Cancer Research Cancer Epigenetics Protein Translation and Cancer
HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
Cancer Research Clinical Cancer Research
Cancer Epidemiology Biomarkers & Prevention Molecular Cancer Therapeutics
Molecular Cancer Research Cancer Prevention Research
Cancer Prevention Journals Portal Cancer Reviews Online
Annual Meeting Education Book Meeting Abstracts Online
This Article
Right arrow Abstract Freely available
Right arrow Full Text (PDF)
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Email this article to a friend
Right arrow Similar articles in this journal
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Right arrow reprints & permissions
Citing Articles
Right arrow Citing Articles via HighWire
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Araten, D. J.
Right arrow Articles by Luzzatto, L.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Araten, D. J.
Right arrow Articles by Luzzatto, L.
[Cancer Research 65, 8111-8117, September 15, 2005]
© 2005 American Association for Cancer Research

Molecular Biology, Pathobiology and Genetics

A Quantitative Measurement of the Human Somatic Mutation Rate

David J. Araten1, David W. Golde1, Rong H. Zhang1, Howard T. Thaler2, Lucia Gargiulo3, Rosario Notaro1,3 and Lucio Luzzatto1,3,4

1 Division of Hematology and 2 Department of Biostatistics, Memorial Sloan-Kettering Cancer Center, New York, New York; 3 Istituto Nazionale per la Ricerca sul Cancro; and 4 University of Genova, Genoa, Italy

Requests for reprints: David J. Araten, Division of Hematology, NYU Cancer Institute, 160 East 34th Street, New York, NY 10016. Phone: 212-731-5186; Fax: 212-731-5540; E-mail: David.Araten{at}nyumc.org.


   Abstract
Top
 Abstract
Introduction
 Materials and Methods
 Results
 Discussion
 References
 
The mutation rate (µ) is a key biological feature of somatic cells that determines risk for malignant transformation, and it has been exceedingly difficult to measure in human cells. For this purpose, a potential sentinel is the X-linked PIG-A gene, because its inactivation causes lack of glycosylphosphatidylinositol-linked membrane proteins. We previously found that the frequency (f) of PIG-A mutant cells can be measured accurately by flow cytometry, even when f is very low. Here we measure both f and µ by culturing B-lymphoblastoid cell lines and first eliminating preexisting PIG-A mutants by flow sorting. After expansion in culture, the frequency of new mutants is determined by flow cytometry using antibodies specific for glycosylphosphatidylinositol-linked proteins (e.g., CD48, CD55, and CD59). The mutation rate is then calculated by the formula µ = f/d, where d is the number of cell divisions occurring in culture. The mean µ in cells from normal donors was 10.6 x 10–7 mutations per cell division (range 2.4 to 29.6 x 10–7). The mean µ was elevated >30-fold in cells from patients with Fanconi anemia (P < 0.0001), and µ varied widely in ataxia-telangiectasia with a mean 4-fold elevation (P = 0.002). In contrast, µ was not significantly different from normal in cells from patients with Nijmegen breakage syndrome. Differences in µ could not be attributed to variations in plating efficiency. The mutation rate in man can now be measured routinely in B-lymphoblastoid cell lines, and it is elevated in cancer predisposition syndromes. This system should be useful in evaluating cancer risk and in the design of preventive strategies.


   Introduction
Top
 Abstract
 Introduction
Materials and Methods
 Results
 Discussion
 References
 
In microbial organisms, the rate of mutation can be measured by Luria-Delbrück fluctuation analysis (1) or in continuous cultures (2, 3). By such methods, the rates of inactivating (forward) mutations in lacI in Escherichia coli and CAN1 in Saccharomyces cerevisiae have been estimated to be 4 x 10–7 and 1 x 10–7, respectively (4, 5). In higher organisms, it is now accepted that accumulation of spontaneous mutations in somatic cells can lead to malignant transformation (6, 7).

Indeed, if mutations in n genes are required for malignant transformation and fx is the frequency of cells with a mutation in a given gene, the probability that any one cell will have all of the mutations required can be expressed as f1·f2·f3...fn. Because fx will be a product of the rate of new mutations in a specific gene per cell division (µx) and the number of cell divisions (d) that have occurred since embryogenesis, this expression can be written as = µ1d·µ2d·µ3d...µnd. Using µ as the geometric mean of the mutation rates for the genes involved, and k as a constant <1, to account for cell death, the probability (P) of a cell becoming malignant can be expressed as P = k(µd)n. Thus, P increases as a power function of both µ and d. Because d increases with age, this formula is consistent with the general increase in cancer rates with age (8). Although µ is well recognized as a critical variable (see discussion in ref. 9), in the 60 years since Luria and Delbrück's landmark study, measuring µ in man remains quite challenging.

We previously described a technique for determining f for the X-linked PIG-A gene in humans (10). The PIG-A gene encodes one of the subunits of an enzyme essential for an early step in the biosynthesis of glycosylphosphatidylinositol (GPI; refs. 11, 12). Deficiency of GPI-linked proteins from the surface of blood cells is characteristic of the human disease paroxysmal nocturnal hemoglobinuria (PNH), and it results from somatic mutations of PIG-A (13). Spontaneously arising PIG-A mutations have been identified in a broad range of cell types (10, 14–16). Mutant cells cannot express the set of proteins that require GPI for attachment to the cell surface (the "GPI-anchored" proteins)—resulting in cells with a "GPI-negative phenotype"—which can coexist with normal cells in stable proportions over a prolonged period of time in humans (17) and mice (18). We now use these properties of PIG-A to directly measure µ in human cells.


