Document Sample
midterm Powered By Docstoc
					MEDG 520 Midterm                                                      Yvonne Bombard

Midterm Sample Questions:

The following are examples of answers to the midterm questions that got %100 or nearly %100. Keep in
mind that these answers are not exhaustive and there is no single correct answer to questions as broad and
open-ended as these. These are meant as illustrative examples only.

Question #1
Part a
        Sample Answer #1 for question 1a
        Sample Answer #2 for question 1a
        Sample Answer #3 for question 1a
Part b
        Sample Answer #1 for question 1b
        Sample Answer #2 for question 1b
        Sample Answer #3 for question 1b
Part c
     Sample Answer #1 for question 1c
         o References
   Sample Answer #2 for question 1c
Marking Outline for Question 1

Question #2
     Sample Answer #1 for question 2
         o References
   Sample Answer #2 for question 2
         o References
Marking Outline for Question 2

Question #3
Part a
    Sample Answer #1 for question 3a
    Sample Answer #2 for question 3a
Part b
    Sample Answer #1 for question 3b
Marking Outline not available

Question #1
There are 3 parts to this question, each worth 1/3 of the total value.
A deletion of 22q11 occurs at a significant frequency in the human population.

a) How do you account for variability of phenotype between individuals who have similar deletions
of this region? Include examples to clarify your answer.
MEDG 520 Midterm                                                        Yvonne Bombard

Sample Answer #1 for question 1a
        The majority (> 85%) of 22q11 deletions (del22q11) are 3.0 Mb in length. A less common
deletion of 1.5 Mb is observed in ~10% of cases. The syndromes observed in the latter deletion are
virtually indistinguishable form the larger deletion; suggesting that the majority of the 30 suspected genes
in the 22q11 region are clustered within this 1.5Mb region. Nevertheless, there is an extensive range of
phenotypic variability in del22q11 patients, including Velo-cardio-facial (VCFS) and DiGeorge (DGS)
syndromes that are characterized by congenital heart defect, cleft palate, facial abnormalities, neonatal
hypocalcemia, T-cell immune deficiency, thymic hypoplasia, learning disabilities and psychiatric illness
(Digillio et al. 2003, Funke et al. 2001). The high incident (1/4000) of del22q11 is attributed to low copy
repeats (LCRs) found in this region. Intra-chromosomal recombinations between the most distal and
proximal LCRs and the two intermediate LCRs may lead to the 3Mb and 1.5Mb deletions respectively.
Since this region is densely packed with several developmentally important genes, even small differences
in deletion between individuals that would disrupt a part of the functional gene or its regulatory elements
may result in phenotype differences (Hatchwell et al, 1998). It is also important to note which allele has
been affected for imprinting purposes. If the deletion occurs in a gene that is maternally expressed, the
phenotype in this individual will vary from some one who has the deletion in the paternal allele. Further,
an involvement of modifier genes elsewhere in the genome may influence expression (Digilio et al. 2003,
Hatchwell et al. 1998). A study by Taddei and colleagues (2001) using mouse models of del22q11 has
demonstrated this point. Mice with identical and mixed genetic backgrounds and heterozygous for the
same deletion (Df1/+), displayed major differences in penetrance of cardiovascular and thymic
phenotypes. Cardiovascular abnormalities were much less frequent (16.1%) in mice with identical
genetic backgrounds compared to the mixed (50%). Similarly, thymic abnormalities were present in
42.5% of mixed versus 11.3% of identical mice. Interestingly, as with humans, the expression of heart
and thymic phenotypes are independent of each other suggesting that these phenotypes may be affected
by independent genetic modifiers.
        Twin and family studies raise other possible mechanisms for the observed variation in phenotype.
There have been reports of monozygotic (MZ) twins that carry the exact deletion, yet, are discordant for
the phenotypes (5/6 MZ twins are highly discordant for a variety of features including cardiac,
developmental, mental and behavioural features (Singh et al. 2002)). Further, because of their genetic
identity, the explanations offered above, may not entirely account for the variations. These differences
may be explained by random genetic events such as differential mosaicism, and ‘second hit’ mutation
events such as replication errors, base changes, additional mutations involving the LCR region (Singh et
al 2002). Most likely however, variation is due to epigenetic differences. It has been suggested that
differential methylation of hemizygous genes between MZ twins may be held accountable for the large
developmental differences observed since such an event would result in differential gene expression as a
result of gene silencing or down regulation. Other factors such as differences in uterine environment
between twins may also have an effect to a lesser degree.

Sample Answer #2 for question 1a
       Deletions of chromosome 22q11 are the most frequent interstitial deletions found in the human
genome with an incidence of 1 in 4000 live births (Scambler, 2000). Deletions at the 22q11 locus are
generally uniform in nature but are characterized by frequently reduced penetrance of the different
phenotypic components and widely variable expressivity (Taddei et al., 2001). I will briefly discuss 8
possible explanations for this observation:
       (1) The simplest case would be that different deletions among patients affect different genes or
regulatory elements and thus have different effects. Even slight differences in the region deleted could
have dramatic effects on phenotype. For example, two deletions might differ in size by only a base, but,
could alter the frame of the deletion region and turn a partially functional protein into a non-functional
one. (2) The phenotype could depend on modifier genes that mediate gene-to-gene interactions.
MEDG 520 Midterm                                                       Yvonne Bombard

Deletions in one part of the region could therefore affect the expression of genes in other intact segments
of the genome. For example, both RanBP1 and Htf9c are thought to be regulated, at least in part, by the
E2F6 gene which lies very close to the 22q11 deletion (Maynard et al, 2002). Thus, loss of one gene will
have affects on several other genes. (3) Deletions could affect regulation at a higher level, such as
chromatin structure. This would be mediated by the deletion of low-copy-number DNA repeats (LCRs)
that are known to be involved in regulating gene expression and might affect the so-called “chromatin
folding code” (Vogt. 1990). The deletion of these LCRs might have unpredictable effects on the
chromatin structure for the entire region affecting the ability of transcription factors to bind to and
transcribe DNA in these regions. (4) The phenotypes could be the result of multiple hits. In other words,
the effect of 22q11 deletions could depend on the presence of other deletions, mutations, or
polymorphisms. There are many such examples in cancer. (5) Epigenetic or environmental effects are
another possibility. Phenotypic discordance between monozygous twins suggests a large non-genetic
component. Examples of this phenomenon are considered in reference to schizophrenia below (section
1c). (6) Similarly, the effect of allelic variations at the haploid locus created by the deletion will be
considered below (section 1b). (7) Lindsay and Baldini (2001) suggest a fascinating self repair
mechanism in which del22q11.2 mice are able to overcome in some cases the developmental delays and
cardiovascular defects associated with the deletion. (8) Finally, both somatic (Consevage et al., 1996)
and germ line (Hatchwell et al., 1998) mosaicism for the deletion have been shown in some cases and
likely explain at least a small fraction of the phenotypic variability observed in the population.

