HfStOpUthOlO@j 1992, 21, 303-313

INVITED REVIEW

Recombinant DNA technology and its diagnostic applications M. J.ARENDS & C.C.BIRD Department of Pathology, University Medical School, Edinburgh, Scotland, UK Accepted for publication 1 3 April 1992

ARENDS M.J.& BIRD C.C.

(1992) Histopathology 21, 303-313

Recombinant DNA technology and its diagnostic applications As yet recombinant DNA technology does not appear to have widespread diagnostic application in pathology. However, it does have a useful role to play in specific circumstances in at least three main areas: a it can provide precise diagnostic information about genetic diseases, allowing appropriate counselling, and indicating future directions for research on therapeutic intervention, e.g. gene therapy: b micro-organisms can be identified more sensitively and specifically,in fresh or fixed tissue samples, and their genomes can be analysed in fine detail, providing information relevant to the aetiology, epidemiology and pathogenesis of many diseases; c in tumour pathology the main application so far has been to resolve diagnostic problems associated with leukaemias and lymphomas, when other diagnostic procedures have been inconclusive. Specific chromosomal translocations, involving recognized genes, are particularly amenable to diagnosis by these means. Diagnostic applications to solid tumours are yet to be identified, although significant insights into tumorigenesis have been obtained, and these may ultimately lead to the development of useful markers for prognostic and therapeutic purposes. Keywords: DNA, recombinant technology, polymerase chain reaction, genetic disorders, tumour diagnosis

Introduction

Structure of the human genome

The contribution of morphological techniques to the diagnosis and management of human disease is well estabfished and likely to remain the mainstay of the pathologist's diagnostic armamentarium for some time to come. The advent of recombinant DNA technology, however, provides an opportunity to define disease processes in more precise molecular terms. Recent developments in this technology have also brought its application within the grasp of most diagnostic pathology laboratories. However, the extent to which these techniques should replace conventional morphological assessments requires careful evaluation and it is important to appreciate their limitations and pitfalls.

Genetic information in cells is distributed within nuclear DNA in 2 3 pairs of chromosomes containing 3.5 x lo9 nucleotide base pairs. More than half of nuclear DNA comprises highly repetitive sequences with no obvious function. The remainder is distributed as functional gene clusters (averaging 1 5 genes in each cluster) with perhaps 3000-5000 gene clusters in each genome. Only a small proportion of the nucleotide sequences within each gene represents actual protein coding sequences (exons)and these are interrupted by variable numbers of intervening sequences (introns)whose function is poorly understood (Figure 1). Flanking each gene are other sequences. Upstream, for example, there are enhancers and promoters that control the level of gene transcription and the site of initiation of RNA synthesis. Downstream of the coding regions there are signals to terminate transcription (poly A tail site: PA). Other sequences are involved in the subsequent processing (splicing) of the primary heterogeneous nuclear RNA

Address for correspondence: Professor C.C.Bird, Department of Pathology, University Medical School. Teviot Place, Edinburgh EH8 9AG. Scotland, UK.

303

304

MJ.Arends and C.C.Bird

Ex1 En

En

C

T

I

-

GU

5‘

; hnRNA

Ex2 In r

I

,

AG

PA

- 1

1

3’

i AAAA

G

UAA

AU G mRNA

Ex3 In1

J AAAA

G L--

I

Protein

N

Figure 1. Diagram of gene structure, transcription and translation. Upstream of the gene there are regulatory sequences including a promoter which determines the site of initiation of RNA synthesis, composed of a CCAAT box (C) and a TATA box (T). and a number of enhancers (En) that bind protein factors which together determine the level of transcription. Downstream of the coding sequences. known as exons (Ex). there is a transcription termination site (PA). which signals the addition of a poly A tail. These direct synthesis of a primary heterogeneous nuclear RNA (bnRNA) transcript, that has a guanosine cap (G) at the 5’ end and poly A tall at the 3‘ end. The htron sequences (In) are flanked by the nucleotides GU and AG at either end, and these direct the splicing out of introns during processing to messenger RNA (mRNA). The start codon for the protein coding sequence Ls always an AUG codon. whereas there are three possible stop codons (UAA. UAG. UGA). This encodes the polypeptide from the amino terminus (N) to the carboxy temlnus (C).

transcript to form the messenger RNA transcript (mRNA).This contains only exon sequences,which pass into the cytoplasm to be translated into the protein product (Figure 1).Expression of genes is temporally and spatially regulated, directing ontogeny and tissue function, and is subject to many internal and external influences’. Some non-coding DNA sequences are repeated in tandem in the genome (e.g. [AGAGGTGGGCAGGTGGIZ~) and are known as mini-satellites or variable number tandem repeats. These show extensive polymorphism (variations in the number of repeats of the same unit sequence) and a majority of individuals are heterozygous. The length of the repeated units varies from two nucleotides upwards (e.g. CACACA . . is a common pattern). These may be used as genetic markers, such as for DNA fingerprinting or loss of heterozygosity (implying allele loss) in tumours.