   Materials and Methods
Top
 Abstract
 Introduction
 Materials and Methods
Results
 Discussion
 References
 
Cell lines. Normal B-lymphoblastoid cell lines (BLCL) were a gift from Dr. Gunther Koehn and Dr. Nathan Ellis (Memorial Sloan-Kettering Cancer Center, New York, NY). Cell lines from patients with cancer predisposition syndromes are summarized in Table 1. BLCLs from patients with PNH (19) were generated as previously described (20). Patients with PNH are somatic cell mosaics, or "somatic chimeras" and harbor GPI+ and GPI– subpopulations of hematopoietic cells: sometimes both can be represented within a BLCL generated from their lymphocytes. Such cell lines were used to set gates for flow cytometry analysis.


View this table:
[in this window]
[in a new window]

  		Table 1. Cell lines used in this study

 
Cell sorting to eliminate preexisting mutants. Cells were stained on ice with a fluorochrome-conjugated antibody that recognizes the GPI-linked protein CD59 (MEM43a, RDI, Flanders, NJ) and then sorted on a FACS Vantage instrument (Becton Dickinson, Franklin Lakes, NJ). Live cells were identified by light scatter characteristics and doublets excluded by pulse width. In order to eliminate preexisting mutants, only the brightest staining 50% of the distribution curve was collected. This step allows side by side comparisons of cell lines derived from individuals of different ages or which have different in vitro replicative histories that might influence mutant frequencies. Using artificial mixtures of GPI+ and GPI– cells, we showed that the cytometer can eliminate the GPI– cells with 95% to 99.6% accuracy. To minimize the chance of mutational drift, we aimed to use at least 106 collected CD59+ cells to establish the starting population. This was achieved in 90 of 103 cell cultures established.

Cell counts and culture conditions. After sorting, live cells were counted by trypan blue exclusion using a hemacytometer. Cells were then grown in flasks in RPMI (Gibco, Grand Island, NY) with 15% FCS (Summit Biotechnology, Ft. Collins, CO), nonessential amino acids (Gibco), L-glutamine (Sigma, St. Louis, MO), and penicillin/streptomycin (Gemini Bio-products, Calabassas, CA) at 37°C, with 100% humidity and 5% CO2. After at least 2 weeks in culture, the cells were recounted and the number of cell divisions (d) calculated as d = log2 (total cell count after expansion / total cell count prior to expansion). In cases where the large volume of culture required reduction, the cells were recounted, split, and returned to culture; here d was calculated as the sum of the d values obtained for each individual expansion.

Preparation of cells for flow cytometry. To eliminate debris, cells were centrifuged over a Ficoll-Hypaque gradient (Amersham-Pharmacia, Uppsala, Sweden), washed, and stained for flow cytometry in a modification of our previously described technique (10). Cells were first stained with a mixture of unconjugated murine antibodies specific for three GPI-linked proteins: CD59 (MEM 43a, RDI), CD55 (MCA 1614, Serotec, Oxford, United Kingdom), and CD48 (MCA1103, Serotec). These antibodies were used at dilutions of ~1:5, 1:10, and 1:400, respectively. The cells were then washed twice and stained with a 1:5 dilution of R-phycoerythrin-conjugated F(ab')2 fragment rabbit anti-mouse immunoglobulin (RAM-PE, DAKO, Glostrup, Denmark), washed twice again and then stained with a 1:10 dilution of a FITC-conjugated antibody specific for HLA-DR, (Becton Dickinson), a non–GPI-anchored transmembrane protein. To ensure that the entire cell population came in contact with the antibodies, we added antibodies prior to resuspension of the cells, which were then briefly recentrifuged and resuspended before incubation on ice. This was done to prevent any cells from being sequestered in any of the staining reactions. Each of the three incubations was done for at least 30 minutes at a concentration of 108 cells/mL. Cells were passed through a 40 µm filter (Becton Dickinson), and propidium iodide (Sigma) was added at a final concentration of 0.15 µg/mL immediately prior to analysis.

Flow cytometry analysis to measure f. Cells were analyzed on a Becton Dickinson FACScan using CellQuest software. Live cells were identified by light scatter characteristics and exclusion of propidium iodide, and were also positively identified by expression of HLA-DR, as indicated by fluorescence of FITC, registering on FL1 (horizontal axis, Fig. 1). The normal (GPI+) cell populations express high levels of the GPI-anchored proteins as determined by phycoerythrin fluorescence, registering typically in the third to fourth decade of FL2 (vertical axis, Fig. 1). With this staining protocol, only cells lacking all three GPI-anchored proteins exhibit low FL2 fluorescence. We set gates using a mixture of GPI– cells from PNH patients along with GPI+ cells. Compensation and detector voltage settings were set such that the GPI– cells fell within the first two decades. The region gates for the GPI– population in the study sample was set in order to capture at least 90% of the control GPI– cells and also so as not to include the tail of the distribution curve of the GPI+ population. f was calculated as the number of gated GPI– cells / number of gated GPI+ cells (the number of GPI– cells was always small enough so as not to significantly affect the denominator of this equation for f). In order to maximize the accuracy of the measurements of f, a median of 1.3 x 106 gated events were counted.



View larger version (11K):
[in this window]
[in a new window]
[Download PPT slide]
  		Figure 1. Calculation of µ. GPI+ cells (upper right quadrant of each panel) express CD48, CD55, CD59—all GPI-anchored proteins—as well as HLA-DR (which is not GPI-anchored). GPI– cells (lower right quadrant of each panel) do not express CD48, CD55 or CD59—but do express HLA-DR. A, flow cytometry dot plot analysis of control cells from a patient with PNH containing both GPI+ and GPI– subpopulations, which are used to establish the gates. B, an ataxia-telangiectasia cell line (GM11254) prior to sorting. GPI– cells are present at a frequency of 36 x 10–6. C, reanalysis immediately after collection of the upper 50th percentile of the distribution curve for CD59 expression shows that these GPI– cells have been completely excluded. This GPI+ population was then placed in culture and expanded from 1 x 106 to 127 x 106 cells (d = 7 cell divisions). D, reanalysis after 27 days in culture. GPI– cells are again present, (f = 25 x 10–6). Because these cells must have arisen by spontaneous mutation in culture, we use the formula µ = f/d to calculate that µ = 36 x 10–7 mutations per cell division. Although there may be a delay between the occurrence of a mutation and its phenotypic expression, in our procedure, a similar lag would occur at the time of sorting and at the time f is measured: thus, the two effects would tend to offset each other.

