Mortality and morbidity in Cushing's disease over 50 years in Stoke-on-Trent, UK: audit and meta-analysis of literature.

CONTEXT: Pituitary ACTH-dependent Cushing's disease (CD) is uncommon, and there are very limited data on long-term mortality. OBJECTIVE: The aim was to summarize what is known about mortality in ACTH-dependent CD, to report on our own data, and to provide a meta-analysis of six other reports that addressed mortality of CD. DESIGN AND METHODS: Vital status of 60 CD patients was recorded as of December 31, 2009, and the standardized mortality ratio (SMR) was calculated and compared with the general population of England and Wales, United Kingdom. A meta-analysis of SMRs from seven studies (including ours) was performed for overall mortality in CD. Where reported (four studies), a similar meta-analysis was performed for those patients whose hypercortisolism was in remission after treatment compared to those patients from the same center with persistent disease. RESULTS: 1. From Stoke-on-Trent, 51 of 60 patients were female, median age at diagnosis was in the range of 36-46 yr, and median follow-up was 15 yr. There were 13 deaths, nine due to cardiovascular disease. Overall SMR for the whole cohort was 4.8 (95% confidence interval, 2.8-8.3) (P < 0001). SMR for vascular disease was 13.8 (7.2-36.5) (P < 0001). For persistent disease (n = 6), SMR was 16 (6.7-38.4) vs. remission (n = 54) SMR of 3.3 (1.7-6.7); after adjustment for age and sex, relative risk of death for persistent disease was 10.7 (2.3-48.6) (P = 0.002). Hypertension and diabetes mellitus were associated with significantly worse survival. 2. Using a random effects model meta-analysis revealed an overall (remission plus persistent disease) SMR of 2.2 (1.45-3.41) (P < 0.001). Pooled SMR was 1.2 (0.45-3.2) (P = not significant) for patients in remission and 5.5 (2.7-11.3) (P = 0.001) for patients with persistent disease. Persistence of disease, older age at diagnosis, and presence of hypertension and diabetes are the main determinants of mortality. CONCLUSIONS: Overall mortality in CD is double that of the general population. However, patients with CD in remission fare much better than those with persistence of hypercortisolism, and they appear not to have an increased mortality rate. Hypertension and diabetes mellitus are risk factors for worse outcome. Because diagnosis and treatment of patients are at a young age, much longer follow-up of patients in remission is required before one can be confident that their mortality outcome is no different from that of the general population, especially because cardiovascular risk factors may persist after successful biochemical control of the disease.

Mortality in Cushing's disease.

The causes of premature death in untreated Cushing's syndrome are vascular disease (myocardial infarction/stroke), uncontrolled diabetes mellitus and complications and infections. Long-term mortality outcome studies on pituitary-dependent Cushing's disease (CD) are limited to six studies in the English language literature. This paper reviews these studies on CD, other causes of Cushing's syndrome being excluded, because CD represents 80% of patients with the syndrome. The period covered by these studies (1970-1990) is when transsphenoidal surgery was well established as primary treatment for CD. Two studies were exclusively from surgical centres and are likely biased in favour of surgically resectable adenomas, so this needs to be borne in mind when interpreting their results. The criteria for remission of hypercortisolism and persistent disease were variable. The overall number of patients in each report is small, and the number of deaths even smaller by epidemiological standards giving very wide confidence intervals to the standardised mortality ratios (SMR). Moreover, follow-up time was relatively short (median 10-12 years) for a disease diagnosed in the patients' late 30s. Notwithstanding the above limitations of retrospective studies, and potential for positive bias, the overall SMR of around 1.5 was not significantly different from the relevant normal population for those patients deemed in remission. However, SMR was significantly worse for those patients with persistent disease. Where it was possible to analyse contributing factors to mortality, the presence of hypertension and diabetes mellitus, in addition to persistence of hypercortisolism, was shown to be significant. It remains possible that an overall SMR in 'cured' patients would be significant given a larger cohort, followed for longer, and with more deaths. What is clearly required is a multicentre prospective cohort study with >30 years' follow-up to answer the question definitively and identify the contributing factors in detail in order to achieve optimum long-term outcome.

Bromocriptine and dopamine mediate independent and synergistic apoptotic pathways in pituitary cells.