Fundamentals of recombinant DNA technology The methods for cloning specific DNA sequences to generate probes for identifying variations or modifica-

tions to the cell genome form the basis of recombinant DNA technology. It is not within the scope of this article to review these techniques in detail2.Recombinant DNA technology depends upon two key steps: a cleavage of DNA at sequence specfficsites by restriction endonucleases and b hgbridization of complementary DNA sequences. Restriction endonucleases often generate restriction fragments with sticky ends (overhanging single strands with complementary sequences) that permit recombination of DNA molecules from different sources. New DNA fragments can be cloned in appropriate vectors, commonly plasmids, that are introduced into bacterial hosts (Figure 2a). Restriction endonucleases recognize specific short DNA sequences which are often palindromic (same sequence in reverse on opposite DNA strand). Unrelated individuals may inherit restriction enzyme recognition sites at different places in the genome, so that cleavage produces DNA fragments of varying lengths, known as restriction fragment length polymorphisms (RpLps), useful as markers at particular regions in the genome. Cloned DNA fragments can be characterized by base pair sequencing. They may also be labelled with radionucleotides or non-radioactive chromogens for use as

Recombinant DNA technology

a

b Foreign DNA to be inserted

- - --

High-molecularweight DNA

- --

Joining

305

Cleave with restriction enzymes

Plasmid

-\

\

Antibiotic-resistance marker

rose gel electrophoresis

I

Ligation

-5sfer

JI

.1

Anti biotic

nitrocellulose to

nitrocellulose Autoradiograph

-

k 32P-labellec

:DNA or R N A

Figure 2. a Gene cloning. The new gene or foreign DNA insert is usually cleaved with the same restriction enzyme as the plasmid vector, thus generating overhanging cohesive ends which hybridize and are ligated. The recombined DNA molecule is introduced into a small proportion of bacterial host cells, which may be selected by growth in the presence of the appropriate antibiotic. b Southern blotting. High molecular weight nuclear DNA is cleaved with three restriction enzymes. The cleaved DNA fragments are separated according to size by agarose gel electrophoresis. The gel is laid on to a nitro-cellulose membrane, and a flow of buffer is set up to transfer the DNA fragments on to the membrane. This is exposed to a radioactively labelled probe, and the complementary DNA fragments that hybridize to the probe will give a signal following autoradiography.

probes. Labelled probes can be used in various hybridization reactions to detect and quantify complementary sequences in extracted DNA by Southern blotting (Figure 2b), or mRNA by Northern blotting, or in tissue sections by in situ hybridization. By varying the conditions under which these hybridization reactions take place (stringency), the specificity and sensitivity of the reactions may be modulated. Libraries of cloned DNA restriction fragments may be prepared from whole nuclear genomic DNA which contain both coding (exon) and non-coding (intron) sequences, and are referred to as genomic libraries. By using reverse transcriptase to

synthesise DNA copies of mRNA that contain only the coding (exon) sequences, complementary DNA (cDNA) libraries can also be prepared. Gene specilk probes may be derived from such libraries or from short singlestranded DNA molecules (oligonucleotides) that are chemically synthesized from known gene or protein sequences.

Polymerase chain reaction The polymerase chain reaction (PCR) can be used to amplify DNA fragments up to 10 kb in length. It requires

306 M.J.Arends and C.C. Bird

a Polymerase chain reaction Denaturation 94"C, 1 min

4

Primer 1

111 Primers anneal 40-6OoC, 2 min 1_1

4

Primer2

Extension 72"C, 3 min

b

+ *

PA

1

PCR

*

1

some knowledge of the base pair sequence of the DNA to be amplified to design primers with complementary sequences that specifically bind to opposite ends of the target sequence and direct its amplification. The method and its numerous applications have been described elsewhere3v4and will not be considered in detail. Briefly, it involves repeated cycles of DNA denaturation (at 94"C), followed by annealing of specific oligonucleotide primers (at 40-60°C) and finally chain extension (at 72OC) by thermostable (Taq) DNA polymerase (Figure 3a). Normally 20-40 cycles are performed in automated devices producing a 105-106 amplification of target DNA within a few hours. Amplified products of known length may be rapidly identified by size fractionation in electrophoretic gels without the need for radioactive probes and blotting techniques (Southern or Northern blots) which take several days. However, the exquisite sensitivity of PCR is also its weakness. Potential contamination of test DNA from extraneous sources or previous reactions must be rigorously excluded and appropriate negative controls included. Template DNA for amplification may be derived from DNA libraries, extracted fresh from tissues or biological samples (e.g. blood, semen, hair roots, smears, fluids),or from paraffin-processed sections. Only minute quantities of template DNA are required to initiate the amplification process and even single cells from in vitro fertilized embryos can provide enough template DNA. Polymerase chain reaction can be used to detect infectious organ-