Sample Answer #3 for question 1a
         There are many ways in which phenotypic variability could occur in 22q11 deletion syndrome
(22q11DS) patients. Differences in deletion size do not appear to be the cause. More than 85% of
patients show the same 3Mb deletion1 and there is no correlation between deletion size and number or
severity of symptoms.2 Imprinting is seen in a microdeletion at 15q11-13. Different genes in this region
are methylated and silenced in either the maternal or paternal copy, which causes a different syndrome
dependant on which copy remains: Prader-Willi syndrome (maternal) and Angelman syndrome
(paternal).3 There does not appear to be a parental difference in the deletions in 22q11DS, but one study
showed a correlation between grey matter volume and parental origin of deletion,2 so some symptoms
may show variability due to imprinting. The deletion could cause chromatin remodelling in the
surrounding region, the extent of which could vary between individuals and affect the expression of genes
around the deleted region. In the Df1 mouse model of 22q11DS, no deficiency is seen when mice have
one chromosome with a deletion and one with a duplication, making this explanation unlikely.4
         Mutations or polymorphisms in genes at other, modifier loci in the genome could affect the
penetrance or expression of a symptom. A simple example is a digenic disease like autosomal dominant
polycystic kidney disease. Mutations in the gene PKD1 cause disease, but an additional mutation of
PKD2 increases severity. Mutations in PKD2 alone do not cause disease.5 Many modifier loci, including
some symptom-specific, could be involved in 22q11DS, creating more variability. There is no way to test
this, as no loci are currently known. In the Df1 mouse model, the strain of mouse has a significant effect
on phenotype. Congenic mouse models show that this difference is not due to alleles within the deleted
region, suggesting modifier loci.1 The syndrome varies even in genetically identical individuals
(monozygotic twins and inbred mice),6 so there are also non-genetic factors. Environmental factors,
before or after birth could affect phenotype. Differential blood flow to monozygotic twins can cause size
differences, or even cardiac anomalies.6 Lastly, there could be stochastic factors affecting the phenotype.

b) The phenotype can include schizophrenia. How might genes involved in schizophrenia be
identified by study of the 22q11 deletion patients?
MEDG 520 Midterm                                                         Yvonne Bombard

Sample Answer #1 for question 1b
        Schizophrenia is a common disorder characterized by cognitive, social and psychiatric
impairments. It is one of the phenotypes exhibited by 22q11deletion syndromes. In a study of 50 adults
with VCFS, Murphy et al. (1999) found 42% of the individuals to have a major psychiatric disorder, with
24% fulfilling the criteria for schizophrenia. Given such high incidents of the disorder among VCFS
recipients, genes involved in schizophrenia may be identified by studying individuals with VCFS.
        Such a study may be conducted by organizing VCFS patients into two groups. Group A should
consist of cases that are diagnosed with schizophrenia as part of the VCFS phenotype and group B of
VCFS cases that do not exhibit schizophrenia (For control purposes, normal individuals should also be
included as group C). It is reasonable to assume that because of the strong variability in phenotype (ie.
presence vs. absence of schizophrenia), small differences in the 3Mb deletion sites will be present. A
normal copy of 22q11 and neighbouring regions may be inserted into a BAC to construct a high
resolution array platform that is specific to this region of interest. An array CGH experiment using this
platform can be conducted to determine the EXACT deletion boundaries. Subtle differences in the regions
of deletion between groups may be highlighted in this manner. Differentially deleted regions must then be
sequenced in order to identify possible susceptibility genes for schizophrenia. Alternatively, gene
expression may be measured for each group using a cDNA array commercially available for chromosome
22. Variation in gene expression between Group A and B may be caused by an additional microdeletion
in group A that is manifested through the schizophrenia phenotype. Finally, to determine if the novel
deletion in Group A contains the culprit gene(s) for schizophrenia, a mouse knockout model exhibiting
schizophrenia in the absence of this region (genes) is necessary.

Sample Answer #2 for question 1b
        Schizophrenia has a known association with the chromosome 22q11 deletion. If it is a primary
association, as opposed to a secondary result of the developmental problems associated with 22q11
deletions, it is possible that the gene or genes responsible for schizophrenia are located within the deletion
region. Linkage studies have identified at least 9 separate linkage regions associated with schizophrenia
including 22q11 (O’Donovan et al., 2003). However, after 15 years of such studies, very few highly
significant loci have been identified. More sensitive approaches are required to further localize the
schizophrenia gene(s). I will consider three of the possible approaches:
        (1) First, we could correlate variability in the 22q11 deletion phenotype with allelic variation in
genes from the deleted region of the normal copy of chromosome 22 by association studies. Genes to
focus on could be identified by expression studies (see below) or putative function. For example, the
gene UFD1L maps to 22q11 and is thought to be involved in neural development (Pizzuti et al, 1997).
        (2) Gene expression studies using microarray analysis to compare 22q11 patients to normal
individuals can be used to detect differentially expressed genes. Differential expression could reflect the
loss of genes or regulatory elements responsible for the phenotype. This approach was used with some
success by Hakak et al. (2001). The authors identified several genes significantly down-regulated in post-
mortem brains of schizophrenics. These results were subsequently replicated (Davis et al. 2003) and one
of the genes turned out to be a neuregulin receptor. The gene for neuregulin (NRG1) has since been
identified by haplotype linkage disequilibrium analysis as strongly associated with schizophrenia. A
similar analysis of 22q11 patients might identify other schizophrenia susceptibility genes.
        (3) A more directed approach might involve finding orthologous genes in mouse models, creating
deletion/knockout mutants, and looking for schizophrenia–like symptoms. This approach worked for the
identification of Tbx1 as the gene responsible for cardiovascular defects in 22q11 mice models (Taddei et
al., 2001). This gene may also be involved in schizophrenia and others could be identified in the same
MEDG 520 Midterm                                                       Yvonne Bombard