Dopamine (DA) agonists are the primary treatment choice for prolactinoma, effectively suppressing prolactin expression and reducing tumour size. However, the intracellular pathway(s) through which either DA or its agonists impact on proliferation or lead to tumour shrinkage are incompletely understood. To identify the mediators in the apoptotic cascades after DA or DA agonist challenges we used a well-characterized model system, the rodent somatolactotroph cell line GH3. In these cells, we show that apoptosis induced by the DA agonist bromocriptine (BC), but not DA, is initiated through activation of the JNK pathway. However, both DA and BC activate the terminal effector caspase, caspase-3. Kinetic studies and chemical inhibitor co-incubation experiments support a role for JNK activation preceding caspase-9 activation in BC challenged cells, however, engagement of these mediators was not apparent in DA challenge cells. These studies suggest that apoptosis induced by BC or DA is mediated through distinct and independent pathways that converge with activation of the terminal caspase, caspase-3. These observations were further reinforced by our findings that DA and BC, in co-incubation experiments, synergistically induce apoptosis. These findings raise the possibility that drugs acting through the same pathway as DA may be clinically beneficial when combined with BC.

ACTH deficiency, higher doses of hydrocortisone replacement, and radiotherapy are independent predictors of mortality in patients with acromegaly.

CONTEXT: A number of retrospective studies report that patients with acromegaly have increased morbidity and premature mortality, with standardized mortality ratios (SMR) of 1.3-3. Many patients with acromegaly develop hypopituitarism as a result of the pituitary adenoma itself or therapies such as surgery and radiotherapy. Pituitary radiotherapy and hypopituitarism have also been associated with an increased SMR. METHODS: Using the West MIDLANDS: Acromegaly database (n = 501; 275 female), we assessed the influence of prior radiotherapy and hypopituitarism (and replacement therapy) on mortality in patients with acromegaly. Median duration of follow-up was 14.0 yr (interquartile range, 7.9-21 yr). RESULTS: All-cause mortality was elevated [SMR, 1.7 (1.4, 2.0); P < 0.001]. On external analysis, prior radiotherapy, ACTH, and gonadotropin deficiency were associated with an elevated SMR [radiotherapy SMR, 2.1 (1.7-2.6); P = 0.006; ACTH deficiency SMR, 2.5 (1.9-3.2); P < 0.0005; and gonadotropin deficiency SMR, 2.1 (1.6-2.7); P = 0.037]. On internal analysis, the relative risk (RR) of mortality was increased in the radiotherapy [RR, 1.8 (1.2-2.8); P = 0.008] and ACTH-deficiency groups [RR, 1.7 (1.2-2.5); P = 0.004], but not in the gonadotropin- or TSH-deficiency groups. In the ACTH-deficient group, increased replacement doses of hydrocortisone greater than 25 mg/d were associated with increased mortality compared to lower doses. CONCLUSIONS: Radiotherapy and ACTH deficiency are significantly associated with increased mortality in patients with acromegaly. In ACTH-deficient patients, a daily dose of more than 25 mg hydrocortisone is associated with increased mortality compared to lower doses. These results have important implications for the treatment of patients with acromegaly and also raise issues as to the optimum hydrocortisone treatment regimens for ACTH-deficient patients.

Pituitary tumours: all silent on the epigenetics front.

Investigation of the epigenome of sporadic pituitary tumours is providing a more detailed understanding of aberrations that characterise this tumour type. Early studies, in this and other tumour types adopted candidate-gene approaches to characterise CpG island methylation as a mechanism responsible for or associated with gene silencing. However, more recently, investigators have adopted approaches that do not require a priori knowledge of the gene and transcript, as example differential display techniques, and also genome-wide, array-based approaches, to 'uncover' or 'unmask' silenced genes. Furthermore, through use of chromatin immunoprecipitation as a selective enrichment technique; we are now beginning to identify modifications that target the underlying histones themselves and that have roles in gene-silencing events. Collectively, these studies provided convincing evidence that change to the tumour epigenome are not simply epiphenomena but have functional consequences in the context of pituitary tumour evolution. Our ability to perform these types of studies has been and is increasingly reliant upon technological advances in the genomics and epigenomics arena. In this context, other more recent advances and developing technologies, and, in particular, next generation or flow cell re-sequencing techniques offer exciting opportunities for our future studies of this tumour type.