RE

C

+

+*

c

1

PCR

+*

1

RE

0

0

Figure 3. a Polymerase chain reaction (PCR).Double-stranded DNA is denatured at high temperature for 1 min. Single-stranded oligonucleotideDNA primers anneal to specific sequences within the template DNA at intermediate temperatures. and Taq DNA polymerase extends the DNA by adding new bases in a complementary sequence to the template, thus generating two double-stranded DNA fragments from the original one. b Restriction endonuclease polymorphism in PCR products (REPPP). A DNA sequence variation (*) is present in the 3' untranslated region of the gene. A short DNA fragment surrounding this sequence variation is amplified using speciftc primers (arrows). Following ampllfication by PCR, the products may be cleaved using the appropriate restriction endonuclease (RE) that recognizes the site created by the presence of the sequence variation (*), generating two fragments. The alternative allele lacks the sequence variation, and thus is not cleaved by the restriction enzyme at that position. c Restriction endonuclease polymorphism in PCR products using an engineered primer. The sequence variation (*) is identified within an intron, but it does not create a new restriction endonuclease site on its own. The upstream primer (arrow) is engineered with a base change ( + ) close to its 3' end. Together the primer base change and the sequence variation ( *) create a new restriction endonuclease site at that position.

+

Recombinant DNA technology

isms, if part of their genomic sequence is known. Modifications to the technique have been developed to permit DNA amplification even where sequence data is incomplete. Polymerase chain reaction can also be used to amplify genes for cloning, and to generate mutant and chimaeric DNA fragments. It can be used to analyse genes for point mutations by amplifying DNA fragments that contain possible mutations which may be subsequently detected by a number of techniques: differential oligoprobe hybridization, temperature gradient gel electrophoresis, single-strand conformational polymorphisms, or sequencing. Polymerase chain reaction can also provide a more rapid method for sequencing DNA. The strand to be sequenced is amplified by asymmetric PCR, in which one primer is present at higher concentration (100:1)than the other, thus generating mostly single-stranded DNA fragments that can be sequenced by chemical cleavage or the dideoxynucleotide method of Sanger4. This approach can be applied directly to amplied DNA fragments from tissues or cells without prior cloning, or to DNA sequences already cloned in plasmid' or phage vectors. Two examples of the advantages of rapidity and accuracy conferred by PCR are HLA typing for transplantation and DNA fingerprintingfor forensic purposes. Allele loss in tumours, also known as loss of heterozygosity, can be studied by PCR. A recently described strategy is based on the rapid isolation and detection of PCR based polymorphisms, which we call restriction endonuclease polymorphisms within PCR products (REPPPs)~.These are also known as 'PCR polymorphisms'-which is a misnomer as it is the restriction endonuclease site that is polymorphic, not the PCR. The DNA sequences of the specific gene of interest are analysed within computer databases for possible nucleotide sequence variations between different clones of the same gene. In particular, nucleotide sequence variations that create or destroy restriction enzyme sites are sought. Primers are designed to flank such variations, and the amplified products are analysed by digestion with the relevant restriction enzyme. The aim is to characterize restriction enzyme cutting sites which are polymorphic and are present in only a proportion of the population. This allows detection of presence or absence of the restriction site, giving two possible alleles, which are within or adjacent to the gene of interest (Figure 3b). Such restriction enzyme polymorphisms may be found at many sites including the convenient non-coding sequences within or immediately adjacent to genes-the 5' or 3' untranslated regions. Some sequence variations may not create or destroy a

307

restriction enzyme site. In this circumstance, primers with modified sequences may be engineered that bind immediately adjacent to the site of sequence variation such that the base change introduced into the primer sequence, together with the naturally occurring sequence polymorphism, create a new restriction enzyme site (Figure 3c). This strategy has been successfully used to define REPPPs for the insulin gene, a candidate Wilms' tumour suppressor gene, and the Ki-ras oncogene5. This approach has several advantages over the detection of RFLPs by Southern blotting as it is faster, easier and quicker to perform. Restriction endonuclease polymorphisms within PCR products may be located within or immediately adjacent to the gene of interest. In contrast, RFLPs may be distantly linked to the gene which may lead to errors associated with recombination events between gene and marker. An REPPP that is located in an untranslated region, which is transcribed to mRNA, may be used to study the relative expression from each parental allele in heterozygousindividuals,using reverse transcription and PCR. Eventually, REPPPs are likely to supersede RFLPs for many types of genetic analysis.