 
Calculation of µ and statistical analysis. Because we eliminated preexisting mutants prior to expansion in culture, the mutation rate could be calculated with the simple formula: µ = f/d, given new mutants arising in each cell division with a probability of µ and given exponential growth of the mutants that arise and exponential growth of the wild-type population (see derivation in ref. 21). Comparisons between the four groups were done with a two-sided nonparametric test, using the mean of the ranks for cell lines with repeat measurements. Of the 50 data points in Fig. 3, 22 represent the average of repeat analyses.



View larger version (9K):
[in this window]
[in a new window]
[Download PPT slide]
  		Figure 3. The mutation rate in normal individuals and patients with inherited cancer predisposition syndromes. Vertical axis, mutation rates (x107) with a change of scale at 100; horizontal lines, means for each group. FA, Fanconi Anemia; NBS, Nijmegen Breakage Syndrome; AT, Ataxia-Telangiectasia.

 
Generation of GPI– cells and PIG-A sequence analysis. At the time of sorting to establish a GPI+ population, GPI– cells (falling within the isotype gate) were also collected, expanded with irradiated feeders in 96-well plates, purified by treatment with 10–7 mol/L Pro-Aerolysin (Protox Biotech, Victoria, BC) and cloned by limiting dilution. This concentration of Pro-Aerolysin was titrated based on the relative survival of known GPI+ and GPI– cells. The GPI– phenotype was verified by lack of expression of CD59. Complementary DNA sequences were amplified using the following PCR primers: (a) forward 5'GAGGACACATCTCTTAACTGG, reverse 5'GTATCACAAAGAGACACGG; (b) forward 5'CCAGAGTTGGTGGAATTCCTGAG, reverse 5'CCCCCAAAAGCAAGGTTATT. Genomic sequences were also amplified with the following primer pairs: (a) for exon 2, forward 5'TGGAATGTGTTTTGTTTCTGAGCTG, reverse 5'CAAGTATTCAACAGCTTTCTATAG; (b) for exon 3, forward 5'TCTCCTTTATGGAGTGCATAGC, reverse 5'AGAAGCAACACACCTAAGG; (c) for exons 4 to 5, forward 5'GTGTAATATAAATAGCAGACTTGG, reverse 5'CACCTGGAAGAGATTTCATCAAC. Amplified sequences were determined by automated sequencing.

Plating efficiency. In order to make sure that we are not underestimating d due to a low plating efficiency in bulk cultures, viable cells were counted by trypan blue exclusion daily for the first week after sorting. The cell counts after day 0 were plotted on the vertical axis on a logarithmic scale with time on the horizontal axis. The cell count was extrapolated back to time zero by linear regression and the intercept on the vertical axis was taken as the minimal number of cells generating the expanded culture if it was less than the actual cell count on day 0. The plating efficiency was calculated as the extrapolated y-intercept divided by the observed cell count on day 0. In cases where the y-intercept was equal to or higher than the starting number of cells, the plating efficiency was considered to approximate 100%. An adjusted value for d was calculated (da) using the y-intercept (or the actual cell count if it was less than the y-intercept) and subsequent cell counts within the first week, in order to take into account both the plating efficiency and any early loss of cells from the culture.


   Results
Top
 Abstract
 Introduction
 Materials and Methods
 Results
Discussion
 References
 
Measurement of f and µ in the PIG-A gene. In pilot studies using flow cytometry we found rare GPI– cells within a normal BLCL culture (from normal donor 1) at a frequency of 8 x 10–6, a number close the frequency of GPI-negative granulocytes and RBC from healthy volunteers (10). We next analyzed AT1ABR, derived from a patient with ataxia-telangiectasia, who is homozygous for the ATM 7636 del 9 mutation (22); here the frequency of GPI– cells was 50 x 10–6. Based on pilot experiments incubating cell lines with FLAER-Alexa488 (fluorescent aerolysin) prior to staining with antibodies specific for GPI-linked proteins, we have shown that the phycoerythrin-negative cell population we observe is also FLAER-negative. This confirms that the appearance of small GPI– cell populations cannot be accounted for by any sequestration of cells during the first two staining reactions.

Because f may be influenced in vivo by the number of cell divisions occurring in lymphocytes and in vitro by the replicative history of the cell line, we proceeded to eliminate preexisting mutants by physical flow sorting, in order to start with a purely wild-type population of cells in all cases, which would then be expanded in vitro followed by determination of f. An illustrative example of this procedure is given in Fig. 1. Based on the frequency of GPI– cells and our data on the purity of the sorting, we estimate that the number of residual unsorted GPI– cells in the population is sufficiently low to allow us to use the simple formula µ = f/d, where f represents the mutant frequency measured at the end of expansion in vitro.

Reproducibility of the determination of µ. From 15 experiments carried out over a period of 2 years, the mean µ value for the cell line from normal donor 1 was 3.4 x 10–7 mutations per cell division. Fourteen values clustered very close to the mean; there was only one outlier value (Fig. 2). A similar analysis was carried out on the cells from a patient with ataxia-telangiectasia, AT1ABR. A homozygous cell line was chosen because its mutant phenotype cannot revert during prolonged culture by intragenic recombination, as has been reported in Bloom's syndrome (23). In 15 cultures established over 10 months, the mean µ was 110 x 10–7 mutations per cell division (Fig. 2).