Loss of neuronatin expression is associated with promoter hypermethylation in pituitary adenoma.

The imprinted gene, neuronatin (NNAT), is one of the most abundant transcripts in the pituitary and is thought to be involved in the development and maturation of this gland. In a recent whole-genome approach, exploiting a pituitary tumour cell line, we identified hypermethylation associated loss of NNAT. In this report, we determined the expression pattern of NNAT in individual cell types of the normal gland and within each of the different pituitary adenoma subtypes. In addition, we determined associations between expression and CpG island methylation and used colony forming efficiency assays (CFE) to gain further insight into the tumour-suppressor function of this gene. Immunohistochemical (IHC) co-localization studies of normal pituitaries showed that each of the hormone secreting cells (GH, PRL, ACTH, FSH and TSH) expressed NNAT. However, 33 out of 47 adenomas comprising, 11 somatotrophinomas, 10 prolactinomas, 12 corticotrophinomas and 14 non-functioning tumours, irrespective of subtype failed to express either NNAT transcript or protein as determined by quantitative real-time RT-PCR and IHC respectively. In normal pituitaries and adenomas that expressed NNAT the promoter-associated CpG island showed characteristics of an imprinted gene where approximately 50% of molecules were densely methylated. However, in the majority of adenomas that showed loss or significantly reduced expression of NNAT, relative to normal pituitaries, the gene-associated CpG island showed significantly increased methylation. Induced expression of NNAT in transfected AtT-20 cells significantly reduced CFE. Collectively, these findings point to an important role for NNAT in the pituitary and perhaps tumour development in this gland.

Medical therapy in patients with acromegaly: predictors of response and comparison of efficacy of dopamine agonists and somatostatin analogues.

CONTEXT: Acromegaly is associated with increased morbidity and mortality. Treatment options include surgery, radiotherapy, and medical therapy. AIMS: The objective of the study was to examine the role of prolactin status, prior surgery, and radiotherapy on the response to medical therapy in patients with acromegaly and assess the relative efficacy of dopamine agonist therapy compared with somatostatin analog therapy. MATERIALS AND METHODS: A total of 276 patients with acromegaly received either dopamine agonists (DA) and/or somatostatin analogs (SSA). One hundred seventy-two patients had received surgery and 73 radiotherapy prior to receiving medical therapy. One hundred ninety-eight of 276 received DA, and 143 of 276 received SSA. GH and IGF-I values at baseline and after 12 months on therapy were analyzed. RESULTS: In the DA group, basal prolactin concentration did not predict response to therapy, GH percent reduction: hyperprolactinemia, 26.7% (-10.4 to 48) vs. normoprolactinemia, 34.8% (0.2-53.2), P = 0.58; IGF-I percent reduction: hyperprolactinemia 30.0% (9.2-43.1) vs. normoprolactinemia 16.8% (4-37), P = 0.45. Prior surgery was not associated with any difference in response to DA: GH percent reduction (P = 0.1) and IGF-I percent reduction (P = 0.08). By contrast, prior radiotherapy was associated with an enhanced efficacy of GH response to DA, P = 0.02. In the SSA group, there was no effect of prior surgery or radiotherapy on response of GH, but radiotherapy was associated with less marked IGF-I percent reduction (P = 0.05). SSA were more potent than DA at decreasing both GH [62.8% (20.7-85%) vs. 42.4% (-6.5 to 68.6), P < 0.008] and IGF-I [SSA 40.4% (0-64.3) vs. 8% (0-40.8), P = 0.05]. CONCLUSIONS: The effects of DA are irrespective of baseline prolactin concentrations. Prior radiotherapy is associated with differences in GH and IGF-I response to DA and SSA therapy.

Monitoring disease activity using GH and IGF-I in the follow-up of 501 patients with acromegaly.