Diagnostic applications The principal diagnostic applications of recombinant DNA technology at present fall into three main categories: diagnosis of genetic disorders,identification of microorganisms and the diagnosis and analysis of tumours. Applications are likely to become more diverse as the human genome project generates greater insights into medical genetics6. DIAGNOSIS OF GENETIC DISORDERS

Broadly speaking, diseases with a genetic basis fall into two main groups. In one, the disease is associated with a clearly identifiable single gene defect (termed unifactorial inheritance), which may be autosomal or sex linked, (usually X-linked). In dominant disorders affected individuals have one defective allele (heterozygousstate). In such individuals disease manifestations may vary in frequency and extent (termed variable penetrance and expression), or may not occur until later in adult lie (e.g. Huntington's disease). In the recessive pattern of inheritance, both alleles must be defective (homozygous state) for the disease to be evident. Carriers of recessive autosomal (e.g. cystic fibrosis) and X-linked (e.g. haemophilia) diseases have only one defective allele and are usually unaffected themselves. The other main category of genetic disease shows familial patterns of inheritance, but the precise genetic

308

M.J.Arendsand C.C.Bird

basis forthese disorders is as yet poorly defined. It is likely that in each disease several gene clusters are involved (termed multifactorial or polygenic inheritance) and that complex interactions occur between genetic and environmental factors. These account for some of the major diseases of man (e.g. diabetes mellitus, atherosclerosis, hypertension, and some psychiatric disorders) and we are only just beginning to unravel their genetic basis'. At present, recombinant DNA technology has most to offer for the diagnosis of unifactorial (single gene) diseases. Over 4000 such disorders are known to exist although only in a handful has the relevant gene mutation been identiied. The potential difficulties in diagnosis must not be underestimated. Several hundred variants of the haemoglobinopathies have been identified, most due to different point mutations in the globin gene, in both protein coding and regulatory sequences. In Duchenne and Becker muscular dystrophies, deletions or duplications may occw in any of the 5 1exons in the dystrophin gene, and in cystic fibrosis over 50 different mutations have been reported. Recombinant DNA technology may be applied to the investigation of individuals suspected of harbouring these diseases as a prenatal measure (employing chorionic villous biopsy) or in adult onset disease by predictive testing (using blood samples). The latter may also be used for the detection of the carrier status. Where genes relevant to the disease have been identified and probes are available,

a

the diagnosis can be made either by direct demonstration of an aberrant gene or indirectly by linkage analysis of WLPs in overlapping or adjacent DNA sequences. In circumstances where the relevant gene has still to be identified, diagnosis can only be made by means of linkage analysis. Until recently, most laboratories have employed Southern blotting for such analyses but increasingly PCR is being applied to a large range of genetic diseases3. This technique has the advantage of increased sensitivity and specificity, and provides a more rapid diagnosis. Furthermore, by using nested and multiple PCR reactions some of the difllculties presented by the multiplicity of mutations in certain diseases can be overcome4. In this country, prenatal diagnosis and predictive testing is most commonly required for the diagnosis of cystic fibrosis, Duchenne muscular dystrophy, Huntington's disease, myotonic dystrophy and adult polycystic kidney disease. Carrier testing on the other hand is most frequently performed for Duchenne and Becker muscular dystrophy, haemophilia and a'-antitrypsin deficiency,and increasingly for cystic fibrosis. It has been calculated there are approximately 2 000 000 carriers of the cystic fibrosis gene in the UK which raises the question whether population screening for this disease should be introduced. There is also the prospect of screening individuals for diseases with multifactorial (polygenic) inheritance patterns, which, numerically, represent by far the greatest health problem. To date

b M

-

a

b

c

d

e

f

g

Plgure 4. a In sltu hybridization showing the presence of HPV 16 DNA' within invasive islands of a squamous carcinoma of the anus. b Amplified HPV 16 DNA from seven cases of CIN 3 (a-g)9. Five are strongly positive and one (d) is weakly positive for HPV 16. One case (f) is negative. A template-freenegative control track (-), and a marker track (M)are included.

