Surgical Endocrinology

By John Lynn and Stephen R. Bloom

Surgical Endocrinology presents the fundamentals and practical application of the subject matter. It discusses the molecular background and physiology of the endocrine systems. It addresses the legal aspects of endocrine surgery such as duty of care, and breach of such duty. Some of the topics covered in the book are the principles of endocrine surgery; molecular structures of organisms; physiology of endocrine system; multiple endocrine neoplasia and molecular genetics; non-surgical management of pituitary tumors; assay formulations and automation technology; syndrome of ectopic hormone production; and anesthesia for pituitary surgery. The classifications of endocrine problems for pregnant women are fully covered. The definition and symptoms of thyrotoxicosis are discussed in detail. The text describes in depth the different histological imaging methods. The anatomy and surgery of anterior pituitary are completely presented. A chapter is devoted to the classification of pituitary adenomas and other tumours. The book can provide useful information to surgeons, doctors, students, and researchers.

• Language: English
• Publisher: Elsevier Science
• Release date: Oct 22, 2013
• ISBN: 9781483193823

Book preview

syndrome

Part 1

Introduction

Outline

Chapter 1: The Principles of Endocrine Surgery

Chapter 2: The Molecular Basis of Endocrinology

Chapter 3: The Physiology of the Endocrine System

Chapter 4: Multiple Endocrine Neoplasia and Molecular Genetics

Chapter 5: Modern Techniques of Hormone Measurement and Their Interpretation

Chapter 6: Paraendocrine Syndromes

Chapter 7: Anaesthetic Considerations in Endocrine Surgery

Chapter 8: Endocrine Problems in Pregnancy

Chapter 9: Modern Histological Imaging Methods

Chapter 10: Legal Aspects of Endocrine Surgery

1

The Principles of Endocrine Surgery

John Lynn

Publisher Summary

This chapter discusses the principles of endocrine surgery. Endocrine surgery is performed in a wide variety of centers throughout the world with excellent results, low morbidity and in some cases no mortality. Surgeons must be convinced of the diagnosis before moving on to any form of procedure. It cannot be emphasized too strongly that the diagnosis must be fully established on endocrinological criteria. A principle of endocrine surgery is to render the patient safe. The classic example is the reduced mortality from pheochromocytomas associated with the use of blockade for the preoperative preparation of the patient. Spironolactone is equally valuable in the management of patients with severe Conns syndrome as are antithyroid drugs in the preparation of patients for thyrotoxicosis and metyrapone in Cushings syndrome. The principles are very similar in all cases, but it may be difficult to decide whether the patient is really safe. A reduced pulse rate and normal thyroid function will give clear evidence in thyrotoxicosis, and lowered blood pressure and normal serum potassium will likewise confirm that the patient is safe in Conns syndrome.

Endocrine surgery is now performed in a wide variety of centres throughout the world with excellent results, low morbidity and in some cases no mortality. These results have been achieved by the adherence to basic principles (five dictates) which are usually applicable to all forms of surgical endocrine problems:

1. Confirm the endocrine diagnosis.

2. Localize the tumour or tumours.

3. Render the patient safe.

4. Does the patient need an operation?

5. What type of operation does the patient need?

The dictates laid out in this list are usually performed in the order given, but on occasions the invasive nature of the localization technique may mean that pharmacological safety measures must be carried out prior to localization. An example of this is the use of α-blockade in a phaeochromocytoma prior to angiography or venous sampling [1]. Pharmacological safety measures may affect the accuracy of localization techniques, therefore the balance between localization techniques and pharmacological safety measures needs careful consideration.

Confirming the diagnosis

Surgeons must be convinced of the diagnosis before moving on to any form of procedure. It cannot be emphasized too strongly that the diagnosis must be fully established on endocrinological criteria; in primary hyperparathyroidism, for example, based on a repeatedly raised serum calcium in the face of an inappropriate parathyroid hormone level or, in the case of a phaeochromocytoma, on repeatedly raised catecholamine levels in a patient with hypertension. Many referred patients seen have not had a proper endocrinological work-up and have been explored unnecessarily due to a wrong diagnosis. This is quite common with hypertensive patients who have been scanned by computed tomography (CT) or magnetic resonance imaging (MRI) and are found to have adrenal tumours. Such tumours are common and are now called ‘incidentalomas’ [2,3]. However, the diagnosis of phaeochromocytoma must be supported by a finding of raised adrenaline and noradrenaline levels, and inappropriate exploration and removal of an incidentaloma in a hypertensive patient will not bring about an improvement in the patient’s condition.

Once the endocrine diagnosis and existence of a tumour has been established, it is essential to clarify whether or not the abnormality is part of a multiple endocrine neoplasia (MEN) syndrome. If it is, then the priority must be which organ to treat first. In MEN type 1 syndrome (MEN 1) the parathyroids should be dealt with before the pancreatic, duodenal or pituitary neoplasm. Conversely, with MEN type 2 (MEN 2) syndrome it is essential to deal with the phaeochromocytomas, whether single or multiple, before medullary carcinoma of the thyroid or parathyroid disease. It is also important to decide whether the source of the problem is due to an ectopic paraendocrine phenomenon. This is particularly true in the ectopic ACTH syndrome which is often associated with bronchial carcinomas, or more rarely a variety of other tumours, e.g. phaeochromocytoma.

The localization of the tumour(s)

There are three main problems: (1) occurrence of incidentalomas (2) whether localization should be omitted and (3) the need to confirm the existence of multiple-gland disease. We are now fortunate in having a plethora of tests available for the localization of endocrine tumours [4], which will be discussed in later chapters. The surgeon must be aware of the pitfalls inherent in these techniques. Widespread use of ultrasound of various parts of the body, CT and MRI have shown a high incidence of benign non-functioning adenomas (incidentalomas) of various organs. This can vary from 10% in the pituitary, 15% in the adrenal, to as high as 30% in the thyroid. These relatively harmless tumours have no endocrine significance and are red herrings! The high incidence of incidentalomas stresses that localization of tumours is not in any way diagnostic.

Localization of endocrine tumours is not always undertaken, and it is not the policy of the author’s group to perform any form of localization prior to the first exploration of the neck for primary hyperparathyroidism. Such scanning techniques are felt to be unnecessary because expert surgery results in a high success rate. However, the situation is completely different for those patients who have had previous parathyroid surgery; then all forms of localization, ranging from thallium—technetium substraction scans, through CT scans, to MRI and angiography, have a major role to play.

Multiple-gland disease must be identified. A classic example of this is the superb result than can be obtained using ¹³¹I-meta-iodobenzylguanidine (¹³¹I-MIBG) which will accurately show whether or not there are single or multiple phaeochromocytomas [5]. However, other scanning techniques, e.g. thallium—technetium scans, are poor at differentiating multiple-gland parathyroid disease.