View larger version (10K):
[in this window]
[in a new window]
[Download PPT slide]
  		Figure 2. Reproducibility of the assay. Repeat mutation rate analysis of a BLCL from a normal donor and from a patient with ataxia-telangiectasia. Vertical axis, mutation rate (x107) with a change in scale at 50; {blacksquare}, results from eight parallel cultures initiated on 1 day.

 
PIG-A mutations in spontaneously arising GPI– cells. Individual clones from GPI– cell lines spontaneously arising within a BLCL population were subjected to sequence analysis of the coding region of PIG-A (19). A PIG-A mutation was found in all cases. For instance, in two clones from a GPI– cell line derived from the cell line from normal donor 1, we found a 121 delT mutation. In 26 GPI– clones from the ataxia-telangiectasia cell line AT1ABR we identified the following mutations: 715 G->T (239G->W, in 15 clones); 979C->T (327Q->Stop, in 1 clone); 1331G->A (444W->Stop, in 1 clone); a truncated PCR product, suggestive of an exon deletion was found in 9 clones. This spectrum of mutations is similar to that described in patients with PNH (19) and in GPI– granulocytes spontaneously arising in normal individuals (10).

Analysis of µ in cancer predisposition syndromes. In eight normal BLCLs, the mean µ value was 10.6 x 10–7 (range 2.4-29.6 x 10–7) mutations per cell division (Fig. 3). In cell lines from patients with Fanconi anemia (representing complementation groups A, B, C, G, and D2), µ was markedly increased (Fig. 3), with a mean of 411 x 10–7 (range 36-1,175 x 10–7, P < 0.0001). In contrast, in cell lines from six patients with Nijmegen breakage syndrome, there was only a slight elevation of µ, with a mean of 14.6 x 10–7 (range 8.8-26 x 10–7; P = not significant for the comparison with the normal group, P < 0.0001 for the comparison with the Fanconi anemia group; Fig. 3). Analysis of cell lines from 30 patients with ataxia-telangiectasia revealed a more complex pattern (Fig. 3): the mean µ value, 40.1 x 10–7, was almost 4-fold elevated compared with the normal (P = 0.002), and 10-fold lower than the Fanconi anemia group (P < 0.0001). There was wide scatter in the values, which overlapped with the µ values for both the normal and the Fanconi anemia groups. Some of the variance in the ataxia-telangiectasia group might have been due to heterogeneity of the ATM mutations; however, within four out of nine pairs of affected siblings (who have identical ATM genotypes), one sibling had a high and the other sibling had a low µ value. We infer that, as for radiation sensitivity in ataxia-telangiectasia (24), other genes must have a modifying influence on µ.

Optimal time in culture for the measurement of µ. It was important to determine whether the measurement of µ would be influenced by the duration of in vitro expansion. In order to verify this, bulk cultures from representative cell lines were sampled and analyzed at multiple time points (Fig. 4). No significant trend over time was seen. Although this finding is consistent with our understanding that GPI– lymphoblastoid cells have neither a growth advantage nor disadvantage, we sought to show this directly. We therefore analyzed longitudinally a single BLCL derived from a patient with PNH. Initially GPI– cells comprised 61% of the cell population. On 13 occasions over the course of 10 months, this value fluctuated between 42% and 89%; but the average value was again 61% over the 10-month period. This confirmed that at least in this type of cell, the PIG-A mutation is growth-neutral. Because time in culture did not affect µ, we subsequently aimed to analyze the culture as soon as we had sufficient numbers of cells for analysis, and a significant number of cell divisions: typically 3 weeks after sorting.



View larger version (12K):
[in this window]
[in a new window]
[Download PPT slide]
  		Figure 4. Consistency of mutation rate measurements as a function of time. Three cell lines were analyzed: a cell line from normal donor 2; cell line GM 13328, from a patient with ataxia-telangiectasia who had an elevated mutation rate; and cell line GM11255, from a patient with ataxia-telangiectasia who had a normal mutation rate. For each cell line, f was determined in the same bulk culture at multiple time points after sorting. The calculated value of µ x 107 (vertical axis) is plotted against the number of days following sorting (horizontal axis). It is seen that allowing additional time in culture—beyond what we have used in our standard experiments—does not significantly alter the value of µ.

 
Effect of plating efficiency on µ. We sought to rule out the possibility that cell death early after the establishment of the culture or a low plating efficiency in bulk culture might result in an underestimate of d, and consequently, an overestimate of µ. We therefore analyzed representative cell lines by simultaneously taking cell counts daily for 1 week after sorting as well as calculating µ. An adjusted value of d (da) was calculated taking into account the plating efficiency as well as any early cell loss from the culture. The plating efficiency ranged from 52% to 100% and was not significantly different in the cell lines from normal donors compared with patients with cancer predisposition syndromes; also, the plating efficiency did not correlate with the calculated value of µ (Table 2). da did not differ from the originally calculated d by more than ±26%, and the adjustment did not have a significant impact on the calculation of µ. Similar to previous experiments, the purity of the sorted cells ranged from 97.6% to 99.6%.


View this table:
[in this window]
[in a new window]

  		Table 2. Plating efficiency and mutation rate

 

   Discussion
Top
 Abstract
 Introduction
 Materials and Methods
 Results
 Discussion
References
 
Here we provide an experimental determination of the mutation rate in humans, which historically has been exceedingly difficult to do. We show that the mutation rate in cells from normal donors is of the order of 2 to 30 x 10–7 mutations per cell division. To our knowledge, this is the first experimental determination of the normal range of the mutation rate in an endogenous gene in nonmalignant human cells.