Recombinant DNA technology

most of these studies have been concerned with rather ill-defined linkages between the major histocompatibility gene complex and relatively uncommon diseases (e.g. ankylosing spondylitis, type I diabetes mellitus). More recent investigations, however, have begun to tackle the much more important disorders like atherosclerosisand their linkage to the LDGreceptor and Apo-B genes. This approach is gaining momentum and will undoubtedly constitute a major growth area in the future. IDENTIFICATION OF MICROORGANISMS

C MPRl 8.9 Kb

T

B

M

Table 1.Identification of microorganisms by polymerase chain reaction Microorganism

Tissues

Viruses BK and JC virus Cytomegalovirus

Epstein-Barr virus

Many different types of infective organism can be identified by recombinant DNA technology. Localization to the cell or site of infection, may be achieved by in situ hybridization using labelled probes hybridizing specifically to the genome of the organism to be detected (e.g. human papill~mavirus)~ (Figure 4a). Amplification of DNA fragments specifically from genomes of infectious organisms using PCR is a highly sensitive and specific

S1

s2

+

7.8 Kb

Human immunodeficiency virus Human B-lymphotrophic virus Human T-cell lymphotrophic virus Hepatitis B virus

Hepatitis C virus Human papilloma virus

Herpes simplex virus Measles Parvovirus Rhinovirus Bacteria Borrelia Chlamydia

4-2 Kb

*

Urine, par&-embedded tissue Urine, peripheral blood, paraffin-embedded tissues, autopsy tissues Tissue biopsy, autopsy tissues, throat swabs Peripheral blood, endoscopy swabs, various tissue samples Peripheral blood, tumour tissue Peripheral blood, tumour tissue Peripheral blood, cord blood, colostrum. liver, paratfinembedded tissues Peripheral blood, liver Cervical lavages and scrapes, paraffin-embedded tissues, various tissues Paraffin-embedded skin biopsies, CSF ParaBBn-embedded tissues Sera, placenta, fetal tissue, paraffin-embedded tissues Swabs and aspirates Ticks, various tissues Urethral secretions, conjunctival samples Sputum, lymph node biopsy, gastric and abscess aspirates, urine, CSF, blood

Mycobacteria

3.7 Kb -b

Neisseria

CSP

Others Helicobacter pylori Malaria Mycoplasma

Gastric biopsies, faeces Blood Throat swabs, bronchoalveolar lavage Sputum, bronchealveolar lavage RNA extracts Peripheral blood Insect vectors, clinical isolates

Pneumocystis Figure 5. Southern blot showing TCR beta gene rearrangement in a case of high grade T-cell non-Hodgkin’slymphoma probed with a TCR beta DNA probeI6. Control (C) DNA extracted from tonsil shows the presence of germline bands of sizes 8.9, 7.8, and 4.2 kb. Tumour (T) DNA shows an extra marker band of 3.7 kb in size due to clonal expansion of Ttells with a unique TCR gene rearrangement. The marker band is not present in samples of peripheral blood (B). marrow (M). or two peripheral blood stem cell harvests (S1 and S2).

309

Rickettsia Toxoplasma Trypanosoma ~

~ _ _ _ _ _

Adapted from Wright & Wynford-Thomas (1990)3.

3 10 M.J.Arends and CLBird

b

a

M

PCR analysis of bc/-Z/lgHJ sequences

T

-

B

s1

s2

3’ primer

Chromosome 14

I

lgHJ

A Breaks

v

Chromosome 18

I

5’ primer

I

Denature DNA into single strands

After repeated cycles, region delineated by primers amplified to unique detectable product of -200 bp

Figure 6. a Translocation of chromosomes 14 and 18. which is common in follicular lymphomas. Break points occur around the IgHJ DNA sequences and the bcl-2 oncogene. Polymerase chain reaction (PCR) primers may be sited at either end of the chromosomal translocation point and generate an amplied product of unique size. b Detection of t( 14:18) translocation by PCR * in a case of B-cell follicular non-HodgkIn’s lymphoma. A unique tumour marker band is seen in the involved lymph node (T). but not in the negative control (-) composed of tonsil DNA.The tumour marker band is present in samples of peripheral blood (B). and two stem cell harvests (S1 and S2) following treatment with chemotherapy. A marker track (M)is included.

method of detection8 (Figure 4b). This is possible with organisms with DNA or RNA genomes, as reverse transcription can be used prior to PCR to convert RNA to cDNA. Polymerase chain reaction is particularly useful because of its increased specificity and sensitivity, and may be applied to a wide range of microorganisms (Table 1).It may also improve the rapidity of identification, and can be used to demonstrate organisms in fixed and para&-embedded tissue as well as in fresh frozen tissueg. Recently it has been demonstrated that the two techniques of in situ hybridization and PCR may be combined to show very low genome copy numbers of HPV in formalin ftxed tissues, both in carcinoma in situ and also in formalin-fixed SiHa cells-a cervical carcinoma derived cell line that contains only a single copy of HPV 16 per celllo. However, there is a potential problem in interpreting the clinical significanceof positive results, when highly sensitive assays capable of detecting only trace amounts of template DNA are employed. The use of recombinant DNA technology to identify