Safety

The third principle of endocrine surgery is to render the patient safe. The classic example is the reduced mortality from phaeochromocytomas associated with the use of α-blockade for preoperative preparation of the patient [6]. Spironolactone (Aldactone) is equally valuable in the management of patients with severe Conn’s syndrome [7], as are antithyroid drugs in the preparation of patients for thyrotoxicosis [8], and metyrapone in Cushing’s syndrome [9]. The principles are very similar in all cases, but it may be difficult to decide whether the patient is really safe. A reduced pulse rate and normal thyroid function will give clear evidence in thyrotoxicosis, and lowered blood pressure and normal serum potassium will likewise confirm that the patient is safe in Conn’s syndrome [7]. With a phaeochromocytoma, it can be quite difficult to determine the end-point of α-blockade and patients are often inadequately blocked when they come to surgery.

The need to operate

The fourth principle of endocrine surgery is to decide whether the patient really needs an operation. This question raises its head in all forms of endocrine surgery, but the classic example is the management of the asymptomatic hypercalcaemic patient with primary hyperparathyroidism (a group of patients that may represent approximately 50% of a referral practice). This is a contentious issue [10]; asymptomatic patients have been seen to obtain the same metabolic benefits from parathyroidectomy as symptomatic patients, and parathyroidectomy is known in such instances to be cost-effective. This approach is adopted by the author’s group; however, many workers would not agree and monitor asymptomatic patients carefully, waiting until their calcium levels become significantly high before considering surgery [10] (see Chapter 28). A similar problem occurs in Conn’s syndrome: this may be quite adequately controlled by spironolactone; however, hypertension is only cured in approximately 56–83% of patients following unilateral adrenalectomy for a true cortical adenoma. These findings have caused some physicians to suggest that long-term treatment with spironolactone may be a satisfactory alternative to surgical intervention − a truly controversial area where the author’s group favours a conservative approach. Similar problems occur in hyperprolactinaemia, due to the efficacy of bromocriptine where the role of pituitary surgery is controversial.

A different aspect of the indications for surgery may be considered in the management of familial medullary carcinoma of the thyroid. There is no doubt that total thyroidectomy is the treatment of choice for patients with medullary carcinoma, but managing the relatives is often difficult. Performing a thyroidectomy at the preclinical stage (when no symptoms are evident) can prevent a later fatal malignancy. High levels of calcitonin, whether stimulated or unstimulated, are indeed an indication for surgery but it is a problem to know at what age these patients should be explored. It is best to be guided by the aggressiveness of the tumour within the family group. It is hoped that, in the future, chromosomal studies will indicate which patients are truly at risk [11].

Surgical tactics

The primary aim is to return the patient to a normal endocrine state. The secondary aim is to deal with the complications produced by excess endocrine abnormality in distant organs. The classic example is the management of primary hyperparathyroidism due to single-gland disease. Here, the removal of the gland will result in a high incidence of normocalcaemia. This may not be easily obtained in patients with parathyroid hyperplasia where it is difficult to judge the extent of resection needed. The management of renal stones, duodenal ulceration and pancreatitis presents a secondary aim in a parathyroid patient. In Graves’ disease, where there is a high incidence of hypothyroidism following subtotal thyroidectomy, the easy replacement with thyroxine obviates this problem, and patients are often willing to accept long-term thyroid replacement due to the advantage afforded of a low incidence of recurrent thyrotoxicosis. On occasions, end-organ removal may be necessary and a good example of this is the inoperable bronchial carcinoma where removal of both adrenals will result in alleviation of a Cushing’s syndrome [12]. This contrasts with the removal of a bronchial carcinoid in the ectopic ACTH syndrome where the end-organ can be left intact. End-organ protection may not always involve surgery; high-dose omeprazole completely protects the stomach (medical gastrectomy) in the Zollinger—Ellison syndrome.

The management of primary hyperparathyroidism is complicated by an approximately 10% incidence of hyperplasia. Hyperplasia provides a similar diagnostic pitfall in an insulinoma which could be confused or, worse still, coexist with nesidioblastosis. These problems of differentiation between isolated or diffuse disease can be helped intraoperatively by pathologists using frozen sections, biochemists using rapid assays or radiologists using intraoperative scanning. In parathyroid disease, the pathologist is able to detect suppressed parathyroids or multiple gland disease. Subtle differences in the density of the parathyroids can be used to advantage through a density test − a test that is grossly under-used in surgery of the parathyroids. It must be emphasized that nothing is really as good as the surgeon’s eye. Endocrine surgeons will usually state that they can obtain a very good evaluation of what to expect from macroscopic findings and that it is rare for microscopic findings to surprise the surgeon. But this does not justify the abandonment of intraoperative frozen section. The frozen section has three major roles: first it confirms that the tissue removed is in fact endocrine tissue; secondly, it helps to differentiate the subtle differences between single-and multiple-gland disease; thirdly it indicates the presence of malignancy, (although this can be very difficult because invasion of vessels can often occur in benign tissue). The presence of secondary deposits often clinches the diagnosis of a malignant tumour.

Intraoperative use of hormonal assays also helps with assessment of whether single- or multiple-gland disease is being dealt with, and a good example is the use of the rapid whole-molecule assay for parathyroid hormone (PTH) during parathyroidectomy. Well-established levels of PTH have now been documented during parathyroidectomy and a drop in PTH level in a difficult exploration may indicate whether single- or multiple-gland disease is involved. This is particularly useful in patients undergoing second or third operations on their necks. Similar principles for hormonal assays during surgery are used in the rapid insulin assay to identify insulinomas of the pancreas and in the rapid calcitonin assay in patients with medullary carcinoma of the thyroid.

In difficult explorations, intraoperative ultrasound may localize intrathyroidal parathyroid adenomas. A similar principle is involved in the use of intraoperative ultrasound in the pancreas. This has been particularly useful for insulinomas which are often small and may be localized using this technique.

One of the major problems in re-exploring patients referred from other centres is the lack of intraoperative documentation. It is important that meticulous documentation of findings is made in all forms of endocrine surgery, whether of the pituitary, the thyroid, the parathyroid or the adrenal. The site of the tumours must be documented and their weight, extent of resection and pathological findings on frozen section clearly recorded. However, quite often these details are not recorded and it is necessary to re-explore patients with little or no information about previous surgery!

