  of breast cancers is so prognostically unfavourable.

One of the potential mechanisms of emerging trisomy in cancer is parthenogenesis. 

You can get a look at our review:

http://www.impactjournals.com/oncoscience/index.php?pii=108

I do not know whether your are aware of the attached article.

Best regards

Robert

Dear all, 

Entosis (not canibalism) was disclosed as a possible mechanism of polyploidization and enabled the long-term survival of cancer cells.

You can get a look at our article: http://journals.plos.org/plosone/article?id=10.1371/journal.pone.0145016

Best regards, 

Jan

Dear Jekaterina,

I am sorry for the late answer to your questions that I am afraid that I was not aware of them.

 We have studied DNA ploidy by flow cytometry and found that many canine mammary tumours were hypodiploid. This was valid both for primary tumours especially carcinomas and their metastases as well as in cell lines. Dogs have 78 chromosomes and we studied the karyotype of the cell lines by counting the chromosomes once. The hypodiploid cell line 27 had a mean value of 56.8 chromosomes but the range was wide (Hellmén 1992). In addition to the main population of cells in the hypodiploid fraction we saw another peak with the double DNA index that we regarded as the corresponding G2 peak. Our interpretation of these findings was that the cells seemed to go through polyploidy that we also have discussed in a paper (Cornelisse et al 1994).

 Regarding EMT and canine mammary tumours, this is an interesting question. EMT appears most probably in the carcinomas but it does not explain the frequent presence of mesenchymal tissues in canine mammary tumours. So far, we have not seen any transdifferentiation neither from epithelial to mesenchymal mammary tumours nor from mesenchymal to epithelial tumours (MET) upon xenotransplants in mice. Interestingly, mammary spindle-cell tumours and mammary osteosarcomas caused tumours with different mesenchymal phenotypes but no ones with epithelial phenotypes (Hellmén et al 2000, Wensman et al 2008).

Best wishes,

Eva

Dear Jan,

Can you develop a little bit your response because I do actually not understand it!

The article I attached is by Michael Ovrholtzer (who unfortunately is not registred on RG).

The firt title of the Introduction of this article is "Cell cannibalism in human tumors: Evidence for Entosis" ...

Did you read this article?

Can you comment this article?

Can you develop your previous answer?

Thank you very much for the time you will provide to me.

Robert

 Dear Robert,

I'm sorry. The comment that I wrote may have been misleading.

We observed two processes of cell engulfment. In one of them was active cannibalic cell and other cell was cannibalized as passive target. This cell was then digested by cannibalic cell (including DNA in nucleus), formation of bird eye structure was observed. This process is by our opinion true cannibalism. The second mechanism of cell engulfment was quite different. The more active cell was the engulfed one and both nuclei in new polyploid cell were rather equal partners (binucleate state was observed), no bird eye structure was formed. This process is by our opinion entosis. Entosis could result in cannibalism, but it is not necessary outcome. When cannibalism did not happen before cell division, engulfed cell can disrupt cell division of host cell (as published in Krajcovic et al.) or it could lead to cooperation between nuclei and it could lead to ploidy increase and to more resistant phenotype.

Cannibalism: https://www.youtube.com/watch?v=ezYTh2heat8

Entosis: https://www.youtube.com/watch?v=HohpRIWVWJI

Best regards,

Jan.

Thank you very much Jan for your crystal-clear response!

Now, I got your point!

Best regards

Robert

Or, as we reported in canine mammary tumors. Initial loss of chromosomes leading to hypodiplody and a proliferation of those cells gave a triploid cell population as their G2 peak. We also found a polyploidy of those cells giving rise to another G2 peak with the double DNA index.