microorganisms is likely to be particularly useful when studying organisms that are difEcult to grow in culture, or are extremely slow in growing (e.g. Mgcobacterium tuberculosis or toxoplasma). This is especially the case where the numbers of organisms are too low for detection by conventional methods (e.g. HIV or Mucobacterium tuberculosis), or where there is no reliable culture system available (e.g. hepatitis B and C viruses, HPV or parnovirus), or if the specimen is fixed and cannot be cultured. Recombinant DNA techniques also have a role in studies of disease pathogenesis (e.g. localization of Epstein-Barr v i r u s DNA in Reed-Stemberg cells in Hodgkin’s disease)’l. Molecular analysis of the genomes of microorganisms allows determination of strain variation including minor changes to gene sequences, which may be functionally important in determining the pathogenicity of the microorganism (e.g. HPV in cervical n e ~ p l a s i a or ) ~ in studying its epidemiology (e.g. mycobacteria in tuberculosis)l2.

Anneal primers I

1

.

0 -

+ polymerase and dNTPs results in chain extension

-

1

Recombinant DNA technology

3 11

Figure 7. M

a

b

c

d

3

Figure 8. a

b

c

d

e

f

M

2

1

Figure 7. Detection of allele loss in colorectal carcinoma using a REPPP within the 3‘ untranslated region of the APC gene. Polymerase chain reaction amplffies an 8 5 0 bp fragment that is cleaved by SspI at either two sites (to produce 135, 135. and 580 bp fragments).or one site (producing 270,and 580 bp fragments),defining a simple two allele polym~rphism’~. A non-informative individual (tracks c and d). homozygous for one allele (arrow 2) is compared with an informative individual (tracks a and b), heterozygous for both alleles (arrows 1 and 2) (a & c =normal colonic mucosa: b & d =colorectal cancer). The upper arrow (3)indicates a constant fragment (580 bp) present in both alleles, which also acts as an internal control for enzymic activity. There is a pattern of allele loss within the tumour resected from the heterozygous individual (track b). The cancer has lost 1 allele (arrow 1) of the APC-associated REPPP. although a faint signal can be discerned due to contamination by stromal fibroblasts, blood vessels and lymphocytes. A marker track (M) is included. Figure 8. Polymerase chain reaction (PCR) detection of a codon 1 2 mutation in the human Ki-ras oncogene in colorectal carcinoma. The PCR mpliAed product from normal Ki-ras exon 1 sequence spans 1 5 7 bp (large arrow) when undigested (tracks a, c and e). and 1 1 4 bp (small arrow) after BstNI digestion (b), due to removal of two small fragments by cleavage at two sites: one at the normal codon 1 2 (engineered by a base change in the adjacent upstream primer), and the other within the downstream primer (also engineered to create an internal control for enzymic activity)20.One carcinoma (d) shows the same proto-oncogene pattern after BstNI digestion, but the other carcinoma (f) shows the presence of a codon 12 mutation that has destroyed the upstream BstNI site giving an additional 143 bp fragment (arrowhead). A marker track (M)is included.

TUMOUR DIAGNOSIS AND ANALYSIS

To date, the use of recombinant DNA technology for tumour diagnosis has been restricted mainly to lymphomas and leukaemias. Lymphomas are characterized by clonal cell expansions exhibiting unique patterns of rearrangement of immunoglobulin genes (in B-cell tumours13),or T-cell receptor genes (in T-cell tumours’*) (Figure 5). Whilst this permits differentiation between most B- and T-cell lymphomas, the application of these techniques is required only in 5 1 0 %of cases where the diagnosis cannot be resolved by morphological and immunocytochemicalmethod@. These lineage markers may also be employed for the identification of bone marrow involvement, the presence of residual or recurrent disease following treatment, or lymphoma cell contamination of stem cells harvested for autologous transplantation16. Some types of leukaemia and lymphoma also show specific chromosomal translocations, which can be

detected by Southern blotting or PCR, representing another form of tumour marker. For example, in some follicular lymphomas there is a t( 14;18) translocation involving the immunoglobulin heavy chain (IgH) gene and the bcl-2 oncogene (Figure 6). In Burkitt’s lymphoma a segment of chromosome 8 containing the c-myc oncogene may be translocated to one of three chromosomal regions: the IgH locus [t(8:14)],the kappa light chain locus [t(8;2)]. or the lambda light chain locus [t(8:22)]. In some small cell non-Hodgkin’s lymphomas and in chronic lymphocytic leukaemia there is a t( 11;14)translocation involving the bcl-1 oncogene and IgH locus. In these three examples, specific oncogenes are translocated to new loci, where expression of the juxtaposed immunoglobulin genes is presumed to activate the translocated oncogene. Another example of this is seen in many cases of chronic myeloid leukaemia and some cases of acute lymphoblastic leukaemia. where there is a t(9;22) translocation involving the c-abl oncogene and bcr sequences. The abnormal chromo-