Although incisions in the neck for parathyroid and thyroid surgery are standard, when dealing with the adrenal the incision will vary. Tumours of less than 5 cm diameter can be easily removed by a posterior approach through a twelfth rib whereas larger tumours usually need an anterior approach. The posterior approach has a lower morbidity than transabdominal and is therefore a great advantage to the patient. However, if there is concern about multiplicity, the ‘gold standard’ for exploration is undoubtedly a transabdominal approach. In the last few years, with improvement of imaging, the author’s group has been able to deal with more and more phaeochromocytomas using a posterior approach. When exploring pancreatic or secondary tumours in the liver, the extent of resection has to be balanced both against morbidity and against improved pharmacological agents to control the symptoms. It may be better to treat a relatively ‘static’ metastatic insulinoma with diazoxide coupled with a debulking exercise rather than just to use more extensive surgery. Similarly, in gastrinomas the use of omeprazole, (often in high dosage), renders unnecessary the need for resection of the stomach. The delicate balance between the use of drugs and surgery is a grey area which still needs considerable thought. This balance should be discussed preoperatively with the referring endocrinologist so a plan can be developed to cover all possible eventualities. It may warrant a preoperative trial of medical therapy, for example to see if diazoxide effectively prevents hypoglycaemia in a patient with an insulinoma, hence allowing some latitude in the surgical approach.

By using these simple rules excellent results can be obtained in endocrine surgery. There will be failures. There may be a resultant persistent hypercalcaemia or persistent hyperinsulinism and when this occurs the basic concepts outlined above must be re-worked. If, on review, the original indications for surgery are correct, the decision taken is based on whether further surgery is necessary or management of the patient can be made through alternative methods. The situation must not drift and, if medical management is not satisfactory, the patient should be aggressively re-imaged in order to find the miscreant tumour or tumours. Failed surgery is no fun for either the patient or the surgeon; postoperative field fibrosis rapidly starts to cause problems and if possible early re-exploration should be performed before obliteration of tissue planes makes further surgery a formidable problem.

References

1. Cockcroft, J. R., Ritter, J. M., Allison, D. J., Causon, R., Brown, M. J. Localisation of extra adrenal catecholamine secreting tumours by selective venous sampling and nuclear magnetic resonance scanning. Postgraduate Medical Journal. 1987; 63:451–453.

2. Geelhoed, G. W., Druy, E. M. Management of the adrenal incidentaloma. Surgery. 1982; 92:866–874.

3. Thompson, N. W., Cheung, P. S.Y. The diagnosis and treatment of functioning and non-functioning adrenocortical neoplasms including incidentalomas. Surgical Clinics of North America. 1987; 67:423–436.

4. Chag, F. L., Wang, C., Imaging endocrine disorders.Clinical Endocrinology and Metabolism. London: Baillière Tindall, 1989.

5. Swensen, S. J., Brown, M. L., Sheps, S. G., et al. ¹³¹I-MIBG in the evaluation of suspected phaeochromocytoma. Mayo Clinic Proceedings. 1985; 60:299–304.

6. Johns, V. J., Jr., Brunjes, S. Phaeochromocytoma. American Journal of Cardiology. 1962; 9:120–125.

7. Mobley, J. E., Headstrong, J. W., Melby, J. C. Primary aldosteronism: Preoperative preparation with spironolactone. Journal of the American Medical Association. 1962; 180:1056.

8. Astwood, E. B. Treatment of hyperthyroidism with thiourea and thiouracil. Journal of the American Medical Association. 1951; 122:1045.

9. Jeffcoate, W. J., Rees, L. H., Tomlin, S., Jones, A. E., Edwards, C. R.W., Besser, G. M., Metyrapone in long-term management of Cushing’s Disease. British Medical Journal 1977; 2:215–247

10. Schulz, D. A., Purnell, D. C. Asymptomatic primary hyperparathyroidism 10 years prospective study. Mayo Clinic Proceedings. 1951; 56:473.

11. Ponder, B. Multiple endocrine neoplasia type 2; the search for the gene continues. British Medical Journal. 1990; 300:484–485.

12. Welbourne, R. B., Manolas, K. J., Khan, O., Galland, R. B. Tumours of the neuroendocrine system (apud cell tumours - APUDOMAS). Current Problems in Surgery. 1984; 21:16.

2

The Molecular Basis of Endocrinology

Simon Wallis

Publisher Summary

This chapter describes the molecular basis of endocrinology. Molecular biology consists of the analysis of organisms and disease in terms of molecular structures and their interrelationships. The flanking regions of the gene are often found to be of importance in determining the pattern and level of expression of the gene. The best understood flanking regions are the promoter where transcription starts and regions further away from the start of the gene that interact with proteins known as transcription factors. Some transcription factors determine the site of expression of the gene, whereas others interact with hormones or are activated by protein kinases controlled by second messengers. This can lead to dramatic changes in the level of expression of some genes especially in response to steroid hormones. The site of expression of a gene may be influenced by factors other than the patterns of expression of tissue-specific transcription factors. The mature mRNA will usually be translated in the cytoplasm, although sometimes mRNA may be produced but not translated. A ribosome binds near the 5-end and moves along the mRNA while producing a protein. Many proteins are synthesized with a highly hydrophobic N terminus that interacts with a set of proteins in the endoplasmic reticulum membrane, causing the translocation of the protein to the lumen of the endoplasmic reticulum.

Overview

Molecular biology consists of the analysis of organisms and disease in terms of molecular structures and their interrelationships. This approach started in the mid-1940s when physicists and chemists turned their attention to biological structures, determining the structure of DNA, myoglobin and insulin. Geneticists studied simple organisms, such as Escherichia coli and its bacteriophages, leading to an understanding of the genetic code, the process of protein synthesis and eventually the discovery of the biological tools of the cloning revolution.

These relatively new techniques are so powerful that purification of proteins, determination of their sequence and isolation of the corresponding gene have become routine in many laboratories. Once the gene has been isolated, it is now usually quite straightforward to express the corresponding protein in E. coli, yeast or higher eukaryotic cells, producing an authentic product in large amounts for further study or even therapeutic use.

A further consequence has been the development of powerful genetic techniques allowing the isolation of a defective gene − ‘reverse genetics’.

There is serious discussion of the possibility of sequencing entire genomes of human beings, mice, fruit flies, eelworms and yeasts. Various bacterial and viral pathogens are also being considered for such projects. It is highly likely that in the next 10–20 years we will know the primary amino acid sequence of every human protein and those of a number of other organisms.