Sample Answer #3 for question 1b
        A candidate gene for schizophrenia at 22q11 would likely be deleted in all 22q11DS patients with
schizophrenia. The critical deletion region could be examined for genes potentially involved in a
neurological disease. If a mutation screen in non-deletion schizophrenia patients detected point
mutations, this would be good indication that gene has a role in schizophrenia.
        Alternatively, genes at other loci could be examined for their role in 22q11DS schizophrenia by
performing microarray analysis on 22q11 patients with and without schizophrenia. One concern would be
that the variability of the syndrome would make it difficult to determine if the results showed genes
involved in schizophrenia, or were due to a different variation. It would be important to try to limit the
variability of other phenotypes as much as possible to prevent this. The best candidates for this type of
analysis would be related individuals, especially monozygotic twins, which appeared to only show a
difference in the schizophrenia phenotype. Analysis of genes differentially expressed in schizophrenia
versus non-schizophrenia patients could give a list of candidate genes for schizophrenia involvement in
22q11DS. These could be genes that are regulated by, or interact with, genes in the 22q11 region in a
pathway important in schizophrenia. Screening of these candidate genes in non-deletion schizophrenics
could determine if these genes had a role in non-22q11DS schizophrenia as well.

c) There are suggestions that epigenetic mechanisms may be involved in the genesis of
schizophrenia. Critically discuss evidence for the role of epigenetics in the development of
schizophrenia and suggest ways to analyze this putative involvement

Sample Answer #1 for question 1c
        The definition of the term epigenetics is “heritable changes in gene expression that occur without
change in DNA sequence” (1). These changes outside the DNA sequence itself may alter the phenotype of
a particular individual when compared to another. Epigenetic effects are thought to be related to DNA
methylation (1), methyl CpG, binding proteins, histone modifying enzymes, chromatin remodeling
factors, transcriptional factors, chromosomal proteins, intra and extracellular events and pre and postnatal
environmental effects (2). DNA methylation is subject to changes by cellular and environmental
influences. Significant differences in degree of methylation are present in different individuals with the
same phenotype and these methylation differences have been suggested to predispose an individual to
different mutations (3) and/or differential gene expression (4). These epigenetics effects have been
postulated to have an important role in complex disorders such as MS, major psychosis including
schizophrenia and inflammatory bowel disease (5). Some studies have suggested an association between
seasonal environmental factors during the pre and perinatal period, HLA-DR1 and schizophrenia. The
suggested hypothesis is that development of schizophrenia may be mediated not only by a direct insult to
the brain but that an alteration of epigenetic status of those genes that may later impact the development
of the brain and that genes that are expected to be expressed may in fact be silenced by abnormal
methylation. (6). In order to evaluate some of the putative epigenetics effects in schizophrenia some
studies have used epigenetic interpretation of genetic association studies of genes such as the serotonin
2A and dopamine D3 receptors. The results of this study showed that there were differences in
methylation and expression between affected and non-affected individuals in these two alleles and that
these differences may have been related to a specific T>C polymorphism in the serotonin and the
dopamine receptor (7). There are a number of methods to test for epigenetics effects. Methylation status
can be tested with sensitive and methylation insensitive restriction enzymes (8), bisulphate-induced
modification of genomic DNA, restriction landmark genomic scanning (RLGS), a method that uses two
dimension electrophoresis to visualize a large number of loci and differentially methylated DNA
fragments (9), methylation sensitive representational difference analysis which can be used for cloning
DNA sequences that are subject to differential DNA methylation (10). Epigenetic effects may also relate
MEDG 520 Midterm                                                         Yvonne Bombard

to chromatin structure and histone acetylation. Gene expression is related to histone hyperacetylation (11)
and histone acetylation can be measured with anti-acetylated histone antibodies and quantitative PCR

1) Wolffe and Matzke. Proc Natl Acad Sci 1999;96:481-486.
2) Nakao et al., Gene 2001;278:25-31.
3) Mancini et al., Am J Hum Genet 1997;61:80-87.
4) Grekova et al., J Neuroimmunol 2000;106:214-219.
5) Petronis A. Trends in Genet. 2001;17:142-145.
6) Narita et al., et al., Am J Psych 2000;157:1173-1175.
7) Petronis A. Neuropsychopharmacol 2000;23:1-12.
8) Frommer et al., PNAS 1992;89:43-49.
9) Kawai et al., Nucleic Acid Res 1993;21:5604-5608.
10) Toyota et al., Cancer Res 1999;59:2307-2312.
11) Fletcher and Hansen. Gene Expres 1996;6:149-188.
12) Coffee et al., Nat Genet 1999;22:98-101.

Sample Answer #2 for question 1c
        In addition to uncertainty about genetic influences in schizophrenia, there is still uncertainty
surrounding the disease mechanism. Some researchers have noted abnormal brain development,
especially cortical abnormalities and lower cortical activity,7 but not in all patients.8 Some patients have a
high density of dopamine receptors, suggesting hyperactivity of the dopamine pathway as a possible
mechanism. In addition, the dopamine receptor DRD2 has been linked to schizophrenia in a subset of
        Epigenetics could be involved in a number of ways: DNA methylation, chromatin acetylation or
remodelling, or otherwise altered gene regulation. Genes involved in schizophrenia could become
differentially methylated or acetylated in different individuals, in response to environmental factors, or by
a stochastic mechanism. This would lead to a difference in gene expression at the corresponding loci.
Methylation at the DRD2 locus has already been shown to correlate with disease presence.9 Gene
expression could be also be altered by inserting viral DNA into the genome at or near a schizophrenia
gene locus. Evidence shows that maternal influenza causes an increased risk of schizophrenia,
presumably through viral integration into in the patient’s genome that disrupts a coding region, or allows
viral production of toxic substances that affect development.7 Transposable elements in the genome,
could cause epimutations by similar mechanisms. Insertion near a schizophrenia gene could cause
increased methylation in the region to silence the retroelement, which could also cause gene silencing.7
Alternately, not all retroelements are perfectly inactivated, so an active retroelement could cause
abnormal gene expression through the use of the retroelement promoter, or silence the gene by blocking
enhancers, sequestering transcriptional machinery or by producing antisense transcripts.10 Since
epigenetic mutations can be inherited, or wiped clean during embryogenesis, these mechanisms could
explain both inherited and sporadic cases.7
        A drawback of examining epigenetics is that it is more difficult than studying traditional genetic
mutations. Currently, to study methylation, one must have an idea of the gene involved, and either use
southern blotting with methylation sensitive enzymes, or sodium bisulfite sequencing to look at
methylation patterns in genes of interest. There is also the difficulty of which regions to look at in
methylation studies, as regulatory regions would be the preferred targets, but these are often poorly
defined.7 One method being developed combines sodium bisulfite treatment with microarray analysis,
making it possible to look at many regions of many genes, without knowing which genes one wants to
look at in advance.11 Chromatin modification can be examined by using antibodies specific for certain
MEDG 520 Midterm                                                        Yvonne Bombard