Even for the measurement of the frequency of mutants in a population of human cells—i.e., to measure f, as distinct from µ—only a few model genes have been used: hprt (25, 26), glycophorin A (GPA; refs 27, 28), HLA (29–31), CD3 (32), and ß globin (33). This list of genes is extremely limited because for any potential "sentinel gene", a wide range of mutants must be viable, with a readily detectable phenotype, which can result from a single inactivating mutation. Using these genes as a sentinel, f was estimated to range from 2 to 30 x 10–6 (reviewed in ref. 33). However, because the respective genes are autosomal, the frequency of GPA and HLA mutants can be determined only in select individuals with specific allele combinations. Estimating the frequency of mutants in the X-linked hprt gene by limiting dilution cloning and fluctuation analysis using this gene have been proposed as a way to measure µ (34), although difficulties with this approach have been pointed out (35).

As a sentinel gene for measuring f, PIG-A has favorable characteristics similar to other candidate sentinel genes, and it has additional advantages that now enable us to measure µ. Like hprt, PIG-A is on the X-chromosome (13), and a mutant phenotype is therefore not complemented by the coexistence of a normal allele. As for GPA, the PIG-A– phenotype is identified by flow cytometry, facilitating elimination of mutants at the beginning of the experiment and rapid analysis of f after expansion in vitro. Because PIG-A is required for the display on the cell membrane of many GPI-linked proteins, by the simultaneous use of multiple antibodies specific for distinct GPI-anchored proteins we can exploit the ability of the flow cytometer to detect low-frequency populations resulting from spontaneous PIG-A mutations, whereas minimizing the number of "false-positive" events (10). Mutants can coexist with wild-type cells in stable proportions in humans (17) and mice (18), and PIG-A mutations are thought by themselves to be neutral with respect to the growth of the cells under most circumstances (36). In contrast to recent studies modeling somatic hypermutation in a particular codon of an immunoglobulin gene (37), a broad spectrum of inactivating mutations in PIG-A can produce the GPI– phenotype (19). We suggest that PIG-A provides a unique and attractive opportunity to measure the relationship between µ and cancer.

Because mutations are stochastic events, random fluctuations can skew any measurement of f, on which our calculation of µ depends. Because such drift effects are more prominent with small populations, we made an effort to minimize them by collecting a large number of GPI+ cells. Specifically, we aimed to have a starting cell population >1/µ, which we achieved in almost all cases. If plating efficiency were to be low, then the effects of drift would be more important and our estimate of d might be too low. However, repeated experiments on representative cell lines taking into account plating efficiency and early cell loss provided results that were consistent with the original analyses, suggesting that this effect is not prominent (Table 2).

Phenotypic lag—the delay between the acquisition of a mutation and the expression of the GPI– phenotype on the cell surface—may allow for a small proportion of the sorted GPI+ cells to harbor a recently acquired PIG-A mutation. A similar effect is expected to result in an underestimation of f after expansion in vitro, and we expect that these effects will offset each other, assuming that the starting cell population is sufficiently large. Theoretically, the effect of phenotypic lag would be less pronounced with prolonged culture, but we did not observe any significant time effect on the calculation of µ (Fig. 4), suggesting that this effect may be small. Mutational drift and phenotypic lag might become important when testing individuals with a very low baseline mutation rate; but this problem can be minimized by increasing the size of the population of sorted GPI+ cells.

In common with techniques using GPA, the measurement of f depends to some extent upon how the flow cytometry gates are set. We used mixtures of GPI– and GPI+ cells to best simulate the appearance of spontaneously arising PIG-A mutants (Fig. 1), and we took considerable care not to include the tail of the distribution curve of the GPI+ population within the GPI– gate. Indeed, our measurements of f in PIG-A in BLCLs and granulocytes from normal individuals (10) are similar to f for hprt, HLA-A, and GPA (33). Interestingly, our experimental measurements of µ in BLCLs from normal individuals are remarkably similar to the theoretical estimate derived by Green et al. (38) for the in vivo mutation rate in lymphocytes. Therefore, whereas we cannot rule out that the process of EBV immortalization or growth in tissue culture media alters the mutation rate, we have found no evidence for this.

Analysis for the simultaneous loss of three GPI-linked proteins probably maximizes the specificity of the analysis, and we have confirmed the presence of PIG-A mutants in spontaneously arising GPI– BLCL cells. Nevertheless, we cannot completely rule out the possibility of "pseudo-mutants" or "phenocopies", such as might occur if the PIG-A gene were to be transcriptionally silenced. However, because we found an inactivating PIG-A mutation in all GPI– clones subjected to sequence analysis, such an occurrence must be uncommon and not likely to significantly affect our estimate of µ.

In Fanconi anemia, previous attempts to show hyper-mutability by analyzing f had yielded conflicting results (39, 40), perhaps because f is not a good surrogate for µ because d is affected by antigenic stimulation of lymphocytes, inflammatory conditions (41), or bone marrow failure—(42) which does occur in Fanconi anemia. We have now shown experimentally that µ is elevated in Fanconi anemia as well as in some patients with ataxia-telangiectasia. In ataxia-telangiectasia, there is a wide range in the values of µ, a result also seen in the distribution of f values in hprt and GPA (27, 43), and we must infer that this disorder is heterogeneous in this respect, perhaps as a result of modifying genes (24). Remarkably, in Nijmegen breakage syndrome, µ is almost normal. In this condition, rearrangements of chromosomes 7 and 14 are common, and even more frequent than in ataxia-telangiectasia (44–46). Perhaps in the pathogenesis of malignancy in Nijmegen breakage syndrome, chromosomal rearrangements are more important, and point mutations that inactive genes such as PIG-A occur at a normal rate. Alternatively, hyper-mutability in Nijmegen breakage syndrome might be elicited only in the presence of certain mutagens.

