(2) |
Dr. Islas often reminds us that cancer is a “reward for
living as long as you do.” Over the
weekend, I came across two studies that prove just that. The first study was conducted by a group of
scientists at the Gene Environment Association Studies (GENEVA). The second study was led by scientists at the
National Cancer Institute (NCI). Both
studies found that large chromosomal abnormalities, some of which are correlated
with increased risk of cancer, can be detected in a fraction of people without
a prior history of cancer. Sampling from
hundreds and thousands of individuals, the scientists also found that
alterations in chromosomes increase with age, specifically after the age of 50,
and may be associated with an increased risk of cancer. (1)
The GENEVA consortium is sponsored by the National Human
Genome Research Institute (NHGRI). The NCI
and the NHGRI are both part of the National Institutes of Health (NIH).
Both studies were published online on May 6, 2012 in Nature Genetics.
Before delving into exactly how the scientists went about their
work, let’s take a step back and talk about the different types of large
chromosomal abnormalities:
Constitutional vs. Acquired (3)
Constitutional: All tissues (“the whole person”) have
the same anomaly. The chromosome error
was present in the embryo. A chromosomal
error occurring in either the maternal or paternal gamete (i.e. error before
fertilization) will result in every cell of the resulting child having the
abnormal karyotype. A chromosomal error occurring
in the zygote (i.e. error after fertilization) will result in a mixture of
abnormal and normal cells in each tissue of the resulting child. (Fig.
1)
Acquired: Only cells in one tissue
have the anomaly, and other cells in other tissues are normal. The chromosome error occurred in a specific
cell type and therefore only affects organs/tissues consisting of that cell
type. (Fig. 1)
Homogenous vs. Mosaic (3)
Homogenous: All cells carry the same anomaly. A constitutional anomaly, having occurred in
either the maternal or paternal gamete, will be found in every cell of the
resulting child and is considered homogenous abnormal. An acquired anomaly, having occurred in a
specific cell type, can be considered homogenous if no normal cell karyotype is
seen in the organ/tissue that cell belongs to.
For example, in leukemia, when an acquired anomaly occurs in bone marrow
cells, these transformed cells can inhibit the growth of normal cells, leading to
a homogenous population of abnormal bone marrow cells.
Mosaic: Only some cells carry the anomaly,
and other cells are normal. Mosaicism is
defined as the coexistence of cells with two or more distinct karyotypes within
an individual.
Chromosomal anomalies have been associated with a variety of
deleterious health conditions, one of the most serious of which is cancer. Cancer cells typically, acquire one or
more chromosomal anomaly, while other cells in the same tissue retain a normal
karyotype. Therefore, cancer patients are
mosaic in that they have a mixture of normal and mutated cells. Both the NCI and GENEVA studies focused on
large, acquired chromosomal anomalies in a mosaic state. More specifically, researchers were
interested in uncovering the type, frequency, and age distribution of these
anomalies and mosaicisms in the general population. (4)
For simplification purposes, from hereon out I will refer to
the “large, acquired chromosomal anomalies in a mosaic state” that was studied
in these two papers as, “chromosome abnormalities.”
Who were sampled?
Scientists from both consortia used DNA extracted from blood
and saliva that was collected in previously conducted genome-wide association studies.
Investigators from the NCI study analyzed 57,853 individuals
from thirteen genome-wide association studies.
There were 31,717 cancer cases and 26,136 cancer-free controls. (5)
Researchers from the GENEVA study analyzed 50,222
individuals from sixteen genome-wide association studies. Participants were of all ages (birth to
elderly), were from several common ancestry groups, had a variety of different
health disease or were healthy controls. (4) (Table 1)
Types of anomalies detected
Investigators from the NCI study
classified three types of chromosomal abnormalities: copy-altering gain, copy-altering
loss, or copy-neutral loss of herterozygosity.
All three event types -- gain, loss, and loss of herterozygosity -- had
to involve greater than 2Mb of DNA in order to be counted as an “abnormality.”
(4)
Researchers
from the GENEVA study classified three slightly different types of chromosomal
abnormalities: deletions, duplications, and uniparental disomy. Uniparental disomy typically occurs during
mitotic crossing over and is defined as having both copies of a chromosome(s) or
chromosome segment from the same
parent resulting in no contribution from the other parent. Whereas the minimum event size threshold was
set at 2Mb for the NCI-study, all three event types in the GENEVA study --- deletions,
duplications, and uniparental disomy -- had to involve greater than 50kb of DNA
in order to be counted as an “abnormality.” (5)
How to detect anomalies?
