For example, mutation in MHS2, which encodes a protein involved in the repair of mismatches that occur during DNA replication, dramatically increases the risk of developing colon cancer. (There are many other examples, like the RB gene which encodes a tumor supressor protein and is correlated to retinoblastoma, thus its name)

My question is: how a gene like MHS2 which participates in a general mechanism of DNA repair increases the risk of a specific type of cancer? Why doesn't it increase the risk of developing cancer in other tissues as well?


I can't fault @WYSIWYG for mentioning the cited Vogelstein article in providing an answer. You point to what seems like a great explanation for why certain cancers arise in some tissues but not others. However, for those who look closely this paper has some serious errors in its derivation of the model, and for good reason it has come under strong fire in the last couple of months.

See this for a really important rebuttal: http://ameyer.me/science/2015/01/02/vogel.html

Unfortunately, the paper provides an elegantly specious explanation for, in part, the longstanding paradox of why cancers more readily arise in some tissues like the intestines and from seemingly identical mechanisms, but do not seem to arise in other tissues. My favorite example, though @El Cid you do list some good ones, are inactivating BRCA1 mutations which normally play a role in repairing double stranded DNA breaks. And yet even when mutated in the germline this only seems to increase cancer risk in the breast, and really only in one gender (females).

So the answer is still quite certainly that the field does not understand this paradox, and the cited Vogelstein paper was a total tragedy of a publication. There have been a number of officially submitted critiques of the paper and the Vogelstein lab has been trying to defend it as best they can. This is however a great example of how big name labs can seem to get anything published in Cell/Nature/Science.

Another thing to consider and @El Cid points to a good example, is the pRb mutations (in a very central tumor suppressor pathway for all cells) cause retinoblastoma and not for instance intestinal or blood cancers very readily, and the retina is not a rapidly dividing tissue. So the Vogelstein paper cannot explain this.

| improve this answer | |
  • $\begingroup$ Retinoblastoma happens in early childhood. This is a carryover of actively dividing cells during embryonic development. pRb is involved on other cancers too. $\endgroup$ – WYSIWYG Mar 7 '15 at 8:26
  • $\begingroup$ From here.... In addition to bladder cancer, somatic mutations in the RB1 gene are associated with many other types of cancer. For example, changes in the RB1 gene have been reported in some cases of lung cancer, breast cancer, a bone cancer known as osteosarcoma, and an aggressive form of skin cancer called melanoma. Somatic RB1 mutations have also been identified in some leukemias, which are cancers of blood-forming cells. Somatic RB1 mutations in cancer cells inactivate pRB so it can no longer regulate cell division effectively. $\endgroup$ – WYSIWYG Mar 7 '15 at 8:31
  • $\begingroup$ Yes, both good points. First you are right, pRb is going to be found other cancers but why it cannot sometimes be found in every Cancer is the paradox. And about retinoblastoma occurring during early age you hit on the most widely cited criticism of the aging-Cancer theory. However, it does overlook an important difference, and that is germline versus somatic mutation. The reason we often associate retinoblastoma with childhood disease is because when they carry a germline mutation, every cell in the tissue carries this, therefore they do not need a high division rate to drive oncogenesis. $\endgroup$ – The Nightman Mar 7 '15 at 14:58
  • $\begingroup$ With germline mutations it is no probably more a question of overall somatic fitness. Every cell carries an oncogenes mutation, and is therefore at a functional disadvantage thereby more likely to allow for the expansion of a newly acquired secondary mutation. $\endgroup$ – The Nightman Mar 7 '15 at 15:00
  • $\begingroup$ Yes it is perhaps not understood why germline pRb mutation is manifested (first?) in Retinoblastoma why can't it cause some other cancer. I am not really sure if pRb's connection with retinoblastoma is anecdotal or whether it predominantly is manifested in this cancer. $\endgroup$ – WYSIWYG Mar 7 '15 at 15:04

You should have a look at this interesting article published earlier this year in Science.

There are two mechanisms by which cells can accumulate mutations in DNA:

  1. Replication errors
  2. External physicochemical agents such as UV, carcinogens etc

In the abovementioned paper, the authors classify cancers into two types:

  1. That arise predominantly because of environmental agents (D-type)
  2. That arise predominantly because of replicative errors (R-type)

Moreover in the same article, the authors point out that the lifetime risk of a cancer is positively correlated with the number of stem cell divisions in the associated tissue. FAP colorectal carcinoma features highest with respect to, number of stem cell divisions and consequently lifetime risk.

Since these cells divide more, they are likely to accumulate more mutations and a defective DNA repair process will worsen the situation. I cannot find a great amount of study on MHS2 gene and its connection with different types of cancers but in general, the above explanation would be applicable to most genes that are involved in DNA repair.

NOTE: As mentioned in the other answer, the abovementioned paper has received some criticism from the scientific community for making far-fetched claims which apparently fits so beautifully with the theory that it seems quite correct. The article however is not retracted. Therefore, it is important to not completely believe in this hypothesis but at the same time keep it in mind considering it as a possibility that needs further verification.

| improve this answer | |

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.