histone modifications. One other method would be to examine the role of retroelements near potential
schizophrenia genes.

Marking Outline for Question 1
  A) How to account for variability:
         a. Modifiers
         b. Somatic events and mosaicism – “self repair”
         c. Environment
         d. Variable deletion size and position effects
         e. Imprinting
         f. Variable expression of loci on remaining copy

   B) How to find schizophrenia gene in 22q11 deletion patient
        a. Define minimal region
        b. Candidate Genes
        c. Expression
                 i. Function – other genes with disease
                ii. Point mutations in non-deleted patients with same phenotype
        d. Mouse model
        e. Allelic expression changes in deleted individuals
                 i. Association studies

   C) Critically discuss epigenetics and suggest ways to analyze
          a. Differences in methylation
          b. Differences in RNA level
          c. Differences in chromatin (histone modifications, DNAse access)

Question #2
You are the head of a large, well-funded genetics laboratory located in a major teaching hospital in
Baltimore. The goal of your laboratory is to identify genes that contribute to prostate cancer.
Epidemiological studies have noted that the incidence of prostate cancer is higher in African
American men than in European American men. The local admixed population includes men of
both backgrounds as well as many who are of both African and European ancestry.

What are the possible study designs you can use to identify prostate cancer genes? Briefly list the
strengths, weaknesses, and pitfalls of each. Indicate which method you would use and why.

Sample Answer #1 for question 2
         After securing appropriate ethical approvals, DNA samples (from families), and recording various
pertinent information such as ethnicity, severity of the disease (stage of cancer), age of onset, response to
various treatments of the admixed and non-mixed populations, an appropriate study design must be
contemplated. A possible study design that most effectively uses an admixture population in order to
identify susceptibility genes is mapping by admixture linkage disequilibrium (MALD). MALD is based
upon the concept that, when admixture occurs between two populations, linkage disequilibrium (LD) is
initially created between all loci that have large allele-frequency differences between the two populations.
With successive admixed generations, the LD between unlinked loci quickly weakens, whereas the LD
MEDG 520 Midterm                                                         Yvonne Bombard

between linked markers persists for many more generations. Thus, a recently admixed population will
have much larger regions of LD between loci than are seen in a standard population (Stephens et al 1994).
If any disease-susceptibility alleles or disease-protective alleles are present in a sufficiently different
frequency distribution in the parental populations, then MALD can be used to map the susceptibility gene
in the admixed populations. The greater the LD in the admixed population will thus theoretically translate
into less demanding requirements for both marker saturation and sample size (Stephens et al 1994).
         MALD is a powerful technique as compared with association-based genome scans which likely
require more than 50,000 markers, MALD requires only 500-2,000 markers. Another strength of MALD
is compared with general association studies that are affected by allelic heterogeneity, MALD has the
advantage of not being deterred by multiple independent mutational events, since only an allele’s ethnic
identity is used in computations (Collins-Schramm et al 2002). A pitfall associated with MALD, however,
is MALD can map only disease-associated alleles that are present in different frequencies in the two
parental populations, and the increased regions of LD may hinder fine-scale mapping (Collins-Schramm
et al 2002). In addition, as with all linkage studies, power may be a problem if sample sizes are not
sufficiently large. Admixture generates linkage disequilibrium that can extend for many centimorgans and
can persist for as many as 20 generations. African American men and European American men are ideal
for MALD-based association studies. Studies have shown that African Americans represent an admixed
population with significant genetic contributions from both African and European ancestors (Chakraborty
et al 1991).
         After using a MALD-based approach to identify polymorphic markers, it is necessary to define the
critical DNA region. Thus, clones of the genes must be generated. Mapped single-copy probes from the
region can then be applied to DNA isolated from a metastatic prostate cancer containing a chromosome
homozygous deletion. It is also possible to find genes from a SNP map or LOH map that may be available
on public databases. Alternatively, it may be possible to screen for a mutation in the admixed sample.
Candidate tumor suppressor genes in this region may be identified in these ways.
         In order to test for association of the candidate gene or variant with prostate cancer, a case-control
association design would be necessary. This would involve genotyping the gene or variant in
chromosomes from prostate cancer patients and age- and ethically-matched controls representing African
Americans, European Americans as well as the admixture. In order to control for population stratification,
detection of stratification is done by genotyping a few unlinked genetic markers. Then, the calculated
association statistic is divided by the average of association statistics across the unlinked markers in order
to correct for population stratification (Kittles et al 2002).
         In order to provide further evidence of the gene’s candidacy, various indications would prove
helpful. This may include finding multiple mutations within the same gene in different people,
demonstrating segregation of the mutation within a family if there is a family history, demonstrating gene
function related to cancer susceptibility, a bioinformatics approach to compare against other species,
examining gene expression, or developing and mouse model.
         An alternative to MALD is the use of loss of heterozygosity (LOH) mapping to identify
polymorphic markers to narrow down a chromosomal region of interest. Loss of heterozygosity and
allelic imbalance can be detected by either radioactive labeling of PCR products with subsequent scoring
of autoradiographs and by FISH. Polymorphic microsatellite loci are the most common marker type used
in these studies. This method has several drawbacks, especially the use of radioactivity and
interpretative/technical problems. The use of fluorescently-labeled primers, automated DNA sequencing
coupled with a computer software package obviates these problems. This technique has the added
advantages of conferring high level of specificity as well as the ability to analyze several microsatellites in
large numbers of cases, simultaneously. It has the disadvantage that FISH is not suitable technology to
use for high-throughput screening.
         Another option is a haplotype–based association study of multilocus linkage disequilibrium which
uses a high density array of biallelic markers, around two informative markers within a chromosomal
MEDG 520 Midterm                                                        Yvonne Bombard