3 12

M.J.Arends and C.C.Bird

chromosomes 5q (APC and/or MCC), 12p (Ki-ras), 18q (DCC), 17p (~531,and other chromosomal loci in varying combinations’ 7. Theloss of one copy of a tumour suppressor gene, such as APC. may be demonstrated either by Southern blotting18, or by analysis of RJPPPS’~(Figure 7). Activation of the oncogene Ki-ras on chromosome 12p, often occurs as a result of specffic mutations at key sites (mainly codons 12/13 or 59/61), and these mutations can be detected by PCRzo(Figure 8). Mutations affecting the p53 gene are often grouped together in ‘hot spots’ in particular exons (usually 5-9) of the gene, and these may be amplified by asymmetric PCR and sequenced, allowing precise definition of the mutations present2’ (Fimre 9).

some 22 that is produced is known as the Philadelphia chromosome. All of these translocations are amenable to detection by recombinant DNA techniques. Lymphomas and leukaemias are relatively uncommon and as yet there are no good clonal or lineage specific markers that are useful for the diagnosis of common solid cancers. Some rarer forms (e.g. retinoblastoma and Wilms’ tumour) do show inheritance of specific single gene defects. By contrast, most common solid tumours tend to demonstrate multiple genetic events, that include both activation of oncogenes and deletion or mutation of tumour suppressor genes. Unfortunately, these genetic changes are not always consistent nor specific. For example, in colorectal tumorigenesis mutations or deletions may involve genes on a

Amplified template

I

1 DNA(s/s) DNA polymerase I 4 dNTPs +

ddNTP

+

ddATP

4

Acrylamide

gel

+

ddlTP

+

ddCPT

*

ddGTP

DNA

sequence of template I

t

T T A G A C C C G A T A A G C C C G C A

P53 EXON 8

b

G

A

T

C

G-A

Figure 9.a Sanger DNA sequencing method. The slngle-stranded DNA template to be sequenced may be generated by a s y ~ ~ ~ e polymerase trIc chain reaction (PCR)in which one primer is present at much greater concentration than the other primer. This template is split into four DNA polymerase reactions each containing one of the four dideoxy nucleotides (ddNTPs). and a labelled primer ( 0 )to initiate DNA synthesis. All reactions require DNA polymerase 1 and the 4 usual dNTPs. Nucleotides are added as the chain is extended until a d d ” is inserted, which lacks an appropriate hydroxyl group, 80 that it cannot form a phosphodiester bond with the next incoming dNTP. Thus, growth of that particular DNA chain stops. With the correct ratio of d d ” to dNTP. a series of labelled strands will result, the lengths of which are dependent on the location of the particular d d ” base relative to the end of the DNA (defined by labelled primer). The resultant labelled fragments are fractionated by size on an acrylamide gel, and the autoradiographic pattern of fragments gives the DNA sequence. b Detection of a p53 mutation in colorectal carcinoma. Sequencing gel of exon 8 following asymmetric PCR and Sanger sequencing. This shows a mutation within the sequence, in which DNA fragments indicate two populations of template DNA,one containing a G and one containing an A nucleotide at the same position. This reflects the presence of two copies of the p53 gene within tumour cells. one of which is mutated (G to A transition), and one of which is not (G).

Recombinant DNA technology

Another feature of colorectal tumorigenesis is that certain genetic events preferentially appear to involve either early or late stages of the adenoma-carcinoma sequence. However, in many instances it is probably the accumulation of genetic events rather than their precise sequence that determines tumour progression1’. Furthermore, some genetic events occur with relatively high frequency in other solid tumours (e.g. p53 mutations in breast and lung cancers). Finally, recombinant DNA techniques can be used to screen high risk cancer groups for the presence of inherited susceptibilitygenes (e.g. APC in familial adenomatous polyposis coli, or p53 in Li-Fraumeni syndrome). Screening may be performed by linkage analysisz2,or by REPPPs, or by detection of specific mutations within tumour suppressor genes. However, the variability and lack of specificity of the genetic changes occurring in sporadic common solid tumours currently restricts their use as definitive histogenetic markers. For the same reasons they cannot yet be used to distinguish between benign and malignant tumours, where such distinctions are critical. On the other hand, the differing patterns of genetic events may explain the varying biological behaviour of tumours, and may ultimately prove more useful as prognostic or therapeutic markers.

Acknowledgements The authors wish to thank the following individuals for contributions of figures: Dr A.H.Wyllie, Dr J.Bubb, Dr P.J.Carder, Mr R.G.Morris, Miss Y.K.Donaldson, Dr J.I.O.Craig and Mr K.Langlands.