CONTEXT: The aims of treatment in patients with acromegaly are to achieve serum GH/IGF-I concentrations associated with cure or normalization of mortality and alleviation of symptoms. OBJECTIVE AND METHODS: Using the West Midlands Acromegaly database (n = 501) we investigated the reliability of basal fasting GH in predicting nadir or mean GH during oral glucose tolerance test (OGTT) or GH day curve (GHDC), respectively, the degree of discordance between disease activity measured by GH and IGF-I values and the effect of radiotherapy on the above relationships. In total 773 OGTT and 507 GHDC were performed. RESULTS: Basal fasting GH was strongly correlated with nadir/mean GH on OGTT/GHDC (r = +0.87, P < 0.0001, r = +0.93, P < 0.0001, respectively). A basal GH < 2.5 microg/l was associated with a nadir/mean GH during OGTT/GHDC < 2.5 microg/l in 98.6% and 88.2% of cases, respectively. Elevated IGF-I was seen in 32.4% and 46.4% of patients with GH nadir values during OGTT < 1 and < 2.5 microg/l, respectively, and in 21.2% and 45.9% of GHDC with mean GH < 1 and < 2.5 microg/l, respectively. Radiotherapy increased the discordance in GH and IGF-I as markers of disease activity at GH < 2.5 microg/l (elevated IGF-I-values when OGTT nadir GH < 2.5 microg/l: radiotherapy 55.5%vs. no radiotherapy 36.9%, P = 0.002). CONCLUSIONS: There is a close relationship between a basal fasting GH < 2.5 microg/l and nadir/mean GH < 2.5 microg/l during OGTT/GHDC. There is a large discordance between disease activity when assessed by GH and IGF-I which is further increased by radiotherapy. These observations illustrate the challenge of defining appropriate biochemical end-points to achieve control of disease and normalization of mortality in acromegaly.

Do polycystic ovaries on ultrasound scan indicate decreased insulin sensitivity in sisters of women with polycystic ovary syndrome?

Polycystic ovary syndrome (PCOS) is associated with hyperandrogenemia, insulin resistance, altered lipid profile, and therefore with subsequently increased risk for metabolic complications such as type 2 diabetes and cardiovascular diseases. It has been reported that sisters of probands with PCOS, who themselves had PCOS or hyperandrogenemia with normal menses, were more insulin resistant than unaffected sisters. We have previously reported that 60% of first-degree relatives (premenopausal mothers and sisters) of PCOS probands had polycystic ovaries (PCO) on ultrascan. Sisters with PCO were more likely to have oligomenorrhea and increased androgen levels than sisters without PCO. The aims of this study were to assess insulin sensitivity status [homeostasis model of assessment, quantitative insulin sensitivity check index, glucose to insulin ratio (G/I)] and lipid profiles in probands with PCOS and their sisters with PCO and without PCO on ultrascan. Mixed model hierarchical regression analysis, to accommodate the nonindependent nature of the subjects (family relationships), with the three groups together did not show significant differences in insulin sensitivity, calculated as quantitative insulin sensitivity check index, homeostasis model of assessment, and G/I for PCOS probands, through sisters with PCO on ultrascan, to sisters without PCO on ultrascan. There was a significant association of measures of insulin sensitivity with body mass index. Lipid parameters did not differ between the groups. These data suggest that presence of PCO on ultrasound scan per se does not predispose to reduced insulin sensitivity in sisters of women with PCOS. Because about 20% of premenopausal women in the general population have PCO on ultrascan, and obesity/overweight is becoming more prevalent, it is important that future studies take full account of the contribution made by obesity to risk factors for metabolic/vascular complications.

Recombinant GH replacement in hypopituitary adults improves endothelial cell function and reduces calculated absolute and relative coronary risk.