Impact on medicine

Molecular biology has affected most fields of medicine with major implications for diagnostics and therapeutics. The exploration of the therapeutic implications is still in its infancy and has very great potential in many fields:

• Cancer

– oncogenes

– cytokines

– chromosome rearrangements

– tumour suppressor genes

• Immunology

– cytokines

– T-cell receptors

– immunoglobulins

– HLA (human leucocyte antigen)

• Endocrinology

– hormone structures

– signalling mechanisms

– receptors

– endocrine tumours

• Genetics

– RFLP linkage analysis

– allele loss in tumours:

colon carcinoma

renal carcinoma

multiple endocrine adenomatosis (MEA) 1

– gene defects

cystic fibrosis

neurofibromatosis I

Duchenne muscular dystrophy

• Therapeutics

– insulin

– somatostatin

– growth hormone

– tissue plasminogen activator

– interferons

– interleukins

– erythropoietin

• Virology

– human T-leukaemia viruses

– human immunodeficiency virus

– cytomegalovirus

– Epstein -Barr virus

Gene expression

Genome structure and gene families

The human genome consists of 3 billion base-pairs of DNA. The average protein is translated from a messenger RNA of approximately 1500 bases. It is believed that a maximum of 10⁵ individual mRNAs are produced in man. This only occupies 10–50% of the genome, much of the remainder being a filler of uncertain function. The sequence of this non–coding DNA is not without interest and importance: much of the material is made up of families of repetitive DNA sequences, short sequences that may be repeated many times in one location and that may make up a sizeable portion of the genome. These sequences are of practical importance because they show more variation between individuals than coding sequences and allow for ‘finger–printing’ of DNA samples for genetic analysis; they are of importance in genetic disease because many deletions of DNA sequences result from unequal recombination between repeat sequences of the same type in different parts of a gene.

Some genes have no close relatives while others belong to gene families; some are linked closely on the same area of the same chromosome whereas others (usually less closely related) may lie on different chromosomes. Relevant examples in the endocrine system are the growth hormone gene family and the glycoprotein hormone (LH, hCG, FSH and TSH) family.

The growth hormone (GH) gene family consists of growth hormone, prolactin and human chorionic somatotrophin (hCS). GH and hCS genes are arranged in a gene cluster on chromosome 17 where there are two GH genes and three hCS genes within an area of 50 kilobases (kb). These genes are thought to have arisen from a single precursor by recent gene duplication and further divergence. The prolactin (PRL) gene is also closely related to GH and hCS but less closely than GH and hCS are related to each other. The prolactin gene is located on a separate chromosome (see Chapter 6) in line with the finding of GH and PRL in fish pituitary, but the much later occurrence of hCS with the evolution of the placenta.

The glycoprotein gene family consists of the common α– subunit which is combined with different (β-subunits to make the various glycoprotein hormones. LH and hCG are closely related, followed by follicle–stimulating hormone (FSH) then thyroid–stimulating hormone (TSH). This relationship is reflected in their chromosomal localization; as with the GH/hCS/PRL family, the closest relationship between luteinizing hormone (LH) and human chorionic gonadotrophic (hCG) is reflected in the clustering of LH and hCG (β-chain genes on a small portion of chromosome 19. In this case there is one LH (β-chain gene and seven clustered hCG (β-chain genes, some of which may be nonfunctional (pseudogenes). Again it is believed that this cluster arose by a process of reduplication and divergence. The less closely related FSH (β-chain lies on chromosome 11 whilst the TSH (β-subunit lies on chromosome 1. The common α–subunit lies on chromosome 6.

Gene structure

A typical gene has the structure shown in Figure 2.1. It is composed of a patchwork of exons (destined to be part of the mature mRNA) separated by introns which do not usually code for any product, the introns are usually removed from the RNA transcript as it is processed to become mRNA.

Figure 2.1 3′ untranslated region

A gene is often found to lie in or close to a region of chromosome which is both relatively rich in CpG dinucleotides (C = cytosine, G = guanosine) and undermethylated. These properties have given rise to the terms ‘CpG–rich’ or ‘methylation–free’ islands. These characteristics have become very important in reverse genetics such as the attempts to find the genes for multiple endocrine adenomatosis (MEA) 1 and 2 and the earlier successful search for the cystic fibrosis gene.

Control regions

Flanking regions of the gene are often found to be of importance in determining the pattern and level of expression of the gene. The best understood flanking regions are the promoter where transcription starts and regions further away from the start of the gene (this is known as the 5′–end whereas the end of the gene is 3′) which interact with proteins known as transcription factors. Some transcription factors determine the site of expression of the gene whereas others interact with hormones (steroids, triiodothyronine or T3 or retinoic acid) or are activated by protein kinases controlled by second messengers (cyclic AMP or diacylglycerol). This can lead to dramatic changes in the level of expression of some genes especially in response to steroid hormones.

The site of expression of a gene may be influenced by factors other than the patterns of expression of tissue–specific transcription factors. Methylation of regions of DNA may influence tissue–specific expression of a gene; an extreme example of this is the inactivated X chromosome which is extensively methylated. Further evidence for the role of methylation comes from the use of agents which reduce DNA methylation, such as azacytidine which can induce the expression of previously inactive genes in tissues or cells in which they are usually silent. The general level of expression of some genes has been found to be controlled by more distant DNA elements known as dominant control loci (DCL) or locus control regions (LCR). This system has so far been characterized in relation to the non–α–globin gene locus on chromosome 11p. A region of DNA some 50 kilobases distant from the locus was identified initially through its accessibility in chromatin to DNAase I at this point. It was later found that this region was necessary to confer normal levels of expression of these genes in transgenic mice. The precise mechanism is unclear; protein factors binding to this region are being pursued. This mechanism may be of importance in other loci, particularly gene clusters.

It is not yet clear if a gene can ever be completely silent or repressed; some evidence obtained by use of highly sensitive polymerase chain reaction (PCR) techniques suggests that all genes are transcribed although in some situations at very low levels indeed.

Transcription

RNA polymerase II binds downstream of the TATA box and initiates transcription of the RNA copy of one strand of the gene. Transcription continues some way downstream of the end of the gene and the transcript is processed while being exported from the nucleus. The 5′–end is modified (capped), introns are removed (splicing) and the 3′–end is trimmed and tailed with 5- to 25-adenosine residues (polyadenylation).

RNA processing

The splicing process involves a complex series of reactions catalysed by a set of small nuclear ribonuclear protein particles (SNRPS pronounced snurps). These recognize sequences at the ends of introns enabling the precise removal of the intron sequence with reconnection of the ends of the two exons. The 5′–sequence is: Exon NNNNN^ guaagunnnnn Intron whereas the 3′–sequence is Intron nnnnnnannnnn(c/u)ncaĝNNNNN Exon. The excised intron forms a lariat in the process (Figure 2.2).

Figure 2.2 3′ untranslated

Mutations of the sequences shown are a common cause of genetic disease leading to abnormal processed RNA transcripts which cannot be translated and which are often unstable leading to low levels of RNA.