region of interest to build haplotypes in case and control samples. The region is amplified by polymerase
chain reaction (PCR) and this fragment is then used as a template for PCR to generate a product
incorporating both of the marker loci. PCR products are then digested and subjected to polyacrylamide gel
electrophoresis. The distinct pattern of bands appearing on the gel is then used to ascribe haplotypes to
each subject. By comparing prostate cancer subjects and unaffected controls, differences in allele,
genotype, and haplotype frequencies of several markers around the gene differences may be found
between case and control subjects. Haplotype-based association studies offer a significant advantage:
genomic regions that can be tested for association without requiring the discovery of the functional
variants. A drawback of this approach, however, is the need for a very dense maker map in the range of
30,000-1,000,000 variants.
        Alternatively, genome-wide scans may be used to identify prostate cancer quantitative trait loci
(QTLs). The product is a dense map of 30,000-500,000 highly polymorphic markers which can be used to
scan the human genome for haplotypes associated with common diseases. This method attempts to
identify additional loci involved in prostate tumorigenesis in DNA samples from prostate cancer patients.
The drawbacks of this approach include the need of large amounts of markers, and may be time
consuming. In addition, interpretive difficulties crop up as it may be difficult to distinguish between true
and false positives using the appropriate statistical methods (Visscher and Haley 2001) . However,
genome-scans achieve a comprehensive set of markers that lend themselves for fine-mapping. It would be
wise to replicate this study in a second, independent population.

Chakraborty et al (1991). ‘Unique’ alleles in admixed populations: a strategy for determining ‘hereditary’
      populations differences of disease frequencies. Ethn Dis 1:245-256.

Collins-Schramm et al (2002) Markers that discriminate between European and African ancestry show
       limited variation within Africa. Hum Genet. 111(6):566-9.

Kittles et al (2002). CYP3A4-V and prostate cancer in African Americans: causal or confounding
        association because of population stratification? Hum Genet. 110(6):553-60.

Paris et al (1999). Association between a CYP3A4 genetic variant and clinical presentation in African-
        American prostate cancer patients. Cancer Epidemiol Biomarkers Prev. 8(10):901-5.

Smith et al (2001). Markers for mapping by admixture linkage disequilibrium in African American and
       Hispanic populations. Am J Hum Genet. 69(5):1080-94.

Stephens et al (1994). Mapping by admixture linkage disequilibrium in human populations: limits and
       guidelines. Am J Hum Genet 69:809-824.

Visscher and Haley (2001) True and false positive peaks in genomewide scans: The long and the short of
       it. Genet Epidemiol. 20(4):409-14.

Xu et al (2001). Linkage and association studies of prostate cancer susceptibility: evidence for linkage at
       8p22-23. Am J Hum Genet. 69(2):341-50

Sample Answer #2 for question 2
         Prostate cancer is the most frequently diagnosed non-cutaneous malignancy among men (Simard
et al., 2003). To date, no conclusive evidence has clearly identified the causative gene or genes. I will
MEDG 520 Midterm                                                        Yvonne Bombard

assess three basic strategies and examples of their usage to identify causative genes: linkage, association,
and expression analyses.
         Linkage makes use of well characterized pedigrees to identify haplotypes that are inherited intact
over several generations. Tavtigian et al (2001) used linkage analysis to identify a locus on chromosome
17p linked to prostate cancer. Positional cloning followed by mutation screening for genes in the region
led to the identification of a gene, ELAC2, along with two variants, Ser217Leu and Ala541Thr, that seem
to confer prostate cancer susceptibility. A potentially useful study could investigate whether these
polymorphisms or others are more common in African populations.
         Linkage analysis, combined with positional cloning, is a very powerful method for the detection of
loci responsible for simple Mendelian phenotypes. Historically it has had a very low false-positive rate
when a stringent LOD-score of 3.0 is used (Risch, 2000). However, linkage tends to identify very large
regions encompassing hundreds or even thousands of genes. It is less effective for the detection of genes
with more subtle effects such as those responsible for most complex common diseases. So far, all genes
identified by linkage and positional cloning, even those for complex diseases, display mendelian or near-
mendelian inheritance (reviewed in Risch, 2000). In the case of complex diseases, linkage analysis tends
to identify rare mutations or polymorphisms unique to only a small subset of the diseased population (ie.
Rare simple causes for otherwise complex diseases).
         Association relies on retention of adjacent DNA variants over many generations (in historic
ancestries) and does not require specific knowledge of pedigrees. Until recently this was limited to a few
thousand DNA variants, but, with the advent of SNP genotyping, characterized variants now number in
the millions. An extension of the basic association analysis is the transmission disequilibrium test (TDT).
TDT is a combined test of linkage and association that examines the transmission of potential disease
alleles from a parent who is heterozygous for a marker to the affected offspring (Cardon and Bell, 2001).
Following the linkage analysis identifying ELAC2 and its common missense mutations (Tavtigian et al.,
2001) numerous studies were conducted to associate those mutations with familial prostate cancer
(reviewed in Simard et al., 2003). Results varied from significantly higher carrier frequency to no
evidence of increased risk at all. Camp and Tavtigian (2002) performed a metanalysis of all association
data published to date and concluded that carriage of the Leu217/Thr541 allele is associated with prostate
cancer but significance depends on the control group used.
         Association studies are generally much more powerful than linkage analyses when it comes to
predicting genetic components of complex diseases. Disease associated regions will be much smaller
than in linkage analysis, often encompassing only one gene or gene fragment. However, there have been
difficulties with reproducing these association studies. This could be the result of poor study designs,
incorrect assumptions about the underlying genetics of the population, or overinterpretation of the data.
To be effective, association studies often require very large sample sizes, of as much as several thousand
patient samples (Cardon and Bell, 2001). Many mutations or polymorphisms in a gene can create the
same phenotype. Each will have its own ancestral haplotypes and thus, reduce the power to detect
associations between the phenotype and a specific allele.
         Expression analysis refers to the use of relatively modern techniques such as DNA microarrays to
study differences in gene expression between normal and affected tissues. These studies while rarely
conclusive can be effective at identifying subsets of potential disease genes. These genes can then be
studied further with more definitive strategies. Rishi et al (2003) recently combined epidemiological
information with expression studies to propose an interesting mechanism for prostate cancer. They found
that African-American men showed reduced expression of two zinc transporters when compared to
European-American men. Previous studies have shown that reduced zinc levels are associated with
prostate cancer. The authors hypothesize that African-Americans display reduced zinc transporter levels
because of their historical origin from a region known for high levels of zinc in the water (the mineral-rich
continent of Africa). If confirmed, this study potentially explains the epidemiological data linking
prostate cancer to race and offers a relatively simple preventative cure (zinc supplements).
MEDG 520 Midterm                                                        Yvonne Bombard