Our demonstration that µ is elevated in patients with cancer predisposition syndromes provides experimental support for the notion that the mutation rate of individuals is associated with their risk of cancer. Considering the formula P = k(µd)n, it is clear that relatively small differences in µ between individuals could have large effects on their relative probability of developing cancer, and our data support this model. Indeed, in patients with ataxia-telangiectasia, as a whole, the increase in µ is only about 4-fold (see Fig. 3), but the relative risk of malignancy in patients with ataxia-telangiectasia has been estimated to be 61 to 184 (47). Having shown that high µ is associated with high cancer risk, we are now in a position to determine whether even smaller increments in µ, e.g., within its range in normal people, may also correlate with an increased risk of cancer. In addition, the method we have described can be used to screen for agents that reduce cancer risk by modulating this most highly relevant biological variable.


   Acknowledgments
 
Grant support: Doris Duke Charitable Foundation Clinical Scientist Development Award, by NIH grant R01 HL56778-06, NIH-CA30388, and NIH-CA109258. This work was also supported in part by grants from Ministero della Sanità, FIRB, and Compagnia di San Paolo, Italy.

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

We thank Martina Serra for technical support, Drs. Martin Lavin and Michael Swift for their advice, and Lilli Zhang, Diane Tabarini, and Ellen Bonfiglio for administrative assistance.

In Memoriam: We pay tribute to the memory of Dr. David W. Golde, our colleague and co-author, who died on August 9, 2004. We remember fondly his scientific acumen and generosity of spirit.


   Footnotes
 
Note: L. Gargiulo is currently at Instituto Toscano Tumori, 50139 Firenze, Italy.

Received 4/ 8/05. Revised 6/ 2/05. Accepted 7/ 1/05.


   References
Top
 Abstract
 Introduction
 Materials and Methods
 Results
 Discussion
 References
 