The splicing reaction may proceed differently in different tissues leading to different mature RNAs and hence different protein products being produced from the same gene in a process known as alternative splicing, the best known example in endocrinology being the calcitonin/CGRP gene in which calcitonin lies on exon 4 and CGRP (calcitonin gene-related peptide) lies on exon 5. In C cells exon 4 is included in the mRNA and its poly(A) addition site is used thus removing CGRP from the mature mRNA, whereas in the nervous system exon 4 is spliced out leaving exon 5 in the mature mRNA which allows translation of CGRP (Figure 2.3).

Figure 2.3 Calcitonin gene alternate splicing: the calcitonin gene contains six exons; exon 4 codes for calcitonin and contains a polyadenylation signal; exon 5 codes for CGRP (calcitonin gene-related peptide) and the next polyadenylation site is in exon 6. Splicing in C cells produces an mRNA containing the first four exons and excludes exons 5 and 6, whereas in nerves exon 4 is excluded by a different tissue–specific splicing pattern, giving an mRNA containing exons 1,2,3,5 (CGRP) and 6. The different mRNAs are translated giving calcitonin or CGRP in association with different flanking peptides

Translation and secretion

The mature mRNA will usually be translated in the cytoplasm although sometimes mRNA may be produced but not translated. A ribosome binds near the 5′-end and moves along the mRNA while producing a protein. Many proteins are synthesized with a highly hydrophobic N terminus (signal sequence) which interacts with a set of proteins (signal recognition particle, SRP) in the endoplasmic reticulum (ER) membrane, causing translocation of the protein to the lumen of the ER. Peptide hormones are partially modified within the ER (glycosylation), transferred to the Golgi body where glycosylation is completed and eventually packaged into secretory granules with other proteins, some of which determine the final proteolytic processing of the hormone (Figure 2.4).

Figure 2.4 The secretory pathway

Peptide hormones are frequently synthesized as a large precursor containing a variety of peptides which are generated by cleavage at paired basic amino acids. The pattern of peptides produced may vary between tissues. The classic example of this form of diversity within the endocrine system is the pro-opiomelanocortin (POMC) gene and its products. The 29-kilodalton POMC precursor is processed in anterior pituitary to a 16–kDa N-terminal fragment, ACTH and βLPH, the cleavages all taking place at paired dibasic amino acids (lysine–argine, Lys–Arg). However, non-pituitary POMC processing may produce a variety of other smaller peptides: γMSH from the N–terminal peptide, α–MSH and CLIP from ACTH, γLPH, (β-endorphin, (βMSH and metenkephalin from (βLPH (βLPH = β-lipotrophic hormone, MSH = melanocyte-stimulating hormone, CLIP = corticotrophin-like intermediate lobe peptide). This pathway occurs in the intermediate lobe in the fetus and in other mammals, as well as in the placenta and possibly other tissues (Figure 2.5).

Figure 2.5 intermediate lobe pituitary, fetal pituitary and other tissues; paired dibasic amino acids

This variety in processing is probably conferred by the expression of processing enzymes of differing specificities. Many peptide hormones have other flanking peptides, some of which have no known biological activity. Peptide hormone receptors follow a similar route but remain anchored in the cell membrane by one or several hydrophobic domains consisting of about 30 amino acids.

Some steroid hormone receptors are targeted to the nucleus by a specific targeting signal which is recognized by nuclear pore proteins, followed by translocation into the nucleus. In some cases this sequence is masked by a second protein which is displaced by conformational changes following hormone binding.

Control points

In such a complex process, it is obvious that there are many potential points where control of output may be exerted.

Transcription rates can be altered quite rapidly, the main examples of this within the endocrine system being shown by the action of various types of steroid hormone. The main immediate action of peptide hormones is the release of preformed stores of hormone from secretory granules or other pools.

The rate of processing to mRNA may be controlled and the accessibility of the mRNA for translation can also be modulated. mRNA stability can be varied as can its translation rate. In addition, the translated protein may be diverted from the secretory pathway to be destroyed in lysosomes.

Signal transduction

All hormones act through receptors which generate diverse signals within the cell allowing transmission of the effect of the hormone to intracellular sites of action. Hydrophilic hormones (polypeptides, peptides, catecholamines, dopamine and serotonin or 5HT) bind to cell surface receptors; these generate an intracellular signal whereas hydrophobic hormones (steroid hormones, thyroxine, calcitriol and reti– noic acid) are able to diffuse across the cell membrane and bind to cytoplasmic or nuclear receptors which have major actions as transcription factors.

G protein–linked receptors

The main group of cell surface receptors that is relevant to endocrinology has been characterized by close interaction with G proteins which mediate further signal generation. The majority of peptide hormones act through such G protein–linked receptors, the major exceptions being insulin, insulin–like growth factors I and II (IGF–I, IGF–II), growth hormone and prolactin. Among G protein-linked receptors there are differences in the balance between the main intracellular signalling routes and the degree of activation or inhibition produced.

The structural characteristics of G protein–linked receptors are well conserved between all members of the group. A constant feature of these receptors is the presence of seven hydrophobic transmembrane domains, a cytoplasmic C–terminal tail and very variable (sequence and size) N-terminal extracellular domain which in many cases is the ligand–binding domain. The transmembrane domains are moderately well conserved between members of the family, particularly those binding related ligands. The transmembrane domains anchor the receptor in the membrane and provide cytoplasmic loops which can contribute to interactions with G proteins or can be the targets of modifications, e.g. phosphorylation, which can lead to desensitization of the receptor. Receptors for peptides and polypeptides have a relatively large extracellular domain which binds the hormone and presumably causes a conformational change leading to enhanced interaction with G proteins. Receptors for catecholamines and dopamine have a very short external domain and the ligand interacts with the loops formed by the transmembrane domains (Figure 2.6).

Figure 2.6 hydrophobic transmembrane domain

The structures of these receptors have been derived by sequencing copy DNA (cDNA) or genomic clones coding for the receptor. These clones are often isolated using a functional assay to detect a full length clone. This either involves expression of an appropriate cDNA library in a mammalian cell line, such as COS cells (a monkey kidney cell line), or expression in xenopus oocytes of micro–injected RNA produced by transcription of an appropriate cDNA library, followed by selection of cells that carry the receptor on their surface by detection of the binding of a labelled ligand. Other methods have relied on the large degree of homology between some members of the family to allow isolation of others either by homology screening or by homology polymerase chain reaction.

G proteins belong to a large group of proteins that bind guanosine triphosphate (GTP) and are involved in many aspects of signalling. Gα, Gβ and Gγ form a trimeric complex and transmit signals either through adenylate cyclase or through phospholipase C. Gα has an inhibitory form Gαi and a stimulatory form Gαs. Each type of G protein has been found to have several related sequences generating potential for differential signal generation. The trimeric complex interacts with the receptor and GTP, resulting in the generation of the active Gα which has bound GTP. This is the effector which can influence adenylate cyclase or phospholipase C activity. The GTP is hydrolysed resulting in inactivation which is followed by binding of Gβγ and displacement of GDP to regenerate the trimeric form ready to interact again with the receptor (Figure 2.7).

