Page 88 - 20dynamics of cancer
P. 88

HISTORY OF THEORIES                                          73

                                      Log 10 (Somatic mutation rate)  -4  2  3 4 6
                                       -2






                                       -6



                                       -8


                                           10          100         1,000        10,000
                                                        Stem cell divisions

                              Figure 4.2  Mutation rate per cell division required to explain observed cancer
                              incidence for various numbers of stages in multistage progression. The x axis
                              shows the number of cell divisions over a lifetime, d. The calculations follow a
                              simple multistage model with constant mutation rate per cell division, the same
                              mutation rate for each transition between stages, and no clonal expansion. Can-
                              cer arises only after the accumulation of n mutations within a single cell lineage.
                              The number attached to each line show the number of stages in progression,
                              n, from classical multistage theory. The shaded area highlights the commonly
                              accepted mutation rate per cell division. I calculated the required mutation
                              rate per cell division, u, by solving for the value of u that satisfies the equation
                                     n−1

                              N(1 −  i=0  P i (ud)) ≈ C, where N is the number of distinct cell lineages in the
                              tissue under study, P i (x) is the Poisson probability of i events given a mean of
                              x events, d is the number of cell divisions per cell lineage over a lifetime in that
                              tissue, and C is the probability that an individual develops cancer in that tissue.
                                                     8
                              For this figure, I used N = 10 and C = 0.05; results change little when varying
                              N up or down by a factor of 10 and when varying C over the range 0.01 − 0.1.
                              See Chapter 6 for the mathematical background.

                                The need for hypermutation seems to be widely accepted (but see
                              Sieber et al. 2005). However, my own calculations of the somatic muta-
                              tion rate required to get several hits contradicts the calculation given by
                              Stein (1991) and the strong conclusions drawn by Loeb (1991) and Beck-
                              man and Loeb (2005) on the sheer implausibility of multiple mutations
                              accumulating in a single cell lineage (see also Calabrese et al. 2004).
                                Figure 4.2 shows that a somatic mutation rate on the order of 10 −5
                              to 10 −6  may be sufficient to explain 4–6 hits. I used a model in which
                              stem cells renew tissues, as happens in colorectal, epidermal, and per-
                              haps several other epithelial tissues, in which most human cancers arise.
   83   84   85   86   87   88   89   90   91   92   93