In many studies, patient samples are divided into cohorts deÞned by age, usually with a cutpoint between 45 and 55 years. The rationale is that younger patients with RVO are less likely to have explanatory vascular risk factors, and may represent a pool of patients in whom genetic risks from thrombophilia could be more important. In most cases, this expectation has been unfounded.

In contrast to sickle cell anemia or HuntingtonÕs disease, but as in diabetes mellitus, a genetic component of RVO would be a complex trait. It has never been proposed that a single gene could determine whether a person will develop RVO. Rather, if there is a genetic component to RVO, it is likely not to be determinative, but to increase risk, by interaction among genes and the environment. The alleles, or versions of genes, that are associated with complex genetic diseases are often common variants. These alleles contribute to the risk of disease expression severity and age of onset.14 A recurring theme in the literature of thrombophilia and RVO is the importance of interactions of thrombophilia with exogenous factors such as oral contraceptive use or trauma.11

3.1 Background for Clinical Genetics

Lack of familiarity of genetic concepts by clinicians has often been identiÞed as an obstacle in progress toward understanding disease pathogenesis.26 Therefore, this section is a brief review of genetic concepts for understanding the relevant literature and associated terminology. Readers familiar with clinical genetics can skip directly to Sect. 3.4.

Of course, understanding molecular genetics begins with the study of deoxyribonucleic acid (DNA), a linear polymer comprising of two strands containing sequences of four nitrogencontaining bases Ð adenine (A), guanine (G), thymine (T), and cytosine (C) (Fig. 3.1).

DNA is the template for its own replication. DNA also serves as a template for the creation of ribonucleic acid (RNA) in a process called transcription. The RNA transcribed from the DNA in turn serves as the template for synthesis of proteins. In transcription, DNA is always read beginning at the 5¢ end and proceeding to the 3¢ end. The two strands of a DNA molecule are held together by hydrogen bonds between paired

Fig. 3.1 Each strand of DNA consists of a backbone of deoxyribose-phosphate sugars with attached purine and pyrimidine bases. The bases show complementarity by forming hydrogen bonds with paired bases in the fellow strand (Redrawn after Della16)

3' 5'

DNA double helix

Deoxyribosephosphate

backbones

Hydrogen bonds

bases. Adenine in one strand always binds to thymine in its fellow strand and likewise for cytosine and guanine. For example, if a section of the sequence of one strand is 5¢-ATGAC-3¢, then its fellow strand at this locus reads 3¢-TACTG-5¢. Binding of complementary strands of DNA or between a single strand of DNA and its complementary RNA strand is termed hybridization.24

The central principle of molecular biology is that DNA is transcribed to RNA, which is in turn translated into protein (Fig. 3.2). In the cell nucleus, transcription of DNA occurs Þrst as a messenger RNA precursor (mRNA) that contains the transcript of the protein-coding DNA sequence (exons), the nonprotein-coding DNA sequence (introns), and untranslated regions adjacent to the 3¢ and 5¢ termini. The transcribed introns are spliced out to yield a mature mRNA product

(Fig. 3.2). In the cell cytoplasm, mature mRNA is translated as three base segments called codons into amino acids that compose the protein product of the gene.60 This intricate process of RNAguided protein synthesis is called translation.

Humans have 23 pairs of chromosomes, each comprised of DNA and associated proteins. Each chromosome contains many genes, or sequences of DNA that code for proteins, as well as sequences of DNA dedicated to regulatory functions such as initiating transcription or translation. There are 3.3 billion base pairs in the human genome.14 The latest estimate on the number of genes in the human genome is 20,000Ð25,000.36 Ninety-nine percent of the genome does not code for proteins.79 These noncoding sections of the genome are broken into classes called introns and intergenic DNA. Introns, comprising 24% of the

Fig. 3.2 DNA is transcribed into an mRNA precursor that is reÞned to mature mRNA by the excision of sequences corresponding to introns and transported out of the nucleus into the cytoplasm. Translation of mature mRNA results in a protein (Redrawn after Della16)

Transcription

Protein

human genome, are segments of DNA adjacent to exons that are initially transcribed into an RNA strand but are then excised from initial RNA transcripts before the Þnal mRNA strand leaves the nucleus for the cytoplasm on the way to protein synthesis (Fig. 3.2). Intergenic DNA, comprising 75% of the genome, remains untranscribed and has regulatory and unknown functions.79 The remaining 1% of DNA composes the exons that code for proteins.12,79

3.2The Role of Polymorphisms in Genetic Studies

The DNA sequences of any two human beings are 99% identical. The 1% of DNA that differs between any two individuals constitutes the genetic basis of certain diseases. In addition, factors that control the expression of the genes contribute to different phenotypes including disease. The purpose of genetic investigations is to determine which genes contribute to disease. On average, a difference in DNA sequence between two persons is found once per 1,200 base pairs. When a variation of the genetic locus is found in at least 1% of a given human population, the variation is termed a polymorphism.

There are several categories of polymorphisms. Single-nucleotide polymorphisms (SNPs) are single substitutions of one base for another at a certain position. A common way of naming an SNP is exempliÞed by c.74C > G. This means that at the level of the coding DNA sequence (indicated by the c. preÞx), at position 74, a cytosine is replaced by a guanine. The Þrst letter characterizes the normal base, and the second letter following the arrowhead signiÞes the mutant base. However, there are many variations in nomenclature for SNPs. For example, some authors refer to mutations not at the level of the coding DNA, but rather at the level of the altered amino acid sequence in the protein product of the mutant gene. In using this nomenclature, the convention is to list the normal amino acid Þrst followed by the codon number in the sequence of the protein followed by the mutant amino acid. Thus, for example,

Ala276Glu means that at codon 276, glutamine has replaced the expected alanine.43

Another important class of polymorphisms is that of short tandem repeats (STRs) or microsatellites. These are strings of repetitive base pair sequences that vary in length between persons. Thus, the alleles are variations in the number of repeats. For example, at a given location, one might Þnd that one person shows CACA [or (CA)2], the next CACACA [or (CA)3], the next CACACACA [or (CA)4], and so on. Nucleotides are repeated in tandem a number of times. SNPs and STRs are scattered at different locations across the human genome, and encyclopedias of such polymorphisms have been compiled. Any part of the human genome can be probed by selecting an SNP or STR located nearby in the genetic map.

Yet another type of polymorphism, more often used in older genetic association studies, is the restriction fragment length polymorphism. Enzymes called restriction endonucleases cleave DNA at sites where certain DNA sequences are detected. There are many restriction endonucleases (Fig. 3.3). By analyzing the length of fragments of DNA produced by the action of restriction endonucleases, one can deduce the presence or absence of polymorphisms that result in cleavage sites.

Figure 3.4 shows an example of a map of genetic markers, in this case short tandem repeats, corresponding to chromosome 11. It also shows their position relative to a gene, PAX6, that is critical in ocular embryogenesis. As this Þgure demonstrates, one can judiciously choose a polymorphism and examine whether it is associated with a disease such as RVO. If there is an association between the polymorphism and the disease phenotype, then neighboring genes of this polymorphism are candidates for investigation of the pathogenesis of the disease. Neighboring genes become candidates because segments of DNA tend to be inherited together as a unit. Conversely, if one suspects that a certain gene is important in causing RVO, one could choose to study a polymorphism found in close proximity to that gene based on the genetic map and then look for associations of that polymorphism with presence or absence of the disease.

The success of such studies depends on how reproducible the classiÞcation of patients is by disease. In studies investigating genetic