OBJECTIVE: Adult GH deficiency (GHD) is linked to endothelial dysfunction and vascular disease. We examined the effect of 12 months of GH therapy on endothelial function, C-reactive protein (CRP) and coronary risk. DESIGN: Open-design intervention study. PATIENTS: Fourteen GH-deficient patients (nonsmokers, without diabetes, hypertension or vascular disease) studied before, 6 months and 12 months after GH therapy. MEASUREMENTS: Flow-mediated dilatation (FMD), carotid intima-media thickness (IMT) thrombomodulin (TM), E-selectin, CRP, lipid profile, blood pressure and anthropometric data were recorded. We used the Framingham equation to calculate coronary risk. RESULTS: FMD improved (7.5 +/- 1.62 vs. 11.93 +/- 1.52, P = 0.038). Overall there was no change in IMT, TM, E-selectin or CRP. The correlation between TM and FMD showed a trend for statistical significance (r = -0.54, P = 0.056). Changes in CRP correlated with change in IGF-1 (r = -0.67, P = 0.012); E-selectin correlated with high density lipoprotein (HDL)-cholesterol (r = -0.60, P = 0.028), triglycerides (r = 0.68, P = 0.01) and waist-to-hip ratio (WHR) (r = 0.71, P = 0.006). Systolic (127.36 +/- 4.47 vs. 120.36 +/- 3.50, P = 0.017) and diastolic (84.71 +/- 2.73 vs. 76.93 +/- 2.03, P = 0.005) blood pressure decreased. HDL-cholesterol increased (0.70 +/- 0.05 vs. 0.93 +/- 0.06, P = 0.001). WHR decreased (0.90 +/- 0.02 to 0.88 +/- 0.02, P = 0.043) without changes in weight or body mass index (BMI). Ten-year absolute (P = 0.009) and relative (P = 0.002) cardiac risk decreased. CONCLUSION: Biophysical test of endothelial function (FMD) improved after 12 months of GH therapy but there was no significant change in biochemical endothelial or inflammatory markers. Calculated coronary risk decreased mainly due to reduction in systolic and diastolic blood pressure and increase in HDL-cholesterol.

Pituitary tumour clonality revisited.

Allelotype analysis and X chromosome inactivation analysis in women enables the assessment of tissue clonality, and has demonstrated that the majority of sporadic human pituitary adenomas are monoclonal. This implies that these tumours arise from de novo somatic genetic change(s) in a single pituitary cell. However, clonality within any given tumour may be multiple or single, multiple tumours arising on the background of hyperplasia may be of identical or different clonality, multiple 'sporadic' tumours in a tissue may be of differing clonal origin, and finally morphology cannot predict genetic makeup. These general principles may also apply to the pituitary so it is simplistic to assume that monoclonality is inevitable and that pituitary tumours cannot be multiclonal in origin. Indeed, these observations would be entirely compatible with the initiating stimulus resulting in hyperplasia of specific cell subtypes in the pituitary giving rise to a number of different clones each with variable potential to develop into a discrete tumour depending on their rate of cell division/rate of apotosis. Stimuli might include pituitary-specific oncogenes, intrapituitary growth factors, or extrapituitary trophic factors (e.g. hypothalamic releasing hormones).

Prediction of recurrence of nonfunctioning pituitary tumours by loss of heterozygosity analysis.

BACKGROUND: Postsurgical regrowth or recurrence of nonfunctioning pituitary adenomas (NFAs) is not uncommon and often requires further surgery or radiotherapy (DXT). Routine postoperative DXT increases the incidence of hypopituitarism, which is associated with increased morbidity and mortality. Identification of genetic abnormalities in the tumour tissue, which can predict recurrence, may allow targeting DXT to the most appropriate patients. DESIGN AND METHODS: We have performed loss of heterozygosity (LOH) analysis on 96 NFAs of which 43 (45%) were recurrent and 53 (55%) were nonrecurrent tumours. Analysis of all tumours was performed on the surgical specimen obtained at the time of first surgery. All tumours underwent allelotyping across nine highly informative microsatellite markers selected on the basis of high LOH frequency in an earlier study involving genome-wide allelotyping. LOH frequency across all microsatellite markers as well as across individual markers was compared between the two cohorts of tumours. RESULTS: LOH frequency in tumours that subsequently recurred was significantly higher across all microsatellite markers as compared to tumours that did not recur (P < 0.05). Allelic loss across one or more microsatellite marker was significantly higher in recurrent tumours (30/43) as compared to their nonrecurrent counterparts (17/53) (P < 0.01). On Poisson regression analysis, the higher LOH frequency in recurrent tumours was independent of the invasiveness of tumours determined radiologically. In addition, LOH at the microsatellite markers D1S215 and D1S459 was significantly higher in tumours that recurred as compared to tumours that did not (32%vs. 3% and 27%vs. 2%, respectively; P < 0.01 for both). No significant difference in LOH frequency between the two tumour groups was evident at the other markers. No association could be demonstrated between the frequency and pattern of LOH and the time to manifest recurrence. CONCLUSIONS: We have shown that it may be possible to predict recurrence of NFAs by LOH analysis of the initial tumour specimen at predefined microsatellite markers, especially on chromosome 1q. This merits further prospective study.