Normally in population studies,
polymerase chain reaction (PCR) is used to determine the presence of
single-nucleotide polymorphism (SNP) and subsequently to detect chromosomal
anomalies. (6) (Fig. 2) However, because researchers in both studies were
focused on detecting large chromosomal abnormities, SNP microarrays were not optimal. Therefore, both the NCI and GENEVA
researchers used a modified SNP-method to detect and map large (>2Mb for
NCI, >50kb for GENEVA) chromosomal abnormalities within a single DNA
sample. This modified method relied on a
having a relatively high frequency (>5-10%) of cells with the same abnormal
karyotype in the presence of other normal cells. (4)
Frequency of acquired mosaicism
The NCI scientists found detectable chromosomal
abnormalities, greater than 2Mb, in 0.87% (517/57,853) of the subjects
analyzed. The most frequent type of
event observed was copy-neutral loss of herterozygosity at 48.2%. Copy-altering gains constituted 15.1% of the
mosaic events, while copy-altering losses occurred at 34.8%. (4) (Table 2)
The GENEVA scientists found detectable chromosomal
abnormalities, greater than 50kb, in 0.8% (404/50,222) of the subjects
analyzed. The most frequent type of
event observed was deletions at 50.4%.
Duplications constituted 15.6% of the mosaic events, while uniparental
disomy occurred at 34.0%. The median
length of observed deletions was 3.8Mb, 34.1Mb for duplications, and 39.8Mb for
uniparental disomy. The fact that the
median lengths of all three classes of chromosomal abnormalities were well over
2Mb – the minimum event threshold set by NCI researchers -- could explain why
both studies yielded very similar results, despite the fact that each consortia
set very different lengths to constitute an “abnormality.” (5)
Frequency of acquired mosaicism
increases with age
Researchers from both studies wanted
to determine the relationship between chromosomal abnormalities and age.
The NCI scientists found that in cancer-free subjects under
the age of 50, frequency of chromosomal abnormalities was low at 0.23%. Frequency of anomalies rose sharply to 1.91% in
cancer-free individuals between the age of 75 and 79. The same pattern, but at a slightly higher
frequency, were seen in individuals with cancer. (4) (Fig. 3)
The GENEVA researchers found that in cancer-free subjects
under the age of 50, frequency of chromosomal abnormalities was low at
<0.5%. Frequency of anomalies rose
sharply to 2.7% in cancer-free individuals over the age of 80. (5)
To further support the finding that chromosomal
abnormalities appear later in life, the GENEVA team carefully analyzed their
only subject who was sampled twice. This
individual was sampled at age 66 and again at age 72. Both DNA samples were obtained from
saliva. GENEVA scientists found no chromosomal
abnormalities detected in the earlier sample, while the later sample contained
five mosaic deletions, each on a different chromosome. (5)
To be thorough, researchers from
both studies corrected for confounding factors, such as genome-wide association
study, sex, DNA source (blood or saliva), and ancestry. In their statistical tests, even after adjusting
for these covariates, the scientists found that age alone was still a highly
significant predictor of mosaic status. (4, 5)
The results indicate that detectable
chromosomal abnormalities increase with age, especially in those over the age
of 50, and that early detection of chromosomal abnormalities could be a decent
predictor for risk of cancer development.
Frequency of acquired mosaicism and
non-hematological cancer (5)
Next, researchers from the NCI study
wanted to determine the relationship between chromosomal abnormalities and non-hematological
cancers.
Investigators from the NCI study
analyzed 14,050 individuals with non-hematological cancer for whom they were
certain DNA was obtained at least one year before diagnosis and before
treatment with radiation or chemotherapy.
Researchers wanted DNA collected at least one year before diagnosis,
because an overall goal of their study was to see whether chromosomal abnormalities
could potentially be used as a marker for cancer. Researchers wanted DNA collected prior to treatment
with radiation or chemotherapy, because they hypothesized that the harmful
effects of both therapies could potentially alter the frequency in chromosomal
abnormalities and skew the data.
The NCI team found that chromosomal
abnormalities were more frequent in individuals with solid tumors (0.94)
compared to cancer-free controls (0.74).
Further analysis of the chromosomal
abnormalities allowed the NCI researchers to map the three most commonly
altered regions. They were 13q
(deletion), 14 (copy-natural loss of herterozygosity), and 20q (deletion). Aberrations in these three regions were
detected in more than 20 individuals and are known to carry well-studied cancer
genes.
NCI researchers observed
copy-neutral loss of herterozygosity in a particular region of chromosome 14,
in many individuals with bladder or kidney.
The specific region of chromosome 14 that underwent loss of
herterozygosity, carries the tumor suppressor gene, MEG3 (maternally expressed gene 3).
NCI researchers also observed deletions
in region 14 on the q arm of chromosome 13 (13q14) in five leukemia cases and
eighteen individuals with solid tumors.
Region 14 of chromosome 13 that underwent deletion, caries the tumor
suppressor genes, DLEU7 (deleted in
leukemia 7) and RB1 (retinoblastoma
1).
NCI researchers observed deletions
in the q arm of chromosome 20 (20q) in two individuals with myeloid leukemia
and those with solid tumors. The 20q
deletion was also seen in cancer-free controls.