         Microarrays offer biologists the kind of information that they could only dream of 10-15 years
ago, a complete snapshot of the state of all biological processes in a tissue or cell for a specific time and
set of conditions. Identifying groups of coexpressed genes should be useful in elucidating pathways,
characterizing gene regulation, and identifying relevant genes for the condition of interest. However this
“guilt by association” has fallen out of favour because of difficulties with reproducibility and achieving
statistical significance (Quackenbush, 2003). Also, microarray experiments are very expensive, generally
preventing the number of replicates required to accurately differentiate random biological and
experimental variability from the actual causative differences in expression. Results generally require
corroborating evidence such as quantitative RT-PCR experiments.
         All three of the approaches outlined have their own strengths and weaknesses. As the head of a
well funded genetics lab, I would use a complementary combination of the three. Expression studies
could be used to identify a list of candidate genes. SNP genotyping could identify haplotypes for these
genes and TDT used to identify alleles associated with the disease in affected families. Knowledge of the
epidemiology would help identify ancestral haplotypes more prevalent in certain populations. Once
potential genes are identified they can be confirmed by gene mutation studies in mice models. Sufficient
traditional linkage studies may already have been conducted but, a good example of the value of
combining approaches was recently illustrated by Schadt et al (2003). The authors used RNA transcript
abundances (from expression studies) as quantitative trait loci (so called eQTLs) in a linkage analysis.
This approach identified gene expression patterns for two distinct sub-types of obesity, and two potential
obesity genes in mice models. This study demonstrated a method for combining linkage and expression
analyses into an approach more powerful than either on their own. I believe these kinds of multi-
disciplinary strategies will be the most effective for understanding and treating complex diseases like
prostate cancer.


Rishi I, Baidouri H, Abbasi JA, Bullard-Dillard R, Kajdacsy-Balla A,
Pestaner JP, Skacel M, Tubbs R, Bagasra O. 2003. Prostate cancer in African American men is
associated with downregulation of zinc transporters. Appl Immunohistochem Mol Morphol. 11(3):253-

Camp NJ, Tavtigian SV. 2002. Meta-analysis of associations of the Ser217Leu and Ala541Thr variants
in ELAC2 (HPC2) and prostate cancer. Am J Hum Genet. 71(6):1475-8.

Cardon LR, Bell JI. 2001. Association study designs for complex diseases. Nat Rev Genet. 2(2):91-9.

Quackenbush J. 2003. Genomics. Microarrays--guilt by association.
Science. 302(5643):240-1.

Risch NJ. 2000. Searching for genetic determinants in the new millennium. Nature. 405(6788):847-56.

Schadt EE, Monks SA, Drake TA, Lusi AJ, Che N, Colinayo V, Ruff TG,
Milligan SB, Lamb JR, Cavet G, Linsley PS, Mao M, Stoughton RB, Friend SH. 2003. Genetics of gene
expression surveyed in maize, mouse and man. Nature. 422(6929):297-302.

Simard J, Dumont M, Labuda D, Sinnett D, Meloche C, El-Alfy M, Berger L, Lees E, Labrie F, Tavtigian
SV. 2003. Prostate cancer susceptibility genes: lessons learned and challenges posed. Endocr Relat
Cancer. 10(2):225-59.
MEDG 520 Midterm                                                       Yvonne Bombard

Tavtigian SV, Simard J, Teng DH, Abtin V, Baumgard M, Beck A, Camp NJ,
Carillo AR, Chen Y, Dayananth P, Desrochers M, Dumont M, Farnham JM, Frank D, Frye C, Ghaffari S,
Gupte JS, Hu R, Iliev D, Janecki T, Kort EN, Laity KE, Leavitt A, Leblanc G, McArthur-Morrison J,
Pederson A, Penn B, Peterson KT, Reid JE, Richards S, Schroeder M, Smith R, Snyder SC, Swedlund B,
Swensen J, Thomas A, Tranchant M, Woodland AM, Labrie F, Skolnick MH, Neuhausen S, Rommens J,
Cannon-Albright LA. 2001. A candidate prostate cancer susceptibility gene at chromosome 17p. Nat
Genet. 27(2):172-80.

Marking Outline for question 2
Explanation of possible methods (max 60% for this part)
Score up to 30 points each for any two of the following that was explained clearly with strengths and
    A) Linkage
    B) TDT or sibpair analysis
    C) Association
    D) LOH mapping in tumors
    E) Animal models
    F) Cytogenetics
    G) Expression methods
    H) Other methods (ex: transfection to find activated oncogenes)
10 point bonus for this section if you discussed admixture mapping (MALD)

Score up to 40 points for your proposed plan:
Plan for gene identification was rational, and explained clearly                            30 points
Plan for gene identification was rational, and explained clearly and likely to be effective 35 points
Plan for gene identification was rational, and explained clearly and innovative             40 points

Question #3
a) Does clonal exhaustion of stem cells ever occur? Describe and discuss the evidence from animal
models and human disease.