Growth hormone and pituitary radiotherapy, but not serum insulin-like growth factor-I concentrations, predict excess mortality in patients with acromegaly.

Increased mortality in patients with acromegaly has been confirmed in a number of retrospective studies, but causative factors and relationship to serum IGF-I remain uncertain. The West Midlands Pituitary database contains details of 419 patients (241 female) with acromegaly. Serum IGF-I data from the Regional Endocrine Laboratory were available for 360 patients (86%). At diagnosis, mean age was 47 yr (range, 12-84) and mean duration of follow-up was 13 yr (0.5-48). Sixty-one percent were treated by surgery and 39% by nonsurgical means. Radiotherapy was used alone or as adjuvant therapy in 50%. All patients were registered with the Office of National Statistics to obtain information on deaths. At the date of analysis (31 December 2001), 95 of the 419 patients had died (43 males), giving a standardized mortality ratio of 1.26 [confidence interval (CI), 1.03-1.54; P = 0.046]. After controlling for age and sex, data indicated that mortality was increased in subjects with posttreatment GH levels more than 2 micro g/liter, compared with those with levels less than 2 micro g/liter [ratio of mortality rates (RR), 1.55 (range, 0.97-2.50); P = 0.068]. By contrast, a much smaller increase was observed for subjects with elevated posttreatment IGF-I levels compared with those with normal levels [RR, 1.20 (range, 0.71-2.03); P = 0.50]. Treatment with radiotherapy was associated with increased mortality [RR, 1.67 (range, 1.09-2.56); P = 0.018], with cerebrovascular disease the predominant cause of death [standardized mortality ratio, 4.42 (range, 2.71-7.22); P = 0.005]. These results confirm the increased mortality in acromegaly and suggest that reduction of GH levels to less than 2 micro g/liter is beneficial in terms of improving long-term outcome. The sole use of IGF-I as a marker for effective treatment of acromegaly is not justified by this data. This study also highlights the potential deleterious effect of radiotherapy.

Influence of the hormonal changes during the normal menstrual cycle in healthy young women on soluble adhesion molecules, plasma homocysteine, free radical markers and lipoprotein fractions.

AIM: Female sex hormones are known to exert a protective role on the vascular endothelial function, but the exact mechanisms of such protection is not known. We aimed to study the possible regulatory role of the female sex hormones changes during the normal menstrual cycle on soluble adhesion molecules E-selectin and ICAM-1, plasma homocyteine, free radical markers and lipoproteins in healthy young women. EXPERIMENTAL DESIGN: a cross sectional study of healthy female volunteers studied during a single normal menstrual cycle at 3 specific time points. SETTING: North Staffordshire Hospitals NHS Trust. SUBJECTS: 20 healthy young menstruating women, aged (mean +/- SEM) 34 +/- 1 years, with normal menstruation, defined as a menstrual cycle of 21-35 days were studied at 3 time points of the same menstrual cycle. First in the early follicular phase (M-phase), at mid-follicular phase (F-phase), and during the luteal phase (L-phase). INTERVENTION: none. MEASUREMENT: serum levels of soluble E-selectin, ICAM-1, plasma homocysteine, vitamin E and malondialdehyde (MDA), as well as lipoprotein fractions were measured at each time points. RESULTS: The mean percentage change for E-selectin between the M-phase and L-phase, F-phase and L-phase were 6% and 4%, respectively, p<0.005, p<0.066. Levels of ICAM-1, vitamin E and malondialdehyde did not vary through the cycle. Homocysteine was not different between M-phase and F-phase (10.39 +/- 0.68 micromol/l vs 10.33 +/- 0.65), nor between M-phase and L-phase (10.39+/-0.68 vs 9.77 +/- 0.75 micromol/l). Although the mean percentage decrease in homocysteine between F- and L-phases was significant (5.36 +/- 0.53%, p=0.029), the absolute decrease in concentrations was not (p=0.07). There were no cyclical changes in total, LDL, HDL cholesterol, triglycerides, apo A-I, apo B or Lp(a). Using a linear regression model, after correction for age, smoking, body mass index (BMI) and waist/hip ratio (WHR), oestrogen levels were the only predictor of E-selectin during the L-phase p<0.005. There were no significant correlations between oestrogen with lipids, apolipoproteins or homocysteine. There was an interesting significant univariate correlation between homocysteine with low-density-lipoprotein (LDL) cholesterol and apo B throughout all phases of the cycle, which persisted after correction for the effects of age, BMI, WHR and smoking history. Multiple regression analysis with all these factors showed homocysteine to be a significant predictor of apo B concentration during M (p=0.030) and L-phases (p=0.023) of the cycle and of LDL cholesterol in the M-phase (p=0.033). CONCLUSION: Female sex hormones may have small, though significant modulating role on E-selectin and homocysteine metabolism in healthy premenopausal women. Furthermore, the correlation between homocysteine, LDL and apo B levels suggests that induction of cholesterol synthesis by homocysteine, shown previously in vitro, may be of relevance in vivo.