  1. Luria SE, Delbrück M. Mutations of bacteria from virus sensitivity to virus resistance. Genetics 1943;28:491–511.[Free Full Text]
  2. Ryan FJ, Wainwright LK. Nuclear segregation and the growth of clones of spontaneous mutants of bacteria. J Gen Microbiol 1954;11:364–79.[Abstract/Free Full Text]
  3. Novick A, Szilard L. Experiments with the chemostat on spontaneous mutations of bacteria. Proc Natl Acad Sci U S A 1950;36:708–19.[Free Full Text]
  4. Drake JW. A constant rate of spontaneous mutation in DNA-based microbes. Proc Natl Acad Sci U S A 1991;88:7160–4.[Abstract/Free Full Text]
  5. Paquin C, Adams J. Frequency of fixation of adaptive mutations is higher in evolving diploid than haploid yeast populations. Nature 1983;307:495–500.[CrossRef]
  6. Knudson AG. Mutation and cancer: a statistical study of retinoblastoma. Proc Natl Acad Sci U S A 1971;68:820–3.[Abstract/Free Full Text]
  7. Kinzler KW, Vogelstein B. Lessons from hereditary colorectal cancer. Cell 1996;87:159–70.[CrossRef][Medline]
  8. DePinho RA. The age of cancer. Nature 2000;408:248–54.[CrossRef][Medline]
  9. Simpson AJG. The natural somatic mutation frequency and human carcinogenesis. Adv Cancer Res 1997;71:209–40.[Medline]
  10. Araten DJ, Nafa K, Pakdeesuwan K, Luzzatto L. Clonal populations of hematopoietic cells with paroxysmal nocturnal hemoglobinuria genotype and phenotype are present in normal individuals. Proc Natl Acad Sci U S A 1999;96:5209–14.[Abstract/Free Full Text]
  11. Miyata T, Takeda J, Iida Y, et al. The cloning of PIG-A, a component in the early step of GPI-anchor biosynthesis. Science 1993;259:1318–20.[Abstract/Free Full Text]
  12. Hillmen P, Bessler M, Mason PJ, Watkins WM, Luzzatto L. Specific defect in N-acetylglucosamine incorporation in the biosynthesis of the glycosylphosphatidylinositol anchor in cloned cell lines from patients with paroxysmal nocturnal hemoglobinuria. Proc Natl Acad Sci U S A 1993;90:5272–6.[Abstract/Free Full Text]
  13. Takeda J, Miyata T, Kawagoe K, et al. Deficiency of the GPI anchor caused by a somatic mutation of the PIG-A gene in paroxysmal nocturnal hemoglobinuria. Cell 1993;73:703–11.[CrossRef][Medline]
  14. Ware RE, Pickens CV, DeCastro CM, Howard TA. Circulating PIG-A mutant T lymphocytes in healthy adults and patients with bone marrow failure syndromes. Exp Hematol 2001;29:1403–9.[CrossRef][Medline]
  15. Chen R, Eshleman JR, Brodsky RA, Medof ME. Glycosylphosphatidylinositol-anchored protein deficiency as a marker of mutation phenotype in cancer. Cancer Res 2001;61:654–8.[Abstract/Free Full Text]
  16. Rawstron AC, Rollinson SJ, Richards S, et al. The PNH phenotype cells that emerge in most patients after CAMPATH-1H therapy are present prior to treatment. Br J Haematol 1999;107:148–53.[CrossRef][Medline]
  17. Araten DJ, Bessler M, McKenzie S, et al. Dynamics of hematopoiesis in paroxysmal nocturnal hemoglobinuria: no evidence for intrinsic growth advantage of PNH clones. Leukemia 2002;16:2243–8.[CrossRef][Medline]
  18. Rosti V, Tremml G, Soares V, Pandolfi PP, Luzzatto L. Murine embryonic stem cells without pig-a gene activity are competent for hematopoiesis with the PNH phenotype but not for clonal expansion. J Clin Invest 1997;100:1028–36.[Medline]
  19. Luzzatto L, Nafa K. Genetics of PNH. In: Young NS, Moss J, editors. PNH and the GPI-linked Proteins. San Diego: Academic Press; 2000. p. 21–47.
  20. Hillmen P, Bessler M, Crawford DH, Luzzatto L. Production and characterization of lymphoblastoid cell lines with the paroxysmal nocturnal hemoglobinuria phenotype. Blood 1993;81:193–9.[Abstract/Free Full Text]
  21. Stent GS, Calendar R. Molecular Genetics. San Francisco, CA: WH Freeman and Company; 1978. p. 152–79.
  22. Savitsky K, Bar-Shira A, Gilad S, et al. A single ataxia telangiectasia gene with a product similar to PI-3 kinase. Science 1995;268:1749–53.[Abstract/Free Full Text]
  23. Ellis NA, Groden J, Ye TZ, et al. The Bloom's syndrome gene product is homologous to RecQ helicases. Cell 1995;83:655–66.[CrossRef][Medline]
  24. Meyn MS, Lu-Kuo JM, Herzing LB. Expression cloning of multiple human cDNAs that complement the phenotypic defects of ataxia-telangiectasia group D fibroblasts. Am J Hum Genet 1993;53:1206–16.[Medline]
  25. Albertini RJ, Castle KL, Borcherding WR. T-cell cloning to detect the mutant 6-thioguanine-resistant lymphocytes present in human peripheral blood. Proc Natl Acad Sci U S A 1982;79:6617–21.[Abstract/Free Full Text]
  26. Morley AA, Trainor KJ, Seshadri R, Ryall RG. Measurement of in vivo mutations in human lymphocytes. Nature 1983;302:155–6.[CrossRef][Medline]
  27. Bigbee WL, Langlois RG, Swift M, Jensen RH. Evidence for an elevated frequency of in vivo somatic cell mutations in ataxia telangiectasia. Am J Hum Genet 1989;44:402–8.[Medline]
  28. Kyoizumi S, Nakamura N, Hakoda M, et al. Detection of somatic mutations at the glycophorin A locus in erythrocytes of atomic bomb survivors using a single beam flow sorter. Cancer Res 1989;49:581–8.[Abstract/Free Full Text]
  29. Grist SA, McCarron M, Kutlaca A, Turner DR, Morley AA. In vivo human somatic mutation: frequency and spectrum with age. Mutat Res 1992;266:189–96.[Medline]
  30. Kavathas P, Bach FH, DeMars R. Gamma ray-induced loss of expression of HLA, glyoxalase I alleles in lymphoblastoid cells. Proc Natl Acad Sci U S A 1980;77:4251–5.[Abstract/Free Full Text]
  31. Pious D, Hawley P, Forest G. Isolation and characterization of HL-A variants in cultured human lymphoid cells. Proc Natl Acad Sci U S A 1973;70:1397–400.[Abstract/Free Full Text]
  32. Umeki S, Suzuki T, Kusunoki Y, Seyama T, Fujita S, Kyoizumi S. Development of a mouse model for studying in vivo T-cell receptor mutations. Mutat Res 1997;393:37–46.[Medline]
  33. Albertini RJ, Nicklas JA, O'Neill JP, Robison SH. In vivo somatic mutations in humans: measurement and analysis. Annu Rev Genet 1990;24:305–26.[Medline]
  34. Seshadri R, Kutlaca RJ, Trainor K, Matthews C, Morley AA. Mutation rate of normal and malignant human lymphocytes. Cancer Res 1987;47:407–9.[Abstract/Free Full Text]
  35. Kendal WS, Frost P. Pitfalls and practice of Luria-Delbruck fluctuation analysis: a review. Cancer Res 1988;48:1060–5.[Abstract/Free Full Text]
  36. Chen R, Nagarajan S, Prince GM, et al. Impaired growth and elevated Fas receptor expression in PIG-A+ stem cells in primary paroxysmal nocturnal hemoglobinuria. J Clin Invest 2000;106:689–96.[Medline]
  37. Bachl J, Carlson C, Gray-Schopfer V, Dessing M, Olsson C. Increased transcription levels induce higher mutation rates in a hypermutating cell line. J Immunol 2001;166:5051–7.[Abstract/Free Full Text]
  38. Green MHL, O'Neill JP, Cole J. Suggestions concerning the relationship between mutant frequency and mutation rate at the hprt locus in human peripheral T-lymphocytes. Mutat Res 1995;334:323–39.[Medline]
  39. Papadopoulo D, Guillhouf C, Mohrenweiser H, Moustacchi E. Hypomutability in Fanconi anemia cells is associated with increased deletion frequency at the HPRT locus. Proc Natl Acad Sci U S A 1990;87:8383–7.[Abstract/Free Full Text]
  40. Sala-Trepat M, Boyse J, Richard P, Papadopoulo D, Moustacchi E. Frequencies of HPRT– lymphocytes and glycophorin A variant erythrocytes in Fanconi anemia patients, their parents, and control donors. Mutat Res 1993;289:115–26.[Medline]
  41. Paganin C, Monos DS, Marshall JD, Frank I, Trinchieri G. Frequency and cytokine profile of HPRT mutant T cells in HIV-infected and healthy donors: implications for T cell proliferation in HIV disease. J Clin Invest 1997;99:663–8.[Medline]
  42. Hattori H, Machii T, Ueda E, Shibano M, Kageyama T, Kitani T. Increased frequency of somatic mutations at glycophorin A loci in patients with aplastic anaemia, myelodysplastic syndrome, and paroxysmal nocturnal haemoglobinuria. Br J Haematol 1997;98:384–91.[CrossRef][Medline]
  43. Cole J, Artlett CF. Cloning efficiency and spontaneous mutant frequency in circulating T-lymphocytes in ataxia-telangiectasia. Int J Radiat Biol 1994;66:S123–31.[Medline]
  44. Aurias A, Dutrillaux B, Buriot D, Lejeune J. High frequencies of inversions and translocations of chromosomes 7 and 14 in ataxia-telangiectasia. Mutat Res 1980;69:369–74.[CrossRef][Medline]
  45. Taalman RDFM, Hustinx TWJ, Weemaes CMR, et al. Further delineation of the Nijmegen breakage syndrome. Am J Med Genet 1989;32:425–31.[CrossRef][Medline]
  46. Tauchi H, Matsuura S, Kobayashi J, Sakamoto SKK. Nijmegen breakage syndrome NBS1 and molecular links to factors for genomic stability. Oncogene 2002;21:8980.
  47. Morrell D, Cromartie E, Swift M. Mortality and cancer incidence in 263 patients with ataxia-telangiectasia. J Natl Cancer Inst 1986;77:89–92.
  48. De Winter JP, Waisfisz Q, Rooimans MA, et al. The Fanconi anaemia group G gene FANCG is identical with XRCC9. Nat Genet 1998;20:281–3.[CrossRef][Medline]
  49. Joenge H, Levitus M, Waisfisz Q, et al. Complementation analysis in Fanconi anemia: assignment of the reference FA-H patient to group A. Am J Hum Genet 2000;67:759–62.[CrossRef][Medline]
  50. Yamashita T, Barber DL, Zhu Y, Wu N, D'Andrea AD. The Fanconi anemia polypeptide FACC is localized to the cytoplasm. Proc Natl Acad Sci U S A 1994;91:6712–6.[Abstract/Free Full Text]
  51. Timmers C, Taniguchi T, Hejna J, et al. Positional cloning of a novel Fanconi anemia gene FANCD2. Mol Cell 2001;7:241–8.[CrossRef][Medline]
  52. Howlett NG, Taniguchi T, Olson S, et al. Biallelic inactivation of BRCA2 in Fanconi anemia. Science 2002;297:606–9.[Abstract/Free Full Text]
  53. Telatar M, Wang Z, Udar N, et al. Ataxia-telangiectasia: mutations in ATM cDNA detected by protein-truncation screening. Am J Hum Genet 1996;59:40–4.[Medline]
  54. Gilad S, Khosravi R, Shkedy D, et al. Predominance of null mutations in ataxia-telangiectasia. Hum Mol Genet 1996;5:433–9.[Abstract/Free Full Text]
  55. Wright J, Teraoka S, Onengut S, et al. A high frequency of distinct ATM gene mutations in ataxia-telangiectasia. Am J Hum Genet 1996;59:839–46.[Medline]