Sample Answer #1 for question 3a
        The unique properties of adult stem cells to self -renew and proliferate for extended periods, make
it tempting to imagine these cells immortal. Realistically however, there are several biological events that
may lead to clonal exhaustion, a few of which will be discussed here.
        Telomeres are short tandem repeats of (TTAGGG)n at the ends of chromosomes and are essential
for maintaining chromosomal integrity. The repeats shorten with each cell division and their loss acts as a
signal to the cell to exit the cell cycle. Telomeres are replenished by the reverse transcriptase enzyme,
telomerase. In a study by Notaro and colleagues (1997), the telomere lengths of peripheral blood
granulocytes in eleven fully engrafted bone marrow transplant recipients were compared with their
corresponding donors. In ten of the eleven recipients, the telomere lengths were found to be significantly
shorter than the donors. This finding implies that telomerase activity is not able to prevent telomere
shortening under demanding proliferation conditions. Loss of telomeres as a result of insufficient
telomerase activity due to stress conditions or mutations in the telomerase gene may result in stem cell
death. Incidentally, in patients with acute myeloblastic leukemia (AML) and acute lymphoblastic
leukemia (ALL), telomerase activity is more than 10 –fold compared to normal hematopoietic cells,
MEDG 520 Midterm                                                         Yvonne Bombard

implying that an increased telomerase activity in leukemia-driven stem cells may be rendering these cells
‘immortal’(Ohyashiki et al., 2002).
         Further, Bummendorf and colleagues (2001) have found that in patients with moderate
depression of blood counts, but a long history of disease, there is a significant decrease of telomere length.
This finding reflects a gradual depletion of the stem cell pool, with progressively fewer stem cells
available to sustain hematopoisis. Consistent loss of telomeres is therefore a mechanism by which the
stem cell pool becomes reduces leading to clonal exhaustion over time.
        The stem cell pool may face exhaustion as a result of molecular events that lead to the disruption
of mechanisms involved in self-renewal regulation. Bmi-1is a member of the Polycomb Group (PcG)
family of transcriptional repressors that control development by the regulation of cell growth and
differentiation genes. Bmi-1 is essential for generation of self- renewing adult HSC. Park and colleagues
(2003), have found that compared to Bmi-1+/+ and Bmi-1+/- littermates ; in the bone marrow of Bmi-1-/-
mice, the frequency of HSCs ranges from 0% to 0.013%, with an average of 0.005% and an overall 10-
fold decrease in HSCs when taking the total number of bone marrow cells into account. Furthermore,
competitive reconstitution experiments using fetal liver cells and bone marrow from Bmi-1-/- mice
contributed transiently to hematopoiesis, with no detectable self renewal of adult HSCs. Gene expression
analysis of Bmi-1-/- mice revealed an alteration in the expression of p16 Ink4a and p19Arf which result in
proliferative arrest and p53-dependent apoptosis respectively. These findings strongly suggest that Bmi-1
may be required for self renewal / maintenance of adult HSCs and in its absence, the stem cell pool will
become exhausted.
        To avoid clonal exhaustion, stem cells must maintain relative quiescence. However alterations in
the circuitry that guides the cell cycle could result in the premature release of stem cells from G0. Cheng
and colleagues (2000) have demonstrated that mice lacking the cyclin-dependent kinase inhibitor p21;
exhibit increased stem cell numbers under homeostatic conditions, but stem cells become exhausted and
unable to self-renew when transplanted.
        The experimental and clinical evidence presented above can only begin to demonstrate the
biological complexity surrounding stem cell renewal properties. It is clear from all of these examples that
any number of changes may result in stem cell exhaustion. Given that stem cells last a long time and
divide frequently, they are more likely to accumulate mutations, some of which may alter the delicate
balance needed for their continual renewal.

Sample Answer #2 for question 3a

         Despite forty years of research, the renewal capacity of stem cells remains somewhat elusive. The
replicative potential of stem cells is clearly greater than is necessary in the duration of any normal
lifetime, as diseases associated with loss of stem cells, such as aplastic anaemia, are rare (1). However,
because the circumstances under which we test the extent of stem cell renewal are largely artificial, we
cannot be certain of their replicative potential beyond a normal lifespan. Some studies suggest that stem
cell self-renewal is limited by the cytokine environment, which differs markedly between embryos and
adults (2). In culture, the loss of replicative capacity of mouse and many human cells may be the result of
transferring the cells from their natural environment to the conditions of cell culture (3).
         Though not conclusive, a large amount of evidence supports the idea that stem cells have a finite
replicative potential. Due to the end-replication problem of DNA, small amounts of telomeric DNA are
lost with each division of a stem cell (4). Transplantation experiments have shown that bone marrow
stem cells can be serially transplanted to irradiated mice three to five times before their regenerative
capacity is severely diminished (5, 6, 7, 8, 9). However, after transplantation, the telomeres in the cells of
recipients are shorter than those of the donors (8, 9). Even in the primary recipient, the hematopoietic
system is not restored to its normal level (1). In addition, embryonic and fetal hematopoietic stem cells
MEDG 520 Midterm                                                        Yvonne Bombard

have been shown to have a higher proliferative potential than adult hematopoietic stem cells, supporting
the hypothesis of a finite replicative potential (10, 11).
         A clue in determining the replicative potential of stem cells may lie in the activity of telomerase,
the enzyme responsible for adding telomeric DNA to chromosome ends to offset telomere loss (4). Stem
cells have telomerase activity; nonetheless, telomere shortening still occurs as stem cells age in both
humans and mice (12, 13). When telomerase deficient mice were used as donors for hematopoietic stem
cell transplantations, only two rounds of transfer were successful, and the rate of telomere shortening
increased (9). The absence of telomerase in human cell cultures also leads to progressive shortening of
telomeres, indicating that telomere length may act as a mitotic clock to limit lifespan (3). In terms of
disease models, the majority of cancer types have been shown to have telomerase activity (14). Cancer
cells are effectively immortal, supporting the role of telomerase in preventing clonal exhaustion.
Nevertheless, cancer cells may have undergone additional changes, such as mutations of tumour
suppressor and cell cycle checkpoint genes, which could also lead to immortality. Patients with the
genetic disease dyskeratosis congenita have strikingly shorter telomeres than normal individuals (3, 15).
Without treatment, the result of the disease is commonly death due to bone marrow failure, implicating a
decreased ability of stem cells to replicate (3, 15). Telomerase appears to slow telomere erosion, allowing
cells to live longer than they would without it, but not indefinitely.

b) Stem Cell transplantation is routinely used as a therapeutic approach to a subset of human
diseases. Recent clinical trials and experiments in animal models provide evidence for or against
the use of Stem Cells in additional situations. Formulate and justify a hypothesis as to which
pathologies will be more amenable to be treated with stem cells in the next ten years.