Epigenetic change in pituitary tumorigenesis.

Throughout the genome CpG dinucleotides are found at one-fifth of their expected frequency and their rarity is further marked by the fact that 70% are methylated. In contrast, CpG islands (CGI), found associated with the promoters of many genes, have maintained their expected frequency of this dinucleotide, and remain unmethylated. Inappropriate methylation of CGIs is associated with histone deacetylation and gene silencing, while methylation of CpGs outside of CGIs is associated with significantly higher mutation rates. Methylation of CGIs is a frequent event in numerous tumour types including those that arise within the pituitary gland. Several studies now show highly frequent methylation of the p16 gene that is significantly associated with loss of cognate protein and that appears to be an early change in pituitary tumorigenesis. Collectively, studies show that somatotrophinomas are an infrequent target for p16 CGI methylation. However, in this pituitary tumour subtype, loss of pRb is associated with either CGI methylation or micro-deletion within the protein-pocket binding domain. As in other tumour types loss of p16 or RB1 appear to be mutually exclusive events in non-functional adenomas and somatotrophinomas respectively. Investigation of the Death Associated Protein Kinase gene shows that loss of its protein (DAPK), a pro-apoptotic molecule, in pituitary tumours is also associated with either methylation or deletion within its associated CGI. In the case of DAPK, however, these changes segregate with invasive pituitary tumours irrespective of tumour subtype. Methylation represents a positive signal that can be detected with exquisite sensitivity; in addition, this change targets multiple genes that show tumour type specificity. Taken together, the detection of DNA methylation changes, using either a panel of predefined marker-islands, or CGI arrays, provides the opportunity to generate "methylation profiles". This new knowledge will increase our understanding of tumour biology and could ultimately aid medical management in these different tumour types, including those of pituitary origin.

Genome-wide amplification and allelotyping of sporadic pituitary adenomas identify novel regions of genetic loss.

Through the use of a candidate gene approach, several previous studies have identified loss of heterozygosity (LOH) at putative tumor-suppressor gene (TSG) loci in sporadic pituitary tumors. This study reports a genome-wide allelotyping by use of 122 microsatellite markers in a large cohort of tumors, consisting of somatotrophinomas and non-functioning adenomas. Samples were first subject to prior whole genome amplification by primer extension pre-amplification (PEP) to circumvent limitations imposed by insufficient DNA for whole-genome analysis with this number of microsatellite markers. The overall mean frequency of loss in invasive tumors was significantly higher than that in their non-invasive counterparts (7 vs. 3% somatotrophinomas; 6 vs. 3% non-functioning adenomas, respectively). Analysis of the mean frequency of LOH, across all markers to individual chromosomal arms, identified 13 chromosomal arms in somatotrophinomas and 10 in non-functioning tumors, with LOH greater than the 99% upper confidence interval calculated for the rate of overall random allelic loss. In the majority of cases, these losses were more frequent in invasive tumors than in their non-invasive counterparts, suggesting these to be markers of tumor progression. Other regions showed similar frequencies of LOH in both invasive and non-invasive tumors, implying these to be early changes in pituitary tumorigenesis. This genome-wide study also revealed chromosomal regions where losses were frequently associated with an individual marker, for example, chromosome arm 1q (LOH > 30%). In some cases, these losses were subtype-specific and were found at a higher frequency in invasive tumors than in their non-invasive counterparts. Identification of these regions of loss provides the first preliminary evidence for the location of novel putative TSGs involved in pituitary tumorigenesis that are, in some cases, subtype-specific. This investigation provides an unbiased estimate of global aberrations in sporadic pituitary tumors as assessed by LOH analysis. The identification of multiple "hotspots" throughout the genome may be a reflection of an unstable chromatin structure that is susceptible to a deletion or epigenetic-mediated gene-silencing events.