Figure 2.7 G protein activation cycle: α, β, γ G protein subunits. GTP = guanosine triphosphate; GDP = guanosine diphosphate

When adenylate cyclase has been activated, cyclic AMP is produced and this, in turn, can activate cyclic AMP-dependent protein kinase which can initiate a signal cascade affecting transcription factors, ion channels and hormone secretion.

Activation of phospholipase C causes hydrolysis of membrane phospholipid generating diacylglycerol (DAG) within the membrane and inositol triphosphate (IP3) within the cytoplasm. DAG causes activation of the phorbol ester sensitive pathway by activation of protein kinase C. This enzyme again exists in multiple forms generating the potential for considerable signalling diversity. Protein kinase C initiates another cascade of signals including the activation of transcription factors. IP3 has as its major effect the activation of an intracellular IP3 receptor which also acts as a calcium channel, allowing sudden increases in the intracellular calcium level. This acts as an important stimulus to exocytosis and hormone secretion. One of the effects of the intracellular Ca²+ increase is to activate calmodulin, a ubiquitous calcium–binding protein which further activates a series of intracellular protein kinases that may regulate enzymes and other processes (Figure 2.8).

Figure 2.8 Phospholipase C and protein kinase C signalling: R = receptor; Gαs = G protein subunit activates PLC; PLC = phospholipase C; PIP2 = phosphatidylinositol diphosphate;IP3 = inositol triphosphate; DAG = diacylglycerol; Ca²+ = calcium ions

Tyrosine kinase receptors

Insulin and IGF–I act through similar receptors which belong to a wider family of receptors characterized by intracellular tyrosine kinase activity.

The insulin receptor is synthesized as a precursor containing α– and β–subunits, two of each associating in the membrane to form a mature receptor. The (β-chains span the cell membrane and carry the tyrosine kinase activity whereas the α–chains are extracellular, linked to the (β-chains by disulphide bonds, and carry the insulin-binding domain. Insulin binding causes tyrosine kinase activation and subsequent receptor autophosphorylation. Several intracellular proteins are phosphorylated at the same time, but there is also evidence accumulating of some signal generation by mechanisms involving phosphoinositol glycan generation.

The IGF–I receptor is very similar to the insulin receptor; it clearly also shares the property of tyrosine kinase activation.

The IGF–II receptor is identical to the mannose-6-phos– phate receptor which has no tyrosine kinase activity; however, IGF–II can also activate the insulin and IGF–I receptor which may actually be responsible for its growth–promoting activity (Figure 2.9).

Figure 2.9 Insulin IGF–I and IGF–II receptors

Other tyrosine kinase receptors are activated by platelet-derived growth factor (PDGF), epidermal growth factor (EGF), CSF I, GM CSF (growth factors).

Growth hormone and prolactin receptors

These are very closely related as are their ligands. Each has a large extracellular ligand-binding domain, a short transmembrane domain and a large cytoplasmic domain which has no homology with any other receptor.

This domain is presumably involved in signal generation within the cell, linking, for example, GH binding with anabolic effects and the secretion of IGF–I. Of further interest is the finding that the serum GH–binding protein is derived from the extra-cellular portion of the GH receptor (Figure 2.10).

Figure 2.10 Growth hormone and prolactin receptors

Atrial natriuretic factor receptor

Atrial natriuretic peptide is a hormone produced by the atrium in response to increased atrial pressure and which subsequently causes sodium excretion by the kidney. The receptor in the kidney is of great interest because it contains an integral guanylate cyclase activity and directly produces cyclic GMP as a second message. There are related forms expressed in the brain but their function as yet remains uncertain (Figure 2.11).

Figure 2.11 ANP receptor

Control of transcription

Genes are transcribed from chromatin within the nucleus to produce an initial RNA transcript which is subsequently modified by splicing and polyadenylation to produce a mature messenger RNA molecule; this is then exported from the nucleus to the cytoplasm where it is translated on ribosomes to produce a protein product.

Transcription is a major site of control within the cell. RNA polymerase II interacts with 5′-control regions immediately upstream of the coding regions of a gene, but this process is extensively modulated by other protein factors known as transcription factors. Transcription factors are major determinants of the tissue distribution of expression (tissue specificity) and level of expression in response to other signals. Transcription factors interact with specific short DNA sequences in the 5′–flanking region, with each other and with RNA polymerase, to enhance transcription; alternatively, some negatively acting transcription factors prevent effective action of RNA polymerase and thus down–regulate transcription.

Transcription factors

Steroid hormone receptors

A major class of transcription factors (TFs) of great relevance to endocrinology is the extensive family of steroid hormone receptors, which has been greatly extended by homology cloning to reveal several relatives with, as yet, unknown ligands and, in some cases, unknown function.

Steroid hormone receptors (SHRs) have been studied for many years as cellular binding sites for radioactively labelled steroid hormones. The oestrogen, progesterone and T3 (triiodothyronine) receptors were cloned in the same year, the first two by classic protein purification and sequencing routes and the latter by the realization that the erbA oncogene coded for a similar receptor which turned out to bind T3. Homology cloning has resulted in the isolation of clones coding for receptors for all the steroid hormones, several additional T3 receptors, the developmental morphogen retinoic acid and several ‘orphan receptors’.

All steroid hormone receptors have considerable structural homology, at the level of protein sequence, and more particularly in the organization of structural domains. All have a highly homologous region now known to be involved in binding to DNA sequence elements within the 5′-flanking regions of responsive genes. This region consists of two loops of about 12 amino acids stabilized by four cysteine molecules coordinated to a single atom of zinc at the base of each loop; this motif is commonly found in TFs and is known as a ‘zinc finger’ (Figure 2.12).

Figure 2.12 Zinc finger DNA–binding domain of oestrogen receptor

The T3 receptor is always firmly bound to chromatin but, on binding, T3 undergoes a change which facilitates transcription. By contrast, the glucocorticoid receptor is mainly cytoplasmic until the ligand is bound, when a nuclear localization signal is thought to be revealed followed by translocation to the nucleus, binding to chromatin and effects on transcription. The oestrogen and progesterone receptors are mainly nuclear but weakly bound as a result of occlusion of the DNA–binding domain by association with a heat shock protein (HSP 90); this binding is abolished following conformational changes induced by ligand binding, resulting in exposure of the DNA–binding domain and subsequent chromatin binding.