Frequency of acquired mosaicism and hematological
cancer
Lastly, researchers from both
studies wanted to determine the relationship between chromosomal abnormalities
and hematological cancer.
Investigators from the NCI study
analyzed 43 individuals with hematological cancer for whom they were certain
DNA was obtained at least one year prior to diagnosis. Once again, researchers wanted DNA collected
at least one year before diagnosis, because an overall goal of their study was
to see whether chromosomal abnormalities could potentially be used as a marker
for cancer. The NCI team found that in
the 43 subjects with hematological cancer, the frequency of detectable chromosomal
abnormalities was 20% for those with acute myeloid lymphoma (AML) and 22% for
those with chronic lymphocytic leukemia (CLL).
Frequency of detectable chromosomal abnormalities in 26,136 cancer-free
controls was a low 0.74%. (4) (Fig. 4)
Using the same DNA collection guidelines for sampling as the
NCI scientists, the GENEVA researchers found that risk of acquiring
hematological cancer was 10-fold higher for subjects with chromosomal
abnormalities compared to individuals without genetic mosaicism. Despite this dramatic finding the scientists
noted that the 10-year incidence rate in individuals with chromosomal
abnormalities was still low at 14.3%. (5)
An accurate measurement of the period of time between the
first appearance of chromosomal abnormalities and the incidence of hematological
cancer could not be obtained. In regards
to the individuals with mosaicism, researchers were unsure whether DNA sampling
was taken at the immediate time chromosomal abnormalities appeared. The GENEVA scientists were however, able to
acquire a very rough estimate of the median time between DNA sampling and
hematological cancer diagnosis at 3.5 years. (5) (Fig. 5)
The results suggest that frequency of mosaic chromosomal
abnormalities is higher in patients with hematological cancers compared to
cancer-free controls, and that chromosomal abnormalities in cancer-free
subjects may raise the risk of hematological caner diagnosis.
Conclusions
Dr. Islas tells us all the time – “cancer
is a disease of old-age.” The two
studies presented here, suggest that the age-related development of large
chromosomal aberrations leading to mosaicism, may be a primary reason as to why
this is so. The NCI and GENEVA-led
papers are the first population-based study of acquire chromosomal
abnormalities in people of all ages, various states of heath, and ancestry.
Findings from both studies indicate
that a substantial amount of DNA from the blood and salivary samples from
subjects with no prior history of cancer may contain large chromosomal
abnormalities. Furthermore, the frequency
of these chromosomal abnormalities is low in individuals under the age of 50,
and then increases dramatically in elderly individuals. Because a majority of the acquired anomalies
tend to occur in genomic regions containing key tumor suppressor genes, genetic
mosaicism may raise the risk of cancer.
Lastly, both NCI and GENEVA scientists observed an increase in
chromosomal abnormalities in subjects with hematological cancers, and that risk
of acquiring hematological cancer was higher in cancer-free individuals with
genetic mosaicism.
Questions & Thoughts
After
reading these two papers, a few questions and thoughts came to mind:
1) The findings of both the NCI and GENEVA
studies are really not surprising. Most
of the acquired mosaic anomalies detected by the modified SNP method appear
with older age simply because they arise more frequently. Early in the quarter, Dr. Islas mentioned how
it is normal for mutations in our genome to occur, but these mutations are often
corrected by DNA repair machinery before replication. As we age, the genome repair proteins
malfunction and regulation of the genes that encode these proteins begin to breakdown. Consequently, mutations are not
corrected before the cell enters replication, which allows any change in the
genome to be copied to all newly synthesized strands of DNA. The mutated DNA is then perpetuated forward
in all of the cell’s descendants. This age-related
decline in genomic maintenance mechanisms can lead to increased mutations and ultimately,
cancer. Another reason why frequency of
detectable genetic anomalies increases with age is due to clonal expansion or the
rapid proliferation of cells containing chromosomal abnormalities.
2) Overall, both studies seem very
promising. The fact that large acquired
mosaic anomalies are correlated with age along with the fact that frequency
of these abnormalities may one day be used as an indicator of cancer risk
certainly makes sense. However, I am
concerned about a few things:
- the
extremely small number of events seen in both studies – 517 out of 57,853 in
the NCI study, and 404 out of 50,222 in the GENEVA study
- although
the genome-wide association studies seems to consist of a relatively diverse
group of individuals with different ages, disease-status, ancestry, etc.
perhaps an even wider array and simply more sampling from different populations
will help to solidify these findings
- the
GENEVA study only had one participant who was sampled twice, perhaps more subjects
who undergo repeated DNA sampling at different time points in their lives will allow researchers to monitor the changes in genetic anomalies in individual participants
as a factor of time
Ultimately, if NCI and GENEVA
conduct additional studies and analyze additional events, then their findings –
that large chromosomal abnormalities are associated with cancer risk and age –
will become much more powerful.
3) This may be useful information for
groups interested in doing their cancer project on age and its correlation with
cancer.