Pituitary tumours are multiclonal from the outset: evidence from a case with dural metastases.

In 1992 a 54-year-old man underwent transsphenoidal adenomectomy to remove a clinically nonfunctioning pituitary adenoma during which there was a transient cerebrospinal fluid (CSF) leak. He received radiotherapy to a small residual remnant. Follow-up magnetic resonance imaging (MRI) scan in 1997 showed an increase in the tumour in the pituitary stalk region and an additional intradural lesion at C1 level. In the absence of neurological symptoms and signs, an observational policy was followed. By 1999 the cervical dural lesion had enlarged and laminectomy was performed, during which three intradural lesions were removed. Histology and immunohistochemistry of the metastases were identical to those of the initial pituitary adenoma. Follow-up MRI scan showed extension of the pituitary remnant above the chiasma, requiring transfrontal surgery. Operation was complicated by secondary brain haemorrhage from which the patient died. Autopsy revealed a small amount of residual tumour at the top of the stalk and several small intradural tumour nodules at the level of the foramen magnum. Genetic analysis of the initial pituitary tumour identified significant allelic losses in keeping with its invasive nature, while that of the metastases indicated a separate clone as shown by retention of alleles lost in the primary tumour. The regrown pituitary tumour also appeared to be of a different clone to the initial tumour and the same as two of three of the first metastases (C1 level). The foramen magnum metastasis showed the same loss of heterozygosity (LOH) pattern as one of the original C1 metastases and the pituitary tumour tissue obtained at autopsy. We speculate that at the initial pituitary surgery, cells seeded into the CSF and grew in the dura. These cells were from a different clone, implying that the original pituitary tumour contained at least two clones, possibly three, providing evidence for the contemporaneous oligoclonality of the original pituitary tumour.

Preferential loss of Death Associated Protein kinase expression in invasive pituitary tumours is associated with either CpG island methylation or homozygous deletion.

Death Associated Protein kinase (DAP kinase) a novel calmodulin-dependent serine/threonine kinase was first identified as a positive mediator of programmed cell death. Loss of DAP kinase expression was first demonstrated in highly metastatic cells, whilst re-expression of the protein resulted in delayed local tumour growth and a decreased incidence of metastasis. Although loss of DAP kinase expression has been reported in several cell lines derived from human malignancies the mechanisms responsible have not been defined. In this study we have examined 32 sporadic pituitary tumours for expression of the DAP kinase protein and transcript. In addition, we examined the methylation and deletion status of the DAP kinase CpG island as possible mechanisms for the inactivation of the DAP kinase gene. Eleven of 32 (34%) tumours had undetectable DAP kinase expression, by Western blot and/or RT-PCR analysis. Loss of DAP kinase expression was significantly (P=0.004) associated with invasive tumours (10 of 17; 59%) compared to their non-invasive (1 of 15; 7%) counterparts. Of 11 tumours that failed to express DAP kinase, five (45%) showed de novo methylation of the CpG island contained within the promoter region, while four (36%) had evidence of homozygous deletion of this region. Statistical analysis showed that loss of DAP kinase expression was significantly (P=<0.001) associated with methylation or deletion of the DAP kinase CpG island. With two exceptions, none of the remaining tumours or five histologically normal post-mortem pituitaries examined had evidence of methylation or deletion within this region. To our knowledge this is the first report that describes two mutually exclusive mechanisms associated with loss of DAP kinase gene expression. In addition, we also show that loss of the DAP kinase protein and associated genetic aberrations preferentially segregates with tumours that show an invasive phenotype.