Pit 1 is an example of a TF which mediates tissue specificity of expression, in this case allowing high level expression of growth hormone in pituitary somatomammotrophs but not elsewhere.

Pit 1 is closely related to other TFs found in species ranging from humans to Drosophila spp. It contains an N–terminal transcription–activating domain rich in serine and threonine. At the C–terminal end, there are two conserved DNA–binding domains. The N–terminal end of the two has great homology with Oct 1,2 and Unc 86 whereas the more C-terminal domain has considerable homology to a consensus homeo domain found in genes that control segmentation in flies and vertebrates.

Pit 1 binds to two DNA elements in the first 100 base-pairs of the 5′-flanking region of the GH gene and stimulates cell–type specific expression of GH and prolactin.

Cyclic AMP and transcription

Cyclic AMP is generated within cells as a result of many different external stimuli transduced through G protein– linked receptors and is usually accompanied by the generation of multiple intracellular signals (DAG, FTP, calcium, protein kinase C activation). This constitutes two major pathways that affect transcription: cyclic AMP activates cyclic AMP–dependent protein kinase A whereas diacyl– glycerol (DAG) activates protein kinase C. These kinases lead to phosphorylation and activation members of a complex family of TFs which bind to DNA as dimers, linked by hydrophobic interaction of a series of leucine residues known as a ‘leucine zipper’. It was initially believed that this was a simple system involving a single factor mediating stimulus and interacting with specific DNA regions. It now appears that the situation is highly complex; the DNA elements recognized by the cAMP and DAG pathways are highly similar.

Figure 2.13 Steroid hormone receptors: the main classes of receptor are shown aligned in reference to the homologous DNA-binding domains

TPA is a potent analogue of DAG. It now appears that the factors recognizing these elements belong to related families which are capable of extensive and complex interactions.

The factors that bind to the CRE are known collectively as CREBs (cyclic AMP response element–binding protein). The CREB proteins are related to the proteins (activating transcription factors – ATFs) of animal cells activated by the E1 A proteins of adenoviruses. At least 10 distinct CREBs have been partially characterized and appear to be the products of different genes.

In a similar manner, it is now clear that there are multiple factors which can bind to the TRE and are known collectively as TREBs. The first to be characterized was API (activator protein 1), a Simian virus (SV40) transactivator. This is related to the Jun oncogene product encoded by a chicken sarcoma virus and to a protein that co-purified with the Fos oncogene product known as Fra (Fos-related antigen).

All these factors have some similarities: all contain the leucine zipper motif believed to mediate dimerization of such proteins and most contain related basic domains believed to interact with DNA. There appears to be some specificity encoded in the structure of the zipper which determines the pattern of dimerization between the various CREBs and TREBs. Jun–Jun dimers and Jun–Fos dimers interact with TREs whereas CREB–CREB or creb– Jun dimers interact with CREs. They all appear to be activated by phosphorylation and are probably inactivated by dephosphorylation.

Tumour formation

Endocrine tumours vary from hyperplastic glands through simple adenomas to metastasizing malignancies, and there is much interest in the cellular and molecular changes underlying these states. Hyperplastic glands are generally polyclonal resulting from proliferation and increased secretion from all cells in response to an external stimulus. Adenomas and malignancies are often monoclonal resulting from the proliferation of a single cell with a growth advantage to produce a tumour. The situation may be complicated by the production of growth factors by the tumour and consequent recruitment of other cell populations.

The increased rate of growth of tumour tissue may be the results of activation, abnormal expression or over-expression of positive factors (oncogenes) or the result of inactivation, decreased expression or deletion of negatively acting factors (tumour suppressor genes).

Oncogenes

Oncogenes now number over 100 and have been discovered by two main routes: either as the cellular equivalent of the transforming gene of an animal retrovirus or other tumour-forming virus or by isolation of a transforming gene from tumour DNA by transfection into NIH 3T3 cells. The normal cellular equivalent of an oncogene is usually found to be part of the signal transduction pathway. The origins of some nuclear oncogenes are less certain but in many cases it is clear that they are related to or identical with transcription factors.

Oncogenes may be related to secreted growth factors which act in an autocrine manner to sustain cell growth, to cell surface receptors especially of the tyrosine kinase family, to G proteins and to other GTP–binding proteins or to transcription factors. For a classification of the oncogenes see Table 2.1.

Table 2.1

Oncogenes

Activation may occur through point mutation which is often found in ras genes resulting in increased activity of the product. An alternative mechanism is translocation so that the oncogene comes under the control of a different promoter resulting in unregulated or higher level expression; this may result in either increased expression of the normal oncogene product or a novel fusion transcript. Finally, the level of expression of an oncogene may be increased by amplification of the copy number of the gene.

In some instances several oncogenes may be active in the same tumour.

It is important to note that, despite the importance of oncogenes in the development of malignancy, many will be expressed in normal tissue or normal proliferating tissue so immunohistochemical demonstration of oncogene expression is rather poor evidence in favour of a pathogenic role. The evidence becomes much stronger if an activating mutation or gene amplification can be demonstrated. However, in some cases the level of expression may be so high as to be persuasive.

p53 mutations lead to a great increase in p53 protein.

Tumour suppressor genes

Control systems frequently operate with combinations of negative and positive factors so it should not be surprising that the loss of negative factors has begun to assume great importance in tumour biology. The first example of such tumour suppressor genes emerged from the study of the childhood cancer retinoblastoma. Such tumours generally occur in susceptible families in which one germline copy of the retinoblastoma (rb) gene is inactivated, and this is followed by loss of the remaining normal gene in the cell that gives rise to the retinoblastoma. These findings are consistent with the notion that the normal role of the rb protein is to down–regulate, or hold in check, the activity of a factor involved in the regulation of cell growth. Since this discovery, it has become clear that a number of genetic tumour syndromes have a similar basis and it has been realized that the loss of tumour suppressor genes can play a major role in the development of cancer. The gene rb has been found to have a role in several tumours in addition to retinoblastoma.

p53 has been found to have a similar role in the rare genetic cancer Li Fraumeni syndrome as well as in bronchial carcinoma and hepatocellular carcinoma.

Three unrelated tumour suppressor genes have been found through the study of hereditary polyposis coli and its progression to colonic carcinoma. One of these, DCC (deleted in colon carcinoma) is of particular interest because it resembles a cell adhesion molecule and probably alters the cell surface properties of the colon cancer cells.

A gene on the short arm of chromosome 3 has been found to be linked to the von Hippel–Lindau syndrome and is usually deleted in the phaeochromocytomas and renal carcinomas found in that condition. The same area, and probably the same tumour suppressor gene, are often deleted in spontaneous renal and small cell lung carcinomas.