References 1. Watson JD. Gilman M, Witkowski J, Zoller M. Recornbinant DNA. 2nd edn. New York Scientific American, 1992. 2. Sambrook J. Fritsch EF, ManiatisT. MolecularCloning: A Laboratoru Manual. 2nd Fdn. Cold Spring Harbour Laboratory Press, 1989. 3. Wright PA, Wynford-Thomas D. The polymerase chain reaction: miracle or mirage? A critical review of its uses and limitations in diagnosis and research. 1. Pathol. 1990: 162; 99-1 17. 4. Innis MA, Gelfand DH. Sninsky JJ. White TJ.PCR P r o t o c o h Guide to Methods and Applications. San Diego: Academic Press, 1990. 5. Hoban PR, Santibanez-Koref MF. Heighway J. A strategy for the rapid isolation of novel PCR based DNA polymorphisms.]. Med. Genet. 1992: In press.

3 13

6. Harper S. The human genome project and medical genetics. 1. Med. Genet. 1992: 29; 1-2. 7. Morris RG, Arends MJ, Bishop PE, Sizer K. Duvall E, Bird CC. Sensitivityof digoxigenin and biotin labelled probes for detection of human papillomavirus by in situ hybridisation. 1. Clin. Pathol. 1990; 43; 800-805. 8. Arends MJ.Identitication of HE’V: in situ hybridisation or polymerase chain reaction?]. Pathol. 1991: 164; 191-193. 9. Arends MJ, Donaldson YK, Duvall E. Wyllie AH. Bird CC. H P V in full thicknesscervicalbiopsies: high prevalence in CIN 2 and CIN 3 detected by a sensitive PCR method. I. Pathol. 1991: 165; 301309. 10. Nuovo GI, MacConneU P. Forde A. Delvenne P. Detection of human papillomavirus DNA in formalin-fured tissues by in situ hybridisation after amplification by polymerase chain reaction. Am. 1, Patho]. 1991: 139; 847-854. 11. Coates PJ. Slavin G, D’Ardenne A]. Persistence of Epstein-Bam virus in Reed-Stemberg cells throughout the course of Hodgkin’s disease. 1. Pathol. 1991; 164; 291-297. 12. Otal I. Martin C, Vincent-Levy-Frebault V, Thierry D. Gicquel B. Restrictionfragment length polymorphism analysis using IS61 10 as an epidemiological marker in tuberculosis. 1. Clin. Microbiol. 1991; 29; 1252-1254. 13. Korsmeyer SJ, Hieter PA, Sharrow SO. Goldman CK. Leder P. Waldman TA. Normal human B cells display ordered light chain gene rearrangements and deletions. 1. Exp. Med. 1982; 156; 97 5-98 5. 14. Flug F. Pelicci P-G, Bonetti F, Knowles DM. Dalla-Favera R. T-cell receptor gene rearrangements as markers of lineage and clonality in T-cell neoplasms. Proc. Natl. Acad. Sci. USA 1985: 82; 34603464. 15. Gledhlll S, Krajewski AS, Dewar AE. Onions DE. Jarrett RF. Analysis of T cell receptor and immunoglobulin gene rearrangements in the diagnosis of Hodgkin’s and non-Hodgkin’s lymphoma. 1. Pathol. 1990: 161; 245-254. 16. Langlands K, Craig JIO,Parker AC. Anthony RS. Molecular determination of minimal residual disease in peripheral blood stem cell harvests. Bone Marrow Transplantation 1990: 5, Suppl 1; 64-65. 17. Fearon ER, Vogelstein B. A genetic model for colorectal tumorigenesis. Cell 1990; 61; 759-767. 18. Ashton-Rickardt PJt Wyllie AH, Bird CC et al. MCC. a candidate f a d i a l polyposis gene in 5q.21, shows frequent allele loss in colorectal and lung cancer. Oncogene 1991: 6; 1881-1886. 19. Heighway J, Hoban PR. Wyllie AH. SspI polymorphism in sequenceencoding 3‘ untranslated region of the APC gene. Nucleic Acids Research 1991: 1 9 (24); 6966. 20. Jiang W, Kahn SM, Guillem JG. Lu S-H. Weinstein IB. Rapid detection of ras oncogenes in human tumors: applications to colon, esophageal and gastric cancer. Oncogene 1989: 4; 923928. 21. Baker SJ,PreisingerAC. JessupJM et al. p53 gene mutations occur in combination with 17p allelic deletions as late events in colorectal tumorigenesis. Cancer Res. 1990; 5Q: 7717-7722. 22. Dunlop MG, Wyllie AH, Steel CM. Piris J. Evans HJ. Linked DNA markers for presymptomatic diagnosis of familial adenomatous polyposis. Lancet 1991; I; 337; 313-316.