This article has been cited by other articles:


Home page
 Proc. Natl. Acad. Sci. USAHome page
D. Dingli, L. Luzzatto, and J. M. Pacheco
Neutral evolution in paroxysmal nocturnal hemoglobinuria
PNAS, November 25, 2008; 105(47): 18496 - 18500.
[Abstract] [Full Text] [PDF]

Home page
 Am. J. Pathol.Home page
C. Hornsby, K. M. Page, and I. Tomlinson
The in Vivo Rate of Somatic Adenomatous Polyposis Coli Mutation
Am. J. Pathol., April 1, 2008; 172(4): 1062 - 1068.
[Abstract] [Full Text] [PDF]

Home page
 Proc. Natl. Acad. Sci. USAHome page
S. Jones, W.-d. Chen, G. Parmigiani, F. Diehl, N. Beerenwinkel, T. Antal, A. Traulsen, M. A. Nowak, C. Siegel, V. E. Velculescu, et al.
Comparative lesion sequencing provides insights into tumor evolution
PNAS, March 18, 2008; 105(11): 4283 - 4288.
[Abstract] [Full Text] [PDF]

Home page
 BloodHome page
J. V. Lopes, J. M. Pacheco, and D. Dingli
Acquired hematopoietic stem-cell disorders and mammalian size
Blood, December 1, 2007; 110(12): 4120 - 4122.
[Abstract] [Full Text] [PDF]

Home page
 Nucleic Acids ResHome page
J. M. Hinz, P. B. Nham, S. S. Urbin, I. M. Jones, and L. H. Thompson
Disparate contributions of the Fanconi anemia pathway and homologous recombination in preventing spontaneous mutagenesis
Nucleic Acids Res., June 28, 2007; 35(11): 3733 - 3740.
[Abstract] [Full Text] [PDF]

Home page
 BloodHome page
D. J. Araten and L. Luzzatto
The mutation rate in PIG-A is normal in patients with paroxysmal nocturnal hemoglobinuria (PNH)
Blood, July 15, 2006; 108(2): 734 - 736.
[Abstract] [Full Text] [PDF]

Home page
 Cancer Res.Home page
Measurement of the Human Mutation Rate
Cancer Res., November 15, 2005; 65(22): 10635 - 10635.
[Full Text]

This Article
Right arrow Abstract Freely available
Right arrow Full Text (PDF)
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Email this article to a friend
Right arrow Similar articles in this journal
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Right arrow reprints & permissions
Citing Articles
Right arrow Citing Articles via HighWire
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Araten, D. J.
Right arrow Articles by Luzzatto, L.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Araten, D. J.
Right arrow Articles by Luzzatto, L.

HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
Cancer Research Clinical Cancer Research
Cancer Epidemiology Biomarkers & Prevention Molecular Cancer Therapeutics
Molecular Cancer Research Cancer Prevention Research
Cancer Prevention Journals Portal Cancer Reviews Online
Annual Meeting Education Book Meeting Abstracts Online