Recently, neurofibromatosis (NFI) was found to be caused by inactivating mutations of a very large gene with significant homology to GAP (GTPase–activating protein) which regulated the activity of GTP–binding proteins, such as ras.

MEA–1 and −2 fall into the same class of defect, although neither gene has been isolated as yet. Given the important role of tyrosine kinases as oncogenes and the relatively recent discovery of phosphatases that can reverse their effects, it is likely that these may emerge as a new category of tumour suppressor gene.

Successful tumour growth requires the recruitment of a vascular supply which may be facilitated by growth factor secretion; the process of spread within a tissue may be enhanced by secretion of proteases which digest surrounding connective tissue (stromelysins). The process of metastasis may be affected by altered cell surface adhesion molecule expression.

In many cases, the process of carcinogenesis is a multistep phenomenon involving various combinations of oncogene activation and tumour suppressor gene inactivation.

Conclusions of relevance to clinical endocrinology

Hypercalcaemia of malignancy

A new peptide hormone (PTH–related peptide, PTHrP[1]) has been discovered by studying this syndrome. A protein was isolated from several solid tumours associated with hypercalcaemia based on its stimulatory effect on renal cyclic AMP production. The N–terminal sequences of the protein and several proteolytic cleavage products were determined and allowed the design of oligonucleotide probes, which were then used to screen a cDNA library made from tumour RNA; this in turn led to the successful isolation of cDNA clones from which the entire protein coding sequence could be determined. The new hormone PTHrP has some homology to PTH at its N terminus which interacts with the PTH receptor, but the remainder is not closely related.

This work has led to the introduction of a number of immunoassays for PTHrP and the detection of PTHrP in the fetus, breast milk, skin and the parathyroid gland, in addition to detection in the solid tumours responsible for hypercalcaemia.

The relatively recent characterization of new peptide hormones such as inhibin, erythropoietin, atrial natriuretic peptide and endothelin owes much to the great technical advances seen in protein chemistry and DNA technology.

G protein abnormalities

A group of acromegalic subjects was found to have raised cyclic AMP levels within the tumour tissue. This subsequently led to the discovery of an abnormally active Gαs protein subunit which resulted from a single base mutation in the DNA of the tumour, but not in DNA elsewhere. Subsequent extension of this study searched for Gαs and Gαi mutations in acromegaly and many other endocrine tumours [2]. The findings in acromegaly were confirmed with 43% of GH adenomas being associated with the Gαs mutation. Inactive Gαs mutants were found mainly in ovarian and adrenocortical tumours.

Similarly, in Albright’s syndrome (a form of PTH resistance or pseudohypoparathyroidism) the defect in renal cyclic AMP response to PTH has been found to be a result of a Gαs subunit mutation [3], in this case leading to a loss of function.

Papillary carcinoma of the thyroid (PTC) has been found in some cases to be caused by the activation of an oncogene. This was discovered by studying the effect of PTC tumour DNA on the phenotype of the NIH 3T3 cell. DNA was transfected into the NIH 3T3 cells and this was followed by the development of a more malignant phenotype. This assay allowed the isolation and characterization of the relevant oncogene, which proved to be ret – originally found in a T–cell lymphoma, colon carcinoma and gastric carcinoma. In PTC the ret oncogene is activated by rearrangement and fusion to the 5′–control regions and N terminus of a second gene of unknown function expressed in the thyroid [4].

A similar study of sporadic parathyroid tumours has revealed an interesting novel growth-stimulating gene. DNA samples from such tumours were digested with restriction enzymes, blotted and probed with PTH (parathyroid hormone) gene probes to allow mapping of the PTH gene. A rearrangement was found which brought a piece of DNA from another region of chromosome 11 close to the promoter regions of the PTH gene. This region has recently been cloned and found to resemble other genes that control the cell cycle. The tumour is heterozygous for this rearrangement, and in addition carries normal copy of the PTH gene allowing continuing PTH production [5].

Neurofibromatosis type I has been known as a common dominantly inherited condition for many years and was eventually localized to an area of chromosome 17 by linkage analysis. Subsequent study of rare patients with chromosome translocations and NFI allowed the finer mapping of the gene and its eventual isolation. This has turned out to code for a huge protein with startling homology to a yeast protein and to the GTPase-activating protein in humans [6].

Von Hippel–Lindau syndrome is a rare condition associated with phaeochromocytomas, angiomas and renal carcinoma. Linkage studies in these families has demonstrated linkage with the oncogene raf I on the short arm of chromosome 3 [7]. Study of tumour DNA (renal carcinoma and phaeochromocytoma) has demonstrated deletions and rearrangements in this region which should allow the isolation of the gene defect which is clearly distinct from raf I in the near future.

Appendix: Definitions

Bacteriophage or phage: larger viruses which infect E. coli and some other bacteria. Often used for high efficiency cloning, especially of larger segments of DNA (up to 25 kb, although recently a phage vector has appeared which can accommodate 100 kb).

Blotting: transfer of molecules from a gel to a paper or nylon support. Different types of blotting are used to identify different molecules: Southern: DNA; Northern: RNA; Western: proteins.

Southern blotting involves several very simple steps: genomic DNA is cut at specific sites using a restriction enzyme producing fragments ranging in size from a few bases to as many as 50 or more kilobases. The fragments are then separated according to size by electrophoresis through an agarose slab gel. The DNA can be visualized after staining with the dye ethidium bromide which causes intense fluorescence on illumination of the DNA gel with ultraviolet light. The gel is then soaked in sodium hydroxide to separate the DNA strands, then in buffer to neutralize the base leaving the DNA single–stranded and ready for capillary transfer to a filter. The DNA is transferred to the filter by allowing a stack of tissues to draw buffer through the gel and across the filter support. The buffer passes through the filter, but the DNA remains bound to the filter in a replica of its position in the gel. The filter is baked to fix the DNA and is then ready for hybridization.

Northern blotting is a similar process in which RNA is size-fractionated in an agarose gel and then transferred to a filter. The main difference is that RNA is already single-stranded so it can be transferred directly after electrophoresis.

Western blotting involves the electrophoretic transfer of proteins from a polyacrylamide gel to a filter usually followed by antibody detection of a protein of interest.

cDNA: copy DNA produced by the action of reverse transcriptase on RNA, commonly an mRNA primed by a small DNA sequence, usually oligo(dT), which starts synthesis on the poly(A) sequence, although a specific or random sequence may be used in some situations. The sequence produced is the complement of the RNA sequence. The resulting RNA/DNA hybrid is converted to double-stranded DNA by further enzymatic steps and is later inserted into an appropriate vector (plasmid or phage).

cDNA library: collection of cDNA molecules cloned into a suitable vector. Such a library should usually start with several million clones to guarantee representation of rare species.

Cloning: set of manipulations involved in inserting a stretch of