Sample Answer #1 for question 3b
         Stem cell transplantation is currently a common treatment for myeloproliferative and other blood
disorders, such as chronic myelogenous leukemia and thalassemia major (16, 17). It is the opinion of this
writer that in the next ten years, stem cell transplantation will most likely be successful in treating
autoimmune diseases, tissue specific damage, and possibly in regenerating solid organs.
          Clinical trials of stem cell transplantation have shown promising results for several autoimmune
diseases, including severe rheumatoid arthritis, juvenile idiopathic arthritis, systemic sclerosis and
systemic lupus erythematosus (18, 19). The idea behind stem cell treatment for autoimmune disease is to
reset the immune system by stem cell-rescue following immunoablation (20, 18, 21). In mouse models,
bone marrow transplantation has been shown to correct the T-cell dysfunction that causes autoimmune
disease (22). Additional support for this type of treatment comes from patients who had coexisting
autoimmune diseases at the time of bone marrow transplantation for hematopoietic or other malignancy.
Following transplantation, the autoimmune diseases of most patients went into remission (23). In most
trials, long term remissions have occurred, indicating good benefits for the patient (18). The treatment
modality appears feasible, and because the procedure involves immunoablation, no immune response
should occur. However, it may take a year or more before normal humoral immunity is completely
restored (20). Some autoimmune diseases, such as systemic sclerosis, have multi-system involvement
that causes greater risk during transplantation (20). Therefore, autoimmune diseases like rheumatoid
arthritis, which have no severe pulmonary, renal or cardiac involvement are more likely to be treated by
stem cell transplantation in the future (20).
         A second potential stem cell treatment could be in repairing infarcted myocardium in individuals
with coronary artery disease. Studies of transplanted human hearts have demonstrated that extracardiac
progenitor cells are capable of migrating to and repopulating damaged myocardium (24). In addition,
when bone marrow cells are transplanted to damaged myocardium in mice, the stem cells can generate
new myocardium (25). While this experimental data seems hopeful, stem cell transplantation has only
been used following surgical revascularisation of myocardium in humans (26). However, because bone
MEDG 520 Midterm                                                        Yvonne Bombard

marrow cells are collected from the patient receiving the transplant, there is no chance for graft rejection,
decreasing the risk of the procedure (27). Transplantation can be carried out by way of a balloon catheter
placed in the affected artery, preventing the need for further surgery (27). A challenge of this therapy is
to determine the best time for transplantation after the infarction so that the stem cells participate in
forming functional muscle rather than causing inflammation (27).
        Liver disease could also be amenable to stem cell therapy, though perhaps less easily than the
aforementioned pathologies. The regenerative capacity of the liver is well known, and three putative
sources of liver stem cells have been identified: bone marrow, oval cells, and hepatocytes themselves
(28). In one experiment, ninety percent of rats that underwent ninety percent hepatectomies showed
significant improvement in liver function after transplantation of primary adult human hepatocytes to the
spleen (29). Several groups have shown the ability to culture and differentiate stem cell-like populations
into hepatic lineages (29). Potentially, patients with severe liver disease could have partial hepatectomies
followed by liver stem cell transplants. Entire organs would not be necessary for transplantation; in this
way, many patients could be treated. Difficulties with this type of transplantation include the delay time
that would be necessary for liver regrowth as well as the potential for immune reaction to the transplanted
cells. Certainly, further clinical trials with all of these diseases will be of interest.

References (sample answer #1 for 3a and answer #2 for 3b)
1.) Van Zant G, Liang Y (2003). The role of stem cells in aging. Exp Hematol 31: 659-672.

2.) Iscove N, Nawa K (1997). Hematopoietic stem cells expand during serial transplantation in vivo
     without apparent exhaustion. Curr Biol 7: 805-808.

3.) Rubin H (2002). The disparity between human cell senescence in vitro and life long replication in
    vivo. Nat Biotech 20: 675-681

4.) Brummendorf T, Rufer N, Holyoake T, Maciejewski J, Barnett M, Eaves C, Eaves A, Young N,
    Lansdorp P (2001). Telomere length dynamics in normal individuals and in patients with
    hematopoietic stem cell-associated disorders. Ann NY Acad Sci 938: 293-304.

5.) Harrison D, Astle C, Delaittre J (1978). Loss of proliferative capacity in immunohemopoietic stem
    cells caused by serial transplantation rather than aging. J Exp Med 147: 1526-1531.

6.) Hellman S, Botnick L, Hannon E, Vigneulle R (1978). Proliferative capacity of murine hematopoietic
    stem cells. PNAS 75: 490-494.

7.) Brecher G, Bookstein N, Redfearn W, Necas E, Pallavicini M, Cronkite E (1993). Self renewal of the
    long-term repopulation stem cell. PNAS 90: 6028-6031.

8.) Brummendorf T, Rufer N, Baerlocher G, Roosnek E, Lansdorp P (2001). Limited telomere shortening
    in hematopoietic stem cells after transplantation. Ann NY Acad Sci 938: 1-8.

9.) Allsopp R, Morin G, DePinho R, Harley C, Weissman I (2003). Telomerase is required to slow
    telomere shortening and extend replicative lifespan of HSCs during serial transplantation. Blood 102:

10.) Lansdorp P, Draowska W, Mayani H (1993). Ontogeny-related changes in proliferative potential of
    human hematopoietic cells. J Exp Med 178: 787-791.
MEDG 520 Midterm                                                      Yvonne Bombard

11.) Lansdorp P (1995). Telomere length and proliferation potential of hematopoietic stem cells. J Cell
    Sci 108: 1-6.

12.) Monroe J, Rothenberg E (1998). Molecular biology of B-cell and T-cell development. NJ: Humana
    Press. 27-42.

13.) Bodnar A, Ouellette M, Frolkis M, Holt S, Chiu C, Morin G, Harley C, Shay J, Lichtsteiner S,
    Wright W (1998). Extension of life-span by introduction of telomerase into normal human cells.
    Science 279: 349-352.

14.) Shay J, Bacchetti S (1997). A survey of telomerase activity in human cancer. Eur J Cancer 33: 787-

15.) Wong J, Collins K (2003). Telomere maintenance and disease. Lancet 362: 983-988.

Marking Outline for question 3
No marking outline available.

Shared By: